Table 4. Covariate-adjusted ANA associations with mortality by cause of death, group analyzed, and analysis performed.
| Cause of Death | Group Analyzed | Analysis Performed | HR (95% CI) a | N/C/C+ |
|---|---|---|---|---|
| All Causes | Enrolled at age ≥75 years | Primary Analysis | 1.27 (0.89, 1.82) | 431/249/53 |
| Exclude participants with ENA or a possible autoimmune disease | 1.51 (1.03, 2.21) | 329/185/38 | ||
| Enrolled at age ≥75 years, | Primary Analysis | 1.99 (1.04, 3.80)b | 77/54/8 | |
| history of cancer | Exclude first 2 years of follow-up | 2.64 (1.09, 6.38)b | 65/43/7 | |
| Exclude participants with ENA or a possible autoimmune disease | 2.14 (1.07, 4.28)b | 60/42/6 | ||
| Enrolled at age ≥75 years, | Primary Analysis | 1.16 (0.78, 1.74) | 237/115/25 | |
| no history of CVD or cancer | Exclude participants with ENA or a possible autoimmune disease | 1.54 (1.01, 2.36) | 189/90/19 | |
| CVD | Enrolled at age ≥75 years | Primary Analysis | 1.69 (0.93, 3.07) | 424/52/13 |
| Exclude first 2 years of follow-up | 1.99 (1.04, 3.78)b | 379/34/9 | ||
| Enrolled at age ≥65 years | Redefine age-at-enrollment strata | 1.90 (1.05, 3.43) | 867/74/18 | |
| Cancer | Enrolled at age <75 years | Primary Analysis | 1.62 (0.71, 3.69) | 2903/75/18 |
| Enrolled at age <65 years | Redefine age-at-enrollment strata | 2.70 (1.02, 7.14) | 2454/39/11 |
Abbreviations: ANA = antinuclear antibodies; HR = hazard ratio; CI = confidence interval; N = number of participants analyzed; C = number of participants who died from the cause of interest; C+ = number of ANA-positive participants who died from the cause of interest; CVD = cardiovascular disease; BMI = body mass index.
a Each HR estimate was derived from a Cox model for cause-specific mortality, fitted within a given group of participants and focused on the effect of baseline ANA status on death from the cause of interest. The continuous event time variable for all-cause mortality was age at death from non-accidental causes, the continuous event time variable for cause-specific mortality was age at death from the cause of interest (either CVD or cancer), and baseline ANA status was represented by a dichotomous indicator variable. All analyses adjusted for age at risk (via the baseline hazard function) and included covariates for sex, race/ethnicity, education, and BMI. Analyses that did not stratify on disease history also adjusted for self-reported history of the disease of interest at baseline. Participants with missing values were excluded.
b Interpret cautiously, as these results are based on fewer than 10 events of interest.