RESULTS
During the study period 38 cases of EG (male 16, female 22) were recorded. Demographic and clinical characteristics of the patients are presented in table 1.
Figure 1 shows the age distribution of patients, according to gender. Sexteen out of 38 (42.1%) patients were in the age 0-3 years, 13 of them (13/16, 81.2%) were females, 5 were infant, 17/38 patients (44.7%) were above the age of 10.
Hematological disease was present in 35/38 (92.1%) patients (table 1). Five out of 38 (13.1%) patients were under treatment for relapsed disease (4/29 for acute leukemia, 1 for Ewing sarcoma).
The infection occurred after an average time of 5.4 months (median 4.8 months, range 0.07-21.8 months) from the diagnosis of the underlying disease or the relapse. Sixteen out of 38 (42.1%) patients were in the induction phase of treatment, 8 in the consolidation phase, 8 in the re-induction phase, 1 patient in the maintenance phase, 2 in the salvage treatment, 2 patients had received allogeneic stem cell transplantation, 1 patient had underwent surgery for treatment of relapsed Ewing sarcoma.
Thirty-seven out of 38 (97.3%) patients had severe neutropenia lasting more than 7 days, with neutrophil count < 0.1 x 109/L in 35 patients (92.1%), 27/38 (71%) patients had received prolonged corticosteroid treatment at the moment of EG. Nine out of 38 (23.7%) patients had iatrogenic diabetes.
Diagnosis. The diagnosis of EG was documented by positive culture of wound swab in 34/38 patients, associated with positive blood culture in 18 patients; by positive blood culture in 4. In total 22/38 (57.9%) patients had bacteremia.
The isolated bacteria was Pseudomonas aeruginosa in 34 cases,Stenotrophomonas maltophilia in 3 and Escherichia coli in 1. In 10 patients a co-infection, with presence of more than one bacteria in the same culture, was documented: 3 Staphylococcus epidermidis, 1 Staphylococcus haemoliticus, 1Corynebacterium striatum, 1 Enterococcus faecalis, 1Enterococcus faecium, 1 Enterococcus faecium +Bacteroides fragilis, 1 Enterobacter cloacae, 1 multidrug resistant Klebsiella pneumoniae.
The EG involved the perineal region in 19/38 (50%) patients, other sites of EG are presented in table 2.
In 22/38 (57.9%) patients of our series the EG was associated with bacteremia, in 14 cases as single lesion, in 8 cases involving multiple sites in the body. Among the 16 patients without bacteremia, the infection was disseminated in 3/16 (18.7%), localized in 13/16 (81.3%).
In total, in 13/38 (34.2%) cases the EG presented as no-bacteriemic single cutaneous lesion.
The different age of presentation as regard to bacteriemic and non bacteriemic forms are presented in table 3. No different age distribution appeared between the bacteriemic and no-bacteriemic group (average age 8.5 years vs 6.8 yrs respectively, ns ); however the group of children with bacteriemic EG and disseminated presentation were apparently older (11.4 yrs) compared to the other groups (table 3).
Sixteen out of 38 (42.1%) patients were receiving antimicrobial prophylaxis or empiric antibacterial treatment, according to local practices, based on amoxicillin/clavulanate in 5 patients, cotrimoxazole in 2, meropenem in 2, ciprofloxacin in 2, ceftriaxone in 2, ceftazidime in 1, clindamicin in 1, ampicillin in 1.
Treatment and outcome. Treatment was based on antibacterial drugs in all children, started after a median of 1.1 days, with a latency of 0 days in 21 patients, 1-3 days in 13 patients, > 3 days in 3. The therapy was delivered for a median time of 20.5 days (range 9-270 days). Eleven patients received monotherapy with anti-Gram negative agent (piperacillin/tazobactam in 5, meropenem in 4, ciprofloxacin and ceftazidime in 1 each), 27 children received combined antimicrobal treatment, with 2 drugs in 22 patients (ceftazidime or cefepime and amikacin in 13, meropenem or imipenem and amikacin in 4, meropenem and ciprofloxacin in 2), ≥3 drugs in 5 cases. Second line antibiotic treatment was needed in 6/11 patients in which a monotherapy approach had been attempted, and in 5/22 children treated with two drug-combined therapy. In 12/37 (32.4%) tested patients the antimicrobal susceptibility revealed MDR bacteria (table 4), 9 of them had received antimicrobal prophylaxis. Ten out of 12 (83.3%) patients with MDR infection presented a bacteremic form of EG (83.3% in the MDR group vs 46.1% in the non-MDR group, p=0.0403 ). The lenght of treatment did not differ between cases with MDR bacteria and non-MDR (average 30.6 days vs 27.2 days, ns ).
Surgical treatment, namely debridement of the necrotic tissue, was performed in 24/38 (63.1%) patients, 18/24 (75%) of these patients had the EG involving more sites in the body and/or bacteriemic disease. In addition five patients underwent Vacuum therapy and 4 patients were treated with hyperbaric chamber. In 5 patients the intervention of plastic surgery with skin graft was needed.
In all patients the recovery from the infection was documented, unfortunately in 5 patients the EG recurred. Nine out of 38 patients, after healing of the lesions, showed deep scars, with muscle atrophy in 2 cases.
The median follow up time was 38.3 months (range 2.9-121.9 months).
Four out of 38 (10.5% mortality rate) patients died, one due to relapse of the EG with KPC co-infection, two due to progression of the underlying disease, the last one died later on due to transplant related mortality after allogeneic stem cell transplantation (SCT). Death occurred after 4, 4.5, 13 and 10 months, respectively. Three out of 4 deaths occurred in the bacteremic infections, with a mortality rate of 13.6% (3/22) in this group. The mortality rate in the non-bacteremic EG was 6.2% (1/16).