Abstract
Background:
People living with HIV (PLWH) experience high rates of mood disorders (major depression and bipolar affective disorder) which in the general population have been associated with non-communicable disease (NCD) risk. We examined whether prevalent mood disorders are associated with incident NCDs and multimorbidity (accumulation of ≥2 NCDs) in PLWH.
Setting:
Adult HIV clinic cohort in Nashville, Tennessee, between 1998–2015.
Methods:
PLWH with ≥1 year of follow-up in the clinic were assessed for cardiovascular disease, metabolic syndrome (any three of hypertension, hyperlipidemia, diabetes, or obesity), chronic kidney and liver disease, non-AIDS-defining cancers, and dementia. Only mood disorders documented during the first year of care were included. Cumulative incidence and adjusted subhazard ratios were calculated for risk of NCDs and multimorbidity with death as a competing risk. Multivariable Cox models estimated mortality risk following multimorbidity.
Results:
Of 4,140 adults, 24% had a mood disorder diagnosed in the first year of care, 51% had ≥1 NCD at baseline, and there were 2588 incident NCDs during the study period. Mood disorders were associated with increased risk of first NCD (adjusted subhazard ratio [aSHR] = 1.29, 95% confidence interval [CI] 1.06–1.57), incident multimorbidity (aSHR ranging from 1.04 to 1.42) and metabolic syndrome (aSHR = 1.29, 95%CI 1.02–1.64). Mood disorders were not conclusively associated with mortality risk following multimorbidity.
Conclusions:
PLWH with mood disorders were at increased risk of incident NCDs and multimorbidity, particularly metabolic syndrome. Focused prevention and treatment of NCDs may reduce the burden of multimorbidity in this high-risk group.
Keywords: HIV, depression, non-communicable disease, multimorbidity, aging, metabolic syndrome
Introduction
With the success of antiretroviral therapy (ART), the burden of morbidity and mortality experienced by people living with HIV (PLWH) is shifting from infections to non-communicable diseases (NCDs)1,2. NCDs including cardiovascular disease, diabetes, certain non-AIDS cancers, liver disease, chronic kidney disease, and osteoporosis are observed at higher rates among PLWH compared to their HIV-negative peers3,4. As PLWH age, they may accumulate these co-morbid conditions and develop multimorbidity – defined as the presence of two or more chronic health conditions. In settings with widely-available ART, reported rates of multimorbidity are high in PLWH, particularly those over the age of 50 years5–8.
Underlying psychiatric disease has been associated with increased NCD risk in the general population, including cardiovascular disease, liver disease, cerebrovascular disease, and renal disease9. One potential mechanism of this risk includes inflammatory changes observed with depression and bipolar disorder. Studies have demonstrated that adults with mood disorders have higher levels of markers of systemic inflammation (including sCD14, CRP and IL-6) and alterations in adaptive immunology10,11. The pathogenesis of NCDs is complex in PLWH: involving inflammatory, immunologic, pharmacologic, and behavioral mechanisms. Depression and bipolar affective disorder are more prevalent in PLWH than in the general population and are associated with an increased risk of death12–15. Studies among PLWH have suggested an association of mood disorders and NCD outcomes, particularly cardiovascular disease16,17. However, the role of mood disorders in exacerbating NCD and multimorbidity risk among PLWH and whether mood disorders are associated with higher risk of mortality among PLWH with multimorbidity have not been well-described to date.
This study sought to examine whether prevalent mood disorders (depression and bipolar disorder) increased risk of incident NCD and multimorbidity in PLWH. We hypothesized that after adjusting for pertinent confounders, the presence of a mood disorder would be independently associated with higher incident NCD and multimorbidity risk.
Methods
This study included adult (age ≥18 years) PLWH enrolled in care at the Vanderbilt Comprehensive Care Clinic (VCCC) between January 1998 and December 2015. The VCCC provides specialty HIV and primary care. It offers on-site mental health care by psychiatrists, psychiatric nurse practitioners, and licensed clinical social workers. Since 2003, research staff have systematically abstracted demographic, clinical, and laboratory data from the electronic medical record as well as diagnoses recorded by medical providers including psychiatric and NCD diagnoses. All NCD diagnoses were reviewed by research staff for validation by documented diagnoses, dates, and/or by medications with supporting objective data such as vital signs (as in the case of obesity and hypertension) or laboratory results (as used for chronic kidney disease and cancer validation). The Vanderbilt Institutional Review Board approved use of these data for research.
Cohort design and study observation times are shown in Figure 1. Patients with less than one year of follow-up were excluded. The start of observation time (baseline) for this study was one year after first clinic visit. Our primary exposure of interest was the presence of a mood disorder such as depression or bipolar affective disorder. Mood disorders are cyclical, and date of mood disorder onset is inherently difficult to obtain in observational research. We therefore included prevalent diagnoses at baseline based upon documentation of depression, bipolar affective disorder, or mood disorder NOS (not otherwise specified) entered into the medical record during the first year of clinic care. Patients with both depression and bipolar affective disorder diagnoses recorded were classified as bipolar affective disorder. We examined any mood disorder as a composite group in our primary analyses. For power and generalizability considerations, we did not exclude patients with other psychiatric diagnoses (such as schizophrenia, anxiety, or post-traumatic stress disorder), many of which are also prevalent among PLWH and may also be associated with mood disorders. Patients with other psychiatric diagnoses and without a mood disorder were included in the unexposed group, and patients with both a mood disorder diagnosis and other psychiatric diagnoses were included in the exposed group. Patients with mood disorders documented after the first year in care were considered unexposed to minimize the risk of reverse causality (development of NCDs leading to depressive disorder). All patients contributed observation time until death; December 31, 2015; or 365 days after their last clinic visit if their last visit was before January 1, 2015, at which point the patient was considered lost to follow-up.
Figure 1.
Flow diagram of cohort creation and observation time for study. Describes inclusion/exclusion criteria for study cohort and baseline time point definition. Observation period for study included person-time from baseline to NCD outcome, death, or censoring. Description of cohorts included for analysis of varying NCD outcome listed.
The NCD outcomes included the eleven following conditions: coronary artery disease (including myocardial infarction, coronary artery disease, ischemic heart disease), cerebrovascular disease, hypertension, peripheral vascular disease, chronic liver disease (cirrhosis, hepatic steatosis, and chronic liver disease), diabetes mellitus, hyperlipidemia, obesity (defined as body mass index [BMI] ≥ 30), stage III-V chronic kidney disease (defined as eGFR <60 mL/min/1.73m2 for ≥3 months duration), dementia, and non-AIDS-defining cancers (excluding basal and squamous cell skin cancers). The earliest date of onset for each condition was used and all diseases were treated as non-reversible, non-repeating conditions. Other possible NCDs such as pulmonary diseases or rheumatologic diseases are not routinely collected for the observational cohort and thus were not included. Our primary outcome of interest was incident multimorbidity, defined as two or more of the eleven possible NCDs. Given the high frequency of hypertension and metabolic disorders observed in the cohort, we also examined a modified metabolic syndrome of multimorbidity which included any three of hypertension, obesity, hyperlipidemia, and diabetes. We included death from any cause as a competing risk in models for the outcome of multimorbidity and modeled risk of mortality as a separate outcome of interest after incident multimorbidity.
We first examined demographic and clinical covariates of patients at baseline (one year after clinic entry), stratified by presence of mood disorder diagnosis. Differences in demographic and baseline clinical covariates were compared by mood disorder exposure groups using Wilcoxon rank sum test for continuous variables and chi-square test for categorical measures.
We examined risk of NCD multimorbidity by calculating cumulative incidence curves and subhazard ratios using time-to-event, competing risk regression for the following incident outcomes: first NCD, multimorbidity with ≥2 NCDs, multimorbidity with ≥3 NCDs, multimorbidity with ≥4 NCDs, and metabolic syndrome. We used age as the unit of time measurement to align risk sets by age and account for confounding by age more directly than by regression adjustment. Multivariable models included mood disorder diagnosis and the preselected covariates of sex, race, ever substance abuse, ever alcohol, ever tobacco, calendar year of clinic entry, time-varying CD4 cell count, time-varying CD4/CD8 ratio, time-varying HIV RNA, and ever hepatitis C infection. Time-varying laboratory values were used to account for potential differences in ART receipt, adherence, and outcomes between patients with and without mood disorders during observation. Multivariable analyses of multimorbidity outcomes (≥2 NCDs, ≥3 NCDs, etc.) also adjusted for number of NCDs at baseline. Analysis of the time to first NCD excluded patients with any NCDs at baseline. As prespecified sensitivity analyses, multivariable models were repeated including (1) only baseline HIV monitoring laboratory values rather than time-varying results and (2) the addition of time-varying, cumulative person-year exposures to psychiatric medications to the adjusted model to account for potential confounding by medication side effects and possible symptom severity during the observation period.
Lastly, to examine whether mood disorders were associated with worse outcomes among patients with multimorbidity, we used Cox proportional hazard models for risk of death following multimorbidity onset (≥2 NCDs). We included all patients with multimorbidity in multivariable models that included age, sex, race, ever substance use, ever tobacco use, ever alcohol use, year of clinic entry, CD4 cell count, CD4/CD8 ratio, and HIV RNA at multimorbidity onset.
Statistical analyses were performed using Stata 12.1 (Stata Corporation, College Station, Texas, USA). All p-values were two-sided.
Results
In total, 5,452 patients entered clinic care between 1998–2015. We excluded 1,312 patients (24%) who had <1 year of clinic follow-up. Our analysis included 4,140 patients with ≥1 year of clinical care, of whom 999 (24%) had documentation of depression (n=649 or 65%) or bipolar affective disorder (n=350 or 35%) diagnoses during their first year in care. Compared to those excluded for <1 year of care, patients included in our analyses were more likely to be female (23 vs. 20%, p=0.02), had a shorter time from HIV diagnosis to clinic entry (median 0.9 vs. 2.5 years, p<0.001), were more likely to have one NCD at clinic entry (37 vs. 33%, p=0.003), and were more likely to ever have a mood disorder diagnosis recorded in their medical record (36 vs. 20%, p<0.001).
The baseline (1 year after clinic entry) characteristics of included patients are shown in Table 1. Patients with prevalent mood disorders were slightly older, more likely to be female, and of white race. They were more likely to ever report injection drug use, substance use, alcohol use, and tobacco use. They had higher CD4 cell count nadir and were less likely to have HIV RNA <400 copies/mL, but they were similar to patients without mood disorders with regards to baseline CD4 cell count and CD4/CD8 ratio. They had higher prevalence of chronic hepatitis C virus infection and higher prevalence of one and ≥2 NCDs at the baseline. They had a higher number of medical visits to clinic in the first year of care and 38% had at least one visit with a mental health specialist at the clinic. Sixty-three percent of all mood disorder patients were treated with an antidepressant at some point in care and 15% died. When we compared baseline characteristics between patients with depression versus those with bipolar affective disorder, the following differences were noted: patients with bipolar affective disorder were slightly younger (median age of 38.5 vs. 40.9 years), had a higher prevalence of history of injection drug use (16 vs. 10%), ever reporting substance use (68 vs. 48%), ever reporting alcohol use (67 vs. 56%), ever reporting tobacco use (74 vs. 56%), hepatitis C virus infection (22 vs. 16%), and treatment with mood stabilizers (13 vs. 2%) and atypical antipsychotics (38 vs. 11%). Baseline CD4 cell count, CD4/CD8 ratio, HIV RNA, prevalent NCDs, and subsequent mortality were similar between patients with depression versus bipolar affective disorder.
Table 1.
Characteristics of patients without and with prevalent mood disorders.
| No Mood Disorder (n=3141) | Mood Disorder (n=999) | P valuea | |
|---|---|---|---|
| Demographic characteristics | |||
| Baselineb age in years, median [IQR] | 38.7 [30.8–45.8] | 40.1 [32.3–46.2] | 0.015 |
| Female sex, N (%) | 686 (22) | 267 (27) | 0.001 |
| Other | 319 (10) | 43 (3) | |
| Other or unknown | 168 (5) | 30 (3) | |
| Behavioral characteristics (ever/never) | |||
| Ever injection drug use | 301 (10) | 122 (12) | 0.017 |
| Ever substance use | 1207 (38) | 549 (55) | <0.001 |
| Ever alcohol use | 1748 (56) | 594 (59) | 0.034 |
| Ever tobacco use | 1489 (47) | 621 (62) | <0.001 |
| HIV clinical history & laboratory measures at baseline | |||
| Years from HIV diagnosis to clinic entry | 0.8 [0.1–5.5] | 1.5 [0.2–7.5] | <0.001 |
| Year of clinic entry | 2006 [2002–2010] | 2006 [2002–2010] | 0.580 |
| CD4 cell count nadir at baseline (cells/μL) | 284 [117–449] | 300 [154–465] | 0.011 |
| CD4 cell count at baseline (cells/μL) | 420 [260–619] | 431 [270–630] | 0.193 |
| CD4/CD8 ratio at baseline | 0.5 [0.3–0.8] | 0.5 [0.3–0.8] | 0.732 |
| Log10 HIV RNA at baseline (copies/ml) | 2.4 [1.7–3.7] | 2.6 [1.7–4.0] | 0.079 |
| HIV RNA <400 at baseline (copies/ml) | 1838 (61) | 552 (57) | 0.018 |
| Follow-up years from baseline | 4.9 [2.3–8.9] | 5.1 [2.4–8.8] | 0.335 |
| Median number of HIV provider visits in 1st year of care | 5 [4–7] | 6 [4–8] | <0.001 |
| Died during observation period | 399 (13) | 153 (15) | 0.034 |
| Co-infections and NCDs at baseline | |||
| Ever hepatitis C infection | 400 (13) | 183 (18) | <0.001 |
| Ever chronic hepatitis B infection | 221 (7) | 79 (8) | 0.354 |
| One prevalent NCD at baseline | 1522 (48) | 572 (57) | <0.001 |
| Two or more prevalent NCDs at baseline | 654 (21) | 274 (27) | <0.001 |
| Psychiatric care | |||
| One or more VCCC mental health visits in 1st year of care | 316 (10) | 379 (38) | <0.001 |
| Any mood stabilizer during observation period | 27 (1) | 57 (6) | <0.001 |
| Any antidepressant use during observation period | 506 (16) | 633 (63) | <0.001 |
| Any anxiolytic use during observation period | 171 (5) | 159 (16) | <0.001 |
| Any antipsychotic use during observation period | 134 (4) | 205 (21) | <0.001 |
Chi square test for comparison of proportions; Wilcoxon rank sum test for comparison of distributions of continuous variables
Baseline indicates at the time of start of observation period (one year after clinic entry).
Abbreviations used:
IQR: interquartile range
NCD: non-communicable disease
VCCC: Vanderbilt Comprehensive Care Clinic
Of all patients included, 1255 (30%) had no prevalent or incident NCDs during the observation period (34% of those patients without a mood disorder and 25% of those patients with a mood disorder). Prevalent or incident multimorbidity with ≥2 NCDs occurred in 470 (47%) patients with baseline mood disorders and 1,219 (39%) of patients without mood disorders overall. Among all incident and prevalent NCD diagnoses, hypertension was the most common, diagnosed in 44% of patients with mood disorders and 38% of patients without mood disorders. This was followed by obesity (43 and 32% in patients with and without mood disorders, respectively), hyperlipidemia (37 and 32%), diabetes (12 and 10%), chronic kidney disease (14 and 13%), coronary heart disease (6 and 4%), non-AIDS cancers (4 and 4%), chronic liver disease (3 and 2%), and cerebrovascular disease (3 and 2%). In total, 37% of patients with mood disorders and 34% without mood disorders met criteria for our modified metabolic syndrome during the observation period.
The unadjusted cumulative incidences of first NCD, ≥2 NCDs, ≥3 NCDs, and metabolic syndrome with age as the unit of time for patients with and without mood disorder diagnosis at baseline are shown in Figure 2a–d, respectively. In multivariable competing risks models for risk of first NCD, prevalent mood disorder at baseline was associated with a 29% increased subhazard for first NCD (adjusted subhazard for mood disorder 1.29 [95%CI 1.06–1.57]) after adjusting for sex, race, ever substance abuse, ever alcohol, ever tobacco, year of clinic entry, time-varying CD4 cell count, time-varying CD4/CD8 ratio, time-varying log10 HIV RNA, and hepatitis C virus co-infection. The unadjusted and adjusted subhazard ratios for mood disorder and the other NCD outcomes are displayed in Figure 3. For most of the multimorbidity outcomes (≥2 NCDs, ≥3 NCDs, ≥4 NCDs, and ≥5 NCDs), a prevalent mood disorder diagnosis at baseline was associated with a 16–35% increased risk in unadjusted analyses, though 95% confidence intervals for some of these analyses included the null. The associations between mood disorder and multimorbidity outcomes were generally only slightly attenuated in the multivariable analyses which adjusted for baseline NCDs as well as the other demographic, behavioral, and clinical variables. Prevalent mood disorder was associated with a 29% increased risk of our metabolic syndrome multimorbidity outcome in adjusted analyses, with 95% confidence intervals suggesting compatibility with 2% to 64% increased risk.
Figure 2a-d.
Cumulative incidence plots for NCD outcomes using age as unit of time and death as competing risk. Curves represent the cumulative proportion of patients at each age with NCD outcomes, stratified by presence of mood disorder at baseline (dashed) or not (solid). a. Cumulative incidence of first NCD; b. Cumulative incidence of multimorbidity with ≥ 2 NCDs; c. Cumulative incidence of multimorbidity with ≥ 3 NCDs; d. Cumulative incidence of metabolic syndrome. As an example, Figure 1a shows that >80% of all patients without any NCDs at clinic entry developed at least one NCD by 50 years.
Figure 3.
Unadjusted and adjusted subhazard ratios for mood disorder and NCD outcomes with 95% confidence intervals, using age as unit of time. Unadjusted subhazard ratios reflect univariate subhazard of NCD outcomes associated with mood disorders (grey boxes). Subhazard of NCD outcomes associated with mood disorders adjusting for sex, race, substance use, alcohol use, tobacco use, year of clinic entry, chronic hepatitis C virus infection as covariates, and time-varying CD4 cell count, CD4.CD8 ratio, and HIV RNA (black boxes).
We performed two planned sensitivity analyses to better characterize the role of potential confounders and the association for mood disorders and NCD outcomes including: (1) using the baseline HIV laboratory values rather than time-varying values and (2) adding time-updated cumulative psychiatric medication exposures to the primary multivariable models. The results of these sensitivity analyses are shown in Table 2. Overall, the point estimates of the subhazard ratios changed minimally in the sensitivity analyses though 95% confidence intervals included the null in models, with the exception of first NCD outcome.
Table 2.
Adjusted subhazard ratios for mood disorder diagnosis and NCD outcomes in primary and sensitivity analysesa
| Primary analysisb aSHR [95% CI] |
Sensitivity Analysis 1c aSHR [95% CI] |
Sensitivity Analysis 2d aSHR [95% CI] |
|
|---|---|---|---|
| 1st NCD | 1.29 [1.06–1.57] | 1.26 [1.05–1.53] | 1.37 [1.11–1.68] |
| ≥ 2 NCDs | 1.04 [0.86–1.26] | 1.07 [0.89–1.29] | 1.07 [0.88–1.31] |
| ≥ 3 NCDs | 1.13 [0.92–1.39] | 1.10 [0.90–1.34] | 1.09 [0.87–1.36] |
| ≥ 4 NCDs | 1.21 [0.94–1.56] | 1.21 [0.95–1.56] | 1.07 [0.80–1.43] |
| ≥ 5 NCDs | 1.42 [0.95–2.13] | 1.42 [0.95–2.12] | 1.18 [0.76–1.85] |
| Metabolic syndrome | 1.29 [1.02–1.64] | 1.19 [0.94–1.50] | 1.26 [0.97–1.64] |
All models also included sex, race, substance use, alcohol use, tobacco use, year of clinic entry, and chronic hepatitis C virus infection as covariates and used age as unit of time. All models for ≥ 2 NCDs also included variable for prevalent NCD at baseline.
Model included time-varying HIV monitoring laboratory values (CD4 cell count, CD4/CD8 ratio, and HIV RNA)
Model included only baseline HIV monitoring laboratory values
Model included time-varying HIV monitoring laboratory values and cumulative psychiatric medication exposure time included in the model.
Abbreviations used:
aSHR: adjusted subhazard ratio
CI: confidence interval
NCD: non-communicable disease
Lastly, we used Cox models to evaluate whether prevalent mood disorders were associated with mortality risk among the 1,139 patients who developed multimorbidity (164 deaths). After adjusting for demographic, behavioral, and HIV laboratory values at the time of multimorbidity, the estimated hazard of death following multimorbidity was 11% higher for those with prevalent mood disorder diagnosis than those without, but confidence intervals for this estimate were wide, suggesting there is little evidence to make strong conclusions (adjusted hazard ratio [aHR]: 1.11 [95% confidence interval: 0.79–1.59]). In the multivariable model, older age (aHR: 1.65 [1.39–1.95]), female sex (aHR: 1.42 [0.98–2.06]), and higher log10 HIV RNA (aHR per 1 log10 increase: 1.21 [1.06–1.38]) appeared to be associated with increased risk of mortality following multimorbidity. Ever reporting alcohol use (aHR: 0.58 [0.41–0.82]) and higher CD4 cell count (aHR per 100 cells/μL increase: 0.85 [0.79–0.92]) were also associated with decreased risk of death.
Discussion
In our study of 4,140 PLWH receiving HIV care, prevalent mood disorders were associated with an increased risk of NCDs and multimorbidity. Diagnosed mood disorders were prevalent in 24% of patients with one year of clinic follow-up and approximately half of all patients included had prevalent or incident multimorbidity with ≥2 NCDs during the study period. Mood disorders were independently associated with increased risk of first NCD and multimorbidity from NCDs, particularly metabolic syndrome. These associations persisted after accounting for differences in immunologic and virologic measures, behavioral risk factors, and medication exposures.
Mood disorders were associated with approximately 30% increased risk of NCDs among patients without any baseline NCDs. In the general population, mood disorders, and depression in particular, have been associated with increased risk of non-communicable diseases including diabetes, cardiovascular disease, hypertension, and chronic lung disease18–22. The reasons for this association are complex and involve overlapping behavioral and social determinants of health. Additionally, alterations in a number of physiologic pathways are observed in both mood disorders and a number of NCDs, including the hypothalamic-pituitary axis, chronic stress, peripheral hormones, and increased measures of systemic inflammation11,23–26. The hypothesis of a shared pathophysiology is also supported by the observation that a number of NCDs are associated with development of depression, and depression is associated with exacerbation of chronic NCDs27,28. The role of mood disorders and risk of NCDs in PLWH has been less defined despite the higher prevalence of mood disorders and NCDs in this population. Importantly, PLWH have higher prevalence of smoking, increased measures of stress and systemic inflammation, and experience adverse social determinants of health that may contribute to increased risk of mood disorders and NCDs29,30. Among PLWH, depressive symptoms have been independently associated with more comorbid conditions in cross-sectional studies and have been prospectively associated with risk of hospitalizations31,32. In our study, prevalent mood disorders were predictive of first NCD after accounting for demographic, behavioral, and immunologic confounders. These results suggest that PLWH who have prevalent mood disorders may be at even higher risk for NCDs and may benefit from additional prevention and targeted treatment programs.
The association we observed between mood disorders and NCDs persisted, though was attenuated, when we looked at development of multiple NCDs and multimorbidity. Multimorbidity from two or more NCDs is a growing clinical challenge in the medical care of older PLWH33–35. Descriptions of multimorbidity in PLWH have focused on cardiovascular diseases, metabolic disorders, and other non-communicable diseases. Psychiatric diseases contribute to decreased quality of life, disability, and polypharmacy in PLWH and are important (though not always included) contributors to multimorbidity36. Furthermore, we demonstrated in this study that prevalent mood disorders remain predictive of increased number of co-morbid NCDs, suggesting a synergistic effect of mood disorders as both a component of and contributor to multimorbidity. Prevalent mood disorders were associated with a 16–35% increased risk of multimorbidity, an association that was attenuated after accounting for confounders. The development of multimorbidity is not independent as some NCDs contribute to risk of others (such as diabetes for cardiovascular disease). The persistence of the approximate 20% increased risk of multimorbidity associated with mood disorders across levels of multimorbidity suggests a lasting association of mood disorders and the overlapping pathophysiology of NCDs.
The most common NCDs observed in our cohort and among patients with multimorbidity were combinations of the components of a metabolic syndrome: hypertension, hyperlipidemia, obesity, and diabetes. Prevalent mood disorders were associated with a 29% increased risk of development of metabolic syndrome, an estimate that was unchanged in multivariable models and only slightly attenuated when cumulative psychiatric medication exposures were added to the model. A number of psychiatric medications, including antidepressants and atypical antipsychotics, have been associated with risk of weight gain and metabolic changes37,38. The association of mood disorders and metabolic syndrome itself was only minimally attenuated after accounting for medication exposure, suggesting that differences in psychiatric medication exposure alone do not account for the excess risk of metabolic multimorbidity observed in patients with mood disorders.
Coordinated mental health care has been shown to benefit HIV outcomes as well non-communicable disease outcomes in the general population39–41. We believe that the integration of psychiatric care within our HIV clinic (whereby receipt of both medical and psychiatric care improve retention in HIV care, other health behaviors, and coordination of therapies), may improve HIV outcomes, although we did not observe a notable difference in mortality after NCD multimorbidity by presence of mood disorder. Further examination and implementation of integrated psychiatric for PLWH is needed to improve HIV, mental health, and non-communicable disease outcomes.
This study is one of the largest longitudinal studies describing mood disorders and risk of NCD multimorbidity in PLWH. However, there are some limitations to consider. First, only documented mood disorder diagnoses in the first year of care were included in this study. Psychiatric diagnoses in the medical record were entered by physicians and mental health providers based upon clinical encounters. Given possible misclassification, we did not examine depression and bipolar affective disorders separately. As cyclical disorders, date of onset of mood disorders is difficult to obtain reliably from diagnoses entered into medical records. We used the one-year period to obtain prevalent diagnoses but did not capture diagnoses documented later in a patient’s medical course that were prevalent at baseline and did not have data on duration or severity of mood disorders at baseline to evaluate how these characteristics affect outcomes. For generalizability, we included patients in exposed and unexposed groups with other psychiatric diagnoses, some of which have also been associated with NCD risk. Further study into how multiple psychiatric diagnoses interact and contribute to NCD risk in PLWH is needed in larger cohort studies. Additionally, by requiring one year in care, our study may have introduced selection bias. While patients with <1 year in care had a lower prevalence of mood disorder diagnoses, our conclusions only reflect differences in NCD outcomes among patients with ≥1 year in care with and without mood disorders. Our analyses attempted to account for important demographic, behavioral, and immunologic confounders. We did not have detailed information on other social determinants of health (such as education, poverty, and trauma) nor detailed longitudinal data on important behaviors that may contribute to unmeasured confounding in our analyses (such as pack-years of tobacco use or time-varying substance use). Lastly, our results are limited by the available NCDs collected in our cohort. We do not systematically abstract data on conditions including chronic lung disease, heart failure, or arthritis which may be important causes of multimorbidity, particularly in patients with co-morbid mood disorders.
In conclusion, mood disorders were independently associated with risk of first NCD and multimorbidity from NCDs in this diverse cohort of PLWH in care. Mood disorders are a prevalent co-morbid condition in PLWH and may contribute to NCD risk due to overlapping and synergistic pathophysiology and risk factors. Further study into the prevention and treatment of NCDs in PLWH with mood disorders is needed.
Acknowledgements
We recognize the efforts of Daniel Rasbach, Paul No, Henry Heaton, James Logan, Asghar Kheshti, Carmen Bofill, William Ben Rogers, and Fernanda Maruri for their contributions toward the collection and validation of non-communicable disease diagnoses and death data from medical records. This study was supported by grant funding by the National Institutes of Health: K23 AI120875 (JLC) and by the NIH-funded Tennessee Center for AIDS Research (P30 AI110527).
Footnotes
Conferences at which data were presented:
- 9th International Workshop on HIV & Aging, New York City, September 2018
- 26th Conference on Retroviruses and Opportunistic Infections, Seattle, March 2019
Conflicts of Interest & Source of Funding: JLC is currently receiving a grant (K23AI112087) as is PFR (K01AI131895), and PFR, BES, MT, SSF, and TRS are supported by a separate grant (P30AI110527) from the National Institutes of Health. JRK has served as a consultant for Gilead. For the remaining authors none were declared.
References
- 1.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 Immune Defic Syndr. 2006;43(1):27–34. [DOI] [PubMed] [Google Scholar]
- 2.Smith CJ, Ryom L, Weber R, et al. Trends in underlying causes of death in people with HIV from 1999 to 2011 (D:A:D): a multicohort collaboration. Lancet. 2014;384(9939):241–248. [DOI] [PubMed] [Google Scholar]
- 3.Rasmussen LD, May MT, Kronborg G, et al. Time trends for risk of severe age-related diseases in individuals with and without HIV infection in Denmark: a nationwide population-based cohort study. Lancet HIV. 2015;2(7):e288–298. [DOI] [PubMed] [Google Scholar]
- 4.Freiberg MS, Chang CC, Kuller LH, et al. HIV infection and the risk of acute myocardial infarction. JAMA Intern Med. 2013;173(8):614–622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kendall CE, Wong J, Taljaard M, et al. A cross-sectional, population-based study measuring comorbidity among people living with HIV in Ontario. BMC Public Health. 2014;14:161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Goulet JL, Fultz SL, Rimland D, et al. Aging and infectious diseases: do patterns of comorbidity vary by HIV status, age, and HIV severity? Clin Infect Dis. 2007;45(12):1593–1601. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hasse B, Ledergerber B, Furrer H, et al. Morbidity and aging in HIV-infected persons: the Swiss HIV cohort study. Clin Infect Dis. 2011;53(11):1130–1139. [DOI] [PubMed] [Google Scholar]
- 8.Schouten J, Wit FW, Stolte IG, et al. Cross-sectional comparison of the prevalence of age-associated comorbidities and their risk factors between HIV-infected and uninfected individuals: the AGEhIV cohort study. Clin Infect Dis. 2014;59(12):1787–1797. [DOI] [PubMed] [Google Scholar]
- 9.Colton CW, Manderscheid RW. Congruencies in increased mortality rates, years of potential life lost, and causes of death among public mental health clients in eight states. Prev Chronic Dis. 2006;3(2):A42. [PMC free article] [PubMed] [Google Scholar]
- 10.Tanaka T, Matsuda T, Hayes LN, et al. Infection and inflammation in schizophrenia and bipolar disorder. Neurosci Res. 2017;115:59–63. [DOI] [PubMed] [Google Scholar]
- 11.Liu CS, Adibfar A, Herrmann N, Gallagher D, Lanctot KL. Evidence for Inflammation-Associated Depression. Curr Top Behav Neurosci. 2017;31:3–30. [DOI] [PubMed] [Google Scholar]
- 12.Nurutdinova D, Chrusciel T, Zeringue A, et al. Mental health disorders and the risk of AIDS-defining illness and death in HIV-infected veterans. AIDS. 2012;26(2):229–234. [DOI] [PubMed] [Google Scholar]
- 13.Do AN, Rosenberg ES, Sullivan PS, et al. Excess burden of depression among HIV-infected persons receiving medical care in the united states: data from the medical monitoring project and the behavioral risk factor surveillance system. PLoS One. 2014;9(3):e92842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.So-Armah K, Gupta SK, Kundu S, et al. Depression and all-cause mortality risk in HIV-infected and HIV-uninfected US veterans: a cohort study. HIV medicine. 2019;20(5):317–329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ciesla JA, Roberts JE. Meta-analysis of the relationship between HIV infection and risk for depressive disorders. Am J Psychiatry. 2001;158(5):725–730. [DOI] [PubMed] [Google Scholar]
- 16.Castilho JL, Shepherd BE, Koethe J, et al. CD4+/CD8+ ratio, age, and risk of serious noncommunicable diseases in HIV-infected adults on antiretroviral therapy. AIDS. 2016;30(6):899–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.White JR, Chang CC, So-Armah KA, et al. Depression and Human Immunodeficiency Virus Infection Are Risk Factors for Incident Heart Failure Among Veterans: Veterans Aging Cohort Study. Circulation. 2015;132(17):1630–1638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lotfaliany M, Bowe SJ, Kowal P, Orellana L, Berk M, Mohebbi M. Depression and chronic diseases: Co-occurrence and communality of risk factors. J Affect Disord. 2018;241:461–468. [DOI] [PubMed] [Google Scholar]
- 19.Goldstein BI, Carnethon MR, Matthews KA, et al. Major Depressive Disorder and Bipolar Disorder Predispose Youth to Accelerated Atherosclerosis and Early Cardiovascular Disease: A Scientific Statement From the American Heart Association. Circulation. 2015;132(10):965–986. [DOI] [PubMed] [Google Scholar]
- 20.Yu M, Zhang X, Lu F, Fang L. Depression and Risk for Diabetes: A Meta-Analysis. Can J Diabetes. 2015;39(4):266–272. [DOI] [PubMed] [Google Scholar]
- 21.Maatouk I, Herzog W, Bohlen F, et al. Association of hypertension with depression and generalized anxiety symptoms in a large population-based sample of older adults. J Hypertens. 2016;34(9):1711–1720. [DOI] [PubMed] [Google Scholar]
- 22.Seldenrijk A, Vogelzangs N, Batelaan NM, Wieman I, van Schaik DJ, Penninx BJ. Depression, anxiety and 6-year risk of cardiovascular disease. J Psychosom Res. 2015;78(2):123–129. [DOI] [PubMed] [Google Scholar]
- 23.Paz-Filho G, Licinio J, Wong ML. Pathophysiological basis of cardiovascular disease and depression: a chicken-and-egg dilemma. Braz J Psychiatry. 2010;32(2):181–191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Halaris A Inflammation-Associated Co-morbidity Between Depression and Cardiovascular Disease. Curr Top Behav Neurosci. 2017;31:45–70. [DOI] [PubMed] [Google Scholar]
- 25.Catena-Dell’Osso M, Rotella F, Dell’Osso A, Fagiolini A, Marazziti D. Inflammation, serotonin and major depression. Curr Drug Targets. 2013;14(5):571–577. [DOI] [PubMed] [Google Scholar]
- 26.Marazziti D, Rutigliano G, Baroni S, Landi P, Dell’Osso L. Metabolic syndrome and major depression. CNS Spectr. 2014;19(4):293–304. [DOI] [PubMed] [Google Scholar]
- 27.Iyer AS, Bhatt SP, Garner JJ, et al. Depression Is Associated with Readmission for Acute Exacerbation of Chronic Obstructive Pulmonary Disease. Ann Am Thorac Soc. 2016;13(2):197–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Jani BD, Mair FS, Roger VL, Weston SA, Jiang R, Chamberlain AM. Comorbid Depression and Heart Failure: A Community Cohort Study. PLoS One. 2016;11(6):e0158570. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Mdodo R, Frazier EL, Dube SR, et al. Cigarette smoking prevalence among adults with HIV compared with the general adult population in the United States: cross-sectional surveys. Ann Intern Med. 2015;162(5):335–344. [DOI] [PubMed] [Google Scholar]
- 30.Deeks SG. HIV infection, inflammation, immunosenescence, and aging. Annu Rev Med. 2011;62:141–155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Havlik RJ, Brennan M, Karpiak SE. Comorbidities and depression in older adults with HIV. Sex Health. 2011;8(4):551–559. [DOI] [PubMed] [Google Scholar]
- 32.Hotton AL, Weber KM, Hershow RC, et al. Prevalence and predictors of hospitalizations among HIV-Infected and at-risk HIV-Uninfected Women. J Acquir Immune Defic Syndr. 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Smit M, Brinkman K, Geerlings S, et al. Future challenges for clinical care of an ageing population infected with HIV: a modelling study. Lancet Infect Dis. 2015;15(7):810–818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wong C, Gange SJ, Moore RD, et al. Multimorbidity Among Persons Living with HIV in the U.S. Clin Infect Dis. 2017. [Google Scholar]
- 35.Rodriguez-Penney AT, Iudicello JE, Riggs PK, et al. Co-morbidities in persons infected with HIV: increased burden with older age and negative effects on health-related quality of life. AIDS Patient Care STDS. 2013;27(1):5–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Millar BM, Starks TJ, Gurung S, Parsons JT. The Impact of Comorbidities, Depression, and Substance Use Problems on Quality of Life Among Older Adults Living With HIV. AIDS Behav. 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Gramaglia C, Gambaro E, Bartolomei G, et al. Increased Risk of Metabolic Syndrome in Antidepressants Users: A Mini Review. Front Psychiatry. 2018;9:621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Pramyothin P, Khaodhiar L. Metabolic syndrome with the atypical antipsychotics. Curr Opin Endocrinol Diabetes Obes. 2010;17(5):460–466. [DOI] [PubMed] [Google Scholar]
- 39.Aggarwal R, Pham M, Dillingham R, McManus KA. Expanded HIV Clinic-Based Mental Health Care Services: Association With Viral Suppression. Open Forum Infect Dis. 2019;6(4):ofz146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Coyle RP, Schneck CD, Morrow M, et al. Engagement in Mental Health Care is Associated with Higher Cumulative Drug Exposure and Adherence to Antiretroviral Therapy. AIDS Behav. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ngo VK, Rubinstein A, Ganju V, et al. Grand challenges: Integrating mental health care into the non-communicable disease agenda. PLoS Med. 2013;10(5):e1001443. [DOI] [PMC free article] [PubMed] [Google Scholar]