Background Invasive cervical carcinoma is preceded by a precancerous phase, cervical intra-epithelial neoplasia (CIN), which can be detected on cervical smears and confirmed by colposcopy and biopsy. Moderate and severe intra-epithelial neoplasia (CIN2 and CIN3) are treated mainly with surgery to prevent progression to invasive carcinoma. Medical methods of preventing the progression or inducing regression of CIN are needed. Retinoids are potent modulators of epithelial cell growth and differentiation and may have potential for the treatment of CIN. Objectives To ascertain whether retinoids can cause regression or prevent progression of CIN. Search strategy Cochrane Gynaecological Cancer Review Group's Specialised Register and Non-Trials Database, Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 2,2007), MEDLINE and EMBASE (June 2007). Selection criteria Randomized controlled trials (RCTs) and non-RCTs of retinoids for treating CIN in women. Data collection and analysis Two authors independently assessed trial quality and extracted data from the trials. Adverse effects information was also collected from the trials. Main results Five RCTs comparing the efficacy of four different retinoids were identified. Two studies examined the effect on CIN2 and CIN3 of retinoids N-(4- hydroxyphenyl) retinamide (fenretinide) (Follen 2001) and 9-cis-retinoic acid (aliretinoin) (Alvarez 2003) given orally and two examined the effect of all-trans-retinoic acid given topically to the cervix (Meyskens 1994; Ruffin 2004). The fifth study investigated the use of 13-cis-retinoic acid ( isotretinoin) given orally in HIV positive patients with CIN1 and condyloma (Robinson 2002). Four studies reported no significant effect of retinoids on the progression to higher grades of CIN, whilst the fifth did not report data on progression. In all studies retinoids had no significant effect on regression of CIN3. Two studies reported that retinoids were associated with regression of CIN2. One reported a greater complete regression of CIN2 over placebo, which was of borderline statistical significance, odds ratio (OR) = 0.5 ( 95% confidence interval (CI) 0.25 to 1.02). The other study reported a non-significant dose- related trend towards increased rates of complete and partial regression compared with placebo. One study reported a significantly worse outcome in women receiving retinoid, OR for regression = 6.00 ( 95% CI 1.00 to 35.91). In general, the retinoid medications were well tolerated. Authors' conclusions The retinoids studied are not effective at causing regression of CIN3 but may have some effect on CIN2. The data on CIN1 is inadequate. Retinoids are not effective at preventing progression of CIN of any grade. At the doses given and duration of treatment studied, the retinoids were reasonably well-tolerated.
