Abstract
Alcohol use was associated with elevated COVID-19 risk in the general population. People with HIV (PWH) have high prevalences of alcohol use. To evaluate the effect of alcohol use on COVID-19 risks among PWH, we estimated the risk of COVID-19 diagnosis and COVID-19-related hospitalization among PWH in routine care at 8 HIV primary care centers that contributed data to the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort according to their alcohol use just prior to the COVID-19 pandemic. The CNICS data repository includes demographic characteristics, clinical diagnoses, and laboratory test results from electronic medical records and other sources. Alcohol use, substance use, and mental health symptoms were self-reported on tablet-based standardized surveys. Alcohol use was categorized according to standard, sex-specific Alcohol Use Disorder Identification Test-Consumption instrument cut-offs. We followed 5,496 PWH (79% male, 48% Black race, median age=53 years) from March 1, 2020 to December 31, 2020. Relative to PWH with no baseline alcohol use, the adjusted hazard ratio (aHR) of COVID-19 diagnosis was 1.09 (95% confidence interval [CI]: 0.78, 1.51) for lower-risk drinking and 1.19 (95%CI: 0.81, 1.73) for unhealthy drinking. The aHR of COVID-19-related hospitalization was 0.82 (95%CI: 0.33, 1.99) for lower-risk drinking and 1.25 (95%CI: 0.50, 3.09) for unhealthy drinking. Results were not modified by recent cocaine or non-prescribed opioid use, depressive symptoms, or diagnoses of alcohol use disorder. The study suggested a slightly increased, but not statistically significant risk of COVID-19 diagnosis and hospitalization associated with unhealthy alcohol use.
Keywords: Alcohol Use, COVID-19, HIV, Substance Use, Mental Disorder
In the first year of the Severe Acute Respiratory Syndrome Cornavirus 2 (SARS-CoV-2) epidemic, globally, over 85 million people were infected and over 2 million people died of coronavirus disease 2019 (COVID-19)[1]. People with COVID-19 and alcohol use disorders (AUD) had higher odds of hospitalization and all-cause mortality[2]. People with substance use disorders (SUD) and mental illnesses also have an increased risk for SARS-CoV-2 infection and poor outcomes following infection[3,4]. Although treated HIV infection without immune suppression does not independently appear to confer increased risk of of adverse COVID-19 outcomes[5], people with HIV (PWH) may be at high risk of COVID-19 morbidity due to the high prevalence of alcohol use, AUD, substance use, SUD, and mental health disorders. Understanding modifiable factors associated with increased risk of SARS-CoV-2 infection and severe COVID-19 may help people make informed decisions about their health and guide policy to protect especially vulnerable groups. The prevelance of low alcohol use (any alcohol use that falls below the threshold of hazardous use: a score ≥3 for women and ≥4 for men on the Alcohol Use Disorders Identification Test – Consumption questions (AUDIT-C)) is about 41% among PWH, and 27% of PWH have recent hazardous alcohol use[6]. Recent alcohol use represents a potentially modifiable factor that might be targeted for intervention among PWH during COVID-19 surges if it is associated with worse COVID-19 outcomes.
This is particularly salient for PWH who have similarly high prevalences of illicit drug use (50%) and drug dependency (12%)[7,8], major depression (36%)[8,9], panic disorder (11%)[10], and AUD (42%)[11]. PWH also have lower socioeconomic status (associated with higher probability of exposure to SARS-CoV-2 through occupation and high-density housing and greater challenges to accessing preventive tools like high-quality masks, rapid SARS-CoV-2 tests, and air purifiers), and higher prevalence of other comorbidities that put them at potentially higher risk of infection if exposed and higher risk of poor outcomes if infected[12].
Because of the unique challenges facing PWH, it is unclear whether the effect of alcohol use on COVID-19 outcomes observed in the general population generalize to PWH. In the general population, chronic unhealthy alcohol use contributes to lower T lymphocyte count, impairing proliferation and cell turnover, and the promotion of pro-inflammatory states, characterized by elevated levels of proinflammatory cytokines[13]. Chronic alcohol use also impairs macrophage functions in lung alveoli and decreases the function and number of natural killer cells, which are responsible for eliminating infected cells[13].
There is limited research on the effects of alcohol use on the risks of COVID-19 among PWH. The goal of this study was to describe the COVID-19 case and hospitalization rates for PWH according to alcohol use just prior to the start of the pandemic. In secondary analyses, we tested whether the effects of alcohol use were modified by substance use, mental illness, or history of AUD.
METHODS
Study sample
We used data from the Centers for AIDS Research (CFAR) Network of Integrated Clinical Systems (CNICS), a prospective clinical cohort study. Full details are available elsewhere[14,15], but briefly, CNICS participants are adults living with HIV who enrolled in clinical care (defined as attending ≥2 clinic visits within a year) at one of eight academic medical centers: Johns Hopkins University (Baltimore, MD), Case Western Reserve University (Cleveland, OH), Fenway Health (Boston, MA), University of Alabama at Birmingham, University of California-San Diego, University of North Carolina at Chapel Hill, University of Washington (Seattle, WA), and University of California-San Francisco. Data were obtained from the CNICS data repository which includes demographic information including race/ethnicity, sex or gender, and probable routes of HIV acquisition self-reported at enrollment. It includes clinical information, including diagnoses, laboratory results, and prescribed medications from electronic medical records and other data sources. Site-specific data were submitted to the CNICS Data Management Core where they went through quality assessment and standardization. Certain outcome data underwent additional protocol-driven medical record review and centralized adjudication by a panel of expert physicians. Patient Reported Outcomes (PROs) were collected for a subset of PWH through tablet-based surveys approximately every 4-6 months in conjunction with a clinic visit. No patient characteristics are used to target PRO collection; PRO completion is determined mainly by clinic flow and patient availability[15,16]. Institutional review boards for every institution have given their approval for CNICS research and this study.
For this study, we included CNICS participants who attended at least one HIV primary care visit between March 1, 2019 and March 1, 2020 and who completed at least one PRO survey during the same period.
Exposure
Alcohol use was self-reported using the United States Alcohol Use Disorders Identification Test-Consumption questions (USAUDIT-C)[17] as part of the routine PROs. The USAUDIT-C is a 3-item questionnaire asking about participants’ average frequency, quantity, and intensity (heavy episodic drinking or “binge” drinking) of alcohol use over the past year. Scores range from 0 to 12. Unhealthy alcohol use was defined using standard cutoffs as a score of ≥3 for women and ≥4 for men (based on birth sex). Lower-risk alcohol use was defined as a score of 1-2 for women and 1-3 for men. Anyone with a score of 0 was classified as having no recent alcohol use. If patients completed multiple PROs during the study period, we used alcohol use reported on the latest PRO (closest prior to March 1, 2020).
Modifiers
Depression symptoms were assessed using a module from the Patient Health Questionnaire (PHQ). PHQ-8 depression module measures depressive symptoms over the past 2 weeks, with scores ranging from 0 to 24. Participants were classified as having depression symptoms (moderate to severe) if they had a PHQ-8 score ≥10. A PHQ-8 score of ≥10 is 88% sensitive and 88% specific for presence of major depressive disorder[18,19]. Recent (prior three months) cocaine and/or non-prescribed opioids were defined as any recent use based on the Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST)[20]. We identified participants who had a diagnosis of AUD prior to March 1, 2020 based on clinical diagnosis codes (ICD-9 codes: 305.0X, 303.X; ICD-10 codes: F10.1X, F10.2X).
Covariates
Age, sex at birth, race/ethnicity, geographic area, and HIV acquisition risk factors were obtained upon CNICS cohort enrollment. We adjusted for CD4 cell count and viral suppression status from the most recent lab results prior to March 1, 2020. We also adjusted for chronic comorbid conditions based on the most recent indicators prior to March 1, 2020. Chronic comorbid conditions included diabetes, hypertension, body mass index (BMI), cigarette smoking, hepatitis C virus (HCV), chronic obstructive pulmonary disease (COPD), high-density lipoprotein (HDL) cholesterol, total cholesterol, estimated glomerular filtration rate (eGFR), and liver fibrosis score (FIB-4). Diabetes was defined based on presence of any of the following: hemoglobin A1c (HbA1c) ≥6.5, a prescription for diabetes-specific medications, or a diagnosis of diabetes (ICD-10 codes: E08-13; ICD-9 codes: 250.XX). Hypertension status was categorized into no hypertension, controlled, and uncontrolled hypertension. No hypertension was defined as both systolic blood pressure <140mmHg and diastolic blood pressure <90mmHg. Controlled hypertension was defined as having normal systolic and diastolic blood pressure and prescriptions of anti-hypertensive medication. Uncontrolled hypertension was defined as either systolic or diastolic blood pressure above the threshold regardless of the presence of antihypertensive medications. Cigarette smoking status (never vs. former vs. recent user) was obtained through self-report on PRO. Panic symptoms and panic disorder wer assessed on the Patient Health Questionnaire-Panic Disorder (PHQ-PD) module on the PRO [21]. Participants were classified as having screened positive for panic symptoms if they had an anxiety attack in the past 4 weeks from PHQ-PD and classified as having screened positive for panic disorder if they then answered “yes” to all subsequent four questions about feelings of panic or fear. Current HCV infection was defined as a positive HCV viral load test or a recent diagnosis of HCV within a year (ICD-9 codes: 070.44, 070.51, 070.54; ICD-10 codes: B18.2, B19.2, B19.9). COPD was defined based on clinical diagnoses (ICD-9 codes: 491.2, 496; ICD-10 codes: J41.X, J42, J44.X). BMI was calculated using baseline height and most recent weight. HDL cholesterol and total cholesterol were obtained from the most recent lab tests. FIB-4 was computed using aspartate aminotransferase, alanine aminotransferase, and platelet count and categorized into ≤1.45 (advanced fibrosis unlikely), >1.45 and ≤3.25 (indeterminate), and >3.25 (likely advanced fibrosis)[22]. eGFR was calculated based on most recent standardized serum creatinine, sex, and age[23].
Outcomes
Outcomes of the study were COVID-19 cases and hospitalizations due to COVID-19 from March 1st, 2020 to December 31st, 2020. There was not systematic surveillance testing for SARS-CoV-2 infection in the first year of the pandemic. COVID-19 cases were identified from provider documented diagnoses (ICD-10 codes) and positive SARS-CoV-2 tests that were done on the basis of clinically compatible symptoms and diagnostic testings. COVID-19 cases and COVID-19 hospitalizations were verified by each site and centrally reviewed and adjudicated by a panel of clinicians[24].
Statistical Analysis
We conducted separate analyses for time to a verified COVID-19 case and time to a COVID-19 hospitalization. We followed participants from March 1, 2020, to the first instance of the outcome of interest, death from any cause, or administrative censoring on December 31, 2020, whichever came first. There were 52 participants who died between March 1 and December 31, 2020. We used Cox proportional hazards models to estimate the hazard ratios of COVID-19 cases and COVID-19 hospitalizations comparing people with unhealthy or lower-risk alcohol use to people with no alcohol use in the year leading up to the start of the pandemic.
We used inverse probability of exposure weighting (IPW) to adjust for baseline covariates defined above: age, sex, race/ethnicity, study site, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, cigarette smoking, CD4 cell count, viral suppression, diabetes, hypertension, BMI, HCV, COPD, HDL cholesterol, total cholesterol, eGFR, and FIB-4. We estimated the weights with a multinomial regression model to get the predicted probability of each participant having their observed alcohol use level based on the above covariates. To minimize residual confounding, we modeled continuous variables (age, CD4 cell count, BMI, eGFR, HDL cholesterol, and total cholesterol) flexibly using restricted quadratic splines with knots at the 5th, 35th, 65th, and 95th percentiles[25]. We calculated the weights for each participant as the inverse of the conditional probability of reporting the alcohol use that they reported, and we stabilized weights by the marginal probability of each participant’s alcohol use category[26].
We used multiple imputation to handle missing data on baseline alcohol use and covariates. Table 1 shows the frequency of missing data. We generated 40 complete datasets using multiple imputation by chained equations based on all covariates, exposure, survival time and indicators for the outcomes, COVID-19 diagnosis and COVID-19 hospitalization[27]. We re-estimated the weights and adjusted hazard ratios (HRs) in each of the imputed datasets and pooled the 40 estimates using Rubin’s rules. P-values were considered statistically significant at P < 0.05. All analyses were performed using R (version 4.2.2).
Table 1.
Participant characteristics, N (%) unless otherwise specified, of 5,496 persons in the CNICS cohort who were in care for HIV and completed PRO survey, March 1, 2019 – March 1, 2020, stratified by alcohol use status
| Alcohol Use | |||||
|---|---|---|---|---|---|
| Total | None | Lower-risk | Unhealthy | Missing | |
| (N=5496) | (N=1833) | (N=1970) | (N=1257) | (N=436) | |
| Female | 1160 (21) | 525 (29) | 303 (15) | 241 (19) | 91 (21) |
| Age in 2020 (years)a | 53 [41, 60] | 55 [47, 62] | 51 [40, 59] | 49 [36, 58] | 53 [43, 60] |
| Race/ethnicity | |||||
| Non-Hispanic White | 2053 (37) | 588 (32) | 774 (39) | 507 (40) | 184 (42) |
| Non-Hispanic Black | 2660 (48) | 1012 (55) | 942 (48) | 545 (43) | 161 (37) |
| Hispanic | 593 (11) | 174 (10) | 183 (9) | 164 (13) | 72 (17) |
| Other/Unknown | 190 (4) | 59 (3) | 71 (4) | 41 (3) | 19 (4) |
| HIV transmission risk factor | |||||
| MSM | 3306 (60) | 877 (48) | 1348 (68) | 809 (64) | 272 (62) |
| IDU | 535 (10) | 274 (15) | 133 (7) | 91 (7) | 37 (9) |
| Insurance status | |||||
| Public Insurance | 1730 (53) | 728 (64) | 546 (47) | 337 (48) | 119 (50) |
| Private Insurance | 1452 (45) | 397 (35) | 591 (51) | 354 (50) | 110 (47) |
| Uninsured or self-paid | 58 (2) | 15 (1) | 25 (2) | 11 (2) | 7 (3) |
| Missing | 2256 | 693 | 808 | 555 | 200 |
| HDL cholesterola | 46 [38, 56] | 44 [37, 54] | 45 [38, 54] | 50 [41, 61] | 45 [37, 57] |
| Missing | 137 | 33 | 44 | 50 | 10 |
| Total cholesterola | 173 [146, 201] | 168 [141, 197] | 173 [148, 202] | 179 [155, 204] | 173 [145, 198] |
| Missing | 116 | 28 | 36 | 44 | 8 |
| COPD | 627 (11) | 301 (16) | 160 (8) | 110 (9) | 56 (13) |
| CD4 Cell Count, cells/mm3,a | 649 [451, 889] | 635 [426, 875] | 659 [467, 893] | 661 [463, 900] | 648 [426, 900] |
| Missing | 8 | 1 | 1 | 5 | 1 |
| Viral Suppression | 5156 (94) | 1717 (94) | 1857 (94) | 1182 (94) | 400 (92) |
| Missing | 6 | 2 | 2 | 2 | 0 |
| HCV | 836 (15) | 411 (22) | 235 (12) | 137 (11) | 53 (12) |
| FIB-4 score | |||||
| ≤1.45 | 4107 (75) | 1255 (69) | 1557 (79) | 981 (78) | 314 (72) |
| 1.46-3.25 | 1233 (23) | 516 (28) | 369 (19) | 242 (19) | 106 (24) |
| >3.25 | 140 (3) | 58 (3) | 39 (2) | 27 (2) | 16 (4) |
| Missing | 16 | 4 | 5 | 7 | 0 |
| Diabetes | 1226 (22) | 559 (31) | 410 (21) | 161 (13) | 96 (22) |
| Hypertension | |||||
| Controlled | 2028 (40) | 771 (45) | 694 (38) | 393 (34) | 170 (42) |
| Uncontrolled | 862 (17) | 336 (20) | 278 (15) | 180 (16) | 68 (17) |
| Missing | 439 | 130 | 159 | 117 | 33 |
| CKD (eGFR), mL/min/1.73 m2<60 | 837 (15) | 400 (22) | 269 (14) | 101 (8) | 67 (15) |
| Missing | 1 | 0 | 0 | 1 | 0 |
| BMI, kg/m2,a | 27.5 [24.2, 31.8] | 27.7 [24.3, 32.4] | 27.5 [24.4, 31.7] | 27.1 [23.8, 31.2] | 27.3 [24.0, 31.5] |
| Cigarette smoking status | |||||
| Former | 1571 (30) | 517 (29) | 599 (31) | 404 (33) | 51 (34) |
| Recent | 1417 (28) | 483 (27) | 482 (25) | 400 (32) | 52 (34) |
| Missing | 345 | 30 | 15 | 16 | 284 |
| Depression, moderate/severe | 856 (16) | 265 (15) | 257 (13) | 218 (18) | 116 (27) |
| Missing | 34 | 12 | 5 | 3 | 14 |
| Panic Disorder | |||||
| Panic Symptoms | 643 (12) | 178 (10) | 230 (12) | 188 (15) | 47 (14) |
| Panic Disorder | 525 (10) | 165 (9) | 166 (9) | 154 (12) | 40 (12) |
| Missing | 173 | 29 | 25 | 12 | 107 |
| Cocaine use | |||||
| Former | 1517 (31) | 585 (33) | 519 (27) | 392 (33) | 21 (25) |
| Recent | 336 (7) | 56 (3) | 114 (6) | 159 (13) | 7 (8) |
| Missing | 537 | 55 | 59 | 70 | 353 |
| Opioids use | |||||
| Former | 778 (16) | 357 (21) | 235 (13) | 178 (15) | 8 (10) |
| Recent | 156 (3) | 49 (3) | 56 (3) | 48 (4) | 3 (4) |
| Missing | 656 | 115 | 97 | 91 | 353 |
Abbreviations: BMI, body mass index; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; FIB-4, fibrosis-4; HCV, hepatitis C virus; HDL, high-density lipoprotein
Median (Q1, Q3)
Secondary Analyses
Alcohol use and particularly hazardous alcohol use commonly co-occurs with other substance use and with mental health disorders. In particular, the correlation between hazardous alcohol use and substance use was strongest for cocaine (14% of PWH in CNICS reporting hazardous alcohol use reported current cocaine use versus only 3% of people reporting no use) versus methamphetamine (11% of people reporting hazardous alcohol use reported current methamphetamine use versus 9% of people reporting no use) or opioids (3% of people reporting hazardous alcohol use reported current opioid use versus 2% of people reporting no use)[6]. In light of the accelerating opioid epidemic[28] and the potential for interactions between opioids and alcohol use[29], we also considered opioid use a potential modifier of interest. Depressive symptoms have been shown to interact with hazardous alcohol use to negatively impact engagement in HIV care[30]. Finally, the effect of recent alcohol use on engagement in HIV care depends on whether patients also had an AUD[31]. To investigate possible modification of the effect of alcohol use on COVID-19 cases, we fit separate, IP-weighted Cox proportional hazards models that included indicators for alcohol use and either substance use (cocaine and non-prescribed opioids), depression, or AUD, and product terms between alcohol use and the modifier. We estimated stratum-specific estimates of the association between alcohol use and infection, and tested for the presence of modification using the p-values of the product terms. The available sample size did not support stratified COVID-19 hospitalization estimates.
To attempt to more directly examine whether alcohol use affected severity of illness among people who were infected with SARS-CoV-2, we also fit unadjusted and adjusted logistic models to estimate the odds of hospitalization among the subset of people who had verified COVID-19 cases. Our primary analyses were conducted in the entire sample because restricting to people with COVID-19 might produced biased estimates if there are factors that are associated with susceptibility to infection and severity of disease[32].
Finally, since FIB-4 and hypertension were measured prior to or concurrent with alcohol use, these clinical indicators could influence individuals to modify their alcohol consumption. Therefore, we present our primary results adjusted for those two variables. However, FIB-4 and hypertension might mediate rather than confound the relationship between alcohol consumption and COVID-19 outcomes (alcohol use leads to liver fibrosis, which might predispose people to worse COVID-19 outcomes). Therefore, we conducted sensitivity analyses excluding FIB-4 and hypertension from the adjustment set in all models.
RESULTS
There were 5496 PWH who attended an HIV primary care visit and completed at least one PRO survey between March 1, 2019 and March 1, 2020. (Among participants who attended an HIV primary care visit in this period, 42% completed at least one PRO survey.) The majority were men (79%) and about half were Black (48%). The median age at the start of follow up was 53 years (interquartile range (IQR): 41, 60). The median CD4 cell count was 649 cells/mm3 (IQR: 451, 889) and 94% of participants had a suppressed viral load. Comorbid conditions were common: 22% of participants had diabetes, 17% had uncontrolled hypertension, 11% had COPD, and 15% had current HCV infection. Seven percent of participants self-reported recent cocaine use and 3% recent opioid use. Depressive symptoms and panic symptoms were common in the study sample: 16% reported moderate to severe depression symptoms, 12% reported a panic attack, and 10% screened positive for panic disorder (Table 1). Thirty-six percent (36%) participants self-reported lower-risk alcohol use and 23% self-reported unhealthy alcohol use.
There were 232 cases of COVID-19 and 33 COVID-19 hospitalizations in our study sample from March 1 to December 31, 2020. In unadjusted analyses, relative to no recent alcohol use, lower-risk and unhealthy alcohol use were associated with decreased hazard of COVID-19 cases, although the association was not statistically significant (hazard ratio (HR) for COVID-19 cases comparing lower-risk alcohol use to no use = 0.89, 95% confidence interval (CI): 0.66, 1.19; HR for unhealthy alcohol use = 0.96, 95% CI: 0.69, 1.33). After adjustment, the associations switched directions and alcohol use was associated with increased hazard of COVID-19 cases, although the association was not statistically significant (adjusted HR (aHR) for lower-risk alcohol use = 1.09, 95% CI: 0.78, 1.51; aHR for unhealthy alcohol use = 1.19, 95% CI: 0.81, 1.73) (Table 2).
Table 2.
Crude and adjusteda hazard ratio of COVID-19 cases and COVID-19 hospitalizations associated with pre-pandemic self-reported alcohol use among people with HIV in the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS), March 1, 2020 – December 31, 2020
| Events | Person- years |
Rates per person-year |
Crude Hazard Ratio |
Adjusted Hazard Ratioa |
|
|---|---|---|---|---|---|
| COVID-19 cases | |||||
| None alcohol use | 89 | 1495 | 0.060 | 1.00 | 1.00 |
| Lower-risk alcohol use | 84 | 1609 | 0.052 | 0.89 (0.66, 1.19) | 1.09 (0.78, 1.51) |
| Unhealthy alcohol use | 59 | 1030 | 0.057 | 0.96 (0.69, 1.33) | 1.19 (0.81, 1.73) |
| COVID-19 hospitalization | |||||
| None alcohol use | 15 | 1515 | 0.010 | 1.00 | 1.00 |
| Lower-risk alcohol use | 9 | 1635 | 0.006 | 0.56 (0.25, 1.28) | 0.82 (0.33, 1.99) |
| Unhealthy alcohol use | 9 | 1044 | 0.009 | 0.87 (0.38, 1.99) | 1.25 (0.50, 3.09) |
Adjusted for baseline age, sex, race/ethnicity, study sites, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, diabetes, hypertension, BMI, cigarette smoking, HCV, COPD, CD4 count, viral suppression status, HDL cholesterol, total cholesterol, eGFR, and FIB-4 score.
Similarly, in unadjusted analyses, lower-risk and unhealthy alcohol use was associated with a lower hazard of COVID-19 hospitalization compared to no use (HR for lower-risk alcohol use = 0.56, 95% CI: 0.25, 1.28; HR for unhealthy alcohol use = 0.82, 95% CI: 0.33, 1.99). After accounting for potential confounders, the aHR of COVID-19 hospitalization comparing lower-risk alcohol use to no use was 0.87 (95% CI: 0.38, 1.99) and the aHR comparing unhealthy alcohol use to no use was 1.25 (95% CI: 0.50, 3.09). None of the above associations were statistically significant (Table 2). Figure 1 and Figure 2 showed unadjusted and adjusted cumulative incidence curves for COVID-19 cases for each stratum of alcohol use from March 1, 2020 to December 31, 2020. There is no evidence that the risk or hazard (rate of change of curves) of COVID-19 within strata of alcohol use was different across the study period.
Figure 1.
Unadjusted cumulative incidence of COVID-19 disease, March 1-December 31, 2020, among 5496 people with HIV enrolled in routine HIV care in the Centers for AIDS Research Network of Integrated Clinical Systems who self-reported their alcohol consumption from March 1, 2019-February 28, 2020, stratified by baseline alcohol use.
Figure 2.
Adjusted cumulative incidence of COVID-19 disease, March 1-December 31, 2020, among 5496 people with HIV enrolled in routine HIV care in the Centers for AIDS Research Network of Integrated Clinical Systems who self-reported their alcohol consumption from March 1, 2019-February 28, 2020, stratified by baseline alcohol use. Adjusted for baseline age, sex, race/ethnicity, study sites, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, diabetes, hypertension, BMI, cigarette smoking, HCV, COPD, CD4 count, viral suppression status, HDL cholesterol, total cholesterol, eGFR, and FIB-4 score
Secondary Analyses
There was no evidence that substance use (cocaine and non-prescribed opioids), depression, or AUD modified the associations between alcohol use and COVID-19 cases (Table 3). There was no evidence that alcohol use was associated with severe illness among people who had diagnosed COVID-19 (Appendix Table 1). Estimates from sensitivity analyses excluding FIB-4 and hypertension from the adjustment set were within 0.03 of primary results (Appendix Table 2, Appendix Table 3).
Table 3.
Adjusteda hazard ratios (aHR) and 95% confidence intervals for the association between alcohol use and COVID-19 disease stratified by cocaine use, opioid use, depression, and alcohol use disorder.
| Lower-risk vs. No Alcohol Use aHR |
P-Valuec | Unhealthy vs. No Alcohol Use aHR |
P-Valued | |
|---|---|---|---|---|
| Cocaine use | ||||
| Never | 1.10 (0.77, 1.59) | (ref.) | 1.07 (0.67, 1.69) | (ref.) |
| Former | 0.43 (0.20, 0.91) | 0.40 | 0.71 (0.38, 1.34) | 0.71 |
| Recent | 0.86 (0.34, 2.21) | 0.25 | 1.05 (0.46, 2.43) | 0.17 |
| Opioid use | ||||
| Never | 1.12 (0.78, 1.61) | (ref.) | 1.21 (0.79, 1.86) | (ref.) |
| Former | 0.88 (0.37, 2.09) | 0.46 | 1.29 (0.61, 2.72) | 0.85 |
| Recent | 1.18 (0.37, 3.78) | 0.15 | b | b |
| Depression | ||||
| No to Mild | 1.00 (0.71, 1.40) | (ref.) | 1.12 (0.74, 1.68) | (ref.) |
| Moderate/Severe | 1.13 (0.54, 2.34) | 0.30 | 1.08 (0.47, 2.46) | 0.48 |
| Alcohol use disorder | ||||
| No | 1.02 (0.73, 1.43) | (ref.) | 1.16 (0.76, 1.78) | (ref.) |
| Yes | 0.97 (0.38, 2.43) | 0.52 | 0.97 (0.52, 1.82) | 0.59 |
Adjusted for baseline age, sex, race/ethnicity, study site, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, diabetes, hypertension, BMI, cigarette smoking, HCV, COPD, CD4 count, viral suppression status, HDL cholesterol, total cholesterol, eGFR, and FIB-4 score
The extreme low hazard ratio was due to no event occur in the stratum of recent opioid use with unhealthy alcohol use.
P-value for interaction term between lower-risk alcohol use and modifier; a low p-value indicates the association between low-risk alcohol (versus no alcohol use) and COVID-19 disease is dependent on the modifier
P-value for interaction term between unhealthy alcohol use and modifier; a low p-value indicates the association between unhealthy alcohol (versus no alcohol use) and COVID-19 disease is dependent on the modifier
DISCUSSION
In this sample of people with HIV during the first 10 months of the COVID-19 pandemic, alcohol use was associated with slightly higher hazards of COVID-19 cases and COVID-19 hospitalization but associations were not statistically significant. Our results are consistent with alcohol use having no effect or a small-to-moderate effect on COVID-19 outcomes; we can probably rule out a large effect of alcohol use on COVID-19 outcomes. Substance use, mental illnesses, or alcohol use disorder did not modify the relationship between pre-pandemic alcohol use and COVID-19 cases.
Prior evidence showed the effects of alcohol use and alcohol use disorder, substance use disorder, and mental illness on COVID-19 infection and mortality. In one prior case-control study from India, alcohol use increased the risk that a COVID-19 infection would be symptomatic [33]. Among people with COVID-19, those with an AUD had 1.5 times the odds of hospitalization and 1.6 times the odds of all-cause mortality[2]. Frequent alcohol consumption among college students was associated with a higher probability diagnosis with COVID-19[34,35]. A retrospective nation-wide cohort study of electronic health records showed that patients with recent diagnoses of SUD have more than eight times the odds of COVID-19 infection and higher rates of hospitalization and death comparing to patients without SUD [3]. Patients with recent diagnoses of depression or schizophrenia had around ten times the odds of developing COVID-19 compared to people without these pre-existing mental illnesses [4].
There are two potential mechanisms by which alcohol use might impact the risk of SARS-CoV-2 infection and COVID-19 hospitalization. First, heavy alcohol consumption can impair immune function by decreasing T lymphocyte and natural killer cell counts, leading to increased susceptibility to and severity of viral infections, including SARS-CoV-2 infections. For example, among PWH, alcohol consumption was associated with greater pneumonia severity[36]. However, since some complications of COVID-19 are a side effect of a robust immune response, it is also plausible that alcohol consumption could be protective for some people[37]. Second, alcohol use could increase the risk of exposure to SARS-CoV-2 through promoting socialization and risky behavior and decreasing adherence to public health guidelines, as compared to individuals who do not drink. Studies on college student alcohol use showed that frequent alcohol consumption, as well as social and conformity motives for drinking, have been associated with a higher probability of exposure to and diagnosis with COVID-19[34,35]. It is also plausible that people might consume alcohol exclusively in their own homes and may not venture out or have as many social contacts because of the deleterious effects of alcohol on social networks[38].
Our results may not be transportable to younger PWH. The median age of our study sample was 53, and our participants have had several decades to establish their drinking behaviors. Thus, the hypothesized effects of alcohol use on COVID through increasing socialization (as seen among college students) might not be as relevant for our study sample. Alcohol use trajectories among PWH in this cohort were quite stable over time when examined pre-pandemic[39]. Alcohol habits are also quite stable among older people in the general population[40]. Older adults were also significantly more likely to adhere to COVID-19 precautionary behaviours such as mask wearing, social distancing, and avoiding crowds than their younger counterparts[41]. Furthermore, participants who are particularly susceptible to adverse health effects of alcohol use (e.g., liver fibrosis, hypertension, and theoretically severe COVID-19) might have already stopped drinking (i.e., the “sick quitter” phenomenon) and be included in this analysis in the no alcohol group. Our results may also not apply to all PWH if the subset of the CNICS cohort who self-reported alcohol use on a PRO in the pre-pandemic period differed from the rest of the cohort on key covariates. However, having an at-risk alcohol use diagnosis only reduced the probability of completing a PRO by 3.2% in an earlier era when completion rates were lower, thus the impact of any selection bias on our results is likely minor if present[16].
We estimated the association between alcohol use measured prior to the pandemic and COVID-19 outcomes up to 22 months later. The most biologically relevant window for assessing the effects of alcohol use on viral infections is unclear[42]; the effects of alcohol may be more immediate or transient. However, if the most important pathways through which alcohol use affects COVID-19 outcomes are short-term, our study is limited in that pre-pandemic alcohol use may not reflect alcohol use during the pandemic. According to a national survey, alcohol consumption from May to June 2020 was 14% higher than during the same period in 2019[43]. Alcoholic beverage sales in March, April, and June 2020 were 11%, 8%, and 6% higher than average sales in the same months in 2018 and 2019[44]. Future research will examine how alcohol use changed for this cohort throughout different stages of the pandemic, and the effects of different measures of alcohol use (e.g., lifetime alcohol consumption or time-updated alcohol consumption) on COVID-19 outcomes.
This analysis covers the early pandemic period, and especially in the first few months, COVID-19 case identification was limited by access to testing. Self-testing and home tests were not widely accessible during our study period. Our COVID-19 case definition was highly specific but likely had low sensitivity. We likely missed almost all asymptomatic SARS-CoV-2 infections and some symptomatic COVID-19 cases. It was plausible that under-detection of COVID-19 cases was differential with respect to alcohol use, and this could have introduced bias into our results. Patients may have also under-reported their alcohol use. Self-report of unhealthy alcohol use in this cohort is estimated to be about 75-80% and specificity is around 99%[45]. If alcohol use was misclassified, it was likely non-differential and independent and thus bias might be towards the null (but because alcohol use is not binary bias might also be away from the null). In this study, we also had limited follow-up (the first 10 months of the pandemic), and we did not conduct stratified COVID-19 hospitalization estimates by AUD, substance use and mental illnesses due to lack of precision. However, it is also encouraging that in a cohort of over five thousand people with HIV, COVID-19 hospitalizations were relatively rare. Longer follow-up time is needed to consider the impact of waves of COVID-19 variants and vaccination status on the association between alcohol use and COVID-19 infections and outcomes.
Our study has several strengths. First, we had validated COVID-19 cases and COVID-19 hospitalization data[46]. Second, we adjusted for a comprehensive set of comorbidities. Finally, we presented the unadjusted and adjusted cumulative incidence curves (Figure 1 & 2), which allowed for use to consider the possibility of time period effects and to make it easier to interpret the hazard ratios.
In conclusion, this study provided insights into the potential association between alcohol use and COVID-19 cases and outcomes among people living with HIV. Our findings suggest that individuals who engage in lower-risk or unhealthy alcohol use may be at slightly increased risk of COVID-19 disease and COVID-19 hospitalization but if there is any effect it is likely to be small.
Appendix:
Appendix Table 4.
Crude and adjusted odds ratios of hospitalization among the subset of people with HIV who had a verified COVID-19 infection comparing lower-risk vs. no alcohol use and unhealthy vs. no alcohol use
| Alcohol Use | Odds Ratio (95% CI) | |
|---|---|---|
| Crude | No Use | 1 (ref.) |
| Lower-Risk | 0.59 (0.24, 1.43) | |
| Unhealthy | 0.88 (0.36, 2.17) | |
| Adjusteda | No Use | 1 (ref.) |
| Lower-Risk | 0.73 (0.28, 1.89) | |
| Unhealthy | 1.07 (0.40, 2.81) |
Adjusted for baseline age, sex, race/ethnicity, study sites, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, diabetes, hypertension, BMI, cigarette smoking, HCV, COPD, CD4 count, viral suppression status, HDL cholesterol, total cholesterol, eGFR, and FIB-4 score.
Appendix Table 5:
Adjusted hazard ratio with and without FIB-4 score and hypertension of COVID-19 cases and COVID-19 hospitalizations associated with pre-pandemic self-reported alcohol use among participants with HIV in CNICS cohort, March 1, 2020 – December 31, 2020.
| Adjusted Hazard Ratio with FIB-4 score and hypertensiona |
Adjusted Hazard Ratio without FIB-4 score and hypertensionb |
|
|---|---|---|
| SARS-Cov-2 infections | ||
| None alcohol use | 1.00 | 1.00 |
| Lower-risk alcohol use | 1.09 (0.78, 1.51) | 1.10 (0.79, 1.52) |
| Unhealthy alcohol use | 1.19 (0.81, 1.73) | 1.18 (0.81, 1.72) |
| COVID-19 hospitalization | ||
| None alcohol use | 1.00 | 1.00 |
| Lower-risk alcohol use | 0.82 (0.33, 1.99) | 0.81 (0.33, 1.98) |
| Unhealthy alcohol use | 1.25 (0.50, 3.09) | 1.27 (0.51, 3.14) |
Adjusted for baseline age, sex, race/ethnicity, study sites, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, diabetes, hypertension, BMI, cigarette smoking, HCV, COPD, CD4 count, viral suppression status, HDL cholesterol, total cholesterol, eGFR, and FIB-4 score.
Adjusted for baseline age, sex, race/ethnicity, study sites, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, diabetes, BMI, cigarette smoking, HCV, COPD, CD4 count, viral suppression status, HDL cholesterol, total cholesterol, and eGFR.
Appendix Table 6:
Stratum-specific adjusteda (without FIB-4 and hypertension) hazard ratios and 95% confidence intervals for the association between alcohol use and COVID-19 disease stratified by cocaine use, opioids use, depression, and alcohol use disorder.
| Variable | Strata | Lower-risk vs. No Alcohol Use aHR (95%CI) |
P-Value of Interaction Term for Lower-risk vs. No Alcohol Use |
Unhealthy vs. No Alcohol Use aHR (95%CI) |
P-Value of Interaction Term for Unhealthy vs. No Alcohol Use |
|---|---|---|---|---|---|
| Cocaine Use | Never | 1.12 (0.77, 1.61) | (ref.) | 1.06 (0.67, 1.67) | (ref.) |
| Former | 0.43 (0.20, 0.93) | 0.40 | 0.72 (0.39, 1.35) | 0.70 | |
| Recent | 0.89 (0.35, 2.28) | 0.24 | 1.06 (0.46, 2.43) | 0.17 | |
| Opioid Use | Never | 1.13 (0.78, 1.62) | (ref.) | 1.21 (0.79, 1.85) | (ref.) |
| Former | 0.90 (0.38, 2.15) | 0.48 | 1.27 (0.60, 2.68) | 0.83 | |
| Recent | 1.18 (0.37, 3.77) | 0.15 | - | - | |
| Depression | No to Mild | 1.01 (0.72, 1.41) | (ref.) | 1.11 (0.75, 1.66) | (ref.) |
| Moderate to Severe | 1.15 (0.55, 2.38) | 0.29 | 1.07 (0.47, 2.48) | 0.48 | |
| Alcohol Use Disorder | No | 1.03 (0.73, 1.45) | (ref.) | 1.15 (0.76, 1.75) | (ref.) |
| Yes | 1.00 (0.40, 2.50) | 0.48 | 0.97 (0.52, 1.81) | 0.58 |
Adjusted for baseline age, sex, race/ethnicity, study sites, HIV transmission risk factors, insurance status, depression, panic disorder, cocaine and opioid use, diabetes, BMI, cigarette smoking, HCV, COPD, CD4 count, viral suppression status, HDL cholesterol, total cholesterol, and eGFR. CI = confidence interval.
- The extreme low hazard ratio was due to no event occur in the stratum of recent opioid use with unhealthy alcohol use.
Footnotes
Declarations: The authors declare that they have no competing interests to report.
REFERENCES
- 1.WHO Coronavirus (COVID-19) Dashboard ∣ WHO Coronavirus (COVID-19) Dashboard With Vaccination Data. https://covid19.who.int/ (accessed 16 Jan2023). [Google Scholar]
- 2.Bailey KL, Sayles H, Campbell J, Khalid N, Anglim M, Ponce J, et al. COVID-19 patients with documented alcohol use disorder or alcohol-related complications are more likely to be hospitalized and have higher all-cause mortality. Alcoholism: Clinical and Experimental Research 2022; 46:1023–1035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wang QQ, Kaelber DC, Xu R, Volkow ND. COVID-19 risk and outcomes in patients with substance use disorders: analyses from electronic health records in the United States. Mol Psychiatry 2021; 26:30–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wang Q, Xu R, Volkow ND. Increased risk of COVID-19 infection and mortality in people with mental disorders: analysis from electronic health records in the United States. World Psychiatry 2021; 20:124–130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sigel K, Swartz T, Golden E, Paranjpe I, Somani S, Richter F, et al. Coronavirus 2019 and People Living With Human Immunodeficiency Virus: Outcomes for Hospitalized Patients in New York City. Clinical Infectious Diseases 2020; 71:2933–2938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Crane HM, McCaul ME, Chander G, Hutton H, Nance RM, Delaney JAC, et al. Prevalence and Factors Associated with Hazardous Alcohol Use Among Persons Living with HIV Across the US in the Current Era of Antiretroviral Treatment. AIDS Behav 2017; 21:1914–1925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Shiau S, Arpadi SM, Yin MT, Martins SS. Patterns of drug use and HIV infection among adults in a nationally representative sample. Addictive Behaviors 2017; 68:39–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bing EG, Burnam MA, Longshore D, Fleishman JA, Sherbourne CD, London AS, et al. Psychiatric Disorders and Drug Use Among Human Immunodeficiency Virus–Infected Adults in the United States. Archives of General Psychiatry 2001; 58:721–728. [DOI] [PubMed] [Google Scholar]
- 9.Uthman OA, Magidson JF, Safren SA, Nachega JB. Depression and Adherence to Antiretroviral Therapy in Low-, Middle- and High-Income Countries: A Systematic Review and Meta-Analysis. Current HIV/AIDS Reports 2014; 11:291–307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Brandt C, Zvolensky MJ, Woods SP, Gonzalez A, Safren SA, O’Cleirigh CM. Anxiety symptoms and disorders among adults living with HIV and AIDS: A critical review and integrative synthesis of the empirical literature. Clin Psychol Rev 2017; 51:164–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Duko B, Ayalew M, Ayano G. The prevalence of alcohol use disorders among people living with HIV/AIDS: a systematic review and meta-analysis. Subst Abuse Treat Prev Policy 2019; 14:52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Schouten J, Wit FW, Stolte IG, Kootstra NA, van der Valk M, Geerlings SE, 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. Clinical Infectious Diseases 2014; 59:1787–1797. [DOI] [PubMed] [Google Scholar]
- 13.Calina D, Hartung T, Mardare I, Mitroi M, Poulas K, Tsatsakis A, et al. COVID-19 pandemic and alcohol consumption: Impacts and interconnections. Toxicol Rep 2021; 8:529–535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kitahata MM, Rodriguez B, Haubrich R, Boswell S, Mathews WC, Lederman MM, et al. Cohort profile: the Centers for AIDS Research Network of Integrated Clinical Systems. International Journal of Epidemiology 2008; 37:948–955. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Crane MH, Lober W, Webster E, Harrington DR, Crane KP, Davis ET, et al. Routine Collection of Patient-Reported Outcomes in an HIV Clinic Setting:The First 100 Patients. Current HIV Research 2007; 5:109–118. [DOI] [PubMed] [Google Scholar]
- 16.Rudolph JE, Cole SR, Edwards JK, Moore R, O’Cleirigh C, Mathews WC, et al. At-Risk Alcohol Use Among HIV-Positive Patients and the Completion of Patient-Reported Outcomes. AIDS Behav 2018; 22:1313–1322. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Higgins-Biddle JC, Babor TF. A review of the Alcohol Use Disorders Identification Test (AUDIT), AUDIT-C, and USAUDIT for screening in the United States: Past issues and future directions. Am J Drug Alcohol Abuse 2018; 44:578–586. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001; 16:606–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kroenke K, Strine TW, Spitzer RL, Williams JB, Berry JT, Mokdad AH. The PHQ-8 as a measure of current depression in the general population. J Affect Disord 2009; 114:163–73. [DOI] [PubMed] [Google Scholar]
- 20.Humeniuk R, Ali R, Babor TF, Farrell M, Formigoni ML, Jittiwutikarn J, et al. Validation of the Alcohol, Smoking And Substance Involvement Screening Test (ASSIST). Addiction 2008; 103:1039–47. [DOI] [PubMed] [Google Scholar]
- 21.Spitzer RL, Kroenke K, Williams JBW, and the Patient Health Questionnaire Primary Care Study Group. Validation and Utility of a Self-report Version of PRIME-MDThe PHQ Primary Care Study. JAMA 1999; 282:1737–1744. [DOI] [PubMed] [Google Scholar]
- 22.Sterling RK, Lissen E, Clumeck N, Sola R, Correa MC, Montaner J, et al. Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology 2006; 43:1317–1325. [DOI] [PubMed] [Google Scholar]
- 23.Levey AS, Stevens LA, Schmid CH, Zhang Y (Lucy), Castro AF, Feldman HI, et al. A New Equation to Estimate Glomerular Filtration Rate. Ann Intern Med 2009; 150:604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bender Ignacio RA, Shapiro AE, Nance RM, Whitney BM, Delaney J, Bamford L, et al. Racial and ethnic disparities in COVID-19 disease incidence independent of comorbidities, among people with HIV in the US. medRxiv : the preprint server for health sciences. 2021; :2021.12.07.21267296. [Google Scholar]
- 25.Howe CJ, Cole SR, Westreich DJ, Greenland S, Napravnik S, Eron JJJ. Splines for trend analysis and continuous confounder control. Epidemiology 2011; 22:874–875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Cole SR, Hernán MA. Constructing Inverse Probability Weights for Marginal Structural Models. American Journal of Epidemiology 2008; 168:656–664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Stuart EA, Azur M, Frangakis C, Leaf P. Multiple Imputation With Large Data Sets: A Case Study of the Children’s Mental Health Initiative. American Journal of Epidemiology 2009; 169:1133–1139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Singh GK, Kim IE, Girmay M, Perry C, Daus GP, Vedamuthu IP, et al. Opioid Epidemic in the United States: Empirical Trends, and A Literature Review of Social Determinants and Epidemiological, Pain Management, and Treatment Patterns. Int J MCH AIDS 2019; 8:89–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Witkiewitz K, Vowles KE. Alcohol and Opioid Use, Co-Use, and Chronic Pain in the Context of the Opioid Epidemic: A Critical Review. Alcoholism: Clinical and Experimental Research 2018; 42:478–488. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Fojo AT, Lesko CR, Calkins KL, Moore RD, McCaul ME, Hutton HE, et al. Do Symptoms of Depression Interact with Substance Use to Affect HIV Continuum of Care Outcomes? AIDS and Behavior 2019; 23:580–591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lesko CR, Hutton HE, Edwards JK, McCaul ME, Fojo AT, Keruly JC, et al. Alcohol Use Disorder and Recent Alcohol Use and HIV Viral Non-Suppression Among People Engaged in HIV Care in an Urban Clinic, 2014-2018. AIDS Behav 2022; 26:1299–1307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Westreich D, Edwards JK, Tennant PWG, Murray EJ, van Smeden M. Choice of Outcome in COVID-19 Studies and Implications for Policy: Mortality and Fatality. American Journal of Epidemiology 2022; 191:282–286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Saurabh S, Verma MK, Gautam V, Kumar N, Jain V, Goel AD, et al. Tobacco, alcohol use and other risk factors for developing symptomatic COVID-19 vs asymptomatic SARS-CoV-2 infection: a case–control study from western Rajasthan, India. Transactions of the Royal Society of Tropical Medicine and Hygiene 2021; 115:820–831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Allen HK, Cohen-Winans S, Armstrong K, Clark NC, Ford MA. COVID-19 exposure and diagnosis among college student drinkers: links to alcohol use behavior, motives, and context. Translational Behavioral Medicine 2021; 11:1348–1353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Zamboanga BL, Blumenthal H, Ramarushton B, Napper LE, Hurlocker MC, Ford K, et al. What pandemic? A multisite study of drinking motives and drinking games participation among college students during a pandemic (COVID-19) academic year. Addictive behaviors 2023; 138:107560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Jolley SE, Alkhafaf Q, Hough C, Welsh DA. Presence of an alcohol use disorder is associated with greater pneumonia severity in hospitalized HIV-infected patients. Lung 2016; 194:755–762. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Romeo J, Wärnberg J, Nova E, Díaz LE, Gómez-Martinez S, Marcos A. Moderate alcohol consumption and the immune system: A review. British Journal of Nutrition 2007; 98:S111–S115. [DOI] [PubMed] [Google Scholar]
- 38.Choi NG, DiNitto DM. Heavy/binge drinking and depressive symptoms in older adults: gender differences. International Journal of Geriatric Psychiatry 2011; 26:860–868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Bilal U, McCaul ME, Crane HM, Mathews WC, Mayer KH, Geng E, et al. Predictors of longitudinal trajectories of alcohol consumption in people with HIV. Alcoholism: Clinical and Experimental Research 2018; 42:561–570. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Sydén L, Wennberg P, Forsell Y, Romelsjö A. Stability and change in alcohol habits of different socio-demographic subgroups-a cohort study. BMC Public Health 2014; 14:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Jehn A, Stackhouse M, Zajacova A. COVID-19 Health Precautions: Identifying Demographic and Socio-Economic Disparities and Changes over Time. Canadian Public Policy 2021; 47:252–264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Barr T, Helms C, Grant K, Messaoudi I. Opposing effects of alcohol on the immune system. Progress in Neuro-Psychopharmacology and Biological Psychiatry 2016; 65:242–251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Pollard MS, Tucker JS, Green HD Jr. Changes in Adult Alcohol Use and Consequences During the COVID-19 Pandemic in the US. JAMA Network Open 2020; 3:e2022942–e2022942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Hu Y, Quigley BM, Taylor D. Human mobility data and machine learning reveal geographic differences in alcohol sales and alcohol outlet visits across US states during COVID-19. Plos one 2021; 16:e0255757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Lesko CR, Keil AP, Moore RD, Chander G, Fojo AT, Lau B. Measurement of Current Substance Use in a Cohort of HIV-Infected Persons in Continuity HIV Care, 2007-2015. Am J Epidemiol 2018; 187:1970–1979. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Shapiro AE, Bender Ignacio RA, Whitney BM, Delaney JA, Nance RM, Bamford L, et al. Factors associated with severity of COVID-19 disease in a multicenter cohort of people with HIV in the United States, March-December 2020. JAIDS Journal of Acquired Immune Deficiency Syndromes Published Online First: 9900.https://journals.lww.com/jaids/Fulltext/9900/Factors_associated_with_severity_of_COVID_19.19.aspx [DOI] [PMC free article] [PubMed] [Google Scholar]