Browse the corpus

Adjunctive Dexamethasone for Tuberculous Meningitis in HIV-Positive Adults. BACKGROUND: Adjunctive glucocorticoids are widely used to treat human immunodeficiency virus (HIV)-associated tuberculous meningitis despite limited data supporting their safety and efficacy. METHODS: We conducted a double-blind, randomized, placebo-controlled trial involving HIV-positive adults (≥18 years of age) with tuberculous meningitis in Vietnam and Indonesia. Participants were randomly assigned to receive a 6-to-8-week tapering course of either dexamethasone or placebo in addition to 12 months of antituberculosis chemotherapy. The primary end point was death from any cause during the 12 months after randomization. RESULTS: A total of 520 adults were randomly assigned to receive either dexamethasone (263 participants) or placebo (257 participants). The median age was 36 years; 255 of 520 participants (49.0%) had never received antiretroviral therapy, and 251 of 484 participants (51.9%) with available data had a baseline CD4 count of 50 cells per cubic millimeter or less. Six participants withdrew from the trial, and five were lost to follow-up. During the 12 months of follow-up, death occurred in 116 of 263 participants (44.1%) in the dexamethasone group and in 126 of 257 participants (49.0%) in the placebo group (hazard ratio, 0.85; 95% confidence interval, 0.66 to 1.10; P = 0.22). Prespecified analyses did not reveal a subgroup that clearly benefited from dexamethasone. The incidence of secondary end-point events, including cases of immune reconstitution inflammatory syndrome during the first 6 months, was similar in the two trial groups. The numbers of participants with at least one serious adverse event were similar in the dexamethasone group (192 of 263 participants [73.0%]) and the placebo group (194 of 257 participants [75.5%]) (P = 0.52). CONCLUSIONS: Among HIV-positive adults with tuberculous meningitis, adjunctive dexamethasone, as compared with placebo, did not confer a benefit with respect to survival or any secondary end point. (Funded by the Wellcome Trust; ACT HIV ClinicalTrials.gov number, NCT03092817.).

The trial methodology, conduct, and analysis are described in the published protocol14 and analysis plan and available at nejm.org.15 The trial was designed and delivered by the investigators, supported by Oxford University Clinical Research Unit Clinical Trials Units in Vietnam and Indonesia. All authors vouch for the data and analysis. The trial was registered at ClinicalTrials.gov (NCT03092817). We recruited participants from the Hospital for Tropical Diseases, and Pham Ngoc Thach Hospital for Tuberculosis and Lung Disease, in Ho Chi Minh City, Vietnam, and from Dr. Cipto Mangunkusumo National Reference Hospital, and Persahabatan National Respiratory Referral Hospital, in Jakarta, Indonesia. Participants were ≥18 years old, HIV-positive (either newly or previously diagnosed) with a clinical diagnosis of tuberculous meningitis (at least 5 days of meningitis symptoms and cerebrospinal fluid abnormalities) with anti-tuberculosis chemotherapy either planned or started by the attending physician. Participants were subsequently classified as having definite, probable, or possible tuberculous meningitis, following published criteria (Supplementary table S1).16 Patients were ineligible if another brain infection was suspected, if >6 consecutive days of anti-tuberculosis chemotherapy or >3 consecutive days of systemic corticosteroids were received before enrolment, or if systemic corticosteroids were considered mandatory or contraindicated for any reason.

ementary table S1).16 Patients were ineligible if another brain infection was suspected, if >6 consecutive days of anti-tuberculosis chemotherapy or >3 consecutive days of systemic corticosteroids were received before enrolment, or if systemic corticosteroids were considered mandatory or contraindicated for any reason. Written informed consent to enter the trial was obtained from all participants or a relative if they were incapacitated. If capacity returned, consent from the participant was obtained. Trial approvals were obtained from local and national ethics and regulatory authorities in Vietnam and Indonesia and the Oxford Tropical Research Ethics Committee in the UK (Supplementary Text S1). An independent data monitoring committee reviewed data at 6-monthly intervals until 303 participants were randomized and annually thereafter. Randomization to dexamethasone or placebo was 1:1, with stratification by participating site and modified Medical Research Council (MRC) tuberculous meningitis severity grade.17 Participants in grade I had a Glasgow Coma Score of 15 (possible range, 3 to 15, with higher scores indicating better status) with no focal neurologic signs; grade II participants had a score of either 11 to 14, or 15 with focal neurological signs; and grade III participants had a score of 10 or less. The randomization list was computer-generated based on random permuted blocks with block size 4 and 6 (probability 0.75 and 0.25). Randomization of participants was performed by trained clinical staff using a web-based software, with 24-hour availability.

cal signs; and grade III participants had a score of 10 or less. The randomization list was computer-generated based on random permuted blocks with block size 4 and 6 (probability 0.75 and 0.25). Randomization of participants was performed by trained clinical staff using a web-based software, with 24-hour availability. All participants received standard-of-care anti-tuberculosis chemotherapy and antiretroviral therapy (ART) according to national guidelines (Supplementary Text S2). In ART-naive participants, ART was started approximately 6–8 weeks after starting anti-tuberculosis chemotherapy. All participants were randomized to dexamethasone or placebo, termed ‘study drug’, following the regimen previously shown to reduce tuberculous meningitis mortality (Supplementary Text S3).3 Blinded study drug packages (fully made-up and labelled treatment packs) contained either dexamethasone or identical placebo. All participants and investigators were blinded to study drug allocation. Participants with grade II or III disease received intravenous treatment for four weeks (0.4 mg per kilogram per day for week 1, 0.3 mg per kilogram per day for week 2, 0.2 mg per kilogram per day for week 3, and 0.1 mg per kilogram per day for week 4) and then oral treatment for four weeks, starting at 4 mg per day and decreasing by 1 mg each week. Participants with grade I disease received three weeks of intravenous therapy (0.3 mg per kilogram per day for week 1, 0.2 mg per kilogram per day for week 2, and 0.1 mg per kilogram per day for week 3) and then oral treatment for three weeks, starting at 3 mg per day and decreasing by 1 mg each week. Adherence to study medication was ensured with the use of supervised drug intake for inpatients, encouraged by detailed instructions at discharge, and medication compliance checks at follow-up visits or phone calls.

k 3) and then oral treatment for three weeks, starting at 3 mg per day and decreasing by 1 mg each week. Adherence to study medication was ensured with the use of supervised drug intake for inpatients, encouraged by detailed instructions at discharge, and medication compliance checks at follow-up visits or phone calls. The primary endpoint was all-cause death over 12 months from randomization. Secondary endpoints were neurological disability (modified Rankin score 3-5; Table S2) at 12 months, neurological immune reconstitution inflammatory syndrome (IRIS) over the first 6 months, and the following endpoints assessed over 12 months from randomization: first new neurological event or death, new acquired immunodeficiency syndrome (AIDS)-defining event or death, HIV-associated malignancy, use of open-label corticosteroid treatment for any reason, requirement for shunt surgery, and serious adverse events.15 Participants underwent trial clinical assessments at baseline, at days 3, 7, 10, 14, 21, and 30, and monthly until month 12. Assessment included Glasgow coma score, modified MRC severity grade, and details of clinical and adverse events and other interventions. Participants were monitored daily whilst in hospital and serious adverse events were reported to local and national regulators. In participants for whom systemic corticosteroids were considered necessary by the treating clinician after randomization, study drug was discontinued (with doses already received remaining blinded) and corticosteroids commenced.

lst in hospital and serious adverse events were reported to local and national regulators. In participants for whom systemic corticosteroids were considered necessary by the treating clinician after randomization, study drug was discontinued (with doses already received remaining blinded) and corticosteroids commenced. Baseline blood tests included full blood count, sodium, potassium, creatinine, alanine transaminase (ALT), bilirubin, hepatitis B and C, CD4 count and HIV viral load. Lumbar cerebrospinal fluid was sampled at baseline and tested for pyogenic bacteria (Gram stain and culture) and cryptococcal antigen. At least 6mls of cerebrospinal fluid (if available) was used for Ziehl-Neelsen smear microscopy, either Xpert MTB/RIF or Xpert MTB/RIF Ultra, and mycobacterial culture (mycobacteria growth indicator tube [MGIT]) following standard procedures.18 Phenotypic drug susceptibility testing was performed using a BACTEC MGIT SIRE kit (Becton, Dickinson; Franklin Lakes, NJ, USA).18 Based on previous trials,3,19 we assumed a target hazard ratio (HR) of 0.69, and a 40% 9-month mortality. We estimated that 520 HIV-positive patients with tuberculous meningitis would be required to achieve 80% power at a two-sided 5% significance level, and allowing for 5% loss-to-follow-up.

Baseline blood tests included full blood count, sodium, potassium, creatinine, alanine transaminase (ALT), bilirubin, hepatitis B and C, CD4 count and HIV viral load. Lumbar cerebrospinal fluid was sampled at baseline and tested for pyogenic bacteria (Gram stain and culture) and cryptococcal antigen. At least 6mls of cerebrospinal fluid (if available) was used for Ziehl-Neelsen smear microscopy, either Xpert MTB/RIF or Xpert MTB/RIF Ultra, and mycobacterial culture (mycobacteria growth indicator tube [MGIT]) following standard procedures.18 Phenotypic drug susceptibility testing was performed using a BACTEC MGIT SIRE kit (Becton, Dickinson; Franklin Lakes, NJ, USA).18 Based on previous trials,3,19 we assumed a target hazard ratio (HR) of 0.69, and a 40% 9-month mortality. We estimated that 520 HIV-positive patients with tuberculous meningitis would be required to achieve 80% power at a two-sided 5% significance level, and allowing for 5% loss-to-follow-up. Unless otherwise stated, the analysis followed a pre-specified published plan (Supplementary Text S4 and S5).15 Briefly, intention-to-treat and per-protocol analyses were performed for the primary and secondary endpoints. The intention-to-treat population included all randomized participants, even if no study drug was received. The per-protocol population included all randomized participants excluding those that subsequently did not meet the inclusion criteria or had exclusion criteria at enrolment, those with a final diagnosis other than tuberculous meningitis, those who received either <7 days of randomized study drug or <30 days of anti-tuberculosis chemotherapy for reasons other than death.

articipants excluding those that subsequently did not meet the inclusion criteria or had exclusion criteria at enrolment, those with a final diagnosis other than tuberculous meningitis, those who received either <7 days of randomized study drug or <30 days of anti-tuberculosis chemotherapy for reasons other than death. Baseline characteristics were summarized by treatment arm for intention-to-treat and per-protocol populations. The primary analysis was a Cox proportional hazards regression model with treatment as the only covariate.15 We tested for heterogeneity by six pre-specified subgroups: MRC severity grade, diagnostic category, leukotriene A4 hydrolase genotype (Supplementary Text S6), anti-tuberculosis drug resistance, ART status at enrolment, and CD4 count. We repeated analyses adjusted for MRC severity grade. Neurological disability at 12 months was compared using proportional odds logistic regression. All other secondary outcomes were compared via Cox proportional hazards models and cumulative event probabilities (nonparametric Kaplan-Meier, and Aalen-Johansen estimates in case death was a competing risk). No correction for multiple testing was made, and confidence intervals should not be interpreted as results from hypothesis tests. The number of individuals with any serious adverse event was compared by chi-squared test. Data were analyzed using the program R (version 4.1.1; R Core Team, 2021).20

We recruited participants from the Hospital for Tropical Diseases, and Pham Ngoc Thach Hospital for Tuberculosis and Lung Disease, in Ho Chi Minh City, Vietnam, and from Dr. Cipto Mangunkusumo National Reference Hospital, and Persahabatan National Respiratory Referral Hospital, in Jakarta, Indonesia. Participants were ≥18 years old, HIV-positive (either newly or previously diagnosed) with a clinical diagnosis of tuberculous meningitis (at least 5 days of meningitis symptoms and cerebrospinal fluid abnormalities) with anti-tuberculosis chemotherapy either planned or started by the attending physician. Participants were subsequently classified as having definite, probable, or possible tuberculous meningitis, following published criteria (Supplementary table S1).16 Patients were ineligible if another brain infection was suspected, if >6 consecutive days of anti-tuberculosis chemotherapy or >3 consecutive days of systemic corticosteroids were received before enrolment, or if systemic corticosteroids were considered mandatory or contraindicated for any reason.

Written informed consent to enter the trial was obtained from all participants or a relative if they were incapacitated. If capacity returned, consent from the participant was obtained. Trial approvals were obtained from local and national ethics and regulatory authorities in Vietnam and Indonesia and the Oxford Tropical Research Ethics Committee in the UK (Supplementary Text S1). An independent data monitoring committee reviewed data at 6-monthly intervals until 303 participants were randomized and annually thereafter.

Randomization to dexamethasone or placebo was 1:1, with stratification by participating site and modified Medical Research Council (MRC) tuberculous meningitis severity grade.17 Participants in grade I had a Glasgow Coma Score of 15 (possible range, 3 to 15, with higher scores indicating better status) with no focal neurologic signs; grade II participants had a score of either 11 to 14, or 15 with focal neurological signs; and grade III participants had a score of 10 or less. The randomization list was computer-generated based on random permuted blocks with block size 4 and 6 (probability 0.75 and 0.25). Randomization of participants was performed by trained clinical staff using a web-based software, with 24-hour availability.

All participants received standard-of-care anti-tuberculosis chemotherapy and antiretroviral therapy (ART) according to national guidelines (Supplementary Text S2). In ART-naive participants, ART was started approximately 6–8 weeks after starting anti-tuberculosis chemotherapy. All participants were randomized to dexamethasone or placebo, termed ‘study drug’, following the regimen previously shown to reduce tuberculous meningitis mortality (Supplementary Text S3).3 Blinded study drug packages (fully made-up and labelled treatment packs) contained either dexamethasone or identical placebo. All participants and investigators were blinded to study drug allocation. Participants with grade II or III disease received intravenous treatment for four weeks (0.4 mg per kilogram per day for week 1, 0.3 mg per kilogram per day for week 2, 0.2 mg per kilogram per day for week 3, and 0.1 mg per kilogram per day for week 4) and then oral treatment for four weeks, starting at 4 mg per day and decreasing by 1 mg each week. Participants with grade I disease received three weeks of intravenous therapy (0.3 mg per kilogram per day for week 1, 0.2 mg per kilogram per day for week 2, and 0.1 mg per kilogram per day for week 3) and then oral treatment for three weeks, starting at 3 mg per day and decreasing by 1 mg each week. Adherence to study medication was ensured with the use of supervised drug intake for inpatients, encouraged by detailed instructions at discharge, and medication compliance checks at follow-up visits or phone calls.

The primary endpoint was all-cause death over 12 months from randomization. Secondary endpoints were neurological disability (modified Rankin score 3-5; Table S2) at 12 months, neurological immune reconstitution inflammatory syndrome (IRIS) over the first 6 months, and the following endpoints assessed over 12 months from randomization: first new neurological event or death, new acquired immunodeficiency syndrome (AIDS)-defining event or death, HIV-associated malignancy, use of open-label corticosteroid treatment for any reason, requirement for shunt surgery, and serious adverse events.15

Participants underwent trial clinical assessments at baseline, at days 3, 7, 10, 14, 21, and 30, and monthly until month 12. Assessment included Glasgow coma score, modified MRC severity grade, and details of clinical and adverse events and other interventions. Participants were monitored daily whilst in hospital and serious adverse events were reported to local and national regulators. In participants for whom systemic corticosteroids were considered necessary by the treating clinician after randomization, study drug was discontinued (with doses already received remaining blinded) and corticosteroids commenced. Baseline blood tests included full blood count, sodium, potassium, creatinine, alanine transaminase (ALT), bilirubin, hepatitis B and C, CD4 count and HIV viral load. Lumbar cerebrospinal fluid was sampled at baseline and tested for pyogenic bacteria (Gram stain and culture) and cryptococcal antigen. At least 6mls of cerebrospinal fluid (if available) was used for Ziehl-Neelsen smear microscopy, either Xpert MTB/RIF or Xpert MTB/RIF Ultra, and mycobacterial culture (mycobacteria growth indicator tube [MGIT]) following standard procedures.18 Phenotypic drug susceptibility testing was performed using a BACTEC MGIT SIRE kit (Becton, Dickinson; Franklin Lakes, NJ, USA).18

Based on previous trials,3,19 we assumed a target hazard ratio (HR) of 0.69, and a 40% 9-month mortality. We estimated that 520 HIV-positive patients with tuberculous meningitis would be required to achieve 80% power at a two-sided 5% significance level, and allowing for 5% loss-to-follow-up. Unless otherwise stated, the analysis followed a pre-specified published plan (Supplementary Text S4 and S5).15 Briefly, intention-to-treat and per-protocol analyses were performed for the primary and secondary endpoints. The intention-to-treat population included all randomized participants, even if no study drug was received. The per-protocol population included all randomized participants excluding those that subsequently did not meet the inclusion criteria or had exclusion criteria at enrolment, those with a final diagnosis other than tuberculous meningitis, those who received either <7 days of randomized study drug or <30 days of anti-tuberculosis chemotherapy for reasons other than death.

From May 25, 2017, to April 29, 2021, 520 adults were randomly assigned to receive either dexamethasone or placebo. Eleven participants (2.1%) did not complete 12 months follow-up due to study withdrawal (six participants) or loss-to-follow-up (five participants). 33 participants were excluded for the per-protocol population (Figure 1). Participant characteristics at baseline were balanced between the study arms (Tables 1 and S3) and were broadly representative of populations of persons with tuberculous meningitis (Table S4). The median age of participants was 36 years (interquartile range [IQR] 30-41). Disease was generally mild or moderate (447/520 [86.0%] grade 1 or 2). 186/520 (35.8%) participants were newly diagnosed with HIV, 255/520 (49.0%) were ART-naïve, and 251/484 (51.9%) had CD4 counts ≤50 per mm3. The enrolment anti-tuberculosis chemotherapy regimens included rifampin in 478/514 (93.0%), isoniazid in 485/514 (94.4%), pyrazinamide in 471/514 (91.6%) and ethambutol in 364/514 (70.8%). Multi-drug resistance was identified in 16 participants (10 dexamethasone; 6 placebo), and treated following national guidelines. There were 116/263 (44.1%) observed deaths in the dexamethasone arm and 126/257 (49.0%) deaths in the placebo arm (HR 0.85, 95% confidence interval [CI] 0.66-1.10; P=0.22) (Figures 2A and 2B; Table 2). Similar results were observed when adjusted for MRC severity grade (Table S5). Heterogeneity of effect was not observed by any pre-specified sub-group in the intention-to-treat (Table 2, Figures S2-8) and the per-protocol populations (Tables S5-6; Figure S9).

terval [CI] 0.66-1.10; P=0.22) (Figures 2A and 2B; Table 2). Similar results were observed when adjusted for MRC severity grade (Table S5). Heterogeneity of effect was not observed by any pre-specified sub-group in the intention-to-treat (Table 2, Figures S2-8) and the per-protocol populations (Tables S5-6; Figure S9). Secondary outcomes occurred with similar frequency across both treatment arms in both the intention-to-treat and the per-protocol populations (Tables S7-8). Between the arms, there were similar frequencies of neurological disability (odds ratio, 1.31 95% CI 0.8-2.14) (Table S9), new neurological events or death (HR 0.85, 95% CI 0.67-1.08) and AIDS-defining events or death (HR 0.87, 95% CI 0.68-1.12). In ART-naïve participants, the median time to starting ART was 31 days (IQR 15-53) in 81 participants who received dexamethasone and 36 days (IQR 16-61) in 66 participants given placebo. Neurological IRIS events occurred during study drug administration in 8/20 (40.0%), and in 11/263 (4.2%) given dexamethasone and 9/257 (3.5%) given placebo (HR 1.11, 95% CI 0.46-2.69). Open-label corticosteroids were prescribed to 70/263 (26.6%) and 68/257 (26.5%) in the dexamethasone and placebo arms respectively, with reasons, and time until use of open-label corticosteroids, provided in Table S10. Time to use of open-label corticosteroid for any reason was similar between study arms (HR 0.97, 95% CI 0.69-1.35). No HIV-associated malignancies occurred by 12 months, and only one participant underwent ventriculoperitoneal shunt insertion.

ons, and time until use of open-label corticosteroids, provided in Table S10. Time to use of open-label corticosteroid for any reason was similar between study arms (HR 0.97, 95% CI 0.69-1.35). No HIV-associated malignancies occurred by 12 months, and only one participant underwent ventriculoperitoneal shunt insertion. No heterogeneity of effect was identified in the planned sub-group analysis of selected secondary outcomes (Tables S11-12, Figures S11-20), with the possible exception of a greater reduction of new neurological events and death associated with dexamethasone in the ‘possible’ tuberculous meningitis diagnosis sub-group of the intention-to-treat population (Table S11).

planned sub-group analysis of selected secondary outcomes (Tables S11-12, Figures S11-20), with the possible exception of a greater reduction of new neurological events and death associated with dexamethasone in the ‘possible’ tuberculous meningitis diagnosis sub-group of the intention-to-treat population (Table S11). The numbers of participants with at least one serious adverse event were similar between dexamethasone (192/263 [73.0%]) and placebo (194/257 [75.5%]) study arms (P=0.52) (Table 3; Tables S13-14). Fewer participants experienced serious neurological adverse events (95/263 [36%] in the dexamethasone arm than in the placebo arm (115/257 [45%]). These serious neurological adverse events (275 events in 210 participants) were predominantly depressed consciousness (N=149) and new focal neurological signs (N=45). Gastrointestinal bleeding events, and other adverse events possibly, probably, or definitely related to corticosteroids, occurred with similar frequency between the study arms (Table S15). Likewise, the frequency of clinical grade 3&4 adverse events (Table S16), adverse events leading to interruptions in anti-tuberculosis chemotherapy or ART (Table S17), and grade 3&4 laboratory abnormalities were similar between study arms (Table S18), with the exception of episodes of high ALT that were more frequent in the dexamethasone (36/263 [13.7%]) than the placebo arm (20/257 [7.8%])).

From May 25, 2017, to April 29, 2021, 520 adults were randomly assigned to receive either dexamethasone or placebo. Eleven participants (2.1%) did not complete 12 months follow-up due to study withdrawal (six participants) or loss-to-follow-up (five participants). 33 participants were excluded for the per-protocol population (Figure 1).

Participant characteristics at baseline were balanced between the study arms (Tables 1 and S3) and were broadly representative of populations of persons with tuberculous meningitis (Table S4). The median age of participants was 36 years (interquartile range [IQR] 30-41). Disease was generally mild or moderate (447/520 [86.0%] grade 1 or 2). 186/520 (35.8%) participants were newly diagnosed with HIV, 255/520 (49.0%) were ART-naïve, and 251/484 (51.9%) had CD4 counts ≤50 per mm3. The enrolment anti-tuberculosis chemotherapy regimens included rifampin in 478/514 (93.0%), isoniazid in 485/514 (94.4%), pyrazinamide in 471/514 (91.6%) and ethambutol in 364/514 (70.8%). Multi-drug resistance was identified in 16 participants (10 dexamethasone; 6 placebo), and treated following national guidelines.

There were 116/263 (44.1%) observed deaths in the dexamethasone arm and 126/257 (49.0%) deaths in the placebo arm (HR 0.85, 95% confidence interval [CI] 0.66-1.10; P=0.22) (Figures 2A and 2B; Table 2). Similar results were observed when adjusted for MRC severity grade (Table S5). Heterogeneity of effect was not observed by any pre-specified sub-group in the intention-to-treat (Table 2, Figures S2-8) and the per-protocol populations (Tables S5-6; Figure S9).

Secondary outcomes occurred with similar frequency across both treatment arms in both the intention-to-treat and the per-protocol populations (Tables S7-8). Between the arms, there were similar frequencies of neurological disability (odds ratio, 1.31 95% CI 0.8-2.14) (Table S9), new neurological events or death (HR 0.85, 95% CI 0.67-1.08) and AIDS-defining events or death (HR 0.87, 95% CI 0.68-1.12). In ART-naïve participants, the median time to starting ART was 31 days (IQR 15-53) in 81 participants who received dexamethasone and 36 days (IQR 16-61) in 66 participants given placebo. Neurological IRIS events occurred during study drug administration in 8/20 (40.0%), and in 11/263 (4.2%) given dexamethasone and 9/257 (3.5%) given placebo (HR 1.11, 95% CI 0.46-2.69). Open-label corticosteroids were prescribed to 70/263 (26.6%) and 68/257 (26.5%) in the dexamethasone and placebo arms respectively, with reasons, and time until use of open-label corticosteroids, provided in Table S10. Time to use of open-label corticosteroid for any reason was similar between study arms (HR 0.97, 95% CI 0.69-1.35). No HIV-associated malignancies occurred by 12 months, and only one participant underwent ventriculoperitoneal shunt insertion. No heterogeneity of effect was identified in the planned sub-group analysis of selected secondary outcomes (Tables S11-12, Figures S11-20), with the possible exception of a greater reduction of new neurological events and death associated with dexamethasone in the ‘possible’ tuberculous meningitis diagnosis sub-group of the intention-to-treat population (Table S11).

The numbers of participants with at least one serious adverse event were similar between dexamethasone (192/263 [73.0%]) and placebo (194/257 [75.5%]) study arms (P=0.52) (Table 3; Tables S13-14). Fewer participants experienced serious neurological adverse events (95/263 [36%] in the dexamethasone arm than in the placebo arm (115/257 [45%]). These serious neurological adverse events (275 events in 210 participants) were predominantly depressed consciousness (N=149) and new focal neurological signs (N=45). Gastrointestinal bleeding events, and other adverse events possibly, probably, or definitely related to corticosteroids, occurred with similar frequency between the study arms (Table S15). Likewise, the frequency of clinical grade 3&4 adverse events (Table S16), adverse events leading to interruptions in anti-tuberculosis chemotherapy or ART (Table S17), and grade 3&4 laboratory abnormalities were similar between study arms (Table S18), with the exception of episodes of high ALT that were more frequent in the dexamethasone (36/263 [13.7%]) than the placebo arm (20/257 [7.8%])).

We were unable to confirm a benefit of adjunctive dexamethasone in HIV-positive adults with tuberculous meningitis on survival or any other pre-specified secondary endpoint. Planned subgroup analyses did not identify a sub-population that clearly benefited from dexamethasone. The frequency of serious adverse events was similar between the arms. In dexamethasone-treated HIV-negative adults with tuberculous meningitis, survival has been associated previously with elevated cerebrospinal fluid inflammatory cytokine concentrations,5 suggesting dexamethasone benefits patients with excessive intracerebral inflammation. Cerebrospinal inflammatory cytokine concentrations are higher in HIV-positive than HIV-negative individuals with tuberculous meningitis,21 yet our trial findings indicate little benefit of adjunctive dexamethasone on survival in this important population. These observations suggest intracerebral inflammation in HIV-associated tuberculous meningitis may be qualitatively different, or the mechanisms leading to death are different, compared to HIV-negative individuals.

al findings indicate little benefit of adjunctive dexamethasone on survival in this important population. These observations suggest intracerebral inflammation in HIV-associated tuberculous meningitis may be qualitatively different, or the mechanisms leading to death are different, compared to HIV-negative individuals. Our study population was profoundly immunosuppressed, with 51.9% presenting with a CD4 count of ≤50 cells/mm3 and 49.0% ART-naïve. These individuals are at risk of developing other opportunistic infections, which might alter the impact of dexamethasone on outcome and increase the risk of adverse events. In addition, many participants were at risk of IRIS after starting ART, the neurological inflammatory complications of which can be fatal or disabling.22 Corticosteroids were previously shown to prevent IRIS in non-neurological tuberculosis,23 but we found dexamethasone did not reduce IRIS incidence or the timing or need for open-label corticosteroid treatment after randomization. However, dexamethasone doses were usually low by the time ART started (median 33 days after randomization), which may have reduced dexamethasone’s ability to prevent IRIS.

losis,23 but we found dexamethasone did not reduce IRIS incidence or the timing or need for open-label corticosteroid treatment after randomization. However, dexamethasone doses were usually low by the time ART started (median 33 days after randomization), which may have reduced dexamethasone’s ability to prevent IRIS. Our study has several limitations. First, we cannot exclude the possibility that a larger trial may have demonstrated a smaller mortality reduction than we hypothesized. The effect size observed (HR 0.85, 95% CI 0.66-1.10) was less than anticipated (HR 0.69), although was similar to the HIV-positive population in our previous trial (relative risk 0.86, 95% CI 0.52-1.41). Second, 138 participants (26.5%) were given open-label corticosteroids at some time during treatment, and although the frequency and timing of their use was similar between arms and initial treatment allocations remained masked, the additional corticosteroids may have obscured outcome differences between arms. Third, the findings may not generalise to better resourced settings, with patients with less advanced HIV and wider access to ventriculoperitoneal shunting for hydrocephalus, a common life-threatening complication of tuberculous meningitis.24

dditional corticosteroids may have obscured outcome differences between arms. Third, the findings may not generalise to better resourced settings, with patients with less advanced HIV and wider access to ventriculoperitoneal shunting for hydrocephalus, a common life-threatening complication of tuberculous meningitis.24 A strength of our trial is that it studies adjunctive corticosteroids exclusively in HIV-positive adults with tuberculous meningitis, increasing the available data from this important population by more than five times. The trial was pragmatic, enrolling all those with suspected tuberculous meningitis requiring anti-tuberculosis drugs and therefore generating findings of real-world relevance, where a high proportion of patients are treated for tuberculous meningitis without bacteriological confirmation. Lastly, the trial had high retention with only 5 participants lost to follow-up. Given the high mortality of HIV-associated tuberculous meningitis, and the frequency of inflammatory intracerebral complications, there is an ongoing need to explore alternative antiinflammatory strategies. These might include more targeted immunosuppression – against tumor necrosis factor α or interleukin-1, for example – which may be superior to corticosteroids for the prevention or treatment of inflammatory complications. Case reports and small case series have demonstrated a potential role for infliximab,25 thalidomide,26 and anakinra,27 in some tuberculous meningitis patients. Evaluation of these agents in clinical trials is required.

ich may be superior to corticosteroids for the prevention or treatment of inflammatory complications. Case reports and small case series have demonstrated a potential role for infliximab,25 thalidomide,26 and anakinra,27 in some tuberculous meningitis patients. Evaluation of these agents in clinical trials is required. In summary, we did not establish a benefit of adjunctive dexamethasone in HIV-positive adults with tuberculous meningitis on survival or any other secondary outcome over 12 months. The mortality associated with tuberculous meningitis in HIV-positive individuals remains unacceptably high, emphasising the global importance of enhanced detection and early treatment of HIV and tuberculosis.