Circulating cell-free DNA level predicts all-cause mortality independent of other predictors in the Health 2000 survey

Abstract

Increased levels of circulating cell-free DNA (cf-DNA) are associated with and predict poor health outcomes. However, its predictive ability for mortality in population-based samples remains understudied. We analysed the capability of cf-DNA to predict all-cause mortality and assessed whether it adds predictive value on top of the other risk factors in the Health 2000 survey (n = 1,257, 46-76 years of age, 15-years-follow-up, 18% deceased). When analysed in a multivariate model with the other factors that independently predicted mortality in the sample (age, gender, self-rated health, smoking and plasma levels of glucose and adiponectin), increases in cf-DNA levels were associated with increased risk of mortality (hazard ratio [HR] for 0.1 µg increase in cf-DNA: 1.017, 95% confidence interval [CI] 1.008-1.026, p = 0.0003). Inclusion of cf-DNA in the model improved the model fit and discrimination. Stratifying the analysis by cardiovascular disease (CVD) status indicated that cf-DNA predicted mortality equally well in individuals with (HR 1.018, 95% CI 1.008-1.026, p = 0.002) and without (HR 1.018, 95% CI 1.001-1.035, p = 0.033) CVD. In conclusion, our study indicates that cf-DNA level predicts mortality in middle-aged and older individuals, also among those with established CVD, and adds significant value to mortality prediction. Our results thus underscore the role of cf-DNA as a viable marker of health.

Figure 1

The estimated survival probabilities according to the baseline cf-DNA level divided into two groups, and pairwise comparison statistics between the groups. Individuals in the highest gender-wise cf-DNA quartile (n = 313, 80 [26%] deceased) were included to the group of “elevated cf-DNA levels” (black line) and all the other individuals (n = 944, 148 [16%] deceased) in the group of “cf-DNA level is in the normal range” (green line).