Javascript must be enabled to continue!
Independent Risk Factors for In-Hospital Mortality in Non-HIV Adult Tuberculous Meningitis: A Large Retrospective Cohort Study from China
View through CrossRef
Abstract
Background:
Tuberculous meningitis (TBM) represents the most severe form of tuberculosis and is associated with high short-term mortality, particularly in low- and middle-income countries. However, robust evidence identifying predictors of in-hospital mortality among non-HIV adult TBM patients remains limited.
Objective:
To identify independent clinical and laboratory predictors of in-hospital mortality among adults with TBM.
Methods:
We conducted a retrospective cohort study including adult patients with suspected TBM admitted to Changsha Central Hospital, Hunan Province, China, between October 1, 2013, and October 1, 2023. TBM cases were classified as definite, probable, or possible using standardized diagnostic criteria. In-hospital mortality was the primary outcome. Candidate predictors were initially screened using least absolute shrinkage and selection operator (LASSO) regression, followed by multivariable analysis with Firth-penalized logistic regression to address potential separation bias. Multicollinearity was assessed using variance inflation factors (VIF). Model discrimination and calibration were evaluated using the area under the receiver operating characteristic curve (AUC), bootstrap calibration, and the Hosmer–Lemeshow goodness-of-fit test.
Results:
A total of 1,346 adult patients were included (32.2% definite, 62.9% probable, and 4.8% possible TBM). The median age was 47 years (interquartile range [IQR], 30–62), and 68.0% were female. Overall in-hospital mortality was 12.26% (165/1,346). Multivariable analysis identified drug-induced hepatitis, hydrocephalus, renal failure, immunodeficiency, positive cerebrospinal fluid (CSF) tuberculosis antibody, elevated procalcitonin (PCT) levels, and advanced age as independent predictors of increased in-hospital mortality. Higher CSF glucose levels showed a protective trend. The final model included 16 predictors, demonstrated minimal multicollinearity (maximum VIF 4.0; mean VIF 1.28), and showed excellent performance (AUC 0.917; bootstrap mean absolute calibration error 0.014; Hosmer–Lemeshow p = 0.098).
Conclusion:
Among non-HIV adult patients with TBM, drug-induced hepatitis, hydrocephalus, renal failure, immunodeficiency, positive CSF tuberculosis antibody, elevated PCT levels, and advanced age are key independent predictors of in-hospital mortality. These findings may assist clinicians in identifying patients at higher risk of in-hospital mortality in high-burden settings.
Springer Science and Business Media LLC
Title: Independent Risk Factors for In-Hospital Mortality in Non-HIV Adult Tuberculous Meningitis: A Large Retrospective Cohort Study from China
Description:
Abstract
Background:
Tuberculous meningitis (TBM) represents the most severe form of tuberculosis and is associated with high short-term mortality, particularly in low- and middle-income countries.
However, robust evidence identifying predictors of in-hospital mortality among non-HIV adult TBM patients remains limited.
Objective:
To identify independent clinical and laboratory predictors of in-hospital mortality among adults with TBM.
Methods:
We conducted a retrospective cohort study including adult patients with suspected TBM admitted to Changsha Central Hospital, Hunan Province, China, between October 1, 2013, and October 1, 2023.
TBM cases were classified as definite, probable, or possible using standardized diagnostic criteria.
In-hospital mortality was the primary outcome.
Candidate predictors were initially screened using least absolute shrinkage and selection operator (LASSO) regression, followed by multivariable analysis with Firth-penalized logistic regression to address potential separation bias.
Multicollinearity was assessed using variance inflation factors (VIF).
Model discrimination and calibration were evaluated using the area under the receiver operating characteristic curve (AUC), bootstrap calibration, and the Hosmer–Lemeshow goodness-of-fit test.
Results:
A total of 1,346 adult patients were included (32.
2% definite, 62.
9% probable, and 4.
8% possible TBM).
The median age was 47 years (interquartile range [IQR], 30–62), and 68.
0% were female.
Overall in-hospital mortality was 12.
26% (165/1,346).
Multivariable analysis identified drug-induced hepatitis, hydrocephalus, renal failure, immunodeficiency, positive cerebrospinal fluid (CSF) tuberculosis antibody, elevated procalcitonin (PCT) levels, and advanced age as independent predictors of increased in-hospital mortality.
Higher CSF glucose levels showed a protective trend.
The final model included 16 predictors, demonstrated minimal multicollinearity (maximum VIF 4.
0; mean VIF 1.
28), and showed excellent performance (AUC 0.
917; bootstrap mean absolute calibration error 0.
014; Hosmer–Lemeshow p = 0.
098).
Conclusion:
Among non-HIV adult patients with TBM, drug-induced hepatitis, hydrocephalus, renal failure, immunodeficiency, positive CSF tuberculosis antibody, elevated PCT levels, and advanced age are key independent predictors of in-hospital mortality.
These findings may assist clinicians in identifying patients at higher risk of in-hospital mortality in high-burden settings.
Related Results
Capítulo 6 – HIV-AIDS, como tratar, o que fazer e o que não fazer durante o tratamento?
Capítulo 6 – HIV-AIDS, como tratar, o que fazer e o que não fazer durante o tratamento?
A infecção pelo vírus do HIV pode ocorrer de diversas maneiras, tendo sua principal forma a via sexual por meio do sexo desprotegido. O vírus do HIV fica em um período de incubação...
Aetiological Evaluation of Patients with Meningitis in a Tertiary Care Hospital of Bangladesh
Aetiological Evaluation of Patients with Meningitis in a Tertiary Care Hospital of Bangladesh
Meningitis is a global health problem as mortality is high and a large proportion of survivors suffered from significant morbidity. The physicians are facing this challenge of emer...
Evolution of Antimicrobial Resistance in Community vs. Hospital-Acquired Infections
Evolution of Antimicrobial Resistance in Community vs. Hospital-Acquired Infections
Abstract
Introduction
Hospitals are high-risk environments for infections. Despite the global recognition of these pathogens, few studies compare microorganisms from community-acqu...
Laboratory-based Evaluation of Wondfo HIV1/2 Rapid Test Kits in the Gambia, December 2020
Laboratory-based Evaluation of Wondfo HIV1/2 Rapid Test Kits in the Gambia, December 2020
Background: HIV rapid diagnosis in The Gambia is mainly done using Determine HIV-1/2 and First Response HIV 1.2.0 or SD Bioline HIV-1/2 3.0 for screening and sero-typing of HIV res...
Impact of HIV/AIDS scale-up on non-HIV priority services in Nyanza Province, Kenya
Impact of HIV/AIDS scale-up on non-HIV priority services in Nyanza Province, Kenya
Background: The HIV pandemic has attracted unprecedented scale-up in resources to curb its escalation and manage those afflicted. Although evidence from developing countries sugges...
Risk factors for intensive care unit admission and mortality among adult meningitis patients
Risk factors for intensive care unit admission and mortality among adult meningitis patients
Objectives: To assess the risk factors for intensive care unit admission and inpatient all-cause mortality among adult meningitis patients.
Method: The retrospective study was cond...
Bacterial Meningitis
Bacterial Meningitis
AbstractBacterial meningitis is bacterial infection of the cerebrospinal fluid within the subarachnoid space. It can be broadly divided into acute or chronic meningitis and communi...
Re-examining the meningitis belt: associations between environmental factors and epidemic meningitis risk across Africa
Re-examining the meningitis belt: associations between environmental factors and epidemic meningitis risk across Africa
AbstractIn 2002 Molesworth et al created a logistic regression model to investigate the spatial distribution of meningitis epidemics in Africa and determine geographical risk. We a...

