Cox Proportional Hazards Analysis to Determine the Predictors of Time-to-Sputum Culture Conversion in Patients with Tuberculosis Open Access

Soni, Tanushree (2017)

Permanent URL:


Background: Tuberculosis continues to be one the deadliest infections in the world, infecting one fourth of the global population and claiming 5000 lives a day. Effective diagnosis and treatment regimens are imperative to reduce the burden of this disease. Sputum culture is used for accurate diagnosis, and conversion of serial sputum cultures from positive to negative is an important indicator of the efficacy of anti-tuberculosis treatment. In this study we determine important predictors for sputum culture conversion in a cohort of patients admitted in a US hospital for the treatment of TB.

Methods and Findings: We abstracted medical records of patients admitted to a TB ward, 1985-2010. We compared patients who did and did not convert with respect to important social and medical characteristics using frequency tables, Kaplan Meier survival curves as well as Cox proportional hazards regression analysis to determine the time to conversion and the predictors of time to conversion. In our studies, we found that 61% of the cohort showed sputum culture conversion; the median time to conversion was 50 days and the maximum time was 56 days. Alcohol consumption, HIV infection IV drug use, race, baseline smear status, treatment with first-line drugs and the calendar period were significantly associated with time to conversion of sputum cultures. Patients with a HIV positive status and use of intravenous drugs had a shorter time to conversion. Treatment with certain drugs also significantly decreased the time to conversion. Cox proportional hazards regression analysis identified gender, baseline smear results, HIV, IV drug use, and race as independent predictors of conversion.

Conclusion: In conclusion, multiple variables can affect sputum culture conversion and should be taken into account for the management of TB patients.

Table of Contents



CHAPTER 1: Introduction. 9

Background. 9

Statement of the Problem.. 10

Purpose of the Study. 11

Significance Statement 11

Acronyms and definitions 11

CHAPTER 2: Literature Survey. 15

Introduction. 15

TB transmission. 15

TB signs and symptoms 16

Pulmonary TB.. 17

Diagnosis of tuberculosis. 18

Current immunodiagnostic test for TB.. 18

Current laboratory based test for TB.. 19

Radiographic Screening. 21

Treatment 23

Adherence to TB treatment 24

Epidemiology. 25

Significance of bacteriological tests in PTB.. 25

Sputum Smear Analysis. 26

Culture analysis. 27

HIV.. 27

Diagnosis of tuberculosis for patients with HIV coinfection. 28

Treatment of tuberculosis for patients with HIV coinfection. 29

Body Mass Index (BMI) 31

Tobacco/Smoking. 32

Alcohol 32

CHAPTER 3: Manuscript 34

Introduction. 34

Methods. 35

Laboratory methods. 35

Definitions. 35

Predictive variables. 36

Outcome variables 37

Statistical analyses 37

Ethical considerations. 38

Results. 38

Bivariate analysis of predictive variables with sputum conversion. 38

Variables associated with time to conversion (TOC) 39

Multivariable Analysis. 40

Adjusted survival curves. 41

Discussion. 41

Conclusion. 43

CHAPTER 4: Conclusion and Public Health Implications. 44


Figures and Tables. 52


SAS Code. 62

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
  • English
Research Field
Committee Chair / Thesis Advisor
Committee Members
Partnering Agencies
Last modified

Primary PDF

Supplemental Files