Objective:

To compare clinical characteristics, imaging, microbiological diagnostic findings, diagnostic delays, and respiratory co-detection patterns in non-HIV immunocompromised and non-immunocompromised patients with pulmonary tuberculosis.

Approach:

• Design and setting: Retrospective study of inpatients with liquid culture-confirmed pulmonary tuberculosis at a tertiary hospital in China from January 2020 through June 2025.
• Groups and measures: Patients meeting specified non-HIV immunocompromise criteria are compared with those without these criteria. Health system delay is measured from first medical contact to the earliest positive respiratory microbiological result; more than 14 days defines long delay.
• Analyses: Clinical, CT, microbiological, and co-detection findings are compared. Multivariable models assess associations between immunocompromised status, time to diagnosis, and long diagnostic delay.

Key Findings:

• Among 578 patients, 71 are immunocompromised and 507 are non-immunocompromised.
• Median health system delay is longer in immunocompromised patients: 90 versus 24 days (P < .001). Long delay occurs in 83.1% versus 69.2% (P = .012).
• Immunocompromised status is independently associated with long delay (adjusted odds ratio, 2.31; 95% CI, 1.03–5.18).
• Immunocompromised patients have more bilateral lesions, pleural effusion, and lymphadenopathy but fewer cavities.
• Sputum acid-fast bacilli smear positivity is numerically lower in immunocompromised patients, without a statistically significant difference; only 29 undergo this test. Bronchoalveolar lavage fluid smear positivity is higher.
• Xpert MTB/RIF positivity exceeds 75% across sputum, bronchoalveolar lavage fluid, and tissue specimens, without significant between-group differences.
• Fungal, viral, and mixed co-detections are more frequent in immunocompromised patients, although the proportion with any co-detection does not differ significantly.

Interpretation:

Non-HIV immunocompromised patients have longer diagnostic delays, less typical radiological presentations, and distinct respiratory co-detection patterns. Specimen selection and molecular testing may facilitate diagnosis; sputum smear findings require caution because testing is limited in this group.

Limitations:

• The retrospective, single-center inpatient design limits generalizability and precludes definitive causal inference.
• Unequal immunocompromised subgroup sizes prevent detailed comparisons by cause of immunosuppression.
• Multiple baseline comparisons lack adjustment for multiple testing.
• Diagnostic selection rationale, medication exposures, and postdischarge tuberculosis treatment initiation are not systematically captured.

Conclusion:

The findings support maintaining suspicion for pulmonary tuberculosis in non-HIV immunocompromised patients and considering appropriate specimen selection and molecular testing to facilitate diagnosis.

Sources:

• Phenotypic atypia, prolonged diagnostic delay and altered co-detection patterns in non-HIV immunocompromised pulmonary tuberculosis