Arch Gynecol Obstet DOI 10.1007/s00404-014-3595-8
GYNECOLOGIC ONCOLOGY
Dietary acrylamide intake and risk of endometrial cancer in prospective cohort studies Youjin Je
Received: 11 September 2014 / Accepted: 9 December 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Acrylamide has been associated with carcinogenicity in experimental animals, but potential health risks of dietary acrylamide intake and endometrial cancer in human are inconclusive. Thus, a meta-analysis of prospective cohort studies was conducted to provide a quantitative assessment of the association between dietary acrylamide intake and endometrial cancer risk. Methods PubMed database was used to identify prospective cohort studies on dietary acrylamide intake and endometrial cancer risk published up to June 2014. Since smoking is an important source of acrylamide and is inversely associated with endometrial cancer risk, the association was examined in women who never smoked as well. Multivariable relative risks (RR) adjusting for potential confounders were combined using random effects models. Results Four large prospective cohort studies were identified, which included 453,355 female participants and 2,019 endometrial cancer cases. There was no association between dietary acrylamide intake and endometrial cancer risk overall [pooled RR for high vs. low intake = 1.10; 95 % confidence interval (CI) 0.91–1.34]. High acrylamide intake, however, was significantly associated with increased risk of endometrial cancer among women who never smoked (pooled RR for high vs. low intake = 1.39; 95 % CI 1.09–1.77). In dose–response analyses, pooled RRs for an increase of 10 lg/day were 1.04 (95 % CI
Y. Je (&) Department of Food and Nutrition, Kyung Hee University, 26 Kyunghee-daero, Dongdaemun-gu, Seoul 130-701, South Korea e-mail: [email protected]
0.97–1.11) among all women and 1.11 (95 % CI 1.04–1.19) among never-smoking women. Conclusions Endometrial cancer risk was not associated with dietary acrylamide intake overall. Among women who never smoked, however, there was a significantly increased endometrial cancer risk in women who consumed high dietary acrylamide. Keywords Acrylamide Endometrial cancer Prospective cohort studies Meta-analysis
Introduction In 1994, acrylamide was evaluated by the International Agency for Research on Cancer (IARC) as a ‘probable human carcinogen (class 2A)’, based on positive evidence from animal studies and inadequate epidemiologic evidence [1]. In addition that the exposure to acrylamide primarily occurs in occupational setting and through tobacco smoke, it is also found in common human foods (e.g., potato crisps, fried potato, French fries, cookies, coffee, etc.) that are mainly formed during high-temperature cooking as part of the Maillard browning reactions [2, 3]. Some animal studies showed positive dose–response relations between acrylamide administered in drinking water and several types of cancers, especially in hormonesensitive organs including uterus [4, 5]. Endometrial cancer, especially most prevalent type-I endometrial cancer associated with unopposed estrogen exposure, has been associated with some dietary factors such as coffee consumption, dietary glycemic load, and alcohol drinking [6– 9], suggesting that dietary modification may lower endometrial cancer risk to some extent. In a meta-analysis of two prospective cohort studies, no significant association
123
Arch Gynecol Obstet
between dietary acrylamide intake and endometrial cancer risk was reported [10]. However, it is possible that residual confounding by smoking exists. Therefore, a comprehensive meta-analysis of prospective cohort studies was conducted to provide a quantitative assessment of the association between dietary acrylamide intake and endometrial cancer risk in all women and never-smoking women, separately.
Methods Search strategy and inclusion criteria PubMed database was used to identify eligible epidemiologic studies published in English through June 2014. The following terms were used in searching: ‘‘(acrylamide, diet, dietary factors or risk factors)’’ combined with ‘‘(corpus uterian or endometrial neoplasms).’’ The reference lists of original and review articles were also reviewed to identify additional eligible studies. Studies were eligible for inclusion if they met the following criteria: (1) a prospective cohort study design; (2) the exposure of interest was dietary acrylamide intake; (3) the outcome of interest was clearly defined as endometrial cancer (preferentially type-I endometrial cancer) incidence; and (4) adjusted relative risks (RRs) with 95 % confidence intervals (CIs). Authors were directly contacted to ask for the full papers when studies were not available [11]. Data extraction The following data were extracted according to the Metaanalysis of Observational Studies in Epidemiology (MOOSE) [12]: first author’s last name, year of publication, country name, cohort name, study period, number of cases, number of subjects and person-years, dietary acrylamide intake (quintiles or quartiles), adjustment factors, and multivariable-adjusted RRs with 95 % CIs across categories of acrylamide intake. Statistical analysis To combine the RRs from individual studies for high versus low categories of dietary acrylamide intake, random effects models that incorporate both within- and betweenstudy variations were used [13]. The study-specific RRs and pooled RR were presented as a forest plot where the size of data markers (squares) corresponds to the inverse of the variance of the natural logarithm of RR from each study, and the diamond shows pooled RR. Statistical heterogeneity among studies was evaluated using the Cochrane’s Q statistic [14], and inconsistency was quantified
123
with the I2 (I-squared) statistic [15]. To examine a dose– response relationship and calculate a pooled RR for an increase of 10 lg/day of dietary acrylamide intake, a generalized least-squares trend (GLST) estimation analysis was used [16, 17]. Subgroup analyses by geographic regions (Europe/USA) and menopausal status were conducted with tests of whether RRs varied by the characteristics using a random effects meta-regression model. A sensitivity analysis was performed by eliminating one study at a time to assess whether the results were driven by a single study. All of the analyses were repeated in never smokers to eliminate the influence of smoking. Finally, publication bias was evaluated through funnel plots (i.e., a plot of study results against precision) and with Begg’s and Egger’s tests [18, 19]. A two-tailed p value of \0.05 was considered statistically significant. All statistical metaanalyses were carried out using Stata/SE version 12.0 software (Stata Corporation, College Station, Texas).
Results Four large prospective cohort studies that met the inclusion criteria were identified, which included 453,355 female participants and 2,019 endometrial cancer cases [11, 20– 22]. The study characteristics included in the meta-analysis are summarized in Table 1. Of the four cohort studies (the Netherlands Cohort Study, NLCS; the Nurses’ Health Study, NHS; the Swedish Mammography Cohort, SMC; and the European Prospective Investigation into Cancer and Nutrition, EPIC cohort), three studies were conducted in the Europe (364,683 women; 1,535 cases) [11, 20, 21], and one study was conducted in the United States (88,672 women; 484 cases) [22]. The NLCS included postmenopausal women only, and the other three studies included both premenopausal and postmenopausal women, which provided separate RRs by menopausal status. Dietary acrylamide intakes of cohort subjects were assessed with food frequency questionnaires (FFQs). For the assessment of dietary acrylamide intake, the NHS used seven FFQs, while the remaining studies used one or two FFQs. In the NHS, cumulative average acrylamide intake was calculated using repeated measurements to best represent long-term dietary exposure to acrylamide for individuals. The four cohort studies also provided data for women who never smoked (1,030 cases). Two cohort studies provided risk estimates of dietary acrylamide intake for type-I endometrial cancer, specifically [11, 22], while the others provided data for overall endometrial cancer only [20, 21]. All studies included in the meta-analysis adjusted for potential confounders such as age, smoking, and body mass index (BMI). Three studies also adjusted for physical activity [20–22].
Netherland (NLCS)
Sweden (SMC)
USA (NHS)
Ten European countries (EPIC)
Hogervorst et al. [20]
Larsson et al. [21]
Wilson et al. [22]
Obo´nSantacann et al. [11]
1992–2010
1980–2006
1987–2007
1986–1997
Study period
Type-I EC
Type-I EC
Overall EC
Overall EC
Outcome
627/301,113 (3,303,26) 350/166,853
Never smokers
257/NA (627,668)
Never smokers All women
484/88,672 (1,386,886)
273/NA (344,580)
Never smokers All women
687/61,226 (1,080,747)
150/NA (9,422)
Never smokers All women
221/2,344 (15,836)
All women
No. of cases/subjects (or person-years)
Q5 (32.1–222.4) vs. Q1 (B14.5)
1.25 (0.79–1.98)
0.97 (0.69–1.36)
1.43 (0.90–2.28)
1.41 (1.01–1.97)
1.20 (0.76–1.90)
Q4 (C29.2) vs. Q1 (\20.5) Q5 (mean 26) vs. Q1 (9)
0.96 (0.76–1.21)
1.99 (1.12–3.52)
1.29 (0.81–2.07)
Relative risk (95 % CI)
Q4 (C28.9) vs. Q1 (\19.9)
Q5 (median 36.8) vs. Q1 (9.5)
High vs. low(ref) acrylamide intake (lg/day)
Age, center, smoking status, OC use, HRT use, total energy intake, BMI, prevalent diabetes, menopause status combined with age at menopause, parity and age at menarche
Age, smoking, BMI, menopausal status/age at menopause/PMH use, parity, OC use, high blood pressure, diabetes, physical activity, caffeine intake, energy intake
Age, education, BMI, parity, age at first birth, age at menarche, age at menopause, use of OC, use of PMH, history of diabetes, smoking status, total physical activity, energy-adjusted carbohydrate intake, and total energy intake
Age, age at menarche, age at menopause, age at first childbirth, parity, duration of OC use, duration of PMH use, BMI, height, current smoking, quantity of smoking, duration of smoking, non-occupational physical activity, energy intake, trans-unsaturated fatty acid intake, carbohydrate intake, alcohol consumption
Adjustment factors
BMI body mass index (kg/m2), EC endometrial cancer, EPIC European Prospective Investigation into Cancer and Nutrition, HRT hormone replacement therapy, NA not applicable, NHS Nurses’ Health Study, NLCS Netherlands Cohort Study, PMH postmenopausal hormone, OC oral contraceptive, SMC Swedish Mammography Cohort
Country (cohort)
References
Table 1 Characteristics of prospective cohort studies included in the meta-analysis of dietary acrylamide intake and endometrial cancer risk
Arch Gynecol Obstet
123
Arch Gynecol Obstet Study
ID
RR (95% CI)
Hogervorst et al, 2007
1.29 (0.81, 2.06)
Larsson et al, 2009
0.96 (0.76, 1.21)
Wilson et al, 2010
1.41 (1.01, 1.97)
Obon-Santacana, 2014
0.97 (0.69, 1.36)
Overall (I-squared = 31.3%, p = 0.224)
1.10 (0.91, 1.34)
NOTE: Weights are from random effects analysis
.5
1
2
3
Relative Risk (95% CI)
Fig. 1 Study-specific and pooled adjusted relative risks of endometrial cancer for high vs. low categories of dietary acrylamide intake. CI confidence interval, RR relative risk
Meta-analysis with all subjects The pooled adjusted RR comparing high (*35.2 lg/day, mean) vs. low (*10.9 lg/day, mean) categories of dietary acrylamide intake was 1.10 (95 % CI 0.91–1.34) with no significant heterogeneity among the studies (p for heterogeneity = 0.22, I2 = 31.3 %) (Fig. 1). The pooled RR from three European studies for high vs. low acrylamide intake was 1.00 (95 % CI 0.84–1.20), and there was no significant difference in RRs by country (p for Europe vs. the United States = 0.22). By menopausal status, postmenopausal women had a pooled RR of 1.15 (95 % CI 0.92–1.43), and premenopausal women had a pooled RR of 1.07 (95 % CI 0.53–1.15), and the difference in RRs by menopausal status was not statistically significant (p = 0.77). The pooled RRs in the sensitivity analysis that eliminated one study at a time did not differ substantially, all of which included a pooled RR of 1.0 in the confidence intervals, indicating that the observed null association was not driven by one single study. The dose–response analysis also showed no association overall (pooled RR for 10 lg/ day = 1.04; 95 % CI 0.97–1.11). Meta-analysis with women who never smoked Among never-smoking women, a significantly increased endometrial cancer risk was found in women consuming high dietary acrylamide. The pooled adjusted RR for high
123
vs. low categories of dietary acrylamide intake was 1.39 (95 % CI 1.09–1.77) with no significant heterogeneity (p for heterogeneity = 0.55, I2 = 0.0 %) (Fig. 2). No significant difference in RRs of studies conducted in Europe and the United States was found (p = 0.91). Based on the results of sensitivity analysis among never smokers, pooled RRs did not differ substantially, ranging from 1.29 to 1.47 (95 % CI 1.11–1.96). In the dose–response analysis, the pooled RR for an increase of 10 lg/day of dietary acrylamide intake was 1.11 (95 % CI 1.04–1.19) in neversmoking women. Publication bias There was no significant evidence of publication bias in the meta-analysis with all subjects (Begg’s p = 0.73; Egger’s p = 0.37) or women who never smoked (Begg’s p = 0.09; Egger’s p = 0.053).
Discussion In this meta-analysis of prospective cohort studies, no significant association between dietary acrylamide intake and endometrial cancer risk was found among all female subjects. However, among women who never smoked, a significantly increased endometrial cancer risk was found in women who consumed high dietary acrylamide. There
Arch Gynecol Obstet Study
ID
RR (95% CI)
Hogervorst et al, 2007
1.99 (1.12, 3.53)
Larsson et al, 2009
1.20 (0.76, 1.90)
Wilson et al, 2010
1.43 (0.90, 2.28)
Obon-Santacana, 2014
1.25 (0.79, 1.98)
Overall (I-squared = 0.0%, p = 0.547)
1.39 (1.09, 1.77)
NOTE: Weights are from random effects analysis
.5
1
2
3
4
Relative Risk (95% CI)
Fig. 2 Study-specific and pooled adjusted relative risks of endometrial cancer for high vs. low categories of dietary acrylamide intake among women who never smoked. CI confidence interval, RR relative risk
were neither evidences of significant heterogeneity among the studies nor publication bias. All of the cohort studies included in the meta-analysis tended to show positive associations among women who never smoked, but only one cohort study showed a significant positive association [20]. When the large prospective data were combined through the meta-analytic technique, among women who never smoked, there was a significantly increased endometrial cancer risk in women who consumed high dietary acrylamide by 39 %. An increment of 10 lg of dietary acrylamide intake daily was associated with 11 % increased endometrial cancer risk among women who never smoked. The positive association was not detected when all subjects were included. Cigarette smoke, an important source of acrylamide exposure, has been associated with lower endometrial cancer risk [23], and thus can act as a negative confounder for the association between dietary acrylamide intake and endometrial cancer risk if unadjusted. All studies included in the meta-analysis adjusted for smoking status in the multivariable models, but residual confounding by smoking remains possible. As a way to remove possible remaining confounding by smoking, a pooled RR of acrylamide intake was calculated in women who never smoked, and there was a significantly increased risk in women who consumed high acrylamide.
The positive association between dietary acrylamide intake and endometrial cancer risk found in never-smoking women is biologically plausible. Acrylamide is metabolized by the CYP2E1 enzyme system to glycidamide, which is a chemically reactive epoxide and mutagen in animals, and there is some evidence of carcinogenicity of acrylamide in experimental animals [4, 5, 24]. In addition, acrylamide may also be carcinogenic through hormonal pathways. Prolonged exposure to excessive estrogens, unopposed by progesterone, has been considered as a major risk factor, resulting in continued stimulation of the endometrium. It is possible that dietary acrylamide may affect activities of enzymes involved in the metabolism of sex steroid hormones and thus have effects on sex hormones [25]. To the best of my knowledge, this is the first metaanalysis of prospective cohort studies to examine dietary acrylamide intake and endometrial cancer risk among women who never smoked, specifically. All of the studies included in the meta-analysis had prospective designs, which had no methodological biases such as recall bias or selection bias found in retrospective studies, and had relatively long follow-up periods. The overall sample size of this meta-analysis was relatively large, and all the cohorts provided separate RRs among all women as well as neversmoking women. Although each study attempted to adjust
123
Arch Gynecol Obstet
for important cofounders in the multivariable models, we cannot rule out the possibility that some unknown or residual confounders may have affected the results given the observational nature of cohort studies. Out of four studies, two studies assessed dietary acrylamide intake, repeatedly [21, 22]. Since the other two studies, however, assessed the dietary acrylamide intake at baseline only [11, 20], the non-differential exposure misclassification may exist, which can attenuate a true relationship between dietary acrylamide intake and endometrial cancer risk. In conclusion, results from the meta-analysis of prospective cohort studies indicate that there was no association between dietary acrylamide intake and endometrial cancer risk overall. Among women who never smoked, however, there was a significantly increased endometrial cancer risk with increasing acrylamide intake. Further prospective studies with repeated measurements of dietary acrylamide intake, careful adjustment for all potential confounders, and subgroup analyses are needed to examine the association of dietary acrylamide intake and endometrial cancer risk. In addition, the association in never smokers could be evaluated in large consortia (data pooling) of studies, and that acrylamide or glycidamide hemoglobin adducts in blood might also help to verify the possible association observed in never-smoking women. Acknowledgments This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2014R1A1A1002736). Conflict of interest interest. Ethical standards or patient data.
The author declares that I have no conflict of
8.
9.
10.
11.
12.
13. 14. 15. 16.
The manuscript does not contain clinical studies
References 1. International Agency for Research on Cancer (IARC) (1994) IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Some industrial chemicals, vol. 60 Lyon, France 2. Stadler RH, Blank I, Varga N, Robert F, Hau J, Guy PA, Robert MC, Riediker S (2002) Acrylamide from Maillard reaction products. Nature 419:449–450 3. Lipworth L, Sonderman JS, Taronea RE, McLaughlina JK (2012) Review of epidemiologic studies of dietary acrylamide intake and the risk of cancer. Eur J Cancer Prev 21:375–386 4. Johnson KA, Gorzinski SJ, Bodner KM, Campbell RA, Wolf CH, Friedman MA, Mast RW (1986) Chronic toxicity and oncogenicity study on acrylamide incorporated in the drinking water of Fischer 344 rats. Toxicol Appl Pharmacol 85:154–168 5. Friedman MA, Dulak LH, Stedham MA (1995) A lifetime oncogenicity study in rats with acrylamide. Fundam Appl Toxicol 27:95–105 6. Je Y, Hankinson SE, Tworoger SS, Devivo I, Giovannucci E (2011) A prospective cohort study of coffee consumption and risk
123
7.
17.
18.
19.
20.
21.
22.
23.
of endometrial cancer over a 26-year follow-up. Cancer Epidemiol Biomarkers Prev 20:2487–2495 Je Y, Giovannucci E (2012) Coffee consumption and risk of endometrial cancer: findings from a large up-to-date meta-analysis. Int J Cancer 131:1700–1710 Galeone C, Augustin LS, Filomeno M, Malerba S, Zucchetto A, Pelucchi C, Montella M, Talamini R, Franceschi S, La Vecchia C (2013) Dietary glycemic index, glycemic load, and the risk of endometrial cancer: a case-control study and meta-analysis. Eur J Cancer Prev 22:38–45 Je Y, DeVivo I, Giovannucci E (2014) Long-term alcohol intake and risk of endometrial cancer in the nurses’ health study. Br J Cancer 111:186–194 Pelucchi CC, La Vecchia C, Bosetti C, Boyle P, Boffetta P (2011) Exposure to acrylamide and human cancer—a review and metaanalysis of epidemiologic studies. Ann Oncol 22:1487–1499 Obo´n-Santacana M, Kaaks R, Slimani N, Lujan-Barroso L, Freisling H, Ferrari P, Dossus L, Chabbert-Buffet N, Baglietto L, Fortner RT, Boeing H, Tjønneland A, Olsen A, Overvad K, Mene´ndez V, Molina-Montes E, Larran˜aga N, Chirlaque MD, Ardanaz E, Khaw KT, Wareham N, Travis RC, Lu Y, Merritt MA, Trichopoulou A, Benetou V, Trichopoulos D, Saieva C, Sieri S, Tumino R, Sacerdote C, Galasso R, Bueno-de-Mesquita HB, Wirfa¨lt E, Ericson U, Idahl A, Ohlson N, Skeie G, Gram IT, Weiderpass E, Onland-Moret NC, Riboli E, Duell EJ (2014) Dietary intake of acrylamide and endometrial cancer risk in the European Prospective Investigation into cancer and nutrition cohort. Br J Cancer 111:987–997 Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB (2000) Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of Observational Studies in Epidemiology (MOOSE) group. JAMA 283:2008–2012 DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7:177–188 Cochran WG (1954) The combination of estimates from different experiments. Biometrics 10:101–129 Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. BMJ 327:557–560 Greenland S, Longnecker MP (1992) Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 135:1301–1309 Orsini N, Bellocco R, Greenland S (2006) Generalized least squares for trend estimation of summarized dose-response data. Stata J 6:40–57 Begg CB, Mazumdar M (1994) Operating characteristics of a rank correlation test for publication bias. Biometrics 50:1088–1101 Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in metaanalysis detected by a simple, graphical test. BMJ 315:629–663 Hogervorst JG, Schouten LJ, Konings EJ, Goldbohm RA, van den Brandt PA (2007) A prospective study of dietary acrylamide intake and the risk of endometrial, ovarian, and breast cancer. Cancer Epidemiol Biomarkers Prev 16:2304–2313 Larsson SC, Hakansson N, Akesson A, Wolk A (2009) Longterm dietary acrylamide intake and risk of endometrial cancer in a prospective cohort of Swedish women. Int J Cancer 124:1196–1199 Wilson KM, Mucci LA, Rosner BA, Willett WC (2010) A prospective study on dietary acrylamide intake and the risk for breast, endometrial, and ovarian cancer. Cancer Epidemiol Biomarkers Prev 19:2503–2515 Zhou B, Yang L, Sun Q, Cong R, Gu H, Tang N, Zhu H, Wang B (2008) Cigarette smoking and the risk of endometrial cancer: a meta-analysis. Am J Med 121:501–508
Arch Gynecol Obstet 24. Hogervorst JG, Baars BJ, Schouten LJ, Konings EJ, Goldbohm RA, van den Brandt PA (2010) The carcinogenicity of dietary acrylamide intake: a comparative discussion of epidemiological and experimental animal research. Crit Rev Toxicol 40:485–512
25. Hogervorst JG, Fortner RT, Mucci LA, Tworoger SS, Eliassen AH, Hankinson SE, Wilson KM (2013) Associations between dietary acrylamide intake and plasma sex hormone levels. Cancer Epidemiol Biomarkers Prev 22:2024–2036
123
GYNECOLOGIC ONCOLOGY
Dietary acrylamide intake and risk of endometrial cancer in prospective cohort studies Youjin Je
Received: 11 September 2014 / Accepted: 9 December 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Acrylamide has been associated with carcinogenicity in experimental animals, but potential health risks of dietary acrylamide intake and endometrial cancer in human are inconclusive. Thus, a meta-analysis of prospective cohort studies was conducted to provide a quantitative assessment of the association between dietary acrylamide intake and endometrial cancer risk. Methods PubMed database was used to identify prospective cohort studies on dietary acrylamide intake and endometrial cancer risk published up to June 2014. Since smoking is an important source of acrylamide and is inversely associated with endometrial cancer risk, the association was examined in women who never smoked as well. Multivariable relative risks (RR) adjusting for potential confounders were combined using random effects models. Results Four large prospective cohort studies were identified, which included 453,355 female participants and 2,019 endometrial cancer cases. There was no association between dietary acrylamide intake and endometrial cancer risk overall [pooled RR for high vs. low intake = 1.10; 95 % confidence interval (CI) 0.91–1.34]. High acrylamide intake, however, was significantly associated with increased risk of endometrial cancer among women who never smoked (pooled RR for high vs. low intake = 1.39; 95 % CI 1.09–1.77). In dose–response analyses, pooled RRs for an increase of 10 lg/day were 1.04 (95 % CI
Y. Je (&) Department of Food and Nutrition, Kyung Hee University, 26 Kyunghee-daero, Dongdaemun-gu, Seoul 130-701, South Korea e-mail: [email protected]
0.97–1.11) among all women and 1.11 (95 % CI 1.04–1.19) among never-smoking women. Conclusions Endometrial cancer risk was not associated with dietary acrylamide intake overall. Among women who never smoked, however, there was a significantly increased endometrial cancer risk in women who consumed high dietary acrylamide. Keywords Acrylamide Endometrial cancer Prospective cohort studies Meta-analysis
Introduction In 1994, acrylamide was evaluated by the International Agency for Research on Cancer (IARC) as a ‘probable human carcinogen (class 2A)’, based on positive evidence from animal studies and inadequate epidemiologic evidence [1]. In addition that the exposure to acrylamide primarily occurs in occupational setting and through tobacco smoke, it is also found in common human foods (e.g., potato crisps, fried potato, French fries, cookies, coffee, etc.) that are mainly formed during high-temperature cooking as part of the Maillard browning reactions [2, 3]. Some animal studies showed positive dose–response relations between acrylamide administered in drinking water and several types of cancers, especially in hormonesensitive organs including uterus [4, 5]. Endometrial cancer, especially most prevalent type-I endometrial cancer associated with unopposed estrogen exposure, has been associated with some dietary factors such as coffee consumption, dietary glycemic load, and alcohol drinking [6– 9], suggesting that dietary modification may lower endometrial cancer risk to some extent. In a meta-analysis of two prospective cohort studies, no significant association
123
Arch Gynecol Obstet
between dietary acrylamide intake and endometrial cancer risk was reported [10]. However, it is possible that residual confounding by smoking exists. Therefore, a comprehensive meta-analysis of prospective cohort studies was conducted to provide a quantitative assessment of the association between dietary acrylamide intake and endometrial cancer risk in all women and never-smoking women, separately.
Methods Search strategy and inclusion criteria PubMed database was used to identify eligible epidemiologic studies published in English through June 2014. The following terms were used in searching: ‘‘(acrylamide, diet, dietary factors or risk factors)’’ combined with ‘‘(corpus uterian or endometrial neoplasms).’’ The reference lists of original and review articles were also reviewed to identify additional eligible studies. Studies were eligible for inclusion if they met the following criteria: (1) a prospective cohort study design; (2) the exposure of interest was dietary acrylamide intake; (3) the outcome of interest was clearly defined as endometrial cancer (preferentially type-I endometrial cancer) incidence; and (4) adjusted relative risks (RRs) with 95 % confidence intervals (CIs). Authors were directly contacted to ask for the full papers when studies were not available [11]. Data extraction The following data were extracted according to the Metaanalysis of Observational Studies in Epidemiology (MOOSE) [12]: first author’s last name, year of publication, country name, cohort name, study period, number of cases, number of subjects and person-years, dietary acrylamide intake (quintiles or quartiles), adjustment factors, and multivariable-adjusted RRs with 95 % CIs across categories of acrylamide intake. Statistical analysis To combine the RRs from individual studies for high versus low categories of dietary acrylamide intake, random effects models that incorporate both within- and betweenstudy variations were used [13]. The study-specific RRs and pooled RR were presented as a forest plot where the size of data markers (squares) corresponds to the inverse of the variance of the natural logarithm of RR from each study, and the diamond shows pooled RR. Statistical heterogeneity among studies was evaluated using the Cochrane’s Q statistic [14], and inconsistency was quantified
123
with the I2 (I-squared) statistic [15]. To examine a dose– response relationship and calculate a pooled RR for an increase of 10 lg/day of dietary acrylamide intake, a generalized least-squares trend (GLST) estimation analysis was used [16, 17]. Subgroup analyses by geographic regions (Europe/USA) and menopausal status were conducted with tests of whether RRs varied by the characteristics using a random effects meta-regression model. A sensitivity analysis was performed by eliminating one study at a time to assess whether the results were driven by a single study. All of the analyses were repeated in never smokers to eliminate the influence of smoking. Finally, publication bias was evaluated through funnel plots (i.e., a plot of study results against precision) and with Begg’s and Egger’s tests [18, 19]. A two-tailed p value of \0.05 was considered statistically significant. All statistical metaanalyses were carried out using Stata/SE version 12.0 software (Stata Corporation, College Station, Texas).
Results Four large prospective cohort studies that met the inclusion criteria were identified, which included 453,355 female participants and 2,019 endometrial cancer cases [11, 20– 22]. The study characteristics included in the meta-analysis are summarized in Table 1. Of the four cohort studies (the Netherlands Cohort Study, NLCS; the Nurses’ Health Study, NHS; the Swedish Mammography Cohort, SMC; and the European Prospective Investigation into Cancer and Nutrition, EPIC cohort), three studies were conducted in the Europe (364,683 women; 1,535 cases) [11, 20, 21], and one study was conducted in the United States (88,672 women; 484 cases) [22]. The NLCS included postmenopausal women only, and the other three studies included both premenopausal and postmenopausal women, which provided separate RRs by menopausal status. Dietary acrylamide intakes of cohort subjects were assessed with food frequency questionnaires (FFQs). For the assessment of dietary acrylamide intake, the NHS used seven FFQs, while the remaining studies used one or two FFQs. In the NHS, cumulative average acrylamide intake was calculated using repeated measurements to best represent long-term dietary exposure to acrylamide for individuals. The four cohort studies also provided data for women who never smoked (1,030 cases). Two cohort studies provided risk estimates of dietary acrylamide intake for type-I endometrial cancer, specifically [11, 22], while the others provided data for overall endometrial cancer only [20, 21]. All studies included in the meta-analysis adjusted for potential confounders such as age, smoking, and body mass index (BMI). Three studies also adjusted for physical activity [20–22].
Netherland (NLCS)
Sweden (SMC)
USA (NHS)
Ten European countries (EPIC)
Hogervorst et al. [20]
Larsson et al. [21]
Wilson et al. [22]
Obo´nSantacann et al. [11]
1992–2010
1980–2006
1987–2007
1986–1997
Study period
Type-I EC
Type-I EC
Overall EC
Overall EC
Outcome
627/301,113 (3,303,26) 350/166,853
Never smokers
257/NA (627,668)
Never smokers All women
484/88,672 (1,386,886)
273/NA (344,580)
Never smokers All women
687/61,226 (1,080,747)
150/NA (9,422)
Never smokers All women
221/2,344 (15,836)
All women
No. of cases/subjects (or person-years)
Q5 (32.1–222.4) vs. Q1 (B14.5)
1.25 (0.79–1.98)
0.97 (0.69–1.36)
1.43 (0.90–2.28)
1.41 (1.01–1.97)
1.20 (0.76–1.90)
Q4 (C29.2) vs. Q1 (\20.5) Q5 (mean 26) vs. Q1 (9)
0.96 (0.76–1.21)
1.99 (1.12–3.52)
1.29 (0.81–2.07)
Relative risk (95 % CI)
Q4 (C28.9) vs. Q1 (\19.9)
Q5 (median 36.8) vs. Q1 (9.5)
High vs. low(ref) acrylamide intake (lg/day)
Age, center, smoking status, OC use, HRT use, total energy intake, BMI, prevalent diabetes, menopause status combined with age at menopause, parity and age at menarche
Age, smoking, BMI, menopausal status/age at menopause/PMH use, parity, OC use, high blood pressure, diabetes, physical activity, caffeine intake, energy intake
Age, education, BMI, parity, age at first birth, age at menarche, age at menopause, use of OC, use of PMH, history of diabetes, smoking status, total physical activity, energy-adjusted carbohydrate intake, and total energy intake
Age, age at menarche, age at menopause, age at first childbirth, parity, duration of OC use, duration of PMH use, BMI, height, current smoking, quantity of smoking, duration of smoking, non-occupational physical activity, energy intake, trans-unsaturated fatty acid intake, carbohydrate intake, alcohol consumption
Adjustment factors
BMI body mass index (kg/m2), EC endometrial cancer, EPIC European Prospective Investigation into Cancer and Nutrition, HRT hormone replacement therapy, NA not applicable, NHS Nurses’ Health Study, NLCS Netherlands Cohort Study, PMH postmenopausal hormone, OC oral contraceptive, SMC Swedish Mammography Cohort
Country (cohort)
References
Table 1 Characteristics of prospective cohort studies included in the meta-analysis of dietary acrylamide intake and endometrial cancer risk
Arch Gynecol Obstet
123
Arch Gynecol Obstet Study
ID
RR (95% CI)
Hogervorst et al, 2007
1.29 (0.81, 2.06)
Larsson et al, 2009
0.96 (0.76, 1.21)
Wilson et al, 2010
1.41 (1.01, 1.97)
Obon-Santacana, 2014
0.97 (0.69, 1.36)
Overall (I-squared = 31.3%, p = 0.224)
1.10 (0.91, 1.34)
NOTE: Weights are from random effects analysis
.5
1
2
3
Relative Risk (95% CI)
Fig. 1 Study-specific and pooled adjusted relative risks of endometrial cancer for high vs. low categories of dietary acrylamide intake. CI confidence interval, RR relative risk
Meta-analysis with all subjects The pooled adjusted RR comparing high (*35.2 lg/day, mean) vs. low (*10.9 lg/day, mean) categories of dietary acrylamide intake was 1.10 (95 % CI 0.91–1.34) with no significant heterogeneity among the studies (p for heterogeneity = 0.22, I2 = 31.3 %) (Fig. 1). The pooled RR from three European studies for high vs. low acrylamide intake was 1.00 (95 % CI 0.84–1.20), and there was no significant difference in RRs by country (p for Europe vs. the United States = 0.22). By menopausal status, postmenopausal women had a pooled RR of 1.15 (95 % CI 0.92–1.43), and premenopausal women had a pooled RR of 1.07 (95 % CI 0.53–1.15), and the difference in RRs by menopausal status was not statistically significant (p = 0.77). The pooled RRs in the sensitivity analysis that eliminated one study at a time did not differ substantially, all of which included a pooled RR of 1.0 in the confidence intervals, indicating that the observed null association was not driven by one single study. The dose–response analysis also showed no association overall (pooled RR for 10 lg/ day = 1.04; 95 % CI 0.97–1.11). Meta-analysis with women who never smoked Among never-smoking women, a significantly increased endometrial cancer risk was found in women consuming high dietary acrylamide. The pooled adjusted RR for high
123
vs. low categories of dietary acrylamide intake was 1.39 (95 % CI 1.09–1.77) with no significant heterogeneity (p for heterogeneity = 0.55, I2 = 0.0 %) (Fig. 2). No significant difference in RRs of studies conducted in Europe and the United States was found (p = 0.91). Based on the results of sensitivity analysis among never smokers, pooled RRs did not differ substantially, ranging from 1.29 to 1.47 (95 % CI 1.11–1.96). In the dose–response analysis, the pooled RR for an increase of 10 lg/day of dietary acrylamide intake was 1.11 (95 % CI 1.04–1.19) in neversmoking women. Publication bias There was no significant evidence of publication bias in the meta-analysis with all subjects (Begg’s p = 0.73; Egger’s p = 0.37) or women who never smoked (Begg’s p = 0.09; Egger’s p = 0.053).
Discussion In this meta-analysis of prospective cohort studies, no significant association between dietary acrylamide intake and endometrial cancer risk was found among all female subjects. However, among women who never smoked, a significantly increased endometrial cancer risk was found in women who consumed high dietary acrylamide. There
Arch Gynecol Obstet Study
ID
RR (95% CI)
Hogervorst et al, 2007
1.99 (1.12, 3.53)
Larsson et al, 2009
1.20 (0.76, 1.90)
Wilson et al, 2010
1.43 (0.90, 2.28)
Obon-Santacana, 2014
1.25 (0.79, 1.98)
Overall (I-squared = 0.0%, p = 0.547)
1.39 (1.09, 1.77)
NOTE: Weights are from random effects analysis
.5
1
2
3
4
Relative Risk (95% CI)
Fig. 2 Study-specific and pooled adjusted relative risks of endometrial cancer for high vs. low categories of dietary acrylamide intake among women who never smoked. CI confidence interval, RR relative risk
were neither evidences of significant heterogeneity among the studies nor publication bias. All of the cohort studies included in the meta-analysis tended to show positive associations among women who never smoked, but only one cohort study showed a significant positive association [20]. When the large prospective data were combined through the meta-analytic technique, among women who never smoked, there was a significantly increased endometrial cancer risk in women who consumed high dietary acrylamide by 39 %. An increment of 10 lg of dietary acrylamide intake daily was associated with 11 % increased endometrial cancer risk among women who never smoked. The positive association was not detected when all subjects were included. Cigarette smoke, an important source of acrylamide exposure, has been associated with lower endometrial cancer risk [23], and thus can act as a negative confounder for the association between dietary acrylamide intake and endometrial cancer risk if unadjusted. All studies included in the meta-analysis adjusted for smoking status in the multivariable models, but residual confounding by smoking remains possible. As a way to remove possible remaining confounding by smoking, a pooled RR of acrylamide intake was calculated in women who never smoked, and there was a significantly increased risk in women who consumed high acrylamide.
The positive association between dietary acrylamide intake and endometrial cancer risk found in never-smoking women is biologically plausible. Acrylamide is metabolized by the CYP2E1 enzyme system to glycidamide, which is a chemically reactive epoxide and mutagen in animals, and there is some evidence of carcinogenicity of acrylamide in experimental animals [4, 5, 24]. In addition, acrylamide may also be carcinogenic through hormonal pathways. Prolonged exposure to excessive estrogens, unopposed by progesterone, has been considered as a major risk factor, resulting in continued stimulation of the endometrium. It is possible that dietary acrylamide may affect activities of enzymes involved in the metabolism of sex steroid hormones and thus have effects on sex hormones [25]. To the best of my knowledge, this is the first metaanalysis of prospective cohort studies to examine dietary acrylamide intake and endometrial cancer risk among women who never smoked, specifically. All of the studies included in the meta-analysis had prospective designs, which had no methodological biases such as recall bias or selection bias found in retrospective studies, and had relatively long follow-up periods. The overall sample size of this meta-analysis was relatively large, and all the cohorts provided separate RRs among all women as well as neversmoking women. Although each study attempted to adjust
123
Arch Gynecol Obstet
for important cofounders in the multivariable models, we cannot rule out the possibility that some unknown or residual confounders may have affected the results given the observational nature of cohort studies. Out of four studies, two studies assessed dietary acrylamide intake, repeatedly [21, 22]. Since the other two studies, however, assessed the dietary acrylamide intake at baseline only [11, 20], the non-differential exposure misclassification may exist, which can attenuate a true relationship between dietary acrylamide intake and endometrial cancer risk. In conclusion, results from the meta-analysis of prospective cohort studies indicate that there was no association between dietary acrylamide intake and endometrial cancer risk overall. Among women who never smoked, however, there was a significantly increased endometrial cancer risk with increasing acrylamide intake. Further prospective studies with repeated measurements of dietary acrylamide intake, careful adjustment for all potential confounders, and subgroup analyses are needed to examine the association of dietary acrylamide intake and endometrial cancer risk. In addition, the association in never smokers could be evaluated in large consortia (data pooling) of studies, and that acrylamide or glycidamide hemoglobin adducts in blood might also help to verify the possible association observed in never-smoking women. Acknowledgments This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2014R1A1A1002736). Conflict of interest interest. Ethical standards or patient data.
The author declares that I have no conflict of
8.
9.
10.
11.
12.
13. 14. 15. 16.
The manuscript does not contain clinical studies
References 1. International Agency for Research on Cancer (IARC) (1994) IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Some industrial chemicals, vol. 60 Lyon, France 2. Stadler RH, Blank I, Varga N, Robert F, Hau J, Guy PA, Robert MC, Riediker S (2002) Acrylamide from Maillard reaction products. Nature 419:449–450 3. Lipworth L, Sonderman JS, Taronea RE, McLaughlina JK (2012) Review of epidemiologic studies of dietary acrylamide intake and the risk of cancer. Eur J Cancer Prev 21:375–386 4. Johnson KA, Gorzinski SJ, Bodner KM, Campbell RA, Wolf CH, Friedman MA, Mast RW (1986) Chronic toxicity and oncogenicity study on acrylamide incorporated in the drinking water of Fischer 344 rats. Toxicol Appl Pharmacol 85:154–168 5. Friedman MA, Dulak LH, Stedham MA (1995) A lifetime oncogenicity study in rats with acrylamide. Fundam Appl Toxicol 27:95–105 6. Je Y, Hankinson SE, Tworoger SS, Devivo I, Giovannucci E (2011) A prospective cohort study of coffee consumption and risk
123
7.
17.
18.
19.
20.
21.
22.
23.
of endometrial cancer over a 26-year follow-up. Cancer Epidemiol Biomarkers Prev 20:2487–2495 Je Y, Giovannucci E (2012) Coffee consumption and risk of endometrial cancer: findings from a large up-to-date meta-analysis. Int J Cancer 131:1700–1710 Galeone C, Augustin LS, Filomeno M, Malerba S, Zucchetto A, Pelucchi C, Montella M, Talamini R, Franceschi S, La Vecchia C (2013) Dietary glycemic index, glycemic load, and the risk of endometrial cancer: a case-control study and meta-analysis. Eur J Cancer Prev 22:38–45 Je Y, DeVivo I, Giovannucci E (2014) Long-term alcohol intake and risk of endometrial cancer in the nurses’ health study. Br J Cancer 111:186–194 Pelucchi CC, La Vecchia C, Bosetti C, Boyle P, Boffetta P (2011) Exposure to acrylamide and human cancer—a review and metaanalysis of epidemiologic studies. Ann Oncol 22:1487–1499 Obo´n-Santacana M, Kaaks R, Slimani N, Lujan-Barroso L, Freisling H, Ferrari P, Dossus L, Chabbert-Buffet N, Baglietto L, Fortner RT, Boeing H, Tjønneland A, Olsen A, Overvad K, Mene´ndez V, Molina-Montes E, Larran˜aga N, Chirlaque MD, Ardanaz E, Khaw KT, Wareham N, Travis RC, Lu Y, Merritt MA, Trichopoulou A, Benetou V, Trichopoulos D, Saieva C, Sieri S, Tumino R, Sacerdote C, Galasso R, Bueno-de-Mesquita HB, Wirfa¨lt E, Ericson U, Idahl A, Ohlson N, Skeie G, Gram IT, Weiderpass E, Onland-Moret NC, Riboli E, Duell EJ (2014) Dietary intake of acrylamide and endometrial cancer risk in the European Prospective Investigation into cancer and nutrition cohort. Br J Cancer 111:987–997 Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB (2000) Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of Observational Studies in Epidemiology (MOOSE) group. JAMA 283:2008–2012 DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7:177–188 Cochran WG (1954) The combination of estimates from different experiments. Biometrics 10:101–129 Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. BMJ 327:557–560 Greenland S, Longnecker MP (1992) Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 135:1301–1309 Orsini N, Bellocco R, Greenland S (2006) Generalized least squares for trend estimation of summarized dose-response data. Stata J 6:40–57 Begg CB, Mazumdar M (1994) Operating characteristics of a rank correlation test for publication bias. Biometrics 50:1088–1101 Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in metaanalysis detected by a simple, graphical test. BMJ 315:629–663 Hogervorst JG, Schouten LJ, Konings EJ, Goldbohm RA, van den Brandt PA (2007) A prospective study of dietary acrylamide intake and the risk of endometrial, ovarian, and breast cancer. Cancer Epidemiol Biomarkers Prev 16:2304–2313 Larsson SC, Hakansson N, Akesson A, Wolk A (2009) Longterm dietary acrylamide intake and risk of endometrial cancer in a prospective cohort of Swedish women. Int J Cancer 124:1196–1199 Wilson KM, Mucci LA, Rosner BA, Willett WC (2010) A prospective study on dietary acrylamide intake and the risk for breast, endometrial, and ovarian cancer. Cancer Epidemiol Biomarkers Prev 19:2503–2515 Zhou B, Yang L, Sun Q, Cong R, Gu H, Tang N, Zhu H, Wang B (2008) Cigarette smoking and the risk of endometrial cancer: a meta-analysis. Am J Med 121:501–508
Arch Gynecol Obstet 24. Hogervorst JG, Baars BJ, Schouten LJ, Konings EJ, Goldbohm RA, van den Brandt PA (2010) The carcinogenicity of dietary acrylamide intake: a comparative discussion of epidemiological and experimental animal research. Crit Rev Toxicol 40:485–512
25. Hogervorst JG, Fortner RT, Mucci LA, Tworoger SS, Eliassen AH, Hankinson SE, Wilson KM (2013) Associations between dietary acrylamide intake and plasma sex hormone levels. Cancer Epidemiol Biomarkers Prev 22:2024–2036
123
