PCP indicates primary care physician.
Cellulitis mimickers diagnosed in the treatment group included erythema chronicum migrans (A); arthropod reaction (B); gout (C); and contact dermatitis (D).
Arakaki RY, Strazzula L, Woo E, Kroshinsky D. The Impact of Dermatology Consultation on Diagnostic Accuracy and Antibiotic Use Among Patients With Suspected Cellulitis Seen at Outpatient Internal Medicine OfficesA Randomized Clinical Trial. JAMA Dermatol. 2014;150(10):1056-1061. doi:10.1001/jamadermatol.2014.1085
Cellulitis is a common and costly problem, often diagnosed in the outpatient setting. Many cutaneous conditions may clinically mimic cellulitis, but little research has been done to assess the magnitude of the problem.
To determine if obtaining dermatology consultations in the outpatient primary care setting could assist in the diagnosis of pseudocellulitic conditions and reduce the rate of unnecessary antibiotic use.
Design, Setting, and Participants
Nonblinded randomized clinical trial of competent adults who were diagnosed as having cellulitis by their primary care physicians (PCPs), conducted at outpatient internal medical primary care offices affiliated with a large academic medical center.
Outpatient dermatology consultation.
Main Outcomes and Measures
Primary outcomes were final diagnosis, antibiotic use, and need for hospitalization.
A total of 29 patients (12 male and 17 female) were enrolled for participation in this trial. Nine patients were randomized to continue with PCP management (control group), and 20 patients were randomized to receive a dermatology consultation (treatment group). Of the 20 patients in the dermatology consultation group, 2 (10%) were diagnosed as having cellulitis. In the control group, all 9 patients were diagnosed as having cellulitis by PCPs, but dermatologist evaluation determined that 6 (67%) of these patients had a psuedocellulitis rather than true infection. All 9 patients (100%) in the control group were treated for cellulitis with antibiotics vs 2 patients (10%) in the treatment group (P < .001). One patient in the control group was hospitalized. All patients in the treatment group reported improvement of their cutaneous condition at the 1-week follow-up examination.
Conclusions and Relevance
Dermatology consultation in the primary care setting improves the diagnostic accuracy of suspected cellulitis and decreases unnecessary antibiotic use in patients with pseudocellulitic conditions. Obtaining an outpatient dermatology consultation may be a cost-effective strategy that improves quality of care.
Cellulitis, an acute infection of the dermis and subcutaneous tissue, is a common problem encountered in both the outpatient and inpatient setting, with an estimated prevalence of 14.5 million cases per year.1 The frequency of cellulitis diagnosis has risen substantially in the ambulatory population over recent decades,2 accounting for nearly $3.7 billion health care dollars overall.1 This superficial infection most commonly affects the lower extremities and is often mediated by streptococcal species,3 although staphylococcal cellulitis, including both methicillin-sensitive Staphylococcus aureus (MSSA) and methicillin resistant Staphylococcus aureus (MRSA), is being reported with increasing frequency.2,4 Despite the frequency of cellulitis diagnosis, there are no gold-standard diagnostic studies or reliable evidence-based diagnostic criteria. Associated physical examination findings include erythema, edema, warmth, and pain; however, all of the aforementioned signs and symptoms are nonspecific and could be present in a number of inflammatory cutaneous conditions.
Because of these nonspecific findings, a multitude of alternative dermatologic conditions may clinically mimic cellulitis. These cellulitis mimickers, also known as pseudocellulitides, are often difficult to recognize, particularly for physicians without dermatological training. As such, the misdiagnosis rate of cellulitis has been estimated to occur in up to 33% of cases in the outpatient setting.5- 8 Given the fact that nearly all patients diagnosed as having cellulitis are prescribed a course of antibiotics, the misdiagnosis of cellulitis is particularly significant in the current health care environment that is challenged by increasing antibiotic resistance and rising health care costs. A limited amount of research has been conducted to assess the magnitude of cellulitis misdiagnosis or address strategies to help correct this considerable problem. Because dermatologists are experts in infectious and inflammatory skin diseases, we hypothesized that outpatient primary care utilization of a dermatology consultation may help identify and treat pseudocellulitides and minimize unnecessary antibiotic use and hospital admissions.
This nonblinded randomized clinical trial was approved by the Partners Institutional Review Board (Partners Human Research Committee). Written informed consent was obtained from all participants. This study was performed in outpatient internal medicine offices affiliated with an academic medical center. The patient population of interest included all adult patients diagnosed as having cellulitis by their outpatient primary care physician (PCP). Patients with postoperative site infections, abscesses, human or animal bites, osteomyelitis, hardware and/or line infections, a history of transplant within 6 months and/or history of acute rejection within 90 days, known use of antithymocyte globulin or campath in the last 6 months, more than 20 mg/d of prednisone for more than 30 days, and patients with abnormal vital signs defined as systolic blood pressure lower than 90 mm Hg, diastolic blood pressure lower than 80 mm Hg, heart rate greater than 90 beats/min or less than 50 beats/min, respiratory rate greater than 20 breaths/min, or temperature exceeding 38.1°C were excluded. Known prisoners, patients who were decisionally impaired, and pregnant women were also excluded. Sample size estimates were obtained using East software (Cytel). A medically important effect size of 0.05 and power of 0.9 yielded an enrollment goal of 248 patients.
The intervention used was an on-site dermatology consultation occurring in the PCP’s office. Participants were randomized to either receive a dermatology consultation (treatment arm) or continue PCP management/standard of care (control arm). Each enrolled patient was evaluated by a single dermatologist who served as the primary investigator (D.K.). The diagnosis and treatment was disclosed to the PCP and patient for those participants randomized to the treatment arm. The diagnosis was recorded but not disclosed for those patients randomized to the control arm. Dermatological diagnoses were based on patient history and physical examination findings.
Primary outcomes included final diagnosis, prescription of antibiotics, need for hospitalization, and clinical outcome. Patients randomized to the treatment arm were required to follow-up in the outpatient dermatology office within 1 week of enrollment. Follow-up for those patients randomized to the control arm was scheduled at the PCP’s discretion. One month following enrollment, a medical chart review was performed on all patients to assess for recurrence, hospitalization, or adverse events.
The 2 patient populations were compared using the Fisher-exact test for proportions and t test for means. Statistical significance was defined as P < .05.
Thirty-one patients were evaluated for enrollment from January to September 2013. A total of 29 patients with a mean age of 59 years were enrolled in this study (Figure 1). Twenty patients (7 male and 13 female) were randomized to a dermatology consultation (treatment arm) and 9 patients (5 male and 4 female) were randomized to PCP management/standard of care (the control arm). Fourteen PCPs participated in the study. There was no statistically significant difference in patient demographics across the 2 populations (Table 1). Of the 20 patients randomized to a dermatology consultation, 2 patients (10%) were diagnosed as having cellulitis, and 5 patients received antibiotics (2 for cellulitis and 3 for erythema chronicum migrans) (Table 2). No patients in the treatment arm required hospitalization. Of the 9 patients randomized to the control arm, all (100%) were diagnosed as having cellulitis by the PCP and were subsequently treated with antibiotics. Evaluation by the dermatologist during enrollment revealed that 6 of 9 patients (67%) randomized to the control arm had a pseudocellulitis rather than true infection. One patient in the control arm was sent to the emergency department by his or her PCP and was subsequently admitted to the hospital for 6 days.
There were no notable differences in risk factor prevalence across both populations (P > .05). The most common cellulitis mimickers in the treatment arm were an eczematous dermatitis (including atopic dermatitis, contact dermatitis, and lichen simplex chronicus), stasis dermatitis, and erythema chronicum migrans (Box). One patient in the treatment group did not return for the 1-week follow-up examination and was unreachable by telephone; however, a medical chart review revealed no additional evaluations, cellulitis diagnoses, or admissions. All patients (100%) in both arms reported a preference for being treated as an outpatient rather than as an inpatient. The original enrollment goal of this study was 248 patients; however, this study was terminated prior to reaching the original enrollment target because interim analysis showed statistical significance. Continuing the study to reach the initial enrollment goals was deemed unethical, given the number of cases of noncellulitis infections in the control group being inappropriately treated. No adverse events occurred in this study.
Eczematous dermatitis (n = 4)
Statis dermatitis (n = 3)
Erythema migrans (n = 3)
Cellulitis (n = 2)
Arthropod reaction (n = 2)
Phytophotodermatitis (n = 1)
Gout (n = 1)
Molluscum contagiosum (n = 1)
Hematoma (n = 1)
Erythema nodosum (n = 1)
Chronic paronychia (n = 1)
Cellulitis is commonly misdiagnosed in the outpatient setting.5,6 The results of this study suggest that dermatology consultation during a patient’s visit to primary care can improve diagnostic accuracy and in turn reduce the rate of antibiotic use in patients presenting with suspected cellulitis. The overall rate of hospitalization for cellulitis in the patient population recruited for this study was low, regardless of the group to which patients were randomized. This suggests that perhaps only patients with nontoxic-appearing cutaneous eruptions were referred by their PCPs for enrollment. However, given that the majority of cases of cellulitis are diagnosed and managed in the outpatient setting,2 the population enrolled in this study accurately captures the burden of outpatient disease.
The differences in the rate of cellulitis diagnosis and antibiotic use between the treatment and control arm highlights the prevalence of conditions that mimic cellulitis and the potential benefit of early involvement of dermatology in the care of patients presenting with suspected cellulitis. There are many possible diagnoses that can resemble cellulitis (Figure 2). Of the patients in the treatment arm, we observed that the most common mimicking conditions were eczema (including atopic dermatitis and contact dermatitis), stasis dermatitis, erythema migrans, and arthropod reaction. This finding is consistent with prior studies that have attempted to estimate the prevalence of pseudocellulitic conditions.5- 8 The remainder of the mimicking conditions that we encountered in the treatment arm of the study do not typically present in a manner that resembles cellulitis; however, they do help to demonstrate the large assortment of conditions that may present in a fashion resembling cellulitis and for which dermatologists may be most familiar identifying.
Although not the primary outcome, the challenge of differentiating cellulitis from pseudocellulitis was also highlighted by the low diagnostic concordance rate (17%) that was observed between the dermatologist and PCP. All of the patients enrolled in this study presented with cutaneous erythema and edema, and none of the patients were febrile or appeared toxic. Therefore, we believe one of the most influential factors in determining a diagnosis of cellulitis is the exclusion of pseudocellulitic conditions. Because the diagnosis of cellulitis and pseudocellulitides in this population were made clinically, we cannot unequivocally ensure diagnostic accuracy. However, previous research has demonstrated that cellulitis and pseudocellulitides cannot be delineated based on patient demographics, risk factor prevalence, or systemic signs of infection6,7; therefore, examination by a dermatologist can be considered the diagnostic gold standard.6,7 In addition, our diagnostic accuracy is supported by the fact that all of the patients in the treatment group, who were not treated for cellulitis, reported symptomatic improvement at the 1-week follow-up examination, cleared their skin condition, did not require hospitalization, and did not represent with cellulitis at the 1-month medical chart review. Finally, all patients diagnosed as having erythema chronicum migrans had positive Lyme IgG serological findings 6 to 8 weeks after enrollment.
Our study highlights how the application of this standard affects both the diagnosis and course of treatment for patients with suspected cellulitis. While our study is limited by a small sample size, we were able to demonstrate with significance that evaluation by a dermatologist can alter the diagnosis and course of treatment received by patients with suspected cellulitis in the outpatient setting. We recognize that the results of this study may not be generalizable to primary care clinics that do not have a dermatologist available for consultation. Nonetheless, previous research has demonstrated improved outcomes in patients with cellulitis who are comanaged by dermatologists and PCPs.6 This study also supports the notion that the availability of a dermatologist in the outpatient primary care setting may be a cost-effective strategy that improves patient safety and quality of care.
Lastly, we believe that dermatologic consultation may be able to have a similar impact on patients with more toxic-appearing cutaneous eruptions who have been admitted for inpatient treatment of cellulitis. As a result of this study, we have initiated a similar project with the goal of assessing the impact of dermatologic consultation on the hospital course of patients who have been admitted for cellulitis.
Corresponding Author: Daniela Kroshinsky, MD, MPH, Department of Dermatology, Massachusetts General Hospital, 50 Staniford St, 200 Boston, MA 02114 (firstname.lastname@example.org).
Published Online: August 20, 2014. doi:10.1001/jamadermatol.2014.1085.
Author Contributions: Drs Arakaki and Kroshinsky had full access to all of 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: Arakaki, Woo, Kroshinsky.
Acquisition, analysis, or interpretation of data: All Authors.
Drafting of the manuscript: Arakaki, Strazzula, Kroshinsky.
Critical revision of the manuscript for important intellectual content: Arakaki, Woo, Kroshinsky.
Statistical analysis: Arakaki, Kroshinsky.
Obtained funding: Kroshinsky.
Administrative, technical, or material support: Arakaki, Kroshinsky.
Study supervision: Kroshinsky.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported in part by The Dermatology Foundation.
Role of the Sponsor: The sponsor had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.