Physical methods for treating fever in children

Plain language summary pending.

Authors' conclusions: 

A few small studies demonstrate that tepid sponging helps to reduce fever in children.

Read the full abstract...

Health workers recommend bathing, sponging, and other physical methods to treat fever in children and to avoid febrile convulsions. We know little about the most effective methods or how these methods compare with commonly used drugs.


To evaluate the benefits and harms of physical cooling methods used for managing fever in children.

Search strategy: 

We searched the Cochrane Infectious Diseases Group's trials register (October 2005), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library Issue 3, 2005), MEDLINE (1966 to October 2005), EMBASE (1988 to October 2005), LILACS (October 2005), CINAHL (1982 to October 2005), Science Citation Index (1981 to October 2005), and reference lists of articles. We also contacted researchers in the field.

Selection criteria: 

Randomized and quasi-randomized controlled trials comparing physical methods with a drug placebo or no treatment in children with fever of presumed infectious origin. We included studies where children in both groups were given an antipyretic drug.

Data collection and analysis: 

Two reviewers independently assessed trial methodological quality. One reviewer extracted data and the other checked the data for accuracy. Results were expressed as risk ratio with 95% confidence intervals for binary outcomes, and mean difference for continuous data.

Main results: 

Seven trials, involving 467 participants, met the inclusion criteria. One small trial (n = 30), comparing physical methods with drug placebo, did not demonstrate a difference in the proportion of children without fever by one hour after treatment in a comparison between physical methods alone and drug placebo. In two studies, where all children received an antipyretic drug, physical methods resulted in a higher proportion of children without fever at one hour (n = 125; risk ratio 11.76; 95% confidence interval 3.39 to 40.79). In a third study (n = 130), which only reported mean change in temperature, no difference was detected. Mild adverse events (shivering and goose pimples) were more common in the physical methods group (3 trials; risk ratio 5.09; 95% confidence interval 1.56 to 16.60).