Linezolid-containing Regimen Lowers Mortality of Rifampicin/multidrug-resistant Tuberculous Meningitis in Shenzhen, China

Background: The morbidity of rifampicin/multidrug-resistant tuberculous meningitis (RR/MDR-TBM) is increasing in many countries and regions in the world. Its mortality is signicantly higher than non rifampicin/multidrug-resistant ones (NRR/MDR-TBM ). This article aims to explore the RR/MDR-TBM related risk factors, and compare the different therapeutic effects to the RR/MDR-TBM patients between linezolid-containing anti-tuberculosis regimen and non linezolid regimen in Shenzhen city. Furthermore, we want to nd a better therapy for pathogen negative TBM with RR/MDR-TBM related risk factors. Methods: 137 cases with conrmed TBM (pathogen positive), who were hospitalized in the Third People’s Hospital of Shenzhen from June 2014 to March 2020, were enrolled in this study, all patients were divided into RR/MDR-TBM group (12 cases) and NRR/MDR-TBM group (125 cases) according to the GeneXpert MTB/RIF and (or) phenotypic drug susceptibility test of CSF (cerebral spinal uid). The risk factors related to RR/MDR-TBM were investigated through comparing the clinical and examination features between the two groups.The mortality of RR/MDR-TBM patients treated with different regimens was analyzed to compare their respective therapeutic effects to the RR/MDR-TBM . P<0.05 differences is considered statistically signicant. Results: difference in other clinical and examination features between the two groups. The mortality of RR/MDR-TBM treated with the linezolid-containing regimen was signicantly lower than that treated with non linezolid regimen (P=0.045). Conclusions: The main related high-risk factor of RR/MDR-TBM is the history of anti-tuberculosis treatment. Linezolid -containing regimen appears to lower the mortality of RR/MDR-TBM signicantly. Therefore, it is be used empirical antituberculosis previously


Background
Tuberculous meningitis (TBM) is the most fatal type of tuberculosis, especially in children or adults coinfected with HIV. Even after standardized anti-tuberculosis treatment, the mortality can still reach as high as 20%-50%, and nearly half of the survivors will develop serious central nervous system sequelaes [1,2].
In recent years, reports of drug-resistant tuberculosis meningitis have been increasing. Compared with non-drug-resistant tuberculosis meningitis, drug-resistant tuberculosis meningitis, especially rifampicin/multidrug-resistant tuberculous meningitis (RR/MDR-TBM) is more lethal and means more medical cost for the patient, so more and more attention has been paid to it [3,4]. China is one of the 30 high burden RR/MDR-TB countries of the world, the estimated number of people with RR/MDR tuberculosis in 2018 is about 66,000 (14% of the world). RR/MDR-TB accounts for 7.1% among the newly diagnosed tuberculosis patients, while among pulmonary tuberculosis patients treated previously ,RR/MDR-TB accounts for 21% [5]. As the most lethal tuberculosis type, the incidence of TBM accounts for about 1% of all tuberculosis [6], and it accounts for about 7.23% of extrapulmonary tuberculosis in China [7], however, the incidence of RR/MDR-TBM and its related risk factors are unknown. In this study, we describe the onset and drug resistance pro e of 137 con rmed TBM patients who were hospitalized in the Third People's Hospital of Shenzhen city from June 2014 to March 2020, as the hospital is the only designated hospital for the treatment of TBM patients in Shenzhen city, its data for TBM can represent the general situation of the TBM onset in Shenzhen. We compared the related characteristics(including demographic, clinical and examination characteristics) between RR/MDR-TBM patients and nonrifampicin/multidrug-resistant tuberculous meningitis (NRR/MDR-TBM) patients to explore the RR/MDR-TBM related risk factors furthermore, we compared the effects of linezolid-containing and non linezolid regimens on the prognosis of RR/MDR-TBM. On this basis, we discuss the empirical treatment regimen for pathogen-negative TBM with RR/MDR-TBM related risk factors we hope it can improve the prognosis of RR/MDR-TBM and pathogen-negative TBM in China.

1: patients enrollment and data collection:
(1):Case enrollment process: This study adopts a retrospective method, we collect the data of 151 tuberculous meningitis patients who were hospitalized and met the diagnostic criteria for tuberculous meningitis from June 2014 to March 2020 , among them, 14 patients who did not perform GeneXpert MTB/RIF and Mycobacterium tuberculosis phenotype drug susceptibility test(DST) were excluded. Finally, 137 patients were enrolled in the study, including 98 males and 39 females, the youngest was 2 years old, and the oldest was 76 years old, with a median of 29 years old. We collect the demographic/clinical and examination characteristics of patients through the HIS and LIS system of the hospital, and con rm the survival status of discharged patients through telephone contact or outpatient visit.
(2)Case entry criteria: meet the diagnostic criteria for de nite tuberculous meningitis [8], that is, in addition to the clinical manifestations of tuberculous meningitis and abnormal cerebrospinal cytology and biochemical examinations, the patient's cerebrospinal uid examination meets at least one of the following two criteria: 1. GeneXpert MTB/RIF (GeneXpert) indicates a positive nucleic acid test for Mycobacterium tuberculosis. 2. The BACTER MGIT960 liquid culture method for mycobacteria culture was positive and was identi ed as Mycobacterium tuberculosis (MTB), at the same time, a phenotypic drug sensitivity test was performed. Only one pathogen-positive CSF result was analyzed for each patient.
(3) Diagnostic criteria for RR/MDR TBM cases: meet the diagnostic criteria for de nite tuberculous meningitis case and meet at least one of the following two criteria: 1.GeneXpert test is positive and indicates resistant to rifampicin(RIF). 2. The cerebrospinal uid MTB culture is positive, and the DST shows that it is resistant to rifampicin ,meanwhile, isoniazid is also resistant or sensitive.
(4) We described the demographic features of the TBM patients ,according to the results of CSF GeneXpert and/or DST, all patients were divided into RR/MDR group and NRR/MDR group,then we compare the differences of the main clinical and examination characteristics between the two groups, on this basis, we explore the main related risk factors of RR/MDR tuberculous meningitis, and compare the differences in mortality between the two groups and the impact of different regimen on the prognosis of RR/MDR-TBM.

Laboratory examination:
(1). Method of obtaining cerebrospinal uid specimen: Lumbar puncture is performed within 24 hours after admission to obtain cerebrospinal uid. Some patients' cerebrospinal uid are acquired through lateral ventricle drainage, the pressure of cerebrospinal uid is measured, and the CSF specimen is submitted at once after acquirement for tests such as routine cytology, biochemistry, acid-fast bacillus (AFB), GeneXpert, MTB culture and species identi cation, Cryptococcus membrane polysaccharide antigen, bacteria and fungi smears and culture , etc. the volume of CSF used for GeneXpert and MTB culture is 1-2ml. 3. The phenotypic drug susceptibility test(DST) of Mycobacterium tuberculosis adopts the absolute concentration method method: according to the literature and operating instructions [12].
4. Acid-fast staining, cytology, biochemistry of CSF and blood routine, blood biochemistry, blood tuberculosis interferon release assay, HIV antibody detection, etc. are operated in accordance with the instruction of the instrument and kit.
3. Anti-tuberculosis treatment regimen: Non-RR/MDR-TBM group: The anti-tuberculosis treatment regimen includes: Isoniazid (10-15mg/Kg/day, the maximum dose is 0.9g/day, intravenous administration is used during hospitalization, and it is changed into oral administration after discharge), Rifampicin (10-15mg/Kg/day, the maximum amount is 0.6g/day, intravenously administered during hospitalization, and changed into oral administration after discharge), Pyrazinamide (25-30mg/Kg/day,1.5-2.0g for adults, oral administration or nasal feeding), ethambutal(15mg/kg/day,0.75-1.0g for adults, oral administration or nasal feeding), some patients were plus streptomycin injection or levo oxacin (or moxi oxacin), the scheduled total course of treatment for NRR/MDR TBM is 12-18 months except for the patient's death. RR/MDR-TBM group: 7 patients were treated with linezolidcontaining regimen (linezolid 0.6g/day, intravenously or orally administration during hospitalization, and changed into oral administration after discharge. The course of linezolid administration was 3 months to 18 months), the regimen include levo oxacin (or moxi oxacin) and the following 2-4 drugs: cycloserine, pyrazinamide, high-dose isoniazid, amikacin, protionamide, etc(the dose is in accordance with the nation's guildline of China for RR/MDR TB). Five patients receive the non-linezolid-containing regimen in the early stage. The regimen included the following 4-6 drugs: levo oxacin (moxi oxacin), pyrazinamide, high-dose isoniazid, rifampicin, amikacin, and protionamide, cycloserine, ethambutol, etc.the scheduled total course of treatmen for RR/MDR TBM is 24 months at least except for the patient's death. In addition to anti-tuberculosis therapy, all patients were treated with glucocorticoids( the treatment course was up to The difference of rifampicin/multidrug resistance rate between the two groups are statistically signi cant, See Table 2.

Comparison of clinical features and cerebrospinal uid examination features between the RR/MDR-
TBM group and the NRR/MDR-TBM group: 50% of the patients in the RR/MDR-TBM group had a history of anti-tuberculosis treatment (6/ 12), while only 10.5% of the patients in the NRR/MDR-TBM group had a history of anti-tuberculosis treatment(6/125), the difference between the two groups was statistically signi cant, other factors including age, gender, type of pulmonary tuberculosis, and HIV infection, diabetes prevalence and clinical manifestations, cerebrospinal uid tests, T lymphocyte subsets, T-SPOT, etc. had no signi cant difference between the two groups, see Table 3.
4. 4.The comparison of mortality between RR/MDR-TBM group and NRR/MDR-TBM group: the difference of mortality between RR/MDR-TBM group and NRR/MDR-TBM group are not statistically signi cant, no matter whether the loss to follow-up are alive or not, see Table 4.

The comparison of mortality between RR/MDR-TBM patients groups treated with different regimens:
Among RR/MDR-TBM patients, the mortality of these patients treated with linezolid -containing regimen in the early stage was signi cantly lower compared with the patients treated without linezolid regimen(P=0.045), see Table 5.

Discussion
The Third People's Hospital of Shenzhen is the only designated hospital for treatment of TBM patients in Shenzhen , Shenzhen is a megacity with a long-term lived population of more than 13 million, while the registered population account for less than 40%, and the native population just account for even far smaller proportion, most of the population are immigrant from other cities, this is also paralled with the proportion of TB patients in their composition of domicile and population. Among the patients included in this study, only 10 patients belonged to registered Shenzhen population (7.3%), and 25 were from other cities of Guangdong province (18.3%). Others were mainly from southern province of China including 13 from Guangxi (9.5%) and 12 from Hunan (8.8%), 11 from Jiangxi (8.0%), 8 from Hubei (5.8%), and 13 from Sichuan (9.5%), 10 from Guizhou (7.3%), 9 from Chongqing (6.6%), etc., 7 people (5.1%) with unknown household registration, the number of cases from other 11 provinces is only 19 people (13.9%). More than half of the patients are migrant workers, and they are relatively young, with a median age of only 29. Therefore, our data can represent the incidence of TBM among the population in southern and southwestern provinces of China better than other researches, especially in the migrant workers in Shenzhen, the reason that the incidence of TBM is higher in the migrant workers in Shenzhen may be due to the migrant worker's low income and hard work, which lead to their poor nutrition condition and lack of break, these nally lower their immunity to resist to Mycobacteria tuberculosis.
China is the world's second largest country with a high burden of RR/ MDR-TB after India, the estimated number of people with RR/MDR tuberculosis in 2018 is about 66,000 (14% of the world). RR/MDR-TB accounts for 7.1% among the newly diagnosed tuberculosis patients, while Among pulmonary tuberculosis patients treated previously, RR/MDR-TB accounts for 21% [5]. As the most lethal tuberculosis type, there have been more and more reports of RR/MDR-TBM in China and abroad in recent years. Some small sample of studies have shown that RR/MDR-TBM accounts for 12%-39.29% of TBM patients in western China, isoniazid resistant TBM reached 64.29% [13,14], even in children (untreated previously for tuberculosis) MDR and XDR accounted for 7.2% [15]. Although the sample size of the above studies is small, it also indicates that RR/MDR-TBM was not rare in China. Many foreign reports also suggest that RR/MDR-TBM accounts for a relatively high proportion of TBM patients [3,4,16,17]. Our research shows that among all patients with positive pathogens and detected for rifampicin resistance, rifampicin resistance rate reached 8.8%, isoniazid resistance (rifampicin sensitive) rate 3.4%, multidrug-resistant rate is 5.1%, and because some patients have not been detected for isoniazid resistance, the ratio of isoniazid resistance and multidrug resistance should been higher. Among previously treated cases, the ratio of RR/MDR-TBM patients is as high as 33.3%, and among new cases, the ratio of RR/MDR-TBM patients is only 5.0% , the difference between them is statistically signi cant. The above results indicate that among the patients with tuberculous meningitis in Shenzhen, the proportion of RR/MDR-TMB patients is higher, and the proportion of RR/MDR-TMB previously treated cases is signi cantly higher than that of new cases, hence, the previous history of anti-tuberculosis treatment should be taken as the main risk factor for RR/MDR-TBM, which is similar to RR/MDR pulmonary tuberculosis.
A series of studies have suggested that the high-risk factors associated with RR/MDR pulmonary tuberculosis include previous history of anti-tuberculosis treatment, HIV infection, diabetes and other factors [4,5,18]. In our study, the proportion of previously treated cases in the meningitis group was up to 50%, while the proportion of previously treated cases in the NRR/MDR-TBM group was only 10.5%. The difference between the two was statistically signi cant, however, the difference in other factors (including HIV infection rate, diabetes prevalence and various clinical manifestations, cerebrospinal uid cytology and biochemical test indicators, etc.)between the two groups was not statistically signi cant, indicating that the previous anti-tuberculosis treatment history is the most important related risk factor for RR/MDR-TBM. Therefore, TBM patients who have a history of anti-tuberculosis treatment must be alert to the possibility of RR/MDR-TBM.
TBM is the most severe type of tuberculosis with the highest mortality, According to different reports, its mortality ranges from 20% to 50% [1,2,19], worsestill, up to 25% of the survivors will develop various central nervous system sequelae, which will seriously lower their life quality and increase long-term mortality [20]. Compared with drug-sensitive TBM, the mortality of various drug-resistant TBM is even higher, for example, the mortality of isoniazid-resistant meningitis is 56%, and the mortality of RR/MDR-TBM treated with rst-line anti-tuberculosis drugs is even as high as 100% [4,21]Therefore, the early diagnosis of tuberculous meningitis and the evaluation of its drug resistance are essential for formulating the correct regimen and reducing the mortality of patients. In this study, there was no signi cant difference in mortality between the RR/MDR-TBM group and the non RR/MDR-TBM (excluding those who were lost to follow-up),there are many patients in the NRR/MDR-TBM group were lost to follow-up in this study ,The reason maybe inclcud: rstly, the contact telephone number has expired or been cancelled because the patient had to returned to his(her) native place due to loss of labor, secondly, it cannot be ruled out that the telephone number was cancelled due to the patients' death, Therefore, we assumed two extreme situations, that is, all patients who were lost to follow-up survived or died (the real situation should locate between these two extreme cases), and the results showed that in any cases, There was no signi cant difference in the mortality between RR/MDR groups and NRR/MDR-TBM group, indicating that after timely and correct anti-tuberculosis treatment, the mortality of RR/MDR-TBM patients was not higher than that of NRR/MDR-TBM.
Among RR/MDR-TBM patients, the mortality rate of the group treated with linezolid containing regimen was signi cantly lower than that of the group treated with non-linezolid regimen, this is consistent with the results of many domestic and foreign related studies ,so our research further supports that linezolid has a good therapeutic effect on RR/MDR-TB, the reasons that linezolid has a good curative effect is , on the one hand related to its strong early bactericidal activity against RR/MDR Mycobacterium tuberculosis, on the other hand, linezolid exibits excellent permeability of the blood-brain barrier, so it can keep the higher drug concentration in the cerebrospinal uid and brain tissue [22,23] For con rmed TBM cases, the anti-tuberculosis treatment regimen can be formulated based on the results of the drug sensitivity test, unfortunately, despite the application of various new detection methods such as GeneXpert RIF/MTB, GeneXpert Ultra, less than 40% of the clinically diagnosed TBM cases (including de nite, probable, and possible cases) without HIV co-infection are pathogen-positive cases (de nite cases) by the above methods, and more than 60% of the cases are pathogen negative(probable or possible cases) [24]. For probable or possible cases, it is generally advised to give empirical anti-tuberculosis treatment as soon as possible, and the empirical anti-tuberculosis treatment usually uses regimen for drug-sensitive TBM (isoniazid, rifampicin, pyrazinamide, ethambutol) [25], but this regimen has very poor effect on drug-resistant TBM, especially RR/MDR-TBM. Even if uoroquinolones are added to the regimen and the dose of rifampicin is increased, it can only decrease the mortality of isoniazid-resistant TBM but cannot reduce the mortality of RR/MDR-TBMD [3,17]. On the contrary, if drug-sensitive TBM patients are treated with RR/MDR-TBM regimen, it may increase drug adverse effects and the economic burden of patients. Therefore, for pathogen negative TBM cases, the risk of drug resistance should be assessed before the initiation of empirical anti-tuberculosis treatment.
Both our and other authors' studies have shown that the proportion of RR/MDR-TBM in the pathogenpositive TBM patients in China is relatively high, especially in the previously treated cases, our data shows it can reach as high as 33%. And we assume that the similar situation should occur in the pathogen-negative TBM patient population, that is, the ratio of RR/MDR-TBM will be signi cantly higher in previously treated cases, and the regimen composed of rst-line anti-tuberculosis drugs will obviously increase the mortality. Therefore, we recommend that for pathogen-negative TBM cases with a history of anti-tuberculsosi , an empirical treatment regimen for RR/MDR-TBM comprised of linezolid and other second-line anti-tuberculosis drugs should be administered from the initiation of treatment, while for pathogen-negative TBM new cases, a regimen comprised of rst-line anti-tuberculosis drugs can be administrated, Once the therapeutic effect is not good, it should be changed into the regimen for RR/MDR-TBM, because it has been proven that no matter the drug sensitivity, the linezolid containing regimen manifests good therapeutic effect to life-threatening TBM. In the group of drugs recommended by WHO for the treatment of multidrug-resistant tuberculosis, drugs that are known to easily penetrate the blood-brain barrier and achieve effective inhibitory concentration include linezolid, uoroquinolones (levo oxacin, moxi oxacin), cycloserine, and pyrazinamide, ethionamide/protionamide, high-dose isoniazid, imipenem, meropenem, etc. (amikacin easily penetrates the blood-brain barrier only when the meninges are in amed, bedaquiline, delamanid, clofazimine still lack relevant data on the permeability of the blood-brain barrier). Therefore, it is necessary to choose the above drugs to constitute an effective anti-tuberculosis regimen [26][27][28]. Recent studies have shown that linezolid has a strong early bactericidal activity and easily penetrates the blood-brain barrier. Its cerebrospinal uid drug concentration/blood drug concentration can reach as high as 80-100% [30], so it should be used as an essential component drug for RR/MDR-TBM treatment regimen. It shows that the use of linezolid-containing regimens can signi cantly reduce the mortality and improve the prognosis of RR/MDR-TBM patients, and the use of linezolid for drug-sensitive patients with severe TMB can also signi cantly improve the prognosis [22,23,27] The limitations of this study: rst of all, it is not clear whether the drug resistance pro le of the pathogennegative TBM patient population is exactly the same as that of the pathogen-positive TBM patient population. However, the prerequisite for clarifying drug resistance is pathogen positive, so maybe there is no very good solution for this question ,it's only a convinced deduction but not a con rmed conclusion. Secondly, the number of RR/MDR-TBM cases in this study is small, and it is necessary to further increase the number of cases for a prospective multicenter cohort study to evaluate the e cacy and adverse effects of linezolid-containing regimens, nally, the high ratio of loss to follow-up of NRR/MDR-TBM patients in this study also makes it di cult to analyze the results, although we have made appropriate assumptions.

Conclusions
In summary, our results shows for the rst time that in recent years, the population of TBM patients in Shenzhen is mainly migrant workers from southern and southwestern China, RR/MDR-TBM accounts for a relatively high proportion among them. The main related high-risk factor of RR/MDR-TBM is the history of anti-tuberculosis treatment. Linezolid containing regimen can decrease the mortality of RR/MDR-TBM signi cantly, therefore, we recommended linezolid-containing regimen as the empiric anti-tuberculosis treatment for pathogen negative TBM with a history of anti-tuberculosis in China.     Figure 1 the enrollment of the patients