Analysis of the differential clinical characteristics of tuberculosis and nontuberculous mycobacterial diseases among AIDS patients

Abstract

Aims

This study aims to systematically compare the clinical characteristics of tuberculosis (TB) and nontuberculous mycobacterial (NTM) diseases in AIDS patients,and to identify independent predictors for differential diagnosis.

Methods

Clinical data of AIDS patients co-infected with TB or NTM at Shanghai Public Health Clinical Center (January 2019 – January 2024) were retrospectively analyzed. Univariate comparisons were performed using t-test, Mann-Whitney U test, χ² or Fisher's exact test, with Bonferroni correction for multiple comparisons. Multivariate binary logistic regression (Enter method) was used to adjust for age, gender, CD4+ T-cell count, C-reactive protein (CRP), procalcitonin, neutrophil count, miliary nodules, and superficial lymphadenopathy.

Results

A total of 494 patients were included (AIDS/TB: 206, AIDS/NTM: 288). The predominant NTM species was Mycobacterium avium (68.2%), followed by Mycobacterium kansasii (13.6%) and Mycobacterium intracellulare (10.9%). After Bonferroni correction, AIDS/NTM patients had significantly higher rates of Pneumocystis jirovecii pneumonia, cytomegalovirus infection, and progressive multifocal leukoencephalopathy (all P < 0.0083). Among clinical symptoms, only enlarged lymph nodes remained significantly more common in the TB group after correction (P = 0.002). Multivariate analysis showed that male gender (adjusted OR for NTM vs. TB = 0.451, 95% CI: 0.208–0.979, P = 0.044), higher CD4+ count (per 10 cells/μL: OR = 0.98, 95% CI: 0.96–0.99, P = 0.005), higher CRP (per 10 mg/L: OR = 0.90, 95% CI: 0.86–0.95, P < 0.001), higher neutrophil count (OR = 0.903 per 1 × 10(Akokuebere et al., 2024⁹)/L, 95% CI: 0.836–0.976, P = 0.010), presence of miliary nodules (OR = 0.079, 95% CI: 0.027–0.233, P < 0.001), and presence of superficial lymphadenopathy (OR = 0.565, 95% CI: 0.327–0.977, P = 0.041) were independently associated with lower odds of NTM disease (i.e., associated with TB). Imaging revealed that only miliary nodules remained significantly more common in TB after Bonferroni correction (P < 0.001).

Conclusions

While AIDS/TB and AIDS/NTM share many clinical similarities, the presence of miliary nodules, superficial lymphadenopathy, higher inflammatory markers (CRP, neutrophils), higher CD4+ count, and male gender favor TB. These findings can assist clinicians in differentiating the two infections when rapid microbiological results are unavailable. The high prevalence of Mycobacterium avium (68.2%) suggests that empirical therapy for suspected NTM in severely immunocompromised AIDS patients in Shanghai should cover the Mycobacterium avium complex (MAC). Prospective multicenter studies with pre-specified outcomes are needed to validate our findings.

1. Introduction

The Mycobacterium genus includes Mycobacterium tuberculosis (MTB), nontuberculous mycobacteria (NTM), and Mycobacterium leprae. In patients with AIDS, mycobacterial infection is a common opportunistic infection, mainly involving infections by Mycobacterium tuberculosis and nontuberculous mycobacteria.

Tuberculosis is a chronic infectious disease caused by Mycobacterium tuberculosis that is transmitted through the respiratory tract and remains one of the top 10 causes of mortality worldwide [1], [2]. HIV is the most powerful risk factor for TB [3], [4], [5]. TB is the most common opportunistic infection and leading cause of mortality in people living with HIV/AIDS(PLWH). According to the 2024 WHO Global Tuberculosis Report, an estimated 10.8 million individuals worldwide were diagnosed with TB in 2023. Of these cases, 6.1% occurred in the HIV-positive population, approximately 161,000 deaths were attributed to HIV-associated tuberculosis among this comorbid cohort. Meta-analysis showed that the HIV/TB co-infection rate was around 23.51% [6], [7]. The incidence of tuberculosis is nearly 14 times higher among PLWH than in the general population, and the mortality rate is two times higher [8].。The clinical presentation of tuberculosis in the AIDS population is different from that of the general population, with atypical symptoms that make diagnosis difficult [5], [9], [10], [11]. The diagnosis of TB in the AIDS population requires a combination of clinical presentation, laboratory tests, imaging, pathogenetic testing, and pathology [8], [10], [12].

NTM are a large group of naturally-occurring mycobacteria, ubiquitous in the environment like soil and water [13], [14], [15]. Over 190 species and subspecies of NTM have been documented, many of which are known to produce disease in humans [16], [17]. In China, the common pathogenic NTM species include M. avium, M. intracellularae, M. kansasii, M. abscessus and M. fortuitum，accounting for 90% of the total cases reported [16]. NTM is known to cause lymphadenitis, skin infections, and catheter-related infections, but pulmonary infection is the most prevalent clinical presentation. Both of NTM and TB cause pulmonary and extra-pulmonary disease, share similar clinical presentations and can coexist [15], [18], [19]. NTM exhibits characteristics such as slow growth, long course, and high resistance to first-line anti-tuberculosis drugs, resulting in a low total care rate [20]. Disseminated NTM infection is often associated with an infection by the HIV. Meningitis, osteomyelitis, soft tissue infection and skin lesions are extrapulmonary sites of disseminated disease that have been reported [13]. The incidence and prevalence of NTM are increasing in many countries，becoming one of the major public health problems that threaten human health worldwide [21], [22], [23]. NTM are pathogens acknowledged worldwide, especially affecting those with compromised immune systems, like HIV/AIDS patients [24], [25]. Among PLWH, disseminated NTM infections are more frequently recognized and Mycobacterium avium-intracellulare complex (MAIC) is the most common species [13], [26].

NTM diseases share clinical signs with TB, causing a clinical dilemma in differentiation between NTM infection and TB. However, the existing diagnostic methods have limitations in terms of early diagnosis or accuracy and are invasive or inconvenient for follow-up examinations [13], [27]. Consequently, we aimed to retrospectively analyze the clinical characteristics of TB and NTM diseases in HIV/AIDS patients to increase physician awareness of the clinical features and harmfulness of these conditions. This study is exploratory and descriptive in nature, with no pre-specified primary outcome. Its purpose is to systematically compare the clinical features between the two groups and generate hypotheses for future prospective studies.

2. Materials and methods

2.1. Research objects

This retrospective study enrolled hospitalized patients diagnosed with AIDS/TB or AIDS/NTM at the Department of Infectious Diseases and Immunology of the Shanghai Public Health Clinical Center between January 2019 and January 2024. Disease diagnostic criteria are consistent with guidelines and expert consensus [4], [13], [16], [28], [29], [30], [31]. The clinical data were retrospectively reviewed. This study was approved by the ethics committee of our hospital.

Inclusion criteria: (1) age ≥ 18 years; (2) confirmed diagnosis of AIDS according to the Chinese Guidelines for the Diagnosis and Treatment of AIDS (2021 Edition); (3) definite diagnosis of either TB or NTM disease based on the diagnostic criteria specified below.

Exclusion criteria: (1) concurrent infection with both TB and NTM during the same disease episode; (2) incomplete medical records (e.g., missing key laboratory, imaging, or follow-up data); (3) failure to meet the microbiological or clinical diagnostic criteria for TB or NTM disease (e.g., only one positive sputum culture without clinical correlation, or isolated culture of environmental NTM species such as M. gordonae without evidence of disease); (4) age < 18 years.

After applying the inclusion and exclusion criteria, a total of 494 patients were included in the final analysis: 288 with AIDS/NTM and 206 with AIDS/TB. No patient with concurrent TB and NTM infection was identified in the final cohort.

The diagnosis of AIDS was made according to the Chinese Guidelines for the Diagnosis and Treatment of AIDS (2021 Edition). All enrolled patients met the diagnostic criteria for AIDS (CD4+ T cell count <200 cells/μL or presence of an AIDS-defining illness).

The diagnosis of TB followed the Chinese national standard “Diagnosis for Pulmonary Tuberculosis (WS 288-2017)” and relevant expert consensus. TB was diagnosed based on at least one of the following: (i) positive acid-fast bacilli smear and/or culture for Mycobacterium tuberculosis from sputum, bronchoalveolar lavage fluid, or extrapulmonary specimens; (ii) positive molecular test (e.g., GeneXpert MTB/RIF) for M. tuberculosis complex; (iii) histopathological evidence of caseating granulomas with positive acid-fast staining.

The diagnosis of NTM disease followed the Chinese “Guidelines for Diagnosis and Treatment of Non-tuberculous Mycobacterial Disease (2020 Edition)” and the ATS/IDSA statement. For pulmonary NTM disease, patients required compatible respiratory symptoms and imaging findings, exclusion of other diseases (including TB), and at least one of the following microbiological criteria: (i) two or more separate sputum cultures positive for the same NTM species; or (ii) one positive culture from bronchial wash or bronchoalveolar lavage; or (iii) positive culture from lung biopsy. For disseminated or extrapulmonary NTM disease, diagnosis required positive NTM culture from blood, bone marrow, or a sterile extrapulmonary site. NTM isolates underwent species identification by molecular methods . Patients with concurrent TB and NTM infection were excluded from the analysis.

2.2. Statistical analysis

Statistical analyses were performed using SPSS for Windows, version 20.0 (SPSS, Chicago, Illinois). Continuous variables were first tested for normality using the Shapiro–Wilk test. Normally distributed continuous variables were reported as mean ± standard deviation (SD) and compared between groups using the independent sample t-test. Non-normally distributed continuous variables were reported as median with interquartile range (IQR) and compared using the Mann-Whitney U exact test (two-tailed). Categorical variables were expressed as counts and percentages (n, %) and compared using the Pearson chi-square test; when any expected cell count was less than 5, Fisher‘s exact test was used instead.

Given the exploratory and descriptive nature of this study, no primary outcome was pre-specified. For univariate baseline comparisons (Table 2, Table 3, Table 4, Table 5), we did not adjust for multiple comparisons. For secondary exploratory analyses, Bonferroni correction was applied; the adjusted significance thresholds are provided in the footnotes of each relevant table.

Table 2.

Incidence of Opportunistic Infections and HIV-Related Tumors Between NTM Disease and TB in the AIDS Population [n (%)].

Comorbidities	AIDS/NTM (n = 288)	AIDS/TB (n = 206)	χ2	P⁎	Significant after Bonferroni
Pneumocystis jirovecii pneumonia	33(11.5)	7(3.4)	10.485	0.001	Yes
Cryptococcal infection	20(6.9)	8(3.9)	2.105	0.147	No
Cytomegalovirus infection	24(8.3)	2(1.0)	13.056	<0.001	Yes
Penicilliosis marneffei infection	12(4.2)	1(0.5)	6.352	0.012	No
Progressive multifocal cerebral white matter lesions	4(1.4)	0(0)	/	0.002†	Yes
HIV-Related Tumors	13(4.5)	8(3.9)	0.117	0.732	No

† Fisher's exact test.

Bonferroni correction: 6 comparisons were made in this table; the adjusted significance threshold is α’ = 0.05 / 6 ≈ 0.0083. Differences with P < 0.0083 are considered statistically significant after correction. Therefore, the incidence of Pneumocystis jirovecii pneumonia, cytomegalovirus retinitis/disseminated infection, and progressive multifocal leukoencephalopathy remained significantly higher in the AIDS/NTM group. The difference in penicilliosis (P = 0.012) is no longer significant after correction.

^⁎P values are uncorrected for multiple comparisons.

Table 3.

Comparison of Clinical Symptoms and Outcomes Between NTM Disease and Tuberculosis in AIDS Patients [n (%)].

Chief Complaint/ Outcome	AIDS/NTM (n = 288)	AIDS/TB (n = 206)	χ2	P⁎	Significant after Bonferroni
Fever	148(51.4%)	123(59.7%)	3.357	0.067	No
Cough	140(48.6%)	94(45.6%)	0.428	0.513	No
Expectoration	94(32.6%)	62(30.1%)	0.359	0.549	No
Fatigue	68(23.6%)	39(18.9%)	1.550	0.213	No
Poor Appetite	53(18.4%)	38(18.4%)	0.000	0.990	No
Weight Loss	42(14.6%)	21(10.2%)	2.079	0.149	No
Diarrhea	25(8.7%)	13(6.3%)	0.950	0.330	No
Shortness of Breath	34(11.8%)	16(7.8%)	2.153	0.142	No
Abdominal Pain	11(3.8%)	16(7.8%)	3.622	0.057	No
Chest Discomfort	17(5.9%)	9(4.4%)	0.567	0.452	No
Abdominal Distension	10(3.5%)	4(1.9%)	1.022	0.312	No
Enlarged Lymph Nodes	4(1.4%)	14(6.8%)	10.001	0.002	Yes
No Obvious Symptoms	17(5.9%)	4(1.9%)	2.915	0.088	No
Prognosis (Death)	42(14.6%)	26(12.6%)	0.389	0.533	No

Bonferroni correction: There are 14 comparisons in this table (13 symptom rows +1 prognosis row). The adjusted significance threshold is α’ = 0.05 / 14 ≈ 0.00357. Only the difference in enlarged lymph nodes (P = 0.002) remains statistically significant after correction. All other P values are >0.00357.

^⁎P values are uncorrected.

Table 4.

Comparison of Laboratory Findings Between NTM and TB Patients in the AIDS Population [Mean ± Standard Deviation].

Laboratory Findings	AIDS/NTM (n = 288)	AIDS/TB (n = 206)	t/Z	P⁎	Significant after Bonferroni
White blood cell count⁎10^9/L	4.85 ± 2.709	5.92 ± 3.882	−3.408	0.001†	Yes
Absolute value of neutrophils⁎10^9/L	3.45 ± 2.401	4.46 ± 3.042	−3.956	<0.001†	Yes
Hemoglobin g/L	100.75 ± 26.250	101.92 ± 22.851	−0.527	0.599	No
Albumin g/L	32.48 ± 6.179	31.59 ± 6.387	1.546	0.123	No
C-reactive protein mg/L	26.80(6.63,58.75)	58.9(20.13,99.65)	−5.155	<0.001	Yes
Erythrocyte sedimentation rate mm/h	64.04 ± 29.074	67.39 ± 32.190	−0.978	0.329	No
Procalcitonin ng/mL	0.12(0.05,0.33)	0.16(0.06,0.61)	−1.997	0.046	No
CD3+ T lymphocyte count cell/μL	457.3(214.71,688.99)	538.0(286.00,730.9)	−1.026	0.305	No
CD4 + T lymphocyte count cell/μL	30.91(8.99,83.10)	70.0(23.75,113.00)	−5.194	<0.001	Yes
CD8 + T lymphocyte count cell/μL	372.21 (185.10，579.84)	403.79(213.25,536.00)	−0.102	0.919	No
CD45 + T lymphocyte count cell/μL	638.76(320.75,949.01)	685.50(384.25,935.08)	−0.547	0.585	No
CD4/CD8 ratio	0.09(0.03,0.21)	0.17(0.09,0.26)	−6.310	<0.001	Yes
HIV-RNA⁎10^4copy/mL	41.40(4.84,1210.02)	64.24(12.25,1210.02)	−1.616	0.106	No

† t-test; all others Mann-Whitney U.

Bonferroni correction: There are 13 comparisons in this table. The adjusted significance threshold is α’ = 0.05 / 13 ≈ 0.00385. Only parameters with P < 0.00385 remain statistically significant after correction. Therefore, white blood cell count (P = 0.001), neutrophil count (P < 0.001), C-reactive protein (P < 0.001), CD4+ T cell count (P < 0.001), and CD4/CD8 ratio (P < 0.001) are significant. Procalcitonin (P = 0.046) is not significant after correction.

^⁎P values are uncorrected. For normally distributed variables: independent t-test; for non-normally distributed: Mann-Whitney U test.

Table 5.

Comparison of Imaging Findings Between NTM and TB Patients in the AIDS Population [n (%)].

Imaging Features	AIDS/NTM (n = 288)	AIDS/TB (n = 206)	χ2	P⁎	Significant after Bonferroni
Patchy and Linear Opacities in the Lungs	197(68.4%)	133(64.6%)	0.798	0.372	No
Pulmonary Nodular	78(27.1%)	56(27.2%)	0.001	0.980	No
Pulmonary Miliary Nodules	4(1.4%)	33(16.0%)	37.101	<0.001	Yes
Pulmonary Cavities	26(9.0%)	25(12.1%)	0.250	0.617	No
Mediastinal Lymphadenopathy	79(27.4%)	60(29.1%)	0.171	0.679	No
Hilar Lymphadenopathy	35(12.2%)	36(17.5%)	2.765	0.096	No
Pleural Effusion	70(24.3%)	74(35.9%)	7.847	0.005	No†
Pericardial Effusion	17(5.9%)	26(12.6%)	6.822	0.009	No†
Abdominal, Retroperitoneal, and Mesenteric Lymphadenopathy	50(17.4%)	38(18.4%)	0.097	0.756	No
Superficial Lymphadenopathy	34(11.8%)	41(19.9%)	6.115	0.013	No†
Pelvic and Abdominal Effusion	8(2.8%)	4(1.9%)	0.354	0.552	No
Intracranial Infectious Lesions	4(1.4%)	4(1.9%)	0.014	0.906	No

Bonferroni correction: There are 12 comparisons in this table. The adjusted significance threshold is α’ = 0.05 / 12 ≈ 0.0042. Only pulmonary miliary nodules (P < 0.001) remain statistically significant after correction. Pleural effusion (P = 0.005), pericardial effusion (P = 0.009), and superficial lymphadenopathy (P = 0.013) are no longer significant after correction.

† Not significant after Bonferroni correction.

^⁎P values are uncorrected.

To identify independent factors associated with NTM versus TB disease, multivariate binary logistic regression analysis was performed with disease type (NTM = 1, TB = 0) as the dependent variable. Variables with a P value <0.05 in univariate analysis or considered clinically important (CD4+ T-cell count, age, gender, C-reactive protein, lymph node enlargement, and pulmonary miliary nodules) were entered as covariates. A forward stepwise (likelihood ratio) method was used for variable selection. Model calibration was assessed using the Hosmer-Lemeshow goodness-of-fit test. Results are presented as adjusted odds ratios (OR) with 95% confidence intervals (CI).

A two-sided P value <0.05 was considered statistically significant in unadjusted analyses, unless otherwise specified for Bonferroni-corrected thresholds.

3. Results

3.1. Baseline characteristics

This study retrospectively enrolled a total of 494 hospitalized patients diagnosed with AIDS/NTM disease and AIDS/TB disease at the Department of Infectious Diseases and Immunology, Shanghai Public Health Clinical Center. Patient enrollment spanned from January 2019 to January 2024.

A total of 288 patients were diagnosed with AIDS/NTM disease, representing 58.3% of the study population. Of these patients, 275 were male (95.5%) and 13 were female (4.5%). The mean age was 43.34 ± 14.00 years. The age range spanned from 18 to 88 years. The average length of hospitalization was 21.42 ± 20.116 days.

Of the study population, 206 patients (41.7%) were diagnosed with AIDS/TB disease, including 186 males (90.3%) and 20 females (9.7%). The mean age was 43.26 ± 13.41 years, with a median age of 41.0 years (interquartile range: 32.0–52.2 years). The average length of hospitalization was 18.46 ± 22.365 days.

A statistically significant difference was observed in the gender distribution between the two groups (χ² = 6.199, P = 0.023), with a higher proportion of males in the AIDS/NTM disease group. No significant differences were found in mean age (t = 0.063, P = 0.950) or average length of hospitalization (t = 1.539, P = 0.124), as presented in Table 1.

Table 1.

Comparison of Baseline Characteristics Between NTM Disease and TB in the AIDS Population [n (%)].

	AIDS/NTM (n = 288)	AIDS/TB (n = 206)	χ2/t	P
Gender Male Female	275(95.5%) 13(4.5%)	186(90.3%) 20(9.7%)	6.199	0.023
Age (Years)	43.34 ± 14.00	43.26 ± 13.41	0.063	0.950
Mean Length of Hospitalization (Days)	21.42 ± 20.116	18.46 ± 22.365	1.539	0.124

4. Comorbidities

The prevalence of NTM disease and tuberculosis, in conjunction with opportunistic infections and HIV-related malignancies among individuals with AIDS, is summarized in Table 2. The proportion of AIDS/TB patients with opportunistic infections was found to be significantly lower than that of AIDS/NTM patients. Statistical analysis revealed that, compared with AIDS/NTM patients, AIDS/TB patients exhibited a significantly lower incidence of Pneumocystis jirovecii pneumonia (χ² = 10.485, P = 0.001), cytomegalovirus infection (χ² = 13.056, P < 0.001), Penicilliosis marneffei infection (χ² = 6.352, P = 0.012), and progressive multifocal leukoencephalopathy (P = 0.002). However, after Bonferroni correction for six comparisons (adjusted α = 0.0083), the difference in Penicilliosis marneffei infection (P = 0.012) was no longer statistically significant. The cryptococcal infection (P = 0.147) and HIV-related tumors (P = 0.732) remained non-significant.

5. Clinical manifestations and prognosis

The clinical features and prognostic outcomes of patients with AIDS/NTM and AIDS/TB are summarized in Table 3. The results analysis revealed that fever was the most common chief complaint in both the AIDS/NTM and AIDS/TB groups, followed by cough and expectoration. Fatigue and anorexia were also frequently reported symptoms. Except for a significantly higher proportion of patients in the AIDS/TB group presenting with lymphadenopathy (χ² = 10.001, P = 0.002), no other clinical symptom or prognosis showed a statistically significant difference between the two groups. After Bonferroni correction for 14 comparisons (α’ = 0.00357), only lymphadenopathy (P = 0.002) remained significant.

6. Laboratory findings

6.1. NTM pathogen profile

Among AIDS/NTM patients, the three most common NTM species were Mycobacterium avium, Mycobacterium kansasii, and Mycobacterium intracellulare. As illustrated in Fig. 1, of the 288 AIDS/NTM patients, 110 had confirmed pathogen typing. Mycobacterium avium was the most prevalent, identified in 75 cases (68.2%), followed by Mycobacterium kansasii with 15 cases (13.6%) and Mycobacterium intracellulare with 12 cases (10.9%). Several rare NTM species were also identified, including Mycobacterium gordonae (3 cases, 2.7%), Mycobacterium chelonae (2 cases, 1.8%), Mycobacterium abscessus (2 cases, 1.8%), Mycobacterium fortuitum (2 cases, 1.8%), and Mycobacterium colombiense (1 case, 0.9%). Co-infections with multiple NTM species were observed in two patients: one infected with both Mycobacterium avium and Mycobacterium kansasii, and another with both Mycobacterium avium and Mycobacterium colombiense. These findings confirm that the predominant pathogens among AIDS/NTM patients are Mycobacterium avium, Mycobacterium kansasii, and Mycobacterium intracellulare.

Fig. 1.

Pathogen Distribution Among 110 Cases of NTM.

6.2. Other laboratory findings

The laboratory findings for NTM disease and TB disease among AIDS patients are presented in Table 4. Statistical analysis reveals that AIDS patients with NTM disease exhibit significantly lower levels of white blood cell count (4.85 ± 2.709 vs. 5.92 ± 3.882, t = −3.408, P = 0.001), absolute neutrophil count (3.45 ± 2.401 vs. 4.46 ± 3.042, t = −3.956, P < 0.001), C-reactive protein [26.80 (6.63, 58.75) vs. 58.9 (20.13, 99.65), Z = −5.155, P < 0.001], procalcitonin [0.12 (0.05, 0.33) vs. 0.16 (0.06, 0.61), Z = −1.997, P = 0.046], CD4+ T lymphocyte count [30.91 (8.99, 83.10) vs. 70.00 (23.75, 113.00), Z = −5.194, P < 0.001], and CD4/CD8 ratio [0.09 (0.03, 0.21) vs. 0.17 (0.09, 0.26), Z = −6.310, P < 0.001] compared to those with TB disease. After Bonferroni correction for 13 comparisons (α’ = 0.00385), white blood cell count, neutrophil count, C-reactive protein, CD4+ T-cell count, and CD4/CD8 ratio remained significantly lower in the AIDS/NTM group (all P < 0.001). Procalcitonin (P = 0.046) was no longer significant.In summary, significant differences exist between AIDS/NTM and AIDS/TB groups in terms of these immunological and inflammatory markers.

7. Imaging examination

The imaging characteristics of NTM and TB diseases in the AIDS population are presented in Table 5. Analysis of the results demonstrates that pulmonary miliary nodules are significantly more prevalent in TB patients (χ² = 37.101, P < 0.001), as are pleural effusion (χ² = 7.847, P = 0.005), pericardial effusion (χ² = 6.822, P = 0.009), and superficial lymphadenopathy (χ² = 6.115, P = 0.013). After Bonferroni correction for 12 comparisons (α’ = 0.0042), only pulmonary miliary nodules remained significantly more prevalent in TB patients (P < 0.001). Pleural effusion (P = 0.005), pericardial effusion (P = 0.009), and superficial lymphadenopathy (P = 0.013) were no longer statistically significant after correction.

8. Multivariate logistic regression analysis

To adjust for all prespecified confounders, we performed binary logistic regression using the Enter method, forcing all candidate variables into the model simultaneously. Candidate variables included CD4+ T-cell count, age, gender, C-reactive protein (CRP), procalcitonin (PCT), neutrophil count, miliary nodules (coded as present vs. absent), and superficial lymphadenopathy (coded as present vs. absent). The results are presented in Table 6.

Table 6.

Multivariate logistic regression analysis (Enter method) of factors associated with NTM disease (vs. TB) in AIDS patients.

Variable	Adjusted OR	95% CI	P value
CD4+ T-cell count (per 1 cell/μL increase) †	0.998	0.996–0.999	0.005
Age (per 1 year increase)	0.999	0.984–1.013	0.842
Gender (male vs. female)	0.451	0.208–0.979	0.044
C-reactive protein (per 1 mg/L increase) †	0.990	0.985–0.995	<0.001
Procalcitonin (per 0.1 ng/mL increase)	1.000	0.979–1.023	0.968
Neutrophil count (per 1 × 109/L increase)	0.903	0.836–0.976	0.010
Miliary nodules (present vs. absent)	0.079	0.027–0.233	<0.001
Superficial lymphadenopathy (present vs. absent)	0.565	0.327–0.977	0.041

† For clinical interpretability, the OR per 10 cells/μL for CD4+ count is 0.98 (95% CI: 0.96–0.99), and per 10 mg/L for CRP is 0.90 (95% CI: 0.86–0.95).

Dependent variable: 1 = AIDS/NTM, 0 = AIDS/TB. All candidate variables were entered simultaneously (Enter method).

Model fit: Hosmer-Lemeshow test χ² = 17.991, df = 8, P = 0.021; Nagelkerke R² = 0.215.

After adjustment, male gender (adjusted OR = 0.451, 95% CI: 0.208–0.979, P = 0.044), higher CD4+ count (OR per 1 cell/μL = 0.998; per 10 cells/μL: 0.98, 95% CI: 0.96–0.99, P = 0.005), higher CRP (OR per 1 mg/L = 0.990; per 10 mg/L: 0.90, 95% CI: 0.86–0.95, P < 0.001), and higher neutrophil count (OR = 0.903 per 1 × 10⁹/L, 95% CI: 0.836–0.976, P = 0.010) were significantly associated with lower odds of NTM disease (i.e., associated with TB). The presence of superficial lymphadenopathy was also associated with lower odds of NTM disease (OR = 0.565, 95% CI: 0.327–0.977, P = 0.041). Notably, the presence of miliary nodules showed a strong association with lower odds of NTM disease (OR = 0.079, 95% CI: 0.027–0.233, P < 0.001), confirming both miliary nodules and superficial lymphadenopathy as independent predictors of TB rather than NTM. Age and procalcitonin were not significantly associated with the outcome (both P > 0.05).

The Hosmer-Lemeshow test indicated a modest lack of fit (χ² = 17.991, df = 8, P = 0.021), which may be attributable to the large sample size and the inclusion of several continuous variables. The Nagelkerke R² was 0.215, suggesting that the model explains approximately 21.5% of the variance in the outcome.

9. Discussion

AIDS patients, due to impaired immune function, are particularly susceptible to opportunistic infections such as TB and NTM infections. These infections not only compromise patients' quality of life but also significantly elevate both hospitalization and mortality rates. The clinical manifestations of NTM disease closely resemble those of TB, making early differentiation challenging even for experienced clinicians. A comprehensive analysis of the clinical characteristics of TB and NTM infections in the AIDS population can provide valuable insights for clinical decision-making and patient management.

This study conducted a comparative analysis of the clinical characteristics of NTM and TB diseases among the AIDS population, focusing on multiple aspects such as general patient information, clinical manifestations, laboratory findings, and imaging features. The results revealed both similarities and certain distinctions in demographic data and clinical features between the two groups.The male predominance in the AIDS/NTM group (95.5%) likely reflects the underlying HIV epidemic among men in Shanghai, as well as possible sex-related differences in healthcare access or environmental exposure to NTM; however, further studies are needed to clarify this observation.Interestingly, although the crude male proportion was higher in the NTM group, multivariate analysis identified male gender as an independent factor associated with TB (adjusted OR for NTM = 0.451). This indicates that the unadjusted gender disparity may be confounded by other covariates, such as immune status and healthcare-seeking behavior.Patients with AIDS/TB have a lower incidence of opportunistic infections compared to those with AIDS/NTM. This difference may be attributed to the generally lower immune status observed in AIDS/NTM patients, as evidenced by laboratory findings showing significantly reduced CD4+ T lymphocyte counts in this group. Individuals with compromised immunity, particularly those with CD4+ T lymphocyte counts below 50 cells/μL, are at increased risk for multiple opportunistic infections.In terms of clinical manifestations, both AIDS/TB and AIDS/NTM infections commonly present with symptoms including fever, cough, expectoration, fatigue, and anorexia. Differentiation between the two diseases based solely on clinical features is challenging. Notably, a higher proportion of patients in the TB group sought medical care primarily due to lymphadenopathy.The atypical nature of clinical manifestations and the high degree of similarity between the two conditions align with findings reported in previous studies [11], [32], [33], [34].Tuberculosis is typically classified into pulmonary tuberculosis and extrapulmonary tuberculosis, with clinical manifestations varying according to the specific type. While pulmonary tuberculosis predominates in the general population, AIDS patients tend to exhibit more variable and extensive lesions. Consequently, their clinical presentations are often atypical compared to those observed in non-AIDS individuals [4], [10].Lymph node tuberculosis represents one of the most common forms of extrapulmonary tuberculosis, with cervical lymph node involvement being the most frequently observed subtype. In AIDS patients, the presence of lymph node fusion, along with abscess formation and sinus tract development, serves as a key indicator of lymph node tuberculosis [10].NTM are opportunistic pathogens that primarily affect individuals with compromised immune systems. Pulmonary involvement is the most common clinical manifestation. In AIDS patients, NTM infections are more likely to progress to disseminated disease, and their clinical presentations often differ from those observed in immunocompetent individuals with pulmonary NTM disease.

In terms of laboratory findings, AIDS/TB patients exhibit higher white blood cell and neutrophil counts, as well as elevated levels of high-sensitivity C-reactive protein and procalcitonin, compared to those with AIDS/NTM disease. Of note, although procalcitonin was significantly lower in the NTM group in univariate analysis (P = 0.046), this difference no longer remained significant after Bonferroni correction (α’ = 0.00385).Additionally, the CD4+ T lymphocyte count and CD4/CD8 ratio are significantly lower in the AIDS/NTM disease group. NTM are considered opportunistic pathogens. While NTM infections are relatively uncommon in immunocompetent individuals, they frequently occur in HIV-infected individuals with severely compromised immunity. Therefore, clinical observations have shown that patients with NTM disease often present with markedly reduced CD4+ T lymphocyte counts.

To adjust for confounding by immune status, we performed multivariate logistic regression. After adjustment, male gender, higher CD4+ count, higher CRP, higher neutrophil count, presence of miliary nodules, and presence of superficial lymphadenopathy remained independently associated with TB (all OR for NTM ≤ 0.565, P < 0.05). These findings confirm that the observed clinical differences are not solely driven by CD4+ levels but reflect pathogen-specific responses. Notably, miliary nodules showed the strongest association with TB (OR = 0.079), followed by superficial lymphadenopathy (OR = 0.565). The Hosmer-Lemeshow test gave a P value of 0.021, indicating a modest lack of calibration, which we acknowledge as a limitation.

Imaging features of AIDS/TB and AIDS/NTM diseases demonstrate significant similarities. Pulmonary involvement typically presents with nodules, cavities, and patchy opacities. Hilar and mediastinal lymphadenopathy is frequently observed. Extrapulmonary manifestations may include multiple enlarged lymph nodes, with some cases exhibiting pericardial effusion or ascites. These radiological findings are consistent with previously published literature [11], [33], [34], [35].Imaging findings alone are insufficient for differential diagnosis between the two conditions. Pulmonary granulomatous nodules are predominantly observed in AIDS/TB patients, and pleural effusion, pericardial effusion, as well as superficial lymphadenopathy, tend to occur more frequently in this group. Although these imaging features lack specificity, they may serve as valuable reference points for clinicians when differentiating between AIDS/TB and AIDS/NTM in clinical practice.Further statistical verification revealed that although the above radiological indicators differed significantly in univariate analysis, after Bonferroni correction for 12 comparisons, only pulmonary miliary nodules remained significantly more prevalent in TB patients (P < 0.001). Pleural effusion, pericardial effusion, and superficial lymphadenopathy failed to maintain statistical significance after correction and should therefore be regarded as merely suggestive trends requiring further validation.

The prevalence of NTM disease in the AIDS population is markedly higher compared to the general population. Clinical manifestations also differ between these groups. Disseminated NTM infection occurs more frequently in AIDS patients, and disseminated NTM disease carries a greater clinical burden than non-disseminated forms, including both pulmonary and extrapulmonary NTM disease [13], [36], [37].Among AIDS patients, particularly those with CD4+ T lymphocyte counts below 50 cells/μL, disseminated NTM disease caused by Mycobacterium avium and Mycobacterium intracellulare is highly prevalent [38].Disseminated NTM infection has the potential to affect multiple organ systems throughout the body. Lesions may disseminate to the lungs, bloodstream, lymph nodes, bone marrow, gastrointestinal tract, urinary tract, and central nervous system [13], [20], [32], [35], [39], [40].Therefore, in clinical practice, diagnosing NTM disease necessitates the collection of specimens from multiple anatomical sites, particularly those exhibiting distinct clinical abnormalities, to obtain etiological evidence for accurate diagnosis.A minority of patients with non-disseminated NTM disease exhibit no apparent clinical symptoms, including respiratory tract manifestations. These cases are either detected incidentally through smear or culture screening of respiratory specimens or diagnosed based on initial presentations of localized lymphadenopathy, findings that are consistent with previous studies [36], [39]. Furthermore, due to their compromised immune systems, AIDS patients are frequently susceptible to co-infections with various pathogens in addition to NTM. Consequently, their clinical presentations tend to be complex and heterogeneous, necessitating careful differential diagnosis in clinical practice.

Because our univariate analyses involved multiple comparisons, we applied Bonferroni correction separately to each table. Most clinically relevant differences (e.g., lower CD4+, CRP, CD4/CD8 ratio in NTM; higher rates of PCP, CMV, PML in NTM; miliary nodules in TB) remained significant after correction. Borderline associations (procalcitonin, pleural effusion, pericardial effusion, superficial lymphadenopathy on imaging) did not survive correction and should be interpreted as exploratory trends.

The pathogen spectrum in our AIDS/NTM cohort showed a strong predominance of Mycobacterium avium (68.2%), followed by Mycobacterium kansasii (13.6%) and Mycobacterium intracellulare (10.9%). This distribution is largely consistent with previous reports from China and other countries. In a multicenter study from Chongqing, Nie et al. also identified MAC as the most common NTM species in HIV-infected patients [37]. Similarly, Guo et al. reported that MAC accounted for approximately 70% of NTM isolates in AIDS patients in Shanghai [42], which aligns closely with our findings. However, the proportion of MAC in our cohort is higher than that reported in some northern Chinese regions, where M. kansasii or rapidly growing mycobacteria (e.g., M. abscessus) are more frequently isolated [16], [41]. This geographic variation may be attributed to differences in soil and water microbiota, climate, and environmental exposure [14], [21].Globally, disseminated MAC disease has long been recognized as a leading opportunistic infection in AIDS patients with profound immunosuppression, particularly when CD4+ T cell counts fall below 50 cells/μL [26], [38]. The median CD4+ count in our NTM group was 30.9 cells/μL, which is well below this threshold, explaining the high prevalence of MAC in our cohort. In contrast, in immunocompetent populations or HIV patients with higher CD4+ counts, other NTM species such as M. kansasii and M. abscessus are more commonly encountered [18], [36].Another contributing factor to the observed predominance of MAC may be diagnostic methodology. Our study relied primarily on culture-based identification followed by molecular confirmation. Studies using molecular methods such as 16S rRNA sequencing or metagenomic next-generation sequencing (mNGS) have reported higher detection rates for fastidious or slow-growing NTM species [43]. However, since MAC is readily cultivable, the high proportion in our cohort likely reflects true epidemiology rather than detection bias.Taken together, our findings underscore that in severely immunocompromised AIDS patients in Shanghai, empirical anti-NTM therapy should cover MAC as the most probable pathogen. Regional differences in NTM species distribution should be considered when interpreting diagnostic results and selecting empirical treatment.

NTM and TB infections pose a significant threat to the health and survival of individuals living with AIDS, representing critical public health concerns that demand heightened attention. Consequently, enhanced management, routine follow-up, and active screening for NTM and TB infections are essential for this vulnerable population.Current diagnostic approaches primarily encompass microscopic smear analysis, isolation and culture of mycobacteria, and molecular diagnostic techniques. Conventional methods such as smear microscopy and culture are associated with low sensitivity and prolonged turnaround times, whereas advanced technologies like GeneXpert and next-generation sequencing require specialized infrastructure and entail substantial economic costs [4], [16], [31], [43].In high tuberculosis burden countries, limited medical resources hinder the identification and classification of mycobacterial species. This study analyzed the clinical features and their distinguishing characteristics, offering valuable insights for differential diagnosis and clinical decision-making.

Several limitations should be acknowledged. First, this was a single-center retrospective study with inherent selection bias and potential missing data. Second, despite adjusting for key confounders, other unmeasured factors (e.g., socioeconomic status, prior anti-mycobacterial treatment, antiretroviral adherence) may have influenced the results. Third, the Hosmer-Lemeshow test for the multivariate model showed a P value of 0.021, indicating modest lack of calibration, which may be due to the large sample size and inclusion of continuous variables without transformation. Fourth, although we applied Bonferroni correction for univariate comparisons, the exploratory nature means that some significant findings may still be due to chance. Fifth, the sample size (494 patients) from a single institution over five years limits generalizability. Future multicenter prospective studies with pre-specified outcomes are needed to validate our findings.

10. Conclusion

In conclusion, after adjusting for immune status and other confounders, male gender, higher CD4+ count, higher CRP, higher neutrophil count, presence of miliary nodules, and presence of superficial lymphadenopathy were independently associated with TB rather than NTM in AIDS patients. These features, although not pathognomonic, can assist clinicians in differentiating NTM from TB when rapid microbiological results are unavailable. The high prevalence of Mycobacterium avium (68.2%) in the AIDS/NTM group suggests that empirical therapy for suspected NTM in severely immunocompromised AIDS patients in Shanghai should cover the Mycobacterium avium complex (MAC). Because of the exploratory nature of this single-center study, prospective multicenter research with pre-specified outcomes is needed to validate these findings.

CRediT authorship contribution statement

Shuibao Xu: Writing – review & editing, Writing – original draft, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization. Junyang Yang: Methodology, Data curation. Jingna Xun: Writing – original draft, Methodology. Renfang Zhang: Supervision, Data curation, Conceptualization. Li Liu: Supervision, Data curation, Conceptualization. Jun Chen: Supervision, Data curation, Conceptualization. Zhenyan Wang: Methodology, Data curation. Tangkai Qi: Data curation. Jianjun Sun: Data curation. Wei Song: Data curation. Yang Tang: Data curation. Youming Chen: Data curation. Yueming Shao: Data curation. Xinyu Zhang: Data curation. Zichen Song: Data curation. Yinzhong Shen: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Hongzhou Lu: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.

Ethical statement

This retrospective study was approved by the Ethics Committee of Shanghai Public Health Clinical Center, Fudan University(Approval No. 2022-S033-01, dated March 30, 2022). Given the retrospective design using anonymized historical clinical data, the requirement for informed consent was waived by the ethics committee. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Ethical approval and consent to participate

This retrospective study was approved by the Ethics Committee of Shanghai Public Health Clinical Center, Fudan University. Given the retrospective design using anonymized historical clinical data, the requirement for informed consent was waived by the ethics committee. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Funding

This research was supported by grants (KY-GW-2022-15) from Shanghai Public Health Clinical Center Research Project; Shanghai Municipal Science and Technology Commission(20MC1920100and 21Y31900400),Shanghai Municipal Health Commission (shslczdzk01102) and Shen Kang Hospital Development Center(SHDC22021317，SHDC22024317).

Declaration of competing interest

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.