The association of telomere length and cancer: a meta-analysis

Abstract

Background: Telomeres shorten with each cell division and are essential for chromosomal stability. Short telomeres in surrogate tissues (e.g., blood cells) are associated with increased cancer risk in several case-control studies, but findings are inconsistent in prospective studies.

Methods: We systematically reviewed studies published prior to August 30, 2010, on the association between telomere length (TL) in surrogate tissues and cancer. There were 27 reports on 13 cancers and/or incident cancer investigating this association. The majority, 16, were retrospective case--control studies, 11 were prospective studies. Meta-analyses were conducted to determine ORs and 95% CIs for these studies.

Results: Studies on bladder, esophageal, gastric, head and neck, ovarian, renal, and overall incident cancer found associations between short telomeres and these cancers. Non-Hodgkin lymphoma, breast, lung, and colorectal cancer reports were inconsistent. Single studies on endometrial, prostate, and skin cancers were null. In a random-effects meta-analysis, short TL was significantly associated with cancer in retrospective studies (pooled OR for the shortest TL quartile compared with the longest: 2.9, 95% CI: 1.75-4.8, P < 0.0001). The pooled OR for prospective studies was 1.16 (95% CI: 0.87-1.54, P = 0.32). All studies combined yielded a pooled OR of 1.96 (95% CI: 1.37-2.81, P = 0.0001) for the association of short TL and cancer.

Conclusion and impact: There is suggestive evidence that short surrogate tissue TL is associated with cancer; the strongest evidence exists for bladder, esophageal, gastric, and renal cancers. Additional prospective studies with consistent methodology are needed to confirm this hypothesis.

Figure 1. Association between Telomere Length and Risk of Cancer in Surrogate Tissues

Association findings for TL are presented categorized into quartiles or tertiles and comparisons between the shortest and longest (referent) group. The graph was created using study-specific, adjusted odds ratios and 95% confidence intervals with SigmaPlot Version 11.0. (37) include two independent case-control series. (11) included bladder, lung, and renal cell carcinoma cases; (14) and (19) reported a significant association of longer TL and risk of breast cancer and NHL, respectively; here we plot the inverse log odds ratios. (37) Case-control study at Lombardi Comprehensive Cancer Center; (38) Chromosome 9-specific OR; (37) Case-control study at Roswell Park Cancer Institute; (25) Prospective EPIC study; (25) Retrospective SEARCH study; (25) Prospective EPIC study; (25) Retrospective SEARCH study; Abbreviations: CRC, colorectal cancer; HNC, head and neck cancer; NHL, non-Hodgkin lymphoma; SCC, squamous cell carcinoma; BCC, basal cell carcinoma.

Figure 2. Meta-Analysis of the Association Between Telomere Length and Cancer

Studies included in the meta-analysis were selected based on the following criteria: DNA derived from blood and buccal cells, and analysis by quartiles of TL among controls. Unadjusted odds ratios were used to conduct analysis (exception: Pooley et al. [25]: ORs are adjusted for batch to avoid confounding). Significant study heterogeneity was present (I²=94.3%, P<0.001), so the random effects meta-analysis was used. The weight percents are from the random effects analysis. The dashed line indicates the OR of the metaanalysis. (14), (19), and (30) reported a significant association of longer TL and risk of breast cancer and NHL, respectively; we used the inverse log odds ratios to compute the summary OR. The forest plot was created using Stata version 11.0 (StataCorp, College Station, TX). Abbreviations: OR, odds ratio; CI, confidence interval; NHL, non-Hodgkin Lymphoma.

Figure 3. Meta-Analysis of Telomere Length and Cancer in Retrospective studies

The plot was created using Stata version 11.0 (StataCorp, College Station, TX). Svenson et al. and Gramatges et al. found an association with longer telomeres; here inverse log odds ratios were used to compute the summary OR. The dashed line indicates the pooled OR of the analysis. Because there was significant heterogeneity described by the I²-statistic between estimates across studies, we present results from the random effects model. Abbreviations: OR, odds ratio; NHL, non-Hodgkin lymphoma.

Figure 4. Meta-Analysis of Telomere Length and Cancer in Prospective studies

The plot was created using Stata version 11.0 (StataCorp, College Station, TX). Lan et al. found an association with longer telomeres and NHL; here inverse log odd ratios were used to compute the summary OR. The weight percents are from the random effects analysis. Heterogeneity was described using the I²-statistic, the approximate proportion of total variability in the point estimates that can be attributed to heterogeneity. The dashed line indicates the pooled estimate of the analysis. Han et al. reported on each skin cancer entity (melanoma, squamous cell, and basal cell carcinoma) separately. We computed and used a summary OR for all skin cancers in this analysis to avoid under-representing heterogeneity. Abbreviations: OR, odds ratio; NHL, non-Hodgkin lymphoma.