Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Association between Long-term Exposure to Outdoor Air Pollution and
Mortality in China: a Cohort Study
Haidong Kan (阚海东阚海东阚海东阚海东), PhDFudan University (复旦大学复旦大学复旦大学复旦大学)
A&WMA International Specialty ConferenceXi’an, ChinaMay 13, 2010
Outline
• Background
• Methods
• Results and discussion
• Conclusion
“Pollution in China” – by Guang LU
Background
Estimated PM10 Concentration in World Cities (pop=100,000+)
Source: World Bank
Background
Annual PM10 levels in 31 provincial capitals of China (µµµµg/m3
,,,,2008)
Background
PM10
(µµµµg/m3)
100China AQS (II)
20WHO AQG
Chinese Burden of Disease from Top 10 Risk FactorsPlus Selected Other Risk Factors
0% 1% 2% 3% 4% 5% 6% 7%
Climate change
Unsafe sex
Physical inactivity
Lead (Pb) pollution
Urban outdoor air pollution
Unsafe water/sanitation
Cholesterol
Low fruit & vegetables
Road traffic accidents*
Overweight
Indoor smoke from solid fuels
Occupational hazards (5 kinds)
Underweight
Tobacco
Blood pressure
Alcohol
Percent of All DALYs in 2000
Smith et al., 2005. Derived from WHO data
Chinese burden of disease from top 10 risk
factors plus selected other risk factors
Outdoor air pollution
Background
A summary of outdoor air pollution epidemiologic studies in China
• Short-term exposure studies: – Time-series/case crossover studies
• Single-city analysis: Beijing, Hong Kong, Shanghai, Shenyang, Taiyuan, Wuhan, etc.
• Multi-city analysis: PAPA, CAPES
– Panel study: Beijing Olympics
• Long-term exposure study– Cross-sectional study: several
– Cohort study: NONE previously
• Intervention study– Beijing Olympics
– Hong Kong
Background
Outline
• Background
• Methods
• Results and discussion
• Conclusion
CNHS-Air Study• China National Hypertension Survey (CNHS)
– Baseline survey in 1991
– Follow-up visit in 1999
– 158,666 participants in 17 provinces of China (including both urban and rural areas)
• CNHS-Air: a retrospective cohort analysis of outdoor air pollution and mortality in China– 70,947 urban participants in 31 Chinese cities
Methods
Previous findings from
the CNHS cohort
JAMA, 2006NEJM, 2005 NEJM, 2009
Method
Health Outcome
• Death (Y/N & follow-up time; ICD 9)
– All-cause non-accidental mortality
– Cardiovascular mortality
– Respiratory mortality
– Lung cancer mortality
– Other cause of cancer (control)
Method
Air pollution exposure
Method
Cities in the CNHS-Air Study
• Period: 1991-1999
• Pollutants: TSP, SO2, NOx (PM10/PM2.5/NO2/O3
not available)
• ZIP code linking residential address to nearest air monitoring station
Covariates
• Baseline examination in 1991
– Age, sex, BMI, physical activity, education, alcohol
intake (never, former and current drinker), and hypertension
– smoking status (current, former and never smoker),
age at starting to smoke, years smoked, cigarettes per day
Method
Statistical analysis
• Proportional hazards regression models, using
SAS version 9.1
• Adjusted for a wide range of covariates
• Single-pollutant & multi-pollutant models
• Stratified analyses by sex, smoking, obesity, and
education
Method
Outline
• Background
• Methods
• Results and discussion
• Conclusion
Descriptive analysis
• 8,319 deaths out of 70,947 subjects
Results
CardiovascularCardiovascularCardiovascularCardiovascular36%36%36%36%
RespiratoryRespiratoryRespiratoryRespiratory11%11%11%11%Lung cancerLung cancerLung cancerLung cancer8%8%8%8%Other cancersOther cancersOther cancersOther cancers16%16%16%16%
Other causesOther causesOther causesOther causes29%29%29%29%
Descriptive analysisResults
Deceased subjects were:
• older
• more likely to be
male or current smokers
• had less physical
activity, and
higher prevalence of hypertension
Descriptive analysis
Results
• Between 1991-1999, the averaged annual concentrations
were 289 µg/m3 for TSP, 74 µg/m3 for SO2, and 50 µg/m3
for NOx
050100150200250300350
TSP SO2 NOx
µg/m3China AQS IChina AQS IIChina AQS IIIOur study
Descriptive analysis
Results
Descriptive analysis
050
100150200250300350
91 92 93 94 95 96 97 98 99 年
ϕg/m3
TSPSO2NOx
• TSP↓↓↓↓5%
• SO2↓↓↓↓ 35%
• NOx↑↑↑↑28%
Results
-1.7 (-4.2, 0.8)
0.2 (-1.0, 1.4)
-0.1 (-0.9, 0.7)
Other causes of cancer
2.6(-1.0, 6.4)
4.2 (2.2, 6.2)
1.0(-0.2, 2.2)
Lung cancer
2.8 (-0.1, 5.7)
3.5 (2.0, 5.0)
0.4(-0.6, 1.3)
Respiratory
2.2%(0.5, 4.0)
3.1 (2.3, 4.0)
0.8(0.3, 1.4)
Cardiovascular
1.4 (0.4, 2.4)
1.9 (1.4, 2.3)
0.2 (-0.1, 0.5)
All-cause
NOxSO2TSPCause of death
% increase of deaths associated with 10μμμμg/m3
increase in air pollutants’ concentrations
PM health effects, compared with the ACS cohort study
CNHS*: 10 μg/m3 increase of PM2.5 was associated with
- 0.6% ↑ of total mortality
- 2.5% ↑ of cardiovascular mortality
- 1.2% ↑ of respiratory mortality
- 3.1% ↑ of lung cancer mortality
*: assuming PM2.5/PM10≈0.65 and PM10/TSP ≈0.5
ACS Cohort:10 μg/m3 increase of PM2.5 was associated with
- 4% ↑ of total mortality
- 6% ↑ of cardiopulmonary mortality
- 8% ↑ of lung cancer mortality
Pope et al, JAMA, 2002
Discussion
Discussion
1.0000
1.0100
1.0200
1.0300
1.0400
1.0500
1.0600
Acute effects Chronic effects
China
Western
Countries
Kan et al, 2005
Results
Single vs. multi-pollutant models
1.6 (95%CI: -1.4, 4.7)1.1 (95%CI: -0.7, 2.9)0.6 (95%CI: -0.4, 1.7)adjusted for SO2
2.3 (95%CI: -0.8, 5.4)1.4 (95%CI: -0.4, 3.3)1.4 (95%CI: 0.3, 2.5)adjusted for TSP
2.8 (95%CI: -0.1, 5.7)2.2 (95%CI: 0.5, 4.0)1.4 (95%CI: 0.4, 2.4)single-pollutantNOX
3.3 (95%CI: 1.9, 4.9)3.0 (95%CI: 2.2, 3.9)1.8 (95%CI: 1.3, 2.3)adjusted for NOX
3.4 (95%CI: 2.0, 4.9)3.1 (95%CI: 2.2, 3.9)1.8 (95%CI: 1.4, 2.3)adjusted for TSP
3.5 (95%CI: 2.0, 5.0)3.1 (95%CI: 2.3, 4.0)1.9 (95%CI: 1.4, 2.3)single-pollutantSO2
0.5 (95%CI: -0.6, 1.6)0.8 (95%CI: 0.2, 1.4)0.0 (95%CI: -0.4, 0.4)adjusted for NOX
0.6 (95%CI: -0.4, 1.7)0.8 (95%CI: 0.2, 1.4)0.0(95%CI: -0.4, 0.4)adjusted for SO2
0.4 (95%CI: -0.6, 1.3)0.8 (95%CI: 0.3, 1.4)0.2 (95%CI: -0.1, 0.5)single-pollutantTSP
Respiratory
mortalityCardiovascular
mortality
Total
mortalityModel
1988 1989 1990 1991 1992 1993 1994 1995
020
40
60
80
NO2
SO2
O3
PM10
Mic
rogra
ms p
er
cubic
metr
e
Year
HALF YEARLY MEAN LEVELS
Fuel restriction on sulphur
50% reduction in SO2
after the intervention
No change in other
pollutantsHedley et al 2002
Hong Kong, 1988 – 95Discussion
Hedley et al, Lancet 2002
-6-5
-4-3
-2-1
0
15-64 65+ 15-64 65+ 15-64 65+
Reductions In Deaths After Sulfur Restriction
All causes Cardiovascular Respiratory
% R
eduction in a
nnual tr
en
d
-1.8%
-2.8%
-1.6%
-2.4%
-4.8%
-4.2%
Discussion
Hedley et al, Lancet 2002
Comparison of SO2’s health effect
Discussion
0.6
(95%CI: -0.4, 1.5)
4.2
(95%CI: 2.2, 6.2)
Lung cancer
mortality
2.0
(95%CI: 0.8, 3.2)
3.5
(95%CI: 2.0, 5.0)
Respiratory
mortality
1.0
(95%CI: 0.2, 1.9)
3.1
(95%CI: 2.3,4.0)
Cardiovascul
ar mortality
1.1
(95%CI: 0.5, 1.7)
1.9
(95%CI: 1.4, 2.3)
Total
mortality
HK
intervention study
Our
cohort study
Stratified analysis
Results
0.550.7 (-1.6, 3.1)<0.013.5 (2.3, 4.6)0.200.5 (-0.3, 1.3)High
0.610.011.7 (0.6, 2.8)0.19<0.011.4 (0.8, 1.9)0.260.490.1 (-0.2, 0.5)LowEducation
0.042.1 (0.1, 4.2)<0.012.2 (1.2, 3.3)0.770.1 (-0.6, 0.8)BMI ≥ 25
0.380.011.4 (0.3, 2.6)0.09<0.011.8 (1.2, 2.4)0.770.050.4 (0.0, 0.8)BMI < 25Obesity
0.420.6 (-0.9, 2.2)0.070.8(-0.1, 1.6)0.230.3 (-0.2, 0.8)Current/past
0.240.012.0 (0.6, 3.4)0.01<0.012.4 (1.8, 3.0)0.680.550.1 (-0.3, 0.6)NeverSmoking
0.022.3 (0.8, 3.9)<0.011.5 (0.7, 2.2)0.970.0 (-0.5, 0.5)Female
0.130.410.6 (-0.8, 2.0)0.18<0.012.3(1.6, 3.0)0.180.070.4 (0.0, 0.9)MaleSex
p for
interaction
pEffect sizep for
interaction
pEffect sizep for
interaction
pEffect size
NOx
SO2TSP
Never-smokers may be more susceptible to air pollution exposure !
Strengths & limitations• Strengths
– Large sample size
– Detailed individual information on potential confounders
• Limitations– TSP and NOx only
– Air pollution exposure on the aggregated city level, rather than individual level
– Lack of some potential confounders, e.g. diabetes
Discussion
Conclusion
• First cohort study of air pollution health effects in China
• Significant associations between air pollutants and mortality from cardiopulmonary diseases and lung cancer
• Independent health risk of SO2 in China
• Preliminary data of susceptible sub-population
Acknowledgement
• Chinese Academy of Medical Sciences: DongfengGu
• Fudan University: Bingheng Chen, Renjie Chen
THANK YOU FOR YOUR ATTENTION
SO2 NOx PM10
PM2.5 CO VOCs
Spatial distribution of air pollution emissions in China (2005)
Method