Melioidosis, Insulin, and the Diabetes Connection

Abstract & Commentary

Synopsis: The medical literature originating from geographic areas where melioidosis is an endemic disease, such as southeast Asia and northern Australia, makes frequent reference to its clinical associations with diabetes mellitus. Which type(s) of diabetes are associated has not been clearly evident from prior case reports. However, the causative organism, Burkholderia pseudomallei, has a unique ability to bind human insulin, and this bacterial property may underlie a remarkable biological and clinical relationship with important implications for diabetics traveling to endemic areas. In addition, the geographic distribution of this disease in 1999 is far wider than we had appreciated in the past.

Source: Lee SC, et al. Melioidosis with adrenal gland abscess. Am J Trop Med Hyg 1999;61:34-36.

A 51-year-old diabetic man from taiwan had traveled to Rangoon, Myanmar (formerly Burma), on a four-day fishing and golfing tour. He had been taking oral hypoglycemic agents since the onset of diabetes mellitus three years previously. Upon arriving home in Taiwan, he immediately experienced fever and chills, which were treated with antipyretic agents. Over the next 10 days, left upper quadrant abdominal pain evolved, so he was admitted to the hospital with fever and left upper quadrant tenderness. He was found to be hyperglycemic and glycosuric with a serum blood glucose of 361 mg/dL. A chest radiograph demonstrated bilateral interstitial lung infiltrates. Initially, typhoid fever was suspected, and he improved over eight days while receiving parenteral cotrimoxazole. At that point, blood cultures that had been taken on admission grew Burkholderia pseudomallei, apparently resistant to ampicillin, cotrimoxazole, and all aminoglycoside agents tested, but sensitive to ceftazidime, which was added to his regimen for the next three weeks. Computerized tomographic examination of his abdomen had demonstrated enlargement and inflammation of the left adrenal gland, consistent with adrenal abscess formation. This finding gradually resolved following three weeks of parenteral combination therapy, plus two months of oral cotrimoxazole, and was associated with normalization of afternoon serum cortisol levels, which had initially been depressed.

Lee and colleagues can only speculate about this patient’s potential sources of infection, which included minor abrasions from walking barefoot around temple areas and inhalation or swallowing of the causative organisms during golfing or fishing activities. Lee et al also note infrequent documentation of adrenal gland suppurative processes previously during documented melioidosis, and the now dominant role for the third generation cephalosporin, ceftazidime, which has become the drug of choice for melioidosis—when both available and affordable. Initial clinical response, in this case, did not correlate with the isolated organism’s apparent lack of sensitivity to cotrimoxazole. Lee et al were also well aware of an apparent association between melioidosis and diabetes in both their case and in prior published reports. What is the actual nature of that unusual clinical association?

Comment by Frank J. Bia, MD, MPH

We owe the initial description of Bacillus mallei to a British Army pathologist, Captain Alfred Whitmore, and his assistant, CS Krishnaswami, working in Rangoon, Burma, early in the 20th century. Glanders is a largely veterinary disease in horses caused by Burkholderia mallei. In 1911, while both men were carrying out autopsy studies in the Rangoon General Hospital, they recorded a hitherto-undescribed glanders-like illness’ among the poor, and particularly among the emaciated morphine addicts of Rangoon. Because these infections were so common among opiate addicts, it was first known as morphine injectors septicemia’ caused by Whitmore’s bacillus.1 The organism gained additional notoriety during both the French and American involvement in Indo-China such that by 1973, more than 340 cases had been described among American soldiers fighting in Vietnam, with helicopter pilots found to be at greater risk of infection from soil blown around by rotor blades. The geographic extent of disease distribution was later found to reach the northern rice-growing regions of Thailand, Australia, and even the Caribbean basin. It continues to expand.

The reviewed case report describes bacteremic pulmonary melioidosis in a tourist from Taiwan who acquired his infection during a rather short four-day excursion in Myanmar, the region where this disease was first documented by Whitmore and associates. This patient was diabetic, and at least one authority estimates "the association with diabetes mellitus is particularly strong, and may increase the relative risk of infection by up to 100-fold."2 A report such as this may be of particular importance to those counseling diabetic travelers to endemic regions.2-5 Yet, what travelers are actually at risk? Why Burkholderia pseudomallei? What are the predisposing factors for diabetic patients? What types of diabetes predispose to potentially severe clinical episodes of melioidosis?

The literature has been less than clear on these issues in the past.4,5 Several thousand cases of melioidosis occur in Thailand each year, with additional cases reported from tropical northern Australia, Malaysia, India, China, and some Pacific islands. The distribution of this organism and reported cases is now far wider and includes parts of Africa, Central and South America, and the Indian subcontinent.

Burkholderia (formerly Pseudomonas) pseudomallei can be isolated from both soil and water, and it is particularly prevalent in rice paddies during the rainy season. Infections seem to be acquired by several routes including inoculation and inhalation. Animal-to-animal or human-to-human transmission are probably exceedingly rare events; therefore, melioidosis is a seasonal disease with a majority of cases occurring during the rainy seasons in those who are regularly in contact with soil and water. Males predominate and the peak age incidence ranges from ages 40 to 60.

In endemic regions, melioidosis accounts for an unusually high proportion of community-acquired sepsis, being the most common source of fatal community- acquired sepsis in the northern territories of Australia. Most patients have an underlying metabolic or disease process such as alcohol abuse, an immunosuppressive disorder, diabetes mellitus, renal disease, liver disease, or pregnancy, although melioidosis does not appear to be an AIDS-associated opportunistic infection. Melioidosis may be localized or disseminated and might only present clinically after many years of bacterial latency.

The majority of culture-documented cases of melioidosis demonstrate bacteremia (60%), but only about 50% demonstrate a clear primary focus of infection. Metastatic abscesses may be found in the lungs, liver, or spleen. Septic shock associated with this organism has a 95% mortality rate. A majority of such patients have abnormal chest radiographs and about 10-20% show cutaneous pustules or subcutaneous abscesses. Cerebral abscesses and aseptic meningitis may also occur. Localized pulmonary melioidosis has been notorious for its being confused with tuberculosis due to upper lobe involvement, but generally without hilar adenopathy and with apical sparing. Thai children with melioidosis had acute suppurative parotitis in a third of cases, and can present with parotid abscess or facial nerve palsy involving the auditory canal.2 There is no asymptomatic carriage of this organism, although it does have the potential for long periods of latency. One intriguing question is what actually constitutes the association of melioidosis with diabetes mellitus?

The answer may lie in better defining the group of diabetics at risk, and in appreciating the unique interactions that occur between B. pseudomallei and insulin. Evidence now indicates that serum insulin may modulate the pathogenesis of B. pseudomallei septicemia. In a series of revealing in vitro and in vivo experiments, Woods and colleagues addressed the diabetes question experimentally and elegantly demonstrated that human insulin markedly both binds to and inhibits the growth of B. pseudomallei. This occurs at serum insulin levels that are within expected physiological ranges. In synthetic media containing recombinant human insulin, the growth of B. pseudomallei was significantly less than in media without insulin. The growth rate of B. pseudomallei was also shown to be more rapid in human serum, which had been depleted of insulin, than it was in control serum. Woods et al showed that recombinant human insulin also binds to B. pseudomallei and reduces its growth rate. (Whether the rapidly acting insulin lispro, or Humalogue, has a similar or different effect has not yet, to my knowledge, been investigated.)

Organisms also grew faster in sera of experimental diabetic rats that were insulin deficient, than they grew in control sera from normal animals. When rats were made diabetic by chemical obliteration of their insulin-secreting beta islet cells, causing experimental type 1 diabetes, they experienced more severe B. pseudomallei infection than did normal control rats. There were higher mean bacterial counts in the lungs of diabetic rats, the majority of which were septicemic, whereas none of the control nondiabetic animals were found to be bacteremic.

These studies are convincing and lend experimental data to support the clinical association of melioidosis with type 1 diabetes, an association now based upon the presence of a specific bacterial insulin receptor, that when bound to insulin, appears to diminish both the growth and virulence of B. pseudomallei. However, this does not explain why type 2 diabetics are also at risk.

Travel medicine consultants should be aware that all diabetics (type-1 and type-2) with the appropriate exposure to infected soil and water, are clearly predisposed to severe melioidosis when physiologic insulin levels are not maintained. Given both this predisposition and the short incubation period for melioidosis, as illustrated by a Taiwanese man who returned home ill after a short visit to Mynamar, consultants should be so informed. When apparently unexplained sepsis follows travel to a location endemic for melioidosis and the presentation suggests typhoid fever, abdominal sepsis, or unexplained upper lobe pneumonia potentially accompanied by cutaneous pustules, the diagnosis may be melioidosis. In addition, a lack of clinical responsiveness to empiric therapy with the fluoroquinolones, ampicillin, penicillin, or the aminoglycosides should also suggest this disease. The current antimicrobial agent of choice appears to be the third generation cephalosporin, ceftazidime (120 mg/kg/d), administered for at least 2-4 weeks. Imipenem (50 mg/kg/d) now appears to be a safe, reasonable, and expensive alternative to ceftazidime. Although type 1 diabetes and insulin deficiency do not fully explain the predisposition of all diabetics for melioidosis, human insulin levels do appear to play a unique role in modulating the pathogenesis of infection and septicemia associated with melioidosis.

Melioidosis in 1999

Several other new and interesting observations in the field of melioidosis were published last year. They include some of the following developments. Dance and associates noted an increase in reported culture-positive cases diagnosed within both England and Wales, with six newly diagnosed cases appearing in 1998 alone, compared with nine over the previous 10 years.8 The appearance of these new cases, originating in Bangladesh, India, and Pakistan, suggests far more melioidosis is present on the Indian subcontinent than was originally thought.

Although not difficult to identify upon isolation, using commercially available laboratory methodology, a new latex agglutination format for rapidly identifying gram-negative bacteria obtained in blood cultures9 as B. pseudomallei may prove useful in preventing critical delays in the isolation, identification, and treatment of melioidosis with septicemia. The propensity for abscess formation during melioidosis should not be underestimated, since visceral abscess formation, and not simply antibiotic resistance, may also account for the common appearance of relapses in the course of this disease.

Munns and Thompson reported two cases of prostatic abscess accounting for relapses, septicemia, and urinary retention.10 However, during a symposium presentation at the 1999 meeting of the Infectious Disease Society of America, reference was made to an additional 33 cases of prostatic abscess occurring among the 400,000 aborigines, who constitute approximately 2% of the total Australian population of 19 million people. Of note, diabetes is common in this population. It is predominantly type 2, which is not caused by insulin deficiency, but is instead associated with insulin resistance. (S94: Currie B. Infectious diseases of aboriginal people in Australia and New Guinea, session 58—Symposium on Infectious Diseases of Australia).

Besides the diabetes issue, travel medicine practitioners should be aware that the risk areas for the acquisition of melioidosis are increasingly clear and variable based upon isolation rates of the causative organism from soil.11,12 It should be possible to review such emerging epidemiological data and advise travelers—particularly diabetic patients—of the risks of acquiring melioidosis during their travel to endemic regions as this information is further refined.


1. Dance DAB, White N. Melioidosis. In: Illustrated History of Tropical Diseases. FEG Cox, ed. London, England: The Wellcome Trust Publishing Co; 1996:72-81.

2. Dance DAB. Melioidosis. In: Tropical Infectious Diseases. Principles, Pathogens and Practice. Guerrant RL, Walker DH, Weller PF, eds. Philadelphia, PA: Churchill Livingston; 1999:430-437.

3. Merianos A, et al. The 1990-1991 outbreak of melioidosis in the Northern Territory of Australia: Epidemiology and environmental studies. Southeast Asian J Trop Med Public Health 1993;24:425-435.

4. Elango S, et al. Parapharyngeal space melioidosis in a diabetic. J Laryngol Otology 1991;105:582-583.

5. Turner MO, et al. Melioidosis in a diabetic sailor. Chest 1994;106:952-954.

6. Woods DE, et al. Interaction of insulin with Pseudomonas pseudomallei. Infect Immun 1993;61: 4045-4050.

7. Simpson AJ, et al. Comparison of imipenem and ceftazidime as therapy for severe melioidosis. Clin Infect Dis 1999;29:381-387.

8. Dance DAB, et al. Imported melioidosis in England and Wales. Lancet 1999;353:208.

9. Dharakuhl T, et al. Rapid identification of Burkholderia pseudomallei in blood cultures by latex agglutination using lipopolysaccharide-specific monoclonal antibody. Am J Trop Med Hyg 1999;61:658-662.

10. Munns J, Thompson L. Melioidosis of the prostate. Aust NZ J Surg 1999;69:154-156.

11. Parry CM, et al. Melioidosis in southern Vietnam: Clinical surveillance and environmental sampling. Clin Infect Dis 1999;29:1323-1326.

12. Vuddhakul V, et al. Epidemiology of Burkholderia pseudomallei in Thailand. Am J Trop Med Hyg 1999;60:458-461.

The pathogenesis of melioidosis, caused by B. pseudomallei, involves each of the following clinical features except:

a. the relatively long incubation period, which is measured in weeks between exposure, infection, and clinical symptoms.

b. the potential for reactivation of latent bacterial infection after many years.

c. a predisposition toward parotid gland abscesses to form in some Thai children.

d. a higher risk of serious infections associated with B. pseudomallei among diabetics.

e. insensitivity of B. pseudomallei to fluoroquinolone antibiotics, such as ciprofloxacin.