Abstract
Background
Human immunodeficiency virus infection (HIV) is a presumed risk factor for severe coronavirus disease 2019 (COVID-19), yet little is known about COVID-19 outcomes in people with HIV (PWH).
Methods
We used the TriNetX database to compare COVID-19 outcomes of PWH and HIV-negative controls aged ≥18 years who sought care in 44 healthcare centers in the United States from January 1 to December 1, 2020. Outcomes of interest were rates of hospitalization (composite of inpatient non-intensive care [ICU] and ICU admissions), mechanical ventilation, severe disease (ICU admission or death), and 30-day mortality.
Results
Of 297 194 confirmed COVID-19 cases, 1638 (0.6%) were HIV-infected, with >83% on antiretroviral therapy (ART) and 48% virally suppressed. Overall, PWH were more commonly younger, male, African American or Hispanic, had more comorbidities, were more symptomatic, and had elevated procalcitonin and interleukin 6. Mortality at 30 days was comparable between the 2 groups (2.9% vs 2.3%, P = .123); however, PWH had higher rates hospitalization (16.5% vs 7.6%, P < .001), ICU admissions (4.2% vs 2.3%, P < .001), and mechanical ventilation (2.4% vs 1.6%, P < .005). Among PWH, hospitalization was independently associated with male gender, being African American, integrase inhibitor use, and low CD4 count; whereas severe disease was predicted by older age (adjusted odds ratio [aOR], 8.33; 95% confidence interval [CI], 1.06–50.00; P = .044) and CD4 <200 cells/mm3 (aOR, 8.33; 95% CI, 1.06–50.00; P = .044).
Conclusions
People with HIV had higher rates of poor COVID-19 outcomes but were not more at risk of death than their non-HIV-infected counterparts. Older age and low CD4 count predicted adverse outcomes.
Keywords: clinical outcomes, COVID-19, HIV
Poor COVID-19 outcomes were frequent among PWH; however, risk of death was comparable with non-HIV-infected controls. Older age, male gender, being Black or African American, and low CD4 count were independently associated with adverse COVID-19 outcomes, regardless of HIV viremia.
As of February 1, 2021, the coronavirus disease 2019 (COVID-19) pandemic caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has resulted in an estimated 100 million confirmed infections and more than 2 million deaths globally [1]. Black and other racial minority heritage, male gender, age >65 years, obesity (body mass index [BMI] ≥30 kg/m2), diabetes mellitus, hypertension, severe cardiopulmonary disease, and other chronic conditions have been reported as contributing to poor COVID-19 outcomes, including increased risk of mortality [2–5].
Compared with the general population, people with human immunodeficiency virus (PWH) are presumed to be at greater risk of severe COVID-19 and adverse clinical outcomes [6, 7]. This has been attributed to the observation that PWH tend to have a higher burden of lifestyle-associated risk factors and underlying comorbidities, in addition to having an already heightened systemic inflammatory state at baseline that could potentially enhance or amplify the viral cytokine release syndrome—also referred to as the “cytokine storm”—that has been described in the setting of COVID-19 [6–11].
Notwithstanding, emerging evidence from case report series, early observational studies, and systematic reviews describing the clinical features of COVID-19 among PWH have yielded mixed findings thus far. Several studies from Europe and North America have observed no substantial differences in morbidity and mortality rates among human immunodeficiency virus (HIV) and non-HIV cohorts of patients hospitalized with COVID-19, with the majority of HIV-infected patients in these studies reported as stable on antiretroviral therapy (ART) and virally suppressed [12–15]. On the other hand, recent data from 2 large population-based studies from South Africa and the United Kingdom found a 2- and 2.6-fold increase in the risk of COVID-19 death, respectively, among PWH compared with their non-HIV-infected counterparts [16, 17]. In the United Kingdom study by Bhaskaran et al [17], black ethnic minority status was associated with a 4.3-fold higher hazard of COVID-19 death among PWH compared with non-HIV-infected controls. However, crucial questions around the associations (if any) between the degree of immunosuppression, level of virologic control, or class of ART and COVID-19 outcomes in PWH have remained largely unanswered and are the focus of ongoing inquiry into the nature of the interactions between HIV and SARS-CoV-2.
In a multicenter study in the United States using TriNetX (a large global federated health research network), Hadi et al [15] had previously observed a higher crude COVID-19 mortality in 404 PWH, compared with their non-HIV-infected counterparts; however, the study used an earlier time cutoff (July 2020) and, importantly, did not examine the relationship between outcomes of interest and CD4 count, virologic control, and class of ART. In this study, we used a higher time cutoff in TriNetX (to include a larger number of patients) to characterize the clinical features and predictors of adverse COVID-19 among PWH in the United States.
MATERIALS AND METHODS
Study Population and Design
We used the TriNetX database to conduct a retrospective study of adult patients aged ≥18 years with SARS-CoV-2 infection (confirmed by polymerase chain reaction [PCR] or serology) who sought care across 44 healthcare organizations in the United States from January 1, 2020 to December 1, 2020. To safeguard protected health information, TriNetX does not include data on participating healthcare organizations (HCOs). Therefore, we are unable to provide geographic or institutional information. However, a participating HCO typically represents a large health center with main and satellite hospitals, specialty care services, and outpatient clinics.
Data Collection and Definitions
We collected clinical data including patient demographics, comorbidities, lifestyle-associated risk factors (smoking, alcohol use, and illicit drug use), vital signs and symptoms at presentation, laboratory findings, medication use, and healthcare services used—ie, outpatient clinic, emergency room, inpatient non-intensive care (ICU) setting, or ICU admission. We additionally collected the last CD4 count and viral load data (measured by HIV-1 ribonucleic acid [RNA] PCR) within the preceding 12 months. A full description of study definitions and variables used to query the TriNetX database and their corresponding International Classification of Diseases, Tenth Revision (ICD-10) codes are provided in the Supplementary Materials.
Our clinical outcomes of interest were the odds of outpatient visit (defined as utilizing ambulatory clinic or emergency department services only), hospitalization (defined as the composite outcome of inpatient non-ICU and ICU service utilization), mechanical ventilation use, severe COVID-19 (ICU admission or death), and mortality rate at 30 days after COVID-19 diagnosis.
We further analyzed COVID-19 severity in the PWH cohort based on hospital services used. Mild disease was defined as having required outpatient services only (ie, ambulatory clinic or emergency department visit). Moderate disease was defined as having required inpatient non-ICU services, whereas severe disease was defined as having required ICU services or death within 30 days of COVID-19 diagnosis.
Statistical Analyses
Summary statistics were generated for all variables at baseline. Continuous data were presented as means and standard deviations and were compared using independent t-tests. Categorical data were presented as frequency and proportions and compared using χ 2 or Fisher’s exact test, as appropriate. To address possible confounders, we balanced cohorts using 1:1 greedy nearest-neighbor propensity score matching by demographics and key comorbidities associated with poor COVID-19 outcomes as outlined in Table 1. Within the PWH cohort, we explored predictors of outcomes (moderate and severe COVID-19) using multinomial logistic regression models with the main effects of age, race, comorbidities, class of ART, CD4 count, and viral load. For outcomes of interest, we calculated odds ratios (ORs) and 95% confidence intervals (CIs), with P < .05 considered statistically significant. All analyses were conducted in the statistical software R version 3.63 (R Core Team, 2020).
Table 1.
Comparison of Patient Demographics, Comorbidities, Social History, and Treatment Between PWH and Non-HIV-Infected COVID-19 Patients Before and After Propensity Score Matching
| Variables | Before Matching | After Matching | ||||
|---|---|---|---|---|---|---|
| HIV (N = 1638) |
Non-HIV (N = 295 556) |
P Value | HIV (N = 1635) | Non-HIV (N = 1609) | P Value | |
| Age, years (mean ± SD) | 43.34 ± 13.59 | 46.48 ± 18.7 | <.001 | 48.34 ± 13.59 | 49.12 ± 14.89 | .116 |
| Gender | ||||||
| Female | 501 (30.6%) | 163 318 (55.3%) | <.001 | 500 (30.6%) | 1135 (69.4%) | 1.000 |
| Male | 1137 (69.4%) | 130 866 (44.3%) | <.001 | 1135 (69.4%) | 1116 (69.4%) | 1.000 |
| Race or Ethnicity | ||||||
| Black or African American | 805 (49.1%) | 55 264 (18.7%) | <.001 | 804 (49.2%) | 791 (49.2%) | .420 |
| White | 573 (35.0%) | 169 681 (57.4%) | <.001 | 572 (35.0%) | 581 (36.1%) | 1.000 |
| Hispanic or Latino | 297 (18.1%) | 44 869 (15.2%) | .001028 | 297 (18.2%) | 304 (18.9%) | .851 |
| Asian | 25 (1.5%) | 7643 (2.6%) | .0088 | 25 (1.5%) | 28 (1.7%) | .303 |
| American Indian or Alaska Native | 6 (0.4%) | 1359 (0.5%) | .7077 | 6 (0.4%) | 5 (0.3%) | .851 |
| Native Hawaiian or other Pacific Islander | 2 (0.1%) | 801 (0.3%) | .3397 | 2 (0.1%) | 1 (0.1%) | .106 |
| Comorbidities | ||||||
| Cardiovascular diseases | 977 (59.6%) | 98 078 (33.2%) | <.001 | 974 (59.6%) | 995 (61.8%) | .198 |
| Diabetes mellitus | 358 (21.9%) | 37 921 (12.8%) | <.001 | 357 (21.8%) | 391 (24.3%) | .104 |
| Obesity (body mass index >30 kg/m2) | 404 (24.7%) | 43 883 (14.8%) | <.001 | 404 (24.7%) | 440 (27.3%) | .095 |
| Liver disease | 179 (10.9%) | 11 714 (4.0%) | <.001 | 179 (10.9%) | 169 (10.5%) | .725 |
| Chronic lower respiratory diseases | 392 (23.9%) | 39 622 (13.4%) | <.001 | 391 (23.9%) | 387 (24.1%) | .960 |
| Neoplasms | 450 (27.5%) | 39 622 (13.4%) | <.001 | 450 (27.5%) | 467 (29.0%) | .323 |
| Chronic kidney disease | 264 (16.1%) | 15 680 (5.3%) | <.001 | 262 (16.0%) | 239 (14.9%) | .382 |
| Asthma | 237 (14.5%) | 24 299 (8.2%) | <.001 | 237 (14.5%) | 228 (14.2%) | .831 |
| Dementia | 22 (1.3%) | 2299 (0.8%) | .014 | 22 (1.3% | 14 (0.9%) | .261 |
| Lifestyle-Associated Risk Factors | ||||||
| Nicotine-related disorders | 301 (18.4%) | 16 292 (5.5%) | <.001 | 300 (18.3%) | 278 (17.3%) | .453 |
| Alcohol-related disorders | 138 (8.4%) | 6174 (2.1%) | <.001 | 138 (8.4%) | 113 (7.0%) | .149 |
| Cocaine-related disorders | 81 (4.9%) | 1272 (0.4%) | <.001 | 81 (5.0%) | 58 (3.6%) | .070 |
| Opioid-related disorders | 77 (4.7%) | 2696 (0.9%) | <.001 | 77 (4.7%) | 58 (3.6%) | .137 |
| ART Regimens | ||||||
| NRTIs | 896 (54.7%) | 895 (54.7%) | ||||
| Tenofovir-based | 653 (39.9%) | 653 (39.9%) | ||||
| NNRTIs | 148 (9.0%) | 148 (9.0%) | ||||
| INSTIs | 781 (47.7%) | 781 (47.7%) | ||||
| Healthcare Service Used | ||||||
| Emergency Department Services | 400 (24.4%) | 46 866 (15.9%) | <.001 | 399 (24.4%) | 372 (23.1%) | .414 |
| Outpatient Office Services | 323 (19.7%) | 50 299 (17.0%) | .004 | 323 (19.8%) | 316 (19.6%) | .782 |
| Inpatient Services | 157 (9.6%) | 12 174 (4.1%) | <.001 | 156 (9.5%) | 116 (7.2%) | .015 |
| Critical Care Services | 57 (3.5%) | 4297 (1.5%) | <.001 | 57 (3.5%) | 50 (3.1%) | .499 |
Bold indicates statistically significant values.
Abbreviations: ART, antiretroviral therapy; COVID-19, coronavirus disease 2019; HIV, human immunodeficiency virus; INSTI, integrase strand transfer inhibitor; NNRTI, nucleos(t)ide reverse-transcriptase inhibitor; NRTI, nucleos(t)ide reverse-transcriptase inhibitor; PWH, people with HIV; SD, standard deviation.
Patient Consent Statement
The study was approved by the Institution Board Review committee at Case Western Reserve University/University Hospitals Cleveland Medical Center. Written informed consent from patients was not required because data from the TriNetX system safeguards patient’s privacy by reporting deidentified data.
RESULTS
Baseline Characteristics of People With Human Immunodeficiency Virus (HIV) and Non-HIV-Infected Patients
Of 297 194 confirmed COVID-19 cases, 1638 (0.6%) were HIV-infected, with >83% on ART. Among PWH, 229 (14.0%) contributed CD4 count data, with 187 (81.6%) reporting CD4 ≥200 cells/mm3. Seven hundred fifty-five (46.1%) contributed viral load data, with 617 (81.7%) being virally suppressed (HIV-1 RNA <20 copies/mL) (Table 1).
At entry into the healthcare system, approximately one fifth of PWH versus non-HIV-infected patients presented in the ambulatory clinic setting (19.5% vs 17.0%; P = .004), whereas one quarter presented to the emergency department (24.4% vs 15.9%, P < .001). Approximately 9.6% and 3.5% of PWH required inpatient services directly upon presentation in the non-ICU and ICU settings, respectively, versus 4.1% and 1.5% of non-HIV-infected patients (P < .001).
Compared with their non-HIV-infected counterparts, PWH patients were more commonly younger (43.34 ± 13.59 vs 46.48 ± 18.7 years, P < .001), male (69.4% vs 44.3%, P < .001), black or African American (49.1% vs 19.7%, P < .001), Hispanic (18.1% vs 15.2%, P < .001), and more likely to have underlying cardiovascular disease (46.8% vs 26.1%, P < .001), obesity (24.7% vs 14.8%, P < .001), diabetes mellitus (21.9% vs 12.8%, P < .001), and other comorbidities. People with HIV were also significantly more likely to have a history of smoking, alcohol, and drug use (all P < .001).
Clinical Presentation and Laboratory Parameters of People With Human Immunodeficiency Virus (HIV) and Non-HIV-Infected Patients
Overall, PWH were more symptomatic at presentation compared with non-HIV-infected patients (Table 2). The most common symptoms were cough (16.9% vs 15.9%, P = .294), difficulty breathing (13.7% vs 8.6%, P < .001), and fever (11.4% vs 8.6%, P < .001), with a smaller proportion of patients exhibiting gastrointestinal and neurological symptoms.
Table 2.
Comparison of Vital Signs and Symptoms at Presentation, Laboratory Findings, and Medical Services Used Between PWH and Non-HIV-Infected COVID-19 Patients Before and After Propensity Score Matching
| Variables | Before Matching | After Matching | ||||
|---|---|---|---|---|---|---|
| HIV (N = 1638) |
Non-HIV (N = 295 556) |
P Value | HIV (N = 1635) |
Non-HIV (N = 1609) |
P Value | |
| Presenting Vital Signs | ||||||
| Systolic blood pressure, mmHg (mean ± SD) | 124.7 ± 19.3 n = 556 |
127.3 ± 19.8 n = 80 404 |
.002 | 124.7 ± 19.3 n = 555 |
128.5 ± 19 n = 547 |
.001 |
| Diastolic blood pressure, mmHg (mean ± SD) | 75.7 ± 12.8 n = 587 |
75.1 ± 12.8 n = 86 147 |
.222 | 75.8 ± 12.8 n = 586 |
77.1 ± 13.5 n = 577 |
.075 |
| Heart rate, beats/min (mean ± SD) | 85.5 ± 16.1 n = 192 |
82.2 ± 15.6 n = 23 460 |
.004 | 85.5 ± 16.1 n = 192 |
82.4 ± 15.3 n = 181 | .058 |
| Respiratory rate, breaths/min (mean ± SD) | 17.3 ± 3.1 n = 237 |
17.0 ± 9.9 n = 23 778 |
.662 | 17.3 ± 3.1 n = 237 |
16.7 ± 3.2 n = 198 |
.038 |
| Oxygen saturation, % (mean ± SD) |
80.7 ± 24.2 n = 120 |
83.0 ± 22.1 n = 12 784 |
.274 | 80.6 ± 24.2 n = 119 |
79.4 ± 25.7 n = 115 |
.726 |
| Symptoms at Presentation | ||||||
| Cough | 277 (16.9%) | 47 080 (15.9%) | .294 | 277 (16.9%) | 273 (17.0%) | 1.000 |
| Difficulty breathing | 224 (13.7%) | 25 438 (8.6%) | <.001 | 224 (13.7%) | 211 (13.1%) | .533 |
| Fever (≥100.4°F) | 186 (11.4%) | 25 553 (8.6%) | <.001 | 186 (11.4%) | 166 (10.3%) | .286 |
| Pain in throat and chest | 119 (7.3%) | 10 957 (3.7%) | <.001 | 118 (7.2%) | 90 (5.6%) | .069 |
| Abnormalities of heart beat | 93 (5.7%) | 7876 (2.7%) | <.001 | 93 (5.7%) | 82 (5.1%) | .504 |
| Nausea and vomiting | 66 (4.0%) | 7753 (2.6%) | <.001 | 66 (4.0%) | 68 (4.2%) | .855 |
| Abdominal and pelvic pain | 68 (4.2%) | 5452 (1.8%) | <.001 | 68 (4.2%) | 52 (3.2%) | .192 |
| Malaise and fatigue | 66 (4.0%) | 13 887 (4.7%) | .223 | 65 (4.0%) | 94 (5.8%) | .017 |
| Headache | 41 (2.5%) | 6669 (2.3%) | .557 | 41 (2.5%) | 36 (2.2%) | .696 |
| Disturbances of smell and taste | 43 (2.6%) | 7545 (2.6%) | .915 | 43 (2.6%) | 52 (3.2%) | .417 |
| Laboratory Parameters | ||||||
| Leukocytes, ×109/L (mean ± SD) | 6.9 ± 4.1 n = 554 |
7.7 ± 26.3 n = 59 895 |
.465 | 6.9 ± 4.1 n = 553 |
8.3 ± 20.6 n = 472 |
.118 |
| Neutrophils, ×109/L (mean ± SD) |
283.5 ± 1977.5 n = 427 |
206.2 ± 1205 n = 37 953 |
.197 | 283.2 ± 1979.8 n = 426 |
279.7 ± 1282.8 n = 305 |
.002 |
| Lymphocytes, ×109/L (mean ± SD) | 24.4 ± 14.2 n = 530 |
20.3 ± 11.7 n = 54 094 |
<.001 | 24.4 ± 14.2 n = 529 |
20.6 ± 12.4 n = 422 |
<.001 |
| Hemoglobin, g/dL (mean ± SD) | 12.6 ± 2.6 n = 582 |
12.9 ± 2.2 n = 62 141 |
<.001 | 12.6 ± 2.6 n = 581 |
12.8 ± 2.5 n = 489 |
.086 |
| Platelets, ×109/L (mean ± SD) | 216.8 ± 91.9 n = 573 |
229.5 ± 93.1 n = 59 360 |
.001 | 216.5 ± 91.7 n = 572 |
231.5 ± 130.9 n = 477 |
.030 |
| Creatinine, mg/dL (mean ± SD) | 1.9 ± 2.9 n = 594 |
1.3 ± 1.7 n = 60 346 |
<.001 | 1.91 ± 2.91 n = 593 |
1.68 ± 2.2 n = 482 |
.155 |
| Glomerular filtration rate, mL/min (mean ± SD) | 71.9 ± 35.5 n = 593 |
78.0 ± 34.8 n = 60 686 |
<.001 | 71.8 ± 35.5 n = 592 |
76.1 ± 38.2 n = 486 |
.058 |
| Prothrombin time, seconds (mean ± SD) | 14.9 ± 5.7 n = 177 |
14.5 ± 6.2 n = 19 708 |
.311 | 14.9 ± 5.7 n = 177 |
15.2 ± 6.2 n = 150 |
.655 |
| International normalized ratio (mean ± SD) | 1.3 ± 0.5 n = 176 |
1.3 ± 0.7 n = 19 598 |
.751 | 1.3 ± 0.5 n = 176 |
1.3 ± 0.6 n = 151 |
.710 |
| Activated partial thromboplastin time, seconds (mean ± SD) | 35.1 ± 11.3 n = 126 |
32.8 ± 11.9 n = 14 010 |
.038 | 35.1 ± 11.3 n = 126 |
33.2 ± 7.0 n = 107 |
.146 |
| C-reactive protein, mg/dL (mean ± SD) | 65.7 ± 73.8 n = 207 |
80.3 ± 86.0 n = 19 572 |
.015 | 65.7 ± 73.8 n = 207 |
85.8 ± 84.8 n = 161 |
.253 |
| D-dimer, ng/mL (mean ± SD) | 276.6 ± 729.4 n = 93 |
308.6 ± 1142.3 n = 9982 |
.788 | 276.6 ± 729.4 n = 93 |
370.5 ± 824.4 n = 78 |
.432 |
| Erythrocytes 106/µL (mean ± SD) | 4.2 ± 0.9 (n = 576) |
4.4 ± 0.8 (n = 59 058) |
<.001 | 4.2 ± 0.9 (n = 575) | 4.5 ± 0.9 (n = 473) | <.001 |
| Erythrocyte sedimentation rate, mm/hour (mean ± SD) | 56.6 ± 35.5 n = 78 |
48.3 ± 32.4 n = 6446 |
.025 | 56.6 ± 35.5 n = 78 |
49.9 ± 30.1 n = 42 |
.300 |
| Lactate, mg/dL (mean ± SD) | 1.7 ± 1.5 n = 198 |
1.6 ± 1.2 n = 16 909 |
.257 | 1.7 ± 1.5 n = 197 |
1.5 ± 0.8 n = 150 |
.100 |
| Lactate dehydrogenase, U/L (mean ± SD) | 393.4 ± 367.8 n = 169 |
410.4 ± 362.9 n = 14 822 |
.546 | 393.5 ± 367.8 n = 169 |
434.3 ± 386.5 n = 142 |
.341 |
| Ferritin, ng/dL (mean ± SD) | 866.7 ± 1000 n = 176 |
884 ± 1890 n = 16 263 |
.901 | 866.7 ± 1000 n = 176 |
1214 ± 1404.2 n = 140 |
.011 |
| Troponin I (cardiac), mg/dL (mean ± SD) | 0.1 ± 0.2 n = 111 |
0.27 ± 3.1 n = 12 029 |
.517 | 0.1 ± 0.2 n = 111 |
0.1 ± 0.1 n = 105 |
.091 |
| Procalcitonin, ng/mL (mean ± SD) | 2.5 ± 10.1 n = 95 |
1.3 ± 5.9 n = 8529 |
.042 | 2.5 ± 10.1 n = 95 |
3.8 ± 14.0 n = 67 |
.492 |
| Interleukin-6, pg/mL (mean ± SD) | 258.5 ± 642.1 n = 17 |
98.2 ± 248.7 n = 2160 |
.010 | 258.5 ± 642.1 n = 17 |
197.4 ± 475.3 n = 16 |
.759 |
| CD4 count <200 cells/mm3 | 187 (11.4%) | 187 (11.4%) | ||||
| CD4 count ≥200 cells/mm3 | 42 (2.6%) | 42 (2.6%) | ||||
| HIV-1 RNA >20 copies/mL | 617 (37.7%) | 616 (37.7%) | ||||
| HIV-1 RNA <20 copies/mL | 138 (8.4%) | 138 (8.4%) |
Bold indicates statistically significant values.
Abbreviations: COVID-19, coronavirus disease 2019; HIV, human immunodeficiency virus; PWH, people with HIV; RNA, ribonucleic acid; SD, standard deviation.
On laboratory parameters, PWH were more commonly anemic (P < .001), thrombocytopenic (P < .001), and had elevated serum creatinine (P < .001), procalcitonin (P = .042), and interleukin (IL)-6 (P = .010) levels. Although C-reactive protein (P < .0148) and erythrocyte sedimentation rate (P < .001) were lower in PWH, serum levels of other markers of acute systemic inflammation, myocardial injury, and coagulopathy were comparable between the 2 cohorts (Table 2).
Disease Severity and Clinical Outcomes at 30 Days in People With Human Immunodeficiency Virus (HIV) and Non-HIV-Infected Patients
Table 3 displays the differences in 30-day outcomes between PWH and non-HIV patients. People with HIV had higher rates of hospitalization (16.5% vs 7.6%, P < .001) and higher rates of severe illness requiring ICU admission (4.2% vs 2.3%, P < .001) and mechanical ventilation (2.4% vs 1.6%, P < .005). Mortality at 30 days was higher among PWH but did not attain statistical significance (2.9% vs 2.3%, P = .123). In propensity score-matched analysis based on demographics and key comorbidities, PWH remained at significantly higher odds of hospitalization (OR, 1.26; 95% CI, 1.04–1.53; P = .023).
Table 3.
Differences in 30-Day Outcomes Between PWH and Non-HIV-Infected COVID-19 Patients
| Outcomes | Before Matching | After Matching | ||||||
|---|---|---|---|---|---|---|---|---|
| HIV | Non-HIV | OR (95% CI) |
P Value | HIV | Non-HIV | OR (95% CI) | P Value | |
| Hospitalization | 270 (16.5%) | 22 398 (7.6%) | 2.41 (2.11–2.74) | <.001 | 269 (16.5%) | 218 (13.5%) | 1.26 (1.04–1.53) | .023 |
| Intensive care services | 68 (4.2%) | 6731 (2.3%) | 1.86 (1.45–2.36) | <.001 | 68 (4.2%) | 71 (4.4%) | 0.94 (0.67–1.32) | .790 |
| Ventilation management | 40 (2.4%) | 4624 (1.6%) | 1.58 (1.13–2.14) | .005 | 40 (2.4%) | 46 (2.9%) | 0.85 (0.55–1.31) | .536 |
| Mortality at 30 days | 47 (2.9%) | 6708 (2.3%) | 1.28 (0.94–1.69) | .1233 | 46 (2.8%) | 61 (3.8%) | 0.74 (0.50–1.08) | .145 |
Bold indicates statistically significant values.
Abbreviations: COVID-19, coronavirus disease 2019; CI, confidence interval; HIV, human immunodeficiency virus; OR, odds ratio; PWH, people with HIV.
aResults shown both before and after propensity score matching.
Predictors of Poor Coronavirus Disease 2019 Outcomes Among People With Human Immunodeficiency Virus
Overall, 511 (31.2%) of PWH had visit data available, 237 (46.4%) of which were classified as having mild COVID-19, 196 (38.4%) had moderate COVID-19, and 78 (15.3%) had severe COVID-19 (Table 4).
Table 4.
Multinomial Logistic Regression Model Results for Predictors of Moderate and Severe COVID-19 Among PWH
| Predictor | Moderate (Hospitalized Non-ICU) N = 196 |
Severe (ICU Admission or Death) N = 78 |
||||||
|---|---|---|---|---|---|---|---|---|
| Estimates | SE | aOR (95% CI) | P Value | Estimates | SE | aOR (95% CI) | P Value | |
| Age | 0.018 | 0.03 | 1.02 (0.96–1.09) | .571 | 0.121 | 0.05 | 1.13 (1.02–1.25) | .017 |
| Male | 1.69 | 0.85 | 5.40 (1.01–28.76) | .048 | 1.66 | 1.15 | 5.25 (0.55–49.9) | .149 |
| Black or African American | 1.66 | 0.76 | 5.28 (1.20–23.18) | .028 | 1.58 | 0.95 | 4.85 (0.75–31.29) | .097 |
| Tenofovir use | 0.863 | 0.86 | 1.55 (0.29–8.43) | .609 | −0.541 | 1.03 | 0.58 (0.08–4.36) | .598 |
| Protease inhibitor use | 1.039 | 1.05 | 2.83 (0.36–22.00) | .321 | 3.25 | 1.17 | 3.42 (0.34–34.04) | .295 |
| Integrase inhibitor use | 3.03 | 1.33 | 20.77 (1.54–280.96) | .026 | 1.59 | 1.34 | 4.93 (0.36–67.57) | .223 |
| CD4 <200 cells/mm3 | −2.56 | 0.88 | 12.50 (2.33–100.00) | .003 | −2.14 | 1.06 | 8.33 (1.06–50.00) | .044 |
| Viral load >20 copies/mL | −0.039 | 0.96 | 0.962 (0.15–6.26) | .968 | 0.217 | 1.22 | 1.24 (0.11–13.69) | .859 |
Bold indicates statistically significant values.
Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; COVID-19, coronavirus disease 2019; ICU, intensive care unit; PWH, people with human immunodeficiency virus; SE, standard error.
In multinomial logistic regression models, moderate COVID-19 in PWH requiring hospitalization was independently associated with being male (adjusted OR [aOR] = 5.40; 95% CI, 1.10–28.18; P = .048), black or African American (aOR = 5.28; 95% CI, 1.20–23.18; P = .028), integrase strand transfer inhibitor (INSTI) use (aOR = 20.77; 95% CI, 1.54–280.96; P = .026), and CD4 < 200 cells/mm3 (aOR = 12.50; 95% CI, 2.33–100.00; P = .003). Severe COVID-19 resulting in ICU admission or death within 30 days of COVID-19 diagnosis was independently associated with older age (aOR = 8.33; 95% CI, 1.06–50.00; P = .044) and CD4 <200 cells/mm3 (aOR = 8.33; 95% CI, 1.06–50.00; P = .044). We found no significant effects of tenofovir use, protease inhibitor (PI) use, or viral load on COVID-19 outcomes among PWH (Table 4).
DISCUSSION
In this analysis of approximately 300 000 confirmed COVID-19 cases (0.6% PWH) from the United States, we found that almost half (47.6%) of PWH presented with mild disease (ie, required outpatient care only), 38.3% presented with moderate disease (ie, required inpatient non-ICU services), and 14.1% had severe disease (ie, required ICU admission or died within 30 days of COVID-19 diagnosis). We found no significant difference in COVID-19 mortality rates between PWH and non-HIV-infected patients. The crude death rates were low in both groups (2.9% vs 2.3%, respectively). Our findings are in broad agreement with population-based studies from other high-resource settings with good healthcare infrastructure, where COVID-19 mortality rates have generally been reported at <5% [6–9, 12–20].
Furthermore, we observed that compared with non-HIV-infected patients, PWH were more symptomatic at presentation, more likely to have severe disease, and used more healthcare resources with higher rates of hospitalizations, mechanical ventilation use, and ICU admissions. The high risk for hospitalization of PWH persisted after propensity score matching by key variables. We hypothesize that the reasons for these findings may be related to the combined effects of the high prevalence of underlying comorbidities coupled with the possibility of a more vigorous manifestation of the cytokine release syndrome in PWH [6–11]. It is notable that PWH in this study had a 2- to 2.5-fold higher serum levels of IL-6 and procalcitonin, respectively, compared with non-HIV-infected patients. The role of IL-6 and other cytokines in COVID-19 pathogenesis has been well described. During the early phases of SARS-CoV-2 infection, activated T lymphocytes and macrophages release IL-1, IL-6, tumor necrosis factor-alpha, monocyte chemoattractant protein-1, and other proinflammatory mediators of the so-called cytokine storm that is characterized by tissue damage, endothelial leakage, and the activation of the complement and coagulation pathways [21, 22]. The centrality of IL-1β and IL-6 in particular in immune dysregulation has warranted the exploration of anticytokine antagonists as novel therapies for severe COVID-19 [22–25].
Our study analyzed risk factors associated with poor COVID-19 outcomes in PWH. Male gender and being black or African American were associated with >5-fold increase in the risk of hospitalization, whereas older age was associated with higher odds of ICU admission or death within 30 days of COVID-19 diagnosis. Importantly, even after adjusting for age and comorbidities, Black ethnic minority status remained independently associated with poor COVID-19 outcomes. These findings are consistent with multiple previous reports from the United States and elsewhere that have highlighted the gender, racial, and ethnic disparities associated with the HIV and COVID-19 pandemics [26, 27]. African American, Hispanic, and other minority populations in the United States continue to be disproportionately impacted by a synergy of persistent social and economic barriers that have limited their access to health services and placed them at heightened risk of poor disease outcomes [28, 29]. Public health efforts aiming to address inequities in HIV and COVID-19 outcomes would benefit from implementing policies and actions prioritizing minority and other vulnerable populations.
Similar to other studies, we did not detect any significant effects of HIV viremia on COVID-19 severity [12–15]; however, advanced HIV disease (CD4 count <200 cells/mm3) and INSTI use were associated with an 8-fold and 20-fold higher risk of poor outcomes, respectively. The association between low CD4 count and poor outcomes was anticipated, but the large effect demonstrated by INSTI use was unexpected. Although it has been widely speculated that some antiretrovirals may have a protective effect against SARS-CoV-2, their role in COVID-19 prevention and treatment remains in dispute. Several HIV-1 PIs (eg, lopinavir, ritonavir, and saquinavir) had demonstrated potent activity against 2 earlier coronaviruses (SARS-CoV and Middle East respiratory syndrome coronavirus) and inhibited their replication in in vitro and animal studies [30–32]; however, a randomized controlled trial of 199 critically ill hospitalized patients failed to show benefit with lopinavir-ritonavir beyond standard of care [33]. Tenofovir and other nucleos(t)ide reverse transcriptase inhibitors (NRTIs) have been shown to inhibit SARS-CoV-2 in in vitro studies by attenuating the effects of IL-8 and IL-10, while promoting the production of interferon-gamma [34–36]. Two recent studies from Spain and South Africa found that compared with other therapies, tenofovir-based ART had a protective effect against COVID-19 death among PWH [16, 37]. In our study, however, NRTI- or PI-based ART did not show mortality benefit among PWH.
Integrase strand transfer inhibitors are preferred first-line ART used in combination with 2 NRTIs and are particularly attractive due to their efficacy, good tolerability profile, and high genetic barrier to resistance [38–40]. However, INSTIs have been associated with significant weight gain in PWH initiating ART [41–43]. There are currently no studies (in vitro or in vivo) describing the direct effects of INSTIs on SARS-CoV-2. We postulate that the adverse effects of INSTIs on COVID-19 in this study likely result indirectly from the metabolic complications of weight gain and increased BMI—a well recognized risk factor for poor COVID-19 outcomes [2, 5].
Our study had several limitations. These included incomplete data on CD4 count, viral load, and other laboratory findings. In addition, we are unable to capture the start date of the symptoms, and as such we cannot exclude the possibility that PWH presented later in the course of COVID-19, which would explain the more frequent symptoms, hospitalizations, and ICU use among PWH with COVID-19. Nonetheless, to the best of our knowledge, this is one of the largest multicenter studies to date from the United States that was sufficiently powered to detect meaningful differences and statistically significant associations between patient characteristics and outcomes of interest.
CONCLUSIONS
In one of the largest multicenter studies to date from the United States, PWH with COVID-19 had a higher burden of comorbidities and life-style associated risk factors and higher rates of hospitalization, mechanical ventilation, and ICU admissions. Despite this, PWH did not appear to be significantly more at risk of death at 30 days after COVID-19 diagnosis compared with their non-HIV-infected counterparts. Male gender, being black or African American, INSTI use, and low CD4 count were independently associated with adverse COVID-19 outcomes, regardless of HIV virologic control. Similar to previous studies, our findings highlight the need for public health efforts and policy to prioritize PWH, minority, and other vulnerable populations in addressing longstanding social and economic inequities that are being further exacerbated by the HIV and COVID-19 pandemics.
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Acknowledgments
Author contributions. G. A. Y., J. A. P., and G. A. M. contributed to study concept and design. All authors contributed to acquisition of data. G. A. Y., J. A. P., and G. A. M. contributed to analysis and interpretation of data. G. A. Y., J. A. P., and G. A. M. drafted the manuscript. All authors contributed to critical revision of the manuscript for important intellectual content. J. A. P. contributed to statistical analysis. G. A. M. obtained funding. K. S. and H. T. contributed to administrative, technical, or material support. G. A. M. supervised the study.
Disclaimer. The contents are solely the responsibility of the authors and do not necessarily represent the official views of University Hospitals Cleveland Medical Center (UHCMC) or the National Institutes of Health (NIH).
Financial support. This publication was made possible through funding support from the UHCMC and the Clinical and Translational Science Collaborative of Cleveland (UL1TR002548) from the National Center for Advancing Translational Sciences (NCATS) component of the NIH and NIH Roadmap for Medical Research.
Potential conflicts of interest. G. A. M. has received grant support from ViiV, Tetraphase, Roche, Vanda, Astellas, and Genentech. She served as a scientific advisor for Gilead, Merck, ViiV/GSK, Theratechnologies, and Jannsen. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Presented in part: Conference on Retroviruses and Opportunistic Infections (CROI), March 2021.
References
- 1.World Health Organization. Coronavirus Disease (COVID-19) Dashboard. Available at: https://covid19.who.int/. Accessed 16 May 2021.
- 2.Zhou F, Yu T, Du R, et al. Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study [published correction appears in Lancet. 2020 Mar 28;395(10229):1038]. Lancet 2020; 395:1054–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wu Z, McGoogan JM. Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: summary of a report of 72314 cases from the Chinese Center for Disease Control And Prevention. JAMA 2020; 323:1239–42. [DOI] [PubMed] [Google Scholar]
- 4.Centers for Disease Control and Prevention. CDC COVID-19 Response Team. Preliminary estimates of the prevalence of selected underlying health conditions among patients with coronavirus disease 2019 - United States, February 12-March 28, 2020. MMWR Morb Mortal Wkly Rep 2020; 69:382–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Williamson EJ, Walker AJ, Bhaskaran K, et al. Factors associated with COVID-19-related death using OpenSAFELY. Nature 2020; 584:430–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Department of Health and Human Services. Interim Guidance for COVID-19 and Persons with HIV. Available at: https://clinicalinfo.hiv.gov/en/guidelines/covid-19-and-persons-hiv-interim-guidance/interim-guidance-covid-19-and-persons-hiv?view=full. Accessed 16 May 2021.
- 7.Shiau S, Krause KD, Valera P, et al. The burden of COVID-19 in people living with HIV: a syndemic perspective. AIDS Behav 2020; 24:2244–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hariyanto TI, Putri C, Frinka P, et al. Human immunodeficiency virus (HIV) and outcomes from coronavirus disease 2019 (COVID-19) pneumonia: a meta-analysis and meta-regression. AIDS Res Hum Retroviruses 2021; doi: 10.1089/AID.2020.0307. [DOI] [PubMed] [Google Scholar]
- 9.Deeks SG, Tracy R, Douek DC. Systemic effects of inflammation on health during chronic HIV infection. Immunity 2013; 39:633–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Yongzhi X. COVID-19-associated cytokine storm syndrome and diagnostic principles: an old and new Issue. Emerg Microbes Infect 2021; 10:266–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pum A, Ennemoser M, Adage T, Kungl AJ. Cytokines and chemokines in SARS-CoV-2 infections-therapeutic strategies targeting cytokine storm. Biomolecules 2021; 11:91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Cooper TJ, Woodward BL, Alom S, Harky A. Coronavirus disease 2019 (COVID-19) outcomes in HIV/AIDS patients: a systematic review. HIV Med 2020; 21:567–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dandachi D, Geiger G, Montgomery MW, et al. Characteristics, comorbidities, and outcomes in a multicenter registry of patients with HIV and coronavirus disease-19. Clin Infect Dis 2020; doi: 10.1093/cid/ciaa1339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sigel K, Swartz T, Golden E, et al. Coronavirus 2019 and people living with human immunodeficiency virus: outcomes for hospitalized patients in New York City. Clin Infect Dis 2020; 71:2933–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hadi YB, Naqvi SFZ, Kupec JT, Sarwari AR. Characteristics and outcomes of COVID-19 in patients with HIV: a multicentre research network study. AIDS 2020; 34:F3–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Boulle A, Davies MA, Hussey H, et al. Risk factors for COVID-19 death in a population cohort study from the Western Cape Province, South Africa [published online ahead of print August 29, 2020]. Clin Infect Dis 2020. doi: 10.1093/cid/ciaa1198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bhaskaran K, Rentsch CT, MacKenna B, et al. HIV infection and COVID-19 death: a population-based cohort analysis of UK primary care data and linked national death registrations within the OpenSAFELY platform. Lancet HIV 2021; 8:e24–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Geretti AM, Stockdale AJ, Kelly SH, et al. Outcomes of COVID-19 related hospitalization among people with HIV in the ISARIC WHO Clinical Characterization Protocol (UK): a prospective observational study [published online ahead of prin October 23, 2020]. Clin Infect Dis 2020. doi: 10.1093/cid/ciaa1605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pastor-Barriuso R, Pérez-Gómez B, Hernán MA, et al. ; ENE-COVID Study Group . Infection fatality risk for SARS-CoV-2 in community dwelling population of Spain: nationwide seroepidemiological study. BMJ 2020; 371:m4509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.COVID-19 National Emergency Response Center, Epidemiology and Case Management Team, Korea Centers for Disease Control and Prevention. Coronavirus disease-19: the first 7,755 cases in the Republic of Korea [published correction appears in Osong Public Health Res Perspect. 2020 Jun;11(3):146]. Osong Public Health Res Perspect 2020; 11:85–90.32257774 [Google Scholar]
- 21.Wiersinga WJ, Rhodes A, Cheng AC, et al. Pathophysiology, transmission, diagnosis, and treatment of coronavirus disease 2019 (COVID-19): a review. JAMA 2020; 324:782–93. [DOI] [PubMed] [Google Scholar]
- 22.Fajgenbaum DC, June CH. Cytokine storm. N Engl J Med 2020; 383:2255–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Huang C, Soleimani J, Herasevich S, et al. Clinical characteristics, treatment, and outcomes of critically ill patients with COVID-19: a scoping review. Mayo Clin Proc 2021; 96:183–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ignatius EH, Wang K, Karaba A, et al. Tocilizumab for the treatment of COVID-19 among hospitalized patients: a matched retrospective cohort analysis. Open Forum Infect Dis 2021; 8:ofaa598. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Stone JH, Frigault MJ, Serling-Boyd NJ, et al. ; BACC Bay Tocilizumab Trial Investigators . Efficacy of tocilizumab in patients hospitalized with Covid-19. N Engl J Med 2020; 383:2333–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Nosyk B, Krebs E, Zang X, et al. “Ending the epidemic” will not happen without addressing racial/ethnic disparities in the United States human immunodeficiency virus epidemic. Clin Infect Dis 2020; 71:2968–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sullivan PS, Satcher Johnson A, Pembleton ES, et al. Epidemiology of HIV in the USA: epidemic burden, inequities, contexts, and responses. Lancet 2021; 397:1095–106. [DOI] [PubMed] [Google Scholar]
- 28.Fields EL, Copeland R, Hopkins E. Same script, different viruses: HIV and COVID-19 in US black communities. Lancet 2021; 397:1040–2. [DOI] [PubMed] [Google Scholar]
- 29.Singer MC, Erickson PI, Badiane L, et al. Syndemics, sex and the city: understanding sexually transmitted diseases in social and cultural context. Soc Sci Med 2006; 63:2010–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Savarino A. Expanding the frontiers of existing antiviral drugs: possible effects of HIV-1 protease inhibitors against SARS and avian influenza. J Clin Virol 2005; 34:170–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Sheahan TP, Sims AC, Leist SR, et al. Comparative therapeutic efficacy of remdesivir and combination lopinavir, ritonavir, and interferon beta against MERS-CoV. Nat Commun 2020; 11:222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Martinez MA. Compounds with therapeutic potential against novel respiratory 2019 coronavirus. Antimicrob Agents Chemother 2020; 64: e00399-20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Cao B, Wang Y, Wen D, et al. A trial of lopinavir-ritonavir in adults hospitalized with severe Covid-19. N Engl J Med 2020; 382:1787–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zídek Z, Franková D, Holý A. Activation by 9-(R)-[2-(phosphonomethoxy)propyl]adenine of chemokine (RANTES, macrophage inflammatory protein 1alpha) and cytokine (tumor necrosis factor alpha, interleukin-10 [IL-10], IL-1beta) production. Antimicrob Agents Chemother 2001; 45:3381–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Chien M, Anderson TK, Jockusch S, et al. Nucleotide analogues as inhibitors of SARS-CoV-2 polymerase, a key drug target for COVID-19. J Proteome Res 2020; 19:4690–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Melchjorsen J, Risør MW, Søgaard OS, et al. Tenofovir selectively regulates production of inflammatory cytokines and shifts the IL-12/IL-10 balance in human primary cells. J Acquir Immune Defic Syndr 2011; 57:265–75. [DOI] [PubMed] [Google Scholar]
- 37.Del Amo J, Polo R, Moreno S, et al. ; The Spanish HIV/COVID-19 Collaboration . Incidence and severity of COVID-19 in HIV-positive persons receiving antiretroviral therapy: a cohort study. Ann Intern Med 2020; 173:536–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Scarsi KK, Havens JP, Podany AT, et al. HIV-1 integrase inhibitors: a comparative review of efficacy and safety. Drugs 2020; 80:1649–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Brenner BG, Wainberg MA. Clinical benefit of dolutegravir in HIV-1 management related to the high genetic barrier to drug resistance. Virus Res 2017; 239:1–9. [DOI] [PubMed] [Google Scholar]
- 40.Shimura K, Kodama E, Sakagami Y, et al. Broad antiretroviral activity and resistance profile of the novel human immunodeficiency virus integrase inhibitor elvitegravir (JTK-303/GS-9137). J Virol 2008; 82:764–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Sax PE, Erlandson KM, Lake JE, et al. Weight gain following initiation of antiretroviral therapy: risk factors in randomized comparative clinical trials. Clin Infect Dis 2020; 71:1379–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Venter WDF, Moorhouse M, Sokhela S, et al. Dolutegravir plus two different prodrugs of tenofovir to treat HIV. N Engl J Med 2019; 381:803–15. [DOI] [PubMed] [Google Scholar]
- 43.Bourgi K, Rebeiro PF, Turner M, et al. Greater weight gain in treatment-naive persons starting dolutegravir-based antiretroviral therapy. Clin Infect Dis 2020; 70:1267–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.