Performance of risk prediction scores for cardiovascular mortality in older persons: External validation of the SCORE OP and appraisal

Development of a prediction model for mortality and cardiovascular outcomes in older adults taking into account AZGP1

Zinc-alpha 2-glycoprotein (AZGP1) is a serum protein with postulated functions in metabolism, cancer and cardiovascular disease. We developed new prediction models for mortality or cardiovascular events investigating the predictive potential of serum AZGP1 in a community-based cohort of older adults. We measured AZGP1 (μg/ml) in stored serum samples of 930 individuals of the Berlin Initiative Study, a prospective, population-based cohort of adults aged ≥ 70. We determined the prognostic potential of 20 knowledge-based predictors including AZGP1 for the outcomes of mortality or the composite endpoint of death and cardiovascular events (stroke, myocardial infarction (MI)) using Cox models; their model fit was evaluated with calibration plots, goodness-of-fit tests and c-indices. During median follow-up of 48.3 months, 70 incident strokes, 38 incident MI and 234 deaths occurred. We found no associations or correlations between AZGP1 and other candidate variables. After multivariable Cox regression with backward-selection AZGP1 remained in both models for mortality (HR = 0.44, 95%CI: 0.24–0.80) and for the composite endpoint (HR = 0.43, 95%CI: 0.23–0.82). Within newly built prediction models, we found that increased AZGP1 levels were predictive for lower risk of mortality and the composite endpoint in older adults. AZGP1 as a predictor warrants further validation in older adults.

Kidney Function as Risk Factor and Predictor of Cardiovascular Outcomes and Mortality Among Older Adults

RATIONALE & OBJECTIVE

Estimated glomerular filtration rate (eGFR) and urinary albumin-creatinine ratio (UACR) are associated with cardiovascular events in the general population but their utility among older adults is unclear. We investigated the associations of eGFR and UACR with stroke, myocardial infarction (MI), and death among older adults.

STUDY DESIGN

Population-based cohort study.

SETTING & PARTICIPANTS

1,581 participants (aged≥70 years) in the Berlin Initiative Study (BIS) without prior stroke or MI.

EXPOSURES & PREDICTORS

Serum creatinine- and cystatin C-based eGFR, UACR categories, and measured GFR (n=436).

OUTCOMES

Stroke, MI, and all-cause mortality.

ANALYTICAL APPROACH

HRs and 95% CIs derived from multivariable-adjusted Cox proportional hazards models for association analyses. Net reclassification improvement (NRI) and C statistic differences comparing the predictive benefit of kidney measures with a traditional cardiovascular risk model.

RESULTS

During a median follow-up of 8.2 years, 193 strokes, 125 MIs, and 531 deaths occurred. Independent of UACR, when GFR was estimated using the creatinine- and cystatin C-based BIS equation, eGFR of 45 to 59mL/min/1.73m² (vs eGFR>60mL/min/1.73m²) was associated with stroke (HR, 2.23; 95% CI, 1.55-3.21) but not MI or all-cause mortality. For those with eGFR<45mL/min/1.73m², the HRs were 1.99 (95% CI, 1.23-3.20) for stroke, 1.38 (95% CI, 0.81-2.36) for MI, and 1.57 (95% CI, 1.20-2.06) for mortality. Compared with UACR<30mg/g, UACR of 30 to 300mg/g was not associated with stroke (HR, 0.91; 95% CI, 0.63-1.33) but was associated with MI (HR, 1.65; 95% CI, 1.09-2.51) and all-cause mortality (HR, 1.63; 95% CI, 1.34-1.98). Prediction analysis for stroke showed significant positive NRI for eGFR calculated using the cystatin C-based Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation and the creatinine- and cystatin C-based BIS and Full Age Spectrum equations. UACR demonstrated significant positive NRIs for MI and mortality.

LIMITATIONS

eGFR and UACR categorization based on single assessments; lack of cause-specific death data.

CONCLUSIONS

eGFR of 45 to 59mL/min/1.73m² without albuminuria was associated with stroke but not MI or all-cause mortality in older adults. In contrast, UACR of 30 to 300mg/g was associated with MI and all-cause mortality but not with stroke. Furthermore, cystatin C-based eGFR improved risk prediction for stroke in this cohort of older adults.

Correction: Performance of risk prediction scores for cardiovascular mortality in older persons: External validation of the SCORE OP and appraisal

[This corrects the article DOI: 10.1371/journal.pone.0231097.].