Statistical Analysis
Baseline characteristics were assessed and descriptive statistics for categorical and continuous variables were presented accordingly. To assess bivariate association with exposure, χ2 tests were used to compare categorical variables, while t tests were used to compare continuous variables by exposure status. Incidence proportions and rates were reported for each outcome of interest. Kaplan-Meier curves were constructed to assess the bivariate association between exposure and study outcomes. Cox proportional hazard modeling was performed to estimate the association between AF/AFL and time to each of the outcomes. Multiple models were designed to assess cardiac and metabolic covariates, in addition to baseline characteristics. This included a base model adjusted for age, sex, military sponsor rank and branch, geographic region, and duty status.
Additional models adjusted for cardiac and metabolic confounders and CCI score. A comprehensive model included the base, cardiac, and metabolic covariates. Multicollinearity between covariates was assessed. Variables with a variance inflation factor > 4 or a tolerance level < 0.1 were added to the models. Cox proportional hazard models were used to estimate the unadjusted and adjusted hazard ratios (HRs) and 95% CIs of the association between AF/AFL and the study outcomes. Data were analyzed using SAS, version 9.4 for Windows.
RESULTS
From 2014 through 2019, 35,539 patients with WPW syndrome were identified in the MHS, 5291 had AF/AFL (14.9%); 19,961 were female (56.2%), the mean (SD) age was 62.9 (18.0) years, and 11,742 were aged ≥ 75 years (33.0%) (Table 1).
There were 4121 (11.6%), 322 (0.9%), and 848 (2.4%) patients with AF, AFL, and both arrhythmias, respectively. The mean (SD) number of cardiology visits was 3.9 (3.0). The mean (SD) baseline CCI score for the AF/AFL subgroup was 5.9 (3.5) vs 3.7 (2.2) for the non-AF/AFL subgroup (P < .001). The most prevalent comorbid conditions were hypertension, hyperlipidemia, chronic obstructive pulmonary disease, and diabetes (P < .001) (Figure 1).
Composite Outcomes
In the overall cohort, during a mean (SD) follow-up time of 3.4 (2.0) years comprising 119,682 total person-years, the components of the composite outcome occurred 6506 times with an incidence rate of 5.44 per 100 person-years. Ventricular tachycardia was the most common event, occurring 3281 times with an incidence rate of 2.74 per 100 person-years. SCA and VF occurred 841 and 135 times with incidence rates of 0.70 and 0.11 per 100 person-years, respectively. Death was the initial event 2249 times with an incidence rate of 1.88 per 100 person-years. Figure 2 shows the Kaplan-Meier curve of cardiac composite outcome by AF/AFL status.
The subgroup with AF/AFL comprised 17,412 total person-years and 1424 cardiac composite incidences compared with 102,270 person years and 5082 incidences in the no AF/AFL group (Table 2). Comparing AF/AFL vs no AF/AFL incidence rates were 8.18 vs 4.97 per 100 person-years, respectively (P < .001). SCA and VF occurred 233 and 38 times and respectively had incidence rates of 1.34 and 0.22 per 100 person-years in the AF/AFL group vs 0.59 and 0.09 per 100 person-years in the no AF/AFL group (P < .001). There were 549 deaths and a 3.15 per 100 person-years incidence rate in the AF/AFL group vs 1700 deaths and a 1.66 incidence rate in the no AF/AFL group (P < .001).
The HR for the composite outcome in the base model was 1.33 (95% CI, 1.26-1.42, P < .001) (Table 3). The association between AF/AFL and the composite outcome remained significant after adjusting for additional metabolic and cardiac covariates. The HRs for the metabolic and cardiac models were 1.30 (95% CI, 1.23-1.38, P < .001) and 1.11 (95% CI, 1.05-1.18, P < .001), respectively. After adjusting for the full model, the HR was 1.12 (95% CI, 1.05-1.19, P < .001).