key: cord-0823107-6ngq9l1x
authors: Laracy, Justin; Zucker, Jason; Castor, Delivette; McMahon, Donald J; Guo, Tai Wei; Yan, Michael; Shalev, Noga; Scherer, Matthew; Gordon, Peter; Sobieszczyk, Magdalena; Yin, Michael T
title: HIV-1 Infection Does Not Change Disease Course or Inflammatory Pattern of SARS-CoV-2-Infected Patients Presenting at a Large Urban Medical Center in New York City
date: 2021-01-28
journal: Open Forum Infect Dis
DOI: 10.1093/ofid/ofab029
sha: c388b1bde0446e8ff63a13ad963413c0bb0a064a
doc_id: 823107
cord_uid: 6ngq9l1x

BACKGROUND: The clinical impact of coronavirus disease 2019 (COVID-19) among people with HIV (PWH) remains unclear. In this retrospective cohort study of COVID-19, we compared clinical outcomes and laboratory parameters among PWH and controls. METHODS: Sixty-eight PWH diagnosed with COVID-19 were matched 1:4 to patients without known HIV diagnosis, drawn from a study population of all patients who were diagnosed with COVID-19 at an academic urban hospital. The primary outcome was death/discharge to hospice within 30 days of hospital presentation. RESULTS: PWH were more likely to be admitted from the emergency department than patients without HIV (91% vs 71%; P = .001). We observed no statistically significant difference between admitted PWH and patients without HIV in terms of 30-day mortality rate (19% vs 13%, respectively) or mechanical ventilation rate (18% vs 20%, respectively). PWH had higher erythrocyte sedimentation rates than controls on admission but did not differ in other inflammatory marker levels or nasopharyngeal/oropharyngeal severe acute respiratory syndrome coronavirus 2 viral load estimated by reverse transcriptase polymerase chain reaction cycle thresholds. CONCLUSIONS: HIV infection status was associated with a higher admission rate; however, among hospitalized patients, PWH did not differ from HIV-uninfected controls by rate of mechanical ventilation or death/discharge to hospice.

The global severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic has led to >54 million cases and 1.3 million deaths worldwide [1] . The clinical manifestations of coronavirus disease 2019 (COVID-19) range from asymptomatic disease to acute respiratory disease syndrome with shock and multiorgan failure [2, 3] . There is ongoing uncertainty about whether HIV infection or degree of HIV-related immune deficiency contributes to severity of COVID-19 and COVID-19related outcomes.

A large cohort study from South Africa recently found that PWH had a 2-fold increased risk of COVID-19 mortality compared with HIV-uninfected people [4] . It has been shown that PWH with CD4 counts <200 cells/mm 3 may be at higher risk for severe outcomes compared with PWH with CD4 counts >500 cells/mm 3 [5] . Other studies have shown that COVID-19 outcomes do not significantly differ by HIV infection status [6] [7] [8] . Indeed, the World Health Organization recognizes that there is currently no evidence to suggest that PWH are at higher risk for contracting SARS-CoV-2 or severe COVID-19 [9] . Some have even postulated that HIV might exert a protective effect against SARS-CoV-2 by diminishing the inflammatory response due to dysfunctional cellular immunity or the antiviral effects of co-incidental use of specific HIV antiretrovirals [10, 11] . We sought to describe the COVID-19 clinical experience and laboratory characteristics of PWH and to evaluate the association between HIV infection status and death/discharge to hospice and other clinical outcomes.

Between March 10, 2020, and June 10, 2020, there were 6098 patients who tested positive for SARS-CoV-2 infection at Columbia University-New York Presbyterian, a quaternary care medical center in Northern Manhattan. From this sample, we conducted a retrospective cohort analysis to evaluate the association between HIV infection status and several clinical outcomes: admission, mechanical ventilation, and death/discharge to hospice within 30 days of hospital admission. We identified all adult patients (≥18 years old) with laboratory-confirmed SARS-CoV-2 infection by real-time reverse transcriptase polymerase chain reaction (RT-PCR) of nasopharyngeal or oropharyngeal swabs. We identified PWH cases (n = 68) and selected matched controls at a ratio of 1 to 4 (n = 272). HIV cases were defined based on HIV diagnostic codes or administration of antiretroviral therapy (ART) in the electronic medical record (EMR). HIV status was confirmed through manual chart review indicating laboratory evidence of HIV or provider documentation of HIV status. Controls were selected from the same COVID-19 population and had no documented HIV status in any of the same EMR data sources used to confirm HIV infection. Controls were matched by age group (<45 years of age, then 10-year bands, and 85 years and above), sex, and the number of days before presentation that the patient was symptomatic (±1 day).

The data were extracted from the EMR and augmented with manually abstracted data. Electronically extracted data included demographics, oxygen rank severity upon arrival (room air, nasal canula, non-rebreather, noninvasive ventilation, invasive ventilation), basic laboratory testing results (ie, complete blood count, basic metabolic panel, liver function tests, coagulation panel), other laboratory parameters per institutional guidelines for management of COVID-19 (ie, erythrocyte sedimentation rate [ESR], C-reactive protein [CRP], lactate dehydrogenase [LDH], ferritin, D-dimer, procalcitonin, and interleukin-6 [IL-6]), presentation, admission and discharge dates, diagnosis codes for comorbid conditions, and potential therapeutic agents for COVID-19. Of note, our institutional protocol for the care of patients with COVID-19 recommended immediate measurement of the above laboratory markers; thus, the values reflect measurements drawn from patients while they were in the emergency department. Additional data abstracted for PWH included a history of AIDS (CD4 ˂200 cells/mm 3 ), prescribed ART during admission, HIV viral load (VL), T-cell panel, and any opportunistic infections diagnosed during the admission. These data were entered into a REDCap database. All data were merged using R studio [12] .

We conducted descriptive analyses comparing PWH cases and matched controls. Continuous variables were described using means and standard deviations. Categorical variables were described with counts and percentages. In bivariate analyses, categorical variables were compared using the Wald chi-square test or Fisher exact test, and continuous variables were compared using the Mann-Whitney-Wilcoxon test or Kruskal-Wallis test as appropriate and visualized in box and whisker plots. Group differences at admission were evaluated with the categorical and continuous variable tests described. The likelihood of admission was calculated using multiple logistic regression adjusting for ethnicity, race, history of hypertension, and history of pulmonary disease. For the effect of HIV status on the risk of requiring mechanical ventilation, we conducted a conditional Cox proportional hazards model using the Fine-Gray extension comparing those who required mechanical ventilation with admitted patients who did not. For the risk of death/ discharge to hospice within 30 days of presentation, we conducted a conditional Cox proportional hazards model with and without discharge within 30 days as a competing risk. Patients were observed for mechanical ventilation or death/discharge to hospice as the primary outcomes, with discharge as a competing risk, or were censored at day 30 after admission if neither of the primary or competing events occurred. The HIV-specific hazard for death/discharge to hospice/mechanical ventilation at any time point, t, is the instantaneous risk of death/discharge to hospice/mechanical ventilation as the first event, conditional on being admitted just before t. Individuals who were discharged first were censored at this time when considering each primary outcome in the hazard analysis. All regression models were fit with all variables that were significantly different by HIV status based on logistic regression analysis and Wald test, or exclusion of the null for 95% confidence intervals. All statistical analyses and data visualization were performed in SAS STAT software, version 13.2 (Cary, NC, USA).

This study was exempt from patient consent, and the Columbia University-New York Presbyterian Institutional Review Board approved the study.

There were 68 PWH and 272 matched controls with positive test results for SARS-CoV-2 by RT-PCR of nasopharyngeal or oropharyngeal swab evaluated during the study period. The overall sample (n = 340) had a mean age (SD) of 58 (12.7) years and was disproportionately male (71%, n = 240) ( Table 1) . There was no difference in age or sex due to the matching criteria, but PWH were more likely to be non-Hispanic Black (41% vs 20%) and less likely to be Hispanic (21% vs 43%; P = .0004). The 3 most common comorbidities among PWH were hypertension (63%), diabetes (37%), and pulmonary disease (26%). PWH had lower body mass index (BMI; 26.9 vs 28.8; P = .02) but were more likely to have a history of hypertension (63% vs 43%; P = .003) and pulmonary disease (26% vs 15%; P = .03). Table 1 Female 100 (29) 20 (29) 80 (29) Racial ethnic identity .0004

Hispanic 130 (38) 14 (21) 116 (43) Non-Hispanic Black 82 (24) 28 (41) 54 (20) Non-Hispanic White 45 (13) 7 (10) 38 (14) Other/not specified 83 (24) 19 (28) 64 (24) Outcomes Admitted to hospital 256 (75) 62 (91) 194 (71) .001

Mechanical ventilation 66 (19) 12 (18) 54 (20) .68

Death or discharge to hospice 48 (14) 13 (19) 35 (13) .19

Diabetes 105 (31) 25 (37) 80 (29) .24

Hypertension 161 (47) 43 (63) 118 (43) .003

Kidney disease 33 (10) 8 (12) 25 (9) .52

Pulmonary 59 (17) 18 (26) 41 (15) .03

Initial temperature, mean ± SD, °F (23) 18 (28) 42 (21) Non-rebreather 28 (10) 6 (9) 22 (11) Noninvasive (25) 15 (22) 69 (25) >40.1 kg/m 2 11 (3) 1 (1) 10 (4) COVID-19 treatment Hydroxychloroquine 141 (57) 31 (52) 110 (59) .35

Remdesivir 6 (2)

2 (3) 4 (2) .59

Tociluzmab 11 (4) 3 (5) 8 (4) .8

Methylprednisolone 39 (16) 12 (20) 27 (14) .30

Laboratory parameters, mean ± SD (%) Compared with matched controls, PWH had lower initial alanine aminotransferase levels (31.1 U/L vs 49.1 U/L; P = .02), a lower initial neutrophil-to-lymphocyte ratio (NLR; 5.39 vs 6.99; P = .01), and higher ESR levels (77.9 mm/h vs 65.8 mm/h; P = .02) ( Table 1 ). There were no differences in oxygen rank severity upon arrival to the emergency department by HIV status (3/68 PWH and 74/272 controls had missing data). Among admitted patients, there were no differences in the use of potential COVID-19 therapeutics between PWH and HIV-negative controls.

PWH were more likely to be admitted to the hospital (91% vs 71%; P = .001) than patients without HIV ( 

There was no significant difference in the rate of mechanical ventilation among PWH and their matched controls (18% vs 20%; P = .68) ( Table 1) , with an unadjusted hazard ratio of 0.8 (95% CI, 0.5-1.6) ( Table 2 ). The competing risk model was not applied to mechanical ventilation, as this always preceded discharge. After adjustment for potential confounding factors, there was still no difference in the rate of mechanical ventilation by HIV status (aHR, 0.7; 95% CI, 0.4-1.5). Variables associated with a higher rate of mechanical ventilation were BMI (aHR, 1.2; 95% CI, 1.0-1.4) and NLR (aHR, 1.5; 95% CI, 1.2-1.8).

A Kaplan-Meier curve representing time to mechanical ventilation by HIV status group is presented in Figure 1A .

There was also no difference in the rate of death/discharge to hospice between PWH and matched controls (19% vs 13%; P = .19) (Table 1) , with an unadjusted hazard ratio of 1.4 (95% CI, 0.7-2.8) ( Table 2 ). Among admitted-only patients, the rate of death/discharge to hospice among PWH and matched controls was 21% and 18%, respectively. After statistical adjustment, PWH had an aHR of 0.9 for death/discharge to hospice (95% CI, 0.3-2.3). Variables associated with a higher rate of death/ discharge to hospice were BMI (aHR, 1.1; 95% CI, 1.0-1.1) and NLR (aHR, 1.1; 95% CI, 1.0-1.1). Although the number of discharges as a competing risk was high, this was not different by HIV status; therefore, the Kaplan-Meier curve representing time to death/discharge to hospice by HIV status group is presented in Figure 1B . Additionally, cumulative incidence curves shown in Supplementary Figure 1 support this observation. 

We found no differences in initial inflammatory markers when stratified by HIV infection and mortality status ( Figure 2 ). We also compared laboratory markers over time in PWH compared with negative controls and found no difference (data not shown). There was no significant difference by HIV status in nasopharyngeal/oropharyngeal SARS-CoV-2 RT-PCR cycle threshold values (E-gene for both Roche cobas SARS-CoC-2 or Cepheid Xpert Xpress) (Table 1) . Additionally, there was no difference in cycle threshold value when stratified by HIV infection and mortality status (Figure 3 ).

Among PWH, there was no association between CD4 cell count on admission, CD4 cell percentage on admission, and use of TDF/TAF with rates of mechanical ventilation. There was also no association between CD4 percentage on admission and use of TDF/TAF with death/discharge to hospice, but PWH with the outcome of death/discharge to hospice had a lower mean CD4 count on admission (206 cells/mm 3 vs 366 cells/mm 3 ; P = .04). Of the 36 PWH with VL measurements during admission for COVID-19, one-fourth (25%) of those with a VL >200 cells/mL required mechanical ventilation vs 9/32 (28%) patients with a VL <200 cells/mL (P = 1.00). Similarly, 2/4 (50%) and 6/32 (19%) PWH with a VL above and below 200 cells/mL, respectively, met the outcome of death/discharge to hospice (P = .38). There were few (4%, n = 3) opportunistic infections diagnosed in PWH. The 3 opportunistic infections in our study were coinfections at the time of COVID-19 diagnosis; they included 2 PWH with oral thrush and 1 PWH with disseminated Mycobacterium avium complex, and they all occurred in patients nonadherent to ART. During the study period, there were 144 PWH either discharged from our emergency department or admitted to our hospital who had negative test results for SARS-CoV-2 who were not included in this study. Their mean age (SD) was 51 (12.9) years, most identified as non-Hispanic and Black, and more than half (51%) had a history of AIDS. Most (86%) of the HIV-positive, SARS-CoV-2-negative patients were prescribed ART, and most (70%) were on an integrase strand transfer inhibitor (INSTI)-based regimen.

Nearly all PWH were prescribed ART during their admission for COVID-19, most (76%) received an INSTI-based regimen, and more than half received either TDF or TAF (Table 1) . Although 18 patients (27%) had a history of AIDS, of the patients with measurements upon admission for COVID-19, the mean CD4 count was 324 cells/mm 3 , 29/42 (69%) had a CD4 count >200 cells/mm 3 , and 35/42 (83%) had a CD4 percentage ≥14. Of the 36 patients with VL measurements during admission, most (89%) achieved viral suppression.

Our institution is a quaternary care medical center in New York City and was an epicenter of the SARS-CoV-2 epidemic during the study period. We found that PWH were more likely to be hospitalized with COVID-19, but once admitted, they had similar rates of mechanical ventilation and death/discharge to hospice compared with demographically matched controls. Higher BMI and higher NLR remained significant predictors of mechanical ventilation and death/discharge to hospice overall. PWH and HIV-negative controls did not differ in their exposure to potential COVID-19 therapeutics.

The higher admission rate among PWH was surprising, as in our sample PWH were similar in age, comorbidity profile, temperature, duration of symptoms, and laboratory measures as their HIV-uninfected comparators. Notably, there was also no difference in oxygen rank severity upon arrival to the emergency department by HIV status. Emergency room providers' perceptions surrounding HIV and historical biases toward admission of HIV-positive patients may have contributed to the difference we observed, as PWH are more likely to be admitted from emergency departments than HIV-negative patients [13] . Providers' anxiety that HIV status may confer poor outcomes from COVID-19 may have also contributed to the higher admission rates among PWH.

The lack of an association between HIV infection status and mechanical ventilation and death/discharge to hospice has been reported by others [14] . It is notable that studies from across Europe and the United States have shown that the risk of SARS-CoV-2 infection and severe COVID-19 in PWH is not due to immunosuppression, but rather mostly due to underlying socioeconomic conditions and comorbidities [7, 8, 15-21]. The 1 notable exception was a population-based study from South Africa that found that PWH had a 2-fold increased risk of COVID-19 mortality compared with HIV-uninfected people [4] . Residual confounding may explain this finding in this cohort, which tried to account for socioeconomic status by adjusting for location within the province but may not have been able to completely adjust for this complex variable. The comorbidities of hypertension, pulmonary disease, and obesity are widely recognized risk factors for severe disease from COVID-19. We found a consistent association between higher BMI and increased rates of mechanical ventilation or death/discharge to hospice in our adjusted model. This finding is consistent with other studies, which also found that elevated BMI is a risk factor for poor outcomes from COVID-19 [22, 23] . Similarly, elevated inflammatory markers and nasopharyngeal SARS-CoV-2 viral loads have been demonstrated to be predictors of poor outcomes from COVID-19 [24, 25] . D-dimer elevation upon admission has been shown to be predictive of critical illness and death in patients with COVID-19, and we found no difference in d-dimer levels between PWH and controls [26] . We explored whether initial levels of all tested inflammatory markers or nasopharyngeal SARS-CoV-2 viral loads differed by HIV status and mortality, but found no differences. However, in adjusted models, we found that NLR was associated with rates of mechanical ventilation and death/discharge to hospice. NLR has been shown to be predictive of severe COVID-19 requiring transfer to an intensive care unit [27] . The biological mechanism underlying this association remains unknown, but a possible explanation is that a sustained reduction in lymphopenia in critically ill patients is associated with nonresolution of inflammation [28] .

In vitro studies suggest that some forms of ART, especially protease inhibitors (PIs) and TDF/TAF, may have activity against SARS-CoV-2 with the following mechanisms: PIs via inhibition of proteases required for polyprotein cleavage into functional subunits and TDF/TAF via inhibition of the RNAdependent RNA polymerase [29] [30] [31] . Despite promising in vitro data, randomized clinical trials in humans with PIs for SARS-CoV-2 infection have been negative [32, 33] . In our study sample, we found that PWH who received TDF/TAF did not differ in risk of mechanical ventilation or death/discharge to hospice.

The strengths of our study include a moderately sized study population with a matched comparison group during the heart of the pandemic in New York City. Limitations include the study being conducted at a single center at a large metropolitan city in the United States, our reliance on manual extraction from the EMR for medical history, insufficient data to match by race/ethnicity or date of diagnosis, and our use of diagnosis codes to identify PWH.

In conclusion, we found that PWH were more likely to be admitted from the emergency department, but once hospitalized they had similar rates of mechanical ventilation and death/discharge to hospice as matched controls.

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. 

World Health Organization. Coronavirus disease 2019 (COVID-19) pandemic

TZp-LENqd4pLWr_6TA-lxPJjQr54oOVg5MrMhOG3FCNMaAuS8EALw_wcB

Accessed 10

Clinical characteristics of 24 asymptomatic infections with COVID-19 screened among close contacts in Nanjing

Outcomes in HIV/COVID-19 Coinfection • ofid • 9

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Risk factors for COVID-19 death in a population cohort study from the Western Cape Province, South Africa

Characteristics, comorbidities, and outcomes in a multicenter registry of patients with HIV and coronavirus disease-19

Covid-19 and people with HIV infection: outcomes for hospitalized patients in New York City

Description of COVID-19 in HIVinfected individuals: a single-centre, prospective cohort

Outcomes among HIVpositive patients hospitalized with COVID-19

Information note on HIV and COVID-19

Incidence and severity of COVID-19 in HIVpositive persons receiving antiretroviral therapy

Why aren't people living with HIV at higher risk for developing severe coronavirus disease 2019 (COVID-19)?

Open source & professional software for data science teams-RStudio

Emergency department (ED) utilization by HIV-infected ED patients in the United States in 2009 and 2010-a national estimation

COVID-19 in hospitalized adults with HIV

COVID-19 in HIV Investigators. COVID-19 in patients with HIV: clinical case series

Clinical features and outcomes of patients with human immunodeficiency virus with COVID-19

Hospitalized patients with COVID-19 and human immunodeficiency virus: a case series

Clinical characteristics, comorbidities and outcomes among persons with HIV hospitalized with coronavirus disease

SARS-CoV-2 and HIV coinfection: clinical experience from Rhode Island, United States

Disproportionate burden of COVID-19 among racial minorities and those in congregate settings among a large cohort of people with HIV

COVID-19 in persons living with HIV-what do we know today? NEJM J Watch

Factors associated with hospital admission and critical illness among 5279 people with coronavirus disease 2019 in New York City: prospective cohort study

Factors associated with COVID-19-related death using OpenSAFELY

Clinical predictors of mortality due to COVID-19 based on an analysis of data of 150 patients from Wuhan, China

Impact of SARS-CoV-2 viral load on risk of intubation and mortality among hospitalized patients with coronavirus disease 2019

COVID-19 and coagulation: bleeding and thrombotic manifestations of SARS-CoV-2 infection

Neutrophil-to-lymphocyte ratio and clinical outcome in COVID-19: a report from the Italian front line

Failure to normalize lymphopenia following trauma is associated with increased mortality

Nucleotide analogues as inhibitors of SARS-CoV-2 polymerase

Coronaviruses-drug discovery and therapeutic options

Candidate drugs against SARS-CoV-2 and COVID-19

A trial of lopinavir-ritonavir in adults hospitalized with severe COVID-19

Efficacy and safety of lopinavir/ritonavir or arbidol in adult patients with mild/moderate COVID-19: an exploratory randomized controlled trial

Financial support. For the research reported in this publication, J.L. was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number T32AI114398.Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.