Melioidosis: An Imported Case From Madagascar


Gianandrea Borgherini, MD, Service de Pneumologie et Maladies Infectieuses, Groupe Hospitalier Sud Réunion, BP 350, 97448 Saint Pierre, La Réunion, France. E-mail:


Burkholderia pseudomallei, the causative agent of melioidosis, is endemic in Southeast Asia and Northern Australia but is a rare pathogen in other parts of the world. No human case of melioidosis has been reported in Madagascar until now. We describe a case of pulmonary melioidosis probably acquired in Madagascar.

Melioidosis, a disease caused by a gram-negative bacterium, Burkholderia pseudomallei, is a major cause of morbidity and mortality in northeast Thailand1 and in northern Australia.2 It is also present in Southeast Asia and in the Indian Subcontinent. Sporadic cases have been reported in other tropical countries such as Mauritius.3

No human cases have been reported in Madagascar. In 2004, a first case imported from Madagascar was recognized in La Réunion,4 a French overseas department, located east of Madagascar. Here, we report a second case of melioidosis probably imported from Madagascar.

Case report

A 58-year-old man came to our hospital with a 3-week story of fever and productive cough. The patient, a retired manager, spent most of his life in France. He traveled for short periods in Tunisia, Turkey, and Mauritius (6 years ago) and had been living in Antananarivo, Madagascar, for the past 5 years. He was a heavy smoker with 35 packs a year and no reported alcohol excess, otherwise, the medical history was unremarkable.

The first symptom appeared at the beginning of March 2005 during a stay in Mahajanga, a city in the northwestern coast of Madagascar, where the patient went for holidays for a couple of weeks. No potential source of contamination was identified.

The patient started abruptly to have high fever with cough in the last days of his stay, and he decided to go back to Antananarivo, where he was admitted to a local hospital. The chest X-ray showed a patchy infiltrate of the right upper and middle lobes.

No blood or sputum cultures were taken, and the sputum acid-fast stain was negative. An intravenous antibiotic treatment with amoxicillin/clavulanate and ciprofloxacin was started, and the patient was treated for 10 days with consequent clinical improvement.

The patient was dismissed with an oral treatment of amoxicillin/clavulanate and levofloxacin. Five days later, the patient was feverish again, and he was referred to our hospital in Saint Pierre, La Réunion on 26th of March.

On examination, he was afebrile and had a productive cough with purulent sputum and rales on the apex. Initial blood tests showed an elevated white blood cell count (12,000 cells/μL, with 73% neutrophils) and C-reactive protein (168 mg/L). No other significant hematological or biochemical abnormalities were found.

On the first 2 days of hospitalization, the patient had no fever and we did not start any antibiotic treatment. On day 2, a bronchoscopy with bronchoalveolar lavage was done. On day 3, the patient started to have high fever. Four blood cultures were drawn; all of them were sterile. On day 5, the culture of bronchoalveolar lavage came positive for a gram-negative rod whose antibiotic susceptibility testing showed a sensitivity to ceftazidime, imipenem, ciprofloxacin, and a resistance to all aminosides and colistin.

The microorganism was identified as B pseudomallei, by use of the API 20 NE identification system. The isolate was sent to a referral center, the Centre de Recherche du Service de Santé des Armées in La Tronche, France, for confirmatory tests. The strain grew on Ashdown’s medium with typical B pseudomallei colony morphology. The phenotypic identification was confirmed, including the lack of arabinose assimilation.

The polymerase chain reaction identification using type III secretion system genes5 was positive for B pseudomallei. Serology performed in accordance with the protocol in use at the Health Protection agency (London, U K)(T. Pitt, personal communication to F.T., 2004) was strongly positive with an immunoglobulin G titer of 16.000.

Imipenem, 500 mg tid, was started. By day 5 of imipenem treatment, the patient was afebrile. A total body computed tomography scan did not show any secondary localization.

The patient received 2 weeks of intravenous antibiotherapy followed by oral trimethoprim/sulfamethoxazole (320/1,600 mg every 12 h) and doxycycline (100 mg every 12 h). He was discharged at day 20 of treatment and then continued the oral maintenance treatment with trimethoprim/sulfametoxazole for 5 more months. Three months after the end of the treatment, the patient is still well.

The possibility of the existence of melioidosis in Madagascar was first raised in 1936 by Girard6 who isolated B pseudomallei from the lymph node of a pig. In 1982, Galimand and Dodin7 reported the presence of B pseudomallei in soil samples coming from the zoo of Antananarivo. Nevertheless, no human case of melioidosis has been reported in Madagascar until now. Recently, a case of septicemic melioidosis in a French expatriate living in Madagascar has been reported in La Réunion,4 but we have no information on the travel history of the patient. It is worth noting that an isolated veterinary case of melioidosis, possibly linked to an outbreak among horses in France, was reported in La Réunion in 1979.7

Our patient neither traveled previously to endemic areas nor was exposed to potential source of contamination as he was living in France. Six years ago, he spent just 1 week, during the dry season, in a holiday resort in Mauritius, where recently an autochthonous case of melioidosis has been described.3 Meanwhile, considering the length and the condition of the stay, it is very unlikely that melioidosis was acquired at that time. Therefore, there is good evidence that the infection has been acquired in Madagascar.

The patient spent his days in Mahajanga in a fishermen’s village where hygiene was quite poor: no clear inoculation event was noted. It is worth noting that in Mahajanga, the rainy season is from December to March. An association between rainfall and melioidosis has long been recognized.2

Diabetes mellitus, excessive alcohol intake, chronic renal disease, chronic lung disease, and thalassemia are recognized risk factors for melioidosis.2,8 In our patient, we cannot exclude a chronic obstructive pulmonary disease as no spirometry was realized; all the other risk factors were absent.

Overall mortality of melioidosis in adults can range from about 50% in Thailand1 to 19% in northern Australia.2 Clinical features are variable, from an acute fulminant septicemia to a chronic debilitating localized infection, the lung being the most commonly affected organ.

In chronic pulmonary melioidosis, upper lobe changes with infiltrates and/or cavitation can be seen on chest X-ray in 95% of cases,6 and that explains why chronic melioidosis is often initially misdiagnosed as tuberculosis.

Antibiotics commonly suggested for treatment of community-acquired pneumonia or septicemia are not, or only partially, effective in melioidosis for which the antibiotics of choice are imipenem and ceftazidime.9

It is possible that in this region, where tuberculosis is hyperendemic and the diagnostic capacities are limited, melioidosis is far more common than expected. Considering that melioidosis can be a fatal disease and needs a specific antibiotic treatment, further investigations are needed to assess the presence of melioidosis.

Meanwhile, physicians managing patients in Madagascar or coming from this country should be alert to the possibility that this organism might cause community-acquired pneumonia and sepsis.

Declaration of interests

The authors state that they have no conflicts of interest.