- Email: [email protected]
Elderly patients are at increased risk for treatment failure in outpatient management of purulent skin infections
Elderly patients are at increased risk for treatment failure in outpatient management of purulent skin infections John P. Haran MD, Eric Wilsterman BS, Tyler Zeoli BS, Francesca L. Beaudoin MD, MS, Jennifer Tjia MD, MSCE, Patricia L. Hibberd MD, PhD PII: DOI: Reference:
S0735-6757(16)30776-8 doi:10.1016/j.ajem.2016.10.060 YAJEM 56233
To appear in:
American Journal of Emergency Medicine
Received date: Revised date: Accepted date:
14 June 2016 24 October 2016 28 October 2016
Please cite this article as: Haran John P., Wilsterman Eric, Zeoli Tyler, Beaudoin Francesca L., Tjia Jennifer, Hibberd Patricia L., Elderly patients are at increased risk for treatment failure in outpatient management of purulent skin infections, American Journal of Emergency Medicine (2016), doi:10.1016/j.ajem.2016.10.060
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Elderly Patients are at Increased Risk for Treatment Failure in Outpatient Management of Acute Cutaneous Abscesses
PT
John P. Haran, MD1, Eric Wilsterman, BS1, Tyler Zeoli, BS1, Francesca L. Beaudoin, MD, MS2, Jennifer Tjia, MD, MSCE 3, and Patricia L. Hibberd, MD, PhD4
SC
RI
Affiliations: From the Department of Emergency Medicine, University of Massachusetts Medical School, Worcester, MA1 Department of Emergency Medicine, Alpert School of Medicine, Brown University, Providence, RI2 Department of Quantitative Health Sciences, University of Massachusetts Medical School, Worcester, MA3 Department of Pediatrics and Division of Global Health, Massachusetts General Hospital, Boston, MA4
NU
Corresponding Author: John P. Haran, MD, 55 Lake Avenue North, Worcester, MA 01655; Phone: 508450-8688; email: [email protected]; fax: 508-421-1490
MA
Meetings: None
D
Grant: This study was designed and carried out at the University of Massachusetts Medical School and was supported by an intradepartmental grant through the Department of Emergency Medicine. PLH is supported by an NIH award 5K24AT003683-09.
TE
Conflicts of Interest: None of the authors listed have any conflict of interest.
AC CE P
Word Count: Abstract 250, Main Article 3,389 Author Contributions: JPH, JT, and PLH conceived the study, designed the trial, and obtained research funding. JPH, EW, and TZ supervised the conduct of the trial and data collection. JPH, EW and TZ recruited participating centers and patients and managed the data, including collection and quality control. JPH and FLB provided statistical advice on study design and analyzed the data; JPH drafted the manuscript, and all authors contributed substantially to its revision. JPH takes responsibility for the paper as a whole.
ACCEPTED MANUSCRIPT Elderly Patients are at Increased Risk for Treatment Failure in Outpatient Management of Purulent Skin Infections
PT
Objective: Current Infectious Disease Society of America (IDSA) guidelines for the management of
RI
purulent skin or soft tissue infections do not account for patient age in treatment recommendations. The
SC
study objective was to determine if age was associated with outpatient treatment failure for purulent
NU
skin infection after adjusting for IDSA treatment guidelines.
MA
Methods: We conducted a multicenter retrospective study of adult patients treated for a purulent skin infection and discharged home from four emergency departments between April and September 2014.
D
Patients were followed for one month to assess for treatment failure (defined as need for a change in
TE
antibiotics, surgical intervention, or hospitalization). We used multivariable logistic regression to examine the role of patient age on treatment failure adjusting for demographic variables (gender, race),
AC CE P
comorbidities and severity of infection.
Results: A total of 467 patients met inclusion criteria (mean age 37.9 years [SD 14.0], 48.2% of whom were women). Overall, 12.4% failed initial therapy. Patients 65 years and older (n=35) were almost 4 times more likely to fail initial ED therapy in follow-up compared with younger patients (adjusted Odds Ratio (OR) 3.87, 95% Confidence Interval (CI) 1.24-12.10). After adjustment, for every 10 years of advancing age there was a 43% increased odds of failing initial treatment (OR 1.43 95% CI 1.09-1.88).
Conclusion: Elderly patients with purulent skin infections, whose providers followed the 2014 IDSA guidelines, were more likely to fail initial treatment than younger patients. This study suggests that there is a need to re-evaluate treatment guidelines in elderly patients.
ACCEPTED MANUSCRIPT Introduction The burden of annual visits and total hospital admissions for skin and soft tissue infections (SSTIs) has
PT
increased dramatically in the past decade.1, 2 Between 1993 and 2005, annual visits for SSTIs in U.S.
RI
emergency departments ED increased from 1.2 million to 3.4 million.3 In response, the Infectious Disease Society of America (IDSA) recently updated their guidelines for the management of skin and soft tissue
SC
infection.4 The 2014 guidelines departed from previous versions by creating three SSTI treatment
NU
categories (mild, moderate, and severe) with specified treatment regimens for both purulent and nonpurulent infection types. These categories of severity are defined mainly by the patient’s medical
MA
comorbidities (such as those with markedly impaired host defenses), recent antibiotic use, and severity of presenting symptoms. Current treatment recommendations do not take a patient’s age into
TE
D
consideration. However, age-related biologic changes such as immune system dysregulation, and physiologic changes such as increased dryness and decreased blood flow to the skin can diminish the
AC CE P
body’s innate ability to fight bacterial infections.5 The objectives of this study were to: 1) describe the frequency with which outpatient treatment failure occurs among older and younger adults treated for a purulent skin infection while comparing the ED treatment to the recommended IDSA guidelines and 2) evaluate whether age is an independent risk factor for treatment failure of purulent SSTI in patients treated in the ED and discharged to home.
ACCEPTED MANUSCRIPT Methods Study design
PT
We conducted a multi-center, retrospective cohort study. This study was approved by the institutional
RI
review board (IRB docket H00007714).
SC
Study setting and population
NU
We identified patients 18 years and older who presented to one urban tertiary care academic center with an annual ED census of 132,000 visits and three community emergency departments with a combined
MA
annual ED census of 100,000 visits for treatment of a purulent skin infection between April and September in 2014. Patients were eligible for participation if they were discharged from the ED or the ED
TE
D
observation unit with a diagnosis of a skin or soft tissue purulent infection pulled from the electronic medical records. All patients were classified as having a suspected purulent infection based on discharge
AC CE P
diagnoses of abscess. Patients were further classified as having a confirmed purulent infection if they underwent an incision and drainage in the ED with document purulent material. Patients were excluded if their purulent skin infection had undergone an incision and drainage within the previous month in the ED, if they were admitted to the hospital for this problem on their initial visit, or if they did not have follow-up visit data. Additionally, we excluded patients with: a post-surgical infection, a purulent skin infection of the oral cavity, hidradenitis suppuritiva, or a Bartholins gland abscess. Both patients with hidradenitis or a Bartholins abscess were excluded because they are not expected to improve without further surgical intervention.
ACCEPTED MANUSCRIPT Data collection In order to reduce the potential for systematic error and to mitigate bias, we followed protocols for the
PT
optimal conduct of retrospective studies.6 Prior to data abstraction activities, we a priori defined the
RI
pertinent predictor and outcome variables to be collected in a standardized manner. Abstractors were uniformly trained by the investigators and blinded to the study objectives and hypotheses. We utilized
SC
two different abstractors for each patient enrolled, the first one collecting data on the initial ED visit,
NU
recording demographic, historical and clinical data, while the second abstractor collected all follow-up visit data that included the principal study outcome variables. Abstractors met regularly with the
MA
investigators to review the coding rules. The investigators performed an interrtater reliability
Variable measurement
TE
D
assessment on a 10% random sample of charts.
AC CE P
Baseline variables were extracted from the initial ED visit including: sociodemographic characteristics, medical history, chief complaint, and previous antibiotic use. Patients already on antibiotics from a source outside of the ED were included (e.g., primary care office). The Charlson comorbidity index (CCI) was calculated and used to rank patient’s medical comorbidities.7, 8 Information pertaining to initial ED presentation, ED course, surgical, and antibiotic treatments used were recorded. Starting from this initial visit, all subsequent medical visits pertaining to this index infection were reviewed to assess for outcomes including: 1) change in antibiotic treatments, 2) unplanned surgical intervention to further treat the purulent skin infection, and 3) inpatient hospitalization or admission to ED observation unit for continued infection treatment. We created a composite outcome of ED treatment failure based on these three outcomes. Follow-up chart review was performed by a different abstractor using electronic medical records recording treatment failure one month after the initial visit. We did not include routine follow-up
ACCEPTED MANUSCRIPT visits (e.g. wick removal, wound checks without a change in care) in the outcome measurement. All patients without documentation of treatment failure but with continued documentation or repeat visits
RI
PT
in the medical record were classified as clinically cured.
Infectious Disease Society of America classification
SC
We used IDSA guidelines to classify the severity of each patient’s purulent skin infection clinical
NU
presentation and the observed ED treatments. Each case was categorized into one of 3 treatment categories of purulent SSTIs: mild, moderate, or severe. In brief, these guidelines define mild as an
MA
uncomplicated infection, a moderate infection as one where the patient has signs of a systemic inflammatory response (SIRS), and a severe infection when patients have either an impaired host
TE
D
defense, have failed previous antibiotic therapy, or have had SIRS plus hypotension. 4, 9 We defined SIRS as the presence of one of the following: heart rate >90 beats per minute, respiratory rate >20 breaths per
AC CE P
minute, temperature >38.0 °C, white blood count > 12,000 cells/mm3 or presence of greater than 10% immature neutrophils. The presence or absence of antibiotics and the class of antibiotics prescribed were used to categorize treatment into mild, moderate, or severe therapy class according to IDSA treatment recommendations.4 We used the 2014 IDSA guidelines to compare the observed antibiotic treatment regimens to the expected according to these guidelines. Incision and drainage was recommended for all purulent skin infections regardless of class. Accordingly, each patient received a score of 1 for mild, 2 for moderate, and 3 for severe for that initial ED visit presentation and a second score for the type of treatments rendered. We then compared the clinical presentation and treatment scores with each other to determine if the observed matched the anticipated treatment class. We categorized patients into expected treatment when the scores equaled each other, under-treated if the observed score was lower than the anticipated score, and over-treated when the observed score was higher than the anticipated
ACCEPTED MANUSCRIPT score. This scoring system has been reported on previously.10
PT
Data analysis
RI
We used chi-square tests to compare categorical variables, and the student’s t-test for continuous variables, between patients with clinical cure and those that failed therapy. Given the large sample size,
SC
data were assumed to be normally distributed. We used multivariable logistic regression analysis to test
NU
whether or not increasing age (the main exposure variable) was associated with treatment failure. To select the set of covariates for the multivariable model, we first determined important a priori covariates
MA
that we thought would have clinical relevance. Patient age was the main variable of interest. Patient demographics including gender and race, patient comorbidities (CCI), and IDSA infection severity class
TE
D
(mild, moderate, or severe), provider adherence to IDSA guidelines as described above (or if they did not by either over or under-treating the patient), and treatment variables such as if an incision and drainage
AC CE P
was performed or if they received any antibiotics were also included. Finally, any covariates with a p<0.10 from our unadjusted bivariate analysis were included. IDSA treatment was categorized into either matched, over, or under-treated while age was categorized into elderly or non-elderly using the cutoff of age ≥ 65 years. We then performed analysis stratified in cases of suspected versus confirmed purulent infection to examine whether or not this modified the effect estimate. To analyze the affect of patients lost to follow-up we performed a sensitivity analysis using best-case and worst-case scenarios where we reran the model with the patients that were lost were coded as either none having failed therapy or all of them failing therapy. We used Stata Version 13.1 (StataCorp LP, College Station, TX) for all analyses.
ACCEPTED MANUSCRIPT Results Characteristics of the study subjects
PT
During the 6-month study period over the four ED sites, there were a total of 2,190 ED visits for a
RI
complaint of a skin or soft tissue infection of which 33.1% were diagnosed with a purulent skin infection and 13.5% were admitted (Figure 1). Among the this cohort there were a total of 627 patients with a
SC
diagnosis of purulent skin infection who were discharged home from the ED. We excluded 28 patients
NU
(4.5%) with an abscess related to a dental infection, 9 patients (1.4%) with a bartholins gland abscess, 6 patients (1.0%) with hidradenitis suppurativa, and 12 patients (1.9%) with a post-surgical infection.
MA
There were 105 patients (18.4%) with an average age of 34.4 years who did not have follow-up data available. The average age of the study sample was 37.9 years, 48.2% were women, with 59.5% of
TE
D
patients seen at the community sites. Among the cohort 362 (77.5%) patients had an incision and drainage performed confirming purulent material and thus were labeled as confirmed cases. We had a
AC CE P
interrtater reliability kappa for all variables used in this analysis of 0.92.
Among all 2,190 ED visits for a SSTI (included both abscess and cellulitis patients), patients 65 years and older had significantly lower prevalence of purulent skin infection diagnosis (11.4% vs. 39.3%, p<0.001) and higher prevalence of admission to the hospital (42.7% vs. 20.5%, p<0.001) for a cutaneous infection. Elderly patients presenting to the ED were 5 times more likely to be diagnosed with a non-purulent bacterial skin infection rather than purulent infection (unadjusted OR 5.0, 95% CI 3.8, 6.8).
Our final study cohort consisted of 467 patients of which 58 (12.4%) patients had a failure of initial ED therapy upon follow-up. A low percentage of patients presented already on antibiotics (15.4%) from a previous non-ED healthcare visit. Patients that failed therapy after their ED visit were older in age, had
ACCEPTED MANUSCRIPT higher CCI scores, and had a purulent skin infection involving the structures of the hand (Table 1). For every 10-year increase in patient age, the odds of failing therapy increased by 22% (OR 1.22, 95% CI
PT
1.01-1.47). With regard to the CCI score, patients with scores of 4 or greater had 13 times the odds of
RI
failing therapy (OR 13.62, 95% CI 1.21-153.12) compared to those with scores of 0. Purulent skin
SC
infection involving the hand were 2.3 increased odds to fail (OR 2.26, 95% CI 1.00-5.26).
NU
Characteristics of initial ED treatment therapies comparing those with treatment failure During the course of their initial ED visit, patients with a purulent infection did not receive an incision
MA
and drainage 22.5% of the time and received antibiotics 72.6% of the time. Intravenous antibiotics were used in 12.8% of patients with 2.6% receiving only one intravenous dose in the ED without being sent
TE
D
home with a prescription. ED providers used antibiotic coverage for Methicillin-resistant Staphylococcus
AC CE P
aureus (MRSA) only 72.6% of the time.
In comparing the treatment options among patients that went on to fail therapy compared to those did not (Table 2), there was no significant difference in the rates of treatment failure among the difference antibiotic classes with the exception of those that received intravenous vancomycin. Patient that received an intravenous dose of vancomycin were at a 3.2 increased odds of failing ED therapy that those that did not receive this antibiotic (OR 3.20, 95% CI 1.45-7.07). Patients that received vancomycin within the ED went home without MRSA antibiotic coverage 31.0% of the time. There was also no difference in treatment failure rates if the ED provider gave intravenous antibiotics before discharge or if the antibiotic class selection included MRSA coverage.
ACCEPTED MANUSCRIPT Comparison with national guidelines Physicians followed the IDSA guidelines for treatment of purulent infections only 40.0% of the time;
PT
patients were over-treated 40.9% of the time and under-treated 20.1% of the time. Among patients
RI
classified by IDSA as having a mild infection requiring only an incision and drainage, they received unnecessary antibiotics 69.9% of the time; while not undergoing and incision and drainage in 23.0% of
SC
cases. The majority of patients had a mild infection according to IDSA guidelines; however, we did not
NU
detect any differences in treatment failure rates among the three IDSA classes (Table2). Additionally, we did not see any differences in treatment failure rates if the provider either followed the IDSA guidelines
TE
Characteristics of elderly patients
D
MA
or if they over or under-treated the patient according to these guidelines.
Elderly patients did not differ from their younger counterparts with regard to the infection location, if
AC CE P
they received an incision and drainage, or if they were placed on antibiotics, either within the ED or at home, as part of their care. Fewer elderly patients had a CCI score of 0 (57.1%) than younger patients, whereas a greater proportion of older patients had a score of ≥3 (11.4%) than their younger counterparts (80.3% and 1.4% respectively). A greater proportion of elderly patients failed ED outpatient therapy upon follow up (25% versus 11.7%, p=0.03). Elderly patients had a 2.5 increased odds for failed initial ED treatment than persons less than 65 years old (OR 2.51, 95% CI 1.06-5.92). Even though patients with higher CCI scores were older we did not observe any significant difference in treatment failure rates with an age CCI score interaction.
ACCEPTED MANUSCRIPT Logistic regression model In our multivariable logistic regression, patient’s age was significantly associated with treatment failure
PT
among both suspected and confirmed cases after adjustment for baseline demographic and clinical
RI
variables that could have affected treatment choice and related outcomes (Table 3). Higher CCI scores and ED use of intravenous vancomycin were also significantly associated with ED treatment failure. For
SC
each increase of patient age by 10 years, the odds of treatment failure increased by 43% to 62% with
NU
patients in their 20s having a 7.0% chance of treatment failure that rose to 27.2% in patients in their 70s or older (Figure 2). We reran this model changing the continuous age covariate into elderly (age ≥ 65
MA
years) and non-elderly patients. Elderly patients of suspected cases had almost 4 times greater odds of failure of initial ED treatment (OR 3.87, 95% CI 1.24-12.10) than non-elderly patients while elderly
TE
D
patients among confirmed cases had 8 times greater odds of failure of initial ED treatment (OR 8.38, 95% CI 2.30-30.49) after adjusting for the other covariates in the model which included IDSA treatment
AC CE P
classification. The significance of elderly patients having greater odds of failing treatment did not change in our sensitivity analysis.
Study Strengths and Limitations
To our knowledge this is the first study to provide evidence for poorer outcomes in elderly patients with purulent skin infections after adjustment for initial treatment guideline adherence. This study builds on prior study demonstrating poor adherence to IDSA guidelines by ED providers.10 The strengths of this study include the quality of the data extractions from multiple ED sites. There are, however, several limitations that must be kept in mind in interpreting our study results. There exists the possibility of selection bias from loss to follow-up however the main affect of age on the outcome did not change during our sensitivity analysis. For example, if the majority of cases lost to follow-up were patients who
ACCEPTED MANUSCRIPT failed treatment, the true effect would still be statistically significant (there still would exist a difference between elderly and non-elderly patients). This study was also limited by reliance of discharge diagnosis
PT
coding of purulent skin infection to identify the study cohort. This may partially explain why 1 in 5
RI
patients did not receive an incision and drainage of their diagnosed purulent skin infection. Providers may have been limited in localizing the purulent skin infection by ultrasound or unsure of their diagnosis.
SC
We did not have a confirmatory test to ensure all patients in the cohort did in fact have a purulent
NU
infection however we observed the same association of age and treatment failure among the cases with suspected and confirmed purulent infections in our stratified analysis. Further, factors such as purulent
MA
skin infection size could not be evaluated due to the low number of charts that had this information recorded. Abscess size may have been a factor in the clinical decision making however size is not part of
Discussion
AC CE P
overestimate the risk.
TE
D
any guidelines on treatment. Finally, since our outcome was common (>10% prevalence) the OR may
Treatment failure of ED patients that present with a purulent skin infection was significantly associated with advanced age with elderly patients having almost 4 times the increased risk of treatment failure compared to their younger counterparts. This risk increased to more than 8 times among patients with confirmed purulent skin infections. There was no difference in failures across the three IDSA classifications (mild, moderate, and severe) nor if the patient was treated following IDSA guidelines or either over or under-treated. The IDSA developed these practice guidelines in order to diagnose and administer effective treatments in a timely fashion. By not including age in their treatment algorithm, the IDSA might be under-treating older adults who carry greater risk of morbidity from bacterial infections.
ACCEPTED MANUSCRIPT Even though elderly patients presented less frequently to the ED with a purulent skin infection their high
PT
treatment failure rate raises concerns for not basing antibiotic adjunct therapy off of age.
RI
In our study, there was a high prevalence of deviation from national guidelines with nearly 1 in 5 patients not receiving an incision and drainage and almost 1 in 4 patients with an uncomplicated mild purulent
SC
skin infection as defined by the IDSA guidelines, receiving unnecessary antibiotics. The IDSA gives a
NU
strong recommendation with level 1 evidence for incision and drainage of any purulent skin infection.4, 11 The high prevalence of purulent skin infections being diagnosed without a procedure to open up the
MA
infection pocket is concerning. Incision and drainage either with or without packing is a painful procedure that ED physicians might avoid in an attempt to improve patient care.12 The increase in
TE
D
overuse of antibiotics for discharged purulent skin infection cases from the ED has been a worrisome trend over the past decade.13 We did not see any association of antibiotic use and treatment failure
AC CE P
among all patients. Antibiotics have been shown unnecessary in the treatment of uncomplicated purulent infections.14-16 Seeing how antibiotic resistance patterns are strongly correlated with antibiotic use patterns, reducing unnecessary outpatient antibiotics among patients with uncomplicated purulent infections would be a critical modality to combat antibiotic resistance.17, 18 Many widely used clinical decision rules incorporate age into their decision algorithms, such as those used for the treatment of pneumonia19, 20 and critically injured adults.21 Age is an important consideration seeing that elderly patients are more susceptible to many types of bacterial infections compared to young adults due to a decline in the host immune defense mechanisms.5 This combined with the fact that skin and soft tissue infections are common in the elderly22 makes defining differential treatment decisions by age essential to reducing treatment failure rates. In our study the failure rates ballooned to almost 30% in patients 70 years and older compared to 6.5% in patients in their 20s. The
ACCEPTED MANUSCRIPT recommendations for mild purulent infections are for an incision and drainage alone without adjunct antibiotic therapy. Recent data from a randomized control trial using trimethoprim-sulfamethoxazone
PT
demonstrated a minimal increase in cure rates for outpatient treatment of uncomplicated skin purulent
RI
skin infections.23 Lacking any subgroup or adjusted analysis, elderly adults may have been one of the groups that benefited the most from adjunct therapy. Instead of widespread antibiotic therapy maybe
NU
SC
targeted therapies to those at greatest risk would reduce the risk/reward of antibiotic use.
Vancomycin as a one-time dose in the ED was also associated with worse patient outcomes. This may be
MA
in part due to the high frequency with which vancomycin is under-dosed within the ED.24 Additionally vancomycin was not followed with MRSA antibiotics for home in almost 1 out of 3 patients treated. In fact
TE
D
when antibiotics were given, MRSA antibiotic coverage occurred in less than 3 out of 4 patients. Given the increased risk of treatment failure seen here with vancomycin use and its association with increased
AC CE P
incidence of vancomycin resistant enterococci and clostridium difficile infection,25 we would recommend ED providers avoid its use as an adjunct in treating patients with a purulent skin infection.
Conclusions
Based on our findings, we recommend that in outpatient treatment of purulent skin infections elderly patients may warrant adjuvant therapy with antibiotics. Further work is needed determine the best treatment modalities for our elderly patients with a purulent skin infection. Treatment guidelines should take the age of the patient into consideration to improve cure rates and possibly target antibiotic therapy to those at increased risk of treatment failure.
ACCEPTED MANUSCRIPT Bibliography 1.
Pallin DJ, Espinola JA, Leung DY, Hooper DC, Camargo CA, Jr. 2009. Epidemiology of dermatitis and
PT
skin infections in United States physicians' offices, 1993-2005. Clin Infect Dis. 49(6):901-7. PMID:
2.
RI
19663690.
Edelsberg J, Taneja C, Zervos M, Haque N, Moore C, Reyes K, Spalding J, Jiang J, Oster G. 2009.
SC
Trends in US hospital admissions for skin and soft tissue infections. Emerg Infect Dis. 15(9):1516-8.
3.
NU
PMID: 19788830; PMCID: 2819854.
Pallin DJ, Egan DJ, Pelletier AJ, Espinola JA, Hooper DC, Camargo CA, Jr. 2008. Increased US
MA
emergency department visits for skin and soft tissue infections, and changes in antibiotic choices, during the emergence of community-associated methicillin-resistant Staphylococcus aureus. Ann Emerg Med.
TE
4.
D
51(3):291-8. PMID: 18222564.
Stevens DL, Bisno AL, Chambers HF, Dellinger EP, Goldstein EJC, Gorbach SL, Hirschmann JV,
AC CE P
Kaplan SL, Montoya JG, Wade JC. 2014. Practice Guidelines for the Diagnosis and Management of Skin and Soft Tissue Infections: 2014 Update by the Infectious Diseases Society of America. Clin Infect Dis. 59(2):147-59. PMID: 24947530. 5.
Ben-Yehuda A, Weksler ME. 1992. Host resistance and the immune system. Clin Geriatr Med.
8(4):701-11. PMID: 1423129. 6.
Kaji AH, Schriger D, Green S. 2014. Looking through the retrospectoscope: reducing bias in
emergency medicine chart review studies. Ann Emerg Med. 64(3):292-8. PMID: 24746846. 7.
Charlson ME, Pompei P, Ales KL, MacKenzie CR. 1987. A new method of classifying prognostic
comorbidities in longitudinal studies: development and validation. J Chronic Dis. 40(5):373-83. PMID: 3558716.
ACCEPTED MANUSCRIPT 8.
Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, Januel JM, Sundararajan V. 2011. Updating
and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge
Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, Schein RM, Sibbald WJ. 1992.
RI
9.
PT
abstracts using data from 6 countries. Am J Epidemiol. 173(6):676-82. PMID: 21330339.
Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The
10.
NU
Care Medicine. Chest. 101(6):1644-55. PMID: 1303622.
SC
ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical
Haran JP, Wu G, Bucci V, Fischer A, Boyer EW, Hibberd PL. 2015. Treatment of bacterial skin
MA
infections in ED observation units: factors influencing prescribing practice. Am J Emerg Med. 33(12):1780-5. PMID: 26381681; PMCID: 4679601.
D
Singer AJ, Talan DA. 2014. Management of skin abscesses in the era of methicillin-resistant
TE
11.
Staphylococcus aureus. N Engl J Med. 370(11):1039-47. PMID: 24620867. O'Malley GF, Dominici P, Giraldo P, Aguilera E, Verma M, Lares C, Burger P, Williams E. 2009.
AC CE P
12.
Routine packing of simple cutaneous abscesses is painful and probably unnecessary. Acad Emerg Med. 16(5):470-3. PMID: 19388915. 13.
Pallin DJ, Camargo CA, Jr., Schuur JD. 2014. Skin infections and antibiotic stewardship: analysis of
emergency department prescribing practices, 2007-2010. West J Emerg Med. 15(3):282-9. PMID: 24868305; PMCID: 4025524. 14.
Rajendran PM, Young D, Maurer T, Chambers H, Perdreau-Remington F, Ro P, Harris H. 2007.
Randomized, double-blind, placebo-controlled trial of cephalexin for treatment of uncomplicated skin abscesses in a population at risk for community-acquired methicillin-resistant Staphylococcus aureus infection. Antimicrob Agents Chemother. 51(11):4044-8. PMID: 17846141; PMCID: 2151464.
ACCEPTED MANUSCRIPT 15.
Schmitz G, Goodwin T, Singer A, Kessler CS, Bruner D, Larrabee H, May L, Luber SD, Williams J,
Bhat R. 2013. The treatment of cutaneous abscesses: comparison of emergency medicine providers'
Fahimi J, Singh A, Frazee BW. 2015. The role of adjunctive antibiotics in the treatment of skin and
RI
16.
PT
practice patterns. West J Emerg Med. 14(1):23-8. PMID: 23447753; PMCID: 3582519.
soft tissue abscesses: a systematic review and meta-analysis. CJEM. 17(4). Epub 2015 Feb 20 PMID:
Cosgrove SE, Carmeli Y. 2003. The Impact of Antimicrobial Resistance on Health and Economic
NU
17.
SC
26013989.
Outcomes. Clin Infect Dis. 36(11):1433-37. PMID: 12766839.
Jernigan DB, Cetron MS, Breiman RF. 1996. Minimizing the impact of drug-resistant Streptococcus
MA
18.
pneumoniae (DRSP). A strategy from the DRSP Working Group. JAMA. 275(3):206-9. PMID: 8604173.
D
Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, Coley CM, Marrie TJ, Kapoor
TE
19.
WN. 1997. A prediction rule to identify low-risk patients with community acquired pneumonia. N Engl J
20.
AC CE P
Med. 336(4):243-50. PMID: 8995086.
Lim WS, van der Eerden MML, R., Boersma WG, Karalus N, Town GI, Lewis SA, Macfarlane JT. 2003.
Defining community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax. 58(5):377-82. PMID: 12728155; PMCID: PMC1746657. 21.
Knaus WA, Draper EA, Wagner DP, Zimmerman JE. 1985. APACHE II: a severity of disease
classification system. Crit Care Med. 13(10):818-29. PMID: 3928249. 22.
Laube S, Farrell AM. 2002. Bacterial skin infections in the elderly: diagnosis and treatment. Drugs
Aging. 19(5):331-42. PMID: 12093320. 23.
Talan DA, Mower WR, Krishnadasan A, Abrahamian FM, Lovecchio F, Karras DJ, Steele MT,
Rothman RE, Hoagland R, Moran GJ. 2016. Trimethoprim-Sulfamethoxazole versus Placebo for Uncomplicated Skin Abscess. N Engl J Med. 374(9):823-32. PMID: 26962903; PMCID: 4851110.
ACCEPTED MANUSCRIPT 24.
Fuller BM, Mohr N, Skrupky L, Mueller K, McCammon C. 2013. Emergency Department
vancomycin use: dosing practices and associated outcomes. J Emerg Med. 44(5):910-8. PMID: 23260465;
Haran JP, Hayward G, Skinner S, Merritt C, Hoaglin DC, Hibberd PL, Lu S, Boyer EW. 2014. Factors
RI
25.
PT
PMCID: 3637841.
influencing the development of antibiotic associated diarrhea in ED patients discharged home: risk of
AC CE P
TE
D
MA
NU
SC
administering IV antibiotics. Am J Emerg Med. 32(10):1195-9. PMID: 25149599.
ACCEPTED MANUSCRIPT Figure 1. Study enrollment flow chart SSTI, skin and soft tissue infections; Confirmed Cases, those patients with and incision and drainage done
SC
Figure 2. Predictive probability of failing ED treatment by age.
RI
PT
confirming a purulent infection
NU
Data is the predicted probability from the multivariable logistic regression model and is presented as
AC CE P
TE
D
MA
means with 95% confidence intervals.
ACCEPTED MANUSCRIPT Tables and Figures
AC CE P
TE
D
MA
NU
SC
RI
PT
Table 1. Characteristics of study patients.* Demographics Clinical Cure Therapy Failure n % n % p value Age (SD) 37.4 (13.4) 41.5 (14.2) 0.03 Female 201 (49.1) 24 (41.4) 0.27 White 208 (82.9) 24 (41.4) 0.18 Hispanic 69 (16.9) 9 (15.5) 0.80 African American 26 (6.4) 6 (10.3) 0.26 Asian 5 (1.2) 0 (0.0) 0.40 Medical History CCI 0 320 (78.2) 47 (81.0) 0.63 CCI 1 60 (14.7) 3 (5.2) 0.05 CCI 2 24 (5.8) 3 (5.2) 0.01 CCI 3 or more 5 (1.2) 5 (8.6) <0.001 IVDA 33 (8.1) 2 (3.4) 0.21 Current Abx. 62 (15.2) 10 (17.2) 0.68 Infection Characteristics Face 49 (12.5) 8 (13.8) 0.53 Trunk 86 (21.0) 9 (15.5) 0.35 Hand 27 (6.6) 8 (13.8) 0.05 Extremity 127 (31.1) 20 (34.5) 0.60 Buttocks 70 (17.1) 10 (17.2) 0.98 Genitals 38 (9.3) 6 (10.3) 0.80 Fever 42 (10.3) 8 (13.8) 0.42 Proximal Streaking 14 (3.4) 3 (5.2) 0.51 * Data is presented as n (percentages) unless otherwise indicated. Abx, antibiotics; CCI, Charlson comorbidity index; IVDA, intravenous drug abuse
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
Table 2. ED treatment characteristics Type Clinical Cure Therapy Failure n % n % p value I&D not done 94 (23.0) 11 (19.0) 0.49 IV Abx in ED 47 (11.5) 13 (22.4) 0.02 IV Abx Only 10 (2.4) 2 (3.4) 0.65 MRSA Coverage 215 (52.6) 31 (53.4) 0.90 ED IV Antibiotics Cephalosporin 3 (0.7) 0 (0.0) 0.51 Clindamycin 4 (1.0) 1 (1.7) 0.61 Fluoroquinolone 6 (1.5) 1 (1.7) 0.88 Penicillin/I 3 (7.3) 7 (9.7) 0.51 Vancomycin 25 (6.1) 10 (17.2) 0.003 Home Antibiotics Azithromycin 1 (0.2) 0 (0.0) 0.71 Bactrim 152 (37.2) 24 (41.4) 0.54 Cephalosporin 110 (27.0) 17 (29.3) 0.70 Clindamycin 37 (9.0) 5 (8.6) 0.92 Doxycycline 26 (6.4) 3 (5.2) 0.73 Flagyl 2 (0.5) 0 (0.0) 0.59 Fluoroquinolone 4 (1.0) 0 (0.0) 0.45 Home Abx None 128 (31.3) 17 (29.3) 0.76 Home Abx 1 203 (49.6) 33 (56.9) 0.30 Home Abx 2 156 (19.1) 8 (13.8) 0.33 IDSA Guidelines Mild 196 (47.9) 30 (51.7) 0.59 Moderate 137 (33.5) 16 (27.6) 0.37 Severe 76 (18.6) 12 (20.7) 0.70 Followed Guidelines 164 (40.1) 18 (31.0) 0.19 Over-Treated 163 (39.9) 28 (48.3) 0.22 Under-Treated 82 (20.0) 12 (20.7) 0.91 Data are presented as n (percentages) unless otherwise indicated. I&D, incision and drainage; Abx, antibiotics; Penicillin/I, penicillin inhibitor combination;
ACCEPTED MANUSCRIPT
PT
Table 3. Factors affecting the risk of treatment failure from multivariable logistic regression Suspected Cases Confirmed Cases Odds Ratio 95% CI Odds Ratio 95% CI Age (10 years) 1.43 (1.09-1.88) 1.62 (1.20-2.20) 6.67 (1.42-31.31) 5.09 (0.85-30.59) CCI Score ≥ 3* Vancomycin 6.78 (2.15-21.33) 7.83 (1.99-30.82))
AC CE P
TE
D
MA
NU
SC
RI
After adjusting for gender, race, Charlson comorbidity index, proximal streaking, severity of infection, location of infection in the hand, severity of infection by IDSA guidelines, if treatment followed IDSA treatment guidelines, if an incision and drainage was performed, and if any antibiotics were given. *Compared to CCI score = 0 as reference; CCI, Charlson comorbidity index
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
Figure 1
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
Figure 2
S0735-6757(16)30776-8 doi:10.1016/j.ajem.2016.10.060 YAJEM 56233
To appear in:
American Journal of Emergency Medicine
Received date: Revised date: Accepted date:
14 June 2016 24 October 2016 28 October 2016
Please cite this article as: Haran John P., Wilsterman Eric, Zeoli Tyler, Beaudoin Francesca L., Tjia Jennifer, Hibberd Patricia L., Elderly patients are at increased risk for treatment failure in outpatient management of purulent skin infections, American Journal of Emergency Medicine (2016), doi:10.1016/j.ajem.2016.10.060
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Elderly Patients are at Increased Risk for Treatment Failure in Outpatient Management of Acute Cutaneous Abscesses
PT
John P. Haran, MD1, Eric Wilsterman, BS1, Tyler Zeoli, BS1, Francesca L. Beaudoin, MD, MS2, Jennifer Tjia, MD, MSCE 3, and Patricia L. Hibberd, MD, PhD4
SC
RI
Affiliations: From the Department of Emergency Medicine, University of Massachusetts Medical School, Worcester, MA1 Department of Emergency Medicine, Alpert School of Medicine, Brown University, Providence, RI2 Department of Quantitative Health Sciences, University of Massachusetts Medical School, Worcester, MA3 Department of Pediatrics and Division of Global Health, Massachusetts General Hospital, Boston, MA4
NU
Corresponding Author: John P. Haran, MD, 55 Lake Avenue North, Worcester, MA 01655; Phone: 508450-8688; email: [email protected]; fax: 508-421-1490
MA
Meetings: None
D
Grant: This study was designed and carried out at the University of Massachusetts Medical School and was supported by an intradepartmental grant through the Department of Emergency Medicine. PLH is supported by an NIH award 5K24AT003683-09.
TE
Conflicts of Interest: None of the authors listed have any conflict of interest.
AC CE P
Word Count: Abstract 250, Main Article 3,389 Author Contributions: JPH, JT, and PLH conceived the study, designed the trial, and obtained research funding. JPH, EW, and TZ supervised the conduct of the trial and data collection. JPH, EW and TZ recruited participating centers and patients and managed the data, including collection and quality control. JPH and FLB provided statistical advice on study design and analyzed the data; JPH drafted the manuscript, and all authors contributed substantially to its revision. JPH takes responsibility for the paper as a whole.
ACCEPTED MANUSCRIPT Elderly Patients are at Increased Risk for Treatment Failure in Outpatient Management of Purulent Skin Infections
PT
Objective: Current Infectious Disease Society of America (IDSA) guidelines for the management of
RI
purulent skin or soft tissue infections do not account for patient age in treatment recommendations. The
SC
study objective was to determine if age was associated with outpatient treatment failure for purulent
NU
skin infection after adjusting for IDSA treatment guidelines.
MA
Methods: We conducted a multicenter retrospective study of adult patients treated for a purulent skin infection and discharged home from four emergency departments between April and September 2014.
D
Patients were followed for one month to assess for treatment failure (defined as need for a change in
TE
antibiotics, surgical intervention, or hospitalization). We used multivariable logistic regression to examine the role of patient age on treatment failure adjusting for demographic variables (gender, race),
AC CE P
comorbidities and severity of infection.
Results: A total of 467 patients met inclusion criteria (mean age 37.9 years [SD 14.0], 48.2% of whom were women). Overall, 12.4% failed initial therapy. Patients 65 years and older (n=35) were almost 4 times more likely to fail initial ED therapy in follow-up compared with younger patients (adjusted Odds Ratio (OR) 3.87, 95% Confidence Interval (CI) 1.24-12.10). After adjustment, for every 10 years of advancing age there was a 43% increased odds of failing initial treatment (OR 1.43 95% CI 1.09-1.88).
Conclusion: Elderly patients with purulent skin infections, whose providers followed the 2014 IDSA guidelines, were more likely to fail initial treatment than younger patients. This study suggests that there is a need to re-evaluate treatment guidelines in elderly patients.
ACCEPTED MANUSCRIPT Introduction The burden of annual visits and total hospital admissions for skin and soft tissue infections (SSTIs) has
PT
increased dramatically in the past decade.1, 2 Between 1993 and 2005, annual visits for SSTIs in U.S.
RI
emergency departments ED increased from 1.2 million to 3.4 million.3 In response, the Infectious Disease Society of America (IDSA) recently updated their guidelines for the management of skin and soft tissue
SC
infection.4 The 2014 guidelines departed from previous versions by creating three SSTI treatment
NU
categories (mild, moderate, and severe) with specified treatment regimens for both purulent and nonpurulent infection types. These categories of severity are defined mainly by the patient’s medical
MA
comorbidities (such as those with markedly impaired host defenses), recent antibiotic use, and severity of presenting symptoms. Current treatment recommendations do not take a patient’s age into
TE
D
consideration. However, age-related biologic changes such as immune system dysregulation, and physiologic changes such as increased dryness and decreased blood flow to the skin can diminish the
AC CE P
body’s innate ability to fight bacterial infections.5 The objectives of this study were to: 1) describe the frequency with which outpatient treatment failure occurs among older and younger adults treated for a purulent skin infection while comparing the ED treatment to the recommended IDSA guidelines and 2) evaluate whether age is an independent risk factor for treatment failure of purulent SSTI in patients treated in the ED and discharged to home.
ACCEPTED MANUSCRIPT Methods Study design
PT
We conducted a multi-center, retrospective cohort study. This study was approved by the institutional
RI
review board (IRB docket H00007714).
SC
Study setting and population
NU
We identified patients 18 years and older who presented to one urban tertiary care academic center with an annual ED census of 132,000 visits and three community emergency departments with a combined
MA
annual ED census of 100,000 visits for treatment of a purulent skin infection between April and September in 2014. Patients were eligible for participation if they were discharged from the ED or the ED
TE
D
observation unit with a diagnosis of a skin or soft tissue purulent infection pulled from the electronic medical records. All patients were classified as having a suspected purulent infection based on discharge
AC CE P
diagnoses of abscess. Patients were further classified as having a confirmed purulent infection if they underwent an incision and drainage in the ED with document purulent material. Patients were excluded if their purulent skin infection had undergone an incision and drainage within the previous month in the ED, if they were admitted to the hospital for this problem on their initial visit, or if they did not have follow-up visit data. Additionally, we excluded patients with: a post-surgical infection, a purulent skin infection of the oral cavity, hidradenitis suppuritiva, or a Bartholins gland abscess. Both patients with hidradenitis or a Bartholins abscess were excluded because they are not expected to improve without further surgical intervention.
ACCEPTED MANUSCRIPT Data collection In order to reduce the potential for systematic error and to mitigate bias, we followed protocols for the
PT
optimal conduct of retrospective studies.6 Prior to data abstraction activities, we a priori defined the
RI
pertinent predictor and outcome variables to be collected in a standardized manner. Abstractors were uniformly trained by the investigators and blinded to the study objectives and hypotheses. We utilized
SC
two different abstractors for each patient enrolled, the first one collecting data on the initial ED visit,
NU
recording demographic, historical and clinical data, while the second abstractor collected all follow-up visit data that included the principal study outcome variables. Abstractors met regularly with the
MA
investigators to review the coding rules. The investigators performed an interrtater reliability
Variable measurement
TE
D
assessment on a 10% random sample of charts.
AC CE P
Baseline variables were extracted from the initial ED visit including: sociodemographic characteristics, medical history, chief complaint, and previous antibiotic use. Patients already on antibiotics from a source outside of the ED were included (e.g., primary care office). The Charlson comorbidity index (CCI) was calculated and used to rank patient’s medical comorbidities.7, 8 Information pertaining to initial ED presentation, ED course, surgical, and antibiotic treatments used were recorded. Starting from this initial visit, all subsequent medical visits pertaining to this index infection were reviewed to assess for outcomes including: 1) change in antibiotic treatments, 2) unplanned surgical intervention to further treat the purulent skin infection, and 3) inpatient hospitalization or admission to ED observation unit for continued infection treatment. We created a composite outcome of ED treatment failure based on these three outcomes. Follow-up chart review was performed by a different abstractor using electronic medical records recording treatment failure one month after the initial visit. We did not include routine follow-up
ACCEPTED MANUSCRIPT visits (e.g. wick removal, wound checks without a change in care) in the outcome measurement. All patients without documentation of treatment failure but with continued documentation or repeat visits
RI
PT
in the medical record were classified as clinically cured.
Infectious Disease Society of America classification
SC
We used IDSA guidelines to classify the severity of each patient’s purulent skin infection clinical
NU
presentation and the observed ED treatments. Each case was categorized into one of 3 treatment categories of purulent SSTIs: mild, moderate, or severe. In brief, these guidelines define mild as an
MA
uncomplicated infection, a moderate infection as one where the patient has signs of a systemic inflammatory response (SIRS), and a severe infection when patients have either an impaired host
TE
D
defense, have failed previous antibiotic therapy, or have had SIRS plus hypotension. 4, 9 We defined SIRS as the presence of one of the following: heart rate >90 beats per minute, respiratory rate >20 breaths per
AC CE P
minute, temperature >38.0 °C, white blood count > 12,000 cells/mm3 or presence of greater than 10% immature neutrophils. The presence or absence of antibiotics and the class of antibiotics prescribed were used to categorize treatment into mild, moderate, or severe therapy class according to IDSA treatment recommendations.4 We used the 2014 IDSA guidelines to compare the observed antibiotic treatment regimens to the expected according to these guidelines. Incision and drainage was recommended for all purulent skin infections regardless of class. Accordingly, each patient received a score of 1 for mild, 2 for moderate, and 3 for severe for that initial ED visit presentation and a second score for the type of treatments rendered. We then compared the clinical presentation and treatment scores with each other to determine if the observed matched the anticipated treatment class. We categorized patients into expected treatment when the scores equaled each other, under-treated if the observed score was lower than the anticipated score, and over-treated when the observed score was higher than the anticipated
ACCEPTED MANUSCRIPT score. This scoring system has been reported on previously.10
PT
Data analysis
RI
We used chi-square tests to compare categorical variables, and the student’s t-test for continuous variables, between patients with clinical cure and those that failed therapy. Given the large sample size,
SC
data were assumed to be normally distributed. We used multivariable logistic regression analysis to test
NU
whether or not increasing age (the main exposure variable) was associated with treatment failure. To select the set of covariates for the multivariable model, we first determined important a priori covariates
MA
that we thought would have clinical relevance. Patient age was the main variable of interest. Patient demographics including gender and race, patient comorbidities (CCI), and IDSA infection severity class
TE
D
(mild, moderate, or severe), provider adherence to IDSA guidelines as described above (or if they did not by either over or under-treating the patient), and treatment variables such as if an incision and drainage
AC CE P
was performed or if they received any antibiotics were also included. Finally, any covariates with a p<0.10 from our unadjusted bivariate analysis were included. IDSA treatment was categorized into either matched, over, or under-treated while age was categorized into elderly or non-elderly using the cutoff of age ≥ 65 years. We then performed analysis stratified in cases of suspected versus confirmed purulent infection to examine whether or not this modified the effect estimate. To analyze the affect of patients lost to follow-up we performed a sensitivity analysis using best-case and worst-case scenarios where we reran the model with the patients that were lost were coded as either none having failed therapy or all of them failing therapy. We used Stata Version 13.1 (StataCorp LP, College Station, TX) for all analyses.
ACCEPTED MANUSCRIPT Results Characteristics of the study subjects
PT
During the 6-month study period over the four ED sites, there were a total of 2,190 ED visits for a
RI
complaint of a skin or soft tissue infection of which 33.1% were diagnosed with a purulent skin infection and 13.5% were admitted (Figure 1). Among the this cohort there were a total of 627 patients with a
SC
diagnosis of purulent skin infection who were discharged home from the ED. We excluded 28 patients
NU
(4.5%) with an abscess related to a dental infection, 9 patients (1.4%) with a bartholins gland abscess, 6 patients (1.0%) with hidradenitis suppurativa, and 12 patients (1.9%) with a post-surgical infection.
MA
There were 105 patients (18.4%) with an average age of 34.4 years who did not have follow-up data available. The average age of the study sample was 37.9 years, 48.2% were women, with 59.5% of
TE
D
patients seen at the community sites. Among the cohort 362 (77.5%) patients had an incision and drainage performed confirming purulent material and thus were labeled as confirmed cases. We had a
AC CE P
interrtater reliability kappa for all variables used in this analysis of 0.92.
Among all 2,190 ED visits for a SSTI (included both abscess and cellulitis patients), patients 65 years and older had significantly lower prevalence of purulent skin infection diagnosis (11.4% vs. 39.3%, p<0.001) and higher prevalence of admission to the hospital (42.7% vs. 20.5%, p<0.001) for a cutaneous infection. Elderly patients presenting to the ED were 5 times more likely to be diagnosed with a non-purulent bacterial skin infection rather than purulent infection (unadjusted OR 5.0, 95% CI 3.8, 6.8).
Our final study cohort consisted of 467 patients of which 58 (12.4%) patients had a failure of initial ED therapy upon follow-up. A low percentage of patients presented already on antibiotics (15.4%) from a previous non-ED healthcare visit. Patients that failed therapy after their ED visit were older in age, had
ACCEPTED MANUSCRIPT higher CCI scores, and had a purulent skin infection involving the structures of the hand (Table 1). For every 10-year increase in patient age, the odds of failing therapy increased by 22% (OR 1.22, 95% CI
PT
1.01-1.47). With regard to the CCI score, patients with scores of 4 or greater had 13 times the odds of
RI
failing therapy (OR 13.62, 95% CI 1.21-153.12) compared to those with scores of 0. Purulent skin
SC
infection involving the hand were 2.3 increased odds to fail (OR 2.26, 95% CI 1.00-5.26).
NU
Characteristics of initial ED treatment therapies comparing those with treatment failure During the course of their initial ED visit, patients with a purulent infection did not receive an incision
MA
and drainage 22.5% of the time and received antibiotics 72.6% of the time. Intravenous antibiotics were used in 12.8% of patients with 2.6% receiving only one intravenous dose in the ED without being sent
TE
D
home with a prescription. ED providers used antibiotic coverage for Methicillin-resistant Staphylococcus
AC CE P
aureus (MRSA) only 72.6% of the time.
In comparing the treatment options among patients that went on to fail therapy compared to those did not (Table 2), there was no significant difference in the rates of treatment failure among the difference antibiotic classes with the exception of those that received intravenous vancomycin. Patient that received an intravenous dose of vancomycin were at a 3.2 increased odds of failing ED therapy that those that did not receive this antibiotic (OR 3.20, 95% CI 1.45-7.07). Patients that received vancomycin within the ED went home without MRSA antibiotic coverage 31.0% of the time. There was also no difference in treatment failure rates if the ED provider gave intravenous antibiotics before discharge or if the antibiotic class selection included MRSA coverage.
ACCEPTED MANUSCRIPT Comparison with national guidelines Physicians followed the IDSA guidelines for treatment of purulent infections only 40.0% of the time;
PT
patients were over-treated 40.9% of the time and under-treated 20.1% of the time. Among patients
RI
classified by IDSA as having a mild infection requiring only an incision and drainage, they received unnecessary antibiotics 69.9% of the time; while not undergoing and incision and drainage in 23.0% of
SC
cases. The majority of patients had a mild infection according to IDSA guidelines; however, we did not
NU
detect any differences in treatment failure rates among the three IDSA classes (Table2). Additionally, we did not see any differences in treatment failure rates if the provider either followed the IDSA guidelines
TE
Characteristics of elderly patients
D
MA
or if they over or under-treated the patient according to these guidelines.
Elderly patients did not differ from their younger counterparts with regard to the infection location, if
AC CE P
they received an incision and drainage, or if they were placed on antibiotics, either within the ED or at home, as part of their care. Fewer elderly patients had a CCI score of 0 (57.1%) than younger patients, whereas a greater proportion of older patients had a score of ≥3 (11.4%) than their younger counterparts (80.3% and 1.4% respectively). A greater proportion of elderly patients failed ED outpatient therapy upon follow up (25% versus 11.7%, p=0.03). Elderly patients had a 2.5 increased odds for failed initial ED treatment than persons less than 65 years old (OR 2.51, 95% CI 1.06-5.92). Even though patients with higher CCI scores were older we did not observe any significant difference in treatment failure rates with an age CCI score interaction.
ACCEPTED MANUSCRIPT Logistic regression model In our multivariable logistic regression, patient’s age was significantly associated with treatment failure
PT
among both suspected and confirmed cases after adjustment for baseline demographic and clinical
RI
variables that could have affected treatment choice and related outcomes (Table 3). Higher CCI scores and ED use of intravenous vancomycin were also significantly associated with ED treatment failure. For
SC
each increase of patient age by 10 years, the odds of treatment failure increased by 43% to 62% with
NU
patients in their 20s having a 7.0% chance of treatment failure that rose to 27.2% in patients in their 70s or older (Figure 2). We reran this model changing the continuous age covariate into elderly (age ≥ 65
MA
years) and non-elderly patients. Elderly patients of suspected cases had almost 4 times greater odds of failure of initial ED treatment (OR 3.87, 95% CI 1.24-12.10) than non-elderly patients while elderly
TE
D
patients among confirmed cases had 8 times greater odds of failure of initial ED treatment (OR 8.38, 95% CI 2.30-30.49) after adjusting for the other covariates in the model which included IDSA treatment
AC CE P
classification. The significance of elderly patients having greater odds of failing treatment did not change in our sensitivity analysis.
Study Strengths and Limitations
To our knowledge this is the first study to provide evidence for poorer outcomes in elderly patients with purulent skin infections after adjustment for initial treatment guideline adherence. This study builds on prior study demonstrating poor adherence to IDSA guidelines by ED providers.10 The strengths of this study include the quality of the data extractions from multiple ED sites. There are, however, several limitations that must be kept in mind in interpreting our study results. There exists the possibility of selection bias from loss to follow-up however the main affect of age on the outcome did not change during our sensitivity analysis. For example, if the majority of cases lost to follow-up were patients who
ACCEPTED MANUSCRIPT failed treatment, the true effect would still be statistically significant (there still would exist a difference between elderly and non-elderly patients). This study was also limited by reliance of discharge diagnosis
PT
coding of purulent skin infection to identify the study cohort. This may partially explain why 1 in 5
RI
patients did not receive an incision and drainage of their diagnosed purulent skin infection. Providers may have been limited in localizing the purulent skin infection by ultrasound or unsure of their diagnosis.
SC
We did not have a confirmatory test to ensure all patients in the cohort did in fact have a purulent
NU
infection however we observed the same association of age and treatment failure among the cases with suspected and confirmed purulent infections in our stratified analysis. Further, factors such as purulent
MA
skin infection size could not be evaluated due to the low number of charts that had this information recorded. Abscess size may have been a factor in the clinical decision making however size is not part of
Discussion
AC CE P
overestimate the risk.
TE
D
any guidelines on treatment. Finally, since our outcome was common (>10% prevalence) the OR may
Treatment failure of ED patients that present with a purulent skin infection was significantly associated with advanced age with elderly patients having almost 4 times the increased risk of treatment failure compared to their younger counterparts. This risk increased to more than 8 times among patients with confirmed purulent skin infections. There was no difference in failures across the three IDSA classifications (mild, moderate, and severe) nor if the patient was treated following IDSA guidelines or either over or under-treated. The IDSA developed these practice guidelines in order to diagnose and administer effective treatments in a timely fashion. By not including age in their treatment algorithm, the IDSA might be under-treating older adults who carry greater risk of morbidity from bacterial infections.
ACCEPTED MANUSCRIPT Even though elderly patients presented less frequently to the ED with a purulent skin infection their high
PT
treatment failure rate raises concerns for not basing antibiotic adjunct therapy off of age.
RI
In our study, there was a high prevalence of deviation from national guidelines with nearly 1 in 5 patients not receiving an incision and drainage and almost 1 in 4 patients with an uncomplicated mild purulent
SC
skin infection as defined by the IDSA guidelines, receiving unnecessary antibiotics. The IDSA gives a
NU
strong recommendation with level 1 evidence for incision and drainage of any purulent skin infection.4, 11 The high prevalence of purulent skin infections being diagnosed without a procedure to open up the
MA
infection pocket is concerning. Incision and drainage either with or without packing is a painful procedure that ED physicians might avoid in an attempt to improve patient care.12 The increase in
TE
D
overuse of antibiotics for discharged purulent skin infection cases from the ED has been a worrisome trend over the past decade.13 We did not see any association of antibiotic use and treatment failure
AC CE P
among all patients. Antibiotics have been shown unnecessary in the treatment of uncomplicated purulent infections.14-16 Seeing how antibiotic resistance patterns are strongly correlated with antibiotic use patterns, reducing unnecessary outpatient antibiotics among patients with uncomplicated purulent infections would be a critical modality to combat antibiotic resistance.17, 18 Many widely used clinical decision rules incorporate age into their decision algorithms, such as those used for the treatment of pneumonia19, 20 and critically injured adults.21 Age is an important consideration seeing that elderly patients are more susceptible to many types of bacterial infections compared to young adults due to a decline in the host immune defense mechanisms.5 This combined with the fact that skin and soft tissue infections are common in the elderly22 makes defining differential treatment decisions by age essential to reducing treatment failure rates. In our study the failure rates ballooned to almost 30% in patients 70 years and older compared to 6.5% in patients in their 20s. The
ACCEPTED MANUSCRIPT recommendations for mild purulent infections are for an incision and drainage alone without adjunct antibiotic therapy. Recent data from a randomized control trial using trimethoprim-sulfamethoxazone
PT
demonstrated a minimal increase in cure rates for outpatient treatment of uncomplicated skin purulent
RI
skin infections.23 Lacking any subgroup or adjusted analysis, elderly adults may have been one of the groups that benefited the most from adjunct therapy. Instead of widespread antibiotic therapy maybe
NU
SC
targeted therapies to those at greatest risk would reduce the risk/reward of antibiotic use.
Vancomycin as a one-time dose in the ED was also associated with worse patient outcomes. This may be
MA
in part due to the high frequency with which vancomycin is under-dosed within the ED.24 Additionally vancomycin was not followed with MRSA antibiotics for home in almost 1 out of 3 patients treated. In fact
TE
D
when antibiotics were given, MRSA antibiotic coverage occurred in less than 3 out of 4 patients. Given the increased risk of treatment failure seen here with vancomycin use and its association with increased
AC CE P
incidence of vancomycin resistant enterococci and clostridium difficile infection,25 we would recommend ED providers avoid its use as an adjunct in treating patients with a purulent skin infection.
Conclusions
Based on our findings, we recommend that in outpatient treatment of purulent skin infections elderly patients may warrant adjuvant therapy with antibiotics. Further work is needed determine the best treatment modalities for our elderly patients with a purulent skin infection. Treatment guidelines should take the age of the patient into consideration to improve cure rates and possibly target antibiotic therapy to those at increased risk of treatment failure.
ACCEPTED MANUSCRIPT Bibliography 1.
Pallin DJ, Espinola JA, Leung DY, Hooper DC, Camargo CA, Jr. 2009. Epidemiology of dermatitis and
PT
skin infections in United States physicians' offices, 1993-2005. Clin Infect Dis. 49(6):901-7. PMID:
2.
RI
19663690.
Edelsberg J, Taneja C, Zervos M, Haque N, Moore C, Reyes K, Spalding J, Jiang J, Oster G. 2009.
SC
Trends in US hospital admissions for skin and soft tissue infections. Emerg Infect Dis. 15(9):1516-8.
3.
NU
PMID: 19788830; PMCID: 2819854.
Pallin DJ, Egan DJ, Pelletier AJ, Espinola JA, Hooper DC, Camargo CA, Jr. 2008. Increased US
MA
emergency department visits for skin and soft tissue infections, and changes in antibiotic choices, during the emergence of community-associated methicillin-resistant Staphylococcus aureus. Ann Emerg Med.
TE
4.
D
51(3):291-8. PMID: 18222564.
Stevens DL, Bisno AL, Chambers HF, Dellinger EP, Goldstein EJC, Gorbach SL, Hirschmann JV,
AC CE P
Kaplan SL, Montoya JG, Wade JC. 2014. Practice Guidelines for the Diagnosis and Management of Skin and Soft Tissue Infections: 2014 Update by the Infectious Diseases Society of America. Clin Infect Dis. 59(2):147-59. PMID: 24947530. 5.
Ben-Yehuda A, Weksler ME. 1992. Host resistance and the immune system. Clin Geriatr Med.
8(4):701-11. PMID: 1423129. 6.
Kaji AH, Schriger D, Green S. 2014. Looking through the retrospectoscope: reducing bias in
emergency medicine chart review studies. Ann Emerg Med. 64(3):292-8. PMID: 24746846. 7.
Charlson ME, Pompei P, Ales KL, MacKenzie CR. 1987. A new method of classifying prognostic
comorbidities in longitudinal studies: development and validation. J Chronic Dis. 40(5):373-83. PMID: 3558716.
ACCEPTED MANUSCRIPT 8.
Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, Januel JM, Sundararajan V. 2011. Updating
and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge
Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, Schein RM, Sibbald WJ. 1992.
RI
9.
PT
abstracts using data from 6 countries. Am J Epidemiol. 173(6):676-82. PMID: 21330339.
Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The
10.
NU
Care Medicine. Chest. 101(6):1644-55. PMID: 1303622.
SC
ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical
Haran JP, Wu G, Bucci V, Fischer A, Boyer EW, Hibberd PL. 2015. Treatment of bacterial skin
MA
infections in ED observation units: factors influencing prescribing practice. Am J Emerg Med. 33(12):1780-5. PMID: 26381681; PMCID: 4679601.
D
Singer AJ, Talan DA. 2014. Management of skin abscesses in the era of methicillin-resistant
TE
11.
Staphylococcus aureus. N Engl J Med. 370(11):1039-47. PMID: 24620867. O'Malley GF, Dominici P, Giraldo P, Aguilera E, Verma M, Lares C, Burger P, Williams E. 2009.
AC CE P
12.
Routine packing of simple cutaneous abscesses is painful and probably unnecessary. Acad Emerg Med. 16(5):470-3. PMID: 19388915. 13.
Pallin DJ, Camargo CA, Jr., Schuur JD. 2014. Skin infections and antibiotic stewardship: analysis of
emergency department prescribing practices, 2007-2010. West J Emerg Med. 15(3):282-9. PMID: 24868305; PMCID: 4025524. 14.
Rajendran PM, Young D, Maurer T, Chambers H, Perdreau-Remington F, Ro P, Harris H. 2007.
Randomized, double-blind, placebo-controlled trial of cephalexin for treatment of uncomplicated skin abscesses in a population at risk for community-acquired methicillin-resistant Staphylococcus aureus infection. Antimicrob Agents Chemother. 51(11):4044-8. PMID: 17846141; PMCID: 2151464.
ACCEPTED MANUSCRIPT 15.
Schmitz G, Goodwin T, Singer A, Kessler CS, Bruner D, Larrabee H, May L, Luber SD, Williams J,
Bhat R. 2013. The treatment of cutaneous abscesses: comparison of emergency medicine providers'
Fahimi J, Singh A, Frazee BW. 2015. The role of adjunctive antibiotics in the treatment of skin and
RI
16.
PT
practice patterns. West J Emerg Med. 14(1):23-8. PMID: 23447753; PMCID: 3582519.
soft tissue abscesses: a systematic review and meta-analysis. CJEM. 17(4). Epub 2015 Feb 20 PMID:
Cosgrove SE, Carmeli Y. 2003. The Impact of Antimicrobial Resistance on Health and Economic
NU
17.
SC
26013989.
Outcomes. Clin Infect Dis. 36(11):1433-37. PMID: 12766839.
Jernigan DB, Cetron MS, Breiman RF. 1996. Minimizing the impact of drug-resistant Streptococcus
MA
18.
pneumoniae (DRSP). A strategy from the DRSP Working Group. JAMA. 275(3):206-9. PMID: 8604173.
D
Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, Coley CM, Marrie TJ, Kapoor
TE
19.
WN. 1997. A prediction rule to identify low-risk patients with community acquired pneumonia. N Engl J
20.
AC CE P
Med. 336(4):243-50. PMID: 8995086.
Lim WS, van der Eerden MML, R., Boersma WG, Karalus N, Town GI, Lewis SA, Macfarlane JT. 2003.
Defining community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax. 58(5):377-82. PMID: 12728155; PMCID: PMC1746657. 21.
Knaus WA, Draper EA, Wagner DP, Zimmerman JE. 1985. APACHE II: a severity of disease
classification system. Crit Care Med. 13(10):818-29. PMID: 3928249. 22.
Laube S, Farrell AM. 2002. Bacterial skin infections in the elderly: diagnosis and treatment. Drugs
Aging. 19(5):331-42. PMID: 12093320. 23.
Talan DA, Mower WR, Krishnadasan A, Abrahamian FM, Lovecchio F, Karras DJ, Steele MT,
Rothman RE, Hoagland R, Moran GJ. 2016. Trimethoprim-Sulfamethoxazole versus Placebo for Uncomplicated Skin Abscess. N Engl J Med. 374(9):823-32. PMID: 26962903; PMCID: 4851110.
ACCEPTED MANUSCRIPT 24.
Fuller BM, Mohr N, Skrupky L, Mueller K, McCammon C. 2013. Emergency Department
vancomycin use: dosing practices and associated outcomes. J Emerg Med. 44(5):910-8. PMID: 23260465;
Haran JP, Hayward G, Skinner S, Merritt C, Hoaglin DC, Hibberd PL, Lu S, Boyer EW. 2014. Factors
RI
25.
PT
PMCID: 3637841.
influencing the development of antibiotic associated diarrhea in ED patients discharged home: risk of
AC CE P
TE
D
MA
NU
SC
administering IV antibiotics. Am J Emerg Med. 32(10):1195-9. PMID: 25149599.
ACCEPTED MANUSCRIPT Figure 1. Study enrollment flow chart SSTI, skin and soft tissue infections; Confirmed Cases, those patients with and incision and drainage done
SC
Figure 2. Predictive probability of failing ED treatment by age.
RI
PT
confirming a purulent infection
NU
Data is the predicted probability from the multivariable logistic regression model and is presented as
AC CE P
TE
D
MA
means with 95% confidence intervals.
ACCEPTED MANUSCRIPT Tables and Figures
AC CE P
TE
D
MA
NU
SC
RI
PT
Table 1. Characteristics of study patients.* Demographics Clinical Cure Therapy Failure n % n % p value Age (SD) 37.4 (13.4) 41.5 (14.2) 0.03 Female 201 (49.1) 24 (41.4) 0.27 White 208 (82.9) 24 (41.4) 0.18 Hispanic 69 (16.9) 9 (15.5) 0.80 African American 26 (6.4) 6 (10.3) 0.26 Asian 5 (1.2) 0 (0.0) 0.40 Medical History CCI 0 320 (78.2) 47 (81.0) 0.63 CCI 1 60 (14.7) 3 (5.2) 0.05 CCI 2 24 (5.8) 3 (5.2) 0.01 CCI 3 or more 5 (1.2) 5 (8.6) <0.001 IVDA 33 (8.1) 2 (3.4) 0.21 Current Abx. 62 (15.2) 10 (17.2) 0.68 Infection Characteristics Face 49 (12.5) 8 (13.8) 0.53 Trunk 86 (21.0) 9 (15.5) 0.35 Hand 27 (6.6) 8 (13.8) 0.05 Extremity 127 (31.1) 20 (34.5) 0.60 Buttocks 70 (17.1) 10 (17.2) 0.98 Genitals 38 (9.3) 6 (10.3) 0.80 Fever 42 (10.3) 8 (13.8) 0.42 Proximal Streaking 14 (3.4) 3 (5.2) 0.51 * Data is presented as n (percentages) unless otherwise indicated. Abx, antibiotics; CCI, Charlson comorbidity index; IVDA, intravenous drug abuse
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
Table 2. ED treatment characteristics Type Clinical Cure Therapy Failure n % n % p value I&D not done 94 (23.0) 11 (19.0) 0.49 IV Abx in ED 47 (11.5) 13 (22.4) 0.02 IV Abx Only 10 (2.4) 2 (3.4) 0.65 MRSA Coverage 215 (52.6) 31 (53.4) 0.90 ED IV Antibiotics Cephalosporin 3 (0.7) 0 (0.0) 0.51 Clindamycin 4 (1.0) 1 (1.7) 0.61 Fluoroquinolone 6 (1.5) 1 (1.7) 0.88 Penicillin/I 3 (7.3) 7 (9.7) 0.51 Vancomycin 25 (6.1) 10 (17.2) 0.003 Home Antibiotics Azithromycin 1 (0.2) 0 (0.0) 0.71 Bactrim 152 (37.2) 24 (41.4) 0.54 Cephalosporin 110 (27.0) 17 (29.3) 0.70 Clindamycin 37 (9.0) 5 (8.6) 0.92 Doxycycline 26 (6.4) 3 (5.2) 0.73 Flagyl 2 (0.5) 0 (0.0) 0.59 Fluoroquinolone 4 (1.0) 0 (0.0) 0.45 Home Abx None 128 (31.3) 17 (29.3) 0.76 Home Abx 1 203 (49.6) 33 (56.9) 0.30 Home Abx 2 156 (19.1) 8 (13.8) 0.33 IDSA Guidelines Mild 196 (47.9) 30 (51.7) 0.59 Moderate 137 (33.5) 16 (27.6) 0.37 Severe 76 (18.6) 12 (20.7) 0.70 Followed Guidelines 164 (40.1) 18 (31.0) 0.19 Over-Treated 163 (39.9) 28 (48.3) 0.22 Under-Treated 82 (20.0) 12 (20.7) 0.91 Data are presented as n (percentages) unless otherwise indicated. I&D, incision and drainage; Abx, antibiotics; Penicillin/I, penicillin inhibitor combination;
ACCEPTED MANUSCRIPT
PT
Table 3. Factors affecting the risk of treatment failure from multivariable logistic regression Suspected Cases Confirmed Cases Odds Ratio 95% CI Odds Ratio 95% CI Age (10 years) 1.43 (1.09-1.88) 1.62 (1.20-2.20) 6.67 (1.42-31.31) 5.09 (0.85-30.59) CCI Score ≥ 3* Vancomycin 6.78 (2.15-21.33) 7.83 (1.99-30.82))
AC CE P
TE
D
MA
NU
SC
RI
After adjusting for gender, race, Charlson comorbidity index, proximal streaking, severity of infection, location of infection in the hand, severity of infection by IDSA guidelines, if treatment followed IDSA treatment guidelines, if an incision and drainage was performed, and if any antibiotics were given. *Compared to CCI score = 0 as reference; CCI, Charlson comorbidity index
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
Figure 1
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
NU
SC
RI
PT
Figure 2