Facilitating meta-analyses by deriving relative effect and precision estimates for alternative comparisons from a set of estimates presented by exposure level or disease category.
ABSTRACT Epidemiological studies relating a particular exposure to a specified disease may present their results in a variety of ways. Often, results are presented as estimated odds ratios (or relative risks) and confidence intervals (CIs) for a number of categories of exposure, for example, by duration or level of exposure, compared with a single reference category, often the unexposed. For systematic literature review, and particularly meta-analysis, estimates for an alternative comparison of the categories, such as any exposure versus none, may be required. Obtaining these alternative comparisons is not straightforward, as the initial set of estimates is correlated. This paper describes a method for estimating these alternative comparisons based on the ideas originally put forward by Greenland and Longnecker, and provides implementations of the method, developed using Microsoft Excel and SAS. Examples of the method based on studies of smoking and cancer are given. The method also deals with results given by categories of disease (such as histological types of a cancer). The method allows the use of a more consistent comparison when summarizing published evidence, thus potentially improving the reliability of a meta-analysis.
- SourceAvailable from: Darren C Greenwood[Show abstract] [Hide abstract]
ABSTRACT: Smoking is estimated to cause about half of all bladder cancer cases. Case-control studies have provided evidence of an inverse association between fruit and vegetable intake and bladder cancer risk. As part of the World Cancer Research/American Institute for Cancer Research Continuous Update Project, we conducted a systematic review and meta-analysis of prospective studies to assess the dose-response relationship between fruit and vegetables and incidence and mortality of bladder cancer. We searched PubMed up to December 2013 for relevant prospective studies. We conducted highest compared with lowest meta-analyses and dose-response meta-analyses using random effects models to estimate summary relative risks (RRs) and 95% confidence intervals (CIs), and used restricted cubic splines to examine possible nonlinear associations. Fifteen prospective studies were included in the review. The summary RR for an increase of 1 serving/day (80 g) were 0.97 (95% CI: 0.95-0.99) I(2) = 0%, eight studies for fruits and vegetables, 0.97 (95% CI: 0.94-1.00, I(2) = 10%, 10 studies) for vegetables and 0.98 (95% CI: 0.96-1.00, I(2) = 0%, 12 studies) for fruits. Results were similar in men and women and in current, former and nonsmokers. Amongst fruits and vegetables subgroups, for citrus fruits the summary RR for the highest compared with the lowest intake was 0.87 (95% CI: 0.76-0.99, I(2) = 0%, eight studies) and for cruciferous vegetables there was evidence of a nonlinear relationship (P = 0.001). The current evidence from cohort studies is not consistent with a role for fruits and vegetables in preventing bladder cancer. © 2014 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.Cancer Medicine 12/2014;
- [Show abstract] [Hide abstract]
ABSTRACT: Background:Alcohol is a risk factor for cancer of the oral cavity, pharynx, oesophagus, colorectum, liver, larynx and female breast, whereas its impact on other cancers remains controversial.Methods:We investigated the effect of alcohol on 23 cancer types through a meta-analytic approach. We used dose-response meta-regression models and investigated potential sources of heterogeneity.Results:A total of 572 studies, including 486 538 cancer cases, were identified. Relative risks (RRs) for heavy drinkers compared with nondrinkers and occasional drinkers were 5.13 for oral and pharyngeal cancer, 4.95 for oesophageal squamous cell carcinoma, 1.44 for colorectal, 2.65 for laryngeal and 1.61 for breast cancer; for those neoplasms there was a clear dose-risk relationship. Heavy drinkers also had a significantly higher risk of cancer of the stomach (RR 1.21), liver (2.07), gallbladder (2.64), pancreas (1.19) and lung (1.15). There was indication of a positive association between alcohol consumption and risk of melanoma and prostate cancer. Alcohol consumption and risk of Hodgkin's and Non-Hodgkin's lymphomas were inversely associated.Conclusions:Alcohol increases risk of cancer of oral cavity and pharynx, oesophagus, colorectum, liver, larynx and female breast. There is accumulating evidence that alcohol drinking is associated with some other cancers such as pancreas and prostate cancer and melanoma.British Journal of Cancer advance online publication, 25 November 2014; doi:10.1038/bjc.2014.579 www.bjcancer.com.British Journal of Cancer 11/2014; · 4.82 Impact Factor
Article: The SAS %METADOSE Macro[Show abstract] [Hide abstract]
ABSTRACT: The %metadose macro is a SAS macro for meta-analysis of linear and nonlinear dose-response relationships. It is used when research reports studying the same dose-response relationship have different exposure or treatment levels. It is a two step macro: First, for each study, it uses either the Greenland method (AJE, 1992) or Hamling method (SIM, 2008) to get estimated cell counts of the 2X2 table adjusted for counfounding, then it estimates the asymptotic correlation between the adjusted log odds ratio estimates for each exposure level relative to the referent level, from which we can get the estimated covariance matrix for these study-specific estimates. After this step, we get a single pooled estimate and its variance estimate across different exposure or treatment levels. Then, meta-analysis is performed analysis for all the studies using the single study-specific trend estimate, in common units across studies. An option also exists to explore and graph non-linearity in the poooled results.