Epidemiological and Clinical Characteristics of Melioidosis Caused by Gentamicin-Susceptible Burkholderia pseudomallei in Sarawak, Malaysia.

BACKGROUND: Burkholderia pseudomallei, the causative agent of melioidosis, is intrinsically resistant to a broad range of antibiotics, including aminoglycosides. In Sarawak, Malaysia, a high proportion of melioidosis cases are caused by gentamicin-susceptible isolates. There are limited epidemiological and clinical data on these infections.
METHODS: We conducted a retrospective study of culture-confirmed melioidosis among adults admitted to Bintulu Hospital in Sarawak, Malaysia, from January 2011 until December 2016.
RESULTS: One hundred forty-eight adults with culture-confirmed melioidosis were identified. Of 129 (87%) tested, 84 (65%) had gentamicin-susceptible B pseudomallei. The average annual incidence of melioidosis was 12.3 per 100 000 population, with marked variation between districts ranging from 5.8 to 29.3 per 100 000 population. Rural districts had higher incidences of melioidosis and overwhelmingly larger proportions of gentamicin-susceptible B pseudomallei infection. Significantly more patients with gentamicin-susceptible infection had no identified risk factors, with diabetes less frequently present in this group. Ninety-eight percent had acute presentations. Pneumonia, reported in 71%, was the most common presentation. Splenic abscesses were found in 54% of those imaged. Bacteremia was present in 88%; septic shock occurred in 47%. Forty-five (35%) patients died. No differences in clinical, laboratory, or outcome characteristics were noted between gentamicin-susceptible and gentamicin-resistant infections.
CONCLUSIONS: Gentamicin-susceptible B pseudomallei infections are common in Sarawak and dominate in the high-incidence rural interior regions. Clinical manifestations and outcomes are the same as for gentamicin-resistant B pseudomallei infections. Further studies are required to determine if all gentamicin-susceptible B pseudomallei infections in Sarawak are clonal and to ascertain their environmental drivers and niches.

Melioidosis is an important cause of infectious disease mortality in the Western Pacific region and in Southeast Asia (SEA) [1, 2]. Although historically known to be endemic to SEA and northern Australia, cases are now increasingly being reported worldwide [3, 4]. Melioidosis is caused by the environmental saprophyte Burkholderia pseudomallei (Bp), a gram-negative bacillus that is intrinsically resistant to a broad range of antibiotics, including aminoglycosides [5].

In 2014, we reported the presence of rare gentamicin-susceptible Bp among a large proportion of clinical isolates from central Sarawak, Malaysian Borneo [6]. Interestingly, all gentamicin-susceptible isolates examined were genetically typed as either multilocus sequence type (ST) 881 or its single-locus variant ST997. The mechanism of the susceptibility to gentamicin was ascertained to be a novel mutation within the gene amrB, which encodes an essential component of the AmrAB-OprA multidrug efflux pump, with the mutation resulting in prevention of the efflux of gentamicin and macrolides that occurs in wild-type Bp.

Limited data are available on the disease burden and the epidemiological and clinical characteristics of melioidosis among adults in Sarawak. Importantly, the epidemiological and clinical characteristics of infection caused by gentamicin-susceptible isolates, and whether differences exist compared to infection with gentamicin-resistant Bp, remain unknown. To answer these questions, we conducted a retrospective study of adults with melioidosis admitted to Bintulu Hospital in Sarawak, Malaysia.

METHODS

Study Site and Population

Bintulu Hospital is a 302-bed hospital that provides medical, surgical and intensive care services to adults and children living in Bintulu Division (consists of Bintulu and Tatau districts, and the subdistrict of Sebauh) and Belaga district (consists of Belaga and Sungai Asap subdistricts) of Kapit Division in Sarawak, Malaysian Borneo. This 32 000-km2 area located within the central region of Sarawak has a total population of approximately 256 000. Although a large proportion of this population resides in Bintulu town (the largest town in the region), rural communities living in traditional longhouses and workers (mainly migrant) living in plantations are present in all districts. This rural population relies mainly on untreated water (eg, from rivers or gravity-fed water supplies from streams) for daily consumption.

We retrospectively identified all culture-confirmed melioidosis cases among adults aged ≥15 years admitted from January 2011 to December 2016 through a manual search of the microbiology laboratory logbooks and electronic database. Medical records of identified cases were retrieved and bacterial, epidemiological, and clinical details were collected using standardized case report forms. In addition, admission chest radiographs of patients with melioidosis pneumonia were retrieved and reviewed.

Case Definitions

Acute melioidosis was diagnosed if symptoms were present for <2 months, while chronic melioidosis was diagnosed if symptoms were present for ≥2 months before admission. Diabetes mellitus was defined by the presence of fasting plasma glucose ≥7.0 mmol/L and/or random plasma glucose ≥11.1 mmol/L, or glycated hemoglobin ≥6.5% [7]. Chronic kidney disease and acute kidney injury were defined according to the Kidney Disease: Improving Global Outcomes (KDIGO) definitions [8, 9]. In this study, melioidosis risk factors assessed included diabetes mellitus, chronic kidney disease, congestive cardiac failure, bronchiectasis, chronic obstructive pulmonary disease, malignancy, immunosuppressive therapy, and acquired or congenital immunodeficiency diseases. Septic shock was defined as the presence of hypotension not responsive to fluid replacement with end organ dysfunction. A melioidosis-active antibiotic was defined as either ceftazidime or a carbapenem.

Microbiological Methods

As part of routine clinical practice in Bintulu Hospital, blood samples collected from patients were subjected to the BACTEC blood culture system according to the manufacturer’s instructions (Becton Dickinson). Positive growth was subcultured onto blood agar, chocolate agar, and MacConkey agar. Samples from other sources (pus, sputum, endotracheal secretions, pleural fluid) were cultured directly on blood agar, chocolate agar, and MacConkey agar. Bp was identified with either API 20NE (bioMérieux) or BBL Crystal Identification Systems (Becton Dickinson). Antibiotic susceptibility of Bp isolates was determined by either Etest (bioMérieux) or Kirby-Bauer disk diffusion test (Becton Dickinson), adapting interpretative criteria from Burkholderia cepacia and Pseudomonas aeruginosa according to Clinical and Laboratory Standard Institute (CLSI) guidelines.

Molecular Characterization of Bacterial Isolates

The presence of nucleotide sequences encoding Burkholderia intracellular motility factor A (BimABm or BimABp) and the Burkholderia thailandensis–like flagellum and chemotaxis and the Yersinia-like fimbrial (YLF) gene clusters were determined using previously published methods [10]. A previously published allele-specific polymerase chain reaction (AS-PCR) assay was used to interrogate Bp isolates for the presence of amrB mutants that confer susceptibility to aminoglycosides and macrolides [6].

Statistical Analysis

Statistical analysis was performed using SPSS Statistics 23. The Mann-Whitney U test was used for numerical variables, and either the χ² test or Fisher exact test was used for categorical variables. Correlations between number of culture-confirmed melioidosis cases and average monthly rainfall was examined using Pearson correlation coefficient test. Population data were obtained from the Malaysian Census Data 2010. Meteorological data were obtained from the Malaysian Meteorological Department.

Ethics Statement

The study was approved by the Malaysian Medical Research Ethics Committee (NMRR-16-1029-31390). All data analyzed were anonymized.

Patient Consent Statement

The study involved a retrospective review of medical records and no intervention in patient care. The study was approved by the Malaysian Medical Research Ethics Committee and not deemed to require consent.

RESULTS

Bacterial Characteristics

During the 6-year study period, 148 adults with culture-confirmed melioidosis were identified. Of these, 129 (87%) had Bp isolates tested for gentamicin susceptibility; all analyses were limited to these patients. Bp was isolated from a single specimen in 112 cases, 2 specimen types in 16 cases, and 3 specimen types in 1 case. Blood was the commonest source of a positive isolate (n = 113). Others included respiratory secretions (n = 20), pus (n = 10), lung aspirate (postmortem) (n = 2), and peritoneal fluid and urine (n = 1 each).

Of the 129 patients, 84 (65%) had infection with gentamicin-susceptible isolates. Antibiotic susceptibility of Bp isolates was determined by disk diffusion in all patients and by Etest in 76; susceptibility findings corresponded in both tests in all cases. The 76 cases tested by Etest included 49 (64%) gentamicin-susceptible isolates (minimum inhibitory concentration [MIC] = 0.5–3.0 μg/mL) and 27 (36%) gentamicin-resistant isolates (MIC = 24–256 μg/mL).

Seventy-eight Bp isolates, including 49 (63%) gentamicin-susceptible and 29 (37%) gentamicin-resistant isolates, had molecular characterization performed. No differences were found between gentamicin-susceptible and gentamicin-resistant isolates, with all isolates possessing the bimA variant and the YLF gene cluster. AS-PCR assay on the isolates confirmed that all 49 gentamicin-susceptible isolates had a mutation at the amrB gene, which conferred the susceptibility.

Population Incidences and Epidemiological Characteristics

The average annual incidence of melioidosis was 12.3 per 100 000 adult population, with the highest incidence recorded in 2016 (19.4 per 100 000 adult population). Marked variation in average annual incidence rates was observed in the different districts, ranging from 5.8 to 29.3 per 100 000 adult population, with the highest incidence rates recorded in Belaga, Tatau, and Sungai Asap (Figure 1). The proportion of gentamicin-susceptible Bp infection was significantly higher in Belaga (23/26 [88%]), Tatau (25/28 [89%]), and Sungai Asap (11/11 [100%]) than in Bintulu district (10/39 [26%]) (P < .001).

Figure 1.

Number of cases by strain susceptibility and district, and the average annual incidence of adult melioidosis in districts in central Sarawak, Malaysia. The map of central Sarawak depicts the number of gentamicin-susceptible and gentamicin-resistant infections in each district or subdistrict. The average annual incidence of adult melioidosis in districts and subdistricts in the region is also shown. Incidences in districts with small numbers of cases (≤5 cases) were not calculated as most melioidosis cases in those districts are admitted to other hospitals and not to Bintulu Hospital. Five patients in the study had no data on address/location recorded.

No correlation was found between the monthly incidence and average rainfall (r = –0.041, P = .73) (Figure 2), and no changes in the gentamicin susceptibility pattern were recorded throughout the study period.

Figure 2.

Distribution of 129 adult melioidosis cases and average rainfall by month. The bar chart shows the distribution of the 129 adult melioidosis cases according to the month of admission. Average monthly rainfall over the 6-year period (January 2011–December 2016) in Bintulu, Sarawak, is shown in the line graph. Cases are categorized into gentamicin-susceptible and gentamicin-resistant infections.

Patient Demographics, Risk Factors, Clinical Manifestations, and Investigations

One hundred fourteen patients (88%) were aged ≤60 years, including 51 (40%) who were ≤40 years (Figure 3). No significant differences in age or sex were noted between gentamicin-susceptible and gentamicin-resistant infections (Table 1). Seventy-eight patients (60%) had at least 1 risk factor for melioidosis, and significantly more patients with gentamicin-susceptible Bp infection had no identified risk factors compared with patients with gentamicin-resistant infection (40/84 [48%] vs 11/45 [24%]; odds ratio {OR}, 2.8 [95% confidence interval {CI}, 1.3–5.9]; P = .01). Diabetes mellitus was the commonest risk factor identified, present in 63 (49%) patients, including 25 (40%) who were newly diagnosed during the admission for melioidosis. Diabetes was present less frequently in those with gentamicin-susceptible infections compared to those with gentamicin-resistant infections (35/84 [42%] vs 28/45 [62%]; OR, 0.4 [95% CI, .3–.9]; P = .03).

Figure 3.

Distribution of the 129 adult melioidosis cases by age and gentamicin susceptibility of isolates. Bar chart shows the distribution of cases according to age. Cases are categorized into gentamicin-susceptible and gentamicin-resistant infections.

Table 1.

Demographic, Presenting Clinical Features on Admission, Subsequent Findings, and Outcome of Adults With Melioidosis Based on Burkholderia pseudomallei Gentamicin Susceptibility

Characteristic	All Patients	Gentamicin-Susceptible B pseudomallei Infection	Gentamicin-Resistant B pseudomallei Infection	P Value
No. of patients	129	84	45
Demography
Age, y, median (IQR)	44.0 (34.0–52.0)	44.0 (32.0–52.8)	45.0 (37.0–51.0)	.94
Male sex	95 (74)	61 (73)	34 (76)	.71
Malaysian citizen	93 (72)	59 (70)	34 (76)	.51
Preexisting conditions/risk factorsa
Diabetes mellitus	63 (49)	35 (42)	28 (62)	.03
Chronic kidney disease	11 (9)	8 (10)	3 (7)	.58
Lymphoma	3 (2)	1 (1)	2 (4)	.24
Congestive cardiac failure	2 (2)	1 (1)	1 (2)	.65
Bronchiectasis	2 (2)	0 (0)	2 (4)	.05
COPD	1 (1)	0 (0)	1 (2)	.17
No identified risk factorb	51 (40)	40 (48)	11 (24)	.01
Duration of illness before admission, d, median (IQR)	7.0 (3.0–13.5)	7.0 (4.0–14.0)	4.0 (2.0–7.0)	.02
Presenting manifestation
Pneumonia	91 (71)	61 (73)	30 (67)	.48
Undifferentiated fever	17 (13)	12 (14)	5 (11)	.61
Soft tissue abscess	12 (9)	9 (11)	3 (7)	.45
Genitourinary	4 (3)	1 (1)	3 (7)	.08
Osteomyelitis	2 (2)	1 (1)	1 (2)	.65
Dacrocystitis	2 (2)	0 (0)	2 (4)	.05
Neurological	1 (1)	0 (0)	1 (2)	.17
Investigations
Hemoglobin, g/dL, median (IQR)	12.2 (10.6–13.6)	12.2 (10.7–13.5)	12.2 (10.4–13.8)	.73
WBC count, × 109/L, median (IQR)	10.4 (7.2–15.1)	10.5 (7.4–14.8)	10.3 (6.8–15.6)	.86
Platelet count, × 109/L, median (IQR)	188 (107–268)	190 (117–269)	175 (93–263)	.94
Urea, mmol/L, median (IQR)	7.0 (4.5–14.0)	6.9 (4.5–14.2)	7.7 (3.9–13.7)	.42
Creatinine, μmol/L, median (IQR)	103 (66–192)	100 (66–184)	107 (67–206)	.85
Abdominal imagingc
Spleen abscesses only	28 (33)	22 (39)	6 (21)	.08
Liver abscesses only	5 (6)	1 (2)	4 (14)	.02
Both spleen and liver abscesses	18 (21)	12 (21)	6 (21)	.93
No liver/spleen abscesses	34 (40)	21 (38)	13 (45)	.51
Subsequent findings and outcome
Bacteremia	113 (88)	74 (88)	39 (87)	.81
Septic shock	61 (47)	36 (43)	25 (56)	.16
Acute kidney injury	57 (44)	35 (42)	22 (49)	.43
Died	45 (35)	27 (32)	18 (40)	.43

Data are presented as No. (%) unless otherwise indicated.

Abbreviations: B pseudomallei, Burkholderia pseudomallei; COPD, chronic obstructive pulmonary disease; IQR, interquartile range; WBC, white blood cell count.

aSome patients had >1 preexisting condition/risk factor.

bData do not include hazardous alcohol use as too few cases had this information recorded.

cEighty-five patients had abdominal ultrasonography (gentamicin-susceptible B pseudomallei infection, n = 56; gentamicin-resistant B pseudomallei infection, n = 29).

Among all presentations, 127 (98%) were acute while only 2 (2%) were chronic. The median duration of symptoms prior to admission was significantly longer in patients with gentamicin-susceptible infection (Table 1). Pneumonia was the most common presentation (71%). Undifferentiated fever, soft tissue abscesses, and pneumonia accounted for 93% of all melioidosis presentations. Neurological presentations were rare (n = 1). No significant differences were noted in the clinical presentations, laboratory parameters, and radiological findings (abdominal ultrasound and chest radiographs) of patients with gentamicin-susceptible and gentamicin-resistant infection (Table 1). Of 85 patients who had abdominal imaging, 51 (60%) had abdominal visceral abscesses; splenic abscesses were present in 54% (46/85) and liver abscesses in 27% (23/85).

Chest radiographs of all 91 melioidosis pneumonia patients were analyzed. All presented with acute melioidosis, including 62%, 87%, and 92% who had symptoms for ≤1, ≤2, and ≤3 weeks, respectively. Fifty-two (57%) had multilobar disease, all involving both lungs. Involvement of a single lobe occurred in 38 (42%): 22 (58%) had discrete patchy infiltrates, whereas 16 (42%) had dense lobar consolidation. The upper lobe was affected in 12 of 38 (32%). Pleural effusions were present in 12 of 91 cases. One patient had a normal chest radiograph.

Clinical Course and Outcome

Forty-five (35%) patients died. Death occurred within a median of 3 (interquartile range, 2–13) days following admission and was unrelated to preadmission illness duration. Mortality was similar in patients from the various districts. Those aged >70 years had a higher mortality (5/7 [71%] vs 40/122 [33%]; P = .04). Mortality was not related to presence of risk factors or to gentamicin susceptibility of isolates: 15 of the 45 (33%) fatal cases had no identified risk factor, and mortality was 32% (27/84) in the gentamicin-susceptible cohort and 40% (18/45) in the gentamicin-resistant cohort (P = .30). Furthermore, mortality in patients without risk factors was similar with both gentamicin-susceptibility patterns (11/40 [28%] vs 4/11 [36%]; P = .45). Overall, no significant differences were noted in the epidemiological or clinical characteristics of fatal gentamicin-susceptible and gentamicin-resistant cases.

Bacteremia was present in 113 (88%) patients overall and in 98% of the 45 fatal cases (P = .01). Septic shock was strongly associated with mortality, recorded in 89% (40/45) of fatal cases but in only 25% (21/84) of survivors (P < .001). Among 61 septic shock patients, the case fatality rate was 66%. Acute kidney injury was significantly associated with mortality, recorded in 34 (76%) fatal cases but in only 23 (27%) survivors (P < .001). Patients with pneumonia had a higher mortality than patients with other presentations (38/91 [42%] vs 7/38 [18%]; OR, 3.2 [95% CI, 1.3–8.0]; P = .01).

Among fatal cases, 84% received ventilatory/intensive care support. Of the 7 patients not receiving intensive care, all were aged >70 years or had severe debilitating comorbidities, and escalation of care was withheld. Only 58% of the fatal cases received a melioidosis-active antibiotic at admission. However, overall, patients who received melioidosis-active antibiotics upon admission had similar mortality to patients who did not receive melioidosis-active antibiotics (26/70 [37%] vs 19/59 [32%]; P = .55). Eight (18%) fatal cases never received a melioidosis-active antibiotic; all succumbed within 3 days of admission and before culture results were known. Only 1 patient overall (with gentamicin-susceptible Bp infection) received gentamicin, in combination with melioidosis-active antibiotics.

Among survivors, complete follow-up data were available in only 58% (49/84) of cases. No relapses were recorded in these patients during the 6-month follow-up period.

DISCUSSION

This study from Sarawak, Malaysian Borneo, compares melioidosis infections caused by rare gentamicin-susceptible Bp with infections caused by the more typical gentamicin-resistant isolates. There were no significant differences in clinical manifestations, severity, or outcomes between infections from gentamicin-susceptible and gentamicin-resistant Bp. What was found were markedly higher incidences of melioidosis and a vastly greater proportion of gentamicin-susceptible Bp in the rural interior regions (Belaga, Tatau, and Sungai Asap districts) in comparison to the coastal Bintulu region, which includes Bintulu town.

In our previous study, we showed that all gentamicin-susceptible Bp isolates from this region and others in Sarawak belonged to a single multilocus sequence type, ST881, or its single-locus variant ST997, whereas gentamicin-resistant Bp isolates included a diversity of sequence types [6]. Whether this still holds true for the isolates in this subsequent study is unknown as the isolates have not been genotyped. Nevertheless, all gentamicin-susceptible isolates that were analyzed were confirmed to possess a single-nucleotide polymorphism mutation in the amrB gene through a previously published AS-PCR assay, inferring that these isolates may in fact belong to ST881 or ST997. In a recent melioidosis study in Kapit, another rural district located in central Sarawak (Figure 1), 92% of infections were due to gentamicin-susceptible isolates, consistent with the findings in the rural areas in our study [11]. In other melioidosis-endemic regions in Australia, Asia, and SEA, including peninsular Malaysia, studies of clinical Bp isolates have demonstrated usually high levels of strain diversity [12-16]. It is hypothesized that the predominance of gentamicin-susceptible Bp seen in rural districts in Sarawak has resulted from expansion of a clonal Bp population, with a single efflux pump mutation preventing the efflux of gentamicin that normally occurs in wild-type Bp. However, whole genome sequencing of current and future strains from across the region will be required to test this hypothesis as it is noted that several other genetic variations have been described to cause gentamicin susceptibility [17]. Furthermore, to date, the gentamicin-susceptible ST881 found in Sarawak has not been reported from other locations, with the exception of a single case in a Chinese traveler returning from working in the Malaysian jungle [18].

While it is recognized that distinct genetic populations of Bp may be spatially clustered and restricted to specific geographical regions [14, 19–21], the reasons for selection and propagation of ST881 to be the predominant genotype in Sarawak remain unknown. Any specific environmental niche for ST881 has yet to be found and indeed environmental sampling in Sarawak to date has been unable to recover ST881, even when using culture media avoiding gentamicin (Y. Podin, unpublished data). The clinical data from this study show no evidence of attenuated virulence with gentamicin-susceptible Bp, consistent with a prior study of another gentamicin-susceptible Bp that retained full virulence in an acute mouse melioidosis model [17]. It is conceivable that gentamicin-susceptible strains are more infective compared to other Bp strains in the Sarawak environment, but a biological reason for this is not apparent. While nonrandom distributions of Bp strains in clinical samples compared to those in environmental samples in previous studies have alluded to this possibility of increased strain infectivity [22, 23], studies from Australia have not found such differential infectivity/virulence between environmental and clinical Bp [24].

The restriction of ST881 to Sarawak is in sharp contrast to the other wild-type STs (gentamicin-resistant Bp) described from Sarawak, which have been found to be widely dispersed through SEA [6]. This is similar to recent studies from Laos and Myanmar [25, 26], which suggest a dynamic process of more recent possibly anthropogenically driven regional dispersal of Bp, in addition to the ancient intercontinental dispersals originally from Australia, where Bp is thought to have arisen [12]. The spread of Bp in the Americas is considered a more recent phenomenon, with recent evidence suggesting that Bp may now be endemic in the United States [4].

That incidence rates are higher in the rural inland areas of Sarawak in comparison to the coastal Bintulu region, and the substantially higher proportion of gentamicin-susceptible Bp in the rural areas is likely to reflect a combination of environmental factors and human host behaviors facilitating infection. The rural regions consist of subsistence indigenous populations, plantation workers with ongoing deforestation for crops such as oil palm, and some large-scale construction projects such as the Bakun dam with concomitant environmental changes and population resettlements. While the link between increased melioidosis and environmental perturbation has been long recognized [27], any specific drivers of emergence and propagation of ST881 remain to be elucidated. Higher overall rates of melioidosis in rural populations are also well recognized and reflect increased exposure to the environment in occupational and daily living activities and likely also ingestion of untreated water containing Bp [28].

In northern Australia and northeast Thailand, 81% and 75% of melioidosis cases, respectively, are reported to occur during the wet season [29, 30]. This contrasts with the lack of seasonal variation in the incidence of melioidosis in Sarawak seen in this study. This may reflect the contrast between the wet/dry tropics of northern Australia and northeast Thailand, where prolonged dry periods occur over some months, and the rainfall pattern in Sarawak (Figure 2).

In Sarawak, 40% of all patients with melioidosis had no identified risk factors and the rate was 48% in those infected with gentamicin-susceptible Bp. Nevertheless, overall, 49% of patients had diabetes as a risk factor, and in 40% of these the diabetes was diagnosed only after admission with melioidosis. The rate of diabetes was consistent with other studies [31] but was significantly lower in those infected with gentamicin-susceptible Bp at 42%. In the recent melioidosis study in Kapit, the absence of risk factors was reported in nearly 50% of cases [11], a figure substantially higher than described in most other studies [31]. The lower rate of diabetes and the higher proportion without identified risk factors in those infected with gentamicin-susceptible Bp may in part reflect the rural population that accounted for the vast majority of these cases, with a higher exposure burden to environmental Bp in the rural areas. Interestingly, these findings correlate well with our previous pediatric melioidosis study, where high incidences of severe and fatal disease were found in children with no risk factors, caused overwhelmingly (97%) by gentamicin-susceptible isolates [32]. One limitation of the present study, however, was that the prevalence of hazardous alcohol use was not determined as most patients did not have data recorded.

Nearly all (98%) patients with melioidosis in Sarawak had acute presentations, and 71% had pneumonia as the presenting diagnosis. Bacteremia and septic shock were found in 88% and 47% of patients, respectively. In other endemic regions, up to 15% of cases have chronic presentations, and pneumonia, recognized as the most common manifestation, is found in 51%–61% of cases [33]. The higher prevalence of acute presentations, pneumonia, bacteremia, and septic shock in Sarawak than in Australia and Thailand may well reflect that those other, less severe cases of melioidosis are going undiagnosed. While undifferentiated fever with positive blood culture and soft tissue infections with culture of Bp from aspirates or pus also occurred, genitourinary, neurological, and cutaneous melioidosis were rare or absent. This spectrum of clinical manifestations is similar to that seen elsewhere in SEA, which is well documented to be distinct from that observed in Australia [24].

Sixty percent (51/85) of those who had abdominal imaging were found to have splenic and/or liver abscesses; splenic abscesses (54% of those imaged) were twice as common as liver abscesses. Thus, the spleen is the second most affected organ (after the lungs). Interestingly, the spleen was also found to be one of the commonest affected organs in children with melioidosis in this region [34]. In comparison, splenic abscesses are detected in only 25% and 5% of melioidosis patients in Thailand and Australia, respectively [29, 35]. The high rate of splenic involvement in Sarawak likely reflects the high prevalence of bacteremia, as does multilobar involvement, the most common manifestation of melioidosis pneumonia in Sarawak, noted in 57% of cases. This was similar to findings from Cambodia but contrasted with findings from northern Australia, where multilobar involvement was present in only 30% of cases [36, 37].

Worldwide, mortality from melioidosis has been reported to be between 9% and 70% [33], and the mortality in this study was 35%. The presence of host risk factors and older age have been reported to be the most important determinants of mortality from melioidosis, and in fact, death from melioidosis in those without risk factors is believed to be extremely uncommon if rapid diagnosis, appropriate antibiotics, and state-of-the-art management of sepsis are available [24, 29]. The relatively high mortality in those without risk factors in Sarawak may be due to several reasons. First, delays in diagnosis and initiation of appropriate antibiotics still occur despite the recognition of central Sarawak as a melioidosis hotspot. Second, although intensive care is available, facilities and expertise are limited compared to those of better-resourced regions. Finally, pneumonia, bacteremia, and septic shock are all associated with increased mortality, and the very high prevalence of these manifestations in those diagnosed with melioidosis in Sarawak underpins the mortality of 35%.

This study has several limitations. First, selective culture media, such as threonine-basal salt solution plus colistin, was not routinely used [38]. This may have resulted in a reduced yield of Bp from nonsterile site specimens and led to fewer diagnoses of less severe/localized melioidosis infections. Second, for antibiotic susceptibility, an adaptive interpretation of the CLSI guidelines was used, as CLSI does not provide specific criteria for interpreting disk diffusion tests for Bp. Recently the European Committee on Antimicrobial Susceptibility Testing has defined specific disk diffusion zone diameter distributions and epidemiological cutoff values for Bp [39, 40]. Nevertheless, the specific interpretation of gentamicin susceptibility for Bp isolates from our study remains valid, given the clear separation of susceptible and resistant isolates on both Etest and disk diffusion. Finally, the study is unable to inform on the usefulness or harmfulness of gentamicin for treatment of gentamicin-susceptible Bp, as only 1 patient in the study received the antibiotic.

CONCLUSIONS

Two-thirds of melioidosis cases in central Sarawak were caused by gentamicin-susceptible Bp, which dominates in the high-incidence rural interior regions. Clinical manifestations and outcomes were the same as for infections with gentamicin-resistant Bp. Extremely high rates of bacteremia, pneumonia, splenic abscesses, and septic shock were observed, with overall mortality of 35%. Further studies are required to determine if all gentamicin-susceptible Bp infections in Sarawak are clonal and to ascertain their environmental drivers and niches.