The Etiology, Antibiotic Therapy and Outcomes of Bacteremic Skin and Soft-Tissue Infections in Onco-Hematological Patients

Objectives: to assess the current epidemiology, antibiotic therapy and outcomes of onco- hematological patients with bacteremic skin and soft-tissue infections (SSTIs), and to identify the risk factors for Gram-negative bacilli (GNB) infection and for early and overall mortality. Methods: episodes of bacteremic SSTIs occurring in cancer patients at two hospitals were prospectively recorded and retrospectively analyzed. Results: Of 164 episodes of bacteremic SSTIs, 53% occurred in patients with solid tumors and 47% with hematological malignancies. GNB represented 45.5% of all episodes, led by Pseudomonas aeruginosa (37.8%). Multidrug resistance rate was 16%. Inadequate empirical antibiotic therapy (IEAT) occurred in 17.7% of episodes, rising to 34.6% in those due to resistant bacteria. Independent risk factors for GNB infection were corticosteroid therapy and skin necrosis. Early and overall case-fatality rates were 12% and 21%, respectively. Risk factors for early mortality were older age, septic shock, and IEAT, and for overall mortality were older age, septic shock and resistant bacteria. Conclusions: GNB bacteremic SSTI was common, particularly if corticosteroid therapy or skin necrosis. IEAT was frequent in resistant bacteria infections. Mortality occurred mainly in older patients with septic shock, resistant bacteria and IEAT. These results might guide empirical antibiotic therapy in this high-risk population.


Introduction
Skin and soft-tissue infections (SSTIs) are an important cause of morbidity and mortality among hospitalized patients and represent a therapeutic challenge for clinicians.Although uncomplicated SSTIs are often managed successfully on an outpatient basis, more severe infections are not uncommon, including a range of clinical presentations such as deep infections requiring surgical intervention, infections with secondary bacteremia and/or sepsis [1], and infections with local complications such as ischemia or tissue necrosis [2].In addition, some specific host conditions including immunosuppression, recent surgery, irradiation or diabetes mellitus may lead to more severe presentations and worse outcomes [3].
In immunosuppressed cancer patients, bacterial SSTIs are often accompanied by bacteremia with severe systemic signs and symptoms of infection.Furthermore, neutropenia can reduce or even conceal local signs of infection, and thus complicate the initial evaluation of the clinical presentation [4].Gram-positives are regarded as the leading cause of SSTIs in the general population.Nevertheless, in immunocompromised patients, such as patients with cancer, and particularly those with neutropenia, Gram-negative bacilli (GNB) may play an important role as causative agents [5].
The last update of the Infectious Diseases Society of America (IDSA) guidelines for the diagnosis and management of SSTIs was published in 2014, almost ten years ago [6].Since then, the global epidemiology of bacterial infections has changed, with an increase in Gramnegatives and a worrisome emergence of antimicrobial resistance [7].These developments are of paramount importance in cancer patients, who are particularly prone to infections by resistant bacteria due to repeated cycles of antibiotics, prolonged hospitalizations and immunosuppression [8].In addition, adequate initial empirical antibiotic therapy is particularly important in this population, because inappropriate empirical antibiotic treatment (IEAT) has been identified as a major risk factor associated with mortality [9,10].
The current IDSA guidelines for the management of patients with febrile neutropenia recommend adding an agent with Gram-positive activity in patients with suspicion of SSTIs [11].However, at a time when resistant Gram-negatives are emerging as an important cause of bacterial infections in cancer patients, this approach may be insufficient.In addition, the current data on SSTIs in cancer patients are very limited [5,12,13].Specifically, little is known about potentially severe SSTIs, such as those accompanied by bacteremia and the risk factors for Gram-negative infection.Furthermore, articles including data of a cohort of patients with well-documented microbiological diagnosis of SSTIs, such as those with bacteremia, are missing.
The aim of this study was to assess the currently poorly understood epidemiology, clinical characteristics, empirical antibiotic therapy and outcomes of bacteremic SSTIs in a cohort of immunocompromised cancer patients and hematopoietic stem cell transplant (HSCT) recipients in the current era of antimicrobial resistance, and to identify the risk factors for Gram-negative infection and for mortality.

Patient Characteristics and Outcomes
We identified 164 episodes of bacteremic SSTIs in our onco-hematological patients.The epidemiological and clinical characteristics and outcomes of all episodes are summarized in Table 1.The underlying disease was more likely to be the presence of solid tumors than hematological malignancies; breast cancer and lymphoma were the most common conditions in each group of patients.Nineteen patients (12%) were HSCT recipients, in most cases allogeneic HSCT (n = 14).Neutropenia was present in 23% of patients, and 38% received concomitant corticosteroid treatment.The limbs were the most frequent primary location of the infection (59.5%).Thirty-three patients presented skin necrosis (20%), and 57% of the episodes were classified as cSSTI.Supplementary Table S1 shows the clinical features, therapeutic management and outcomes of patients with necrotizing fasciitis.Septic shock at BSI onset was present in 13% of the patients; 7% required intensive care unit (ICU) admission, and early and overall case-fatality rates were 12% and 21%, respectively.

Risk Factors for Gram-Negative Infection
Table 3 displays the risk factors for Gram-negative infection by univariate and multivariate analysis.Polymicrobial episodes have been excluded to allow the analysis of risk factors for GNB infection to be more homogeneous.The presence of necrosis and the use of concomitant corticosteroids were significantly associated with Gram-negative infection in the univariate and multivariate analyses (OR 2.4, 95% CI 1.03-6, and OR 2.4, 95% CI 1.2-5.1,respectively).

Setting, Patients and Study Design
This study was carried out at two university referral hospitals in Barcelona, Spain: the Institut Català d'Oncologia, a 100-bed center for adult cancer patients, and the Hospital Clínic i Provincial, a 1000-bed general hospital.At each institution, all patients with bacteremia are identified daily by the Microbiology laboratory, and visited afterwards by an infectious disease physician who provides medical advice when necessary.Thus, data are prospectively collected in a specific database as part of standard infectious disease management.
We included in the study those adult patients (≥18 years old) with solid tumors, hematological malignancies or HSCT recipients presenting one episode of bacteremic SSTIs from January 2010 to June 2021.The objectives of the study were to assess the etiology, clinical characteristics, empirical antibiotic therapy, and outcomes of onco-hematological patients with bacteremic SSTIs, and to identify the risk factors for Gram-negative infection and for early (7-day) and overall 30-day case fatality rates.For this purpose, we compared the episodes caused by Gram-negatives with those caused by Gram-positives, and we also compared patients who died with those who survived.
The Clinical Research Ethics Committees at both hospitals approved the study.To protect privacy, personal information was encrypted in the electronic database.The need for informed consent was waived by the ethics committees due to the retrospective nature of the data analysis, and because no intervention was carried out.

•
SSTIs were defined as complicated (cSSTIs) if deep subcutaneous tissues were involved and/or surgery was necessary in addition to antimicrobial therapy [14].

•
Necrotizing fasciitis was defined when the infection involved the superficial fascia, the subcutaneous tissue planes and underlying muscles [2].

•
Co-morbidities were defined as the presence of one or more of the following conditions: chronic obstructive pulmonary disease, diabetes mellitus, chronic heart disease, chronic kidney disease (eGFR < 60 mL/min), chronic liver disease and cerebrovascular disease.

•
Bloodstream infection (BSI) was classified as nosocomial-acquired, healthcare-related, or community-acquired in accordance with previously reported criteria [15].

•
Previous corticosteroid treatment was defined as the administration of 20 mg of prednisone, or equivalent dosing, for at least four weeks within the last 30 days from BSI onset.

•
Septic shock was defined as a systolic blood pressure <90 mmHg that was unresponsive to fluid treatment and that required vasoactive drug therapy.

•
MDR strains were defined based on previously described criteria [16].

•
Initial empirical antibiotic therapy was considered inadequate if the treatment regimen did not include at least one antibiotic active in vitro against the infecting microorganism.

•
The early and overall case-fatality rates were defined as death from any cause within seven and 30 days of bacteremia onset, respectively.

Blood Sample Collection and Laboratory Equipment
Two sets of two 8-10 mL blood samples (Bactec PlusAerobic and Anaerobic, BD) were taken at 30 min apart from all patients who presented with fever ≥38 • C or when BSI was suspected due to clinical signs or symptoms.Either a BACTEC 9240 apparatus (from 2006 to May 2010) or a BACTEC-FX apparatus (from May 2010, BD Microbiology Systems) were used to process blood samples with an incubation period of five days.Positive blood samples were subcultured onto chocolate agar.

Identification Methods
Identification and antibiotic susceptibility of GNB, Enterococcus spp.and S. aureus were performed using commercially available MicroScan panels from the Walkaway automated system (Beckman-Coulter).Identification of other Streptococcus spp. was performed by standard biochemical testing and antibiotic susceptibility with commercially available Sensititre panels (TREK Diagnostic System).Anaerobes were identified by standard biochemical methods, and antibiotic susceptibility by the E-test method (BioMérieux).From November 2012 onward, identification was also performed by Matrix-Assisted Laser Desorption Ionisation (MALDI-TOF; Bruker Daltonics).Finally, susceptibility or resistance to antimicrobial agents were defined according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) recommendations [17].

Statistical Evaluation
The Mann-Whitney U test or the Student t-test were used to compare continuous variables.Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated, as appropriate.Categorical variables were analyzed by the chi-squared test or Fisher's exact test using two-by-k contingency tables.All p-values were two-tailed and statistical significance was considered at p < 0.05.Multivariate conditional logistic regression analysis of factors potentially associated with Gram-negative infection and with mortality included all statistically significant variables in the univariate analysis, together with sex, age, and all clinically relevant variables regardless of whether they were statistically significant.
Analyses were performed by stepwise logistic regression modelling in the SPSS statistics Version 27.0 (SPSS institute Inc., Chicago, IL, USA) [18].

Discussion
In our large cohort of bacteremic SSTIs in onco-hematological patients, we found that the patients affected were mostly non-neutropenic patients with breast cancer and lymphoma.Gram-negatives remained an important cause of infection, particularly in patients presenting with necrosis and concomitant use of steroids.Mortality occurred mainly in older patients with septic shock at presentation who received IEAT, often due to infection by resistant bacteria.
In our study, the rates of bacteremic SSTI were similar in patients with solid tumors and with hematological malignancies, and only a quarter presented neutropenia.This finding highlights the importance of local factors such as the primary site of the neoplasm with prior surgery and additional local lymphadenectomy and/or radiotherapy in patients with solid tumors, and, to a lesser extent, in patients with lymphoma, as notable risk factors for infection [5].Other risk factors that predispose to SSTI are the presence of mucositis, epithelial damage and myelosuppression, as well as prolonged use of central venous catheters, skin microbiome dysbiosis, poor wound healing, comorbidities, malnutrition and the use of other immunosuppressants such as corticosteroids [4].
In line with previous studies involving both the general non-immunocompromised population and patients with cancer [4,14,19,20], we found that S. aureus was the leading cause of SSTI.However, the rate of Gram-negative infection was particularly high, accounting for almost half of the episodes, rising slightly to 56% in patients with complicated SSTI.Bearing in mind that a large number of patients had breast cancer and that SSTIs in these patients are usually caused by Gram-positives (mainly S. pyogenes), we performed a subanalysis of the cohort excluding these patients, and found that the rate of GNB infection remained high (44%).
In addition, more than 15% of the Gram-negative isolates were MDR.These results are in concordance with the increase in Gram-negatives as a leading cause of bacterial infection in cancer patients in recent decades, and with the emergence of antibacterial resistance among them [8,10,21,22].This is a matter of special concern because the currently available guidelines for the management of SSTI and febrile neutropenia do not include specific recommendations for immunosuppressed cancer patients at risk of MDR-GNB SSTIs.In fact, the IDSA guidelines recommend adding an agent with Gram-positive activity in neutropenic patients with SSTI, but offer no specific guidance for MDR-GNB [11].In our study, the presence of necrosis and the use of concomitant corticosteroids were significantly associated with an increased risk of GNB infection.To our knowledge, this is the first study evaluating the risk factors for GNB SSTIs in cancer patients.
More than half of the patients presented complicated bacteremic SSTI, including the 30 cases that required drainage or surgery.This percentage is higher than those reported previously in cancer patients and in the general non-immunocompromised population [2,5,6].In addition, a high proportion of patients debuted with septic shock and presented mortality rates that exceeded the expected figures for this type of infection.This could be explained by the fact that all patients included in our study had concomitant bacteremia, a finding not frequently reported in other studies [5,13].
Five of the seven episodes of necrotizing fasciitis were caused by GNB (two were polymicrobial).Although empirical antibiotic therapy was adequate in almost all patients (86%), the overall case fatality rate reached 43%.Of note, three patients were admitted to the ICU and only three underwent surgery.Necrotizing fasciitis is a life-threatening disease in which surgical treatment is essential to achieving a better prognosis [23,24].The low rate of surgical debridement in our study is consistent with other reports of lower percentages of surgical treatment among immunocompromised patients compared to their non-immunocompromised peers [12,25].One plausible explanation for the fact that cancer patients are less likely to undergo surgery could be the fear of the high intraoperative mortality in frail patients with an underlying malignancy land who often present cytopenias.
IEAT was mainly associated with infection due to MDR bacteria, and was associated with an increased risk of death.Since older age and septic shock are not variables in which physicians can intervene, empirical antibiotic therapy remains the only modifiable variable to improve outcomes.In the current era of widespread antimicrobial resistance, updated recommendations for guiding empirical antibiotic therapy in immunocompromised patients with cancer are urgently needed.
This study has some limitations that should be acknowledged.First, some information may have been lost due to its retrospective design, and we may not have adequately controlled for certain confounders.Second, all the patients included in the study had concomitant bacteremia, and so patients with SSTI and no microbiological documentation and those without bacteremia were excluded.Nevertheless, this design allows the identification of patients at higher risk of severe sepsis and worse outcomes, in whom prompt adequate empirical antibiotic treatment and aggressive management are mandatory.Third, the study population may be heterogeneous due to inherent differences in the underlying disease.On the other hand, the main strength of this study is that it describes the largest cohort of cancer patients with bacteremic SSTI, using a multicenter design, even though the participating centers are located in the same geographical area.
In conclusion, we found that bacteremic SSTI due to GNB was common, particularly in patients with concurrent corticosteroid therapy and presenting skin necrosis.P. aeruginosa was the leading cause.Mortality occurred mainly in older patients with septic shock at presentation who received inadequate empirical antibiotic therapy, often due to infection by resistant bacteria.In the current era of widespread antimicrobial resistance, updated recommendations for guiding empirical antibiotic therapy in cancer patients with SSTIs are warranted.

Table 1 .
Epidemiological and clinical characteristics and outcomes of 164 episodes of bacteremic skin and soft tissue infections.
a SSTI: skin and soft-tissue infection.

Table 2 .
Causative organisms of 164 episodes of bacteremic SSTIs in cancer patients and hematopoietic stem cell transplant recipients.

Table 3 .
Risk factors for Gram-negative infection in the univariate and multivariate analyses.
a COPD: chronic obstructive pulmonary disease.b IEAT: inadequate empirical antibiotic therapy.

Table 4 .
Risk factors for 7-day mortality in the univariate and multivariate analyses.

Table 5 .
Risk factors for 30-day mortality in the univariate and multivariate analyses.