Acta Oncologica Vol.

30 No. 5 1991

FROM THE EPIDEMIOLOGY UNIT, CANCER INSTITUTE ( WIA), MADRAS, INDIA.

Acta Oncol Downloaded from informahealthcare.com by University of South Carolina on 08/01/13 For personal use only.

RISK FACTORS FOR FEMALE BREAST CANCER A hospital-based case-control study in Madras, India C. K. GAJALAKSHMI and V. SHANTA

Abstract Five hundred and thirty-one histologically confirmed breast carcinoma cases examined from 1983 to 1986 inclusive at the Cancer Institute (WIA), Madras, India, were matched for age, socioeconomic class and menopausal status with an equal number of controls. Cancer patients without diseases in breast, gynaecological organs or endocrine glands were used as controls. Risk factors for breast cancer were analyzed separately in the premenopausal and the postmenopausal groups. In neither group was there significant association between age at menarche and breast cancer risk. Single women had higher risk than married women. Nulliparity was found to be a risk factor in premenopausal women only. The relative risk increased with age at marriage and age at first birth. A three-fold risk was voted in both pre- and postmenopausal groups when the interval between age at first birth and menarche was more than 12 years and also in women who attained menopause between the age of 44-49. Key words: Breast cancer, risk factors, casecontrol study.

Cancer of the breast is the second most common cancer among females in Madras and Bangalore and the most common female cancer in Bombay, India. Incidence rates for breast cancer for whole of India are not available but valuable information can be derived from the three population-based cancer registries in Bombay, Bangalore and Madras. The incidence rates (age adjusted for world population) for Bombay, Bangalore and Madras are 25.6, 15.8 and 20.1 respectively per 100 000 female population in 1985 ( I ) . These figures are similar to that seen in Singapore Indians (28.2), but considerably lower than those reported from Western countries (2). Breast cancer constituted between 9.9-21.0% of all female cancers in different hospital-based cancer registries (in the National Cancer Registry project) in India with the lowest figure in Dibrugarh (9.9'%,)and the highest in Bombay (21.0%) ( I ) . The levelling off of breast cancer rates around menopause was first noted by Clemmesen in 1948 (3). He

reported that the age specific incidence curve for breast cancer in many countries rose till the age of 45, then levelled off up to 55 and thereafter increased again. DeWaard et al. (4) advanced a hypothesis on the existence of two types of breast cancer with different aetiology. According to this hypothesis ovarian oestrogenic activity is of aetiological importance in premenopausal group, whereas oestrogen of adrenocortical origin is important in postmenopausal group. Anderson (5) reported that the risk increased in first-degree relatives of premenopausal breast cancer patients but not in relatives of postmenopausal breast cancer patients, suggesting a genetic difference between premenopausal and postmenopausal breast cancer. Craig et al. (6) found that a positive family history of breast cancer and a first child birth at age twenty-five or older were significantly associated with early onset cases (before age 45) while in late onset cases (age 45+), lactation was associated with increased risk. Two recent studies revealed that lactation had protective effect against breast cancer in premenopausal women only (7, 8). These studies support the view that premenopausal women differ from postmenopausal women concerning risk factors for breast cancer. In view of the above observed differences between premenopausal and postmenopausal breast cancer, we have analyzed the risk factors in the two groups separately. Material and Methods The hospital-based cancer registry of the Cancer Institute (WIA), Madras, India, has existed since 1955. Its -

Presented at Indian Society of Oncology Conference in Bangdlore. India. in February 1988. Submitted 28 February 1989. Accepted for publication 21 August 1990. 569

570

C. K . GAJAL.AKSHMI A N 0 V . SHANTA

Table 1 Distribution ?f cases and controls hy socioeconomic class Cases

Premenopausal group Class 1 I1 111 IV V

Acta Oncol Downloaded from informahealthcare.com by University of South Carolina on 08/01/13 For personal use only.

Total Postmenopausal group Class I I1 111 IV V Total

Controls

n

'%

n

1%)

163 31 22 8

68.5 13.0 9.2 3.4 5.9

166 28 25 6 13

69.7 11.8 10.5 2.5 5.5

14

238

199 49 20 12 13 293

238

61.9 16.7 6.8 4. I 4.4

206 44 18

13 12

70.3 15.0 6.1 4.4 4.1

293

main task is to maintain detailed information of patients admitted for treatment and of their follow-up. Data on demographic factors, women's reproductive history and other suspected risk factors necessary for the analysis were abstracted from the medical records. We have reviewed 53 1 histologically confirmed new breast carcinoma cases covering the four-year period from 1983 to 1986, and compared them with 531 hospital controls who had no disease in breasts, gynaecological organs or endocrine glands. Cases were matched with controls for age, socioeconomic class and menopausal status. The age at diagnosis was taken as the current age. The controls were matched to the age of respective cases &- 1 year. Forty-five per cent of cases and controls belonged to the premenopausal group (238 cases and 238 controls) and 55% to the postmenopausal group (293 cases and 293 controls). Table 1 shows the distribution of cases and controls by socioeconomic class. Age at menarche, marital status, age at marriage, age at first birth, interval between menarche and age at first birth, number of children and age at menopause were the risk factors analyzed separately in premenopausal and postmenopausal groups. For each variable the difference between cases and controls was tested by applying xz-test on contingency tables and when appropriate by the x2-test on trend (9). For odds ratio (OR), the 95% confidence intervals (CI) were computed (10). Results Out of 53 I histologically confirmed breast carcinoma cases examined from 1983 to 1986 at the Cancer Institute

(WIA), 238 were premenopausal and 293 postmenopausal. The distribution of risk factors in premenopausal group is shown in Table 2. The average age at menarche was similar in cases ( 13.7 years) and controls (13.8 years). There was no significant difference noted in the distribution of age at menarche nor in linear trend ( x 2 for trend = I . I). Single women had 5.3 times higher risk of developing breast cancer than married women. Nulliparity was found to be a risk factor with an OR of 2.9 (1.6-5.2)compared to parous women. There was a difference in the average age at marriage for cases (17.8 years) and controls (16.7years). As the age at marriage increased, the relative risk also increased steadily and reached a maximum OR of 2.4, when the age at marriage was more than 20 years and also showed a significant linear trend ( x 2 for trend = 10.2). In parous women, the relative risk increased with increase in age a t first birth. With a reference level of age at first birth below 17 years, age at first birth between 23 and 25 years gave an OR of 2.1 (1.0-4.1). age at first birth more than 25 years an OR of 2.6 (1.2-5.3)and nulliparous women an OR of 4.6 (2.3-9.4). A significant linear trend was observed ( x 2 for trend = 8.4). When the difference between age at first birth and menarche increased, the relative risk also increased and reached an OR of 3.1 (1.4-6.7) when the interval was more than 12 years. The linear trend was statistically significant ( x 2 for trend = 13.2). Among parous women, women with more than four children was used as reference level. We found that women with one or two children had an OR of 1.8 (1.1-3.1) which was less than the risk seen in nulliparous women (OR = 4.3).Women with three to four children had an OR of 1.5 which was not statistically significant whereas a significant linear trend was noted ( x 2 for trend = 4.7). Distribution of risk factors in postmenopausal cases and controls is given in Table 3. Average age at menarche of cases (13.9 years) and controls (13.7years) were very similar. Single women had 5.6 times higher relative risk of breast cancer than married women. Nulliparous women had an O R of 1.2compared to parous women which was not significantly different from unity. The average age at marriage of cases (16.8 years) and controls ( 15 years) differed significantly. O R was 3.8 (2.16.8) when the age at marriage was more than 20 years compared to an age at marriage of less than 15 years. The relative risk increased slightly with age at first birth until 25 years (less than 2-fold). With a reference level of age at first birth below 17 years, women with age at first birth above 25 years had an OR of 5.6 (2.5-12.5) and nulliparous women an OR of 1.9 (1.1-3.4). The interval between age a t first birth and menarche greater than 12 years gave an O R of 3.3 (1.5-7.3) compared to an interval less than 4 years.

57 1

RISK FACTORS FOR BREAST CANCER

Table 2 Distribution of risk jactors in prewienopuusul group

Factors

Cases n

Controls n

All

238

238

120 89 29 13.7 yrs 223

OR

95%) CI

Ill 91 36 13.8 yrs)

I* 0.9 0.7

0.6- 1.3 0.4- 1.3

15

235 3

I# 5.3*

1.7-16.3

197 41

222 16

I# 2.9*

1.6-5.2

41 75 61 46 17.8 yrs

65 88 52 30 16.7 yrs)

If 1.3 1.9* 2.4*

0.8-2.2 1.1-3.2 1.3-4.4

Trend (1.d.f.)

Age at menarche ( yrs)

< 14 14 & 15 > 15

Acta Oncol Downloaded from informahealthcare.com by University of South Carolina on 08/01/13 For personal use only.

( Mean

Marital status Married Single Nulliparity No Yes Age at marriage ( yrs) < 15 15-17 18-20 > 20 ( Mean Age at first birth (yrs) < I7 17-19 20-22 23-25 > 25 Nulliparous Dif. between first birth and menarche (yrs) 12 Number of children Nulliparous 1-2 3 -4 r 4

I* 1.4

30 53 57 30 27 41

54 67 56 26 19 16

2.1* 2.6* 4.6*

0.8-2.5 1 .O-3.3 I .O-4.1 1.2-5.3 2.3-9.4

43 60 42 28 24

66 90 32 22 12

I f I .o 2.0* 2.0 3.1*

0.6-1.7 1.1-3.7 1.0-3.8 1.4-6.7

41 85 79 33

16 78 89 55

4.3* I .8* 1.5

2.1-8.6 1.1-3.1 0.9-2.5

OR = Odds ratio; * = Significant at p < 0.05;

#

1.8*

10.2*

8.4* ( Excluding

nulliparous)

13.2*

4.7* ( Excluding

nulliparous)

1*

= Ref.

Nulliparous women had an OR of I .5 (not significant), women with less than three children an OR of 1.8 (1.22.7) and women with three or four children an OR of 1.2 (not significant) when women with more than four children were used as reference group. The average age at menopause was 45.8 years and 44.5 years in the study group and the control group respectively. Breast cancer patients attained menopause on average a year later than controls. About 3-fold relative risk was estimated when the age at menopause ws 44-46 and 47-49 years compared to less than 41 years. The following risk factors showed statistically significant linear trends: age at marriage, age at first birth, interval

1.1

group

between age at first birth and menarche, number of children and age at menopause.

Discussion Lilienfeld ( 1 1) reported excess breast cancer incidence among single women after 35 t o 40 years of age. In the present study, single women in the postmenopausal group showed a higher risk in relation to married women than those in the premenopausal group. In both menopausal groups the breast cancer risk increased with age at marriage, and the association was stronger in the postmenopausal group. The increased risk

512

C.

K. GAJALAKSHMI AND V. SHANTA

Table 3 Distribution of risk factors in postmenopausal group

Acta Oncol Downloaded from informahealthcare.com by University of South Carolina on 08/01/13 For personal use only.

~

Factors

Cases n

Controls n

All Age at menarche (yrs) < I4 14 & 15 > I5 (Mean Marital status Married Single Nulliparit y No Yes Age at marriage (yrs) < I5 15-17 18-20 > 20 (Mean Age at first birth (yrs) < I7 17-19 20 - 22 23-25 > 25 Nulliparous Diff. between first birth and menarche (yrs) 12 Number of children Nulliparous 1-2 3 -4 >4 Age at menopause ( yrs) 49 (Mean

293

293

138 I09

154

OR = Odds ratio;

100

~

~~

OR

95% CI

Trend ( 1.d.f.)

I# 1.2 1.3

0.8- 1.7 0.8-2.1

1.7

13.9 yrs

39 13.7 yrs)

277 16

290 3

1# 5.6*

1.8-17.0

252 41

259 34

I# 1.2

0.8-2.0

75 99

146 84

64

40

I# 2.3* 3.1* 3.8*

1.5-3.4 1.9-5.0 2.1 -6.8

46

39 16.8 yrs

20 15.0 yrs)

49 85 61 29 28 41

78 85 56 32 8 34

If+ I .6* 1.7* 1.4 5.6* 1.9*

1.0-2.5 1.0-2.9 0.8-2.7 2.5-12.5 1.1-3.4

80 79 44 25 24

99 88 46 17 9

1f+ 1.1 1.2 1.8 3.3'

0.7-1.7 0.7-2.0 0.9-3.6 1.5-7.3

41 86 75 91

34 61 82 1 I6

1.5 1.8* 1.2 I*

88 26 52 56 71 44.5 yrs)

I* 1.6 2.8' 2.9* 1.4

48 23 79 89 54 45.8 yrs

* = Significant at p < 0.05;

f+

= Ref.

with increasing age at first birth in both menopausal groups as noted in the present study is consistent with the findings of Paffenbarger et al. (12) and Lipnick et al. (13). However, Stavraky et al. (14) found increased risk only in postmenopausal women and Wynder et al. ( 15) in pre- and perimenopausal women. Other studies wherein menopausal status had not been considered have also shown increased risk with increasing age at first birth (16-20).

0.9-2.6 1.2-2.7 0.8-1.8

0.8-3.1 1.7-4.5 1.8-4.7 0.8-2.3

31.6*

11.5*

(Excluding nulliparous)

8.8*

7.0* (Excluding nulliparous)

5.5'

group

From previous studies it has often been reported that the risk is increased when the age at first birth is above 35 years. In our material higher risk was noted when the age at first birth was more than 25. The risk of breast cancer in nulliparous women was higher than in women whose age at first birth was more than 25 years in the premenopausal group whereas the reverse was found in the postmenopausal group. Wynder et al. (21) indicated that

573

Acta Oncol Downloaded from informahealthcare.com by University of South Carolina on 08/01/13 For personal use only.

RISK FACTORS FOR BREAST CANCER

the breast cancer patients tend to marry late and consequently have their pregnancy later. In India, compared to Western countries, women tend to marry early (within 2-3 years after menarche) which naturally leads to an early age at first birth and more children. MacMahon et al. (16) have suggested that a dual effect of the first pregnancy explains the higher risk seen in women with a late first birth. Older women are more likely to have cells with malignant transformation in the breast at the time of pregnancy. Hormonal stimulation of the breast tissue during pregnancy can often promote the transformed cells. The protective effect of early full-term pregnancy is probably related to prevention of tumour initiation. Our data showed an inverse relationship between parity and breast cancer risk as expected from previous reports. MacMahon et al. (17) have suggested that the inverse relationship between parity and breast cancer risk is mainly due to the reduction in breast cancer risk associated with early age at first birth. However, it has also been reported that parity has protective effect on breast cancer independent of age at first birth (13,20,22-28). In the premenopausal group in our study, nulliparous women had a higher risk relative to women with more than four children and a 3-fold risk relative to all parous women. In the postmenopausal group, on the contrary, the risk in nulliparous women was not significantly different from unity when compared to women with more than four children or all parous women. Nulliparity was found to be a risk factor only in premenopausal women in our data. Wynder et al. (15) found that nulliparity was a risk factor only perimenopausally and Rosero-Bixby et al. (28) observed it to be so in women younger than 45 years of age. A long interval between age at menarche and age at first pregnancy increased the risk in both menopausal groups (14). A 3-fold risk of developing breast cancer was noted in our study in both menopausal groups when the interval between age at first birth and menarche was more than 12 years relative to an interval less than 4 years. The risk of breast cancer increases with increasing age at menopause (12- 14,20,29-31). Trichopolous et al. (32) found that women whose menopause occurred before the age of 45 had only one-half of the breast cancer risk of those whose menopause occurred after the age of 55. In our study, a 3-fold risk was found in women whose menopause occurred between 44-49 years compared to women who attained menopause before the age of 41. Recent investigators have reported similarity of most risk factors of breast cancer in both pre- and postmenopausal women (12, 13.33). Our study also showed that most of the risk factors analyzed had similar trend in both pre- and postmenopausal women. However, nulliparity was found to be a risk factor in the premenopausal group only.

Corresponding author: Dr. C. K. Gajalakshmi, Epidemiology Unit, Cancer Institute (WIA), Madras 600 020, Tamilnadu. India.

REFERENCES I . (ANON 1985) Annual Report of the National Cancer Registry. A Project of Indian Council of Medical Research, New Delhi, India, 1985. 2. Water House J, Muir C, Shanmugarathnam K, Powell J, eds. Cancer incidence in five continents, vol. IV, Lyon: IARC, 1982. 3. Clemmesen J. Carcinoma of the breast: Results from statistical research. Br J Radio1 1948; 21: 583-90. 4. DeWaard F, Baanders-Van Halewijn EA, Huizinga J. The bimodal age distribution of patients with mammary carcinoma. Cancer 1964; 17: 141 -52. 5. Anderson DE. A genetic study of human breast cancer. JNCl 1972; 48: 1029-34. 6. Craig T, Camstock GW, Geiser GB. Epidemiologic comparison of breast cancer patients with early and late onset of malignancy and general population controls. JNCl 1974; 53: 1577-81. 7. Byers T, Graham S, Rzepka, Marshall J. Lactation and breast cancer. Am J Epidemiol 1985; 121: 664-74. 8 McTiernan A, Thomas DB. Evidence for a protective effect of lactation on risk of breast cancer in young women. Am J Epidemiol 1986; 124: 353-8. 9 Armitage P. Test for linear trends in proportions and frequencies. Biometrics 1955; 1 1 : 375-86. 10. Breslow NE, Day NE. Statistical methods in cancer research, Vol. 1-The analysis of case control studies. Lyon: IARC (Scientific Publication No. 32). 1980. 1 I . Lilienfeld AM. The relationship of cancer of female breast to artificial menopauseand marital status. Cancer 1956; 9: 927-34. 12. Paffenbarger RS, Kampert JB. Chang HG. Characteristics that predict risk of breast cancer before and after the menopause. Am J Epidemiol 1980; 112: 258-68. 13. Lipnick R, Speizer FE, Bain C, et al. A case control study of risk indicators among women with premenopausal and early postmenopausal breast cancer. Cancer 1984; 53: 1020- 4. 14. Stavraky K, Emmons S. Breast cancer in premenopausal and postmenopausal women. JNCI 1974; 53: 647-54. 15. Wynder EL, MacCornack FA, Stellman SD. The epidemiology of breast cancer in 785 United States Caucasian women. Cancer 1978; 41: 2341 -54. 16. MacMahon B, Cole P, Brown J. Etiology of human breast cancer: A review: JNCl 1973; 50: 21 -42. 17. MacMahon B, Cole P, Lin TM, et al. Age at first birth and breast cancer risk. Bull WHO 1970; 43: 209-21. 18. Ravnihar B, MacMahon B, Lindtner J. Epidemiologic features of breast cancer in Slovenia, 1965-1967. Eur J Cancer 1971; 7: 295-306. 19. Valaoras VG, MacMahon B, Trichopoulos, Polychronopoulou A. Lactation and reproductive histories of breast cancer patients in greater Athens. Int J Cancer 1969; 4: 350-63. 20. Helmrich SP, Shapiro S, Rosenberg L, et al. Risk Factors for breast cancer. Am J Epidemiol 1983; 117: 35-45. 21. Wynder EL, Bross IJ, Hirayama T. A study of the epidemiology of cancer of the breast. Cancer 1960; 13: 559-601. 22. Lilienfeld AM, Coombs J, Bross IDJ. Chamberlain A. Marital and reproductive experience in a community wide epidemiological study of breast cancer. Johns Hopkins Med J 1975; 136: 157-62.

514

C. K . GAJALAKSHMI AND V. SHANTA

Acta Oncol Downloaded from informahealthcare.com by University of South Carolina on 08/01/13 For personal use only.

23. Soini I. Risk factors of breast cancer in Finland. Int J Epidemiol 1977; 6: 365-73. 24. Tulinius H, Day NE, Johannesson G, Bjarnason 0, Gonzales M. Reproductive factors and risk for breast cancer in Iceland. Int J Cancer 1978; 21: 724-30. 25. Thein-Hlang, Thein-Maung-Myint. Risk factors for breast cancer in Burma. Int J Cancer 1978; 21: 432-7. 26. Adami HO, Hansen J, Jung B, Rimsten AJ. Age at first birth, parity and risk of breast cancer in a Swedish population. Br J Cancer 1980; 42: 651-8. 27. Layde PM, Webster LA, Baughman AL, Wingo PA, Rubin GL, Ory HW. Reproductive history and breast cancer risk. Am J Epidemiol 1986; 124: 516. 28. Rosero-Bixby L, Oberle MW, Lee NC. Reproductive history

29.

30. 31. 32. 33.

and breast cancer in a population of high fertility, Costa Rica, 1984-1985. Int J Cancer 1987; 40: 747-54. Brinton LA, Williams RR, Hoover RN, Stegens NL, Feinlab M, Fraumeni JF. Breast cancer risk factors among screening program participants. JNCI 1979; 62: 37-44. Kvale G, Heuch I. Menstrual factors and breast cancer risk. Cancer 1988; 62: 1625-31. Hiatt RA, Fireman BH. Smoking, menopause and breast cancer. JNCI 1986; 76: 833-8. Trichopolous D, MacMahon B, Cole P. Menopause and breast cancer risk. JNCI 1972; 48: 605-13. Frank1 G. Risk Factors in breast cancer: Are they important, are they the same in pre- and postmenopausal breast cancer patients? Oncology 1980; 37: 41-5.

Risk factors for female breast cancer. A hospital-based case-control study in Madras, India.

Five hundred and thirty-one histologically confirmed breast carcinoma cases examined from 1983 to 1986 inclusive at the Cancer Institute (WIA), Madras...
409KB Sizes 0 Downloads 0 Views