Melioidosis in 6 Tsunami Survivors in Southern Thailand

Abstract

Background. Six cases of melioidosis were identified in survivors of the 26 December 2004 tsunami who were admitted to Takuapa General Hospital in Phangnga, a region in southern Thailand where melioidosis is not endemic. All 6 cases were associated with aspiration, and 4 were also associated with laceration.

Methods. We compared the clinical, laboratory, and radiographic findings and the outcomes for these 6 patients with those for 22 patients with aspiration-related melioidosis acquired during 1987–2003 in a melioidosis-endemic region in northeast Thailand. Results of tests for detection of Burkholderia pseudomallei in soil specimens from Phangnga and from northeast Thailand were compared.

Results. The 6 patients (age range, 25–65 years) presented with signs and symptoms of pneumonia 3–38 days (median duration, 6.5 days) after the tsunami. Chest radiograph findings at the onset of pneumonia were abnormal in all cases; 1 patient developed a lung abscess. B. pseudomallei was grown in blood cultures in 3 cases and in cultures of respiratory secretions in 4 cases. Two patients required ventilation and inotropes; 1 patient died. Compared with tsunami survivors, patients with aspiration-related melioidosis in northeast Thailand had a shorter interval (median duration, 1 day) between aspiration and onset of pneumonia; were more likely to exhibit shock, respiratory failure, renal failure, and/or altered consciousness (P = .03); and had a higher in-hospital mortality (64% [14 of 22 patients]; P = .07). These differences may be related to the severity of the near-drowning episode, the inhalation of sea water versus fresh water, the size of bacterial inoculum, and the possible acquisition (among tsunami survivors) of B. pseudomallei via laceration. Only 3 (0.8%) of 360 soil samples from Phangnga were positive for B. pseudomallei, compared with 26 (20%) of 133 samples from northeast Thailand (P < .0001).

Conclusions. Tsunami survivors are at increased risk of melioidosis if they are injured in an environment containing B. pseudomallei.

Melioidosis is a serious infection caused by the gram-negative bacillus Burkholderia pseudomallei. The disease is reported most commonly in Southeast Asia and the Northern Territory of Australia [1, 2]. The number of reported cases is likely to be a considerable underestimate of the true number, because diagnosis relies on culture confirmation, a technique that is not available for large areas of this populous region. B. pseudomallei is present in soil and water in areas of endemicity [3], and infection is acquired by contamination of breaks in the skin or by inhalation. The clinical manifestations of disease are broad ranging, but the most frequent clinical picture at presentation is a septicemic illness, often associated with bacterial dissemination to distant sites [1]. Pneumonia occurs in ∼50% of patients in settings where melioidosis is endemic. Persons who are affected usually have ⩾1 predisposing factor associated with impaired immunity, the most common being diabetes mellitus and/or renal failure [1]. The case fatality rate is ∼50% among Thai adults and ∼19% among persons in Northern Territory, Australia [1, 2]. The incubation period may be very prolonged, with a maximum recorded interval between exposure and clinical manifestation of 62 years [4, 5].

Takuapa General Hospital in Phangnga, a province in southern Thailand, was one of the major referral centers for people injured in this area during the tsunami of 26 December 2004. This hospital saw ∼1000 patients on the day of the tsunami. Six tsunami survivors subsequently developed pneumonia caused by melioidosis. Melioidosis is uncommon in southern Thailand, although occasional cases do occur. A search of computerized patient records revealed that 9 cases of melioidosis have been diagnosed at this hospital during the past 6 years (none of which occurred during the 12-month period before the tsunami). This raises the possibility that individuals affected by the recent Asian tsunami are at increased risk of melioidosis. Here, we report the epidemiological, clinical, microbiological, and radiographic features of these patients. We also compare and contrast features of this group with those of 22 individuals with aspiration-related melioidosis acquired in Ubon Ratchathani, a melioidosis-endemic area of northeast Thailand.

Patients and Methods

Setting. In Thailand, the greatest effects of the tsunami were seen in the southern provinces of Phangnga, Krabi, Phuket, Ranong, Trang, and Satun (figure 1). Most damage occurred in Phangnga, where 4221 people died, 1802 people went missing, and 5597 people were injured [6]. This province is situated on the western coast of southern Thailand facing the Andaman Sea and has a registered population of ∼240,000 people. It is an agricultural region comprised mainly of rubber plantations. Tourism is an important source of revenue and employment, with ∼100 hotels in an area that includes Khao Lak National Park. Takuapa General Hospital provides health care to the population of western Phangnga, with a catchment area stretching >60 km, from Ban Bang Wan in the north to Lak Kan in the south. It is a 177-bed facility with a 6-bed intensive care unit, 4 operating theaters, an emergency department, and medical, surgical, and pediatric wards. Departments of radiology, biochemistry, and microbiology are located at the hospital.

Figure 1

A, Map of Thailand highlighting the southern coastline affected by the tsunami (red) and the location of Ubon Ratchathani Province (blue).

Patients. On 30 December 2004, an inpatient at Takuapa Hospital who had aspirated and sustained a laceration during the tsunami developed severe pneumonia, and cultures of 2 blood samples obtained on that day became positive for B. pseudomallei. Increased awareness for further possible cases resulted from discussions between clinical teams and the microbiology department. A further 5 patients received a diagnosis of B. pseudomallei pneumonia between 31 December 2004 and 7 February 2005. Patient characteristics, vital signs, physical examination findings, and results of routine laboratory tests were recorded in all cases by a combination of prospective review during patient care and a retrospective review of hospital records. The first 5 cases have been reported to the Program for Monitoring Emerging Diseases (ProMED) e-mail list [7], and the World Health Organization, Thailand, was informed as soon as the 5 cases were confirmed.

Microbiological evaluation. Blood specimens, samples of respiratory secretions, and wound swabs were obtained for culture on clinical grounds at the discretion of the primary physicians. Respiratory secretions (sputum or endotracheal suction samples) were collected into a sterile container. A 5-µL loopful of sample was plated directly onto each of 3 agar plates (MacConkey agar, blood agar, and chocolate agar). Wound swabs were transported in bacterial transport medium and streaked onto MacConkey agar and blood agar. All plates were incubated at 37°C in air and examined daily for 2 days. Blood cultures were performed using the automated Bactec 9120 System and Bactec Plus Aerobic/F Medium bottles (BD Biosciences). Bottles were routinely subcultured on days 2 and 7 and at intervening periods if the bottles were flagged as being positive. B. pseudomallei was identified by standard methods, as described elsewhere [8]. Susceptibilities to ceftazidime, imipenem, and doxycycline were evaluated by the disk diffusion assay, and susceptibility to trimethoprim-sulfamethoxazole was determined by the Etest (AB Biodisk). Interpretive standards were based on NCCLS guidelines [9].

Cohort of patients in northeast Thailand with melioidosis after aspiration. Sappasithiprasong Hospital is a 1000-bed Provincial General Hospital in Ubon Ratchathani in northeast Thailand (figure 1). This rice-growing region is a melioidosis-endemic area, where B. pseudomallei accounts for ∼20% of all cases of community-acquired septicemia. All patients with melioidosis admitted to this hospital since 1986 have been included in prospective clinical and therapeutic studies. We have prospectively studied >2500 patients with culture-proven melioidosis. Aspiration during a near-drowning episode is a risk factor for melioidosis, because soil and environmental water in this region commonly contain B. pseudomallei. We identified 22 culture-confirmed melioidosis cases during 1987–2003 that were associated with a near-drowning episode.

Environmental sampling in Phangnga. A total of 360 soil samples were collected from 18 distinct sampling sites in Phangnga during a 2-day period in January 2005 (figure 2). Of these, 140 samples were from 7 sites unaffected by the tsunami, and 220 were from 11 sites with major damage due to the tsunami. At each site, 100-g samples were collected at 2 depths (10 cm and 30 cm) at each of 10 points within a radius of 100 m. In areas unaffected by the tsunami, sampling was performed on damp soil around fresh or brackish water. In tsunami-affected areas where the previous topography of the land had been highly disrupted, soil was sampled from around water collections. All sampling was performed within 2 km of the sea. Tidal sea water was not sampled because B. pseudomallei has not previously been identified from this source. The soil was processed and B. pseudomallei was recovered as described elsewhere [10].

Figure 2

Map of Takuapa District within Phangnga province, showing the area damaged by the tsunami, the location of the 6 patients who developed melioidosis after the tsunami struck, and the soil sampling sites.

The results of the Phangnga soil survey were compared with results of a similarly conducted unpublished survey performed during June and July 1998, when 133 soil samples from 19 sites in northeast Thailand were cultured. All sites were flooded rice paddies, the majority of which were sampled after ploughing but before planting. Samples of damp soil were taken at 2 depths (10 cm and 30 cm) from around the edge of the paddy field.

Statistical analysis. Statistical analysis was performed using the Intercooled Stata statistical program, version 8.0 (Stata).

Results

Patient characteristics. The median age of the 6 patients at Takuapa Hospital was 41 years (range, 25–65 years) (table 1). Four patients were men, all of whom were laborers. Five patients were permanent residents of Phangnga, and 1 (patient 4) was a contract laborer and temporary resident who was from northeast Thailand. Three patients were known to be diabetic. All 6 patients had a history of aspiration during the tsunami, and 4 patients had tsunami-related lacerations. Five patients presented to Takuapa Hospital on the day of the tsunami; 2 were discharged, 1 was admitted for treatment of aspiration, and 2 were admitted for treatment of severe lacerations.

Table 1

Demographic characteristics and clinical features at hospital admission for 6 Thai patients with melioidosis after the 26 December 2004 tsunami.

Clinical features and related data. The median incubation period between the aspiration and/or laceration episode and the onset of features of pneumonia was 6.5 days (range, 3–38 days). All patients had cough, sputum production, and crackles on auscultation. Features of the 6 patients with pneumonia and culture-confirmed melioidosis are summarized in table 1.

Laboratory assessment. The median leukocyte count was 10,100 leukocytes/mm³ (range, 8800–22,700 leukocytes/mm³). The hemoglobin level ranged from 6.6 to 14.6 g/dL (median level, 11.9 g/dL), with the lowest level observed in a patient with blood loss from a severe lower-limb laceration. The serum bicarbonate concentration ranged from 11.0 to 21.0 mmol/L (median concentration, 17.6 mmol/L; normal range, 25–28 mmol/L). Platelet counts, serum potassium levels, blood urea nitrogen levels, and creatinine levels were within the normal range for patients 1–5, but blood urea nitrogen and creatinine levels were increased for patient 6 (19.3 mmol/L and 353.6 µmol/L, respectively). Patients 1 and 2 had high blood glucose levels at hospital admission (17.5 and 6.4 mmol/L, respectively), consistent with a known history of diabetes mellitus. Liver function tests were not performed. Arterial blood gas data measured at admission for patient 1, who presented to the hospital with septic shock and had metabolic acidosis after intubation, were as follows: pH, 7.25; partial pressure of CO₂, 2.7 kPa; and partial pressure of O₂, 13.6 kPa.

Microbiological assessment. Identification of B. pseudomallei from any sample was considered to be diagnostic for melioidosis. B. pseudomallei was grown in cultures of blood obtained from 3 persons and in cultures of respiratory secretions obtained from 4 persons (table 2). Results of disk diffusion tests revealed that all isolates were susceptible to ceftazidime, imipenem, doxycyline, and amoxicillin-clavulanate. Results of Etests showed that all isolates were susceptible to trimethoprim-sulfamethoxazole (MIC, <0.75 mg/L in all cases).

Table 2

Results of culture on days 1–13 after the 26 December 2004 tsunami for 6 Thai patients with melioidosis.

Radiographic assessment. Findings on all chest radiographs were abnormal at the time of presentation with pneumonia (figure 3). Chest radiography performed a second time for patient 3 (a total of 4 days after the onset of pneumonia) revealed an opacity in the left-middle zone, which, by day 11, contained an obvious cavity (figure 3).

Figure 3

Chest radiographs for 6 patients with melioidosis after the 26 December 2004 tsunami that were obtained on the day of pneumonia onset (day 1) or on day 11 of pneumonia.

Treatment and outcome. Five of 6 patients survived. One survivor (patient 1) presented with severe pneumonia with respiratory distress and septic shock (blood pressure, 60/35 mm Hg; O₂ saturation [by pulse oximetry], 80%). Dopamine therapy was required for 3 days, and positive-pressure ventilation was used for 7 days. The patient who died (patient 6) developed features of pneumonia 38 days after the tsunami but did not present to the hospital until day 43, at which time she had respiratory failure and was in shock; 6 h after admission, she had a cardiac arrest. Antimicrobial therapy with agents to which B. pseudomallei is usually susceptible (ceftazidime or a carbapenem) was started on days 2, 3, 5, 4, 1, and 5 after the clinical onset of pneumonia in patients 1–6, respectively. A single dose of imipenem was given to the patient who died, and parenteral therapy was given for 12–15 days to the rest of the patients, followed by a switch to oral trimethoprim-sulfamethoxazole plus doxycycline to complete a 20-week course of treatment.

Three of 4 patients who sustained lacerations during the tsunami developed complicated wounds. Patient 1 and 6 required wound debridement. Patient 5 sustained a severe laceration on the right leg, in which several muscles were torn and the anterior and posterior tibial arteries were damaged. There was severe bleeding, and the wound was extensively contaminated with mud. At presentation to the hospital, the limb was immobilized by use of a splint, a dressing was applied, and parenteral treatment with cloxacillin, gentamicin, and metronidazole was given. Cloxacillin and gentamicin were replaced by cefotaxime after 48 h. The wound was initially debrided in the operating theater, but by 30 December, the right foot was cold, the dorsalis pedis pulse was absent, and the right leg was foul smelling and gangrenous, at which time amputation below the knee was performed. Features of pneumonia followed 6 days after amputation.

Comparison of melioidosis occurring after tsunami-related aspiration and after accidental aspiration in northeast Thailand. During 1987–2003, a total of 22 patients were admitted to Sappasithiprasong Hospital with culture-confirmed melioidosis following a tsunami-related aspiration event or a near-drowning episode. All accidents occurred inland in the province of Ubon Ratchathani (figure 1). Twenty patients had no history of concomitant laceration. Two patients aspirated as a result of an automobile traffic accident; one had no other injuries, but the other sustained multiple injuries, including trauma to the spine and abdomen. The median age of these patients was 20 years. Nine patients were children aged <8 years, and the remaining patients were rice farmers (n = 11) or laborers (n = 2). Three people had preexisting risk factors for melioidosis (all 3 individuals had renal impairment, and 1 was also diabetic), and 1 person was pregnant.

Despite the brief period between aspiration and presentation in most cases, 14 patients (64%) were febrile at the time of admission. Chest radiograph findings were recorded for 21 of 22 patients at the time of presentation, all of which showed abnormalities. A comparison of characteristics of patients who developed melioidosis after the tsunami with those of patients in Ubon Ratchathani showed broad similarity in terms of risk factors for melioidosis, chest-related signs, chest radiography findings, and leukocyte count at the time of admission (table 3). Microbiological findings were also similar; the intervals between the onset of pneumonia and the first culture positive for B. pseudomallei were roughly equivalent, and, importantly, the proportions of patients who were bacteremic were similar. However, there were several striking differences. The patients in Ubon Ratchathani tended to be younger and to develop pneumonia more rapidly following aspiration than did the patients in Phangnga (P = .02 and P < .001, respectively). Patients in Ubon Ratchathani were more likely to have severe disease, as judged by the presence of shock, respiratory failure, renal failure, and/or altered consciousness (P = .03). They were also more likely to die: 14 (64%) of 22 patients in Ubon Ratchathani died, compared with 1 (17%) of 6 patients in Phangnga (P = .07).

Table 3

Comparison of characteristics between patients who developed melioidosis after the 26 December 2004 tsunami and patients who developed melioidosis after accidental aspiration in northeast Thailand.

Environmental sampling. Three (0.8%) of 360 soil samples (1 each from 3 [17%] of 18 separate sites) in Phangnga tested positive for B. pseudomallei; 2 of the 3 samples were from soil unaffected by the tsunami (figure 2). In contrast, 26 (20%) of 133 samples (from a total of 8 [42%] of 19 sites) in northeast Thailand—an area of high endemicity—tested positive for B. pseudomallei (P < .0001 for individual samples). The number of organisms per gram of soil was very low in Phangnga (2 cfu/g of soil at 1 site, and <1 cfu/g at 2 sites), compared with a geometric mean concentration of 315 cfu/g of soil (range, 10–17,000 cfu/g of soil; 95% CI, 136–733 cfu/g of soil) for the 26 samples from northeast Thailand that tested positive (P = .005).

Discussion

Melioidosis following aspiration or a near-drowning episode is well recognized [11–13]. Approximately 1% of patients presenting to Sappasithiprasong Hospital in northeast Thailand with culture-proven melioidosis develop disease associated with aspiration. Inhalation of aerosols during heavy monsoonal rain and wind probably plays an important role in causing melioidosis in the Top End of the Northern Territory of Australia [14]. Inhalation of contaminated aerosols created during helicopter takeoff and landing was considered to be a cause of melioidosis in Army personnel involved in the Vietnam War. Direct delivery of B. pseudomallei into the lungs may result in the enhanced ability of this pathogen to cause disease; alternatively, the bacterial inoculum suspended in an aerosol may be higher than that delivered during an inoculation event. We also recognize that 4 of the 6 patients in this study with tsunami-related cases may have developed melioidosis as a result of laceration and wound inoculation. This may be particularly relevant to patient 6, who presented to the hospital 43 days after the tsunami.

A comparison of the 6 persons with tsunami-related melioidosis with the individuals who acquired melioidosis through a near-drowning episode in northeast Thailand shows that patients in the northeast have a more severe form of disease and a higher rate of complications and death. Possible reasons for these differences includes the severity of the near-drowning event, the type of water aspirated (sea water vs. brackish irrigation water), and the higher bacterial densities in soil and environmental water. It is also possible that ⩾1 of the tsunami-related cases resulted from inoculation rather than from inhalation, which could have affected the time to presentation to the hospital and the severity of the disease.

Many thousands of people aspirated and/or had lacerations during the tsunami. A key question is how many individuals may be at future risk for melioidosis or other infections. Case reports highlight the risk of infection with predictable aspiration-related organisms, such as Aeromonas species [15], and with unusual organisms, including environmental fungi [16]. Tetanus has also been linked to aspiration in 1 case [17]. However, pneumonia is not inevitable after a near-drowning episode, as demonstrated by 2 retrospective studies. A study from Singapore of 17 patients reported that none developed pneumonia [18], and a study of 125 patients in Holland reported that 14.7% who had clinical findings during initial examination developed pneumonia [19]. The specific risk for melioidosis among people who experienced the tsunami will depend on the presence and concentration of B. pseudomallei in the soil, combined with the extent of aspiration or wound inoculation. Soil sampling has not been performed across Asia, but cases of disease are an indicator of B. pseudomallei in the environment. Melioidosis in indigenous people or in returning travelers has been reported for Sri Lanka [20], southern India [21, 22], the southwest Indian Ocean island of Mauritius [23], Myanmar [24], and Indonesia [25]. B. pseudomallei has been identified in soil samples in Malaysia, where disease is well recognized [26–28]. In Thailand, the incidence of melioidosis and the rate of B. pseudomallei isolation from the soil is closely associated, with high prevalences of soil samples with the organism and of disease in the northeast and low prevalences in the central and southern regions [10, 29]. Our finding that the bacterial concentration in soil specimens is greater in the northeast, compared with the south, is consistent with previously published findings [10, 29]. Taken together, these studies indicate that there is a risk of melioidosis for tsunami survivors over a wide geographical area. More cases have been reported via ProMED, including 2 suspected melioidosis cases in hospitalized patients in Meulaboh, Indonesia [30], 4 cases in Europeans who were injured in southern Thailand (plus 1 case in a fifth individual who was culture positive in the absence of features of melioidosis) [31], and 1 case in an Australian injured in Sri Lanka [32].

We predict that the number of reported cases of melioidosis in tsunami survivors worldwide will increase during the coming years. Increased awareness of this potential problem may reduce mortality and morbidity among tsunami survivors.

Acknowledgments

We thank the staff at Takuapa Hospital, Phangnga, especially Drs. Sudaporn Praikanarat, Charlermchai Praikanarat, Punyasiri Jindakul, Somgamon Boonyuen, Pimporn Suksathien, Sommon Jindakul, Chansak Watcharong, Orawin Luecha, Somchai Buasod, Somboon Sukhumcampee, and Somsak Choksuchat. We also thank the staff of Sappasithiprasong Hospital, Ubon Ratchathani. For help with environmental sampling, we thank the personnel of the Royal Thai Fleet base at Naval Third Area Command and the staff at the Wellcome Trust-Mahidol University-Oxford Tropical Medicine Research Program, in particular, Mr. Gumphol Wongsuvan, Miss Premjit Amornchai, and Mr. Sayan Langla.

Financial support. The Wellcome Trust (Career Development Award in Clinical Tropical Medicine to S.J.P.).

Potential conflicts of interest. All authors: no conflicts.

Author notes

a

W. Chierakul and W. Winothai contributed equally to this article.