Evidence strengths and limitations

Evidence strengths and limitations

The quality and relevance of the evidence comparing the efficacy and safety of 0.2% glyceryl trinitrate ointment to 0.4% glyceryl trinitrate ointment is very limited for the following reasons:

  • No trials were identified that were specifically designed to detect statistical differences between the 0.2% and 0.4% glyceryl trinitrate ointments, such as a randomised controlled trial (RCT) appropriately powered to detect between-group differences or a non-inferiority trial.

  • The 3 RCTs that recruited adults, and the 1 pilot RCT that recruited children, were not statistically powered to detect differences in glyceryl trinitrate strength. This means that clinically significant differences in strength-related efficacy and headache side effects may have been missed.

  • The 3 RCTs that recruited adults were small. In 2 of the trials, 47 (Scholefield et al. 2003) and 23 patients (Carapeti et al. 1999) were in the 0.2% glyceryl trinitrate arms; however, patient numbers were not reported in the third trial (Bailey et al. 2002). The pilot RCT (Simpson et al. 2003) was even smaller (n=15, all participants).

  • The duration of follow-up in the adult studies (8 to 10 weeks) was very short. This is a major limitation as fissures can heal and relapse over various time frames. This is demonstrated in Scholefield et al. (2003), in which the fissures of 37.5% of patients who were given placebo showed healing at 8 weeks. Because of the short-term follow-up, the efficacy estimates do not take into account relapse rates after 8 weeks.

  • The only trial (Carapeti et al. 1999) that performed statistical tests comparing glyceryl trinitrate strengths did not compare 0.2% ointment with 0.4% ointment. It compared 0.2% ointment against 0.2% ointment with weekly 0.1% increments to a maximum 0.6%, a non-standard treatment regimen. Non-significant findings between glyceryl trinitrate strengths in this study may be because of a lack of statistical power to detect such differences.

  • Assessment of safety and adverse events was generally restricted to the 8-week treatment period.

  • The trials used different definitions of chronic anal fissure. Results from Scholefield et al. (2003) indicated that healing rates were influenced by the definition of anal fissure chronicity used.

  • Only 1 of the 4 RCTs explicitly described that treatment allocation had been concealed (Carapeti et al. 1999).

A more general limitation to consider is that the method of applying glyceryl trinitrate ointment is imprecise and so, the actual strength of glyceryl trinitrate may vary greatly from patient to patient and application to application, even when using the same strength of ointment. This may confound any association between glyceryl trinitrate strength and fissure healing or between glyceryl trinitrate strength and key side effects such as headache. In addition, it is not known whether applying less ointment of the same strength, rather than reducing the strength of ointment applied, might have an effect on the incidence or headache.