The authors concluded that early conversion to oral antibacterial treatment for moderate to severe community-acquired pneumonia appeared to be as effective as traditional intravenous treatment and was associated with fewer drug-related adverse events and shorter hospital stays. Limitations of the review data and its interpretation suggest that the authors' conclusions may not be reliable.

To assess the effectiveness and safety of early switch strategies compared to conventional intravenous antibacterial treatment for moderate to severe community-acquired pneumonia (CAP).

MEDLINE/PubMed and Scopus were searched in February 2008 without language restriction. The reference lists of articles retrieved were handsearched. Search terms were reported.

Randomised controlled trials (RCTs) comparing an early switch from intravenous to oral antibacterial treatment with conventional intravenous treatment for hospitalised patients with moderate to severe CAP were eligible for inclusion. The primary outcomes of the review were treatment success (cure or clinical improvement), all-cause mortality and length of hospital stay; secondary outcomes were also specified. Criteria for the diagnoses of pneumonia and severe pneumonia were detailed in the review. The study population was defined as severely ill if over 75 per cent had severe pneumonia on hospital admission. An early switch was defined as administration of oral antibiotics to clinically improving patients who had been treated with intravenous antibiotics for two to four days.

The mean/median age of included participants ranged from 54 to 70 years. Most were patients male (where stated). Half of the studies included patients with severe CAP (where stated). Most studies administered the same type of antibacterials to both groups, commonly including second or third generation cephalosporins. In all cases the duration of treatment was the same in both groups. Most studies applied precise criteria for conversion to oral treatment and/or specified a specific day for the switch. Few described criteria for hospital discharge. Criteria for hospitalisation and for assessing CAP severity differed.

Two reviewers independently selected studies for inclusion.

Validity assessment was conducted using a modified Jadad scale (Moher 1998) to evaluate description/method of randomisation and blinding, and description of exclusion criteria. The authors did not state how the assessment was performed.

Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for dichotomous outcomes. Mean differences were calculated for continuous outcomes. Outcomes were extracted using intention-to-treat (ITT) principles, and/or among patients who were clinically evaluable (CE) or microbiologically evaluable (ME) at follow up. Two reviewers independently extracted the data.

ORs were combined using a Mantel-Haenszel fixed-effect model. Continuous data were combined to calculate weighted mean differences (WMDs), with 95% CIs. Where there was significant heterogeneity (assessed using the I² statistic), a DerSimonian random-effects model was used. Subgroup analyses were conducted for the primary outcomes to examine whether effects differed according to illness severity.

Six RCTs were included (n=1,219): four scored 3 points for quality; two scored 2; and one scored 1. None were blinded. In all studies prognostic factors in the two groups were comparable at baseline.

There were no statistically significant differences between early switch and conventional treatment groups in terms of treatment success (three RCTs, n=987 intention-to-treat patients), mortality (five RCTs, n=619 intention-to-treat patients) or recurrent infection rates (five RCTs, n=385 clinically or microbiologically evaluable patients).

Early switch treatment was associated with a significantly shorter hospital stay (WMD -3.34 days, 95% CI: -4.42 to -2.25; five RCTs, n=600 clinically evaluable patients, I²=88.5%) and fewer drug-related adverse events (OR 0.65, 95% CI: 0.48 to 0.89; five RCTs) or withdrawals for adverse events (OR 0.49, 95% CI: 0.27 to 0.89; four RCTs). Subgroup analyses in clinically evaluable patients with severe CAP showed similar results. There were insufficient data for the assessment of pathogen eradication.

Early conversion to oral antibacterial treatment for moderate to severe community-acquired pneumonia appeared to be as effective as traditional intravenous treatment and was associated with fewer drug-related adverse events and shorter hospital stay.

The objectives and inclusion criteria of the review were clear. Relevant sources were searched for studies without language restriction. As no specific attempts were made to locate unpublished studies, there was a risk of publication bias. Publication bias was not formally assessed. Steps were taken to reduce the risk of error and bias by having more than one reviewer conduct study selection, data extraction and validity assessment. Few details were reported on the quality characteristics of individual studies (such as allocation concealment and number of withdrawals). Study validity was not explicitly taken into account in the interpretation of findings. Suitable statistical techniques were used to combine studies. Appropriate methods were used to assess both statistical and clinical heterogeneity. Differences between the studies were discussed in the text, as were other limitations that might have affected the generalisability and reliability of the findings. The review was limited by several factors: there were few studies; sample sizes were small; and there was significant statistical heterogeneity for the only outcome with significant findings for which the results of the heterogeneity assessment were reported (length of hospital stay). The authors appeared to suggest that lack of evidence that the treatments differ equated to evidence that they were equally effective. Overall, the limitations of the review data and its interpretation suggest that the authors' conclusions may not be reliable.

Practice: The authors stated that the evidence appeared to support guideline recommendations for early switch strategies in clinically stable hospital patients with CAP, even those with more severe forms of CAP.

Research: The authors stated that further research was needed to compare rates of recurrent infection associated with early switch and traditional intravenous treatment in patients with CAP.

No external funding sources.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.