Antiretroviral therapy (ART) has transformed HIV into a chronic, manageable condition. However, the decrease in AIDS-related mortality has been accompanied by an increase in the prevalence of noncommunicable chronic diseases, including cardiovascular disease (CVD). In fact, CVD is a leading cause of death among HIV-infected adults.1–3 A 2012 meta-analysis reported a 1.6–2.0 times greater risk of CVD in HIV-infected versus HIV-uninfected adults.4 A portion of the elevated HIV-CVD risk has been attributed to the HIV virus,5–7 ART,4,8,9 and both traditional10 and nontraditional (eg, renal disease and hepatitis C)11,12 CVD risk factors. Even so, excess CVD risk is still observed among HIV-infected adults after controlling for HIV viral load, ART use, and CVD risk factors13 and is observed in those with optimal CVD risk profiles.14 Other factors believed to contribute to the excess CVD risk in HIV include altered immune, coagulation, and endothelial function.15 Although these biological mechanisms improve with ART, markers of these factors do not fully normalize among virologically suppressed HIV-infected adults.16–22 Thus, there is a need to identify novel, modifiable risk factors for HIV-CVD with the potential to improve these dysregulated mechanisms.
General population research provides 3 lines of evidence suggesting that insomnia may be a novel risk factor for HIV-CVD. First, insomnia is an independent risk factor for CVD in the general population. Meta-analytic findings show that adults with elevated insomnia symptoms have an increased risk of CVD [risk ratio (RR) = 1.45],23 including coronary heart disease (RR = 1.28), myocardial infarction (RR = 1.41), stroke (RR = 1.55), and CVD-related mortality (RR = 1.33), after adjustment for traditional CVD risk factors.24 In addition, a meta-analysis examining individual insomnia symptoms found that difficulty initiating sleep, difficulty maintaining sleep, and nonrestorative sleep were associated with increased CVD risk (RRs = 1.27, 1.11, and 1.18, respectively).25 Moreover, recent evidence suggests that the atherogenic effects of insomnia are independent from that of sleep apnea.26,27
Second, intriguing evidence suggests that insomnia may be a causal risk factor for CVD in the general population. The largest randomized control trial of behavioral interventions for insomnia by Irwin et al28 found that older adults with insomnia treated with cognitive behavioral therapy for insomnia, versus an educational sleep seminar, had a significantly reduced risk (odds ratio = 0.26) of elevated C-reactive protein (3.0 mg/L), an inflammatory marker predictive of incident CVD.29 The authors concluded that cognitive behavioral therapy for insomnia is a viable option for modifying an inflammatory marker of CVD risk.
Third, insomnia may be a novel driver of putative biological mechanisms underlying HIV-CVD. In addition, insomnia has been associated with altered immune function,30 systemic inflammation,31 and altered coagulation.32–38
Despite the extensive literature supporting an insomnia–CVD association in the general population and the high insomnia/sleep disturbance prevalence among HIV-infected adults,39,40 to the best of our knowledge, no study has examined whether indicators of insomnia predict incident CVD in HIV-infected adults. Thus, the study purpose was to test whether insomnia symptoms independently predict incident CVD events among HIV-infected veterans.
Study Design and Setting
We used the HIV-infected sample from the Veterans Aging Cohort Study (VACS) Survey Cohort—a prospectively enrolled observational longitudinal survey cohort of HIV-infected veterans and 1:1 matched uninfected veterans on age, sex, race/ethnicity, and clinical site.41 The VACS Survey Cohort undergoes continuous enrollment, beginning in June of 2002. We defined the baseline period as -ever to +6 months of the enrollment date. All participants were followed until the occurrence of a CVD event, death, or the last date of our follow-up period (December 25, 2014). From the total HIV-infected sample of the VACS Survey Cohort (N = 3714), we excluded (1) veterans with pre-existing CVD International Classification of Diseases, Ninth Revision (ICD-9) codes for acute myocardial infarction, unstable angina, cardiovascular revascularization, stroke or transient ischemic attack, peripheral vascular disease, or heart failure in the VA, Medicare, and Medicaid administrative data or VA fee-for-service data during our baseline period (n = 523); (2) veterans missing insomnia symptoms data (n = 40); (3) veterans missing more than 2 items on the Patient Health Questionnaire-9 (PHQ-9; n = 40); and (4) veterans with coding errors (ie, negative follow-up time; n = 3). Our final sample consisted of 3108 HIV-infected veterans.
Measures and Procedures
Baseline Insomnia Symptoms
Insomnia symptoms were assessed at enrollment by the insomnia symptom item of the VACS HIV Symptom Index—a 20-item, self-report questionnaire assessing the frequency and bother of common symptoms in HIV-infected adults exposed to multidrug ART and protease inhibitors.42 Participants were asked to indicate what response best described their experience of each symptom over the past 4 weeks using the following options: 0 = “I do not have this symptom” or “I have this symptom and…,” 1 = “it doesn't bother me,” 2 = “it bothers me a little,” 3 = “it bothers me,” and 4 = “it bothers me a lot.” We used responses to the insomnia item—“Difficulty falling or staying asleep?”—to create a 5-level insomnia symptoms variable. From this variable, 4 dummy coded variables were created with the “no difficulty falling or staying asleep” group as the reference category. Although our insomnia symptoms measure is a single-item, 1-time assessment, data from the VACS Survey Cohort suggest that this measure is moderately stable over time and thus likely captures participants' long-term exposure to insomnia symptoms. Specifically, when correlating responses to this item across VACS Survey Cohort follow-ups occurring between 2002 and 2011, we observed medium positive correlations (rs range: 0.44–0.60; all Ps < 0.001) between adjacent follow-ups.
Incident CVD Events
CVD events were identified using VA and Medicare administrative data and VA fee-for-service data. We defined an incident CVD event as first occurrence of at least 1 inpatient ICD-9 code for acute myocardial infarction or stroke or 1 inpatient ICD-9 Clinical Modification (ICD-9-CM) code or Current Procedural Terminology code for coronary artery revascularization (ie, percutaneous coronary intervention or coronary artery bypass graft).
Five groups of covariates were included, all of which were determined either using self-report questionnaire data collected at baseline or using routine clinical care data in the electronic medical record abstracted for the encounter closest to the participant's enrollment date during the baseline period (ie, -ever to +6 months of enrollment date). The specific ICD-9, ICD-9-CM, and Current Procedural Terminology codes used in variable definitions are presented in Table 5, Supplemental Digital Content, http://links.lww.com/QAI/B279. First, demographic factors were age, sex (0 = male and 1 = female), and race/ethnicity (0 = white, Hispanic, or other and 1 = African American).
Second, cardiovascular risk factors were hypertension, diabetes, body mass index (BMI), smoking, total cholesterol, and statin use. Blood pressure was defined as the average of the last 3 outpatient clinical measurements collected during routine clinical care closest to the enrollment date during the baseline period.43,44 Hypertension was categorized as no hypertension [blood pressure <140/90 mm Hg and no antihypertensive medication (reference category)], controlled hypertension (<140/90 mm Hg with antihypertensive medication), or uncontrolled hypertension (≥140/90 mm Hg). Diabetes (yes/no) was identified by a previously validated metric incorporating glucose measurements, diabetes medication use, and/or at least 1 inpatient or 2 outpatient ICD-9 codes for diabetes.45,46 BMI (kg/m2) was defined by one outpatient clinical measurement collected during routine clinical care closest to the enrollment date during the baseline period and modeled as a continuous variable. Smoking [never (reference category), former, and current smoker] was assessed by self-report at enrollment. Total cholesterol was defined by one outpatient clinical measurement collected during routine clinical care closest to the enrollment date during the baseline period and modeled as a continuous variable. Baseline statin use (yes/no) was defined as a filled prescription receipt for a 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor at the time of enrollment.
Third, additional potential confounders were hepatitis C infection, renal disease, anemia, alcohol use, and cocaine use. Hepatitis C infection (yes/no) was identified by a positive hepatitis C virus antibody test or at least 1 inpatient or 2 outpatient ICD-9 codes for hepatitis C infection.47 Renal disease was defined by estimated glomerular filtration rate,48 and anemia was defined using a hemoglobin; both were based on a measurement collected during routine clinical care closest to the enrollment date during the baseline period and were modeled as a continuous variables. Alcohol use was assessed by the Alcohol Use Disorders Identification Test (AUDIT-C) at baseline and alcohol abuse/dependence ICD-9 codes during the baseline period and dichotomized as no current use or nonhazardous use versus hazardous use (AUDIT-C ≥ 4) or alcohol abuse/dependence disorder. Cocaine use was assessed by self-report at baseline and cocaine use disorder ICD-9 codes during the baseline period and dichotomized as never tried or no use in past 12 months versus use in the past 12 months or cocaine abuse/dependence disorder.
Fourth, HIV-specific factors were HIV-1 RNA viral load, CD4+ T-cell count, and ART regimen. HIV-1 RNA viral load and CD4+ T-cell count were defined by outpatient clinical measurements collected during routine clinical care closest to the enrollment date during the baseline period and modeled as continuous variables. Baseline ART regimen (yes/no) was defined by a filled prescription receipt for any ART (−180 days to +7 days of enrollment date).
Finally, insomnia-related variables were nonbenzodiazepine sleep medication use, depressive symptoms, and antidepressant medication use. Nonbenzodiazepine sleep medication use (yes/no) was defined by a filled prescription receipt closest to baseline (-ever to +180 days of enrollment date) for the following medications: zolpidem, zaleplon, eszopiclone, and indiplon. Depressive symptoms were assessed by the PHQ-9 at baseline. Antidepressant medication use was defined by a filled prescription receipt for an antidepressant medication closest to baseline (-ever to +180 days of enrollment date). We computed 3 dichotomous variables (yes/no) based on the antidepressant medication type—serotonin reuptake inhibitor (SSRI), tricyclic antidepressant (TCA), and miscellaneous other antidepressant.
To examine the association between baseline insomnia symptoms and incident CVD events among HIV-infected veterans, first, we constructed Kaplan–Meier event-free survival curves to illustrate the time from enrollment to first CVD event for each of the 5 levels of our insomnia symptoms variable. Second, we constructed Cox proportional hazards regression models to estimate the hazard ratio (HR) and 95% confidence intervals (CIs) for each level of the insomnia symptoms variable (reference category: “no difficulty falling or staying asleep”), given our interest in the most clinically relevant insomnia symptoms group (“bothers a lot”) and the novelty of our research question in the HIV population. We also performed a likelihood ratio test of 2 Cox models (one with the insomnia symptoms variable and the other without) to present the overall significance of the insomnia symptoms variable.
Five Cox proportional hazards regression models were constructed to hierarchically add the main covariates to create model 4, our primary model, as follows: model 0—unadjusted (insomnia symptoms dummy coded variables alone predicting incident CVD); model 1—model 0 + demographics (age, sex, and race/ethnicity); model 2—model 1 + CVD risk factors (hypertension, diabetes, BMI, smoking, total cholesterol, and statin use); model 3—model 2 + additional potential confounders (hepatitis C infection, renal disease, anemia, alcohol use, and cocaine use); and model 4—model 3 + HIV-specific factors (HIV-1 RNA level, CD4+ T-cell count, and ART regimen).
An additional 4 exploratory Cox proportional hazards regression models were constructed to examine the potential influence of insomnia-related variables. The first 3 Cox models hierarchically and individually added insomnia-related variables to model 4 to examine the influence of each variable on model 4. Three versions of model 5 were constructed as follows: model 5a—model 4 + nonbenzodiazepine sleep medication use; model 5b—model 4 + PHQ-9 total score; and model 5c—model 4 + SSRI use, TCA use, and miscellaneous other antidepressant use. The fourth Cox model included all of the covariates in 1 model as follows: model 6—model 4 + nonbenzodiazepine sleep medication use, PHQ-9 total score, SSRI use, TCA use, and miscellaneous other antidepressant use.
We also conducted 4 exploratory sensitivity analyses. First, to examine the influence of adjusting for ART medications associated with insomnia or disruption of the metabolism of insomnia medications on our primary model,49,50 we re-ran 2 separate model 4's. The first replaced the ART regimen variable with an efavirenz use variable (yes: n = 787; no: n = 2321) and the second replaced the ART regiment variable with a protease inhibitor use variable (yes: n = 1457; no: n = 1651). Second, to examine the influence of omitting item 3 of the PHQ-9 (“trouble falling or staying asleep, or sleeping too much”) on models including depression, we removed the item from the PHQ-9 total score calculation and re-ran model 5b and model 6.
Multiple imputations using chained equations with 5 separate imputed data sets were generated based on predictive mean matching using the Hmisc library of R programming language. Cox survival models were fitted in each imputed data set and then combined to obtain pooled HRs and standard errors. All analyses were performed using R software (version 3.3.3; www.r-project.org). To account for inflation in type 1 error associated with multiple tests, we considered P < 0.01 as statistically significant.
Participant characteristics are shown in Table 1. The median age was 49 years. Most participants were men (97%), and two-thirds were African American (66%). The distribution of participants and incident CVD events across insomnia symptom categories is displayed in Table 2. At enrollment, most HIV-infected veterans endorsed some level of insomnia symptoms (59%). During a median follow-up time of 10.8 years (first quartile: 6.7, third quartile: 11.8), 267 incident CVD (8.6%) events occurred.
Baseline Insomnia Symptoms and Incident CVD Events
Figure 1 presents Kaplan–Meier survival curves illustrating the time to first CVD event for each of the 5 insomnia symptom categories. For HIV-infected veterans, the “bothers a lot” survival curve consistently separates from those for the other 4 categories around the sixth year of follow-up. Although visually supporting a difference in time to CVD event between “bothers a lot” and the other 4 categories, this is not reflected in the omnibus test of the survival curves (log-rank χ2 = 5.2, df = 4, P = 0.267).
The 5 Cox models constructed to hierarchically create our primary model 4 examining the association between baseline insomnia symptoms and incident CVD events revealed that the test of overall significance across the 4 insomnia symptoms dummy coded variables was not significant in models 0–4 (P value range: 0.069–0.304; Table 3). These omnibus null results were driven by the lack of associations between incident CVD events and the dummy coded variables comparing the “doesn't bother,” “bothers a little,” and “bothers” groups to the “no difficulty falling or staying asleep” group in model 0–4 (P value range: 0.305–0.971). By contrast, the dummy coded variable comparing the “bothers a lot” group to the “no difficulty falling or staying asleep” group fell just short of significant in model 0 (P = 0.034) and was significantly associated with incident CVD in models 1–4 (HR range: 1.45–1.70; P value range: 0.003–0.007). To illustrate, our primary model 4 results indicate that HIV-infected veterans bothered a lot by difficulty falling or staying asleep had a 66% greater risk of incident CVD events than HIV-infected veterans without these symptoms, independent of demographics, CVD risk factors, additional potential confounders, and HIV-specific factors.
The 4 additional exploratory Cox models constructed to hierarchically examine the potential influence of insomnia-related variables revealed similar results to models 0–4 for the overall significance across the 4 insomnia symptoms dummy coded variables in models 5a–5c and model 6 (P value range: 0.147–0.591; Table 3) and the dummy coded variables comparing the “doesn't bother,” “bothers a little,” and “bothers” groups to the “no difficulty falling or staying asleep” group in model 5a–5a and model 6 (P value range: 0.360–0.933). Similar to model 4, the dummy coded variable comparing the “bothers a lot” group to the “no difficulty falling or staying asleep” group remained significantly associated with incident CVD after including nonbenzodiazepine sleep medication use in model 5a (62% greater risk; P = 0.009). However, the “bothers a lot” and “no difficulty falling or staying asleep” group comparison fell short of significance after including PHQ-9 total score in model 5b (51% greater risk; P = 0.060), antidepressant medication use in model 5c (51% greater risk; P = 0.031), or all 3 insomnia-related variables in model 6 (42% greater risk, P = 0.115).
Results of the 4 exploratory sensitivity analyses are presented in Table 4, Supplemental Digital Content, http://links.lww.com/QAI/B279. Our first and second sensitivity analyses, in which ART regimen was replaced with either efavirenz use or protease inhibitor use in model 4, yielded nearly identical results (“bothers a lot” HR in model 4 with efavirenz use = 1.68, 95% CI: 1.17 to 2.39, P = 0.005; “bothers a lot” HR in model 4 with protease inhibitor use = 1.67, 95% CI: 1.17 to 2.39, P = 0.004) to our original model 4. Our third and fourth sensitivity analyses, in which the sleep disturbance item was omitted from the PHQ-9 total score calculation in models 5b and 6, also produced nearly identical results (“bothers a lot” HR in model 5b = 1.53, 95% CI: 1.02 to 2.32, P = 0.042; bothers a lot HR in model 6 = 1.43, 95% CI: 0.94 to 2.17, P = 0.093) to our original models 5b and 6. These findings indicate that (1) adjusting for ART medications with recognized effects on insomnia or insomnia medications and (2) removing the sleep disturbance item from the depressive symptoms measure had a negligible influence on the relationships of interest.
In this large prospective cohort study using the VACS Survey Cohort data, we found that insomnia symptoms were associated with incident CVD among HIV-infected veterans. After adjustment for demographics, CVD risk factors, additional potential confounders (hepatitis C infection, renal disease, anemia, alcohol use, and cocaine use), and HIV-specific factors (HIV-1 RNA level, CD4+ T-cell count, and ART regimen), HIV-infected veterans bothered a lot by difficulty falling or staying asleep exhibited a 66% greater risk of an incident CVD event than HIV-infected veterans without these symptoms. This association persisted after individually adjusting for nonbenzodiazepine sleep medication use (62% greater risk). However, although the hazard ratios were similar in magnitude (HRs = 1.51), the association between highly bothersome insomnia symptoms and incident CVD fell short of significance after individually adjusting for depressive symptoms or antidepressant medication use. We did not observe associations between the other insomnia symptoms category comparisons (“doesn't bother,” “bothers a little,” and “bothers” versus “no difficulty falling or staying asleep”) and incident CVD events. Taken together, the present results suggest that highly bothersome insomnia symptoms may be an independent predictor of incident CVD events among HIV-infected adults, although the overlap with depression and its treatment needs further investigation.
To the best of our knowledge, this is the first examination of the insomnia-CVD relationship in people with HIV. We observed a threshold effect, with only the highest level of insomnia symptoms exhibiting an elevated CVD risk. The leading potential explanation for this finding is that (1) the “bothers a lot” group may have the highest rate of insomnia disorder and that (2) it is insomnia disorder, but not subclinical insomnia symptoms, that is related to elevated CVD risk. Future studies are needed to evaluate this possibility.
The biological plausibility of the insomnia–CVD relationship in the HIV population has preliminary support. Insomnia symptoms have been associated with inflammation-related single nucleotide polymorphisms51,52 and elevated circulating levels of interleukin-6 and C-reactive protein,53 both of which are implicated in atherosclerosis among HIV-infected adults.54 Additional support for the insomnia–CVD relationship can be drawn from the general population literature, as 3 meta-analyses have linked insomnia symptoms with future CVD.23–25 In addition, general population evidence supports associations between insomnia symptoms and biological mechanisms underlying CVD development, including systemic inflammation,31 altered coagulation,32–38 and endothelial dysfunction.55–57
There are study limitations that warrant consideration. First, our insomnia symptoms assessment was obtained only at baseline and relied on a single item. Although the VACS HIV Symptom Index's42 insomnia item does assess difficulty initiating or maintaining sleep and bothersomeness, it misses other insomnia symptoms (eg, early morning awakening and functional impairment), and it does not measure frequency or duration. These content validity issues could lead to misclassification (eg, coding a veteran with insomnia symptoms as not having them), which could contribute to effect size underestimation. That said, most of the general population studies have used single-item insomnia assessments (most often assessing difficulty falling asleep) and have observed positive associations with CVD-related outcomes.24 In addition, the insomnia symptoms item does not capture chronicity of insomnia. Future research should examine the potential influence of insomnia duration or chronic versus acute insomnia on incident CVD risk. Second, because we used a veteran sample comprising mostly men, it is unclear whether our results extend to HIV-infected women. Third, because of the timeframe of this study, we could not assess for newer ART therapies linked to insomnia namely, rilpivirine and dolutegravir. Fourth, because of the absence of a strong assessment, we were unable to adjust our models for sleep apnea—a sleep disorder with an estimated prevalence of 2.9% among veterans58 and an established independent risk factor for incident CVD.59 Fifth, we were unable to include posttraumatic stress disorder (PTSD) as a potential confounder. Some of the symptoms of PTSD have strong conceptual and measurement overlap with the insomnia symptoms of interest here (ie, nightmares and difficulty sleeping due to alterations in arousal and reactivity)60 and the available PTSD assessments do not allow for the separation of insomnia and PTSD symptoms. Future research should examine include comprehensive measures of insomnia and PTSD symptoms to tease apart the potential influence of these overlapping factors on incident CVD.
In summary, we provide intriguing evidence that highly bothersome insomnia symptoms may be an independent predictor of incident CVD events among HIV-infected adults. Future prospective studies with comprehensive insomnia assessments using nonveteran samples are needed to fully characterize the insomnia–CVD relationship and to identify its underlying mechanisms in the HIV population. Ultimately, such research could lead to the identification of a novel, modifiable risk factor for HIV-CVD and could inform the development of new primary prevention programs that include insomnia interventions to prevent CVD in people with HIV.
1. Rodger AJ, Lodwick R, Schechter M, et al. Mortality in well controlled HIV
in the continuous antiretroviral therapy arms of the SMART and ESPRIT trials compared with the general population. AIDS. 2013;27:973–979.
2. Trickey A, May MT, Vehreschild J, et al. Cause-specific mortality in HIV
-positive patients who survived ten years after starting antiretroviral therapy. PLoS One. 2016;11:e0160460.
3. Palella FJ Jr, Baker RK, Moorman AC, et al. Mortality in the highly active antiretroviral therapy era: changing causes of death and disease in the HIV
outpatient study. J Acquir Immun Defic Syndr. 2006;43:27–34.
4. Islam FM, Wu J, Jansson J, et al. Relative risk of cardiovascular disease
among people living with HIV
: a systematic review and meta-analysis. HIV
5. Lang S, Mary-Krause M, Simon A, et al. HIV
replication and immune status are independent predictors of the risk of myocardial infarction in HIV
-infected individuals. Clin Infect Dis. 2012;55:600–607.
6. Triant VA, Regan S, Lee H, et al. Association of immunologic and virologic factors with myocardial infarction rates in a US healthcare system. J Acquir Immun Defic Syndr. 2010;55:615–619.
7. Sico JJ, Chang CC, So-Armah K, et al. HIV
status and the risk of ischemic stroke among men. Neurology. 2015;84:1933–1940.
8. Friis-Moller N, Reiss P, Sabin CA, et al. Class of antiretroviral drugs and the risk of myocardial infarction. N Engl J Med. 2007;356:1723–1735.
9. Friis-Moller N, Sabin CA, Weber R, et al. Combination antiretroviral therapy and the risk of myocardial infarction. N Engl J Med. 2003;349:1993–2003.
10. Currier JS, Lundgren JD, Carr A, et al. Epidemiological evidence for cardiovascular disease
-infected patients and relationship to highly active antiretroviral therapy. Circulation. 2008;118:e29–e35.
11. Freiberg MS, Chang CC, Skanderson M, et al. The risk of incident coronary heart disease among veterans
with and without HIV
and hepatitis C. Circ Cardiovasc Qual Outcomes. 2011;4:425–432.
12. George E, Lucas GM, Nadkarni GN, et al. Kidney function and the risk of cardiovascular events in HIV
-1-infected patients. AIDS. 2010;24:387–394.
13. Freiberg MS, Chang CC, Kuller LH, et al. HIV
infection and the risk of acute myocardial infarction. JAMA Intern Med. 2013;173:614–622.
14. Paisible AL, Chang CC, So-Armah KA, et al. HIV
infection, cardiovascular disease
risk factor profile, and risk for acute myocardial infarction. J Acquir Immune Defic Syndr. 2015;68:209–216.
15. So-Armah K, Freiberg MS. Cardiovascular disease
risk in an aging HIV
population: not just a question of biology. Curr Opin HIV
16. Neuhaus J, Jacobs DR Jr, Baker JV, et al. Markers of inflammation, coagulation, and renal function are elevated in adults with HIV
infection. J Infect Dis. 2010;201:1788–1795.
17. Bahrami H, Budoff M, Haberlen SA, et al. Inflammatory markers associated with subclinical coronary artery disease: the multicenter AIDS cohort study. J Am Heart Assoc. 2016;5:e003371.
18. Beltrán LM, Muñoz Hernández R, de Pablo Bernal RS, et al. Reduced sTWEAK and increased sCD163 levels in HIV
-infected patients: modulation by antiretroviral treatment, HIV
replication and HCV co-infection. PLoS One. 2014;9:e90541.
19. Funderburg NT. Markers of coagulation and inflammation often remain elevated in ART-treated HIV
-infected patients. Curr Opin HIV
20. Hearps AC, Maisa A, Cheng WJ, et al. HIV
infection induces age-related changes to monocytes and innate immune activation in young men that persist despite combination antiretroviral therapy. AIDS. 2012;26:843–853.
21. Karim R, Mack WJ, Kono N, et al. T-cell activation, both pre- and post-HAART levels, correlates with carotid artery stiffness over 6.5 years among HIV
-infected women in the WIHS. J Acquir Immun Defic Syndr. 2014;67:349–356.
22. Kaplan RC, Sinclair E, Landay AL, et al. T cell activation and senescence predict subclinical carotid artery disease in HIV
-infected women. J Infect Dis. 2011;203:452–463.
23. Sofi F, Cesari F, Casini A, et al. Insomnia
and risk of cardiovascular disease
: a meta-analysis. Eur J Prev Cardiol. 2014;21:57–64.
24. Li M, Zhang XW, Hou WS, et al. Insomnia
and risk of cardiovascular disease
: a meta-analysis of cohort studies. Int J Cardiol. 2014;176:1044–1047.
25. He Q, Zhang P, Li G, et al. The association between insomnia
symptoms and risk of cardio-cerebral vascular events: a meta-analysis of prospective cohort studies. Eur J Prev Cardiol. 2017;24:1071–1082.
26. Ogilvie RP, Lakshminarayan K, Iber C, et al. Joint effects of OSA and self-reported sleepiness on incident CHD and stroke. Sleep Med. 2018;44:32–37.
27. Bertisch SM, Pollock BD, Mittleman MA, et al. Insomnia
with objective short sleep duration and risk of incident cardiovascular disease
and all-cause mortality: Sleep Heart Health Study. Sleep. 2018;41:1–9.
28. Irwin MR, Olmstead R, Carrillo C, et al. Cognitive behavioral therapy vs. Tai Chi for late life insomnia
and inflammatory risk: a randomized controlled comparative efficacy trial. Sleep. 2014;37:1543–1552.
29. Helfand M, Buckley D, Fleming C, et al. U.S. Preventive Services Task Force Evidence Syntheses, Formerly Systematic Evidence Reviews. Screening for Intermediate Risk Factors for Coronary Heart Disease. Rockville, MD: Agency for Healthcare Research and Quality (US); 2009.
30. Besedovsky L, Lange T, Born J. Sleep and immune function. Pflugers Arch. 2012;463:121–137.
31. Irwin MR, Olmstead R, Carroll JE. Sleep disturbance
, sleep duration, and inflammation: a systematic review and meta-analysis of cohort studies and experimental sleep deprivation. Biol Psychiatry. 2016;80:40–52.
32. Jackowska M, Steptoe A. Sleep and future cardiovascular risk: prospective analysis from the English Longitudinal Study of Ageing. Sleep Med. 2015;16:768–774.
33. Mausbach BT, Ancoli-Israel S, von Känel R, et al. Sleep disturbance
, norepinephrine, and D-dimer are all related in elderly caregivers of people with Alzheimer disease. Sleep. 2006;29:1347–1352.
34. von Känel R, Ancoli-Israel S, Dimsdale JE, et al. Sleep and biomarkers of atherosclerosis in elderly Alzheimer caregivers and controls. Gerontology. 2010;56:41–50.
35. von Känel R, Dimsdale JE, Ancoli-Israel S, et al. Poor sleep is associated with higher plasma proinflammatory cytokine interleukin-6 and procoagulant marker fibrin D-dimer in older caregivers of people with Alzheimer's disease. J Am Geriatr Soc. 2006;54:431–437.
36. Matthews KA, Zheng HY, Kravitz HM, et al. Are inflammatory and coagulation biomarkers related to sleep characteristics in mid-life women?: study of women's health across the Nation Sleep Study. Sleep. 2010;33:1649–1655.
37. Prather AA, Epel ES, Cohen BE, et al. Gender differences in the prospective associations of self-reported sleep quality with biomarkers of systemic inflammation and coagulation: findings from the Heart and Soul Study. J Psychiatr Res. 2013;47:1228–1235.
38. Suarez EC. Self-reported symptoms of sleep disturbance
and inflammation, coagulation, insulin resistance and psychosocial distress: evidence for gender disparity. Brain Behav Immun. 2008;22:960–968.
39. Reid S, Dwyer J. Insomnia
infection: a systematic review of prevalence, correlates, and management. Psychosom Med. 2005;67:260–269.
40. Wu J, Wu H, Lu C, et al. Self-reported sleep disturbances in HIV
-infected people: a meta-analysis of prevalence and moderators. Sleep Med. 2015;16:901–907.
41. Justice AC, Dombrowski E, Conigliaro J, et al. Veterans
Aging Cohort Study (VACS
): overview and description. Med Care. 2006;44(8 suppl 2):S13–S24.
42. Justice AC, Holmes W, Gifford AL, et al. Development and validation of a self-completed HIV
symptom index. J Clin Epidemiol. 2001;54(suppl 1):S77–S90.
43. Armah KA, Chang CC, Baker JV, et al. Prehypertension, hypertension, and the risk of acute myocardial infarction in HIV
-infected and -uninfected veterans
. Clin Infect Dis. 2014;58:121–129.
44. Chobanian AV, Bakris GL, Black HR, et al. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension. 2003;42:1206–1252.
45. Butt AA, Fultz SL, Kwoh CK, et al. Risk of diabetes in HIV
pre- and post-HAART and the role of HCV coinfection. Hepatology. 2004;40:115–119.
46. Butt AA, McGinnis K, Rodriguez-Barradas MC, et al. HIV
infection and the risk of diabetes mellitus. AIDS. 2009;23:1227–1234.
47. Butt AA, Xiaoqiang W, Budoff M, et al. Hepatitis C virus infection and the risk of coronary disease. Clin Infect Dis. 2009;49:225–232.
48. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002;39(2 suppl 1):S1–S266.
49. Kenedi CA, Goforth HW. A systematic review of the psychiatric side-effects of efavirenz. AIDS Behav. 2011;15:1803–1818.
50. Omonuwa TS, Goforth HW, Preud'homme X, et al. The pharmacologic management of insomnia
in patients with HIV
. J Clin Sleep Med. 2009;5:251–262.
51. Gay CL, Zak RS, Lerdal A, et al. Cytokine polymorphisms and plasma levels are associated with sleep onset insomnia
in adults living with HIV
/AIDS. Brain Behav Immun. 2015;47:58–65.
52. Lee KA, Gay C, Pullinger CR, et al. Cytokine polymorphisms are associated with poor sleep maintenance in adults living with human immunodeficiency virus/acquired immunodeficiency syndrome. Sleep. 2014;37:453–463.
53. Wirth MD, Jaggers JR, Dudgeon WD, et al. Association of markers of inflammation with sleep and physical activity among people living with HIV
or AIDS. AIDS Behav. 2015;19:1098–1107.
54. Vos AG, Idris NS, Barth RE, et al. Pro-inflammatory markers in relation to cardiovascular disease
infection. A systematic review. PLoS One. 2016;11:e0147484.
55. Routledge FS, Dunbar SB, Higgins M, et al. Insomnia
symptoms are associated with abnormal endothelial function. J Cardiovasc Nurs. 2017;32:78–85.
56. Cooper DC, Ziegler MG, Milic MS, et al. Endothelial function and sleep: associations of flow-mediated dilation with perceived sleep quality and rapid eye movement (REM) sleep. J Sleep Res. 2014;23:84–93.
57. Aziz M, Ali SS, Das S, et al. Association of subjective and objective sleep duration as well as sleep quality with non-invasive markers of sub-clinical cardiovascular disease
(CVD): a systematic review. J Atheroscler Thromb. 2017;24:208–226.
58. Sharafkhaneh A, Richardson P, Hirshkowitz M. Sleep apnea in a high risk population: a study of Veterans
Health Administration beneficiaries. Sleep Med. 2004;5:345–350.
59. Xie C, Zhu R, Tian Y, et al. Association of obstructive sleep apnoea with the risk of vascular outcomes and all-cause mortality: a meta-analysis. BMJ Open. 2017;7:e013983.
60. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed (DSM-5). Arlington, VA: American Psychiatric Association; 2013.