We evaluated the diagnostic performance of a simple and label-free pathogen enrichment method using homobifunctional imidoesters (HI) and a microfluidic system, called the SLIM assay, followed by real-time PCR from cerebrospinal fluid (CSF) in human immunodeficiency virus (HIV)-uninfected patients with suspected tuberculous meningitis (TBM). Patients with suspected TBM were prospectively enrolled in a tertiary hospital in an intermediate tuberculosis (TB)-burden country during a 30-month period.
KEYWORDS: SLIM assay, Xpert MTB/RIF, tuberculous meningitis
ABSTRACT
We evaluated the diagnostic performance of a simple and label-free pathogen enrichment method using homobifunctional imidoesters (HI) and a microfluidic system, called the SLIM assay, followed by real-time PCR from cerebrospinal fluid (CSF) in human immunodeficiency virus (HIV)-uninfected patients with suspected tuberculous meningitis (TBM). Patients with suspected TBM were prospectively enrolled in a tertiary hospital in an intermediate tuberculosis (TB)-burden country during a 30-month period. TBM was classified according to the uniform case definition. Definite and probable TBM were regarded as the reference standards for TBM, and possible TBM and not-TBM as the reference standards for not-TBM. Of 72 HIV-uninfected patients with suspected TBM, 10 were diagnosed with definite (n = 2) and probable (n = 8) TBM by the uniform case definition. The sensitivity of the SLIM assay was 100% (95% confidence interval [CI], 69 to 100%) compared with definite or probable TBM, and it was superior to those of mycobacterial culture (20% [95% CI, 3 to 56%]) and the Xpert MTB/RIF assay (0% [95% CI, 0 to 31%]). Of 21 possible TBM and 41 not-TBM patients by the uniform case definition, 5 possible TBM and no not-TBM patients gave positive results in the SLIM assay. The specificity of the SLIM assay was 92% (95% CI, 82 to 97%; 5/62). We demonstrated that the SLIM assay had a very high sensitivity and specificity with small samples of 10 cases of definite or probable TBM. Further studies are needed to confirm this finding and to compare the SLIM assay with mycobacterial culture, Xpert MTB/RIF, and Xpert MTB/RIF Ultra assays in a larger prospective cohort of patients with suspected TBM, including both HIV-infected and HIV-uninfected cases.
INTRODUCTION
Tuberculous meningitis (TBM) is the most severe form of Mycobacterium tuberculosis infection and accounts for about 1% of all tuberculosis cases (1). The clinical course of TBM can deteriorate rapidly, and the mortality rate is high in the absence of prompt diagnosis and treatment (2). The gold standard for diagnosing tuberculosis is mycobacterial culture, but it takes 2 to 8 weeks to get the results, and the sensitivity is relatively low because of the paucibacillary nature of TBM (2). In 2013, the World Health Organization (WHO) endorsed the Xpert MTB/RIF assay as the initial diagnostic test for suspected TBM patients in preference to other tests, such as mycobacterial culture and M. tuberculosis PCR (3). The 43 to 72% sensitivity of the Xpert MTB/RIF assay was still not high enough to rule out TBM (4–7). In a recent report, the sensitivity of the Xpert MTB/RIF Ultra assay (70 to 95%) was superior to those of the Xpert MTB/RIF assay (43 to 45%) and mycobacterial culture (43 to 45%) (7). However, only human immunodeficiency virus (HIV)-infected patients with suspected TBM were enrolled in that study. So, further studies are needed in HIV-uninfected patients because of the lower bacterial loads in those patients (4, 5).
Due to these limitations of current diagnostic tests, clinicians usually have problems diagnosing TBM, especially in HIV-uninfected patients. Hence, it is important to develop a new diagnostic tool to rapidly rule out TBM. We have developed a new assay for diagnosing TBM, involving simple and label-free pathogen enrichment by homobifunctional imidoesters (HIs) using a microfluidic system, called the SLIM assay (8). We have evaluated the diagnostic performance of the SLIM assay followed by real-time PCR in HIV-uninfected patients with suspected TBM compared with the uniform case definition (9) in a country with an intermediate tuberculosis (TB) burden and low HIV burden.
MATERIALS AND METHODS
Study population.
All patients with suspected TBM who consented to the use of their cerebrospinal fluid (CSF) for additional tests, such as the SLIM assay (Fig. 1), were prospectively enrolled at a 2,700-bed tertiary-care facility in Seoul, South Korea, from March 2015 to August 2017. The HIV antibody screening test was performed on all participants. Decisions regarding anti-TB treatment were made by the attending physicians (Y. S. Koo, S.-B. Jeon, and S.-A. Lee) based on each patient’s initial clinical features, blood tests, image findings, and CSF profiles. The results of the SLIM assays were concealed from the attending physicians to avoid bias. The study protocol was approved by the Institutional Review Board of our hospital.
FIG 1.
The SLIM assay for diagnosing tuberculous meningitis from CSF.
Study design and definition.
CSF studies, including white blood cell count, protein, glucose, and brain imaging (computed tomography or magnetic resonance imaging), were performed in all patients, and the patients also underwent Xpert MTB/RIF and SLIM assays for diagnosing TBM. We categorized the patients into definite, probable, possible, and not-TBM by the uniform case definition (9). According to this criterion, definite TBM meant microbiologically confirmed TBM (excluding the results of the SLIM assays), and probable (≥12 points when imaging is available, but without microbiological evidence) and possible TBM (6 to 11 points when imaging is available, with exclusion of alternative diagnoses) were determined by the diagnostic scoring system.
The SLIM assay.
The principle of the SLIM assay, and its components, have been described previously (8). Briefly, it is based on combining a microfluidic platform with a low-cost thin film and homobifunctional imidoester (HI) reagents to enrich and extract M. tuberculosis. After the platform modification step with O2 plasma and 3-aminopropyl diethoxymethylsilane (APDMS), about 1 ml of noncentrifuged CSF sample is mixed with dimethyl pimelimidate (DMP) solution (100 mg/ml) as a HI reagent and injected into the microfluidic platform using a syringe pump. HI reagents can capture bacterial surfaces by electrostatic coupling without the need for detergent or bulky instruments. Then, the enriched samples are extracted with a mixture of DMP and lysis buffer in the same platform, and the extracted DNA is captured by covalent bonding with the DMP. The final elution volumes of genomic DNA are 100 μl and they are stored at −20°C for subsequent conventional PCR (Fig. 1). We used 1 ml of each noncentrifuged CSF for the SLIM assay to compare the sensitivity of the Xpert MTB/RIF assay using the same volume of CSF.
Amplification of M. tuberculosis DNA.
DNA amplification was performed by both endpoint PCR and real-time PCR. The primers for targeting rpoB, katG, and IS6110 were used as described previously (10, 11). The PCR process consisted of an initial denaturation step at 95°C for 15 min; 45 cycles of 95°C for 30 s, 56°C (for IS6110), 60°C (for katG), or 63°C (for rpoB) for 30 s, and 72°C for 30 s; and a final elongation step at 72°C for 5 min. Samples of DNA (5 μl) were amplified in a total volume of 25 μl containing 10× PCR buffer (Qiagen), 2.5 mM MgCl2, 0.25 mM deoxynucleotide triphosphate, 25 pmol of each primer, and 1 unit of Taq DNA polymerase (Qiagen). Gel electrophoresis was used to separate the PCR products on a 2% agarose gel containing ethidium bromide (EtBr). The gel was visualized using a GelDoc System (Clinx Science Instruments). For real-time PCR, the following procedure was modified from the AriaMx real-time PCR instrument protocol (Agilent Technologies). A 5-μl sample of DNA was amplified in a total volume of 20 μl containing 10 μl of 2× Brilliant III SYBR green quantitative PCR (qPCR) master mix, 25 pmol of each primer, and deionized water. An initial preincubation cycle of 95°C for 10 min was followed by 40 cycles of 95°C for 10 s, 56°C (for IS6110) or 63°C (for katG and rpoB) for 20 s, and 72°C for 20 s, and by a cooling step of 40°C for 30 s. The SYBR green signals of the amplified products were acquired using an AriaMx real-time PCR system (Agilent Technologies).
To determine drug resistance, the PCR products of the rpoB gene were purified using an Expin PCR SV kit (GeneAll, Republic of Korea). The purified samples were directly sequenced with the forward sequencing primer by BigDyeTerminal chemistry. The DNA sequencing reaction mixtures were electrophoresed using an ABI 3730XL DNA analyzer (Applied Biosystems, USA) at the Macrogen, Inc., sequencing center (Seoul, Republic of Korea).
Mycobacterial culture and the Xpert MTB/RIF assay.
For culture identification of mycobacteria, at least 2-ml CSF samples were inoculated in liquid (Bactec MGIT 960) and solid (Ogawa media) culture media. All cultures were performed for at least 8 weeks. The Xpert MTB/RIF assay (Cepheid, Sunnyvale, CA) was performed following the manufacturer’s protocol and previous reports (12, 13). Each noncentrifuged CSF sample (1 ml) was mixed with 1 ml of Xpert MTB/RIF sample reagent and incubated at room temperature for 15 min. After incubation, the mixture was transferred to the cartridge provided, and the cartridge was operated according to the manufacturer’s protocol.
Statistical analysis.
Categorical variables were compared using the χ2 or Fisher’s exact test, as appropriate, and continuous variables were analyzed using Student's t test or the Mann-Whitney U test, as appropriate, to compare the baseline characteristics of TBM and not-TBM patients. We compared the diagnostic performance of the SLIM assay with that of conventional diagnostic tests using McNemar’s test. All tests of significance were two-tailed, and P values of <0.05 were considered statistically significant. Calculations were performed using SPSS Statistics for Windows, version 21.0 (IBM Corp., Armonk, NY).
RESULTS
Study population.
All participants were HIV-uninfected patients with negative results of HIV antibody screening tests. After excluding 15 patients with insufficient volumes of CSF, a total of 72 HIV-uninfected patients with suspected TBM were finally analyzed. Of these patients, 31 were diagnosed with TBM by the uniform case definition (9), as follows: definite (n = 2), probable (n = 8), or possible (n = 21) TBM. The remaining 41 patients were diagnosed as not-TBM (bacterial [n = 5], viral [n = 20], fungal [n = 4], aseptic meningitis [n = 3], and other [n = 9]). Only the two definite TBM patients gave positive results from mycobacterial culture of the CSF.
The clinical characteristics of all the patients are shown in Table 1. We compared the clinical characteristics of the TBM group, including definite and probable TBM (n = 10) and possible TBM (n = 21), and the not-TBM (n = 41) group according to the uniform case definition. Diabetes mellitus was more common in the TBM group than in the not-TBM group (40% versus 2%; P = 0.004). Also, symptom duration was longer in the TBM group (median, 15.5 versus 7.0 days; P = 0.005), and focal neurologic deficits were more common (70% versus 27%; P = 0.02). There were no statistically significant differences in other clinical characteristics between the two groups.
TABLE 1.
Clinical characteristics of all patients with suspected tuberculous meningitis classified by the uniform case definitiona
| Patient characteristic | Case definition |
P valueb | |||
|---|---|---|---|---|---|
| Definite TBM (n = 2) | Probable TBM (n = 8) | Possible TBM (n = 21) | Not TBM (n = 41) | ||
| Age (median [IQR]) (yrs) | 64 | 58 (46–71) | 50 (36–65) | 49 (34–72) | 0.33 |
| Male | 2 (100) | 6 (75) | 16 (76) | 27 (66) | 0.47 |
| Underlying diseases | |||||
| Human immunodeficiency virus | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |
| Solid tumor | 1 (50) | 1 (13) | 0 (0) | 8 (20) | >0.99 |
| Hematologic malignancy | 1 (50) | 0 (0) | 0 (0) | 0 (0) | 0.20 |
| Organ transplantation | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |
| Diabetes mellitus | 1 (50) | 3 (38) | 0 (0) | 1 (2) | 0.004 |
| Liver cirrhosis | 0 (0) | 2 (25) | 0 (0) | 1 (2) | 0.09 |
| Chronic kidney disease | 0 (0) | 0 (0) | 0 (0) | 1 (2) | >0.99 |
| Rheumatologic disease | 0 (0) | 0 (0) | 0 (0) | 1 (2) | >0.99 |
| Previous history of tuberculosis | 0 (0) | 3 (38) | 1 (5) | 2 (5) | 0.046 |
| Clinical symptoms | |||||
| Symptom duration (median [IQR]) (days) | 16.5 | 15.5 (7.8–27.5) | 8.0 (5.0–13.0) | 7.0 (3.0–11.0) | 0.005 |
| Symptom duration of ≥5 days | 2 (100) | 8 (100) | 16 (76) | 23 (56) | 0.009 |
| Systemic symptoms suggestive of TB | 0 (0) | 1 (13) | 0 (0) | 0 (0) | 0.20 |
| Recent close contact with PTB patients | 0 (0) | 0 (0) | 1 (5) | 1 (2) | >0.99 |
| Focal neurological deficit | 1 (50) | 6 (75) | 10 (48) | 11 (27) | 0.02 |
| Cranial nerve palsy | 0 (0) | 1 (13) | 1 (5) | 0 (0) | 0.20 |
| Altered mental status | 1 (50) | 2 (25) | 10 (48) | 17 (42) | 0.72 |
| CSF profile | |||||
| Clear appearance | 1 (50) | 0 (0) | 0 (0) | 2 (5) | 0.49 |
| White blood cell count (median [IQR]) (no./µl) | 595.5 | 156.5 (33.5–352.8) | 140.0 (65.5–350.0) | 82.0 (30.5–237.5) | 0.36 |
| Lymphocyte (median [IQR]) (%) | 88.5 | 67.5 (31.5–87.8) | 78.0 (62.0–87.5) | 63.0 (40.5–85.5) | 0.42 |
| Protein concentration (median [IQR]) (mg/dl) | 299.1 | 136.3 (47.7–370.1) | 134.0 (91.7–182.0) | 83.8 (39.2–153.0) | 0.06 |
| CSF/plasma glucose ratio, median (IQR) | 0.41 | 0.34 (0.21–0.41) | 0.43 (0.40–0.52) | 0.51 (0.40–0.61) | 0.004 |
| CSF, ADA (median [IQR]) (IU/liter) | 17 | 7.7 (2.3–14.4) | 9.6 (6.9–11.4) | 3.7 (2.0–6.5)c | 0.01 |
| CSF, AFB stain | 0 (0) | 0 (0) | 2/20 (10)d | 0/38 (0)e | |
| CSF, mycobacterial culture | 2 (100) | 0 (0) | 0/20 (0)d | 0/38 (0)e | 0.04 |
| CSF, M. tuberculosis PCR | 0 (0) | 0/7 (0) | 0/20 (0) | 0/35 (0) | |
| CSF, Xpert MTB/RIF assay | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |
| Brain imaging results | |||||
| Hydrocephalus | 0 (0) | 3 (38) | 2 (10) | 4 (10) | 0.13 |
| Basal meningeal enhancement | 0 (0) | 3 (38) | 0 (0) | 0 (0) | 0.006 |
| Tuberculoma | 0 (0) | 2 (25) | 0 (0) | 0 (0) | 0.04 |
| Infarct | 0 (0) | 2 (25) | 3 (14) | 2 (5) | 0.17 |
| Precontrast basal hyperdensity | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |
| Total diagnostic score by a uniform case definition (median [IQR]) | 13 | 12 (12–13) | 9 (8–10) | 4 (3–7) | <0.001 |
Data are given as the number (percentage) of patients unless otherwise indicated. TBM, tuberculous meningitis; IQR, interquartile range; TB, tuberculosis; PTB, pulmonary tuberculosis; CSF, cerebrospinal fluid; ADA, adenosine deaminase; AFB, acid-fast bacilli.
P values were evaluated by comparing the TBM group and the not-TBM group. The TBM group included definite and probable TBM patients.
ADA tests were done in 40 of 41 not-TBM patients.
AFB staining and mycobacterial culture tests were available for 20 of the 21 possible TBM patients.
AFB staining and mycobacterial culture tests were available for 38 of the 41 not-TBM patients.
CSF profiles and brain imaging.
All of the study participants underwent CSF studies and brain imaging. When we compared the CSF profiles in the TBM and not-TBM group, there were statistically significant differences in glucose/plasma glucose ratio (0.35 versus 0.51; P = 0.004) and adenosine deaminase (ADA) (10.6 versus 3.7 IU/liter; P = 0.01). Acid-fast bacilli (AFB) microscopy, MTB PCR, and the Xpert MTB/RIF assay of the CSF samples yielded negative results in all tested patients. In brain imaging, basal meningeal enhancement (30% versus 0%; P = 0.006) and tuberculoma (20% versus 0%; P = 0.04) were more common in the TBM group. There were no significant differences between the two groups in other CSF profiles or brain imaging (Table 1).
Diagnostic performance of the SLIM assay.
The diagnostic performances of the SLIM assay and conventional tests are presented in Table 2. The sensitivity of the SLIM assay was 100% (95% confidence interval [CI], 69 to 100) compared with definite or probable TBM by the uniform case definition, and the SLIM assay was more sensitive than mycobacterial culture (20% [95% CI, 3 to 56], P = 0.008) and the Xpert MTB/RIF assay (0% [95% CI, 0 to 31]; P = 0.002) (Table 2).
TABLE 2.
Diagnostic performances for tuberculous meningitis of the SLIM assay, mycobacterial culture, and the Xpert MTB/RIF assay
| Diagnostic performance metricb | Metric value (95% CI; n/N) fora
: |
||
|---|---|---|---|
| SLIM assay | Mycobacterial culture | Xpert MTB/RIF assay | |
| Sensitivity for diagnosis of definite and probable TBM | 100 (69–100; 10/10)c,d | 20 (3–56; 2/10)c | 0 (0–31; 0/10)d |
| Specificity for diagnosis of not-TBM | 100 (91–100; 0/41) | 100 (89–100; 0/38) | 100 (91–100; 0/41) |
| Specificity for diagnosis of possible and not-TBM | 92 (82–97; 5/62)e,f | 100 (92–100; 0/58)e | 100 (94–100; 0/62)f |
CI, confidence interval; n/N, number of patients with a positive test result/number of patients tested.
TBM diagnosis categorized by the uniform case definition.
P value = 0.008 between SLIM assay and mycobacterial culture.
P value = 0.002 between SLIM assay and Xpert MTB/RIF assay.
P value = 0.13 between SLIM assay and mycobacterial culture.
P value = 0.06 between SLIM assay and Xpert MTB/RIF assay.
None of the 41 patients with not-TBM gave positive results in the SLIM assay (specificity, 100%; 95% CI, 91 to 100). Of the 21 possible TBM patients by the uniform case definition, 5 gave positive results in the SLIM assay. If we assume that these 5 patients were false positives, then the diagnostic specificity of the SLIM assay was 92% (95% CI, 82 to 97) for ruling in possible and not-TBM diagnoses (Table 2). Of these 5 patients, 4 were treated with anti-TB medication for possible TBM for 6 to 9 months, and they recovered without any complications or sequelae after the treatment. The remaining patient received anti-TB treatment but was not followed up after 3 months of anti-TB treatment. The clinical characteristics of SLIM-positive (n = 5) and SLIM-negative (n = 16) patients with possible TBM are shown in Table S1. However, there were no significant differences in the clinical characteristics between SLIM-positive and SLIM-negative patients with possible TBM.
Sequencing of Mycobacterium tuberculosis DNA extracted by the SLIM assay.
We also carried out Sanger sequencing to confirm the presence of M. tuberculosis DNA in the 15 CSF samples that gave positive results in the SLIM assay followed by real-time M. tuberculosis PCR. IS6110 (an insertion element found exclusively in the M. tuberculosis complex), and the rpoB (β subunit of RNA polymerase), and katG (the catalase-peroxidase gene) genes were tested. All of the CSF samples were positive for M. tuberculosis DNA. In addition, none of the samples contained mutations in rpoB, which are associated with drug resistance.
DISCUSSION
In this study, we evaluated the diagnostic performance of the newly developed SLIM assay in 72 suspected TBM patients. The sensitivity of the SLIM assay was 100% compared with definite or probable TBM by the uniform case definition (9). The sensitivity of the SLIM assay was superior to that of other conventional diagnostic tests, such as mycobacterial culture (20%) and the Xpert MTB/RIF assay (0%), without significant loss of specificity. The sensitivities of the conventional diagnostic tests in our study were much lower than those in previous studies (mycobacterial culture, 30 to 67%; Xpert MTB/RIF, assay 43 to 72%; MTB PCR, 41%, respectively), which were performed mainly in HIV-infected patients (4, 5, 7, 14). In previous studies, the sensitivities of mycobacterial culture and the Xpert MTB/RIF assay in HIV-uninfected patients were lower than those in HIV-infected patients due to the paucibacillary nature of HIV-uninfected TBM patients (4, 5). However, the previous study showed that HIV infection status was not independently associated with microbiologically confirmed (definite) TBM (odds ratio [OR], 2.07; 95% CI, 0.71 to 5.99; P = 0.18) in multivariate analysis (15). Therefore, it is uncertain how much HIV-uninfected status contributed to low sensitivities of the conventional tests in our study. It is worth noting that we used 1 ml of noncentrifuged CSF for the Xpert MTB/RIF assay, while the previous studies used 0.2 ml – 2.0 ml centrifuged CSF (6, 15). Therefore, the low volume of noncentrifuged CSF may contribute to low sensitivity of the Xpert MTB/RIF assay, while the SLIM assay, which does not need bulky instruments, revealed promising results from the same volume of noncentrifuged CSF.
We developed the innovative SLIM assay and tried to use it to detect M. tuberculosis in CSF for diagnosing TBM. In general, only a small volume of CSF is actually used for the conventional assays. From a given sample of CSF, about 100 µl for MTB PCR and 500 µl for the Xpert MTB/RIF assay are usually used, and the remaining fluid is discarded. On the other hand, the SLIM assay can concentrate the entire volume of each CSF sample, and DNA is extracted by the microfluidic chip from the entire volume of available CSF. In addition, our system does not require additional instruments or specific antibodies or probes for sample processing. Furthermore, the assembled microfluidic chips that we used to streamline sample processing are potentially scalable at low cost, and this would increase the surface area for pathogen capture without increasing the size of the device (8). Thus, we can perform this test in a simple and efficient way, and we expect that it will be usable in resource-limited countries.
In our study, 5 of 21 possible TBM patients by the uniform case definition gave positive results both in the SLIM assay and by Sanger sequencing. We might consider these results to be false positives resulting from contamination by DNA fragments of dead M. tuberculosis in patients who had been treated for TB or to be due to assay error (16). However, these 5 patients did not have histories of previous TB. In addition, because the SLIM assay uses a single platform for enrichment and extraction of pathogens, the risk of external contamination should be reduced (8). Thus, we think that the likelihood of false positives is extremely low. If, in spite of this low possibility, we assume that these 5 patients gave false-positive results due to verification bias, then the diagnostic specificity of the SLIM assay for ruling in definite or probable TBM based on the uniform case definition would be 92% (95% CI, 82 to 97%). Therefore, the sensitivity of the SLIM assay would be higher than those of other conventional tests for TBM without significant loss of the specificity. Interestingly, clinicians actually started anti-TB treatment in all of the enrolled patients who were eventually classified as definite, probable, and possible TBM. Of these 31 TBM patients (2 definite, 8 probable, and 21 possible TBM), 2 of the possible TBM patients (with negative results for the SLIM assay) stopped anti-TB treatment early because of adverse events and did not show any clinical deterioration, and 2 other possible TBM patients (one with a positive result in the SLIM assay and one with a negative result) were not followed up after the start of anti-TB treatment. Therefore, we assume that a substantial proportion of the patients who fall into the “possible” TBM category are exposed to unnecessary anti-TB drug toxicity in real practice because there are no rapid, sensitive tests to rule out TBM. Further studies of the clinical outcomes of empirical anti-TB treatment are needed based on rapid, sensitive tests for TBM, such as the SLIM assay and the Xpert MTB/RIF Ultra assay. In addition, we proposed a new composite standard that includes these rapid, ultrasensitive tests in patients with suspected TBM for the use of future clinical studies and improving patient care.
Our study has some limitations. First, it included only a small number of patients who agreed to additional tests of their CSF in a single large medical center, and 15 of the original 87 patients were excluded due to insufficient CSF volume. So, there could be selection bias. The small sample size and biased recruitment may partially explain our low detection rates by mycobacterial culture and the Xpert MTB/RIF assay. Second, there is no gold standard for rapidly diagnosing TBM. There is the uniform case definition for TBM, but this diagnostic criterion is not adequate to rule out or rule in TBM in the real clinical setting. In our study, 5 of 21 possible TBM patients by the uniform case definition gave positive results in the SLIM assay. However, it is not clear whether these five positive cases were true or false positives because gold standard tests alone would underdiagnose TBM cases. We performed variable numbers of tandem repeat genotyping with five M. tuberculosis loci to evaluate the possibility of cross-contamination leading to false-positive results of the SLIM assay. Unfortunately, we could not obtain meaningful results because there were no available mycobacterial culture isolates from these patients (data not shown). Finally, we did not perform the Xpert MTB/RIF Ultra assay in our cohort, and this test may have higher sensitivity than current diagnostic tests (7). Thus, we could not directly compare the diagnostic performance of the SLIM assay with this assay.
In conclusion, clinicians have problems with diagnosing TBM, especially in HIV-uninfected patients, because of the low sensitivity of conventional diagnostic tests. The newly developed SLIM assay followed by real-time PCR demonstrated a very high sensitivity and specificity with small samples from 10 cases of definite or probable TBM. Therefore, we could expect the SLIM assay to be used in resource-limited countries because of its simple and efficient nature; however, this system is not currently available in these countries. Further studies are needed to confirm this finding and to compare SLIM with mycobacterial culture, Xpert MTB/RIF, and Xpert MTB/RIF Ultra assays in a larger prospective cohort of patients with suspected TBM, including both HIV-infected and HIV-uninfected cases.
Supplementary Material
ACKNOWLEDGMENTS
This work was supported by grants from the National Research Foundation of Korea funded by the Ministry of Science, ICT & Future Planning (grants NRF-2018R1D1A1A09082099 and NRF-2017R1A2B4005288) and was also supported by grants (2018-7034 and 2018-7040) from the Asan Institute for Life Sciences, Asan Medical Center, Seoul, Republic of Korea.
We have no potential conflicts of interest.
Footnotes
Supplemental material for this article may be found at https://doi.org/10.1128/JCM.01975-18.
REFERENCES
- 1.Leonard JM. 2017. Central nervous system tuberculosis, p 331–341. In Schlossberg D. (ed), Tuberculosis and nontuberculous mycobacterial infections, 7th ed ASM Press, Washington, DC. [Google Scholar]
- 2.Rock RB, Olin M, Baker CA, Molitor TW, Peterson PK. 2008. Central nervous system tuberculosis: pathogenesis and clinical aspects. Clin Microbiol Rev 21:243–261. doi: 10.1128/CMR.00042-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.World Health Organization. 2013. Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary TB in adults and children. World Health Organization, Geneva, Switzerland. [Google Scholar]
- 4.Patel VB, Theron G, Lenders L, Matinyena B, Connolly C, Singh R, Coovadia Y, Ndung’u T, Dheda K. 2013. Diagnostic accuracy of quantitative PCR (Xpert MTB/RIF) for tuberculous meningitis in a high burden setting: a prospective study. PLoS Med 10:e1001536. doi: 10.1371/journal.pmed.1001536. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Nhu NTQ, Heemskerk D, Thu DDA, Chau TTH, Mai NTH, Nghia HDT, Loc PP, Ha DTM, Merson L, Thinh TTV, Day J, Chau N. v V, Wolbers M, Farrar J, Caws M. 2014. Evaluation of GeneXpert MTB/RIF for diagnosis of tuberculous meningitis. J Clin Microbiol 52:226–233. doi: 10.1128/JCM.01834-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bahr NC, Tugume L, Rajasingham R, Kiggundu R, Williams DA, Morawski B, Alland D, Meya DB, Rhein J, Boulware DR. 2015. Improved diagnostic sensitivity for tuberculous meningitis with Xpert(®) MTB/RIF of centrifuged CSF. Int J Tuber Lung Dis 19:1209–1215. doi: 10.5588/ijtld.15.0253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bahr NC, Nuwagira E, Evans EE, Cresswell FV, Bystrom PV, Byamukama A, Bridge SC, Bangdiwala AS, Meya DB, Denkinger CM, Muzoora C, Boulware DR, Williams DA, Taseera K, Nyehangane D, Ivan M, Orikiriza P, Rhein J, Hullsiek KH, Musubire A, Pastick K, Nabeta P, Mwesigye J, Rajasingham R. 2018. Diagnostic accuracy of Xpert MTB/RIF Ultra for tuberculous meningitis in HIV-infected adults: a prospective cohort study. Lancet Infect Dis 18:68–75. doi: 10.1016/S1473-3099(17)30474-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jin CE, Koo B, Lee EY, Kim JY, Kim SH, Shin Y. 2018. Simple and label-free pathogen enrichment via homobifunctional imidoesters using a microfluidic (SLIM) system for ultrasensitive pathogen detection in various clinical specimens. Biosens Bioelectron 111:66–73. doi: 10.1016/j.bios.2018.04.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Marais S, Thwaites G, Schoeman JF, Torok ME, Misra UK, Prasad K, Donald PR, Wilkinson RJ, Marais BJ. 2010. Tuberculous meningitis: a uniform case definition for use in clinical research. The Lancet Infect Dis 10:803–812. doi: 10.1016/S1473-3099(10)70138-9. [DOI] [PubMed] [Google Scholar]
- 10.Piatek AS, Tyagi S, Pol AC, Telenti A, Miller LP, Kramer FR, Alland D. 1998. Molecular beacon sequence analysis for detecting drug resistance in Mycobacterium tuberculosis. Nat Biotechnol 16:359–363. doi: 10.1038/nbt0498-359. [DOI] [PubMed] [Google Scholar]
- 11.Wada T, Maeda S, Tamaru A, Imai S, Hase A, Kobayashi K. 2004. Dual-probe assay for rapid detection of drug-resistant Mycobacterium tuberculosis by real-time PCR. J Clin Microbiol 42:5277–5285. doi: 10.1128/JCM.42.11.5277-5285.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.World Health Organization. 2014. Xpert MTB/RIF implementation manual: technical and operational “how-to”: practical considerations. World Health Organization, Geneva, Switzerland. [PubMed] [Google Scholar]
- 13.Denkinger CM, Schumacher SG, Boehme CC, Dendukuri N, Pai M, Steingart KR. 2014. Xpert MTB/RIF assay for the diagnosis of extrapulmonary tuberculosis: a systematic review and meta-analysis. Eur Respir J 44:435–446. doi: 10.1183/09031936.00007814. [DOI] [PubMed] [Google Scholar]
- 14.Rufai SB, Singh A, Singh J, Kumar P, Sankar MM, Singh S, Nigam K, Sharma D, Gupta AK, Rawat VS. 2017. Diagnostic usefulness of Xpert MTB/RIF assay for detection of tuberculous meningitis using cerebrospinal fluid. J Infect 75:125–131. doi: 10.1016/j.jinf.2017.04.010. [DOI] [PubMed] [Google Scholar]
- 15.Heemskerk AD, Donovan J, Thu DDA, Marais S, Chaidir L, Dung VTM, Centner CM, Ha VTN, Annisa J, Dian S, Bovijn L, Mai NTH, Phu NH, Chau NVV, Ganiem AR, Van CT, Geskus RB, Thuong NTT, Ruslami R, Meintjes G, van Crevel R, Wilkinson RJ, Thwaites GE. 2018. Improving the microbiological diagnosis of tuberculous meningitis: a prospective, international, multicentre comparison of conventional and modified Ziehl-Neelsen stain, GeneXpert, and culture of cerebrospinal fluid. J Infect 77:509–515. doi: 10.1016/j.jinf.2018.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Theron G, Venter R, Smith L, Esmail A, Randall P, Sood V, Oelfese S, Calligaro G, Warren R, Dheda K. 2018. False-positive Xpert MTB/RIF results in retested patients with previous tuberculosis: frequency, profile, and prospective clinical outcomes. J Clin Microbiol 56. doi: 10.1128/jcm.01696-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.