EPIDEMIOLOGIC TRENDS OF HEAD AND NECK CANCERSshodhganga.inflibnet.ac.in/bitstream/10603/2363/12/12_chapter 3.pdf · tobacco smoking and heavy alcohol consumption, with low intake

3. EPIDEMIOLOGIC TRENDS OF

HEAD AND NECK CANCERS

25

3.0 EPIDEMIOLOGIC TRENDS OF HEAD AND NECK CANCERS

3.1 Introduction

Head and neck cancers are among the 10 most common cancers globally. In India, it

accounts for one fourth of male cancers and one tenth of female cancers (Yeole 2001).

The sub-site predilection of head and neck squamous cell carcinoma (HNSCC) reflects

the risk profile of a community and there are suggestions that these are changing over

time.

The risk factors in different countries are different. This leads to difference in risk of

HNSCC and sub-site predilection. With changes in risk profile, both the incidence and the

sub-site predilection may change. Among females the age-adjusted rates of India are the

highest in the world (Parkin 1997; Parkin 1998; Sankaranarayanan, Masuyer et al. 1998;

Parkin 2002). In India, oral cavity (ICD-9 141,143-5) is the predominant site, whereas

pharynx (ICD-9 146,148,149) is the common site in France (Coleman 1993). The

difference in disease predilection may be a reflection of the prevailing risk habits of the

population (Sankaranarayanan 1998; Menvielle 2004).

Globally the major cause for oral cancer is the use of tobacco products (Johnson 2001;

Winn 2001 ). Consumption of alcohol is synergistic as a risk factor; it is super-

multiplicative for the mouth, additive for the larynx and between additive and

multiplicative for the esophagus (Fioretti 1999 ). Tobacco smoke, smokeless tobacco,

alcohol, marijuana use, deficient diet, Human Papilloma virus infection have been

suggested as risk factors for head and neck cancers (Hashibe 2000; Iype 2001; Balaram

2002). The practice of different habits in different populations might account for the

difference in subsite predilection and the difference in the incidence observed.

3.2 Hypothesis

Changes in risk habits over the years may reflect in the incidence and sub-site

predilection of head and neck cancer.

3.3 Objectives

1. To determine the trends in incidence of head and neck squamous cell carcinoma

over a time period in India.

26

2. To compare change in the trends of head and neck squamous cell carcinoma of

India and US populations.

3. To evaluate etiologic basis for the change in trends (if any) of head and neck

squamous cell carcinoma.

The overall goal of this part of the project is to determine whether there is a change in the

trends of head and neck cancers in India and to find out if they are comparable between

urban and rural populations and those of the developed nations. We hypothesize that the

changes in risk habits and urbanization of rural population may cause alterations in the

patterns of HNSCC, like the age of onset and subsite predilection. To test the hypothesis

we conducted an in-depth analysis of cancer registry data representing one urban and one

rural population in India. We also compared the incidence trends in head and neck

cancers with data from developed countries to frame an etiological hypothesis if any.

Additionally we have investigated whether or not there has been a change in the risk

factor profile in comparison to historical data to the present time, and to determine

whether there are significant numbers of patients without any established risk factors in

our hospital set up.

3.4 Background

Head and Neck Squamous cell carcinoma (HNSCC), which constitutes the most common

upper aero digestive tract carcinomas, has a distinct geographical predilection. It accounts

for 3% of all cancers in the developed countries like USA whereas in developing

countries like India it accounts for 30% of all cancers.

3.4.1 Variation in incidence

3.4.1.1 Geographic

The age-adjusted rates of head and neck cancers are highest in countries like France,

India, Brazil, and USA (Blacks) (Fig. 3.1). The age adjusted rates of head and neck

cancers range from 45 in France to 5 in places like Costa Rica, Kuwait and Japan (Parkin

2002). There is also a distinct site predilection in different countries.

27

As far as oral cancers are concerned, the AAR range from as high as 21.5 among males in

South Karachi, Pakistan to as low as 0.8 in Ecuador, Quito (Curado 2007) (Fig 3.2).

AAR - Countries

<21.5 -Pakistan, India, France, Brazil, Spain, Australia,

<10.0 - USA Blacks, Germany, Thailand, Italy, Russia, USA Whites, Canada,

Switzerland

<5.0 – Netherlands, Finland, Ireland, Japan, Iceland, Sweden, Singapore, China

<2.5- Oman, Korea, Uganda, Malaysia, Algeria, Zimbabawe, Ecuador.

Among females it is as high as 19.3 in South Karachi, Pakistan to as low as 0.1 in Chile.

<19.3- Pakistan, India, Australia

<5.0 – Switzerland, USA Blacks, USA Whites, Thailand, Brazil, Canada, Switzerland

<2.5 - Austria, France, Germany, Italy, Malaysia, Spain, Finland, Japan

<1.0 – Korea, Israel, China, Algeria, Chile

3.4.2 Variations in risk factors

3.4.2.1 Smoking

Smoking of tobacco in the forms of cigarettes, bidis, cigars and cheroots and loose

tobacco in pipes is common. And previous literature gives good evidence of dose

response relationship for both tobacco and alcohol (Blot 1988; Blot 1992). Bidi smoking

is more hazardous than cigarette smoking. Reverse smoking with the lighted end inside

28

the mouth which is practised in rural Andhra Pradesh is linked with hard palate cancers

(Gupta 1980; van der Eb 1993).

3.4.2.2 Smokeless tobacco

It is more in use globally, being placed into contact with the mucous membrane. The practice of

placing snuff in the lower labial sulcus though has become less popular, but still in practice in

Scandinavia and USA is associated with oral cancers (Blot 1977; Winn 1981). Tobacco flakes is

used for chewing by North Americans (Johnson 2003). In Cuba, 82% of oral cancer cases were

attributable to tobacco smoking & 19% to smoking cigars or pipe only (Garrote 2001). Over 80%

of oral cancers in the United States (Blot 1988) and in Europe (Negri 1993) are attributable to

tobacco smoking and heavy alcohol consumption, with low intake of fresh fruit and vegetables

playing a smaller role. In the UK, cigarette, cigar and pipe smoking are the main forms of tobacco

use and all are causes of oral cancer (Warnakulasuriya 2005).

The primary cause of the very high incidence of oral cancer in South Asia is the widespread habit

of chewing betel quid (or paan) and related areca nut use (Bedi 1996). In India tobacco is mostly

combined with other ingredients like betel leaf, arecanut, slaked lime, etc.

Table 3.1: Common forms of oral smokeless tobacco in developing countries (Johnson 2003)

Habit Ingredients Population

Pan/paan/betel quid Areca nut, betel leaf, slaked lime, catechu,

condiments, with or without tobacco

Indian subcontinent, South-east Asia, Papua,

New Guinea, part of South America

Khaini Tobacco and lime Bihar –India

Mishri Burned tobacco Maharashtra-India

Zarda Boiled tobacco India and Arab countries

Gadakhu Tobacco and molasses Central India

Mawa Tobaco, lime and areca Bhavnagar –India

Nass Tobacco, ash, cotton or sesame oil Central Asia, Iran, Afghanistan, Pakistan

Naswar/niswar Tobacco, lime, indigo, cardamom, oil,

menthol,etc.

Central Asia, Iran, Afghanistan, Pakistan

Shammah Tobacco, ash, lime Saudi Arabia

Toombak Tobacco and sodium bicarbonate Sudan

29

Chewing of areca nut is common in India, SriLanka, Pakistan, Bangladesh, Myanmar,

Thailand, Cambodia, Malaysia, Indonesia, Singapore, Philippines, New Guinea, Taiwan

and China and in people emigrating from these countries.

A relation between oral cancer and sexual habits also exist, but information is limited

(Maden 1992; Schwartz 1998) A deficient diet also predisposes towards the development

of oral cancers and it has been estimated that this may be responsible for 10-15% of cases

in Europe (La Vecchia 1997). Use of Marijuana is linked to the development of oral

cancers in USA (Zhang 1999). Human Papilloma Viruses especially HPV-16 is the most

common type associated with oral cancers (Park 1991; Yeudall 1992; Sexton 1993;

Woods 1993; Mao 1995 ).

3.4.3 Relationship to the anatomical site of cancer

There is a strong association between the site of cancer and the site where the quid is

placed regularly. Pooling of carcinogens in saliva give rise to cancers in the floor of the

mouth and ventral and lateral tongue.

Smoking is more strongly associated with soft palate and larynx cancers and alcohol with

floor of the mouth and tongue lesions (Boffetta 1992). Lip cancer is most common in fair-

skinned races, particularly in rural areas and in men who work out of doors. Intraoral

cancer in western countries most commonly affects the lateral borders of the tongue and

the floor of the mouth, followed by the buccal mucosa, mandibular alveolus, retromolar

region and soft palate. By comparison, the hard palate and dorsum of the tongue are the

lowest risk sites. In the high-risk areas of south Asia, the buccal, retromolar and

commissural mucosae are the most prone sites. Cancers of the tongue as well as buccal

mucosa have been noted to be quite common in India, attributed to the local custom of

chewing pan, betel leaf with tobacco. In India, smoking is the most common form of

tobacco consumption in urban male population in India whereas in rural population,

chewing tobacco is more common. There are evidences to suggest that the tobacco habits

in populations are changing over time (National Sample Survey Organization NSS 50th

Round (1993-1994) 1998)

3.4.4 Age distribution

In most countries the incidence of oral cancer is low in people below the age of 45

(Ferlay 2001). The incidence of oral cancer increases with age, though the pattern differs

in different countries and with different risk factors. Over the last few decades there have

30

been reports suggesting an increase in incidence among young adults (<45 years). There

is an alarming rise in the incidence of oral cancers among young men due to the heavy

abuse of smokeless tobacco (Johnson 1991; Levi 1994) Davis and Severson (Davis 1987)

were among the first who pointed out an increasing incidence of tongue cancer in young

adults in the United States – with data from the 1970s to 1980s. They reported a 13-fold

increase in those aged 10-29 years (although very small numbers and no significant

trend), a 1.8-fold increase in those aged 30-39 years, no change in the 40-49 age-group

and a 1.2-fold increase in those aged 50 and over. These increases were all accounted for

by rises in male incidence. SEER data analysis by Schantz et al revealed a 7% increase in

tongue cancer incidence from 1973 to 1984 in those under 40 years, which then remains

stable through to the 1990s, while in those over 40 years the incidence increased only

marginally (Schantz 2002). Despite these increases in younger adults, the incidence rates

for older adults remain significantly higher in all of these studies. The explanations for

the increases in younger populations are generally unclear, with various hypotheses

around behavioral risk factors proposed.

3.4.5 Ethnic distribution

Ethnicity strongly influences prevalence due to social and cultural practices. Where such

habits represent risk factors, their continuation by emigrants from high prevalence regions

to other parts of the world results in comparatively high cancer incidence in immigrant

communities (Swerdlow 1995).

3.4.6 Gender distribution

In all countries men are affected almost twice as often as women, probably due to their

higher indulgence in risk factors such as alcohol and tobacco consumption, for intraoral

cancer and sunlight for lip cancer. The incidence of tongue and other intraoral cancer for

women is, however, greater than or equal to that for men in high prevalence areas such as

India, where chewing and smoking are also common among women. Men and women are

also almost equally affected in some of the ethnic groups in Singapore, Denmark and

Hawaii. The disparity between the sexes has decreased progressively especially in

younger age groups which probably could be due to social acceptance of habits among

females. Globally highest rates have been reported from South Karachi of Pakistan and

Trivandrum of India in men (Curado 2007). For females, the highest rate reported has

been in South Karachi of Pakistan and Karunagapally, Poona, Chennai, Trivandrum of

31

India (Curado 2007). In developing countries like India, oral cancer ranks the first most

common cancer in males and third most common cancer in rural Indian females (Mathew

2007).

3.4.7 Trends

Globally, overall there continues to be an increase of newly diagnosed oral cancers

particularly tongue cancers (Parkin 2001; Ferlay 2004; Parkin 2005) The rising trend in

tongue cancer in young men in western countries, is thought to be due to marked

increases in alcohol consumption, perhaps combined with increased use of oral smokeless

tobacco products, especially in the USA and Nordic countries (Macfarlane 1987;

Franchesci 1994; Shiboski 2000; Schantz 2002).

3.4.8 Migrant studies

Migrant studies related to oral cancer are few. McCredie and colleagues found migrants

from the Middle East to Australia have lower rates than native Australians (McCredie

1994). Migrants from Asia and China to London, UK have also been shown in several

descriptive case-series studies to bring with them a higher risk of oral cancer compared to

others in the area (Swerdlow 1995; Warnakulasuriya 1999). This risk was shown to be

passed down generations, and this was considered to be a result of cultural behaviors

including tobacco and betel quid use (Swerdlow 1995; Warnakulasuriya 1999;

Warnakulasuriya 2002). In one another survey it was reported that 80% of the

Bangladeshi community living in London reported using betel quid chewing while also

adopting the cigarette behaviors of the native Londoners (Ahmed 1997).

3.5 Materials & Methods

3.5.1 Population studied

Objective 1: The tumor registry data was taken from an urban South Indian cancer

registry (Chennai-1986-98) and from a rural (Barshi-1988-98) registry, which has data for

a long period since 1982 and 1987 respectively to identify the change in trends of head

and neck cancers. Chennai, a South Indian urban cancer registry caters to an area of

170sq.km and a population of 4.2 millions. The rural registry data was available only

from Barshi. It is the first rural registry in India that covers a total population of around

0.4 million in 346 villages spread over 3713 sq.km. in Solapur district of Maharashtra.

The National Cancer Registry Programme of the Indian Council of Medical Research

(ICMR) monitors these cancer registry data. The data from these two registries were used

to analyze change in trends within the country.

Objective 2: To compare change in trends from India to that of developed nation, data

was taken from the Surveillance, Epidemiology, and End Results (SEER) Program Public

Use (1973-97), National Cancer Institute, Cancer Surveillance Research Program, Cancer

Statistics Branch, of USA, released in April 2000. This data encompass 10% of the U.S.

population. Comparable data from India for this time period was obtained from Mumbai

Cancer Registry. Acceptable levels of reliability and completeness of data have been

reported for the Mumbai cancer registry (Yeole 2001). Mumbai cancer registry

established in 1963 became part of National Cancer Registry Program (NCRP) since

1981-82. It covers the area under the city of Greater Bombay with the area of about 603

sq.km. The data for this registry was taken from Cancer Incidence in 5 Continents vol. IV

to VIII (1973-97). The Descriptive Epidemiology Group of IARC (The International

Agency for Research on Cancer) monitors these data.

Objective 3: To investigate the risk factor profile of head and neck cancers, a

retrospective study was done with data obtained from AIMS Head and Neck tumour

board records between 1st January 2004 and 31st December 2006.

This retrospective study constituted data of 611 Head and Neck cancer patients.

Demographics, cancer sub-site and risk factors were recorded for each patient. The

number of tongue cancer patients with and without risk factors were statistically analysed,

using the Chi-Square test, against various other groups of sub-sites in the Head and Neck

region.

3.5.2 Statistical method adopted

Objectives 1 and 2

In this study we used Age Adjusted incidence Rate (AAR) for analyzing the trends.

Age Specific Incidence Rate (AspR) refers to the rate obtained by dividing the total

number of cancer cases by the corresponding estimated population in that age group and

sex/ site/ geographic area/ time period and multiplying by 100,000.

ASpR = New cases of cancer of a particular year in the given age group x 100,000

Estimated population of the same year for the given age group

32

33

As age increases the incidence of cancers also increases, therefore with an increase in the

median age of a population the cancer incidence also increases in the community. In order

to make the rates comparable between different populations, the age specific incidence

rates are adjusted to the five-year age distribution of the world standard population

(Plummer 1997) to obtain the age-adjusted rates (AAR). Age adjusted rates are calculated

for all age groups as well as for the young (20-44 years) age groups in this study and for

males and females separately.

AAR = Σ (ASpR) X No. of persons in Std. world population in that 5 yr. age group

100,000

Joinpoint regression program is a statistical software for analyzing the trends using

joinpoint models, where different lines are connected together at the joinpoints. The

joinpoint regression analysis involves fitting a series of joined straight lines on a log scale

to the trends in the annual age-adjusted cancer incidence rates. Line segments are joined

at points called joinpoints. Each joinpoint denotes a statistically significant (P = .05)

change in trend. The software takes trend data and fits the simplest joinpoint model that

the data allow. The program starts with the minimum number of joinpoint (e.g. 0

joinpoints, which is a straight line) and tests whether more joinpoints are statistically

significant and must be added to the model (up to that maximum number). This enables

us to test that an apparent change in trend is statistically significant. This is estimated by

the method of Least Squares and the regression equation is stated as Y = mX + b, where

Y is the outcome or the dependent variable- the age adjusted rate, X is the independent

variable- year of diagnosis, m is the annual percentage change, b is the Y-intercept. The

tests of significance use the Monte Carlo permutation method. In the final model, each

joinpoint indicates a statistically significant change in trend, and an estimated annual

percentage change (EAPC) is computed for each of those trends by means of generalized

linear models. A maximum number of 3 joinpoints was allowed for estimations. Once the

line segments are established, the estimated annual percent change is used to describe and

test the statistical significance of the trends in the model. Testing the hypothesis (two-

sided P value < .05) that the annual percent change is equal to zero is equivalent to testing

the hypothesis that the trend in the cancer rates is neither increasing nor decreasing.

Annual Percent Change (APC) gives a complete picture of the trend over time. A short

segment rising at a steep rate may not be statistically significant. Hence to summarize the

trend over a fixed predetermined interval, Average Annual Percent Change (AAPC) is

34

used which computes the summary measure over the fixed interval. This is advantageous

especially when the data is from a small geographic area like Barshi. Since each gives

different perspective, both APC and AAPC are given in the tables. Joinpoint software

version 3.3.1 was used.

Cumulative incidence risk percent and the lifetime risk of developing head and neck

cancers and other subsites were calculated for the urban and rural registry data.

Cumulative risk (Cu R) refers to the probability that the person will develop a particular

cancer during a certain age period in the absence of any other cause of death. The

Cumulative Rate is an approximation of the cumulative risk. It is obtained by adding the

annual age-specific incidence rates for each five-year age interval (for whatever age

group is to be used to calculate the cumulative risk) multiplied by 5 (representing the five

year age interval) times 100 divided by 100000

CuR = 5xΣ(ASpR) x100

100,000

The cumulative incidence rate has a useful probabilistic interpretation. This also removes

the arbitrariness of choosing a standard population. The confidence intervals are

calculated using the formula CIRP+/- (1.96X S.E.). The probability of getting a specific

cancer, expressed in terms of one in every ‘n’ persons is computed by reciprocating

estimated cumulative incidence rate expressed as a percentage.

Head and neck cancers in our study I constitute cancers of the oral cavity, pharynx, and

larynx (ICD-9, 141, 143-146, 148, 149, 161). As there are reports to suggest that the

incidence of tongue cancers is increasing in young adults (Schantz 2002; Stimson and

Guo-Pei 2002), we have separated tongue (ICD-9, 141) from the rest of oral cavity (ICD-

9, 143-5) in our analyses.

Head and neck cancers in our study II include cancers of the tongue, mouth, tonsil,

oropharynx, pyriform sinus, hypopharynx, pharynx unspecified, larynx and excludes

salivary glands, nasopharynx, eye/orbit. Mumbai: ICD-10, C 01-06, C 09-14, USA: ICD-

O-3, C 019-069, C 090-109, C 129, C 130-139, C 140, 142-148. See ICD coding in

annexure I.

35

Objective 3

The statistical test used to analyse the data was the Chi-Square test using the Epi-info

software. The number of tongue cancer patients with and without risk factors were

analyzed against those with cancers of the other head and neck sub sites.

In our retrospective study, 611 HNSCC patients, who presented at AIMS Hospital,

Cochin (Kerala) between 1st January 2004 and 31st December 2006, formed our study

group. The data was obtained from the department’s ‘Tumour Board’ records, which is a

collaboration of each patient’s full history and examination. The data collected for each

patient included:

• Demographics

• Risk factors (Smoking, Alcohol, Paan chewing, Pre-malignant lesions or None)

• Tumour sub-site

• Treatment received

• Follow-up status, where recorded.

The inclusion criteria, for the statistical analysis, were that the patient must have a cancer

of the oral cavity, oropharynx, hypopharynx or larynx. Oral cavity cancers included those

of tongue, buccal mucosa, floor of mouth, retromolar trigone, hard palate, lip and

alveolus. Oropharynx cancers included those of the soft palate, tonsils and other

unspecified oropharyngeal sites. Hypopharnyx cancers included those of the lateral and

posterior pharyngeal walls, pyriform sinus, post cricoid and other unspecified

hypopharyngeal sites. Larynx cancers included those of the glottis, supraglottis and other

unspecified laryngeal sites.

Patients excluded were those who suffered from any other head and neck cancer. Risk

factors were recorded if the patient had the habit at the time of, or within 1 month prior to,

their diagnosis.

The statistical test used to analyse the data was the Chi-Square test. The number of tongue

cancer patients with and without risk factors were analysed against those with cancers of

the previously mentioned sites. Our hypothesis was that there is significant difference

between the number of tongue cancer cases that present without the known associated

risk factors when compared to other subsites. A p value of 0.05 or less determined

significance.

3.6 Results

3.6.1 Objective 1

To determine the trends in incidence of head and neck squamous cell carcinoma over a

time period in India.

The urban registry recorded a total of 6857 head and neck cancers during the time period

1986 to 1998, out of which 4777 were males (23% of all male cancers), and 2080 were

females (9% of female cancers). In the rural registry, a total of 325 head and neck cancers

were registered during the time period 1988 to 1998, out of which 272 were males (28%

of all male cancers) and 53 were females (5% of all female cancers).

The overall male: female ratio of head and neck cancers in urban population is 2:1and in

rural population is 5:1. Table 3.2 gives the subsite distribution of cancers.

Table 3.2: Sub site distribution of total head and neck cancers in urban and rural Indian populations

Urban Rural

Male Female Young

male

Young

female

Male Female Young

male

Young

female

Site

n % n % n % n % n % n % n % n %

Tongue 1019 21 298 14 123 22 47 13 44 16 12 23 6 15 4 29

Oral

cavity

1252 26 1175 57 140 25 181 48 59 22 21 40 15 39 2 14

Larynx 863 18 104 5 100 18 24 6 43 16 5 9 5 13 3 21

Pharynx 1643 34 503 24 198 35 122 33 126 46 15 28 13 33 5 36

36

The AAR of the head and neck cancers and its subsites were analyzed in all and young

age groups in both males and females separately.

Age: To investigate whether there is any change in the incidence of head and neck

cancers in the young adults the AAR for age group 20-44 years were analyzed. In the

urban population, among males, 12% of the total head and neck cancers were from

younger age group (20-44 years) and among females it was 18%. In the rural population,

the contribution of young males was 14.5% and young females were 26%. It is interesting

to note that the young females contribute to nearly one fourth of all rural female head and

neck cancers. Among young adults, the male: female ratio in urban population was 1.5:1

and in rural population it was 3:1. In contrast in patients over 45 years old the male:

female ratio was 2.5:1 in the urban and 6:1 in the rural community.

Subsite: The sub-site distribution of all head and neck cancers in all ages and in young

adults is given in table.2. Oral cavity including tongue was the predominant site in the

urban and rural population except in rural males, where pharynx was the predominant

site. The age specific incidence rates of head and neck cancers showed the peak incidence

in 70-74 years in both urban and rural populations (Fig. 3.3)

020406080

100120140160180

0-4

10 to

14

20 to

24

30 to

3440

-4450

-5460

-6470

-74

1986-881989-911992-941995-98

Age groups

Age

Spe

cific

Inci

denc

e R

ate

Figure 3.3: Age Specific Incidence Rates of Head and Neck Cancers in India

The joinpoint analysis of the trends in the age-adjusted cancer incidence allows us to

more accurately interpret changes over time and, more importantly, to determine if those

changes are statistically significant. Figures 3.4 through 3.7 graphically display the results

of the joinpoint analysis shown in Tables 3.3 and 3.4.

37

38

The table below shows the best joinpoint fit for each site and the APC & AAPC of the

urban and rural registry data for all ages.

Table 3.3: Joinpoint analysis of the trends in the Age Adjusted Cancer Incidence

Rates of Urban and Rural registries -all ages

APC = Annual Percent Change; AAPC = Average Annual Percent Change

Chennai males-Joinpoint analysis 1986-98 trend 1 trend 2 trend 3 Cancer site: years APC years APC years APC

Head & Neck 1986-98 -1 Tongue 1986-98 -1.1

Other oral cavity 1986-98 -2.3 pharynx 1986-98 -1.76 larynx 1986-98 -0.21

Chennai females-Joinpoint analysis 1986-98 trend 1 trend 2 trend 3 Cancer site: years APC years APC years APC

Head & Neck 1986-98 -3.8

Tongue 1986-98 2.7

Other oral cavity 1986-98 -5.5

pharynx 1986-98 -3.5

larynx 1986-98 -1.4

Barshi males-Joinpoint analysis 1988-98 trend 1 trend 2 trend 3 Cancer site: years APC years APC years APC

Head & Neck 1988-98 -7.4

Tongue 1988-98 -4.3

Other oral cavity 1988-98 -0.2 1991-94 -37.11994-97 44.3

pharynx 1988-98 -8.2

larynx 1988-98 -9.6 Barshi females-Joinpoint analysis 1988-98 trend 1 trend 2 trend 3 Cancer site: years APC years APC years APC

Head & Neck 1988-98 7.7

Tongue 1988-98 24.3

Other oral cavity 1988-98 5.5

pharynx 1988-98 -3.3

larynx 1988-98 *61.9

AAPC=-3.3 last 10 observations

Fig.3.4. Trends in AAR of Head and Neck cancers and subsites – Chennai all ages

a b

c d

e f

g h

i j

a: Head and Neck cancers males, b: Head and Neck cancers females, c.Tongue cancers male, d. Tongue cancers female

e. Other oral cavity cancers male, f. Other oral cavity cancers female, g. Pharynx cancers male, h. Pharynx cancers female

i. Larynx cancers male, j. Larynx cancers female

Fig.3.5. Trends in AAR of Head and Neck cancers and subsites – Barshi all ages

a b

c d

e f

g h

i j




39

Table 3.4 shows the best joinpoint fit for each site and the APC & AAPC of the urban and

rural registry data for the younger age group.

Table 3.4: Joinpoint analysis of the trends in the Age Adjusted Cancer Incidence Rates of

Urban and Rural registries –young

Chennai males-Joinpoint analysis 1986-98 trend 1 trend 2 Cancer site: years APC years APC Head & Neck 1986-98 -3.8 Tongue 1986-98 -7 Other oral cavity 1986-98 -2.3 Pharynx 1986-98 -3.7 Larynx 1986-98 -2.3 Chennai females-Joinpoint analysis 1986-98 trend 1 trend 2 trend 3 trend 4 Cancer site: years APC years APC years APC years APC Head & Neck* 1986-98 19.2 1988-92 -15.3 1992-95 2.4 1995-97 -11.2 Tongue 1986-98 0.8 Other oral cavity 1986-98 -7.8 pharynx 1986-98 -7.7

larynx 1986-98 -

10.6 Barshi males-Joinpoint analysis 1988-98 trend 1 trend 2 Cancer site: years APC years APC Head & Neck 1988-98 -2.7 Tongue 1988-98 -24.5 Other oral cavity 1988-98 29.2 pharynx 1988-98 -45.3 larynx 1988-98 71.8 Barshi females-Joinpoint analysis 1988-98 trend 1 trend 2 Cancer site: years APC years APC Head & Neck 1988-98 50 Tongue 1988-98 50.8 Other oral cavity 1988-98 29.2 pharynx 1988-98 -3.3 larynx 1988-98 53.8

* - AAPC=-8.8 last 10 observations

Fig.3.6. Trends in AAR of Head and Neck cancers and subsites – Chennai young

a b

c d

e f

g h

i j




Fig.3.7. Trends in AAR of Head and Neck cancers and subsites – Barshi young

a b

c d

e f

g h

i j




40

Overall there was a reduction in the AAR of head and neck cancers in both urban and

rural population with APC of –1 among urban males, –3.8 among urban females and –7.4

among rural males. However among rural females the AAR of Head and Neck squamous

cell carcinoma had increased during the study period.

Among males, there was a decrease in the trends of oral cavity cancers (oral cavity

cancers other than tongue) with APC of –2.3 among urban males and –5.5 among urban

females. Among rural males it decreased during the time periods 1988-91, 1991-94 and

increased during 1994-97 with the AAPC of –3.3. Among rural females there was an

increase with the APC of 5.5. With respect to tongue cancers, it was found that there was

a decrease among urban and rural males but an increase among urban and rural females.

There was a decrease in the incidence of pharyngeal and laryngeal cancers in the urban

and rural population except for an increase among rural females.

Analysis of the incidence trends in young urban population revealed that there was a

decrease in the incidence of all cancers except for an increase in tongue cancers among

females. Among rural males there was a decrease in the incidence of head and neck,

tongue and pharynx cancers and an increase in other oral cavity and larynx cancers.

Among rural females, there was an increase in the incidence of all cancers of interest

except pharynx cancers, which showed a decrease. The increased AAPC in Barshi might

be due to the lower number of cancer cases reported.

The probability estimates showed that among urban population, one out of every 27men

and one out of every 64 women would contract head and neck cancer in their life span

(Table 3.5). Estimation of the cumulative incidence rate percentage showed that 3.7% of

the urban male population is estimated to develop head and neck cancer in their life span;

out of this 0.2% belong to 20-44 years age group. Of the urban female population 1.6% is

estimated to develop head and neck cancer in their life span; out of this 0.13% belong to

the younger age group.

Among rural population, one out of every 54 men and one out of every 494 women would

contract head and neck cancers in their lifespan. Estimation of the cumulative incidence

rate percentage showed that 2% of the rural male population is estimated to develop head

and neck cancer in their life span; out of this 0.2 % belongs to 20-44 years age group. Of

the rural female population 0.3% is estimated to develop head and neck cancer in their

life span; out of this 0.1 % belong to the younger age group. The probability estimates of

the sub sites are also given in Table 3.5.

Table 3.5: Probability estimates of the sub sites in urban and rural populations

LTR CIRP

Urban Rural Urban Urban Rural Rural Site

M F M F M F M F

Head &

neck 27 64 54 494

4%

(3.433-

4.036)

1.6%

(1.444-

1.716)

2%

(1.565-

2.438)

0.34%

(0.203-

0.463)

Tongue 136 458 459 1250

.76%

(0.657-

0.864)

.23%

(0.199-

0.259)

0.33%

(0.188-

0.483)

0.07%

(0.007-

0.138)

Oral cavity 102 111 422 1068

1%

(0.902-

1.102)

.94%

(0.816-

1.065)

0.33%

(0.213-

0.522)

0.12%

(0.079-

0.162)

Larynx 150 138

6 -* -*

.69%

(0.611-

0.766)

.08%

(0.064-

.092)

0.24%

(0.210-

0.451)

0.003%

(-.012-

.066)

Pharynx 80 313 130 1304

1.27%

(1.154-

1.413)

.3%

(0.287-

0.386)

0.97%

(0.676-

1.261)

0.11%

(0.055-

0.168)

*- Not enough cases for reliable estimation

LTR-Life time risk; CIRP-Cumulative Incidence Rate Percent 95% Confidence Intervals of the difference

for CIRP, the lower and upper limits are given in brackets.

3.6.2 Objective 2

To compare the change in trends of head and neck squamous cell carcinoma of India and

US populations.

41

In this study the trends in the age adjusted rates of the head and neck cancers and its

subsites from Mumbai cancer registry were compared with that of SEER data. Figures 3.8

through 3.11 graphically display the results of the joinpoint analysis shown in Tables 3.6

and 3.7. The table below shows the best joinpoint fit for each site and the APC & AAPC

of the Mumbai and SEER registry data for all ages and younger age group respectively.


Rates of Mumbai and SEER database -all ages

42

ALL AGES Mumbai males

Joinpoint analysis 1973-97 trend 1 trend 2 Cancer site years APC years APC

Head & Neck 1973-97 -9.4 Tongue 1973-97 -14

Other oral cavity 1973-97 -0.6 pharynx 1973-97 -12.6 larynx 1973-97 -13.2

Mumbai females Joinpoint analysis 1973-97 trend 1 trend 2 Cancer site years APC years APC

Head & Neck 1973-97 -10.5

Tongue 1973-97 -13.8


pharynx 1973-97 -12.6

larynx 1973-97 -17.7 SEER males Joinpoint analysis 1973-97

trend 1 trend 2 Cancer site years APC years APC

Head & Neck 1973-97 3.0 1983-97 -7.8

Tongue 1973-97 6.1 1983-97 -0.8 Other oral cavity 1973-97 3.7 1983-97 -11.6

pharynx 1973-97 9.27 1983-97 -10.01

larynx 1973-97 0.3 1983-97 -11.2 SEER females Joinpoint analysis 1973-97

trend 1 trend 2 Cancer site years APC years APC

Head & Neck 1973-97 -2

Tongue 1973-97 1.49


pharynx 1973-97 -3.6

larynx 1973-97 1.5

AAPC= -2.5 last 10 observations AAPC= +2.6 last 10 observations AAPC= -4.3 last 10 observations AAPC= -0.8 last 10 observations AAPC= -5.6 last 10 observations

Fig.3.8. Trends in AAR of Head and Neck cancers and subsites – Mumbai all ages

a b

c d

e f

g h

i j




X‐ axis‐‐‐ Year: 1=1973‐77, 2=1978‐82, 3=1983‐87, 4=1988‐92, 5= 1993‐97

Fig.3.9. Trends in AAR of Head and Neck cancers and subsites – SEER all ages

a b

c d

e f

g h

i j




X‐ axis‐‐‐ Year: 1=1973‐77, 2=1978‐82, 3=1983‐87, 4=1988‐92, 5= 1993‐97


Rates of Mumbai and SEER database -young

Mumbai males Joinpoint analysis 1973-97

trend 1 trend 2 Cancer site: years APC years APC

Head & Neck 1973-97 -7.7 Tongue 1973-97 -11.7

Other oral cavity 1973-97 1.2

pharynx 1973-97 4.6 1983-97 -36 larynx 1973-97 -15.6

Mumbai females Joinpoint analysis 1973-97 trend 1 trend 2 Cancer site: years APC years APC

Head & Neck 1973-97 -7.39

Tongue 1973-97 3.89


pharynx 1973-97 -8.11

larynx 1973-97 -4.32 SEER males Joinpoint analysis 1973-97


Head & Neck 1973-97 16.7

Tongue 1973-97 45.7 1983-97 -1.8 Other oral cavity 1973-97 36.96

pharynx 1973-97 53.68

larynx 1973-97 -14.39 1983-97 -0.21 SEER females Joinpoint analysis 1973-97


Head & Neck 1973-97 2.3

Tongue 1973-97 16.9


pharynx 1973-97 1.63 1983-97 -98.93

larynx 1973-97 -11.1

AAPC = +19.6 last 10 observations

AAPC= -18.1 last 10 observations



Analysis of the Mumbai registry data showed that there was an overall decrease in the

incidence of head and neck cancers and the subsites during the study period. SEER data

analysis also showed that there was a decrease in the incidence of head and neck cancers

43

Fig.3.10. Trends in AAR of Head and Neck cancers and subsites – Mumbai young

a b

c d

e f

g h

i j




X‐ axis‐‐‐ Year: 1=1973‐77, 2=1978‐82, 3=1983‐87, 4=1988‐92, 5= 1993‐97

Fig.3.11. Trends in AAR of Head and Neck cancers and subsites – SEER young

a b

c d

e f

g h

i j




X‐ axis‐‐‐ Year: 1=1973‐77, 2=1978‐82, 3=1983‐87, 4=1988‐92, 5= 1993‐97

44

except for the increase in the incidence of tongue cancers both among males and females

and larynx cancers among females.

Analysis of the Mumbai data among the younger age group showed similar results as that

of all ages except for the increase in incidence of other oral cavity among males and

tongue cancers among females.

SEER data analysis among the younger age group showed that there was an increase in

the incidence of head and neck especially tongue cancers both among males and females.

Other oral cavity and pharynx cancers showed an increase among young males and a

decrease among young females. There was a decrease in the incidence of laryngeal

cancers.

Proportion:

The SEER data showed that the Head and Neck Cancers formed around 6% (1977-82) to

4% (1993-97) of all cancers among males and around 2% (1977-82) to 1.7% (1993-97)

among females, whereas in Mumbai, they formed around 31% (1977-82) to 25% (1993-

97) of all cancers among males and 12% (1977-82) to 9% (1993-97) among females.

Cumulative rate and risk

Mumbai data showed that one out of every 16 men and one out of every 48 women would

contract head and neck cancers in their life span, whereas SEER data showed that one out

of every 25 men and one out of every 83 women would contract head and neck cancers in

their life span.

3.6.3 Objective 3

To evaluate etiologic basis for the change in trends (if any) of head and neck squamous

cell carcinoma.

The risk factors and cancer sub-site information was extracted from the raw data and the

information has been summarised in Table 3.8, and then used in the statistical analysis

(Table 3.9)

A total of 611 head and neck cancer patients (males=473-77%, females=138-33%)

reported at our hospital during the study period. Forty nine percent patients presented

with risk habits (n=298, males=93%, females=7%). Patients with cancers of tongue, oral

cavity, oropharynx, hypopharynx and larynx were included in the study. One hundred and

45

twenty nine patients with other head and neck cancers were excluded from the study.

Tongue cancer made up 151 patients of our study group, i.e. 25% of the patients. The

second most prevalent cancer at this institute within the last few years was cancer of the

glottis (82 patients), making up 13% of head and neck cancers. The raw data showed

quite a high proportion of tongue cancer patients who presented without any risk factors

(48%).

Table 3.8: Summarised risk factors vs. generalized cancer sub-sites

Risk Factors

Sub-sites

Sub-site

Proportion Presence % Absence %

Tongue 31% 78

(m=75, f=3) 52

73

(m=51, f=22) 48

Other Oral

Cavity 28%

73

(m=58, f=15) 54

62

(m=38, f=24) 46

Oropharyn

x 8%

29

(m=27, f=2) 78

8

(m=6, f=2) 22

Hypophary

nx 9%

28

(m=26, f=2) 62

17

(m=7, f=10) 38

Larynx 24% 68

(m=68, f=0) 60

46

(m=42, f=4) 40

Total 100% 276

(m=254, f=22)

57 206

(m=144, f=62)

43

Excluded

data

22

(m=22, f=0)

107

(m=53, f=54)

m-males, f-females

Chi-square analysis of tongue cancers against cancers of other subsites were analysed

using the Epi-info software. However, following the statistical analysis of tongue cancers

against cancers of the remainder of the oral cavity (135 patients), a P value of 0.77 was

obtained. This tells us that the number of tongue cancers presenting without any risk

factors, when compared to the rest of the oral cavity, is not of significance. Statistical

analysis of tongue cancers against the hypopharnyx (45 patients) and larynx (114

patients) also showed insignificance. However, a P value of 0.0058 was obtained when

analysing tongue cancer against the oropharnygeal cancers (37 patients). This tells us that

the prevalence of tongue cancer patients without risk factors is of a significant value,

when compared to oropharyngeal cancers.

Table.3.9: Statistical analysis results of tongue cancers against other subsites

46

Tongue vs Odds ratio P value

Oral cavity 0.91 0.77

Oropharynx 0.29 0.0058

Hypopharynx 0.65 0.28

Larynx 0.72 0.24

3.7 Discussion

In this study it is shown that that there is an overall reduction in the incidence of head and

neck cancers.

This is consistent with the previous studies published on head and neck cancer trends. A

slow and steady decline in the trend of head and neck cancers is seen in India (Yeole

2003 ). A study on the data from Mumbai cancer registry for 15 years from 1986-2000 on

time trends in the incidence and the risk of developing oral cancer has shown a decreasing

trend which was significant among males (Sunny 2004). In our study II it is shown that

the head and neck cancers showed a decline in the trend and accounted for 25% of male

cancers and 12% of female cancers in Mumbai. The overall incidence of head and neck

cancers was found to be stable among Americans during the time period 1973-97

(Schantz 2002). This reduction in head and neck cancer trends was not observed in young

47

Americans and rural Indian females. The increased incidence observed in rural females

might be due to very few cases reported there.

What is alarming is a consistent increase in the incidence of tongue cancers observed in

females- Indian urban, rural & SEER in both all age groups and younger age groups, in

young Mumbai females and in SEER males. Such increase was not found among Indian

males.

Oral cavity cancers showed a decline in the incidence in the study groups except for an

increase among rural females and young rural, Mumbai and SEER males. And there is an

overall reduction in the incidence of laryngeal cancers in all the study groups. This was

not observed among rural females and young rural males and young Americans, both

males and females as far as laryngeal cancers are concerned. With respect to pharyngeal

cancers there is a uniform reduction in the incidence except for an increase shown by

young SEER males.

The observed trends are a "tip-of-the-iceberg" phenomenon. Several factors like

prevalence of risk habits, awareness, screening, diagnosis and completeness of reporting

can affect the reported incidence of cancer.

There are strong epidemiological evidences to link tobacco habits with the occurrence of

head and neck cancers. A wide variety of tobacco habits like smoking, chewing, snuffing,

using burnt tobacco as powder or paste are prevalent in India, which is more so in the

rural population than in their urban counterparts (National Sample Survey Organization

1998). Among them, bidi smoking, tobacco chewing and cigarette smoking, in that order,

account for a large majority of head and neck cancers (Sanghvi 1989). Smoking is most

common form of tobacco consumption among males and chewing among females

(National Sample Survey Organization NSS 50th Round(1993-1994) 1998). Tobacco

related cancers (TRC) account for nearly 48.2% of all cancers in Indian men and 20.1%

in women (Gajalakshmi 1996). Other than tobacco, alcohol abuse and viral infection have

been implicated in head and neck carcinogenesis. In India, National Sample Survey

Organization (NSSO) collects data on the prevalence of tobacco use. The data on tobacco

consumption over the periods 1987-88 to 1993-94 has shown that it has decreased in both

urban and rural males and females. Contrary to the popular belief that the tobacco

consumption is increasing, this data shows that it has decreased in all sectors. According

to the National Health Interview Survey the prevalence of cigarette smoking and alcohol

48

use among Americans have markedly decreased (US Department of Health and Human

Services 1994; US Department of Health and Human Services 1998; The Gallup

Organization 2001). Possibly the decreased consumption could account for the decrease

in head and neck cancers, especially that of oral cavity. Though there is a decreasing

trend, the incidence of oral cavity cancers in India, is still one of the highest in the world

(Franceschi 2000). One of the possible reasons could be the easy availability of tobacco

products. Women appear to have a higher prevalence of chewing in many rural areas due

to the beliefs that tobacco has many magical and medicinal properties; keeping the mouth

clean, getting rid of a foul smell, curing toothache, controlling morning sickness, and

minimizing labor pains.

The sub-site predilection table shows that the reduction in incidence is obvious in tumors

where tobacco is the primary risk factor such as oral cavity (chewing of tobacco) and

larynx (smoking). Similar trends were observed in other parts of India (Coleman 1993;

Sunny 2004). The decreasing trend may be attributable to a decrease in the usage of pan

and tobacco. The increased incidence of oral cavity cancers in rural females and young

rural and Mumbai males might be attributed to the common habit of chewing in rural

population especially females and young Mumbai males. There are many reports from

Mumbai stating the increased prevalence of chewing among young males. In a study on

use of tobacco products among students and school personnel in India, it was found to be

14.1% among students-17.3% boys and 9.8% girls (Sinha 2007).

There are reports from Europe and United State of America to suggest that head and neck

cancers are increasing among young, particularly in tongue (Schantz 2002; Llewellyn

2004). Young Americans are diagnosed with increased rates of oral and tongue cancers

attributed to the use of smokeless tobacco and marijuana. This finding was not observed

in our study though there was an increased incidence of tongue cancers among both urban

and rural females as well as in young females in Mumbai.

There are not many reports from India about the awareness of these cancers among the

general public. Few reports from developed countries have shown an alarming lack of

awareness of oral cancer among the public (Warnakulasuriya 1999). In India

(unpublished data), there seems to be increased awareness of oral cancer and its factors,

which is discussed elaborately in chapter 5.

49

There is increased incidence of cancers in rural population especially females

emphasizing the need to target that population for health education. The increased

incidence might be due to the very few number of cases reported there. Indian data

showed an increase of tongue cancers among females and oral cavity cancers among

Mumbai males.

And in our study III, it was found that 48% of tongue cancers did not have any

established risk factors. It is fair to say from the statistical analysis that the number of

tongue cancer patients presenting without risk factors are not of a significant level,

especially when compared to the remainder of the oral cavity except oropharynx. But the

raw results of tongue cancer alone show quite a striking number of cases (48%) that

presented to the hospital without any of the known associated risk factors. As for the

possible aetiological factors for tongue cancers, firstly it should be reminded that 52% of

tongue cancer patients did indeed have one or more of the known associated risk factors.

Out of these, 62 were smokers, and from those 29 also drank alcohol. This suggests a

strong link between the risk factors of smoking and alcohol with tongue cancer. Alcohol

on its own seems not be a strong contributory factor, but its synergistic affect with

smoking is now widely accepted. Only 8 tongue cancer patients had the habit of chewing.

Paan chewing is generally associated with buccal mucosa cancers. Our results also show a

strong association between paan chewing as a habit and cancer of the buccal mucosa,

with 18 out of the 35 patients having the habit. This can be explained by the fact that the

paan is held in the buccal sulcus for an extensive period of time, so allowing the

carcinogens to act on the buccal mucosa more so than other regions of the oral cavity.

For the remaining 48% of the tongue cancer patients, who did not present with any risk

factors, alternative aetiological factors, if any, need to be proposed. This may suggest that

risk factor for these cancers may be different from that of other head and neck sub-sites.

Diverse risk factors other than tobacco, which could reflect changes in diet, nutritional

habits, food processing, sexual habits, need to be investigated for this subsite. In a study

by Shiboski et al in 2005, a paralleled increasing trend in tonsillar and tongue cancers was

found, against a general decline of other oral cavity cancers. Knowing that the association

between human papillomavirus (HPV) infection and tonsillar carcinoma is now well

established, this led to the suggestion that there may be a link between HPV and tongue

cancer (Shiboski 2005). However, a more recent study by Pintos et al concluded that this

aetiologic link is less clear for non-tonsillar oral cancers (Pintos 2008).

50

Diet, in terms of cancer prevention, has much ongoing research. The general consensus is

that the anti-oxidants available in fresh fruits and vegetables play a significant role in

protecting against cancer. These include vitamin C, vitamin E, carotenoids, and many

other phytochemicals. Little has been suggested regarding any particular dietary

ingredients that may be cancer causing. Those ingredients that have been suggested are

linked to cancers elsewhere in the body, for example high doses of soya and the increased

risk of estrogen-responsive cancers, such as breast or endometrial cancer. Any dietary

factors (other than paan, tobacco and alcohol) that may be associated with oral cavity

cancers are still to be found.

This was not a case-control study, but instead a study in which one cancer was compared

to a group of others. Carrying out this study with a control group of patients with the

same risk factors but not suffering from cancer may have resulted in a p value that would

have suggested a significant relationship. Hence, this study can be improved upon if it

were to be conducted prospectively, looking into the detailed dietary history of all

patients who present with oral cavity cancers, in an attempt to find a common ingredient

that could be proposed as a possible alternative aetiological factor in tongue cancer. Such

a study can then be expanded upon by investigating treatment outcomes and follow-up

details of each patient to find out whether there is any significant difference in the

prognosis of tongue cancer patients with and without risk factors.

3.8 Conclusion

A trend is emerging showing that the type of oral cancers that patients present with are

changing, with a definite increase in the number of patients presenting with tongue

cancer. We are also seeing that a considerable number of these patients are presenting

without the known associated risk factors. However, what might be causing these cancers

is still to be proven, with HPV and dietary factors being at the forefront of alternative

aetiological factors.

Documents

EPIDEMIOLOGIC TRENDS OF HEAD AND NECK CANCERSshodhganga.inflibnet.ac.in/bitstream/10603/2363/12/12_chapter 3.pdf · tobacco smoking and heavy alcohol consumption, with low intake