Evidence review: efficacy

Cochrane systematic review: glyceryl trinitrate compared with placebo

A Cochrane systematic review (assessed as up-to-date September 2011) of non-surgical therapy for anal fissure concluded that glyceryl trinitrate (pooled strengths, range unclear) was found to be marginally, but statistically significantly, better than placebo in healing chronic anal fissure (48.9% compared with 35.5%, p<0.0009) but late recurrence was common, occurring in approximately 50% of people whose fissures were initially cured. This was based on the pooled results from 18 randomised controlled trials (RCTs; 1315 patients), of which 4 included only children (165 children); all studies only looked at chronic anal fissure. In children, the statistically significant benefit of glyceryl trinitrate was lost when a study with an abnormally low placebo response was excluded.

Cochrane systematic review: glyceryl trinitrate strength comparisons

The Cochrane review identified 4 RCTs and pooled the results to compare 'high' with 'low' strength topical glyceryl trinitrate ointment to treat chronic anal fissure in adults and children (n=324). High or low cut-offs were not defined in the review but glyceryl trinitrate ointment strength from individual trials included in the meta-analysis ranged from 0.05% to 0.6%. It found no statistically significant difference for fissure healing (pooled odds ratio (OR) favoured low-strength glyceryl trinitrate compared with high strength: OR 0.91, 95% CI 0.57 to 1.45).

Importantly, this finding does not represent a direct 0.2% with 0.4% glyceryl trinitrate comparison. Also, no pooled estimate of effect was reported for pain reduction or adverse events due to headaches. Similarly, the meta-analysis included pooled results from a study in children (which used 0.05% and 0.1% glyceryl trinitrate strengths) and adults (range 0.1% to 0.6% glyceryl trinitrate), which may have confused the individual glyceryl trinitrate strength response relationships in these distinct groups.

The 4 studies underlying the Cochrane review are summarised below.

Randomised controlled trial by Scholefield et al. (2003)

Scholefield et al. (2003) conducted a parallel group, double-blind, multicentre RCT in 200 adults (mean age 43 years, standard deviation 13 years) with chronic anal fissure (duration of symptoms lasting more than 6 weeks). Participants applied placebo, 0.1%, 0.2% or 0.4% glyceryl trinitrate ointment (1 cm to a site just inside the anus at the junction of the perianal skin and the anal canal) twice daily at approximately 12-hour intervals for 8 weeks. All patients were also instructed to follow a high-fibre diet as well as being given advice on perianal hygiene.

The primary outcome was complete fissure healing at 8 weeks assessed by visual inspection and measurement of the length and width of the fissure. Secondary outcomes included pain intensity on defecation and pain intensity overall, both assessed by a visual analogue score. Maximal anal resting pressure was also recorded in a subset of 38 patients at 1 centre. Patients were withdrawn after 4 weeks if there was a clear lack of efficacy or a complete resolution of symptoms.

The intention-to-treat analysis included 181 patients who had used some of the study medication and were assessed for healing at the end of the study. It did not compare healing outcomes between the different glyceryl trinitrate strengths, only against placebo.

The analysis reported healing rates at 8 weeks of 37.5% (18/48) using placebo, 46.9% (23/49) using 0.1% glyceryl trinitrate ointment, 40.4% (19/47) using 0.2% glyceryl trinitrate ointment, and 54.1% (20/37) using 0.4% glyceryl trinitrate ointment. Each glyceryl trinitrate ointment strength was not statistically significantly different from placebo, which was also the case for all glyceryl trinitrate strengths combined (individual p values for each glyceryl trinitrate strength compared with placebo not given, global test for trend of all glyceryl trinitrate strengths compared with placebo p=0.4).

The differences between pain scores from baseline to 2, 4, 6 and 8 weeks (for pain on defecation and overall pain scores) were not statistically significantly different for each glyceryl trinitrate strength compared with placebo (figures not reported). Reduction in anal pressure at 8 weeks appeared to increase with increased strength of glyceryl trinitrate. However, this relationship was not statistically significant (pooled glyceryl trinitrate groups compared with placebo p=0.77).

Because of the high healing rates in the placebo arm of the trial (37.5%, 18/48), the authors suspected their study population was a mix of acute (more likely to heal spontaneously) and chronic fissure. A secondary analysis restricted results to patients showing 2 or more recognised features of chronicity (sentinel skin tag, hypertrophied anal papillae, exposed internal anal sphincter, fibrotic lateral fissure or fibrotic anal sphincter). This reduced the intention-to-treat sample from 181 to 145, and found statistically significantly higher healing rates for 0.1% glyceryl trinitrate ointment (50.0%, 21/42 p=0.05) and 0.4% glyceryl trinitrate ointment (56.7%, 17/30 p=0.03) compared with placebo (24.3%, 9/37) and for the effect of the trend of all glyceryl trinitrate groups combined compared with placebo (p=0.03). Interestingly, the 0.2% glyceryl trinitrate group was the only strength in this sub-analysis that did not show a statistically significant difference from placebo (36.1%, 13/36 p=0.91). It is important to note, however, that the numbers of participants in each group were small.

In summary, under restricted criteria for fissure chronicity, albeit only assessed in 145 patients, the combined glyceryl trinitrate groups (0.1%, 0.2% and 0.4%) achieved 47% healing rates at 8 weeks, which was statistically significantly better than placebo (p=0.03). However, on its own, the 0.2% glyceryl trinitrate strength was not found to be statistically significantly better than placebo and noticeably bucked the apparent dose-response trend.

Randomised controlled trial by Bailey et al. (2002)

Bailey et al. (2002) conducted a parallel group, double-blind, multicentre RCT in 304 adults (mean age 42 years, range 19 to 81 years) with chronic anal fissure (fissure symptoms for more than 30 days). Participants were randomised to apply placebo, 0.1%, 0.2% or 0.4% glyceryl trinitrate ointment to the distal anal canal and anus twice daily or 3 times daily for up to 8 weeks using a strength measuring device to standardise the delivery of 374 mg of ointment.

Fissure symptom duration at enrolment ranged from 4 weeks to 2 years but was balanced between groups. Overall, 20.7% (63/304) dropped out of the study before 8 weeks but this did not differ significantly between trial groups (p=0.25).

The statistical analysis did not directly compare the healing outcomes between the glyceryl trinitrate groups, only with placebo. The patient numbers in each trial arm were not reported. The intention-to-treat analysis showed no significant difference in fissure healing at 8 weeks between glyceryl trinitrate groups combined compared with placebo (p value not significant), or when individual glyceryl trinitrate ointment strengths (0.1%, 0.2% or 0.4%) were compared with placebo. Similarly, no significant difference was found between twice-daily and 3-times daily administration of the ointment (p value not significant). For the twice-daily dosage schedule, fissure healing was reported in 50% of people on placebo compared with 31%, 26% and 39% of people using the 0.1%, 0.2% and 0.4% glyceryl trinitrate ointment strengths respectively. The placebo healing rate was markedly high.

Additional outcomes including decrease in pain intensity, worst pain, pain at defecation and frequency of headache were reported for pooled 0.4% glyceryl trinitrate strength only (twice daily and 3 times daily pooled).

Randomised controlled trial by Carapeti et al. (1999)

Carapeti et al. (1999) conducted a small double-blind RCT that included 70 adults (median age 35 to 36 years, range 21 to 81 years across all groups) with chronic anal fissure (3 months or longer with features of chronicity, such as fibrosis of the base of the ulcer or an associated sentinel pile). Participants were randomised to receive 8 weeks of treatment with placebo, 0.2% glyceryl trinitrate ointment 3 times daily, or 0.2% glyceryl trinitrate ointment 3 times daily with weekly 0.1% increments to a maximum 0.6%. This was applied to the skin of the anal verge around the anal opening.

Fissure healing was assessed by clinical examination, anal manometry and laser Doppler flowmetry. If healing occurred in the 8-week treatment period, the patients were followed up at 3, 6 and 12 months, or sooner if recurrent symptoms developed. The study did attempt to statistically compare the 2 glyceryl trinitrate strengths for selected secondary outcomes. However, the study may not have been statistically powered to detect anything but large differences between strengths. Moderate to small differences may have been missed.

After 10 weeks (8 weeks treatment plus 2 weeks after), the primary outcome of healing of the fissure, was reported in 32% (7/22) of patients using placebo; 65% (15/23) using 0.2% glyceryl trinitrate ointment and 70% (16/23) using an escalating strength of glyceryl trinitrate ointment (p=0.008 placebo compared with combined glyceryl trinitrate healing rate). No statistical test compared 0.2% glyceryl trinitrate with the escalating strength for healing of fissure at 8 weeks (the primary outcome). Median time to healing was 8 weeks (range 4 to 10 weeks). A higher proportion of fissures had healed after 6 weeks treatment in the escalating glyceryl trinitrate group (39%, 9/23) compared with the 0.2% glyceryl trinitrate group (22%, 5/23) but this was not statistically significant (p=0.33). There was no statistically significant difference in mean pain score between the 2 glyceryl trinitrate groups from baseline to 8 weeks (p=0.7). In addition, although statistically significant reductions from the pre-treatment pain score were reported in each glyceryl trinitrate group (both p<0.0001), the average pain score after glyceryl trinitrate (combined) did not differ from placebo (p=0.4). Similarly, no significant difference was found for anodermal blood flow between the 2 glyceryl trinitrate groups (p=0.2), or when the combined average of the 2 glyceryl trinitrate groups was compared with placebo (p=0.5).

There was a reduction in maximal anal sphincter resting pressure from pre-treatment to 8 weeks for all 3 treatment groups (results displayed graphically). However, no significant differences were reported between the 2 glyceryl trinitrate strengths (p=0.7%).

No significant difference was found for recurrence of anal fissure after initial healing during a median follow-up period of 9 months (range 6 to 14 months) for the 2 glyceryl trinitrate groups (p=0.7). A total of 33% (5/15) of fissures recurred in the 0.2% glyceryl trinitrate group compared with 25% (4/16) in the escalating strength group. Recurrence for placebo was 43% (3/7); no p value was reported.

Pilot randomised controlled trial by Simpson et al. (2003)

Simpson et al. (2003) conducted a pilot double-blind RCT in only 15 children (median ages 4.5 and 7 years, range 3 to 13 years for both trial arms) with chronic anal fissure (visible presence of fissure, evidence of fibrosis at the base and persistence of symptoms for more than 3 months). The children were randomised to receive 0.05% (n=7) or 0.1% (n=8) glyceryl trinitrate ointment twice daily, applied to the distal end of the anal canal for 8 weeks. Their parents were shown how to apply the ointment correctly.

The researchers stated they used lower strengths of ointment (0.05% and 0.1%) to minimise the risk of headache in children because a high incidence of headache had been reported in adults using 0.2% glyceryl trinitrate ointment. Most children were taking laxatives at study enrolment (13/15); this medication was not altered.

At 8 weeks, the fissures of all 7 children using 0.05% glyceryl trinitrate ointment, and of 5 out of 8 children using 0.1% glyceryl trinitrate ointment, had healed and 'symptoms had settled' (no significant difference, no p value reported). The fissures in 2 of the 3 children whose fissures had not healed using 0.1% ointment healed after a second 8-week treatment; the fissure in 1 child had not healed at 16 weeks, but became asymptomatic with no evidence of fissure at 1-year follow-up. This child was receiving chemotherapy for leukaemia at enrolment. No fissure recurrence was found in the remaining 14 children followed up for a median 4.5 months (range 2 to 12 months).