We retrospectively reviewed the medical records of hospitalized patients who were diagnosed with “NTM disease” between January 1, 2013 and December 31, 2020 at a large, tertiary, urban, teaching hospital of Peking Union Medical College Hospital in Beijing, China. Data collected included patient characteristics, symptoms, medical histories, radiological findings, laboratory data since hospital admission, NTM species, clinical exacerbations, and comorbidities, which were obtained at the time of diagnosis and during follow-up.

Patients fulfilling criteria for one of the three major clinical syndromes of NTM infections based on a combination of clinical, radiographic, and microbiological criteria according to the guidelines of the American Thoracic Society and the Infectious Disease Society of America (ATS/IDSA), published in 2007 and updated in 2020 (9, 13), met inclusion for this study. At the same time, all the patients were treated for NTM disease with antibiotics according to guideline recommendation excluding colonization or contamination with NTM (13, 17).

In our study, the three major clinical syndromes caused by NTM infection depend on clinical presentation and the diagnostic criteria of the official ATS/IDSA statement:

Patients were divided into immunocompetent and immunocompromised categories according to their general state, medical history, medication, and immune status (18). Immunocompromised patients were defined broadly, including patients with acquired immunodeficiency syndrome (AIDS/HIV), those under continuous (>3 months) or high-dose (>0.5 mg/kg/day) corticosteroid or immunosuppressive therapy, solid-organ transplant recipients, those with rheumatic autoimmune diseases, solid-organ malignancies requiring chemotherapy in the last 5 years, those with any history of hematological malignancies, or those with primary immune deficiencies. Previous tuberculosis (TB) refers to patients who have previous treatment for active TB disease in the past history and were currently cured.

Charlson comorbidity index score was used to assess the comorbidities and known to be associated with mortality. The score includes myocardial infarction, diabetes mellitus, cancer, chronic pulmonary disease, liver disease, chronic renal failure, cerebrovascular accident, peptic ulcer, autoimmune disease, cirrhosis, leukemia, AIDS (19).

The date of NTM diagnosis was designated as the date of the first positive NTM culture. Findings and interval changes in clinical symptoms, bacteriological identification, and chest CTs during follow-up (from the initial screening or diagnostic images to the last observation) were assessed by physicians and radiologists together. The NTM exacerbations were judged based on the interval recurrence of symptoms and any changes in chest imaging during 1-year follow-up. We evaluated risk factors for all-cause mortality in NTM patients and calculated time-to-event as the time in days from the index of diagnosis until the date of death or end of follow-up (December 31, 2021), whichever occurred first for mortality analysis.

Pathology detection methods included mycobacterium culture, acid-fast smear staining, immunological detection, DNA detection for NTM and mycobacterium tuberculosis (MTB). Samples sent for examination included sputum, peripheral blood, tracheobronchial aspirates, bronchoalveolar lavage fluid, stool, pus, wound secretions, pleural effusions, ascites fluid, subcutaneous nodules, skin tissue, joint fluid, lymph node biopsies, subcutaneous tissue or bone marrow.

Diagnostic steps were as follows: (1) mycobacterium culture: the fully automatic BACTEC MGIT 960 Mycobacterium culture system (Becton, Dickinson U.K.Limited, Berkshire, UK, BD) was used to identify mycobacteria. Cultures can be positive after 2–4 weeks of culture but were reported negative after 42 days of no growth. Samples of culture-positive mycobacteria were screened, smeared and inoculated with Mycobacterium Roche Medium for preliminary mycobacterium identification (20). It is used to distinguish patients with and without mycobacterial infection (2) immunological detection method: after mycobacterium detection on culture, the immunological monitoring method and the MPB64 antigen colloidal gold method were used to monitor for antigens of MTB. If this antigen monitoring was negative, the sample was considered to be NTM. After the initial differentiation, molecular testing was performed to determine the type of mycobacteria (3). Molecular detection: NTM or MTB were then confirmed by polymerase chain reaction (PCR) and NTM species identification was carried out by chip hybridization or DNA sequencing. NTM species identification was performed using gene chip (CapitalBio Corporation, Chengdu, China). The DNA microarray chip method was used to identify the common clinical mycobacteria species. We performed rpoB DNA sequencing to identify the rare NTM species. The DNA sequencing service was provided by gene sequencing company (RuiBiotech Corporation, Beijing, China).

This study conformed to the Declaration of Helsinki and was approved by the medical ethics committee of the host institution (ethics approval number: S-K1673).

Statistical analysis was conducted using SPSS 22.0 software (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA). The Kolmogorov-Smimov method was used to test whether the data conformed to a normal distribution. Continuous variables with normal distribution were expressed as means ± standard deviations (SD), while continuous variables with non-normal distribution were expressed as medians with interquartile ranges. Categorical variables were expressed using frequencies and percentages. A non-parametric test (the Mann–Whitney U test or Kruskal–Wallis test) was used for analyzing data with non-normal distributions or heterogeneity of variances. Categorical variables were used with the Chi-square test or Fisher's exact test.

We analyzed demographic, clinical, and laboratory data to identify predictors of all-cause mortality among patients with NTM. Specifically, we considered age, gender, body mass index (BMI), comorbidities, and lymphocyte count, immunosuppression, disseminated NTM infection, and hemoglobin (Hgb) levels. The univariate and multivariate association of demographic, clinical, and laboratory test factors with risk of mortality was assessed using Cox proportional hazards regression models. Hazard ratios (HRs) were calculated using the Cox proportional hazard model. Statistical significance was assessed at P < 0.05. Kaplan-Meier survival curves were calculated, with the significance of difference in the survival distribution function estimated using the log rank test.

A total of 480 suspected NTM patients completed species identification between January 1, 2013 and December 31, 2020 and we found a total of 38 NTM species. After reviewing microbiological results, clinical characteristics, and radiologic findings, there were 118 NTM patients who met ATS/IDSA NTM disease diagnostic criteria and had complete medical records were included for further analysis (Figure 1).

Among the 118 patients, 11 NTM species were identified. The most common species were M. intracellulare (45/118, 38.1%), followed by M. abscessus (26/118, 22.0%), M. avium (13/118, 11.0%) and M. kansasii (10/118, 8.5%). Others included M. gordonae (4/118, 3.4%), M. fortuitum (4/118, 3.4%) and coinfection (11/118, 9.32%). The rare NTM included M. xenopi (1/118, 0.8%), M. szulgai (1/118, 0.8%), M. colombiense (1/118, 0.8%), M. paraseoulense (1/118, 0.8%) and M. lentiflavum (1/118, 0.8%). These rare NTM species were all found in immunocompromised patients. We compared the percentage of species between immunosuppressed and immunocompetent hosts. There was no significant difference in NTM species identification between immunosuppressed and immunocompetent patients (P = 0.224), and the percentage of mycobacteria species were shown in Figure 2.

Patients were classified according to their immune status, as shown in Table 1. Among them, primary immunodeficiency diseases were found in 10 patients: three patients with cellular immune dysfunction, five patients with adult-onset immunodeficiency, one patient with MonoMAC syndrome, and one patient with hypoimmunoglobulinemia.

A comparison of patient characteristics and clinical manifestations between immunocompetent and immunocompromised patients with NTM infections is shown in Table 2. NTM patients in immunocompromised group had more comorbidities compared to immunocompetent group: chronic kidney disease (17.2 vs. 3.7%) (P = 0.02), fever (87.5 vs. 66.7%) (P = 0.007), bone and joint pain (25 vs. 5.6%) (P = 0.004), disseminated NTM infection (35.9 vs. 11.1%) (P = 0.002). Immunocompromised patients also had more frequent exacerbations than immunocompetent patients (59.4 vs. 27.8%) (P = 0.001).

The mean Charlson Comorbidity Index (CCI) score (19) was 2.38 in total patients, and 1.2 in immunocompetent patients (vs. 3.4 in immunocompromised patients, P < 0.001) (Figure 3). 37.3% (44/118) of those patients had a CCI score of 3 or higher.

In 118 NTM patient, 30 (25.4%) patients died during the follow-up period. Among them, 8 (14.8%) cases and 22 (34.3%) cases were from immunocompetent and immunocompromised patients, respectively. The median age at death was 50 years. The median length of follow-up was 41 month. The overall mortality rate was 9.6 per 100 person-years. Cumulative mortality was 16.9% at 1 years, 24.5% at 3 years and 25.4% at 5 years in total patients, 7.4% at 1 years, 12.9% at 3 years and 12.9% at 5 years in immunocompetent patients, and 25% at 1 years, 34.3% at 3 years and 35.9% at 5 years in immunocompromised patients, respectively. The all-cause mortality rate is higher and earlier in immunocompromised patients than immunocompetent patients.

Of the 30 patients, 23 (76.7%) patients died at the end of the observation whose primary cause of death was thought to be NTM disease or related to NTM infection. Among them, nine patients had respiratory failure due to pulmonary multiple infections including NTM: five cases with aspergillus, one case with actinomycete, one case with staphylococcus aureus, one case with Burkholderia cepacia, one case with pseudomonas aeruginosa. Other causes of death included one patient died of acute myocardial infarction, one died of gastrointestinal bleeding, three died of malignant tumor, two died of pulmonary alveolar proteinosis.

A Kaplan-Meier survival curve analysis showed that all-cause mortality in immunocompromised patients was significantly higher than that in immunocompetent patients (Figure 4). There is a statistically significant difference between curves (reference category = immunocompetent, HR 3.537, 95%CI 1.526–8.362, P = 0.003, log-rank) in crude survival comparison.

The CD4⁺ T lymphocyte counts, as well as lymphocyte, T lymphocyte and B lymphocyte counts (/μL), were significantly lower in non-survivors compared with survivors (Figure 5). These data were measured at diagnosis of NTM disease.

Cox proportional hazard regression analyses was conducted to assess risk factors related to all-cause mortality, including baseline factors (age, gender, BMI, and CCI), immunosuppression, disseminated NTM infection, CD4⁺ T lymphocyte count, B lymphocyte count, and Hgb at the time of NTM diagnosis. The univariate model included each factor separately, and the multivariate models included these potential predictors simultaneously. CD4⁺ T lymphocyte count were correlated with B lymphocyte count (Pearson correlation coefficient = 0.364, P = 0.001) and immunosuppression, so they were analyzed separately. In univariable analysis, the prognostic factors that appeared to be significantly associated with mortality included immunosuppression, disseminated NTM infection, CD4⁺ T lymphocyte count, B lymphocyte count, and Hgb. In multivariate analysis, after adjust age, gender, BMI, CCI, disseminated NTM infection, the result showed that decreased CD4⁺ T and B lymphocyte count and decreased Hgb were all independent predictors of mortality in NTM patients (Table 3).

Several limitations in this study should be acknowledged. This was a single-center retrospective study, which has its inherent limitations. The sample size included in our study was relatively small, and the result may only be representative of a subset of the population. Additionally, only about a quarter of NTM patients were able to complete analysis due to a lack of complete medical records. This was a conscious choice on our part, as we wanted to maximize the number of patients with complete data, rather than the total number of patients alone. NTM infections are quite rare, so being able to find over 100 cases with full data for analysis is still remarkable and has certain representativeness. Finally, we didn't include factors such as drug combinations, duration of medications, or patient compliance during NTM infection, which may affect patient outcomes. Future work should include prospective, controlled studies to evaluate prognosis and confirm our findings.