Fatal urosepsis due to delayed diagnosis of genitourinary melioidosis¹⁾

Eight weeks prior to being admitted as an inpatient, the 54-year-old male patient had taken a four-week trip to Thailand during the rainy season, where he also bathed in a rice pond. Two weeks after his return, he presented himself to a local urologist with urinary retention and painful urination. After catheterization, he was given a low oral dose of doxycycline (100 mg, daily), being suspected of Chlamydia trachomatis urethritis due to positive serology. With persistent fever and arthralgia, productive cough and “neurological problems”, the patient was admitted to a regional hospital’s inpatient department. His case history showed diabetes mellitus type 2, a condition following a unilateral orchidectomy five years prior to the admission, and alcohol abuse with five to six glasses of beer a day.

The physical examination revealed a reduced general condition at a temperature of 40°C and a pulse of 100/min. as well as pale skin color. In addition, the physical examination findings, including auscultatoric, percutatory and palpatory examinations of the thorax, abdomen and lymph nodes, were inconspicuous. Exploratory examinations did not identify any focal-neurological event.

A chest X-ray showed streaky bronchopneumonic consolidation, radiating from the left hilum to the left hemithorax; an abdominal ultrasonography revealed hepatosplenomegaly.

The blood test done upon admission showed inconspicuous parameters [e.g., 6.2×10⁶/mL leukocytes (4.5–11.5)], with the exception of an erythrocyte sedimentation of 84/90 mm; a CRP test was not done. Aside from leukocyturia, microscopic Gram-negative, rod-shaped bacteria were detected in the urine. The isolate was identified at the local laboratory as Burkholderia cepacia and found to be fully sensitive to ciprofloxacin.

With a suspected diagnosis of pneumonia and urinary tract infection, the antibiotic treatment was empirically changed from doxycycline to cefotiam i.v. (2 g 1-0-1) and gentamicin i.v. (80 mg 1-0-1, with monitoring of the level). Upon receipt of the urine analysis, the antibiotic treatment was adjusted to ciprofloxacin (250 mg 1-0-1) based on the antibiogram. With this treatment in place, the patient’s fever dropped, and he seemed to recover clinically.

A few days later, he deteriorated clinically: his fever returned, and he presented leukocytosis (19.9×10⁶/mL), thrombocytosis (830×10⁶/mL) and anemia (Hb 8.5 g/dL). The C-reactive protein was given semiquantitatively as significantly elevated (+++) for this measurement. Malaria was ruled out by way of a “thick blood smear” and blood smear. The CT scan of the thorax showed bilateral multiple lung abscesses or, as a differential diagnosis, septic emboli; at this point, the antibiotic treatment was recalculated and changed to imipenem/cilastatin (500 mg 1-1-1).

Due to the development of an acute respiratory distress syndrome, the patient was moved to the intensive care unit of a university hospital. At the hospital, it was possible to isolate Gram-negative, rod-shaped bacteria from the blood culture, sputum and pus, which was identified, biochemically (Api 20 NE bioMérieux, Nürtingen, Germany) and by means of 16S-rRNA gene sequencing, as B. pseudomallei (sequence identical to B. pseudomallei strain 1026b (Accession: U91839.1) via 1,488 base pairs). The isolate was sensitive to amoxicillin/clavulanic acid, imipenem, ceftazidime and ciprofloxacin, but the minimum inhibitory concentration of ciprofloxacin at 0.75 μg/mL was close to the resistance limit (1 μg/mL).

Under intensive therapy, the patient died three days after being admitted to the university hospital (about at the same time that the pathogen was identified), from septic shock and respiratory failure. The autopsy found multiple abscesses of the lungs, spleen, kidneys, prostate, groin and paravertebral block (in the latter case, affecting the TH 6/7 segment). Despite formalin fixation and waxing, it was still possible to detect in the autopsy tissue B. pseudomallei from the lung, adrenal gland, spleen, residual testis, prostate, heart, intercostal space and paravertebral abscesses by means of rpsU and fliC-PCR [1]. The species identity was confirmed by sequence analysis, which also revealed sequence identity with the B. pseudomallei strain 1026b (Accession: U73848.1). In very densely infected prostate abscesses (Figure 1), the pathogen was detected by way of fluorescence in situ hybridization (FISH) [2] (Figure 2).

Figure 1

Granulomatous infiltrate in prostate tissue (hematoxylin-eosin, 100x).

Figure 2

B. pseudomallei cluster in infected cells in prostate tissue in FISH (fluorescence overlay, 630x).

Blue: DAPI-stained DNA of the nuclei. Yellow (e.g., in the white circle): B. pseudomallei in the superposition of the green fluorescent eubacterial-specific probe and the red-fluorescent B. mallei/pseudomallei probe.

Retrospectively, in case of initially leading urological symptoms, one must assume a genitourinary focus that formed through neighboring tissue (per continuitatem) via the urinary tract, for example, in the prostate from which B. pseudomallei spread systemically through the blood stream. Burkholderia (B.) pseudomallei, the pathogen of melioidosis, is a risk level 3 bacterium predominantly indigenous to Southeast Asia and northern Australia [3]. The primary manifestation of melioidosis, at 50%, is pneumonia; in another 8% of cases, as observed by us, secondary pneumonia occurs with an alternatively localized primary focus [4]. Severe genitourinary infections have been described, in particular, septic prostatitis and prostatic abscesses [5–7]. Risk factors include diabetes mellitus, and alcoholism [5, 7]. When coming into contact with fresh water in Thailand, as happened in this case, an ascending urinary tract infection caused by the pathogen is plausible, because B. pseudomallei occurs endemically in rice paddies in Southeast Asia [8]. As a result of the initial overvaluation of the C. trachomatis serology, whose diagnostic value must generally be questioned critically due to its regularly minor significance with respect to the time of infection and activity of the infection [9], an advanced pathogen diagnosis and thus the diagnosis itself were delayed.

Due to the rarity of genitourinary B. pseudomallei infection and the resulting lack of experience, delays in diagnosis or misdiagnoses will occur, despite an indicative constellation of findings. The most common Gram-negative, nonfermenting rod-shaped bacterium that is to be considered in terms of the differential diagnosis, causes severe urinary tract infections, and can be confused in terms of the culture morphology with B. pseudomallei is Pseudomonas aeruginosa. Burkholderia pseudomallei is much rarer; however, as described here, it can result in serious or lethal outcomes.

In the case presented by us, too, it is very likely that a misdiagnosis in connection with the initial species identification of the pathogen isolate delayed a potentially life-saving diagnosis. Although the initial urinary tract isolate was no longer available for re-identification, it apparently involved a misidentification of B. pseudomallei as B. cepacia, a risk level 2 pathogen that occurs more commonly with respiratory infections in cystic fibrosis patients. Unfortunately, the unusual result of differentiation did not trigger subsequent testing. Given the known limited reliability of biochemical identification systems [10–12] for species identification within the Burkholderia genus in general and given the lack of experience – due to how extremely rare it is – with the identification of Burkholderia pseudomallei isolates in particular, further test methods should be used in case of Burkholderia isolates from patients with no known cystic fibrosis and indicative case history so as to rule out B. pseudomallei.

Therefore, it is important to take into account the rare but nonetheless life-threatening differential diagnosis of melioidosis in the differential diagnostic consideration. Despite existing risk factors and a stay in the Southeast Asian tropics, the differential diagnosis of a genitourinary B. pseudomallei infection was considered too late in this case, as had also previously been observed impressively with tsunami victims [13]. In cases of a suspected B. pseudomallei infection, species identification can be realized using one of several molecular methods described [14–16], of which 16S rRNA gene sequencing [15] has meanwhile become relatively widely adopted in routine diagnostics. MALDI-TOF-MS (Matrix-Assisted-Laser-Desorption-/Ionization-Time-of-Flight-Mass-Spectrometry) has, in the recent past, been evaluated as a new and innovative method for identifying B. pseudomallei [17], after previous studies had already demonstrated promising results in terms of the MALDI-TOF-MS-based identification of pathogens of the B. cepacia complex [18–20].

In addition to the diagnostic difficulties that delayed the identification of B. pseudomallei, this case serves as an example that proves the risk related to the persistence of nonfermenting, Gram-negative rod-shaped bacteria in connection with low-dose antibiotic therapy. The ciprofloxacin dose of 250 mg 1-0-1, which is at the lower end of the spectrum, led to an initial clinical improvement by eradicating the formally still sensitive pathogen from the body fluids, but also caused pathogen persistence in the tissue abscesses.

In cases of severe, acute melioidosis, the patient should be given – over several weeks! – intravenous ceftazidime + trimethoprim-sulfamethoxazole or meropenem [21, 22], as recommended by an Australian-Brazilian team of authors in a guideline based on the evaluation of various therapeutic studies [21]. Following the acute treatment, the authors recommend eradication therapy over 12–20 weeks, which can be performed with different regimes consisting of trimethoprim-sulfamethoxazole, doxycycline or chloramphenicol (the latter is not recommended in Germany due to its severe unwanted side effects) [21]. More recent recommendations have been clearly in favor of trimethoprim-sulfamethoxazole – in case of resistance, an amoxicillin-clavulanic acid combination, which is less effective, may be considered as a substitute. In the event of an inadequate response to initial therapy with ceftazidime, treatment should be switched to meropenem, because carbapenems, at least in vitro, lead to a more rapid eradication [23]. In the case described here, the switch to imipenem/cilastatin came too late to save the patient’s life.

As a conclusion for diagnostic practice, one should consider the rare differential diagnosis of an abscess-forming B. pseudomallei infection in connection with refractory urinary tract infections caused by nonfermenting, Gram-negative rod-shaped bacteria following a stay in the tropics of Southeast Asia, particularly if risk factors, such as diabetes mellitus and alcoholism, exist. Molecular methods [1, 2, 13–15] or MALDI-TOF-MS [17] should be used to confirm the diagnosis; PCR-based techniques [1, 13–15] can be considered also for primary material [2]. In contrast, the simple and inexpensive FISH method, with the exception of very high bacterial densities, as observed in our case in the prostate tissue of the patient, should be used regularly only for the species identification of cultural material. A positive Chlamydia serology, which has limited diagnostic value per se, should not be allowed to delay further diagnostics. In case of severe infections with deep tissue involvement and incipient abscess formation, antibiotic doses should be set, in consultation with clinical colleagues, at the higher end of the spectrum in order to prevent pathogen resistance and the concomitant resurgence of the infection, which would quickly become unmanageable. International recommendations [21] should be considered.