ORIGINAL RESEARCH
Infectious Disease & Tuberculosis

Factors Associated With Unsuccessful Treatment Outcome of Pulmonary Tuberculosis in Kota Bharu, Kelantan

Background and Context

Tuberculosis (TB) remains one of the most significant communicable disease challenges facing Malaysia and the broader Southeast Asian region. Despite substantial investments in TB control infrastructure, directly observed therapy (DOTS) programmes, and contact tracing systems, Malaysia has struggled to achieve the World Health Organisation’s target of a treatment success rate exceeding 90% among notified TB cases. National surveillance data from the mid-2000s revealed a troubling pattern: the overall treatment success rate for pulmonary TB had declined from 78% in 2000 to approximately 48% in 2006, raising serious questions about the adequacy of existing control strategies.

Kelantan, one of Malaysia’s northeastern states, has historically reported a substantial TB burden, attributed in part to its geographical position near the Thai border, population mobility patterns, and the prevalence of co-morbid conditions — particularly diabetes mellitus and HIV infection — that compromise immune function and TB treatment response. Kota Bharu, the state capital, serves as the primary hub for TB diagnosis and treatment in the region, making it an important site for investigating the determinants of treatment outcomes.

Understanding the factors that contribute to unsuccessful treatment outcomes — defined in this study as treatment failure or death during TB treatment — is essential for designing targeted interventions to improve case management and reduce TB-related morbidity and mortality.

Study Design and Methods

This retrospective cohort study examined all notified and treated pulmonary TB cases in Kota Bharu district between January 2006 and December 2007. Case data were extracted from the National Tuberculosis Information System (TBIS), a web-based surveillance application maintained by the Ministry of Health Malaysia. Of 765 patients initially identified, 472 cases met the inclusion criteria and were analysed.

The study population comprised 316 males and 156 females with a mean age of 45 years (standard deviation 17.9 years). Univariate comparisons and multiple logistic regression analyses were performed to identify factors independently associated with unsuccessful treatment outcomes. Variables examined included sociodemographic characteristics (age, sex, educational level, employment status, family income), clinical factors (co-existence of extrapulmonary TB, smoking status, co-morbid diabetes mellitus, HIV status, sputum culture results, chest X-ray findings), and health system factors (duration of diagnostic delay).

Key Results

The multiple logistic regression model demonstrated good fit (Hosmer-Lemeshow test p-value = 0.37; area under the Receiver Operating Characteristics curve = 0.78; overall percentage correct = 92.9%). No multicollinearity problems were detected among the independent variables. Among all factors examined, HIV-positive status emerged as the most powerful independent predictor of unsuccessful treatment outcome, a finding consistent with the well-established immunological relationship between HIV and TB.

The TB-HIV Co-Infection Challenge

The strong association between HIV positivity and treatment failure in this study reflects a broader national and global concern about TB-HIV co-infection. HIV compromises cell-mediated immunity, which is the primary host defence against Mycobacterium tuberculosis. Co-infected patients experience higher rates of treatment failure, relapse, and mortality compared to HIV-negative TB patients, even when receiving appropriate anti-TB therapy.

In Malaysia, TB-HIV collaborative activities have been strengthened since the early 2000s, including routine HIV testing for all notified TB patients, early initiation of antiretroviral therapy (ART) for co-infected individuals, and cotrimoxazole preventive therapy. However, the findings from this study suggest that at the time of the research, these measures had not yet achieved optimal outcomes in the Kota Bharu setting, possibly due to late HIV diagnosis, delays in ART initiation, or challenges in coordinating care between TB and HIV treatment programmes.

The Role of Diabetes as a Comorbidity

The association between diabetes mellitus and poor TB treatment outcomes observed in this study is particularly relevant to the Malaysian context, where diabetes prevalence ranks among the highest in the Asia-Pacific region. Diabetes impairs both innate and adaptive immune responses, increases susceptibility to TB infection, and may reduce the effectiveness of anti-TB medications through altered drug metabolism and pharmacokinetics.

The co-existence of TB and diabetes in Malaysia represents a significant public health intersection that requires integrated management strategies. Screening all TB patients for diabetes (and vice versa), optimising glycaemic control during TB treatment, and ensuring close monitoring of treatment response in diabetic TB patients are recommended approaches to improving outcomes in this population.

Implications for TB Control in Malaysia

The study’s findings carry direct implications for TB control policy in Malaysia. First, they underscore the importance of rapid TB diagnosis to minimise diagnostic delay, which was independently associated with poor outcomes. Second, they highlight the need for strengthened TB-HIV collaborative activities, including earlier HIV testing and ART initiation. Third, they suggest that TB treatment programmes should incorporate routine diabetes screening and management as a core component of care. Finally, the socioeconomic risk factors identified (low education, unemployment, low income) point to the need for patient support interventions — including transportation assistance, nutritional support, and treatment adherence counselling — to address the social determinants of treatment success.

Limitations

This study relied on retrospective data from a single district, which may limit generalisability to other regions of Malaysia with different demographic, epidemiological, and health system characteristics. The use of registry-based data meant that some variables of interest (medication adherence patterns, specific HIV treatment regimens, glycaemic control levels) were not available for analysis. Additionally, the two-year study period (2006–2007) may not reflect current treatment outcomes, given subsequent improvements in TB and HIV care protocols.

← Back to MJPHM Home