Cellulitis Disposition: Discharge versus Admission

Authors: Annette Mueller, MD (EM Resident Physician, SUNY Downstate/ Kings County Hospital Center, Brooklyn, NY), Andrew Sweeny, MD (EM Attending Physician, Kings County Hospital Center, Brooklyn, NY) // Reviewed by: Mark Ramzy, DO, EMT-P (@MRamzyDO); Alex Koyfman, MD (@EMHighAK); Brit Long, MD (@long_brit)


Case 1 – A 58-year-old man with a history of hypertension, presents with redness, pain, and swelling of his right lower leg for one week. He reports that he sustained a small cut to the leg while doing yard work two days prior. The patient otherwise feels well and denies fever, malaise, or chills. He has no history of prior skin infection, abscess, or immunocompromised state.  On physical exam his vital signs are normal, and the patient is afebrile. He is well appearing and in no distress.  On his anterior right lower leg is a 10cm by 4 cm area that is erythematous, poorly demarcated, warm to the touch, and mildly tender to palpation. There is no induration, fluctuance, skin breakdown, or purulent discharge. Bedside ultrasound reveals cobble stoning under the area of erythema without evidence of fluid collection.

Case 2 – An undomiciled 46-year-old man with a history of daily alcohol use presents with pain and swelling to his left lower leg for 5 days.  He occasionally has pain to this leg but has never sought medical care.  He denies any history of prior skin infection, abscess, or immunocompromised state. On physical exam, his vital signs are normal, and he is afebrile. He is disheveled appearing with poor hygiene and smells of alcohol.  On his anterior left lower leg is a 10 cm by 4 cm area that is erythematous, poorly demarcated, warm to the touch, and mildly tender to palpation. There is no induration, fluctuance, skin breakdown, or purulent discharge. Bedside ultrasound reveals cobble stoning under the area of erythema without evidence of fluid collection.

Case 3 – A 40-year-old female with a history of breast cancer, currently undergoing chemotherapy, presents with right leg pain and swelling.  The patient’s husband noticed a rash on her right lower leg that developed over the last week and is worsening.  Patient felt fever and malaise today. Her last chemotherapy session was 5 days ago.  On physical exam her vitals are blood pressure 102/70, heart rate 104, respiratory rate 20, and temperature 102.5F. On her anterior right lower leg is a 10 cm by 4 cm area that is erythematous, poorly demarcated, warm to the touch, and mildly tender to palpation. There is no induration, fluctuance, skin breakdown, or purulent discharge.  Bedside ultrasound reveals cobble stoning under the area of erythema without evidence of fluid collection. Right lower extremity venous ultrasound does not show evidence of a venous thromboembolism.


Acute bacterial skin and soft tissue infections (ABSSTIs, also SSTIs) is a common diagnosis in the Emergency Department (ED), with over 2 million annual ED visits and estimated admission rate between 13.9% and 17%. Frequently the patients’ disposition including discharge, observation, or admission based on optimal treatment and resource utilization is a challenge. [1,2]

This post will focus on treatment pathways and disposition decision making once a patient has been diagnosed with cellulitis. Orbital cellulitis has been previously discussed on emDOCs here and as part of our EM@3AM series here.  Diabetic foot infections and surgical wound infections will not be discussed here as their management is different.

Cellulitis: Definition, Typical features, and Risk factors

Bacterial Causes of Cellulitis

The differential for cellulitis is wide because it shares similar features to other diseases including allergic reaction, insect bites or stings, lymphangitis, folliculitis, osteomyelitis, and deep vein thrombosis. We have previously covered mimics to cellulitis on emDOCs, the post can be found here.

Cellulitis involving several sites, such as the neck, buttocks, groin, and hand, are concerning for several potentially deadly infections including Lemierre’s syndrome, Fournier’s gangrene, and flexor tenosynovitis, respectively. While not necessarily a requirement for admission, cellulitis in these areas should prompt consideration of these deadly infections.

Cellulitis is primarily a clinical diagnosis based on appearance and symptoms. Lab work in most cases is not useful. Imaging such as X-ray and CT can be helpful in evaluating for deeper infections like necrotizing fasciitis or pyomyositis. [6] If concern for a necrotizing infection is present, consultation with the surgical specialist is recommended. Additionally, Point of Care Ultrasound (POCUS), may show cobble stoning of subcutaneous tissue due to accumulation of fluid. For a detailed discussion on utilizing POCUS for skin and soft tissue infections see this previous emDOCs post, found here.

The problem: Scarce evidence-based guidelines for the management of cellulitis

Once cellulitis is diagnosed clinically, the clinician needs to develop a treatment pathway. The Infectious Disease Society of America (IDSA) advises on selection of appropriate antimicrobials, elevation of the affected area, and if possible, treatment of any predisposing underlying conditions. Due to a lack of evidence-based guidance on disposition from the ED, the decision between outpatient and inpatient care frequently creates a dilemma for clinicians. The absence of a widely accepted classification system for the severity of cellulitis leads to significant variability in treatment regimens, based on physicians’ assessments and preferences. [7]

In 2014, IDSA issued guidelines for the management of skin and soft tissue infections that differentiated between mild, moderate and severe non-purulent cellulitis. [5]

  • MILD: ‘Typical cases of cellulitis’ without systemic signs of infections should be discharged with an oral antimicrobial agent against Streptococci. The treatment recommendation is oral Penicillin, Cephalosporin, Dicloxacillin or Clindamycin.
  • MODERATE: ‘Typical cases of cellulitis’ with systemic signs of infection should be covered against Streptococci and MSSA. This recommendation was based on a 96% of success rate of treatment with β-lactams. The treatment recommendations are single antibiotic treatment with IV Penicillin, Ceftriaxone, Cefazolin or Clindamycin.
    Per IDSA, an exception from this recommendation is that patients with penetrating trauma, evidence of MRSA infection ‘elsewhere’ or nasal colonization, or history of IVDU should receive broad-spectrum antimicrobial coverage against Streptococci and MRSA with Vancomycin and either Piperacillin-tazobactam or Imipenem/ Meropenem.
  • SEVERE: Severe systemic signs of infection (temperature >38°C, heart rate >90, respiratory rate >24 or abnormal WBC), who have clinical signs of deeper infection such as bullae, skin sloughing. After first ruling out necrotizing fasciitis, these patients should be treated empirically against Streptococci and MRSA with Vancomycin and Piperacillin-tazobactam.

Haran et al. found in 2018 that “Patients who were under-treated based on IDSA guidelines were over two times more likely to fail treatment.” [8]

Patients who fall into the mild category may be treated with oral antibiotics in an outpatient setting. Certain mild cases may also require hospitalization such as inability to take oral medication, severe immunocompromise, high risk for poor medication adherence, or if no improvement with oral antibiotic therapy. [5]

Claeys et al. showed that there were several severity scoring systems to aid in the determination of patient acuity and proper level of care (including Clinical Resource Efficiency Support Team (CREST), Standardized Early Warning Score (SEWS)), but none of them had been robustly validated and specifically CREST tended to over-estimate required level of care. [10]

In regard to ED observation units, the American College of Emergency Physicians stated in their Policy Resource and Education Paper ‘State of the Art: Observation Units in the Emergency Department” that cellulitis accounted for approximately 8% of observation patients with 85% discharged within 23 hours. [11]

ACEP recommended in the same source several exclusion criteria for placement on observation, which comprise:

  • Severe pain
  • Tissue necrosis
  • Neck abscess
  • Peripheral vascular disease
  • Foreign bodies
  • Bite wounds
  • Specific locations (hand, orbit, joints, scrotum, neck)


In addition, immunocompromised patients with diabetes mellitus, cancer, HIV, or on immunosuppressants should likely be admitted. The ACEP paper also states the following are risk factors for which patients may require admission: IVDU, history of MRSA, advanced age, female gender, and an elevated WBC over 15,000. [11]  However, these only be used as a part of the clinical picture and do not necessitate admission if the patient is well appearing, non-toxic, can follow-up, and will be compliant.

Question: What are other risk factors for treatment failures?

Multiple studies have analyzed risk factors for ‘treatment failures,’ defined as a necessary change in route or antibiotic dose of the initially prescribed regimen or upgrade from ED observation to full medicine admission. Significant additional identified predictors of oral antibiotic treatment failure included elevated lactate, [12] recent or chronic cellulitis [7, 13], and comorbid medical conditions including acute kidney injury, chronic kidney disease, chronic edema or lymphedema, diabetes, or liver disease. [10]  

A prospective study published in 2014 included 497 patients, with 102 patients experiencing a failure of treatment. Authors found fever > 38 C at triage (OR 4.3), chronic leg ulcers (OR 2.5), chronic edema/lymphedema (OR 2.5), prior cellulitis in the same area (OR 2.1), and cellulitis at wound site (OR 1.9) to be associated with treatment failure. [7]  However, immunocompromised patients including those with HIV, cancer, and renal or liver disease were not well represented; there was a wide range of antibiotics prescribed; and the MRSA rate was 13.2%, which is less than most communities. [7] 

Question: Who needs to receive parenteral antibiotics?

Several studies have shown that in the absence of sepsis or decreased oral absorption, oral antibiotics have sufficiently high bioavailability. Additionally, they are non-inferior to intravenous antibiotics for uncomplicated skin and soft tissue infections. [14,15]  However, the ISDA’s recommendations leave only very little space for oral antibiotics. As previously discussed, only patients with mild cellulitis, without systemic symptoms, who are able to tolerate PO treatment, are not immunocompromised, have not yet tried outpatient antibiotics, and are likely to be adherent to therapy are recommended to be managed with oral antibiotics. [5]

Hamill et al advocate that in absence of specific exclusion criteria, the default approach to treat moderate to severe ABSSTI should be PO antibiotics. Specific criteria against PO antibiotics include septic shock, severe gastroenteritis/ diarrhea/ vomiting/ malabsorption, immunocompromised state (chemotherapy, cancer, steroids), location in hand or face (eye), high likelihood of non-compliance or follow-up in case of complications. [15] Hamill et all state that in the ED IV antibiotics seem to be administered inappropriately frequent. They list multiple physician biases, including prescribing culture, size of infection, fever, inflammatory markers, convenience, inexperience, attending preference, patient expectations, and fear of litigation. [15]

Question: What about outpatient parenteral antibiotic therapy?

Parenteral antibiotics don’t always automatically rule out discharge and outpatient management. Several studies describe the performance of outpatient management of parenteral antibiotic therapy (OPAT). Those studies mostly originate in Europe and Canada and found OPAT to be safe and have a low rate of treatment failures and high patient satisfaction. Although OPAT appeared to be an attractive alternative to hospitalization, limitations include that some healthcare systems don’t have an OPAT program for ED patients and that in places that offered OPAT, a large proportion of patients returned to the ED for repeat antibiotic administration because these programs were only open during the weekday. [9, 16]

Take Home Points

  • A lack of evidence-based guidance on disposition from the ED results in significant variability in treatment regimens. IDSA issued guidelines for the management of skin and soft tissue infections differentiate between mild, moderate and severe non-purulent cellulitis. Patients with ‘mild’ cases of cellulitis, without systemic signs of infection, are recommended for discharge with oral antibiotics.
  • The most prevalent reason for the admission of cellulitis patients is the recommendation of intravenous antibiotics for moderate to severe cellulitis per IDSA guidelines.
  • In addition, providers need to consider a long list of additional risk factors for outpatient treatment failure, including severe immunocompromise, kidney or liver disease, diabetes, previous MRSA exposure or cellulitis, inability to take oral medication, high risk for poor compliance, or previous outpatient treatment failure.
  • Two options that could be considered to reduce admissions are 1) outpatient treatment with intravenous antibiotic or 2) early substitution of IV with PO antibiotics for moderate to severe cases. However, both strategies have limitations and are currently not supported by generally accepted guidelines.

Case Resolutions

Case 1 – This patient falls in the ‘MILD’ category in the ISDA classification, which is described as ‘typical cases of cellulitis’ without systemic signs of infections that should be discharged with an oral antimicrobial agent against streptococci. The treatment recommendation is an oral penicillin, cephalosporin, dicloxacillin, or clindamycin. This patient does not fulfill criteria for the exceptions (failed outpatient treatment, cannot tolerate PO, severely immunocompromised state, or risk of poor therapy adherence). Her past medical history is unremarkable (no kidney disease, PVD, DM or liver disease). he location of the patient’s cellulitis is not a high-risk location and the patient has no prior or chronic cellulitis.

Case 2 – This patient also falls in the ‘MILD’ category in the ISDA classification of ABSSTIs. However, per history, the patient is at risk of poor adherence to therapy. He also has additional risk factors for possible chronic cellulitis. Admission should be considered for this patient.

Case 3 – This patient falls in the ‘SEVERE’ category in the ISDA classification, due to her abnormal vitals: she is tachycardic and febrile. She is also immunocompromised as she is currently undergoing chemotherapy. The patient should receive vancomycin and piperacillin/tazobactam empirically, along with inpatient admission.

References/Further Reading:

  1. Linder KA, et al. Cellulitis.  JAMA. 2017; PMID: 28535238
  2. Weng QY, et al. Costs and Consequences Associated With Misdiagnosed Lower Extremity Cellulitis. JAMA Dermatol. 2017; PMID: 27806170
  3. Quirke M, et al. Risk factors for nonpurulent leg cellulitis: a systematic review and meta-analysis. Br J Dermatol. 2017; PMID: 27864837
  4. Sullivan T, et al. Diagnosis and management of cellulitis. Clin Med (Lond). 2018; PMID: 29626022
  5. Stevens DL, et al. Practice guidelines for the diagnosis and management of skin and soft tissue infections: 2014 update by the Infectious Diseases Society of America [published correction appears in Clin Infect Dis. 2015 May 1;60(9):1448. Dosage error in article text]. Clin Infect Dis. 2014; PMID: 24973422
  6. Bystritsky R, et al. Cellulitis and Soft Tissue Infections [published correction appears in Ann Intern Med. 2020 May 19;172(10):708]. Ann Intern Med. 2018; PMID: 29404597
  7. Peterson D, et al. Predictors of failure of empiric outpatient antibiotic therapy in emergency department patients with uncomplicated cellulitis. Acad Emerg Med. 2014; PMID: 24842503
  8. Haran JP, et al. Deviating from IDSA treatment guidelines for non-purulent skin infections increases the risk of treatment failure in emergency department patients [published online ahead of print, 2018 Dec 5]. Epidemiol Infect. 2018; PMID: 30516120
  9. Talan DA, et al. Factors associated with decision to hospitalize emergency department patients with skin and soft tissue infection. West J Emerg Med. 2015; PMID: 25671016
  10. Claeys KC, et al. Development of a Risk-Scoring Tool to Determine Appropriate Level of Care in Acute Bacterial Skin and Skin Structure Infections in an Acute Healthcare Setting. Infectious Diseases and Therapy 2018; PMID: 30244362
  11. (ACEP 3/2019) ACoEM. State of the art: observation units in the emergency department.https://www.acep.org/globalassets/new-pdfs/preps/state-of-the-art—observation-units-in-the-ed—prep.pdf
  12. Volz KA, et al. Identifying patients with cellulitis who are likely to require inpatient admission after a stay in an ED observation unit. Am J Emerg Med. 2013; PMID: 23158603
  13. Yadav K, et al. Predictors of Oral Antibiotic Treatment Failure for Nonpurulent Skin and Soft Tissue Infections in the Emergency Department. Acad Emerg Med. 2019; PMID: 29869364
  14. Aboltins CA, et al. Oral versus parenteral antimicrobials for the treatment of cellulitis: a randomized non-inferiority trial. J Antimicrob Chemother. 2015; PMID: 2533616
  15. Hamill LM, et al. Picking the low-hanging fruit: Why not choose oral antibiotics for skin and soft-tissue infections in the emergency department. Emerg Med Australas. 2019; PMID: 31456350
  16. Yadav K, et al. Evaluation of an emergency department to outpatient parenteral antibiotic therapy program for cellulitis. Am J Emerg Med. 2019; PMID: 30824277


Leave a Reply

Your email address will not be published. Required fields are marked *