Utility of hematological parameters to predict tuberculosis disease among PLHIV: A retrospective cohort study

INTRODUCTION

Tuberculosis (TB) remains a formidable global health challenge, exacerbated by its deadly synergy with the human immunodeficiency virus (HIV) pandemic.^[1] The compromised immune system in individuals living with HIV renders them highly susceptible to active TB disease, underscoring the urgent need for early diagnosis and prompt treatment.^[2] However, conventional diagnostic methods for TB, such as clinical presentation, radiological findings, and microbiological tests, can be time-consuming, costly, and lack sensitivity, particularly in advanced immunosuppression.^[3]

Hematological parameters, including hemoglobin levels, platelet counts, white blood cell differentials, and derived ratios, have emerged as potential markers for predicting TB and its monitoring.^[4] Several studies have reported alterations in these parameters among people living with HIV (PLHIV) with active TB, such as anemia, thrombocytopenia, and leukocytosis.^[5,6] These hematological abnormalities may reflect the underlying inflammatory response, immune dysregulation, and disease progression associated with TB and HIV co-infection.^[7]

Among the hematological parameters, the monocyte-tolymphocyte ratio (MLR) and neutrophil-to-lymphocyte ratio (NLR) have garnered increasing attention as potential biomarkers for TB and other infectious diseases.^[8,9] These ratios, derived from the differential white blood cell counts, may provide insights into the host’s immune response to TB infection and could aid in risk stratification and disease monitoring.^[10]

In the context of HIV and TB co-infection, hematological parameters could serve as valuable adjuncts to existing diagnostic methods, particularly in resource-limited settings where access to advanced diagnostic tools is restricted.^[11] Moreover, identifying potential hematological biomarkers associated with TB disease in PLHIV could contribute to our understanding of the pathogenesis and disease progression, ultimately informing strategies for early intervention and improved patient management.^[12]

Several studies have explored the utility of hematological parameters in predicting TB disease and survival outcomes in PLHIV. For instance, Gatechompol et al. found that a higher MLR and lower hemoglobin levels were associated with an increased risk of TB after initiating antiretroviral therapy (ART).^[13,14]

Despite these promising findings, there is a need for further research to validate the diagnostic and prognostic value of hematological parameters in different settings and populations. In addition, exploring the potential synergistic effects of combining multiple hematological markers could enhance the predictive accuracy for TB disease and survival outcomes in PLHIV.^[15]

By incorporating the assessment of hematological parameters into existing HIV/TB programs, the National AIDS Control Program can leverage readily available and cost-effective tools to enhance early detection, risk stratification, and targeted interventions, ultimately contributing to improved management and outcomes for PLHIV with TB co-infection. With this context, the present study aimed to assess the utility of hematological parameters, such as hemoglobin levels, platelet counts, white blood cell differentials, and derived ratios (MLR and NLR), in predicting TB disease among PLHIV, and to evaluate the association between hematological parameters in PLHIV with TB co-infection, and also to investigate the potential of combining multiple hematological markers to enhance the predictive accuracy for TB disease in PLHIV.

MATERIALS AND METHODS

TB definition and diagnosis

For this study, TB was defined as either bacteriologically confirmed or clinically diagnosed PTB.^[16] Bacteriologically confirmed TB included cases with positive sputum smear microscopy for acid-fast bacilli (AFB), culture-positive for Mycobacterium TB complex, or positive GeneXpert Mycobacterium tuberculosis/rifampicin (MTB/RIF) assay. Clinically diagnosed TB cases were defined based on suggestive clinical features (such as persistent cough, fever, night sweats, and weight loss), radiological findings consistent with TB (cavities, infiltrates, pleural effusion on chest X-ray, or computed tomography scan), or strong clinical suspicion by a physician leading to initiation of a full course of anti-TB treatment. The diagnosis algorithm for TB in PLHIV included symptom screening at each visit using the World Health Organization (WHO) four-symptom screen (current cough, fever, night sweats, or weight loss). Symptomatic patients underwent sputum collection for AFB smear microscopy and GeneXpert MTB/RIF testing, along with chest X-ray. PLHIV with advanced HIV disease (CD4 count <200 cells/μL) or those not on ART underwent TB screening with GeneXpert MTB/RIF. Anti-TB treatment was initiated based on bacteriological confirmation or strong clinical suspicion, per national guidelines, with patients closely monitored for treatment response and adverse effects. In the present study, only PTB cases were included, with extrapulmonary TB excluded from the analysis.

The figure 1 shows the participant selection process [Figure 1].

Close

Figure 1:

Participants selection process. PLHIV: People living with human immunodeficiency virus, HIV: Human immunodeficiency virus

Export to PPT

RESULTS

Table 1 provides the baseline characteristics of the study population, which consisted of 813 individuals. The mean age of the participants was 39.7 ± 13.7 years. The majority of the participants were male (67.7%). Most of the participants (79%) had a CD4 count >350 cells/μL, and 87.6% were in WHO Clinical Stage 1. Around 41.5% of the participants received cotrimoxazole preventive therapy (CPT), and 70% received isoniazid preventive therapy (IPT). The majority (97%) were on a first-line ART regimen. The mean baseline hemoglobin level was 11.5 ± 2.48 g/dL. Regarding anemia status, 32.3% had no anemia, 26.9% had mild anemia, 31.9% had moderate anemia, and 8.9% had severe anemia. Most of the participants (92.4%) had a working functional status [Table 1].

Table 2 shows a comparison of hematological profiles of PLHIV patients with and without TB. This table compares the hematological profiles of PLHIV with PTB and without PTB. The mean hemoglobin level was found to be lower in PLHIV with PTB (10.2 ± 2.1 g/dL) compared to those without PTB (11.9 ± 2.43 g/dL). The mean platelet count and mean CD4 count were almost marginally similar between the two groups (P = 1.00 and 0.643). PLHIV with TB had a higher mean white blood cell count (7.06 ± 3.62 × 10³/μL vs. 6.5 ± 2.4 × 10³/μL, P = 0.011) and a higher mean absolute neutrophil count (4.61 ± 2.9 × 10³/μL vs. 3.78 ± 1.94 × 10³/μL, P value < 0.001) compared to those without TB. PLHIV with TB had a lower mean MCV (79 ± 9.9 vs. 82.3 ± 11.07, P = 0.001) and higher median MLR ratio (0.286 ± 0.274 vs. 0.158 ± 0.151, P = 0.001) and NLR ratio (2.60 ± 2.26 vs. 1.71 ± 1.16, P = 0.001) compared to those without TB [Table 2]. Of the 813 total participants, 310 had greater MLR (>0.23) while 503 had lower MLR (≤0.23). Among those with greater MLR, 153 (49.4%) were TB positive, and 157 (50.6%) were TB negative. In contrast, among those with lower MLR, only 52 (10.3%) were TB positive, while 451 (89.7%) were TB negative. This difference in TB positivity between MLR groups was statistically significant (χ² = 155, df = 1, P < 0.001).

Table 3 provides modeling analysis using logistic regression analysis and prediction analysis. This table presents the results of two logistic regression models for predicting the presence of TB [Table 3].

Model 2

Model 2 aimed to predict the TB status among PLHIV using hematological parameters, especially anemia status (mild, moderate, and severe vs. no anemia) and NLR ratio (high vs. low). The model had an R² of 0.122, indicating that it explained 12.2% of the variation in the outcome. Compared to those without anemia, the odds of having TB were 3.72 times higher for those with mild anemia (95% CI: 2.13– 6.49, P < 0.001), 7.35 times higher for those with moderate anemia (95% CI: 4.35–12.42, P < 0.001), and 7 times higher for those with severe anemia (95% CI: 3.59–13.63, P < 0.001). Individuals with a high NLR ratio had 2.27 times higher odds of having TB than those with a low NLR ratio (95% CI: 1.62–3.20, P < 0.001). The model had an accuracy of 71%, an AUC of 73.4%, a specificity of 75.8%, and a sensitivity of 56.6% [Table 4].

Table 4: Modeling analysis by logistic regression analysis and prediction analysis for NLR and anemia (Model 2).

Predictor	COR (95% CI)	aOR (95% CI)	P-value
NLR status
High versus low	2.80 (2.04–3.86)	2.27 (1.62–3.20)	<0.001
Anemia status
Mild versus no anemia	3.94 (2.29–6.77)	3.72 (2.13–6.49)	<0.001
Moderate versus no anemia	8.19 (4.93–13.60)	7.35 (4.35–12.42)	<0.001
Severe versus no anemia	8.12 (4.25–15.51)	7.00 (3.59–13.63)	<0.001

Model fit: McFadden’s R²=0.122, Adjusted R²=0.110 Model performance: AUC=0.734, Accuracy=0.710, Sensitivity=0.566, Specificity=0.758. COR: Crude odds ratio, aOR: Adjusted odds ratio, CI: Confidence interval, NLR: Neutrophil-to-lymphocyte ratio, AUC: Area under the ROC curve. Adjusted for age, gender, baseline CD4 count, WHO clinical stage, cotrimoxazole preventive therapy, isoniazid preventive therapy, and antiretroviral therapy regimen

Figure 2 shows the receiver operating characteristic curves of Model 1 (MLR and Anemia) [Figure 2] and Figure 3 presents the curves for Model 2 (NLR and Anemia)[Figure 3], both used for predicting tuberculosis in PLHIV.

Close

Figure 2:

Receiver operating characteristic curves of model-1 (monocyte-to-lymphocyte ratio and anemia) for predicting tuberculosis in people living with human immunodeficiency virus.

Export to PPT

Close

Figure 3:

Receiver operating characteristic curves of model-2 (neutrophil-to-lymphocyte ratio and anemia) for predicting tuberculosis in people living with human immunodeficiency virus.

Export to PPT

DISCUSSION

This retrospective cohort study aimed to assess the utility of hematological parameters in predicting TB disease among PLHIV attending an ART center. The findings of this study contribute to a growing body of evidence highlighting the potential value of routinely available hematological markers in the management of HIV and TB co-infection.

A key finding of this study was the significant association between the MLR and the presence of TB disease in PLHIV. The logistic regression analysis (Model 1) revealed that individuals with a higher MLR ratio had an 8-fold increased odds of having TB compared to those with a lower MLR ratio (adjusted OR = 8.13, 95% CI: 5.54–11.93, P < 0.001). This is consistent with previous studies that have reported an elevated MLR in PLHIV with active TB disease.^[7,8,17] The MLR is a derivative of the differential white blood cell count, reflecting the balance between monocytes and lymphocytes. An elevated MLR may indicate an ongoing inflammatory response and immune dysregulation associated with TB infection.^[18,19]

The study also examined the NLR as a potential biomarker for predicting TB disease in PLHIV. In Model 2, individuals with a high NLR (>2) had 2.27 times higher odds of having TB compared to those with a low NLR (95% CI: 1.62–3.20, P < 0.001). This finding is consistent with previous studies that have reported an elevated NLR in PLHIV with active TB disease.^[9,16] The NLR reflects the balance between neutrophils and lymphocytes, with an elevated ratio indicating an ongoing inflammatory response and immune dysregulation.^[18]

While both MLR and NLR are derived from the differential white blood cell count, the observed difference in their predictive performance for TB disease in PLHIV is noteworthy. Model 1, which incorporated MLR and anemia status, demonstrated superior model fit (McFadden’s R² = 0.242) and predictive performance (AUC = 0.816, accuracy = 0.801) compared to Model 2, which included NLR and anemia status (McFadden’s R² = 0.122, AUC = 0.734, accuracy = 0.710). This suggests that the MLR may be a more robust predictor of TB disease in PLHIV than the NLR.

Another key finding of this study was the strong association between anemia and TB disease in PLHIV. The logistic regression models revealed that individuals with mild, moderate, or severe anemia had significantly higher odds of having TB compared to those without anemia (Model 1 and Model 2). These results align with previous research indicating a high prevalence of anemia among PLHIV with active TB disease.^[5,14,20,21]

The observed association between anemia and TB disease in PLHIV can be explained by several potential mechanisms. Chronic inflammation and immune activation associated with HIV and TB co-infection can lead to anemia through various pathways, including impaired erythropoiesis, increased red blood cell destruction, and blood loss.^[6] In addition, nutrient deficiencies, which are common in PLHIV, can contribute to the development of anemia.^[11] Anemia may also be an indicator of disease severity and advanced immunosuppression, which could increase the risk of TB reactivation or progression.^[12]

The observed differences in other hematological parameters, such as white blood cell counts, absolute neutrophil counts, and red blood cell parameters (e.g., MCV) between PLHIV with and without TB, further support the potential diagnostic utility of these readily available markers. In the present study, PLHIV with TB had significantly higher mean white blood cell counts (7.06 ± 3.62 × 10^3/μL vs. 6.5 ± 2.4 × 10^3/μL, p-value = 0.011) and mean absolute neutrophil counts (4.61 ± 2.9 × 10^3/μl vs. 3.78 ± 1.94 × 10^3/μl, P < 0.001) compared to those without TB.^[4,7,18] In addition, PLHIV with TB had a lower mean MCV (79 ± 9.9 vs. 82.3 ± 11.07, P value = 0.001), which may indicate a higher prevalence of microcytic anemia in this group.^[22,23]

These findings collectively highlight the potential value of routinely available hematological parameters, particularly the MLR and anemia status, as adjunctive tools in the diagnosis and management of TB disease in PLHIV. While further research is needed to validate and refine these findings, the integration of these cost-effective markers into existing diagnostic algorithms and clinical decision-making processes could potentially improve case detection, risk stratification, and targeted interventions for this vulnerable patient population, especially in resource-limited settings.

CONCLUSIONS

This retrospective cohort study has provided valuable insights into the potential utility of routinely available hematological parameters in predicting TB among PLHIV. The findings highlight the strong associations between MLR, anemia, and the presence of TB disease in this vulnerable population.

However, several limitations should be acknowledged. The retrospective design limits causal inference and may introduce information bias. As a single-center, hospital-based study, our findings may not be generalizable to the broader PLHIV population across different settings. In addition, we were unable to account for all potential confounders, such as detailed nutritional status, certain co-infections, and ART adherence that might influence both hematological parameters and TB risk.

Despite these limitations, by incorporating the assessment of hematological parameters into existing HIV/TB programs, healthcare providers can leverage readily available and cost-effective tools to enhance early detection, risk stratification, and targeted interventions. Future prospective, multi-center studies with comprehensive assessment of confounding factors are warranted to validate these findings before widespread implementation. Nevertheless, this study provides a foundation for the potential role of hematological markers in improving the management and outcomes for this vulnerable patient population.