Three controlled trials (n=93) were included: two parallel-group trials (n=70) and one crossover trial (n=23).
In terms of study quality, none of the trials reported the method of randomisation. One trial used triple-blinding; the other two trials were open label. Two trials analysed patients on an intention-to-treat basis; the other performed a per protocol-based analysis.
New skin cancers.
The two studies comparing acitretin with placebo or no treatment found a significant reduction in the number of new skin cancers compared with control: 18 with acitretin versus 2 with control (P=0.009) in one trial and 46 SCCs versus 65 SCCs during the drug-free period (P=0.002) in the crossover trial. A statistically significant reduction in the numbers developing BCC was also noted (P=0.045). One of the trials also found a significant reduction in the number of patients developing new skin cancer (2 out of 19 versus 9 out of 19 for the control, P=0.01). The other trial found a reduction in the number of patients with skin cancer during acitretin treatment compared with the drug-free period, but the statistical significance was not reported (6 developed SCCs during acitretin treatment, while 15 developed SCCs during the drug-free period).
The third trial found no significant difference between high- and low-dose acitretin in the number of new malignant lesions. However, no details on the actual numbers of tumours were available.
Two trials assessed the reduction in premalignant lesions with acitretin. One found a reduction of 13.4% with acitretin versus an increase of 28.2% with the control (P=0.008), while the other found reductions with both high- and low-dose acitretin after 2 months (P<0.0001) compared with the situation before acitretin treatment.
The rates of withdrawal due to adverse events were relatively high: 21% in one trial, 39% in a second, and 7% with high-dose and 8% with low-dose acitretin in the third. Adverse events leading to withdrawal included rash, hyperlipidaemia, dysphagia due to stomach cancer, headaches, paronychia, elevated serum liver enzymes, musculoskeletal complaints, gastritis and mucocutaneous effects. Cheilitis was common (70 to 100%) in patients continuing with treatment. Other tolerated adverse effects included alopecia (44 to 47%), headache (40%), myalgia (20 to 35%), rash (30%), photosensitivity (30%), dry eyes (30%), palmoplantar desquamation (20%), epistaxis (20%), nail changes (15%) and pruritus (10%).
None of the trials reported a worsening of renal or liver function.