How Long is Long Enough?

Abstract & Commentary

By Allan J. Wilke, MD, Residency Program Director, Associate Professor of Family Medicine, University of Alabama at Birmingham School of Medicine—Huntsville Regional Medical Campus. Dr. Wilke reports no financial relationship to this field of study.

Synopsis: Three days of amoxicillin is as effective as five for mild-to-moderately severe community-acquired pneumonia.

Source: el Moussaoui R, et al. Effectiveness of discontinuing antibiotic treatment after three days versus eight days in mild to moderate-severe community acquired pneumonia: randomised, double blind study. BMJ. 2006;332:1355-1361.

El Moussaoui and colleagues from Amsterdam enrolled 186 patients in this non-inferiority study. The inclusion criteria were age ≥ 18 years with clinical signs of pneumonia, temperature ≥ 38° C (100.4° F), new infiltrate consistent with pneumonia on chest x-ray (CXR), and a Pneumonia Severity Index (PSI)1 ≤ 110. They also included patients ≥ 65 years who were afebrile, if they had signs of pneumonia, and a positive CXR. There were many exclusion criteria including pregnancy, allergy to amoxicillin, immunocompromised state, admittance to intensive care, substantial pleural effusion on CXR, and suspicion of aspiration or staphylococcal pneumonia. The patients all had a complete history and physical at baseline. Routine laboratory testing, serum antibody testing for Mycoplasma pneumoniae, Legionella pneumophila, Chlamydia pneumoniae, and respiratory viruses, and sputum cultures were performed at baseline. The antibody studies and CXR were repeated at day 10. All patients received intravenous amoxicillin (a common treatment for mild-to-moderately severe CAP in the Netherlands) for 72 hours and then were re-evaluated. Those who showed improvement in symptoms (decreases in shortness of breath, cough, sputum production, color of sputum, and fever) and who could take oral medication, were randomized to oral amoxicillin 750 mg or placebo 3 times daily for 5 days. Follow up was at days 7, 10, 14, and 28.

The primary outcome measure was clinical cure rate at day 10. Analysis was per protocol and intention to treat. They did subgroup analysis of those individuals in whom a bacterial cause was diagnosed. After eliminating subjects who withdrew their consent, who did not meet inclusion criteria, or who violated protocol, there were 119 patients available for randomization. They ranged in age from 40 to 74 years (median, 57) and were predominantly male (60%). Despite randomization, there were more smokers (55% vs 27%) in the 3-day treatment group than in the 8-day group, and the 3-day group had more severe symptoms at admission. A pathogen was identified in 64 subjects (54%). Streptococcus pneumoniae was the most commonly cultured bacteria (37/64, 58%), followed by Haemophilus influenzae (10/64, 16%). Atypical bacterium (C. pneumoniae, M. pneumoniae, and L. pneumophila) accounted for just one case. Per protocol cure rates were 93% in both treatment groups. By intention-to-treat analysis, both groups had 89% cure rates. At day 28, the per protocol cure rates were 90% for the 3-day group and 88% for the 8-day group, a non-significant difference. Similarly, the cure rates in the intention-to-treat analysis were 84% and 78%, respectively. Mild adverse side effects were reported more frequently in the 8-day group (11 % vs 21%).


The putative benefits of a shorter course of antibiotic therapy are better compliance, reduced expense (cost of medication, treatment of adverse side effects), and reduced induction of antibiotic resistance. At what point does shortening the course decrease the cure rate? For amoxicillin at least, this study shows that a 3-day course is not inferior to an 8-day course, even though the 3-day group had more smokers and worse pneumonia severity indices. Intravenous amoxicillin is not available in the United States, although IV ampicillin is available. However, in this country, IV ampicillin is not the drug usually chosen for empiric therapy of an inpatient with CAP. A respiratory fluoroquinolone (moxifloxacin, gatifloxacin, levofloxacin, or gemifloxacin), an advance macrolide (azithromycin or clarithromycin), and a β-lactam (which would include high-dose amoxicillin and high-dose amoxicillin-clavulanic acid) in various combinations are usually chosen. Respiratory fluoroquinolones and macrolides are frequently recommended, because they have activity against pneumococci and atypical bacteria. Outpatient therapy would include these drugs and erythromycin, doxycycline, and clindamycin.2

There have been other studies of antibiotic duration in pneumonia. A 2004 study in India compared 3 days vs 5 days of oral amoxicillin in 2188 children with pneumonia that could be treated at home. Clinical cure rates for both regimens were almost 90% with no difference in clinical failures or relapses.3 In a recent study,4 439 children were randomized to 3 days of usual-dose oral amoxicillin (45 mg/kg/day) or double dose. There were no differences in treatment failure (about 5% in both groups at day 5). Five days of levofloxacin 750 mg a day had a cure rate of 92.4%, which compared favorably with levofloxacin 500 mg daily for 10 days (91.1%).5 You will note that these rates are not much different from those obtained with high-dose amoxicillin. It should be noted that 20-30% of patients in the levofloxacin study dropped out, often for adverse side effects, and were not included in the analysis. By the way, there is good evidence that hospitalized patients with CAP who have received IV antibiotics and who are stable, do not have to stay in the hospital for observation when switched from IV to oral antibiotics.6

If duration of therapy for pneumonia is in question, so is choice of initial antibiotic. Back in my day, the pneumococcus was uniformly sensitive to penicillin; in my 900-bed hospital, now only 56% of it is. Amoxicillin--clavulanic acid is at 94%, ceftriaxone at 99%, and levofloxacin at 98%. This seems to argue for dumping the "bubble gum medicine," but maybe the Dutch are on to something. In this study, S. pneumoniae accounted for 57% of the identified pathogens, H. influenzae for 16%, and atypicals for < 1%, which begs the question: "Do we really need to cover for atypicals in our patients with pneumonia?" Aspa and colleagues7 prospectively looked at 30-day mortality from pneumococcal CAP in patients who were prescribed β-lactam monotherapy, macrolide monotherapy, β-lactam + macrolide, levofloxacin alone or in combination, or other combinations (including those with aminoglycosides). The 30-day overall survival rate was 85%. The choice of initial antibiotic or combination of antibiotic was not associated with mortality.

The PSI is frequently bundled in personal digital assistant software, such as MedRules,8 which is freely downloadable. The tool assigns points to the following variables: age, gender, nursing home resident, neoplastic disease, liver disease, congestive heart failure, cerebrovascular disease, renal disease, altered mental status, respiratory rate ≥ 30/min, systolic blood pressure < 90 mm Hg, temperature < 35°C or ≥ 40°C, pulse ≥ 125/min, arterial pH < 7.35, blood urea nitrogen ≥ 30 mg/dL, sodium < 130 mmoL/liter, glucose ≥ 250 mg/dl, hematocrit < 30%, partial pressure of oxygen < 60 mm Hg, and pleural effusion. Based on the total points, the patient's risk of mortality is then estimated. It can aid you in deciding who can be treated at home and who needs to be admitted.


1. Fine MJ, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med. 1997;336:243-250.

2. Mandell LA, et al. Update of practice guidelines for the management of community-acquired pneumonia in immunocompetent adults. Clin Infect Dis. 2003;37:1405-1433.

3. Agarwal G, et al. Three day versus five day treatment with amoxicillin for non-severe pneumonia in young children: a multicentre randomised controlled trial. BMJ. 2004;328:791-796. Erratum in: BMJ. 2004;328:1066.

4. Hazir T, et al. Comparison of standard versus double dose of amoxicillin in the treatment of non-severe pneumonia in children aged 2-59 months: A multi-centre, double blind, randomized controlled trial in Pakistan. Arch Dis Child. 2006 Mar 17; [Epub ahead of print].

5. Dunbar LM, et al. High-dose, short-course levofloxacin for community-acquired pneumonia: a new treatment paradigm. Clin Infect Dis. 2003;37:752-760. Erratum in: Clin Infect Dis. 2003;37:1147.

6. Nathan RV, et al. In-hospital observation after antibiotic switch in pneumonia: a national evaluation. Am J Med. 2006;119:512-518.

7. Aspa J, et al. Impact of initial antibiotic choice on mortality from pneumococcal pneumonia. Eur Respir J. 2006;27:1010-1019.

8. Accessed August 21, 2006.