Increased Isolation of Methicillin-Resistant Staphylococcus aureus in Pediatric Head and Neck Abscesses

Objective  To compare the proportion of community-associated, methicillin-resistant Staphylococcus aureus (MRSA) infections in pediatric head and neck abscesses between 2 study periods.

Design  Retrospective case review.

Setting  Tertiary care pediatric otolaryngology practice.

Patients  Pediatric patients with head and neck abscesses presenting over 2 separate 2.5-year intervals: July 1999 through December 2001 and January 2002 through June 2004.

Interventions  Incision and drainage of abscess.

Main Outcome Measures  Type and antimicrobial susceptibility of cultured organisms.

Results  We identified 21 abscesses in 19 patients from July 1999 through December 2001 and 32 abscesses in 32 patients from January 2002 through June 2004. Of the 21 abscesses in the first study period, 15 demonstrated pathogen growth compared with 29 of 32 abscesses in the second study period. In the first period, 6 (40%) of 15 abscesses yielded S aureus compared with 17 (58.6%) of 29 abscesses in the second period. The proportion of abscesses yielding MRSA increased from 0% (0/6) in the first study period to 64.7% (11/17) in the second study period (P<.01). All MRSA infections were considered to be community acquired.

Conclusions  Our study demonstrates a statistically significant rise in the proportion of community-associated MRSA infections of the head and neck in the pediatric population at our institution. For communities where similar microbial recovery patterns exist, we suggest that a culture be obtained as soon as possible in a child presenting with a head and neck abscess to identify the organism. Until that time, the best empirical treatment is clindamycin, with other agents available if warranted by culture and sensitivity results. A treatment algorithm is presented.

Over the past several years, infections caused by methicillin-resistant Staphylococcus aureus (MRSA) have become endemic in most hospitals and usually affect patients with established risk factors.¹ More recently, however, MRSA infections have been described in patients without established risk factors who are living in the community. These have been referred to as community-acquired or community-associated MRSA (CA-MRSA) infections. A CA-MRSA infection is defined as an MRSA isolate recovered from a clinical culture from a patient residing in the community with no established risk factors for MRSA infections. These risk factors include (1) the isolation of MRSA 2 or more days after hospitalization; (2) a history of hospitalization, surgery, dialysis, or residence in a long-term care facility within 1 year before the MRSA culture date; (3) the presence of a permanent indwelling catheter or percutaneous medical device (eg, tracheostomy tube, gastrostomy tube, or Foley catheter) at the time of culture; or (4) previous isolation of MRSA.²

At the University of Chicago Children's Hospital, Chicago, Ill, Herold and colleagues³ reported an increase in the incidence of CA-MRSA infections from 10 per 100 000 admissions from 1988 through 1990 to 259 per 100 000 admissions from 1993 through 1995. This increased incidence persisted between 1998 and 1999 at the same institution.⁴ In another study, Buckingham and colleagues⁵ reviewed cohorts of MRSA isolates from Le Bonheur Children's Medical Center, Memphis, Tenn, between January 2000 and June 2002, and found a significant increase in the proportion of these infections considered to be CA-MRSA, from 38% to 63%.

Culture results obtained from pediatric patients presenting with head and neck abscesses often yield S aureus as the responsible pathogen.⁶ The recent increase in CA-MRSA infections among pediatric patients has implications for the choice of antimicrobial treatment given to these patients on admission. A report of head and neck infections in children from the Otolaryngology Department at the University of Texas in Houston found MRSA infections in 7 patients in a 4-month period, 6 of which were sensitive to clindamycin.⁷

In light of these findings, we were interested in reviewing the bacteriologic features and susceptibility patterns of head and neck abscesses in our pediatric patients at the University of Chicago Children's Hospital. We therefore performed a retrospective review of all pediatric patients presenting with head and neck abscesses, for whom the pediatric otolaryngology service was consulted and intervened surgically, and compared the proportion of CA-MRSA between 2 different study periods. Our working hypothesis was that we would find an increasing proportion of CA-MRSA in head and neck abscesses in the pediatric population over time.

Using data from the surgical logs of the pediatric otolaryngologists at the University of Chicago from January 1999 through June 2004, we identified all pediatric patients who underwent incision and drainage of head and neck abscesses during that study period. The consulting patterns at our hospital are such that most head and neck abscesses are managed by the pediatric otolaryngology team. We then chose to compare the culture findings from 2 different 2.5-year intervals (ie, that between July 1999 and December 2001 with that between January 2002 and June 2004). Inclusion criteria included patients younger than 18 years who presented with a head and neck abscess and were treated with incision and drainage. Exclusion criteria included patients with a neck mass that did not show evidence of abscess in the preoperative period and suspected abscesses from which pus was not expressed at the time of surgery. Patients with recurrent branchial cleft cysts were not included. The protocol was approved by the University of Chicago institutional review board.

We gathered the following information from the medical records of the identified patients: age, sex, race, date of admission, date of surgery, location of abscess, preoperative diagnosis, antimicrobial treatment prior to surgery, previous hospitalization, previous exposure to antibiotics, and underlying medical conditions. We also collected the date of culture and the organism that grew from culture. From those culture specimens yielding S aureus, we collected information regarding specific antimicrobial sensitivities and thus characterized the organisms isolated as MRSA or methicillin-susceptible S aureus (MSSA). We also identified the susceptibilities of the isolated S aureus organisms to other antimicrobials and the results of the double disk diffusion test (D test) when performed.

Patients with MRSA-infected neck abscesses were classified as having “nosocomially” acquired MRSA if the specimen was obtained 48 hours or longer after admission to the hospital. Conversely, they were classified as having CA-MRSA if the specimen was obtained within the first 48 hours of hospitalization. Patients were identified as having risk factors for MRSA if they had any of the following characteristics: prolonged hospitalization, history of invasive or surgical procedures including indwelling catheters or endotracheal tubes, or prolonged or recurrent exposure to antibiotics. This information was obtained by reviewing the hospital medical charts.

The surgical logs and radiology reports were reviewed for technique of drainage and the location of the abscess. Of the 21 abscesses during the first study period, 13 were documented by computerized tomography (CT), 2 by ultrasonography (US), 1 by both CT and US, and 5 by clinical examination only. Of the 32 abscesses during the second study period, 25 were documented by CT, 6 by both US and CT, and 1 by clinical examination only. The location of the abscesses were assigned as follows: anterior triangle (included submandibular area), posterior triangle, submental, masseter, cheek, parotid, preauricular, postauricular, parapharyngeal, and prevertebral. The prevertebral and parapharyngeal abscesses were all incised and drained intraorally. When possible, needle aspiration was performed prior to incision and drainage to obtain purulent material for culture and minimize contamination. When no pus was obtained, a cotton-tipped applicator was used to obtain a culture after incision. No antiseptic was used intraorally prior to incision and drainage. Thus, although intraoral contamination of these specimens was a possibility, it has no bearing on our conclusions because none of the intraorally drained abscesses yielded MRSA. The other abscesses were all drained through appropriate external incisions after appropriate sterilization of the surgical site. A needle aspirate was also attempted, followed by incision and drainage. Pus was sent for culture either from the needle aspirate or after incision and drainage. The cavities were explored bluntly and loculations were broken, followed by copious irrigation with sterile isotonic sodium chloride solution. All cavities were drained by either Penrose drains or iodoform gauze packed into the wound. The drains were removed postoperatively, usually within 48 to 72 hours, depending on the amount of drainage observed.

To test the susceptibility of the isolates, the Clinical Microbiology Laboratory at the University of Chicago Hospitals uses the Vitek system (bioMériux Vitek, Inc, Hazelwood, Mo). During the time of the review, S aureus isolates were tested with penicillin, methicillin, clindamycin, erythromycin, gentamicin, rifampin, trimethoprim-sulfamethoxazole, ciprofloxacin, and vancomycin.

An important issue surrounding the use of clindamycin for the treatment of MRSA infections is the possible risk of treatment failure if the infection is caused by erythromycin-resistant S aureus with the potential for selecting for clindamycin resistance. The presence of erythromycin-inducible clindamycin resistance can be clinically determined using a double-disk diffusion method or D test. In a positive D test result, when the disk diffusion method is used to determine susceptibility, a distorted “D-shaped” (rather than circular) zone of inhibition is observed around disks impregnated with 16-membered macrolides, lincosamides, or streptogramin-B antibiotics if an erythromycin disk is placed nearby, suggesting inducible clindamycin resistance.

The percentages of different parameters were compared between the 2 study periods with a χ² test using the Georgetown University χ² calculator (available at http://www.georgetown.edu/faculty/ballc/webtools/web_chi.html).

We identified a total of 21 abscesses in 19 patients from July 1999 through December 2001 (1 patient presented 3 different times) and 32 abscesses in 32 patients from January 2002 through June 2004. The characteristics of the patients in each group were similar in regard to race and sex (Table 1). A larger percentage of patients in the second study period were younger than 5 years (χ² = 4.17; P<.05). Abscesses from all of the MRSA-positive patients identified were cultured within the first 48 hours of hospitalization and were therefore classified as CA-MRSA. None of these MRSA-positive patients had undergone invasive procedures, endotracheal intubation, or prolonged hospitalization prior to culture. None of the patients had underlying chronic diseases. One patient had a history of MRSA-positive stool at 5 days old. Two of 11 patients with MRSA abscesses had been treated with antibiotics for more than 72 hours prior to incision and drainage of the abscess.

Culture results in 15 (71%) of 21 abscesses in the first period demonstrated growth of a pathogen, and 29 (91%) of the 32 abscesses yielded a pathogen during the second period (Table 2). Of the culture-positive abscesses, 40% yielded S aureus during the first study period compared with 59% of abscesses during the second study period. Of the S aureus–positive abscesses, 100% were sensitive to methicillin in the first study period compared with 35% in the second study period. Therefore, there was a significant increase in the proportion of MRSA among abscesses yielding S aureus, from 0% during the first study period to 65% during the second study period (χ² = 7.44; P<.01). When the percentage of MRSA-yielding abscesses out of the total number of abscesses reviewed was calculated, there was also an increase from 0% during the first study period to 34% during the second study period (χ² = 9.1; P<.01).

Locations of the abscesses are detailed in Table 3. There were significantly more abscesses in the posterior triangle during the second study period compared with the first (31.3% vs 4.8%; χ² = 5.4; P<.03) and significantly fewer masseter/cheek abscesses (14.3% vs 0%; χ² = 4.84; P<.05). The rest of the proportions were not significantly different. The 11 abscesses that yielded MRSA were all during the second study period and were distributed as follows: 4 in the anterior triangle, 6 in the posterior triangle, and 1 in the submental area. For comparison, 6 abscesses yielded MSSA during the first study period (2 anterior triangle, 1 posterior triangle, 1 postauricular, 1 parapharyngeal, and 1 parotid) and 6 abscesses yielded MSSA during the second study period (3 anterior triangle, 1 postauricular, 1 parotid, and 1 prevertebral).

Susceptibility patterns of the MRSA isolates are given in Table 4. All isolates were sensitive to clindamycin, gentamicin, rifampin, and vancomycin. Of the 11 MRSA isolates, 10 were found to be erythromycin resistant and clindamycin sensitive and only 1 isolate was erythromycin sensitive and clindamycin sensitive. A D test was performed only on 4 of the erythromycin-resistant MRSA isolates, and in all these cases, the results were negative. All D tests were performed after June 2003, during the latter part of our review. All 8 MRSA isolates tested for their susceptibility to trimethoprim-sulfamethoxazole were sensitive. Of the non-MRSA isolates for which sensitivity testing was performed during the first study period, all were sensitive to clindamycin except for 1 isolate of coagulase-negative Staphylococcus species. During the second study period, all non-MRSA isolates for which sensitivity testing was performed were sensitive to clindamycin except for 1 MSSA isolate that had intermediate sensitivity.

Evidence that MRSA is a growing clinical problem has been reported in the literature with increasing frequency over the past decade or two.^3,8,9 Recently, concern has emerged regarding the increasing incidence of CA-MRSA infections presenting in the pediatric population.^3,7 In this comparison of all pediatric head and neck abscesses over a 5-year study period, we report an overall increase in the percentage of abscesses yielding MRSA from 0% to 34% and a 64.7% increase in the incidence when considering only abscesses that yielded S aureus from the first half of the study period compared with the second half. In fact, none of the head and neck abscesses presenting during the first study period were positive for MRSA. When looking at the classification of MRSA strains as being community associated or hospital acquired, 100% of the MRSA cases reported in this review were community associated. When looking at the location of the abscesses, they were distributed throughout the head and neck area during both study periods. The significance of the increase in the number of abscesses in the posterior triangle during the second study period compared with the first is unclear. Furthermore, all MRSA-yielding abscesses were located in the anterior and posterior triangles and the submental area. Because of the small numbers in this retrospective series, no generalizations about locations likely to yield MRSA can be made. None of the children presenting had any risk factors for MRSA, including long-term exposure to antibiotics. Newer reports exploring the phenomenon of increasing CA-MRSA among pediatric patients have examined additional risk factors including attendance at day care and the presence of household contacts with known MRSA.^10,11 We were unable to evaluate such additional risk factors because of the inherent limitations of performing a retrospective review. Additional studies have used genetic and molecular testing of MRSA isolates to demonstrate distinct patterns differentiating hospital-acquired strains from community-acquired strains.^12,13 This information supports the idea that the spread of MRSA in the pediatric community has not been caused by the spread of nosocomial strains but is a separate phenomenon.

The increasing incidence of CA-MRSA infections in the pediatric population has important clinical implications when approaching a patient presenting with a head and neck abscess. Often, one must prescribe an antibiotic before culture results are available, and now it is clear that consideration should be given to cover MRSA. This might not apply to all locales around the country and would depend on the local microbial recovery patterns. β-Lactamase-resistant antibiotics are the traditional choice to treat community-acquired S aureus infections, but in view of the new resistance patterns at our institution and the increased incidence of MRSA in our community, these agents are no longer a viable empirical first choice of therapy for our patients. These tenants of empirical coverage are even more important when the infection is still in the early stages of cellulitis without abscess formation in which case surgical drainage and acquisition of purulent specimen for culture are not yet possible. Thus, practitioners must be cognizant of the prevalence of CA-MRSA in their communities and its patterns of antimicrobial susceptibility and adjust empirical treatment accordingly.

Some authors have advocated intravenous antibiotic therapy in select stable children with deep neck abscesses (parapharyngeal and retropharyngeal) and have reserved incision and drainage for those cases in which initial conservative management fails.^14,15 In the reported series, the success rates of conservative management ranged from 75% to 91% of the cases. We have tended to approach all abscesses with incision and drainage because this seems to minimize time in the hospital and accelerates resolution. In cases in which conservative treatment is considered and empirical antibiotic coverage is the initial therapeutic approach, it is even more important to be aware of the local microbial patterns and the susceptibility of prevalent organisms encountered in pediatric head and neck infections.

Several agents are available for use as oral therapy for CA-MRSA infections including clindamycin, trimethoprim-sulfamethoxazole, tetracyclines, and linezolid. Clindamycin is a bacteriostatic lincosamide antimicrobial that has activity against gram-positive organisms including S aureus, Streptococcus pneumoniae, Streptococcus pyogenes, and anaerobes. Favorable pharmacokinetics, the intraphagocytic concentrations, and the ability of clindamycin to inhibit staphylococcal toxins all play a role in ameliorating the clinical signs and symptoms of MRSA infections when treated with this agent.¹⁶ While most hospital-acquired MRSA strains have been resistant to clindamycin, CA-MRSA has been generally susceptible, particularly in the pediatric age group.³ An important issue surrounding the use of clindamycin for MRSA infections is the possible risk of treatment failure if the infection is caused by erythromycin-resistant S aureus with the potential for selecting for clindamycin resistance. Target modification mediated by erythromycin ribosomal methylase (erm) genes methylate a site on the 23S ribosomal RNA that binds not only macrolides but also lincosamides and streptogramin B antibiotics. Resistance to these 3 classes of antibiotics is referred to as the macrolide-lincosamide-streptogramin B (MLS_B) phenotype and is common among S aureus isolates.¹⁷ Expression of this phenotype can be constitutive or inducible. When constitutive, the isolates are resistant to all MLS_B antibiotics including clindamycin. When inducible, the isolates will test resistant to 14-membered (eg, erythromycin) and 15-membered (eg, azithromycin) macrolides only, but can be induced to become resistant to lincosamides (eg, clindamycin) and streptogramin B antibiotics in the presence of a strong inducer of methylase synthesis, raising the concern that a constitutive MLS_B phenotype could be selected during clindamycin therapy. Fortunately, the presence of erythromycin-inducible clindamycin resistance can be clinically determined using a double-disk diffusion method or D test.¹⁸ Panagea and colleagues¹⁹ found a high rate of in vitro erm-induced clindamycin resistance with MRSA isolates.¹⁹ In our review, 10 of 11 MRSA isolates were erythromycin resistant and clindamycin susceptible and only 1 was sensitive to both antibiotics. D test results were available for 4 of the 10 erythromycin-resistant, clindamycin-susceptible isolates and were negative in all instances, suggesting the absence of erythromycin-inducible clindamycin resistance in our small sample. The prevalence of the inducible MLS_B phenotype among CA-MRSA isolates varies between geographic regions and has been reported in as many as 56% of erythromycin-resistant, clindamycin-susceptible S aureus isolates in Baltimore, Md,¹⁸ to as few as in 8% of erythromycin-resistant, clindamycin-susceptible MRSA isolates in Houston, Tex.²⁰ Thus, knowledge of these trends in the community is essential to guide appropriate antimicrobial coverage choices.

Based on these findings, we recommend the following paradigm for the treatment of head and neck abscesses in children who present with a community-associated infection in areas where CA-MRSA is often recovered (Figure): initiate treatment with clindamycin and obtain a sample for culture through either needle aspiration or definitive incision and drainage if possible. If the culture yields an organism other than MRSA, then coverage should be modified according to the type of organism and susceptibility results. In our series, clindamycin provided good coverage in all except 2 cases of non-MRSA organisms during both study periods. If the culture yields MRSA that is sensitive to clindamycin and erythromycin, definitive treatment is continued using clindamycin. If the MRSA isolate is erythromycin resistant and clindamycin susceptible, a D test should be performed. If the test result is negative, then treatment is continued with clindamycin. If the D test result is positive, then clindamycin treatment should be discontinued because of the potential of emerging resistance to this agent. Alternative agents available for use would then include trimethoprim-sulfamethoxazole, tetracycline, or linezolid. If the culture yields an MRSA isolate resistant to clindamycin, treatment should be modified according to the susceptibility results and might include any of the previously mentioned antimicrobial agents or vancomycin.

Trimethoprim-sulfamethoxazole has been recommended by the Committee on Infectious Diseases of the American Academy of Pediatrics as useful therapy for mild skin and soft-tissue infections caused by CA-MRSA.²¹ There is, however, some reluctance to use this agent because of limited clinical experience and some concerns about reliability of this antibiotic in treating CA-MRSA infections.²² In our study, all MRSA isolates tested were sensitive to trimethoprim-sulfamethoxazole. The use of tetracyclines in clinical practice has traditionally been limited because of the prevalence of resistant pathogens and the availability of more effective antibiotics to treat these pathogens. With the advent of the CA-MRSA epidemic, however, tetracyclines can be considered as alternative oral treatment for mildly to moderately ill patients with skin and soft-tissue infections. Linezolid represents a unique class of antimicrobials, with a mechanism of action that differs from that of the MLS_B antibiotics; thus, cross-resistance with these antibiotics does not occur. It has broad in vitro activity against antibiotic-susceptible and antibiotic-resistant gram-positive bacteria, including MRSA, and has been compared with vancomycin for the therapy of skin and soft-tissue infections and pneumonia caused by MRSA. Linezolid is almost 100% orally bioavailable, allowing the switch from intravenous to oral therapy without a concern about continued clinical efficacy. Among its disadvantages are its high cost and possibly significant adverse effects such as reversible thrombocytopenia, anemia, and neutropenia requiring weekly monitoring of these parameters during therapy.²³ Vancomycin is also available and has been the drug of choice for the treatment of MRSA infections. However, it is only in intravenous form and has potentially serious adverse effects including the “red man syndrome” (related to nonimmune-mediated histamine release) and nephrotoxic effects. Furthermore, serum levels have to be monitored during administration, and thus it should only be considered in children with skin and soft-tissue infections when they also have accompanying systemic manifestations. Since this is rarely the case in community-associated head and neck abscesses, the requirement for vancomycin should be rare. Finally, one should not forget that clindamycin also can have adverse effects such as rash, diarrhea (3.5% incidence), nausea, and rarely, Clostridium difficile–associated colitis (1 in 100 000 cases), which might progress into fulminant colitis with ileus, toxic megacolon, perforation, and death.

In conclusion, our data support the notion that CA-MRSA infections are on the rise and are an important part of the differential diagnosis when approaching a pediatric patient with a head and neck abscess, especially in locales where similar microbial recovery patterns exist. It is important to consider the rising incidence of MRSA when choosing the antibiotic to prescribe while waiting for culture results and sensitivities. Cultures are critical to determine the organism and its possible resistance patterns. Thus, incision and drainage should be performed as soon as possible to determine the organism involved and provide definitive treatment. If this is not feasible, then a needle aspirate can be performed for culture while waiting to go to the operating room for definitive drainage. The best initial empirical choice is clindamycin, with trimethoprim-sulfamethoxazole, tetracycline, and linezolid available in case of clindamycin resistance. Vancomycin should be reserved only for severe cases with systemic manifestations and when the organism is clindamycin resistant or has the potential to become so.

Correspondence: Fuad M. Baroody, MD, Section of Otolaryngology–Head and Neck Surgery, The University of Chicago, 5841 S Maryland Ave, MC1035, Chicago, IL 60637 (fbaroody@surgery.bsd.uchicago.edu).

Submitted for Publication: December 17, 2005; final revision received May 13, 2006; accepted July 9, 2006.

Author Contributions: Drs Ossowski and Baroody had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Chun, Suskind, and Baroody. Acquisition of data: Ossowski and Baroody. Analysis and interpretation of data: Ossowski, Chun, Suskind, and Baroody. Drafting of the manuscript: Ossowski, Suskind, and Baroody. Critical revision of the manuscript for important intellectual content: Chun and Baroody. Statistical analysis: Baroody. Study supervision: Suskind and Baroody.

Financial Disclosure: None reported.