Comorbidity indices in observational studies on cardiovascular risk.

Age-and-sex alone showed higher AUROCs than any single comorbidity index alone for all four cardiovascular outcomes in both age groups studied.

• Study population: all individuals in Sweden born 1936–1975, n=4,454,895 individuals, followed up for outcomes 2014–2019.
• Two age groups were studied: 40–64 years and 65–79 years.
• In the age group 40–64 years, AUROCs for age-and-sex alone were: CHD (0.739), MI (0.736), HF (0.730), and stroke (0.685).
• AUROCs for age-and-sex alone were higher than those for the NMI alone and numbers of filled prescriptions alone in all analyses.

The cardiovascular risk factor indicator variable set (CV-IV) alone showed a higher AUROC for heart failure than the Nordic Multimorbidity Index (NMI) alone or numbers of filled prescriptions alone.

• CV-IV alone AUROC for HF: 0.694 (95% CI 0.690 to 0.697).
• NMI alone AUROC for HF: 0.643 (95% CI 0.639 to 0.647).
• Numbers of filled prescriptions alone AUROC for HF: 0.671 (95% CI 0.667 to 0.675).
• CV-IV consisted of a set of five cardiovascular risk factors as indicator variables.

The highest AUROC observed across all analyses was for CV-IV adjusted for age-and-sex predicting heart failure.

• CV-IV adjusted for age-and-sex AUROC for HF: 0.781 (95% CI 0.778 to 0.784).
• Numbers of filled prescriptions adjusted for age-and-sex AUROC for HF: 0.776 (95% CI 0.773 to 0.780).
• NMI adjusted for age-and-sex AUROC for HF: 0.771 (95% CI 0.768 to 0.774).
• All three comorbidity measures improved upon age-and-sex alone (AUROC 0.730) for HF prediction in the 40–64 age group.

CV-IV adjusted for age-and-sex showed a higher AUROC than age-and-sex alone for all four cardiovascular outcomes in both age groups.

• This improvement was observed for CHD, MI, HF, and stroke outcomes.
• The finding held in both the 40–64 and 65–79 year age groups.
• No other single comorbidity measure alone outperformed age-and-sex alone for any outcome.

Numbers of filled prescriptions performed comparably to more complex comorbidity indices, suggesting its potential use as a proxy measure for comorbidity.

• Numbers of filled prescriptions adjusted for age-and-sex yielded an AUROC of 0.776 (95% CI 0.773 to 0.780) for HF, close to CV-IV (0.781) and NMI (0.771).
• The authors note 'similar results were obtained for numbers of filled prescriptions implying possible use as a proxy measure for comorbidity.'
• Filled prescription data was drawn from Swedish healthcare registries with a look-back period for diagnosis data from 2010 to 2014.

The study used a large nationwide Swedish registry-based population with a defined look-back period for comorbidity ascertainment and prospective outcome follow-up.

• Study population: n=4,454,895 individuals born between 1936 and 1975.
• Diagnosis data look-back period: 2010–2014.
• Outcome follow-up period: 2014–2019.
• Comorbidity measures evaluated included the Nordic Multimorbidity Index (NMI), CV-IV, inpatient care-based measures, and numbers of filled prescriptions.
• Outcomes evaluated were coronary heart disease (CHD), myocardial infarction (MI), heart failure (HF), and stroke.