Abstract
Objectives
We aimed to evaluate the relation between menstrual and reproductive factors, exogenous hormones, and risk of pancreatic cancer (PC).
Methods
Eleven case-control studies within the International Pancreatic Cancer Case-Control Consortium (PanC4) took part in the present study, including in total 2,838 case and 4,748 control women. Pooled estimates of odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated using a two-step logistic regression model and adjusting for relevant covariates.
Results
An inverse OR was observed in women who reported having had hysterectomy (ORyesvs.no 0.78, 95%CI 0.67–0.91), remaining significant in post-menopausal women and never-smoking women, adjusted for potential PC confounders. A mutually-adjusted model with the joint effect for hormone replace therapy (HRT) and hysterectomy showed significant inverse associations with PC in women who reported having had hysterectomy with HRT use (OR 0.64, 95%CI 0.48–0.84).
Conclusion
Our large pooled analysis suggests that women who have had a hysterectomy may have reduced risk of PC. However, we cannot rule out that the reduced risk could be due to factors or indications for having had a hysterectomy. Further investigation of risk according to HRT use and reason for hysterectomy may be necessary.
Keywords: Pancreatic cancer, menstrual and reproductive factors, exogenous hormones, hysterectomy, consortium
INTRODUCTION
Pancreatic cancer is the twelfth most common cancer in the world1, but has among the poorest survival of all cancers. The aggressive nature of the disease and the lack of early markers or effective treatment options results in the lowest 5-year survival rate (3–7%) of all cancers in the US2,3. About 95% of pancreatic cancers are ductal adenocarcinomas (PC). Tobacco smoking is the main risk factor for PC and explains about 20% of the risk in a population where prevalence of smoking is 30%4,5. ABO non-O blood group, obesity, long-term type 2 diabetes, family history of PC, histories of pancreatitis and possibly heavy alcohol consumption, familial rare inherited mutations in BRCA2, p16 and other genes, and common variants in at least eight genetic loci are other known PC risk factors6–8.
PC incidence is somewhat higher in men than in women. In the US, between 2007 and 2011, the sex ratio ranged from 1.3 (for ages 40–44), to 1.1 (for ages 85 and over)9. In Europe, the estimated sex ratio in 2012 is highest at 1.9, for ages 40–44, and is 1.1 at ages 75 and over10. Some studies, using castrated rats, showed that administration of sex steroids inhibits the development and growth of preneoplastic lesions of the pancreas11,12. Motivated by these observations, and under the hypothesis that greater exposure to female sex hormones (through early menarche, later menopause, high number of pregnancies, and having a history of hormone use) decreases the risk of PC, several epidemiological studies have examined possible risk associations with menstrual and reproductive factors, and hormone use, but with inconsistent results. A review paper on reproductive factors and PC13, two meta-analyses on parity14,15, and a recent meta-analysis16 attempted to make clear the relations between these factors and PC risk. Comparing and summarizing previous evidence, however, is not a simple task. Inconsistencies in results may arise from different categorization and reference categories of exposure variables, different adjustment or confounding variables, and various study designs and target populations. Further, many of the previous studies were limited by small numbers of cases, and some were limited by inadequate adjustment for smoking, the primary risk factor for pancreatic cancer.
The aim of the present study was to assess whether or not menstrual or reproductive factors or hormone use are associated with risk of developing PC. Pooled individual analyses of eleven case-control studies in the PanC4 consortium allowed us to obtain precise estimates of risks and to analyze the associations in detail.
MATERIALS AND METHODS
Studies
Eleven case-control studies with information on menstrual and reproductive factors and hormone use were available within the PanC4 consortium17–26. At a minimum, the studies were able to provide information on age at menarche and age at menopause. The total number of women available in the combined data set were 2,838 with pancreatic cancer and 4,748 controls. The Surveillance of Environmental Aspects Related to Cancer in Humans (SEARCH), Toronto and Shanghai-I studies included proxy responders, accounting for 14.5% of the case and 4.4% of the control women (Table 1).
Table 1.
Studies and core variables by case and control status in PanC4 women.
| Cases n (%) |
Controls n (%) |
Source of controls |
Type of interview |
||
|---|---|---|---|---|---|
| Pancreatic cancer | 2838 (37.4) | 4748 (62.6) | |||
| Study |
Principal Investigator |
||||
| North America | |||||
| MDACC (2004–2008) | Li, D | 257 (9.1) | 288 (6.1) | Hospital visitors | Direct |
| MSKCC (2003–2010) | Olson, S | 424 (14.9) | 242 (5.1) | Hospital visitors | Direct |
| Toronto (2003–2006) | Gallinger, S | 217 (7.7) (6.5% proxies) |
134 (2.8) |
Population | Direct/ Proxy (4%) |
| UCSF (1995–1999) | Bracci, PM | 237 (8.4) | 818 (17.2) | Population | Direct |
| Europe | |||||
| Central Europe study; Poland, Czech Republic, Slovakia (2004–2009) |
Scélo, G Holcatova, I |
361 (12.7) | 435 (9.2) | Population | Direct |
| Greece (1990–1992) | Lagiou, P | 66 (2.3) | 132 (2.8) | Hospital, Hospital visitors |
Direct |
| Milan (1982–1999) | La Vecchia, C | 133 (4.7) | 409 (8.6) | Hospital | Direct |
| Italy (1991–2008) | Serraino, D | 148 (5.2) | 304 (6.4) | Hospital | Direct |
| China | |||||
| Shanghai-I (1990–1993) | Ji, B-T | 187 (6.6) (29.4% proxies) |
701 (14.8) (5.6% proxies) |
Population | Direct / Proxy (10.6%) |
| Shanghai-II (2006–2011) | Risch, HA Gao, Y-T |
445 (15.7) | 464 (9.8) | Population | Direct |
| International | |||||
| SEARCH; Canada, The Netherlands, Poland, Australia (1983–1989) |
Bueno-de- Mesquita, HB; Miller, AB; Baghurst, PA; Zatonski, W |
363 (12.8) (58.4% proxies) |
821 (17.3) (16.8% proxies) |
Population | Direct / Proxy (29.6%) |
| Age (years) | |||||
| <45 | 112 (4.0) | 329 (6.9) | |||
| 45–49 | 148 (5.2) | 282 (5.9) | |||
| 50–54 | 228 (8.0) | 488 (10.3) | |||
| 55–59 | 371 (13.1) | 675 (14.2) | |||
| 60–64 | 466 (16.4) | 746 (15.7) | |||
| 65–69 | 544 (19.2) | 832 (17.5) | |||
| 70–75 | 504 (17.8) | 778 (16.4) | |||
| ≥75 | 465 (16.4) | 618 (13.0) | |||
| Race | |||||
| Non-Hispanic White | 1555 (67.6) | 2839 (65.6) | |||
| Non-Hispanic Black | 46 (2.0) | 54 (1.2) | |||
| Hispanic | 18 (0.8) | 53 (1.2) | |||
| Asian/Pacific Islander | 665 (28.9) | 1221 (28.2) | |||
| American Indian/Alaskan native | 2 (0.1) | 3 (0.1) | |||
| Other | 12 (0.5) | 24 (0.6) | |||
| Missing | 1 (0.04) | 133 (3.1) | |||
| Education | |||||
| ≤8th grade | 833 (29.4) | 1708 (36.0) | |||
| 9th – 11th grade | 455 (16.0) | 715 (15.1) | |||
| 12th grade/high school graduate | 607 (21.4) | 850 (17.9) | |||
| Some college/college graduate | 581 (20.5) | 990 (20.9) | |||
| ≥1 year of graduate school | 341 (12.0) | 454 (9.6) | |||
| Missing | 21 (0.7) | 31 (0.7) | |||
| BMI (kg/m2) | |||||
| <20 | 319 (11.2) | 623 (13.1) | |||
| ≥20–<25 | 1298 (45.7) | 2361 (49.7) | |||
| ≥25–<30 | 801 (28.2) | 1204 (25.4) | |||
| ≥30 | 391 (13.8) | 511 (10.8) | |||
| missing | 29 (1.0) | 49 (1.0) | |||
| Tobacco smoking | |||||
| Never | 1730 (61.0) | 3250 (68.5) | |||
| Current smokers (cigarettes/day) | |||||
| <20 | 335 (11.8) | 486 (10.2) | |||
| ≥20 | 142 (5.0) | 113 (2.4) | |||
| Ex-smokers (years since quitting) | |||||
| <10 | 141 (5.0) | 223 (4.7) | |||
| ≥10 | 417 (14.7) | 622 (13.1) | |||
| Missing | 73 (2.6) | 54 (1.1) | |||
| Alcohol drinking (drinks/day)1 | |||||
| 0 to <1 | 1675 (69.4) | 3367 (74.7) | |||
| 1 to <4 | 317 (13.1) | 690 (15.3) | |||
| >=4 | 104 (4.3) | 180 (4.0) | |||
| missing | 318 (13.2) | 269 (6.0) | |||
| History of diabetes | |||||
| No | 2269 (80.0) | 4391 (92.5) | |||
| Yes | 561 (19.8) | 357 (7.5) | |||
| missing | 8 (0.3) | ||||
| History of pancreatitis2 | |||||
| No | 1987 (73.9) | 3912 (88.0) | |||
| Yes | 123 (4.6) | 49 (1.1) | |||
| missing | 580 (21.6) | 483 (10.9) | |||
MDACC: MD Anderson Cancer Center; MSKCC: Memorial Sloan Kettering Cancer Center; UCSF: University of California, San Francisco; SEARCH: Surveillance of Environmental Aspects Related to Cancer in Humans
Information not available in the MSKCC study
Information not available in the Italian study
Exposure variables
Questions about menstrual and reproductive factors and exogenous hormone were generally similar across all the studies; however, full harmonization of the data was performed with the collaboration of the study investigators. Variables included in the analysis were: reported age at menarche, age at menopause, type of menopause (natural or surgical), history of oophorectomy, hysterectomy, age at hysterectomy, number of pregnancies, number of births (including live and stillbirths), age at first birth, number of abortions (including induced and spontaneous), history of oral contraceptive (OC) use, duration of OC use, use of menopause hormone replacement therapy (HRT), and duration of HRT. We also calculated the cumulative lifetime number of menstrual cycles, by adapting the index proposed by Chavez-MacGregor27 as follow: for postmenopausal women, we calculated the difference between age at menopause and age at menarche; for each birth and stillbirth we subtracted cycles during 36 weeks, and 12 weeks for each abortion. Menstrual cycles absent while under OC use were assumed to last 28 days duration. For pre or perimenopausal women, we used age at recruitment instead of age at menopause. Missing age at menopause was imputed using the study-specific mean age at menopause or in case of both ovaries removed, the age at surgery.
Statistical analysis
Two-stage models were used to estimate pooled odds ratios (OR) between menstrual and reproductive factors, hormone use and PC risk. At the first stage, for each study, the association between each factor and PC risk was assessed by estimating the OR and 95% confidence interval (CI) using study-specific logistic regression models28. All models in the first step were adjusted by age (<45, 45–49,50–54, 55–59, 60–64, 65–69, 70–75, ≥75 years), education (≤ 8th grade, 9th-11th grade, 12th grade or high school graduates, some college or college graduate, ≥1 year of graduate school), usual body mass index (BMI, <20, 20 to <25, 25 to <30, ≥30 kg/m2), history of non-gestational diabetes mellitus, cigarette smoking (never-smokers, current smokers < 20 cigarettes/day, current smokers ≥20 cigarettes/day, ex-smokers < 10 years, ex-smokers ≥10 years), alcohol (no info available, 0 to <1 drink/day, 1 to < 4 drinks/day, ≥4 drinks/day), race (Non-Hispanic white, non-Hispanic black, Hispanic, Asian/Pacific Islander, American Indian/Alaskan native) and center (for multicentric studies). History of pancreatitis was not included as an adjustment covariate in the present analysis because some of the studies observed no controls with this exposure. To account for possible differences in hormone levels during pregnancies, number of pregnancies was included when lifetime cumulative cycles was evaluated. At the second stage, pooled effect estimates across studies were calculated using random effects meta-analysis29. To evaluate study-based heterogeneity, we calculated the χ2 statistic and the index I-square30. Galbraith plots were used to examine sources of heterogeneity31, and sensitivity analyses excluding the study/studies identified with Galbraith plots were performed to evaluate study influence on pooled ORs. Studies that contributed significantly to heterogeneity in the pooled estimates for reproductive factors were excluded from the analysis of that factor. In order to account for all sources of hormone exposure at the same time, mutually adjusted models were evaluated after possible collinearity between exposure variables was assessed. Effect-measure modification by tobacco use (never, current, former), BMI (under+normal weight vs. obese+overweight), and histories of diabetes were evaluated using the likelihood-ratio statistic. All analyses were additionally examined restricted to postmenopausal women and to never smokers. Finally, sensitivity analyses excluding the proxy respondents and stratifying by source of control participants (population- or hospital-based) from the mutually adjusted models were also performed.
RESULTS
Table 1 gives study characteristics and distributions of the core variables according to PC status. Case subjects were older than controls, reported higher BMIs, and were more likely to be current or ex-smokers. Greater proportions of women who had had diabetes or pancreatitis were observed among cases than in controls.
To account for possible differences in the exposure variables by race/ethnicity, distributions were evaluated and only minor differences in age at menarche were observed between Asian and non-Asian ethnicities; however these differences were not statistically significant (data not shown).
All models of menstrual and reproductive factors and hormone use variables were adjusted for attained age, education, race, usual BMI, cigarette smoking, histories of diabetes and use of alcohol, and center. A statistically significant inverse association was observed in women who reported having had hysterectomy (OR 0.78, 95%CI 0.67–0.91), remaining consistent in postmenopausal and never-smoking women (Table 2). Earlier hysterectomy (≤41 years) showed non-significant inverse associations with PC risk (OR≤37vs>52 0.87; 0.53–1.42; OR>37&≤41vs>52 0.62; 0.37–1.05) (Table 2). The observed OR for HRT users was 0.82 (95%CI 0.68–0.98) (Table 2).
Table 2.
Menstrual, reproductive factors, and hormone use in relation to PC risk in PanC4 women.
| Cases n (%) |
Controls n (%) |
Pooled OR (95%CI)1 |
Pooled OR (95%CI)1 postmenopausal |
Pooled OR (95%CI)2 never smokers |
|
|---|---|---|---|---|---|
| MENSTRUAL FACTORS | |||||
| Age at menarche (years) | |||||
| <123 | 343 (12.1) | 571 (12.0) | 0.98 (0.79–1.21) | 0.96 (0.76–1.21) | 0.91 (0.68–1.21) |
| 12 | 467 (16.5) | 783 (16.5) | reference | reference | Reference |
| 13 | 634 (22.3) | 949 (20.0) | 1.03 (0.86–1.24) | 1.05 (0.87–1.26) | 1.01 (0.80–1.29) |
| 14 | 475 (16.7) | 871 (18.3) | 0.92 (0.72–1.17) | 0.92 (0.73–1.17) | 0.84 (0.62–1.12) |
| >14 | 734 (25.9) | 1441 (30.4) | 0.85 (0.66–1.08) | 0.85 (0.66–1.10) | 0.85 (0.65–1.01) |
| Missing | 185 (6.5) | 133 (2.8) | |||
| Type of menopause4,5 | |||||
| Natural menopause | 2053 (78.6) | 3214 (78.7) | Reference | Reference | |
| Surgical menopause | 504 (19.3) | 832 (20.4) | 0.91 (0.77–1.08) | 0.91 (0.74–1.11) | |
| missing | 56 (2.1) | 39 (0.9) | |||
| Age at menopause (years)5 | |||||
| <=39 | 209 (7.7) | 352 (8.1) | reference | reference | |
| 40–44 | 290 (10.7) | 492 (11.3) | 0.99 (0.75–1.31) | 1.19 (0.81–1.77) | |
| 45–49 | 704 (25.9) | 1120 (25.8) | 1.26 (0.98–1.62) | 1.41 (0.92–2.15) | |
| 50–54 | 1013 (37.3) | 1495 (34.4) | 1.27 (0.97–1.66) | 1.46 (0.96–2.20) | |
| >=55 | 243 (8.9) | 353 (8.1) | 1.32 (0.93–1.87) | 1.58 (0.86–2.88) | |
| Missing | 255 (9.4) | 532 (12.3) | |||
|
Lifetime cumulative menstrual cycles (cycles)6,7 |
|||||
| <=333 | 239 (14.1) | 474 (18.6) | reference | reference | Reference |
| >333–383 | 329 (19.4) | 552 (21.6) | 1.01 (0.74–1.38) | 0.96 (0.75–1.24) | 1.01 (0.62–1.63) |
| >383–419 | 378 (22.3) | 536 (21.0) | 1.13 (0.87–1.46) | 1.08 (0.83–1.40) | 1.07 (0.70–1.63) |
| >419–455 | 379 (22.3) | 502 (19.7) | 0.98 (0.75–1.29) | 0.98 (0.74–1.28) | 1.03 (0.75–1.43) |
| >455 | 336 (19.8) | 468 (18.3) | 0.74 (0.54–1.03) | 0.75 (0.53–1.06) | 0.68 (0.48–0.97) |
| Missing | 37 (2.2) | 23 (0.9) | |||
| Oophorectomy8 | |||||
| No | 1512 (75.3) | 2309 (75.9) | Reference | Reference | Reference |
| 1 ovary removed | 126 (6.3) | 177 (5.8) | 0.96 (0.72–1.27) | 0.96 (0.71–1.29) | 0.99 (0.67–1.45) |
| 2 ovaries removed | 279 (13.9) | 442 (14.5) | 0.83 (0.64–1.06) | 0.84 (0.65–1.08) | 0.77 (0.58–1.02) |
| Missing | 90 (4.5) | 114 (3.8) | |||
| Hysterectomy8 | |||||
| No | 1361 (67.8) | 1930 (63.5) | Reference | Reference | Reference |
| Yes | 562 (28.0) | 925 (30.4) | 0.78 (0.67–0.91) | 0.82 (0.70–0.97) | 0.74 (0.60–0.91) |
| Missing | 84 (4.2) | 187 (6.2) | |||
|
Age at hysterectomy (years)8,9 |
|||||
| <=37 | 128 (24.3) | 181 (18.7) | 0.87 (0.53–1.42) | 0.96 (0.57–1.62) | 0.86 (0.48–1.55) |
| >37–41 | 64 (12.1) | 158 (16.3) | 0.62 (0.37–1.05) | 0.66 (0.38–1.14) | 0.51 (0.26–0.98) |
| >41–46 | 109 (20.7) | 167 (17.2) | 1.01 (0.53–1.91) | 1.01 (0.51–2.01) | 1.40 (0.79–2.47) |
| >46–52 | 90 (17.1) | 163 (16.8) | 0.95 (0.58–1.54) | 1.01 (0.60–1.70) | 0.84 (0.47–1.51) |
| >52 | 83 (15.8) | 150 (15.5) | reference | reference | Reference |
| Missing | 53 (10.1) | 151 (15.6) | |||
|
Hysterectomy & Oophorectomy8 |
|||||
| None of them | 1265 (63.0) | 1830 (60.2) | Reference | Reference | Reference |
| Oophorectomy | 54 (2.7) | 73 (2.40) | 0.91 (0.60–1.39) | 0.81 (0.62–1.05) | 0.83 (0.63–1.08) |
| Hysterectomy | 206 (10.3) | 366 (12.0) | 0.77 (0.60–1.00) | 0.85 (0.64–1.14) | 0.83 (0.61–1.13) |
| Oophorectomy & Hysterectomy | 348 (17.3) | 540 (17.8) | 0.79 (0.62–1.00) | 0.85 (0.55–1.30) | 0.73 (0.33–1.64) |
| Missing | 134 (6.7) | 233 (7.7) | |||
| REPRODUCTIVE FACTORS | |||||
| Number of pregnancies10 | |||||
| Never pregnant | 233 (9.0) | 461 (10.3) | Reference | Reference | Reference |
| 1 | 294 (11.4) | 536 (12.0) | 1.05 (0.77–1.43) | 1.01 (0.73–1.40) | 1.06 (0.77–1.45) |
| 2 | 641 (24.8) | 1,043 (23.4) | 1.25 (0.86–1.83) | 1.31 (0.89–1.95) | 1.35 (0.87–1.68) |
| 3 | 562 (21.8) | 885 (19.8) | 1.16 (0.92–1.45) | 1.19 (0.93–1.52) | 1.22 (0.89–1.68) |
| >=4 | 808 (31.3) | 1,501 (33.6) | 1.10 (0.79–1.52) | 1.13 (0.80–1.59) | 1.23 (0.83–1.82) |
| Missing | 43 (1.7) | 36 (0.8) | |||
|
Number of births (life births and stillbirth)10 |
|||||
| Never pregnant | 175 (6.8) | 258 (5.8) | 0.93 (0.62–1.39) | 0.89 (0.58–1.37) | 0.83 (0.60–1.16) |
| Nulliparous | 32 (1.2) | 94 (2.1) | 0.49 (0.29–0.83) | 0.47 (0.28–0.81) | 0.50 (0.22–1.13) |
| 1 | 493 (19.1) | 811 (18.2) | Reference | Reference | reference |
| 2 | 788 (30.5) | 1347 (30.2) | 0.92 (0.77–1.10) | 0.94 (0.78–1.15) | 0.85 (0.70–1.03) |
| 3 | 518 (20.1) | 799 (17.9) | 1.06 (0.86–1.29) | 1.05 (0.83–1.32) | 1.02 (0.82–1.26) |
| >=4 | 473 (18.3) | 895 (20.1) | 1.05 (0.83–1.32) | 1.05 (0.81–1.37) | 0.94 (0.74–1.18) |
| Missing | 102 (4.0) | 256 (5.7) | |||
| Age at first birth (years)10,11 | |||||
| <20 | 214 (11.0) | 370 (10.10) | Reference | Reference | Reference |
| 20–23 | 656 (33.7) | 1185 (32.3) | 0.98 (0.78–1.24) | 1.00 (0.75–1.35) | 1.02 (0.75–1.39) |
| 24–27 | 509 (26.2) | 1034 (28.2) | 0.90 (0.70–1.15) | 0.97 (0.69–1.35) | 0.88 (0.63–1.22) |
| >27 | 455 (23.4) | 801 (21.8) | 1.01 (0.65–1.55) | 1.04 (0.64–1.69) | 1.07 (0.76–1.52) |
| Missing | 110 (5.7) | 277 (7.6) | |||
|
Number of abortions (induced and pontaneous)10 |
|||||
| Never pregnant | 175 (6.9) | 258 (5.8) | Reference | Reference | Reference |
| 0 | 1361 (52.7) | 2204 (49.4) | 1.26 (0.82–1.96) | 1.40 (0.90–2.07) | 1.40 (0.91–2.41) |
| 1 | 562 (21.8) | 1010 (22.7) | 0.98 (0.75–1.27) | 1.09 (0.80–1.48) | 1.03 (0.67–1.60) |
| 2 | 259 (10.0) | 475 (10.7) | 0.85 (0.55–1.33) | 0.99 (0.60–1.64) | 0.97 (0.53–1.80) |
| >=3 | 122 (4.7) | 257 (5.8) | 0.88 (0.58–1.35) | 1.06 (0.68–1.65) | 0.93 (0.51–1.69) |
| Missing | 102 (3.9) | 256 (5.7) | |||
|
EXOGENOUS HORMONE USE |
|||||
| OC use12 | |||||
| Non-user | 1863 (67.2) | 3266 (70.8) | Reference | Reference | Reference |
| Yes | 660 (23.8) | 1177 (25.5) | 0.83 (0.69–1.01) | 0.82 (0.65–1.02) | 0.86 (0.67–1.09) |
| Missing | 249 (9.0) | 173 (3.8) | |||
| OC duration (months)12,13,14 | |||||
| Non-user | 1568 (73.8) | 2574 (71.3) | Reference | Reference | Reference |
| <12 | 78 (3.7) | 176 (4.9) | 0.77 (0.49–1.21) | 0.79 (0.52–1.19) | 0.76 (0.41–1.40) |
| 12–<36 | 131 (6.2) | 225 (6.2) | 0.90 (0.67–1.20) | 0.88 (0.65–1.19) | 1.30 (0.87–1.94) |
| ≥36 | 332 (15.6) | 627 (17.4) | 0.84 (0.67–1.06) | 0.84 (0.63–1.11) | 0.85 (0.64–1.13) |
| Missing | 15 (0.7) | 9 (0.25) | |||
| HRT use12, 15 | |||||
| No | 1395 (65.2) | 2251 (65.2) | Reference | Reference | Reference |
| Yes | 632 (29.5) | 1105 (32.0) | 0.82 (0.68–0.98) | 0.79 (0.65–0.97) | 0.78 (0.58–1.05) |
| Missing | 113 (5.3) | 95 (2.8) | |||
| HRT duration (months)12,16 | |||||
| Non-user | 1395 (68.8) | 2251 (67.1) | Reference | Reference | Reference |
| <12 | 95 (4.7) | 184 (5.5) | 0.86 (0.55–1.33) | 0.88 (0.56–1.38) | 0.92 (0.61–1.39) |
| 12–<36 | 128 (6.3) | 182 (5.4) | 1.04 (0.72–1.49) | 1.04 (0.71–1.53) | 0.94 (0.61–1.46) |
| ≥36 | 393 (19.4) | 703 (21.0) | 0.81 (0.66–1.00) | 0.80 (0.63–1.01) | 0.75 (0.56–1.00) |
| Missing | 16 (0.8) | 36 (1.1) | |||
OC: Oral contraceptive use; HRT: hormone replace therapy
Adjusted for: age in 5-year categories, education, BMI, tobacco smoking, diabetes (yes/no), alcohol and center
Adjusted for: age in 5-year categories, education, BMI, diabetes (yes/no), alcohol and center
Shanghai-I was not included in this category because there were 0 cases and 2 controls
Information not available in Shanghai-I study
In post-menopausal women
Information not available in MDACC, SEARCH, Toronto and Greek studies. UCSF was excluded due to heterogeneity (I2=62%)and its influence in the pooled estimates
Adjusted for age 5 years categories, education, BMI, tobacco smoking, diabetes (yes/no), alcohol, center and number of pregnancies
Information not available in the Milan, Greek, Shanghai-I and Shanghai-II studies
Central Europe study excluded due to high heterogeneity (I2: 53.4%) and its influence in the pooled estimations
Information on parity not available in the MDACC study
In parous women
Information not available in the Greek study
Shanghai - I study was not included due to the proportion of non-users and short-users (<6 months) was ≥95%
Toronto study was not included due to the lack of controls in the highest category
Shanghai-I study was not included due to the proportion of non-users was ≥95%. Italy+Milan studies were excluded due to heterogeneity (I2=58.2%) and its influence in the pooled estimates.
Milan, Italy, Shanghai-I and Shanghai-II were not included due to the proportion of non-users and short-users (<6 months) was ≥90%
Non-significant ORs below unity were observed in women who had later menarche (OR>14vs<12 0.85, 95%CI 0.66–1.08), bilateral oophorectomy (OR 0.83, 95%CI 0.64–1.06), used OC (OR 0.83, 95%CI 0.69–1.01), reported long-term use of HRT (OR≥36vsnon-users 0.81, 95%CI 0.66–1.00) or had high number of menstrual cycles (OR>455vs≤333 0.74, 95%CI 0.54–1.06) (Table 2). Although the association of high number of menstrual cycles was non-significant, in the subset of never-smokers the pooled OR was 0.68 (95%CI 0.60–0.91) (Table 2). An elevated but non-significant OR was observed in women with later menopause (OR≥55vs≤39 1.32, 95%CI 0.93–1.87) (P trend=0.63). Number of pregnancies, number of births and age at first birth showed no associations with PC risk (Table 2). We also evaluated a mutually adjusted model including all sources of hormone exposure: exogenous hormones, lifetime cumulative menstrual cycles (as a summary variable for endogenous hormone exposure); and gynecological surgery: hysterectomy and oophorectomy. Since not all the studies collected all the information mentioned previously, two separate models were analyzed. Studies that contributed significantly to heterogeneity in the pooled estimates in table 2 were excluded for the mutually adjusted model. The first model included lifetime cumulative menstrual cycles, HRT and OC use; the studies that provided information were MSKCC, Central Europe and Shanghai-II. A non-significant pooled OR below unity was observed in women with more than 455 cycles compared with women with at most 333 cycles (OR>455vs≤333 0.75, 95%CI 0.51–1.12); the pooled OR for HRT users was 0.87 (95%CI 0.55–1.24). A null effect was observed for OC users (Model 1; Table 3). The second mutually adjusted model included MDACC, MSKCC, Toronto, UCSF, Central Europe and SEARCH studies. For this model, the joint effect of hysterectomy and HRT use was evaluated according to the following categories: use of neither, HRT use alone, hysterectomy alone, and both hysterectomy plus HRT use (Model 2; Table 3). This joint effect variable was evaluated because the frequency of HRT use was three times as high in women who had had hysterectomies than not (data not shown). History of oophorectomy and OC use were also included in the model (Model 2; Table 3). The joint effect analysis of hysterectomy and HRT use showed that HRT use without hysterectomy conveyed a non-significant inverse association with PC (ORHRTvsnone 0.84, 95%CI 0.64–1.10), while a significant inverse association with PC risk was observed in women with hysterectomy without HRT use (ORhysterectomyvsnone 0.70, 95%CI 0.54–0.92) which was somewhat lower (ORhysterectomy+HRTvsneither 0.64, 95%CI 0.48–0.84) for women also taking HRT. No statistically significant association was observed in relation to oophorectomy. A sensitivity analysis was performed, where the joint effect was considered only when HRT use started at the same time or after the hysterectomy, however, the results did not change (data not shown). A forest plot of the study-specific and the pooled ORs for the joint effect of hysterectomy and use of HRT and PC risk is presented in Figure 1.
Table 3.
Mutually adjusted models for cumulative number of menstrual cycles, hormone use, gynecological surgery, and PC risk in PanC4 women.
| Cases/ controls |
Pooled OR (95%CI) | Pooled OR (95%CI) postmenopausal |
Pooled OR (95%CI) never smokers |
||
|---|---|---|---|---|---|
| Model 11,2 |
Lifetime cumulative menstrual cycles (cycles) |
||||
| <=333 | 161/135 | reference | reference | reference | |
| >333–383 | 208/215 | 0.81 (0.58–1.14) | 0.81 (0.58–1.15) | 0.71 (0.48–1.06) | |
| >383–419 | 273/247 | 0.93 (0.66–1.31) | 0.94 (0.67–1.32) | 0.81 (0.54–1.21) | |
| >419–455 | 288/273 | 0.85 (0.60–1.19) | 0.85 (0.60–1.20) | 0.79 (0.52–1.18) | |
| >455 | 267/252 | 0.75 (0.51–1.12) | 0.76 (0.52–1.11) | 0.65 (0.42–1.01) | |
| Missing | 33/19 | ||||
| HRT use | |||||
| No | 995/912 | reference | reference | reference | |
| Yes | 185/166 | 0.87 (0.55–1.39) | 0.87 (0.55–1.39) | 0.87 (0.53–1.42) | |
| Missing | 60/63 | ||||
| OC use | |||||
| Non-user | 796/785 | reference | reference | reference | |
| Yes | 231/223 | 0.95 (0.72–1.24) | 0.94 (0.72–1.23) | 0.95 (0.68–1.32) | |
| Missing | 203/133 | ||||
| Model 2 3,4 | Hysterectomy & HRT use | ||||
| Neither | 865/1114 | reference | reference | reference | |
| HRT use alone | 303/492 | 0.84 (0.64–1.10) | 0.84 (0.63–1.12) | 0.73 (0.46–1.15) | |
| Hysterectomy alone | 234/323 | 0.70 (0.54–0.92) | 0.80 (0.60–1.07) | 0.57 (0.39–0.83) | |
| Hysterectomy and HRT use | 287/534 | 0.64 (0.48–0.84) | 0.65 (0.48–0.88) | 0.55 (0.37–0.83) | |
| Missing | 170/275 | ||||
| Oophorectomy | |||||
| No | 1611/1144 | reference | reference | reference | |
| 1 ovary removed | 113/147 | 1.13 (0.81–1.58) | 1.08 (0.77–1.53) | 1.34 (0.82–2.19) | |
| 2 ovaries Removed | 269/413 | 1.12 (0.83–1.52) | 1.09 (0.80–1.48) | 1.13 (0.77–1.66) | |
| Missing | 88/101 | ||||
| OC use | |||||
| Non-user | 1046/1576 | reference | reference | reference | |
| Yes | 565/990 | 0.78 (0.61–0.99) | 0.76 (0.57–1.01) | 0.83 (0.59–1.15) | |
| Missing | 248/172 | ||||
OC: Oral contraceptive use; HRT: Hormone replace therapy
Adjuster for: age in 5-year categories, education, BMI, tobacco smoking, diabetes (yes/no), alcohol, center and number of pregnancies // Never smokers model did not include tobacco smoking variable
Studies included MSKCC, Central Europe, and Shanghai-II
Adjusted for: age in 5-year categories, education, BMI, tobacco smoking, diabetes (yes/no), alcohol and center // Never smokers model did not include tobacco smoking variable
Studies included MDACC, MSKCC, Toronto, UCSF, Central Europe, and SEARCH
Figure 1.
Study-specific and pooled OR estimates for the associations of hysterectomy and HRT use with risk of PC in PanC4 women (Model 2; Table 3)
Footnote for figure 1: Adjusted for: age in 5-year categories, education, BMI, cigarette smoking, diabetes, alcohol, center, oophorectomy and OC use.
Box sizes are weighted by the inverse of the variance; HRT: hormone replace therapy; OC: oral contraceptive
We found no evidence for effect-measure modification of the relation between each factor and PC risk by cigarette smoking, BMI, and history of diabetes (all likelihood ratio statistic p-values>0.05) (data not shown). Sensitivity analyses in never-smoking women and in postmenopausal women were performed, but in general, except as noted above, results did not change (Table 2 and 3). No major differences in pooled OR estimates were observed when proxy-respondents were excluded from the mutually adjusted models or when models were stratified by the source of control participants (data not shown).
DISCUSSION
The present pooled data analysis of eleven case-control studies, including 2,838 pancreatic cancer case women and 4,748 controls, allowed us to estimate more precisely possible relations between menstrual and reproductive factors, hormone use, and PC risk. Our results suggest that undergoing hysterectomy may significantly reduce the risk of developing PC by 22%.
A non-statistically significant OR below unity was observed with high lifetime cumulative number of menstrual cycles (>455 cycles). When the analysis was restricted to never-smokers, this association became statistically significant. Lifetime cumulative menstrual cycles is an index for total exposure to endogenous hormones. Since hormonal levels differ during pregnancy, we subtracted from the calculation 36 and 12 weeks for each birth and abortion, respectively, and included the number of pregnancies as an adjustment variable; and hormone use (HRT and OC) was also included in the model. Chavez-MacGregor27 also excluded from the calculation a 6-week absence of cycles for women who reported lactation, and they accounted for menstrual cycle irregularity. Unfortunately, this information was not available in the included PanC4 studies. Because smoking is an important risk factor for PC and may also affect sex hormone levels32, we carried out additional analyses limited to never-smokers, though similar patterns of relative risk were observed. Obese and overweight persons have an increased risk of PC33, and high BMI is positively associated with high estrogens levels34, but in our analysis, different levels of BMI did not alter our OR estimations. Further analyses were also performed in postmenopausal women, and similar estimates of risk and 95%CIs were observed.
In the published literature, seven previous studies of PC collected information on oophorectomy and hysterectomy; and all case-control studies that collected this information18–20,35 were included in the present pooled analysis. Hysterectomy prevalence in PanC4 cases is almost 28% (American studies: 33%, European studies: 21%). In the American cohort studies hysterectomy prevalence in PC cases is approximately 40%36,37, while in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort the prevalence was almost 15%38. The Iowa Women's health study (IWHS) cohort showed a statistically significant increase in risk for both hysterectomy (hazard ratio (HR): 1.37, 95%CI: 1.02–1.82) and bilateral oophorectomy (HR: 1.43, 95%CI: 1.01–2.00)39. The EPIC cohort observed a no effect of bilateral oophorectomy and hysterectomy with ovarian conservation on PC risk38. The Cancer Prevention Study-II Nutrition Cohort found that hysterectomy with ovarian conservation was associated with an increased risk of PC (OR 1.48, 95%CI 1.03–2.14), although no effect of hysterectomy with bilateral oophorectomy on PC was observed (OR 0.97, 95%CI 0.69–1.37)36. A recent meta-analysis observed16 an inverse association in relation to hysterectomy and PC risk in case-control studies (OR 0.77, 95%CI 0.64–0.94) and no association in cohort studies. This result in case-control studies agreed with our general findings on hysterectomy (OR 0.78, 95%CI 0.67–0.91). Discrepancies between cohort studies and case-controls studies might be explained by selection bias in hospital-based (HB) case-control studies, however, we observed that the association between hysterectomy and PC was borderline in the population-based (PB) case-control studies (PB: OR: 0.84, 95%CI 0.69–1.01; HB: OR: 0.70, 95%CI 0.54–0.90) and results by source of controls were not considered heterogeneous (Wald statistic for heterogeneity: 1.21, p-value=0.271).
The role of HRT in relation to PC risk is not clear. While two cohort37,40 showed non-significant inverse associations, one cohort study41 and a case-control studies42 showed non-significant increases in risk. Finally, two studies showed relative risk estimates for HRT close to unity38,39. All remaining case-control studies were included in our pooled analysis18–20,35. None of these studies, or our pooled analyses, were able to distinguish the type of hormone therapy used, for example, whether combinations of estrogen and progestin were used, or estrogen alone. Use patterns for type of hormone therapy have changed over the past decades, with eras of estrogen alone and estrogen plus progestin43. The Women’s Health Initiative (WHI) randomized controlled trial findings support the hypothesis that estrogen alone or estrogen plus progestin may not have the same biological effects. The WHI observed that women who reported the combination of estrogen and progestin had an increased risk of breast cancer and a non-significant decrease in the risk of endometrial cancer44; while taking estrogen alone did not show an increase of breast cancer in women with hysterectomy45. Thus, estrogen-only HRT is usually recommended for women who have had hysterectomy43. To evaluate the influence of changes in the consumption patterns of HRT, we analyzed the use of HRT by study year (before 2002 vs. after 2002); and we observed no differences in the pooled risk estimate.
In PanC4, information on HRT type was not available; however, we did observe that the frequency of HRT use was three times as high among women who reported hysterectomy than not. Further, almost 57% of women started the HRT treatment at the age of the hysterectomy, and 30% after hysterectomy. We observed a 36% lower risk in comparison to women with intact uteri and who had not used HRT. It is possible that women who have had hysterectomy and used HRT may have lower PC risk because of other factors besides hysterectomy and HRT.
Some diseases, for which the treatment is hysterectomy, are directly related with increased female hormone levels. Almost 70% of American women46 during their fertile life will have fibroids. In many cases, fibroids do not cause symptoms, however, they can cause abnormal bleeding, and pelvic pressure, for which hysterectomy is the recommended treatment46,47. Among the factors that can contribute to fibroid growth are elevated levels of estrogens and progesterone47. Further, women diagnosed with hyperplasia tend to undergo hysterectomy surgery48. Hyperplasia is associated with higher levels of estrogens and insufficient levels of progesterone, and may evolve to endometrial cancer48. Unfortunately, the specific reason for having a hysterectomy was not available in PanC4 data, so we could not verify if the observed protective effect that was shown with hysterectomy was due to underlying elevated estrogen levels in the mentioned diseases or to other factors related to having a hysterectomy.
Lifetime cumulative menstrual cycles is an index that attempts to summarize reproductive information, including information on menarche, menopause, pregnancies, OC and lactation. Our analytical approach allowed us to evaluate the index including these factors, with the exception of lactation. Three previous studies of pancreatic cancer have evaluated the index, but each study used a different calculation and different adjustment variables38,39,49. All the studies, including our analysis, found no association between lifetime cumulative menstrual cycles and PC risk, suggesting little or no effect of female hormones on PC risk.
In light of the inconsistent results between the studies on menstrual and reproductive factors, hormone use and PC risk, it is worth noting that previous studies had diverse study designs, target populations and confounder and adjustment variables; specifically, adjustments for smoking and alcohol. Furthermore, risk estimates for hysterectomy and for oophorectomy were inconsistent in our analysis with those provided by cohort studies36,38,39, possibly caused by selection bias in case-control studies.
The main weakness of our analysis is that not all of the studies collected all of the information on menstrual and reproductive factors and hormone use. Thus, mutually adjusted models could be obtained in only 4 or 6 studies, depending on the included variables, and a model that contained all collected factors in all studies was not possible. Another weakness of our analyses is that for some variables such as age at menopause, and for OC and HRT durations, we received categorical variables from some of the studies, so cut points for these variables had to be based on the received information, or the information could not be included in the calculation of lifetime cumulative menstrual cycles. Also, time periods of studies varies from the 1980’s to 2011, however, low heterogeneity between studies in each pooled OR estimation was observed. Even so, the PanC4 consortium includes a large dataset which allowed us to adjust for cigarette smoking, alcohol intake, and other potential confounding variables, and with sufficient power to estimate ORs across strata of major PC risk factors, for example, smoking, BMI, and diabetes.
In conclusion, our pooled analysis found no associations between age at menarche, menopause, lifetime cumulative menstrual cycles, oophorectomy, parity, history of OC use, and PC risk, but suggests that women who have had hysterectomy may be at lower risk of PC. Further investigations by type and formulation of HRT and reason for hysterectomy could clarify the role, if any, of hysterectomy in relation to PC risk.
Acknowledgments
Funding: This work was supported by the Red Temática de Investigación Cooperativa en Cáncer (RTICC) [Rd06/0020/0091 and Rd12/0036/0018] and by the Generalitat de Catalunya. The Central Europe study was supported by the Grant Agency of Ministry of Health of the Czech Republic (IGA MZ ČR 8090-3 & 9422-3). The center of Brno, Czech Republic – within the Central Europe study – was supported by MH CZ - DRO (MMCI, 00209805) and RECAMO CZ.1.05/2.1.00/03.0101. The Italy and Milan studies were supported by the Italian Association for Cancer Research (AIRC grant 10068), and the Italian Foundation for Cancer Research (FIRC). The MD Anderson Cancer Center study was supported by the NIH Research Project Grant Program (R01 CA098380). The Pancreatic Cancer Family Registry at the Memorial Sloan Kettering Cancer Center (MSKCC) was supported by the Prevention, Control, and Population Research Goldstein Award; the Society of MSKCC; and the Geoffrey Beene Cancer Research Fund. The University of California, San Francisco (UCSF) study work was supported in part by National Cancer Institute grants CA059706, CA108370, CA109767, CA089726, CA098889, and by the Rombauer Pancreatic Cancer Research Fund. Cancer incidence data collection in the UCSF study was supported by the California Department of Public Health, the National Cancer Institute’s Surveillance, Epidemiology and End Results Program contract N01-PC-35136 awarded to the Northern California Cancer Center. The Montreal investigation in the Surveillance of Environmental Aspects Related to Cancer in Humans study was supported by the Cancer Research Society, the Toronto contribution was supported by the National Cancer Institute of Canada, and the Netherlands contribution was supported by the Dutch Ministry of Public Health, Welfare and Sports (formerly Welfare, Health and Culture). The Shanghai Pancreas Cancer Study (study II) was supported by U.S. National Institute of Health grant 5R01 CA114421, by grants from the Science and Technology Commission of Shanghai Municipality, and from the Shanghai Cancer Institute (SB10-06). In all of the study support, the sponsors had no roles in the collection, analysis, or interpretation of study data.
Footnotes
Conflict of interest: The authors have no conflicts of interest to report
Contributor Information
Leila Lujan-Barroso, Unit of Nutrition and Cancer, Cancer Epidemiology Research Program, Catalan Institute of Oncology (ICO), Bellvitge Biomedical Research Institute (IDIBELL), Barcelona, Spain.
Wei Zhang, Department of Epidemiology, Shanghai Cancer Institute and Jiao Tong University, Shanghai, China.
Sara H. Olson, Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY.
Yu-Tang Gao, Department of Epidemiology, Shanghai Cancer Institute and Jiao Tong University, Shanghai, China.
Herbert Yu, Epidemiology Program, University of Hawaii Cancer Center, Honolulu, HI, USA.
Peter A. Baghurst, Public health, Women’s and Children’s Hospital, Adelaide, SA, Australia.
Paige M. Bracci, University of California, San Francisco, San Francisco, CA, USA.
H. Bas Bueno-de-Mesquita, National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands; Department of Gastroenterology and Hepatology, University Medical Center Utrecht, Utrecht, The Netherlands; Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, London, UK; Department of Social and Preventive Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia.
Lenka Foretová, Department of Cancer Epidemiology and Genetics, Masaryk Memorial Cancer Institute, Institute and MF MU, Brno, Czech Republic.
Steven Gallinger, University Health Network, Department of Surgery, University of Toronto, Toronto, Canada.
Ivana Holcatova, Institute of Hygiene and Epidemiology, 1st Faculty of Medicine, Charles University in Prague, Prague, Czech Republic.
Vladimír Janout, Department of Preventive Medicine, Faculty of Medicine, Palacky University, Olomouc, Czech Republic.
Bu-Tian Ji, National Cancer Institute, Bethesda, MD, USA.
Robert C. Kurtz, Department of Medicine, Memorial Sloan Kattering Cancer Center, New York, NY, USA.
Carlo La Vecchia, Department of Clinical Sciences and Community Health, Università degli Studi di Milano, Milan, Italy.
Pagona Lagiou, Department of Hygiene, Epidemiology and Medical Statistics, School of Medicine, University of Athens, Greece; Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA.
Donghui Li, M.D. Anderson Cancer Center, University of Texas, Houston, TX, USA.
Anthony B. Miller, Dalla Lana School of Public Health, University of Toronto, Toronto, Canada.
Diego Serraino, Unit of Epidemiology and Biostatistics, CRO Aviano, National Cancer Institute, IRCCS, Aviano, Italy.
Witold Zatonski, Cancer Center & Institute of Oncology, Warsaw, Poland.
Harvey A. Risch, Department of Chronic Disease Epidemiology, Yale School of Public Health, New Haven, CT, USA.
Eric J. Duell, Unit of Nutrition and Cancer, Cancer Epidemiology Research Program, Catalan Institute of Oncology (ICO), Bellvitge Biomedical Research Institute (IDIBELL), Barcelona, Spain.
Reference
- 1.Worldwide Data - Cancer Facts & Figures. [Accessed October 1, 2014];[World Cancer Research Fund International] 2014 Oct 1; Available at: http://www.wcrf.org/int/cancer-facts-figures/worldwide-data.
- 2.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin. 2015;65:5–29. doi: 10.3322/caac.21254. [DOI] [PubMed] [Google Scholar]
- 3.Gong Z, Holly EA, Bracci PM. Survival in population-based pancreatic cancer patients: san francisco bay area, 1995–1999. Am J Epidemiol. 2011;174:1373–1381. doi: 10.1093/aje/kwr267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Rosato V, Polesel J, Bosetti C, et al. Population attributable risk for pancreatic cancer in northern Italy. Pancreas. 2015;44:216–220. doi: 10.1097/MPA.0000000000000251. [DOI] [PubMed] [Google Scholar]
- 5.Iodice S, Gandini S, Maisonneuve P, et al. Tobacco and the risk of pancreatic cancer: a review and meta-analysis. Langenbecks Arch Surg. 2008;393:535–545. doi: 10.1007/s00423-007-0266-2. [DOI] [PubMed] [Google Scholar]
- 6.Yeo TP, Lowenfels AB. Demographics and epidemiology of pancreatic cancer. Cancer J. 2012;18:477–484. doi: 10.1097/PPO.0b013e3182756803. [DOI] [PubMed] [Google Scholar]
- 7.Anderson KE, Mack TM, Silverman DT. In: Cancer epidemiology and prevention. Schottenfeld D, Fraumeni JF, editors. Vol. 3. New York: Oxford University Press; 2006. pp. 721–762. [Google Scholar]
- 8.Wolpin BM, Rizzato C, Kraft P, et al. Genome-wide association study identifies multiple susceptibility loci for pancreatic. Cancer Nat Genet. 2014;46:994–1000. doi: 10.1038/ng.3052. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. [Accessed February 19, 2015];Surveilleance, Epidemiology, and End Results Program. [Fast stats] 2015 Feb 19; Available at: http://seer.cancer.gov/faststats/
- 10. [Accessed February 19, 2015];GLOBACAN 2012: Estimated Cancer Incidence, Moratality and Prevalence Worldwide in 2012. [International Agency for Research on Cancer] 2015 Feb 19; Available at: http://globocan.iarc.fr/Pages/age-specific_table_sel.aspx.
- 11.Sumi C, Longnecker DS, Roebuck BD, et al. Inhibitory effects of estrogen and castration on the early stage of pancreatic carcinogenesis in fischer rats treated with azaserine. Cancer Res. 1989;49:2332–2336. [PubMed] [Google Scholar]
- 12.Sumi C, Brinck-Johnsen T, Longnecker DS. Inhibition of a transplantable pancreatic carcinoma by castration and estradiol administration in rats. Cancer Res. 1989;49:6687–6692. [PubMed] [Google Scholar]
- 13.Wahi MM, Shah N, Schrock CE, et al. Reproductive factors and risk of pancreatic cancer in women: a review of the literature. Ann Epidemiol. 2009;19:103–111. doi: 10.1016/j.annepidem.2008.11.003. [DOI] [PubMed] [Google Scholar]
- 14.Guan H-B, Wu L, Wu Q-J, et al. Parity and pancreatic cancer risk: a dose-response meta-analysis of epidemiologic studies. PloS One. 2014;9:e92738. doi: 10.1371/journal.pone.0092738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhu B, Zou L, Han J, et al. Parity and pancreatic cancer risk: evidence from a meta-analysis of twenty epidemiologic studies. Sci Rep. 2014;4:5313. doi: 10.1038/srep05313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tang B, Lv J, Li Y, et al. Relationship between female hormonal and menstrual factors and pancreatic cancer: a meta-analysis of observational studies. Medicine (Baltimore) 2015;94:e177. doi: 10.1097/MD.0000000000000177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hassan MM, Bondy ML, Wolff RA, et al. Risk factors for pancreatic cancer: case-control study. Am J Gastroenterol. 2007;102:2696–2707. doi: 10.1111/j.1572-0241.2007.01510.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lucenteforte E, Zucchetto A, Bosetti C, et al. Reproductive and hormonal factors and pancreatic cancer risk in women. Pancreas. 2011;40:460–463. doi: 10.1097/MPA.0b013e31820bf986. [DOI] [PubMed] [Google Scholar]
- 19.Duell EJ, Holly EA. Reproductive and menstrual risk factors for pancreatic cancer: a population-based study of San Francisco Bay Area women. Am J Epidemiol. 2005;161:741–747. doi: 10.1093/aje/kwi104. [DOI] [PubMed] [Google Scholar]
- 20.Duell EJ, Maisonneuve P, Baghurst PA, et al. Menstrual and reproductive factors and pancreatic cancer in the SEARCH program of the IARC. Cancer Causes Control. 2009;20:1757–1762. doi: 10.1007/s10552-009-9408-x. [DOI] [PubMed] [Google Scholar]
- 21.Anderson LN, Cotterchio M, Gallinger S. Lifestyle, dietary, and medical history factors associated with pancreatic cancer risk in Ontario, Canada. Cancer Causes Control. 2009;20:825–834. doi: 10.1007/s10552-009-9303-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Olson SH, Orlow I, Simon J, et al. Allergies, variants in IL-4 and IL-4R alpha genes, and risk of pancreatic cancer. Cancer Detect Prev. 2007;31:345–351. doi: 10.1016/j.cdp.2007.10.002. [DOI] [PubMed] [Google Scholar]
- 23.Ji BT, Hatch MC, Chow WH, et al. Anthropometric and reproductive factors and the risk of pancreatic cancer: a case-control study in Shanghai, China. Int J Cancer. 1996;66:432–437. doi: 10.1002/(SICI)1097-0215(19960516)66:4<432::AID-IJC4>3.0.CO;2-X. [DOI] [PubMed] [Google Scholar]
- 24.Risch HA, Lu L, Kidd MS, et al. Helicobacter pylori seropositivities and risk of pancreatic carcinoma. Cancer Epidemiol Biomarkers Prev. 2014;23:172–178. doi: 10.1158/1055-9965.EPI-13-0447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Urayama KY, Holcatova I, Janout V, et al. Body mass index and body size in early adulthood and risk of pancreatic cancer in a central European multicenter case-control study. Int J Cancer. 2011;129:2875–2884. doi: 10.1002/ijc.25959. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kalapothaki V, Tzonou A, Hsieh CC, et al. Tobacco, ethanol, coffee, pancreatitis, diabetes mellitus, and cholelithiasis as risk factors for pancreatic carcinoma. Cancer Causes Control. 1993;4:375–382. doi: 10.1007/BF00051341. [DOI] [PubMed] [Google Scholar]
- 27.Chavez-MacGregor M, Elias SG, Onland-Moret NC, et al. Postmenopausal breast cancer risk and cumulative number of menstrual cycles. Cancer Epidemiol Biomarkers Prev. 2005;14:799–804. doi: 10.1158/1055-9965.EPI-04-0465. [DOI] [PubMed] [Google Scholar]
- 28.Breslow NE, Day NE. Statistical methods in cancer research. The analysis of case-controls studies. Vol. 1. Lyon, France: IARC; 1980. [PubMed] [Google Scholar]
- 29.Smith-Warner SA, Spiegelman D, Ritz J, et al. Methods for pooling results of epidemiologic studies: the pooling project of prospective studies of diet and cancer. Am J Epidemiol. 2006;163:1053–1064. doi: 10.1093/aje/kwj127. [DOI] [PubMed] [Google Scholar]
- 30.Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539–1558. doi: 10.1002/sim.1186. [DOI] [PubMed] [Google Scholar]
- 31.Galbraith RF. Graphical display of estimates having differing standard errors. Technometrics. 1988;30:271–281. [Google Scholar]
- 32.Kapoor D, Jones TH. Smoking and hormones in health and endocrine disorders. Eur J Endocrinol. 2005;152:491–499. doi: 10.1530/eje.1.01867. [DOI] [PubMed] [Google Scholar]
- 33.Preziosi G, Oben JA, Fusai G. Obesity and pancreatic cancer. Surg Oncol. 2014;23:61–71. doi: 10.1016/j.suronc.2014.02.003. [DOI] [PubMed] [Google Scholar]
- 34.Olson MB, Shaw LJ, Kaizar EE, et al. Obesity distribution and reproductive hormone levels in women: a report from the NHLBI-sponsored WISE study. J Womens Health (Larchmt) 2006;15:836–842. doi: 10.1089/jwh.2006.15.836. [DOI] [PubMed] [Google Scholar]
- 35.Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. Anthropometric and reproductive variables and exocrine carcinoma of the pancreas: a population-based case-control study in the Netherlands. Int J Cancer. 1992;52:24–29. doi: 10.1002/ijc.2910520106. [DOI] [PubMed] [Google Scholar]
- 36.Gaudet MM, Gapstur SM, Sun J, et al. Oophorectomy and hysterectomy and cancer incidence in the cancer prevention study-II nutrition cohort. Obstet Gynecol. 2014;123:1247–1255. doi: 10.1097/AOG.0000000000000270. [DOI] [PubMed] [Google Scholar]
- 37.Lee E, Horn-Ross PL, Rull RP, et al. Reproductive factors, exogenous hormones, and pancreatic cancer risk in the CTS. Am J Epidemiol. 2013;178:1403–1413. doi: 10.1093/aje/kwt154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Duell EJ, Travier N, Lujan-Barroso L, et al. Menstrual and reproductive factors in women, genetic variation in CYP17A1, and pancreatic cancer risk in the European prospective investigation into cancer and nutrition (EPIC) cohort. Int J Cancer. 2013;132:2164–2175. doi: 10.1002/ijc.27875. [DOI] [PubMed] [Google Scholar]
- 39.Prizment AE, Anderson KE, Hong C-P, et al. Pancreatic cancer incidence in relation to female reproductive factors: iowa women’s health study. JOP. 2007;8:16–27. [PubMed] [Google Scholar]
- 40.Navarro Silvera SA, Miller AB, Rohan TE. Hormonal and reproductive factors and pancreatic cancer risk: a prospective cohort study. Pancreas. 2005;30:369–374. doi: 10.1097/01.mpa.0000160301.59319.ba. [DOI] [PubMed] [Google Scholar]
- 41.Skinner HG, Michaud DS, Colditz GA, et al. Parity, reproductive factors, and the risk of pancreatic cancer in women. Cancer Epidemiol Biomarkers Prev. 2003;12:433–438. [PubMed] [Google Scholar]
- 42.Lo AC, Soliman AS, El-Ghawalby N, et al. Lifestyle, occupational, and reproductive factors in relation to pancreatic cancer risk. Pancreas. 2007;35:120–129. doi: 10.1097/mpa.0b013e318053e7d3. [DOI] [PubMed] [Google Scholar]
- 43.Lacey JV, Colditz GA, Schottenfeld D. In: Cancer Epidemiology and Prevention. Schottenfeld D, Fraumeni JF, editors. Vol. 3. New York: Oxford University Press; 2006. pp. 468–488. [Google Scholar]
- 44.Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the women’s health initiative randomized controlled trial. JAMA. 2002;288:321–333. doi: 10.1001/jama.288.3.321. [DOI] [PubMed] [Google Scholar]
- 45.Anderson GL, Limacher M, Assaf AR, et al. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the women’s health initiative randomized controlled trial. JAMA. 2004;291:1701–1712. doi: 10.1001/jama.291.14.1701. [DOI] [PubMed] [Google Scholar]
- 46.NIH Fact Sheets - Uterine Fibroids. [Accessed May 18, 2015];[National Institutes of Health] 2013 May 29; Available at: http://report.nih.gov/nihfactsheets/viewfactsheet.aspx?csid=50.
- 47.Hysterectomy. [Accessed May 18, 2015];[Medscape] 2015 Jan 23; Available at: http://emedicine.medscape.com/article/267273-overview#a0103.
- 48.Lacey JV, Chia VM, Rush BB, et al. Incidence rates of endometrial hyperplasia, endometrial cancer and hysterectomy from 1980 to 2003 within a large prepaid health plan. Int J Cancer. 2012;131:1921–1929. doi: 10.1002/ijc.27457. [DOI] [PubMed] [Google Scholar]
- 49.Fernandez E, La Vecchia C, D’Avanzo B, et al. Menstrual and reproductive factors and pancreatic cancer risk in women. Int J Cancer. 1995;62:11–14. doi: 10.1002/ijc.2910620104. [DOI] [PubMed] [Google Scholar]