Summary of main results
(1) For double versus single freeze technique cryotherapy, the evidence suggests that cryotherapy should be used with a double freeze technique rather than single freeze in order to reduce the risk of residual disease within 12 months, although statistical significance was not reached. The single freeze technique had higher treatment failure rates.
(2) Laser ablation demonstrated no overall difference in residual disease after treatment for CIN compared with cryotherapy. Cryosurgery appears to have a lower success rate but the majority of authors used a single freeze thaw technique. Creasman (Creasman 1984) demonstrated that using a double freeze-thaw-freeze technique improves results towards those achieved by destructive and excisional methods. However, analysis of results demonstrated that there was no significant difference for the treatment of CIN1 and 2; laser ablation appeared to be better, but not significantly so, for treating CIN3. The clinician's choice of treatment of low grade disease must therefore be influenced by the side effects related to the treatments.
Laser ablation was associated with significantly fewer vasomotor symptoms and less malodorous discharge or inadequate colposcopy at follow up compared with cryotherapy. No other statistical differences were observed in any other side effects, although there may be more peri-operative pain and bleeding for laser ablation. Since the number of events was low, this needs to be explored further.
(3) Four trials compared laser conisation and knife conisation (Bostofte 1986; Kristensen 1990; Larsson 1982; Mathevet 1994). For the two trials that evaluated residual disease after laser conisation or knife conisation, no significant difference was observed between the two groups. There was also no evidence of a difference between the two interventions for primary and secondary haemorrhage. Significant thermal artefact prevented interpretation of resection margins in 38% of laser cones compared to none in the knife cones, which was statistically significant. Laser conisation produced significantly fewer inadequate colposcopes (transformation zone seen in its entirety) at follow up and cervical stenosis was significantly less common after this treatment.
(4) Only the trial of Partington 1989 compared laser conisation with laser ablation for ectocervical lesions. There was no significant difference with respect to residual disease at follow up, peri-operative severe bleeding, secondary haemorrhage or inadequate colposcopy at follow up.
(5) Six trials compared laser conisation with large loop excision of the transformation zone (LLETZ) (Crompton 1994; Mathevet 1994; Oyesanya 1993; Paraskevaidis 1994; Santos 1996; Vejerslev 1999). There was no significant difference with respect to residual disease at follow up, peri-operative severe pain, secondary haemorrhage, significant thermal artefact, inadequate colposcopy or cervical stenosis. However, laser conisation takes significantly longer to perform, has a significantly higher rate of perioperative bleeding and produces a greater depth of thermal artefact.
(6) Laser ablation compared to LLETZ was evaluated by four trials. Alvarez 1994 was included in the comparison but its methodology differed from the trials of Dey 2002, Gunasekera 1990 and Mitchell 1998. The Alvarez 1994 trial performed LLETZ on all the patients randomised to that group whereas laser ablation was only performed if colposcopic directed biopsies were performed. There was no difference in residual disease rates between the two treatments. There was no significant difference in the risk of primary or secondary haemorrhage or peri-operative severe pain.
(7) For knife cone biopsy compared to loop excision, (a) six randomised trials evaluated knife cone biopsy and loop excision (Duggan 1999, Giacalone 1999, Girardi 1994, Mathevet 1994, Sadek 2000, Takac 1999). The trials found that there was no evidence of a difference between the two interventions on residual disease rate.
(b) Measuring primary haemorrhage, the trials of Giacalone 1999, Duggan 1999, Mathevet 1994 found that there was no statistical difference in inadequate colposcopy rates between knife conisation and loop excision. There was also no clear evidence that there was any difference in primary haemorrhage or cervical stenosis rates.
(8) For radical diathermy versus LLETZ, there was no significant difference between these two modalities with regards to the side effects reported, with exception of significantly increased vaginal pain in those undergoing radical diathermy. Residual disease rates were not an outcome measure in the single trial identified.
(9) For haemostatic sutures, there was no evidence that haemostatic sutures were significantly different for the risk of primary haemorrhage or cervical stenosis compared to using no routine sutures or vaginal packing in the two included trials (Gilbert 1989; Kristensen 1990). Use of haemostatic sutures did however increase the risk of secondary haemorrhage, dysmenorrhoea and inadequate follow-up colposcopy.
(10) One trial compared the use of bipolar electrocautery scissors with a monopolar energy scalpel during LLETZ (Cherchi 2002). Bipolar electrocautery scissors were associated with a significant reduction in perioperative bleeding and duration of the procedure but no change in the rate of primary haemorrhage.
(11) One trial compared the use of LEEP versus cryotherapy (Chirenje 2001). This trial found that women who received the loop electrosurgical excisional procedure (LEEP) had significantly lower rates of watery discharge and residual disease at 12-month follow up but an increased risk of secondary haemorrhage and offensive discharge. There was no significant difference in the rates of primary haemorrhage, residual disease at six months or peri-operative severe pain.
(12) One trial compared pure cut settings versus blend settings for LLETZ (Nagar 2004) and found no significant difference in the rates of residual disease between the settings but a reduced depth of thermal artefact at the deep stromal margin in women whose surgeon used a pure cut setting for LLETZ.
(13) Two trials compared LLETZ and needle excision of the transformation zone (NETZ) (Panoskaltsis 2004a; Sadek 2000) but reported on different outcomes. There was no significant difference between the techniques in terms of perioperative pain, bleeding requiring vaginal packing or cervical stenosis at follow up. LLETZ was associated with a reduction in peri-operative blood loss but an increase in residual disease rates at 36-month follow up. There was no difference in residual disease rates for NETZ compared to knife conisation.
Overall completeness and applicability of evidence
The incidence of treatment failures following surgical treatment of CIN has been demonstrated by case series reports, as illustrated in the Background section, to be low. The reports from randomised and non-randomised studies suggest that most surgical treatments have around 90% success rate. In these circumstances, several thousand women would have to be treated to demonstrate a significant difference between two techniques. The vast majority of RCTs evaluating the differences in treatment success are grossly underpowered to demonstrate a significant difference between treatment techniques and no real conclusions can be drawn on differences of treatment effect. The largest of these studies recruited 498 participants (Mitchell 1998) and the smallest recruited 40 women (Cherchi 2002; Paraskevaidis 1994). It might be the case that if a well-conducted mega-trial was conducted no difference in treatment effect would be demonstrated. The RCTs and meta-analyses have demonstrated some clear differences in morbidity and these should be considered as significant outcomes when deciding upon optimum management.
The trials compare different interventions and report different outcomes, which limits the analyses and means that many outcome measures include only one trial per treatment pairing.
Quality of the evidence
In total, 29 trials were included in this review. A total of 5441 women participated of whom 4509 were analysed. We have used a pragmatic approach to the RCTs included in the comparisons. Slight variations of surgical technique occur in some of the comparisons, which reflects the differences in clinical practice. If we considered that these differences did not seriously alter the intervention compared with the other interventions in the comparison, then the trial was considered in the same analysis. For example, when we compared laser ablation to cryotherapy, we included trials using single and double freeze techniques.
Many analyses included only one or two randomised trials due to the different outcome measures chosen and reported in the trials. This limits the conclusions which may be drawn from some of the analyses. Furthermore, the method of randomisation in many of the trials was not optimised so that the results might be prone to bias due to inherent methodological flaws in these trials.
Potential biases in the review process
A comprehensive search was performed, including a thorough search of the grey literature, and all studies were sifted and data extracted by at least two review authors working independently. We restricted the included studies to RCTs as they provide the strongest level of evidence available. Hence, we have attempted to reduce bias in the review process.
The greatest threat to the validity of the review is likely to be the possibility of publication bias. That is, studies that did not find the treatment to be effective may not have been published. We were unable to assess this possibility as the analyses were restricted to meta-analyses of a small number of trials or single trials.
Agreements and disagreements with other studies or reviews
The conclusions reflect the previous findings from the original Cochrane review by the authors. Furthermore, a Canadian group published an independent systematic review on the same subject and the findings were the same as the original review (Nuovo 2000). The review by Nuovo 2000 used similar methodologies as the original Cochrane review and used quasi-randomised trials as well as gold standard RCTs within their meta-analyses.
The single RCT by Dey 2002 almost demonstrated a significant reduction in treatment failures with LLETZ compared to laser ablation, in contrast to other studies. This trial included HPV testing as well as cytology for screening for treatment failures, which enhances the detection of disease.