Objective
To offer a general figure of the available data on the relation between alcohol intake and risk of endometriosis, we conducted a systematic review and a metaanalysis of studies published up to May 2012.
Study Design
We carried out a literature search of all case-control and cohort studies published as original articles in English up to May 2012. Only those papers that were published as full-length articles were considered. Pooled estimates of the relative risks (RRs) and the corresponding 95% confidence intervals (CIs) were calculated using fixed or, when significant heterogeneity among estimates emerged, random effects models. A total of 15 studies were identified for the review.
Results
The summary estimate was 1.24 (95% CI, 1.12–1.36) for any alcohol intake vs no alcohol intake. Considering the results of the analyses of infrequent, moderate/regular, and heavy alcohol intake vs no alcohol intake, the summary RR estimates were, respectively, 1.14 (95% CI, 0.86–1.52), 1.23 (95% CI, 1.08–1.40), and 1.19 (95% CI, 0.99–1.43). Three studies reported separate results for current and former drinkers, and the summary RR were 1.42 (95% CI, 1.14–1.76) and 1.09 (95% CI, 0.83–1.43), respectively.
Conclusion
The present metaanalysis provides evidence for an association between alcohol consumption and endometriosis risk. Further studies are needed to clarify whether alcohol consumption may exacerbate an existing disease or could be related to the severity of the disease.
Endometriosis is an estrogen-dependent, chronic inflammatory gynecological condition. Despite the high prevalence that has been estimated between 6-10% in reproductive-age women, and the recognized economic burden associated with the disease, potential modifiable risk factors of the disease are still to be completely elucidated.
Alcohol has been consistently found to increase the risk of developing estrogen-dependent diseases such as breast cancer. It has been demonstrated that alcohol intake increases circulating bioavailable estrogen level and this is believed to be one of the mechanisms underlying the association between alcohol consumption and estrogen-dependent diseases. Alcohol increases aromatase activity, ie, the conversion of testosterone to estrogens that results in reduced testosterone and increased estrogens. Alcohol may also interact with luteinizing hormone production from the pituitary gland, resulting in increased estradiol release from the ovaries.
On the other hand, long-term alcohol intake may also affect immune function and may regulate production of proinflammatory cytokines. A significant association between alcohol consumption and selected chronic inflammatory diseases has been observed.
Along these lines, several studies have analyzed the association between alcohol drinking and risk of endometriosis but conflicting results have been published regarding the potential effect. To offer a general figure of the available data on the relation between alcohol intake and risk of endometriosis, we conducted a systematic review and a metaanalysis of studies published up to May 2012.
Materials and Methods
Identification of studies
We carried out a literature search of all case-control and cohort studies published as original articles in English up to May 2012. We searched the electronic databases MEDLINE (1966 through May 31, 2012), EMBASE (1985 through May 31, 2012), and Science Citation Index Expanded (1945 through May 31, 2012) using the Medical Subject Heading terms “diet” or “nutrition” or “alcohol” or “vitamin” or “fat” or “vegetable,” combined with “endometriosis.” Only those papers that were published as full-length articles in English language were considered. Furthermore, we reviewed reference lists of retrieved articles to search for other pertinent studies.
Two authors reviewed the papers and independently selected the articles eligible for the systematic review. Studies were selected for review if they met all of the following criteria:
- ▪
case-control or cohort study reporting original data;
- ▪
diagnosis of endometriosis was clinically and/or histologically based;
- ▪
number or percentage of subjects with and without endometriosis according to alcohol intake were provided;
- ▪
full-length articles, published in English.
If multiple published reports from a same study were available, we included only the one with the most detailed information, or published more recently.
One study was excluded because the level of alcohol intake was provided as Michigan Alcoholism Screening Test score <5, 5-6, or ≥7, where a Michigan Alcoholism Screening Test score >5 is suggestive of alcoholism. Thus, for this study the reference category would not be comparable with other studies included (ie, we could not disentangle subjects with no or low alcohol consumption). The selection procedure is shown in Figure 1 .
Data collection for metaanalysis
Data were extracted independently by 2 investigators and discrepancies were resolved by discussion. For each study, the following information was collected: first author’s last name; year of publication; country of origin; study design; number of subjects; age, if available; category amounts of alcohol intake, if available; site of endometriosis for cases, if available; relative risks (RRs), hazard ratios or odds ratios (ORs) of endometriosis and corresponding 95% confidence intervals (CIs) for every category of alcohol intake; and covariates adjusted in the statistical analysis.
Statistical analysis
We combined the RR estimates from each study. We computed unadjusted RR from the exposure distributions of cases and controls as reported in the publications. For the study by Heilier et al, we used the method proposed by Hamling et al, which allows combining the estimates originally shown in the paper, changing the reference category and taking into account the correlation between categories. For the study by Signorello et al, we summed the data concerning fertile and infertile controls ( Figures 2-4 ) other than for the subgroup analyses ( Figure 6 ). Three separate estimates were given in the Figures 2, and 4-6 for the study by Hemmings et al, according to different subgroups of surgery type (ie, diagnostic laparoscopy, tubal ligation, or hysterectomy).
The inverse variance method was used to pool the RR. A fixed effect model was used as the combination method. To assess the heterogeneity across studies we conducted a test based on the χ 2 distribution. When heterogeneity was significant (ie, P < .05), the pooled estimate was calculated using random effects model analysis. The Egger test and funnel plot were used to detect publication bias. In this graph, the effect measure estimates are plotted against corresponding sample size and, if there is no publication bias, it should have the shape of a funnel with a wide dispersion of results among small studies and a narrower range of results for large ones.
Subgroups analyses
We performed the subgroups’ analysis by type of controls (fertile, infertile, both, or not specified).
The statistical analyses were performed using software (STATA, version 10.0; StataCorp LP, College Station, TX).
Results
Figure 1 shows a flowchart for selection of articles. A total of 15 studies were identified for the review.
The main methodological characteristics of identified papers are presented in Table 1 . Most of them were retrospective case-control studies. One cohort study was identified. A total of 5 studies were conducted in the United States, 4 in Europe, 2 in Canada, 1 in the United States and Canada, 1 in Japan, 1 in Australia, and 1 in Taiwan. Simple “No” or “Yes” questions were used in 6 studies to evaluate alcohol intake.
Study | Country | Study design | Cases | Controls | Sample size cases/controls | Age, y | Confounding factors considered in analysis | Alcohol consumption assessment as reported by authors |
---|---|---|---|---|---|---|---|---|
Berubé et al, 1998 | Canada | Case-control on baseline data collected in a prospective study | Infertile women for minimal or mild endometriosis (laparoscopically diagnosed) | Women with unexplained infertility | 329/262 | 20-39 | 0, 1-2, 3-8, ≥9 drinks/mo | |
Buck Louis et al, 2007 | United States | Hospital-based case-control in cohort of women undergoing laparoscopy | Women with endometriosis in cohort undergoing laparoscopy for any gynecologic indication including sterilization | Women without endometriosis from same cohort of cases | 32/52 | 18-40 | In utero exposure, age, parity, smoking habit, caffeine intake | None, 1-4, ≥5 drinks/mo |
Eskenazi et al, 2002 | Italy | Cohort study | Women with endometriosis confirmed by surgery or ultrasound examination | Women without endometriosis confirmed by surgery or negative ultrasound examination | 19/277 | ≤30 in 1976 | Yes for polytomous analysis | Never, former, current |
Grodstein et al, 1994 | United States, Canada | Hospital-based case-control | Women with primary infertility due to endometriosis | Fertile women | 180/3833 | Age, center, smoking habits, lifetime no. of sexual partners, contraception, BMI, exercise, coffee | No consumption, moderate (≤100 g/wk), heavier (>100 g/wk) | |
Heilier et al, 2007 | Belgium | Matched case-control | Women with PE or DEN | Women with no clinical suspicion of PE or DEN; without infertility, pelvic pain, dysmenorrhea; with normal pelvic examination and vaginal echography; with serum CA-125 <35 U/mL | 88 (PE), 88 (DEN)/88 | None | Never, <1 time/wk, several times/wk, every day | |
Hemmings et al, 2004 | Canada | Hospital-based case-control | Women with endometriotic lesions at time of surgery (surgery for diagnosis, fertility-regulating surgery, hysterectomy) | Women with no evidence of endometriotic lesion at surgery (surgery for diagnosis, fertility-regulating surgery, hysterectomy) | 896/1881 | Premenopausal age | None, <7, ≥7 drinks/wk | |
Huang et al, 2010 | Taiwan | Case-control | Women with endometriosis | Women without endometriosis | 28/29 | Mean: cases = 34.3, controls = 36.2 | Backward selection of confounders | No, yes |
Marino et al, 2009 | United States | Case-control | Women with surgically confirmed endometriosis from GH cooperative | Women without endometriosis randomly selected from list of GH during same period | 341/742 | 18-49 | None | Never, former, current |
Matalliotakis et al, 2008 | United States | Case-control in retrospective review | Women with pelvic endometriosis who had undergone laparoscopy or laparotomy for pelvic pain or infertility within 6 y | Infertile women (tubal or male factor infertility) | 535/200 | 15-56 | None | No/yes |
Nagle et al, 2009 | Australia | Case-control | Women with surgically confirmed endometriosis | Women without endometriosis | 268/244 | 18-55 | None | Never/occasional, infrequent, regular (defined by author and not otherwise specified) |
Parazzini et al, 2004 | Italy | Hospital-based case-control (from 2 studies) | Women with laparoscopically confirmed endometriosis | Women admitted for acute nongynecological, nonhormonal, nonneoplastic conditions | 504/504 | 20-65 | Age, calendar year, education, parity, BMI, study | <0.5, 0.5-8, ≥8 drinks/wk (thresholds based on tertiles of control group) |
Pauwels et al, 2001 | Belgium, The Netherlands | Case-control | Infertile endometriosis women | Mechanical infertile women | 42/27 | 24-42 | Age, BMI, ovulatory disfunction, smoking pattern, caffeine consumption | <6, ≥6 drinks/wk |
Signorello et al, 1997 | United States | Hospital-based case-control | Women with infertility-associated endometriosis | 89 fertile women and 47 infertile women both without endometriosis | 50/(89 and 47) | 23-44 | Age, education, height, weight, regularity of menstrual cycle, exercise smoking | None, any, <once/wk, ≥once/wk; alcohol consumption was assessed for overall consumption and for each type of drink: beer, liquor, red wine, white wine |
Trabert et al, 2011 | United States | Population-based case-control | Women with endometriosis ( ICD-9 = 617.0, .5, .8, .9) from GH cooperative | Women without endometriosis from GH during same period | 284/660 | 18-49 | None | Never, present, former |
Tsukino et al, 2005 | Japan | Case-control | Women with stage II-IV endometriosis | Women without endometriosis or stage I endometriosis | 58/81 | 20-45 | Menstrual regularity, average cycle (d) | None/occasionally, weekly, daily |
Table 2 reports the cutoffs of alcohol drinking in different studies according to the classification levels used in this metaanalysis. Cutoffs used in different studies were fairly homogeneous, except for a somewhat lower cutoff for moderate/regular alcohol consumption in the study by Parazzini et al and for heavy alcohol consumption in the study by Berubé et al.
Study | Infrequent | Moderate/regular | Heavy |
---|---|---|---|
Berubé et al, 1998 | 1-2 drinks/mo | 3-8 drinks/mo | ≥9 drinks/mo |
Buck Louis et al, 2007 | 1-4 drinks/mo | ≥5 drinks/mo | |
Grodstein et al, 1994 | ≤100 g/wk = <1 drink/d = <30 drinks/mo | >100 g/wk = ≥1 drink/d = ≥30 drinks/mo | |
Heilier et al, 2007 | <Once/wk = <4 drinks/mo | Several times/wk = 4-29 drinks/mo | Every day = ≥30 drinks/mo |
Hemmings et al, 2004 | <7 drinks/wk = <30 drinks/mo | ≥7 drinks/wk = ≥30 drinks/mo | |
Nagle et al, 2009 | a | a | |
Parazzini et al, 2004 | 0.5-8 drinks/wk b | ≥8 drinks/wk b | |
Pauwels et al, 2001 | ≥6 drinks/wk = ≥24 drinks/mo | ||
Signorello et al, 1997 | <Once/wk = <4 drinks/mo | ≥Once/wk = ≥4 drinks/mo | |
Tsukino et al, 2005 | Weekly | Daily |
a Categories defined “infrequent” and “regular” by authors
b Tertile of intake, reference category: <0.5 drinks/wk, pure alcohol content was assumed in each type of drink (125 mL wine = 333 mL beer = 30 mL spirits).
Systematic review
Figure 2 shows the study-specific and pooled RR for any vs no alcohol intake. All but 2 of the RR estimates were above unity ranging from 0.34 to 2.28. The summary estimate was 1.24 (95% CI, 1.12–1.36). In a sensitivity analysis, we excluded results from 1 study that provided prevalence ORs, and we found a summary RR of 1.23 (95% CI, 1.11–1.36).
Figures 3-5 present, respectively, the results of the analyses of infrequent, moderate/regular, and heavy alcohol intake vs no alcohol intake. The summary RR estimates were, respectively, 1.14 (95% CI, 0.86–1.52), 1.23 (95% CI, 1.08–1.40), and 1.19 (95% CI, 0.99–1.43). We also performed 2 sensitivity analyses. One by excluding the study where heavy drinkers were women consuming relatively low amounts (ie, ≥9 drinks/mo) as compared to other studies ; the estimate for heavy vs no alcohol consumption was 1.14 (95% CI, 0.93–1.39). Another sensitivity analysis was performed by excluding the study where moderate/regular drinkers were women consuming relatively low amounts (ie, 0.5-8 drinks/wk) as compared to other studies ; the estimate for moderate/regular vs no alcohol consumption was 1.29 (95% CI, 1.12–1.49). When we performed subgroup analysis by type of controls, the estimates were 1.35 (95% CI, 0.97–1.87) for studies based on infertile controls, 1.50 (95% CI, 1.15–1.95) for those of fertile controls, and 1.19 (95% CI, 1.06–1.33) for those of both fertile and infertile or unspecified controls.
Only 1 study analyzed separately the role of different types of alcoholic beverages on endometriosis risk. White wine consumption (OR, 3.5; 95% CI, 1.2–10.4; and OR, 1.5; 95% CI, 0.4–4.9 [fertile and infertile controls, respectively]) but not red wine (OR, 1.3; 95% CI, 0.5–3.7; and OR, 0.8; 95% CI, 0.3–2.5 [fertile and infertile controls, respectively]), beer (OR, 1.0; 95% CI, 0.4–2.5; and OR, 0.8; 95% CI, 0.3–2.4 [fertile and infertile controls, respectively]), and liquor (OR, 0.9; 95% CI, 0.9–2.2; and OR, 1.8; 95% CI, 0.6–5.3 [fertile and infertile controls, respectively]) consumption were associated with the risk of endometriosis.
Figure 7 shows the funnel plot for any vs no alcohol consumption. There was no asymmetry in the funnel plot, thus suggesting the absence of publication bias; the Egger test was not significant ( P = .902).