EURURO Vol. 72 No. 5

752 860

the genetic variants on RCC risk (p = 0.08). Furthermore,
results from MR-Egger regression estimated an intercept of
–0.043 (95% CI = –0.133 to 0.047, p = 0.4), suggesting no
significant evidence for directional pleiotropy (Supplemen-
tary Fig. 3).
In analyses restricted to individual histologic subtypes,
comparable associations were observed for each of the
telomere length associated variants across RCC subtype
(Supplementary Table 2). Likewise, similar telomere length
associated GRS associations were observed for clear cell RCC
(OR = 1.93 per predicted kilobase increase, 95% CI = 1.50–
2.49, p < 0.0001; Supplementary Fig. 4), papillary RCC
(OR = 1.96, 95% CI = 1.01–3.81, p = 0.046; Supplementary
Fig. 5), and chromophobe RCC (OR = 2.37, 95% CI = 0.78–7.17,
p = 0.13; Supplementary Fig. 6), although the latter finding
did not reach statistical significance. Analyses conducted
across strata of sex, body mass index, history of hyperten-
sion, and smoking status did not identify statistically
significant evidence of effect modification by these factors
(Supplementary Figs. 7–10).
4. Discussion
Our findings suggest that an excess of telomere length-
related variants is associated with RCC risk and, in
aggregate, a genetic risk score predicting longer telomere
length in peripheral blood leukocytes is strongly associated
with increased RCC risk. The association between longer
genetically-predicted telomere length and RCC risk
remained statistically significant even after removing two
telomere length associated variants highly correlated with
GWAS-identified RCC risk variants from the telomere length
GRS, indicating additional telomere length-associated SNPs
are associated with RCC risk beyond these two potentially
influential SNPs. We observed no significant differences in
the overall telomere length GRS and RCC association across
common RCC subtypes, although our power to detect
heterogeneity in associations across subtypes was limited.
Future studies with larger collections of chromophobe and
papillary RCC cases are needed to confirm these associa-
tions with telomere length variants by subtype.
With 10 784 RCC cases and 20 406 cancer-free controls,
this study is the largest to date to assess the relationship
between telomere length and RCC risk. Rather than directly
measuring leukocyte telomere length, our study used
genetic variants highly associated with leukocyte telomere
length as a surrogate of telomere length to assess the
relationship with RCC risk. Our genetic approach has several
advantages; it is not susceptible to potential biases due to the
timing of specimen collection in relation to diagnosis,
potential confounding, or differences in preanalytical speci-
men processing.
While many lines of evidence in our analysis suggest a
clear and robust association between longer telomere length
and RCC risk, perhaps the main limitation of our approach is
in estimating the magnitude of this association. The telomere
length-associated variants used in this analysis originated
from GWAS studies of leukocyte telomere length, where
telomere length was measured by qPCR . These
studies then use correlations between qPCR-measured
telomere length and Southern blot from other laboratories
to extrapolate the base pair change in telomere length
associated with each variant allele. While these conversions
might not be entirely accurate, we chose to use kilobase
change in telomere length as weights in our telomere length
GRS to facilitate combining variants discovered in different
studies into a homogenous telomere length GRS. As such,
measurement error may be present in the reported effect
estimates; however, the association p values remain valid.
Renal epithelial cell telomere length would perhaps be
the best means to assess the relationship between telomere
length and RCC risk. Ideally, genetic surrogates of renal
epithelial cell telomere length would be available as
instruments in our current analysis, but as of publication
no genetic variants have been reported to be associated with
renal cell telomere length. A prior study has demonstrated
that telomere length measurements in leukocytes and
nonmalignant renal tissue are correlated, with a Pearson
correlation coefficient of 0.44 This relationship
between leukocyte telomere length and renal cell telomere
length suggests the most likely biological mechanism linking
increased leukocyte telomere length to RCC risk may be
longer correlated renal epithelial cell telomere length.
Longer renal telomere length may promote renal tumor
growth by increasing replicative potential of renal epithelial
cells, although further studies are needed to confirm this
hypothesis and alternative explanations are possible. If
validated, our findings indicating longer telomere length as a
risk factor for RCC may inform clinicians of potential RCC
risks associated with administering prolonged treatments
with telomerase activating properties (eg, androgen therapy)
particularly in high-risk RCC populations. Additionally,
telomere length GRSs, in combination with other genetic,
clinical, and risk factor data, may hold future clinical value for
the development and application of RCC risk prediction
models in support of a “precision prevention” paradigm of
targeted disease prevention.
5. Conclusions
Our investigation adds to the growing body of evidence
indicating some aspect of telomere length is important for
the development of a variety of common cancer types
suggesting clinicians weigh the potential increases in
cancer risk when considering treatments with telomerase
activating properties. Future studies are needed to decipher
which components of telomere biology, whether it be
telomere length, telomerase activity, or an altogether
unknown mechanism, are biologically important in onco-
genesis. Such mechanistic insight will lead to improved risk
modeling and identify potentially promising targets for
drug development.
Author contributions: Mark P. Purdue had full access to all the data in the
study and takes responsibility for the integrity of the data and the
accuracy of the data analysis.
Study concept and design: Machiela, Hofmann, Carreras-Torres, Brown,
Brennan, Chanock, Scelo, Purdue.
E U R O P E A N U R O L O GY 7 2 ( 2 0 17 ) 74 7 – 7 5 4 752