Associations Between HIV and Severe Mpox in an Atlanta Cohort

Abstract

The global outbreak of mpox in 2022 was the largest in history to occur outside the continent of Africa [1, 2]. As of 1 July 2023, there have been 88 122 confirmed global mpox cases, with 148 confirmed deaths [3]. Of more than 30 000 total mpox cases occurring in the United States, approximately one-third were in the South. Most cases were reported between July and September 2022 [4].

During the 2022 mpox outbreak, an estimated 95% of cases globally occurred in men who have sex with men [5] and 40% of cases occurred in people with HIV (PWH). More severe mpox disease has been reported in PWH [6–9], particularly in those with lower CD4⁺ T-cell counts [9–14] and nonsuppressed human immunodeficiency virus (HIV) loads [10, 11, 15]. In the United States, PWH accounted for 38% of total mpox cases, but 94% of mpox case fatalities [4].

Mpox in the Southeastern United States disproportionately affected black individuals [16, 17]. While representing 14% of the total US population, black persons accounted for 33% of US mpox cases and 86% of deaths [4]. The Southeastern United States also has the highest rates of people with stage 3 HIV infection (AIDS) of any US region [18]. Black individuals with HIV who live in the southeastern United States are particularly vulnerable to mpox-related morbidity and mortality and are underrepresented in the current mpox literature.

Here, we describe the clinical characteristics and outcomes of persons with mpox across 3 health care systems in Atlanta, Georgia and identify individual and HIV-related factors associated with severe mpox, as defined by the Centers for Disease Control and Prevention (CDC) [19].

METHODS

Study Design and Population

We identified a cohort of all individuals diagnosed with mpox at 4 academic medical centers and their affiliated clinics in Atlanta, GA between 1 June 2022 and 7 October 2022. A mpox case was defined by a positive mpox virus (MPXV) polymerase chain reaction (PCR) swab from any anatomic site. Subjects were identified using laboratory result database records at each medical center. Data was manually collected from the electronic medical record using a standardized data abstraction tool. Each chart review extended through the conclusion of the individual's mpox disease course (recovery or death). This study was reviewed by the Emory University Institutional Review Board and determined to be exempt from review.

Covariates

Clinical characteristic variables included medical comorbidities, history of sexually transmitted infections (STIs), STIs diagnosed within 1 month before or after mpox diagnosis, whether an individual was screened for STIs within 1 month of their mpox diagnosis, suspected mpox exposure, and mpox vaccination status at the time of diagnosis. Among PWH, additional variables collected included the CD4⁺ T-cell count and HIV load closest to the date of mpox diagnosis, within the timeframe from 1 year prior to mpox diagnosis to 1 month after mpox diagnosis. For people without HIV (PWoH), additional variables collected included HIV preexposure prophylaxis (PrEP) usage and whether individuals not currently prescribed PrEP were referred to a PrEP clinic within 1 month of mpox diagnosis.

Data on mpox vaccination status was collected both from the available electronic medical records and from the statewide Georgia Immunization Registry database. Jynneos vaccine doses were available at 3 of the clinical sites included in this analysis starting on 27 July 2022, and at the fourth site starting on 4 August 2022.

Outcomes

The primary outcome was severe mpox, as defined by the US CDC definition: “when a patient has conditions such as hemorrhagic disease; a large number of lesions such that they are confluent; necrotic lesions; severe lymphadenopathy that can be necrotizing or obstructing (such as in airways); involvement of multiple organ systems and associated comorbidities (for example, pulmonary involvement with nodular lesions; sepsis; encephalitis; myocarditis; ocular or periorbital infections); or other conditions requiring hospitalization.” [19]. In this study, associated comorbidities included the following complications when they were attributed to mpox disease: adrenal insufficiency, acute kidney injury, bacteremia, colitis (not including proctitis), diabetic ketoacidosis, encephalitis, epididymo-orchitis, epiglottitis, gastrointestinal bleeding requiring blood transfusion, myocarditis, ocular involvement, osteomyelitis, pleural effusion, and pulmonary nodules. Sepsis, in the absence of one of the above conditions, was not included in the severe mpox composite outcome. Additional outcomes recorded included hospitalization, intensive care unit admission, hospital length of stay, and each individual criterion for severe mpox. The date of any prescribed antiviral treatments (tecovirimat, cidofovir, brincidofovir, vaccinia immunoglobulin) was also collected. Specific definitions used for each covariate and outcome can be found in Supplementary Table 1.

Statistical Analysis

Descriptive statistics were used to summarize data on demographics, clinical characteristics, and mpox outcomes. Univariate associations between predictors of interest and the outcome of severe mpox were examined using Fisher exact test for binary variables and 2-sample t tests for continuous variables (eg, age). Missing data were noted where applicable.

We conducted 2 multivariable logistic regression analyses. The first was built to estimate the association between HIV status and severe mpox in the entire cohort, controlling for covariates of interest that were selected on the basis of univariate analysis and clinical relevance. These included: age, race, gender identity, presence of non-HIV immunosuppression, health insurance status, housing status, and mpox vaccination status. The second model was built to include only PWH with known CD4⁺ T-cell count and HIV load, to estimate the association between HIV-related immunosuppression (measured by CD4⁺ T-cell count) and severe mpox. For this model, additional covariates of interest included HIV load and engagement in HIV care, with the latter defined as at least 1 HIV clinic visit in the 6 months prior to mpox diagnosis. Data analysis was performed using SAS/STAT software (SAS Institute, Inc, 2019, SAS/STAT 15.1 User's Guide.).

RESULTS

Population, Exposures, and Codiagnoses

Of 395 individuals diagnosed with mpox in the study period, the median age at diagnosis was 35 years (interquartile range [IQR], 30–41 years). There were 384 (97.2%) cisgender men, 10 (2.5%) transgender women, and 1 (0.3%) cisgender woman; 335 (84.8%) individuals identified as black, 18 (4.6%) as white, 3 (0.8%) as Asian, 2 (0.5%) as multiracial, 1 (0.3%) as American Indian/Alaskan Native, and 36 (9.1%) had no available data on race. Of the total, 21 (5.3%) were Hispanic/Latinx. Being sexually active with cisgender male partners was reported by 342 individuals; 86.6% of the total population and 96.1% of those for whom sexual activities were known. There were 30 (7.6%) unhoused individuals at the time of diagnosis. There were 229 (58.0%) individuals without health insurance (Table 1).

Of the total, 324 (82.0%) individuals were PWH, 66 (16.7%) were PWoH, and 5 (1.3%) had unknown HIV status. Of 242 PWH with a known CD4⁺ T-cell count at the time of mpox diagnosis, 20 (8.3%) had < 100 cells (c)/µL, 19 (7.9%) had 100–200 c/µL, 51 (21.1%) had 201–350 c/µL, and 152 (46.9%) had > 350 c/µL. Of 257 PWH with a known HIV load at the time of mpox diagnosis, 138 (42.6%) had < 50 copies/mL, 29 (9.0%) had 51–200 copies/mL, 12 (4.7%) had 200–1000 copies/mL, and 78 (30.4%) had > 1000 copies/mL. Of the 90 (35.0%) individuals with HIV load >200 copies/mL, 10 (11.1%) were new HIV diagnoses and of the 80 with previously known HIV, 33 (41.3%) were engaged in care with at least 1 clinic visit with a primary HIV provider in the 6 months preceding their mpox diagnosis. Non-HIV immunosuppressive comorbidities were present in 19 (4.8%) people and chronic dermatologic conditions in 7 (1.8%).

In this cohort, 332 (84.1%) individuals had not received any vaccination against mpox at the time of mpox diagnosis; 30 (7.6%) individuals had received 1 dose of the Jynneos (MVA-BN) vaccine prior to mpox diagnosis, with a median interval of 6 days between vaccination and mpox diagnosis (IQR, 4–12 days). One (0.3%) individual had completed the 2-dose Jynneos vaccination series, although this individual received the second Jynneos dose on the same date as his mpox diagnosis. No patients in this cohort had received the ACAM2000 vaccine. Vaccination status was unknown in 32 (8.1%) individuals.

In total, 340 (86.1%) individuals had a suspected mpox exposure documented in their medical record. Of those cases, in 309 (90.9%) sexual contact was reported as the suspected mode of transmission. In 20 (5.9%) cases, nonsexual close contact was the suspected mode of transmission.

One or more STI diagnoses were made within 1 month of mpox diagnosis in 124 (31.4%) individuals. This included 64 (16.2%) individuals diagnosed with syphilis, 35 (8.9%) gonorrhea, and 22 (5.6%) chlamydia. Screening for syphilis, gonorrhea, and chlamydia within 1 month of mpox diagnosis was performed in 73.7%, 71.4%, and 70.9% of individuals, respectively. Screening for all 3 was performed in 118 (54.6%) of 216 individuals who were diagnosed with mpox in an emergency department and 118 (73.8%) of 160 individuals diagnosed at an HIV clinic (Supplementary Table 2). A new diagnosis of HIV was made in 10 (2.5%) individuals. Screening for HIV within 1 month of mpox diagnosis was performed in 51.5% of PWoH. Swabs were sent for herpes simplex virus PCR testing at the time of mpox testing in 103 (26.1%) cases, and 13 (12.6%) of those tests were positive (Supplementary Table 6).

Among PWoH, 19 (28.8%) were using HIV PrEP at the time of mpox diagnosis. Among eligible PWoH not taking HIV PrEP, 9 (19.1%) were prescribed PrEP or referred for PrEP care within 1 month of mpox diagnosis.

Clinical Presentation

Mpox clinical presentation data are summarized in Supplementary Table 3. Rash was the most reported initial symptom (present on the day of initial symptom onset), occurring in 297 (75.2%) individuals, followed by fevers/chills in 178 (45.1%), lymphadenopathy in 88 (17.8%), rectal pain in 87 (17.6%), fatigue in 86 (17.4%), and sore throat in 77 (15.6%). The median duration of time from symptom onset to mpox testing was 5 days (IQR, 3–7 days). One or more health care visits for mpox symptoms, without mpox testing being performed, occurred in 91 (23.0%) individuals prior to the visit when mpox testing was conducted. At the time of mpox testing, the most common symptom reported by patients was rash in 366 (92.7%) individuals. In 103 (26.1%) individuals the rash was localized and in 248 (62.8%) the rash was disseminated (involving 2 or more body segments). Mucosal involvement was present at the time of mpox testing in 141 (35.7%) individuals, including anorectal involvement in 89 (22.5%) individuals, oral/pharyngeal involvement in 82 (20.8%), urethral involvement in 26 (6.6%), ocular involvement in 3 (0.8%), and nasal mucosal involvement in 2 (0.5%). Multiple mucosal sites were involved in 32 (8.1%) individuals.

Disease Course and Outcomes

Severe mpox occurred in 77 (19.5%) individuals; 6 (9.1%) PWoH, 31 (18.6%) PWH with HIV load ≤ 200 copies/mL and 39 (43.3%) PWH with HIV load > 200 copies/mL. Severe mpox occurred in 36 (23.7%) PWH with CD4⁺ > 350 c/µL, 14 (27.5%) PWH with CD4⁺ 201–350 c/µL, 4 (21.1%) PWH with CD4⁺ 100–200 c/µL, and 16 (80.0%) PWH with CD4⁺ < 100 c/µL. Hospitalization occurred in 66 (16.7%) individuals, including 6 (9.1%) PWoH, 22 (13.2%) PWH with HIV load ≤ 200 copies/mL, and 37 (41.1%) PWH with HIV load > 200 copies/mL. Mpox complications occurred in 106 (26.8%) individuals. The most common mpox complication was bacterial superinfection of mpox lesions (including cellulitis, soft tissue abscess, anorectal abscess, pharyngeal abscess, and bacteremia), which occurred in 6 (9.1%) PWoH, 11 (6.7%) PWH with HIV load ≤ 200 copies/mL, and 16 (17.8%) PWH with HIV load > 200 copies/mL. Gastrointestinal complications occurred in 1 (1.5%) PWoH, 7 (4.2%) PWH with HIV load ≤ 200 copies/mL, and 8 (8.9%) PWH with HIV load > 200 copies/mL. There was 1 death in the cohort, occurring in a PWH with HIV load > 200 copies/mL and CD4⁺ < 100 c/µL (Table 2 and Supplementary Table 7).

Antiviral treatment with tecovirimat was prescribed for 112 (28.4%) individuals, including 56 (72.7%) of the 77 individuals with severe mpox. Tecovirimat was prescribed to 3 (4.5%) PWoH, 55 (32.9%) PWH with HIV load ≤ 200 copies/mL, and 51 (56.7%) PWH with HIV load > 200 copies/mL. No other antiviral medications directed against MPXV were prescribed to any individuals in our cohort. Additional therapeutics for symptomatic management of mpox are summarized in Supplementary Table 4.

Multivariable Analysis

In multivariable analysis of the full cohort, older age, HIV infection, non-HIV immunosuppressive comorbidities, and unhoused status were associated with a higher incidence of severe mpox (Table 3). PWH had 2.52 (95% confidence interval [CI], 1.01–6.27) times higher odds of developing severe mpox compared to PWoH. Race, gender identity, health insurance status, and mpox vaccination status were not found to be significantly associated with severe mpox.

A second multivariable analysis included 242 PWH with known CD4⁺ T-cell count and HIV load. Older age showed a statistically significant but low-amplitude association with severe mpox (odds ratio [OR], 1.04; 95% CI, 1.00–1.08). PWH with HIV load > 200 copies/mL had 2.10 (95% CI, 1.00–4.39) times higher odds of severe mpox compared with those with HIV load ≤ 200 copies/mL. Not being engaged in HIV care was also associated with higher odds of severe mpox (OR, 3.47; 95% CI, 1.80–6.69). Lower CD4⁺ T-cell count showed a significant univariate association with severe mpox but was not found to be significantly associated with severe mpox in multivariable analysis (OR, 1.60; 95% CI, .68–3.79; Table 3).

DISCUSSION

We analyzed a cohort of 395 individuals in Atlanta, Georgia who were predominantly black and living with HIV and found that HIV-positive status, especially when HIV load was >200 copies/mL, was associated with severe mpox. Mpox complications were common in PWH with nonsuppressed HIV loads, but mortality was low in our cohort.

Our cohort describes a patient population distinct from those previously described in the mpox literature, with 82% of cases occurring in PWH and 85% occurring in black individuals and emphasizes the disproportionate burden the 2022 mpox outbreak had on these groups in the Southeastern United States. Mpox was also more common in black individuals in other urban centers in the Southeast [16, 17], a region where the mpox and HIV epidemics collided. In this region, new cases of HIV are also disproportionally reported in black individuals [18] and highlights how inequities in social determinants of health in the setting of ongoing institutionalized racism lead to barriers to prevention, diagnosis, and treatment of multiple infectious diseases.

Our findings add to previous reports that identified uncontrolled HIV as an important risk factor for the development of severe mpox disease [9–15]. We found that HIV infection, especially with a viral load >200 copies/mL, was associated with severe mpox. Although lower CD4⁺ T-cell counts were significantly associated with more severe mpox in univariate modeling, our multivariable model did not show a significant association between lower CD4⁺ T-cell counts and severe mpox when accounting for the effects of HIV load and other covariates. A multivariable analysis from a large international cohort of individuals with mpox found immunocompromised status, such as HIV with low CD4⁺ T-cell count, to be significantly associated with increased odds of hospitalization and death [9]. A Spanish mpox case series of 448 PWH found HIV viremia (> 1000 copies/mL) to be the only risk factor out of age, sex, HIV load, and CD4⁺ T-cell count (CD4⁺ >350 vs <350 c/µL) that was independently associated with severe mpox in a multivariable model [15].

Our findings suggest that the immune response to MPXV infection may be significantly impaired in the setting of nonsuppressed HIV viremia. HIV viremia impairs the function of multiple immune cell types, notably including CD4⁺ T cells [20], CD8⁺ T cells [21], B cells [22], and natural killer (NK) cells [23], each of which has been shown in animal models to play an important role in combatting orthopoxvirus pathogenesis [24–27]. Our findings are consistent with the principle that immune dysfunction in HIV extends beyond just decreases in T-cell counts. Additional studies will be needed to further understand the interplay of HIV, host immune function, and mpox pathogenesis. An aggressive approach to mpox monitoring and treatment is especially important in PWH with nonsuppressed HIV loads.

Unhoused status was associated with severe mpox in our multivariable analysis of the full cohort (analysis 1) but did not meet statistical significance in the PWH cohort (analysis 2). Our cohort included 30 (7.6%) unhoused individuals and is insufficient to draw conclusions on the effects of unhoused status on mpox disease severity, although it is reasonable to speculate that the barriers to health care access that are associated with experiencing homelessness likely favored worse outcomes. Unhoused persons in our cohort presented to care a median of 9 days from symptom onset, whereas housed individuals presented a median of 6 days from symptom onset. Further study is warranted to explore mechanisms by which experiencing homelessness may affect mpox presentation and disease course.

One individual died in this cohort, who was a PWH with high HIV load (1.2 million copies/mL) and low CD4⁺ T-cell count (81 [4%] c/µL). The SHARE-NET international cohort of PWH with CD4⁺ T-cell counts ≤ 350 c/µL reported a 7% overall mortality and a 15% mortality for subjects with CD4⁺ T-cell count <200 c/µL [10]. Our mpox case fatality rate was considerably lower both for CD4⁺ T-cell count ≤ 350 c/µL (1/90 = 1.1%) and CD4⁺ T-cell count <200 c/µL (1/39 = 2.6%). Potential explanations for the lower mortality in this cohort could include faster diagnosis, greater access to care, more aggressive supportive care, or higher use of antiviral treatment with tecovirimat (73% of individuals with severe mpox in our cohort received tecovirimat compared to 16% of individuals in the SHARE-NET cohort). While tecovirimat was the first-line antiviral treatment and was widely used in severe mpox cases in our cohort, randomized clinical trials demonstrating efficacy in mpox disease, such as the ACTG A5418 STOMP trial, are needed to inform clinical practice.

Our cohort demonstrated a high frequency of STI codiagnoses within 30 days of positive mpox testing. Screening rates for STIs were less than 75%. Among PwOH, only 52% were screened for HIV, with 10 (29%) of those screening tests returning positive. The high rate of coincident STIs demonstrates the need to protocolize universal STI screening at the time all individuals undergo testing for mpox, in addition to utilizing this time to discuss HIV PrEP and doxycycline postexposure prophylaxis (doxyPEP).

This study has several limitations. First, our method of data collection relied on information documented in the medical charts of individuals in our cohort, which may underestimate the prevalence of some reported findings. Patient evaluation and mpox testing practices occurred in a real-world setting and thus changed over time in response to the rapidly emerging mpox outbreak and may not have been standardized across all sites and time points. Second, electronic medical records were shared between several, but not all, regional health systems, so it is likely that some people lacked background or follow-up information if they sought care in outside health systems. Third, our cohort included a small number of individuals with CD4⁺ T-cell count <100 c/µL (n = 20), which may have affected the lack of an association observed between lower CD4⁺ T-cell counts and severe mpox in the multivariable model of PWH. There was a moderate degree of collinearity between HIV load >200 copies/mL and CD4⁺ T-cell count <200 c/µL which may have dampened the association of CD4⁺ T-cell count and severe mpox. Ongoing study of both HIV load and CD4⁺ T-cell count as predictors of mpox disease severity in PWH is warranted. Fourth, since our cohort included only people diagnosed with mpox, we are unable to comment on differences in exposures, medical history, HIV indices, mpox immunization status, and other factors which may have differed between our cohort population and individuals who did not develop mpox disease.

The 2022 mpox epidemic disproportionately affected people who are black and those with HIV in the Southeastern United States. People with HIV with nonsuppressed viral loads had more mpox complications, hospitalizations, and protracted disease courses than those without HIV or with HIV with suppressed viral loads in multivariable modeling. PWH with elevated HIV loads who are diagnosed with mpox warrant particularly aggressive monitoring and treatment. These findings also highlight the paramount importance of ongoing efforts to optimize the HIV care continuum to engage and retain PWH in care to achieve virologic suppression. In the Southeast United States, factors including stigma and institutionalized racism allowed the mpox and HIV epidemics to exploit the same social and structural fault lines and disproportionately harm vulnerable communities.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Acknowledgments. We acknowledge the full Emory Monkeypox Patient Analysis Series (MAPS) Study Group.

Author contributions. B. M. A. and V. D. C. contributed study design, chart reviews, data analysis and statistics, manuscript drafting and revision, and leadership/oversight. J. Y. S., E. J. C., and B. K. T. contributed study design, chart reviews, manuscript drafting and revision, and leadership/oversight. A. A., D. J. G., V. C. M., M. L. N., P. A. R., and A. K. contributed study design, chart reviews, and manuscript drafting and revision. A. M. A. contributed study design, chart reviews, data analysis and statistics, and manuscript drafting and revision. J. A. C., J. T. J., Z. W., and K. W. contributed study design, and manuscript drafting and revision. B. H. and B. S. performed chart reviews. S. K. contributed study design and chart reviews. C. F. K. and A. N. S. contributed study design, manuscript drafting and revision, and leadership/oversight. R. H. L. contributed data analysis and statistics, and manuscript drafting and revision.

Data availability. Raw data is not publicly available.

Financial support. This work was supported by the Emory Center for AIDS Research (grant number P30AI050409).

Supplement sponsorship. This article appears as part of the supplement “Mpox: Challenges and Opportunities Following the Global 2022 Outbreak,” sponsored by the Centers for Disease Control and Prevention (Atlanta, GA).

References

Author notes