Does adjunctive corticosteroid and aspirin therapy improve the outcome of tuberculous meningitis?
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.246278
Source of Support: None, Conflict of Interest: None
Keywords: Antiplatelet drugs, aspirin, mortality, steroid, stroke, tuberculous meningitis
The worldwide prevalence of tuberculosis is 159 per 100,000 population, and 35% of these patients reside in China and India. The prevalence of tuberculosis in India is 211/100,000 population, which amounts to approximately 2,532,000 patients in India. Ten percent of extrapulmonary tuberculosis involves the central nervous system (CNS), which has a higher mortality and morbidity. The clinical manifestations of tuberculous meningitis (TBM) are attributable to diverse pathological changes such as meningitis, encephalitis, hydrocephalus, tuberculoma, infarction, and arachnoiditis occurring at different time points in isolation or in various combinations. The treatment of central nervous system (CNS) tuberculosis is challenging and is compounded by limited CNS penetration of antituberculous drugs, paradoxical worsening, infarction, hydrocephalus, and above all, emerging drug resistance.,,, An unpredictable course and outcome of TBM are attributed to variations in the virulence of Mycobacterium tuberculosis and a variety of host factors such as the immune status of the patient and the varied inflammatory response. In TBM, stroke has been reported in up to 60% of the patients and is associated with a poor outcome.,, Stroke in TBM occurs because of involvement of the adventitia, media, and rarely intima, affecting small, medium, and large-sized vessels. Moreover, there may be stretching of blood vessels due to hydrocephalus and raised intracranial pressure (ICP); at later stages, blood vessels may be compressed by the presence of organized exudates. Corticosteroids have been used since long in this condition because of their anti-inflammatory and immune-modulatory effects.,, Thalidomide is a tumor necrotic factor (TNF) α antagonist and has been found to be effective in experimental models of TBM; however, human trials evaluating this drug had to be prematurely terminated because of its side effects. Aspirin was reported to reduce mortality in TBM in an open labeled trial. In children with TBM,100mg aspirin seemed to be effective in improving the outcome. Corticosteroids have been found to be effective in a randomized controlled trial (RCT) in reducing mortality. When the mortality and poor outcome were combined, there was, however, no significant benefit in using corticosteroids. There may be a group of TBM patients who respond well to corticosteroids while the others may not. There is no study comparing the role of corticosteroid and aspirin independently, and in combination, in patients with TBM. We have undertaken this study to evaluate the effect of adjunctive aspirin or aspirin plus corticosteroid in reducing death and in improving the outcome in patients suffering from TBM.
This is a retrospective analysis of TBM patients receiving adjunctive treatment; group I received aspirin alone, group II received a combination of both aspirin and prednisolone, and finally, group III was without these adjunctive treatments. The patients were selected from a prospectively maintained registry of TBM over the last 6 years. Some of these patients have been included in our earlier publications.,
TBM was diagnosed on the basis of the characteristic clinical features, computed tomography (CT) or magnetic resonance imaging (MRI) findings, and cerebrospinal fluid (CSF) examination
a. The essential criteria included:
1. The symptoms of meningitis: fever, headache, and vomiting for 2 weeks or more in patients in whom malaria as well as septic and fungal meningitis were excluded.
b. Supportive criteria:
1. Cranial imaging (CT/MRI) evidence of exudate, hydrocephalus, infarction, or tuberculoma in isolation or in combination
2. Cerebrospinal fluid revealing lymphocytic pleocytosis, raised protein, and a sterile culture for bacteria and fungus
3. Evidence of extra CNS tuberculosis.
The patient was considered to be suffering from TBM if there was presence of essential criteria with any two of the supportive criteria. The presence of acid-fast bacilli (AFB) in the cerebrospinal fluid (CSF) smear or culture, or detectable by polymerase chain reaction (PCR), and/or the presence of IgM antibody against the bacterium was considered as definite evidence of TBM.
Patients with prior antituberculous treatment for 4 weeks or more, before their enrollment in the study, or the presence of acid peptic disease, gastric hemorrhage, bleeding diathesis, liver, kidney, or heart failure, pregnancy, malignancy, organ transplantation, human immunodeficiency virus (HIV) infection, or the history of already being on treatment with corticosteroid or antiplatelet therapy, or having a drug allergy, were excluded.
The demographic details of the patients were recorded. The duration of illness, headache, vomiting, fever, seizures, diplopia, visual loss, and focal weakness were noted. A history of tuberculosis and HIV status were recorded. The presence of altered sensorium was noted, and the level of consciousness was assessed by the Glasgow Coma Scale (GCS). The presence of papilloedema and cranial nerve palsy were noted. The muscle tone and tendon reflexes were categorized as normal, increased, or reduced. The muscle power was graded as being normal, or having partial, or complete weakness. Evidence of extra CNS tuberculosis, such as tuberculosis of the lymph node, bone, lung, and joint, were noted. Based on the radiological findings and the GCS score, the severity of meningitis was graded into stage I (meningitis only), stage II (meningitis with focal deficit or a Glasgow Coma Scale (GCS) score of 11–14), and stage III (meningitis with a GCS score <11).
The blood counts, hemoglobin, erythrocyte sedimentation rate, serum chemistry, and HIV serology were done. The electrocardiogram and radiograph of the chest were carried out. Lumbar CSF was analyzed for cell, protein, and glucose. A CSF smear was examined for bacteria and fungus, and the CSF was cultured for AFB, fungus, and other bacteria. CSF polymerase chain reaction (PCR) was done for AFB and the IgM enzyme linked immunosorbent assay (ELISA) was done for M. tuberculosis. Cranial MRI was done using 1.5-Tesla [T] (in patients enrolled before 2010) or a 3-T scanner (Signa GE medical system, Wisconsin, USA). T1, T2, fluid-attenuated inversion recovery (FLAIR), diffusion-weighted imaging (DWI), and T1 contrast images were obtained. In critically ill patients in whom the MRI scan was not possible, a CT scan was performed. The presence of exudates, hydrocephalus, infarction, and tuberculoma were noted. The cranial MRI/CT scan was repeated at 3 months or earlier, if indicated.
The patients were treated with standard antituberculous treatment –rifampicin 10 mg/kg (maximum of 450 mg/day), isoniazid 5 mg/kg (maximum 300 mg/day), pyrazinamide 25 mg/kg (maximum 1500 mg/day), and ethambutol 15 mg/kg (maximum 800 mg/day).
Prednisolone 0.5 mg/kg (maximum 40 mg/day) was prescribed for 1 month and then tapered in the next 4 weeks. Aspirin was given daily after food at a dose of 150 mg in adults, and the dose was adjusted in children. In unconscious patients, these drugs were administered through a nasogastric tube. Aspirin was continued for 6 months. The patients were given the required calories, fluid, and electrolytes. Seizures were controlled by antiepileptic drugs. Raised intracranial pressure was managed by carbonic acid anhydrase inhibitor or 20% mannitol 100 ml intravenously. Ventriculoperitoneal shunt was done in patients having hydrocephalus with deteriorating consciousness.
Categorization of the patients
The patients were categorized as those belonging to group I – aspirin, group II – aspirin plus prednisolone, and group III –without these adjunctive treatments.
The patients were followed up at a 3 month interval clinically and the MRI was also repeated. Death and functional outcome after 3 months of treatment were noted. The functional outcome was assessed on a 0–20-point Barthel index (BI) scale and was categorized as poor status (BI score <12), partial recovery (BI score = 12–19) and complete recovery (BI score = 20)., On repeat MRI, the occurrence of infarction, an increase in the preexisting lesions, and the appearance of new lesions were noted, and their presence was considered as paradoxical radiological worsening.
During the treatment, drug-induced hepatitis was noted, and in case it occurred, hepatotoxic antituberculous drugs were withdrawn and restarted when serum glutamic pyruvic transaminase (SGPT) levels reached below twice the normal value. The side effects of aspirin and prednisolone were also noted.
The baseline clinical, MRI, and laboratory parameters between group I, group II, and group III were compared using chi-square test for categorical variables and analysis of variance (ANOVA) for continuous variables with Tuckey's corrections. On chi-square test, if the variable was significant, multiple comparisons were done. The death, functional outcome, and paradoxical response among the three treatment arms were also compared by the chi-square test, followed by multiple comparisons. The predictors of mortality were evaluated using multivariate logistic regression analysis including the variables with the lowest P value on univariate analysis. The outcome at 3 months was also evaluated using logistic regression analysis, and the patients were categorized as having a poor (poor recovery and death) or a good outcome (partial and complete recovery). Kaplan–Meier survival estimate was used to display the survival of patients at discharge and then at 3 months in the three treatment arms. The relative risk of death in the three treatment arms was evaluated by Cox regression analysis. Statistical analysis was done using the Statistical Package for the Social Sciences (SPSS), version 16 software and GraphPad prism 5. A variable having a two-tailed P value of <0.05 was considered significant.
One hundred and thirty-five patients with TBM were included, whose median age was 32 years and 67 (49.6%) of them were female patients. The duration of illness ranged between 16 and 122 (median, 45) days. The majority of patients were in stage II [74 (54.8%)] and stage III [21 (15.6%)], and the remaining 40 (29.6%) were in stage I meningitis. Seizures were present in 45 (36%) patients, which were focal in 17 and generalized in 28. Focal weakness was present in 39 (28.9%) patients, and included hemiparesis in 30 patients, paraparesis in 7 patients, and quadriparesis and monoparesis in 1 patient each. Cranial nerves were involved in 49 (36.3%) patients, and included optic nerve involvement in 18, external ophthalmoplegia in 39, and isolated sixth nerve palsy in 20 patients. The median GCS score was 13 (range: 3–15). Consciousness was impaired in 98 (57%) patients and 10 had a GCS score <8.
The CSF findings were abnormal in all the patients; the median CSF protein level was 124 mg/dl (range, 10–495), the median sugar level was 39 mg/dl (range, 5–228), and the cell count was 125/mm3 (range, 5–1500). Magnetic resonance imaging (MRI) was done in 115 patients, computed tomographic (CT) scan in 38 patients, and either CT or MRI was performed in 134 patients. The CT/MRI findings revealed hydrocephalus in 51 (38.1%), exudates in 28 (20.9%), tuberculoma in 57 (42.5%), and infarctions in 44 (32.8%) patients.
Comparison of treatment arms
There were 44 patients in group I, 50 in group II, and 41 in group III. The age, gender, duration of illness, GCS score, seizures, cranial nerve palsy, and CSF pleocytosis were not significantly different in groups I, II, and III. There was no difference in the severity of meningitis in groups I and III, but group II patients had more severe meningitis compared to groups I and III (P = 0.002). Focal weakness was also more frequent in group II (40%) compared to group I (25%) and group III (19.5%), but it was not statistically significant (P = 0.08). The CSF protein level was higher in group III compared to group I (P < 0.001) but there was no difference between group I and group III. The details are summarized in [Table 1].
At 3 months, 31 (23%) patients died; 8 (18.2%) in group I, 9 (18%) in group II, and 14 (34.1%) in group III. Patients without adjunctive therapy (group III) had a higher mortality whereas those on aspirin with corticosteroids had lower death rates despite having more severe meningitis. The difference, however, was not statistically significant. The absolute risk reduction in group II was 16.1 and that in group I was 15.9. The number needed to treat was 6.2 each in group II and group I, respectively.
Out of 31 deaths, 24 patients died in the hospital and 7 died after discharge. The causes of death were raised intracranial pressure and brain herniation in 11, sepsis in 13, and undetermined in 7 patients. One hundred and four (77%) patients were followed up for 3 months; 38 (36.5%) had complete, 26 (25%) had partial, and 40 (38.5%) had a poor recovery. The proportion of patients with complete recovery was higher in group II (40%) compared to group I (25%) and group III (17.1%). The functional outcome, however, was not significantly different in the three groups (P = 0.09). The details of the outcome are presented in [Table 2]. On multivariate analysis, death was related to the GCS score (OR, 0.80; 95% CI,0.64–1.00; P = 0.05) after adjustment of the stage of meningitis and treatment arms. On Cox regression analysis, the survival at 3 months was better in group II after adjusting for the GCS score and the stage of TBM (HR, 1.55;95% CI, 0.96–2.49; P = 0.07) [Figure 1]. The cumulative survival at 3 months was better in the group II (combined group) followed by group I (aspirin), and group III (no adjunctive treatment). On binary multivariate logistic regression analysis, a poor outcome at 3 months was related to the admission GCS score (OR, 0.77; 95% CI, 0.62–0.96; P = 0.02] after adjusting for the stage of TBM and the adjunctive treatments given.
Thirty-one patients had epigastric pain and/or vomiting, 2 had gastric hemorrhage, and 43 had drug-induced hepatitis. None of the patients had renal impairment, as evidenced by their serum creatinine level. Aspirin was withdrawn in both of those patients who had a gastric hemorrhage. Drug-induced hepatitis was more common in group II (44%) compared to group I (22.7%) and group III (26.8%) patients. In patients with drug-induced hepatitis, the hepatotoxic ATT was stopped till the liver function tests became normal, and thereafter, ATT was reintroduced sequentially. The above-mentioned side effects were not significantly different in the 3 groups [Table 3].
In the present study, survival was insignificantly better in group II and group I compared to group III. A higher proportion of patients recovered completely in group II (40%) compared to group I (25%) and group III (17.1%). These effects occurred despite more severe meningitis in group II, and may have been due to an additive benefit of using aspirin and corticosteroid as adjunctive therapy in TBM. This is the first report comparing the effect in patients with TBM, of administering with ATT, with either aspirin alone, or with aspirin plus prednisolone, and comparing the effect of these medications with the effects obtained in a group of patients where no adjunctive therapies were given. The role of corticosteroids in TBM has been evaluated in seven trials. In a class I study from Vietnam, 545 adult patients with TBM were randomly assigned to dexamethasone and a placebo. At the end of 6 months, there was a significant reduction in mortality in the dexamethasone group compared to the placebo group (31.8% vs 41.3%). However, on combining the numbers of patients who either died or were disabled, no significant benefit (44.2% vs 49.5%) of the dexamethasone therapy could be ascertained. In a Cochrane review, six out of seven trials favored the fact that the administration of corticosteroids resulted in reduction of mortality,,,,,, and only one trial reported a reduced mortality in the control group. The pooled analysis of these trials revealed a 22% risk reduction in death, a 10% absolute risk reduction, and the number-needed-to-treat (NNT) was 10. In three trials, both death and neurological disability were reported.,,, The largest randomized controlled trial on dexamethasone, however, did not reveal any benefit (in terms of either death or disability) in the intention-to-treat analysis. Moreover, the frequency of stroke was not reduced by administration of dexamethasone.
In TBM, not only the presence of raised intracranial pressure, hydrocephalus, exudates, and tuberculoma, but infarction also influences the outcome. A prothrombotic state resulting in a high frequency of deep vein thrombosis has been reported in patients with pulmonary tuberculosis. In TBM, both increased procoagulant and decreased anticoagulant activities have been reported in 16 children with stage II and stage III meningitis. There was reduced level of protein S and fibrinolytic activity; increased factor VIII as well as plasminogen activator inhibitor; and, increased platelet count, which were more marked in the stage III than in the stage II TBM. These parameters normalized after 1 month of anti-tuberculous therapy. Aspirin has antiplatelet, anti-aggregation, anti-inflammatory, and antioxidant properties, and has been effective in the secondary prophylaxis of ischemic stroke as well as in the primary stroke prevention in women.,, The antithrombotic effect of aspirin is attributable to the inhibition of thromboxane A2, which is mediated by the acetylation and inactivation of cyclo-oxygenase. In an open-labeled randomized controlled study on 118 patients with TBM, death was significantly reduced in the aspirin group compared to controls (21.7% vs 43.4%). The absolute risk reduction following aspirin was 22%. The frequency of stroke at a three -month follow-up, though lower in the aspirin group (43.3% vs 24.2%), did not achieve statistical significance. In another trial on 146 children with TBM, aspirin, in dosages of 75 mg and 100 mg, was compared with a placebo. The study revealed no significant effect on death or disability. Ten percent of the patients developed hemiplegia during the follow-up period; 4% in the placebo, 9% in the low-dose, and none in the high-dose aspirin group. The group receiving high-dose aspirin had more severe meningitis and a higher proportion of hemiplegia on admission. Even then, the outcome was not worse than that found in the other groups suggesting that some beneficial effects of aspirin in TBM exist. In the present study, drug induced hepatotoxicity (DIH) was more common in group II (44%) patients compared to group I (22.7%) and group III (26.8%) patients. Group II patients also had more severe meningitis. The higher frequency of DIH in group II is unlikely due to aspirin or prednisolone, as DIH improved on withdrawal of hepatotoxic antituberculous drugs.
The present study has a number of limitations; it is based on a small sample size, has a retrospective design, and has a referral bias of a tertiary care center that may have received patients in a more severe and complicated stage of the disease. However, this study for the first time compared two adjunctive therapies, and also compared these therapies with a group that was devoid of adjunctive therapy. All patients were personally evaluated by the investigators and the clinical results were further confirmed by both CT/MRI and laboratory findings.
From this study, it can be concluded that corticosteroid and aspirin combination may have a favorable role in TBM by reducing death and disability. This finding needs to be unequivocally confirmed in an randomized controlled trial.
We thank Mr. Shakti Kumar for secretarial help.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3]