- Email: [email protected]
Diabetic foot infection in hospitalized adults
J Infect Chemother xxx (2016) 1e7
Contents lists available at ScienceDirect
Journal of Infection and Chemotherapy journal homepage: http://www.elsevier.com/locate/jic
Original article
Diabetic foot infection in hospitalized adults David E. Katz a, N. Deborah Friedman b, Evgenia Ostrovski c, Dor Ravid d, Nadav Amrami d, Dori Avivi d, Bethlehem Mengesha d, Ronit Zaidenstein d, e, Tsilia Lazarovitch f, Mor Dadon d, Dror Marchaim d, e, * a
Department of Internal Medicine D, Shaare Zedek Medical Center, Jerusalem, Israel Department of Infectious Diseases, Barwon Health, Victoria, Australia c Department of Internal Medicine D, Assaf Harofeh Medical Center, Zerifin, Israel d Unit of Infectious Diseases, Assaf Harofeh Medical Center, Zerifin, Israel e Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel f Clinical Microbiology Laboratory, Assaf Harofeh Medical Center, Zerifin, Israel b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 October 2015 Received in revised form 15 November 2015 Accepted 16 December 2015 Available online xxx
Background: Acute infections of the diabetic foot (DFI) are a common and complex condition. Patients are generally managed in the ambulatory setting and epidemiological data pertaining to hospitalized patients is lacking. The aim of this study was to analyze the epidemiology, microbiology and outcomes of hospitalized patients with DFI, who are managed at a referral center equipped with hyperbaric oxygen (HBO) therapy. Methods: A retrospective cohort study of adult patients admitted to a tertiary referral center with DFI over a six-month period in 2013 was undertaken. Predictors of clinical outcomes and efficacy of treatment modalities were analyzed by Cox regression. Results: Sixty-one patients with DFI were identified. Most patients were elderly (67 ± 13 years), with long-standing (17 ± 9 years), poorly controlled (HbA1c 9 ± 3%) diabetes. Most patients had polymicrobial infection (80%); specifically, anaerobic (39%) and multi or extensively-drug resistant organisms (61%). Administration of appropriate antimicrobials was delayed for >48 h in 83%. Advanced age was associated with worse outcomes. Sicker patients with severe peripheral vascular disease were managed with HBO. The use of HBO was associated with higher costs and increased functional deterioration, and did not prevent future limb amputation. Conclusions: Our study illustrates the descriptive epidemiology of hospitalized adults with DFI predominantly of polymicrobial etiology. MDROs and anaerobic organisms are common causative pathogens, and appropriate antibiotics were frequently delayed. HBO treatment may delay the need for limb amputation, but not obviate this eventual outcome. © 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Keywords: DFI Skin and soft tissue infections Diabetes Hyperbaric oxygen SSTI Antimicrobial resistance
1. Introduction Diabetes mellitus affects 8e13% of the world population, and according to national data from the United States (US), has a prevalence of 27% in patients over 65 years [1]. Among patients with diabetes, diabetic foot infection (DFI) is one of the most common complications with up to 10% of diabetic patients
* Corresponding author. Unit of Infectious Diseases, AssafHarofeh Medical Center, Zerifin, 70300, Israel. Tel.: þ972 8 977 9049; fax: þ972 8 977 9043. E-mail address: [email protected] (D. Marchaim).
developing a DFI [1,2]. DFI presents with a range of signs and symptoms, and the severity of the DFI can range from a superficial infection of the nail bed to deeper infections involving bone (osteomyelitis) [3]. Although DFI is predominantly managed in the outpatient setting, in the developed world, DFI is the most common admission diagnosis for diabetic patients [4], and management in these cases frequently involves partial or total amputation of the affected lower limb. DFI is the most common reason for non-traumatic amputations in developed countries [2], with data from the US showing that up to 60% of all non-traumatic adult limb amputations in 2010 were related to DFI [1]. Survival following limb amputation for DFI is less
http://dx.doi.org/10.1016/j.jiac.2015.12.007 1341-321X/© 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
2
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
than 50% within 3 years after surgery, and only 40% at 5 years after surgery [2]. Partial or total limb amputation for DFI predicts further amputation in 20% of patients within 2 years of the procedure, and also predicts development of another DFI in the initially unaffected limb in 50% of patients [2]. Therefore, amputation is both a marker for serious DFI and of an unfavorable outcome. Treatment with hyperbaric oxygen (HBO) is one of few existing modalities for treating DFI in order to avoid or delay amputation [5]. However, there is ongoing debate regarding the independent role of HBO treatment in effectively reducing amputation rates over time [6]. Although the clinical and microbiological diagnosis of DFI may be both challenging and complicated [7], early and precise diagnosis is important for both effective infection management and for reducing the need for subsequent amputation [3,6]. Microbiological diagnosis is critical as it allows the clinician to tailor antibiotic therapy to the causative organism [8]. However, in many cases, the antimicrobial management for DFI is empiric as there may be no appropriate site from which to obtain a culture, superficial skin cultures correlate with the ‘true offending pathogen’ less than 30% of the time, and patients may also present already taking long-term antibiotics, which further decreases the culture yield [6,9]. Importantly, inappropriate empiric broad-spectrum antimicrobial therapy can result in unfavorable outcomes for individual patients and contribute to the broader harmful ecological impact of antimicrobial therapy [10]. The epidemiology of DFI in Israel has not been comprehensively researched [9,11,12]. Moreover, internationally, the epidemiological features of patients with DFI managed as inpatients is also lacking [8]. The goals of this study were to perform a descriptive analysis of the epidemiological and microbiologic features of patients with DFI admitted to a referral acute-care facility, to determine risk factors for a worse clinical course (i.e., death, amputation, re-admission, and functional deterioration), and to analyze the controlled effect of HBO treatment on the clinical course of patients with DFI. 2. Materials and methods 2.1. Study setting Assaf Harofeh Medical Center (AHMC) is an 815-bed tertiary care university-affiliated facility located in southern-central Israel. AHMC contains a state-of-the-art hyperbaric medicine centre (with dual chambers with capacity for 37 patients), and a multidisciplinary team with expertise in the management of DFI [13]. This study was approved by the institutional review board at AHMC according to the declaration of Helsinki statement of ethical principles for conducting human research. A retrospective review of charts of patients consecutively admitted to AHMC from May through November 2013 was undertaken as part of a broader project on infectious syndromes among hospitalized patients. During this time period, 8132 charts were reviewed. Patients over 18 years of age with a diagnosis of DFI based on their discharge records and ICD-9 discharge codes, were included. The diagnosis of DFI was per accepted definitions [6] and in consultation with a senior infectious disease physician. Microbiology samples for culture were obtained by tissue biopsy or collected via sterile syringe in the case of direct pus draining from a deep sinus tract [9]. 2.2. Definitions Multidrug resistant organisms (MDR) included; extendedspectrum b-lactamase producing Enterobacteriaceae (ESBLs), methicillin-resistant Staphylococcus aureus (MRSA), carbapenem-
susceptible Pseudomonas aeruginosa, and carbapenem-susceptible Acinetobacter baumannii. Extensively-drug resistant organisms (XDR) included; vancomycin-resistant enterococci (VRE), carbapenem-resistant Enterobacteriaceae (CRE), carbapenem-nonsusceptible Pseudomonas aeruginosa, and carbapenem-nonsusceptible Acinetobacter baumannii. 3. Data collection The following data was extracted from patient medical records; patient demographics, details of the DFI, history and duration of diabetes, co-morbid illnesses (including Charlson's scores [14]), functional status, previous hospital admissions, surgical procedures and use of invasive devices (e.g. urinary catheters, feeding tubes) prior to admission, medications, hospital length of stay, details of HBO treatment, vital signs, physical examination findings, laboratory results including bacterial culture results, and the results of imaging of the affected foot. In addition, post-discharge follow-up data were collected on both surgical procedures, and microbiology results up to 6 months after the hospitalization. Mortality data was extracted from national records of the Israeli ministry of interior. The University of Texas (UT) diabetic wound classification incorporating wound stage and appearance was calculated for all patients to estimate the severity of the DFI [15]. 3.1. Data management and analysis The data was entered and processed using SPSS version 22.0 (2014). Both descriptive and univariable analyses were performed. Multivariable models were created for four outcomes: 1) decrease in functional status following the DFI index event, 2) readmission to the acute care hospital within 6 months, 3) partial or total amputation of the limb affected resulting from the index DFI, and 4) increased cost (admissions costing in excess of 30,000 New Israeli Shekels [NIS], which equals ~$7700). Cost analyses were based on the total costs that were billed to insurance companies, based on the rates set by the Israeli ministry of health. These costs included; inpatient care, invasive procedures (including surgery), days of hospitalization (based on the type of unit), and imaging, set at fixed rates as determined by the Israeli ministry of health. HBO costs, if present, were deducted. An analysis was also performed comparing treatment groups (HBO versus no HBO). All analyses were performed by accepted methods. Categorical variables were analyzed by the chi-square test or Fisher's exact test; continuous variables were analyzed by the student's t-Test or ManneWhitney U test. Multivariable modeling was performed for each outcome, using Cox regression. Multivariable analysis of patient characteristics for those patients treated with HBO was performed using logistic regression. All parameters with a P-value less than or equal to 0.1 were incorporated to the multivariable analyses. 4. Results Of the 8132 consecutive patient discharge notes reviewed during the 6-month study period, we identified 61 adult patients with DFI. All DFIs were present on admission. The study population was male predominant, of mean age 67 years and more than half were functionally impaired at baseline, with high Charlson's scores [14]. Most patients had long-term diabetes under poor control and had been diagnozed with diabetic microvascular disease (Table 1). Approximately one-third of patients were
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
3
Table 1 Baseline epidemiologic features among hospitalized patients with DFI. Parameter Demographics Age, years, mean ± SD Female gender Elderly (over 65 years) Admitting Unit
Medicine Surgery Adult ICU
Parameters related to diabetes Diabetic microvascular disease (neuropathy/retinopathy/ nephropathy) Years of diabetes at DFI onset, mean ± SD HbA1c, percent, mean ± SD Acute and chronic conditions on admission Dependent Functional status Cognitive impairment Rapidly fatal McCabe score [34] Dementia Chronic lung disease Congestive heart failure Ischemic heart disease Malignancy (any type, past or present) Peripheral vascular disease Smoker Charlson's weighted index co-morbidity, median (range) Charlson's combined condition score, median (range) Charlson's 10-year survival probability, percent, median (range) Immunosuppressive states or conditions Glucocorticoid therapyb in the previous month Post transplantation (solid organ or bone marrow) Exposure to healthcare settings and environments before infection Resident of long-term care facility Hospitalized in the previous 3 months Past limb amputation Chronic hemodialysis Invasive procedure (including surgeries) in the previous 3 months Permanent urinary catheter Severity of illness indices at time of infection Reduced consciousness Severe sepsis, septic shock or multi-organ failure (as opposed to sepsis syndrome) Acute renal failure Antimicrobial therapy Antibiotics in the preceding 3 months Number of antibiotics administered in the preceding 3 months, median (range) Days from last antibiotic administered, median (range) 48 h delay in initiating appropriate therapy (per in-vitro report of microbiology laboratory) Days to initiation of appropriate antibiotics, median (range) Duration (in days) of antimicrobial treatment course for the index infectious episode, median (range) Most common empiric antimicrobial regimens Beta-lactam beta-lactamase inhibitor combinations Cephalosporins Fluoroquinolones Glycopeptides Clindamycin Metronidazole Most common maintenance antimicrobial regimens Beta-lactam beta-lactamase inhibitor combinations Cephalosporins Carbapenems Fluoroquinolones Glycopeptides Metronidazole a b
Valid percenta
Number 67 ± 13 22 35 19 41 1
36 57 31 67 2
55
90
17 ± 9 9±3 37 9 4 8 17 15 19 7 32 20 4 (1e9) 7 (1e13) 8 (0e96)
61 15 7 13 28 25 31 12 53 33
5 2
8 3
6 22 17 8 16 2
10 36 28 13 26 3
4 6
7 10
15
25
27 1 (0e3)
44
1 (0e33) 29
83
6 (0e44) 16 (0e55)
9 40 17 5 4 26
15 66 28 8 7 43
8 17 2 11 4 7
13 28 3 18 7 12
Prevalence calculated only for patients with data. Glucocorticoid therapy in the previous month was defined as receipt of an equivalent dose of at least 16 mg of prednisolone for 5 days or more.
recently hospitalized and nearly half of patients were prescribed antibiotics in the preceding three months (Table 1). Over half of patients had a UT diabetic wound classification score over IIB [15] (Table 2).
4.1. Microbiological results Forty of 61 patients (66%) had positive microbiological results from specimens collected. Most infections were polymicrobial, and
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
4
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
Table 2 University of Texas (UT) Diabetic wound classification score [15] for hospitalized patients with DFI. Wound grade
Number
Valid percenta
0eB 0eD IeA IeB IeC IeD IIeB IIeD IIIeB IIIeC IIIeD
6 1 1 14 2 1 5 16 8 1 6
9 1.5 1.5 20.6 2.9 1.5 7.4 23.5 11.8 1.5 8.8
a
Prevalence calculated only for patients with data.
over 60% of patients isolated a multidrug resistant (MDR) or extensively drug-resistant (XDR) bacteria (Table 3). 4.1.1. Antimicrobial therapy Patients received a median of 16 days of antimicrobial therapy (up to 55 days) and were commonly managed with combinations of agents both as empiric therapy and as consolidative regimens. Cephalosporins, metronidazole fluoroquinolones and beta-lactam/ beta-lactamase inhibitor combinations were most commonly prescribed (Table 1). 4.2. Patient outcomes Patient mortality was low, but functional deterioration occurred in one-third of patients (Table 4). More than half of patients were re-hospitalized, and one-quarter required amputation in the next 6 months. The median length of hospital stay was 10 days, with a maximum length of stay of 65 days. Median hospitalization costs were in excess of $6500 (Table 4). The multivariable models of patients' outcomes revealed that empiric use of metronidazole and treatment with HBO, were associated with worse outcomes. Increasing age was an independent risk factor for re-admission within 6 months after the index hospitalization due to DFI, and PVD was associated with increased costs (Table 5). 4.3. Hyperbaric oxygen (HBO) treatment Sixteen patients were treated with HBO per protocol [13]. Patients who underwent HBO were more likely to have PVD, a lower Charlson's survival probability [14], and over one-third of cases were dialysis-dependent. Higher numbers of amputations, increased hospital length of stay and overall costs (excluding HBOassociated costs) were observed in the patients who underwent HBO (Table 6). 5. Discussion This study describes a cohort of 61 patients with DFI who were hospitalized in a tertiary-care academic medical center. AHMC has a multi-disciplinary team for management of DFI including onsite HBO. The patient cohort consisted of an elderly group of patients, with functional impairment at baseline, a median 10-year survival probability of only 8%, and extensive exposure to antimicrobials, the healthcare environment, and procedures (Table 1). The suboptimal outcomes of patients, specifically, deterioration in functional status, future additional hospitalizations and eventual amputation of the affected limb were substantial among this hospitalized group of patients (Table 4).
Table 3 Microbiological results of 40 hospitalized patients with DFI. Parameter
Number
Valida percent
Monomicrobial Polymicrobial Anaerobic Patients with bacteremia ESBLb MDR organismsc XDR organismsd
8 32 16 3 5 18 7
20 80 39 8 12 44 17
Bacteria isolated Enterococcus faecalis Corynebacterium spp. Staphylococcus aureus MRSAe MSSAf Streptococcus dysgalactiae Streptococcus pyogenes Staphylococcus epidermidis Streptococcus agalactiae Streptococcus viridians group Peptococcus spp. Escherichia coli Pseudomonas aeruginosa Proteus mirabilis Klebsiella pneumoniae Morganella morganii Acinetobacter baumanii Enterobacter cloacae complex Citrobacter koseri Citrobacter freundii Alcaligenes faecalis Klebsiella spp. Providencia rettgeri Acinetobacter haemilyticus Aeromonas hydrophila Enterobacter spp. Klebsiella oxytoca Proteus penneri Providencia spp. Providencia stuartii Pseudomonas putida Raoultella ornithinolytica Serratia marcescens Bacteroides fragilis Prevotella oralis Bacteroides ovatus Bacteroides thetaiotaomicron Prevotella melaninogenica Bacteroides caccae Bacteroides stercoris Bacteroides uniformis Fusobacterium varium
Bacteria type Gram positives
Number 9 8
Gram negatives
Anaerobic bacteria
1 5 4 3 1 1 1 1 9 9 8 7 7 5 4 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1 8 5 4 3 3 1 1 1 1
a
Prevalence calculated for the 40 patients with positive microbiological results. ESBL ¼ extended-spectrum b-lactamase producing Enterobacteriaceae. MDR ¼ multidrug resistant organisms include: ESBLs, methicillin-resistant Staphylococcus aureus (MRSA), carbapenem-susceptible Pseudomonas aeruginosa, and carbapenem-susceptible Acinetobacter baumannii. d XDR ¼ extensively-drug resistant organisms include: vancomycin-resistant enterococci (VRE), carbapenem-resistant Enterobacteriaceae (CRE), carbapenemnonsusceptible Pseudomonas aeruginosa, carbapenem-nonsusceptible Acinetobacter baumannii. e MRSA ¼ methicillin-resistant Staphylococcus aureus. f MSSA ¼ methicillin-susceptible Staphylococcus aureus. b
c
The microbiology results in this study are worthy of note. MDR and XDR organisms (i.e., MDROs) were isolated in over 60% of patients. A French study found that 18% of patients with DFI had a MDRO isolated [16], while Kandemir et al. found that 40% of microbiological isolates from patients with DFI contained an MDRO [17]. In contrast to this, Citron et al. reported low rates of MDROs and found that ertapenem and piperacillin-tazobactam were each active against >98% of enteric gram-negative rods, methicillinsensitive S. aureus, and anaerobes from a study of over 400 patients with DFI in the United States [18]. The presence of MDROs led
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
5
Table 4 Outcomes of hospitalized patients with DFI. Parameter In hospital mortality 30 day mortality 90 day mortality Functional deterioration Discharged to long-term care facility after being admitted from home Additional hospitalizations in the 6 months following the index infection Amputation 6 months post infection onset Length of hospital stay from infection initiation to discharge after excluding those deceased, days, median (range) Hospital costs from infection initiation to discharge after excluding those deceased, US dollar, median (range) a
Valid percenta
Number 2 1 5 17 7 34 13 10 (1e65) $6648 ($561e$69,376)
3 2 8 29 13 61 23
Prevalence calculated only for patients with data.
Table 5 Multivariable models of DFI outcomes. Variable
Treated with HBO Anaerobic pathogen Empiric regimen containing metronidazole Chronic renal failure PVD Charlson's comorbidity combined condition score [14] Age in years Days on antibiotics
Additional hospitalizations in next 6 months
Amputation in next 6 months
Direct costs > 30,000 NIS ($7700) from infection to discharge
OR
CI-95%
OR
P value
CI-95%
OR
P value
CI-95%
0.02 0.09 0.1 0.93 0.6 0.05
1.4e111 0.8e50 0.6e68 0.1e14.4 0.1e34 1.1e2.3
3.9
0.1
0.75e20
2e48 0.96e14
12.4 6.2 6.6 1.1 2.1 1.5
4.4
0.04
1.1e17.5
7.3
0.01
1.6e33
0.96 1.1
0.09 0.04
0.9e1.1 1.01e1.11
9.7 3.7
1.1
P value
0.005 0.06
0.005
1.03e1.17
Note. NIS ¼ new Israeli Shequels. HBO ¼ hyperbaric oxygen. PVD ¼ peripheral vascular disease. Th bold parameters are those with statistically significant associations. Table 6 Epidemiologic factors and infection outcomes of patients with DFI who were treated with hyperbaric oxygen. Parameter Demographics Age, years, mean ± SD Female gender Elderly (over 65 years) Acute and chronic conditions on admission Dependent functional status Cognitive impairment Rapidly fatal McCabe score [34] Peripheral vascular disease Years of diabetes, mean ± SD Hb A1C, percent, mean ± SD Charlson's combined condition score [14] Charlson's 10-year survival probability [14] Exposure to healthcare settings and environments before infection Resident of long-term care facility Past amputation Chronic hemodialysis Invasive procedure in the previous 3 months Severity of illness indices at time of infection Severe sepsis, septic shock or multi-organ failure ICU stay in current hospitalization Development of acute renal failure Outcomes In hospital mortality 30 day mortality 90 day mortality Functional deterioration Discharged to long-term care facility after being admitted from home Amputation in the 6 months following the onset of infection Length of hospital stay from infection diagnosis to discharge after excluding those deceased, mean ± SD Cost from infection diagnosis to discharge, US dollars, mean ± SD
Treated with hyperbaric oxygen Not treated with hyperbaric oxygen OR (n ¼ 16), N (%) (n ¼ 45), N (%)
CI-95%
P value
66 ± 9 8 (50) 7 (44)
67 ± 14 14 (31) 28 (62)
10 (63) 0 (0) 0 (0) 13 (81) 17 ± 10 9±3 8±2 7 ± 18
27 (60) 4 (9) 4 (9) 19 (42) 17 ± 10 9±2 7±3 23 ± 34
1.1 0.9 0.9 5.9
0.3e3.6 >0.99 0.8e1.0 0.6 0.8e1.0 0.6 1.5e23.8 <0.001 0.8 0.9 0.2 <0.001
0 (0) 6 (38) 6 (38) 3 (19)
6 (13) 11 (24) 2 (4) 13 (29)
0.9 1.9 13.0 0.6
0.8e1.0 0.3 0.6e6.3 0.3 2.3e73.6 <0.001 0.1e2.3 0.5
0 (0) 0 (0) 4 (25)
6 (13) 1 (2) 11 (24)
0 (0) 0 (0) 1 (6) 4 (25) 1 (6) 9 (56) 21 ± 18
2 (4) 1 (2) 4 (9) 13 (30) 6 (15) 4 (10) 11 ± 8
17,499 ± 17,360
7224 ± 5081
2.2 0.7e7.1 0.5 0.2e1.5
0.9 0.8e1.0 1.0 0.9e1.0 1.0 0.3e3.9 1.0 1.0 0.7 0.8 0.4 11.6
0.9e1.0 0.9e1.0 0.1e6.6 0.2e2.8 0.1e3.3 2.8e48.3
0.7 0.2 0.3
0.3 >0.99 >0.99 >0.99 >0.99 >0.99 0.8 0.7 <0.001 <0.001 <0.001
The bold parameters are those with statistically significant associations.
to a median delay of 6 days before commencing appropriate antibiotic therapy in this study. In over 80% of patients, appropriate treatment was delayed for over 48 h. Delay in appropriate antibiotic commencement is the strongest independent predictor for worse
outcomes in septic patients [19]. Ensuring earlier collection of appropriate specimens for culture will likely reduce inappropriate antimicrobial exposure, MDRO colonization pressure, and may improve patient clinical outcomes [7,20,21].
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
6
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
There are several factors which were identified that may have contributed to the high prevalence of MDROs among these patients. As found by others, previous exposure to antimicrobials was present in over 40% of patients, which would contribute to selective pressure [17,22]. In addition, among patients in this study there was a high prevalence of established risk factors for MDROs including; older age, complex co-morbidities, institutionalization, functional dependency, in addition to healthcare exposure in a country with high endemic rates of MDROs in hospitals [20,23e28]. Other studies have also found that previous hospitalization and healthcare exposures were risk factors for MDROs [16,17]. Interestingly and in keeping with other studies, many patients were exposured to both long-term care and to hemodialysis [29]. Moreover, the specific overuse of cephalosporins, fluoroquinolones and metronidazole (Table 1) have been found to be predictors of the emergence of MDROs [29]. We also observed a high prevalence of polymicrobial infections, and frequent isolation of anaerobic pathogens, as the causative pathogens of DFI (Table 3). These findings are in keeping with data from a large DFI cohort with over 80% polymicrobial infections, and nearly half growing only anaerobes [18]. The prevalence of anaerobic pathogens might even be underestimated, due to suboptimal anaerobic conditions in specimen transport and processing. Although empiric treatment with metronidazole was found to be an independent risk factor for future amputations and for high medical costs (Table 5), this was a confounder, as empiric treatment including metronidazole was prescribed to patients with higher severity of illness (data not shown). Guidelines for appropriate management of DFI recommend broad-spectrum empiric treatment including anaerobic and MDRO coverage [5]. Proper culture acquisition is however vital prior to substantial exposure to antibiotics to allow for appropriate adjustment of the treatment regimen, which is frequently prolonged in deep or invasive DFI (associated with osteomyelitis) [6]. Sixteen patients (26%) were treated with HBO. This small number does not allow for a significant analysis of patient risk factors and clinical outcomes. The independent association of HBO treatment with future amputations (Table 5) relates primarily to an inability to adjust adequately for confounding effects. We were however able to conclude that patients at AHMC that were referred for HBO had a higher burden of underlying diseases and both poor functional status and a low 10 year survival probability compared with patients not managed with HBO (Table 6). These patients referred for HBO likely had a DFI that required amputation during this hospital stay. HBO was likely being used as a “last-ditch” effort to prevent amputation. Our results suggest that HBO may delay inevitable amputation in this group of patients with DFI and severe peripheral vascular disease. Research by Faglia and colleagues revealed that adjunctive HBO reduced amputations in patients with ischemic diabetic foot ulcers, while other research has found that use of HBO neither improved the likelihood that a wound would heal nor prevented amputation [30,31]. Furthermore, some believe that HBO should not be utilized in DFI until large-scale, adequately blinded, controlled, and powered randomized studies have clearly demonstrated efficacy and cost effectiveness in the healing of ulcers and the prevention of major amputation [32]. Unfortunately, as we observed in our study, delaying amputation with HBO resulted in prolonged inpatient care and increased costs. The utility and value of HBO in DFI therefore needs to be further explored. The limitations of this study are that it is retrospective, and therefore, calculation of wound and comorbidity scores may have been inaccurate. Moreover, while we did not capture specific data regarding patients being assigned a diagnosis of osteomyelitis, our calculation of the UT wound score we believe has enabled an estimate of wound severity and depth. In addition, we acknowledge
that comprehensive cost analyses ought to also include the cost of social services, home care, rehabilitation and subsequent episodes of care for DFI. While the numbers of this study are relatively small, it provide an insight into the epidemiology of inpatients with DFI, while capturing and analyzing outcomes such as functional deterioration following the event, further hospitalizations and amputations in an aging diabetic cohort. DFI places a large fiscal burden on the healthcare system [33]. We identified a median cost of hospitalization of $6,648, but with costs ranging as high as $69,376. Larger prospective caseecontrol and cohort studies will be able to identify optimal treatments including the use of HBO and the optimal timing of surgery if required. Nonetheless this study provides a rationale for the early institution of broad-spectrum (including anaerobic) antimicrobial empiric therapy for acute DFI, while obtaining appropriate samples early for microbiological testing. 6. Financial support This study was partially supported financially by Merck Inc. The company had no access or involvement in data collection, data interpretation, and was not involved in the process of drafting this manuscript. 7. Potential conflicts of interest All authors report no conflicts of interest relevant to this article. Acknowledgments None. References [1] Association AD, editor. Statistics about diabetes: data from the national diabetes statistics report; 2014. 2014. [2] Zandman-Goddard G, Feldbrin Z, Ovadia S, Zubkov T, Lipkin A, Wainstein J. A multi-disciplinary approach to diabetic foot patientsean organizational model for the treatment of leg complications in diabetic patients. Harefuah 2011;150:593e5. 616. [3] Bader MS. Diabetic foot infection. Am Fam Physician 2008;78:71e9. [4] Lipsky BA. Osteomyelitis of the foot in diabetic patients. Clin Infect Dis 1997;25:1318e26. [5] Peters EJ, Lipsky BA, Berendt AR, Embil JM, Lavery LA, Senneville E, et al. A systematic review of the effectiveness of interventions in the management of infection in the diabetic foot. Diabetes Metab Res Rev 2012;28(Suppl. 1): 142e62. [6] Lipsky BA, Berendt AR, Cornia PB, Pile JC, Peters EJ, Armstrong DG, et al. 2012 infectious diseases society of America clinical practice guideline for the diagnosis and treatment of diabetic foot infections. Clin Infect Dis 2012;54: e132e73. [7] Giurini JM, Lyons TE. Diabetic foot complications: diagnosis and management. Int J Low Extrem Wounds 2005;4:171e82. [8] Lipsky BA. A current approach to diabetic foot infections. Curr Infect Dis Rep 1999;1:253e60. [9] Slater RA, Lazarovitch T, Boldur I, Ramot Y, Buchs A, Weiss M, et al. Swab cultures accurately identify bacterial pathogens in diabetic foot wounds not involving bone. Diabet Med 2004;21:705e9. [10] Dellit TH, Owens RC, McGowan Jr JE, Gerding DN, Weinstein RA, Burke JP, et al. Infectious diseases society of America and the society for healthcare epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship. Clin Infect Dis 2007;44:159e77. [11] Tamir E. Treating the diabetic ulcer: practical approach and general concepts. Isr Med Assoc J 2007;9:610e5. [12] Weiss A, Karpf A, Luger E, Schmilowitz H, Dekel S, Shapira I. Long-term antibiotic treatment in geriatric diabetic foot infection. J Med 1998;29: 365e73. [13] Efrati S, Gall N, Bergan J, Fishlev G, Bass A, Berman S, et al. Hyperbaric oxygen, oxidative stress, no bioavailability and ulcer oxygenation in diabetic patients. Undersea Hyperb Med 2009;36:1e12. [14] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373e83.
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7 [15] Lavery LA, Armstrong DG, Harkless LB. Classification of diabetic foot wounds. J Foot Ankle Surg 1996;35:528e31. [16] Hartemann-Heurtier A, Robert J, Jacqueminet S, Ha Van G, Golmard JL, Jarlier V, et al. Diabetic foot ulcer and multidrug-resistant organisms: risk factors and impact. Diabet Med 2004;21:710e5. [17] Kandemir O, Akbay E, Sahin E, Milcan A, Gen R. Risk factors for infection of the diabetic foot with multi-antibiotic resistant microorganisms. J Infect 2007;54: 439e45. [18] Citron DM, Goldstein EJ, Merriam CV, Lipsky BA, Abramson MA. Bacteriology of moderate-to-severe diabetic foot infections and in vitro activity of antimicrobial agents. J Clin Microbiol 2007;45:2819e28. [19] Paul M, Shani V, Muchtar E, Kariv G, Robenshtok E, Leibovici L. Systematic review and meta-analysis of the efficacy of appropriate empiric antibiotic therapy for sepsis. Antimicrob Agents Chemother 2010;54:4851e63. [20] Bonten MJ. Colonization pressure: a critical parameter in the epidemiology of antibiotic-resistant bacteria. Crit Care 2012;16:142. [21] Edmonds M, Foster A. The use of antibiotics in the diabetic foot. Am J Surg 2004;187:25Se8S. [22] Baquero F, Negri MC, Morosini MI, Blazquez J. Antibiotic-selective environments. Clin Infect Dis 1998;27(Suppl. 1):S5e11. [23] Schwaber MJ, Carmeli Y. An ongoing national intervention to contain the spread of carbapenem-resistant enterobacteriaceae. Clin Infect Dis 2014;58: 697e703. [24] Abbo A, Carmeli Y, Navon-Venezia S, Siegman-Igra Y, Schwaber MJ. Impact of multi-drug-resistant acinetobacter baumannii on clinical outcomes. Eur J Clin Microbiol Infect Dis 2007;26:793e800. [25] Bogan C, Kaye KS, Chopra T, Hayakawa K, Pogue JM, Lephart PR, et al. Outcomes of carbapenem-resistant enterobacteriaceae isolation: matched analysis. Am J Infect Control 2014;42:612e20.
7
[26] Cosgrove SE, Qi Y, Kaye KS, Harbarth S, Karchmer AW, Carmeli Y. The impact of methicillin resistance in staphylococcus aureus bacteremia on patient outcomes: mortality, length of stay, and hospital charges. Infect Control Hosp Epidemiol 2005;26:166e74. [27] Schwaber MJ, Carmeli Y. Antimicrobial resistance and patient outcomes: the hazards of adjustment. Crit Care 2006;10:164. [28] Schwaber MJ, Navon-Venezia S, Kaye KS, Ben-Ami R, Schwartz D, Carmeli Y. Clinical and economic impact of bacteremia with extended- spectrum-betalactamase-producing enterobacteriaceae. Antimicrob Agents Chemother 2006;50:1257e62. [29] Tacconelli E, De Angelis G, Cataldo MA, Mantengoli E, Spanu T, Pan A, et al. Antibiotic usage and risk of colonization and infection with antibioticresistant bacteria: a hospital population-based study. Antimicrob Agents Chemother 2009;53:4264e9. [30] Faglia E, Favales F, Aldeghi A, Calia P, Quarantiello A, Oriani G, et al. Adjunctive systemic hyperbaric oxygen therapy in treatment of severe prevalently ischemic diabetic foot ulcer. A randomized study. Diabetes Care 1996;19: 1338e43. [31] Margolis DJ, Gupta J, Hoffstad O, Papdopoulos M, Glick HA, Thom SR, et al. Lack of effectiveness of hyperbaric oxygen therapy for the treatment of diabetic foot ulcer and the prevention of amputation: a cohort study. Diabetes Care 2013;36:1961e6. [32] Berendt AR. Counterpoint: hyperbaric oxygen for diabetic foot wounds is not effective. Clin Infect Dis 2006;43:193e8. [33] Boulton AJ, Vileikyte L, Ragnarson-Tennvall G, Apelqvist J. The global burden of diabetic foot disease. Lancet 2005;366:1719e24. [34] Bion JF, Edlin SA, Ramsay G, McCabe S, Ledingham IM. Validation of a prognostic score in critically ill patients undergoing transport. Br Med J 1985;291: 432e4 (Clin Res Ed).
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
Contents lists available at ScienceDirect
Journal of Infection and Chemotherapy journal homepage: http://www.elsevier.com/locate/jic
Original article
Diabetic foot infection in hospitalized adults David E. Katz a, N. Deborah Friedman b, Evgenia Ostrovski c, Dor Ravid d, Nadav Amrami d, Dori Avivi d, Bethlehem Mengesha d, Ronit Zaidenstein d, e, Tsilia Lazarovitch f, Mor Dadon d, Dror Marchaim d, e, * a
Department of Internal Medicine D, Shaare Zedek Medical Center, Jerusalem, Israel Department of Infectious Diseases, Barwon Health, Victoria, Australia c Department of Internal Medicine D, Assaf Harofeh Medical Center, Zerifin, Israel d Unit of Infectious Diseases, Assaf Harofeh Medical Center, Zerifin, Israel e Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel f Clinical Microbiology Laboratory, Assaf Harofeh Medical Center, Zerifin, Israel b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 October 2015 Received in revised form 15 November 2015 Accepted 16 December 2015 Available online xxx
Background: Acute infections of the diabetic foot (DFI) are a common and complex condition. Patients are generally managed in the ambulatory setting and epidemiological data pertaining to hospitalized patients is lacking. The aim of this study was to analyze the epidemiology, microbiology and outcomes of hospitalized patients with DFI, who are managed at a referral center equipped with hyperbaric oxygen (HBO) therapy. Methods: A retrospective cohort study of adult patients admitted to a tertiary referral center with DFI over a six-month period in 2013 was undertaken. Predictors of clinical outcomes and efficacy of treatment modalities were analyzed by Cox regression. Results: Sixty-one patients with DFI were identified. Most patients were elderly (67 ± 13 years), with long-standing (17 ± 9 years), poorly controlled (HbA1c 9 ± 3%) diabetes. Most patients had polymicrobial infection (80%); specifically, anaerobic (39%) and multi or extensively-drug resistant organisms (61%). Administration of appropriate antimicrobials was delayed for >48 h in 83%. Advanced age was associated with worse outcomes. Sicker patients with severe peripheral vascular disease were managed with HBO. The use of HBO was associated with higher costs and increased functional deterioration, and did not prevent future limb amputation. Conclusions: Our study illustrates the descriptive epidemiology of hospitalized adults with DFI predominantly of polymicrobial etiology. MDROs and anaerobic organisms are common causative pathogens, and appropriate antibiotics were frequently delayed. HBO treatment may delay the need for limb amputation, but not obviate this eventual outcome. © 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Keywords: DFI Skin and soft tissue infections Diabetes Hyperbaric oxygen SSTI Antimicrobial resistance
1. Introduction Diabetes mellitus affects 8e13% of the world population, and according to national data from the United States (US), has a prevalence of 27% in patients over 65 years [1]. Among patients with diabetes, diabetic foot infection (DFI) is one of the most common complications with up to 10% of diabetic patients
* Corresponding author. Unit of Infectious Diseases, AssafHarofeh Medical Center, Zerifin, 70300, Israel. Tel.: þ972 8 977 9049; fax: þ972 8 977 9043. E-mail address: [email protected] (D. Marchaim).
developing a DFI [1,2]. DFI presents with a range of signs and symptoms, and the severity of the DFI can range from a superficial infection of the nail bed to deeper infections involving bone (osteomyelitis) [3]. Although DFI is predominantly managed in the outpatient setting, in the developed world, DFI is the most common admission diagnosis for diabetic patients [4], and management in these cases frequently involves partial or total amputation of the affected lower limb. DFI is the most common reason for non-traumatic amputations in developed countries [2], with data from the US showing that up to 60% of all non-traumatic adult limb amputations in 2010 were related to DFI [1]. Survival following limb amputation for DFI is less
http://dx.doi.org/10.1016/j.jiac.2015.12.007 1341-321X/© 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
2
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
than 50% within 3 years after surgery, and only 40% at 5 years after surgery [2]. Partial or total limb amputation for DFI predicts further amputation in 20% of patients within 2 years of the procedure, and also predicts development of another DFI in the initially unaffected limb in 50% of patients [2]. Therefore, amputation is both a marker for serious DFI and of an unfavorable outcome. Treatment with hyperbaric oxygen (HBO) is one of few existing modalities for treating DFI in order to avoid or delay amputation [5]. However, there is ongoing debate regarding the independent role of HBO treatment in effectively reducing amputation rates over time [6]. Although the clinical and microbiological diagnosis of DFI may be both challenging and complicated [7], early and precise diagnosis is important for both effective infection management and for reducing the need for subsequent amputation [3,6]. Microbiological diagnosis is critical as it allows the clinician to tailor antibiotic therapy to the causative organism [8]. However, in many cases, the antimicrobial management for DFI is empiric as there may be no appropriate site from which to obtain a culture, superficial skin cultures correlate with the ‘true offending pathogen’ less than 30% of the time, and patients may also present already taking long-term antibiotics, which further decreases the culture yield [6,9]. Importantly, inappropriate empiric broad-spectrum antimicrobial therapy can result in unfavorable outcomes for individual patients and contribute to the broader harmful ecological impact of antimicrobial therapy [10]. The epidemiology of DFI in Israel has not been comprehensively researched [9,11,12]. Moreover, internationally, the epidemiological features of patients with DFI managed as inpatients is also lacking [8]. The goals of this study were to perform a descriptive analysis of the epidemiological and microbiologic features of patients with DFI admitted to a referral acute-care facility, to determine risk factors for a worse clinical course (i.e., death, amputation, re-admission, and functional deterioration), and to analyze the controlled effect of HBO treatment on the clinical course of patients with DFI. 2. Materials and methods 2.1. Study setting Assaf Harofeh Medical Center (AHMC) is an 815-bed tertiary care university-affiliated facility located in southern-central Israel. AHMC contains a state-of-the-art hyperbaric medicine centre (with dual chambers with capacity for 37 patients), and a multidisciplinary team with expertise in the management of DFI [13]. This study was approved by the institutional review board at AHMC according to the declaration of Helsinki statement of ethical principles for conducting human research. A retrospective review of charts of patients consecutively admitted to AHMC from May through November 2013 was undertaken as part of a broader project on infectious syndromes among hospitalized patients. During this time period, 8132 charts were reviewed. Patients over 18 years of age with a diagnosis of DFI based on their discharge records and ICD-9 discharge codes, were included. The diagnosis of DFI was per accepted definitions [6] and in consultation with a senior infectious disease physician. Microbiology samples for culture were obtained by tissue biopsy or collected via sterile syringe in the case of direct pus draining from a deep sinus tract [9]. 2.2. Definitions Multidrug resistant organisms (MDR) included; extendedspectrum b-lactamase producing Enterobacteriaceae (ESBLs), methicillin-resistant Staphylococcus aureus (MRSA), carbapenem-
susceptible Pseudomonas aeruginosa, and carbapenem-susceptible Acinetobacter baumannii. Extensively-drug resistant organisms (XDR) included; vancomycin-resistant enterococci (VRE), carbapenem-resistant Enterobacteriaceae (CRE), carbapenem-nonsusceptible Pseudomonas aeruginosa, and carbapenem-nonsusceptible Acinetobacter baumannii. 3. Data collection The following data was extracted from patient medical records; patient demographics, details of the DFI, history and duration of diabetes, co-morbid illnesses (including Charlson's scores [14]), functional status, previous hospital admissions, surgical procedures and use of invasive devices (e.g. urinary catheters, feeding tubes) prior to admission, medications, hospital length of stay, details of HBO treatment, vital signs, physical examination findings, laboratory results including bacterial culture results, and the results of imaging of the affected foot. In addition, post-discharge follow-up data were collected on both surgical procedures, and microbiology results up to 6 months after the hospitalization. Mortality data was extracted from national records of the Israeli ministry of interior. The University of Texas (UT) diabetic wound classification incorporating wound stage and appearance was calculated for all patients to estimate the severity of the DFI [15]. 3.1. Data management and analysis The data was entered and processed using SPSS version 22.0 (2014). Both descriptive and univariable analyses were performed. Multivariable models were created for four outcomes: 1) decrease in functional status following the DFI index event, 2) readmission to the acute care hospital within 6 months, 3) partial or total amputation of the limb affected resulting from the index DFI, and 4) increased cost (admissions costing in excess of 30,000 New Israeli Shekels [NIS], which equals ~$7700). Cost analyses were based on the total costs that were billed to insurance companies, based on the rates set by the Israeli ministry of health. These costs included; inpatient care, invasive procedures (including surgery), days of hospitalization (based on the type of unit), and imaging, set at fixed rates as determined by the Israeli ministry of health. HBO costs, if present, were deducted. An analysis was also performed comparing treatment groups (HBO versus no HBO). All analyses were performed by accepted methods. Categorical variables were analyzed by the chi-square test or Fisher's exact test; continuous variables were analyzed by the student's t-Test or ManneWhitney U test. Multivariable modeling was performed for each outcome, using Cox regression. Multivariable analysis of patient characteristics for those patients treated with HBO was performed using logistic regression. All parameters with a P-value less than or equal to 0.1 were incorporated to the multivariable analyses. 4. Results Of the 8132 consecutive patient discharge notes reviewed during the 6-month study period, we identified 61 adult patients with DFI. All DFIs were present on admission. The study population was male predominant, of mean age 67 years and more than half were functionally impaired at baseline, with high Charlson's scores [14]. Most patients had long-term diabetes under poor control and had been diagnozed with diabetic microvascular disease (Table 1). Approximately one-third of patients were
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
3
Table 1 Baseline epidemiologic features among hospitalized patients with DFI. Parameter Demographics Age, years, mean ± SD Female gender Elderly (over 65 years) Admitting Unit
Medicine Surgery Adult ICU
Parameters related to diabetes Diabetic microvascular disease (neuropathy/retinopathy/ nephropathy) Years of diabetes at DFI onset, mean ± SD HbA1c, percent, mean ± SD Acute and chronic conditions on admission Dependent Functional status Cognitive impairment Rapidly fatal McCabe score [34] Dementia Chronic lung disease Congestive heart failure Ischemic heart disease Malignancy (any type, past or present) Peripheral vascular disease Smoker Charlson's weighted index co-morbidity, median (range) Charlson's combined condition score, median (range) Charlson's 10-year survival probability, percent, median (range) Immunosuppressive states or conditions Glucocorticoid therapyb in the previous month Post transplantation (solid organ or bone marrow) Exposure to healthcare settings and environments before infection Resident of long-term care facility Hospitalized in the previous 3 months Past limb amputation Chronic hemodialysis Invasive procedure (including surgeries) in the previous 3 months Permanent urinary catheter Severity of illness indices at time of infection Reduced consciousness Severe sepsis, septic shock or multi-organ failure (as opposed to sepsis syndrome) Acute renal failure Antimicrobial therapy Antibiotics in the preceding 3 months Number of antibiotics administered in the preceding 3 months, median (range) Days from last antibiotic administered, median (range) 48 h delay in initiating appropriate therapy (per in-vitro report of microbiology laboratory) Days to initiation of appropriate antibiotics, median (range) Duration (in days) of antimicrobial treatment course for the index infectious episode, median (range) Most common empiric antimicrobial regimens Beta-lactam beta-lactamase inhibitor combinations Cephalosporins Fluoroquinolones Glycopeptides Clindamycin Metronidazole Most common maintenance antimicrobial regimens Beta-lactam beta-lactamase inhibitor combinations Cephalosporins Carbapenems Fluoroquinolones Glycopeptides Metronidazole a b
Valid percenta
Number 67 ± 13 22 35 19 41 1
36 57 31 67 2
55
90
17 ± 9 9±3 37 9 4 8 17 15 19 7 32 20 4 (1e9) 7 (1e13) 8 (0e96)
61 15 7 13 28 25 31 12 53 33
5 2
8 3
6 22 17 8 16 2
10 36 28 13 26 3
4 6
7 10
15
25
27 1 (0e3)
44
1 (0e33) 29
83
6 (0e44) 16 (0e55)
9 40 17 5 4 26
15 66 28 8 7 43
8 17 2 11 4 7
13 28 3 18 7 12
Prevalence calculated only for patients with data. Glucocorticoid therapy in the previous month was defined as receipt of an equivalent dose of at least 16 mg of prednisolone for 5 days or more.
recently hospitalized and nearly half of patients were prescribed antibiotics in the preceding three months (Table 1). Over half of patients had a UT diabetic wound classification score over IIB [15] (Table 2).
4.1. Microbiological results Forty of 61 patients (66%) had positive microbiological results from specimens collected. Most infections were polymicrobial, and
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
4
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
Table 2 University of Texas (UT) Diabetic wound classification score [15] for hospitalized patients with DFI. Wound grade
Number
Valid percenta
0eB 0eD IeA IeB IeC IeD IIeB IIeD IIIeB IIIeC IIIeD
6 1 1 14 2 1 5 16 8 1 6
9 1.5 1.5 20.6 2.9 1.5 7.4 23.5 11.8 1.5 8.8
a
Prevalence calculated only for patients with data.
over 60% of patients isolated a multidrug resistant (MDR) or extensively drug-resistant (XDR) bacteria (Table 3). 4.1.1. Antimicrobial therapy Patients received a median of 16 days of antimicrobial therapy (up to 55 days) and were commonly managed with combinations of agents both as empiric therapy and as consolidative regimens. Cephalosporins, metronidazole fluoroquinolones and beta-lactam/ beta-lactamase inhibitor combinations were most commonly prescribed (Table 1). 4.2. Patient outcomes Patient mortality was low, but functional deterioration occurred in one-third of patients (Table 4). More than half of patients were re-hospitalized, and one-quarter required amputation in the next 6 months. The median length of hospital stay was 10 days, with a maximum length of stay of 65 days. Median hospitalization costs were in excess of $6500 (Table 4). The multivariable models of patients' outcomes revealed that empiric use of metronidazole and treatment with HBO, were associated with worse outcomes. Increasing age was an independent risk factor for re-admission within 6 months after the index hospitalization due to DFI, and PVD was associated with increased costs (Table 5). 4.3. Hyperbaric oxygen (HBO) treatment Sixteen patients were treated with HBO per protocol [13]. Patients who underwent HBO were more likely to have PVD, a lower Charlson's survival probability [14], and over one-third of cases were dialysis-dependent. Higher numbers of amputations, increased hospital length of stay and overall costs (excluding HBOassociated costs) were observed in the patients who underwent HBO (Table 6). 5. Discussion This study describes a cohort of 61 patients with DFI who were hospitalized in a tertiary-care academic medical center. AHMC has a multi-disciplinary team for management of DFI including onsite HBO. The patient cohort consisted of an elderly group of patients, with functional impairment at baseline, a median 10-year survival probability of only 8%, and extensive exposure to antimicrobials, the healthcare environment, and procedures (Table 1). The suboptimal outcomes of patients, specifically, deterioration in functional status, future additional hospitalizations and eventual amputation of the affected limb were substantial among this hospitalized group of patients (Table 4).
Table 3 Microbiological results of 40 hospitalized patients with DFI. Parameter
Number
Valida percent
Monomicrobial Polymicrobial Anaerobic Patients with bacteremia ESBLb MDR organismsc XDR organismsd
8 32 16 3 5 18 7
20 80 39 8 12 44 17
Bacteria isolated Enterococcus faecalis Corynebacterium spp. Staphylococcus aureus MRSAe MSSAf Streptococcus dysgalactiae Streptococcus pyogenes Staphylococcus epidermidis Streptococcus agalactiae Streptococcus viridians group Peptococcus spp. Escherichia coli Pseudomonas aeruginosa Proteus mirabilis Klebsiella pneumoniae Morganella morganii Acinetobacter baumanii Enterobacter cloacae complex Citrobacter koseri Citrobacter freundii Alcaligenes faecalis Klebsiella spp. Providencia rettgeri Acinetobacter haemilyticus Aeromonas hydrophila Enterobacter spp. Klebsiella oxytoca Proteus penneri Providencia spp. Providencia stuartii Pseudomonas putida Raoultella ornithinolytica Serratia marcescens Bacteroides fragilis Prevotella oralis Bacteroides ovatus Bacteroides thetaiotaomicron Prevotella melaninogenica Bacteroides caccae Bacteroides stercoris Bacteroides uniformis Fusobacterium varium
Bacteria type Gram positives
Number 9 8
Gram negatives
Anaerobic bacteria
1 5 4 3 1 1 1 1 9 9 8 7 7 5 4 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1 8 5 4 3 3 1 1 1 1
a
Prevalence calculated for the 40 patients with positive microbiological results. ESBL ¼ extended-spectrum b-lactamase producing Enterobacteriaceae. MDR ¼ multidrug resistant organisms include: ESBLs, methicillin-resistant Staphylococcus aureus (MRSA), carbapenem-susceptible Pseudomonas aeruginosa, and carbapenem-susceptible Acinetobacter baumannii. d XDR ¼ extensively-drug resistant organisms include: vancomycin-resistant enterococci (VRE), carbapenem-resistant Enterobacteriaceae (CRE), carbapenemnonsusceptible Pseudomonas aeruginosa, carbapenem-nonsusceptible Acinetobacter baumannii. e MRSA ¼ methicillin-resistant Staphylococcus aureus. f MSSA ¼ methicillin-susceptible Staphylococcus aureus. b
c
The microbiology results in this study are worthy of note. MDR and XDR organisms (i.e., MDROs) were isolated in over 60% of patients. A French study found that 18% of patients with DFI had a MDRO isolated [16], while Kandemir et al. found that 40% of microbiological isolates from patients with DFI contained an MDRO [17]. In contrast to this, Citron et al. reported low rates of MDROs and found that ertapenem and piperacillin-tazobactam were each active against >98% of enteric gram-negative rods, methicillinsensitive S. aureus, and anaerobes from a study of over 400 patients with DFI in the United States [18]. The presence of MDROs led
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
5
Table 4 Outcomes of hospitalized patients with DFI. Parameter In hospital mortality 30 day mortality 90 day mortality Functional deterioration Discharged to long-term care facility after being admitted from home Additional hospitalizations in the 6 months following the index infection Amputation 6 months post infection onset Length of hospital stay from infection initiation to discharge after excluding those deceased, days, median (range) Hospital costs from infection initiation to discharge after excluding those deceased, US dollar, median (range) a
Valid percenta
Number 2 1 5 17 7 34 13 10 (1e65) $6648 ($561e$69,376)
3 2 8 29 13 61 23
Prevalence calculated only for patients with data.
Table 5 Multivariable models of DFI outcomes. Variable
Treated with HBO Anaerobic pathogen Empiric regimen containing metronidazole Chronic renal failure PVD Charlson's comorbidity combined condition score [14] Age in years Days on antibiotics
Additional hospitalizations in next 6 months
Amputation in next 6 months
Direct costs > 30,000 NIS ($7700) from infection to discharge
OR
CI-95%
OR
P value
CI-95%
OR
P value
CI-95%
0.02 0.09 0.1 0.93 0.6 0.05
1.4e111 0.8e50 0.6e68 0.1e14.4 0.1e34 1.1e2.3
3.9
0.1
0.75e20
2e48 0.96e14
12.4 6.2 6.6 1.1 2.1 1.5
4.4
0.04
1.1e17.5
7.3
0.01
1.6e33
0.96 1.1
0.09 0.04
0.9e1.1 1.01e1.11
9.7 3.7
1.1
P value
0.005 0.06
0.005
1.03e1.17
Note. NIS ¼ new Israeli Shequels. HBO ¼ hyperbaric oxygen. PVD ¼ peripheral vascular disease. Th bold parameters are those with statistically significant associations. Table 6 Epidemiologic factors and infection outcomes of patients with DFI who were treated with hyperbaric oxygen. Parameter Demographics Age, years, mean ± SD Female gender Elderly (over 65 years) Acute and chronic conditions on admission Dependent functional status Cognitive impairment Rapidly fatal McCabe score [34] Peripheral vascular disease Years of diabetes, mean ± SD Hb A1C, percent, mean ± SD Charlson's combined condition score [14] Charlson's 10-year survival probability [14] Exposure to healthcare settings and environments before infection Resident of long-term care facility Past amputation Chronic hemodialysis Invasive procedure in the previous 3 months Severity of illness indices at time of infection Severe sepsis, septic shock or multi-organ failure ICU stay in current hospitalization Development of acute renal failure Outcomes In hospital mortality 30 day mortality 90 day mortality Functional deterioration Discharged to long-term care facility after being admitted from home Amputation in the 6 months following the onset of infection Length of hospital stay from infection diagnosis to discharge after excluding those deceased, mean ± SD Cost from infection diagnosis to discharge, US dollars, mean ± SD
Treated with hyperbaric oxygen Not treated with hyperbaric oxygen OR (n ¼ 16), N (%) (n ¼ 45), N (%)
CI-95%
P value
66 ± 9 8 (50) 7 (44)
67 ± 14 14 (31) 28 (62)
10 (63) 0 (0) 0 (0) 13 (81) 17 ± 10 9±3 8±2 7 ± 18
27 (60) 4 (9) 4 (9) 19 (42) 17 ± 10 9±2 7±3 23 ± 34
1.1 0.9 0.9 5.9
0.3e3.6 >0.99 0.8e1.0 0.6 0.8e1.0 0.6 1.5e23.8 <0.001 0.8 0.9 0.2 <0.001
0 (0) 6 (38) 6 (38) 3 (19)
6 (13) 11 (24) 2 (4) 13 (29)
0.9 1.9 13.0 0.6
0.8e1.0 0.3 0.6e6.3 0.3 2.3e73.6 <0.001 0.1e2.3 0.5
0 (0) 0 (0) 4 (25)
6 (13) 1 (2) 11 (24)
0 (0) 0 (0) 1 (6) 4 (25) 1 (6) 9 (56) 21 ± 18
2 (4) 1 (2) 4 (9) 13 (30) 6 (15) 4 (10) 11 ± 8
17,499 ± 17,360
7224 ± 5081
2.2 0.7e7.1 0.5 0.2e1.5
0.9 0.8e1.0 1.0 0.9e1.0 1.0 0.3e3.9 1.0 1.0 0.7 0.8 0.4 11.6
0.9e1.0 0.9e1.0 0.1e6.6 0.2e2.8 0.1e3.3 2.8e48.3
0.7 0.2 0.3
0.3 >0.99 >0.99 >0.99 >0.99 >0.99 0.8 0.7 <0.001 <0.001 <0.001
The bold parameters are those with statistically significant associations.
to a median delay of 6 days before commencing appropriate antibiotic therapy in this study. In over 80% of patients, appropriate treatment was delayed for over 48 h. Delay in appropriate antibiotic commencement is the strongest independent predictor for worse
outcomes in septic patients [19]. Ensuring earlier collection of appropriate specimens for culture will likely reduce inappropriate antimicrobial exposure, MDRO colonization pressure, and may improve patient clinical outcomes [7,20,21].
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
6
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7
There are several factors which were identified that may have contributed to the high prevalence of MDROs among these patients. As found by others, previous exposure to antimicrobials was present in over 40% of patients, which would contribute to selective pressure [17,22]. In addition, among patients in this study there was a high prevalence of established risk factors for MDROs including; older age, complex co-morbidities, institutionalization, functional dependency, in addition to healthcare exposure in a country with high endemic rates of MDROs in hospitals [20,23e28]. Other studies have also found that previous hospitalization and healthcare exposures were risk factors for MDROs [16,17]. Interestingly and in keeping with other studies, many patients were exposured to both long-term care and to hemodialysis [29]. Moreover, the specific overuse of cephalosporins, fluoroquinolones and metronidazole (Table 1) have been found to be predictors of the emergence of MDROs [29]. We also observed a high prevalence of polymicrobial infections, and frequent isolation of anaerobic pathogens, as the causative pathogens of DFI (Table 3). These findings are in keeping with data from a large DFI cohort with over 80% polymicrobial infections, and nearly half growing only anaerobes [18]. The prevalence of anaerobic pathogens might even be underestimated, due to suboptimal anaerobic conditions in specimen transport and processing. Although empiric treatment with metronidazole was found to be an independent risk factor for future amputations and for high medical costs (Table 5), this was a confounder, as empiric treatment including metronidazole was prescribed to patients with higher severity of illness (data not shown). Guidelines for appropriate management of DFI recommend broad-spectrum empiric treatment including anaerobic and MDRO coverage [5]. Proper culture acquisition is however vital prior to substantial exposure to antibiotics to allow for appropriate adjustment of the treatment regimen, which is frequently prolonged in deep or invasive DFI (associated with osteomyelitis) [6]. Sixteen patients (26%) were treated with HBO. This small number does not allow for a significant analysis of patient risk factors and clinical outcomes. The independent association of HBO treatment with future amputations (Table 5) relates primarily to an inability to adjust adequately for confounding effects. We were however able to conclude that patients at AHMC that were referred for HBO had a higher burden of underlying diseases and both poor functional status and a low 10 year survival probability compared with patients not managed with HBO (Table 6). These patients referred for HBO likely had a DFI that required amputation during this hospital stay. HBO was likely being used as a “last-ditch” effort to prevent amputation. Our results suggest that HBO may delay inevitable amputation in this group of patients with DFI and severe peripheral vascular disease. Research by Faglia and colleagues revealed that adjunctive HBO reduced amputations in patients with ischemic diabetic foot ulcers, while other research has found that use of HBO neither improved the likelihood that a wound would heal nor prevented amputation [30,31]. Furthermore, some believe that HBO should not be utilized in DFI until large-scale, adequately blinded, controlled, and powered randomized studies have clearly demonstrated efficacy and cost effectiveness in the healing of ulcers and the prevention of major amputation [32]. Unfortunately, as we observed in our study, delaying amputation with HBO resulted in prolonged inpatient care and increased costs. The utility and value of HBO in DFI therefore needs to be further explored. The limitations of this study are that it is retrospective, and therefore, calculation of wound and comorbidity scores may have been inaccurate. Moreover, while we did not capture specific data regarding patients being assigned a diagnosis of osteomyelitis, our calculation of the UT wound score we believe has enabled an estimate of wound severity and depth. In addition, we acknowledge
that comprehensive cost analyses ought to also include the cost of social services, home care, rehabilitation and subsequent episodes of care for DFI. While the numbers of this study are relatively small, it provide an insight into the epidemiology of inpatients with DFI, while capturing and analyzing outcomes such as functional deterioration following the event, further hospitalizations and amputations in an aging diabetic cohort. DFI places a large fiscal burden on the healthcare system [33]. We identified a median cost of hospitalization of $6,648, but with costs ranging as high as $69,376. Larger prospective caseecontrol and cohort studies will be able to identify optimal treatments including the use of HBO and the optimal timing of surgery if required. Nonetheless this study provides a rationale for the early institution of broad-spectrum (including anaerobic) antimicrobial empiric therapy for acute DFI, while obtaining appropriate samples early for microbiological testing. 6. Financial support This study was partially supported financially by Merck Inc. The company had no access or involvement in data collection, data interpretation, and was not involved in the process of drafting this manuscript. 7. Potential conflicts of interest All authors report no conflicts of interest relevant to this article. Acknowledgments None. References [1] Association AD, editor. Statistics about diabetes: data from the national diabetes statistics report; 2014. 2014. [2] Zandman-Goddard G, Feldbrin Z, Ovadia S, Zubkov T, Lipkin A, Wainstein J. A multi-disciplinary approach to diabetic foot patientsean organizational model for the treatment of leg complications in diabetic patients. Harefuah 2011;150:593e5. 616. [3] Bader MS. Diabetic foot infection. Am Fam Physician 2008;78:71e9. [4] Lipsky BA. Osteomyelitis of the foot in diabetic patients. Clin Infect Dis 1997;25:1318e26. [5] Peters EJ, Lipsky BA, Berendt AR, Embil JM, Lavery LA, Senneville E, et al. A systematic review of the effectiveness of interventions in the management of infection in the diabetic foot. Diabetes Metab Res Rev 2012;28(Suppl. 1): 142e62. [6] Lipsky BA, Berendt AR, Cornia PB, Pile JC, Peters EJ, Armstrong DG, et al. 2012 infectious diseases society of America clinical practice guideline for the diagnosis and treatment of diabetic foot infections. Clin Infect Dis 2012;54: e132e73. [7] Giurini JM, Lyons TE. Diabetic foot complications: diagnosis and management. Int J Low Extrem Wounds 2005;4:171e82. [8] Lipsky BA. A current approach to diabetic foot infections. Curr Infect Dis Rep 1999;1:253e60. [9] Slater RA, Lazarovitch T, Boldur I, Ramot Y, Buchs A, Weiss M, et al. Swab cultures accurately identify bacterial pathogens in diabetic foot wounds not involving bone. Diabet Med 2004;21:705e9. [10] Dellit TH, Owens RC, McGowan Jr JE, Gerding DN, Weinstein RA, Burke JP, et al. Infectious diseases society of America and the society for healthcare epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship. Clin Infect Dis 2007;44:159e77. [11] Tamir E. Treating the diabetic ulcer: practical approach and general concepts. Isr Med Assoc J 2007;9:610e5. [12] Weiss A, Karpf A, Luger E, Schmilowitz H, Dekel S, Shapira I. Long-term antibiotic treatment in geriatric diabetic foot infection. J Med 1998;29: 365e73. [13] Efrati S, Gall N, Bergan J, Fishlev G, Bass A, Berman S, et al. Hyperbaric oxygen, oxidative stress, no bioavailability and ulcer oxygenation in diabetic patients. Undersea Hyperb Med 2009;36:1e12. [14] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373e83.
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007
D.E. Katz et al. / J Infect Chemother xxx (2016) 1e7 [15] Lavery LA, Armstrong DG, Harkless LB. Classification of diabetic foot wounds. J Foot Ankle Surg 1996;35:528e31. [16] Hartemann-Heurtier A, Robert J, Jacqueminet S, Ha Van G, Golmard JL, Jarlier V, et al. Diabetic foot ulcer and multidrug-resistant organisms: risk factors and impact. Diabet Med 2004;21:710e5. [17] Kandemir O, Akbay E, Sahin E, Milcan A, Gen R. Risk factors for infection of the diabetic foot with multi-antibiotic resistant microorganisms. J Infect 2007;54: 439e45. [18] Citron DM, Goldstein EJ, Merriam CV, Lipsky BA, Abramson MA. Bacteriology of moderate-to-severe diabetic foot infections and in vitro activity of antimicrobial agents. J Clin Microbiol 2007;45:2819e28. [19] Paul M, Shani V, Muchtar E, Kariv G, Robenshtok E, Leibovici L. Systematic review and meta-analysis of the efficacy of appropriate empiric antibiotic therapy for sepsis. Antimicrob Agents Chemother 2010;54:4851e63. [20] Bonten MJ. Colonization pressure: a critical parameter in the epidemiology of antibiotic-resistant bacteria. Crit Care 2012;16:142. [21] Edmonds M, Foster A. The use of antibiotics in the diabetic foot. Am J Surg 2004;187:25Se8S. [22] Baquero F, Negri MC, Morosini MI, Blazquez J. Antibiotic-selective environments. Clin Infect Dis 1998;27(Suppl. 1):S5e11. [23] Schwaber MJ, Carmeli Y. An ongoing national intervention to contain the spread of carbapenem-resistant enterobacteriaceae. Clin Infect Dis 2014;58: 697e703. [24] Abbo A, Carmeli Y, Navon-Venezia S, Siegman-Igra Y, Schwaber MJ. Impact of multi-drug-resistant acinetobacter baumannii on clinical outcomes. Eur J Clin Microbiol Infect Dis 2007;26:793e800. [25] Bogan C, Kaye KS, Chopra T, Hayakawa K, Pogue JM, Lephart PR, et al. Outcomes of carbapenem-resistant enterobacteriaceae isolation: matched analysis. Am J Infect Control 2014;42:612e20.
7
[26] Cosgrove SE, Qi Y, Kaye KS, Harbarth S, Karchmer AW, Carmeli Y. The impact of methicillin resistance in staphylococcus aureus bacteremia on patient outcomes: mortality, length of stay, and hospital charges. Infect Control Hosp Epidemiol 2005;26:166e74. [27] Schwaber MJ, Carmeli Y. Antimicrobial resistance and patient outcomes: the hazards of adjustment. Crit Care 2006;10:164. [28] Schwaber MJ, Navon-Venezia S, Kaye KS, Ben-Ami R, Schwartz D, Carmeli Y. Clinical and economic impact of bacteremia with extended- spectrum-betalactamase-producing enterobacteriaceae. Antimicrob Agents Chemother 2006;50:1257e62. [29] Tacconelli E, De Angelis G, Cataldo MA, Mantengoli E, Spanu T, Pan A, et al. Antibiotic usage and risk of colonization and infection with antibioticresistant bacteria: a hospital population-based study. Antimicrob Agents Chemother 2009;53:4264e9. [30] Faglia E, Favales F, Aldeghi A, Calia P, Quarantiello A, Oriani G, et al. Adjunctive systemic hyperbaric oxygen therapy in treatment of severe prevalently ischemic diabetic foot ulcer. A randomized study. Diabetes Care 1996;19: 1338e43. [31] Margolis DJ, Gupta J, Hoffstad O, Papdopoulos M, Glick HA, Thom SR, et al. Lack of effectiveness of hyperbaric oxygen therapy for the treatment of diabetic foot ulcer and the prevention of amputation: a cohort study. Diabetes Care 2013;36:1961e6. [32] Berendt AR. Counterpoint: hyperbaric oxygen for diabetic foot wounds is not effective. Clin Infect Dis 2006;43:193e8. [33] Boulton AJ, Vileikyte L, Ragnarson-Tennvall G, Apelqvist J. The global burden of diabetic foot disease. Lancet 2005;366:1719e24. [34] Bion JF, Edlin SA, Ramsay G, McCabe S, Ledingham IM. Validation of a prognostic score in critically ill patients undergoing transport. Br Med J 1985;291: 432e4 (Clin Res Ed).
Please cite this article in press as: Katz DE, et al., Diabetic foot infection in hospitalized adults, J Infect Chemother (2016), http://dx.doi.org/ 10.1016/j.jiac.2015.12.007