Article, Pediatrics

Pediatric abscess characteristics associated with hospital admission from the ED

Unlabelled image American Journal of Emergency Medicine (2011) 29, 1013-1018

Original Contribution

Pediatric abscess characteristics associated with hospital admission from the ED^?

Michael W. Sauer MD ^a,?, Daniel A. Hirsh MD ^a,b,c, Harold K. Simon MD, MBA ^a,b,c,

Shervin A. Kharazmi MD ^a,c, Jesse J. Sturm MD, MPH ^a,b,c

^aDepartment of Pediatrics, Emory University, Atlanta, GA 30303, USA

^bDepartment of Emergency Medicine, Emory University, Atlanta, GA 30303, USA

^cPediatric Emergency Medicine, Children’s Healthcare of Atlanta, Atlanta, GA 30329, USA

Received 20 April 2010; revised 26 May 2010; accepted 27 May 2010

Abstract

Objective: To determine the characteristics of pediatric soft tissue abscesses that result in hospital admission.

Methods: All visits for soft tissue abscesses to the study emergency department (ED) were examined during 2008.

Detailed records were reviewed to determine ED disposition, Abscess size, location, presence of fever, duration of symptoms, previous antibiotic therapy, prior ED visit(s), and wound and blood culture results. Data were analyzed to determine which of these characteristics were associated with hospital admission from the ED.

Results: Six hundred twenty-two patients met the inclusion criteria. One hundred thirteen (18%) patients were admitted to the hospital and 509 (82%) were discharged home. Compared to those sent home, abscesses resulting in admission were more likely to be located in the genital area (odds ratio [OR], 3.08; 95% confidence interval [CI], 1.37-6.90), breast (OR, 4.8; 95% CI, 1.08-21.4), or face

(OR, 4.39; 95% CI, 1.86-10.3), and were more likely to be larger than 3 cm (OR, 3.66, 95% CI, 2.10-6.36). Patients who were admitted to the hospital were also more likely to have fever (OR, 5.93; 95% CI, 3.4-10.3) and have had a prior ED visit with the same complaint (OR, 3.81; 95% CI, 1.77-8.2). Seventy-seven percent of abscesses that were cultured were positive for methicillin- resistant Staphylococcus aureus.

Conclusions: Size and location (especially those in the genital region, breast, and face), appear to be associated with admission for pediatric abscesses. History of fever and previous ED visit also appear to be associated with hospital admission. Obtaining blood cultures for pediatric abscesses is likely of little clinical benefit.

^? Presented in part at the 2010 Pediatric Academic Societies’ Annual Meeting on May 1, 2010, in Vancouver, Canada.

* Corresponding author. Emory University School of Medicine,

Atlanta, GA 30303, USA. Tel.: +1 404 778 1440; fax: +1 404 778 1401.

E-mail address: [email protected] (M.W. Sauer).

Introduction

Soft tissue abscesses in children are a common presenting complaint to emergency departments (EDs) [1,2]. Several studies have shown that the incidence of methicillin-resistant Staphylococcus aureus (MRSA) skin infections is increasing

in the pediatric population and is leading to increased ED visits [3-7]. Currently, there is no consensus about management, although some studies suggest incision and drainage of superficial abscesses without antibiotics may be sufficient therapy in an outpatient setting [8-11].

The American Academy of Pediatrics and the Centers for Disease Control and Prevention have suggested algorithms for managing children with suspected MRSA skin infections in the outpatient setting [12,13]. Admission to the hospital is recommended for patients with moderate infections or in situations where there is concern about patient compliance or follow-up. Moderate infections are defined as those with fever, ill appearance, or with infections covering extensive surface area. In this recom- mendation, the definitions of extensive area of involvement or systemic symptoms are not clearly defined. Further- more, other factors that may be important in determining whether to admit a patient (such as body site, previous antibiotic use, previous visit for similar complaint, or age of patient) are not addressed in these guidelines. Few studies have systematically looked at characteristics of abscesses in pediatric patients associated with admission to the hospital. One recent study found that a significant predictor of hospitalization was having a lesion greater than 5 cm [14].

The objective of this study was to determine the characteristics of soft tissue abscesses in children that are associated with hospital admission. Specifically, the study sought to determine whether abscess location or size, patient age, presence of fever, recent outpatient antibiotic failures, or multiple visits for the same concern were important factors in whether or not a patient was admitted. Further- more, the study analyzed the results of white blood cell counts and C-reactive protein levels in patients seen in the ED for abscess care, as well as results of wound and blood cultures.

An additional goal was to develop a further understanding of abscess management in the ED at one institution and provide data for potential future development of more specific evidence-based management guidelines. The guide- lines that currently exist are vague and do not address potentially important issues such as patient age, abscess location and size, and failed antibiotic use. Therefore, how practitioners in the ED manage soft tissue abscesses in children is not standardized and not always consistent with existing guidelines. There is need for greater understanding of factors that are associated with admission.

Methods

Population

This study was conducted at the ED of an urban tertiary care pediatric hospital with 60, 000 visits annually.

Outcome measures

All visits to the study ED for abscesses during a 12-month period from January 1, 2008, through December 31, 2008, were retrospectively reviewed. All records for patients from age 2 months through 18 years that contained the word “abscess” in the diagnosis field of the electronic medical record were identified for review. These terms were taken from medical coding discharge diagnosis fields where alternative terms such as boil, carbuncle, and furuncle were recoded as “abscess.” The following abscesses were excluded from this study: peritonsillar, intra-abdominal, intracranial, orbital, and cervical abscesses. For the purpose of this study, these abscesses were excluded because these infections are nearly universally admitted without ED incision and drainage at this institution. Visits were categorized as admissions to the inpatient ward, intensive care unit, operating room, or as Discharge home from the ED. visit characteristics for the ED visit and patient demographics including patient age, weight, sex, race, and insurance status were recorded. A standard data form was used to collect ED as well as subsequent hospital data when a patient was not discharged home. ED characteristics were abscess location and size (diameter), length of symptoms, presence of fever (N38.0?C), duration of fever, prior antibiotic use for this abscess, antibiotics prescribed upon discharge, previous ED visits within the past 5 days for the same complaint, and whether incision and drainage was performed in the ED. Laboratory data collected included WBC count, CRP level, and blood and wound culture results. Blood culture results were deemed a contaminant if the growth was skin flora and not the same organism as in the abscess culture. Bacterial cultivation and methicillin resis- tance were performed using methods recommended by the Clinical and Laboratory Standards Institute [15]. For admitted children, further data were collected on whether

the patient was taken to the operating room for drainage.

Abscesses were grouped by anatomic location. The groups were axillary, breast, buttock, dental, extremity, face, genital, scalp, and trunk. Data for abscess location, size, incision and drainage, previous antibiotic use, presence of fever, and duration of symptoms were extracted from the electronic record based on the notations in the ED physician charts. A single reviewer primarily extracted the data and a second reviewer (blinded to whether the patient was admitted or not) extracted data on 20% of randomly selected records to ensure consistency.

Statistical analysis

Data were analyzed using ?2 and t tests where appropriate to compare the characteristics between admitted and discharged patients for statistical significance; P b .05 was considered significant. Interpretation of reliability was conducted using ? score for categorical variables and Pearson r correlation coefficient for continuous variables.

To control for differences in the patient populations of admitted and discharged patients, we created a binary logistic regression model with the outcome of hospital admission. Predictor variables in the model included age less than 2 years, abscess size greater than 3 cm (in largest dimension), presence of fever (N38.0?C), prior antibiotic use for treatment of the abscess, previous ED visit for the abscess, and whether or not incision and drainage was performed in the ED. These variables were coded as dichotomous outcomes where the presence of the character- istics was assigned a value of 1. To incorporate the location of the abscess in the model, a series of dummy variables were created. Extremity abscesses were the reference category for the dummy variables and therefore this was left out of the model. Of the included variables, only 1 record of 622 had data missing and this record was excluded from regression analysis.

Some variables were intentionally left out of model (WBC and CRP) as many values in these variable fields were missing because not all patients in the study had blood work done. Regression diagnostics (C-statistic and Hosmer- Lemeshow goodness-of-Fit test) were performed on the above model. For the models, we then determined which variables might have significant interaction terms and added each category of interaction terms to the models. The interaction terms were not significant and the trends in the probabilities did not change and we therefore present our models without interaction terms. Odds ratios and 95% confidence intervals are shown to represent the strength of the associations for regression model output. All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS, version 15.0, Chicago, Ill).

Results

During the study period, 622 pediatric patients meeting inclusion criteria were identified based on the study method. Of these, 113 (18%) were admitted to the hospital and 509 (82%) were discharged home. There were no significant differences in patient demographics between the admitted and discharged groups (Table 1). The patients in the study had a median age of 42 months (interquartile range, 18-122) and median weight of 16 kg (interquartile range, 11-39). Measures of interpreter reliability in the randomly selected audit of extracted variables yielded a mean ? score of 0.84 and a Pearson r correlation of 0.947 (P b .001).

In an unadjusted analysis, the presence of fever, parental report of prior antibiotic treatment of the skin infection, previous ED visit for the same complaint, size greater than 3 cm, and age less than 2 years old were statistically more likely in the group admitted to the hospital (Table 2). Incision and drainage performed in the ED was less common in those patients who were admitted. Thirty-two percent of patients who were admitted underwent abscess incision and drainage,

Admitted (n = 113)			Discharged (n = 509)	P
Male	57	(50.4%)	252 (49.5%)	.909
Female	56	(49.6%)	257 (50.5%)
Median age in months	33		46	.420
Race
Caucasian	38	(33.6%)	134 (26.3%)	.241
African American	71	(62.8%)	346 (68.0%)	.162
Other	4	(3.5%)	29 (5.7%)	.556
Median weight in kilograms	14		17.3	.120
Insurance type
Private	36	(31.9%)	139 (27.3%)	.624
Medicaid	64	(56.6%)	284 (55.8%)	.890
Self-pay	12	(10.6%)	74 (14.5%)	.438
None	1	(0.9%)	12 (2.4%)	.240
P b .05 considered significant.

whereas 57% of discharged patients had their abscess drained in the ED. Sixty-two percent of admitted patients were taken to the operating room for incision and drainage. Of those patients on oral antibiotics before the ED visit, patients were most commonly on trimethoprim/sulfamethox- azole (13% of all those admitted vs 13% of all those discharged) and clindamycin (20% of those admitted vs 8% discharged). Upon admission, patients were continued on

Table 1 Characteristics of subjects

clindaymcin 82% of the time.

The most common anatomic locations of abscesses presenting to the ED were extremities (30% of all abscesses) and buttocks (25%). Abscess locations and

Table 2 Patient clinical characteristics associated with hospital admission for abscess

Admitted (n = 113)		Discharged (n = 509)	P
Presence of fever ^a	69 (61.1%)	120 (23.6%)	b.001
Prior antibiotic use	57 (50.4%)	140 (27.5%)	b.001
Previous ED visit	27 (23.9%)	42 (8.3%)	b.001
Size N3 cm	65 (57.5%)	157 (30.8%)	b.001
Age b2 y old	49 (43.4%)	155 (30.5%)	.008
Incision and drainage in ED	35 (31.0%)	289 (56.8%)	b.001
Mean duration of symptoms	4.6 (3.4)	4.4 (5.9)	.793
in days (SD)
Location
Axillary	2 (1.8%)	16 (3.1%)	.425
Breast	5 (4.4%)	5 (1.0%)	.008
Buttock	21 (18.6%)	135 (26.5%)	.078
Dental	5 (4.4%)	45 (8.8%)	.118
Extremity	21 (18.6%)	167 (32.8%)	.003
Face	20 (17.7%)	35 (6.9%)	<.001
Genital	29 (25.7%)	44 (8.6%)	<.001
Scalp	2 (1.8%)	23 (4.5%)	.175
Trunk	8 (7.1%)	39 (7.7%)	.832
P value b .05 considered significant (items in boldface). ^a Fever defined as T N 38.0?C.

their association with admission are shown in Table 2. In the unadjusted analysis, abscesses located in the genital region, face, or breast were significantly more likely to result in hospital admission. Extremity abscesses were less likely to be admitted.

Table 4 Results of laboratory evaluation in patients with abscess presenting to the ED

In the regression model, after controlling for differences between the groups, pediatric patients with fever, previous ED visit for same complaint, or abscess size greater than 3 cm were more likely to be admitted to the hospital from the ED. The Hosmer-Lemeshow goodness-of-fit test for our models allowed us to accept the null hypothesis that our model was a good fit [16]. Prior antibiotic use and patient age were not significantly associated with admission in this model (Table 3). Those patients who had incision and drainage performed in the ED were more likely to be discharged home than their counterparts. Anatomic locations that were associated with admission in this model again were genital, face, and breast.

As seen in Table 4, only 10% of discharged patients had blood work done, whereas 91% of admitted patients had blood work done. Admitted patients had average WBC counts of 16 and CRP levels of 6; both of which were significantly higher than those of discharged patients (P b 0.001). Blood cultures were obtained in 70 of 113 (62%) of admitted patients and 42 of 509 (8%) of discharged patients. Excluding a small number (4) of cultures suspected to be

Table 3 Patient characteristics and odds ratios for hospital admission for abscess

Model variable ^a	Odds ratio	95% Confidence interval ^c
Presence of fever	5.93	(3.4-10.3)
Prior antibiotic use	1.59	(0.91-2.81)
Previous ED visit	3.81	(1.77-8.20)
Size N3 cm	3.66	(2.10-6.36)
Age b2 y old	0.93	(0.53-1.65)
Incision and drainage	0.25	(0.15-0.44)
during ED visit
Location
Axillary	0.55	(0.1-3.04)
Breast	4.80	(1.08-21.4)
Buttock	0.75	(0.35-1.62)
Dental	0.64	(0.21-1.97)
Extremity ^b	-	-
Face	4.39	(1.86-10.3)
Genital	3.08	(1.37-6.90)
Scalp	0.57	(0.10-3.15)
Trunk	1.06	(0.37-2.99)
Hosmer-Lemeshow goodness-of-fit test P value is .415, suggesting good model fit. ^a Binary logistic regression model with outcome of hospital admission and dichotomous predictor variables of fever (N38?C), prior antibiotic use, previous ED visit; for abscess care, size greater than 3 cm, age less than 2 years, and location. ^b Extremity location used as the reference location in the model. ^c Odds ratios provided with 95% confidence interval (items in boldface considered significant).

Admitted (n = 113)		Discharged (n = 509)	Reference range	P
Mean WBC	16.3	11.8	5.0-19.5	b.001
count in
thousand/mL
(SD) ^a
Mean CRP in	6.7	2.5	b1.0	b.001
mg/dL (SD) ^a
Positive blood	0	0	N/A	1.000
culture ^b
Wound culture	67 (68.4%)	259 (79.9%)	N/A	.010
positive for
MRSA (%) ^c
P b .05 considered significant (items in boldface). ^a WBC and CRP obtained in 103 (91.2%) of Admitted group and 53 (10.4%) of discharged group. ^b Blood cultures obtained in 70 (61.9%) of admitted group and 42 (8.3%) of discharged group. ^c wound cultures obtained in 98 (86.7%) of admitted group and 324 (63.7%) of discharged group.

skin contaminants, there were no positive blood cultures in either group.

When abscess cultures were obtained, MRSA accounted for 68% of admitted abscesses and 80% of those discharged, yielding a total MRSA prevalence of 77% in the population who had cultures. Only 64% of discharged patients had wound cultures obtained, whereas 87% of admitted patients had wound culture results available. Only 3 wound cultures in admitted patients and 1 in discharged patients demonstrated MRSA-inducible clindamycin resistance (positive D-test).

Discussion

Studies have looked at a limited number of characteristics of abscesses associated with hospitalization, but the criteria used to determine admission have not been well established in the pediatric population. Existing guidelines are overly broad and do not address potentially significant factors that may influence clinical decision making. Previous studies have been done in adults or patient populations where MRSA prevalence is low [17,18]. The total MRSA prevalence of 77% is higher than that reported in several recent studies [2,3,6].

This is the largest study done to date on the characteristics of children hospitalized for soft tissue abscesses. This study was able to show that several factors are associated with hospital admission. Location, size greater than 3 cm, and fever were independently associated with admission.

Specifically, independent of other factors, anatomic locations such as genital, face, and breast were most important. The reasons for this were likely multifactorial.

Children with infections in these locations underwent drainage in the ED less often. Presumably, this is related to physician Comfort level and need for subspecialty Surgical consultation.

In the unadjusted analysis, admitted patients were also more likely to have been on a course of outpatient antibiotics and more likely to have a prior ED visit. These patients were likely thought to have “failed” Outpatient therapy and were in need of inpatient management to treat their abscesses. Most commonly, admitted patients were previously on clindamycin, which is the most common antibiotic pre- scribed for superficial skin infections at the study institution. However, in regression model, these variables were no longer significant.

Pediatric patients with fever were also more likely to be admitted to the hospital. Although it cannot be known precisely why the treating physician admitted the patient, it is possible that the physician interpreted the fever to represent a more serious degree of infection. However, 31% of patients had fever, but this was not reflected in any positive blood culture results. Admitted patients also had increased WBC counts and CRP levels relative to their counterparts, although the clinical significance of this is unclear, given that none of the blood cultures were positive. Given the rarity with which blood work was done in patients who were discharged home, it is difficult to make any conclusions about the utility of obtaining blood work. There were 4 cultures in this study that were thought to be skin contaminants, with a false positive rate of 3.6%, similar to baseline contaminant rates in the literature [19]. The utility of obtaining a blood culture for patients presenting with soft tissue abscesses is likely of little clinical benefit. In a scenario where there is an extremely low likelihood of a true positive result and the potential cost of a false positive is high (ie, repeated blood culture, unnecessary Intravenous antibiotics, return to the ED for reevaluation), blood cultures may not be necessary. This is the first study to examine the rate of blood culture results in pediatric soft tissue abscesses.

Incision and drainage has been shown previously to be the standard of care for the treatment of abscesses [9-11]. In the study population, incision and drainage in the ED was found to be less common in those patients who were admitted. This again was possibly related to physician comfort level and ability to drain complicated abscesses in the ED. Most admitted patients were taken to the operating room for incision and drainage.

There were limitations to this study. The retrospective study design limited the data to that documented by ED physicians. Abscess size, degree of fever, prior antibiotic use, previous ER visit, and duration of symptoms were nearly always clear from physician documentation. In blinded review, the correlation coefficient of 0.947 and ? score of 0.84 reflected “almost perfect” agreement in data extraction [20].

Also, the retrospective nature of the study leads to inherent biases as differences between the groups cannot be

controlled. However, the demographic variables between the 2 groups are similar and the logistic regression model was able to control for other differences to the best of our ability. Because there are no clear guidelines on hospital admission, whether or not to admit a patient, perform incision and drainage, or check labs, doing so was up to the discretion of the treating physician. These data reflect the clinical practice of one institution. It is therefore possible that admission decisions here reflect institutional or regional biases and the results of our analysis may not be generalizable to other institutions. Although the criteria for hospitalization are unclear and may not correlate perfectly with disease severity, most patients hospitalized had operative drainage suggesting that admission was warranted and these data can likely be generalized. In addition, the physicians treating the patients were not aware of the subsequent study so results likely actually reflect practices of physicians in a high-

volume regional pediatric emergency center.

In conclusion, this study was able to show that several factors are associated with increased rates of hospital admission for management of pediatric superficial abscesses. These include location of the abscess on the face, breast, or genital region, size greater than 3 cm, and presence of fever. Although the data only represent practice at one large institution, the current management strategies/guidelines are overly broad and these data may be used as a basis for more formal and specific Evidence-based guidelines. These predictors may be helpful in developing more clearly defined management algorithms and designing further prospective studies for children presenting to the ED with abscesses. Despite the presence of fever, blood cultures are very unlikely to be positive in patients with soft tissue abscesses and obtaining them may not be necessary. The value of obtaining routine blood work in the management of skin abscesses is unclear.

References

Pallin DJ, Egan DJ, Pelletier AJ, et al. Increased US 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 2008; 51:291-8.
Kairam N, Silverman ME, Salo DF, et al. Cutaneous methicillin- resistant Staphylococcus aureus in a suburban community hospital pediatric emergency department. J Emerg Med 2009 July [epub ahead of print].
Moran GJ, Krishnadasan A, Gorwitz RJ, et al. Methicillin-resistant Staphylococcus aureus infections among patients in the emergency department. N Engl J Med 2006;355:666-74.
Kaplan SL. Treatment of community-associated methicillin-resistant

Staphylococcus aureus infections. Pediatr Infect Dis 2005;24:457-8.

Kaplan SL. Community-acquired methicillin-resistant Staphylococcus

aureus infections in children. Semin Pediatr Infect Dis 2006;17:113-9.

Herold BC, Immergluck LC, Maranan MC, et al. Community-acquired methicillin-resistant Staphylococcus aureus in children with no identified predisposing risk. J American Medical Association 1998; 279:593-8.
Purcell K, Fergie J. Epidemic of community-acquired methicillin- resistant Staphylococcus aureus infections: a 14-year study at Driscoll Children’s Hospital. Arch Pediatr Adolesc Med 2005;159:980-5.
Kilbane BJ, Reynolds SL. Emergency department management of community-acquired methicillin-resistant Staphylococcus aureus. Ped Emerg Care 2008;24(2):109-14.
Duong M, Markwell S, Peter J, Barenkamp S. Randomized, controlled trial of antibiotics in the management of community-acquired skin abscesses in the pediatric patient. Ann Emerg Med 2010;55(5):401-7.
Llera JL, Levy RC. Treatment of Cutaneous abscess: a double blind clinical study. Ann Emerg Med 1985;14:15-9.
Rajendran P, Young D, Maurer T, et al. Randomized, double-blind, placebo-controlled trial of cephalexin for treatment of uncomplicated skin abscesses in a population at risk of community-acquired methicillin-resistant Staphylococcus aureus infection. Antimicrob Agents Chemother 2007;55:4044-8.
Baker CJ, Frenck Jr RW. Change in management of skin/soft tissue infections needed. AAP News 2004;25:105.
outpatient management of skin and soft tissue infections in the era of community-associated MRSA. Centers for Disease Control and Prevention 2007.
Lee MC, Rios A, Aten MF, et al. Management and outcome of children with skin and soft tissue abscesses caused by community-acquired methicillin-resistant Staphylococcus aureus. Pediatr Infect Dis J 2004; 23:123-7.
CLSI. Performance standards for antimicrobial susceptibility testing. 18th informational supplement, vol. 27. Wayne (Pa): Clinical and Laboratory Standards Institute; 2008.
Hosmer DW, Lemeshow S. Assessing the Fit of the Model. Applied Logistic Regression. Hoboken (NJ): John Wiley & Sons Inc; 2000.

p. 143-200.

Nalmas S, Bishburg E, Shah M, Chan T. Skin and soft tissue abscess: 1 year’s experience. J Cutan Med Surg 2009;13(5):257-61.
Tanir G, Tonbul A, Tuygun N, Aydemir C, Ertan U. Soft tissue infections in children: a retrospective analysis of 242 hospitalized patients. Jpn J Infect Dis 2006;59:258-60.
Bekeris L, Tworek J, Walsh M, Valenstein PN. Trends in blood culture contamination: a College of American Pathologists Q- Tracks study of 356 institutions. Arch Pathol Lab Med 2005;129: 1222-5.
Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;3(1):159-74.

AJEM