Neurocognitive Impairment Among Cryptococcal Meningitis Survivors in Uganda, a Prospective Cohort Study

Abstract

Cryptococcal meningitis stands as the primary cause of meningitis among adults with HIV, accounting for 19% of advanced HIV-related mortality [1, 2]. In Africa, where the burden of cryptococcal meningitis mortality is the highest, 10-week mortality approaches 30% in clinical trial settings, often exceeding 70% at 1 year in routine care [2–5]. Neurological abnormalities are highly prevalent at the time of cryptococcal meningitis diagnosis, with up to 27% of individuals exhibiting altered mental status, defined as a Glasgow Coma Scale score <15, which is independently associated with short- and long-term mortality risk [3, 6, 7]. Seizures occur in as many as 19% of cryptococcal meningitis cases and are associated with increased 10-week mortality [8]. Additionally, cranial nerve neuropathies, visual disturbances, and hearing loss are frequently observed manifestations of cryptococcal meningitis, often present at the time of diagnosis [9–13].

Among cryptococcal meningitis survivors, the extent of long-term neurological impairment and disability is not well characterized [4, 14]. Carlson et al. previously demonstrated that in HIV-associated cryptococcal meningitis, 89% of survivors exhibited impaired neurocognitive function at 1 month post-treatment [15]. The proportion of persons impaired decreased to 59% at 12 weeks and further to 41% at 1 year [15]. Although there was an overall improvement in global neurocognitive function over time, 20% of survivors still exhibited severe neurocognitive impairment at 1 year [15]. Notably, severe impairment persisted in executive and gross motor function, with some residual deficits observed in verbal learning and motor speed [15]. While global neurocognitive impairment improved over time, a substantial proportion of individuals continued to demonstrate varying degrees of impairment despite treatment for cryptococcal meningitis and optimization of antiretroviral therapy (ART) [15].

The current understanding of risk factors for the development of neurocognitive impairment in cryptococcal meningitis remains unclear. Carlson et al. were unable to identify any baseline clinical or laboratory characteristics associated with the development of neurocognitive impairment in an ART-naïve cohort of cryptococcal meningitis survivors [15]. Surprisingly, traditional indicators of poor prognosis in cryptococcal meningitis, such as altered mental status, raised intracranial pressure, high fungal burden, and low cerebrospinal fluid (CSF) white cell count, were not correlated with neurocognitive function [15]. We sought to better understand the impact of cryptococcal meningitis and identify underlying risk factors contributing to neurocognitive impairment in HIV-associated cryptococcal meningitis survivors.

METHODS

Study Design

We conducted a prospective cohort study nested within the Adjunctive Sertraline for the Treatment of HIV-Associated Cryptococcal Meningitis (ASTRO-CM) randomized clinical trial [16]. The ASTRO-CM trial was a double-blinded, placebo-controlled trial that enrolled participants presenting at Mulago National Referral Hospital, Kampala, and Mbarara Regional Referral Hospital, Mbarara, Uganda, between May 2015 and March 2017. Participants were ≥18 years of age, had HIV, and had been diagnosed with an index episode of cryptococcal meningitis based on a positive serum and CSF cryptococcal antigen (CrAg) lateral flow assay (IMMY, Norman, OK, USA). Participants were randomized to receive standard antifungal therapy with 0.7–1 mg/kg/d of amphotericin B deoxycholate and fluconazole 800 mg/d plus either adjunctive sertraline or a matched placebo. Sertraline was administered at 400 mg/d for 2 weeks, followed by 200 mg/d for 12 weeks, followed by a 3-week taper. Antiretroviral therapy was initiated at week 6 in ART-naïve individuals. Those already on ART at presentation were continued on their ART with adherence counseling and assessed at week 6 to see if they needed to be switched to second-line therapy per Ugandan HIV guidelines. Baseline is defined as the time of cryptococcal meningitis diagnosis. Participants were prospectively followed for 18 weeks.

Neurocognitive Assessment

Participants underwent a neurocognitive assessment at 12 weeks. The Neurocognitive Assessment consisted of a functional evaluation using the Karnofsky Performance Score [17], screening for depressive symptoms with the Center for Epidemiologic Studies Depression (CES-D) score [18–20], and screening for HIV dementia using the International HIV Dementia Scale (IHDS) [21]. A CES-D score of ≥16 indicated a higher likelihood of significant depressive symptoms, and an IHDS score of ≤10 indicated greater cognitive impairment. The neuropsychological testing consisted of a battery of 8 tests aimed at evaluating neurocognitive function [22], including the WHO-UCLA auditory verbal learning test total (immediate) and delayed recall (assessing verbal learning and memory) [23, 24]; verbal fluency (for language fluency) [25, 26]; grooved pegboard (for evaluating manual dexterity, fine motor control, and speed) [23, 25, 26]; finger tapping (for assessing motor speed) [26]; symbol digit modalities test (for measuring speed of processing and concentration) [27]; color trails 1 (for evaluating speed of processing and attention) [23, 25]; and color trails 2 (for assessing executive function) [23, 25] (Supplementary Table 1). The neurocognitive assessment was conducted by a trained study nurse in the participants' preferred language (English or Luganda). Some participants started the exam but were unable to complete all 8 of the neuropsychological tests. All available data were collected.

Biomarker Analysis

We analyzed the concentrations of cytokines and chemokines in CSF at baseline in a subgroup of participants. At the time of sample collection, CSF samples underwent centrifugation at 400 g for 4 minutes at 4°C, after which the supernatant was collected and stored at −80°C. Forty-five different cytokines and chemokines were measured using the multiplex Human Luminex Discovery Assay (R&D Systems, Minneapolis, MN, USA).

Statistical Analysis

We utilized data from 2 cohorts for imputation and standardization of neurocognitive scores. First, a reference cohort was established, composed of 496 participants with HIV-associated cryptococcal meningitis from our prior randomized clinical trials, all of whom underwent a neurocognitive assessment at 12 weeks [16, 28, 29]. The means and standard deviations for each of the 8 neuropsychological tests were computed using the reference cohort data. Second, data from a previously published cohort of 100 HIV-negative Ugandans were used to standardize the z-scores [30].

Neurocognitive scores were calculated for individuals who completed ≥5 of the 8 neuropsychological tests. Participants who completed at least 5 but fewer than 8 of the neuropsychological tests were assigned a score of 1.5 standard deviations below the mean of the reference cohort of HIV-associated cryptococcal meningitis Ugandans for the missing tests. Raw scores for each domain were standardized to the HIV-negative Ugandan cohort, without adjustment for age or education, to obtain a test-specific z-score. All z-scores were transformed so that a higher z-score reflected better performance. The z-scores for each cognitive domain were averaged to calculate the Quantitative Neurocognitive Performance Z-score (QNPZ-8). The resulting QNPZ-8 score represents the overall cognitive performance of the individual across the 8 domains. A QNPZ-8 score close to 0 indicates performance similar to the HIV-negative Ugandan cohort's average, while positive scores indicate above-average performance and negative scores indicate below-average performance. Participants were classified by QNPZ-8 score as having mild (QNPZ-8 ≥−1), moderate (−2 < QNPZ-8 < –1), or severe (QNPZ-8 ≤−2) neurocognitive impairment compared with the cohort of HIV-negative Ugandan adults. We compared differences in baseline demographic, clinical, and laboratory variables by impairment categories using chi-square and Kruskal-Wallis tests as appropriate. We compared CSF biomarkers (transformed on the log₂ scale to improve linearity) by impairment categories using linear regression and Kruskal-Wallis tests (to account for nonlinearity). We used linear regression to compare QNPZ-8 scores between subgroups and logistic regression to examine covariates associated with moderate or severe (QNPZ-8 <–1) and severe (QNPZ-8 <–2) impairment. Adjusted models accounted for the baseline covariates GCS <15, ART status, weight, and receipt of sertraline after enrollment. A Spearman correlation analysis was conducted to examine the relationship between IHDS and QNPZ-8. Statistical significance was assessed using a 2-tailed P value, with significance set at P < .05.

Study Participant Consent and Ethics

The ASTRO-CM trial had ethical approval from the Mulago Hospital Ethics Committee, Uganda National Drug Authority, Uganda National Council of Science and Technology, and the University of Minnesota institutional review board. All study participants provided written informed consent for participation in ASTRO-CM. In situations where participants had diminished capacity to consent, surrogate consent was obtained until capacity was regained and the study participant was able to provide informed consent themselves.

RESULTS

Between May 2015 and March 2017, a total of 460 individuals were enrolled in the ASTRO-CM trial. Among the 460 participants enrolled, 48% (219/460) died and were censored before week 12, while 3% (12/460) were alive but censored before week 12. Of the 229 people alive at week 12, 83% (191/229) underwent a neurocognitive assessment. Among those who underwent a neurocognitive assessment, 80% (152/229) completed ≥5 of the 8 neuropsychological tests, while 20% (39/229) completed <5 tests (Supplementary Table 2). The final analysis cohort consists of the 152 participants who completed ≥5 of the 8 neuropsychological tests.

The raw scores for the 8 neuropsychological tests are summarized by mean and standard deviation (Table 1, Figure 1). We imputed scores for each individual test with missing values as follows: auditory verbal learning test total (n = 1), auditory verbal learning recall (n = 1), verbal fluency (n = 0), grooved pegboard (n = 6), finger tapping (n = 3), symbol digit modalities test (n = 17), color trials 1 (n = 30), and color trials 2 (n = 35). The test-specific mean z-scores were in the mild impairment range (−1 ≤ z-score < 0) for verbal fluency and grooved pegboard; moderate impairment range (−2 < z-score < –1) for auditory verbal learning total and recall, finger tapping, symbol digit modalities test, and color trails 1; and severe impairment range (z-score ≤−2) for color trails 2. The overall QNPZ-8 score (mean [SD], –1.4 [0.82]; n = 152) denotes moderate neurocognitive impairment at 12 weeks among cryptococcal meningitis survivors.

Figure 1.

Neurocognitive Z-scores by individual domain score and overall QNPZ-8. Individual domain z-scores are represented as mean and 95% CI. Z-scores have been adjusted so that higher scores reflect better performance. QNPZ-8 score is represented as mean and 95% CI and a boxplot of the median and interquartile range. Abbreviations: AVLT, auditory verbal learning recall; QNPZ-8, composite quantitative neurocognitive performance score.

Open in new tabDownload slide

Among study participants, 37% (57/152) exhibited mild impairment (QNPZ-8 ≥−1), 37% (56/152) displayed moderate impairment (−2 < QNPZ-8 < –1), and 26% (39/152) demonstrated severe impairment (QNPZ-8 ≤−2). At 12 weeks, there was a notable difference in Karnofsky Score among QNPZ-8 impairment categories: individuals with mild impairment had a median Karnofsky Score (interquartile range [IQR]) of 80 (70–90), those with moderate impairment scored 80 (70–80), and those with severe impairment scored 70 (70–80; P < .001) (Table 1). There were no differences in CES-D score across QNPZ-8 impairment categories (P = .90). The IHDS score differed across QNPZ-8 impairment categories: Individuals with mild impairment had a median IHDS score (IQR) of 11 (9–12), those with moderate impairment had a median IHDS score (IQR) of 10 (8–11), and those with severe impairment had an IHDS score (IQR) of 8 (7–10; P < .001).

We assessed differences in baseline demographics by QNPZ-8 impairment categories (Table 2). We found no differences in age or gender across impairment categories. There was a significant difference in median (IQR) baseline weight between mild (56 [50–61] kg), moderate (55 [50–60] kg), and severe (50 [47–60] kg) impairment (P = .03). There were no differences in ART status or CD4 cell count across impairment categories. We found a statistically significant difference in baseline Glasgow Coma Scale score <15 between mild (21%), moderate (36%), and severe (49%) neurocognitive impairment at 12 weeks (P = .03). We found no differences in significant baseline clinical variables such as seizures, serum sodium, hemoglobin, and CSF characteristics including opening pressure, cryptococcal quantitative culture, and white cell count across QNPZ-8 impairment categories. The proportions randomized to receive sertraline were similar across the QNPZ-8 impairment categories (P = .46). We also generally found no significant differences in baseline CSF cytokines or chemokines across QNPZ-8 impairment categories (Supplementary Table 3).

Assessment of overall QNPZ-8 scores by subgroups revealed statistically significant differences based on education level, Glasgow Coma Scale, and IHDS in both unadjusted and adjusted models (Table 3). QNPZ-8 scores were not significantly different for age, biological sex, CD4 cell count, or CES-D score. Those with ≤7 years of education had lower QNPZ-8 scores (education ≤7 years: QNPZ-8 score, −1.7; 95% CI, −1.9 to −1.5; vs education >7 years: QNPZ-8 score, −1.0; 95% CI, −1.2 to −0.8; P < .001). With unadjusted logistic regression analyses, participants with ≤7 years of education were 6 times more likely to have moderate or severe impairment (odds ratio [OR], 6.13; 95% CI, 2.96–12.68; P < .001) and 3 times more likely to have severe impairment (OR, 3.17; 95% CI, 1.41–7.11; P = .005) compared with those with >7 years of education (Table 4). In the multivariable analyses, while education <7 years still trended toward higher odds of severe impairment, it was no longer statistically significant in the adjusted model.

Study participants with a baseline Glasgow Coma Scale score <15 had significantly lower 12-week QNPZ-8 scores compared with those with a Glasgow Coma Scale score of 15 (GCS <15: QNPZ-8 score, −1.65; 95% CI, −1.86 to −1.43; vs GCS = 15: QNPZ-8 score, −1.26; 95% CI, −1.42 to −1.10; P = .006) (Table 3). Furthermore, a Glasgow Coma Scale score <15 was associated with close to a 3-fold increase in the odds of moderate or severe neurocognitive impairment (OR, 2.61; 95% CI, 1.23–5.57; P = .013) and severe neurocognitive impairment (OR, 2.41; 95% CI, 1.14–5.09; P = .022) compared with those with a Glasgow Coma Scale score of 15 (Table 4). The increased risk of neurocognitive impairment in those with a GCS <15 remained statistically significant in multivariate analyses for both moderate or severe neurocognitive impairment and severe neurocognitive impairment.

We found that IHDS score was rank-correlated with QNPZ-8 score (r = 0.46; P < .0001). Individuals with an IHDS score ≤10 had lower mean QNPZ-8 scores (QNPZ-8 score mean, −1.7; 95% CI, −1.8 to −1.5) than those with an IHDS score >10 (QNPZ-8 score mean, −1.0; 95% CI, −1.1 to −0.8; P < .001) (Table 3). Individuals with an IHDS score ≤10 compared with those with an IHDS score ≥10 were 3 times more likely to have moderate or severe neurocognitive impairment (OR, 2.85; 95% CI, 1.44–5.65; P = .003) and 5 times more likely to have severe neurocognitive impairment (OR, 4.86; 95% CI, 1.89–12.50; P = .001) (Table 4). In multivariate analysis, the association between IHDS and impairment category remained statistically significant only for severe neurocognitive impairment.

As we did not determine neurocognitive performance until 12 weeks, we assessed differences in baseline demographics between individuals who were censored before 12 weeks and those who survived to 12 weeks for the neurocognitive assessment (Supplementary Table 4). Persons who were censored (ie, died) before 12 weeks were more likely to present at cryptococcal meningitis diagnosis with a Glasgow Coma Scale score <15, lower CD4 T-cell count, lower hemoglobin, lower sodium, lower weight and mid-upper arm circumference, higher CSF cryptococcal quantitative culture, and lower CSF white blood cell count compared with those who survived to 12 weeks for a neurocognitive assessment. Second, we assessed differences in baseline demographics between those who completed <5 of the neuropsychological tests and those who completed ≥5 tests (Supplementary Table 5). Individuals who completed <5 of the neuropsychological tests were more likely to present with a baseline Glasgow Coma Scale score <15 and have higher baseline CSF opening pressures. At 12 weeks, individuals who completed <5 neuropsychological tests compared with those who completed ≥5 tests had a lower median (IQR) Karnofsky score (55 [50–60] vs 80 [70–80]; P < .001), higher median CES-D score (22 [16–31] vs 13 [7–21]; P = .002) and lower median IHDS scores (7 [5–8] vs 10 [8–11]; P < .001) (Supplementary Table 6).

DISCUSSION

In HIV-associated cryptococcal meningitis, we found that survivors at 12 weeks exhibited moderate global neurocognitive impairment compared with HIV-negative Ugandan adults. Survivors demonstrated deficits across all neurocognitive domains, with executive function, attention, and motor speed showing the most pronounced impairments. Individuals with severe neurocognitive impairment at cryptococcal meningitis diagnosis more often presented with lower weight, a Glasgow Coma Scale score <15, and ≤7 years of education. Baseline factors such as age, sex, ART history, CD4 count, seizures, CSF opening pressure, CSF quantitative cryptococcal culture, and CSF white cell count did not significantly differ across neurocognitive impairment categories. At 12 weeks, lower QNPZ-8 scores were observed in individuals with ≤7 years of education, a Glasgow Coma Scale score <15, or an IHDS ≤10.

Education ≤7 years was independently associated with moderate or severe neurocognitive impairment. Education is a well-known factor influencing cognitive ability, with lower levels of education identified as a risk for the development of young-onset dementia, Alzheimer's disease, and vascular dementia [31, 32]. Similarly, in HIV infection, fewer years of education has been independently associated with the development of HIV-associated neurocognitive disorder [33]. We found that in HIV-associated cryptococcal meningitis, individuals with ≤7 years of education were 6 times more likely to have moderate neurocognitive impairment and 3 times more likely to have severe neurocognitive impairment at 12 weeks in the unadjusted model. Most neuropsychological tests utilize normative data adjusted for both age and education to ensure that test scores accurately reflect true cognitive abilities rather than educational background. This adjustment allows for more accurate evaluation of neurocognitive impairment across individuals with varying levels of educational backgrounds. Therefore, it is possible that our results may potentially overestimate neurocognitive impairment in our study population. That said, studies looking at HIV-associated neurocognitive disorder have shown that even with adjustment, lower education remains independently associated with the risk of developing HIV-associated neurocognitive disorder [33].

The relationship between education and cognitive performance is well documented; however, our findings highlight the significant role that education plays in the development of moderate to severe neurocognitive impairment following cryptococcal meningitis in people with HIV. Notably, the impact of education on neurocognitive impairment after cryptococcal meningitis has not been previously described. We hypothesize that education serves as a proxy for multiple factors that may contribute to lower neurocognitive scores in cryptococcal meningitis, particularly through its association with cognitive reserve. Cognitive reserve refers to the brain's ability to withstand neuropathological damage while still maintaining cognitive function [31, 33, 34]. This concept is particularly relevant in HIV infection, where individuals with lower cognitive reserve may be more susceptible to the development of neurocognitive impairment [34]. In the context of HIV-associated cryptococcal meningitis, individuals with lower levels of education may be more susceptible to the neurotoxic effects of infection due to reduced cognitive reserve. Second, lower levels of education might have impacted the quality of results obtained in individual neurophysiological tests. Among the entire cohort, of the 8 neuropsychological tests, assessments involving more complex cognitive tasks, like color trials 1 and 2, exhibited higher rates of missing scores, necessitating imputation. Lastly, lower educational attainment may be associated with lower socioeconomic status and barriers to accessing timely medical care. This could lead to delayed diagnosis, exacerbating the neurological impact of cryptococcal meningitis, potentially explaining the higher frequency of participants with severe neurocognitive impairment presenting with a Glasgow Coma Scale score <15 observed in our cohort.

Altered mental status, indicated by a Glasgow Coma Scale score <15, is consistently associated with higher mortality risk in cryptococcal meningitis [7]. However, the precise pathophysiology of altered mental status in cryptococcal meningitis is not well understood. Adjusted analyses revealed that a baseline Glasgow Coma Scale score <15 was associated with a 4-fold increased risk of severe neurocognitive impairment at 12 weeks. Additionally, individuals with baseline altered mental status were less likely to complete ≥5 of the neuropsychological tests, and among those who did complete ≥5 tests, a higher percentage exhibited severe impairment. Lofgren et al. previously identified disparities in CSF cytokines and chemokines among individuals presenting with altered mental status, suggesting a dysfunctional immune response as the pathogenesis of delirium rather than a greater burden of infection [35]. Moreover, in people with HIV, CSF inflammatory cytokines and chemokines have been found to increase with worsening neurocognitive impairment [36]. Additionally, after critical illness, individuals with delirium often experience long-term cognitive impairment across multiple cognitive domains, potentially due to neuronal apoptosis and brain atrophy [37]. Despite our study not detecting differences in baseline CSF cytokines and chemokines across neurocognitive impairment categories, it is plausible that altered mental status serves as an indicator of neuronal injury and death leading to severe neurocognitive impairment.

In addition to evaluating baseline clinical factors associated with neurocognitive impairment, we utilized the International HIV Dementia Scale (IHDS) to screen for neurocognitive impairment. The IHDS is a validated tool designed for non-neurologists to screen individuals at risk for HIV dementia. An IHDS score of ≤10 demonstrates 80% sensitivity and 55% specificity for identifying individuals with HIV-associated dementia, who would then require a comprehensive neuropsychological assessment for confirmation [21]. Importantly, the IHDS has not been validated for individuals with known or past central nervous system opportunistic infections [21]. In our study, we observed a strong correlation between lower IHDS score and worse neurocognitive impairment among cryptococcal meningitis survivors at 12 weeks. Among those who completed 5 or more of the neuropsychological tests, an IHDS ≤10 was associated with severe neurocognitive impairment (QNPZ-8 ≤−2). Of those who completed <5 neuropsychological tests and had an IHDS score done, all scored ≤10. Our findings demonstrate that the IHDS is a brief and easy-to-administer screening test that can effectively identify individuals with HIV-associated cryptococcal meningitis who are at higher risk of severe neurocognitive impairment.

In HIV-associated cryptococcal meningitis, Carlson et al. had previously reported a high prevalence of depressive symptomology (CES-D Score ≥16), with 73% exhibiting symptoms at 4 weeks [15]. At 12 weeks, we observed a significant decrease in the prevalence of depressive symptomology to 40%, which was distributed similarly across neurocognitive impairment groupings. We have previously shown that the administration of sertraline did not lead to a significant reduction in depression within the ASTRO-CM cohort [38]. However, among individuals who completed <5 of the neuropsychological tests at 12 weeks and had a CES-D assessment, 75% had depressive symptomology, with scores indicating moderate to severe depression. Although we did not find depressive symptomology to be associated with neurocognitive impairment at 12 weeks, those who were not able to complete the full neurocognitive assessment and were not included in the analyses more frequently displayed depressive symptomology. Depressive symptomology is very prevalent early after the diagnosis of cryptococcal meningitis, decreasing over time, likely secondary to both effective cryptococcal management and the start of ART. Screening for depressive symptomology is useful in identifying individuals who may benefit from more intensive interventions, such as targeted counseling, therapy, or pharmacological treatments, ultimately improving their overall well-being and treatment outcomes. However, whether any pharmacological therapy has benefits for persons with HIV living in Africa is unknown [39].

Our study has several limitations. First, 17% of our study participants who survived at 12 weeks did not complete the neurocognitive assessment. Those who did not complete the neurocognitive assessment had a lower Karnofsky score, higher CES-D score, and lower IHDS score, all indicative of more severe neurocognitive impairment. Thus, the estimates provided herein would likely be worse with inclusion of those 17%. Second, our neurocognitive scores were not adjusted for age and education, thus preventing us from discerning whether the scores might have been influenced by educational background. This is a change from our prior methodology [15, 40] due to the relatively small numbers in the HIV-negative control group within each age and education subgroup used to standardize the z-score [30]. Moreover, our analysis was confined to the 12-week time point, without access to precryptococcal meningitis neurocognitive scores. Given the lack of baseline QPNZ-8 scores, we cannot definitively link education to the development or progression of cognitive impairment over time. Furthermore, we were unable to ascertain the specific impact of HIV vs cryptococcal meningitis on neurocognitive impairment. Lastly, this cohort was conducted before flucytosine became available in Africa through the efforts of UNITAID [41]; thus participants received amphotericin B deoxycholate with fluconazole as induction therapy. Whether more potent induction therapies would improve neurocognitive outcomes is an open question for the future.

In conclusion, in HIV-associated cryptococcal meningitis, moderate global neurocognitive impairment is prevalent at 12 weeks, with deficits observed across all cognitive domains. Importantly, education and altered mental status emerged as significant independent predictors of severe neurocognitive impairment, highlighting their crucial role in cognitive outcomes following cryptococcal meningitis. Utilizing educational level or Glasgow Coma Scale score <15 at the time of cryptococcal meningitis diagnosis could help identify individuals at risk of severe neurocognitive impairment and allow for targeted interventions and support. Additionally, the IHDS proved a valuable, rapid, and easy-to-administer screening tool in identifying individuals at higher risk of severe neurocognitive impairment in cryptococcal meningitis. Furthermore, our study contributes to the understanding of depressive symptomology dynamics post-treatment, emphasizing the need for screening and tailored interventions to address mental health concerns in HIV-associated cryptococcal meningitis. Future studies are needed to better elucidate the complex interplay between HIV, cryptococcal meningitis, and neurocognitive outcomes, ultimately informing more effective interventions and improving outcomes.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Acknowledgments

Author contributions. Laura Nsangi and Mahsa Abassi were involved in conceptualization of the manuscript, data analysis and interpretation, and writing of the initial draft. Kathy Huppler Hullsiek and Biyue Dai were involved in the formal statistical analysis and reviewing and editing of the manuscript. Alice Namudde, Grace Menya, Kenneth Ssebambulidde, Lillian Tugume, Edwin Nuwagira, Joshua Rhein, Darlisha A. Williams, and Abdu K. Musubire were involved in data curation, manuscript review, and editing. David R. Boulware and David B. Meya were involved in supervision, project administration, and manuscript review and editing.

Financial support. This work was supported by the National Institute of Neurological Disorders and Stroke (K23NS122601 to M.A. and R01NS086312), the Fogarty International Center (R01NS086312, R25TW009345), the National Institute of Allergy and Infectious Diseases (T32AI055433), and the Northern/Pacific Universities Global Health Research Training Consortium, Fogarty International Center (R25TW009345).

Collaborators. Edward Mpoza, Reuben Kiggundu, Katelyn A. Pastick, Kenneth Ssebambulidde, Andrew Akampurira, Darlisha A. Williams, Ananta S. Bangdiwala, Abdu K. Musubire, Melanie R. Nicol, Cynthia Ahimbisibwe, Florence Kugonza, Carolyne Namuju, Kiiza K. Tadeo, Paul Kirumira, Michael Okirwoth, Tonny Luggya, Julian Kaboggoza, Eva Laker, Stewart Walukaga, Emily E. Evans, Anna Stadelman, Andrew G. Flynn, Ayako W. Fujita, Richard Kwizera, Sarah M. Lofgren, Fiona V. Cresswell, Bozena M. Morawski.

References

Author notes