The efficacy and safety of polyethylene glycol 3350 in the treatment of constipation in children

Pediatric Nursing, May-June, 2004 by Margo A. Kinservik, Margaret M. Friedhoff

Critique of the Evidence

The strength of all of the studies reviewed in this article is that they were all prospective studies, which are considered stronger than retrospective studies. All of the studies also had clear exclusion criteria, which eliminated children who had past gastrointestinal surgery and organic disease. All used a convenience sample of children who presented to their gastroenterology practices with constipation. A weakness of convenience sampling is that this group of children may not be representative of all children with constipation, since not all children with constipation are referred to sub-specialists. The race of the children in the study was only cited in the Gremse et al. (2002) study. Therefore, it was difficult to assess whether the children in the studies represented the entire population of children with constipation.

The observational comparison study by Loening-Baucke (2002) was weakened by the selection criteria for the laxative used. Instead of a random selection, the MiraLax was given to the children in the study who had refused to take milk of magnesia in previous treatment trials. However, since the study was evaluating the acceptability of the MiraLax, this selection process was probably valid. Observational studies may be unintentionally biased by the researcher's preconceptions, which could weaken the results of the study (Polit & Hungler, 1995). Loening-Baucke (2002) did state in her paper that a randomized controlled trial comparing MiraLax and milk of magnesia has been started.

Three of the studies were rather short in duration (3 days, 8 weeks, and 14 days), which provides evidence about the initial effects of the laxative, but does not demonstrate the long-term ramifications of its use. Another weakness of four of the studies was their small sample size. Only the study by Pashankar et al. (2003) had a sample size of over 45 patients. However, despite the small sample size, statistically significant differences were found in most studies.

The study with the strongest design was the prospective, double-blind, parallel, randomized study of disimpaction using PEG 3350 by Youssef et al. (2002). A double-blind study removes the observer bias, which could have been present in the other studies. This study randomized four different doses of the medication to determine the most effective dose for fecal disimpaction. All dosage groups had the same number of subjects. Although the laxative was only given for 3 days in this study, it was appropriate since the studying was only trying to determine the dose needed for disimpaction, not continued laxative therapy.

None of the investigators in the first four studies acknowledged any support from the manufacturer of MiraLax (Braintree Laboratories), nor did they report any other relationship with the manufacturer. This is a strength of those studies, since sponsorship by a drug manufacturer can lead to bias in the design and interpretation of the research study. Only the research by Pashankar et al. (2003) reported financial assistance by Braintree Laboratories.