|Meta-analysis: the use of non-steroidal anti-inflammatory drugs and pancreatic cancer risk for different exposure categories
|Capurso G, Schunemann H J, Terrenato I, Moretti A, Koch M, Muti P, Capurso L, Delle Fave G
The review determined the relationship between the use of aspirin/non-steroidal anti-inflammatory drugs (ASA/NSAIDs) and the risk of developing pancreatic cancer at different exposure categories. The authors' conclusion, that no association between ASA/NSAIDs and pancreatic ductal adenocarcinoma was found, is likely to be reliable but further studies, particularly in European populations, are needed to clarify the findings.
To evaluate the relationship between the use of aspirin and non-steroidal anti-inflammatory drugs (ASA/NSAIDs) and the risk of developing pancreatic ductal adenocarcinoma (PDAC) at different exposure categories.
MEDLINE, EMBASE and Scopus (all from inception until April 2007) and the Cochrane Database of Systematic Reviews (up to November 2006) were searched without language restrictions; the search terms were reported. To identify further relevant articles, a citation search was conducted using Scopus, proceedings from conferences of the American Gastroenterology Association (2001 to 2007) and the American Society of Clinical Oncology (2001 to 2006) were searched, and the references of identified papers were checked.
Eligible study designs were randomised controlled trials (RCTs), cohort or case-control studies. Studies that examined exposure to ASA (including low-dose ASA) and/or NSAIDs (defined as NSAIDs other than ASA) and reported the incidence of pancreatic cancer or death related to pancreatic cancer were eligible for inclusion. Five of the included studies evaluated ASA exposure; the remaining studies evaluated both ASA and NSAID exposure. Exposure was categorised as high, intermediate or low use. The primary outcomes included incidence, mortality and risk of pancreatic cancer, general cancer mortality, general cancer and digestive cancer risk. Studies were required to report the odds ratio (OR) or risk ratio with confidence intervals (CIs), and original raw data sufficient to evaluate the hypothesized effect. Most of the included studies were conducted in the USA; one study was conducted in the UK. The overall use of ASA/NSAIDs in the study populations ranged from 43 to 72% in the U.S. studies and was approximately 20% in the UK study.
Two reviewers independently selected papers for the review. Any disagreements were resolved through discussion with a third reviewer.
Assessment of study quality
Only cohort and case studies were evaluated for quality. Each study was assessed against the following criteria: an explicit statement of the research question, description of cases and controls, definition of exposure, information on data collection, confounders and methods of analysis.
Two reviewers independently assessed the quality of the included studies.
In addition to the standard information relating to study design, population characteristics and intervention, incidence and mortality rates were extracted in order to calculate the OR. In addition, the authors extracted details of known covariates. Contact with the primary authors was made in an attempt to obtain missing information where necessary. Full details of the data extracted were given in the paper.
Two reviewers independently extracted the data from the primary studies. Any disagreements were resolved by consultation with a third reviewer.
Methods of synthesis
The pooled OR and corresponding 95% CIs were calculated using random-effects meta-analysis. Publication bias was assessed using a funnel plot. Statistical heterogeneity was assessed using the the I2 statistic. A value of 25% or lower was considered to indicate low heterogeneity, whereas a value of 75% or more was considered to indicate substantial heterogeneity. The sensitivity analysis was defined a priori. Aspects relating to differences in exposure, population, types of outcome measure and study methodology were explored as potential sources of differences between the studies.
Results of the review
Eight studies were included in the review: 1 RCT, 4 cohort studies and 3 case-control studies. Seven studies (n=6,301) were included in the main analysis.
In 7studies the outcome was blinded to exposure. Four studies reported adequate group description, had sufficient duration of follow-up and conducted a dose-response analysis. Three studies reported comparison of confounders and had a loss to follow-up greater than 25%. Two studies adjusted for confounding factors and treatment was ascertained in one study.
No statistically significant association of ASA/NSAIDs exposure was found with PDAC in the low-exposure (OR 0.99, 95% CI: 0.83, 1.19), intermediate-exposure (OR 1.11, 95% CI: 0.84, 1.47) or high-exposure (OR 1.09, 95% CI: 0.67, 1.75) groups. Considerable heterogeneity was found: I2 ranged from 60 to 86%. Sensitivity analysis by ASA use lone, study design, or gender (women only) did not significantly alter these findings. A small protective effect was found in the low-exposure group for men (OR 0.79, 95% CI: 0.71, 0.88). Funnel plots indicated possible publication bias in all three categories (data not reported).
No association between ASA/NSAIDs and PDAC was found, although the large baseline exposure in North American controls may have obscured the results. Further studies, particularly in Europe, are needed to clarify the findings.
The review was supported by inclusion criteria relating to the study design, intervention and outcome. Several relevant sources were searched without language restrictions to locate pertinent papers. However, there was no apparent attempt to locate unpublished material, which means that relevant studies might have been missed. Steps were taken to minimise reviewer error or bias in all stages of the review process. The validity of the non-randomised studies was assessed and the results reported. Standard statistical methods were used to pool the data and potential sources of heterogeneity were assessed. However, pooling in the face of such heterogeneity may be questionable. The authors' conclusion is likely to be reliable, although the results might not be generalisable to populations outside of the USA.
Implications of the review for practice and research
Practice: The authors did not state any implications for practice.
Research: The authors suggest that, owing to the high baseline use of ASA/NSAIDs in North America, any association between the use of ASA/NSAIDs and PDAC should also be evaluated in prospective or well-designed retrospective studies in European populations, where both obesity and analgesic use are less common than in the USA. In addition, the authors suggest that a pooled analysis of RCTs of ASA/NSAIDs for the prevention of cancer and cardiovascular disease, taking account of known risk factors for pancreatic cancer, should be undertaken.
Italian Ministry of University; University of Roma 'La Sapienza'; European Comission Grant.
Capurso G, Schunemann H J, Terrenato I, Moretti A, Koch M, Muti P, Capurso L, Delle Fave G. Meta-analysis: the use of non-steroidal anti-inflammatory drugs and pancreatic cancer risk for different exposure categories. Alimentary Pharmacology and Therapeutics 2007; 26(8): 1089-1099
Subject indexing assigned by NLM
Anti-Inflammatory Agents, Non-Steroidal /administration & Carcinoma, Pancreatic Ductal; Dose-Response Relationship, Drug; Female; Humans; Male; Outcome Assessment (Health Care); Pancreatic Neoplasms /drug therapy /mortality; Risk Factors; dosage /adverse effects
Date bibliographic record published
Date abstract record published
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.