Association between leukocyte telomere length and glioma risk: a case-control study

Abstract

Background: Compelling epidemiological evidence indicates that alterations of telomere length are associated with risks of many malignancies in a tumor-specific manner, such as lung cancer, breast cancer, and non-Hodgkin's lymphoma. However, the association between leukocyte telomere length and glioma risk has not been investigated.

Methods: Relative telomere length (RTL) of peripheral blood leukocytes from 467 glioma patients and 467 healthy controls, matched by age and sex, was measured using the real-time PCR-based method in a case-control study. An unconditional multivariate logistic regression model was applied to estimate the association between RTL and glioma risk.

Results: Glioma patients showed notably longer RTL than controls (median, 0.555 vs 0.444; P > .04). RTL was negatively correlated with age in both cases (ρ = -0.430; P < .001) and controls (ρ = -0.388; P < .001). After adjusting for age, sex, smoking status and family history of cancer, multivariate logistic regression analysis showed that there was a U-shaped association between RTL and glioma risk (P for nonlinearity <.001). Compared with individuals in the second tertile of RTL, the odds ratios (95% CI) for participants in the first and third tertiles were 2.16 (range, 1.52-3.09) and 3.51 (range, 2.45-5.00), respectively. Stratified analysis showed that the association between RTL and glioma risk was not modulated by major host characteristics.

Conclusions: Our study demonstrates for the first time that either shorter or longer RTL in peripheral blood leukocytes is associated with increased glioma risk, which warrants further investigation in the future.

Keywords: glioma risk; logistic regression model; real-time PCR; telomere length.

Fig. 1.

Spearman's correlation analysis between relative telomere length and age in glioma cases (A) and healthy controls (B). There were negative correlations between RTL and age in participants of both groups.

Fig. 2.

Risk of glioma as estimated by selected characteristics. Odds radios (ORs) were calculated by logistic regression analysis with adjustments for age (years, continuous variable), sex, smoking status, family history of cancer, and relative telomere length, where appropriate. Solid squares indicated study-specific ORs; horizontal lines indicated 95% confidence intervals; dotted vertical line showed the OR value of reference.

Fig. 3.

Dose-dependent effect of RTL on glioma risk. RTL was transformed into natural logarithm and then standardized using the mean and standard deviation of RTL in control group. There was a U-shaped relationship between RTL and glioma risk (P for nonlinearity <.001).