|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 22, No 13 (July 1), 2004: pp. 2754-2755 © 2004 American Society of Clinical Oncology. DOI: 10.1200/JCO.2004.99.064
In Reply:Cancer Genetics Program, Division of Hematology/Oncology, Department of Medicine, Department of Preventive Medicine, and Robert H. Lurie Comprehensive Cancer Center, Feinberg School of Medicine, Northwestern University, Chicago, IL We appreciate Dr Lai's insightful comment about the issue of population heterogeneity in our meta-analysis of seven case-control studies.1 There is always some heterogeneity of results of studies included in a meta-analysis, and it is unlikely that true effects are identical in varying populations of patients, especially when both European and US populations are included. The results of our first meta-analysis1 have been updated in a subsequent report.2 As Dr Lai indicates, the main analysis in the first report found that with seven studies, the odds ratio for any type of mutation, using allele as the unit of analysis, was 1.26, P = .0062 (the article erroneously indicated P < .001) with a 95% CI of 1.07 to 1.49 using the fixed-effects model. The Breslow-Day test is significant (P < .025), and the analysis of the same data using the random-effects model is not significant (P = .38). The corresponding analyses for the second meta-analysis indicate an odds ratio of 1.24, P = .0005, 95% CI of 1.10 to 1.40. The Breslow-Day test for heterogeneity for the 12 studies is marginally significant, with P = .048; for a significant effect using the random-effects model, P = .072. If we remove the first positive report (defined as the earliest study to achieve P < .05), an approach commonly used for genetic-association studies,3 the meta-analysis for the remaining 11 studies has an odds ratio of 1.17, P = .022, 95% CI 1.02 to 1.34; the Breslow-Day test for heterogeneity has P = .12, and the random-effects analysis for a significant effect has P = .20. Several points need to be considered when assessing the validity of a meta-analysis: (1) is there evidence that support the hypothesized effect, and did the hypothesis precede rather than follow the analysis? In our case, the hypothesis was based on the preceding in vitro evidence that epithelial cells that contain TGFBR1*6A grow more rapidly than cells that contain TGFBR1 when treated with transforming growth factor ß, one of the most potent naturally occurring inhibitors of cell growth.4,5 (2) Is the magnitude of the effect large? Our meta-analysis of 12 case-control studies shows that overall cancer risk is increased by 19% for TGFBR1*6A heterozygotes and 70% for TGFBR1*6A homozygotes. Breast cancer risk is increased by 38% for TGFBR1*6A carriers, which is in the same range as the increased breast cancer risk observed in women who take hormone replacement therapy.6 TGFBR1*6A homozygotes have a 102% increased risk of colorectal cancer and 169% increased risk of ovarian cancer, a non-negligible effect. (3) Is there a publication bias? We addressed this issue in the first meta-analysis to the best of our ability using the funnel-plot analysis suggested by Egger et al.7 We didn't find any evidence of publication bias.1 Our second meta-analysis of 12 studies included five additional unpublished studies, and only three studies showed an association between TGFBR1*6A and cancer,2 further ruling out a publication bias as an explanation for our results. In summary, our data are in line with those recently reported by Lohmueller et al3 in their meta-analysis of genetic-association studies. Although there was significant heterogeneity in our meta-analyses, evidence for heterogeneity disappeared after we removed the first positive study. Thus, there is only limited evidence of heterogeneity across the 12 studies, and population-specific effects seem unlikely to explain the association between TGFBR1*6A and cancer. Authors' Disclosures of Potential Conflicts of Interest The authors indicated no potential conflicts of interest. REFERENCES
1. Kaklamani VG, Hou N, Bian Y, et al: TGFBR1*6A and Cancer Risk: A Meta-Analysis of Seven Case-Control Studies. J Clin Oncol 21:3236-3243, 2003
2. Pasche B, Kaklamani VG, Hou N, et al: TGFBR1*6A and Cancer: A Meta-Analysis of 12 Case-Control Studies. J Clin Oncol 22:756-758, 2004 3. Lohmueller KE, Pearce CL, Pike M, et al: Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet 33:177-182, 2003[CrossRef][Medline]
4. Pasche B, Kolachana P, Nafa K, et al: T beta R-I(6A) is a candidate tumor susceptibility allele. Cancer Res 59:5678-5682, 1999 5. Chen T, de Vries EG, Hollema H, et al: Structural alterations of transforming growth factor-beta receptor genes in human cervical carcinoma. Int J Cancer 82:43-51, 1999[CrossRef][Medline]
6. Nelson HD, Humphrey LL, Nygren P, et al: Postmenopausal Hormone Replacement Therapy: Scientific Review. JAMA 288:872-881, 2002
7. Egger M, Davey SG, Schneider M, et al: Bias in meta-analysis detected by a simple, graphical test. BMJ 315:629-634, 1997
Related Article
Related Correspondence
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
