Content uploaded by Ming Lee
Author content
All content in this area was uploaded by Ming Lee on Mar 09, 2021
Content may be subject to copyright.
1
Clinical outcomes of patients with and without HIV hospitalised with COVID-19 in England during
the early stages of the pandemic: a matched retrospective multicentre analysis (RECEDE-C19
study)
Ming Jie Lee1, Colette Smith2, Sam T Douthwaite, Sarah Fidler4, Naomi Fitzgerald1, Lynsey Goodwin5,
Lisa Hamzah6, Ranjababu Kulasegaram1, Sarah Lawrence5, Julianne Lwanga1, Rebecca Marchant6,
Chloe Orkin7, Adrian Palfreeman8, Padmini Parthasarathi6, Manish Pareek8, Kyle Ring4, Hamed
Sharaf5, Eliana Shekarchi-Khanghahi7, Rebecca Simons1, Luke B Snell9, Jhia Jiat Teh4, John Thornhill7,
Clare van Halsema5, Marie Williamson7, Martin Wiselka8, Julie Fox1, Achyuta Nori1
1. Harrison Wing, Guys’ and St Thomas’ NHS Foundation Trust, London, UK
2. Institute for Global Health, UCL, London, United Kingdom
3. Department of Virology, Guys’ and St Thomas’ NHS Foundation Trust, London, UK
4. Imperial College London, Department of Infectious Disease and Imperial College NIHR BRC,
Imperial College NHS Trust, London, UK
5. North Manchester General Hospital, Manchester, UK
6. Department of HIV, St George’s Hospital, London, UK
7. Barts Health NHS Trust, London, UK
8. University hospitals of Leicester, Leicester, UK
9. Centre for Clinical Infection & Diagnostics Research, King’s College London, London, UK
Contact author:
Dr Ming Lee
Harrison Wing, Guy’s and St Thomas Hospital NHS Foundation Trust, Great Maze Pond, London SE1
9RT, United Kingdom
Telephone: +44 207 188 2662
Email: minglee@doctors.org.uk
Running title: Analysis of HIV/COVID-19 outcomes
Meetings presented: AIDS 2020: Virtual, June 2020
Funding
This study has not received any funding sources. The corresponding author had full access to all data
and had the final responsibility for the decision to submit for publication.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
2
Research in context
Evidence before this study: Epidemiological data for people living with HIV (PLWH) have shown mixed
conclusions on the effect of HIV status on COVID-19 outcomes. Case series and matched case-control
studies have shown little evidence that COVID-19 outcomes in PLWH differ from HIV-negative
individuals. None of the matched studies evaluated socioeconomic deprivation, non-mortality
outcomes and were all conducted within the USA.
The OPENSAFELY study analysed primary care data to conclude that HIV-positive status was associated
with nearly three-fold (adjusted hazard ratio (aHR) 2.90, 95% confidence interval (CI) 1.96, 4.30)
increased risk of COVID-19 related mortality. The ISARIC analyses similarly reported findings from a
large multi-centre prospective cohort of UK hospitalised patients demonstrating an adjusted HR 1.49
(95% CI 0.99, 2.25) of increased 28-day mortality. Although these studies include large national
datasets, limitations include inaccurate or incomplete primary care coding for HIV, non-matched
populations, a lack of data on CD4 count, viral load, or HIV treatment available, and it is unclear if
PLWH with poorer outcomes were over-represented. Mortality endpoints may miss non-mortality
outcomes such as time to recovery and long-term disability.
Added value of this study:
We report the largest matched cohort study outside the USA, showing HIV status alone was not
associated with a difference in time to clinical improvement or hospital discharge, neither were
unadjusted mortality rates different between cohorts. Instead, greater baseline frailty and higher
proportion of active malignancy in the PLWH cohort were associated with worse outcomes. This study
was able to provide direct comparison to account not just for baseline demographics and key
comorbidities, but also socioeconomic deprivation, baseline frailty, and providing HIV disease
characteristics and HIV treatment detail. We were also able to analyse time to clinical improvement
or hospital discharge as a primary non-mortality outcome.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
3
Implications of all the available evidence:
Differences in clinical outcomes of COVID-19 hospitalisations in PLWH may be due to other important
factors including increased frailty and comorbidities such as malignancies, which are more prevalent
in PLWH, an ageing population in the era of effective ART. The presence of these comorbidities and
other risk factors rather than HIV-status alone should be considered for the prognosis of poorer
outcomes in severe COVID-19 infections or prioritization for vaccinations.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
4
Abstract
Background
It is unclear from epidemiological data for COVID-19 infections, whether people living with HIV (PLWH)
have a different outcome compared to the HIV-negative population. We conducted a multi-centre,
retrospective matched cohort study of SARS-CoV-2 PCR-positive hospital inpatients analysed by HIV-
status.
Methods
HIV-negative patients were matched to PLWH admitted to hospital from 1st February 2020 to 31st May
2020 up to a 3:1 ratio by: hospital site, SARS-CoV-2 test date +/- 7 days, age +/- 5 years, gender, and
index of multiple deprivation decile (IMDD) +/- 1. The primary objective was clinical improvement (≥2-
point improvement on a 7-point ordinal scale) or hospital discharge by day 28, whichever was earlier.
Results
68 PLWH and 181 HIV-negative comparators were included. After adjustment for ethnicity, frailty,
baseline hypoxia, duration of symptoms prior to baseline, body mass index (BMI) categories, and
comorbidities (hypertension, chronic cardiac disease, chronic lung disease, active malignancy,
diabetes, and chronic renal disease), the effect size of HIV-status was not associated with time to
clinical improvement or discharge from hospital (aHR 0.70, 95%CI 0.43, 1.17; p=0.18), despite
unadjusted hazards of PLWH achieving the primary outcome being 43% lower (p=0.005). Baseline
frailty (aHR=0.79; 95%CI 0.65, 0.95; p=0.011), malignancy (aHR=0.37; 95%CI 0.17, 0.82; p=0.014)
remained associated with poorer outcomes. PLWH were more likely of black and minority ethnicities
(75.0% vs 48.6%, p=0.0002), higher median clinical frailty score (3 IQR 2-5 vs 2 IQR 1-4, p=0.0069),
higher proportion of active malignancy (14.4% vs 9.9%, p=0.29). Median CD4 count of PLWH was
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
5
352cells/µL (IQR 235-619) and 95.7% had suppressed viral loads <200copies/mL, 63/68 (92.3%) were
taking antiretroviral therapy.
Conclusions
Differences in clinical outcomes of COVID-19 hospitalisations in PLWH may be due to other important
factors including increased frailty and comorbidities such as malignancies, rather than HIV-status
alone.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
6
Background
SARS-CoV-2 infection is estimated to cause mild to moderate disease (coronavirus disease 2019 or
COVID-19) in about 80% of people but can cause severe lower respiratory tract infection in
approximately 15-20%, particularly among those in high-risk groups, defined by advanced age (≥65
years), ethnicity (African and Asian) presence of comorbidities (e.g., cardiopulmonary disease,
diabetes mellitus) or obesity1–4.
Epidemiological data for people living with HIV (PLWH) have shown mixed conclusions on the effect
of HIV status on COVID-19 outcomes. Case series2,5–7 and matched case-control studies8–11 have shown
little evidence that COVID-19 outcomes in PLWH differ from HIV-negative individuals. None of the
matched studies evaluated socioeconomic deprivation, non-mortality outcomes and were all
conducted within the USA.
The OPENSAFELY12 study analysed primary care data to conclude that HIV-positive status was
associated with nearly three-fold (adjusted hazard ratio (aHR) 2.90, 95% confidence interval (CI) 1.96,
4.30) increased risk of COVID-19 related mortality. The ISARIC13 analyses similarly reported findings
from a large multi-centre prospective cohort of UK hospitalised patients demonstrating an adjusted
HR 1.49 (95% CI 0.99, 2.25) of increased 28-day mortality. Although these studies include large
national datasets, limitations include inaccurate or incomplete primary care coding for HIV, non-
matched populations, a lack of data on CD4 count, viral load, or HIV treatment available, and it is
unclear if PLWH with poorer outcomes were over-represented. Mortality endpoints may miss non-
mortality outcomes such as time to recovery and long-term disability.
To estimate the epidemiological effect of HIV status and other confounding variables on the outcomes
of PLWH hospitalised with COVID-19, we conducted a multi-centre, retrospective matched analysis of
people living with and without HIV hospitalised with COVID-19 across England.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
7
Methods
Study design
RECEDE-C19 is a multicentre retrospective matched cohort study, analysing outcomes in people living
with HIV admitted with COVID-19 to a matched cohort of HIV-negative individuals admitted with
COVID-19 up to a 1:3 ratio. Ethical approval was granted by the UK Health Research Authority (REC
reference 20/HRA/2278).
Study setting and participants
Six large hospital trusts across England were included: four in London (Imperial College Healthcare
NHS Foundation Trust, Barts Health NHS Trust, St George’s Hospital NHS Foundation Trust, and Guy’s
and St Thomas Hospitals NHS Foundation Trust), the Pennine Acute Hospitals NHS Trust in Greater
Manchester, and the University Hospitals of Leicester NHS Trust in Leicester.
Inclusion criteria: PLWH aged 18 and older admitted to hospital with a confirmed diagnosis of COVID-
19 were identified from the 1st February 2020 to the 31st May 2020. The comparator cohort of HIV-
negative individuals was identified from hospitalised patients with a presumed or confirmed negative
HIV status and a confirmed diagnosis of COVID-19.
Exclusion criteria: Patients diagnosed with COVID-19 not requiring admission at time of presentation.
Cohort matching
Selection of matched comparator patients: HIV-negative individuals were matched to each PLWH
identified by the following criteria - hospital site, SARS-CoV-2 test date within seven days, age within
five years, same gender, and within one decile of each PLWH’s index of multiple deprivation decile
(IMDD) based on post code, a surrogate marker of socioeconomic status. Ethnicity data are often
poorly defined and collected14, thus were not included in the matching criteria; IMDD was used instead
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
8
to account for geographical deprivation. Investigators blinded to patient outcomes in the comparator
cohort, selected up to a ratio of three comparator patients meeting the above criteria to each PLWH.
Outcomes and data sources
The primary outcome was defined as clinical improvement or discharge from hospital by day 28,
whichever was earlier. Clinical improvement was defined as an improvement of two points or more
from the baseline status, on a seven-point ordinal scale described by Cao et al15 from 1: not
hospitalised with resumption of normal activities’; to 7: palliation or death (Categories presented in
Supplementary table 1).
Secondary outcomes include 28-day mortality, time to death, length of hospitalisation, proportion of
patients requiring high dependency or intensive care admission, requirements for organ support
including mechanical ventilation or renal replacement therapy, and laboratory markers both at
baseline and most abnormal results during hospitalisation. The data were collected by researchers
from the direct care team reviewing electronic patient records.
Study definitions
Confirmed COVID-19 diagnosis was defined with detectable SARS-CoV-2 RNA by RT-PCR from a
combined nose and throat swab or other respiratory samples. Deprivation scores was determined by
mapping the patient’s post codes to the English index of multiple deprivation score16 by decile, 1 the
least affluent, to 10 the most affluent decile. Patient ethnicities are self-reported and coded as part of
the patient demographics records.
The clinical frailty score (CFS)17 is a seven-point scale for assessment of frailty from a scale of 1 (very
fit) to 9 (terminally ill) (Supplementary table 1). If not recorded at the time of admission, the CFS score
was retrospectively applied based on the patient’s documented activities of daily living from the
admission history or physiotherapy documentation.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
9
Study baseline was defined as date of admission (if admission reason was COVID-19 related) or date
of first detectable SARS-CoV-2 result (if admission reason was not COVID-19 related), patients were
included even if their admission were not COVID-19 related as nosocomial SARS-CoV-2 transmission
remains a significant concern18. Baseline HIV viral load results were recorded if available at time of
admission or the most recent result within six months of admission, and baseline CD4 count, CD4
percentage, CD4:CD8 ratio was recorded from the first result available during or most recent to the
admission date. Immunosuppressed PLWH were defined as those with CD4 counts ≤200 cells/µL and
CD4 percentage ≤14%. Number of comorbidities was summed from the 18 comorbidities categories
recorded (List of comorbidities presented in Supplementary Table 1).
End of study was defined as completion of the follow up period at day 28 from baseline, or if the
patient is discharged from hospital, or death where the cause of death was recorded if known.
Study size and statistical analysis
The study was initially designed to evaluate non-inferiority of the primary outcome for PLWH
compared to HIV-negative individuals, with a non-inferiority margin for the hazard ratio set at 0.81.
Assuming HIV-status did not have any impact on the primary outcome, 20% of the cohort would either
die or not recover by day 28, and the median duration to improvement was 6 days in the remaining
80%, data simulations indicated that 50 PLWH and 100 HIV-negative comparators were required to
provide 80% power to show non-inferiority.
The primary outcome was assessed after patients had reached day 28, with failure to reach clinical
improvement, hospital discharge or death before day 28 considered as right-censored at day 28.
Univariable and multivariable Cox proportional hazards regression was performed, stratified by
matching cluster. Multiple imputation was used to account for missing data across 20 simulated
datasets, and results were combined using Rubin’s rules. It was decided a priori to include the
following covariates in the multivariable model, as they have been shown previously to be associated
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
10
with COVID-19 outcomes: ethnicity, clinical frailty score, body mass index (BMI), hypoxia at admission,
days with symptoms at admission, hypertension, diabetes, and chronic kidney disease. Other
covariates were included if the associated p-value as <0.2 in univariable analysis. Matched factors
were not included as covariates.
A number of sensitivity analyses were performed to assess the robustness of the association between
HIV status and time to a two-point improvement or discharge: (1) complete case analysis excluding
those with missing data (n=217); (2) excluding COVID-related factors (baseline hypoxia at admission,
duration of symptoms at admission); (3) excluding ethnicity; (4) additionally adjusting for age to
account for any residual confounding (as clusters were matched to within 5 years); (5) adjusting for
the matching variables instead of stratification.
Secondary endpoints included the time to death, which was investigated using standard survival
analysis. As numbers of events were small, only univariable analysis was performed. Additional
secondary outcomes were summarised using number and percentage or median (inter-quartile range)
as appropriate and compared between groups using a chi-squared test, Fisher’s exact test or Mann
Whitney U-test. Multiplicity in hypothesis testing was not accounted for, so results from secondary
analyses should be seen as indicative findings. Analyses were performed using SAS, Version 9.4 (SAS
Institute Inc, Cary, NC) and Stata, Version 14 (Statacorp, College Station, Texas).
Results
Patient population, characteristics, and comorbidities at baseline
6612 people with COVID-19 between 1st February to 31st May 2020 were hospitalised across the
RECEDE-C19 study sites. 69 patients (1.04%) had known HIV-positive status. No HIV-negative
comparators within the same site were available for one PLWH, the remaining 68 PLWH were matched
to between one to three HIV-negative comparator patients (Supplementary table 2); in total 181
patients were included in the comparator cohort for analysis.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
11
Baseline characteristics data are displayed in Table 1 and further baseline characteristics are
presented in Supplementary Table 3. Reported smoking and excess alcohol use was not significantly
different across groups, and although more PLWH reported recreational drug use, this field was poorly
documented across both cohorts.
Baseline symptoms and laboratory markers
Prior to their COVID-19 diagnosis, PLWH had similar duration of COVID-19 associated symptoms onset
to the HIV-negative cohort (median 6 vs 7 days, p=0.23) (Table 1), and the proportion of patients
presenting with hypoxia (peripheral oxygen saturations <94% on air) were similar (55.9% vs 57.5%,
p=0.82). 83.8% of PLWH and 86.7% of HIV-negative individuals were admitted for COVID-19 related
reasons, and the remainder had incidental COVID-19 diagnoses during admissions. Symptoms
reported at time of diagnosis were similar between PLWH and HIV-negative patients (Supplementary
table 4).
With regards to laboratory markers, PLWH were more likely to be anaemic, had a lower white cell
count, and higher lymphocyte count at baseline than HIV-negative patients (Table 1). Platelet count,
estimated glomerular filtration rate (eGFR), C-reactive protein (CRP), and alanine transaminase were
similar between cohorts (Supplementary table 4).
Baseline HIV disease characteristics in PLWH cohort
Table 2 summarises the characteristics of HIV infection and ARVs for the PLWH cohort. The median
CD4 count of PLWH was 352cells/µL (IQR 235-619), median time since HIV diagnosis 14.8 years, and
95.7% had suppressed viral loads <200 copies/mL. 63/68 (92.6%) were receiving ARVs at baseline.
Of the immunosuppressed patients with HIV, the median time since HIV diagnosis was shorter at 9
years, and the median CD4 count was 83 (IQR 76, 139). 4/5 (80%) had suppressed viral loads and were
receiving ARVs.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
12
Clinical outcomes
After adjustment for ethnicity, frailty, baseline hypoxia, duration of symptoms prior to baseline, BMI
categories, and comorbidities including hypertension, chronic cardiac disease, chronic lung disease,
active malignancy, diabetes, and chronic renal disease, HIV status was not associated with time to
clinical improvement or discharge (aHR 0.70, 95%CI 0.43, 1.17; p=0.18) (Table 3) The unadjusted
cumulative hazard of patients reaching the primary outcome was 43% lower in PLWH than HIV-
negative patients (p=0.005) (Figure 1), and the difference in outcomes were associated with greater
baseline clinical frailty scores (aHR=0.79; 95%CI 0.65, 0.95; p=0.011) and malignancy (aHR=0.37; 95%CI
0.17, 0.82; p=0.014) after adjustment. There was a trend for patients at extreme BMI categories being
associated with lower hazards to achieving the primary outcome (BMI <25: aHR 0.46; 95% CI 0.21,
0.99; p=0.047).
HIV status was not associated with difference in mortality rates (HR 1.18, 95%CI 0.54, 2.60, p=0.68)
(Table 3, Figure 2). The sensitivity analyses (Table 5) generally found consistent results. A complete
case analysis (n=217) led to greater attenuation (aHR 0.90; 95%CI 0.51, 1.59; p=0.72). When not
adjusting for ethnicity, the association between HIV and the primary outcome then reached statistical
significance at the 5% level (aHR 0.62; 95%CI 0.39, 0.96; p=0.031), suggesting the contribution of
confounding effect of ethnicity was not fully explained by matching for geographical deprivation.
Analyses adjusted for the matching variables instead of stratifying by clusters led to an attenuated
association (aHR=0.90; 95%CI 0.63, 1.30; p=0.58).
Secondary outcomes are summarized in Table 4. PLWH had longer overall duration of hospitalisations
from COVID-19 diagnosis (median 10 vs 7.5days, p=0.0061). A higher proportion of PLWH required
mechanical ventilation (23.5% vs 17.1%, p=0.25) during admission but did not meet significance, and
if admitted to critical care, the median duration in critical care was similar between groups (14 vs 15;
p=0.83) (Supplementary table 4). A minority of patients received COVID-19 specific trial medications,
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
13
4 PLWH (2 high-dose steroids, 1 tocilizumab, 1 remdesivir), and 5 HIV-negative patients (2 high-dose
steroids, 2 tocilizumab, 1 lopinavir/ritonavir).
PLWH were more likely to have a lower nadir haemoglobin level during admission than HIV-negative
patients (103 vs 116, p=0.0029), however there was no significant difference between the most
abnormal results for white cell count, lymphocyte count, CRP (Table 4), platelet count, eGFR, and ALT
results during admission (Supplementary table 4).
In the immunosuppressed PLWH subgroup and their comparators (Table 4), there were 2 deaths in
the HIV-negative comparators, and zero deaths in immunosuppressed PLWH. Immunosuppressed
PLWH had longer duration of hospitalisations and time to primary outcome although not achieving
significance. Proportion requiring ITU/HDU care (40.0% vs 50.0%, p=1.00), mechanical ventilation
(40.0% vs 41.7%, p=1.00), or initiation of renal replacement therapy requirements (20.0% vs 8.3%,
p=0.515) were similar across both immunosuppressed PLWH and their matched comparator groups.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
14
Discussion
Key results and interpretation
We report the largest matched cohort study outside the USA, showing HIV status alone was not
associated with a difference in time to clinical improvement or hospital discharge, neither were
unadjusted mortality rates different between cohorts. Instead, greater baseline frailty and higher
proportion of active malignancy in the PLWH cohort were associated with worse outcomes.
The sample size was based on non-inferiority to demonstrate less than a 20% reduction in rate of
improvement. Although the overall adjusted HR for HIV status did not demonstrate strong evidence
of a difference in outcomes, the confidence interval was wide with 95% of PLWH being between 57%
less likely and 17% more likely to achieve clinical improvement compared to HIV-negative patients.
This introduces uncertainty in this estimate, which could be overcome with larger studies with
additional power.
The association with age, obesity, co-morbidities and male gender and worse COVID-19 outcomes
have been well described1,2,19. There was a trend for BMI extremes to be associated with poorer
outcomes, however the confidence intervals were wide with borderline significance for lower BMI
categories, likely reflecting incomplete data fields.
There is a disproportionate impact on COVID-19 in people of BAME backgrounds, even when age, sex
and comorbidities are adjusted for4. When considering matching criteria, ethnicity was not included
as ethnicity has been previously reported to be poorly defined or collected14, IMDD was used instead
to address the contribution of geographical deprivation to COVID-19 outcomes. Ethnicity was
subsequently adjusted for in the multivariable analysis and explored further in the sensitivity analyses.
The change in significance level when ethnicity was removed from the multivariable model suggests
the study size was underpowered to fully explain the confounding effect of ethnicity on COVID-19
outcomes by geographical deprivation alone, or other possible reasons for poorer outcomes of COVID-
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
15
19 infections in people of BAME backgrounds were contributing to inequalities in outcomes by
ethnicity; these may include cultural factors, poor housing, overcrowding, likelihood of low-paid
essential jobs, and increased prevalence of comorbidities amongst people of BAME backgrounds20.
PLWH were frailer, as measured by the CFS, compared to their matched HIV-comparators and greater
frailty predicted worse outcomes. The CFS has been recommended by UK national guidelines for
COVID-1921 for assessment of all adults on admission, and while not validated in those younger than
65 years, the COPE study showed the CFS was predictive of increased mortality and longer duration
of hospitalisation, even after adjustment for age and comorbidities, including in patients under 6522.
The CFS may be useful as a tool to guide decision making even in PLWH admitted with COVID-19.
Comparison with other studies
This study adds to the growing literature describing the complex interplay between HIV and COVID-
195–7,9,10,12,13,23–25. The ISARIC findings13 suggests PLWH had an age-adjusted 47% increased risk of
mortality by day 28, which increased to 100% after adjustment. The UK population-wide OPENSAFELY
database showed a 3.8-fold higher risk of COVID-19 death in PLWH, but both studies were unmatched,
unable to include HIV markers, antiretroviral therapy use, and possibility of incomplete or misclassified
HIV coding. Similarly, HIV was associated with higher hazards for mortality in South Africa26 and the
USA10. In comparison, we did not show a difference in time to clinical improvement or discharge from
hospital once confounders were adjusted for in cohorts matched for age, gender and deprivation. This
study was also able to adjust for baseline frailty, co-morbidities, and provide details of HIV disease
characteristics. Compared to ISARIC13, the matched IMDD criteria in this study may have ameliorated
the inequalities in health access contributing to early mortality outcomes not measured in the ISARAIC
protocol.
Immunosuppression and COVID-19
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
16
In this study, immunosuppressed PLWH did not appear to have poorer outcomes or higher risk of
deaths compared to PLWH with higher CD4 percentages. Because the numbers were small, and most
patients were receiving ARVs, these findings may not be generalizable to populations with lower
median CD4 counts and lower proportions not receiving ARVs. Reports of COVID-19 in
immunosuppressed patients are still limited5,7,9,27, a recent study suggest a CD4 cell count <350cells/µL
were associated with severity of COVID27. While the OPENSAFELY platform showed patients with other
immunosuppressive conditions a 2.21 adjusted increased risk for COVID-19 related mortality28, and a
meta-analysis of eight studies similarly showed a 3.29 increased risk of severe COVID-19 disease29,
these studies did not provide details of HIV infection or treatment in PLWH included in the
immunodeficiency categories. Larger studies with detailed markers of immunity are required to
further evaluate the risk of severe COVID-19 disease in immunosuppressed PLWH.
Strengths and Limitations
This study is the largest matched cohort analysis of PLWH and HIV-negative individuals hospitalised
with COVID-19 outside the USA and is able to provide direct comparison to account not just for
baseline demographics and key comorbidities, but also socioeconomic deprivation, baseline frailty,
and providing HIV disease characteristics and HIV treatment detail. We were also able to analyse time
to clinical improvement or hospital discharge as a primary non-mortality outcome.
There were limitations to this study. Firstly, the study was restricted to hospitalised patients only,
which may introduce collider bias, where variables which lead to hospitalisation may also directly
affect outcomes, caution is required in generalising the findings beyond hospitalised patients. Sites in
this study were predominantly within the greater London area. London was the epicenter of the first
peak of COVID-19 hospitalisations in the UK, and the data were likely to be reflective of hospitalised
cases across the UK. 10.5% of patients included in the HIV-negative cohort were presumed negative
due to the lack of availability of a HIV test result within the previous 12 months. These patients were
included as the estimated numbers of people living with undiagnosed HIV in the UK have dramatically
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
17
fallen, with 6% of all PLWH estimated to have undiagnosed HIV in 201930, the probability of an
undiagnosed PLWH included in the HIV-negative cohort was low. Data were retrospectively collected
and there were fields with missing data such as BMI, this may have resulted in bias as patients who
were discharged early may not have results available, on the other extreme patients may be too
unwell to have their height or weight measured accurately. Follow up was ended at time of discharge
and whether patients were able to fully resume usual activities were only able to be assessed on day
of discharge, if patients subsequently improved or deteriorated with follow up complications these
outcomes may not be fully captured. We were also not able to include complications or mortality
outcomes beyond 28 days and are unable to comment on long-term complications of COVID-19.
Finally, the study ordinal scale endpoints described by Cao et al15 was originally described for use in
therapeutic trials, the use of the ordinal scale in this retrospective observation study allowed finer
inspection of non-mortality outcomes, compared to larger sample sizes and longer follow-up required
to account for the lag in mortality data.
Conclusions
Differences in clinical outcomes of COVID-19 hospitalisations in PLWH may be due to other important
factors including increased frailty and comorbidities such as malignancies, which are more prevalent
in PLWH, an ageing population in the era of effective ART. The presence of these comorbidities and
other risk factors rather than HIV-status alone should be considered for the prognosis of poorer
outcomes in severe COVID-19 infections or prioritization for vaccinations.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
18
Acknowledgements
This study used data collected in the routine care of NHS patients, by the NHS staff involved in their
care and we acknowledge both patients and staff in their contributions to this study. We acknowledge
the support of Alice Sharp, Elizabeth Bruna, Marie-Rose Dwek, Jo Bagshaw and Shirin Hussein.
Conflicts of interests declarations
MJL has received grants and honoraria from Gilead Sciences and Viiv Healthcare not related to this
work. SF has received research grants to her institution from NIH, MRC, BMGF. JT has received support
for virtual conference registration from ViiV Healthcare and research grants from the Medical
Research Council and the British HIV Association not related to this work. CvH has received educational
grants, conference support and advisory board fees from ViiV Healthcare, Gilead Sciences, MSC not
related to this work. MP reports grants and personal fees from Gilead Sciences and personal fees from
QIAGEN, outside the submitted work. MP is supported by a NIHR Development and Skills
Enhancement Award (NIHR301192) and in receipt of funding from UKRI / MRC (MR/V027549/1). He
acknowledges the support from UKRI, the NIHR Leicester BRC and NIHR ARC East Midlands. No other
competing interests, financial relationships with any organisations that might have an interest in the
submitted work, or other relationships or activities that could appear to have influenced the
submitted work have been reported by other authors.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
19
References
1 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. Lancet 2020; 395: 1054–62.
2 Richardson S, Hirsch JS, Narasimhan M, et al. Presenting Characteristics, Comorbidities, and
Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. JAMA
2020; published online April 22. DOI:10.1001/jama.2020.6775.
3 Docherty AB, Harrison EM, Green CA, et al. Features of 20 133 UK patients in hospital with
covid-19 using the ISARIC WHO Clinical Characterisation Protocol: Prospective observational
cohort study. BMJ 2020; 369. DOI:10.1136/bmj.m1985.
4 Sze S, Pan D, Nevill CR, et al. Ethnicity and clinical outcomes in COVID-19: A systematic review
and meta-analysis. EClinicalMedicine 2020; 0: 100630.
5 Blanco JL, Ambrosioni J, Garcia F, et al. COVID-19 in patients with HIV: clinical case series.
Lancet HIV. 2020; 7: e314–6.
6 Inciarte A, Gonzalez-Cordon A, Rojas J, et al. Clinical characteristics, risk factors, and incidence
of symptomatic coronavirus disease 2019 in a large cohort of adults living with HIV: a single-
center, prospective observational study. AIDS 2020; 34: 1775–80.
7 Härter G, Spinner CD, Roider J, et al. COVID-19 in people living with human immunodeficiency
virus: A case series of 33 patients. DOI:10.1101/2020.04.28.20073767.
8 Karmen-Tuohy S, Carlucci PM, Zervou FN, et al. Outcomes Among HIV-Positive Patients
Hospitalized With COVID-19. J Acquir Immune Defic Syndr 2020; 85: 6–10.
9 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; published online June 28. DOI:10.1093/cid/ciaa880.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
20
10 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.
DOI:10.1097/QAD.0000000000002666.
11 Stoeckle K, Johnston CD, Jannat-Khah DP, et al. COVID-19 in hospitalized adults with HIV.
Open Forum Infect Dis 2020; 7. DOI:10.1093/ofid/ofaa327.
12 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 2020; 0. DOI:10.1016/S2352-3018(20)30305-2.
13 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. Clin Infect Dis 2020; published online Oct 23.
DOI:10.1093/cid/ciaa1605.
14 Bokor-Billmann T, Langan EA, Billmann F. The reporting of race and/or ethnicity in the
medical literature: a retrospective bibliometric analysis confirmed room for improvement. J
Clin Epidemiol 2020; 119: 1–6.
15 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; published online March 18. DOI:10.1056/NEJMoa2001282.
16 Ministry of Housing Communities & Local Government. English indices of deprivation 2019.
http://imd-by-postcode.opendatacommunities.org/imd/2019.
17 Rockwood K, Song X, MacKnight C, et al. A global clinical measure of fitness and frailty in
elderly people. CMAJ 2005; 173: 489–95.
18 Wake RM, Morgan M, Choi J, Winn S. Reducing nosocomial transmission of COVID-19:
Implementation of a COVID-19 triage system. Clin Med J R Coll Physicians London 2020; 20:
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
21
E141–5.
19 Liang W, Guan W, Chen R, et al. Cancer patients in SARS-CoV-2 infection: a nationwide
analysis in China. Lancet Oncol. 2020; 21: 335–7.
20 Khunti K, Singh AK, Pareek M, Hanif W. Is ethnicity linked to incidence or outcomes of covid-
19? BMJ. 2020; 369. DOI:10.1136/bmj.m1548.
21 NICE. COVID-19 rapid guideline: critical care in adults. Natl Inst Heal Care Excell 2020; : 2020.
22 Hewitt J, Carter B, Vilches-Moraga A, et al. The effect of frailty on survival in patients with
COVID-19 (COPE): a multicentre, European, observational cohort study. Lancet Public Heal
2020; 5: e444–51.
23 Mondi A, Cimini E, Colavita F, et al. COVID-19 in people living with HIV: Clinical implications of
dynamics of the immune response to SARS-CoV-2. J Med Virol 2020. DOI:10.1002/jmv.26556.
24 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.
25 Mirzaei H, McFarland W, Karamouzian M, Sharifi H. COVID-19 Among People Living with HIV:
A Systematic Review. AIDS Behav. 2020. DOI:10.1007/s10461-020-02983-2.
26 Boulle A, Davies M-A, Hussey H, et al. Risk factors for COVID-19 death in a population cohort
study from the Western Cape Province, South Africa. Clin Infect Dis 2020; published online
Aug 29. DOI:10.1093/cid/ciaa1198.
27 Hoffmann C, Casado JL, Härter G, et al. Immune deficiency is a risk factor for severe COVID-19
in people living with HIV. HIV Med 2020. DOI:10.1111/hiv.13037.
28 Williamson EJ, Walker AJ, Bhaskaran K, et al. Factors associated with COVID-19-related death
using OpenSAFELY. Nature 2020; 584: 430–6.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
22
29 Gao Y, Chen Y, Liu M, Shi S, Tian J. Impacts of immunosuppression and immunodeficiency on
COVID-19: A systematic review and meta-analysis. J. Infect. 2020; 81: e93–5.
30 Public Health England. Trends in HIV testing, new diagnoses and people receiving HIV-related
care in the United Kingdom: data to the end of December 2019. 2020
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment
(accessed Nov 4, 2020).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
23
Tables and Figures
Table 1. Baseline patient characteristics and investigations stratified by HIV-status
HIV-positive individuals
HIV-negative individuals
N = 68
% / IQR
N = 181
% / IQR
p-value
Median Age
57
50, 63
56
51, 62
†
Median Index of Multiple Deprivation Decile
3
2, 4
3
2, 4
†
Gender
†
Cis Female
26
38.2%
67
37.0%
Cis Male
42
61.8%
114
63.0%
Median Body Mass Index
27.7 (n=52)
23.9, 32.3
29.4 (n = 115)
24.7, 34.3
0.19
Median Clinical Frailty Score
3
2, 5
2
1, 4
0.0069
Ethnicity
-
White
17
25.0%
78
47.0%
Black African
39
57.4%
33
18.2%
Black Caribbean
1
1.5%
14
7.7%
Asian
4
5.9%
18
10.8%
Mixed or Other ethnic groups not listed
7
10.3%
23
13.9%
Not documented
0
0.0%
15
8.3%
Ethnicity categories
0.0002
Black and minority ethnicities
51
75.0%
88
48.6%
White
17
25.0%
93
51.4%
Comorbidities
Hypertension
35
51.5%
74
40.9%
0.13
Diabetes Mellitus (Type 1 or Type 2)
18
26.5%
53
29.3%
0.66
History of diabetic complications
6
8.8%
19
10.5%
0.70
Asthma
3
4.4%
24
13.3%
0.045
Chronic pulmonary disease
6
8.8%
21
11.6%
0.53
Chronic cardiac disease
12
17.6%
22
12.1%
0.26
Liver disease (Child-Pugh score B or C)
3
4.4%
1
0.6%
0.031
Chronic Hepatitis B
1
1.5%
4
2.2%
0.71
Chronic Hepatitis C (untreated)
2
2.9%
1
0.6%
0.12
Chronic neurological disorder
9
13.2%
19
10.5%
0.54
Mental health disorder
12
17.6%
30
16.6%
0.84
Active malignancy
10
14.7%
18
9.9%
0.29
Chronic haematological disorder
6
8.8%
9
5.0%
0.26
Rheumatological disease
0
0.0%
19
10.5%
0.0054
Dementia
4
5.9%
10
5.5%
1.00
Malnutrition
0
0.0%
5
2.8%
0.33
Chronic Kidney Disease (stage 3 or worse)
24
35.3%
23
12.7%
<0.0001
End stage renal failure requiring dialysis
13
19.2%
9
5.0%
0.0005
Number of comorbidities
2
1, 3
2
1, 3
0.16
Median duration from symptom onset (days)
at baseline
6 (n=47)
1, 10
7 (n=173)
3, 10
0.23
Hypoxia on presentation (oxygen saturations
<94%)
38
55.9%
104
57.5%
0.82
Admission reason related to COVID-19
57
83.8%
157
86.7%
0.58
Baseline investigations
White cell count (x109/L)
6.1 (n=66)
4.7, 8.4
7.5 (n=177)
5.5, 9.8
0.016
Lymphocytes (x109/L)
1.2 (n=65)
0.85, 1.6
0.9 (n=178)
0.64, 1.3
0.0008
C-Reactive Protein (mg/L)
108.5 (n=68)
44, 178.5
92 (n=176)
44, 192.4
0.92
Abbreviations: IQR – Inter-quartile range
† Matched demographics between cohorts
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
24
Table 2. Characteristics of HIV infection in PLWH cohort
N /
Median
% / IQR
All PLWH (n=68)
Median time since HIV diagnosis (years) (n=59)
14.8
10.2, 18.8
Median CD4 count at time of COVID-19 diagnosis (cells/µL) (n=67)
352
235, 619
Median CD4:CD8 ratio (n=52)
0.88
0.4, 1.2
Median CD4 percentage (n=57)
30
21, 36
Patients with suppressed HIV viral load <200copies/ml (n=69)
66
95.7%
At time of COVID-19 diagnosis, number of patients on antiretroviral regimens
containing the following:
Tenofovir disoproxil
17
27.0%
Tenofovir alafenamide
21
33.3%
Integrase strand transfer inhibitors
30
47.6%
Protease inhibitors
21
33.3%
Non-nucleoside reverse transcriptase inhibitors
17
27.0%
Number of patients not receiving antiretrovirals
5
7.4%
PLWH with CD4 counts < 200cells/µL & CD4 percentage < 14% (n=5)
Median time since HIV diagnosis (years) (n=5)
9
3, 12
Median CD4 count at time of COVID-19 diagnosis (cells/µL) (n=5)
83
76, 139
Median CD4:CD8 ratio (n=5)
0.2
0.17, 0.2
Median CD4 percentage (n=5)
12
10, 12
Patients with suppressed HIV viral load <200copies/ml (n=5)
4
80.0%
At time of COVID-19 diagnosis, number of patients on antiretroviral regimens
containing the following:
Tenofovir disoproxil
0
0.0%
Tenofovir alafenamide
1
20.0%
Integrase strand transfer inhibitor
2
40.0%
Protease inhibitors
1
20.0%
Non-nucleoside reverse transcriptase inhibitor
1
20.0%
Not on antiretrovirals
1
20.0%
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
25
Table 3 – Primary outcomes and mortality outcomes analysis
Univariable
Multivariable
Variable
Reference variable
HR
95% CI
p-value
HR
95% CI
p-value
Primary outcome analysis – Factors associated with time to clinical improvement or discharge
HIV-positive status
HIV-negative status
0.57
0.39, 0.85
0.005
0.70
0.43, 1.17
0.18
Ethnicity - BAME
White
0.59
0.39, 0.89
0.012
0.86
0.52, 1.42
0.55
Clinical frailty score - Per
1 higher
0.74
0.63, 0.86
<0.0001
0.79
0.65, 0.95
0.011
BMI (kg/m2)
25 - 30
1.00
1.00
<25
0.49
0.26, 0.96
0.12
0.46
0.21, 0.99
0.047
30-35
0.96
0.49, 1.90
0.99
0.47, 2.11
0.98
>35
0.65
0.32, 1.32
0.65
0.29, 1.48
0.30
Hypoxic at admission
Not hypoxic at
admission
0.80
0.54, 1.18
0.27
0.67
0.41, 1.09
0.10
Days with symptoms at
admission – Per 1 day
longer
1.02
0.98, 1.06
0.28
1.00
0.95, 1.04
0.94
Hypertension
No comorbidity
0.70
0.46, 1.06
0.094
0.88
0.52, 1.47
0.63
Chronic cardiac disease
No comorbidity
0.49
0.24, 0.99
0.048
0.77
0.34, 1.74
0.53
Chronic lung disease
No comorbidity
0.76
0.38, 1.55
0.45
1.08
0.48, 2.41
0.85
Asthma
No comorbidity
1.37
0.75, 2.50
0.31
Neurological condition
No comorbidity
1.17
0.62, 2.21
0.62
Active malignancy
No comorbidity
0.38
0.19, 0.77
0.007
0.37
0.17, 0.82
0.014
Diabetes
No comorbidity
0.79
0.50, 1.22
0.29
0.73
0.43, 1.25
0.26
Rheumatological disease
No comorbidity
1.57
0.74, 3.33
0.24
Chronic renal disease
No comorbidity
0.51
0.29, 0.90
0.019
0.79
0.40, 1.58
0.51
Mortality analysis – Factors associated with time to death
HIV-positive status
HIV-negative status
1.18
0.54, 2.60
0.68
Ethnicity - BAME
White
2.29
0.90, 5.86
0.083
Clinical frailty score - Per
1 higher
1.24
0.97, 1.60
0.092
Hypoxic at admission
Not hypoxic at
admission
2.08
0.69, 6.26
0.20
Days with symptoms at
admission – Per 1 day
longer
0.98
0.89, 1.09
0.72
BMI (kg/m2)
25 - 30
1.00
<25
1.17
0.30, 4.55
0.72
30-35
1.28
0.22, 7.27
>35
2.56
0.43, 15.2
Comorbidities - Per 1
additional
1.16
0.89, 1.52
0.28
Hypertension
No comorbidity
1.46
0.55, 3.85
0.45
Chronic cardiac disease
No comorbidity
1.66
0.65, 4.28
0.29
Chronic lung disease
No comorbidity
2.38
0.75, 2.55
0.14
Asthma
No comorbidity
0.80
0.19, 3.40
0.77
Neurological condition
No comorbidity
0.64
0.15, 2.78
0.56
Active malignancy
No comorbidity
2.59
0.80, 8.40
0.11
Diabetes
No comorbidity
1.00
0.44, 2.27
1.00
Rheumatological disease
No comorbidity
3.00
0.25, 35.8
0.39
Chronic renal disease
No comorbidity
1.48
0.61, 3.55
0.39
Abbreviations: HR=hazard ratio; CI=confidence interval; BAME=Black, Asian and Minority Ethnicities; BMI= Body Mass Index.
Results from Cox proportional hazards model stratified by matching clusters, with missing data accounted for using multiple
imputation with chained equations (20 simulated datasets combined using Rubin’s rules). Clinical centre, date of admission,
gender, age and IMD decile were not included as co-variates as these were matching variables.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
26
Table 4. Secondary outcomes by day 28 following hospitalisation, stratified by HIV-status
Abbreviations: IQR - Inter-quartile range; ECMO – Extra-corporeal membrane oxygenation; ITU – Intensive Care Unit;
HDU – High Dependency Unit
HIV-positive individuals
HIV-negative individuals
p-values
N = 68
% / IQR
N = 181
% / IQR
Patients achieving primary outcome
(≥ 2 points improvement or discharge from hospital)
47
69.1%
138
76.2%
0.25
Recorded outcome by day 28
0.266
Left hospital alive
45
66.2%
133
73.4%
Remained inpatient on ward
5
7.4%
8
4.4%
Remained inpatient in critical care (ICU + HDU)
5
7.4%
5
2.8%
Death
13
19.1%
35
19.3%
Median duration of hospitalisation (days)
10 (n=65)
6, 23
7.5 (n=178)
4, 14
0.0061
Required mechanical ventilation during admission
16
23.5%
31
17.1%
0.25
Received trial drug or specific COVID-19 therapy
4
5.8%
5
2.8%
-
Most abnormal investigation result during admission
Peak White cell count (x109/L)
10.3 (n=56)
6.9, 14.3
9.9 (n=170)
6.6, 14.9
0.739
Nadir Lymphocytes count (x109/L)
0.8 (n=54)
0.5, 1.0
0.7(n=172)
0.5, 1.0
0.694
Peak C-Reactive Protein (mg/L)
191 (n=54)
106, 315
165 (n=172)
79, 287
0.247
Disease severity on 7-point scale by day 28
0.379
1 - Not hospitalised with resumption of normal activities
11
16.2%
31
17.1%
2 - Not hospitalised but unable to resume normal
activities
34
50.0%%
102
56.4%
3 - Hospitalised not requiring supplemental oxygen
5
7.4%
5
2.8%
4 - Hospitalised requiring supplemental oxygen
1
1.5%
4
2.2%
5 - Hospitalised requiring nasal high-flow oxygen
therapy, non-invasive ventilation, or both
0
0.0%
0
0.0%
6 - Hospitalised, requiring ECMO, invasive mechanical
ventilation, or both
4
5.9%%
4
2.2%
7 – Death or palliation
13
19.9%
35
19.3%
PLWH with CD4 counts <200cells/µL & CD4 percentage < 14% (n=5) and HIV-negative matched comparators (n=12)
Death by day 28
0
0.0%
2
16.7%
0.515
Median length of hospitalisation (days)
11
10, 16
8.5
2, 22
0.460
Median time to improvement or discharge (days)
11
7, 19
7.5
2, 12
0.296
Patients requiring ITU/HDU level care
2
40.0%
6
50.0%
1.000
Patients requiring mechanical ventilation
2
40.0%
5
41.7%
1.000
Patients with a new requirement for renal replacement
therapy during admission
1
20.0%
1
8.3%
0.515
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
27
Table 5 – Sensitivity analyses considering the association between HIV status and time to improvement or discharge
Adjustments made for:
Change from primary
analysis
HR
PLWH vs
HIV-
95% CI
p-value
Primary analyses
Unadjusted
-
0.57
0.39, 0.85
0.005
Ethnicity, CFS, hypoxia, duration of
symptoms, number of comorbidities
-
0.66
0.42, 1.04
0.074
Ethnicity, CFS, hypoxia, days symptoms,
hypertension, cardiac, pulmonary, active
malignancy, diabetes, CKD
Consider presence of
specific comorbidities
0.70
0.43, 1.17
0.18
Sensitivity analyses
Ethnicity, CFS, number of comorbidities
Exclude COVID-related
factors (hypoxia,
duration of symptoms)
0.68
0.43, 1.06
0.088
CFS, hypoxia, days symptoms, number
of comorbidities
Exclude ethnicity
0.62
0.39, 0.96
0.031
Age*, Ethnicity, CFS, hypoxia, days
symptoms, number of comorbidities
Add age
0.67
0.42, 1.06
0.088
All results from Cox proportional hazards model, stratified by matching cluster and using multiple imputation with chained
equations to account for missing data (20 simulations, results combined using Rubin’s rules).
PLWH=people living with HIV; HIV-=HIV negative, HR=hazard ratio; CI=confidence interval; CFS=clinical frailty score; CKD=chronic
kidney disease
*Additionally adjusting for age to account for any residual confounding (clusters were age-matched to within 5 years)
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
28
Figures
Figure 1. – Time to clinical improvement or discharge
Figure 2 – Time to death
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3771328
