Upload
doantuyen
View
217
Download
0
Embed Size (px)
Citation preview
A CASE-CONTROL STUDY OF AMBIENT HEAT
EXPOSURE AND UROLITHIASIS AMONG OUTDOOR
WORKERS IN A SHIPBUILDING COMPANY,
GUANGZHOU, CHINA
Haiming Luo
Bachelor of Medicine, Master of Science in Public Health
Submitted in fulfilment of the requirements for the degree of
Master of Applied Science (Research)
School of Public Health and Social Work
Queensland University of Technology
July 2012
I
KEY WORDS
Ambient heat exposure, Case-control study, Conditional logistic regression,
Guangzhou, Outdoor worker, Urolithiasis
II
III
ABSTRACT
Higher ambient temperatures will increase heat stress on workers, leading to impacts
upon their individual health and productivity. In particular, research has indicated that
higher ambient temperatures can increase the prevalence of urolithiasis. This thesis
examines the relationship between ambient heat exposure and urolithiasis among
outdoor workers in a shipbuilding company in Guangzhou, China, and makes
recommendations for minimising the possible impacts of high ambient temperatures
on urolithiasis.
A retrospective 1:4 matched case-control study was performed to investigate the
association between ambient heat exposure and urolithiasis. Ambient heat exposure
was characterised by total exposure time, type of work, department and length of
service. The data were obtained from the affiliated hospital of the shipbuilding
company under study for the period 2003 to 2010. A conditional logistic regression
model was used to estimate the association between heat exposure and urolithiasis.
This study found that the odds ratio (OR) of urolithiasis for total exposure time was
1.5 (95% confidence interval (CI): 1.2–1.8). Eight types of work in the shipbuilding
company were investigated, including welder, assembler, production security and
quality inspector, planing machine operator, spray painter, gas-cutting worker and
indoor employee. Five out of eight types of work had significantly higher risks for
IV
urolithiasis, and four of the five mainly consisted of outdoors work with ORs of 4.4
(95% CI: 1.7–11.4) for spray painter, 3.8 (95% CI: 1.9–7.2) for welder, 2.7 (95% CI:
1.4–5.0) for production security and quality inspector, and 2.2 (95% CI: 1.1–4.3) for
assembler, compared to the reference group (indoor employee). Workers with
abnormal blood pressure (hypertension) were more likely to have urolithiasis with an
OR of 1.6 (95% CI: 1.0–2.5) compared to those without hypertension.
This study contributes to the understanding of the association between ambient heat
exposure and urolithiasis among outdoor workers in China. In the context of global
climate change, this is particularly important because rising temperatures are expected
to increase the prevalence of urolithiasis among outdoor workers, putting greater
pressure on productivity, occupational health management and health care systems.
The results of this study have clear implications for public health policy and planning,
as they indicate that more attention is required to protect outdoor workers from heat-
related urolithiasis.
V
TABLE OF CONTENTS
KEY WORDS ................................................................................................................ I
ABSTRACT ................................................................................................................. III
LIST OF TABLES ....................................................................................................... IX
LIST OF FIGURES ...................................................................................................... X
LIST OF ABBREVIATIONS ...................................................................................... XI
STATEMENT OF AUTHORSHIP ........................................................................... XII
ACKNOWLEDGEMENTS ...................................................................................... XIII
CHAPTER 1: INTRODUCTION .................................................................................. 1
1.1 Overview .............................................................................................................. 1
1.2 Knowledge gaps ................................................................................................... 6
1.3 Hypotheses ........................................................................................................... 6
CHAPTER 2: LITERATURE REVIEW ....................................................................... 9
2.1 Impacts of climate change .................................................................................... 9
2.2 Climate change and occupational health ............................................................ 10
2.3 Workplace heat exposure and occupational health ............................................ 11
2.4 Heat exposure and urolithiasis ........................................................................... 12
2.5 Study designs assessing the effects of occupational heat exposure on urolithiasis
.................................................................................................................................. 13
2.6 Risk factors for heat-induced urolithiasis........................................................... 20
VI
2.7 Knowledge gaps ................................................................................................. 23
CHAPTER 3: STUDY DESIGN AND METHODS ................................................... 33
3.1 Aims and objectives ........................................................................................... 33
3.1.1 Aims............................................................................................................. 33
3.1.2 Specific objectives ....................................................................................... 33
3.2 Research design and data collection................................................................... 33
3.2.1 Research design ........................................................................................... 33
3.2.2 Study setting ................................................................................................ 34
3.2.3 Data collection ............................................................................................. 37
3.2.4 Criteria for selection .................................................................................... 38
3.3 Data analysis ...................................................................................................... 43
3.3.1 Descriptive statistics .................................................................................... 43
3.3.2 Univariate analysis ...................................................................................... 44
3.3.3 Selecting variables for modelling process ................................................... 44
3.3.4 Bivariate and multivariable modelling ........................................................ 45
CHAPTER 4: RESULTS ............................................................................................. 53
4.1 Descriptive statistics and univariate analyses .................................................... 53
4.2 Overlapping effect .............................................................................................. 58
4.3 Association between risk factors and urolithiasis .............................................. 59
VII
4.4 Sensitivity test of the multivariable model ......................................................... 61
CHAPTER 5: DISCUSSION AND CONCLUSION .................................................. 65
5.1 Key findings in this study .................................................................................. 65
5.2 Public health implications .................................................................................. 65
5.3 Recommendations .............................................................................................. 67
5.4 Alternative explanations ..................................................................................... 71
5.4.1 Chance ......................................................................................................... 71
5.4.2 Bias .............................................................................................................. 72
5.5 Comparison with other studies ........................................................................... 73
5.5.1 Previous studies on ambient heat and urolithiasis ....................................... 73
5.5.2 General comparisons of study designs and results ...................................... 75
5.5.3 Association between total exposure time and urolithiasis ........................... 77
5.5.4 Association between the type of work and urolithiasis ............................... 78
5.5.5 Association between hypertension and urolithiasis ..................................... 85
5.6 Strengths and limitations .................................................................................... 86
5.7 Directions for future research ............................................................................. 88
5.8 Conclusions ........................................................................................................ 90
REFERENCES ............................................................................................................ 91
APPENDIX 1: Tables of Results for Statistical Analyses ......................................... 107
APPENDIX 2: Ethics Application Exempt ............................................................... 121
VIII
APPENDIX 3: GIHI Approval Letter........................................................................ 123
APPENDIX 4: Information Collection Form to OHSOs ........................................... 125
IX
LIST OF TABLES
Table 1. Characteristics of previous studies assessing the effects of heat exposure on
urolithiasis .................................................................................................................... 26
Table 2. Metabolic evaluation for explaining heat-induced urolithiasis...................... 30
Table 3. Descriptive statistics and univariate analysis of variables ............................. 57
Table 4. Overlapping effect between type of work and department ............................ 58
Table 5. Odds ratios of risk factors of urolithiasis in the conditional logistic regression
model............................................................................................................................ 60
Table 6. Odds ratios of risk factors for urolithiasis in the conditional logistic
regression model for a sensitivity test .......................................................................... 62
Table 7. Threshold WBGT levels for different workloads among men normally
clothed, acclimatized, physically fit and in good health .............................................. 80
Table 8. Modification of threshold WBGT level by different conditions ................... 80
Table 9. Threshold WBGT limits ................................................................................ 82
Table 10. Non-climate factors impacting on heat stress of different types of work and
relevant odds ratios of urolithiasis ............................................................................... 84
Table 11. Summary of strengths and limitations of this study .................................... 88
X
LIST OF FIGURES
Figure 1. Atmosphere-Ocean General Circulation Model projections of surface
warming ....................................................................................................................... 2
Figure 2. Scatter plot of prevalence rates for kidney stones in the United States by
state ................................................................................................................................ 3
Figure 3. Predicted growth in high-risk area (stone belt; risk ratio ≥ 1.2) vs. time 2000
(yellow), 2050 (orange), and 2095 (red); linear model. At 2000 41% of the population
within a high-risk zone, 56% at 2050, and 95% at 2095, based on year 2000
population distribution ................................................................................................... 4
Figure 4. The location of Guangzhou City (the area of yellow colour) ....................... 35
Figure 5. Mean monthly temperature in Guangzhou City ........................................... 36
Figure 6. Bar graph of percentages by three levels of total exposure time .................. 54
Figure 7. Bar graph of percentages by three levels of total exposure time among eight
types of work................................................................................................................ 55
Figure 8. Odds ratios and 95% confidence intervals for seven different types of work
...................................................................................................................................... 61
XI
LIST OF ABBREVIATIONS
BP: Blood Pressure
CI: Confidence Interval
DALYs: Disability Adjusted Life Years
ECG: Electrocardiogram
GEE: Generalized Estimating Equations
GHG: Greenhouse Gas
25-HCC: 25-Hydroxycholecalciferol
iPTH: intact Parathyroid Hormone
IPCC: Intergovernmental Panel on Climate Change
ISO: International Organization for Standardization
OHSO: Occupational Health and Safety Officer
OR: Odds Ratio
PPE: Personal Protective Equipment
UV: Ultraviolet
WBGT: Wet Bulb Globe Temperature
WHO: World Health Organisation
25(OH)D: 25-hydroxyvitamin D
XII
STATEMENT OF AUTHORSHIP
The work contained in this thesis has not been previously submitted to meet
requirements for an award at this or any other higher education institution. To the best
of my knowledge and belief, the thesis contains no material previously published or
written by another person except where due reference is made.
Signature:
Date: 30th
July 2012
XIII
ACKNOWLEDGEMENTS
I would like to thank my three supervisors Prof. Shilu Tong (Principal Supervisor), Dr.
Cameron Hurst (Associate Supervisor) and Dr. Lyle Turner (Associate Supervisor).
Prof. Tong, his enthusiasm, insight and dedication to the advancement in
environmental and public health research are inspirational. His academic, empirical
and prompt advice and comments on my study were highly valuable. I really
appreciate his guidance and support for my study design, data collection, analyses and
thesis writing. Dr. Hurst, as an experienced biostatistician, had given me valuable
statistical advice on data analysis and writing of my thesis. I feel grateful for his
support on statistical data analysis and professional encouragement to help me move
forward. Dr. Turner had shared with me his research experience, his patience and
willingness to help during my study. I thank him for his support and friendship.
I would like to acknowledge the School of Public Health and Social Work, Institute of
Health and Biomedical Innovation, QUT, for funding this study and for all the support
in my research.
I would like to acknowledge Guangdong Institute of Health Inspection for providing
the health surveillance data in this study.
XIV
I would also like to thank my family and friends for their encouragement and
emotional support, and particularly my husband, for encouraging me to pursue higher
degrees in research. I greatly appreciate all their unconditional love and support
throughout my study.
1
CHAPTER 1: INTRODUCTION
1.1 Overview
Prolonged exposure to a hot work environment can bring about heat-induced
disorders, such as dehydration, heat rash, fatigue, heat cramps, syncope, and
heatstroke [1]. Symptomatic exhaustion and clinical diseases, particularly urolithiasis,
can also occur as a result of excessive heat exposure [2, 3]. Some factors may increase
or decrease individual susceptibility (e.g. age, obesity and pre-existing medical
conditions), and socio-economic status may also play a role in how heat exposure will
impact on workers‟ health [4].
There is clear evidence that at current levels of economic growth, global emissions of
greenhouse gas (GHG) will continue to grow over the coming few decades [5]. In the
most recent Intergovernmental Panel on Climate Change (IPCC) Assessment Report
in 2007, it was projected that global average temperatures will increase by 1.1–6.4 °C
by 2100 (with the most likely range being 2 to 4 °C) (See Figure 1) [6]. This would
result in the warmest period on the Earth for at least the last 100,000 years [7].
Climate change is likely to impact on occupational health and safety across all sectors
of industry [4]. For example, outdoor workers will face hotter conditions and
increased exposure to higher ambient temperatures.
2
Figure 1. Atmosphere-Ocean General Circulation Model projections of surface
warming (Source: IPCC, 2007 [6])
Urolithiasis, or kidney stone disease, is a common disease across the world [8]. The
average global prevalence of urolithiasis was 3.3% in the 1980s and 5.6% in the
1990s [9, 10]. The highest prevalence rates were for uranium workers in eastern
Tennessee (18.5%) and adults in northeast Thailand (16.9%) [11, 12]. Although the
underlying cause for urolithiasis is still not clear, a great number of studies have
suggested ambient temperature as a contributor to kidney stone formation [2]. The
prevalence of urolithiasis has shown noticeable geographic variability. In the US, the
southeast has been reported to have as much as a 50% higher prevalence than the
northwest. Mean annual temperature has been estimated to account for 70% or more
of this variability, with other risk factors such as age, gender, race, diet, family history,
social class and sunlight index found to potentially account for the remainder [13].
Results by Fakheri and Goldfarb [2] demonstrated that mean annual temperature was
3
positively correlated with the prevalence of urolithiasis and men were more
vulnerable than women to heat-induced urolithiasis (See Figure 2).
Figure 2. Scatter plot of prevalence rates for kidney stones in the United States by
state (Source: Fakheri and Goldfarb, 2011 [2])
In recent years, urolithiasis has attracted much attention because of climate change [2,
13]. It has been projected that the nationwide prevalence of urolithiasis attributable to
climate change in the United States will probably increase from 5.2% in 1988–1994 to
10.4% (linear model) by 2050 [13] (See Figure 3).
4
Figure 3. Predicted growth in high-risk area (stone belt; risk ratio ≥ 1.2) vs. time 2000
(yellow), 2050 (orange), and 2095 (red); linear model. At 2000 41% of the population
within a high-risk zone, 56% at 2050, and 95% at 2095, based on year 2000
population distribution. (Source: Brikowski et al., 2008 [13])
Heat is an important occupational health issue because its effects are likely to impair
workers‟ health and consequently their productivity [14]. Outdoor workers in tropical
and subtropical areas, particularly in low and middle-income countries, are at highest
risk of outdoor heat exposure, because outdoor workers in these countries engage in
heavy physical activities in hot environment [15]. Occupational heat exposure as a
possible contributor to urolithiasis formation was firstly reported in 1945 by Pierce et
al. [16]. The researchers found that the incidence of urolithiasis among American
troops in the desert area was two times higher than in the mountainous area. Since
then, a number of epidemiological studies have been conducted to explore the
association between occupational heat exposure and urolithiasis among engineers [17],
lifeguards [18], marathon runners [19], machinists [20] and steel workers [21]. Most
of these studies were carried out in developed countries.
5
An increase in workplace heat exposure due to climate change is likely to create a
range of occupational health problem including urolithiasis. However, the potential
impacts of climate change on occupational safety and workers‟ health remain largely
unknown currently, with most previous research focusing on the health impacts of the
general population rather than on the working population [4, 22]. This is in part due to
the lack of useful data relevant to this relatively new area of occupational health
research [4].
China is a rapidly developing country that has experienced a general trend of
temperature increases since the late 1980s, in line with the global pattern [23]. China
has lots of labour-intensive industries, e.g. shipbuilding, construction, stevedoring
services and mining, and it has a large proportion of the world‟s manufacturing
workers. Many of these workers are exposed to ambient heat in poor working
conditions, with heavy workloads and low incomes. They are generally from poor
families with little education, and receive no or very little training about occupational
health. They have extremely low awareness and understanding of how to prevent
urolithiasis at the workplace. They usually have no regular urolithiasis screening
programs, and no adequate water breaks and free drinking water supplies.
6
1.2 Knowledge gaps
Up to now, most previous studies have examined the association between
occupational ambient heat exposure and urolithiasis in developed countries [16-20,
24-27]. Very few studies have explored the impacts of ambient heat on urolithiasis
among working populations in developing countries, such as China. Additionally,
most of these studies were descriptive or cross-sectional in design. Therefore,
previous studies have been unable to make a causal inference because both health and
ambient temperature data were collected at the same time. Thus, analytical designs
such as a case-control or cohort design should be employed if feasible. This study
aimed to determine the association between ambient heat exposure and urolithiasis
among workers in a large shipbuilding company in Guangzhou, Guangdong Province,
China, using a case-control study design.
1.3 Hypotheses
Hypotheses for this study are:
1. H1: There is an association between ambient heat exposure and urolithiasis
among outdoor workers in China.
H0: There is no association between ambient heat exposure and urolithiasis
among outdoor workers in China.
7
2. H1: Workers with longer exposure time, certain types of work or relevant
cardiovascular diseases are more likely to have urolithiasis.
H0: There is no association of exposure time, certain types of work or relevant
cardiovascular diseases with urolithiasis.
This thesis is composed of five chapters. The first chapter is an introduction; the
second is a literature review of occupational heat exposure and urolithiasis; the third
describes the design and methods used in this research; the fourth presents the results;
and, the fifth contains discussions, recommendations, and conclusions.
8
9
CHAPTER 2: LITERATURE REVIEW
2.1 Impacts of climate change
There is a global scientific consensus that the Earth is warming, due to GHG
emissions caused by human activity [6]. The IPCC has concluded that “warming of
the climate system is unequivocal, as is now evident from observations of increases in
global average air and ocean temperatures, widespread melting of snow and ice, and
rising global average sea level” [6]. In the last century, the Earth has warmed by
approximately 0.75 °C and global sea levels have risen by over 4 cm. Eleven of the
twelve years from 1995–2006 were ranked among the twelve warmest years on
records globally since 1850 [6].
Climate change is one of the most critical issues facing the global community [7].
Because the climate is changing rapidly, it has the potential to significantly impact the
global environment, society and economy [28]. The impacts of climate change will be
broad and complex, and may have far reaching consequences on all parts of society
[4]. Climate change has been increasingly recognised as the biggest global health
threat of the 21st century [29]. A great number of studies have indicated that climate
change can affect human health directly and indirectly. Direct health impacts include
heat-induced mortality and morbidity, such as deaths and illnesses due to heat-waves
10
[30-32]. Indirect impacts include climate-mediated changes in the incidence of
infectious diseases, such as geographical and seasonal changes in incidences of
malaria, dengue, tick-borne viral disease and schistosomiasis [33-35], and the
prevalence of chronic diseases, such as urolithiasis, lung disease and heart diseases [2,
3, 22].
2.2 Climate change and occupational health
Climate change is likely to impact on occupational health and safety across all sectors
of industry [4]. For example, outdoor workers will face hotter conditions and
increased exposure to higher ambient temperatures. The predicted increase in the
intensity of ultraviolet (UV) radiation may also present further hazards to outdoor
workers [1].
As strategies to promote effective climate change adaptation are now being
implemented, the occupational health community should play an important role in
both understanding the impacts of climate change and minimising these impacts [4].
How is climate change related to occupational safety and workers‟ health? Will the
impacts of climate change on the working population be different to impacts on the
general population? Many questions remain to be answered. Recently, a framework
has been developed by Schulte and Chun [1] to help identify how climate change can
affect occupational health and safety. The hazards were grouped into seven categories:
11
(1) increased ambient temperature; (2) UV radiation; (3) air pollution (resulting from
increased temperatures, ozone levels and airborne particles); (4) extreme weather; (5)
expanded vector habitat; (6) industrial transitions and emerging technologies; and (7)
changes in the built environment [1]. However, uncertainty remains in attributing
occupation-related health effects to climate change because of the lack of long-term
and high-quality datasets. Additionally, the role of a range of confounders (e.g. socio-
economic factors) remains unclear [30].
2.3 Workplace heat exposure and occupational health
An increase in workplace heat exposure due to climate change is likely to create
occupational health risks and have a significant impact on workers‟ productivity. The
human body maintains a core body temperature of between 36–37.2 °C. A person
carrying out physical activities creates metabolic heat inside the body, which needs to
be transferred to the outside environment in order to avoid an increase in the core
body temperature. The heat transfer between the human body and outside
environment depends on both the ambient air temperature and the type of clothing. If
ambient temperatures exceed 35 °C, the human body can only maintain normal core
body temperature by sweat evaporation [22].
Effective measures (e.g. air conditioning) to reduce workplace heat exposure can be
practical for indoor environments, but such control is much more difficult in outdoor
12
environments. To avoid midday work capacity loss, some people may choose to work
at night time or work during cooler parts of each day. However, night work is not
possible for many workers who must work during daylight [22].
2.4 Heat exposure and urolithiasis
Dehydration induced by sweating is one main mechanism for heat exposure causing
urolithiasis [33]. Continuous sweating leads to body dehydration represented by the
loss of extracellular fluid. This induces an increase in serum osmolality which in turn
causes vasopressin secretion by the posterior pituitary. This leads to reduced urinary
volume and increased urinary concentration, including the concentration of relatively
insoluble salts. When the concentration of these salts exceeds their upper limit of
solubility, the salts precipitate and form solid crystals that develop into stones [36].
If heat exposure and sunlight are experienced together, another mechanism for
causing urolithiasis exists besides dehydration. Increased urinary calcium excretion
attributed to sunlight accelerates the formation of urolithiasis [35]. Sunlight exposure
facilitates the generation of serum 25-hydroxycholecalciferol (25-HCC), which can be
converted to 25-hydroxyvitamin D (25(OH)D) in the kidneys [35]. 25(OH)D is
associated with increasing urinary calcium excretion [37, 38]. Excessive urinary
calcium excretion (hypercalciuria) is an important contributor to urolithiasis in the
pathogenesis process [39].
13
2.5 Study designs assessing the effects of occupational heat exposure on
urolithiasis
To understand previous studies on the effects of occupational heat exposure on
urolithiasis and find knowledge gaps in this area, a literature search was conducted in
April 2011, using the electronic databases PubMed, Scopus, ScienceDirect, and ISI
Web of Science. The search was limited to English language articles published in
peer-reviewed journals from all years up to December 2011. The key words used in
the literature search included: heat, temperature, urolithiasis, kidney stone,
nephrolithiasis, occupation and workers. References and citations of the articles
identified were manually checked to ensure that all relevant published literature was
included.
In total, ten studies [16-21, 24-27] were identified through the literature search. These
studies were conducted in different parts of the world, including the countries of
Brazil [21], Japan [27], Italy [20], Singapore [26], Scotland [25], Israel [18], the
United Kingdom [24], the United States [16, 19], regions of the Persian Gulf [24], the
Mediterranean, the Middle East and the Far East [17]. Nine of these studies were
conducted in developed countries except for one which was carried out in a
developing country Brazil. Their publication dates ranged from 1945 to 2005.
14
A number of studies have assessed the effects of heat exposure on urolithiasis. Most
of these studies have reported a positive association between heat exposure and
urolithiasis [16, 18-21, 24-26]. However, some researchers found no association, such
as one Japanese study performed by Iguchi et al. [27] showing administrative workers
without heat exposure had a higher urolithiasis prevalence than farmers and lumber
workers with heat exposure.
Through the literature review, various study designs have been applied to explore the
relationship between heat exposure and urolithiasis. Two case studies described that
increased heat exposure seems to raise the incidence of urolithiasis, and showed a
potential relationship between heat exposure and urolithiasis [16, 25]. Six cross-
sectional studies explored a possible association between heat exposure and
urolithiasis among outdoor occupations (i.e. lifeguards, marathon runners, quarry
drilling and crusher workers, quarry truck and loader drivers and postal deliverymen),
and indoor occupations (i.e. engineers, steel plant workers and glass plant workers)
[18-21, 26]. One cohort study analysed the association between sunlight exposure and
hypercalciuria and found that increased sunlight exposure increased the excretion of
urine calcium. This was relevant to the association between heat exposure and
urolithiasis [24]. Another cohort study focused on studying the pattern of incidence of
urolithiasis among engineers exposed to heat in the engine room, and found that the
incidence of urolithiasis decreased after the temperature decreased in the engine room
[17].
15
Two case studies were from the United States [16] and the United Kingdom [25]. The
first study published in 1945 was performed by Pierce et al. [16], which evaluated the
detailed information in medical records of 61 newly proven urolithiasis cases among
American troops in an overseas desert area. They found that the number of urolithiasis
cases increased during the hot months and decreased during the cool months. They
stated that the incidence of urolithiasis among soldiers in the desert area was twice as
high as that for those in the mountainous area, however the exact incidence was not
demonstrated due to military security [16]. Ferrie et al. [25] performed another case
study on examining the profiles of 47 urolithiasis patients from a clinic in the
Glasgow area, Scotland. They found that 27 out of 47 (57.5%) patients in total had an
occupational history of hot-metal exposure [25]. These two studies described the
possible association between heat exposure and urolithiasis.
Six cross-sectional studies were based separately in Israel [18], the United States [19],
Japan [27], Singapore [26], Italy [20], and Brazil [21], with the first four focusing on
outdoor workers and the latter two on indoor workers.
Better et al. [18] examined the prevalence of urolithiasis between 45 lifeguards and 50
people from the general population with similar age, sex and season in Israel. They
found that lifeguards had a ten-fold greater risk of urolithiasis than the general
population. Milvy et al. [19] compared the prevalence of urolithiasis between 1977
16
New York marathon runners and the matched general population (from census data)
in the United States. They reported that 19 of the 1,832 marathon runners had
urolithiasis, and the prevalence of urolithiasis among marathon runners was 4.5 times
higher than that of an age-matched group from census data. Pin et al. [26] carried out
a cross-sectional study in Singapore comparing the prevalence of urolithiasis between
several outdoor and indoor occupations involving different levels of physical activity.
They found that outdoor workers had five times higher prevalence of urolithiasis than
indoor workers. However, there was no statistical difference in the prevalence of
urolithiasis between physically active and inactive workers. Iguchi et al. [27] carried
out a survey among 1,972 randomly selected inhabitants of Kaizuka, Japan aged 20 to
59. They found the prevalence of urolithiasis in administrative workers was 19.6%,
which was significantly higher than any other occupations including outdoor workers
(e.g. farmers and lumber workers). Their result was inconsistent with most other
relevant studies [16-20, 24-26]. However, farmers and lumber workers consisted of a
very low fraction (1.3%) of the total respondents, which was referred to as a small
sample size of 25 men and women in total. This was a potential reason for explaining
why the study did not found any cases within the farmers and lumber workers group.
These four studies [18, 19, 26, 27] assessed the association between ambient heat
exposure and urolithiasis. However, the first two did not independently examine the
impact of ambient heat but interact with other risk factors. For example, lifeguards
were exposed to strong sunlight and ambient heat together, and marathon runners
17
undertook intensive physical activities when exposed to ambient heat [18, 19].
Meanwhile, the latter two studies may have suffered from biases in the homogeneity
of their subjects. For instance, Pin et al. [26] conducted their study by combining
different kinds of workers together including quarry drilling and crusher workers,
quarry truck and loader drivers and postal deliverymen as outdoor workers. These
workers were from different kinds of industries where occupational heat and sunlight
exposure differed markedly, and they also had different diet and health status. Similar
biases may exist between farmers and lumber workers in the study performed by
Iguchi et al. [27].
Borghi et al. [20] and Atan et al. [21] carried out cross-sectional studies on heat-
exposed and non-heat-exposed workers in a steel plant and a glass plant, respectively.
Both studies controlled many risk factors, including age, race, family history, diet,
and relevant diseases between the heat-exposed group and the non-heat-exposed
group. Statistical analyses showed significant differences between the two groups [20,
21, 40]. In Borghi et al.‟s study [20], male glass plant workers chronically exposed to
heat stress had more than three times higher risk of urolithiasis than those working at
room temperature. In Atan et al.‟s study [21], heat-exposed male workers were found
to have a ten-fold greater risk of urolithiasis than non-heat-exposed workers [21].
These two were the most recent studies on indoor workplace heat exposure and
urolithiasis. Both studies [20, 21] examined the changes in urinary biochemistry
which was mainly attributed to low urinary volume and led to calcium salts
18
supersaturation. They demonstrated the association between indoor heat exposure and
urolithiasis.
Parry et al. [24] conducted a cohort study on two groups of soldiers transported from
the United Kingdom to the Persian Gulf during both the cold and hot seasons.
However, this study only showed the possible association between ambient heat
exposure and hypercalciuria. They compared the difference of urinary calcium levels
among the same group of soldiers before and after being transported for ten days.
They also compared urinary calcium levels between the two groups at the time when
the summer-transported group arrived in Persian Gulf for 10 days and the winter-
transported group had already stationed in Persian Gulf for 8 months. A significant
increase of urinary calcium excretion was only observed in the summer-transported
group ten days after arrival. Parry et al. [24] argued that the difference between the
winter transported group and the summer transported group was most likely due to
increased exposure to sunlight rather than higher temperatures and decreased
humidity, because other researchers observed that ultraviolet light exposure increased
the excretion of calcium, and the summer transported group received a much longer
period of extra sunlight (5.22 hours per day) than the winter transported group (1.73
hours per day) with rises of 19.5 °C and 12 °C, respectively. Although it was reported
that 2 of 91 soldiers followed up for three years had urolithiasis, the incidence of
urolithiasis among soldiers in Service during the same period in the United Kingdom
or that among the general population was not provided for comparison. Thus, this
19
cohort study mainly indicates the association between sunlight and hypercalciuria, but
also suggests the possible association between both sunlight and ambient heat
exposure and urolithiasis.
Another cohort study was carried out by Blacklock et al. [17] to analyse the pattern of
incidence of urolithiasis in the British Navy over seven years (1958–1964). They
found that the incidence of urolithiasis was higher among engineers who were often
exposed to heat in the engine room. However, their incidence of urolithiasis
significantly declined year by year after air-conditioners were introduced. Thus, it was
considered that the decline in urolithiasis was associated with decreased heat exposure.
However, the study did not report how many engineers were lost to follow up during
the study period, and whether the incidence of urolithiasis in different years had been
modified according to the changed number of follow-up engineers. Nevertheless, this
study provided useful evidence for the association between indoor heat exposure and
urolithiasis.
In total, nine studies indicated that the prevalence of urolithiasis was higher among
groups with heat exposure at workplaces than those without [16-21, 24-26], while
only one study failed to identify such an association [27]. The relative risk of
urolithiasis was found to be between two and ten times higher among heat-exposed
workers than non-heat-exposed ones (See Table 1).
20
2.6 Risk factors for heat-induced urolithiasis
Some cross-sectional studies [18, 20, 21] carried out metabolic evaluation to support
the association between heat exposure and urolithiasis. Risk factors such as length of
service, age, and social class have also been taken into consideration in these studies.
For example, age was a proven risk factor in many epidemiological studies on
urolithiasis [41-43] and was usually controlled for research purposes.
Three studies attempted to demonstrate the mechanism of heat-induced urolithiasis
formation in relation to biochemical changes in urine generated by heat exposure [18,
20, 21]. The association between metabolic changes and urolithiasis has been proven
through clinical studies [44, 45]. The following epidemiological studies also found a
link between biochemical changes and heat exposure. Atan et al. [21] performed a
urine metabolic evaluation among 59 workers without urolithiasis (34 heat-exposed
workers and 25 non-heat-exposed workers), finding significantly low urine volume
and hypocitraturia for heat-exposed workers. Borghi et al. [20] conducted a three-day
examination on 8-hour shift workers without urolithiasis in a glass plant, which
consisted of 21 randomly selected workers with and 21 workers without heat exposure.
Various indexes were tested for metabolic changes in urine samples, including uric
acid, pH value and specific gravity. A high level of uric acid, lower pH values and a
higher specific gravity were found in the heat-exposed group. Better et al. [18]
estimated calcium metabolism using blood and urine samples from 45 lifeguards and
21
20 controls from the general population matched with sex, age and season. Serum and
urine changes were measured, and hyperuricemia, higher serum level of 25-HCC, and
lower serum level of intact parathyroid hormone (iPTH) were observed among
lifeguards.
These studies provided information on biochemical changes associated with
dehydration and calcium excretion. Dehydration can cause low urine volume, high
specific gravity, high uric acid, hyperuricemia, and low pH [2]. Urinary calcium
excretion increase was attributed to high levels of serum 25-HCC induced by
increased sunlight exposure, being accompanied with low iPTH and a decreased
Mg/Ca ratio [34]. Dehydration and urinary calcium excretion increase were two well-
recognised risk factors for urolithiasis formation [46].
Length of service is an indicator for the cumulative amount of occupational heat
exposure if subjects have an equivalent dose of daily exposure. It is especially reliable
for stable jobs if someone continues doing the same type of work for a very long
period of time. There are mixed reports about the association between the length of
service and urolithiasis. For example, Atan et al. [21] found no significant difference
in the length of service between heat-exposed and non-heat-exposed groups. However,
Better et al. [18] found that lifeguards with a longer length of service were more likely
to suffer from urolithiasis than those with shorter periods of service.
22
Social class is another possible risk factor for urolithiasis in studies examining heat-
induced urolithiasis. Social class was considered to be linked with protein
consumption, viz. higher social classes having greater amount of protein consumption
than lower social classes [47]. Greater amounts of protein consumption was also
reported associated with urolithiasis formation [48]. However, Ferrie et al. [25]
studied the subjects‟ social classes and found no link between higher social class and
urolithiasis. In addition, a recent study reported that urolithiasis has spread to all
social classes as social and economic status has grown steadily [49]. Therefore,
influence of differing social classes on the prevalence of urolithiasis remains to be
determined [50-52].
Overall, the above risk factors provided important information for studying the
association between heat exposure and urolithiasis. Biochemical changes verified the
possible biomechanism in the impact of heat exposure on urolithiasis in previous
studies [18, 20, 21]. Length of service was a risk factor more closely related with
occupational heat exposure than other risk factors [18, 21], so it may be a useful
indicator for future research on occupational heat exposure and urolithiasis. Social
class used to be considered as an important predictor for urolithiasis [47] but it has
become less so in recent studies [49].
23
2.7 Knowledge gaps
Previous studies have shown that workers worked with exogenous heat stress
(generally indoors), marathon runners, lifeguards, soldiers and outdoor workers
(quarry drilling and crusher workers, quarry truck and loader drivers and postal
deliverymen) exposed to sunlight had as two to ten-fold increased risk of urolithiasis
compared with the general population [16-21, 24-26]. A variety of designs have been
used to examine any difference in the prevalence of urolithiasis between heat-exposed
and non-heat-exposed populations in many parts of the world [16-21, 24-26]. Some
studies also conducted a metabolic evaluation to provide supportive evidence for
verifying the causality between heat exposure and urolithiasis [18, 20, 21, 24], or
examined other risk factors such as social class and length of service [18, 25]. These
studies provided a basic pattern of the association between heat exposure and
urolithiasis, however these studies do have some limitations.
Firstly, a cross-sectional design was commonly used to identify the possible
association between occupational heat exposure and urolithiasis, but very few studies
employed an analytical design (e.g. cohort or case-control), making it more difficult
to draw causal inference.
Secondly, very few studies examined the effects of ambient heat exposure on
urolithiasis among outdoor working populations. Previous studies had studied both
24
lifeguards [18] and marathon runners [19] , but they might not be representative
samples for the general outdoor working populations. They were special working
populations exposed to ambient heat, high intensity sunlight with bare skin and/or
intensive physical activity at the same time. While Pin et al. [26] examined the
relationship between outdoor heat exposure (e.g. quarry drilling and crusher workers,
quarry truck and loader drivers and postal deliverymen) and urolithiasis, this study
was conducted in Singapore, which is a tropical, developed country. It is unlikely
their findings could be generalised to developing countries like China.
Thirdly, more attention should be paid to studying the effects of heat exposure on
urolithiasis among working populations in developing countries. Most previous
studies on occupational heat exposure and urolithiasis were carried out in developed
countries including the United States [16, 19], the United Kingdom [17, 24, 25], Israel
[18], Singapore [26], Italy [20] and Japan [27]. Only one is from a developing country
(Brazil) [21]. However, the situation in developing countries may be worse than
developed countries due to lower awareness of occupational health and lower income.
Therefore, the risk of occupational heat exposure for urolithiasis in many developing
countries remains to be determined.
Climate change is likely to put additional stress on ambient-heat-exposed working
populations [53, 54]. Based on this literature review, heat exposure may play an
important role in urolithiasis genesis. To the best of our knowledge, no study has
25
examined the association between ambient heat exposure and urolithiasis in a
developing country in Asia. It is necessary to conduct a study in China as it has the
largest manufacturing working population in the world, with large proportion of
workers exposed to ambient heat through industries such as shipbuilding, construction
and stevedoring service.
26
Table 1. Characteristics of previous studies assessing the effects of heat exposure on urolithiasis
Study and location Research Design Analysis methods Subject Exposure time Key findings
Atan, et al., 2005 [21],
Brazil
Cross-sectional Fisher‟s Exact Test and
Chi-square Test
10,326 male employees
working for more than 3
years at a large steel
plant
At least 3 years Workers exposed to high temperature had a higher
prevalence of urolithiasis (8.0%), ten times greater than
that for other workers (0.9%) not exposed to heat
(P<0.01).
Iguchi, et al., 1996 [27],
Japan
Cross-sectional Chi-square test,
Wilcoxon‟s rank-sum
test, and one-way
factorial analysis of
variance (ANOVA)
1,972 respondents of
3000 inhabitants
randomly selected from
20 to 59 years old in
Kaizuka City, Japan
Not mentioned The prevalence of urolithiasis among male administrative
workers was significantly higher than any other
occupational group, such as farmers and lumber workers.
Farmers and lumber workers without recorded stones
made up of the sample at 1.5%.
Borghi, et al., 1993
[20], Italy
Cross-sectional Chi-square test and
Student‟s t-test
236 man machinists
chronically exposed to a
hot environment in a
glass plant in Parma,
Italy and 165 workers
not exposed to heat in
the same factory
8hr/d, at least 6
months for 5
years
The prevalence of nephrolithiasis among heat exposed
machinists (8.5%) was three-fold higher than those not
exposed to heat (2.4%).
27
Pin, et al., 1992 [26],
Singapore
Cross-sectional Not mentioned 406 male workers in
several occupations,
including outdoor
workers, e.g. quarry
drilling and crusher
workers, quarry truck
and loader drivers and
postal deliverymen;
indoor workers, e.g.
postal clerks and
hospital maintenance
workers
The outdoor workers had increased prevalence compared
with indoor workers (5.2% vs 0.85%, P<0.05).
Ferrie, et al., 1984 [25],
Scotland
Case study Descriptive statistic
including counts and
percentages
47 patients with
urolithiasis
(38 males and 9
females)
Hot-metal
exposure from 1
to 50 years
(average 19.0)
Individuals with an occupational history of hot metals
process exposure were more likely to develop
urolithiasis.
28
Milvy, et al., 1981 [19],
the United States
Cross-sectional Not mentioned 1893 respondents of the
male participants in the
1977 New York City
Marathon
Not mentioned The prevalence among the marathon runners was 4.5
times greater than that of the matched population; the
prevalence of aberrant urine among respondents with
urolithiasis was nearly five times significantly higher
than that of those without urolithiasis.
Better, et al., 1980 [18],
Israel
Cross-sectional Student t-test 45 randomly selected
lifeguards out of a total
number of 120, and 50
people from the general
population matched for
sex, age and season
At least 8 hr/dy,
6 mo/yr
Lifeguards had an approximately ten-fold higher
prevalence of urolithiasis than the general population,
and lifeguards with urolithiasis had significantly longer
length of service than lifeguards without urolithiasis.
Parry, et al.,1975 [24],
the United Kingdom
and the Persian Gulf
Cohort Not mentioned Two groups of soldiers
(40 males in winter and
51 males in summer)
were transported from
the United Kingdom to
the Persian Gulf in cold
and hot seasons,
respectively
The group transported from the United Kingdom to the
Persian Gulf in summer had statistical significant
hypercalciuria while those in winter had not, which was
most likely to be caused by increased sunlight exposure.
29
Blacklock, et al., 1965
[17], the Britain, the
Mediterranean, the
Middle East and the Far
East
Cohort Not mentioned The navy soldiers in
Service from 1958 to
1964 in British Royal
Navy
Not mentioned Engineers worked in the engine room of the ship had
higher incidence of urolithiasis; the incidence of
urolithiasis was declined year after year probably because
of the increasing use of air-conditioning in engine rooms
in ships.
Pierce, et al.,1945 [16],
an overseas desert
Case study Descriptive statistic
including counts
61 diagnosed
urolithiasis cases among
a command of
American troops in an
overseas desert area
Not mentioned There was a relatively short period for urolithiasis
formation in the desert area; the prevalence of urolithiasis
among soldiers in the desert area was two times higher
than in the mountainous area.
30
Table 2. Metabolic evaluation for explaining heat-induced urolithiasis
Reference Subject
Selection criteria for
metabolic evaluation
Risk factors Key findings Comment
Atan, et
al., 2005
[21]
34 heat-exposed
workers and 25
non-heat-
exposed workers
Volunteer workers
without urolithiasis
Calcium, creatinine, uric
acid in serum; volume,
calcium, uric acid, citrate
and oxalate in 24 hours urine
Heat-exposed workers had greater
frequency of hypocitraturia
(P=0.03) and lower urinary
volume (P=0.01).
Hypocitraturia and low urinary volume
were proved risk factors for urolithiasis;
the significant difference on hypercitraturia
and low urinary volume between heat-
exposed workers and non-heat-exposed
workers supported the causal association
between heat exposure and urolithiasis.
Borghi, et
al., 1993
[20]
21 heat-exposed
workers and 21
non-heat-
exposed workers
Randomly selected;
without urolithiais,
family history, gout or
medications; examining
social class and dietary
habits
Volume, sodium, potassium,
calcium and magnesium,
chloride, creatine, uric acid,
phosphate, sulphate, oxalate,
citrate, ammonium, specific
gravity, pH, calcium
oxalate, calcium phosphate
and struvite
Fluid intake was significantly
different between heat-exposed
workers and non-heat-exposed
workers; uric acid was
significantly greater; pH was
significantly lower; specific
gravity was significantly higher.
Although fluid intake was greater among
heat-exposed workers, they still had higher
uric acid, lower pH and higher specific
gravity which were risk factors for
urolithiasis formation. This was supportive
evidence for the association between heat
exposure and urolithiasis.
31
Better, el
al., 1980
[18]
45 lifeguards
were randomly
selected from
120 lifeguards in
Israel.
They were randomly
selected out of the
general population in
Israel, matched for sex,
age and season.
Serum 25-HCC, mean urine
calcium, hyperuricemia,
hypercalciuria, iPTH level,
daily urine volume and mean
Mg/Ca ratio
Lifeguards had significantly
higher level of serum 25-HCC,
mean urine calcium,
hyperuricemia and hypercalciuria
than the general population; iPTH
level, daily urine volume and
mean Mg/Ca ratio were
significantly lower than the
general population.
Results of the metabolic evaluation were
high risk factors for urolithiasis formation,
so they were supportive to the association
between sunlight exposure (heat) and
urolithiasis.
32
33
CHAPTER 3: STUDY DESIGN AND METHODS
3.1 Aims and objectives
3.1.1 Aims
This study examined the association between ambient heat exposure and urolithiasis among
outdoor workers; and recommends strategies for minimising the impacts of ambient heat
exposure on urolithiasis among outdoor workers in China.
3.1.2 Specific objectives
Describe general demographic characteristics among cases and controls;
Quantify the relationship between outdoor heat exposure and urolithiasis;
Recommend strategies for minimising the impacts of ambient heat exposure on
urolithiasis among outdoor working populations in China.
3.2 Research design and data collection
3.2.1 Research design
A retrospective 1:4 matched case-control study was performed to investigate the association
between high ambient temperature exposure and urolithiasis among outdoor workers. The
statistical power of the study can be increased by selecting more than one control per case,
34
however, there is usually little marginal increase in precision from increasing the ratio of
controls to cases beyond four [55, 56]. Thus 1:4 as the case/control ratio was selected for this
study. To calculate sample size Dupont‟s method was used [57] for sample size calculation
for this matched case-control study, and the minimum effect size is 163 (sets) (α=0.05,β
=0.2). The calculation was carried out after a literature review for estimating the expected
prevalence in the general population in Guangzhou, with an estimated result for the
prevalence of urolithiasis (4.9% for both sex) derived from a nearby city (Shenzhen) [58-60].
Due to little information available on this topic in literature, it was decided to use as many
cases as possible and 1:4 case/control ratio to maximise the statistical power in this pilot
study.
In order to collect necessary background information of subjects, a small investigation was
conducted on all occupational health and safety officers (OHSOs) in the company studied
using forms with brief questions (See APPENDIX Ⅳ Information Collection Form to
OHSOs). This helped to estimate the hottest period in summer at the workplace studied,
lengths for ambient heat exposure for different types of work and their workloads, and to
collect information out of health surveillance data including the highest temperature records,
the supply of free drinking water, supplies of free meals, the arrangement of water breaks,
and the historical records for urolithiasis treatments from 2003 to 2010.
3.2.2 Study setting
Research setting: A long established shipbuilding company in Guangzhou City,
Guangdong Province, China
35
Guangzhou is the capital city of Guangdong Province in the southeast of China (See Figure
4). It is the third largest city in China, with a resident population of approximately 13 million
in 2010 according to the official statistical data [61]. It locates next to Hongkong with a total
area of 7434 km2
(22°26‟ to 23°56‟ north latitude, and 112°57‟ to 114°3‟ east longitude) [62].
As a subtropical city influenced by the Asian monsoon, Guangzhou has long summers that
are hot and humid, and short winters that are mild and dry. The Chinese meteorological
definition of summer is the period with average daily temperatures over 22 °C of at least 5
consecutive days. Summers of Guangzhou begin from May to October with mean monthly
temperature range between 25.9 to 28.6 °C [63] (See Figure 5). The mean annual temperature
of Guangzhou is 21.9 °C according to annual reports from Guangdong Meteorological
Bureau [63].
Figure 4. The location of Guangzhou City (the area of yellow colour)
36
Figure 5. Mean monthly temperature in Guangzhou City
Source: Guangdong Meteorological Bureau, 2010 [63]
The shipbuilding company in this study is over 160 years old and is located in the
southeast of Guangzhou City. The average number of employees is approximately 1, 600 and
includes workers building and repairing ships in the open boatyard and employees working in
offices. Outdoor workers are exposed to high levels of heat from the sun. The highest
temperature is experienced during summer from two to four o‟clock in the afternoons
according to the records of the company‟s OHSOs. The highest temperatures records at
different locations varied, e.g. an outdoor cement floor temperature up to 42 °C, a deck
temperature up to 45 °C, and a cabin temperature up to 52 °C. According to the records from
its affiliated hospital (suggested by OHSOs), there are on average one to two outdoor workers
with transient effects of heat-related illnesses such as heat cramps or even heat fainting and
heatstroke during hot days in this company.
37
3.2.3 Data collection
Ethics Application Exempt has been granted by the QUT University Human Research Ethics
Committee (See APPENDIX ⅡEthics Application Exempt). Guangdong Institute of Health
Inspection (GIHI) has also approved this study (See APPENDIX Ⅲ GIHI Approval Letter).
Researchers of this study promised to inform the management and workforce the research
findings and provide recommendations for the prevention of urolithiasis.
The health surveillance of workers was based on health checks performed according to
occupational health regulations in China [64]. Paper records of health surveillance data
covering 2003–2010 (available in 2003, 2005, 2007, 2008, 2009 and 2010) were collected
from the affiliated hospital of the shipbuilding company by occupational health inspectors at
Guangdong Institute of Health Inspection (GIHI). For the evaluation of the association
between occupational risk factors and urolithiasis among workers in the shipbuilding
company, researchers of this study worked with GIHI together to transfer the collected hard
copy data into an electronic version (without persons‟ names), and used the data for this study.
As these data were results of occupational health inspections, they were stored and assessed
within offices of GIHI according to the requirement of the approval from GIHI. Researchers
of this study therefore performed all data access and statistical analysis for the study on a
computer based in the GIHI office.
Prior to 2007, this company carried out health surveillance biennially, but changed to an
annual health check following this date. The total number of workers examined in the health
surveillance was 3288, which are twice as the annual average number of employees. This
number was made up of not only workers and employees continuously working during 2003
38
to 2010, but also older workers and employees retiring, and new workers and employees
recruited during the same period of time. The company had expanded its workforce during
2009 to 2010 because of an increase in business volume. About half of these workers were
normally exposed to heat at work.
Data collected in the health surveillance system include: health surveillance ID, age, sex,
birth place, the type of work, the department, the length of service, kidney ultrasonography
results, electrocardiogram (ECG), blood pressure (BP), urine test, blood test, hepatitis B virus
infection test, and health check years. Parts of the above information, which were believed to
be closely relevant to urolithiasis and heat exposure, were selected for statistical analyses in
this study.
To investigate the ambient heat exposure of workers, OHSOs in the shipbuilding company
were invited to estimate average exposure time per day for different types of work based on
their observation and experience. The total exposure time for subjects (both cases and
controls) was then calculated according to the average exposure time per day and a worker‟s
length of service.
3.2.4 Criteria for selection
Cases were determined by the positive results of kidney ultrasonography. Ultrasound is an
effective and commonly used screening test for urolithiasis with a sensitivity of 98% and
specificity of 74% [51]. Controls were selected from those health check records with negative
results of kidney ultrasonography, with the criteria for selection as follows: the same sex and
similar age (± 1 year old) matched with cases in the closest health check years. The order of
39
this selection procedure was sex, age, health check year, and health surveillance ID. For
example, when a male case was defined as 30 years old with health check in 2003, the
matched controls were selected from other 30 year old males with health check in 2003; this
selection process began from the male of the earliest health surveillance ID to the one of the
latest ID in 2003, e.g. from 440049000001 to 440049000002, then 440049000003. If the
health check file of an earlier health surveillance ID was lacking of some important
information, e.g. without information on type of work, department or BP, the next completed
health record file would be chosen as a control. When there was a lack of controls of 30 year
olds in 2003, controls amongst 30 year olds in 2005 would be checked instead. The rest were
selected in the same manner. If there was a lack of controls amongst 30 years old in all health
check years, controls amongst 31 or 29 year olds in the same health check year would be
checked beginning from the smallest health surveillance ID. A health surveillance ID could
only be chosen once in the process of subject selection. This avoided the situation of selecting
controls from cases that had passed out stones or had their stone removed surgically.
Missing values in the data set were less than 5%. A number of strategies were used to deal
with these missing values. Firstly, as the health surveillance data covered six years, some
missing values could be imputed by other years‟ health check records of the same subjects.
For example, if one‟s age was missing in the file of 2005 health check, the same information
from his or her health check files in 2003 and 2007 could be examined, and then his or her
age in 2005 could be calculated. Secondly, the definition of certain binary variables without
numerical results was clarified in the individual health check files. For the binary variables in
this study such as BP, besides those with the exact blood pressure results recorded, some
health check files in 2003 and 2005 did not contain the exact numeric BP results but only
reported “abnormal blood pressure” or “hypertension” or “normal” in the conclusion columns
40
of health checks. Due to this situation, this study defined that those subjects with “abnormal
blood pressure” or “hypertension” recorded in the conclusions of health check files belong to
subjects of “hypertension”, and those with only “normal” in the conclusion columns of health
checks were subjects of “normal blood pressure”. Thirdly, certain variables were excluded if
they had over 10% of missing values. For example, although urine test was thought to be one
of the good non-invasive measurements for potential urolithiasis screening [65, 66], the
records of urine test were not included in this study because about 30% of these values were
missing.
The data were examined to identify outliers. Normal Q-Q plots were used to examine
whether there were outliers in the data collected. No outliers were identified in these data
through a preliminary analysis.
Age and sex were chosen as the matched criteria for the following reasons. Age is a
recognised risk factor for urolithiasis. Epidemiological studies have found that older people
are more likely to have urolithiasis [27, 67-69]. Sex was chosen because males have been
found to have a higher prevalence of urolithiasis than females [10, 69-71]. Although a recent
study reported no difference in the prevalence of urolithiasis in males [68], sex was still
chosen as a matched criterion.
In terms of diet, the company provided at least two meals per day on workdays, cases and
controls were considered that had similar diet. All cases and controls in the study were born
in Guangdong Province, and they had the same race. Because of the physical requirement for
heavy workloads, it is to be expected that the prevalence of metabolic diseases would be
lower in workers of the shipbuilding company than the general population.
41
Possible risk factors in this study included type of work, department, total exposure time,
ECG and BP. The reasons for choosing these factors are explained below.
Type of work and department were the key occupational risk factors, and were used to
characterise the type of ambient heat exposure. These variables could not only reflect the
association between ambient heat exposure and urolithiasis, but could also categorise high
risk groups of outdoor workers. In this study, the subjects fell into eight work categories
including spray painter, smelter, welder, production security and quality inspector, assembler,
planing machine operator, gas-cutting worker and indoor employees. Six of these categories
worked mainly outdoors except for smelter and indoor employees who usually worked indoor.
Indoor employees were defined as employees working in rooms equipped with air-
conditioning. They consisted of executive staff in the company, teachers in its affiliated
kindergarten, primary school, middle school, and the medical staff in its affiliated hospital.
Department was also a risk factor, reflecting the association between work conditions
relevant to heat exposure and urolithiasis, however it was not used to define type of work due
to the fact that workers moved between departments regularly. For example, because the
shipbuilding and ship-repair departments shared workers from the same types of work
(including smelter, welder, production security and quality inspector, assembler, planing
machine operator, gas-cutting worker but not spray painter who was located only in ship-
repair department), when the volume of ships needing repair increased, the number of
workers in the ship-repair department would increase due to a movement of workers from the
shipbuilding department. As a result, the numbers of workers in the shipbuilding department
would also decrease. Given that the volume of workers in the ship-repair business was not
constant because clients would drop in when their ships were in need of repair due to
42
accidents or corrosion, the numbers of workers in shipbuilding and ship-repair departments
could often fluctuate. Hence, while it was hard to specify fixed numbers for these two
departments, fixed numbers could be defined for types of work according to the fact that the
movement between job titles was seldom because the workers in this company were technical
workers who had to be trained and certified before going on duty. Therefore, department was
used in this study as the supplement for type of work. Four categories of departments were
used in this study, including shipbuilding, ship-repair, production security and indoor
departments. The production security department consisted of several sections including
production security, quality security, manufacture management and occupational safety
security. These sections had similar period of ambient heat exposure during daily work.
Indoor departments included the human resource department, the administrative office, the
sales department, the accounting department, the affiliated kindergarten, the affiliated
primary school, the middle school, the technical school and the affiliated hospital.
Length of service, average exposure time and total exposure time were used to determine the
extent of ambient heat exposure. These data were collected from the health surveillance data
directly. For cases, length of service was determined by the year when the worker entered the
company and the year of first diagnosis according to kidney ultrasonography results. Average
exposure time was estimated based mainly on types of work and OHSOs‟ daily observation
and experience. Thus, average exposure time was expressed in terms of hours per workday.
Total exposure time was calculated by multiplying the adjusted length of service by average
exposure time under some necessary unit conversion (e.g. converting „hour‟ into „day‟, then
into „year‟). According to the actual situation of 5 workdays per week and 6 hot months per
year in Guangzhou, every adjusted year of length of service was around 130 workdays of heat
exposure. The calculation formula was 365 × (5/7) × (6/12) ≈ 130 days. Average exposure
43
time (hours) was multiplied by adjusted length of service (days) to get the result of total
exposure time (hours), and then the result was converted from hours into years.
Although the length of service could partly indicate cumulative heat exposure, it was not as
direct as total exposure time. Because total exposure time estimated the entire amount of heat
exposure for every worker or employee, the length of service only recorded the whole length
of employment from entering the company to the diagnosed time.
ECG result is a medical test that detects cardiac (heart) abnormalities by measuring the
electrical activity generated by the heart as it contracts. It is the most commonly performed
cardiac test reflecting heart diseases. Abnormal BP (hypertension) in this study was defined
as systolic blood pressure 140 mmHg or diastolic 90 mmHg or both. There are existing
occupational health policies that if workers have serious heart problems or/and hypertension,
they might be at risk when exposed to high levels of heat [49]. This means that heart disease
and hypertension are associated with heat exposure. Meanwhile, hypertension has also been
reported to be associated with urolithiasis [65]. So it is possible that abnormal ECG and
hypertension were risk factors for heat-related urolithiasis. Hence, the ECG and BP results
were included in this study as risk factors for occupational urolithiasis formation.
3.3 Data analysis
3.3.1 Descriptive statistics
44
The characteristics (e.g. sex, age, type of work, department, length of service, average
exposure time and total exposure time) of cases and controls were examined using descriptive
statistics, which described the association between ambient heat exposure and urolithiasis.
3.3.2 Univariate analysis
As the data in this study were of a 1:4 design, they could not be simply analysed the same as
the 1:1 matched design; the strata of 1 case to 4 controls had to be taken into consideration. In
SPSS, the Generalized Estimating Equations (GEE) suite offers analysis within clusters
which can also fit to analyse the 1: n matched case-control data; and it can analyse variables
of different categories (e.g. scale variables, binary variables and nominal variables). This
study used GEE for the univariate analysis of scale variables (age, length of service, average
exposure time and total exposure time), binary variables (sex, BP and ECG) and nominal
variables (type of work and department). In order to understand more about the distribution
of different levels of total exposure time between cases and controls, total exposure time was
divided equally into three range groups representing low, medium and high levels of
exposure, and percentages of these three levels were compared between cases and controls.
The distribution (percentages) of low, medium and high levels of total exposure time among
eight types of work was also described.
3.3.3 Selecting variables for modelling process
Because total exposure time is a measure of average exposure time and length of service, they
were strongly correlated with each other. So it was necessary to select one of them for
entering into the multivariable model. Total exposure time was selected because it derives
45
from ambient heat exposure and length of service and directly represents ambient heat
exposure.
As type of work and department provided similar occupational information from two
different layers, it was possible that they would suffer from collinearity when put into the
model together. In order to check whether there was collinearity with each other, both type of
work and department were put into the conditional logistic regression model together. If there
was collinearity, type of work would be chosen to put into the multivariable model with other
risk factors together for the main analyses, as type of work was more constant and could
provide more detailed information than department. Department could be used to test the
sensitivity of the final multivariable model.
3.3.4 Bivariate and multivariable modelling
A conditional logistic regression analysis by fitting a Cox regression model was used in the
modelling process. Model building used the purposeful selection of covariates approach [72].
3.3.4.1 Conditional logistic regression
The matched study is an important special case of the stratified case-control study. The
rationale for matched studies is discussed by Breslow and Day [73], Kleinbaum, Kupper, and
Morgenstern [72], Schlesselman [74], Kelsey, Thompson, and Evans [75] and Rothman and
Greenland [76], discussed the rationale for matched studies. In matched studies, subjects are
stratified on the basis of variables supposed to be associated with the outcome, so they do not
interfere with the covariates being studied. By design, a matched case-control study controls
46
for the effect that is matched on. An assumption of matched design is that the effect of
covariates is expected to be the same across strata.
Conditional logistic regression is useful in investigating the association between an outcome
and a set of explanatory variables in a matched case-control study. The outcome in such a
study is whether the subject is a case or a control.
When there is one case and one control in a matched set, the matching is 1:1. A “trick” can be
used to allow a standard logistic regression approach to fit a conditional logistic regression.
For the ith
matched set, let ui to be the covariate vectors for the case and let vi1,…, vini be the
covariate vectors for the ni controls. The likelihood for the M matched sets is given by
M
in
j ij
i
i
v
u
11
)'exp(
)'exp()(
for the 1:1 matching, the likelihood is reduced to
M
i ii
i
vu
u
1 1 )'exp()'exp(
)'exp()(
by dividing the numerator and the denominator by exp (v‟i1β), one obtains
)'exp(1
)'exp()(
1
1
1 ii
iiM
i vu
vu
47
When there is one case and multiple controls in a matched set, the matching is 1: n, Cox
regression can then be adapted to fit the conditional logistic regression model [43].
Cox regression is a widely used method, which was firstly introduced by Cox in 1972. It is
often used on censored survival or other time-to-event data for identifying differences in
survival due to treatment and prognostic factors in clinical studies.
The basic model is as the following:
λ (t, Xi) = λ0(t) exp (β‟ Xi)
for i =1,…, N.
N is the number of individuals in the study; T is the failure time and censored or not. So
its observed vector is (ti, δi, Xi). It assumes that the hazard function for failure time T for an
individual i with covariate vector X‟i= (x1i , x2i ,…,xki ,…,xKi) is the equation above [73].
The Cox model is a semi-parametric model since it does not specify the form of λ0(t).
However, it specifies the hazard ratio for any two individuals with covariate vectors X1 and
X2 [73], given that the ratio does not depend on time:
)]('exp[)'exp()(
)'exp()(
),(
),(21
20
10
2
1
t
t
t
t
(1)
48
Therefore, this model is a proportional hazard regression model. It assumes that the failure
rates of any two individuals are proportional throughout the whole survival experience. If
X1=1, X2= 0, it becomes:
)'exp()(
)'exp()(
)0,(
),(
0
0
t
t
t
t
(2)
λ0 (t) is often termed the baseline hazard as it may be regarded as the hazard function of all
covariates are zero.
The model above is articulated for the purposes of modelling time-to-event data. However,
the use of this model for the conditional logistic regression needs modifications, and the
adapted Cox regression model can be used in matched case-control studies for estimating
proportional odds, rather than the proportional hazard. In this case the odds are substituted
into equation (1) and (2).
In SPSS 18.0, the Cox proportional hazard regression can be adapted to fit a conditional
logistic regression for 1: n matched case-control studies.
3.3.4.2 Purposeful selection of covariates
When selecting variables for inclusion in the regression model, there were two main aspects
to consider. Firstly, theoretically for the sake of obtaining a numerically stable and easily
generalizable model, the number of selected variables should be minimized. Secondly, from
the perspective of epidemiologic methodologists, in order to have a possible complete control
49
of confounding [11, 77], all clinically and intuitively relevant variables should be included
into the model.
Hosmer [72] proposed a six-step approach to select covariates to build a multivariable model
which is appropriate for meeting the objectives illustrated above. This approach is called
“Purposeful Selection of Covariates” and while being mainly empirically driven, it allows the
researcher the discretion of ensuring that clinically important covariates are considered. It is a
method completely controlled by the researcher according to the research aim as opposed to
other statistical methods such as stepwise and best subsets selection [68]. This approach is
used here to build an appropriate conditional logistic regression model.
Step 1. Bivariate analysis
A bivariate analysis of each variable was conducted as the first step of the selection. This
produced crude odds ratios (ORs) estimates measuring the association between individual
risk factor and urolithiasis. Every variable was individually added to a conditional logistic
regression model to test whether it had a significant effect resulting in a table of crude ORs
and their corresponding P-values and confidence intervals.
Variables with „P-value<0.25‟ were added to enter an initial multivariable model in Step 2;
but those with „P-value>0.25‟ were excluded from bivariate analysis.
This screening criterion used in Step 1 for the variable selection was based on previous
studies on linear regression and logistic regression [68]. It showed that the traditional
inclusion criterion of P-value<0.05 often failed to identify variables known to be important
50
[12, 78]. The higher P-value of 0.25 allowed the inclusion of all statistically significant
variables and those with the potential to be important confounders [68].
Step 2. Multivariable modelling
The second step was to fit an initial multivariable model with a set of covariates chosen from
Step 1; in this case, covariates with P-value<0.05 were kept in the model, and those with P-
value>0.05 were excluded.
Step 3. Reintroduce variables excluded from the initial multivariable model
The third step was to check whether the variables excluded from the initial multivariable
model are confounders, or have become significant in the new model. Also, at this step, it
was decided, according to the research aim, whether variables should be added to the model.
Following Step 2, covariates with P-value<0.05 were kept; and in this step, all variables with
P-value>0.05 were reintroduced one at a time into the model, and re-analysed to see whether
they were confounders or statistically significant.
A covariate was considered a confounder if its addition to the model resulted in a change≥20%
in the coefficient of covariates kept in the model. Confounders and variables important to the
research aim were kept in the model as well for the following steps.
Step 4. Reintroduce variables excluded from bivariate analysis
51
The fourth step was to check whether the variables excluded from bivariate analysis were
confounders, or had become significant in the new model. Also, at this step any covariates
important to the research aims were added to the model.
After adjusting the multivariable model in Step3, variables excluded after bivariate analysis
were reintroduced into the model one at a time. This helped in identifying covariates that
were not significant by themselves but contributed to other present covariates, or were
identified as a confounder (as defined in step 3).
Step 5. Check for effect modification
The fifth step was to check whether the covariates remaining in the multivariable model
interacted with each other. That is, whether one covariate acted as an effect modifier of one
or more of the other covariates. An interaction was considered significant if the interaction
had a corresponding P-value<0.05. Only two way interactions were considered.
Step 6. Model adequacy
After checking the interaction of variables in Step 5, the final multivariable model was built.
Once the covariates to be included in the model were determined, the adequacy of the model
needed to be assessed. In the case of logistic regression models, this was done by simply
assessing overall model significance.
52
After the main multivariable model was built, a sensitivity test using the same process was
conducted but replacing variables with variables of similar information, for example,
replacing “type of work” with “department”. Then the adequacy of the model was tested.
In this study, statistical analysis was performed using the SPSS 18.0 software package; in all
cases, two sided tests and a significance level of 0.05 were used.
53
CHAPTER 4: RESULTS
4.1 Descriptive statistics and univariate analyses
In this study, there were 190 cases and 760 matched controls, with 83% of males among
cases and controls and the age range of total subjects being from 20 to 59 years. Age was also
matched well among cases and controls (38±11 years). Details are described in Table 3.
Length of service, average exposure time and total exposure time were three potential risk
factors indicating occupational ambient heat exposure. All these three were significantly
longer among cases compared with controls. Length of service of subjects ranged from 3 to
38 years, and was significantly longer among cases (17±11 years) than controls (16±11 years)
(P<0.01). Average exposure time ranged from 0.5 hours to 8 hours, and it was significantly
longer among cases (7±3 hours) than controls (5±3 hours) (P<0.01). Total exposure time
ranged from 0.0 years to 4.5 years, and it was also statistically longer among cases (1.6±1.4
years) than controls (0.9±1.1 years) (P<0.01). For sub-group comparisons, cases and controls
were allocated into three levels of total exposure time: 0.0 to 1.5 years for low level exposure,
1.5 to 3.0 years for medium level exposure, and 3.0 to 4.5 years for high level exposure (See
Figure 6). 25.8% of cases had a high level exposure compared to 11.6% of controls (χ2=24.9,
P < 0.00), which indicated that cases had a significantly greater percentage of high level
exposure than controls.
54
Figure 6. Bar graph of percentages by three levels of total exposure time
Types of work included eight different categories, such as 297 indoor employees (31.3% of
total subjects), 16 gas-cutting workers (1.7%) and 12 planing machine operators (1.3%).
Among cases, 53 welders (27.9%), 41 assemblers (21.6%) and 16 smelters (8.4%) were three
larger types of work. Among controls, 270 indoor employees (35.5%) were the largest, 162
assemblers (21.3%) the second, and 8 planing machine operators (1.1%) the smallest. There
was statistically significant difference between cases and controls in “type of work” as a
whole (P<0.01). In order to show the difference in total exposure time among types of work,
cases and controls were put into three groups (See the previous paragraph) using three levels
of total exposure time (0.0 to 1.5 years for low level exposure, 1.5 to 3.0 years for medium
level exposure, and 3.0 to 4.5 years for high level exposure) for each type of work. The
results showed that cases with outdoor types of work usually had a medium to high level heat
exposure (See Figure 7).
55
Figure 7. Bar graph of percentages by three levels of total exposure time among eight types
of work
Four categories of departments in this study were analysed, namely shipbuilding, ship-repair,
production security and indoor departments. The shipbuilding department contained 37.6% of
the total subjects, with indoor, production security and repair departments representing for
31.9%, 16.7% and 13.8% of the total, respectively. Among cases, 86 workers (45.3%) were
from the shipbuilding department, with 40 workers (21.1%) from ship-repair, 33 workers
(17.4%) from production security, and 31employees (16.3%) from indoor departments.
Among controls, employees from indoor department and workers from the shipbuilding
department accounted for 35.8% and 35.7% of the cases, respectively. There was a
statistically significant difference of the percentage composition of departments between
cases and controls (P<0.01) (See Table 3).
56
In addition, statistical analyses showed that there was no significant difference in abnormal
ECG and hypertension between cases and controls. It was considered that heat exposure
could induce and exacerbate both heart diseases and hypertension according to the literature.
Hypertension was also reported to be associated with urolithiasis [65]. However, the
descriptive analyses suggested that abnormal ECG and hypertension might not be risk factors
of urolithiasis in this particular working population.
57
Table 3. Descriptive statistics and univariate analysis of variables
Case Control P-value
M ± SD M ± SD
Age 37.8 ± 11.4 37.8 ± 11.4 0.45
Length of service (yr) 16.7 ± 11.3 15.9 ± 11.0 0.00**
Average exposure time (hr) 6.6 ±2.6 5.0 ± 3.4 0.00**
Total exposure time (yr) 1.6 ± 1.4 0.9 ± 1.1 0.00**
Sex Frequency (%) Frequency (%) 1.00
Male 158 (83.2) 632 (83.2)
Female 32 (16.8) 128 (16.8)
Type of work
0.00**
Welder 53 (27.9) 125 (16.5)
Assembler 41 (21.6) 162 (21.3)
Production security and quality inspector 33 (17.4) 126 (16.6)
Smelter 16 (8.4) 32 (4.2)
Planing machine operator 4 (2.1) 8 (1.1)
Spray painter 12 (6.3) 25 (3.3)
Gas-cutting worker 4 (2.1) 12 (1.6)
Indoor 27 (14.2) 270 (35.5)
Department
0.00**
Shipbuilding 86 (45.3) 271 (35.7)
Ship-repair 40 (21.1) 91 (12.0)
Production security 33 (17.4) 126 (16.6)
Indoor 31 (16.3) 272 (35.8)
ECG
0.66
Abnormal 39 (20.5) 146 (19.2)
Normal 151 (79.5) 614 (80.8)
BP
0.08
Abnormal 42 (23.7) 129 (17.0)
Normal 148 (77.9) 631 (83.0)
** P-value < 0.01
58
4.2 Overlapping effect
When “type of work” and “department” were entered into a conditional logistic regression
model, there was an obvious overlapping effect which led to the absence of production
security department (See Table 4). The reason was that the “production security department”
provided the same information as “production security and quality inspector” when both put
into one model together. This was another reason why “type of work” and “department” were
not included in the same model.
Table 4. Overlapping effect between type of work and department
Covariates Degree of freedom OR [95% CI] P-value
Type of work 7 0.00
Welder 1 4.4 [2.0–9.4] 0.00
Assembler 1 2.6 [1.2–5.8] 0.02
Production security and quality inspector 1 3.8 [2.0–6.9] 0.00
Smelter 1 4.4 [1.7–11.2] 0.00
Planing machine operator 1 6.0 [1.4–25.6] 0.02
Spray painter 1 4.9 [1.6–14.6] 0.00
Gas-cutting worker 1 3.4 [0.9–13.4] 0.07
Department 2a 0.10
Shipbuilding 1 1.4 [0.7–2.8] 0.34
Ship-repair 1 2.1 [1.0–4.4] 0.05
a. Degree of freedom reduced because of constant or linearly dependent covariates
b. Constant or Linearly Dependent Covariates S=Stratum effect. Production security
department = Production security and quality inspector (type of work) + S.
59
4.3 Association between risk factors and urolithiasis
Due to a strong correlation, it was decided that type of work and department should be
examined in two separate models. Therefore, the model using type of work was performed
for the main analysis in this section, and the model using department was run as a sensitivity
test in section 4.4.
In Step 1 of the Purposeful Selection of Covariates, four risk factors, including type of work,
total exposure time, ECG and BP, were put into bivariate analyses one at a time.
Three risk factors (except ECG) with statistical significance (P<0.25) from Step 1 were
added into an initial multivariable model (Step 2), which was statistically significant (χ2=1.0,
df=9, P<0.01). All three covariates were of statistical significance (P<0.05). Thus type of
work, total exposure time and BP were added to the new model.
In Step 3, no variable was reintroduced into the new model because no variable was
excluded in Step 2. In Step 4, ECG was reintroduced into the model, and it showed no
statistical significance itself in the model. Also, it was not considered as a confounder for the
present covariates (type of work, total exposure time and BP). Thus ECG was still excluded
from the model.
Step 5 was an effect modification step which means the check for two way interaction
between every pair of covariates in the model. No significant interaction was found in this
step. In Step 6, the final accepted model included three covariates. Their ORs, 95% CIs and
P-values are presented in Table 5.
60
Table 5. Odds ratios of risk factors of urolithiasis in the conditional logistic regression model
Covariate Crude1 OR
[95%CI]
P-value
(P<0.25)
Adjusted2 OR
[95%CI]
P-value
(P<0.05)
Type of work 0.00 0.00**
Welder 6.1 [3.4–10.9] 0.00 3.7 [1.9–7.2] 0.00**
Assembler 3.6 [2.0–6.6] 0.00 2.2 [1.1–4.3] 0.03*
PSQI3 3.5 [1.9–6.3] 0.00 2.7 [1.4–3.0] 0.00**
Smelter 6.4 [2.9–13.9] 0.00 4.0 [1.8–9.2] 0.00**
PMO4 7.2 [1.9–27.4] 0.00 4.0 [0.9–16.6] 0.06
Spray Painter 6.9 [2.7–17.7] 0.00 4.4 [1.7–11.4] 0.00**
Gas-cutting worker 4.3 [1.3–14.6] 0.02 2.6 [0.7–9.1] 0.15
Indoor employee REFERENT
Total exposure time (yr) 1.7 [1.5–2.1] 0.00 1.5 [1.2–1.8] 0.00**
BP 1.5[1.0–2.2] 0.08 1.6 [1.0–2.5] 0.04*
ECG 1.1 [0.7–1.6] 0.68 N/A (not entering the final model)
** P-value < 0.01; * P-value < 0.05.
1Crude ORs were results from the bivariate analysis in Step 1.
2Adjusted ORs were results from the final multivariable model in Step 6.
3PSQI: Production security and quality inspector.
4PMO: Planing machine operator.
When adjusted for total exposure time and BP, the odds of urolithiasis for welder were 3.7
times higher than indoor employees. Similarly, the ORs of smelter, spray painter, production
security and quality inspector, and assembler were adjusted. They suggested that the odds of
smelter, welder, spray painter, production security and quality inspector, and assembler of
having urolithiasis were increased by 302%, 273%, 293%, 168% and 116% compared to
those of indoor employees (See Table 5 and Figure 8). The OR of total exposure time was 1.5,
meaning that if the total exposure time increased by one more year, the odds of having
urolithiasis would increase by 50%. The OR of BP was 1.6, which meant that if workers or
employees had hypertension, their odds of urolithiasis increased by 60% compared to those
having normal blood pressure.
61
Figure 8. Odds ratios and 95% confidence intervals for seven different types of work
*PMO: planing machine operator; PSQI: production security and quality inspector; referent is
the red line represented for indoor employees.
4.4 Sensitivity test of the multivariable model
As “department” provides similar occupation information as “type of work”, “department”
was used to replace “type of work” to test the multivariable model built in the previous
process.
There were three covariates included in the model under the category of “department” (χ2
=55.05, df=5, P<0.05) (See Table 6). The adjusted ORs of total exposure time and BP
62
changed little when adjusting for “department” instead of “type of work” in the modelling
procedure.
Table 6. Odds ratios of risk factors for urolithiasis in the conditional logistic regression
model for a sensitivity test
Covariate Crude OR
[95%CI]
P-value
(P<0.25)
Adjusted OR
[95%CI]
P-value
(P<0.05)
Department 0.00* 0.08
Shipbuilding 3.5 [2.2–5.8] 0.00* 1.9 [1.0–3.6] 0.04
Ship-repair 5.0 [2.8–9.0] 0.00* 2.5 [1.2–5.2] 0.01
Production Security 2.8 [1.6–5.0] 0.00* 2.0 [1.1–3.7] 0.03
Indoor REFERENT
Total exposure time (yr) 1.7 [1.5–2.1] 0.00 1.5 [1.2–1.9] 0.00**
BP 1.5 [1.0–2.2] 0.08 1.6 [1.0–2.5] 0.03*
ECG 1.1 [0.7–1.6] 0.68 N/A (not entering the final model)
** P-value < 0.01; * P-value < 0.05.
In conclusion, the analysis used descriptive, univariate and bivariate analyses, and a
conditional logistic regression modelling process to explore associations of risk factors with
urolithiasis. Descriptive and univariate analyses showed that cases had significantly longer
length of service, average exposure time and total exposure time than controls. Compared to
controls, cases also had a greater percentage of medium and high levels of total exposure time.
The percentages among different types of work were significantly different between cases
and controls, as were the percentages among different departments.
In the final conditional logistic regression model, results showed that the OR of urolithiasis
for total exposure time was 1.5 (95% CI: 1.2–1.8) in the multivariable model (Table 5). This
implies that if the total exposure time increased by one more year, the OR of urolithiasis
increased by 50%. Five types of work (spray painter, smelter, welder, production security and
63
quality inspector, and assembler) had higher odds of urolithiasis compared with indoor
employees, which meant that workers of these five types of work were significantly more
likely to have urolithiasis. The values of adjusted ORs and 95% CIs for the workers were:
spray painter (OR=4.4, 95% CI: 1.7–11.4), smelter (OR=4.0, 95% CI: 1.8–9.2), welder
(OR=3.7, 95% CI: 1.9–7.2), production security and quality inspector (OR=2.7, 95% CI: 1.4–
3.0), and assembler (OR=2.2, 95% CI: 1.1–4.3). The OR of urolithiasis for hypertension was
1.6 (95% CI: 1.0–2.5) in the multivariable model (see Table 5). These data imply that
workers with hypertension were significantly more likely to have urolithiasis compared to
workers with normal blood pressure levels.
64
65
CHAPTER 5: DISCUSSION AND CONCLUSION
5.1 Key findings in this study
This study demonstrated that there was a clear association between ambient heat exposure
and urolithiasis among outdoor workers in Guangzhou, China. Workers with longer lengths
of total exposure time were more likely to have urolithiasis. Workers with outdoor work were
more likely to have urolithiasis than those in other types of work. Workers with hypertension
were also more likely to have urolithiasis compared to those with normal BP levels.
Therefore, this study provides evidence that ambient heat could contribute to increasing
urolithiasis prevalence, and therefore, outdoor workers are more likely to develop urolithiasis
compared to indoor employees.
5.2 Public health implications
This study found a significant association between ambient heat exposure and urolithiasis
among shipbuilding workers in a subtropical area in China. The results have several major
implications for both the planning and implementation of public health interventions.
Firstly, health surveillance data can be an important source of reporting health outcomes (e.g.
urolithiasis) resulted from ambient heat exposure, and it can help to identify high risk
industries and occupations. This provides useful information for decision-makers to develop
intervention strategies to minimise adverse health outcomes from ambient heat exposure. For
example, to recommend training on heat-induced urolithiasis prevention, to detect the
66
incidence of urolithiasis in high risk industries, and to implement governmental notification
programs if the incidence of urolithiasis continue to increase due to lacking of necessary
occupational health management performed by employers. These strategies will promote
spontaneous prevention of heat-induced urolithiasis in both outdoor and indoor workers. In a
long run, this can help to reduce the substantial financial burden to workers, industries and
health care system.
Secondly, this study suggests some characteristics for identifying the high risk types of work
in subtropical or tropical areas. For example, workers in long uniforms and full-covered
personal protective equipment (PPE) with long total exposure time to ambient heat are more
likely to have urolithiasis; occupations working with a heat source, heavy workloads and
fully-covered PPE are also at high risk; and workers with hypertension are susceptible to
urolithiasis. Therefore, after the research findings and recommendations were informed,
decision makers at a local level can use these characteristics to identify high risk groups for
implementing interventions, e.g. availability of free water supplies, reasonable water breaks
and working time schedules, and decreased workloads during hot days.
Thirdly, the findings of this study show that an annual health check program should include a
urolithiasis examination for outdoor working populations. This study was based on health
surveillance data collected from a shipbuilding company in Guangzhou, China. However,
urolithiasis-relevant medical examinations (e.g. kidney ultrasonography and abdominal X-ray)
were not an essential item for the health check program of shipbuilding workers according to
the official health surveillance requirements in China [66]. The shipbuilding company in this
study was one of a few companies that had their workers undertake regular health checks
with a urolithiasis examination, and it provided important medical information on heat-
67
induced urolithiasis among outdoor workers. Hence, it is necessary to advocate the
incorporation of a urolithiasis examination in the health check program for outdoor workers.
More attention should be paid to preventing adverse health outcomes from long-term ambient
heat exposure.
5.3 Recommendations
From the findings of this study, preventative intervention should be taken to minimise the
adverse impacts of ambient heat exposure on urolithiasis among outdoor workers. There are
some practical recommendations on workplace and personal interventions and they are listed
as the follows.
Firstly, training is necessary to educate field supervisors and outdoor workers to recognize,
reduce and prevent ambient heat impacts on urolithiasis, especially for those at higher risk.
Some knowledge and skills are important for protecting outdoor workers from heat impacts
(e.g. effectively replenishing water such as drinking small amounts of water frequently before
they feel thirsty [79], avoiding working under the sun during the hottest period of hot days,
organising water breaks for rest and rehydration in cool areas and reducing workloads during
hot days).
Secondly, reducing workloads in hot months is an important measure for urolithiasis
prevention. According to “Threshold Wet Bulb Globe Temperature (WBGT) Levels”, the
heavier the workloads are, the lower the threshold levels should be [80, 81]. This means that
workers are more vulnerable to temperature increases when they have heavier workloads.
Because heavier workloads require more energy, such work usually causes an increase in
68
body temperature in a short period of time. This can stimulate the thermoregulation system of
the body to balance the core body temperature through elevated skin blood flow, sweating
and respiration. Thus heavy workloads will increase water loss and aggravate dehydration
status facilitating urolithiasis [79]. Workers with heavier physical work under ambient heat
were more likely to have urolithiasis. Hence, it is necessary to reduce workloads in hot
months. One method of achieving this is to shorten work time. This can be integrated with
working time schedule adjustment in hot months, which includes avoiding working during
the hottest period of the day (e.g. 1 to 4 pm) and establishing work/rest schedules appropriate
for the current heat indices [79]. An appropriate work/rest regimen can also be beneficial. For
example, Attia et al. [82] performed a field study on heat stress and recovery welders and
found that the optimization work/rest regimen was work (2h)/rest (2h) because welders had
almost full recovery from heat stress, while welders of work (2h)/rest (1h) or work (1h)/rest
(1h) did not reach complete recovery during resting period.
Thirdly, it is necessary to encourage appropriate fluid intake [83, 84], because adequate
hydration is a primary element in preventing urolithiasis [83, 85]. This includes the proper
kinds of and the appropriate amount of liquid consumption. Evidence has shown that if
acclimatisation is good and sufficient liquids are consumed, the effect of heat on urolithiasis
is minimal [86]. However, the proper amount of liquid intake is yet to be determined. Since
normal urine is supersaturated with various stone forming salts, it is considered that urine
contains potent inhibitors for controlling crystal formation, aggregation and subsequent stone
formation [70]. Urolithiasis formation is usually determined by how much the urine is
supersaturated with respect to calcium oxalate and calcium phosphate on the one hand, and
the activity of various inhibitors on the other [71]. It is possible that too much fluid intake
will reduce the inhibitory activity sufficiently to negate the beneficial effect of reduced
69
saturation. This was recognised in an epidemiological study performed by Tucak et al. [41]
which reported that people consuming less than 2 litres of liquid every day had the highest
incidence of urolithiasis, people with more than 2 litres was the second, and people with
around 2 litres had the lowest incidence of urolithiasis. In terms of the general population, 2
litres of liquid intake every day is recommended. However, for the working population under
ambient heat, it is better to measure their hydration status as they lose more water through
perspiration, and their liquid intake will therefore be greater than that of the general
population.
In order to adequately replenish water loss and encourage appropriate amount of fluid intake
for heat-exposed workers, hydration status needs to be measured. According to the
experience from Australian mining industry, measurement of urine gravity is a practical way
for identifying hydration status [87, 88]. It can be easily performed before work and after
work with the Hydrate 1TM
System [89] or with specific gravity test strips [90]. This
measurement indicates whether workers have condensed urine or not, which suggests the
needs to take more fluid or not. This measurement can also provide information to
supervisors on managing the fluid supply and the frequency of intermittent water breaks [79].
Specific gravity of urine measurement has been implemented in the mining industry for a
long period of time, and has successfully prevented miners exposed to underground heat from
heat stroke and urolithiasis in Australia [87, 88].
Fourthly, access to shade or cooling areas such as air-conditioned rooms or shaded areas with
fans is a common but important method [79] for preventing the core body temperature from
reaching over-heating status. The human body must maintain a core temperature between
36.0–37.2 °C to keep functioning properly [91], and if the core body temperature increases,
70
the body will deal with heat stress through thermoregulation, which includes elevated skin
blood flow, sweating and respiration. When water loss generated by thermoregulation is not
compensated in time, dehydration can lead to acute effects such as exhaustion and even heat
stroke, and if the human body sweats over an extended period, it can result in chronic
dehydration and a decrease in the volume of urine. These are important risk factors for
urolithiasis. Thus, providing access to shade or cooling areas can also be an important
interventions for preventing urolithiasis by minimising dehydration.
Moreover, there are recommendations for urolithiasis prevention at an individual level, which
can be supplemented through workplace training programs. Poor personal hygiene habits can
fail to protect them from getting unnecessary chemicals or heavy metals of renal toxicity into
their body. For example, it may increase the potential cadmium absorption if spray painters
do not change their clothes and make sure their hands and mouths free of cadmium-
containing paint pigments before drinking water or having meals at the workplace. Besides,
individual recommendations should include suggestions on an increased dietary intake of
calcium. Lower calcium intake levels can stimulate 25(OH)D mediated absorption and
urinary excretion of calcium, while increased dietary intake of calcium is not associated with
a greater prevalence of urolithiasis [92, 93]. Recommendations should also contain
suggestions to increase appropriate levels and kinds of fluid consumption. A low urine
volume was considered to be an important risk factor for calcium stone formation, so a high
fluid intake may be beneficial in the management of urolithiasis. Distilled water is
recommended to be consumed, but one might anticipate adverse effects of various kinds of
other fluids (e.g. beer, mineral water, and soft drinks). Beer may contain considerable
amounts of calcium, oxalate and guanosine, which is metabolized almost completely to uric
acid [94-96]. Bicarbonate alkaline water and mineral water should also be avoided because
71
drinking bicarbonate alkaline water with a high content of calcium leads to an increase in the
urinary calcium level by 50% [97]. It was found that replacing one litre of the usual fluid
intake with mineral water elevated the urinary calcium for normal subjects as well as
urolithiasis cases [98, 99]. The high oxalate content of soft drinks (e.g. apple juice and
grapefruit juice) are common examples for adverse effects [100]. Moreover,
recommendations should cover the promotion of healthy behaviours that will benefit
urolithiasis prevention, e.g. abatement or cessation of smoking, lower consumption of animal
protein and fat, and higher consumption of vegetables [101, 102].
5.4 Alternative explanations
5.4.1 Chance
In this study, the association between ambient heat exposure and urolithiasis was examined,
taking into account total exposure time, type of work, ECG and BP. These findings are
unlikely to be explained by chance because of the following reasons:
Firstly, all the statistically significant tests were driven by the hypotheses. All available data
were analysed and a whole spectrum of research findings was presented.
Secondly, there is consistent evidence that urolithiasis is associated with ambient heat
exposure among outdoor workers. Such evidence was not only observed for total heat
exposure time, but also through the association with different types of work with different
workloads and PPE.
72
Thirdly, rigorous statistical approaches were used to examine the association between
ambient heat exposure and urolithiasis. The multivariable logistic regression models using
“types of work” or “departments” produced similar results, which strengthened the validity of
the research findings.
5.4.2 Bias
Information bias could occur in the process of data collection, although the health
surveillance data from the affiliated hospital of the company was considered reliable as they
were from original hard copies of health check files in this study. Firstly, measurement error
in urolithiasis diagnosis could be a source of information bias. Kidney ultrasonography
results from different health exam years were used as the main medical evidence for
urolithiasis diagnosis in this study. The kidney ultrasonography results were measured in the
same hospital but from different years (from 2003 to 2010). This means that there is a
possibility of diagnostic bias because the ultrasonic tests were conducted by different
practitioners using a variety of somascopes. Secondly, bias may have occurred from the strict
selection criteria for cases, which may have excluded some potential correct cases. For
example, three cases with less than three years of employment were not selected for analysis,
because an estimated time for generating urolithiasis among the general population was
longer than three years [68]. Although these cases could still be associated with ambient heat
exposure at work due to individual differences, the influence from the three excluded cases is
impossible to be significant when compared to the other 190 cases. Therefore, information
bias of such kind is unlikely to change the results remarkably. Thirdly, information bias
might also affect calculation for total exposure time and the estimation of average exposure
time, because both of these indices were not from actual measurement but based on OHSOs‟
73
observation and experience. However, they were a useful estimation when the actual
measurement was difficult to carry out for a chronic disease which required a long period of
follow-up observation, and for a project with unidentifiable secondary data which were not
able to find those exact subjects for further information on previous ambient heat exposure.
5.5 Comparison with other studies
5.5.1 Previous studies on ambient heat and urolithiasis
Urolithiasis constitutes a major health problem worldwide and many studies report the role of
heat as a significant risk factor for urolithiasis [12]. These studies were all based on the
hypothesized mechanism that heat induced urolithiasis is attributed to water loss from
sweating. The decrease of extracellular fluid causes an increase in serum osmolality which
increased the vasopressin, leading to low urinary volume and concentrated urine.
Concentrated urine contains higher levels of relatively insoluble salts, e.g. calcium oxalate,
and these salts precipitate out of solution when their upper limits of solubility are exceeded,
forming urolithiasis. Sunlight exposure is an alternative explanation for the association
between heat and urolithiasis, as it has been related to hypercalciuria [24]. The association
between heat exposure and urolithiasis has been examined from different aspects: some
observed the association over geographical variation in epidemiological studies [12], some
investigated the association on temporal variation in the same geographic area by season
[103], while the others reported the association with general analyses [83].
Geographical variation in the prevalence of urolithiasis in previous national studies has been
observed. Sourcie et al. [12] mapped out the prevalence of urolithiasis based on nationwide
74
survey data in the United States, and found that the trend of increasing prevalence was from
North to South and from West to East. The authors also found a higher risk in the hottest and
sunniest locations. Similar results were found in Australia [104], Iran [75], and Turkey [105].
Studies have shown that the role of heat in the pathogenesis of urolithiasis has seasonal
variation. Robertson et al. [103] performed one of the first studies to investigate the seasonal
variation on the incidence of urolithiasis in Leeds, Great Britain. They analysed 24-hour urine
samples and found that the values of calcium and oxalate were significantly higher in summer,
and their changes were correlated with heat and hours of sunlight. Studies have also been
performed in many other countries on the seasonal variation of urolithiasis in regards to
temperature difference, including Finland [105], Kuwait [106], Iraq [107], Saudi Arabia
[108], Iran [109], Japan [110], Taiwan [79], the United States [111] and Italy [112], with all
reporting similar results.
Temporal changes over time in the same geographical area were also observed. For example,
Stamatelou et al. [102] performed a study in the United States which compared the
prevalence of urolithiasis between 1988–1994 and 1976–1980. The prevalence of urolithiasis
increased for both males and females when temperature increased. Similar results were also
found in the United Kingdom [113], Italy [43, 97], Japan [104], and Germany [83].
Some studies have examined the association between occupational heat exposure and
urolithiasis. Pierce et al. [16] reported in 1945 that the incidence of urolithiasis among
American troops deployed in desert areas was two-fold higher than those in mountainous
areas. Later, Blacklock et al. examined the British Royal Navy on the incidence of
urolithiasis among different occupations, and found that engineers in the hot engine room had
75
a higher incidence of urolithiasis [17]. Another study was conducted with British troops who
were transported from the United Kingdom to the Persian Gulf, and found that the group
transported in summer had a significantly higher level of urine calcium excretion than those
transported in winter. This had been attributed to longer hours of sunlight exposure and
higher temperatures [24]. A study in Italy and a study in Brazil found that workers exposed to
heat from glass production or furnaces in the steel plant had a much higher risk of urolithiasis
than those who were not exposed to heat [20, 21]. Pin et al. [26] reported that some outdoor
occupations had a higher prevalence of urolithiasis than indoor occupations in Singapore.
The above studies have provided a body of literature supporting the potential contribution of
heat to an increased prevalence of urolithiasis. Although studies were from different countries
demonstrated the possible role of heat in the development of urolithiasis, the quantified
association of heat exposure with urolithiasis in specific groups or occupations remains
unclear. In order to explore the quantified association with different occupations, this study
focused on outdoor workers in a shipbuilding company. The comparisons with previous
studies are indicated in the following section.
5.5.2 General comparisons of study designs and results
Impacts of ambient heat exposures were combined with effects of the sunlight in many
studies [16, 18, 19, 24, 26]. For example, the earliest study on the association between
ambient heat exposure and urolithiasis stated that American troops stationed in a “desert area”
had two-fold higher prevalence of urolithiasis than those in mountainous areas [16]. However,
this study did not provide an exact prevalence or any statistical analyses. It did not clearly
verify the association between the prevalence or incidence of American troops in desert areas
76
and those in mountainous areas, nor the contribution of ambient heat or sunlight. In another
study on British troops, the group deployed from the United Kingdom to the Persian Gulf in
summer had a higher risk of hypercalciuria while those in winter did not; the results were
considered to be associated with longer sunlight exposure and higher temperature [24]. The
investigators in this study examined the association between the increased exposure to
sunlight and hypercalciuria. In a cross-sectional study performed by Better et al. [18] in Israel,
the risk of urolithiasis in lifeguards was ten-fold greater than that in the general population,
and metabolic changes in blood and urine were examined for evidence of urolithiasis
formation; again it was thought the increase should be due to both the intensive sunlight
exposure and heat on an almost completely bare body. Thus, the independent effect of
sunlight and ambient heat had not been thoroughly examined.
Sunlight exposure influences the serum concentration of 25(OH)D, which facilitates the
process of urolithiasis formation by inducing hypercalciuria [24]. Studies in Turkey and
Jordan showed a strong relationship with clothing and serum 25(OH)D, which decreased for
women in western clothing compared to those wearing the traditional hijab or completely
veiled in niqab [111, 114]. Thus skin covered by clothes played a predominant role in
determining sunlight-induced 25(OH)D increase which was associated with urolithiasis
formation. In comparison, the impact from sunlight exposure on workers in this study had
been minimised. As shipbuilding workers in this study had to be in long uniforms and put on
PPE (e.g. safety helmet, mask, gloves and safety shoes) on workdays due to occupational
health and safety purpose, most of their skin was protected from exposure to sunlight. This
minimised the impact from direct sun exposure on urolithiasis through increasing 25(OH)D
level, so this study was able to analyse the independent association of ambient heat exposure
and urolithiasis.
77
The causal relationship between occupational ambient heat exposure and urolithiasis has been
described in previous literature in general terms only. For instance, a cross-sectional study
conducted in Singapore reported higher prevalence of urolithiasis in outdoor workers
(including quarry drilling and crusher workers, quarry truck and loader drivers and postal
deliverymen) than indoor workers based on a questionnaire methodology [26]. The study
only controlled for social class as a confounding factor and used a simple descriptive
statistical method for analysing the differences in prevalence between outdoor workers and
indoor workers. The authors stated that urolithiasis resulted from prolonged physical activity
under extreme tropical heat with low fluid intake. However, they failed to account for several
important confounders. In order to verify the association between occupational ambient heat
exposure and urolithiasis, careful consideration of other risk factors is required.
In this study, a matched 1:4 case-control design was used, and age and sex were used as two
matching criteria. More reliable data from medical health checks were collected instead of
questionnaires. For statistical analyses, a conditional logistic regression model was used for
analysing the association between heat exposure and urolithiasis. The risk factors considered
in this study included length of ambient heat exposure, different types of work with various
ambient heat exposure levels, and health status relevant to heat exposure such as heart
diseases and hypertension. The results of this study were consistent with those of previous
research.
5.5.3 Association between total exposure time and urolithiasis
78
Several studies have investigated the association between length of service and urolithiasis,
but only two studies found a statistically significant difference in length of service between
heat-exposed and non-heat-exposed groups [18, 20, 21, 26]. Pin et al. [26] only provided the
average working time per week for both outdoor workers and indoor employees, and found a
statistical significant association between outdoor workers and urolithiasis. However, they
did not quantify ambient heat exposure time for describing this association. Better et al. [18]
performed a cross-sectional study among lifeguards in Israel, and reported on the average
length of service among cases which was significantly longer than that of the general
population. Their results were limited because average working time and length of service
alone were not able to provide all the relevant heat exposure information solely. In order to
reflect the cumulative amount of ambient heat exposure, this study used total exposure time
(average exposure time multiplied by adjusted length of service) for assessing the ambient
heat exposure for workers and employees in the shipbuilding company. Total exposure time
could be a good indicator because it contains information derived from both average
exposure time and length of service. This study showed that total exposure time was an
important risk factor for urolithiasis formation among subjects (See Table 5).
5.5.4 Association between the type of work and urolithiasis
Previous studies indicated that occupations in hot environments had an increased risk of
urolithiasis [17, 18, 20, 21, 26]. However, they often combined the effects of ambient heat
exposure and sunlight exposure together. By contrast, this study minimised the effect of
sunlight exposure on urolithiasis in the model, as was stated in 5.3.2. Hence, it was possible
for us to analyse the independent effect of ambient heat exposure on urolithiasis.
79
In general, heat stress on workers can be affected by climatic and non-climatic factors which
include workloads, clothing and high temperature operations, etc. [108, 115, 116]. For
outdoor workers in this study, the main climatic factor was ambient heat exposure during hot
months and this was determined by the climate itself. However, non-climatic factors varied
according to different types of work. Threshold WBGT is used to measure the highest
tolerable limit for health and safety of high temperature work.
For example, if workers undertake heavy work, their threshold WBGT should be reduced by
4 °C given the threshold level of 30 °C for light work when male workers normally clothed,
acclimatized, physically fit and in good health [80] (See Table 7). Meanwhile, this threshold
WBGT will be modified according to different conditions. If workers undertake heavy work
wearing impermeable full-length coats, it is recommended that threshold WBGT levels
should be reduced by another 4 °C [80] (See Table 8). In this study, spray painters were in
this situation because they were allocated to the heavy workload category by OHSOs and had
to wear impermeable full-length coats and other PPE including full face mask respirators,
chemical protective gloves and safety shoes. The possible explanations were that spray
painters had to hold high pressure spray guns to spray hulls involving stretching, bending and
twisting of their bodies on unshaded scaffolds [115], which required quite big amount of
calories that met the criterion of heavy work (See Table 7); and spray painters had to wear
impermeable PPE to avoid the exposure of high speed paint pigments and aerosols from
spraying [97]. Thus even under the same ambient temperatures in hot months, ambient heat
stress might impact on spray painters more severely than other types of work (OR=4.4,
95% CI: 1.7–11.4).
80
Table 7. Threshold WBGT levels for different workloads among men normally clothed,
acclimatized, physically fit and in good health
Workload WBGT, °C
Light work (less than 200 kcal/hr) 30
Moderate work (201–300 kcal/hr) 28
Heavy work (301–400 kcal/hr) 26
Very heavy work (above 400 kcal/hr) 25
Source: Yousef, M., et al. [80]
Table 8. Modification of threshold WBGT level by different conditions
Condition Modification of WBGT
Unacclimatized Subtract 2 °C
Unfit Subtract 2 °C
Obese Subtract 1–2 °C
Old age Subtract 1–2 °C
Females Subtract 1 °C
Clothing
shorts or seminude Add 2 °C
impermeable jackets Subtract 2 °C
impermeable full-length coats Subtract 4 °C
impermeable completely enclosed suits Subtract 5 °C
Increased air velocity above 1.5 mps, as long as air
temperature is below 35
Add 2 °C
Source: Yousef, M., et al. [80]
81
Working with exogenous heat generated by equipment or tools is another non-climatic factor.
Atan et al. [21] reported that steel workers exposed to heat from furnaces (ranging from 40–
1500 °C) had higher prevalence of urolithiasis than those not exposed. In this study, smelter
workers who worked beside furnaces also had a higher risk of urolithiasis (OR=4.0, 95% CI:
1.8–9.2). Moreover, the risk of urolithiasis was also high among welders (OR=3.7, 95% CI:
1.9–7.2) who had to face exogenous heat stress of 8.33 °C (15 °F) when welding [93], with
estimated heavy workloads by OHSOs and relevant PPE, e.g. full-length coats, welding mask
and welding gloves (See Table 10).
The occupations of assembler and production security and quality inspector are generally
regarded as to be of lower risk for urolithiasis, and few previous studies had studied these two
groups. However, this study found that there was a mild but statistically significant higher
risk of urolithiasis for assemblers (OR=2.2, 95% CI: 1.1–4.3) and production security and
quality inspectors (OR=2.7, 95% CI: 1.4–3.0), with common PPE and workloads estimated to
be heavy and moderate, respectively (See Table 10). These higher risks of urolithiasis
suggested that ambient heat exposure in these working groups could currently be
underestimated. These workers should take care to consume adequate water and take regular
water breaks for health protection purposes.
Two other types of work, planing machine operator and gas-cutting worker, showed no
increased risk of urolithiasis (See Table 5). The main task for planing machine operators and
gas-cutting workers is to groove metal for welding and to perform metal cutting with an
oxyacetylene cutting torch, respectively. The results were reasonable because their workloads
were regarded as light; and only 5 hours of their work time (as estimated by OHSOs) were
occupied by physical and strenuous work, which would also increase the threshold WBGT
82
limits (See Table 9). Although the work environment for planing machine operators and gas-
cutting workers was shaded in concrete work sheds, these workers were still exposed to
ambient heat as no air-conditioners were installed. However, their work environment is
regarded as much more comfortable than many other types of work (e.g. spray painters,
welders and assemblers) operating outdoors or inside closed steel cabins heated by the sun on
hot days.
Table 9. Threshold WBGT limits
Work-rest regimen
Workload
Light Moderate Heavy
Continuous work 30.0 26.7 25.0
75% work+25% rest; each hour 30.6 28.0 25.9
50% work+50% rest; each hour 31.4 29.4 27.9
25% work+75% rest; each hour 32.2 31.1 30.0
Source: ACGIH (1996) [81]
Literature also suggested some potential factors that might pose the risk of urolithiasis to
certain types of workers besides ambient heat. For welders, cadmium exposure may be a
possible reason for an increasing risk of urolithiasis. Earlier studies reported that workers
exposed to cadmium had a higher prevalence of urolithiasis [117, 118]. Trevisan et al. [119]
published a clinical case report of urolithiasis on a welder who was exposed to cadmium at
work. A recent study performed by Ding et al. [120] in China reported that 6 out of 103
railway welders had urine cadmium levels exceed the Chinese reference value and 17% of
samples exceed the threshold limit value of the concentration of airborne cadmium. However,
cadmium had not been detected in the air samples of welding fumes at workplace and
shipbuilding material including the hull steel and welding rods (wires) in the 2010 technical
83
report of the company carried out by an certified organization providing service of
construction project occupational hazard evaluations (Guangdong Province Hospital for
Occupational Disease Prevention and Treatment, unpublished data). Thus cadmium exposure
should not be the main reason for higher risk of urolithiasis among welders in this study.
For spray painters, paint and solvent exposure may be a possible causal factor for urolithiasis
formation. Laerum et al. [121] reported that railroad shopmen exposed to oxalic acid which
was used in a repainting and cleaning process for railroad cars was associated with
urolithiasis. Vitayavirasuk et al. [122] reported that spray painters working in automobile
body repair shops had significantly higher urine cadmium level than the matched general
population because cadmium was components of paint pigments. Increased cadmium level
was a risk factor for urolithiasis [94, 118]. However, Cadmium-containing paints are not
widely used any more in many countries including China. Meanwhile, spray painters in this
study wore full face mask respirators, impermeable full-length coats, chemical protective
gloves and safety shoes that should have been able to protect them from severe exposure to
paint and relevant solvents during the ship-repair process, which was much better than the
situations that railroad shopmen exposed to oxalic acid without efficient PPE and spray
painters with aerosol-removing respirators reported by Laerum et al. [121] and Vitayavirasuk
et al. [122]. Thus, paint and solvent exposure should not be the main reason for high
prevalence of urolithiasis among spray painters.
In this study, five types of work (i.e. spray painter, welders, assemblers and smelter,
production security and quality inspector) had a higher risk of urolithiasis than indoor
employees. These are the types of work with long impermeable suits and fully-covered PPE,
moderate to heavy workloads under ambient heat exposure, and/or working with extra heat
84
from equipment or tools (See Table 10). The results of this study indicate that these types of
work may be more susceptible to urolithiasis, and require more attention in regards to
preventative care.
Table 10. Non-climate factors impacting on heat stress of different types of work and
relevant odds ratios of urolithiasis
Type of work OR
Non-climate factors
Clothing Workloads Heat source
Spray painter 4.4
Impermeable full-length coats, full face mask
respirator, chemical protective gloves , and
safety shoes
Heavy –
Welder 3.7
Full-length coats, welding mask, welding
gloves, safety helmet, and safety shoes
Heavy
Welding
machine
Assembler 2.2
Full-length coats, gloves, safety helmet, and
safety shoes
Heavy –
Production
security and
quality inspector
2.7
Full-length coats, safety helmet, and safety
shoes
Moderate –
Smelter 4.0
Impermeable full-length coats, smelter helmet,
smelter gloves, and safety shoes
Heavy Furnace
Planing machine
operator
– Full-length coats, safety helmet, and safety
shoes
Light –
Gas-cutting
workers
– Full-length coats, safety helmet, and safety
shoes
Light
Gas-cutting
machine
85
5.5.5 Association between hypertension and urolithiasis
Evidence has shown that workers exposed to prolonged heat are at high risk of hypertension
[123, 124]. People with severe hypertension are not allowed to be recruited into heat exposed
occupations in many countries, including China. Those hypertensive subjects in this study
were more likely to develop hypertension after recruitment.
An eight-year follow-up cohort study on workers from the Olivetti factory performed by
Cappuccio et al. [125] reported that hypertension was a predictor for urolithiasis. This was
consistent with the finding in this study. It was demonstrated that people with hypertension
caused calcium homeostasis alteration which induced urolithiasis formation [51, 125-128].
This was the possible mechanism for explaining the result found in this study that workers
with hypertension were more likely to have urolithiasis. Outdoor heat exposure might
increase the risk of hypertension and urolithiasis, and hypertension could also increase the
risk of urolithiasis.
The results of this study indicate that outdoor workers had higher risk of urolithiasis than
indoor workers and so did workers with longer total exposure time than those with shorter
total heat exposure time. More attention should be paid to minimise the impact of ambient
heat exposure on urolithiasis, and to develop preventative strategies to protect outdoor
workers from the potential effects of climate change.
86
5.6 Strengths and limitations
This study has five major strengths. Firstly, to our knowledge, it was the first case-control
study investigating the association between ambient heat exposure and urolithiasis between
shipbuilding workers and employees. Secondly, the datasets used in this study are quite
comprehensive with less than 5% missing values of health surveillance data. Thirdly, the data
were quite reliable as they were collected from the affiliated hospital of the company by GIHI
and derived from regular health surveillance. Fourthly, the study was able to examine the
independent effect of ambient heat exposure on urolithiasis after the confounding effects of
sunlight were minimised. Finally, this study controlled for most well-known confounding
factors, including age, sex, length of service (longer than three years), diet (at least two meals
per day were provided by the company), birth places (all of them were from Guangdong
Province) and heat exposure-related health status (hypertension and heart disease).
There are also several limitations in this study. Firstly, this study did not have information on
family history [40], obesity [129], social class [51], metabolic diseases [68], smoking and
drinking behaviours [94, 95, 130], which have been related to urolithiasis according to the
literature. Secondly, records for water drinking behaviour [65] of workers and employees
were not available because this study was mainly based on secondary health surveillance data.
Thirdly, the control selection process was not completely random, because health
surveillance files containing more information were prioritised, particular when matching by
same sex and age. However, after strictly matched 4 controls for 1 case by sex and age, there
were not a lot of choices for each set especially for male controls over 45 years old and
female controls of all age. Thus the control selection process should not significantly
influence the results. Fourthly, this study did not deal with the potential false positive results
87
generated from the ultrasound specificity of 74% for urolithiasis screening because of the
time limit for a master‟s thesis. However, subjects with abnormal ultrasonography results in
the health checks would be required to re-examine by soma scopes for ensuring the validity
of the results and the re-examine results would also be attached to the health check files (as
mentioned by OHSOs). Not all cases were noticed that they were with the re-examine results
attached, but false positive results were rare according to the attached re-examine results in
the collected health surveillance data. So the potential positive results were considered that
they should not consist of a big percentage that generated important influence. Fifthly, this
study was lacking of actual heat exposure measurements due to mainly being based on
historical secondary data. However, according to literature, there were mainly two kinds of
occupational hazard exposure associated with kidney stones, which included heat exposure
and cadmium exposure. Potential cadmium exposure should not be the key risk factor for
outdoor workers in this study, and this has been discussed in Section 5.5.4. Ambient heat can
be interpreted to be the main key risk factor for outdoor workers. In addition, the movement
of workers between different types of work was not particularly controlled but it was really
rare for technical workers moving frequently from one certified type of work to other types as
shown in the data collected in this study, and in actual practice. The interpretation of the
findings in this study still should be treated with caution and further studies are necessary to
confirm our findings. A summary of the strengths and limitations of this study can be found
in Table 11.
88
Table 11. Summary of strengths and limitations of this study
Order Strengths Limitations
1
The first case-control study on the association
between occupational heat exposure and
urolithiasis
There was no information on family history,
obesity, social class, metabolic diseases, smoking
and drinking behaviours
2 Missing values were less than 5%
Did not deal with the potential false positive
results due to the time limit
3 Based on the actual health check data
The control selection process was not completely
random
4
Examined the independent health impact from
heat
There was no information on water drinking
behaviours
5 Controlled for most of the confounding factors
There was lacking of actual heat exposure
measurements
6
Did not control movement between different
types of work
5.7 Directions for future research
This study used a 1:4 matched case-control study to examine the association between ambient
heat exposure and urolithiasis for 950 workers and employees in a big shipbuilding company
in China. Results showed that ambient heat exposure increased the risk of urolithiasis. In the
context of climate change, global warming will probably exacerbate the impacts of ambient
heat on urolithiasis. Thus, it is necessary to conduct further research in order to improve the
understanding of this relationship.
Firstly, improved data collection should be carried out. In this study, some potential
confounders such as family history of urolithiasis, social class and urine volume of workers
89
and indoor employees were not available. It has been shown that an established personal or
family history of urolithiasis can increase the risk of urolithiasis by 50% [131]. Also, a low
urine volume is an important risk factor for urolithiasis in temperate climates [132]. Future
researchers can use first-hand data collection approaches for additional details of individual
information (e.g. family history, social class and metabolic diseases) through interviews or
questionnaires undertaken of workers.
Secondly, some risk factors could be better measured and observed. WBGT is a commonly
used index for the evaluation of a heat stress at threshold limits [81], and it should be set in
parallel to health check years for representing ambient temperature. The WBGT is a weighted
average of the three sensors on an area heat stress monitor: natural wet bulb temperatures
(WB), globe temperature (GT), and air temperature (dry bulb temperature, DBT). The
formula for outdoor measurements is WBGT=0.7WB + 0.2GT + 0.1DBT. This study was
based on secondary data collection, and it collected records of temperature from the official
meteorological bureau and records of health outcomes from the studied company. In future
research, ambient temperature for each workplace could also be measured by WBGT.
Similarly, important individual characteristics such as water consumption behaviour could
also be observed and recorded. Exposure time could be recorded by researchers‟ observation
or through collection of modified attendance sheets with self-reported ambient heat exposure.
Moreover, a stronger causal inference study design (e.g. a cohort study) could be employed to
evaluate the association between ambient heat exposure and urolithiasis. In this study, the
average exposure time (the main exposure index) was estimated by OHSOs, but whether the
OHSOs‟ estimations were precise enough to estimate the actual exposure still needs to be
confirmed in future studies. In a cohort study, average exposure time can be directly recorded,
90
allowing health outcomes to be prospectively measured. A cost-benefit analysis of
preventative interventions on high risk groups could also be performed.
5.8 Conclusions
This study examined the relationship between ambient heat exposure and urolithiasis among
shipbuilding workers and indoor employees in Guangzhou, China. A conditional logistic
regression model was implemented in describing the odds ratios of various risk factors (e.g.
different types of work, total exposure time and hypertension). The study found that
shipbuilding workers with longer total exposure time, hypertension, and of certain types of
work including spray painter, smelter, welder, production security and quality inspector and
assembler were at higher risk of urolithiasis. Four of the above five occupations worked
mainly outdoors except for smelters with furnaces indoors. There were several possible
reasons for why these types of work may be impacted by ambient heat stress more than others.
These included heavier workloads, fully-covered PPE (for spray painters and welders), and
working with exogenous source of heat (for welders).
Findings in this study may have important public health implications in the control and
prevention of urolithiasis associated with ambient heat exposure among outdoor working
populations. These findings provide information for decision-makers to prioritise potential
industries similar to shipbuilding affected by ambient heat for public health interventions, and
support practical recommendations that could be incorporated into relevant industrial
regulations.
91
REFERENCES
1. Schulte, P. and H. Chun, Climate change and occupational safety and health:
establishing a preliminary framework. Environmental and Industrial Hygiene, 2009.
6(9): p. 542-554.
2. Fakheri, R.J. and D.S. Goldfarb, Ambient temperature as a contributor to kidney stone
formation: implications of global warming. Kidney International, 2011: p. 1178-1185.
3. Romero, V., H. Akpinar, and D.G. Assimos, Kidney stones: a global picture of
prevalence, incidence, and associated risk factors. Reviews in Urology, 2010. 12(2-3):
p. e86-e96.
4. Adisesh, A., E. Robinson, and A.D. Curran, Climate change: enabling a better
working Britain for the next 100 years. Occupational Medicine, 2011. 61(5): p. 292-
294.
5. IPCC, Climate Change 2007: The physical science basis: contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. 2007: Cambridge University Press.
6. IPCC, Climate change 2007: Synthesis report. Contribution of Working Groups I, II
and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. 2007, Geneva, Switzerland.
7. IPCC. Climate Change 2007: Impacts, Adaptaion and Vulnerability. Contribution of
Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on
92
Climate Change. 2008 [cited 2008 September]; Available from:
http://www.ipcc.ch/ipccreports/ar4-wg3.htm.
8. Trinchieri, A., Epidemiological trends in urolithiasis: impact on our health care
systems. Urological Research, 2006. 34(2): p. 151-156.
9. Amato, M., M. Lusini, and F. Nelli, Epidemiology of nephrolithiasis today. Urologia
Internationalis, 2004. 72(1): p. 1-5.
10. Serio, A. and A. Fraioli, Epidemiology of nephrolithiasis. Nephron, 1999. 81(1): p.
26-30.
11. Yanagawa, M., et al., Incidence of urolithiasis in northeast Thailand. International
Journal of Urology, 1997. 4(6): p. 537-540.
12. Thun, M.J. and S. Schober, Urolithiasis in Tennessee: an occupational window into a
regional problem. American Journal of Public Health, 1991. 81(5): p. 587-591.
13. Brikowski, T., Y. Lotan, and M. Pearle, Climate-related increase in the prevalence of
urolithiasis in the United States. Proceedings of the National Academy of Sciences,
2008. 105(28): p. 9841-9846.
14. Kjellstrom, T., et al., The direct impact of climate change on regional labor
productivity. Archives of Environmental & Occupational Health, 2009. 64(4): p. 217-
227.
15. Kjellstrom, T., Climate change, direct heat exposure, health and well-being in low
and middle-income countries. Global Health Action, 2009. 2.
93
16. Pierce, L. and B. Bloom, Observations on urolithiasis among American troops in a
desert area. The Journal of Urology, 1945. 54: p. 466-470.
17. Blacklock, N.J., The pattern of urolithiasis in the Royal Navy. Journal of the Royal
Naval Medical Service, 1965. 51(2): p. 99-111.
18. Better, O., et al., Increased incidence of nephrolithiasis in lifeguards in Israel.
Advances in Experimental Medicine and Biology 1980. 128: p. 467-472.
19. Milvy, P., E. Colt, and J. Thornton, A high incidence of urolithiasis in male marathon
runners. The Journal of Sports Medicine and Physical Fitness, 1981. 21(3): p. 295-
298.
20. Borghi, L., et al., Hot occupation and nephrolithiasis. The Journal of Urology, 1993.
150(6): p. 1757-1760.
21. Atan, L., et al., High kidney stone risk in men working in steel industry at hot
temperatures. Urology, 2005. 65(5): p. 858-861.
22. Kjellstrom, T., I. Holmer, and B. Lemke, Workplace heat stress, health and
productivity–an increasing challenge for low and middle-income countries during
climate change, in Global Health Action. 2009, Co-Action Publishing.
23. Qian, W. and Y. Zhu, Climate change in China from 1880 to 1998 and its impact on
the environmental condition. Climatic Change, 2001. 50(4): p. 419-444.
24. Parry, E.S. and I.S. Lister, Sunlight and hypercalciuria. Lancet, 1975. 1(7915): p.
1063-1065.
94
25. Ferrie, B. and R. Scott, Occupation and urinary tract stone disease. Urology, 1984.
24(5): p. 443-445.
26. Pin, N.T., N.Y. Ling, and L.H. Siang, Dehydration from outdoor work and urinary
stones in a tropical environment. Occupational Medicine, 1992. 42(1): p. 30-32.
27. Iguchi, M., et al., Prevalence of urolithiasis in Kaizuka City, Japan - An
epidemiologic study of urinary stones. International Journal of Urology, 1996. 3(3): p.
175-179.
28. Stern, N.H., The economics of climate change: the Stern review. 2007: Cambridge
University Press.
29. Costello, A., et al., Managing the health effects of climate change: Lancet and
University College London Institute for Global Health Commission. Lancet, 2009.
373(9676): p. 1693-1733.
30. Patz, J., et al., Impact of regional climate change on human health. Nature, 2005.
438(7066): p. 310-317.
31. McMichael, A.J., R.E. Woodruff, and S. Hales, Climate change and human health:
present and future risks. Lancet, 2006. 367(9513): p. 859-869.
32. Haines, A., et al., Climate change and human health: impacts, vulnerability, and
mitigation. Lancet, 2006. 367(9528): p. 2101-2109.
33. Stuart, R.O., 2nd, et al., Seasonal variations in urinary risk factors among patients
with nephrolithiasis. The Journal of Lithotripsy & Stone Disease, 1991. 3(1): p. 18-27.
95
34. Shakhssalim, N., et al., An assessment of parathyroid hormone, calcitonin, 1,25
(OH)2 vitamin D3, estradiol and testosterone in men with active calcium stone
disease and evaluation of its biochemical risk factors. Urological Research, 2011.
39(1): p. 1-7.
35. Ding, C., et al., Serum levels of vitamin D, sunlight exposure, and knee cartilage loss
in older adults: the Tasmanian older adult cohort study. Arthritis & Rheumatism,
2009. 60(5): p. 1381-1389.
36. Coe, F.L., J.H. Parks, and J.R. Asplin, The pathogenesis and treatment of kidney
stones. New England Journal of Medicine, 1992. 327(16): p. 1141-1152.
37. Giannini, S., et al., Possible link between vitamin D and hyperoxaluria in patients
with renal stone disease. Clinical Science, 1993. 84(1): p. 51-54.
38. Jarrar, K., et al., Relationship between 1, 25-dihydroxyvitamin-D, calcium and uric
acid in urinary stone formers. Urologia Internationalis, 1996. 56(1): p. 16-20.
39. Robertson, W., Pathophysiology of stone formation. Urologia Internationalis, 1986.
41(5): p. 329-333.
40. Ljunghall, S., et al., Family history of renal stones in recurrent stone patients. British
Journal of Urology, 1985. 57(4): p. 370-374.
41. Tucak, A., et al., The incidence and risk factors of urolithiasis in active working
population of the Osijek community: an epidemiological study. Periodicum
Biologorum, 2000. 102(4): p. 431-436.
42. Asper, R., Epidemiology and socioeconomic aspects of urolithiasis. Urological
Research, 1984. 12(1): p. 1-5.
96
43. Hesse, A., et al., Study on the prevalence and incidence of urolithiasis in Germany
comparing the years 1979 vs. 2000. European Urology, 2003. 44(6): p. 709-713.
44. Sellaturay, S. and C. Fry, The metabolic basis for urolithiasis. Surgery (Oxford), 2008.
26(4): p. 136-140.
45. Park, S., Medical management of urinary stone disease. Expert Opinion on
Pharmacotherapy, 2007. 8(8): p. 1117-1125.
46. National Institute for Occupational Safety and Health (NIOSH), Working in hot
environments. 1986, NIOSH: Cincinnati, Ohio. p. 86-112.
47. Breslau, N.A., et al., Relationship of animal protein-rich diet to kidney-stone
formation and calcium-metabolism. Journal of Clinical Endocrinology & Metabolism,
1988. 66(1): p. 140-146.
48. Assimos, D.G. and R.P. Holmes, Role of diet in the therapy of urolithiasis. Urologic
Clinics of North America, 2000. 27(2): p. 255-268.
49. Trinchieri, A., Epidemiology of urolithiasis. Arch Ital Urol Androl, 1996. 68(4): p.
203-249.
50. Curtin, J. and M. Sampson, Greenhouse effect and renal calculi. Lancet, 1989.
2(8671): p. 1110.
51. Robertson, W.G., M. Peacock, and M. Baker, Studies on the prevalence and
epidemiology of urinary stone disease in men in Leeds. British Journal of Urology,
1983. 55(6): p. 595-598.
97
52. La Vecchia, C., et al., Education, prevalence of disease, and frequency of health care
utilisation. The 1983 Italian National Health Survey. Journal of Epidemiology and
Community Health, 1987. 41(2): p. 161-165.
53. Gonzalez, P., A world vulnerable to climate change. WorldView, 2008. 21(2): p. 6-34.
54. Haines, A. and J.A. Patz, Health effects of climate change. JAMA: the journal of the
American Medical Association, 2004. 291(1): p. 99-103.
55. Ury, H.K., Efficiency of case-control studies with multiple controls per case:
continuous or dichotomous data. Biometrics, 1975: p. 643-649.
56. Miettinen, O.S., Individual matching with multiple controls in the case of all-or-none
responses. Biometrics, 1969: p. 339-355.
57. Dupont, W.D., Power calculations for matched case-control studies. Biometrics,
1988. 44(4): p. 1157-1168.
58. Peng, J., et al., A study on urolithiasis prevalence characteristics of resident
population in Shenzhen Special Zone. Journal of Mathematical Medicine, 2000.
13(006): p. 516-517.
59. Mi, H. and Y. Deng, Epidemiology Characteristics of Urolithiasis in China. Chinese
Journal of Urology, 2003. 24(010): p. 715-716.
60. Xu, S., et al., A survey report of urolithiasis prevalence rate in Shenzhen City.
Chinese Journal of Urology, 1999. 20(011): p. 655-657.
98
61. Statistics Bureau of Guangdong Province, Communiqué of Statistics Bureau of
Guangdong Province on Major Figures of the 2010 Population Census, Statistics
Bureau of Guangdong Province, Editor. 2011.
62. Qingguo, Z., et al., Effects of air pollution on neonatal prematurity in guangzhou of
china: a time-series study. Environmental Health, 2011. 10(2).
63. Guangdong Meteorological Bureau. 2010 Guangzhou City climate report. 2010;
Available from: http://www.grmc.gov.cn/article/2010/12/30/article_5518.html.
64. The National People's Congress Standing Committee, Law of the People's Republic of
China on the Prevention and Control of Occupational Diseases. 2002.
65. Ozsvath, D.L., Fluoride and environmental health: A review. Reviews in
Environmental Science and Biotechnology, 2009. 8(1): p. 59-79.
66. Standing Committee of the National Occupaitonal Diseases Diagnosis, Guideline of
Occupational Health Surveillance. 2007: People's Republic of China.
67. Yoshida, O. and Y. Okada, Epidemiology of urolithiasis in Japan: a chronological
and geographical study. Urologia Internationalis, 1990. 45(2): p. 104-111.
68. Johri, N., et al., An update and practical guide to renal stone management. Nephron
Clinical Practice, 2010. 116(3): p. c159-c171.
69. Trinchieri, A., Epidemiology of urolithiasis: an update. Clinical Cases in Mineral and
Bone Metabolism, 2008. 5(2): p. 101-106.
70. Robertson, W., M. Peacock, and B. Nordin, Inhibitors of the growth and aggregation
of calcium oxalate crystals in vitro. Clinica Chimica Acta, 1973. 43(1): p. 31-37.
99
71. Robertson, W.G., et al., Saturation-inhibition index as a measure of the risk of
calcium oxalate stone formation in the urinary tract. New England Journal of
Medicine, 1976. 294(5): p. 249-252.
72. Hosmer, D.W. and S. Lemeshow, Applied logistic regression. Vol. 354. 2000: Wiley-
Interscience.
73. Tack, I., Effects of water consumption on kidney function and excretion. Nutrition
Today, 2010. 45(6 SUPPL.): p. S37-S40.
74. Ramello, A., C. Vitale, and M. Marangella, Epidemiology of nephrolithiasis. Journal
of Nephrology, 2001. 13: p. e86-e96.
75. Safarinejad, M.R., Adult urolithiasis in a population-based study in Iran: prevalence,
incidence, and associated risk factors. Urological Research, 2007. 35(2): p. 73-82.
76. Torres, C., et al., Decreased kidney function of unknown cause in Nicaragua: a
community-based survey. American Journal of Kidney Diseases, 2010. 55(3): p. 485-
496.
77. Yasui, T., et al., Prevalence and epidemiological characteristics of urolithiasis in
Japan: national trends between 1965 and 2005. Urology, 2008. 71(2): p. 209-213.
78. Attar, K., et al., The secret of the phantom stone: A case report. International Urology
and Nephrology, 2004. 36(1): p. 27-28.
79. Chen, Y.-K., et al., Seasonal Variations in Urinary Calculi Attacks and Their
Association With Climate: a Population Based Study. The Journal of Urology, 2008.
179(2): p. 564-569.
100
80. Yousef, M., S. Sagawa, and K. Shiraki, Heat stress: a threat to health and safety.
Journal of UOEH, 1986. 8(3): p. 355-364.
81. Parsons, K., Heat stress standard ISO 7243 and its global application. Industrial
Health, 2006. 44(3): p. 368-379.
82. Attia, M. and P. Engel, A field study of thermal stress and recovery using
thermoregulatory behavioral and physiological indicators. International Archives of
Occupational and Environmental Health, 1980. 47(1): p. 21-33.
83. Elomaa, I., et al., Seasonal variation of urinary calcium and oxalate excretion, serum
25 (OH) D3 and albumin level in relation to renal stone formation. Scandinavian
Journal of Urology and Nephrology, 1982. 16(2): p. 155-161.
84. Frank, M. and A. De Vries, Prevention of urolithiasis. Education to adequate fluid
intake in a new town situated in the Judean Desert Mountains. Archives of
Environmental Health, 1966. 13(5): p. 625-630.
85. Pak, C.Y.C., et al., Evidence justifying a high fluid intake in treatment of
nephrolithiasis. Annals of Internal Medicine, 1980. 93(1): p. 36-39.
86. Embon, O.M., G.A. Rose, and T. Rosenbaum, Chronic dehydration stone disease.
British Journal of Urology, 1990. 66(4): p. 357-362.
87. Brake, D. and G. Bates, Fluid losses and hydration status of industrial workers under
thermal stress working extended shifts. Occupational and Environmental Medicine,
2003. 60(2): p. 90-96.
101
88. Miller, V. and G. Bates, Hydration of outdoor workers in north-west Australia.
Journal of Occupational Health and Safety Autralia and New Zealand, 2007. 23(1): p.
79-88.
89. Dilworth, J.P. and J.W. Segura, Compliance with Use of the Hydrate 1(tm) System by
Patients with Treated Urolithiasis. Journal of Endourology, 1993. 7(3): p. 197-199.
90. Wakefield, B., et al., Monitoring hydration status in elderly veterans. Western Journal
of Nursing Research, 2002. 24(2): p. 132-142.
91. Kerslake, D., The stress of hot environments. 1972: Cambridge University Press.
92. Strazzullo, P. and M. Mancini, Hypertension, calcium metabolism, and
nephrolithiasis. The American Journal of the Medical Sciences, 1994. 307: p. S102-
S106.
93. Sowers, M.F.R., et al., Prevalence of renal stones in a population-based study with
dietary calcium, oxalate, and medication exposures. American Journal of
Epidemiology, 1998. 147(10): p. 914-920.
94. Crowe, J., B.W. de Joode, and C. Wesseling, A pilot field evaluation on heat stress in
sugarcane workers in Costa Rica: What to do next? Global Health Action, 2009. 2.
95. Lall, S. and J. Weiss, Industrial Competitiveness: the challenge for Pakistan. Asian
Development Bank Institute-Pakistan Resident Mission Research Paper, Islamabad,
2004.
96. Malchaire, J., et al., Criteria for estimating acceptable exposure times in hot working
environments: a review. International Archives of Occupational and Environmental
Health, 2000. 73(4): p. 215-220.
102
97. Singh, P. and R. Kiran, Are we overstressing water quality in urinary stone disease?
International Urology and Nephrology, 1993. 25(1): p. 29-36.
98. Caudarella, R., et al., Comparative study of the influence of 3 types of mineral water
in patients with idiopathic calcium lithiasis. The Journal of Urology, 1998. 159(3): p.
658-663.
99. Coen, G., et al., Urinary composition and lithogenic risk in normal subjects following
oligomineral versus bicarbonate-alkaline high calcium mineral water intake.
Urologia Internationalis, 2001. 67(1): p. 49-53.
100. Curhan, G.C., et al., A Prospective Study of the Intake of Vitamins C and B6, and the
Risk of Kidney Stones in Men. The Journal of Urology, 1996. 155(6): p. 1847-1851.
101. Goldfarb, M., S., Diet and nephrolithiasis. Annual Review of Medicine, 1994. 45(1):
p. 235-243.
102. Stamatelou, K., et al., Time trends in reported prevalence of kidney stones in the
United States: 1976–1994. Kidney International, 2003. 63(5): p. 1817-1823.
103. Robertson, W., et al., Seasonal variations in the composition of urine in relation to
calcium stone-formation. Clinical Science and Molecular Medicine, 1975. 49(6): p.
597-602.
104. Baker, P.W., et al., Influence of season, age, and sex on renal stone formation in
South Australia. The Medical Journal of Australia, 1993. 159(6): p. 390-392.
105. Akinci, M., T. Esen, and S. Tellaloğlu, Urinary stone disease in Turkey: an updated
epidemiological study. European Urology, 1991. 20(3): p. 200-203.
103
106. Salem, S. and E.L.Z. Abu, The incidence of renal colic and calculi in Kuwait. An
epidemiological study. Le Journal Médical Libanais, 1969. 22(6): p. 747-755.
107. AL-Dabbagh, T. and K. Fahadi, Seasonal variations in the incidence of ureteric colic.
British Journal of Urology, 1977. 49(4): p. 269-275.
108. Al-Hadramy, M., Seasonal variations of urinary stone colic in Arabia. The Journal of
the Pakistan Medical Association, 1997. 47(11): p. 281-285.
109. Basiri, A., et al., Monthly variations of urinary stone colic in Iran and its relationship
to the fasting month of Ramadan. Journal Pakistan Medical Association, 2004. 54(1):
p. 6-7.
110. Fujita, K., Weather and the incidence of urinary stone colic in Tokyo. International
Journal of Biometeorology, 1987. 31(2): p. 141-146.
111. Chauhan, V., et al., Effect of season, age, and gender on renal colic incidence. The
American Journal of Emergency Medicine, 2004. 22(7): p. 560-563.
112. Chen, Y.K., et al., Seasonal variations in urinary calculi attacks and their association
with climate: a population based study. The Journal of Urology, 2008. 179(2): p. 564-
569.
113. Robertson, W., et al., Epidemiological risk factors in calcium stone disease.
Scandinavian journal of urology and nephrology. Supplementum, 1980. 53: p. 15-30.
114. Boscolo-Berto, R., et al., Do weather conditions influence the onset of renal colic? A
novel approach to analysis. Urologia Internationalis, 2008. 80(1): p. 19-25.
104
115. Wisniewsky, Z., B. Armstrong, and J.G. Brockis, The pattern of urinary calculus in
Western Australia. Urinary Calculus, 1981: p. 47-55.
116. Stamatelou, K.K., et al., Time trends in reported prevalence of kidney stones in the
United States: 1976–19941. Kidney International, 2003. 63(5): p. 1817-1823.
117. Järup, L. and C.G. Elinder, Incidence of renal stones among cadmium exposed battery
workers. Br J Ind Med, 1993. 50(7): p. 598-602.
118. Swaddiwudhipong, W., et al., An association between urinary cadmium and urinary
stone disease in persons living in cadmium-contaminated villages in northwestern
Thailand: A population study. Environmental Research, 2011. 111(4): p. 579-583.
119. Trevisan, A. and C. Gardin, Nephrolithiasis in a worker with cadmium exposure in
the past. International Archives of Occupational and Environmental Health, 2005.
78(8): p. 670-672.
120. Ding, X., et al., Cadmium-induced renal tubular dysfunction in a group of welders.
Occupational Medicine, 2011. 61(4): p. 277-279.
121. Laerum, E. and S. Aarseth, Urolithiasis in railroad shopmen in relation to oxalic acid
exposure at work. Scandinavian Journal of Work, Environment & Health, 1985. 11(2):
p. 97-100.
122. Vitayavirasuk, B., S. Junhom, and P. Tantisaeranee, Exposure to lead, cadmium and
chromium among spray painters in automobile body repair shops. Journal of
occupational health, 2005. 47(6): p. 518-522.
123. Kloetzel, K., et al., Relationship between hypertension and prolonged exposure to
heat. Journal of Occupational and Environmental Medicine, 1973. 15(11): p. 878-880.
105
124. Niimi, Y., et al., Effect of heat stress on muscle sympathetic nerve activity in humans.
Journal of the Autonomic Nervous System, 1997. 63(1-2): p. 61-67.
125. Cappuccio, F.P., et al., A prospective study of hypertension and the incidence of
kidney stones in men. Journal of Hypertension, 1999. 17(7): p. 1017-1022.
126. Olapade-Olaopa, E., et al., Chronic dehydration and symptomatic upper urinary tract
stones in young adults in Ibadan, Nigeria. West African Journal of Medicine, 2004.
23(2): p. 146-150.
127. Gallagher, J., et al., Intestinal calcium absorption and serum vitamin D metabolites in
normal subjects and osteoporotic patients: effect of age and dietary calcium. Journal
of Clinical Investigation, 1979. 64(3): p. 729-736.
128. Lips, P., Vitamin D deficiency and secondary hyperparathyroidism in the elderly:
consequences for bone loss and fractures and therapeutic implications. Endocrine
Reviews, 2001. 22(4): p. 477-501.
129. Scales, C.D., et al., Changing gender prevalence of stone disease. The Journal of
Urology, 2007. 177(3): p. 979-982.
130. Malchaire, J., et al., Criteria for estimating acceptable exposure times in hot working
environments: a review. International Archives of Occupational and Environmental
Health, 2000. 73(4): p. 215-220.
131. Morton, A.R., E.A. Iliescu, and J.W.L. Wilson, Nephrology: 1. Investigation and
treatment of recurrent kidney stones. Canadian Medical Association Journal, 2002.
166(2): p. 213-218.
106
132. Fleisch, H., Inhibitors and promoters of stone formation. Kidney International, 1978.
13(5): p. 361-371.
107
APPENDIX 1: Tables of Results for Statistical Analyses
The multivariable modelling process with conditional logistic regression model
The multivariable modelling process (excluding department)
Step 1. Bivariate analysis (P<0.25*)
1. Type of work
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
556.7 51.8 7 0.00 54.9 7 0.00 54.9 7 0.00
a. Beginning Block Number 1. Method = Enter
Variables in the Equation
Covariate P OR 95%CI
Type of work 0.00*
Welder 0.00* 6.1 (3.4, 10.9)
Assembler 0.00* 3.7 (2.0, 6.6)
Production security and
quality inspector
0.00* 3.5 (1.9, 6.4)
Smelter 0.00* 6.4 (2.9, 13.9)
Planing machine
operator
0.00* 7.2 (1.9, 27.4)
Spray painter 0.00* 6.9 (2.7, 17.7)
Gas-cutting worker 0.02* 4.3 (1.3, 14.6)
108
2. Total exposure time
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
567.0 46.0 1 0.00* 44.6 1 0.00 44.6 1 0.00
a. Beginning Block Number 1. Method = Enter
Variables in the Equation
Covariate P OR 95.0% CI
Total exposure time (yr) 0.00* 1.7 (1.5, 2.1)
3. Electrocardiogram
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
611.4 0.2 1 0.68 0.2 1 0.68 0.2 1 0.68
a. Beginning Block Number 1. Method = Enter
Variables in the Equation
Covariate P OR 95.0% CI
Electrocardiogram 0.68 1.1 (0.7, 1.6)
109
4. Blood pressure
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
608.6 3.1 1 0.08 3.0 1 0.08 3.0 1 0.08
a. Beginning Block Number 1. Method = Enter
Variables in the Equation
Covariate P OR 95.0% CI
Blood Pressure 0.08* 1.5 (0.9, 2.2)
Step 2.An initial multivariable model(P<0.05*)
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
538.9 71.0 9 0.00* 72.7 9 0.00 72.7 9 0.00
a. Beginning Block Number 1. Method = Enter
110
Variables in the Equation
Covariate P OR 95.0% CI
Type Of Work 0.00*
Welder 0.00* 3.7 (1.9, 7.2)
Assembler 0.03 2.2 (1.1, 4.3)
Production security and quality
inspector
0.00* 2.7 (1.4, 5.0)
Smelter 0.00* 4.0 (1.8, 9.2)
Planing machine operator 0.06 3.9 (0.9, 16.6)
Spray painter 0.00* 4.4 (1.7, 11.4)
Gas-cutting worker 0.15 2.6 (0.7, 9.1)
Total exposure time (yr) 0.00* 1.5 (1.2, 1.8)
BP 0.04* 1.6 (1.0, 2.5)
Step 3. Reintroduce variables excluded from the initial multivariable model
There is no variable excluded from the initial multivariable model.
Step4. Reintroduce variables excluded from the bivariate analysis
Reintroduce ECG (not a confounder)
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
538.9 71.0 10 0.00 72.7 10 0.00 72.7 10 0.00
a. Beginning Block Number 1. Method = Enter
111
Variables in the Equation
Covariate P OR 95.0% CI
Type Of Work 0.00
Welder 0.00 3.7 (1.9, 7.2)
Assembler 0.03 2.2 (1.1, 4.3 )
Production security and quality
inspector
0.00 2.7 (1.5, 5.0)
Smelter 0.00 4.0 (1.8, 9.2)
Planing machine operator 0.06 3.9 (0.9, 16.6)
Spray painter 0.00 4.4 (1.7, 11.5)
Gas-cutting worker 0.14 2.6 (0.7, 9.1)
Total exposure time (yr) 0.00 1.5 (1.2, 1.8)
BP 0.04 1.6 (1.0, 2.6)
ECG 0.93 1.0 (0.7, 1.6)
Step 5. Interaction checking
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
520.3 91.2 24 0.00 91.3 24 0.00 91.3 24 0.00
a. Beginning Block Number 1. Method = Enter
112
Variables in the Equation
Covariate P OR 95.0% CI
Type Of Work 0.10
Welder 0.25 1.7 (0.7, 4.0)
Assembler 0.89 0.9 (0.4, 2.3)
Production security and quality
inspector
0.18 1.9 (0.8, 4.7)
Smelter 0.02 4.3 (1.3, 14.6)
Planing machine operator 0.17 5.4 (0.5, 58.8)
Spray painter 0.06 3.5 (0.9, 13.1)
Gas-cutting worker 0.68 1.7 (0.1, 18.9)
Total exposure time (yr) 0.32 0.5 (0.1, 1.9)
BP 0.98 1.0 (0.4, 3.0)
ECG 0.92 1.0 (0.7, 1.6)
Total exposure time (yr)*Type Of Work 0.46
Total exposure time (yr)*Welder 0.05 3.8 (1.0, 14.5)
Total exposure time (yr)*Assembler 0.08 3.4 (0.9, 13.1)
Total exposure time (yr)*Production
security and quality inspector
0.16 2.9 (0.7, 13.3)
Total exposure time(yr)*Smelter 0.23 2.3 (0.6, 9.3)
Total exposure time (yr)*Planing
machine operator
0.27 2.3 (0.5, 10.7)
Total exposure time(yr)*Spray painter 0.21 2.5 (0.6, 10.6)
Total exposure time (yr)*Gas-cutting
worker
0.23 3.4 (0.5, 24.1)
BP*Type Of Work 0.66
BP*Welder 0.62 1.5 (0.3, 6.7)
BP*Assembler 0.09 3.4 (0.8, 13.5)
113
BP*Production security and quality
inspector
0.94 0.9 (0.2, 4.8)
BP*Smelter 0.49 1.9 (0.4, 10.9)
BP*Planing machine operator 0.72 0.6 (0.0, 13.3)
BP*Spray painter 0.26 3.1 (0.4, 23.4)
BP*Gas-cutting worker 0.97 0.0 (0.0, 1.2)
Step 6.The final model
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
538.9 71.0 9 0.00 72.7 9 0.00 72.7 9 0.00
a. Beginning Block Number 1. Method = Enter
114
Variables in the Equation
Covariate P OR 95.0% CI
Type Of Work 0.00*
Welder 0.00* 3.7 (1.9, 7.2)
Assembler 0.03* 2.2 (1.1, 4.3)
Production security and quality
inspector
0.00* 2.7 (1.4, 5.0)
Smelter 0.00* 4.0 (1.8, 9.2)
Planing machine operator 0.06 3.9 (0.9, 16.6)
Spray painter 0.00* 4.4 (1.7, 11.4)
Gas-cutting worker 0.15 2.6 (0.7, 9.1)
Total exposure time (yr) 0.00* 1.5 (1.2, 1.8)
BP 0.04* 1.6 (1.0, 2.5)
The multivariable modelling process (a sensitivity test)
Step 1. Bivariate analysis (P<0.25*)
1. Department
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step
Change From Previous
Block
Χ2 df P Χ2
df P Χ2 df P
572.0 37.0 3 0.00 39.6 3 0.00 39.6 3 0.00
a. Beginning Block Number 1. Method = Enter
115
Variables in the Equation
Covariate P. OR 95.0% CI
Department 0.00
Shipbuilding 0.00 3.5 (2.2, 5.8)
Ship-repair 0.00 5.0 (2.8, 9.0)
Production security 0.00 2.8 (1.6, 5.0 )
2. Total exposure time
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
567.0 46.0 1 0.00* 44.6 1 0.00 44.6 1 0.00
a. Beginning Block Number 1. Method = Enter
Variables in the Equation
Covariate P OR 95.0% CI
Total exposure time (yr) 0.00* 1.7 (1.5, 2.1)
3. Electrocardiogram
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
611.4 0.18 1 0.68 0.2 1 0.68 0.2 1 0.68
a. Beginning Block Number 1. Method = Enter
Variables in the Equation
116
Covariate P OR 95.0% CI
ECG 0.68 1.1 (0.7, 1.6)
4. Blood pressure
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
608.6 3.1 1 0.08 3.0 1 0.08 3.0 1 0.08
a. Beginning Block Number 1. Method = Enter
Variables in the Equation
Covariate P OR 95.0% CI
BP 0.08* 1.5 (0.9, 2.2 )
Step 2.An initial multivariable model(P<0.05*)
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step
Change From Previous
Block
Χ2 df P Χ2
df P Χ2 df P
555.9 55.1 5 0.00 55.7 5 0.00 55.7 5 0.00
a. Beginning Block Number 1. Method = Enter
117
Variables in the Equation
Covariate P OR 95.0% CI
Department 0.08
Shipbuilding 0.04 1.9 (1.0, 3.6)
Ship-repair 0.01 2.5 (1.2, 5.2)
Production security 0.03 2.0 (1.1, 3.7)
Total exposure time (yr) 0.00* 1.5 (1.2, 1.8)
BP 0.03* 1.6 (1.0, 2.5)
According to our research aim, department is one of the important risk factor so it was
decided not to exclude it from the model.
Step 3. Reintroduce variables excluded from the initial multivariable model
There was no variable excluded from step 2
Step 4. Reintroduce variables excluded from the bivariate analysis
ECG (not a confounder)
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
555.9 55.1 6 0.00 55.7 6 0.00 55.7 6 0.00
a. Beginning Block Number 1. Method = Enter
118
Variables in the Equation
Covariate P OR 95.0% CI
Department 0.08
Shipbuilding 0.04 1.9 (1.0, 3.6)
Ship-repair 0.01 2.5 (1.2, 5.2)
Production security 0.03 2.0 (1.1, 3.7)
Total exposure time (yr) 0.00* 1.5 (1.2, 1.8)
BP 0.03* 1.6 (1.0, 2.5)
ECG 0.94 1.0 (0.7, 1.5)
Step 5. Interaction checking
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
546.1 67.0 12 0.00 65.5 12 0.00 65.5 12 0.00
a. Beginning Block Number 1. Method = Enter
119
Variables in the Equation
Covariate P OR 95.0% CI
Department 0.01
Shipbuilding 0.21 1.6 (0.8, 3.4)
Ship-repair 0.00 4.2 (1.8, 10.2)
Production security 0.07 2.3 (0.9, 5.5)
Total exposure time (yr) 0.83 .8 (0.2, 4.4)
BP 0.17 1.9 (0.8, 4.6 )
Department*Total exposure time (yr) 0.45
Shipbuilding*Total exposure time (yr) 0.46 1.9 (0.4, 10.1)
Ship-repair*Total exposure time (yr) 0.68 1.4 (0.3, 7.7 )
Production security*Total exposure time
(yr)
0.62 1.6 (0.3, 9.6)
BP*Department 0.24
BP*Shipbuilding 0.78 .8 (0.2, 3.2)
BP*Ship-repair 0.14 .3 (0.1, 1.5)
BP*Production security 0.25 .4 (0.1, 1.9)
BP*Total exposure time (yr) 0.30 1.2 (0.8, 1.9)
Step 6.The final model
Omnibus Tests of Model Coefficientsa
-2 Log
Likelihood
Overall (score) Change From Previous Step Change From Previous Block
Χ2 df P Χ2
df P Χ2 df P
555.9 55.1 5 0.00 55.7 5 0.00 55.7 5 0.00
a. Beginning Block Number 1. Method = Enter
120
Variables in the Equation
Covariate P OR 95.0% CI
Department 0.08
Shipbuilding 0.04 1.9 (1.0, 3.6)
Ship-repair 0.01 2.5 (1.2, 5.2 )
Production security 0.03 2.0 (1.1, 3.7)
Total exposure time (yr) 0.00* 1.5 (1.2, 1.8 )
BP 0.03* 1.6 (1.0, 2.5)
121
APPENDIX 2: Ethics Application Exempt
Dear Ms Haiming Luo
Project Title: The impacts of high ambient temperature on the prevalence of
urolithiasis among outdoor workers in Guangzhou, China
Ethics Category: Human
Status: Exempt
Exempt Number: 1000001329
This email is to advise that your application has been reviewed by the Chair, University
Human Research Ethics Committee (UHREC) and deemed exempt from the need for
UHREC review, approval and monitoring in conformity with sections 5.1.22 and 5.1.23 of
the National Statement on Ethical Conduct in Human Research (2007).
Please note that since this exemption has been granted, responsibility for ensuring that the
project is conducted in accord with the National Statement, with relevant legislation and with
QUT policies still rests with you, the investigator, and responsibility for monitoring
compliance rests with your Supervisor and/or Head of School. Please inform your Supervisor
and/or Head of School of any changes to the study protocol, also informing.
UHREC, via the Research Ethics Unit, if the study protocol changes in ways that might affect
this exemption, for example altering risks or the usage of personal information.
Please also note you are required to keep an auditable record of any human research that is
exempted from ethical review as per section 5.2.9 of the National Statement.
Please note that exemption is not equivalent to approval and therefore care must be taken to
accurately describe the conditions under which this study has been reviewed. UHREC
recommends the following statement be used when drafting manuscripts for publication:
"The QUT University Human Research Ethics Committee assessed this research as meeting
the conditions for exemption from HREC review and approval in accordance with section
5.1.22 of the National Statement on Ethical Conduct in Human Research (2007)."
Should you have any further queries please do not hesitate to contact the Research Ethics
Unit on 3138 5123.
122
Regards,
Janette Lamb on behalf of the Chair UHREC
Research Ethics Unit | Office of Research
Level 4 | 88 Musk Avenue | Kelvin Grove
p: +61 7 3138 5123
123
APPENDIX 3: GIHI Approval Letter
124
125
APPENDIX 4: Information Collection Form to OHSOs
1. When was the period of highest temperature at the workplace in summer from 2003 to
2010 according the company‟s historical records? What were the highest temperature
records of different sites?
2. Could you estimate the outdoor working time (hours/day) and relevant workloads (light,
medium and heavy) for the following types of work? Welder, assembler, production
security and quality inspector, smelter, planing machine operator, spray painter and gas
cutting worker.
3. Did the company supply free drinking water to workers in shipbuilding and ship-repair
departments, and arrange regular water breaks for them especially in summer?
4. Did the company provide free meals for workers and employees in the company? If yes,
how many times?
5. Had the company ever noticed about the prevalence of urolithiasis according to the health
check files of 2003 to 2010? If yes, had any treatment ever been provided? What kinds of
treatment and at what time?
Participant Signature: Time:
Investigator Signature: Time: