Individual co-morbidity profiles, AF + co-morbid count, and multi-morbidity in AF cohorts
The individual comorbidity profiles for AF prevalence and incidence are shown in Table 2. With the exception of sleep apnea, the prevalence of comorbidities was higher for the elderly than the younger populations (<0.0001). The prevalence ratios of the older (> 65 years) to the younger (< 65 years) were: 2.42 (95%CI 2.379-2.420) for congestive heart failure; 1.37 (95%CI 1.366-1.375) for hypertension; 1.91 (95%CI 1.868-1.906) for diabetes mellitus; 2.03 (95%CI 1.976-2.027) for history of stroke and thromboembolic events; 2.42 (95%CI 2.375-2.423) for vascular disease; 1.51 (95%CI 1.493-1.514) for valvular disease; 2.29 (95%CI 2.257-2.294) for coronary artery disease; and 3.49 (95%CI 3.545-3.639) for chronic kidney disease. For sleep apnea, the elderly had about 50% lower prevalence and incidence compared to those age < 65, with a prevalence ratio of 0.49 (95%CI 0.474-0.491).
Table 3 provides data for individual comorbidity profiles in relation to prevalent and incident AF. For Medicare patients aged ≥65, the prevalence of stroke, hypertension and valvular disease were significantly higher for females than males. There was no significant difference for congestive heart failure between males and females. Other comorbidities were significantly higher for males. For those age <65 (Commercial), there was no significant difference between males and females for valvular disease. Other comorbidities were significantly higher for males than females, except for history of stroke, which was significantly higher for females than males.
The relationship between the AF + co-morbidity count is displayed in fig 2 (parts a and b) for incident cases, and appears to be an inverted U-shape. Pure incident AF targets are greatly underestimated when one relies only on ICD 10 codes. Therefore, the use of both pharmacy and medical claims is greatest for AF targets in the absence of co-morbid history. In addition, the peak of AF + co-morbidity count is achieved in the range of 2 to 5 depending on the criterion used and case prevalence/incidence. The relationship for prevalent AF targets was slightly different from incident AF targets (fig 2 parts c and d). Relative to incident cases, prevalent cases are greatly underestimated on the basis of ICD 10 claims for pure AF and those with lower co-morbidity count, then, overestimated in the higher range of co-morbidity count. The cumulative incidence and prevalence were higher the overall population relative to the age > 65 year cohort.
The prevalence of multimorbidity in AF patients aged ≥65 using medical and pharmacy claim databases was 11.73%, when defined as AF with ≥1 other chronic long term condition. This figure was 10.14% if defined as AF plus ≥2 chronic long term conditions (Table 4). Both prevalence figures are higher than if relying on ICD claim codes alone, which would underestimate multimorbidity in this cohort. Similar patterns were seen for incidence of multimorbidity.