Clinical Outcomes and Mortality Predictors of Melioidosis: A Four-Year Cohort Study from Malaysia

1. Introduction

Melioidosis is a potentially fatal tropical infection caused by the gram-negative bacterium Burkholderia pseudomallei [1]. The disease is endemic to Southeast Asia and Northern Australia [1,2]. Transmission typically occurs through skin inoculation, inhalation or ingestion of contaminated soil or water, leading to a wide spectrum of clinical manifestations ranging from localised abscesses to septic shock and even death [1]. Despite advances in treatments and critical care, melioidosis continues to be associated with significant mortality [3,4,5] and is estimated to reach 165,000 cases per year, of which 89,000 die [6].

Malaysia is an upper-middle income Southeast Asian country with a population of approximately 34 million. It comprises Peninsular Malaysia and the states of Sabah and Sarawak on the island of Borneo. The country has a tropical equatorial climate characterised by high humidity, heavy seasonal rainfall and extensive agricultural activity including rice farming and plantation work. These environmental conditions are conducive for the survival of Burkholderia pseudomallei. [3,6,7]

Malaysia is recognised as an endemic country, with increasing numbers of culture confirmed cases reported annually [8]. However, melioidosis is not a notifiable disease nationally; hence, the true disease burden is likely underestimated [8]. Several Malaysian states have published data describing the epidemiology, clinical characteristics and mortality predictors of melioidosis, reporting consistently high case fatality rates [3,4,6,9,10,11].

Negeri Sembilan is a state located in central Peninsular Malaysia and comprises extensive agricultural land as well as rural areas that may predispose its population to environmental exposure. The state comprises seven districts, namely Seremban, Port Dickson, Jempol, Jelebu, Kuala Pilah, Rembau and Tampin. Although a recent study examined the epidemiological pattern of melioidosis in the state, and its association with meteorological factors [12], data on clinical outcomes and predictors of mortality remain limited.

This study aimed to describe the sociodemographic characteristics of patients, clinical features, mortality rate, and mortality predictors of culture-confirmed melioidosis cases admitted to Hospital Tuanku Ja’afar, Negeri Sembilan’s sole tertiary referral centre, in order to better inform local clinical management and risk stratification.

2. Materials and Methods

2.1. Study Design and Setting

This was a retrospective cohort study conducted at the Hospital Tuanku Ja’afar (HTJS), a tertiary teaching hospital located in the city of Seremban, Negeri Sembilan, Malaysia. The study reviewed all patients with culture confirmed Burkholderia pseudomallei infection that were diagnosed between 1st January 2021 and 31st December 2024.

2.2. Study Population and Case Definition

All patients aged 12 years and above who were admitted to HTJS with a culture confirmed diagnosis of melioidosis were included. Patients whose records and/or culture results that could not be traced were excluded from this study . Cases were defined as patients with culture-positive melioidosis isolated from any clinical specimen processed within the HTJS microbiology laboratory during the study period.

2.3. Data Sources and Collection

Eligible patients were identified through the hospital microbiology database. Clinical data comprising of patients’ demographics (age, sex, ethnicity, occupation, residential address), clinical characteristics (comorbidities, presenting symptoms, disease manifestations, complications), microbiological data (source of culture-positive specimens), treatment details (type of antibiotics administered, time to initiation and duration) and outcomes (recovery, length of hospital stay, mortality) were retrieved from electronic medical records and case notes. Data collection was performed using a standardized case report form adapted from melioidosis registries from the states of Sabah and Kedah.

2.4. Study Variables

Independent variables include clinical findings such as symptoms and organ involvement, and laboratory findings such as deranged renal function. The dependent variable is mortality, defined as a patient who has died from this episode of melioidosis infection. Associations between selected risk factors and mortality were also explored in the analysis.

2.5. Data Analyses

Continuous variables were categorised where clinically appropriate. As this study utilised secondary retrospective data, missing data were addressed using multiple imputation via fully conditional specification (multiple imputation by chained equations), with regression models specified according to variable type, generating five imputed datasets.

Descriptive analysis was conducted to describe the sociodemographic characteristics, comorbidities, clinical features, and outcomes of patients with melioidosis admitted to Hospital Tuanku Ja’afar. Case fatality rate was calculated based on in-hospital mortality.

Univariate and multivariate analyses were performed using logistic regression to identify independent predictors of mortality. Variables with p <0.05 in univariate analysis were entered into the multivariate logistic regression model. Clinically related variables were consolidated where appropriate to reduce collinearity and improve model stability. To avoid omission of known confounders, variables that were not found to be significant in univariate analysis (p ≥ 0.05) but were deemed to be clinically significant, such as history of smoking, were included in the multivariate analysis. Specifically, deranged systolic blood pressure and deranged heart rate were considered under the composite variable 'septic shock'; liver abscess and spleen abscess were considered components of 'multiorgan failure' and were not analyzed separately. Fever, generalized weakness, shortness of breath and abscess formation, although statistically significant on univariate analysis, were considered nonspecific clinical presentations with limited independent contribution to mortality risk beyond that captured by the physiological and organ-involvement variables, and were therefore not included into the multivariate model. Deranged white cell count, while not independently significant (p = 0.120), was retained due to its clinical relevance to sepsis recognition. Statistical significance was defined as p<0.05. Results were reported as odds ratios (OR) with 95% confidence intervals (CI).

Data collected was analyzed using SPSS version 25 (IBM Corp., Armonk, NY, USA).

3. Results

3.1. Patient Characteristics

3.1.1. Patient Sociodemography

A total of 107 patients with culture-confirmed melioidosis were included in the study. The median age was 51.9 years (range 21–81 years). The majority of patients were male (83.2%).

By ethnicity, most patients were Malay (62.6%), followed by Indian (20.6%), Chinese (9.3%), and other ethnic groups (7.5%) as detailed in Table 1.

Geographically, most patients were residents of Seremban district (81.3%), where Hospital Tuanku Ja’afar is located. Other districts represented included Rembau (6.5%), Jelebu (5.6%), Tampin (2.8%), Kuala Pilah (1.9%), Port Dickson (0.9%) and Jempol (0.9%).

3.1.2. Prevalence of Risk Factors

Diabetes mellitus was the most common comorbidity, present in 78.5% of patients. Other comorbidities included chronic kidney disease (12.1%), chronic lung disease (7.5%), immunosuppression (6.5%), malignancy (6.5%) and congestive cardiac failure (5.6%). Common risk factors present in the patient population were smoking (27.1%) and alcohol use (7.5%)

Environmental exposure data were incompletely documented. 3.7% of patients had occupational exposure from farming and plantations. Other non-farming/plantation occupational exposure to soil or water was recorded in 10.3% of patients. Recreational exposure to soil like gardening was reported in 3.7% of patients, while 0.9% reported recreational water exposure. Detailed results of the risk factors of mortality are shown in Table 2.

3.2. Clinical Data and Outcomes

Clinical Outcomes

Of the 107 patients, 71 (66.4%) survived and 36 (33.6%) died. Patients who died were predominantly male (83.3%) and of Malay ethnicity (77.8%), with a mean age of 56.3 ± 13.5 years, compared to 49.7 ± 14.4 years among survivors. Diabetes mellitus was the most prevalent comorbidity in this group (69.4%), followed by chronic kidney disease (16.7%) and malignancy (13.9%). Clinically, patients who died presented with significantly higher rates of septic shock (83.3% vs 22.5% in survivors), generalized weakness (47.2%), and shortness of breath (47.2%); while localised abscess formation was notably less common among those who died (2.8% vs 23.9% in survivors), consistent with a more disseminated and severe disease phenotype. Multiorgan failure occurred in 69.4% of non-survivors and bacteremia was documented in 88.9%. With regard to laboratory parameters, there were high rates of renal impairment (94.1%) and deranged liver function (87.5%).

Among the 36 deaths, 10 patients (27.8%) died before melioidosis-specific antimicrobial therapy could be initiated. Although some of these patients had received broad-spectrum antibiotics, these did not include ceftazidime or meropenem, which are recommended first-line antimicrobials for melioidosis [13].

On univariate analysis, several demographic, clinical and laboratory variables were significantly associated with mortality. These included increasing age (p = 0.027) and malignancy (p = 0.047). Presenting symptoms significantly associated with mortality included fever (p = 0.030), shortness of breath (p = 0.010), abscess formation (p = 0.023) and generalized weakness (p = 0.004). Vital sign abnormalities associated with mortality included hypotension (p = 0.006) and reduced peripheral oxygen saturation requiring oxygen supplementation (p = 0.019).

Several laboratory abnormalities were also significantly associated with mortality, including deranged lymphocytes (p = 0.013), deranged platelets (p = 0.024), deranged renal function (p = 0.014), and deranged liver function (p = 0.009).

Organ involvement significantly associated with mortality included pneumonia (p = 0.010), liver abscess (p = 0.015), splenic abscess (p = 0.013), septic shock (p < 0.001) and multiorgan failure (p < 0.001). Detailed results of univariate analysis are shown in the Supplementary Table S1, and significant variables are shown in Table 3.

On multivariate logistic regression analysis, septic shock remained the only independent predictor of mortality (adjusted OR 10.95, 95% CI 2.83–42.37, p = 0.001). Deranged renal function approached statistical significance (OR 7.60, 95% CI 1.00–57.75, p = 0.050). Other variables, including age, malignancy, smoking status, oxygen supplementation, deranged liver function, pneumonia and multiorgan failure, were not statistically significant after adjustment for confounders. Detailed results of the regression analysis are shown in Table 4.

4. Discussion

This study provides one of the first contemporary clinical descriptions of culture-confirmed melioidosis in the state of Negeri Sembilan, Malaysia and demonstrates a disease profile consistent with other Malaysian states. The condition predominantly affected middle-aged men with diabetes, and was associated with severe systemic complications and high mortality.

Diabetes mellitus was the most common comorbidity (77.2%), reinforcing its established role as the principal host risk factor for melioidosis in endemic regions [3,4,5,7,9,10,14]. Similar prevalence has been reported in Thailand and Australia [5,14]. It was suggested that elevated glucose levels delays identification and recognition of B. pseudomallei surface structures, thus impairing the host innate immune system [15]. In diabetes, impaired neutrophil and macrophage function compromises bacterial clearance, further predisposing to invasive disease [1]. Given the rising prevalence of diabetes in Malaysia and around the world, this comorbidity likely contributes to the rising prevalence and mortality of melioidosis [6,16]

Marked male predominance in our cohort parallels regional data, where men account for up to 60–86% of cases [3,4,7,9,10,14,17]. This trend is commonly attributed to greater occupational and environmental exposure to contaminated soil and surface water, in particular agricultural work [8]. However, this association was not elicited in our study mainly due to incomplete documentation in the clinical notes.

Clinically, pneumonia and bacteraemia were the most common presentations. High rates of septic shock (43.9%) and multiorgan failure (30.7%) suggest advanced disease at presentation. The overall case fatality rate of 31.6% aligns with reports from other Malaysian states like Kelantan (32.9%), Kedah (33.8%), Sarawak (35%) and Pahang (54%) [3,4,7,9,10,11]. Similar mortality rates have been reported in neighbouring Thailand, where case fatality typically ranges from 25–39% [5,18,19], reflecting the significant disease burden in endemic rural regions. In contrast, Singapore has reported lower mortality of approximately 20.9% [17], likely reflecting earlier diagnosis, improved healthcare access, and more consistent critical care support.

Nevertheless, mortality in Malaysia and the wider region remains substantially higher than in Australia, where rates have fallen to approximately 12% due to early recognition, standardized management protocols, and well-established intensive care infrastructure [14]. These regional differences suggest that outcomes in melioidosis are strongly influenced not only by disease severity but also by healthcare system factors such as diagnostic capacity, clinician awareness, and timely initiation of appropriate antimicrobial therapy.

Notably, nearly one-third of deaths occurred before initiation of treatment, underscoring delays in recognition and the rapid progression of severe melioidosis. This finding suggests potential delays in clinical recognition, rapid disease progression, or late presentation to tertiary care. Melioidosis frequently mimics other causes of community-acquired sepsis and pneumonia, and microbiological confirmation may be delayed, thereby postponing melioidosis-specific antimicrobial treatment with ceftazidime, meropenem, or trimethoprim-sulfamethoxazole [13]. These findings highlight the importance of maintaining a high index of suspicion, particularly in diabetic patients presenting with severe community-acquired sepsis.

Septic shock emerged as the sole independent predictor of mortality, increasing the odds of death tenfold. This finding is consistent with other studies identifying septic shock as a significant predictor of mortality [3,4,5]. Variations in additional predictors across studies may reflect differences in sample size, comorbidity burden, and timing of presentation. These findings emphasise that early identification of sepsis and prompt initiation of appropriate antimicrobial therapy remain critical to improving survival. In endemic regions, diabetic patients presenting with severe sepsis or pneumonia may warrant early empiric coverage for melioidosis while awaiting culture confirmation.

Although melioidosis incidence in Malaysia has been associated with rainfall and agricultural exposure [3,7,12], environmental factors could not be reliably analysed in this study due to incomplete documentation. Given the predominantly agricultural landscape surrounding parts of Negeri Sembilan, environmental exposure likely remains an important but under-characterised contributor of transmission.

The continued high mortality observed in Malaysian settings also highlights broader health system considerations. As Hospital Tuanku Ja’afar is the sole tertiary referral centre in the state, more severe cases may be overrepresented, potentially inflating mortality estimates. Strengthening diagnostic capacity in peripheral hospitals, improving early referral pathways, and enhancing clinician awareness may reduce progression to septic shock before tertiary admission. Furthermore, melioidosis is not currently a notifiable disease in Malaysia, limiting national surveillance and coordinated public health response. Consideration of mandatory notification in endemic regions may improve disease burden estimation, resource allocation, and targeted preventive strategies.

Preventive efforts should focus on high-risk populations, particularly individuals with poorly controlled diabetes engaged in agricultural or soil-related activities. Public health education and improved diabetes management may reduce severe outcomes. Given projected increases in diabetes prevalence and climate change, the burden of melioidosis in endemic regions may persist or increase without proactive intervention [16,20].

Limitations and Future Recommendations

Our study was limited by its retrospective design and reliance on hospital records, which have led to incomplete data capture. Long-term sequelae of melioidosis were not assessed in this study due to the retrospective design and one-time nature of hospital data, which limited the ability to evaluate patient outcomes beyond the acute hospitalisation period. We were also unable to analyse potential environmental links that were highlighted in papers from other states [3,4,9,12] due to missing data.

The relatively small sample size and single-center nature may limit the generalisability of findings. With a larger study population, parameters such as deranged renal function (p value = 0.05) may carry higher significance. Hospital Tuanku Ja’afar being a tertiary centre where severe cases from peripheral hospitals are referred to may contribute to selection bias.

Nevertheless, it provides valuable baseline data for Negeri Sembilan and highlights key areas for intervention. Future multicenter studies with prospective data collection could better define local risk factors and evaluate interventions to reduce mortality.

5. Conclusions

In conclusion, melioidosis in Negeri Sembilan mirrors patterns seen in other endemic states in Malaysia, with middle-aged diabetic men disproportionately affected and septic shock representing the ultimate predictor of death. Early recognition, timely initiation of empiric antibiotic therapy and robust sepsis management are critical to reduce mortality in this highly fatal tropical infection.