Diagnostic value of the neutrophil lymphocyte ratio in discrimination between tuberculosis and bacterial community acquired pneumonia: A meta-analysis.

Horieh Shojaan: Conceptualization, Investigation, Data curation. Niusha Kalami: Investigation. Masoumeh Ghasempour Alamdari: Investigation. Seyed Matin Emami Alorizy: Investigation, Writing – original draft. Arshin Ghaedi: Writing – review & editing. Aida Bazrgar: Writing – original draft, Writing – review & editing. Monireh Khanzadeh: Conceptualization, Writing – original draft. Brandon Lucke-Wold: Writing – review & editing. Shokoufeh Khanzadeh: Conceptualization, Supervision.

This review was reported and conducted based on Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. Two independent researchers conducted a thorough literature search up to July 20, 2023 using Scopus, PubMed, and Web of Science. Adequate observational studies were included that assessed the relationship between NLR levels and TB were recognized. The following search phrases were displayed, along with free text words and mesh terms: “Tuberculosis,” “community-acquired pneumonia,” “neutrophil to lymphocyte ratio,” and “NLR.” To find any other possible missing articles, included literature was also subjected to a manual reference search. Two independent researchers evaluated the relevant articles based on the title or abstract of each record. Following the first screening step, full texts were reviewed based on the eligibility criteria. Differences of opinion were settled via debate or via contact with a tie breaking third author. Studies were deemed eligible for meta -analysis if they met the following requirements: (1) the mean concentration and its standard difference (SD) of the NLR level could be acquired from original studies for both B-CAP and TB patients or could be estimated using the indirect methods offered; (2) provided data for the NLR's standardized mean difference (SMD) and 95% confidence limit (CI); (3) the involvement of NLR in distinction between TB and B-CAP. The exclusion criteria were as follows: (1) comments, reviews, or meta -analysis lacking original data for meta -analysis; (2) research lacking appropriate data for abstraction; and (3) publications that were duplicated or irrelevant. In addition, we conducted a qualitative review on studies, investigating different aspects of association between NLR and TB, but not included in quantitative analysis, for the reasons such as incomplete data and different control groups. Following data was extracted from included studies: first author, study sample size, publication year, mean and SD of NLR in each group, the bacterial pathogens, infecting patients in CAP group, and research country as well as design was among the variables that were extracted. The Newcastle-Ottawa Quality Assessment Scale (NOS) was used to assess the included study's quality. For observational studies, three aspects of this scale were assessed to include a total of nine points: comparability, study selection, and outcome. Publications with a score of 7 to 9, 3 to 6, and 0 to 3 were classified as high, moderate, or low quality, respectively. STATA software (version 12.0; Stata Corporation) was used for all analyses. To calculate the pooled effect estimates, data were gathered as means ± standard deviation (SD). For analyzing the relationship between TB and NLR levels, a meta -analysis was undertaken using standardized mean difference (SMD) and 95% confidence limit (CI). Because of between-study heterogeneities, the random effects model was applied for all meta -analyses. To evaluate the diagnostic significance of NLR, we used the “metandi” command, which calculated the positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio (DOR), and pooled specificity and sensitivity. A summary receiver operating characteristic (SROC) curve was also developed. We used sensitivity analysis to assess the stability of the results by excluding each study individually from analysis. The inter-study heterogeneity was examined using Cochran's Q test and I 2 statistic. High heterogeneity, moderate heterogeneity, low heterogeneity, and insignificant heterogeneity were all denoted by I 2 values of 76–100%, 51–75%, 26–50%, and 0–25%, respectively. We also performed subgroup analyses based on different accessible factors, such as study continent (Africa vs. Asia), sample size (≥200 vs. <2 0 0), and NOS score (high quality vs. moderate quality). This was done to further evaluate the probable cause of inter-study variability. Visual assessment of the funnel plot symmetry, as well as Eggers and Beggs’ tests were conducted to determine publication bias. A P value<0.05 suggested significant publication bias.

PRISMA flow diagram was shown in Fig. 1 . After multiple stages of screening of the 564 studies found in the database, ten papers were included in the meta -analysis [23] , [24] , [25] , [26] , [27] , [28] , [29] , [30] , [16] . Fig. 1 Flow chart of search and study selection. Flow chart of search and study selection. In total, 37 articles were included, among them, ten articles were included in the meta -analysis. This includes 730 patients with TB and 725 patients with B-CAP. All of them were retrospective. Three of them were conducted in South Korea [26] , [30] , [16] . Others were conducted in Tunisia [29] , India [27] , Turkey [24] , Ethiopia [23] , Egypt [25] , China and Bangladesh [28] . The sample size ranged from 60 [27] to 269 [26] . Three studies had large sample sizes [26] , [30] , [16] and others were small studies. Table 1 shows the overall characteristics of the included articles. Table 1 General characteristics of included studies. Author Year Design Country Bacterial pathogens TB group B-CAP group Se Sp NOS Score N NLR N NLR Um 2012 P Korea Not declared 112 3.67± 2.12 94 14.64± 9.72 91 82 8 El-Emshaty 2017 R Egypt SP: 41.1%, SA: 22.2%, KP: 13.3%, Enterobacter spp.: 5.6%, Citrobacter spp.: 3.4%, Moraxella catarrhalis:2.2%, Pseudomonas aeruginosa: 2.2%, Haemophilus influenza: 2.2%, Mixed SA with Gram negative: 2.2%, Other Gram negative:5.6%, 68 3.50± 1.43 90 11.85± 3.29 91 100 7 Yoon 2013 P Korea SP: 62.96%, KP: 14.81%, Escherichia coli: 7.4%, methicillin-sensitive SA: 3.7%, Acinetobacter baumannii: 3.7 %, methicillin-resistant SA: 3.7%, and Stenotro-phomonas maltophila: 3.1% 112 3.67± 2.12 94 14.64± 9.72 91 82 8 Tabaubi 2018 R Tunisia Not declared 105 5.30± 3.00 80 8.10± 7.00 – – 7 Berhane 2019 R Ethiopia Pseudomonas aeruginosa: 37%, KP: 23.3%, and SP: 17.8% 73 4.63± 3.93 73 2.96± 2.34 – – 7 Jeon 2019 P Korea Not declared 110 4.37± 2.68 159 12.42± 9.26 84 79 6 Kumar 2020 R India SP: 55.55%, KP:11.11%, Escherichia coli 16.66%, Methicillin-Sensitive SA: 11.11%, and Methicillin-Resistant SA: 5.5% 30 3.73± 2.32 30 14.43± 9.23 – – 7 Bozdemir 2021 R Turkey Not declared 50 0.62± 0.36 50 8.02± 5.61 – – 6 Mosleh 2022 R Bangladesh Not declared 70 3.16± 1.90 85 5.02± 1.20 – – 7 Chai 2023 R China Not declared 60 3.47± 1.42 70 2.05± 1.02 73 76 6 N: Number; NLR: Neutrophil to lymphocyte ratio; TB: Tuberculosis, B-CAP: Bacterial community acquired pneumonia; Se: Sensitivity; Sp: Specificity; NOS: Newcastle-Ottawa Scale; R:Retrospective; P: Prospective; SP: Streptococcus pneumonia; SA: Staphylococcus aureus; KP: Klebsiella pneumonia General characteristics of included studies. We found that patients with TB had significantly lower levels of NLR compared to those with B-CAP (SMD = -1.09, 95 %CI = −1.78- −0.40, P = 0.002, Fig. 2 .A). Fig. 2 Forest plots. A) Meta-analysis of differences in NLR level between patients with TB and those with B-CAP; B) Subgroup analysis of differences in NLR level between patients with TB and those with B-CAP, according to sample size; C) Subgroup analysis of differences in NLR level between patients with TB and those with B-CAP, according to NOS score; D) Subgroup analysis of differences in NLR level between patients with TB and those with B-CAP, according to study continent. Forest plots. A) Meta-analysis of differences in NLR level between patients with TB and those with B-CAP; B) Subgroup analysis of differences in NLR level between patients with TB and those with B-CAP, according to sample size; C) Subgroup analysis of differences in NLR level between patients with TB and those with B-CAP, according to NOS score; D) Subgroup analysis of differences in NLR level between patients with TB and those with B-CAP, according to study continent. In the subgroup analysis according to sample size, we found that patients with TB had significantly lower level of NLR compared to those with B-CAP in particular for the large studies (SMD = -1.44, 95 %CI = −1.81, −1.08, P < 0.001). This pattern was not maintained in the small studies (SMD = -0.94, 95 %CI = -1.98, 0.08, P = 0.07, Fig. 2 .B). In the subgroup analysis based on quality, we found that patients with TB had significantly lower levels of NLR compared to those with B-CAP seen in the high quality studies (SMD = -1.25, 95 %CI = -2.07, −0.42, P = 0.003). This pattern was not seen in studies with moderate quality (SMD = -0.86, 95 %CI = -2.30, 0.57, P = 0.23, Fig. 2 .C). In the subgroup analysis according to study continent, we found that patients with TB had significantly lower level of NLR compared to those with B-CAP in studies performed in Asia (SMD = −1.37, 95 %CI = −2.13, −0.61, P < 0.001), but not those in African populations (SMD = −0.02, 95 %CI = -1.06, 1.02, P = 0.97, Fig. 2 .D.). As seen in Fig. 3 and Table S1 , no significant change in final result is seen after excluding every single study individually from analysis. It suggested the reliability of our findings. Fig. 3 Sensitivity analysis of differences in NLR level between patients with TB and those with B-CAP. Sensitivity analysis of differences in NLR level between patients with TB and those with B-CAP. As seen in Fig. 4 . A, the funnel plot for studying publication bias also did not appear promising, because several points existed outside the funnel, which challenges the aforementioned un-biasedness of the data. However, asymmetry of the funnel plot alone does not accurately predict publication bias [31] . So we used advanced statistical analysis and found that there was no significant publication bias among the included studies (Egger’s test P = 0.48, Begg’s test P = 0.72). Fig. 4 Funnel plot assessing publication bias. Funnel plot assessing publication bias. The pooled sensitivity was 0.86 (95% CI = 0.80, 0.91), and the pooled specificity was 0.88 (95% CI = 0.69, 0.95). The pooled positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio (DOR) of NLR were 7.27 (95 %CI = 2.53, 20.94), 0.14 (95 %CI = 0.08, 0.24), and 49.22 (95 %CI = 11.38, 212.93), respectively ( Fig. 4 . B). Twenty articles were included in qualitative review. Table S2 shows the general characteristics of these studies. Due to the challenge of estimating the respiratory isolation period in instances with pulmonary tuberculosis (PTB) disorder, Ramirez-Hidalgo et al. [32] conducted a retrospective evaluation of 229 smear-positive patients to establish a predictive score considering risk factors related to positive sputum cultures. This was done after four weeks of therapy. Multivariate analyses revealed strong associations between persistent positive cultures and NLR > 3.5 (OR, 2.22; 95% CI, 1.24–3.99; P = 0.008), number of affected lung lobules > 2 (OR, 1.95; 95% CI, 1.08–3.54; P = 0.027), smoking (OR, 2.44; 95% CI, 1.36–4.37; P = 0.003), and fever at time of consultation (OR, 1.87; 95% CI, 1.02–3.41; P = 0.042). Moreover, for each variable, one score was allocated. Scoring four points represented a high probability (82%) of cultures remaining positive after four weeks of medication. Similarly, Suryana et al. [33] carried out a case-control study of 62 patients to detect factors associated with the delayed conversion of sputum after treatment. NLR ≥ 5.065 showed a positive association with an increased risk of sputum conversion in bivariate analysis (OR, 8.23; 95% CI, 2.482–27.317; P < 0.001). However, after confounding factors adjustment, this association disappeared (OR, 3.370; 95% CI, 0.714–15.910; P = 0.125). NLR was one of the parameters used to discriminate uninfected (No TBI) and latently infected contacts (LTBI) from active TB patients. Higher values of white blood cells and neutrophils were observed in active TB patients compared to contact groups. Interestingly, this pattern contrasted with circulating lymphocytes (P < 0.0001). Therefore, NLR displayed higher levels in patients with active TB (P < 0.0001) [34] . Likewise, another investigation examined various pre-treatment laboratory components among instances with and without PTB. Regarding NLR, values were higher in the TB group compared to the controls (5.06 vs. 1.89, P = 0.001) with an AUC of 0.88 (78% specificity and 79% sensitivity) [35] . Sulastri et al. [36] discussed levels of neutrophils, lymphocytes, and NLR in cases with PTB with or without HIV co-infection (TB/HIV). Neutrophils and Lymphocytes were lower in TB/HIV group compared to PTB (P = 0.003, 0.001, respectively). Statistical significance was noticed among NLR for TB/HIV compared with PTB patients without HIV co-infection (6.05 ± 2.67 and 5.16 ± 1.88, P = 0.041). Moreover, NLR for TB screening among HIV-infected instances was inspected in two prospective studies. Miyahara et al. [37] found the NLR to be a predictive modality to assess the risk of TB in HIV-infected individuals. Statistically, high NLR ratios (>2) were related to higher TB risks of development (adjusted hazard ratio (aHR), 2.19; 95% CI, 1.23–3.90; P = 0.007). Gersh et al. (7) screened PTB in adults with anti-retroviral-treated HIV. For NLR ≥ 1.5 patients where isoniazid was initiated before enrollment, statistical analysis (% (95% CI)) was as follows: 40 (5–85) for sensitivity, 80 (76–84) for specificity, 3 (0–10) for positive predictive value, and 99 (97–100) for negative predictive value. In an appraisal of 70 PTB cases and 50 healthy individuals, the associations between the severity of PTB and inflammatory biomarkers were investigated. Based on the disease spread on chest radiography, PTB cases were categorized from mild to moderate to advanced stages. NLR of the PTB cohort displayed elevated amounts compared to controls (3.9 ± 2.6 vs. 1.6 ± 0.3, P < 0.001). This ratio was 4.7 ± 2.8 in advanced PTB verse 3.1 ± 2.2 in the mild to moderate group (P = 0.009) [8] . Ardalan et al. [38] found higher amounts of NLR in PTB than in controls (4.65 vs. 1.83, P  0.05). Chen et al. [39] analyzed laboratory blood results, including NLR, among non-spinal and spinal tuberculosis cases. Comparisons were also conducted in spinal tuberculosis cases considering the Oswestry disability index (ODI) scores. Instances with scores ≥ 20 were classified as a high ODI group, and scores < 20 as a low ODI group. The level of NLR showed statistical differences in non-spinal and spinal tuberculosis cohorts (2.84 ± 1.68 vs. 3.69 ± 3.20, respectively, P < 0.001). Moreover, NLR values were elevated in the high ODI group compared to the low ODI group (4.80 ± 7.89 VS. 3.51 ± 3.27, P = 0.044). Seventy-four patients and 54 healthy control patients were retrospectively reviewed to assess NLR diagnostic value in TB spondylitis instances. Accordingly, NLR was significantly higher in patients with TB than in controls (3.24 ± 1.38 vs. 1.58 ± 0.59, P < 0.001) [40] . Additionally, this ratio was used as a potential determinant for distinguishing between pyogenic spinal infection (PSI) and spinal tuberculosis (STB). Lower values of NLR were attained for STB patients (median (IQR)): 3.85 (2.70–5.71)) than those in PSI cases (median (IQR): 10.82 (6.79– 17.62)) with a p-value of < 0.001. The cut-off value of 6.742 was optimal suggesting that the NLR level is important in discriminating PSI and STB [41] . NLR was also utilized as a biomarker to differentiate the etiology of exudative pleural effusions (PEs) between malignancy (MPEs) and TB (TBPEs). This ratio showed statistical differences between these two causes (4.40 (±9.75) for TBPEs vs. 4.94 (±4.19) for MPEs, P = 0.001). Regarding NLR quartiles (q), the distribution of cases with MPEs was higher in q2 (2.3–4.95) and q3 (>4.95, P = 0.0003) [42] . Regarding PTB retreatment, the estimation of NLR revealed that ratios ≥ 2.53 prior to therapy were predictive of subsequent retreatment (OR, 1.994; 95% CI, 1.116–3.564; P = 0.020). Furthermore, in the multivariate analysis, this association remained significant (OR, 2.409; 95% CI, 1.212–4.788; P = 0.012) [43] . Another study evaluated NLR levels among patients with active and subclinical TB, which was defined as utilizing microbiologic or radiologic assays to detect TB in symptom-free individuals [44] . Levels were higher in active TB cases compared to the subclinical TB cohort (4.83 ± 14.56 vs. 2.08 ± 1.03, P < 0.001) [45] . Han et al. [46] established pre-treatment NLR as an efficient prognostic biomarker for overall outcomes in miliary TB. By comparing non-survivors to survivors during hospitalization, NLR was markedly higher in non-survivors (13.6 ± 13.8 vs. 7.8 ± 5.9, P = 0.010). Furthermore, elevated NLR was statistically associated with poor clinical outcomes, including acute respiratory distress syndrome (ARDS) development (adjusted odds ratio (aOR), 1.19; 95% CI, 1.05–1.34; P = 0.008), in-hospital death (adjusted hazard ratio (aHR), 1.08; 95% CI, 1.03–1.13; P = 0.002), and one-year mortality (aHR, 1.08; 95% CI, 1.05–1.13; P < 0.001). The one-year survival rate declined among patients with NLR ≥ 5 when assessed with the Kaplan-Meier curves (log-rank test, P = 0.005). In one retrospective investigation, the PDE NLR values at presentation were higher in the MSSA group compared to others (p < 0.001). To determine TB/NTM species, analyses displayed that for every one-point rise in the PDE NLR at day zero, OR decreased by 5% (OR, 0.95; 95% CI, 0.92–0.97; P < 0.001) [47] . Two further studies analyzed NLR as a prognostic and diagnostic method in the pediatric and adolescent populations. NLR was compared within four groups of children with TB disease, TB infection, non-TB lower respiratory tract infection (nTB-LRTI), and TB exposure. Compared to TB exposed individuals ((median (IQR)): 0.81 (0.64, 1.29)), nTB-LRTI (0.31 (0.11, 0.97)), and TB infected individuals (1.08 (0.67, 1.50)), levels of NLR in children with TB disease was the highest (2.03 (1.23, 2.23)) [48] . Evaluation of this factor in Tanzanian adolescents with negative interferon-gamma release assays (IGRAs), who enrolled in a trial of booster injection series of DAR-901 [49] , revealed no significant differences between IGRA-positive converters and IGRA-negative controls (0.88 vs. 1.02, P = 0.08) [50] . The link between immunonutritional condition and cavitation in the lung was an object of interest in another retrospective evaluation. NLR ≥ 5 and levels of serum albumin < 3 g/dl achieved significant association with pulmonary cavitation in TB patients (P = 0.014, 0.025, respectively) [51] . Assessing the relation of NLR and Acid-Fast Bacilli (AFB) TB, it was found that higher NLR levels wer seen in positive AFB TB in comparison to negative AFB TB (10.20 ± 9.53 vs. 2.47 ± 1.56, P < 0.001). This was due to the lower percentages of neutrophils in negative AFB TB patients than in positive cases (55.02 ± 9.80 vs. 81.18 ± 8.52, P < 0.001). On the other hand, lymphocytes of negative AFB TB showed higher percentages than positive AFB TB (28.69 ± 12.01 vs. 12.72 ± 7.51, P < 0.001) [52] . In an investigation of 1489 patients, the discriminative power of various blood markers, including NLR, was examined for early differentiation of genitourinary TB (GUTB) from PTB. Considering NLR, statistical significance was observed among patients with culture results including negative sputum and positive urine, sputum−&urine+, (10.0 ± 23.6), and patients with positive sputum and negative urine, sputum+&urine−, (32.5 ± 27.7, P = 0.024) [53] . With respect to extra-pulmonary tuberculosis (EPTB), NLR was a valuable prognostic factor in indicating poor prognosis (PP) and relapse. NLR ≥ 3 was related to relapse in EPTB cases (P = 0.027) [54] . Moreover, NLR ≥ 2.7 showed a predictive value for PP (HR = 1.7; CI95% 1.1–2.9; P = 0.048) [55] . Disease activity in TB and diabetic mellitus patients (TB-DM) was also determined by calculating NLR. Consequently, TB-DM patients with NLR < 2.9 were deemed immunocompetent [56] .

In this systematic review and meta -analysis, we discovered that patients with TB had a significantly lower NLR level compared to those with B-CAP. Thus, these findings suggest the NLR's underlying efficacy in predicting TB. However, there was not any study on the role of NLR in acute tuberculous pneumonia, so a further study investigating this association would be very interesting.

In this systematic review and meta -analysis, we discovered that patients with TB had a significantly lower NLR levels compared to those with B-CAP. The prognosis and severity of several infectious diseases like B-CAP [57] and bacteremia [58] , [59] are related to the NLR levels; as recent studies have indicated. It has been demonstrated that increased NLR values can perform as an indicator of inflammation and worse outcomes [60] . In addition, it has been reported that NLR is a valuable biomarker for assessing the therapeutic response of patients with systemic inflammation [61] . Inflammation plays a crucial role in the pathogenesis of tuberculosis and B-CAP. Yoon et al. found that a NLR < 7 could be helpful in differentiating between tuberculosis and Bacterial-CAP in adults. They discovered that a NLR < 7 was more accurate than a CRP < 7 mg/dL (another inflammatory biomarker) for distinguishing between tuberclosis and bacterial CAP [16] . According to the findings seen in subgroup analysis based on continent, patients with TB had significantly lower level of NLR compared to those with B-CAP for studies performed on Asian patients, but not on African population. NLR can be retrieved quickly by routine laboratory testing, and this straightforward ratio can be determined in various health clinics. It has been demonstrated that the NLR plays a prognostic role in several diseases, such as infectious diseases, chronic obstructive pulmonary disease (COPD), metabolic syndrome, subdural hemorrhage, endometriosis, and end-stage renal disease, Behcet disease, cardiovascular disorders, and cancers like colorectal, kidney, lung, and breast cancers [62] , [63] , [64] , [65] , [66] , [67] , [68] , [69] . In the majority of cancers and inflammatory disorders, the physiological responses of innate immunity to inflammation and malignancies result in an increase in neutrophils and a decrease in lymphocyte numbers [70] . Horne et al. demonstrated that NLR could be a more accurate predictor of inflammation than leukocyte counts [71] . Higher NLR levels were detected in active PTB, which were linked to persistent infection and considerable pulmonary destruction [72] . Therefore these findings resulted in the broad application of NLR levels in the diagnosis and prognosis of inflammatory diseases [73] . Increased neutrophils and decreased lymphocyte levels are normal response of immune reactivity to stress [16] . Neutrophils are the predominant cells engaged in inflammation and immunological responses against bacterial infection [74] . They resume the inflammatory response, which can cause tissue damage [75] . Variation in baseline immune response may be a predisposing factor making certain individuals more susceptible to the downsides of infection. Neutrophils are the main implicated white blood cell in the early phase of M. tuberculosis infection, participating in granuloma formation [76] or pulmonary destruction [77] . High neutrophil levels are associated with increased mortality [78] , positive M. tuberculosis sputum [79] , and delayed negative smear conversion [80] . Furthermore, lymphocytes, especially T-lymphocyte subsets, are important immune cells active against TB infections. Studies have shown that active TB patients had significantly lower percentages of peripheral blood T-cell subsets (CD4 + and CD8 + ), as well as total and central memory CD4 + T cells compared to healthy controls [81] . In the subgroup analysis according to study continent, we found that patients with TB had significantly lower level of NLR compared to those with B-CAP in studies performed on Asian patients, but not on African population. It is possible that the data obtained for African population in biased due to lower number of currently available sample sizes. It also makes it difficult to generalize NLR as TB diagnostic test as high burden areas are not showing similar patterns. Although meta -analyses generally enhance the quality of the available evidence, inherent limitations of our study should be acknowledged. First, due to the insufficient data provided in individual studies, the effect of certain immunosuppressive medications (such as nonsteroidal anti-inflammatory medicines or glucocorticoids) on NLR was not reviewed. Furthermore, NLR is probably impacted by other systemic TB comorbidities, which may restrict its clinical use in individuals with concurrent illnesses. Additionally, baseline immune response of individuals may influence outcome more so than pathology alone. In addition, a high level of heterogeneity was found between our included studies. It could be due to different study design, different populations, and different methods for TB diagnosis. In addition, B-CAP in itself is a disease spectrum with different pathogens, and as seen in Table 1 , the bacterial pathogens, infecting patients in CAP group were different among studies, which could effect on NLR and be another source of heterogeneity. Finally, since eight out of the ten included studies were conducted in Asia, a potential racial bias could not be ruled out. On the other hand, to the best of our knowledge, this is the first meta -analysis to look at the links between NLR and TB. Additional strengths of our analysis include the relatively large number of cases and controls, the quality of the selected articles, and the meticulous methodological approach.

Worldwide, tuberculosis (TB) remains a major health concern. Even within the last two decades, it has been recognized as a global health emergency [1] . In 2020, it was anticipated that TB would rank as the second leading cause of death from a single infectious agent after COVID-19 [2] . TB is an airborne disease caused by the Mycobacterium tuberculosis. Infection happens when an individual breathes in droplet nuclei containing tubercle bacilli and these droplets go to the lung alveoli. Additionally, numerous studies carried out in nations with high tuberculosis burdens have demonstrated that Mycobacterium tuberculosis (MTB) is commonly the cause of bacterial community-acquired pneumonia (B-CAP); however, it is often challenging to differentiate TB and B-CAP during the initial diagnostic stage [3] , [4] . Symptoms of TB can be categorized into constitutional and pulmonary categories [5] . Tuberculosis is an infectious disorder whose progression and prognosis rely on the host's immunological response [6] . Both TB and B-CAP pathophysiology are characterized mainly by the inflammatory response [7] . TB generates self-limited hematological abnormalities [8] . White blood cell (WBC) populations are essential in the systemic inflammatory response to infection [9] , [10] . WBC count is a standard marker of inflammation. Several recent studies have examined the possible involvement of varied leukocyte ratios in chronic inflammatory disorders [11] , [12] , [13] , [14] . Hematological abnormalities are more frequently seen in patients with severe tuberculosis, as Yaranal et al. have shown. They found that patients with severe tuberculosis were typically found to have anemia, increased ESR, and thrombocytosis [15] . The physiological immune response of circulating leukocytes to stress is an increase in neutrophils and a decrease in lymphocytes [16] . The peripheral blood neutrophil to lymphocyte ratio (NLR), which indicates a combination of circulating neutrophil and lymphocyte counts, has been described as a representative indication of systemic inflammatory response that correlates with the prognosis of numerous acute or chronic disorders [17] . Increased levels of NLR is a unique predictor of inflammation in several clinical disorders, including thyroid autoimmunity [18] , cardiac arrhythmia [19] , irritable bowel syndrome [20] , COVID-19 [21] , and type 2 DM [22] . It is crucial for public health to distinguish TB from B-CAP; patients with TB need rapid isolation and urgent anti-TB therapy. However, it might be challenging to determine TB from B-CAP solely based on the physical exam, history, and radiological results [23] . Therefore, using an easily calculable laboratory marker like NLR might be beneficial in discriminating TB from bacterial CAP. We conducted this systematic review and meta -analysis to evaluate current literature on role of NLR in TB.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.