A prospective evaluation of the Flacker-Kiely One Year Mortality Score and the added value of NT-proBNP

Abstract

Objective: Several mortality indices have been constructed to aid clinical decision making in older adults. We aimed to prospectively validate the Flacker-Kiely (FK) mortality index in a Norwegian nursing home cohort, which has not been done before, and explore whether NT-ProBNP could improve its discriminatory power.

Methods: We performed a cohort/mortality study. From November 2017 to July 2018, physicians in all public long-term nursing homes in Bergen, Norway, scored residents according to the original Flacker Kiely index. Mortality data were derived from the Norwegian Cause of Death Registry and NT-ProBNP values were obtained from routinely collected blood chemistry. An alternative FK index using the NT-ProBNP-value as a marker for the presence of heart failure was constructed (FK NT-ProBNP index). The ProBNP cut-off value was selected based on a Cox regression model (“dead/alive 1 year”/” NT-ProBNP (Ng/l)”, where the value with the highest Youden index was identified. We judged index performance by using c-statistics.

Results: Both the original FK index and the constructed FK NT-ProBNP index discriminated between risk strata. The FK NT-ProBNP index yielded a C-index of 0.66 compared to 0.62 for the original FK index. Optimal discriminatory power was shown with a NT-ProBNP cut-off value of 1,595 Ng/l as heart failure criterion, and FK NT-ProBNP score 6.6.

Conclusions: The prospective mortality estimation ability of the FK-index was comparable to previous retrospective studies. The inclusion of NT-ProBNP as a heart failure criterion strengthen the discriminatory power and utility of the index, both in clinic and administration.