Chronic anal fissure: botulinum toxin type A injection

Botulinum toxin injection compared with placebo or lidocaine

The combined odds ratio (OR) from the 3 RCTs (Maria et al. 1998, Colak et al. 2002 and Siproudhis et al. 2003) identified in the Cochrane systematic review favoured botulinum toxin type A over placebo but the difference was not statistically significant (pooled OR for non-healing of fissure with botulinum toxin type A injection compared with placebo or lidocaine 0.29, 95% confidence interval [CI] 0.02 to 3.61, p=0.34). The crude combined fissure healing rate was 56.3% using botulinum toxin type A compared with 29.2% using placebo or lidocaine pomade.

The 3 studies used botulinum toxin type A injections (2 used Botox, 1 used Dysport) of between 20 and 25 bioequivalent units and followed up participants for between 1 and 3 months to assess fissure healing.

The Cochrane review reported on the risk of bias in each of its included studies. It assessed Maria et al. (1998) as having a low risk of bias from appropriate allocation concealment and intention-to-treat analysis but a high risk of bias because of its short follow-up period (1–2 months). Allocation concealment was not used in Colak et al. (2002), and there was a high risk of bias caused by selective reporting of results and a short follow-up period (2 months). Siproudhis et al. (2003) had adequate allocation concealment but high risk of bias because it lacked an intention-to-treat analysis and had a short follow-up period (12 weeks).

Yiannakopoulou et al. (2012) identified the same 3 RCTs and noted that they differed in the volume of solution and preparation of botulinum toxin type A injection used. In addition, 1 of the trials (Siproudhis et al. 2003) appeared underpowered to detect anything other than very large differences in efficacy. This may have inhibited the ability of the pooled meta-analysis to detect a statistically significant difference between botulinum toxin type A injection and placebo.

Botulinum toxin injection compared with topical glyceryl trinitrate (GTN) or isosorbide dinitrate ointment

The combined odds ratio (6 RCTs, 334 patients) in the Cochrane systematic review favoured botulinum toxin type A over GTN ointment (mostly 0.2%) or isosorbide dinitrate 1% ointment but the difference was not statistically significant (pooled OR for non-healing of fissure with botulinum toxin type A injection compared with GTN or isosorbide dinitrate ointment 0.56, 95% CI 0.20 to 1.57, p=0.27). The crude combined fissure healing rate was 73.5% using botulinum toxin type A injection compared with 62.5% using GTN ointment (5 RCTs) or isosorbide dinitrate ointment (1 RCT).

The 6 studies used botulinum injections of between 20 and 30 units of Botox or 90 units of Dysport (with the exception of 1 study using Botox at a very low dose of 5 units). This was compared with 0.2% GTN ointment applied 2 or 3 times a day (Brisinda et al. 1999, De Nardi et al. 2006, Uluutku et al. 2001, Brisinda et al. 2007), 0.3% GTN ointment applied 3 times a day (Gecim et al. 2001 [unpublished data only]), and 1% isosorbide dinitrate applied 6 times a day (Festen et al. 2009). Follow-up varied significantly from 30 days to 36 months. However, 4 of the 6 studies had a maximum follow-up period of 4 months or less.

The individual study results were not consistent, with some studies finding the results were in favour of botulinum toxin type A, while others favoured GTN.

Noticeably, none of the studies identified compared botulinum toxin type A with 0.4% GTN ointment, the only licensed treatment for chronic anal fissure in the UK. The unlicensed 0.2% GTN ointment was the main comparator in the majority of cases.

The Cochrane review reported that Brisinda et al. (1999) had a low risk of bias in relation to allocation concealment and blinding of participants, personnel and outcome assessment, but a high risk of bias because of a short follow-up period (2 months). The risk of bias in Gecim (2001) (unpublished data only), was generally unclear or high, because allocation concealment was not used and follow-up duration was variable (3–18 months). De Nardi et al. (2006) had an unclear risk of bias because of a lack of allocation concealment but a low risk of bias because no participants were reported to have dropped out of the study and it had a long follow-up period (36 months).

Uluutku et al. (2001) had a low number of participants who dropped out of the study but a high risk of bias because of a short follow-up period (30 days) and an unclear risk of bias because of a lack of allocation concealment. In Brisinda et al. (2007) allocation concealment was unclear although single blinding of the clinical assessors was used (patients were aware of their treatment). The study showed a high risk of bias because of a short follow-up period (2 months).

Festen et al. (2009) was assessed as having a low risk of bias because of allocation concealment but a high risk of bias because of a lack of statistical power and a high proportion of participants who dropped out (more than 30%). The follow-up period was also relatively short (up to 4 months).

Botulinum toxin injection compared with lateral internal sphincterotomy

The combined odds ratio (5 RCTs, 365 patients) in the Cochrane systematic review showed that surgery was much more likely to result in fissure healing than botulinum toxin type A injection; a statistically significant difference was seen (pooled OR for non-healing of fissure with botulinum toxin type A compared with surgery 7.20, 95% CI 3.97 to 13.07, p<0.00001). The crude combined fissure healing rate in this meta-analysis was 59.0% with botulinum toxin type A injection and 89.3% with lateral internal sphincterotomy.

The 5 studies used botulinum toxin type A injection at varying doses of between 10 and 25 units as well as based on weight (0.3 units per kg). Follow-up was generally longer than studies included in the other meta-analyses in the Cochrane systematic review and ranged from 18 weeks to 3 years. The number of participants who dropped out was variable and this may have been a significant source of potential bias in some studies.

The Cochrane review reported on the risk of bias in each included study. Nasr et al. (2010) was generally considered to have a high risk of bias because it did not use allocation concealment, the number of participants who dropped out was not reported, it was not blinded and it had a short follow-up period (18 weeks). Although Mentes et al. (2003) did not use allocation concealment, it was generally considered to be at low risk of bias because it used an intention-to-treat analysis, reported that no participants dropped out of the study and had a 6-month follow-up period. Iswariah et al. (2005) did not use allocation concealment and reported that 6 participants dropped out of the study (high risk of bias). There was a low risk of bias from its 26-week follow-up period. Suknaic et al. (2008) did not use allocation concealment and more than 10% of participants dropped out. The 6-month follow-up period posed a low risk of bias. Allocation concealment was not possible in Arroyo et al. (2005), no participants were reported to have dropped out of the study and the study follow-up period was long (3-years).

Meta-analysis of 4 (Mentes et al. 2003; Arroyo et al. 2005; Suknaic et al. 2008; Nasr et al. 2010) of the 5 studies (321 participants) in the Cochrane review found that surgery was associated with a statistically significantly higher risk of minor incontinence than botulinum type A injection (pooled OR 0.11, 95% CI 0.02 to 0.46, p=0.0028). Minor incontinence was described as synonymous with incontinence to flatus or anal seepage. No cases of minor incontinence were reported in the botulinum toxin type A injection trial arms of the 4 included trials compared with a crude rate of 9% in the lateral internal sphincterotomy arms.

Recurrence, dose and toxin type

The Cochrane review reported the rate of recurrence of healed fissure after using botulinum toxin type A exceeded 50% after 1 year in 1 RCT (Arroyo et al. 2005) and exceeded 40% after 42 months in a case series (Minguez et al. 2002).

Yiannakopoulou et al. (2012) included 18 studies providing recurrence data. It reported follow-up periods ranging from 6 months to 5 years with recurrence in the botulinum arm ranging from 0% (24 months) to 52.5% (5-year follow-up).

The Cochrane review reported that neither the dose nor the type of botulinum toxin type A had been found to alter fissure healing rates. However, this conclusion was based on relatively few, small and short-term trials.

Both the Cochrane review and Yiannakopoulou et al. (2012) identified a single small RCT of 50 participants (Maria et al. 2000) as the evidence for optimum site of injection. It showed injections of botulinum toxin type A either side of the anterior midline resulted in improved lowering of resting anal pressure and produced earlier healing results than injections either side of the posterior midline. This study had a mean follow-up period of 18 months and reported healing rates of 60% (15/25) in the posterior midline group and 88% (22/25) in the anterior midline group. The study reported allocation concealment and was double blinded, but had a short follow-up period (2 months). It was also small and the results have not been replicated in larger studies so may not be reliable.

Overall healing rates

The overall healing rate with botulinum toxin type A injection was 76.8% in the Cochrane review; however, this included 2 studies with abnormally high healing rates (greater than 90%). When these outliers were excluded, the average healing rate was 67.5%.

The review highlighted that botulinum toxin efficacy varied, working much better in some studies than others. They pointed to possible different effects, with the site of injection as a potential influencing factor (Maria et al. 2000), although other factors may be involved.