摘要: Drugs were given by nasogastric tube in patients who were unable to swallow。 Resistance to at least isoniazid and rifampicin had a significant effect on outcome: all of the patients with MDR TBM died, and the median time to death was 12 days (range, 162 days)。 Resistance to isoniazid, with or w......
Oxford University Clinical Research Unit, Hospital for Tropical Diseases, Pham Ngoc Thach Hospital for Tuberculosis and Lung Disease
Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam
Centre for Tropical Medicine, Nuffield Department of Clinical Medicine, John Radcliffe Hospital, Oxford, United Kingdom
Tuberculous meningitis (TBM) caused by Mycobacterium tuberculosis resistant to 1 or more antituberculosis drugs is an increasingly common clinical problem, although the impact on outcome is uncertain.
We performed a prospective study of 180 Vietnamese adults admitted consecutively for TBM. M. tuberculosis was cultured from the cerebrospinal fluid (CSF) of all patients and was tested for susceptibility to first-line antituberculosis drugs. Presenting clinical features, time to CSF bacterial clearance, clinical response to treatment, and 9-month morbidity and mortality were compared between adults infected with susceptible and those infected with drug-resistant organisms.
Of 180 isolates, 72 (40.0%) were resistant to at least 1 antituberculosis drug, and 10 (5.6%) were resistant to at least isoniazid and rifampicin. Isoniazid and/or streptomycin resistance was associated with slower CSF bacterial clearance but not with any differences in clinical response or outcome. Combined isoniazid and rifampicin resistance was strongly predictive of death (relative risk of death, 11.63 [95% confidence interval, 5.2126.32]) and was independently associated with human immunodeficiency virus infection.
Isoniazid and/or streptomycin resistance probably has no detrimental effect on the outcome of TBM when patients are treated with first-line antituberculosis drugs, but combined isoniazid and rifampicin resistance is strongly predictive of death.
The World Health Organization recently estimated that, between 1999 and 2002, 10% of all clinical isolates of Mycobacterium tuberculosis were resistant to 1 or more first-line antituberculosis drugs . The impact of resistance on treatment outcome varies according to which drug (or combination of drugs) is resistant and reflects the different but complementary role each agent plays in the treatment of tuberculosis . For example, resistance to isoniazid only probably has minimal effect on the outcome of pulmonary tuberculosis as long as the regimen contains rifampicin, but resistance to both isoniazid and rifampicin dramatically reduces the likelihood of being cured of pulmonary tuberculosis . This suggests that rifampicin plays a critical role in short-course regimens for the treatment of pulmonary tuberculosis, but this may not be true for other forms of the disease. There is evidence to suggest that the relative contributions of each drug to a successful outcome may be different for tuberculous meningitis (TBM), the most lethal form of infection with M. tuberculosis. This may be because each drug differs in its ability to cross the blood-brain barrier and achieve sufficient intracerebral concentrations to kill bacilli. Isoniazid and pyrazinamide pass freely into the cerebrospinal fluid (CSF), whereas CSF concentrations of rifampicin are probably <10% of serum concentrations although CSF concentrations of rifampicin may correspond to that of the free, unbound fraction in plasma, and there is evidence to suggest that only the unbound fraction is active in the treatment of pulmonary tuberculosis . Streptomycin and ethambutol, the other first-line treatment agents, probably only penetrate the blood-brain barrier in the presence of inflammation during the early stages of treatment [4, 6].
Few studies have been conducted examining the relationship between different patterns of drug resistance and their effect on treatment response and the outcome of TBM. The potent early bactericidal effect of isoniazid, together with excellent CSF penetration, suggests that TBM caused by organisms resistant to this drug may be comparatively harder to treat. We previously conducted a prospective study that failed to show a difference in in-hospital survival between TBM caused by fully sensitive organisms and TBM caused by isoniazid-resistant organisms , but the sample was small (16/56 isoniazid resistant) and longer follow-up data and morbidity in survivors were not reported. More recently, we showed that resistance to isoniazid was associated with significantly longer times to CSF bacterial clearance , implying an attenuated bactericidal response in individuals infected with isoniazid-resistant organisms that might influence outcome.
There are more convincing data from cases of TBM caused by M. tuberculosis resistant to at least isoniazid and rifampicin (multidrug resistant ). A recent retrospective study of 30 South African patients with MDR TBM reported an in-hospital case-fatality rate of 57% (17/30), with significant functional impairment in most of the survivors . Another study described how 7 of 8 patients with HIV-associated MDR TBM died within 16 weeks of the initiation of treatment in the United States (the eighth was lost to follow-up) .
A large prospective study is required to assess the longer-term impact of drug resistance on morbidity and mortality from TBM and to define the patterns of resistance that may require early intervention with second-line antituberculosis drugs, such as a fluoroquinolone. Comparison of presenting clinical features and response to treatment between individuals infected with bacilli of differing drug susceptibilities may also identify features that are predictive of drug resistance and that may assist in diagnosis.
PARTICIPANTS, MATERIALS, AND METHODS
Setting and study participants.
Study participants were recruited from 2 centers in Ho Chi Minh City, Vietnam: Pham Ngoc Thach Hospital for Tuberculosis and Lung Disease (PNT) and the Hospital for Tropical Diseases (HTD). These 500-bed hospitals serve the local community and act as tertiary referral centers for severe tuberculosis (PNT) and infectious diseases (HTD) in southern Vietnam.
Adults (defined as those >14 years old) admitted consecutively to these centers between September 2000 and April 2003 who had clinical meningitis (defined as nuchal rigidity and abnormal cerebrospinal parameters) and had M. tuberculosis isolated from the CSF were eligible to enter the study. Standardized clinical assessment with prospective documentation in individual study notes occurred throughout this period, although between April 2001 and April 2003 all adults were recruited into a randomized, placebo-controlled trial of dexamethasone for the treatment of TBM; details of the methods of recruitment, treatment, and primary outcome assessment of these adults have been reported elsewhere . Adults admitted before April 2001 did not receive adjunctive corticosteroids, and clinical data were recorded as part of a different investigation .
The ethics and scientific committees of both hospitals, the Health Services of Ho Chi Minh City, and the Oxford Clinical Research Ethics Committee approved the study protocols. Written, informed consent to participate in each study was obtained from all participants or their relatives.
It is routine clinical practice in both hospitals for patients with TBM to have serial lumbar punctures, to improve rates of bacteriological diagnosis and assess response to treatment. CSF cell counts and biochemical analysis were performed for each sample by standard methods, and samples were stained and cultured by standard methods for pyogenic bacteria, fungi, and mycobacteria. M. tuberculosis isolates were tested for susceptibility to isoniazid, rifampicin, ethambutol, and streptomycin by the proportion method . If the number of colonies growing on drug-containing medium (isoniazid, 0.2 mg/L; rifampicin, 1.0 mg/L; ethambutol, 2.5 mg/L; and streptomycin, 2.0 mg/L) was 1% of that growing on drug-free medium, the isolate was considered to be resistant. All patients were tested for antibodies to HIV (Determine HIV1/2; Abbott), with positive results confirmed by Western blot.
Those patients untreated previously for tuberculosis received 3 months of daily oral isoniazid (5 mg/kg), oral rifampicin (10 mg/kg), oral pyrazinamide (25 mg/kg; maximum, 2 g/day), and intramuscular streptomycin (20 mg/kg; maximum, 1 g/day), followed by 6 months of oral isoniazid, rifampicin, and pyrazinamide at the same dosages. Ethambutol (20 mg/kg; maximum, 1.2 g/day) was substituted for streptomycin for those infected with HIV and was added to the regimen for 3 months for those treated previously for tuberculosis. Drugs were given by nasogastric tube in patients who were unable to swallow. None of the patients received antiretroviral drugs or any second-line antituberculosis drugs.
Assessment of outcome.
A daily round of all inpatients with TBM by the principal investigator (G.E.T.) ensured uniformity of management between sites and accurate prospective recording of clinical data in individual study notes. Particular attention was given to recording daily the Glasgow coma score , the highest daily temperature, and the onset of new focal neurological signs. Time to coma resolution was defined as the interval (in days) between randomization and the time when a Glasgow coma score of 15 was reached and sustained for >2 consecutive days. Time to fever clearance was defined as the interval between randomization and the observation of a maximum daily temperature of <37.5°C for >5 consecutive days. Relapse was defined as either (1) the onset of new focal neurological signs or (2) a decrease in Glasgow coma score of 2 points for 2 days after >7 days of clinical stability or after any improvement at any time after randomization. All definitions were set a priori. Disability was assessed after 1, 2, 6, and 9 months of treatment by the Rankin scale and "simple questions" score . All data were recorded prospectively in individual study notes, were entered into an electronic database (Microsoft FoxPro; version 6.0) as soon as each patient completed follow-up, and were double-checked before analysis.
Kaplan-Meier estimates were used to display the survival experiences of patients infected with fully sensitive or drug-resistant M. tuberculosis, and the log-rank test was used to assess the equality of the survival distributions. Data from patients lost to follow-up were censored at the time of their last recorded outcome. Relative risk of death between the groups was calculated by Cox regression. Death or severe disability after 9 months was compared as a combined outcome by the 2 test, and the odds ratio (OR) of the outcome was calculated by logistic regression. The last recorded disability score was taken as the 9-month score in those who failed to complete follow-up. Times to fever clearance, coma resolution, and hospital discharge were also compared by Kaplan-Meier estimates and the log rank test. Data from patients who died before discharge, fever clearance, or coma resolution were censored at the time of death.
Clinical variables associated with specified resistance patterns were assessed by univariate and multivariate analysis. The Kruskal-Wallis test was used to compare continuous parameters between 2 or more groups of patients; the 2 test with Yates's correction (or Fisher's exact test) was used for comparison of categorical variables. Variables identified in univariate analysis as being associated (P < .1) with the outcome variable (resistance pattern) were then incorporated into multivariate logistic regression analysis. A forward stepwise variable-selection procedure was used to identify independent predictors of resistance (P to enter, .05; P to remove, .1). All analyses were performed by use of SPSS software (version 10.0; SPSS).
M. tuberculosis was isolated from the CSF of 180 patients with TBM. Of these, 60 (33.3%) died before completing 9 months of treatment, and the median length of follow-up in survivors was 273 days (range, 29407 days). One patient was lost to follow-up; this person was severely disabled when last seen after 29 days of treatment. An HIV test was performed for 178 of the 180 patients, and infection was confirmed in 40 (22.2%).
Comparison of baseline variables between the various patient groups (defined on the basis of the susceptibility of the organism isolated from the CSF) showed no significant differences in duration of symptoms before diagnosis and in Medical Research Council (MRC) grade (table 1). However, there were nonsignificant associations between the patients with MDR TBM and younger age, male sex, and HIV infection. The latter was strongly associated with a worse outcome . Death and recovery in survivors was compared across groups (figure 1 and table 2). Resistance to at least isoniazid and rifampicin had a significant effect on outcome: all of the patients with MDR TBM died, and the median time to death was 12 days (range, 162 days); comparison with the patients infected with fully sensitive organisms revealed a relative risk of death of 11.63 (95% confidence interval [CI], 5.2126.32) (table 3). Resistance to isoniazid, with or without streptomycin resistance, did not have a significant effect on outcome after 9 months.
Other measures of treatment response provide supportive evidence to suggest that isoniazid resistance, with or without streptomycin resistance, does not have a deleterious clinical effect on treatment response. There were no significant differences between times to fever clearance, coma resolution, and hospital discharge in the patients infected with drug-resistant organisms and those infected with fully sensitive organisms, although there was a trend toward a longer time to coma resolution in the patients infected with organisms resistant to 1 or more drugs (P = .054), particularly in the patients infected with organisms resistant to both isoniazid and streptomycin (P = .050) (table 4). All but 1 of the 10 patients with MDR TBM died before any of these outcomes occurred; therefore, comparison of these outcome measures with those in patients infected with organisms of different susceptibilities was not possible.
Drug resistance was not associated with an increased proportion of patients who relapsed after starting treatment. Of the 108 patients infected with fully sensitive organisms, 17 (15.7%) relapsed, compared with 13 (18.1%) of the 72 patients infected with organisms resistant to 1 or more drugs (P = .688). The limited availability of brain imaging meant that the cause of relapse (e.g., hydrocephalus, infarction, or tuberculoma enlargement) could be defined in a only few patients. Only 1 adult had a microbiologically confirmed relapse: M. tuberculosis resistant to both isoniazid and streptomycin was isolated from the CSF before treatment and then again after 255 days of treatment. However, the patient had stopped taking rifampicin and pyrazinamide after 60 days of treatment, because of presumed drug-induced hepatotoxicity.
However, drug resistance did affect the rate of CSF bacterial clearance (figure 3). After 3 days of treatment, there was a trend toward a greater proportion of CSF cultures to be positive in the patients infected with drug-resistant organisms (for fully sensitive organisms, 12/28 [41.4%]; for organisms resistant to isoniazid and/or streptomycin, 12/16 [75.0%]; for MDR organisms, 1/2 [50.0%]; P = .096). After 7 and 30 days of treatment, these differences were statistically significant, with greater proportions of positive cultures in the patients infected with drug-resistant organisms (figure 3B). Two patients with MDR TBM had CSF cultured after 30 days of treatment; both cultures were positive.
Presenting clinical features predictive of a patient's being infected with MDR organisms were assessed by univariate and then multivariate analysis. Only HIV infection was independently predictive of MDR TBM (OR, 12.35 [95% CI, 1.23125.00]; P = .034). There were no significant associations between presenting clinical features and any other resistance pattern (data not shown).
The present study assessed prospectively the effect of drug resistance on response to therapy and outcome in 180 Vietnamese adults treated consecutively for TBM. Of these, 72 (40.0%) had TBM caused by M. tuberculosis resistant to 1 or more first-line antituberculosis drugs. Important prognostic variables, such as MRC grade at presentation and duration of symptoms before diagnosis, were similar across the patient groups (defined on the basis of the susceptibility of the organism isolated from the CSF), although a comparatively higher proportion of patients infected with organisms resistant to 2 or more drugs were infected with HIV. This may have influenced treatment response in these patients, because HIV infection is an independent risk factor for death in Vietnamese adults with TBM . However, outcome after 9 months of treatment was significantly worse only in the patients with MDR TBM. The potent early bactericidal effect of isoniazid  and the ease with which it penetrates the blood-brain barrier  suggests that TBM caused by organisms resistant to isoniazid only (or in combination with streptomycin) might also be more difficult to treat.
The present study compared early clinical response to treatment as well as 9-month morbidity and mortality and failed to show a significant association between isoniazid resistance (with or without streptomycin resistance) and any clinical outcome measure, despite showing that drug-resistant organisms take significantly longer to clear from the CSF. This is a surprising and important finding, and there are a number of possible explanations. The outcome of TBM is believed to be dependent on the severity of the intracerebral inflammatory response, which may be provoked by the release of mycobacterial products after the initiation of bactericidal drugs. We have recently shown that early immunosuppression with adjunctive dexamethasone improves survival from TBM . Our present data suggest that resistance to bactericidal isoniazid and/or streptomycin may delay bacterial killing and, therefore, slow the release of mycobacterial products. This may attenuate the intracerebral inflammatory response and may explain the apparent lack of effect of isoniazid resistance on outcome. However, no differences were observed between the kinetics of CSF inflammatory markers in TBM caused by fully sensitive and drug-resistant organisms.
Alternatively, it is possible that the intracerebral concentrations of isoniazid exceeded the MIC of drug-resistant organisms and continued to exert a bactericidal effect. The present study assessed in vitro resistance by the proportion method using a single drug concentration; variations in levels of resistance were not assessed. This is an important omission, because these data would have allowed the roles played by each drug in the treatment of TBM to be characterized in greater detail. In particular, there is considerable variation in the isoniazid MIC of clinical isolates of M. tuberculosis, and low and high levels of resistance have been recognized . Ellard et al. showed that, 4 h after administration of an oral dose of 9 mg/kg isoniazid, the mean CSF concentration was 3.2 mg/L , 10 times the MIC of most isoniazid-resistant strains of M. tuberculosis  but possibly insufficient to kill highly resistant organisms. However, there are no convincing data from numerous studies of the treatment of pulmonary tuberculosis to suggest that isoniazid can have an effect on drug-resistant organisms and, without more-precise data on the isoniazid MIC, the present study does not support this explanation. Instead, we suggest that the most plausible explanation for the observed lack of effect of isoniazid resistance on outcome is that the relative roles played by isoniazid, rifampicin, and pyrazinamide are similar in the treatment of TBM and pulmonary tuberculosis. The dramatic effect of MDR on the outcome of TBM suggests that, as with pulmonary tuberculosis, rifampicin is a critical component of successful treatment, regardless of concerns about low intracerebral drug concentrations.
Given that all of the patients with MDR TBM in the present study died within the first 2 months of treatment and that intervention with second-line antituberculosis drugs (such as a fluoroquinolone, ethionomide, and cycloserine) probably improves the outcome of MDR TBM [16, 17], methods for the early detection of MDR organisms in the CSF are desperately required. Sadly, the present study suggests that presenting clinical features are poorly predictive: only HIV infection was independently associated with MDR TBM. Likewise, it was not possible to identify any clinical markers of poor treatment response (changes in serial CSF parameters, for example) that identified TBM caused by MDR or more-susceptible M. tuberculosis, although the rapid deaths of many of the patients with MDR TBM limited the available data. A positive CSF culture after 7 or 30 days of treatment might be predictive of MDR TBM, but the bacilli grow too slowly for this to be clinically useful. Future studies may show that differences in the kinetics of CSF molecules known to be of prognostic importance in TBM, such as lactate and glucose , may be more helpful in identifying persons with MDR TBM. However, until these data become available, alternative methods need to be taken into consideration. Amplification and detection of mutations in the M. tuberculosis rpoB gene that predict phenotypic rifampicin resistance has been shown to be useful for the early prediction of MDR pulmonary disease [18, 19] but has not yet been assessed for use with CSF.
In conclusion, M. tuberculosis resistant to isoniazid and/or streptomycin probably has no detrimental effect on the outcome of TBM when patients are treated with a rifampicin-containing first-line antituberculosis drug regimen. However, M. tuberculosis resistant to at least isoniazid and rifampicin has a devastating effect on the outcome of TBM. The performance of new diagnostic methods and the efficacy of second-line antituberculosis drugs for the treatment of MDR TBM require urgent assessment.
We thank all of the doctors and nurses at Pham Ngoc Thach Hospital for Tuberculosis and Lung Disease (PNT) and the Hospital for Tropical Diseases (Ho Chi Minh City, Vietnam) who cared for the patients. We also thank the administrative and laboratory staff of PNT, in particular, Ms. Dai Viet Hoa, Dr. Mai Nguyet Thu Huyen, Mr. Tran Huu Loc, and Ms. Pham Hoang Anh, who performed the drug-susceptibility testing.
1. World Health Organization (WHO). Global tuberculosis control: surveillance, planning, financing. Geneva: WHO, 2004. First citation in article
2. Mitchison DA. Role of individual drugs in the chemotherapy of tuberculosis. Int J Tuberc Lung Dis 2000; 4:796806. First citation in article
3. Mitchison DA, Nunn AJ. Influence of initial drug resistance on the response to short-course chemotherapy of pulmonary tuberculosis. Am Rev Respir Dis 1986; 133:42330. First citation in article
4. Ellard GA, Humphries MJ, Allen BW. Cerebrospinal fluid drug concentrations and the treatment of tuberculous meningitis. Am Rev Respir Dis 1993; 148:6505. First citation in article
5. Donald PR, Sirgel FA, Botha FJ, et al. The early bactericidal activity of isoniazid related to its dose size in pulmonary tuberculosis. Am J Respir Crit Care Med 1997; 156:895900. First citation in article
6. Bobrowitz ID. Ethambutol in tuberculous meningitis. Chest 1972; 61:62932. First citation in article
7. Thwaites GE, Chau TT, Caws M, et al. Isoniazid resistance, mycobacterial genotype and outcome in Vietnamese adults with tuberculous meningitis. Int J Tuberc Lung Dis 2002; 6:86571. First citation in article
8. Thwaites GE, Caws M, Chau TT, et al. Comparison of conventional bacteriology with nucleic acid amplification (amplified mycobacterium direct test) for diagnosis of tuberculous meningitis before and after inception of antituberculosis chemotherapy. J Clin Microbiol 2004; 42:9961002. First citation in article
9. Patel VB, Padayatchi N, Bhigjee AI, et al. Multidrug-resistant tuberculous meningitis in KwaZulu-Natal, South Africa. Clin Infect Dis 2004; 38:8516. First citation in article
10. Daikos GL, Cleary T, Rodriguez A, Fischl MA. Multidrug-resistant tuberculous meningitis in patients with AIDS. Int J Tuberc Lung Dis 2003; 7:3948. First citation in article
11. Thwaites GE, Nguyen DB, Nguyen HD, et al. Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults. N Engl J Med 2004; 351:174151. First citation in article
12. Thwaites GE, Simmons CP, Than Ha Quyen N, et al. Pathophysiology and prognosis in Vietnamese adults with tuberculous meningitis. J Infect Dis 2003; 188:110515. First citation in article
13. Collins CH, Grange JM, Yates MD. Tuberculous bacteriology: organisation and practice. 2nd ed. Oxford: Butterworth-Heinemann, 1997. First citation in article
14. Teasdale G, Jennett B. Assessment of coma and impaired consciousness: a practical scale. Lancet 1974; 2:814. First citation in article
15. van Soolingen D, de Haas PE, van Doorn HR, Kuijper E, Rinder H, Borgdorff MW. Mutations at amino acid position 315 of the katG gene are associated with high-level resistance to isoniazid, other drug resistance, and successful transmission of Mycobacterium tuberculosis in The Netherlands. J Infect Dis 2000; 182:178890. First citation in article
16. DeVincenzo JP, Berning SE, Peloquin CA, Husson RN. Multidrug-resistant tuberculosis meningitis: clinical problems and concentrations of second-line antituberculous medications. Ann Pharmacother 1999; 33:11848. First citation in article
17. Berning SE, Cherry TA, Iseman MD. Novel treatment of meningitis caused by multidrug-resistant Mycobacterium tuberculosis with intrathecal levofloxacin and amikacin: case report. Clin Infect Dis 2001; 32:6436. First citation in article
18. Mokrousov I, Otten T, Vyshnevskiy B, Narvskaya O. Allele-specific rpoB PCR assays for detection of rifampin-resistant Mycobacterium tuberculosis in sputum smears. Antimicrob Agents Chemother 2003; 47:22315. First citation in article
19. Johansen IS, Lundgren B, Sosnovskaja A, Thomsen Vs VO. Direct detection of multidrug-resistant Mycobacterium tuberculosis in clinical specimens in low- and high-incidence countries by line probe assay. J Clin Microbiol 2003; 41:44546. First citation in article