5. Therapy of actinic keratosis – Actinic keratosis and squamous cell carcinoma of the skin

Cryosurgery vs. 5-fluorouracil 0.5% with salicylic acid 10% solution.

Population and study design: One randomized study compared 5-FU 0.5% with SA 10% solution with cryosurgery. Sixty-six patients with 4-10 hyperkeratotic lesions (grade II-III according to Olsen) on the face and hairless scalp were included. Individuals with iatrogenic immunosuppression were excluded [142].

Interventions: 5-FU 0.5% with SA 10% was applied as a solution lesion-directed once daily for a maximum of 6 weeks. If poorly tolerated, reduction to three times weekly administration was allowed. Cryosurgery was performed with liquid nitrogen as a spray procedure with a double freeze-thaw cycle, with the duration of icing left to the individual investigator. It could be repeated after 3 weeks, which was also at the discretion of the investigator. The study was unblinded. There was interindividual randomization to one of the two arms in a 1:1 ratio.

Results: The mean lesion reduction per patient at 14 weeks was 5.2 for 5-FU 0.5% with SA 10% and 5.7 for cryosurgery. In the first arm, 33.3% of patients showed complete and 51.5% partial (>75%) healing of all lesions after 98 days. For cryosurgery, these results were 25.0% (complete) and 62.5% (partial). Histological remission rates were lesion-related 62.1% for 5-FU 0.5% with SA 10% and 41.9% for cryosurgery. The recurrence rate (proportion of patients with at least one recurrence) 6 months after the end of therapy was higher for cryosurgery than for 5-FU 0.5% with SA 10% (84.8% vs. 39.4%). Overall response assessed by the investigator and patients was very similar in both arms. Similarly, the cosmetic outcome of both therapies was predominantly rated as "good" or "very good". Side effects tended to be observed more frequently with 5-FU 0.5% with SA 10% and included erythema, crusting and burning. Serious treatment-mediated events occurred in 24.2% of cases, compared with 6.1% with cryosurgery.

Assessment: In terms of mean lesion reduction and complete or partial healing of all lesions, the efficacy of both interventions was comparable with a slightly significant advantage of topical therapy in terms of complete response. Of note, there was a significant effect in the reduction of recurrences of initially healed lesions of 5-FU 0.5% with SA 10%, which may indicate a better long-term efficacy of this agent compared to cryosurgery. This was in contrast to a lower rate and side effects with cryosurgery, which should be considered in the harm-benefit assessment of this comparison. Because of the open-label study design without blinding and a small number of participants, the work was downgraded to an evidence level of 3. Because of the small number of participants and some end points that were not significantly different, superiority of 5-FU 0.5% with SA 10% over cryosurgery remains unclear.

Cryosurgery vs. CO₂ laser

Population and study design: Zane et al. studied 200 patients with a maximum of 4 lesions on the face and hairless scalp without signs of light-induced skin damage or field cancerization. A total of 543 AK of all clinical grades (I-III according to Olsen) were included.

Interventions: Randomization was interindividual 1:1 to either cryosurgery with liquid nitrogen for 10-20 seconds or single treatment with an ablative CO₂ laser [143].

Results: Lesion-wise, the complete healing rate at 3 months was 78.2% for cryosurgery and 72.4% for CO₂ laser. Regarding this endpoint, the subgroup of thicker AK (grade III) responded better to cryosurgery (80.2%) than to CO₂ laser (59.8%). Patient-wise, a greater proportion of patients treated with cryosurgery showed a complete response (71.6% vs. 65.3% for CO₂ laser), although it remains unclear whether the case number calculation was based on individuals or lesions. This superiority was also evident after a 12-month observation period (53% vs. 14% with CO₂ laser). Global cosmetic outcome was rated similarly by blinded investigators and patients for both interventions on a four-point scale (poor, moderate, good, excellent) and was excellent in 50.7% and 48.4% (investigator) and 58.9% and 50.0% (patient), respectively (cryosurgery vs. CO₂ laser, respectively). Subjectively, patient satisfaction was significantly higher on a three-point scale for cryosurgery. Common local side effects were erythema, edema, hemorrhage, blisters, erosion, crusts (cryosurgery), and erosions and crusts (CO₂ laser).

DESIGN: In this study,CO₂ laser ablation was inferior to cryosurgery with respect to the efficacy endpoints of lesion-related healing rate, complete response and partial response. However, the difference in effects was rather small and not consistently significant. Similarly, the cosmetic outcome of both interventions as assessed by the investigators was not significantly different. Only patient satisfaction was significantly higher for cryosurgery. However, the study design was open, which specifically affects the ability to assess subjective endpoints. Therefore, the study was downgraded to evidence level 3.

 

Cryosurgery vs. 5-fluorouracil 5% cream

Population and study design: Krawtchenko et al. conducted an unblinded, three-arm, randomized trial. Seventy-five subjects with at least 5 histologically confirmed AK in an area no larger than 50 cm² on the head, neck, or décolleté were included. Immunosuppressed individuals were excluded [140].

Intervention: Patients were randomized interindividually 1:1:1 to 5-FU 5% cream twice daily for 4 weeks, imiquimod 5% cream three times weekly for at least 8 h for 4 weeks, or spray liquid nitrogen cryosurgery (20-40 seconds per lesion). The cryosurgery and imiquimod interventions could be repeated once after 2 and 4 weeks, respectively.

Results: Complete clinical healing of all lesions was observed for 5-FU 5% in 96% of patients (23/24), for cryosurgery in 68% (17/25) and for imiquimod 5% in 85% (22/26) (p=0.03). Histopathologically confirmed complete healing rates per patient were 67% (5-FU 5%), 32% (cryosurgery) and 73% (imiquimod 5%; p=0.02), respectively. At 12 months after the end of the study, 54% of all patients treated with 5-FU 5% and 28% of all patients treated with cryosurgery still showed a sustained response of the lesions. The overall treatment field was still unremarkable after this time in 33% (5-FU 5%) and 4% (cryosurgery) of all study participants. Cosmetic outcome at 12 months was similarly poor for both interventions and was considered excellent in only 4% of cases. No serious adverse events were observed. Normal skin appearance after therapy was observed more frequently in the 5-FU 0.5% arm than under cryosurgery (58% vs. 16%). Other adverse events were similarly distributed, although more precise numbers were not reported.

Assessment: In this study, efficacy benefits of 5-FU 5% cream were demonstrated with respect to the endpoints of complete clinical and histological healing of all lesions, with the difference in treatment effect appearing clinically relevant (96% vs. 68% and 67% vs. 32%, respectively). Because cryosurgery is a primarily lesion-directed therapy and 5-FU 5% is a primarily field-directed therapy, the two modalities have limited comparability in terms of field efficacy. Nevertheless, the results highlight that cryosurgery is likely inferior to field-directed procedures in the presence of multiple lesions or field cancerization. This is also evident in the assessment of long-term outcomes at 12 months. However, bias is possible due to the open study design. Therefore, the study was downgraded to evidence level 3.

Cryosurgery vs. imiquimod 5% cream

Two studies evaluated cryosurgery versus imiquimod 5% cream [139][140]. A randomized, prospective, controlled trial from Australia compared cryosurgery treatment in 36 patients versus imiquimod therapy in 35 patients for 10 or more AK on the face and scalp [139]. Response was assessed 12 months after initiation of therapy in terms of lesion-related complete cure rate and patient-related complete or partial response. The complete patient-related response for repeated cryosurgery for up to 4 cycles at 3-month intervals was 85%, significantly higher than 66.9% for 2 cycles of imiquimod 5% cream for self-treatment three times per week for 3 to 4 weeks. The adverse event rate was significantly higher with blistering, redness, scaling, crusting and hypopigmentation in 54.8% of patients for cryosurgery than in 24% after Imiquimod. Therapy discontinuation occurred in 13.9% in the cryosurgery group and 28.6% in the imiquimod group, although this is inconsistent with the observed side effects. Evidence of treatment adherence between the two groups was not mentioned. The authors stated that both therapeutic modalities showed high response rates, with imiquimod treatment showing a more favorable side effect profile and better aesthetic results [139].

Another study compared cryosurgery with imiquimod 5% cream and 5-FU 5% cream [140]. In this study, cryosurgery showed lower response rates than either topical therapy and was inferior in terms of complete healing of the entire treatment field at 12 months. The study reported complete clinical healing rates of 68% for cryosurgery, 96% for 5-FU and 85% for imiquimod. In contrast, histological examination showed complete healing rates of only 32%, 67% and 73%, respectively. Long-term complete healing at 12 months of total treatment field was reported at 4%, 33% and 73%, respectively, which highlights the long-term efficacy of field therapy (here, 5-FU and imiquimod) versus lesion-directed therapy (here, cryosurgery) [140].

Cryosurgery vs. MAL red light PDT.

Two studies compared cryosurgery with MAL red light PDT [144][141]. Morton et al. performed an open-label, intraindividual study in 119 patients with a total of 1501 non-hyperkeratotic lesions on the face or hairless scalp (at least 3 per patient) [141]. Kaufmann et al. compared both interventions intraindividually in 121 patients with a total of 1343 non-hyperkeratotic lesions (at least 4 per patient). However, only extrafacial lesions on extremities, neck or trunk were treated [144]. In both studies, one treatment field was treated with MAL cream (160 mg/g) and then incubated for 3 h with an occlusive dressing. Illumination was with a narrow spectrum red light source at a wavelength of 630 nm and a dose of 37 J/cm². The other treatment field was treated with cryosurgery using liquid nitrogen at two freeze-thaw cycles. If the therapeutic success was insufficient, the treatment was repeated after 12 weeks.

After 12 weeks, the mean lesion reduction was slightly better for MAL-PDT than for cryosurgery (86.9% vs. 76.2%; p<0.001). However, both interventions were relatively similar for this endpoint at 24 weeks (89.1% vs. 86.1%; p=0.20). Also lesion-related, a high proportion of AK were completely healed after 24 weeks in both treatment arms, namely 85.8% for MAL-PDT and 82.5% for cryosurgery. However, in terms of global cosmetic outcome, MAL-PDT showed advantages on a four-point scale (poor, moderate, good, excellent), with 77.2% of the investigators classifying the result as excellent after 24 weeks. In comparison, this was the case in only 49.7% of patients treated with cryosurgery. Therefore, MAL-PDT was also preferred as therapy. Regarding tolerability, there were rather minor differences, although slight advantages of MAL-PDT were observed here as well. Nevertheless, both therapies were well tolerated and side effects were mild to moderate [141].

Kaufmann et al. reported similarly high mean lesion reduction rates after 24 weeks, which were 78% for MAL-PDT and 88% for cryosurgery. The latter therapy was significantly superior even with respect to this efficacy endpoint (p=0.002). However, the cosmetic outcome was rated significantly worse by both the practitioner and the patient. Similarly, 59% of patients would prefer MAL-PDT therapy in the future, but only 25% cryosurgery. Both interventions were well tolerated without serious therapy-mediated side effects [144].

In summary, both intraindividual studies showed similar high efficacy values for cryosurgery of more than 80% lesion reduction. In contrast, the cosmetic results were worse than those of MAL-PDT, which were rated significantly better in both studies. Nevertheless, the open study design results in a risk of bias in the global assessment of the cosmetic outcome. In addition, measures of dispersion such as standard deviations were not consistently reported in the study by Morton et al. 2006, which is why this study was downgraded to evidence level 3. Other studies were not included in the comparison of these two interventions because of ambiguities regarding effectiveness endpoints [146][147].

A total of six independent studies were identified that compared the use of diclofenac sodium 3% gel with vehicle for efficacy. Of these six studies, one investigated efficacy in organ transplant recipients undergoing immunosuppression [172], which is reported separately.

Population and study design: The oldest study by McEwan in 1997 compared the active ingredient in the carrier gel hyaluronic acid with vehicle in a 1:1 randomized, double-blind, monocentric study. Of the 130 participants included, 73 were male and 57 were female; the age range was 48 to 87 years. AKs of varying severity were included [169]. Scaling of AK reported in 62 of the 65 AKs in the verum group suggests that more severe AK manifestations were also treated. 28 of the 65 documented AKs were on the hands. The study by Wolf (2011) was based on a multicenter, randomized, double-blind, placebo-controlled, parallel-group design in which 58 patients were included in the active group and 59 patients in the control group [173]. Another study by Rivers in 2002 included 195 patients (73% male) with at least 5 AK in a randomized, double-blind, placebo-controlled multicenter trial with parallel group design. Participants were randomized to a total of 4 different groups in a 1:1:1:1 scheme, which differed in treatment duration [170]. The multicenter study by Gebauer 2003 had a randomized, double-blind, placebo-controlled parallel group design and included 150 patients (89 men, 61 women) with a mean age of 68 years (range 27-87 years). 73 of the participants were randomized to the active group and 77 to the placebo group [167]. In another study, the effect of diclofenac sodium 3% gel was compared with the effect of imiquimod 5% cream twice a week and the application of pure base gels (vehicle) in 61 patients [166]. This also resulted in comparative data from diclofenac sodium 3% gel to drug-free foundation, but these are presented under the comparison to imiquimod.

Interventions: In McEwan's study, diclofenac sodium 3% gel or vehicle was applied only twice daily for 8-24 weeks to areas of varying size and thereby in varying amounts. In addition, sunscreen was applied during the day. In the study by Wolf et al., patients applied 0.5 g of diclofenac sodium 3% gel twice daily per 5 cm² of area for 90 days. In the third study by Rivers, 4 parallel groups with different intervention schedules were present: Group 1 received 0.5 g of 3% diclofenac sodium gel twice daily for 30 days (n=49), and Group 2 received the same therapy but for 60 days (n=48). Group 3 received placebo for a duration of 30 d, and group 4 again received placebo for 60 d (n=49 each). In Gebauer's study, patients applied 0.25 g of gel as active substance or vehicle to a defined area of 5 cm² twice daily for 12 weeks. For more detailed intervention details of the study by Akarsu (2011).

Results: McEwan et al. reported complete healing in 29% and 17% of participants, respectively (diclofenac vs. placebo; p=0.14) and partial healing of at least 75% in 38% and 45% of participants, respectively (p=0.18). Side effects were higher (29%) in the treatment group compared to 5% with baseline only. These were mostly skin irritations. The difference was statistically significant [169].

In the study by Wolf et al., a complete patient-related response (operationalized by a TLNS of 0) in 50% of the intervention group versus 20% in the control group demonstrated the superiority of diclofenac sodium 3% gel (p<0.001). The investigators' assessment of global efficacy using IGII as "completely improved“ was higher at 47% vs. 19% than the patients' assessment of the same endpoint using PGII at 41% vs. 17% (diclofenac vs. placebo, respectively). In terms of side effects, local skin symptoms occurred in 90% and 81% of cases, respectively. The most common were pruritus (55% vs. 49%), irritation (34% vs. 20%) and skin dryness (36% vs. 17%, respectively, diclofenac vs. placebo) [173].

Of the 4 groups in the third study by Rivers 2003, the 60-day verum group showed clear superiority over the 60-day placebo group in terms of efficacy. A complete patient-related response (operationalized by a TLNS of 0) was achieved by 33% taking diclofenac sodium 3% gel and by 10% taking placebo, respectively (p<0.005). The global effect assessment by IGII by the investigators as "completely improved“ was 31% vs. 10% (p<0.05), similar to the patients' assessment of 29% vs. 10% (PGII; p<0.05). Compliance of the groups was comparable. A total of 10 serious adverse events occurred in 7 participants in the active group, including pruritus, local irritation, paresthesia, rash, edema and contact dermatitis. Of these, 6 side effects were possibly attributable to treatment [170].

In the Gebauer 2003 study, complete healing of lesions per patient occurred in 38% of the intervention group and in 10% of the placebo group (p=0.002). A more than 50% reduction of lesions was present after 16 weeks in more than half of the participants in the diclofenac group (65% vs. 29% under placebo). The difference was statistically highly significant. At 16 weeks, the mean number of lesions also decreased more in the verum than placebo group, by 6.2±7.5 (56.1% reduction) vs. 2.4±4.3 (23.6% reduction). Side effects were predominantly mild and mostly included pruritus, erythema, edema and scaling. 19% of cases with pruritus, 18% with dry skin and 12% of those with skin irritation were reported as severe in this regard [167].

For the results of the study by Akarsu (2011), see here.

Assessment: The collected assessment of the studies is complicated by the different observation times and different treatment protocols. In the study by McEwan et al., no significant superiority to vehicle was observed with respect to complete or partial response. In this collective, a proportion of lesions were on the back of the hands/hands. However, local side effects were significantly more frequent [169]. These results are in contrast to the other studies, in which significantly improved efficacy endpoints were visible versus vehicle in each case. The strength of the effect was moderate. This is also supported by the meta-analysis by Vegter and Tolley, in which lower values were observed for diclofenac sodium 3% gel compared to other interventions with regard to the effectiveness endpoint complete patient-related response [131]. This is in contrast to good cosmetic results and good tolerability. A long application time of 60-90 days implies good therapy adherence. The Wolf 2001 study was downgraded to evidence level 3 due to unclear randomization and asymmetric distribution of study dropouts [173].

Diclofenac sodium 3% gel vs placebo (immunosuppressed)

Population and study design: In 32 organ transplant patients with immunosuppression (31 men, 3 women aged 49-77), diclofenac sodium 3% gel was compared with placebo in a randomized, double-blind, placebo-controlled multicenter trial [172]. 24 of the patients were randomized to the active group and 8 to the control group. Of the 32 patients, 6 had liver transplantation, 18 had kidney transplantation, and 8 had heart transplantation. The evaluation was done after 20 weeks and after 24 months.

Intervention: Application of diclofenac sodium 3% gel or placebo twice daily for 16 weeks.

Results: Complete healing was seen in 41% of patients in the verum group vs. 0% of the placebo group, regardless of which organ had been transplanted. Partial healing of at least 75% was seen in 59% in the active group vs. 16.7% in the control group. Differences according to the organ originally transplanted were present, but the number of cases appear too small for these differences to be meaningful enough and could not have been caused by other, intraindividual factors. One exception, however, is the 18 heart transplanted patients who showed 100% partial improvement on verum and 0% on placebo. The mean reduction in the number of lesions was 53% vs. 17% (verum vs. placebo). The recurrence rate was 55% after a mean of 9.3 months, and side effects reported were mild to moderate erythema, scaling, pruritus, skin irritation and edema.

Rating: Due to the loss of patients and ambiguities in randomization, there is a risk of bias, which is why the level of evidence was downgraded to 3. However, from the reported end-points of effectivity, there is a large effect of the intervention compared with placebo; the difference in complete response was most pronounced in the heart transplant subgroup. For evidence-based recommendations on the use of diclofenac sodium 3% gel in organ transplant recipients, see see Therapy for Immunosuppression and Organ Transplantation.

Diclofenac sodium 3% gel vs. imiquimod 5% cream

For more information, see here.

Diclofenac sodium 3% gel vs. 5-fluorouracil 5% cream

A randomized parallel group trial with 28 participants examined this intervention [171]. For more details comparing these interventions, see see 5-fluorouracil 5%.

Diclofenac sodium 3% gel vs 5-fluorouracil 0.5% with salicylic acid 10% solution.

Population and study design: Stockfleth et al. investigated both interventions in a randomized phase IV study in direct comparison. A total of 502 patients from 33 centers in Germany, Spain and the United Kingdom were randomized. The median age of the participants was 75 years (range 34-96 years), 85.2% were male. The median lesion number per patient was 6 (range 4-9) [178]. The treatment area included a contiguous area of 25 cm2 on capillitium and face.

Interventions: The study population was randomized in a 1:1 ratio to one of the two therapies. IMB 0.015% gel was applied once daily for three consecutive days. After 8 weeks, treatment was allowed to be repeated if response was inadequate or new lesions appeared. Diclofenac sodium 3% gel was applied twice daily for 90 days.

Results: After the first treatment cycle (8 weeks for IMB, 17 weeks for diclofenac sodium 3% gel), the patient-specific complete healing rate was 34.5% (IMB) vs. 23.5% (diclofenac). The difference was statistically significant (p=0.006). After two cycles (17 weeks), the values were 45.1% vs. 23.5% (p<0.001). Patient-related partial response after the first cycle of treatment was 57.6% (IMB) vs. 43.3% (diclofenac; p=0.001). The percent lesion reduction after the first cycle was 69.5% vs. 57.7% (p<0.001) and after the second cycle was 77.2% vs. 57.7% (p<0.001). After two cycles, scores for this endpoint were 69.0% vs. 43.3% (p<0.001). Patient satisfaction was significantly higher for IMB at 17 weeks (p<0.001). The rate of adverse events was slightly higher for IMB than for diclofenac (49% vs. 41%). In contrast, the rate of treatment discontinuation due to side effects was slightly higher for diclofenac (6%) than for IMB (2%). The following adverse events were reported in this study (IMB vs. diclofenac): erythema 19% vs. 11.5%, crusting/scaling 8.1% vs. 2.6%, pain 6.5% vs. 3.4%.

Assessment: The study was methodologically well executed and without major risks for bias. The only point of criticism remains the lack of blinding. Regarding the efficacy endpoints, there was an advantage for IMB. However, this therapy was allowed twice, whereas the treatment with diclofenac was allowed only once. Considering the effectiveness scores, it appears that much of the effectiveness superiority of IMB in this study was actually caused by the second treatment cycle. Side effects were observed more frequently for IMB. In contrast, the rate of treatment discontinuation due to side effects was slightly higher for diclofenac, which may be due to the longer treatment duration of 90 days in addition to local side effects. Therefore, this study shows that good treatment adherence is necessary for diclofenac. Despite the efficacy advantage of IMB in this study, this topical is not recommended because of the current dormant approval in Europe and the possible increased incidence of NMSC and melanoma in the treatment area (see see).

Population and study design: Two independent studies were found comparing 5-FU 0.5% with SA 10% solution versus placebo. The first study examined 470 patients with 4-10 lesions on the face or hairless scalp. The primary endpoint was histological healing of a predefined AK lesion 8 weeks after the end of therapy. AK of grades I and II according to the Olsen classification were included. The study design was inter-individual and double-blind (grade I-II according to Olsen) [175]. The second study included 166 patients with 4-10 lesions on the face or hairless scalp in an area of 25 cm² with field cancerization. The primary endpoint of this study was complete clinical healing 8 weeks after the end of therapy. Grade I and II AK were included. The study was double-blind [196].

Interventions: In the first study, patients were randomized 2:1:2 to 5-FU 0.5% with SA 10%, vehicle (placebo) or diclofenac sodium 3% in hyaluronic acid 2.5% gel, which were indistinguishable externally. Therapy was applied once daily (5-FU 0.5% with SA 10% or placebo) or twice daily (diclofenac sodium 3% in hyaluronic acid 2.5% gel) for 12 weeks or until lesions healed. Treatment was lesion-directed; that is, the product was applied directly to clinically manifest AK with a brush applicator or the finger [175]. In the second study, randomization was 2:1 to 5-FU 0.5% with SA 10% or vehicle. Application was once daily over a contiguous area of no more than 25 cm², in which 4-10 AK or signs of field cancerization were present, and thus field directed, in contrast to the previous study [196]. In both papers, reduction of active therapy to three times weekly (5-FU 0.5% with SA 10%) or once daily (diclofenac sodium 3% in hyaluronic acid 2.5% gel) was possible.

Results: The histologically proven reduction rate at the end of the study was 72.0% (5-FU 0.5% with SA 10%) and 44.8% (vehicle), respectively, and the mean lesion reduction was equally significantly higher in the group of patients treated with 5-FU 0.5% with SA 10%. The number of patients with complete healing of all lesions at 20 weeks was significantly better in the active arm (55.4% vs. 15.1%). 74.5% of all lesions treated with 5-FU 0.5% with SA 10% had healed, compared with only 35.5% in the control group (vehicle). Global assessment of treatment success after 20 weeks was also highest in the 5-FU 0.5% with SA 10% group. 92.0% of investigators and 93.2% of patients gave an assessment as "very good" or "good". In contrast, local side effects such as burning and inflammatory reactions were also observed more frequently in the active arm. These were mostly mild to moderate. Itching occurred in both groups. Serious adverse events were recorded in 1.1% (5-FU 0.5% with SA 10%) and 4.1% (vehicle), respectively, but were not assessed as treatment-mediated [175]. A follow-up study reported later observation time points at 6 and 12 months within the same study population [180]. Based on healed lesions at week 20, lesion-related healing rates of 85.8% (5-FU 0.5% with SA 10%) versus 79.8% (placebo) were reported at 12 months (p=0.04419). Similarly, global assessment of treatment success at 6 and 12 months was significantly higher in active therapy, with 94.7% of treated patients recommending therapy with 5-FU 0.5% with SA 10%. However, local reactions such as burning (81.3% vs. 57.4%) or inflammatory reactions (70.3% vs. 22.3%) were again more frequently reported [180].

In the primary field-directed therapy study, the number of patients with complete healing of all lesions 8 weeks after the end of treatment was 49.5% (5-FU 0.5% with SA 10%) versus 18.2% (placebo) in the intention-to-treat analysis. Similarly, a reduction of at least 75% of clinically visible lesions was achieved in 69.5% vs. 34.6%. Both endpoints were significantly different. The percent lesion reduction relative to the start of therapy was 78.0% for active therapy vs. 46.9% with placebo. Global treatment response was also rated as "very good" or "good" by the investigator in 90.2% of patients treated with 5-FU 0.5% with SA 10%. Side effects of active therapy reported were erythema (88.9%), pain (69.4%) and skin irritation (59.3%), all of which occurred more frequently than with placebo [196].

Assessment: One strength of these studies is a long follow-up period of results over 12 months. In the follow-up study, lesion-related healing rates based on healed lesions at week 20, were determined at 12 months. However, the strength of the effect of treatment with 5-FU 0.5% with SA 10% in solution appears to progressively decrease over the observation period. This is particularly evident at 12 months, where the lesion-related healing rates of active therapy and placebo and vehicle were close (85.8% vs. 79.8%, respectively). The high healing rate in the control arm could be due to either some efficacy of vehicle or a high natural remission rate (see (see Indication and Natural History). Therefore, further studies are needed to further investigate the long-term efficacy of 5-FU 0.5% with SA 10% in solution.

5-Fluorouracil 0.5% with salicylic acid 10% solution vs. cryosurgery

Population and study design: There is one randomized trial comparing 5-FU 0.5% with SA 10% with cryosurgery. Sixty-six patients with 4-10 hyperkeratotic lesions (grade II-III according to Olsen classification) on the face and hairless scalp were included. Individuals with iatrogenic immunosuppression were excluded [142].

Interventions: 5-FU 0.5% with SA 10% was applied as a solution lesion-directed once daily for a maximum of 6 weeks. If poorly tolerated, reduction to three times weekly administration was allowed. Cryosurgery was performed with liquid nitrogen as a spray procedure with a double freeze-thaw cycle, with the duration of icing left to the individual investigator. It could be repeated after 3 weeks, which was also at the discretion of the investigator. The study was not blinded. There was interindividual randomization to one of the two arms in a 1:1 ratio.

Results: The median lesion reduction per patient at 14 weeks was 5.2 for 5-FU 0.5% with SA 10% and 5.7 for cryosurgery. In the first arm, 33.3% showed complete and 51.5% partial (>75%) healing of all lesions after 98 days. For cryosurgery, it was 25.0% (complete) and 62.5% (partial). Histological remission rates were lesion-related 62.1% for 5-FU 0.5% and 41.9% for cryosurgery. The recurrence rate (proportion of patients with at least one recurrence) 6 months after the end of therapy was higher for cryosurgery than for 5-FU 0.5% with SA 10% (84.8% vs 39.4%). The overall response assessed by the investigator and patients was very similar in both arms. Similarly, the cosmetic outcome of both therapies was predominantly rated as "good" or "very good". Side effects tended to be more common with 5-FU 0.5% with SA 10% and included erythema, crusting and burning. Serious treatment-mediated events occurred in 24.2% of cases compared with only 6.1% with cryosurgery [142].

Assessment: In terms of mean lesion reduction and complete or partial healing of all lesions, the effectiveness of both interventions was comparable. Of note, there was a significant effect in the reduction of recurrences of originally healed lesions of 5-FU 0.5% with SA 10%, which may indicate a better long-term efficacy of this agent compared to cryosurgery. This was in contrast to a lower rate and side effects with cryosurgery, which should be considered in the harm-benefit assessment of this comparison. Because of the open-label study design, the work was downgraded to an evidence level of 3.

5-Fluorouracil 0.5% with salicylic acid 10% solution vs diclofenac sodium 3% gel

Population and study design: One study examined these interventions in a three-arm, interindividual design [175].

Interventions: Randomization was 2:1:2 to 5-FU 0.5% with SA 10%, vehicle (placebo), or diclofenac sodium 3% in hyaluronic acid 2.5% gel.

Results: The histologically proven reduction rate at study end was 72.0% (5-FU 0.5% with SA 10%) vs. 59.1% (diclofenac sodium 3% in hyaluronic acid 2.5% gel), and the clinically measured reduction rate was 74.5% (5-FU 0.5% with SA 10%) vs. 54.6% (diclofenac sodium 3% in hyaluronic acid 2.5% gel). The number of study patients with complete healing was 55.4% (5-FU 0.5% with SA 10%) vs. 32.0% (diclofenac sodium 3% in hyaluronic acid 2.5% gel). Lesion-wise, 74.5% and 54.6% of all AK had healed at 20 weeks, respectively. Global treatment response was consistently rated better by investigators for 5-FU 0.5% with SA 10%. Thus, in 92.0%, outcomes were found to be "very good" or "good"; in the diclofenac arm, this was the case in 73.8%. Similarly distributed was the global response assessed by patients (93.2% vs. 81.6%). In contrast, the rate of adverse events was higher in the experimental arm (95.2%) than in the active diclofenac arm (76.8%). Serious adverse events were observed in 1.1% (5-FU 0.5% with SA 10%) and 4.9% (diclofenac sodium 3% in hyaluronic acid 2.5% gel), respectively, but were not judged to be treatment-mediated. Thirty-five patients terminated the study prematurely, of whom 14 were treated with 5-FU 0.5% with SA 10% and 16 with diclofenac sodium 3% in hyaluronic acid 2.5% gel [175]. At 12 months, 85.8% and 81.0% of lesions originally healed at week 20 were still healed, respectively. Significant clinical improvement ("good/very good“) was present in 93.2% and 81.6%, respectively, according to the investigators here (5-FU 0.5% with SA 10% vs. diclofenac sodium 3% in hyaluronic acid 2.5% gel). Patient assessment here was slightly lower for both groups [180].

Assessment: 5-FU 0.5% with SA 10% was significantly superior to diclofenac sodium 3% in hyaluronic acid 2.5% gel after 20 weeks with respect to the investigated efficacy endpoints.

Population and study design: Several independent studies were found comparing MAL-PDT versus placebo [237][236][235]. They studied 100 [237] or 80 [236] patients with 4-10 lesions on face or hairless scalp. Non-pigmented, non-hyperkeratotic grade I and II AK according to the Olsen classification were included. The study design was inter-individual and double-blind in both cases. Another study included 115 patients with identical parameters [235].

Interventions: Patients were randomized 1:1 to MAL-PDT or vehicle (placebo). After 3 hours of incubation, LED light was used to illuminate and treatment was repeated after 1 week [237][236][235].

Results: After 3 months, the complete lesion response rate was recorded. This was 86.2% (MAL) vs. 52.2% (placebo) [237] and 83.3% vs. 28.7% [235]. The high proportion of placebo healing rate in the Pariser et al. study was explained by lesion pretreatment. On a patient basis, 59.2% vs. 14.9% healed in the Pariser et al. study and 68.4% vs. 6.9% in the Szeimies et al. study. Main side effects in both studies were pain, erythema, and burning in the active groups.

MAL-PDT vs. placebo (immunosuppressed)

Population and study design: A prospective, monocentric, randomized, double-blind, placebo-controlled intraindividual comparative study was conducted in 17 organ transplant patients under chronic immunosuppression (129 mild to moderate AK) [252].

Interventions: Two lesion-bearing areas received two PDTs one week apart with white light (75 J/cm²). One area was treated with MAL, the other with placebo cream.

Results: The complete lesion-related response rate was evaluated 16 weeks after the second treatment; it was 90.3% for the MAL site and 0% for the placebo site. Side effects were not precisely quantified. However, pain, erythema, edema and crusts were more pronounced in the verum region [252]. For evidence-based recommendations on MAL-PDT in immunosuppressed individuals, see Therapy in Immunosuppression or Organ Transplantation.

MAL-PDT vs. cryosurgery

For more detailed information comparing these interventions, see cryosurgery.

MAL-PDT vs. ALA-PDT

Population and study design: One study examined these interventions for mild to moderate AK of the face and capillitium in a three-arm, interindividual design [233]. This was a multicenter, randomized, investigator-blinded, placebo-controlled trial of BF-200 ALA-containing gel (7.8%) vs MAL (16%) vs placebo. Illuminations could be either red light from LED lights or broad spectrum lamps [233]. Another monocentric study compared the efficacy of an ALA-PDT with that of a MAL-PDT in a field-directed approach, double-blind in a half-side trial ("split-scalp“ design) [230].

Interventions: Randomization was 3:3:1 to MAL-PDT, ALA-PDT, or ALA vehicle (placebo) [233]. 1:1 comparison intraindividually in the study by Moloney et al. [230].

Results: In the study by Dirschka et al., clinical complete healing of AK on a per patient basis at 3 months was 78.2% for BF-200 ALA vs. 17.1% placebo (p<0.0001) and 64.2% for MAL (p<0.05). Healing rates on the face or forehead were better than on the capillitium for both sensitizers. Statistically relevant differences in cosmetic outcome, incidence of side effects and mean pain score (VAS) were not evident between the two sensitizers [233]. At 12 months, the overall healing rate for BF-200 ALA was 47% (although further inclusion of a placebo-controlled comparator in this evaluation), and that for MAL was 36% [232]. In the subgroup analysis at 12 months, the healing rate using the LED lighting systems was significantly better than for the broad spectrum lamps 60.3% vs. 56.3% for BF-200 ALA and 59.3 vs. 50.7 for MAL [233]. Moloney et al. achieved complete healing of treated AK fields with ALA in 40% of cases (MAL 46.7%). The mean reduction in the number of AK from baseline was 6.2 for ALA vs. 5.6 for MAL (no statistically significant difference) [230].

MAL-PDT vs. diclofenac sodium 3% gel

Population and study design: Zane et al. conducted an open-label, monocentric, randomized trial comparing these interventions in 200 patients with a total of 1,674 lesions. There were 58 women in this cohort, and the age range was 42-93 years. Patients had at least 5 AK of all grades on the face or hairless scalp and signs of actinic damage [174].

Intervention: Interindividual randomization was 1:1 to either MAL-PDT in a standard conventional manner (MAL cream 160 mg/g, 3 hours incubation, illumination with red light at a dose of 37 J/cm²) or treatment with diclofenac sodium 3% in hyaluronic acid 2.5% gel twice daily for 90 days. In case of insufficient success, MAL-PDT could be repeated once after 3 months.

Results: After 3 months, 68% of patients treated with MAL-PDT but only 27% of those treated with diclofenac sodium 3% in hyaluronic acid 2.5% gel showed complete response of all AK. Lesion-related healing rates were 85.9% (MAL-PDT) and 51.8% (diclofenac sodium 3% in hyaluronic acid 2.5% gel). The cosmetic outcome was also judged to be significantly better for MAL-PDT compared to diclofenac sodium 3% in hyaluronic acid 2.5% gel by both the respective investigator and the patients themselves. However, the study was not blinded. Subjective patient satisfaction was also better with treatment with MAL-PDT. Here, 59% of all patients rated the outcome of the intervention as excellent in contrast to only 6% under diclofenac sodium 3% in hyaluronic acid 2.5% gel. Interestingly, this study also compared the cost-effectiveness of the two therapies. Again, advantages were seen for MAL-PDT, which achieved a complete response at a lower cost than diclofenac sodium 3% in hyaluronic acid 2.5% gel [174].

Assessment: Because of the open study design, there may be bias from the patient satisfaction and cosmetic outcome endpoints. There were also differences in lesion pretreatment between the two study arms (see evidence tables). In addition, there is a high potential for bias due to the high proportion of hyperkeratotic AK (moderate/thick) and sequential PDT therapy preceded by curettage for scaling/crusting. Therefore, the study was downgraded to evidence level 3.

MAL-PDT vs. ingenol mebutate 0.015% gel

One study was identified [203]. For comparison of these interventions, see see chapter Ingenolmebutate.

MAL-PDT vs. colchicine 0.5% cream

Miola et al. studied MAL-red light PDT versus colchicine 0.5% cream in 36 patients with AK on the forearms [254]. The two interventions were randomized 1:1 in intraindividual comparison. The patient-related complete healing rate was 17% (colchicine) versus 19% (MAL-PDT), and the partial healing rate was 44% versus 67%. The lesion-related healing rate was 45% for colchicine and 40% for MAL-PDT. Side effects were at a similar incidence in both arms. The low healing rates in this study demonstrate that AK is more difficult to treat on the forearms than on the face or capillitium. Both interventions were similar in terms of efficacy endpoints, but colchicine is not approved as a topical treatment for AK.

MAL-PDT vs. 5-fluorouracil 5% cream vs. imiquimod 5% cream vs. ingenol mebutate 0.015% gel

For a comparison of this intervention, see 5-Fluorouracil 5% [184].

Szeimies et al. evaluated BF-200 ALA-PDT versus vehicle-PDT in a multicenter, randomized, double-blind, placebo-controlled, two-arm study [231]. 122 patients (105 men, 17 women) with a mean age of 71 years (age range 57-85 years) with 4-8 mild to moderate lesions were randomized 2:1 to therapy with BF-200 ALA-PDT (81 participants) or vehicle-PDT (41 participants). Three hours after gel application, lesions were irradiated. To represent clinical practice within Europe, two different light sources were used: first, a narrow-spectrum source with an emission spectrum between 590 and 670 nm and a recommended light dose of 37 J/cm², and second, an incoherent broad-spectrum source with emissions between 580 and 1400 nm at a light dose of 170 J/cm². The light sources varied depending on the participating study center and were not randomized. If necessary, a PDT cycle was allowed to be repeated 12 weeks after the first therapy. Already after the first treatment BF-200 ALA-PDT was significantly superior to vehicle PDT (49% vs. 11%). Overall, the complete cure rate after one or two cycles of therapy was 64% vs. 11% on a patient-by-patient basis. The difference was highly statistically significant. Lesion-wise, the proportion of completely healed AK was 81% in the ALA-PDT group and 22% in the vehicle-PDT group after both cycles of therapy. Interestingly, patient- and lesion-related healing rates were consistently higher for narrow-spectrum source illumination than for broad-spectrum source illumination. In 49% of participants in the intervention group and in 27% of the control group, the overall cosmetic result was rated as very good or good by the investigator. However, it was classified as unsatisfactory in 22% in the control group, compared with only 4% in the ALA-PDT group. In addition, skin quality improved in the BF-200 ALA-PDT group in terms of roughness, dryness, scaling, and hyperpigmentation. No side effects occurred solely as a result of the application of the gel before illumination. Local side effects such as pain, itching or burning during and after illumination were more frequent in patients irradiated with narrow spectrum [231].

Dirschka et al. conducted a multicenter, randomized, placebo-controlled, investigator-blinded phase III study of 570 patients with 4-8 mild-to-moderate AK (Olsen grade I or II) on the face and hairless scalp [233]. 84% of participants were male, and the median age was 71.0 years (range 39-87 years). Participants were randomized in a 3:3:1 ratio to BF-200 ALA-PDT (n=248), MAL-PDT (n=247) and placebo-PDT (n=76) arms. Illumination was provided by several different light sources. A follow-up study summarized the results from two randomized trials with a total of 663 patients (104 women, age range: 39-87 years), one of which included the population originally described by Dirschka et al. [232]. The proportion of patients with completely healed lesions was significantly superior to the vehicle group after three months in the BF-200 ALA-PDT group (78.2% vs. 17.1%; p<0.0001). Treatment of lesions on the face and forehead was more effective than on the hairless scalp. The percentage of completely healed lesions at three months was also higher after treatment with BF-200 ALA-PDT than under vehicle, with a rate of 90.4%, although relatively good results were also observed (37.1%). Similarly, the cosmetic outcome was better after treatment with BF-200 ALA-PDT. In 43.1% of cases, BF-200 ALA-PDT was rated as "very good/good“, compared with 36.4% of cases with vehicle-PDT. Side effects occurred more frequently in the BF-200 ALA group (96.4%) than in the placebo group (72.4%) in the study. Redness (78.2%), burning (85.9%), and pain (69.4%) were reported most frequently and were also the most severe. After a follow-up of 12 months, 47% of patients remained with complete healing and no recurrence. However, the cosmetic outcome was rated similarly in both groups at the time points. At 6 months, it was 39.7% ("very good“) and 43.1% ("good“) for BF-200 ALA-PDT and 34.8% ("very good“) and 44.1% ("good“) for vehicle, respectively. After 12 months, values of 38.9% ("very good“) and 45.0% ("good“) were still observed for BF-200 ALA-PDT and for 32.8% ("very good“) and 46.9% ("good“) for vehicle, respectively.

Reinhold et al. investigated the efficacy, safety as well as the cosmetic outcome of BF-200 ALA-PDT in a multicenter, randomized, double-blind, placebo-controlled phase III study using a red light lamp of a wavelength of 635 nm and a light dose of 37 J/cm² [234] . The photosensitizer was applied field-directed to an area of approximately 20 cm². A total of 94 patients with 4-8 mild to moderate lesions on the face and/or scalp participated. Of these, 87 patients (79 men, 8 women) with a mean age of 71.6 years (standard deviation 6.4 years) were randomized: 55 to the BF-200 ALA-PDT group and 32 to the vehicle-PDT group. If lesions persisted three months after initial therapy, PDT was repeated once. Overall, treatment with ALA-PDT was significantly superior to treatment with vehicle. After two cycles of PDT, the complete response rate was 91% in the active compared to 22% in the control group (p<0.0001). The proportion of patients with at least 75% healed lesions was also significantly significantly higher in the active group (94% vs. 25%; p<0.0001). In terms of lesions, 94.3% of all treated AK healed completely in the BF-200 ALA-PDT group, but this number was only 32.9% in the control arm (p<0.0001). A higher response of lesions treated with active therapy was also confirmed histopathologically (78% vs. 22%; p<0.0001). Side effects occurred in all patients in the BF-200 ALA-PDT group (100%) and in only 69% in the vehicle-PDT group. The most commonly reported side effects included pain at the application site (96.4% and 50.0%, respectively), skin redness, and itching. Side effects were described as mild to moderate in intensity. On average, pain was reported as significantly more severe in the active than in the vehicle group, during both the first and second treatments (mean BF-200 ALA VAS score: 1st treatment 5.5, 2nd treatment 5.8; vehicle group: 1st treatment 0.9, 2nd treatment 0.3). This difference was not statistically significant. The overall cosmetic result of BF-200 ALA-PDT was rated as "very good“ in 59% of cases and "good“ in 31%. This was also reflected in the subjective satisfaction of the participants, as 91% in the BF-200 ALA-PDT group were "very satisfied“ or "satisfied“ (45% for vehicle PDT) [234].

ALA-PDT vs MAL-PDT

Study design and population: Three studies (including one follow-up study) compared the efficacy and safety of 5-ALA-PDT with MAL-PDT. Dirschka et al. conducted a multicenter, randomized, placebo-controlled, investigator-blinded phase 3 study of 570 patients with 4-8 mild-to-moderate AK on the face and hairless scalp. 84% of patients were male, and the median age was 71.0 years (range, 39-87 years). Patients were randomized in a 3:3:1 ratio to BF-200 ALA-PDT (N=248), MAL-PDT (N=247), or vehicle-PDT (N=76) [233]. The follow-up study pooled results from two randomized trials with a total of 663 patients (104 women, age range: 39-87 years), one of which included the population originally described by Dirschka et al. [232]. A randomized, double-blind, intraindividual, monocentric study by Moloney et al. examined 16 men with a mean age of 71 years (age range: 59-87 years). Multiple lesions of all grades including hyperkeratotic AK were included. The two treatment fields per patient were randomized to either MAL-PDT or ALA-PDT therapy [230].

Intervention: In the studies by Dirschka et al., patients were randomized to BF-200 ALA gel (7.8%), MAL cream (160 mg/g), or vehicle gel followed by illumination by various light sources (narrow spectrum around 630 nm with 37 J/cm², broad spectrum between 580-1400 nm with 170 J/cm², spectrum 600-750 nm with 100 J/cm²) [232][233]. In the study by Moloney et al., randomization was to application of MAL for 3 hours or a 20% ALA cream for 5 hours followed by illumination of a light source with 580-740 nm at a dose of 50 J/cm² [230].

Results and assesment: In Dirschka et al., the percentage of patients with completely healed lesions at 3 months in the BF-200-ALA group was significantly superior to the MAL group (78.2% vs. 64.2%; p<0.05). At 12-month follow-up, this proportion was also higher for BF-200 ALA (BF-200 ALA: 47%, MAL-PDT: 36%). Both treatments were each more effective for lesions on the face and forehead than on the scalp. The percentage of healed lesions at three months was also higher after treatment with BF-200 ALA than with MAL (71%-83.2%), with a rate of 87%-90.4%. In the study by Moloney et al., the percentage of completely healed treatment fields was 40% with ALA-PDT and 46.7% with MAL-PDT. The mean reduction of lesions here was higher in the ALA-PDT group (6.2±1.9 vs. 5.6±3.2 for MAL-PDT). However, the difference was not statistically significant. The overall cosmetic outcome of treatment with MAL-PDT was rated as "very good/good“ in 45.2% and with BF-200 ALA-PDT as "very good/good“ in 43.1%. After a follow-up of 6 months, it was again rated similarly in both groups (ALA-PDT: 39.7%-43.1%, MAL-PDT: 42.6%). Side effects occurred more frequently in the MAL-PDT group (98.0%) than in the BF-200 ALA-PDT group (96.4%) in the study by Dirschka et al. Redness, burning, and pain were reported most frequently and were also the most severe. In the study by Moloney et al., only mild redness and erosion occurred in 2 patients. Pain (measured by VAS scale) occurred less frequently in the MAL-PDT group (4.0±3.58) than in the BF-200 ALA-PDT group (4.8±3.61) in the studies by Dirschka et al. This difference was not statistically significant. These results were confirmed in the study by Moloney et al., in which mean VAS score at 12 minutes was 3.9 in the ALA-PDT group and 2.2 in the MAL-PDT group. The difference was statistically significant. Participants felt discomfort for a longer time when treated with ALA-PDT, and two-thirds preferred treatment with MAL-PDT.

PDT with conventional 5-ALA formulation vs. placebo

A 20% ALA formulation is not approved in Germany, unlike in the USA (Levulan Kerastick®). Illumination with blue light is also not recommended. Nevertheless, the following studies are discussed for the sake of completeness.

Jeffes et al. conducted a multicenter, randomized, single-blinded (investigator-blinded) vehicle-controlled, intraindividual study of 36 patients (30 men, 6 women; mean age 68.8 years, range 38-100 years) to evaluate the safety and efficacy of 20% ALA PDT at various doses of blue light (2, 5, and 10 J/cm²) in the treatment of single or multiple AK on the face and scalp (minimum 4) [225]. In each patient, two lesions were treated with vehicle and two other lesions were treated with 20% ALA. The incubation period was 14 to 18 hours. Subsequently, all lesions in a patient were irradiated with the same dose randomized per patient. After 8 weeks, the rate of completely healed lesions was significantly higher for ALA-PDT than for vehicle-PDT (66% vs. 17%; p<0.001). It continued to increase significantly to 85% for active therapy at 16 weeks, whereas it dropped to 6% for vehicle. However, there was no difference in the rate of at least 50% healed lesions (17% for both treatments at 8 weeks). After 16 weeks, it was only 6% for ALA-PDT-treated lesions. In 46% of the study population, both ALA-PDT-treated lesions were completely healed after 8 weeks (vehicle-PDT 6%). Regarding randomized light doses, treatment with the highest dose of 10 J/cm² was found to be most effective. Under this, 88% of all treated lesions showed a complete response after 8 weeks. The dose of 5 J/cm² performed second best, with a complete response rate of 62%. In general, thin (grade I) lesions responded better to therapy than thicker (grade II) lesions. Burning and stinging during PDT were significantly more common for 20% ALA than for vehicle (89% vs 3%), with burning reported as severe in 18% of treated patients. Furthermore, pain (43%) and itching (39%) were more common in active therapy [225].

The randomized, vehicle-controlled, investigator-blinded, multicenter study by Pariser et al. compared the effect of a short incubation period of ALA-PDT with vehicle in the treatment of AK on the face and scalp [226]. For this, a total of 236 participants (211 men, 23 women; mean age: 68 years, range 40-88 years) with 6-20 Olsen grade 1-2 lesions on the face or scalp were randomized interindividually to 5 different treatment arms: large area application of 20% ALA for 1, 2 or 3 hours followed by illumination with blue light (n=47, n=48, n=47), spot application of 20% ALA for 2 hours followed by illumination with blue light (n=46) or application of vehicle followed by illumination with blue light (n=46). The majority of patients had Fitzpatrick skin type II or III (44% and 43%, respectively), with 6% each having skin type I. Overall, the ALA-PDT groups performed significantly better in healing AK compared with the vehicle group. The median rate of completely healed lesions in the active treatment arms (large-area application and 1, 2, or 3 hours of incubation, and spot application and 2 hours of incubation) was significantly superior to the vehicle-PDT group at 12 weeks (ALA-PDT: 68-79%, vehicle-PDT: 7%). In each case, the highest median rate of completely healed lesions was achieved by the ALA-PDT group with 3 hours of incubation and the group with 2 hours of incubation and spot application, respectively (57.1%±37.0 and 57.1%±43.8). The 2-hour incubation group showed a median reduction of 52.5%±37.2, with 1-hour incubation 35.7%±42.0, and the vehicle group 5.7%±33.5. Likewise, rates of patients with completely healed lesions at 12 weeks were significantly higher in the intervention groups than in the vehicle PDT group (ALA PDT group: 17%-30%, vehicle PDT group: 2%; p=0.0041). Here, the longest incubation time of 3 hours achieved the highest complete response rate of 17%, the group with an incubation time of 2 hours had a rate of 14.6%, with one-hour incubation time only 6.4% remained, and the group with punctual application and 2 hours incubation time 8.7% compared to 0% with vehicle. Regarding the rate of patients with at least 75% of lesions completely healed, a similar trend of increasing efficacy with longer incubation time could be observed in each cohort: 31.9% (3 hours), 27.1% (2 hours), 21.3% (1 hour), 28.3% (spot and 2 hours) and 2.2% (vehicle). The differences were significant in each case to the vehicle group. 79% of all patients in the ALA-PDT groups rated the treatment outcome as satisfactory, whereas only 35% in the vehicle group. Regarding tolerability, 63.8% of patients with 1-hour incubation, 79.2% with 2-hour incubation, 78.7% with 3-hour incubation, and 58.2% with 2-hour incubation and spot application rated stinging and burning during treatment as moderate or severe. The incidence of skin redness increased in all treatment groups shortly after illumination compared with baseline measurements, but was more severe in the ALA groups than in the vehicle group. The increase was highest in the 3-hour incubation group (61.7%), followed by the 2-hour incubation group (58.3%). In the 1-hour incubation group, 38.3% of patients had skin redness, compared with 41.3% in the 2-hour incubation and spot application group and 6.5% in the vehicle group. The incidence of edema was highest in patients with the longer incubation times. All participants in the intervention groups showed an increase in skin scaling and skin dryness at a visit between 24 and 48 hours after initial treatment compared with baseline. The evidence level of the study was downgraded to 3 because of ambiguities in randomization.

Piacquadio et al. reported the results of two identical phase III studies within one publication [227]. These were two multicenter, randomized, placebo-controlled, investigator-blinded trials in a parallel-group design with 243 patients (203 men and 40 women aged 34-89 years). Inclusion criteria were 4-15 discrete AK on the face or hairless scalp. Overall, participants were randomized to lesion-directed application of 20% ALA (n=181) or vehicle (n=62). This was followed by illumination with blue light within 14 to 18 hours. Treatment with ALA-PDT was significantly more effective and superior to the vehicle group in terms of efficacy. 66% of participants in the ALA-PDT group had completely healed lesions after 8 weeks (complete response). In the vehicle group, only 11% had completely healed AK. After another month, the rate of patients with completely healed lesions still increased to 89% in the intervention group and decreased to 13% in the vehicle group. Also, at 8 weeks, 77% of patients in the intervention group and 18% in the vehicle group had at least 75% healed lesions; at 12 weeks, the rate increased to 89% in the ALA-PDT group and decreased to 13% in the vehicle group. Lesion-wise, 83% of AK were completely healed at 8 weeks in the active group, and as many as 91% at 12 weeks. In the vehicle group, however, only 31% of lesions had healed after 8 weeks and 25% after 12 weeks. A total of 113 adverse events were observed, of which 92% were classified as mild or moderate. The most common side effects included skin redness and edema at the treated sites and stinging and burning during illumination. In addition, crusting (49%), itching (30%), and scaling (31%) were reported. 6.6% of patients in the intervention group and 3.2% in the vehicle group reported headache, 5.0% and 1.6%, respectively, reported injury in the treatment area, and 1.7% in the ALA-PDT group reported hypertension and hypertension. The evidence level of the study was downgraded to 3 because of ambiguities in blinding.

Schmieder at al compared 20% ALA-PDT with vehicle-PDT with blue light illumination in a multicenter, randomized, investigator-blinded, vehicle-controlled phase II trial [228]. As a special feature, they investigated whether occlusion improved treatment efficacy. Seventy patients with at least 4 AK on the upper extremities (hands, arms), including 45 men and 25 women with a mean age of 64 years (range 44-83 years), participated in the study. Each patient was randomized to treatment with either ALA or vehicle (interindividual). Then, the left and right arms were each randomized to occlusive or non-occlusive application of the assigned substances (intraindividual). The intervention consisted of application of 20% ALA hydrochloride or vehicle to the dorsum of the hand or dorsal forearms, respectively, for 3 hours, followed by illumination with blue light. If lesions persisted after 8 weeks, treatment was repeated. Treatment with ALA-PDT generally achieved significantly better results with respect to all end points compared with vehicle-PDT. In the additional comparison of occlusive vs. non-occlusive treatment, the former was again significantly more effective. The proportion of patients with completely healed lesions at 12 weeks was 34.3% for ALA-PDT with occlusion and only 20.0% without occlusion compared with 0% and 2.9%, respectively, for vehicle. The proportion of participants with at least 75% of lesions healed also showed a very similar trend (ALA-PDT with occlusion 60% and without occlusion 42.9%; vehicle-PDT with occlusion 8.6% and without occlusion 5.7%). The median lesion reduction rate was consistently higher in patients treated with ALA-PDT than with vehicle-PDT. Again, there were slight benefits after occlusion of ALA (88.7% vs. 70% without occlusion). Subjectively, participants were also satisfied with ALA-PDT in occlusion; 83% here rated the improvement in treated AK as moderate or excellent. Side effects occurred in all treatment arms. Skin redness increased in both groups after blue light treatment, but more markedly in patients treated with ALA-PDT (100% vs. 88.6%). Scaling and dry skin were also observed significantly more frequently under ALA-PDT (91.4% with vs. 85.7% without occlusion) than under vehicle-PDT (71.4% with vs. 68.6% with occlusion). Additional erysipelas and muscle pain occurred in 3% of participants treated with ALA-PDT. The study was downgraded to an evidence level of 3 because of ambiguities in treatment assignment and randomization.

Taub et al. compared 20% 5-ALA with vehicle in a randomized, blinded, intra-individual, vehicle-controlled trial [229]. The mean age of the 15 patients (11 women, 4 men) was 55.8 years with a standard deviation of 9.4 years. Participants had at least 4 lesions each on the dorsum of the hand or the dorsal aspect of the forearm of both sides. Treatment was given twice at eight-week intervals and consisted of therapy with 20%-ALA hydrochloride or vehicle on one side at a time (randomized) followed by illumination with blue light. Treatment with 20% 5-ALA was superior in this study. The percent mean reduction of lesions here was 58.4% (±22.2), statistically significantly higher than in vehicle-treated lesions (24.8±20.6%; p=0.0004). In 73% of extremities treated with ALA-PDT, the number of lesions decreased by half, but this number was only 13% with vehicle. This difference was also highly statistically significant. Side effects 48 hours after illumination such as skin redness, edema, stinging and burning were more frequent and severe in the active group. Overall, 86.7% of all participants were satisfied with the treatment. The level of evidence was downgraded to an evidence level of 3 because of ambiguities in randomization and a small number of cases.

Brian-Jiang et al. conducted a randomized trial of 20% ALA formulation versus placebo on the arms and dorsum of the hands, respectively. A total of 269 patients with 4-15 lesions (grade I-II according to Olsen) were randomized 1:1 to ALA-PDT or placebo (188 men, mean age 68 years) [256]. The patient-related complete healing rate at 12 weeks was 31% versus 13% with placebo (p<0.0001). In percentage terms, the mean healing rate was 69.1+37.4% for ALA-PDT and 29.9%+51.5% for placebo. Side effects were significantly more common with ALA-PDT, with burning (93% vs. 17%), erythema (91% vs. 58%), edema (30-40% with ALA-PDT), and scaling (76% with ALA-PDT) recorded most frequently.

PDT with conventional 5-ALA formulation vs. trichloroacetic acid

For a comparison of these interventions, see see chemoexfoliation.

PDT with conventional 5-ALA formulation vs. 5-fluorouracil 5% cream

Study design and population: Smith et al. compared the efficacy and tolerability of ALA-PDT with blue light or laser illumination versus 5-FU in a study population of 36 patients (29 men, 7 women) with a mean age of 61 years. Treatments were field-directed. The face or hairless scalp was treated either with ALA topically for 1 hour followed by blue light or laser light, or with 0.5% 5-FU. The number of lesions per patient and the degree of AK included were not reported in detail [186].

Intervention: Group 1: application of 20% ALA for 1 hour followed by blue light illumination; group 2: application of 20% ALA for 1 hour followed by pulsed dye laser at 595 nm; group 3: application of 0.5% 5-FU cream once or twice daily for a total of 4 weeks.

Results: The percentage of patients with completely healed lesions was highest under 5-FU and ALA with blue light, 50% each. Only 8% of patients in the ALA with laser illumination group showed a complete response. Regarding the proportion of patients with at least 75% completely healed lesions, treatments with 5-FU and ALA with blue light also achieved similar and at the same time the most effective results (75% each); in the treatment with ALA plus dye laser (group 2), only 8% of patients had at least 75% healed lesions. Global response was classified as "complete/almost complete“ in 8% with 5-FU, 17% with ALA plus blue light, and 8% with ALA plus dye laser. Global response was classified as moderate in 58% (5-FU), 33% (ALA plus blue light), and 42% (ALA plus dye laser). Skin redness was most common. Patients in the 5-FU group had the highest incidence of skin redness on average. Local skin reactions such as crusts and erosions were likewise most frequently observed in the 5-FU group and were less pronounced in the other two treatment arms. Overall, ALA plus blue light or plus dye laser illumination was better tolerated than treatment with 5-FU. The level of evidence was downgraded to an evidence level of 3 because of ambiguities in randomization and a small number of cases.

Population and study design: Alirezai et al. studied the efficacy of isotretinoin (13-cis-retinoic acid) in 0.1% cream base in 100 patients in France. Patients had at least 5 AK on hairless scalp, face or arms including hands [282].

Intervention: As the intervention, 0.3-0.5 g of isotretinoin 0.1% cream was applied twice daily to the field to be treated for 24 weeks and compared interindividually with vehicle.

Results: There was a significant mean lesion reduction of 3.9 (isotretinoin) versus 1.7 (vehicle) on the face. Complete or partial healing of lesions was observed here in 66% (isotretinoin) versus 45% (vehicle). For lesions on the hairless scalp, mean lesion reductions were 4.1 (isotretinoin) vs. 3.6 (vehicle). This difference was not significant. Similarly, there was no significant change in lesions on the arms or hands, with a mean lesion reduction of 2.9 (isotretinoin) vs. 1.0 (vehicle). Signs of local skin irritation such as erythema, scaling, burning, or itching were more frequent in the isotretinoin arm [282].

Assessment: In summary, this study showed significant improvement over placebo in only two efficacy endpoints examined, whereas lesion reduction on the capillitium and arms or hands was not significantly different. The strength of the effect in terms of mean lesion reduction appears rather small.

Tretinoin 0.05% cream vs. arotinoid methyl sulfone 0.05% cream

Population and study design: Another study compared a topical formulation of tretinoin (all-trans retinoic acid) with the third-generation retinoid arotinoid methyl sulfone (Ro 14-9706, see above) in 26 patients with more than three lesions on the face [283].

Intervention: This was an intra-individual hemiparesis comparison of tretinoin 0.05% cream with Ro 14-9706 0.05% cream applied to both sides of the face twice daily for 16 weeks.

Results: After 16 weeks, the mean lesion reduction was 37.8% for Ro 14-9706 and 30.3% for tretinoin (significant compared with therapy initiation). Complete healing of all lesions occurred in 0% (Ro 14-9706) and 8% (tretinoin) of patients. A partial response was observed in 48% (Ro 14-9706) and 40% (tretinoin), with no significant difference in efficacy between the two faces. Treatment with Ro 14-9706 was better tolerated in terms of redness and scaling of the treatment areas. In contrast, therapy with tretinoin had to be interrupted or reduced in 58% of cases due to adverse side effects [283].

Assessment: Due to a missing placebo arm of the study, a robust assessment of the effectiveness of the two interventions investigated is possible only to a limited extent. Furthermore, the endpoint "complete healing of all lesions“ reaches low values of 0-8% compared to other interventions, suggesting a rather low effect size. This is in contrast to a high percentage of therapy interruptions or dose reductions of tretinoin.

Bercovitch et al. studied a combination of tretinoin 0.05% cream with 5-fluorouracil 5% cream in 19 patients with multiple AK [289]. However, this study was excluded from an evidence-based recommendation because of the combination of two different agents.

Adapalene 0.1% and 0.3% gel vs placebo

Population and study design: A three-arm, interindividual randomized, controlled, investigator-blinded study compared adapalene gel in 90 Caucasian patients (Fitzpatrick skin types I and II) with 5 to 25 AK or solar lentigines with a minimal clinical spread of 2 mm per lesion [284].

Intervention: Patients were randomized in equal proportions to adapalene 0.1% gel, adapalene 0.3% gel, or vehicle. Treatment was initially applied once daily. If well tolerated, it was increased to twice daily application after 4 weeks and therapy was continued for a maximum of 9 months.

Results: At the end of therapy, there was a mean absolute lesion reduction of 0.5 (adapalene 0.1%) and 2.5 (adapalene 0.3%), whereas a lesion increase of 1.5 was observed with placebo. Global response after IGII was significantly better for adapalene 0.3% gel versus placebo at 3, 6, and 9 months and significantly better for adapalene 0.1% versus placebo at 1 and 6 months. Overall, 62% (adapalene 0.1%) and 66% (adapalene 0.3%) of treated patients were evaluated as completely ("clear“), almost completely ("marked“), or moderately ("moderate“) healed. This rate compared with 34% in the placebo arm. No serious side effects of adapalene were observed, with 3 patients terminating the study prematurely. Erythema, scaling, dryness, burning and pruritus were recorded more frequently in the two active arms and were consistently classified as mild. Mild dermatitis of the treated area occurred most frequently, in 40% (adapalene 0.1%), 25% (adapalene 0.3%) and 9% (placebo) [284].

Rating: Because the study selectively reported endpoints, it was downgraded to evidence level 3. Overall, adapalene at both concentrations showed significantly better efficacy with respect to the efficacy endpoints studied compared with placebo. However, the magnitude of the effect appears rather small, especially when measured by the absolute mean lesion reduction. This is in contrast to more frequent, albeit mild, side effects in both active arms, so that further studies are needed for a harm-benefit assessment.

Etretinate vs. placebo

Population and study design: The retinoid etretinate was compared with placebo in systemic administration in one study. It was a randomized, double-blind, crossover study. Fifty patients with histologically confirmed AK were included. However, information on the number and location of lesions is lacking. Also, it remains unclear whether patients with field cancerization or immunosuppression were within the studied population [285].

Intervention: Patients received either etretinate at a dose of 75 mg once daily or placebo for 2 months, followed by the previously unadministered therapy in each case (crossover study).

Results: After the first 2 months, the proportion of patients with complete healingwas 22.7% (5/22) on etretinate compared with 0% on placebo. A partial response was defined as at least 50% size reduction in at least 75% of all treated lesions. At 2 months, 63.6% (14/22) of patients treated with etretinate but only 4.3% (1/23) of participants on placebo achieved a partial response. At the end of therapy, of 44 patients who fulfilled the protocol regularly with etretinate, 37 (84%) showed complete healing of all lesions or a partial response compared with only 2 patients (5%) on placebo. Adverse effects of etretinate were dry mouth (86.1%), unspecified rash with pruritus (15.9%), increased skin scaling (70.4%), and nausea (4.5%). Dose reduction was carried out in 17 patients, and 6 patients terminated the study prematurely due to adverse events [285].

Assessment: The study was downgraded to an evidence level of 3 due to ambiguities in randomization and allocation of treatment arms. In addition, endpoints were not fully reported (reporting bias). Etretinate is not available in Germany due to unfavorable pharmacokinetic properties such as strong lipophilicity or a long half-life.