Mucocutaneous and Soft Tissue Infections Caused by Xanthomonas maltophilia: A New Spectrum

Methods

Patients

We reviewed the records of the clinical microbiology laboratory of The University of Texas M.D. Anderson Cancer Center from October 1991 to December 1992 for all cultures positive for X. maltophilia. We also reviewed the medical records of all patients with positive cultures from mucocutaneous surfaces and soft tissues.

Routine microbiological methods were used. Xanthomonas maltophilia was identified by the Vitek-AMS (BioMerieux; Hazelwood, Missouri) or the API-NFT (BioMerieux) recognition systems.

Definitions

We defined infection as positive cultures obtained from a site that manifested clinical signs of infection. No evidence could suggest other pathogens, including herpesvirus, isolated from the same site. We excluded patients who had phlebitis but had no signs of cellulitis. Patients who were infected with X. maltophilia that was isolated from the study sites but who did not meet the inclusion criteria were considered to be colonized and were thus excluded.

We used definitions suggested by the Food and Drug Administration and Infectious Diseases Society of America for skin-skin structure infections [22]. Briefly, cellulitis was defined as warmth, erythema, and induration of skin or subcutaneous tissue or both, with or without pain. Mucocutaneous infection was defined as an ulcer associated with pain, swelling, and erythema. Response to therapy was defined as disappearance of all signs and symptoms of infection. Neutropenia was defined as a neutrophil count of less than 1000 cells/µL.

Results

Xanthomonas maltophilia was isolated from various sites in 237 patients during the study period, and 114 patients were judged to have true X. maltophilia infections. Sixteen of the 237 patients fulfilled the inclusion criteria of our study. We included one additional patient who only had positive blood cultures for X. maltophilia and multiple nontraumatic subcutaneous nodules with no breakdown of overlying skin. There was no clustering of cases by either time or space. We divided the patients into two groups on the basis of the most likely route of acquisition of the infection: those with hematogenously spread metastatic lesions (6 patients) and those who probably acquired the infection by direct inoculation through the mucocutaneous surfaces (11 patients). This latter group had negative blood cultures and was further subdivided into patients with primary cellulitis (5 patients) and those with mucocutaneous infections (6 patients). The clinicopathologic characteristics of these patients are shown in Table 1.

Metastatic Cellulitis

Five of six patients with metastatic cellulitis had leukemia refractory to treatment; in one patient, malignancy had been recently diagnosed, and the infection developed during the first course of chemotherapy. All patients had received broad-spectrum antibiotics within 10 days before the onset of their infections. In the patients receiving ciprofloxacin and ceftazidime, organisms causing the infections were susceptible to the antibiotics when infection occurred. All patients were hospitalized with severe neutropenia (<100 neutrophils/µL) and fever.

The lesions consisted of hard, nonfluctuant, tender skin or subcutaneous nodules in five patients (Figure 1). The nodules ranged in size from 0.5 × 1 cm to 5 × 6 cm and were characterized by warmth and a violaceous erythema of the overlying skin. In all five patients, tender areas of cellulitis surrounded the nodules or appeared in areas distant from the nodules. Culture of biopsy tissue from one of these distant cellulitic areas showed X. maltophilia. Black central necrosis (2 patients) or ulcerations (2 patients) developed in some of the nodules and cellulitic areas. The skin lesions involved the extremities, scalp, back, and abdomen in decreasing frequency. These lesions increased in size and number in four patients over a median of 2 days. One patient had a nodule (5 × 6 cm) of the right upper arm with surrounding cellulitis (3 cm in width); during the next 10 days, an area of cellulitis over the abdomen developed and increased in size. A second patient had a nodule (2 × 3 cm) of the right thigh with surrounding cellulitis (4 cm in width). Areas of cellulitis developed that increased in size over the next 3 days on the left thigh (as large as 2 × 2 cm), bilateral lower extremities (as large as 2 × 3 cm), back (as large as 3 × 4 cm), and scalp (as large as 2 × 3 cm). A third patient had a nodule of the left shin and cellulitis of the lateral aspect of the left lower extremity. The cellulitis increased in size over the next day to cover the area from the dorsum of the left foot to above the knee. A fourth patient had one nodule of the right lower leg (0.5 × 1 cm) and two nodules of the scalp (0.5 × 1 cm) with cellulitis involving the left lower leg and the left big toe. During the next day, one more nodule of the left lower extremity (1 × 1 cm) and two more nodules of the scalp (1 × 1 cm) developed with increasing cellulitis of both lower extremities. The fifth patient had a single nodule of the right elbow with surrounding cellulitis.

Four of these patients developed septic shock, with confusion (3 patients), multiorgan failure (3 patients), and pneumonia (4 patients). Pathologic examination of biopsy tissue from three of the five patients with nodular lesions showed swelling of the endothelial cells and inflammatory infiltrates of the subcutaneous tissue and dermis (Figure 2). Two patients had necrotizing inflammation of the dermis.

The sixth patient with metastatic cellulitis had multiple erythematous macular lesions involving the left lower extremity that were typical of ecthyma gangrenosum. During the next day, multiple new ecthymatous lesions of both lower extremities, the right upper extremity, and the scalp developed; they ranged in size from 1 × 1 cm to 5 × 6 cm. Septic shock with confusion, multiorgan failure, and pneumonia developed. Examination of biopsy tissue from this patient showed characteristic pathologic changes (consisting of infiltration of the dermal blood vessel walls, perivascular connective tissue, and papillary dermis with gram-negative rods) and fibrin thrombi within superficial and deep vessels with local necrosis of dermis, dermal blood vessels, and eccrine sweat coils (Figure 2).

Two patients with metastatic cellulitis responded to trimethoprim-sulfamethoxazole when they recovered from myelosuppression. The remaining four patients died of active X. maltophilia infection and persistent neutropenia a median of 10 days after the onset of skin lesions.

Mucocutaneous Infections

The mucocutaneous lesions consisted of infected ulcers of the gingiva, lip, and buccal mucosa. Five of the six patients with mucocutaneous infections were febrile and neutropenic. All six were given chemotherapy for cancers that had recurred, and they received one or more broad-spectrum antibiotics within 10 days of the onset of their infections. In those patients receiving β-lactam antibiotics, the X. maltophilia organisms were resistant. On the other hand, the organisms were susceptible to ciprofloxacin in the three patients receiving this antibiotic. Two of the three patients who responded to treatment with trimethoprim-sulfamethoxazole recovered from myelosuppression. The third patient responded despite persistent neutropenia. The remaining three patients died of causes unrelated to their persistent infection. Two of these patients had persistent neutropenia.

Primary Cellulitis

None of the five patients with primary cellulitis had leukemia, and none received antibiotics. Only two were receiving chemotherapy and had neutropenia; four were outpatients. All patients had tender cellulitis without clear demarcation of the borders. The lesion surrounded a neck mass in one patient and produced an exudate at catheter insertion sites in four patients. Response in the three patients receiving antimicrobial treatment, two of whom were neutropenic and recovered from myelosuppression, coincided with the removal of the catheter. Removal of the catheter was the only therapeutic maneuver required to achieve response in the remaining two patients.

Discussion

Several new findings emerged from our study: 1) Mucocutaneous and soft tissue infections caused by X. maltophilia are not uncommon; 2) X. maltophilia can cause metastatic nodular skin lesions that mimic disseminated fungal infections [this type of lesion has not been described previously]; 3) X. maltophilia causes serious morbidity and high mortality in patients with metastatic skin nodules; 4) X. maltophilia can cause superinfections in patients receiving broad-spectrum β-lactam or quinolone antibiotics to which the organisms are susceptible at the time of the infections; and 5) catheter removal contributes to a favorable outcome in patients with catheter-associated X. maltophilia cellulitis without bacteremia. Xanthomonas maltophilia should be added to the differential diagnosis of mucocutaneous or soft tissue infection in patients with cancer. Trimethoprim-sulfamethoxazole with or without ticarcillin-clavulanate is the current treatment of choice for culture-proven infections, but early empiric therapy as a result of a high index of suspicion may improve outcome. In addition, metastatic skin infections and, to a certain degree, mucocutaneous infections in neutropenic patients with cancer seem to be poor prognostic signs because many of these patients died of their infections and of causes that were probably secondary to their severe immunosuppression.

We identified X. maltophilia involvement of the skin and soft tissues in 15% (17 of 114) of patients with X. maltophilia infections. The infection usually developed in chronically debilitated patients with prolonged hospitalization who were receiving chemotherapy, usually for advanced, recurrent hematologic malignancies, and who were receiving broad-spectrum antibiotics. In one case, however, the infection developed early in the hospitalization. In addition, most patients with primary cellulitis were outpatients who were not receiving antimicrobial therapy. Because X. maltophilia is considered to be a nosocomially acquired pathogen [1-6, 8], these facts may indicate that it can colonize inpatients rapidly and for prolonged periods of time.

In our study, we described three different clinical entities of mucocutaneous and soft tissue infections caused by X. maltophilia on the basis of their probable route of acquisition. The nodular lesions caused by X. maltophilia, which constituted all but one of the metastatic infections in our series, have never been described previously. The development of these nodules was associated with serious morbidity and high mortality. In addition, the nodular lesions were similar to those seen in disseminated fungal infections. The only previously described metastatic skin and soft tissue infections caused by X. maltophilia have consisted of ecthyma gangrenosum in five patients and gangrenous cellulitis in one patient [3, 4, 19-21]. The nodular lesions in our patients differed from ecthyma gangrenosum by their striking nodularity, the absence of bulla or vesicle formation, the associated pain and tenderness, and the tender areas of cellulitis distant from the nodules. In addition, necrosis was not universally present. Histologically, the differences were significant for the absence of microbial invasion and for the necrosis of vessel walls and vascular thrombosis. Subcutaneous nodules have been reported in Pseudomonas aeruginosa infections [23-30]. However, most of the nodules caused by P. aeruginosa were deep-seated abscesses that were drained and that often healed with scarring. Histopathologic examination of one of these pseudomonal nodules showed dense infiltration by neutrophils of the subcutis, with a microscopic diagnosis of suppurative panniculitis [27]. It is unclear whether any of the nodular lesions in our patients would have formed abscesses given the setting of neutropenia; however, two patients recovered from myelosuppression without evidence of suppuration. Localized wound infections with X. maltophilia have been reported in 52 patients [1, 5, 9-18, 31]. The diagnostic criteria used, however, and the etiologic role of the organism were not clarified in 32 of these patients [5, 14, 17, 31]. In one study, X. maltophilia was recovered from several malignant lesions [31]. Primary cellulitis has been described in seven patients [3, 16], umbilical cellulitis has been reported to be associated with the application of contaminated disinfectant solution in two neonates [16], and central venous catheter-associated cellulitis has been reported in five patients with cancer [3]. However, all five patients had concomitant bacteremia.

Superinfections with X. maltophilia have been associated almost exclusively with imipenem-cilastatin [1]. We found no preponderance of infected patients receiving this antibiotic despite its widespread use in our hospital. Many patients in our study developed the infection around the time of administration of a quinolone or another broad-spectrum β-lactam or even aminoglycoside therapy. It is interesting that in most cases, the organisms were sensitive to the antibiotics that these patients were receiving.

The presence of neutropenia seemed to particularly predispose patients to developing metastatic and mucocutaneous infections. Therefore, it is not surprising that recovery from myelosuppression played a major role in the response to therapy in these cases; trimethoprim-sulfamethoxazole with or without ticarcillin-clavulanate was the antibiotic combination most consistently associated with a favorable outcome in concordance with in vitro findings [32]. Trimethoprim-sulfamethoxazole plus ticarcillin-clavulanate or moxalactam and single-agent moxalactam have been reported to result in a good outcome in treating patients with skin lesions caused by X. maltophilia[4, 19, 20]. However, information relative to recovery from neutropenia was not readily available in most patients [3, 20, 21]. Removal of catheters with or without systemic therapy seemed to be associated with a favorable outcome in patients with cellulitis and catheter-associated X. maltophilia bacteremia [3]. In our study, patients responded to catheter removal with or without antimicrobial therapy for primary cellulitis in the absence of bacteremia.

In conclusion, X. maltophilia infection should be included in the differential diagnosis of any mucocutaneous or soft tissue infection in patients with cancer, particularly if the patient is severely neutropenic and febrile and is receiving broad-spectrum antibiotics. Nodular skin lesions seen in such patients are similar to systemic fungal infections. Because disseminated X. maltophilia and fungal infections develop in the same patient population, awareness of the role of X. maltophilia in causing these infections has important therapeutic implications. Maintaining a high index of suspicion is particularly important given the serious morbidity and high mortality associated with the disseminated X. maltophilia infections found in our study and the paucity of therapeutic alternatives against these infections. Early empiric therapy using trimethoprim-sulfamethoxazole with or without ticarcillin-clavulanate may improve the outcome in patients with such disseminated X. maltophilia infections.