Chromobacterium violaceum sepsis, a rarely reported phenomenon, has a high mortality rate. We report a unique case of C. violaceum sepsis in an infant. A 4-month-old girl presented to our institution with fever, pustular skin lesions, and distended abdomen, as well as diminished activity and mental status. Radiological investigation revealed brain, lung, and hepatic abscesses. The infant was successfully treated with trimethoprim-sulfamethoxazole and ciprofloxacin.
Chromobacterium violaceum, a saprophytic bacterium, was first discovered in water buffalo by Wooley in 1905 . It is unique to tropical and subtropical climates and is found between latitudes 35°N and 35°S. This includes Malaysia, where this facultatively anaerobic gram-negative rod bacterium was first seen in humans in 1927 . Sepsis with this organism is often found with concomitant immune deficits, most frequently chronic granulomatous disease (CGD) . There have been fewer than 40 cases of human infection reported, and several have come from the southeastern United States, primarily Florida. We review the treatment of C. violaceum sepsis in what is, to our knowledge, the youngest child yet described with this disease.
Case report. A previously healthy 4-month-old girl from rural southeast Georgia was transferred to our institution. Two weeks before admission, the infant was diagnosed with fever and an upper respiratory tract infection. After 2 separate local hospital admissions, the infant was transferred with unresolved fever, decreased activity and mental status, impending respiratory failure, abdominal distention, and anemia.
At admission, the patient was febrile, with a temperature of 38.9°C (102°F). Her other vital signs included a heart rate of 180 beats/min, a respiratory rate of 80 breaths/min, and a blood pressure of 115/49 mm Hg. The child was lethargic and displayed a diminished response to painful stimuli. Generalized facial edema was noted, and pulmonary auscultation revealed diffuse rales bilaterally. Also of note was a soft, distended abdomen with cyanotic areas around the umbilicus. The liver could be percussed and palpated 3 cm below the right costal margin, but the spleen was not palpable and no fluid wave was appreciated. The skin had multiple small pustular lesions with an erythematous base over the abdomen and both upper and lower extremities (figure 1). Petechiae and evidence of thrombophlebitis were absent.
The patient's initial peripheral WBC count of 37,000 cells/mm3 increased after 24 h to 42,000 cells/mm3 with 85% neutrophils, 11% lymphocytes, and 2% band forms. A complete blood count revealed microcytic anemia (hemoglobin level, 7.0 g/dL; hematocrit, 21%; and platelet count, 77,000 cells/mm3). Electrolyte levels remained within normal limits for the duration of the illness. A liver profile was obtained and revealed the following levels: total bilirubin, 1.3 mg/dL; direct bilirubin, 0.7 mg/dL; alkaline phosphatase, 122 U/L; aspartate aminotransferase, 53 U/L; alanine aminotransferase, 19 U/L; and albumin 1.4 g/dL. Evaluation of the CSF revealed the following levels: WBC count, 9 cells/mm3 (3% polymorphonuclear leukocytes, 89% lymphocytes, and 8% monocytes/macrophages); glucose, 50 mg/dL (serum glucose, 73 mg/dL); and protein, 127 mg/dL. Urine and stool culture results were negative. Analysis of blood cultures yielded a gram-negative rod bacterium that was later identified as C. violaceum, as did cultures of pustule and sputum specimens. Imaging studies included a radiogram and CT of the chest that showed probable pneumatoceles and abscess formation. CT of the abdomen and brain revealed multiple hepatic and brain abscesses (figures 2 and 3).
During the initial 2-day hospitalization by her primary-care physician, the patient was treated with iv ceftriaxone. After discharge, the patient continued to be treated at home with amoxicillin and acetaminophen. However, she remained febrile and was readmitted to hospital 3 days later, where she was again treated with ceftriaxone. Blood drawn during the second admission was cultured and grew a gram-negative rod bacterium after 48 h. When the patient was admitted to our institution 3 days later, she was initially treated with a regimen that included vancomycin, gentamicin, and metronidazole. When C. violaceum was identified, piperacillin-tazobactam and trimethoprim-sulfamethoxazole (TMP-SMX) were empirically added, replacing vancomycin and metronidazole. Once bacterial sensitivities returned, gentamicin was discontinued because of in vitro resistance, and ciprofloxacin was substituted for piperacillin-tazobactam.
The sensitivity profile of the C. violaceum isolate revealed in vitro resistance to cephalosporins and other β-lactam compounds, excluding aztreonam (MIC <8 µg/mL) and imipenem (MIC <1 µg/mL). However, the organism was susceptible to the fluoroquinolones ciprofloxacin (MIC <1 µg/mL) and levofloxacin (MIC <2 µg/mL), along with trimethoprim (MIC <2 µg/mL). It was only intermediately susceptible to gentamicin (MIC 4 µg/mL) and tetracycline (MIC <4 µg/mL). Because of the rarity of patients with C. violaceum infection, it is unclear how in vitro breakpoints translate to in vivo response.
Because of the nature of the patient's infection, studies to rule out an immunodeficiency state were also deemed necessary. Laboratory investigation, including the use of repeat nitro blue tetrazolium testing, revealed no evidence of CGD, polymorphonuclear leukocyte glucose-6-phosphate-dehydrogenase deficiency (G6PD), or any other immunodeficiency, including HIV infection.
During her hospital course, the patient developed disseminated intravascular coagulation, acute renal failure, and acute respiratory distress syndrome. She also required vasopressors as well as mechanical ventilation before she began to recover. After discharge, she continued a regimen of first iv, then oral TMP-SMX, at appropriate doses, for 6 weeks. Both hepatic and brain abscesses regressed with treatment. Despite the use of ciprofloxacin, there was no evidence of joint pathology or other complications.
Discussion. C. violaceum is a long gram-negative bacillus that is motile and facultatively anaerobic. Medium containing tryptophan is required to grow the organism, which is usually detectable after 18–24 h. When cultured, the organism produces large, smooth, convex colonies that are a violet-black color. The organism's appellation derives from its striking purple pigment, violacein, which is soluble in alcohol but insoluble in water. These pigmented strains are identified by their fermentative degradation of carbohydrates . Nonpigmented strains do exist; however, they have infected humans only twice [5, 6]. Consistent findings of C. violaceum were found in 4 different cultures of specimens from sterile sites in this patient.
The incidence of C. violaceum sepsis is low, and documented human infection is rare. Most cases have occurred during the summer, and they have been reported from such disparate locales as Australia, southeast Asia, India, Argentina, and the southeastern United States. They have shown similarities in both presentation and severity of illness [3, 6–13]. Mortality rates >60% have been reported [7, 14]. The constellation of findings attributed to C. violaceum sepsis has most commonly included fever, hepatic abscesses, and skin lesions. Orbital and periorbital cellulitis, osteomyelitis, and meningitis are less frequently reported [3, 15, 16]. Brain abscesses secondary to C. violaceum sepsis have been reported, but only once, and the patient did not survive . This patient, who presented in early autumn, survived C. violaceum sepsis despite multiple brain abscesses as well as the more common complications of fever, hepatic and lung abscesses, and skin lesions.
The patient we studied is similar to others in that the infant lived in the southeastern United States, specifically rural southeastern Georgia. C. violaceum has a wide distribution in soil and water in tropical and subtropical areas. Ponte and Jenkins  posited that inoculation may occur when injured or broken skin is exposed to stagnant water. Although Macher and others  and Mamlok et al.  have found C. violaceum to be more common in patients with CGD and polymorphonuclear leukocyte G6PD, our patients had neither of those conditions. In previous cases of C. violaceum sepsis, individuals have ranged in age from a 20-month-old aboriginal Australian boy to a 56-year-old man from Cairns, Australia [11, 19]. The 4-month-old infant we describe had no history of exposure to still water or mucocutaneous infection.
Because of the rarity of C. violaceum sepsis, treatment has not been conclusively defined. C. violaceum is known to be generally resistant to cephalosporins and penicillins [17, 20]. Indeed, our patient received both ceftriaxone and ampicillin without improvement of symptoms. Significant improvement came when the sensitivities were known and the patient was started on ciprofloxacin and TMP-SMX. This outcome is consistent with the extensive in vitro studies of Aldridge et al. , which found ciprofloxacin to be the most active compound against C. violaceum.
The virulence of C. violaceum is attributed partly to endotoxin and to inadequate host defense. The host defense against the organism depends on adequate levels of superoxide dismutase and catalase . This explains the susceptibility of patients with CGD and polymorphonuclear leukocyte G6PD deficiency to C. violaceum infection. If C. violaceum infection is strongly suspected, therapy with TMP-SMX and a fluoroquinolone should be started promptly. Long-term treatment is needed to fully eradicate the organism and resolve potentially fatal abscesses in the brain and other viscera.
This case illustrates several important aspects of both the presentation and treatment of C. violaceum. It is unusual because of the infant's survival, her young age, the lack of stagnant water exposure, the presence of brain abscesses, and the absence of a known immunodeficiency. It is important for physicians in tropical and subtropical regions to consider this infection as part of the differential diagnosis of sepsis with gram-negative rod bacteremia, especially with a history of exposure to stagnant water.
Analysis of in vitro data suggests that of available antibiotics, fluoroquinolones are the most active against C. violaceum. Our patient survived a usually fatal infection after being treated with an antibiotic regimen that included ciprofloxacin. Ciprofloxacin has the most gram-negative coverage of the quinolones presently available; however, any quinolone with good coverage could be considered. In this patient, a carbapenem would be an option within the treatment regimen, and meropenem would also be a good option because it penetrates the CNS. Taken together, these facts suggest that clinicians faced with patients critically ill from C. violaceum sepsis should consider adding a quinolone to the antibiotic regimen, even for infants.