Magical thinking in modern medicine: IV antibiotics for cellulitis

“I’ve been on these oral antibiotics for 36 hours and my cellulitis isn’t improving. My doctor sent me in because I NEED IV antibiotics.”

“This patient has pretty mild cellulitis, but he does have a fever, so I think we should go with IV antibiotics.”

“That is a pretty big cellulitis. There is no way it is going to improve with just oral antibiotics.”

Among physicians and patients alike, it is generally accepted that IV antibiotics are better than oral. They are stronger. They will work faster. They will save the day when oral antibiotics have failed. But do the bacteria floating around in the soft tissues of your leg really care (or even know) whether the antibiotics entered your system through a vein or via the stomach?

Our devotion to the IV has always seemed a little silly to me. Outside of a few truly emergent infections, when getting antibiotics to the source a few minutes earlier might truly help, it seems obvious that oral and intravenous antibiotics should be fairly similar. Yes, the patient has to be able to tolerate oral intake. And yes, the chosen agent has to be effectively absorbed from the GI tract. But most patients can swallow, and most of the antibiotics we prescribe are well absorbed orally, especially for cellulitis.

Bioavailability

There is a lot of nerdy pharmacologic stuff that one can talk about, such as time-dependant versus concentration-dependant killing, tissue penetration, post-antibiotic affect, and various host factors that can alter the efficacy of antibiotics. Although pathophysiology and pharmacology are interesting, they are frequently misleading in medicine, and lose importance when we have randomized trials to guide our management (which I will discuss in a moment).

However, it might be worth taking a moment to discuss bioavailability, if only because it always seems to come up. By definition, bioavailability is the proportion of a drug that enters the circulation after it is introduced to the body. Bioavailability is variable. It depends on a number of factors, such as co-administration with food, gastric pH, intestinal function, and drug stability in the GI tract.

Bioavailability is just the percentage of medication that enters the circulation. A low bioavailability does not necessarily mean the medication will be ineffective. A drug with an oral bioavailability of 10% might work great if the patient is able to tolerate a 1 gram dose orally, but only 1 mg needs to reach the circulation for the desired effect. For example, azithromycin has a poor oral bioavailability (37%) but is effective orally in a number of conditions including pneumonia. (MacGregor 1997)

That being said, the first line medications for cellulitis all have excellent bioavailability. Cephalexin is 90-100% bioavailable, clindamycin 90%, and doxycycline >90% when taken with food. (MacGregor 1997) The following table from MacGregor (1997) summarizes the bioavailability of many commonly used antibiotics.

Randomized Controlled Trials

The best evidence that we have comes from randomized trials that directly compare intravenous to oral antibiotics in cellulitis. There are 4 of these trials, and they are worth reviewing in a little more detail, but the quick summary is that oral antibiotics looked better than IV antibiotics in all 4 trials.

Aboltins CA, Hutchinson AF, Sinnappu RN. Oral versus parenteral antimicrobials for the treatment of cellulitis: a randomized non-inferiority trial. The Journal of antimicrobial chemotherapy. 2015; 70(2):581-6. PMID: 25336165

Methods

• Randomized, controlled, non-inferiority trial at a single center in Australia
• Patients were included if they had acute (<5 days) cellulitis and the emergency physician thought they required IV antibiotics (because of severity of disease or progression despite oral antibiotics).
• Patients were randomized to either oral therapy with cephalexin 1 gram QID or parenteral therapy with cefazolin 2 grams BID.
• Oral antibiotics were given for 10 days. In the parenteral group, IV antibiotics were given until the cellulitis was no longer progressing and the patient was afebrile, then they were switched over to oral therapy for the rest of the 10 day course.
• The primary outcome was time until no further advancement of the area of cellulitis.

Results

• They included 47 patients.
• 43% had already received a course of antibiotics prior to being enrolled in the study.
• The primary outcome (mean time to no further advancement) was 1.29 days in the oral group and 1.78 days in the IV group (mean difference 0.49 days; 95% CI 1.02 to -0.04 days). Oral was non-inferior to IV, but not quite statistically superior.
• Treatment failure was higher, but not statistically so in the IV group (5 vs 1, p=0.1)
• There was no difference in complications.
• Patients rated the two treatments as similar in convenience, effectiveness, and overall satisfaction.

Caveats

• A small, single center trial.
• The oral group was somewhat sicker at the outset (more fever,  tachycardia, patients with high white cell count, and high CRP), which could have biased the results in favour of the IV treatment group.
• The trial was not blinded and the primary outcome is pretty subjective. Although I would usually think this would bias in favour of the intravenous group, any researcher tackling this question probably questions the dogma of intravenous superiority.

---

Bernard P, Chosidow O, Vaillant L. Oral pristinamycin versus standard penicillin regimen to treat erysipelas in adults: randomised, non-inferiority, open trial. BMJ (Clinical research ed.). 2002; 325(7369):864. PMID: 12386036 [free full text]

Methods

• A multi-center, open-label, randomized, non-inferiority trial at 22 French hospitals (dermatology wards, not emergency departments).
• Included adult patients with erysipelas (well delineated inflammation for less than 5 days with fever or chills).
• Excluded patients with suspected necrotizing infections, abscess, bites, HIV, immunosuppressants, prior antibiotics exposure, or allergies to the study drugs.
• Patients were randomized to either oral pristinamycin 1 gram TID for 14 days or IV benzylpenicillin 18 million units a day divided in 6 infusions.
• The primary outcome was clinical cure rate (afebrile and rash completely resolved) at follow up approximately 1 month later.

Results

• They included 289 patients. There was a lot of loss to follow up (22% in the oral group and 27% in the IV group).
• For the primary outcome, the cure rate was 81% with oral treatment and 67% with IV in the per-protocol analysis. The intention to treat analysis also favoured the oral treatment (cure rate 65% vs 53%). This numbers allow the conclusion that oral therapy is statistically non-inferior to IV therapy.
• Mean duration of hospital stay was the same for the two groups (11 days!!).
• There were more side effects (most GI) in the oral group (28% vs 17%). In the penicillin group there was one case of leukopenia and one case of erythema multiforme.

Caveats

• Again, this trial was not-blinded.
• The trial was stopped early because the interim analysis proved that oral treatment was non-inferior to IV.
• Comparing two different classes of antibiotic makes the comparison more complex than simply “IV versus oral”.
• The patients are probably different than those I see, as they were admitted to dermatology wards. However, they were clearly sick, as fever was an inclusion criteria, and if oral therapy is good enough for sicker patients, it is almost certainly good enough for healthier outpatients.

---

Bernard P, Plantin P, Roger H. Roxithromycin versus penicillin in the treatment of erysipelas in adults: a comparative study. The British journal of dermatology. 1992; 127(2):155-9. PMID: 1390144

Methods

• A prospective, multi-center, randomized trial.
• They included patients over 15 years old admitted to hospital with non-gangrenous skin and soft tissue infections. Patients had to have fever and an elevated white count.
• Patients were excluded if they were pregnant, had a known allergy to macrolides or beta-lactams, had liver or renal disease, used antibiotics in the prior 7 days, had a skin infection of the head, or had an alternative diagnosis such as abscess or DVT.
• Patients were randomised to either oral treatment with roxithromycin (150 mg BID) or intravenously with penicillin (2.5 million units 8 times daily). Antibiotics were continued until the patient was afebrile for 10 days.
• They don’t explicitly state a primary outcome.

Results

• The included 72 patients.
• The average length of treatment was 13 days in both groups. (The IV group transitioned to oral after an average of 6 days).
• Cure rates were 84% with roxithromycin orally and 76% with IV penicillin (p=0.43).
• There were no adverse events with roxithromycin but 2 rashes with penicillin.

Caveats

• This study was not blinded.
• Comparing two different classes of antibiotic makes the comparison more complex than simply “IV versus oral”.
• I am not familiar with roxithmycin, so can’t really comment on the dose.

---

Jorup-Rönström C, Britton S, Gavlevik A, Gunnarsson K, Redman AC. The course, costs and complications of oral versus intravenous penicillin therapy of erysipelas. Infection. 1984;12(6):390-4. PMID: 6394505

Methods

• This is a single-center, open-label, randomized, controlled trial.
• Patients were included if they were admitted to hospital for erysipelas (a high fever plus a red, indurated plaque with a distinct border).
• Patients were excluded if they had a temperature less than 38.5 celsius, prior antibiotics therapy, sepsis, abscess, gangrene, diarrhea, vomiting, or were younger than 15 years old.

• There were 4 groups
  • Oral phenoxymethyl penicillin (0.8 + 0.8 + 1.6 grams). Adjusted to 1.6 grams TID for patients over 90 kg
  • Oral phenoxymethyl penicillin (0.8 + 0.8 + 1.6 grams) plus flucloxacillin (0.5 + 0.5 + 1 gram). Adjusted to phenoxymethyl penicillin 1.6 grams TID and flucloxacillin 1 gram TID.
  • Intravenous benzylpenicillin 3 grams TID. Adjusted to QID.
  • Intravenous benzylpenicillin 3 grams TID plus cloxacillin 2 grams TID. Both adjusted to QID
• For penicillin allergic patients, they compared clindamycin 300 mg TID orally to 600 mg TID intravenously.
• Patients in the IV group transitioned to oral treatment after 2 consecutive temperatures below 37.5 C. Treatment was given for a minimum of 10 days.
• There is not a clearly stated primary outcome.

Results

• They included 73 patients (but 233 were eligible).
• Median fever duration was 2 days with oral treatment versus 3 days with IV.
• Mean hospital stay was 5 day with oral vs 6 days with IV (p<0.05).
• Median sick leave was 13 days with oral vs 16 with IV.
• Recurrence was 11% with oral vs 16% with IV.
• Daily cost in the oral group was $2 USD vs $27 USD in the IV group.

Caveats

• Again, this was an unblinded trial.
• There is a risk of selection bias because of the large number of eligible patients not included.
• They had 4 groups, but only present the data as oral versus intravenous for some reason.

---

There are only 4 RCTs. They are all small, and they all have their flaws. However, there is not even a hint of intravenous superiority here. In fact, in every single one of these trials, oral therapy looks better than intravenous therapy. A Cochrane meta-analysis of the 2 Bernard papers and the paper by Jorup-Rönström concluded that oral therapy was actually superior to intravenous therapy (RR 0.85, 95% CI 0.73 to 0.98). (Kilburn 2010) The patients included in these trials tended to be sicker than the average emergency department cellulitis patient (with fever, systemic symptoms, or previous antibiotics failure), which makes me pretty confident the non-inferiority of oral antibiotics will extrapolate well to a more generalized population. It would be nice to see more research, but rather than trying to prove that oral therapy is non-inferior to IV, considering the added costs and harms of IV therapy, I think future research should consider oral therapy as the gold standard, and try to demonstrate that intravenous therapy is actually superior.

Oral antibiotics in other conditions

Although the data on intravenous versus oral antibiotics is relatively scarce for cellulitis, IV and oral routes have been compared in a number of other infectious diseases. In pediatric pneumonia, oral antibiotics have been shown to be equivalent to IV in 4 large multicenter RCTs. (Addo-Yobo 2004; Atkinson 2007; Hazir 2008; Agweyu 2015) There are also 4 RCTs that demonstrate that oral therapy is as effective as IV in adults with pneumonia, although in 2 of those trials the first couple days of therapy was given IV in both groups. (Vogel 1991; Siegel 1996; Castro-Guardiola 2001; Oosterheert 2006) One RCT demonstrated equivalent treatment failures but less drug toxicity when using oral therapy to treat endocarditis as compared to intravenous therapy. (Heldman 1996) There are also Cochrane reviews that conclude that oral antibiotics are as good as IV in febrile neutropenia, chronic osteomyelitis, and pediatric pyelonephritis. (Conterno 2013; Vidal 2013; Strohmeier 2014)

These trials are just the tip of the iceberg. There are numerous other studies, across a wide range of infectious diseases, and all seem to come to the same conclusion: oral antibiotics are just as good as intravenous. In all my searching (although my search is in no way systematic) I have yet to find a trial where IV antibiotics were actually better than oral.

Harms

There are harms associated with IV antibiotics. Even just a single dose of IV antibiotics in the emergency department has been associated with a significant increase in antibiotic associated diarrhea. (Haran 2014) Furthermore, intravenous therapy can result in phlebitis, thrombosis, extravasation injury, localized infection, and bacteremia. (Li 2015) The overall risk of bacteremia is low with peripheral IV access, but it increases with length of therapy, and we can expect between 0.2 and 2 cases for every 1,000 intravenous catheter days. (Maki 2006, Edgeworth 2009). In order to facilitate outpatient intravenous therapy, physicians often choose medications that are dosed once daily, which are needlessly broad spectrum. Furthermore, even when done as an outpatient, intravenous therapy is significantly more expensive than oral therapy. (Li 2015)

Summary

Putting this all together, I think it is pretty clear that oral antibiotics should be used for the vast majority of cellulitis patients. Even patients who have already been on oral antibiotics seem to fare great when randomized to cephalexin. (Aboltins 2015) In fact, oral antibiotics seem to be the right choice for almost every infectious disease that has been studied. That makes sense, considering that the bacterium living in your tissues remains ignorant of the antibiotic’s port of entry.

Clearly there are times when intravenous therapy is required. If a patient cannot swallow. If the required dose cannot be tolerated orally. If oral antibiotics cannot be absorbed, either because of the chemical structure of the antibiotic, or because of intestinal problems, which often occur in the critically ill. Or in emergent scenarios, when rapidly achieving peak antibiotic levels might matter. However, these represent a minority of clinical scenarios, especially when discussing cellulitis.

It is time that we dispel the magical thinking that surrounds IV antibiotics.

Newer post on oral vs IV antibiotics: Oral antibiotics are equivalent to IV (again) – the OVIVA trial (Li 2019)

Related question: do you need antibiotics at all in dog bites?

Other EBM Resources

Evidence based medicine is easy

The EBM bibliography

Evidence based medicine resources

EBM deep dives

References

Addo-Yobo E, Chisaka N, Hassan M. Oral amoxicillin versus injectable penicillin for severe pneumonia in children aged 3 to 59 months: a randomised multicentre equivalency study. Lancet (London, England). ; 364(9440):1141-8. PMID: 15451221

Agweyu A, Gathara D, Oliwa J. Oral amoxicillin versus benzyl penicillin for severe pneumonia among kenyan children: a pragmatic randomized controlled noninferiority trial. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2015; 60(8):1216-24. PMID: 25550349

Atkinson M, Lakhanpaul M, Smyth A. Comparison of oral amoxicillin and intravenous benzyl penicillin for community acquired pneumonia in children (PIVOT trial): a multicentre pragmatic randomised controlled equivalence trial. Thorax. 2007; 62(12):1102-6. PMID: 17567657

Edgeworth J. Intravascular catheter infections. The Journal of hospital infection. 2009; 73(4):323-30. PMID: 19699555

Castro-Guardiola A, Viejo-Rodríguez AL, Soler-Simon S. Efficacy and safety of oral and early-switch therapy for community-acquired pneumonia: a randomized controlled trial. The American journal of medicine. 2001; 111(5):367-74. PMID: 11583639

Conterno LO, Turchi MD. Antibiotics for treating chronic osteomyelitis in adults. The Cochrane database of systematic reviews. 2013; PMID: pubmed

Haran JP, Hayward G, Skinner S. Factors influencing the development of antibiotic associated diarrhea in ED patients discharged home: risk of administering IV antibiotics. The American journal of emergency medicine. 2014; 32(10):1195-9. PMID: 25149599

Hazir T, Fox LM, Nisar YB. Ambulatory short-course high-dose oral amoxicillin for treatment of severe pneumonia in children: a randomised equivalency trial. Lancet (London, England). 2008; 371(9606):49-56. PMID: 18177775

Heldman AW, Hartert TV, Ray SC. Oral antibiotic treatment of right-sided staphylococcal endocarditis in injection drug users: prospective randomized comparison with parenteral therapy. The American journal of medicine. 1996; 101(1):68-76. PMID: pubmed

Kilburn SA, Featherstone P, Higgins B, Brindle R. Interventions for cellulitis and erysipelas. The Cochrane database of systematic reviews. 2010; PMID: 20556757

Li HK, Agweyu A, English M, Bejon P. An unsupported preference for intravenous antibiotics. PLoS medicine. 2015; 12(5):e1001825. PMID: 25992781

MacGregor RR, Graziani AL. Oral administration of antibiotics: a rational alternative to the parenteral route. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 1997; 24(3):457-67. PMID: 9114201

Maki DG, Kluger DM, Crnich CJ. The risk of bloodstream infection in adults with different intravascular devices: a systematic review of 200 published prospective studies. Mayo Clinic proceedings. 2006; 81(9):1159-71. PMID: 16970212

Oosterheert JJ, Bonten MJ, Schneider MM. Effectiveness of early switch from intravenous to oral antibiotics in severe community acquired pneumonia: multicentre randomised trial. BMJ (Clinical research ed.). 2006; 333(7580):1193. PMID: 17090560

Siegel RE, Halpern NA, Almenoff PL, Lee A, Cashin R, Greene JG. A prospective randomized study of inpatient iv. antibiotics for community-acquired pneumonia. The optimal duration of therapy. Chest. 1996; 110(4):965-71. PMID: 8874253

Strohmeier Y, Hodson EM, Willis NS, Webster AC, Craig JC. Antibiotics for acute pyelonephritis in children. The Cochrane database of systematic reviews. 2014; PMID: 25066627

Vidal L, Ben Dor I, Paul M. Oral versus intravenous antibiotic treatment for febrile neutropenia in cancer patients. The Cochrane database of systematic reviews. 2013; PMID: 24105485

Vogel F, Lode H. The use of oral temafloxacin compared with a parenteral cephalosporin in hospitalized patients with pneumonia. The Journal of antimicrobial chemotherapy. 1991; 28 Suppl C:81-6. PMID: 1664833