Author + information
- Mayank Sardana, MBBS,
- Priscilla Y. Hsue, MD,
- Zian H. Tseng, MD, MAS,
- Eric Vittinghoff, PhD,
- Gregory Nah, MA,
- Thomas A. Dewland, MD and
- Gregory M. Marcus, MD, MAS∗ (, )@gregorymmarcus
- ↵∗Division of Cardiology, Department of Medicine, University of California San Francisco, 505 Parnassus Avenue, M-1180B, Box 0124, San Francisco, California 94143-0124
With effective antiretroviral therapy, the life expectancy of individuals infected with human immunodeficiency virus (HIV) infection has increased. However, patients living with HIV infection remain at increased risk of cardiovascular diseases and sudden cardiac death (1). Although available evidence provides conflicting information regarding a possible relationship between HIV and atrial fibrillation (AF) (2,3), no previous study has determined if HIV infection predicts incident AF.
We used the Healthcare Cost and Utilization Project (HCUP) California State Databases to identify all California state residents (age ≥21 years) who received care in an outpatient surgery unit, emergency department, or inpatient hospital unit between January 1, 2005, and December 31, 2011. Patients entered the cohort at their first health care encounter and were followed prospectively using the unique patient identifiers for censoring events: incident diagnosis of AF or atrial flutter, or inpatient death, or end of follow-up. HIV infection was analyzed as a time-updated variable using Cox proportional hazards models.
Among 17,293,971 patients (mean age 49 ± 18 years, 56% women, 18,242 with HIV) followed for a median 4.7 years (interquartile range: 2.7 to 6.1 years), 625,167 new diagnoses of AF occurred. Those with HIV were younger, were more likely to be men and of black race, and exhibited more cardiovascular risk factors. Incidence rates of AF in HIV-positive patients were higher than in patients without HIV (18.2 vs. 8.9 per 1,000 person-years of follow-up). After adjustment for age, sex, race, income level, number of health care encounters during the study period, obesity, hypertension, diabetes, obstructive sleep apnea, coronary artery disease, congestive heart failure, valvular heart disease, chronic kidney disease, cigarette smoking, and alcohol abuse, HIV infection was associated with an increased risk for AF (hazard ratio: 1.46; 95% confidence interval [CI]: 1.38 to 1.55; p < 0.0001) (Figure 1). In exploratory multivariable-adjusted Cox models, baseline and incident HIV infection, when included as separate variables, were both associated with incident AF (hazard ratios: 1.35; 95% CI: 1.24 to 1.47; and 1.57; 95% CI: 1.45 to 1.71; respectively). In interaction analyses, the relative risk of incident AF with HIV was significantly higher in younger patients, blacks, Hispanics, and those without hypertension, diabetes, or alcohol abuse.
In this analysis leveraging all inpatient, emergency room, and ambulatory surgery visits in California spanning 7 years, HIV infection was independently associated with an increased risk of AF. The magnitude of the risk conferred by HIV was similar to most established AF risk factors. Only 1 prior study has investigated the relationship between HIV infection and atrial arrhythmias, but was limited to a cross-sectional, double-cohort study from a single institution (2). In that study, no relationship was observed between HIV and atrial arrhythmias after adjusting for potential confounders. In contrast, in our current study, all patients originated from the same cohort, or “study base” (hence, no bias relevant to internal validity can be attributed to a selection of specific patients as “controls”), and all were prospectively followed for incident AF.
Our study has 2 broad implications. First, the mechanisms underlying the association of HIV with incident AF may be a fruitful area of investigation regarding understanding both the pathophysiology of HIV as well as etiologies of atrial arrhythmias. Second, because AF can be asymptomatic and stroke may be the first manifestation of it, our findings have immediate practical implications for HIV-positive patients. These data provide compelling evidence that those with HIV should indeed be considered at high risk for the disease.
Ours is the first study investigating HIV infection as a risk factor for AF in a longitudinal analysis. Whereas the broad inclusion criteria, utilizing a common study base, and large sample size all support the validity of our results and make overfitting of multivariable models less likely, our study has several limitations. HCUP relies on physician coding practices. Prior analyses relying on similar coding for HIV infection and AF have been shown to be reliable (4), but it is possible that HIV seroconversion predated the coding for HIV infection. We were unable to evaluate measures of HIV severity or receipt of antiretroviral medications, constraining our conclusions regarding the mechanisms underlying our observations. Furthermore, due to the absence of outpatient clinical encounters in HCUP, we cannot exclude selection bias.
In summary, HIV infection was associated with a significantly increased risk of incident AF, suggesting that microbial infection can influence AF risk. Future studies can determine whether antiretroviral therapies mitigate that risk.
Please note: The Healthcare Cost and Utilization Project is sponsored by the Agency for Healthcare Research and Quality. Dr. Marcus has received research funding from Medtronic, Jawbone Health, Baylis Medical, and Eight Sleep; and has served as a consultant for and holds equity in InCarda. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2019 American College of Cardiology Foundation
- Sanders J.M.,
- Steverson A.B.,
- Pawlowski A.E.,
- et al.
- Hsu J.C.,
- Li Y.,
- Marcus G.M.,
- et al.
- Antoniou T.,
- Zagorski B.,
- Loutfy M.R.,
- Strike C.,
- Glazier R.H.