Clinical utility of Leicester Cough Score as a correlate of sputum conversion and inflammatory resolution in pulmonary Tuberculosis during Intensive phase therapy

INTRODUCTION

Tuberculosis (TB) remains a leading infectious cause of mortality globally, with an estimated 10 million new cases and 1.4 million deaths.^[1] Effective treatment is available for drug-susceptible pulmonary TB, yet monitoring a patient’s response to therapy remains challenging. Standard practice relies on bacteriological indicators, such as sputum smear microscopy and culture, to evaluate the early treatment response. However, these tools have significant limitations. Sputum smears have low sensitivity and cannot differentiate between viable and non-viable bacilli, while culture, although more sensitive, is time-consuming and resource-intensive.^[2-4] Moreover, even culture conversion at 2 months is not an absolute predictor of final treatment success, especially in subsets like drug-resistant TB.^[5,6] There is thus an urgent need for alternative or adjunct biomarkers that are rapid, safe, and reflective of disease activity to guide clinical decisions during TB therapy.^[7]

Clinical symptoms, particularly cough, present a potential complementary indicator of treatment response. Cough is the most common symptom of pulmonary TB and is routinely assessed during therapy. Active pulmonary TB typically presents with a persistent cough, among other constitutional symptoms.^[8] Moreover, cough frequency and severity tend to diminish with effective anti-TB treatment, with objective studies showing significant reductions in cough episodes within the 1^st weeks of therapy.^[9] However, until recently, symptom monitoring in TB has been largely qualitative and subjective, relying on patient recall and basic checklists. This subjective approach can be unreliable due to variability in patient perception and recall bias.^[10]

Furthermore, symptomatic improvement may not always parallel bacteriological clearance, highlighting the need to better understand the relationship between patient-reported symptoms and objective disease markers (evidence shows cough and other clinical symptoms often improve independently of culture or smear results).^[11] To systematically quantify symptom changes, validated patient-reported outcome measures are needed (studies using standardized quality of life (QoL) instruments demonstrate measurable symptomatic change).^[12]

The Leicester cough questionnaire (LCQ) was originally developed to measure the impact of chronic cough on QoL in a standardized manner.^[13] It consists of 19 items across physical, psychological, and social domains, each scored on a 7-point Likert scale, yielding a total score ranging from 3 (worst cough-related QoL) to 21 (best).^[12] The LCQ is a valid, internally consistent, and responsive instrument in chronic cough populations.^[13] Importantly, the LCQ has also been applied in TB contexts. Turner et al.^[14] validated the LCQ in patients with pulmonary TB and found it to be a reliable measure of cough-specific QoL in this population, responsive to early treatment effects. Their study demonstrated substantial improvement in LCQ scores within the first few weeks of therapy, supporting its potential utility as an outcome measure in TB. Another study reported that cough- and sputum-related QoL (measured by LCQ and a related questionnaire) was significantly impaired in active TB and improved after 2 months of treatment.^[12] These findings suggest that a Leicester cough score (LCS), derived from the total LCQ score, could serve as a practical tool to monitor symptomatic response during TB therapy.

Despite these indications, the role of LCS as a quantitative monitoring tool and its relationship with bacteriological and inflammatory markers in TB has not been fully elucidated. In particular, it remains unclear whether improvements in LCS correlate with conventional microbiological indicators such as sputum smear grade, time to culture positivity (TTP), or sputum culture conversion, and whether baseline LCS has any prognostic relevance for early bacteriological response. Given the subjective nature of cough and the biological heterogeneity of pulmonary TB, the ability of baseline cough-related QoL measures to predict microbiological outcomes warrants careful evaluation.

In this study, we aimed to investigate the utility of the LCS in adult patients with pulmonary TB during the intensive phase of treatment. Specifically, our objectives were (1) to assess changes in LCS over the intensive phase of therapy, (2) to examine the relationship between LCS and microbiological and inflammatory indicators of disease (including sputum smear grading, TTP, and serum C-reactive protein [CRP] levels), and (3) to evaluate whether baseline LCS is associated with sputum culture conversion at 2 months. We hypothesized that LCS would show significant improvement with effective therapy and explored whether baseline LCS demonstrated any association with early bacteriological response, recognizing that this association might be limited. By evaluating LCS in this context, we seek to clarify its role as a patient-reported outcome measure for symptom monitoring rather than as a surrogate marker of microbiological treatment response in TB care.

MATERIALS AND METHODS

RESULTS

ROC analysis of baseline LCS

ROC analysis demonstrated that baseline LCS had poor discriminatory ability for predicting sputum culture conversion at 2 months (AUC = 0.563, 95% CI 0.40-0.72; p = 0.43 against AUC = 0.5). The optimal cut-off value (≥7.55) yielded low specificity, limiting clinical utility [Table 6 and Figure 1]. These findings indicate that baseline LCS is not a reliable predictor of early bacteriological response.

Table 6: Diagnostic performance of baseline LCS for predicting sputum culture conversion (n=56).

Baseline LCS cut-off	Sensitivity (%)	Specificity (%)	Youden’s J
≥7.55	78.7	33.3	0.453
≥7.70	76.6	44.4	0.444
≥7.85	72.3	44.4	0.417
≥10.25	42.6	55.6	0.382

LCS: Leicester Cough score

Close

Figure 1:

ROC curve for baseline LCS predicting 2-month sputum conversion (AUC 0.563), illustrating poor discrimination.

Export to PPT

DISCUSSION

Our study demonstrates that cough-related QoL improves markedly during TB therapy, even when traditional microbiological or inflammatory markers do not change in parallel. We observed a significant rise in LCQ scores following treatment initiation, consistent with prior reports. For example, Suzuki et al.^[12] showed that mean LCQ scores increased by ~2.28 points after 2 months of therapy, regardless of smear grade or lung cavitation. Likewise, Turner et al.^[14] validated the LCQ in pulmonary TB and reported a median 5.1-point increase in LCQ after just 2 weeks of treatment (p = 0.003). Together, these findings and our results confirm that the LCQ is a responsive instrument for capturing symptomatic improvement during TB therapy, supporting its use as a patient-reported treatment-monitoring tool.

Our findings align with international literature on cough and TB care. Prospective and retrospective studies have consistently shown that TB patients experience substantial relief of cough and sputum symptoms soon after effective therapy begins. Suzuki et al.^[12] observed improved cough-related QoL post-treatment in all patients, independent of initial disease severity. Similarly, Turner et al.^[14] demonstrated high reliability (Cronbach’s a = 0.93) and validity of the LCQ in TB, with early and substantial score improvements. These parallels suggest that the LCQ behaves in TB much as it does in other chronic cough conditions, capturing real changes in symptom burden over time. In non-TB populations, the LCQ has been validated in chronic respiratory diseases such as bronchiectasis, nontuberculous mycobacterial lung disease, and COPD as a sensitive measure of cough impact.^[10,15] Takao et al.,^[15] for example, found the LCQ to be a useful health-related QoL tool in nontuberculous mycobacterial lung disease.

Beyond cough-specific instruments, symptom-based clinical scores have long been used to monitor TB patients. The WHO-endorsed Bandim TB score is a well-known example. Wejse et al.^[16] developed a composite symptom-and-sign score that declined steadily with successful treatment; in their cohort of 698 patients, 96% of survivors had lower scores at treatment completion than at baseline, and higher baseline scores predicted mortality. Rudolf et al.^[17] subsequently refined this tool (TBscore II) and demonstrated significant score reductions at 2 and 6 months of therapy (Cohen’s d ~0.5–0.7), with failure to improve by ≥25% at 2 months predicting treatment failure. These symptom-based tools reinforce the concept that patient-reported or clinician-assessed symptoms often improve with therapy and can provide meaningful information on the clinical course. Our findings are consistent with this paradigm in showing substantial symptomatic improvement measured by LCQ. However, unlike composite symptom scores designed with prognostic intent, the LCQ appears less suited for predicting microbiological endpoints.

Importantly, we observed no significant correlation between baseline LCQ scores and measures of bacterial burden or systemic inflammation. This finding is consistent with prior evidence. Miranda et al.,^[18] for example, reported that baseline CRP levels did not reliably distinguish patients with higher sputum Mycobacterium tuberculosis (MTB) loads or predict early culture positivity at treatment initiation.

The lack of correlation between cough-related QoL and microbiological markers likely reflects biological heterogeneity. Cough arises from airway irritation, inflammation, and structural lung involvement, processes that do not scale linearly with bacillary load. The “Making cough count” report highlighted that cough frequency is a poor proxy for bacterial shedding: some patients with infrequent cough released large quantities of MTB, whereas others with frequent cough did not.^[10] By extension, patient-reported cough impact, as measured by the LCQ, may diverge substantially from objective microbiological measures. Thus, the absence of association between LCQ and smear grade, TTP, or CRP in our study is consistent with the view that cough-related QoL scores capture a different dimension of disease, namely symptom experience, than laboratory-based markers.

We also observed that LCQ scores improved even among patients who remained culture-positive at the end of the intensive phase. In other words, cough-related QoL improved regardless of whether sputum conversion occurred within 2 months. This mirrors the findings of Suzuki et al.^[12] who reported that cough and sputum-related QoL improved “regardless of clinical characteristics.” In clinical practice, many patients experience symptomatic relief soon after starting therapy, as bacterial replication is suppressed and inflammation begins to subside. Such early symptomatic improvement may precede, or occur independently of, bacteriological clearance. Turner et al.^[14] similarly demonstrated substantial subjective improvement within the first 2 weeks of therapy. These observations emphasize that LCQ primarily reflects treatment-related symptomatic benefit rather than sterilizing activity. Clinically, this implies that patient-perceived improvement does not guarantee rapid culture conversion, and conversely, the absence of a prominent cough does not necessarily indicate microbiological cure.

Consistent with this dissociation, the predictive performance of baseline LCQ for early sputum culture conversion was poor in our cohort (AUC ~0.5-0.6). Baseline cough severity did not meaningfully distinguish converters from non-converters. This likely reflects the multifactorial nature of cough, which is influenced by airway hyperreactivity, smoking status, coexisting lung pathology, and individual perception, in addition to infection-related factors. In contrast, sputum conversion is driven primarily by bacillary burden, drug susceptibility, host immune response, and treatment adherence. As a result, a patient with severe cough-related impairment may convert as rapidly as one with milder symptoms. Our findings are consistent with reports cautioning against the use of cough metrics as surrogate markers of microbiological activity. Williams et al.^[19] demonstrated that objective cough counts were poorly associated with MTB output in exhaled air samples, suggesting that neither subjective nor objective cough measures reliably reflect bacillary load or transmission risk. The low AUC observed in our study reinforces that baseline LCQ should not be considered a stand-alone predictor of microbiological outcome.

The strengths of this study include its prospective design and systematic use of a validated patient-reported outcome measure from treatment initiation. The use of a culturally adapted LCQ and standardized microbiological assessments adds to the reliability of the findings. However, several limitations should be acknowledged. The sample size was modest, limiting statistical power and precision, particularly for predictive analyses. LCQ assessments were limited to baseline and the end of the intensive phase, precluding evaluation of longer-term symptom trajectories. The questionnaire was not formally revalidated in the study population, and objective cough monitoring was not performed. Potential confounders such as smoking status, diabetes, or chronic airway disease were not systematically captured, and selection bias may have occurred if severely ill patients were less able to complete questionnaires. These limitations constrain interpretation and underscore that our findings should be viewed as exploratory.

Overall, the LCQ appears useful for documenting patient-perceived symptomatic improvement during TB therapy but should not replace microbiological markers in guiding clinical decision-making. Improvement in cough-related QoL may provide reassurance to patients and clinicians that therapy is producing symptomatic benefit, even when cultures remain positive. However, reliance on LCQ to predict early sputum conversion would be inappropriate. Our study adds to emerging evidence that comprehensive TB monitoring should integrate patient-reported outcomes with established laboratory measures. Future research may explore combining subjective cough assessments with objective cough monitoring and biomarker panels to develop more holistic approaches to treatment monitoring in TB.

CONCLUSION

The LCS proved to be a highly responsive, feasible, and patient-centered tool for monitoring symptomatic improvement during the intensive phase of anti-TB therapy in routine DOT settings. The majority of patients experienced clinically meaningful improvement in cough-related QoL, with >90% exceeding the MCID. Follow-up LCS and the magnitude of improvement were significantly greater among patients who achieved sputum culture conversion, indicating that symptomatic recovery parallels effective treatment response.

However, baseline LCS demonstrated limited association with microbiological burden, systemic inflammation, or early bacteriological response and showed poor discriminatory ability for predicting 2-month sputum culture conversion. Baseline LCS should therefore not be used as a stand-alone predictor of early sputum conversion or treatment success. These findings highlight the biological dissociation between patient-reported cough severity and bacillary clearance during TB therapy.

Overall, LCS is best positioned as a complementary patient-reported outcome measure for symptom monitoring, patient counseling, and assessment of treatment-related QoL improvement, rather than as a surrogate marker of microbiological response. Future studies with larger, multicenter cohorts, longer follow-up, and integration of objective cough monitoring or biomarker panels may help clarify the role of cough assessment within comprehensive TB treatment-monitoring frameworks.