The Darwin Prospective Melioidosis Study: a 30-year prospective, observational investigation

Background: The global distribution of melioidosis is under considerable scrutiny, with both unmasking of endemic disease in African and Pacific nations and evidence of more recent dispersal in the Americas. Because of the high incidence of disease in tropical northern Australia, The Darwin Prospective Melioidosis Study commenced in October, 1989. We present epidemiology, clinical features, outcomes, and bacterial genomics from this 30-year study, highlighting changes in the past decade. 

Methods: The present study was a prospective analysis of epidemiological, clinical, and laboratory data for all culture-confirmed melioidosis cases from the tropical Northern Territory of Australia from Oct 1, 1989, until Sept 30, 2019. Cases were identified on the basis of culture-confirmed melioidosis, a laboratory-notifiable disease in the Northern Territory of Australia. Patients who were culture-positive were included in the study. Multivariable analysis determined predictors of clinical presentations and outcome. Incidence, survival, and cluster analyses were facilitated by population and rainfall data and genotyping of Burkholderia pseudomallei, including multilocus sequence typing and whole-genome sequencing. 

Findings: There were 1148 individuals with culture-confirmed melioidosis, of whom 133 (12%) died. Median age was 50 years (IQR 38–60), 48 (4%) study participants were children younger than 15 years of age, 721 (63%) were male individuals, and 600 (52%) Indigenous Australians. All but 186 (16%) had clinical risk factors, 513 (45%) had diabetes, and 455 (40%) hazardous alcohol use. Only three (2%) of 133 fatalities had no identified risk. Pneumonia was the most common presentation occurring in 595 (52%) patients. Bacteraemia occurred in 633 (56%) of 1135 patients, septic shock in 240 (21%) patients, and 180 (16%) patients required mechanical ventilation. Cases correlated with rainfall, with 80% of infections occurring during the wet season (November to April). Median annual incidence was 20·5 cases per 100 000 people; the highest annual incidence in Indigenous Australians was 103·6 per 100 000 in 2011–12. Over the 30 years, annual incidences increased, as did the proportion of patients with diabetes, although mortality decreased to 17 (6%) of 278 patients over the past 5 years. Genotyping of B pseudomallei confirmed case clusters linked to environmental sources and defined evolving and new sequence types.

Interpretation: Melioidosis is an opportunistic infection with a diverse spectrum of clinical presentations and severity. With early diagnosis, specific antimicrobial therapy, and state-of-the-art intensive care, mortality can be reduced to less than 10%. However, mortality remains much higher in the many endemic regions where health resources remain scarce. Genotyping of B pseudomallei informs evolving local and global epidemiology. 

Funding: The Australian National Health and Medical Research Council.