|
Oral contraceptives and colorectal cancer risk: a meta-analysis |
Fernandez E, La Vecchia C, Balducci A, Chatenoud L, Franceschi S, Negri E |
|
|
Authors' objectives To investigate the association between oral contraceptive (OC) use and the risk of colorectal cancer.
Searching MEDLINE and Cancerlit were searched for papers published up to June 2000. The search strategy employed synonyms of the following: 'neoplasms', 'tumours', or 'cancer of colon and/or rectum'; 'exogenous female hormones', 'oral contraceptives' or 'oestro-progestins'. The authors also reviewed reference lists in relevant papers, and contacted colleagues to update the papers included in the IARC Monograph (see Other Publications of Related Interest no.1) and a previous review (see Other Publications of Related Interest no.2). Only full papers published in the English language were considered for the review.
Study selection Study designs of evaluations included in the reviewEpidemiological studies on colorectal cancer, published as full reports in the English language, were considered. Cohort and case-control studies were included in the review. Studies were only eligible if information had been obtained from each woman participating in the study.
Specific interventions included in the reviewStudies involving any OC were eligible for inclusion in this review. The studies had to include quantitative information on OC use, and the OCs had to be distinguishable from hormone replacement and other hormonal therapies.
Participants included in the reviewWomen, aged mainly between 55 and 60 years. No other demographic details were provided in the review.
Outcomes assessed in the reviewThe incidence of cancer of the colon or rectum, or both. The outcome measure was the relative risk of cancer in women who had used OC, compared with those who had never used OC.
How were decisions on the relevance of primary studies made?The authors do not state how the papers were selected for the review, or how many of the reviewers performed the selection.
Assessment of study quality The authors did not assign a quality score to each study. No studies were excluded a priori for weakness of design or data quality. The papers were not assessed for validity.
Data extraction The following details were extracted for each study: study design, number of patients (cases and controls or person-years), prevalence of OC use, and control of confounding. Two authors abstracted the relative risks (RRs) and confidence intervals (CIs) from published papers, giving preference to those estimates adjusted for multiple confounding factors. When multivariate RRs were unavailable, these were computed from the exposure distribution, as given in the articles.
Methods of synthesis How were the studies combined?The studies were combined in a meta-analysis. The summary estimates of the RR and 95% CIs were derived using fixed-effect models. The weighted average of the estimated RRs was computed by giving each study a weight proportional to its precision. The summary estimates were calculated separately for the cohort and case-control studies, as well as in combination. Publication bias was investigated using funnel plots and Egger's test (see Other Publications of Related Interest no.3).
How were differences between studies investigated?Heterogeneity was evaluated using a chi-squared test for heterogeneity and the Galbraith plot (see Other Publications of Related Interest no.4).
Results of the review A total of 20 studies were included in the review: 6 cohort studies (n=238,862) and 14 case-control studies (total number of cases (2188) and controls (9592)).
A total of 8 case-control and 4 cohort studies provided data for calculating the RR of colorectal cancer for 'ever' versus 'never use' of OCs. After excluding one study that greatly increased the heterogeneity of these studies, the pooled RR from case-control studies was 0.72 (95% CI: 0.61, 0.85; chi-squared 12.59, d.f.=6, p=0.05). The pooled estimate from the cohort studies was 0.84 (95% CI: 0.72, 0.97; chi-squared 4.18, d.f.=3, p=0.24). The pooled estimate from all studies was 0.82 (95% CI: 0.74, 0.93). There was no heterogeneity between the case-control and the cohort studies.
For colon cancer (2 cohort and 9 case-control studies), the summary RR was 0.83 (95% CI: 0.74, 0.95; chi-squared 13.32, d.f.=10, p=0.21). For rectal cancer (1 cohort and 5 case-control studies), the summary RR was 0.74 (95% CI: 0.59, 0.93; chi-squared 11.23, d.f.=5, p=0.05). Duration of use was not related to a decrease in risk of colorectal cancer, although data from 2 studies indicated that the apparent protection was stronger for women who had used OCs more recently (RR 0.46, 95% CI: 0.30, 0.71).
Authors' conclusions The available data suggested that OC use was inversely related to the risk of colorectal cancer. Some aspects, however, remained undefined; these included the risk profile associated with duration and how recent the OCs were used, and the possibility of confounding. The issue of causal inference for the observed association is therefore still open to discussion.
CRD commentary This review addressed an appropriate question using well-defined inclusion and exclusion criteria. The search strategy included only two databases and only English language papers were considered. Therefore, the possibility of publication bias cannot be excluded. The authors state that publication bias was evaluated but no results were presented or discussed. The validity of the studies was not assessed. In addition, only a limited amount of information on the primary studies was presented in the review.
The methods used for the meta-analysis appear to have been appropriate. The pooled estimates of effect calculated were subject to heterogeneity, with not all primary studies finding a inverse relationship between OC use and colorectal cancer. Thus, the authors' conclusions should be interpreted with caution.
Implications of the review for practice and research Practice: The issue of causal inference for the observed association is still open to discussion.
Research: The issue of causal inference for the observed association is still open to discussion.
Funding Italian Association for Cancer Research, Milan.
Bibliographic details Fernandez E, La Vecchia C, Balducci A, Chatenoud L, Franceschi S, Negri E. Oral contraceptives and colorectal cancer risk: a meta-analysis. British Journal of Cancer 2001; 84(5): 722-727 Other publications of related interest 1. IARC Monographs on the evaluation of carcinogenic risks to humans. Vol 72. Hormonal contraception and postmenopausal hormone therapy. Lyon: IARC; 1999. p.180-6, 294-338. 2. Franceschi S, la Vecchia C. Oral contraceptive and risk of colorectal tumors. A review of epidemiologic studies. Contraception 1998;58:335-43.
3. Egger M, Davey-Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. Br Med J 1997;315:629-34. 4. Galbraith RF. A note on graphical presentation of estimated odds ratios from several clinical trials. Statist Med 1988;7:889-94.
Indexing Status Subject indexing assigned by NLM MeSH Case-Control Studies; Cohort Studies; Colorectal Neoplasms /epidemiology; Contraceptives, Oral; Female; Humans; Risk Factors AccessionNumber 12001000869 Date bibliographic record published 31/05/2002 Date abstract record published 31/05/2002 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|