Implications for future studies

General points that future investigators should bear in mind include using a double-blind randomised study design wherever possible, in order to minimise bias and confounding factors. If investigators wish to use a crossover study design, extreme care must be taken because of the potentially long wash-out periods for most AK therapies. Lastly, investigators should report results of ITT analysis. Without taking into account those subjects who drop out of the study, results can be misleading.

While any given individual AK lesions has a high probability of resolving spontaneously, a patient with extensive involvement most probably has a much lower probability of all his/her AK lesions resolving spontaneously - an issue unique to AK. Thus, studies should either take into account the high correlation of multiple AK lesions if they choose to use number of lesions as their unit of analysis, or results should be stratified by severity of AK if persons cleared is used as the unit of analysis. The later outcome is likely to be of more interest because it is clinically more relevant.

Another precaution unique to AK studies is to ensure that the outcome measure is reliable. Weinstock et a/.112 reported their experience in counting numbers of AK lesions, a commonly used technique. They found the outcome measure to be unreliable, most likely because of

Table 28.1 Summary of therapies for actinic keratoses (AK) and Bowen's disease (BD). Unless indicated, all interventions listed pertain to both AK and BD

• We designated therapies to be "beneficial" If they met the Quality of Evidence level I (see Box 28.1) and had an efficacy of at least 75%.

• Therapies were "likely to be beneficial" if they met the Quality of Evidence levels II-I or Il-ii and had an efficacy of at least 60%.

• Therapies were of "unknown effectiveness" if the Quality of Evidence level was II-iii, III or IV. Therapies were "unlikely to be beneficial" if they met a quality of evidence level of I and had an efficacy rate of less than 30%.

Beneficial

Likely to be beneficial

Unknown effectiveness

Unlikely to be beneficial

Actinic keratoses

• Topical 5-FU

• Topical retinoids

• Chemical peels

• Masoprocol

• Intralesional interferon

• Cryotherapy

• Radiotherapy

• Dermabrasion

• Topical imiquimod

• Diclofenac

• Tubercidin

• Curaderm

• Topical interferon

Bowen's disease

• Intralesional interferon

• Radiotherapy

• Topical imiquimod

• Oral retinoids

• Intralesional bleomycin

• None

ED&C, electrodessication and curettage; 5-FU, 5-fluorouracil; PDT, photodynamic therapy

the spectrum of clinical features. Discussion of discrepancies among investigators enhanced the reliability of the counts, but substantial variation remain. Thus, investigators should test the reliability of their outcome measure before proceeding with the therapeutic part of a study.

0 0

Post a comment