key: cord-0803247-2eukgq79
authors: Winston, Alan; De Francesco, Davide; Post, Frank; Boffito, Marta; Vera, Jaime; Williams, Ian; Anderson, Jane; Mallon, Patrick W.G.; Sabin, Caroline A.
title: Comorbidity indices in people with HIV and considerations for coronavirus disease 2019 outcomes
date: 2020-10-01
journal: AIDS
DOI: 10.1097/qad.0000000000002606
sha: 489df527e45908c0b587c6497071920cb905518e
doc_id: 803247
cord_uid: 2eukgq79

To determine comorbidity indices in people with HIV (PWH) and lifestyle-similar HIV-negative controls. DESIGN: Cross-sectional analysis of the Pharmacokinetic and clinical Observations in PeoPle over fiftY cohort study in the United Kingdom and Ireland. METHODS: The Elixhauser Comorbidity Index (ECI), Charlson Comorbidity Index and the Comorbidity Burden Index were compared between older PWH and HIV-negative controls using the Mann–Whitney U test; the magnitude of the difference between groups was quantified using the r effect size. RESULTS: The 699 PWH and 304 HIV-negative controls were predominantly male (87.5% vs. 64.0%), white (86.3% vs. 90.0%) and had median ages of 57 and 58 years, respectively. Among PWH, the median (interquartile range) CD4(+) T-cell count was 624 (475, 811) cells/μl; 98.7% were on antiretroviral therapy. The median (interquartile range) ECI was 0 (0, 8) and 0 (−3, 1), Charlson Comorbidity Index was 2 (1, 5) and 1 (0, 1) and Comorbidity Burden Index 8.6 (2.2, 16.8) and 5.9 (0.6, 10.8), respectively. While all three indices were significantly higher in PWH than in controls (P < 0.001 for each), the magnitude of the differences between the two groups were small to medium, with effect sizes (95% confidence interval) of 0.21 (0.16, 0.27), 0.38 (0.32, 0.42) and 0.18 (0.11, 0.23), respectively. CONCLUSION: These three comorbidity indices are higher in PWH compared with HIV-negative controls, although the magnitude of differences between groups were small. Differences in the ECI, reportedly associated with poorer coronavirus disease 2019 outcomes, were driven by more individuals with HIV being within the higher end of the range.

Mortality rates among people with coronavirus disease 2019 (COVID-19) are challenging to ascertain. For example, many cases may not be captured if individuals have only relatively mild symptoms where people may neither seek nor require medical attention, and cases may not be virologically confirmed with Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) infection. In virologically confirmed cases of COVID-19, individuals with multimorbidity and older age appear to have a higher risk of mortality [1] . Specific morbidities such as hypertension, diabetes and cardiovascular disease have all been associated with poor COVID-19 outcomes [2] .

Measurement of the presence of multiple comorbidities, or multimorbidity, in an individual can be undertaken via several classification systems. The Elixhauser Comorbidity Index (ECI) [3] and Charlson Comorbidity Index (CCI) [4] are two such classification systems which are significantly associated with both in-hospital and all-cause mortality [5] . A greater ECI has been reported to be associated with higher rates of referral for palliative care and in mortality for people hospitalized with COVID-19 [6, 7] .

Higher rates of noninfectious comorbidities have been reported in people with HIV (PWH) compared with lifestyle-similar individuals without HIV [8] . The presence of such comorbidities may theoretically place PWH at higher risk of morbidity and mortality from COVID-19. However, to date no studies have specifically reported higher rates of mortality from COVID-19 in PWH compared with mortality rates from COVID-19 in HIV-negative individuals. Our aim was to determine comorbidity indices in PWH and in lifestyle-similar HIVnegative controls and consider implications of any differences that may exist for the expected outcomes of COVID-19.

PWH and HIV-negative individuals aged at least 50 years recruited in the Pharmacokinetic and Clinical Observations in People Over Fifty (POPPY) study between April 2013 and January 2016 were included. As previously described in detail [9] , PWH were recruited from eight HIV outpatient clinics in the United Kingdom/Ireland; HIV-negative controls were selected from sexual health centres affiliated to the HIV clinics and were frequency matched on sex, ethnicity, sexual orientation and geographical location to the PWH.

Participants underwent a detailed assessment of comorbidities via a structured interview with trained staff who, where possible, also reviewed hospital notes to validate the presence of comorbidities [10] . Information provided regarding medical history as well as reasons for any healthcare utilisation and use of (nonantiretroviral) medication over the previous year were used to determine the presence of comorbidities to calculate the ECI [3] and the CCI [4] with scores for these indices obtained using weights proposed by van Walraven et al. [11] and Quan et al. [12] , respectively.

To assess a comorbidity index associated with patient reported outcomes, we utilised the comorbidity burden index (CBI) which is based on a list of 65 comorbidities (Supplementary material, http://links.lww.com/QAD/ B779) with weights empirically derived to reflect the effect of those comorbidities on physical health, measured using the physical health summary scale obtained from the Short Form Health Survey questionnaire [13, 14] , and quality of life. Weights were derived using a series of median regression models with physical health quality of life as outcome and each possible pair of comorbidities (one pair at a time) and their interaction as independent variables. The weight of each comorbidity to calculate the CBI was obtained by averaging all the standardised regression coefficients for that comorbidity. Furthermore, regression coefficients for interactions terms were converted into weights in a similar way [15] .

We compared the ECI (theoretical range: À19-89), CCI (theoretical range: 0-24) and CBI (theoretical range: À14-110) between PWH and HIV-negative controls using the Mann-Whitney U test and quantified the magnitude of the difference between groups using the r effect size (interpreted as small if r < 0.3, medium if r between 0.3 and 0.5, and large if r > 0.5).

A total of 699 PWH and 304 HIV-negative controls were included (Table 1) . Participants were predominantly male (87.5% vs. 64.0% in PWH and controls, respectively), of white ethnicity (86.3% vs. 90.0%) and of homosexual or bisexual orientation (78.4% vs. 47.4%). Median (interquartile range: IQR) age was 57 (53, 62) years in PWH and 58 (53, 63) years in HIV-negative controls. Among PWH, the median (IQR) CD4 þ T-cell count was 624 (475, 811) cells/ml, 98.7% were on antiretroviral therapy and 92.1% had a suppressed viral load (<50 copies/ml).

Median (IQR) ECI was 0 (0, 8) in PWH and 0 (À3, 1) in HIV-negative controls (Table 1 and Fig. 1 ). While the median ECI in PWH and HIV-negative controls was similar, overall, PWH had a greater ECI compared with controls (P < 0.001) with a small effect size r [95% confidence interval (CI)] ¼ 0.21 (0.16, 0.27). CCI also differed significantly in the two groups (P < 0.001).

Median (IQR) CCI was 2 (1, 5) and 1 (0, 1) in PWH and controls, respectively, with a medium effect size r (95% CI) ¼ 0.38 (0.32, 0.42). PWH reported a significantly higher CBI [median (IQR) ¼ 8.6 (2.2, 16.8)] compared with HIV-negative controls [median (IQR) ¼ 5.9 (0.6, 10.8), P < 0.001], effect size r (95% CI) was small: 0.18 (0.11, 0.23).

We assessed three comorbidity indices in PWH and controls and observed these indices to be higher in PWH when compared with our control population. The ECI has been associated with poorer COVID-19 outcomes. Of interest, although the ECI differed between PWH and our control population, median values were similar with the difference between the groups being driven by the range of values, with more individuals with HIV being within the higher end of the range. So far, little is known regarding the pathogenesis and clinical outcomes of COVID-19 in PWH. To date, available data are limited to case-reports or case-series and a higher incidence of COVID-19 in PWH or higher rates of mortality in PWH [16] have not been observed.

Assessment of potential factors for outcomes of COVID-19 in PWH include factors with both positive and negative considerations. Some factors may be associated with positive outcomes in PWH. With regards to incidence of COVID-19, PWH may carry out social distancing and/or isolation more strictly and at an earlier time-point in any given regional epidemic, as individuals with PWH are aware they have an underlying preexisting clinical condition. This could in certain settings reduce the number of PWH acquiring SARS-CoV-2. With regards to outcomes of COVID-19, in settings with widespread access to antiretroviral therapy, the majority of PWH are on virologically suppressive antiretroviral therapy and have CD4 þ lymphocyte counts within normal ranges and therefore, may not be at higher risk of more severe COVID-19 [17] . Lastly, some antiretroviral agents to treat HIV may have intrinsic antiviral activity against SARS-CoV-2. Importantly, lopinavir/ritonavir is noted to have antiviral activity against SARS-CoV-2 in laboratory models and could theoretically work as prevention for SARS-CoV-2 acquisition or have beneficial effects on disease outcomes. Ongoing studies are investigating these possibilities however no positive results have been reported thus far [18] .

Conversely, several theoretical factors could be associated with poorer outcomes of COVID-19 in PWH. Although, in general, PWH undergo immune-restoration with antiretroviral therapy, it remains unknown whether immune system function returns to levels in HIV-negative individuals and the possibility remains that individuals with low-nadir CD4 þ lymphocyte counts and other historical legacy effects of HIV-infection may be at greater risk of adverse outcomes from COVID-19. In addition, given PWH have higher rates of noninfectious comorbidities and higher comorbidity indices as we have observed, a greater risk of adverse outcomes from COVID-19 could be expected in PWH.

Although the comorbidity indices we have measured are greater in PWH compared with our controls, the differences were generally driven by a skewed distribution in a subset of PWH. Because of regular clinical follow-up, the capture of the presence of comorbidities is likely to be greater and more accurate in PWH compared with our control population. Although both PWH and controls in our study are part of a cohort study specifically designed to capture comorbidity details, biases are likely to be present. In general, PWH have clinical monitoring undertaken on at least a 6-monthly basis. This monitoring may ascertain the presence of noninfectious comorbidities more accurately when compared with control populations, the majority of whom would not be attending for regular healthcare monitoring. Therefore, the presence of comorbidities and the comorbidity indices we have calculated could be a truer reflection in PWH and may be an underestimate in our control population.

We have described three comorbidity indices in PWH and controls and highlighted their potential importance. Further conclusions are currently limited by a lack of data on COVID-19 incidence or outcomes within our population. Ongoing monitoring and careful management of comorbidities in PWH are essential with vigilance on outcome measures of COVID-19 among PWH over time needed.

Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study

Clinical and laboratory predictors of in-hospital mortality in patients with COVID-19: a cohort study in Wuhan, China

Comorbidity measures for use with administrative data

A new method of classifying prognostic comorbidity in longitudinal studies: development and validation

Prediction of survival of critically ill patients by admission comorbidity

Characteristics, symptom management and outcomes of 101 patients with COVID-19 referred for hospital palliative care

Clinical characteristics and predictors of outcomes of hospitalised patients with COVID-19 in a London NHS Trust: a retrospective cohort study. London: School of Public Health, Imperial College London COVID-19

Cross-sectional comparison of the prevalence of age-associated comorbidities and their risk factors between HIV-infected and uninfected individuals: the AGEhIV cohort study

Cohort profile: the Pharmacokinetic and clinical Observations in PeoPle over fiftY (POPPY) study

Patterns of co-occurring comorbidities in people living with HIV

A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data

Updating and validating the Charlson Comorbidity Index and score for risk adjustment in hospital discharge abstracts using data from 6 countries

The MOS 36-item short-form health survey (SF-36): I. Conceptual framework and item selection

SF-36 physical and mental summary scales: a user's manual

Development and validation of a comorbidity index for people living with HIV and its ability to predict frailty and mortality

COVID-19 in patients with HIV: clinical case series

SARS-CoV-2 pandemic expanding in sub-Saharan Africa: considerations for COVID-19 in people living with HIV

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

We would like to thank all participants and staff involved in the POPPY study. Authors contributions: A.W., D.D.F. and C.A.S. contributed to the concept and design of this analysis and wrote the first draft of this article. D.D.F. and C.A.S. undertook the data analysis. All authors substantially contributed to the writing of the second and further drafts of this article and read and approved the final article.

There are no conflicts of interest.