Predictive Value of Nutritional Status in Sputum Culture Conversion Among Patients with Nontuberculous Mycobacterial Pulmonary Disease: A Retrospective Cohort Study

Introduction

The incidence of nontuberculous mycobacterial pulmonary disease (NTM-PD) has been increasing worldwide in recent years,¹ becoming an important public health concern.^2,3 Its clinical manifestations commonly include chronic cough, sputum production, abnormal chest imaging findings,⁴ and repeated positive sputum cultures. In severe cases, NTM-PD can lead to structural lung damage and functional decline, thereby significantly impairing patients’ quality of life and prognosis.⁵ With population aging and the growing number of individuals with impaired immunity, the diagnosis and treatment of NTM-PD are becoming increasingly challenging.^6,7

Previous studies have primarily focused on pathogen distribution and antimicrobial susceptibility.⁸ For example, in Europe and North America, Mycobacterium avium complex (MAC) and Mycobacterium kansasii (M. kansasii) are the predominant pathogens,⁹ whereas in Asian countries, Mycobacterium intracellulare (M. intracellulare) and Mycobacterium abscessus are more prevalent.¹⁰ In recent years, however, researchers have proposed that beyond microbiological factors, host-related characteristics, particularly nutritional status, may play a critical role in disease progression and sputum culture conversion. Previous studies have suggested that nutritional status, particularly body mass index (BMI), is associated with disease severity and treatment outcomes in NTM-PD.^11–15. There is a lack of integrative analyses incorporating multiple nutritional parameters.

Despite increasing recognition of the role of nutritional status in NTM-PD, existing studies have primarily focused on single indicators such as BMI, with limited integration of multiple nutritional parameters. Moreover, the relationship between nutritional status and sputum culture conversion, a key indicator of treatment response, remains insufficiently characterized. We therefore hypothesize that a combination of nutritional indicators, including BMI, hemoglobin, albumin, and total protein, may provide improved predictive value for sputum culture conversion compared with individual markers. Accordingly, the present study aimed to (1) characterize the distribution of pathogenic species in NTM-PD, (2) evaluate the association between nutritional status and sputum culture conversion, and (3) explore the predictive performance of a multivariable model incorporating nutritional indicators. This study seeks to provide additional evidence to support risk stratification in clinical practice.

Methods

Study Design and Participants

This study was a single-center retrospective cohort study. Patients diagnosed with and treated for NTM-PD at Weifang Second People’s Hospital between January 2016 and September 2024 were consecutively included. All cases were diagnosed based on clinical symptoms, chest imaging findings, and the isolation of nontuberculous mycobacteria from sputum or bronchoalveolar lavage fluid (BALF), in accordance with the diagnostic criteria of the American Thoracic Society (ATS), European Respiratory Society (ERS), European Society of Clinical Microbiology and Infectious Diseases (ESCMID), and Infectious Diseases Society of America (IDSA).^16,17 The study was conducted in compliance with the principles of the Declaration of Helsinki (2013 revision) and was approved by the Ethics Committee of Weifang Second People’s Hospital (approval number: KY2023-010-01).¹⁸ As this was a retrospective cohort study, all data were obtained from the electronic medical record system and laboratory results. No identifiable personal information was included, and the Ethics Committee approved a waiver of informed consent.

Inclusion criteria were: (1) age ≥18 years, and (2) complete clinical and follow-up data. Exclusion criteria were: (1) co-infection with Mycobacterium tuberculosis or other confirmed pulmonary fungal infections; (2) severe hepatic or renal insufficiency or hematological disorders; and (3) incomplete follow-up data or loss to follow-up.

Sample Size Calculation

The primary outcome of this study was sputum culture conversion. Sample size estimation was based on the principle of events per variable (EPV) for logistic regression analysis. According to the method proposed by Riley et al, at least 10–15 outcome events per candidate predictor variable are required to ensure model stability.¹⁹ In this study, approximately 12 candidate predictors were planned (including age, body mass index [BMI], hemoglobin, albumin, total protein, erythrocyte sedimentation rate [ESR], smoking history, place of residence, and comorbidities). Therefore, at least 120–180 outcome events (sputum conversion or non-conversion) were required. Based on previous literature and pilot data from our center, the sputum conversion rate in NTM-PD patients is estimated at 25–30%.²⁰ Accordingly, the total sample size required was estimated to be over 400 cases. Ultimately, 416 patients were enrolled, including 109 in the sputum conversion group and 307 in the non-conversion group, thereby meeting the prespecified statistical power requirements.

Data Collection

Data were extracted from the electronic medical record system, including demographic characteristics, nutritional indicators, inflammatory markers, microbiological findings, and comorbidities. Demographic and clinical data included age, sex, marital status, place of residence, and history of smoking or alcohol consumption. Nutritional indicators included BMI, hemoglobin, albumin, total protein, lipid profile, and blood glucose. Inflammatory markers included ESR and lymphocyte count. Microbiological analyses included all nontuberculous mycobacterial isolates isolated from sputum or BALF cultures, and their distribution frequencies were recorded. Clinical symptoms assessed included cough, sputum production, hemoptysis, fever, night sweats, fatigue, anorexia, and weight loss. Radiological follow-up was performed using chest computed tomography (CT) to compare lesion changes before and after treatment, evaluating absorption or progression.

Grouping and Primary Outcome

Nutritional status was assessed using BMI, serum albumin, total protein, and hemoglobin. Malnutrition was defined as the presence of at least one of the following: BMI <18.5 kg/m², serum albumin <40 g/L, or hemoglobin <120 g/L.^21–23 Patients were divided into malnutrition and well-nourished groups accordingly, and further stratified into sputum conversion and non-conversion groups based on follow-up sputum smear or culture results. The primary endpoints included the distribution of pathogenic species, differences by nutritional status, and predictors of sputum culture conversion, with subsequent development of a predictive model. This simplified definition was adopted due to the retrospective nature of the study and the limited availability of standardized nutritional assessment tools such as GLIM criteria. Although this approach may introduce potential misclassification bias, it reflects routinely available clinical indicators in real-world settings.

Statistical Analysis

All statistical analyses were conducted using SPSS version 26.0 (IBM Corporation, Armonk, NY, USA) and R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were expressed as median (interquartile range, IQR) and compared between groups using the Mann–Whitney U-test. Categorical variables were expressed as frequency and percentage, and compared using the chi-square (χ²) test. Variables with p <0.05 in univariable analysis were entered into multivariable logistic regression to calculate odds ratios (ORs) and 95% confidence intervals (CIs), identifying independent predictors of sputum conversion (This threshold was predefined based on conventional statistical practice). Given the exploratory nature of the study, no formal adjustment for multiple comparisons was performed. Receiver operating characteristic (ROC) curves were generated to compare predictive performance between single factors and the combined model, evaluated by area under the curve (AUC), sensitivity, and specificity. The combined model was constructed using multivariable logistic regression, incorporating age, BMI, hemoglobin, and total protein as independent predictors identified in the analysis. Internal validation was performed using bootstrap resampling (1000 iterations) to assess model stability and potential overfitting.

Kaplan–Meier analysis was performed to evaluate the time-dependent pattern of sputum culture conversion according to nutritional status. Sputum culture conversion was defined as the first documented negative sputum culture during follow-up. The median follow-up time was 9.0 months (interquartile range [IQR]: 6.0–12.0 months). Time to conversion was calculated from the initiation of treatment to the date of sputum culture conversion. Patients who did not achieve culture conversion were censored at the date of last follow-up. Kaplan–Meier curves were constructed to estimate the probability of remaining culture-positive over time, and differences between groups were compared using the Log rank test. A lower curve indicated an earlier and higher probability of sputum culture conversion. All time-to-event analyses were conducted with a two-sided significance level of 0.05. Cumulative sputum culture conversion at clinically relevant time points during follow-up was derived from the Kaplan–Meier estimates. All statistical analyses were conducted following standard methodological practices to ensure robustness and reproducibility.

Results

Pathogen Spectrum in NTM-PD

A total of 423 NTM pathogenic isolates were identified from all enrolled patients. Among them, M. intracellulare was the most common species (317 isolates, 78.6%), representing the dominant pathogen. The next most frequent species were M. kansasii (43 isolates, 10.7%) and M. abscessus (24 isolates, 6.0%).

Less frequently detected species included Mycobacterium terrene (M. terrene, 12 isolates, 3.0%), Mycobacterium fortuitum (M. fortuitum, 10 isolates, 2.5%), and Mycobacterium avium (M. avium, 10 isolates, 2.5%). In addition, Mycobacterium colombiense (M. colombiense), Mycobacterium chimaera (M. chimaera), Mycobacterium malmoense (M. malmoense), Mycobacterium margaritense (M. margaritense), Mycobacterium szulgai (M. szulgai), Mycobacterium scrofulaceum (M. scrofulaceum), Mycobacterium aarhusense (M. aarhusense), and Mycobacterium lentiflavum (M. lentiflavum) were each isolated only once (0.2%), considered sporadic species (Table 1). Overall, the pathogen spectrum was highly concentrated, with M. intracellulare predominating, whereas other species were sporadically distributed, suggesting specific pathogen characteristics in this population or environment.

Table 1 Distribution of Pathogen Species Isolated from Clinical Samples

Association Between Nutritional Status and Pathogen Distribution

In the well-nourished group (n = 137, Figure 1A), 120 isolates were identified. M. intracellulare remained dominant (94 isolates, 78.3%), followed by M. kansasii (19 isolates, 15.8%) and M. abscessus (11 isolates, 9.2%). Other species, such as M. terrene, M. fortuitum, M. avium, M. malmoense, and M. colombiense, were rare (≤2 isolates). In the malnourished group (n = 279, Figure 1B), 303 isolates were isolated, with M. intracellulare still predominant (223 isolates, 73.6%), followed by M. kansasii (24 isolates, 7.9%) and M. abscessus (13 isolates, 4.3%). Other species such as M. terrene, M. fortuitum, and M. avium accounted for <3%. Comparison between groups indicated a relatively higher proportion of M. intracellulare in malnourished patients, whereas low-frequency species (M. terrene, M. fortuitum, M. avium) were slightly more frequent in the well-nourished group. A trend toward differences in pathogen distribution was observed between groups; however, these findings should be interpreted with caution.

Figure 1 Distribution of pathogenic species stratified by nutritional status. (A) Well-nourished group (n = 137); (B) Malnourished group (n = 279). The distribution of Mycobacterium species is shown for each group. M. intracellulare was the predominant species in both groups, followed by M. kansasii and M. abscessus.

Baseline Characteristics of Patients by Nutritional Status

Among 416 patients, 279 were malnourished and 137 were well-nourished (Table 2 and Table 3). Compared with the well-nourished group, the malnourished group was older (65 vs. 57 years, P < 0.001), had a higher proportion of males (62.7% vs. 46.0%, P = 0.001), and more often lived in rural areas (74.9% vs. 55.5%, P < 0.001). Nutritional indices including BMI, hemoglobin, albumin, and total protein were significantly lower in the malnourished group (all P < 0.001). ESR was markedly higher (36 vs. 10 mm/h, P < 0.001), suggesting greater inflammation. In terms of pathogens, both groups were dominated by M. intracellulare, but its proportion was higher in malnourished patients (78.8% vs. 66.2%, P = 0.043), whereas M. avium and M. kansasii were more frequent in well-nourished patients. Regarding comorbidities, the malnourished group showed higher rates of hyperlipidemia, malignancy, rheumatic immune diseases, chronic bronchitis, and interstitial pulmonary fibrosis (all P < 0.05).

Table 2 Baseline Characteristics of Patients Stratified by Nutritional Status (Continuous Variables)

Table 3 Baseline Characteristics of Patients Stratified by Nutritional Status (Categorical Variables)

Clinical Symptoms by Nutritional Status

Significant differences were observed between groups (Table 4). Malnourished patients more frequently had sputum production (87.5% vs. 79.6%, P = 0.035), fever (39.8% vs. 13.9%, P < 0.001), fatigue (32.3% vs. 18.3%, P = 0.003), anorexia (46.2% vs. 18.3%, P < 0.001), and weight loss (11.5% vs. 2.2%, P = 0.001). No significant differences were found for cough, hemoptysis, night sweats, chest tightness, or chest pain (all P > 0.05). Overall, malnourished patients presented with more wasting symptoms (anorexia, weight loss, fatigue) and infection-related manifestations (fever, sputum production), suggesting that poor nutrition may aggravate susceptibility and clinical presentation.

Table 4 Comparison of Clinical Symptoms Between Malnourished and Well-Nourished Groups

Comparison of Nutritional Indices by Sputum Culture Conversion

Sputum culture conversion was primarily assessed at 3 months after treatment initiation, with extended follow-up up to 12 months to evaluate cumulative conversion over time. As shown in Figure 2, during the 3-month follow-up, patients achieving sputum culture conversion had significantly higher BMI, albumin, and hemoglobin levels compared with non-converters (all P < 0.001).

Figure 2 Comparison of nutritional indices between sputum conversion and non-conversion groups. (A) BMI (kg/m2); (B) Albumin (g/L); (C) Hemoglobin (g/L). ****P < 0.0001.

Baseline Features Associated with Sputum Culture Conversion

Compared with the non-conversion group (n = 307), the sputum conversion group (n = 109) was younger (55 vs. 65 years, P < 0.001) and had higher blood pressure (SBP: 127 vs. 122 mmHg, P = 0.018; DBP: 80 vs. 75 mmHg, P = 0.001). Nutritional indices, including BMI, hemoglobin, and albumin, were all significantly higher in converters (all P < 0.001). Total protein, cholesterol, and triglyceride levels were also higher, whereas ESR was lower (10 vs. 33 mm/h, P < 0.001), indicating a milder inflammatory response. Comorbidity analysis showed lower rates of malignancy (0.9% vs. 6.5%, P = 0.022), rheumatic immune diseases (0% vs. 5.9%, P = 0.010), and chronic bronchitis (4.6% vs. 11.7%, P = 0.032) in converters. Rural residence was more frequent in non-converters (73.0% vs. 56.0%, P = 0.001). Positive inflammatory markers were also more common in non-converters (73.3% vs. 46.8%, P < 0.001, Table 5).

Table 5 Baseline Characteristics and Clinical Features Between Sputum-Negative and Non-Negative Groups

Independent Risk Factors for Sputum Culture Non-Conversion

Malnutrition (OR ≈ 2.0–2.5), diabetes (OR ≈ 1.8–2.2), and age >65 years (OR ≈ 1.5–2.0) were independent risk factors for sputum non-conversion. Multivariate logistic regression (Tables S1 and 6) identified age (OR = 1.048, 95% CI: 1.012–1.085, P = 0.009), BMI (OR = 0.746, 95% CI: 0.668–0.835, P < 0.001), hemoglobin (OR = 0.936, 95% CI: 0.906–0.967, P < 0.001), and total protein (OR = 0.920, 95% CI: 0.853–0.992, P = 0.031) as independent predictors of sputum conversion. Advanced age was a risk factor, whereas higher BMI, hemoglobin, and total protein were protective. Other factors, including blood pressure, albumin, ESR, smoking, residence, and comorbidities, were not significant (all P > 0.05).

Table 6 Multivariate Logistic Regression Analysis of Factors Associated with Sputum Conversion

ROC Curve Analysis for Predictive Performance

As shown in Table 7 and Figure 3, in single-variable models, hemoglobin (AUC = 0.833, 95% CI: 0.792–0.874) and BMI (AUC = 0.808, 95% CI: 0.764–0.852) had the best predictive performance, whereas total protein (AUC = 0.723, 95% CI: 0.671–0.774) and age (AUC = 0.710, 95% CI: 0.652–0.768) were less predictive. The combined model, incorporating age, BMI, hemoglobin, and total protein, demonstrated improved predictive performance compared with individual predictors. By contrast, the combined model integrating multiple indices achieved the highest performance (AUC = 0.925, 95% CI: 0.899–0.952), with sensitivity 89.3% and specificity 78.0%, significantly outperforming any single predictor. This indicates that the combined model has higher clinical utility in predicting sputum culture conversion.

Table 7 ROC Curve Analysis of Predictors for Sputum Conversion

Figure 3 Receiver operating characteristic (ROC) curves for predicting sputum culture conversion. ROC curves were generated for individual predictors (age, BMI, hemoglobin, and total protein) and the combined model. The combined model (Alliance) included age, BMI, hemoglobin, and total protein. The area under the curve (AUC) was used to evaluate predictive performance.

Time-Dependent Sputum Culture Conversion According to Nutritional Status

To further characterize the time-dependent pattern of sputum culture conversion, Kaplan–Meier analysis was performed to assess the probability of remaining culture-positive according to nutritional status (Figure 4). During a median follow-up of 9.0 months (IQR: 6.0–12.0 months), patients in the well-nourished group showed a significantly higher probability of sputum culture conversion. Throughout the 12-month follow-up, patients in the well-nourished group exhibited a significantly lower probability of remaining culture-positive compared with those in the malnourished group, indicating earlier and more frequent sputum culture conversion.

Figure 4 Kaplan–Meier curves showing the probability of remaining culture-positive stratified by nutritional status. Sputum culture conversion was defined as the event. Patients without conversion were censored at the last follow-up. A lower curve indicates a higher and earlier probability of sputum culture conversion.

The separation between the two curves became evident within the early months after treatment initiation and progressively widened over time. By 12 months, the proportion of patients remaining culture-positive was markedly higher in the malnourished group, whereas a substantially greater proportion of well-nourished patients had achieved sputum culture conversion. The difference in culture conversion dynamics between the two groups was statistically significant (Log rank test, P < 0.0001).

Discussion

Current Status of Research on NTM-PD

NTM-PD has shown an increasing trend globally and in China in recent years.²⁴ The pathogen spectrum is complex, primarily including Mycobacterium avium complex (MAC), M. intracellulare, M. abscessus, and M. kansasii, with geographic and environmental exposure differences.²⁵ Previous studies have mainly focused on pathogen identification, drug susceptibility, imaging features, and immunophenotypes, whereas the relationship between nutritional status (eg, BMI, serum protein/albumin levels) and sputum culture conversion or treatment outcomes has been less frequently reported. In several studies in China, elderly or malnourished patients were noted to have lower treatment or microbiological sputum conversion rates compared with well-nourished patients, though specific indicators and predictive models vary widely.²⁶ For example, Kang et al²⁷ noted in a systematic review that “elderly and malnourished” patients represent a population with lower microbiological cure rates after treatment completion.

Comparison with Existing Literature

Consistent with our findings, Kang et al reported that malnourished NTM-PD patients in multiple regions of China had lower treatment or microbiological sputum conversion rates, and elderly patients had reduced cure rates.^27–29 Moon et al³⁰ found that lower BMI in MAC-PD patients was associated with failure to achieve sputum conversion, although hemoglobin and other indicators did not emerge as strong independent factors in their sample. Regarding nutritional assessment tools, Chung et al³¹ recently evaluated the utility of the short-form Mini Nutritional Assessment (MNA-SF) in elderly NTM-PD patients, finding moderate correlations with BMI, CT lesion severity, and other disease severity indicators, supporting the link between nutritional status and disease severity/outcomes. In our study, Follow-up imaging showed that improved cases exhibited lesion shrinkage and absorption of inflammatory consolidation, whereas progression cases demonstrated lesion enlargement and aggravation of fibrotic strands (Figure S1). Total protein and hemoglobin were included in the model as protective factors, consistent with previous studies, but we additionally found that their combined predictive performance was superior, a finding rarely reported in prior literature.

Potential Mechanistic Explanations

The strong association between malnutrition and failure to achieve sputum conversion may reflect multiple mechanisms. First, malnutrition is associated with insufficient protein and energy intake, resulting in decreased serum proteins (including albumin and total protein) and impaired immune function, particularly cell-mediated immunity, reducing pathogen clearance.³² Second, decreased hemoglobin may cause tissue hypoxia, impairing local pulmonary immune responses, including macrophage phagocytosis and reactive oxygen species-mediated bacterial killing.³³ Third, advanced age contributes to immunosenescence and a higher prevalence of comorbidities (eg, chronic bronchitis, interstitial lung disease), which may further delay or hinder pathogen clearance. Additionally, recent experimental and basic research suggests that protein-energy restriction can exacerbate MAC-PD progression by disrupting lipid metabolism and impairing immune and pulmonary tissue repair functions.³⁴ Collectively, these mechanisms may explain the lower sputum conversion rates in malnourished patients.

The observed predominance of Mycobacterium intracellulare in malnourished patients may reflect a potential interaction between host nutritional status and pathogen susceptibility. Malnutrition has been associated with impaired immune responses, which may influence host–pathogen dynamics and contribute to differential pathogen distribution. Although this relationship requires further investigation, it suggests that nutritional status may play a role in shaping the microbial profile in NTM-PD.

Clinical Significance and Applications

This study provides additional insight into the association between nutritional status and sputum culture conversion in NTM-PD. Our study highlights that nutritional status in NTM-PD patients not only reflects overall health but is closely associated with sputum culture conversion. Early screening of nutritional indices may help identify high-risk patients and guide individualized treatment and care management.³⁵ In particular, routine monitoring of BMI, hemoglobin, and total protein can serve as important supplemental prognostic indicators. The combined predictive model demonstrated high sensitivity and specificity in our study, suggesting its potential application in clinical risk stratification and follow-up management; however, further validation is required before clinical application. These findings may assist clinicians in identifying patients at higher risk of delayed sputum culture conversion and support individualized risk stratification in NTM-PD.

Limitations and Future Directions

Despite a comprehensive analysis of nutritional indices, pathogen spectrum, clinical symptoms, and sputum outcomes, this study has limitations. Some regression coefficients appeared relatively large, which may reflect sparse data or quasi-separation, a known issue in multivariable logistic regression when certain categories have limited observations. However, the direction and statistical significance of the key predictors remained consistent. First, it was a single-center retrospective study, which may introduce selection bias. The study population was derived from a single-center cohort, which may limit the generalizability of the findings to other regions or healthcare settings. Second, follow-up was limited to short-term sputum culture conversion, without systematic evaluation of long-term outcomes such as recurrence, pulmonary function decline, or mortality. Third, some confounding factors were not fully included, such as differences in treatment regimens, medication adherence, environmental exposures (eg, water and soil pathogens), and immune cell function (eg, T-cell subsets). Future research could adopt prospective, multicenter designs with larger sample sizes, incorporating immune function indices, microbiome analysis, and genetic susceptibility, to establish more dynamic and generalizable predictive models. Interventional studies on nutritional support (dietary or supplementation) to improve sputum conversion rates are also warranted. The use of a simplified nutritional classification may have introduced misclassification bias. Important treatment-related factors, including regimen selection, duration of therapy, and treatment adherence, were not consistently available and therefore were not included in the analysis. These variables may act as potential confounders influencing sputum culture conversion. Although internal validation was performed, external validation was not conducted, which may limit the generalizability and robustness of the predictive model. The model should be considered exploratory and hypothesis-generating rather than definitive. The model requires further validation before clinical application.

Conclusion

The observed predominance of M. intracellulare in malnourished patients in this cohort may indicate a potential association between host nutritional status and pathogen distribution. Malnutrition has been associated with impaired immune responses, which may alter host–pathogen interactions and influence susceptibility to specific mycobacterial species. However, given the observational design of this study, this finding should be interpreted with caution and does not imply a causal relationship. Further studies are required to clarify the underlying mechanisms. Nevertheless, these findings suggest that nutritional status may be a contributing factor in shaping the microbial profile in patients with NTM-PD. The model demonstrates potential utility for risk stratification; however, further validation is required before clinical application.