Association Between the Multidimensional Prognostic Index and Mortality During 15 Years of Follow-up in the InCHIANTI Study

Background: Multidimensional Prognostic Index (MPI) is recognized as a prognostic tool in hospitalized patients, but data on the value of MPI in community-dwelling older persons are limited. Using data from a representative cohort of community-dwelling persons, we tested the hypothesis that MPI explains mortality during 15 years of follow-up. Methods: A standardized comprehensive geriatric assessment was used to calculate the MPI and to categorize participants in low-, moderate-, and high-risk classes. The results were reported as hazard ratios (HRs) and the accuracy was evaluated with the area under the curve (AUC), with 95% confidence intervals (CIs) and the C-index. We also reported the median survival time by standard age groups. Results: All 1453 participants (mean age 68.9 years, women = 55.8%) enrolled in the InCHIANTI study at baseline were included. Compared to low-risk group, participants in moderate (HR = 2.10; 95% CI: 1.73-2.55) and high-risk MPI group (HR = 4.94; 95% CI: 3.91-6.24) had significantly higher mortality risk. The C-index of the model containing age, sex, and MPI was 82.1, indicating a very good accuracy of this model in explaining mortality. Additionally, the time-dependent AUC indicated that the accuracy of the model incorporating MPI to age and sex was excellent (>85.0) during the whole follow-up period. Compared to participants in the low-risk MPI group across different age groups, those in moderate- and high-risk groups survived 2.9-7.0 years less and 4.3-8.9 years less, respectively. Conclusions: In community-dwelling individuals, higher MPI values are associated with higher risk of all-cause mortality with a dose-response effect.