A.01.015-Aluminum Oxide (topical) vs. Other Cleansers (topical)

There was one small five-arm trial, Trial 239 ²⁸ that studied the cleansers or abradants aluminum oxide, polystyrene, polyethylene, polyester, and pumice in patients with moderate to severe acne. Patients were not randomized, physicians were not blinded, there were no losses to followup, and the arms were not comparable. Outcomes focused on non-inflammatory lesions; differences were not statistically significant.

B.01.01-Azelaic Acid (topical) vs. Inert (topical)

There were two trials: one small British trial, Trial 77, ³¹ and one medium Greek trial, Trial 187. ³² The Greek trial studied subjects with moderate acne; the British trial matched subjects, but the acne severity cannot be categorized. Both trials were randomized and measured changes in inflammatory and non-inflammatory lesions.

Both trials showed statistically significant reductions in inflammatory and non-inflammatory lesions at 2-3 months in subjects treated with azelaic acid. Trial 187 also showed a significant improvement in physician rating of overall change.

Side effects, including burning, erythema, itching, and scaling were reported in small numbers in both treatment and placebo groups.

B.01.02-Azelaic Acid (topical) vs. Benzoyl Peroxide (topical)

There was one large (N = 309) trial, Trial 52. ³³ Limited information about the subjects and methodology was provided. Although significant changes from baseline were reported in both treatment groups, neither treatment was superior in the short-, medium-, or long-term (maximum followup = 24 weeks). No between-group P-values were reported. The authors reported that benzoyl peroxide produced a more rapid effect.

No specific information about side effects was reported, although the authors stated that azelaic acid applications were "better tolerated."

B.01.03-Azelaic Acid (topical) vs. Tretinoin (topical)

See E.05.04.

B.01.04-Azelaic acid (topical) vs tetracycline (oral)

See H.08.05.

B.02.01-Azelaic Acid/Glycolic Acid (topical) vs. Tretinoin/Vehicle (topical)

See E.05.04.

B.03.01-Salicylic Acid (topical) vs. Inert

There were two trials, a large matched-control trial of 114 patients, Trial 95, ³⁴ of 2 percent salicylic acid and a medium parallel trial, Trial 294, ³⁵ of 1.5 percent salicylic acid. Trial 95 had serious methodological flaws as almost half the subjects departed from protocol at some point secondary to side effects.

Both trials reported greater improvements from baseline in the group treated with salicylic acid. In the matched-control trial, these differences were statistically significant at 12 weeks but not at 4 weeks. In Trial 294, salicylic acid led to an apparently greater improvement in facial oiliness, but comparison statistics were not provided.

The parallel trial reported local reactions (skin and eye burning) in 10 subjects in the salicylic acid group and none in the placebo group. No P-values were provided. No side effects data were reported in the large trial.

B.03.02-Salicylic Acid (topical) vs. Benzoyl Peroxide (topical)

See D.01.04.

B.04.01-Sulfur/resorcinol (topical) vs. benzoyl peroxide (topical)

See D.01.05.

B.04.02-Sulfur/Resorcinol (topical) vs. Vitamin A (topical)

See E.06.02.

B.05.01-Polythionate/Quaternary Ammonia Complex of Polythionate (topical) vs. Polythionate (topical)

There was one medium trial, Trial 86, ³⁶ where acne severity could not be categorized. Although treatment duration was not specified, data were given at 1 year, at which time overall change was comparable between the two groups. The authors reported that there were no side effects.

B.05.02-Sulfurated Lime (topical) vs. Vehicle (vehicle)

There was one medium trial, Trial 101, ³⁷ of patients with moderate acne. The single efficacy outcome, subjectively assigned severity change, was statistically significantly better in the treatment group, with two-thirds moderately or excellently improved versus one-quarter in the vehicle group. No side effects data were reported.

B.06.01-Komed^® (Sulfur/Salicylic Acid/Resorcinol) (topical) vs. Inert

There was one large (N=51) split-face trial, Trial 191, ³⁸ of subjects with mild, moderate, and severe acne. Subjects were a referral population and there were no reported losses to followup. The one outcome was physicians' assessment of superiority of one face-side versus the other. Komed^® was rated as better on both sides of the face, although comparison statistics were not provided. No data about side effects were reported.

B.06.02-Komed^® (Sulfur/Salicylic Acid/Resorcinol) (topical) vs. Proprietary Formula (Sulfur/Salicylic Acid/Resorcinol) (topical)

There was one large (N=51) split-face trial, Trial 191, ³⁸ of subjects with mild, moderate, and severe acne. Subjects were a referral population and there were no reported losses to followup. Treatment was administered for 16 weeks, with followup for up to 36 weeks. The outcome measured was overall change in acne by physician assessment. The one outcome was physicians' assessment of superiority of one face-side versus the other. Komed^® was ated as about the same as the proprietary formula, although the numbers are small. No data about side effects were reported.

B.07.01-Aluminum Chloride (topical) vs. Vehicle

There was one small split-face trial, Trial 165, ³⁹ of patients with mild to moderate acne. Reported data show that patients were blinded. The single outcome, assessment of overall change, was not statistically significantly different. No side effect data were reported.

B.07.02-Aluminum Chlorhydroxide/Sulfur (topical) vs. Inert

There was one large (N = 776) multi-center trial, Trial 228. ⁴⁰ The trial was conducted at multiple sites in the United Kingdom and represents pooled data from multiple trials conducted in general practices and dermatologists' offices without specific information about how these data were combined. Characteristics of the subjects and specific methodological information were not reported. No specific information about blinding of subjects or assessors was reported. At 5 weeks, 142 of 201 subjects treated with aluminum hlorhydroxide/sulfur were improved, by physician assessment of overall change in acne, compared to 76 of 151 treated with vehicle alone. This was statistically significant at p<0.001. No information about side effects was reported.

B.07.03-Aluminum Chlorhydroxide/Sulfur (topical) vs. Neomycin/Methylprednisolone

See N.02.06.

B.07.04-Gluconolactone (topical) vs. Inert

There was one large (N = 150) trial, Trial 176, ⁴¹ of patients with moderate to severe acne. Although subjects in the two groups were reportedly similar in terms of acne duration at enrollment, no data were provided to support this. Although acne severity was assessed at baseline, it was not clear that the groups were comparable. From the lesion count data provided, gluconolactone led to a statistically significant decrease in inflammatory and total lesion counts, with respect to baseline, although only the change in inflammatory lesions was statistically significantly better than the change in subjects treated with placebo. Side effects reported were dryness, scaling, burning, stinging, and redness. These effects were more frequently reported in the gluconolactone group. No P-values were reported.

B.07.05-Gluconolactone (topical) vs. Benzoyl Peroxide (topical)

See D.01.03.

B.07.06-Soybean Liposome/Glycolic Acid (topical) vs. Soybean Liposome (topical)

There was one small split-face, multi-center trial, Trial 246, ⁴² of mild-to-moderately affected patients, notable for no losses to followup, and for a pharmaceutical-employee co-author. Both treatments were effective on comedones, papules, and pustules, with the combined treatment being statistically significantly better in each lesion type. There were more withdrawals from the combination group, but the number (2) was small.

B.07.07-Nicotinamide (topical) vs. Clindamycin (topical)

See C.01.02.

C.01.01-Clindamycin (topical) vs. Inert

There were eight trials addressing the effectiveness of topical clindamycin preparations (clindamycin phosphate or clindamycin hydrochloride) versus inert substances (placebos or vehicles): four large (Trial 20, ⁴³ Trial 33, ⁴⁴ Trial 99, ⁴⁵ and Trial 214 ⁴⁶ ), two medium (Trial 199, ⁴⁷ Trial 217, ⁴⁸ ), and two small (Trial 327 ⁴⁹ and Trial 144 ⁵⁰ ). These trials included subjects with mild to severe acne and an approximate mean age of 20 years. Seven of the trials were conducted in the USA and one (Trial 33) was conducted in Oslo, Norway. Four trials studied a dose of 1 percent applied twice daily, one studied a dose of 1 percent before bed, and three reported no dose information.

Subjects treated with clindamycin had similar reductions in lesion counts to those treated with inactive substances. Clindamycin tended to produce greater reductions in inflammatory lesions and greater reductions when counts were assessed at later time points, but these results were rarely statistically significant. Clindamycin was more consistently superior to placebo when global measures of acne severity or improvement were measured by subject or physician report. There were limited data reported regarding side effects. These data tended to show similar rates of cutaneous side effects and diarrhea between subjects treated with clindamycin and inactive preparations, although insufficient data were reported for statistical significance. There was one case (Trial 20) of diarrhea attributed to clindamycin; the patient was withdrawn, and it was not pseudomembranous colitis. One patient, in Trial 20, dropped out because of local burning.

C.01.02-Clindamycin (topical) vs. Nicotinamide (topical)

There was one medium multi-center trial, Trial 305, ⁵¹ of patients with moderate acne. The trial was compromised by major losses to followup. While both preparations produced improvement in severity and lesion counts, there were no statistically significant differences between them for any of the outcomes measured. There were no reported differences in local reactions between the two treatments.

C.01.03-Clindamycin (topical) vs. Clindamycin (topical)

There were six parallel trials that reported on the relative effectiveness of different topical clindamycin preparations. Three trials (Trial 20 ⁴³ Trial 143, ⁵² Trial 229 ⁵³ ) tested clindamycin phosphate versus clindamycin hydrochloride. Four trials (Trial 99, ⁴⁵ Trial 131, ⁵⁴ Trial 229, ⁵³ Trial 267 ⁵⁵ ) tested clindamycin in different vehicles such as gel, lotion, and solution. All were parallel trials of subjects with acne severity ranging from mild to severe. Two of the trials (Trial 143, Trial 229) used a split-face design.

Trials of topical clindamycin phosphate versus topical clindamycin hydrochloride showed no significant differences in severity or lesion counts between the two preparations.

Trials of topical clindamycin in different vehicles failed to show differences in effectiveness whether the vehicle was a gel, lotion, or solution. Trial 229 showed slightly greater improvement with the Ionax^® preparation versus the E-Solve^® preparation.

There were no significant differences in side effects between various topical clindamycin preparations. Trial 131 showed slightly less drying and erythema associated with use of lotion versus solution. Across the trials, there were two cases of mild diarrhea that did not lead to discontinuation of treatment.

C.01.04-Clindamycin (topical) vs. Erythromycin (topical)

There were five parallel trials involving subjects that compared topical preparations of clindamycin to topical preparations of erythromycin. There were four large trials, Trial 158 ⁵⁶ (120 patients), Trial 208 ⁵⁷ (109 patients), Trial 244 ⁵⁸ (120 patients), and Trial 303 ⁵⁹ (178 patients). There was one medium trial, Trial 326. ⁶⁰ These trials generally compared 1 percent clindamycin to 2 percent erythromycin, although one trial (Trial 326 ⁶⁰ ) used 1.5 percent clindamycin and 2 trials (Trial 303, ⁵⁹ Trial 326 ⁶⁰ ) used 1.5 percent erythromycin. All the trials enrolled subjects with mild to severe acne. Trial 158 had blinded ascertainment, but neither physicians nor patients were blinded. In Trial 208, patients were not blinded. Trial 244 had no losses to followup, but the enroller was not blinded, and arms were not comparable.

Although several trials documented differences between the preparations for certain outcomes at certain time points, these differences occurred in favor of each treatment within individual trials and none of the trials showed consistent differences overall. A review of all five trials fails to show that one treatment is superior to the other.

There was very limited information on side effects reported in these trials.

C.01.05-Clindamycin (topical) vs. Tetracycline (topical)

There were two parallel trials that compared the effectiveness of topical clindamycin (1 percent twice daily to topical tetracycline (no dose specified in either trial). One medium trial, Trial 261 ⁶¹ included patients with mild to moderate acne, the other medium trial was of patients with mild acne, Trial 291. ⁶²

Although differences in specific lesion counts or changes in lesion counts were insignificant or inconsistent, both trials showed a significant difference in favor of clindamycin in their overall measures of acne severity or improvement. In Trial 261, overall improvement was rated as "poor" in 2 of 22 in the clindamycin group and 13 of 27 in the tetracycline group. In Trial 291, acne severity at 8 weeks was rated as poor in 1 of 44 in the clindamycin group and 11 of 43 in the tetracycline group.

There were no differences in side effects in either trial, although information was limited.

C.01.06-Clindamycin (topical) vs. Benzoyl Peroxide (topical)

There was one large multi-center, multi-arm trial, Trial 214, ⁴⁶ and two small trials, Trial 324 ⁶³ and Trial 329, ⁶⁴ of patients with moderate acne. All three trials used 1 percent clindamycin and 5 percent benzoyl peroxide. Trial 324 and Trial 329 studied twice daily application and Trial 214 studied before-bed application.

All three trials demonstrated significant reductions in inflammatory and non-inflammatory lesions from baseline for both treatments. There was no statistically significant difference between the two at any time point except at 12 weeks in Trial 324 when a significantly greater reduction in non-inflammatory lesions was noted in the benzoyl peroxide group. Similarly, in Trial 329, benzoyl peroxide led to a statistically significantly greater reduction in severity, with respect to baseline, than did clindamycin, at 10 weeks.

Except for statistically significantly greater irritancy score for benzoyl peroxide in Trial 329, data about side effects in these trials did not support a statistically significant difference in local reactions such as erythema, peeling, dryness, burning, pruritus, or oiliness.

C.01.07-Clindamycin (topical) vs. Clindamycin/Benzoyl Peroxide (topical)

There was one large (N=393) parallel, multi-center, multi-arm trial, Trial 214, ⁴⁶ that reported on the relative effectiveness of topical clindamycin (1 percent) versus topical clindamycin (1 percent)/benzoyl peroxide (5 percent) applied before-bed for treatment of moderate acne. Both preparations produced significant reductions from baseline of inflammatory and non-inflammatory lesions. These reductions were greater for the combination of clindamycin/benzoyl peroxide than for clindamycin alone. More subjects had their acne rated as good or excellent in the combination group relative to the clindamycin alone group.

Side effects reported were erythema, peeling, burning, dryness, and pruritus. Peeling was reported relatively more often in subjects using the clindamycin/benzoyl peroxide preparation.

C.01.08-Clindamycin (topical) vs. Erythromycin/Benzoyl Peroxide (topical)

See D.05.02.

C.01.09-Clindamycin (topical) vs. Erythromycin/Zinc (topical)

See N.08.03.

C.01.10-Clindamycin (topical) vs. Tretinoin/Clindamycin (topical)

See G.02.01.

C.02.01-Clindamycin/Benzoyl Peroxide (topical) vs. Inert

See D.04.01.

C.02.02-Clindamycin/Benzoyl Peroxide (topical) vs. Clindamycin (topical)

See C.01.07.

C.02.03-Clindamycin/Benzoyl Peroxide (topical) vs. Benzoyl Peroxide (topical)

There was one large parallel multi-center, multi-arm trial, Trial 214, ⁴⁶ and one medium trial, Trial 329, ⁶⁴ of patients with moderate acne. In Trial 214, the combination produced statistically significantly greater reductions from baseline of inflammatory and non-inflammatory lesions than benzoyl peroxide alone. However, in Trial 329, the number of lesions appeared comparable (no P-values were provided), and severity was improved to a statistically significant extent with respect to clindamycin, not with respect to benzoyl peroxide. A larger proportion of subjects had their acne rated as good or excellent in the combination group. Side effects reported with relatively equal frequency in both groups were erythema, peeling, burning, dryness, and pruritus.

C.03.01-Erythromycin (topical) vs. Inert

There were four large parallel trials (Trial 184, ⁶⁵ Trial 205, ⁶⁶ Trial 274, ⁶⁷ and Trial 89 ⁶⁸ ), two medium trials (Trial 277 ⁶⁹ and Trial 39 ⁷⁰ ), and two small studies (Trial 157 ⁷¹ and Trial 285 ⁷² ) of patients with mild to moderate acne.

Six trials showed statistically significant improvement in inflammatory lesions by erythromycin. Trial 157 reported that more patients treated with erythromycin assessed their acne as "better" than did those treated with vehicle, but P-values were not provided. Trial 277 failed to show a difference between the preparations in terms of inter-treatment lesion change at 8 weeks.

There were no reported differences in side effects between topical erythromycin and its vehicle.

C.03.02-Erythromycin (topical) vs. Clindamycin (topical)

See C.01.04.

C.03.03-Erythromycin (topical) vs. Erythromycin/Zinc (topical)

See N.08.02.

C.04.01-Fusidic Acid (topical) vs. Inert

There were two medium trials, Trial 151 ⁷⁴ and Trial 315, ⁷⁵ of patients with mild to moderate or non-categorizable. No statistically significant differences between topical fusidic acid and inactive preparations were observed.

C.04.02-Fusidic Acid (topical) vs. Doxycycline (oral)

See H.03.02.

C.04.03-Fusidic Acid (topical) vs. Sodium Fusidate (oral)

See H.09.04.

C.05.005-Meclocycline (topical) vs. Inert (topical)

There were two large two-arm trials, Trial 194 ⁷⁶ and Trial 195, ⁷⁶ of patients with moderate to severe acne. Both trials had no losses to followup, but no randomization and inconsistent ascertainment across arms. There was a statistically significantly greater decrease in inflammatory lesions at 5 and 11 weeks in the meclocycline group. The side effect of follicular staining was noted in about 16 percent of anti-bacterial-treated patients, but none in the vehicle group.

C.05.01-Meclocycline (topical) vs. Benzoyl Peroxide (topical)

There was one large (N=102) trial, Trial 30, ⁷⁷ of young adults with moderate acne. There were no losses to followup, but the outcome assessor was not blinded. Comedone count was statistically significantly fewer in the benzoyl peroxide group, but not for papules and pustules. There were three cases of skin discoloration split between both groups.

C.05.02-Meclocycline (topical) vs. Meclocycline/Benzoyl Peroxide (topical)

See D.06.01.

C.05.03-Meclocycline (topical) vs. Tetracycline (oral)

See H.08.06.

C.06.01-Metronidazole (topical) vs. Inert

There was one medium trial, Trial 328, ⁷⁸ of patients with mild to moderate acne. No significant differences between topical metronidazole and placebo were observed in this trial.

C.07.01-Neomycin (topical) vs. Keratolytics/Anti-inflammatory (topical)

There was one medium multi-arm trial, Trial 287, ⁷⁹ with patients whose acne status cannot be categorized, whose reported execution differed from the reported design. The following proportions of subjects were classified as improved in each group: neomycin, 10 of 20; vehicle, 14 of 26; hydrocortisone, 14 of 24; sulfur, 11 of 23; resorcinol, 17 of 24; combination, 21 of 32. Differences were not statistically significant.

C.08.01-Tetracycline (topical) vs. Inert

There were four large parallel trials: Trial 10 ⁸⁰ (moderate severity), Trial 23 ⁸¹ (moderate), Trial 40 ⁸² (cannot determine severity), and Trial 313 ⁸³ (mild to moderate severity). The first three used topical tetracycline 0.5 percent twice daily. The fourth trial used topical tetracycline 2.2 percent twice daily.

The three trials using 0.5 percent reported no statistically significant differences between topical tetracycline and inactive preparations. In the fourth trial, 25 patients treated with tetracycline were classified as "better" compared to 18 in the placebo group.

In the trial using 2.2 percent, tetracycline was associated with more skin discoloration, placebo with more exfoliation.

C.08.02-Tetracycline (topical) vs. Clindamycin (topical)

See C.01.05.

C.08.03-Tetracycline (topical) vs. Benzoyl Peroxide (topical)

There was one large (N=69, 45 females, age range 12-33 years) parallel trial, Trial 255, ⁸⁴ comparing the effectiveness of topical tetracycline to topical 5 percent benzoyl peroxide applied once per day for treatment of moderate acne. Outcomes assessed at 8 and 12 weeks were acne severity (by the Leeds method) on back, face, and chest as well as counts of inflammatory and non-inflammatory lesions.

No significant differences were observed between topical tetracycline and topical benzoyl peroxide in any of the outcomes assessed.

C.08.04-Tetracycline (topical) vs. Povidone-Iodine/Tetracycline (topical)

See H.08.14.

C.08.05-Tetracycline (topical) vs. Oxytetracycline (oral)

See H.06.05.

C.08.06-Tetracycline (topical) vs. Tetracycline (oral)

See H.08.04.

C.09.01-Triclosan (topical) vs. Inert

There was one medium trial, Trial 114, ⁸² of patients with mild to moderate acne. The trial was conducted using male subjects at a military hospital.

No significant differences were observed between topical triclosan and placebo in any of the outcomes assessed (patients' assessment of change and lesion counts). No side effect data were reported.

C.09.02-Triclosan (topical) vs. Hexachlorophene (topical)

There was one medium cross-over trial, Trial 21 ⁸⁵ (N=34). The trial reported the involvement of a pharmaceutical company. Outcomes assessed at 6 and 12 weeks were overall change as assessed by the subjects and by a physician.

There were no significant differences reported for either of the outcomes (patients' assessment of change and lesion counts) between topical triclosan and topical hexachlorophene. Almost all patients in both arms experienced local drying, scaling, or itching.

C.09.03 -Triclosan/propylene Phenoxetol (topical) vs. Vehicle vs. Triclosan

There was one medium trial, Trial 40, ⁸² of patients with mild to moderate acne. The trial was conducted using male subjects at a military hospital. No significant differences were observed between topical triclosan and a topical preparation of triclosan and propylene phenoxetol in any of the outcomes assessed (patients' assessment of change and lesion counts). No side effect data were reported.

D.01.01-Benzoyl Peroxide (topical) vs. Inert

There were seven parallel trials. Trial 241 ⁸⁶ , Trial 173, ⁸⁷ and Trial 314 ⁸⁸ were medium, two-arm trials; Trial 176 ⁴¹ was a medium, three-arm trial; Trial 177 ⁸⁹ and Trial 97 ⁹⁰ were medium, four-arm studies, and Trial 214 ⁴⁶ was a large, four-arm trial. Each studied the 5 percent concentration of benzoyl peroxide, except for Trial 314, which used 20 percent. Two of these (Trial 173, Trial 314) were specifically designed for this comparison. In each of the others, benzoyl peroxide was used as the control treatment for other active preparations. The trials were conducted in Britain (Trial 173, Trial 177), Australia (Trial 176), and the USA (Trial 214, Trial 241). Four of the trials studied twice daily application of the treatment (Trial 173, Trial 177, Trial 241, and Trial 314), one studied before-bed application (Trial 214), one (Trial 97) had patients applying treatment between once and four times daily, and the other gave no details about how the treatment was used. Acne severity ranged from mild to severe in the trials.

All trials showed improvements from baseline in their measures of acne severity. These changes were consistently superior to the effects of placebo at a statistically significant level.

Side effects such as erythema, peeling, dryness, burning, pruritus, stinging, itching, and soreness were reported more frequently in groups treated with benzoyl peroxide. In Trial 176, of moderately-to-severely affected patients, local side effects were statistically more frequent than in the vehicle group.

D.01.02-Benzoyl Peroxide (topical) vs. Azelaic Acid (topical)

See B.01.02.

D.01.03-Benzoyl Peroxide (topical) vs. Gluconolactone (topical)

There was one large trial, Trial 176, ⁴¹ of patients with moderate to severe acne. Although subjects in the two groups were reportedly similar in terms of acne duration at enrollment, no data were provided to support this. Although acne severity was assessed at baseline, it was not clear that the groups were comparable. The decrease was statistically significantly greater in benzoyl peroxide for inflammatory and non-inflammatory lesions. Side effects reported were dryness, scaling, burning, stinging, and redness. These effects were less frequently reported in the gluconolactone group. No P-values were reported.

D.01.04-Benzoyl Peroxide (topical) vs. Salicylic Acid (topical)

There was one medium cross-over trial, Trial 304, ⁹¹ of mild acne. There were no losses to followup or important protocol departures reported. However, the subjects, enrollers, and outcomes assessors were not blinded to treatment assignment. There were no statistically significant differences in change in comedones between the preparations in either phase of the trial. Information regarding side effects was not reported.

D.01.05-Benzoyl Peroxide (topical) vs. Sulfur/Resorcinol (topical)

There was one large (N = 189) trial, Trial 272, ⁹² where acne severity could not be categorized. There was very limited information about methodology reported. A greater reduction in total lesions was observed in the benzoyl peroxide group versus the sulfur/resorcinol group, but this difference was not statistically significant. No data about side effects were reported.

D.01.06-Benzoyl Peroxide (topical) vs. Clindamycin Phosphate (topical)

See C.01.06.

D.01.07-Benzoyl Peroxide (topical) vs. Meclocycline (topical)

See C.05.01.

D.01.08-Benzoyl Peroxide (topical) vs. Tetracycline (topical)

See C.08.03.

D.01.09-Benzoyl Peroxide (topical) vs. Benzoyl Peroxide (topical)

There were seven trials that compared various preparations of topical benzoyl peroxide to one another. These comparisons were drawn between benzoyl peroxide in different vehicles (alcohol, water, acetone, gel, lotion) and different concentrations (2.5 percent, 5 percent, and 10 percent).

Trial 351 ⁹³ was a large two-arm trial comparing 2.5 percent with a 5 percent preparation. There were no losses to followup, but physicians were not blinded. Trial 242 ⁸⁶ was a medium trial of patients with mild to moderate acne that compared the effectiveness of topical benzoyl peroxide (2.5 percent) and topical benzoyl peroxide (5 percent) applied twice daily. There were no reported losses to followup. Trial 243 ⁸⁶ was a medium trial of patients with mild to moderate acne that compared the effectiveness of topical benzoyl peroxide (2.5 percent) and topical benzoyl peroxide (10 percent) applied twice daily. There were no reported losses to followup. Trial 221 ⁹⁴ was a large multi-center trial of patients (mean age 46 years) with moderate acne, with daily application of topical 5 percent benzoyl peroxide in either an acetone or alcohol base. This was part of a trial comparing both of these preparations to tretinoin. Subjects, enrollers, and outcomes assessors were not blinded to treatment assignment. Trial 118 ⁹⁵ was a medium split-face design where acne severity could not be categorized. The trial compared the effectiveness of topical 5 percent benzoyl peroxide in an alcohol base to the same medication in a water base. Trial 76 ⁹⁶ was a medium matched-control trial of patients with moderate acne, with topical 5 percent benzoyl peroxide in a gel and with the same medication in a lotion. Limited information about methodology was reported. Trial 149 ⁹⁷ was a large five-arm trial of benzoyl peroxide 5 percent and 10 percent alone and with tretinoin.

Of the many outcomes based on lesion counts, there were no statistically significant differences among these different treatments, and there was no dose-response effect demonstrated.

Side effects were numerically almost identical in the 2.5 percent and 5 percent concentration groups in Trial 351. In Trial 243, side effects were more prevalent in the 10 percent group, although P-values were not reported. In Trial 242, side effects were not reported. In Trial 221, side effects were reported and were similar in the two groups. In Trial 118, side effects were reported. Burning and erythema occurred with similar frequency between the groups. Scaling and stinging were reported more frequently with the alcohol-based product. In Trial 76, erythema and scaling were reported with similar frequency in both groups. In Trial 149, there were statistically significantly more side effects in the 10 percent group than in the 5 percent group.

D.01.10-Benzoyl Peroxide (topical) vs. Benzoyl Peroxide/Urea (topical)

There was one medium split-face trial, Trial 278, ⁹⁸ of patients with moderate acne. Patient education to minimize improper application and translocation was documented in the report. There were minimal losses to followup and only minor protocol departures. There was no statistically significant difference between the preparations for inflammatory and total lesion counts as well as change in inflammatory lesions. No information about side effects was reported.

D.01.11-Benzoyl Peroxide (opical) vs. Benzoyl Peroxide/Meclocycline

See D.06.01.

D.01.12-Benzoyl Peroxide (topical) vs. Benzoyl Peroxide/Metronidazole (topical)

See D.07.02.

D.01.13-Benzoyl Peroxide (topical) vs. Tretinoin (topical)

See E.05.02.

D.01.14-Benzoyl Peroxide (topical) vs. Vitamin A (topical)

See E.06.02.

D.01.15-Benzoyl Peroxide (topical) vs. Oxytetracycline (oral)

See H.06.02.

D.01.16-Benzoyl Peroxide (topical) vs. Isotretinoin (oral)

See E.02.02.

D.01.17-Benzoyl Peroxide (topical) vs. Miconazole/Benzoyl Peroxide (topical)

See D.08.01.

D.01.18-Benzoyl Peroxide (topical) vs. Chloramphenicol/Hydrocortisone/Sulfur (topical)

See N.04.05.

D.01.19-Benzoyl Peroxide (topical) vs. Isolutrol (topical)

See N.01.03.

D.02.01-Chloramphenicol/Sulfur (topical) vs. Chloramphenicol/Hydrocortisone/Sulfur (topical)

See N.04.03.

D.03.01-Benzoyl Peroxide/Glycolic Acid/Zinc Sulfur (topical) vs. Inert

There was one medium two-site trial, Trial 310, ⁹⁹ of patients with mild to moderate acne. Data showed that patients and enrollers were blinded to treatment assignment. The performance of outcomes assessors was monitored by a dermatologist. The trial treatment and placebo produced statistically significant improvements in lesion counts in the short (4 weeks) and long (12 weeks) term. However, there was no statistically significant difference between the trial treatment and placebo at either followup time.

Side effects of scaling or tightness of the skin were recorded at 4 and 12 weeks. Although these effects occurred more frequently in the group treated with the trial medication, the difference between this and placebo was not statistically significant.

D.03.02-Benzoyl Peroxide/Hydroxyquinolone/Hydrocortisone (topical) vs. Benzoyl Peroxide/Hydroxyquinolone (topical)

There were two trials: one medium cross-over trial, Trial 16, ¹⁰⁰ of patients with mild to moderate acne, and one medium four-arm trial, Trial 177, ⁸⁹ of patients whose acne status could not be categorized. Severity in Trial 177 was statistically significantly improved in every arm, including vehicle. Severity was improved in both arms in Trial 16, without P-values reported. No between-arms P-values were reported in either trial. Inflammatory and cyst counts in Trial 16 were improved in both arms, but no comparison P-values were reported. No data on side effects were reported.

D.03.03-Benzoyl Peroxide/Chlorhydroxyquinolone/Hydrocortisone (topical) vs. Benzoyl Peroxide/Chlorhydroxyquinolone

There was one large (N=196) four-arm trial, Trial 97, ⁹⁰ of patients with mild to moderate acne. Overall change and lesion counts improved in all arms, but only statistically significantly greater than vehicle arm. It is not clear if the components beyond benzoyl peroxide added efficacy. There were no side effects data reported.

D.03.04-Benzoyl Peroxide/Sulfur (topical) vs. Sulfur (topical)

There was one large trial (N=113), Trial 82, ¹⁰¹ of patients with mild, moderate, and severe acne. Subjects and enrollers were blinded to treatment assignment. At 2 weeks, 46 of 56 subjects in the benzoyl peroxide/sulfur group were assessed as either "good' or "excellent" versus 22 of 57 in the sulfur group. This difference was statistically significant at p<0.001. Skin irritation was reported in five subjects in the benzoyl peroxide/sulfur group and two in the sulfur group.

D.04.01-Clindamycin/Benzoyl Peroxide (topical) vs. Inert

There was one large (N=393) multi-center, multi-arm trial, Trial 214 ⁴⁶ of patients with moderate acne. The clindamycin/benzoyl peroxide preparation produced statistically significant reductions from baseline of inflammatory and non-inflammatory lesions; the vehicle did not. More subjects had their acne rated as good or excellent in the clindamycin/benzoyl peroxide group relative to the vehicle group. Side effects reported were erythema, peeling, burning, dryness, and pruritus. Peeling, particularly, was reported relatively more often in subjects using the clindamycin/benzoyl peroxide preparation.

D.04.02-Clindamycin/Benzoyl Peroxide (topical) vs. Clindamycin (topical)

See C.01.07.

D.04.03-Clindamycin/Benzoyl Peroxide (topical) vs. Benzoyl Peroxide (topical)

See C.02.03.

D.05.01-Erythromycin/Benzoyl Peroxide (topical) vs. Inert

There was one large four-arm trial, Trial 53 ¹⁰² and one medium trial, Trial 310, ⁹⁹ of patients with mild to moderate acne. Trial 53 had inconsistent ascertainment across arms. In Trial 310, a pharmaceutical company is listed as a co-author. The performance of assessors was monitored by a dermatologist. Data showed that subjects and enrollers were blinded, and there were no reported losses to followup. In Trial 53, erythromycin and benzoyl peroxide, and their combination, produced statistically significant decreases in inflammatory lesions. The combination had a numerically larger effect than either constituent, although statistics were not provided for that comparison. The effect was smaller, and statistically significant only at 10 weeks for non-inflammatory lesions. In Trial 310, both benzoyl peroxide/erythromycin and placebo produced significant reductions from baseline in lesion counts at 4 and 12 weeks. However, neither preparation was superior to the other at either time point. In Trial 310, scaling or tightness was reported more often in the group treated with benzoyl peroxide/erythromycin.

D.05.02-Erythromycin/Benzoyl Peroxide (topical) vs. Clindamycin (topical)

There was one large (N=205) multi-center trial, Trial 260, ¹⁰³ of young and older adult women with mild to moderate acne. A pharmaceutical company employee was the primary author. Inter-site differences in baseline severity were reported to significantly influence the results and were controlled for in the analysis. Data showed that the subjects, enrollers, and assessors were blinded to treatment assignment, but assignment was deterministic.

Counts of comedones at 4 and 10 weeks and counts of inflammatory lesions at 10 weeks were statistically significantly lower in the group treated with benzoyl peroxide/erythromycin than in the group treated with clindamycin phosphate. Overall change scores were similar in both groups at all timepoints. There were few reported side effects. Three subjects in the benzoyl peroxide/erythromycin group withdrew from the trial compared to none in the clindamycin phosphate group.

D.05.03-Erythromycin/Benzoyl Peroxide (topical) vs. Benzoyl Peroxide/Glycolic Acid/Zinc Sulfur (topical)

There was one medium trial, Trial 310, ⁹⁹ of patients with mild to moderate acne. A pharmaceutical company employee was listed as a co-author. The performance of assessors was monitored by a dermatologist. Data showed that subjects and enrollers were blinded, and there were no reported losses to followup. Both benzoyl peroxide/erythromycin and placebo produced significant reductions from baseline in lesion counts at 4 and 12 weeks. Benzoyl peroxide/erythromycin produced greater reduction in inflammatory lesions at 4 and 12 weeks than benzoyl peroxide/glycolic acid/zinc sulfur. Scaling or tightness was reported more often in the group treated with benzoyl peroxide/erythromycin.

D.05.04-Erythromycin/Benzoyl Peroxide (topical) vs. Zinc/Erythromycin (topical)

There was one medium trial, Trial 62, ¹⁰⁴ of patients whose acne severity could not be categorized. There were no significant differences between the two preparations for physicians' perceived changes in comedones. Benzoyl peroxide/erythromycin produced greater reductions in inflammatory lesions and superior clinical assessment scores than did zinc/erythromycin. A greater number of subjects improved by "50 percent or more" in the benzoyl peroxide/erythromycin group at all time points. Dryness and irritation was reported by four subjects in the zinc/erythromycin group and by one subject in the benzoyl peroxide/erythromycin group.

D.05.05-Erythromycin/Benzoyl Peroxide (topical) vs. Group Therapy

See N.06.03.

D.05.06-Erythromycin/Benzoyl Peroxide (topical) vs. Relaxation Training/Benzoyl Peroxide/Erythromycin

See N.06.03.

D.06.01-Meclocycline/Benzoyl Peroxide (topical) vs. Benzoyl Peroxide (topical)

There was one large (N=102) trial, Trial 30, ⁷⁷ of young adults with moderate acne. There were no losses to followup, but the outcome assessor was not blinded. Comedo count was statistically significantly fewer in the benzoyl peroxide-only group, but not for papules and pustules. There were two cases of skin discoloration split between both groups.

D.07.01-Metronidazole/Benzoyl Peroxide (topical) vs. Inert

There was one medium trial, Trial 121, ¹⁰⁵ of patients with moderate to severe acne. A statistically significantly greater number of subjects in the benzoyl peroxide/metronidazole group were improved by self-assessment and by photographic assessment than in the vehicle group. No evidence about side effects was reported.

D.07.02-Metronidazole/Benzoyl Peroxide (topical) vs. Benzoyl Peroxide (topical)

There was one medium trial, Trial 121, ¹⁰⁵ of patients with moderate to severe acne. There were no reported losses to followup. A statistically significantly greater number of subjects in the benzoyl peroxide/metronidazole group were improved by self-assessment and by photographic assessment than in the benzoyl peroxide group. No data about side effects were reported.

D.07.03-Metronidazole/Benzoyl Peroxide (topical) vs. Oxytetracycline (oral)

See H.06.03.

D.08.01-Miconazole/Benzoyl Peroxide (topical) vs. Benzoyl Peroxide (topical)

There were three trials, Trial 85, ¹⁰⁶ Trial 111, ¹⁰⁷ and Trial 231, ¹⁰⁸ of patients of mostly late adolescents with moderate to severe acne. While the largest trial (Trial 85) had no losses to followup, the second largest (Trial 111) did, and both had pharmaceutical industry employees as co-authors. All three compared the same concentrations of miconazole and benzoyl peroxide, all applied twice daily.

Against non-inflammatory lesions, in large-sized Trial 85, not only did both treatments reduce the count within each arm (although P-values are not provided), but the anti-fungal combination was more effective than benzoyl peroxide. Both treatments conferred what the authors in Trial 111 called a "statistically significant" rate at 4 and 12 weeks, although no P-value was provided. In Trial 231, there was a similar response of comedones; the effect appears large, but no P-value was provided.

In the comparison against inflammatory lesions, the results were similar to those seen in non-inflammatory lesions: within arms, a statistically significant response, but no difference between arms. There were similar results for total lesions. However, although a medium-sized trial, Trial 111 is one of the few trials that stratified results by gender, and females got a better response in terms of total lesion from the combination treatment. In Trial 231, the effect of both treatments was large, but no P-values were reported.

Patients' and physicians' overall assessment in Trial 231 gave the combination statistically significantly better results than benzoyl peroxide alone. In the same trial, physician's and patients' overall assessment were numerically similar, but no P-values were reported.

The side effect of irritation was worse in the benzoyl peroxide only arm in Trial 85, about 10 percent versus 1 percent. While the outcome assessment was different in Trial 111, the conclusion was the same: that benzoyl peroxide alone caused more irritation. The rate was comparable between males and females in that trial. The side effect rates reported in Trial 231 were about 10 percent for both arms.

D.08.02-Miconazole/Benzoyl Peroxide (topical) vs. Isoconazole/Benzoyl Peroxide (topical)

There was one small split-face trial, Trial 335, ¹⁰⁹ of mostly females with mild to moderate acne. The isoconazole arm led to a statistically significant decrease in closed comedones, while the miconazole did not, but the difference between the two arms was not statistically significantly different. This pattern was the same for open comedones, papules, and pustules. The miconazole preparation led to a statistically significant decrease in pustules, and open comedones. The rates of erythema, scaling, and itching were about 50 percent in both groups. Comparison P-values were not provided.

E.01.01-Adapalene (topical) vs. Adapalene (topical)

There were three trials, Trial 78 ¹¹⁰ and Trial 338, ¹¹¹ comparing 0.1 percent versus 0.03 percent, and Trial 92, ¹¹² a split-face trial comparing immediate versus delayed adapalene. Patients' acne severity varied from mild to severe; there were moderately severely affected patients in all trials. Although statistics with respect to baseline are not available, both concentrations reduced total lesion counts. The higher concentration reduced Leeds severity more in Trial 78, although this trial had major losses to followup. The higher concentration also yielded a higher percent change in inflammatory lesion count in Trial 338. In the split-face trial, Trial 92, erythema and dryness were low and the same.

E.01.02-Adapalene (topical) vs. Tretinoin (topical)

There were 11 trials. Seven trials had two-arm design: Trial 49, ¹¹³ Trial 80, ¹¹⁰ and Trial 93 ¹¹⁴ (split-face), Trial 63, ¹¹⁵ Trial 100, ¹¹⁶ Trial 140, ¹¹⁷ and Trial 306. ¹¹⁸ These all involved the same concentration of tretinoin, 0.025 percent. There were four multi-arm trials: Trial 78, ¹¹⁰ comparing doses 0.1 percent, 0.03 percent, and tretinoin 0.025 percent with adapalene; Trial 119, ¹¹⁹ a seven-arm trial with a range of concentrations from 0.025 percent to 0.1 percent in different vehicles; Trial 338, ¹¹¹ a three-arm trial of 0.03 percent to 0.025 percent; and Trial 120, ¹¹⁹ a six-arm, split-face trial, with tretinoin concentration of 0.025 percent to 0.1 percent and vehicles (gel and cream). The total is over 1,700 patients, with about one-third in Trial 63. The patients studied varied in severity from mild to severe; all trials had moderately affected subjects. Most of the trials were demonstrably randomized.

Across outcomes, the two treatments were mostly comparable in efficacy. The effect was about 50 percent reduction in total lesions, requiring between 4 weeks (Trial 78) and 12 weeks (Trial 338) to achieve this effect. Adapalene had greater intra-patient percent change in non-inflammatory lesions in one trial (Trial 306) and in inflammatory lesions in two (Trial 306 and Trial 338).

Adapalene fared better in terms of side effects. Tretinoin had more patients with desquamation (a score of 1 versus adapalene's 0) and slightly more total adverse skin symptoms (Trial 93), more scaling (Trial 306, Trial 80) or dryness (Trial 80) or "tightness" (Trial 49), more burning (Trial 80, Trial 49) at 2 and at 12 weeks, more "adverse reactions" (Trial 63), and more patients discontinuing therapy because of adverse effects. In one trial, adapalene caused more burning or itching (Trial 100). In a split-face trial (Trial 93), patients were more often intolerant of tretinoin; and of patients asked their preferences after trial completion, more favored adapalene or the method of application (P = 0.001), although patients were not blinded and the trial was funded by adapalene's producers.

E.02.01-Isotretinoin (topical) vs. Inert (topical)

There were three trials, Trial 54, ¹²⁰ Trial 173, ⁸⁷ and Trial 202 ¹²¹ (the last two being three-arm trials), of patients with mostly mild to moderate severity. Topical isotretinoin proved to be statistically significantly more effective (although the authors make the claim without providing the actual significance levels) in terms of Cook or Leeds severity change, inflammatory lesions, and non-inflammatory lesions, starting at 4 weeks and extending to 14 weeks. In Trial 202, there was no statistical difference with respect to papules or to comedones. The effect was on the order of 40 percent change in severity or lesion count. In terms of adverse effects, there was no statistically significant difference between the therapies.

E.02.02-Isotretinoin (topical) vs. Benzoyl Peroxide (topical)

There was one comparison, Trial 173, ⁸⁷ two arms of a three-arm trial. In this trial of 77 mild and moderately affected patients, benzoyl peroxide conferred mildly better improvement (P = 0.01). For both treatments, Leeds severity dropped to a mean of 1 after 4 weeks, and 0 after 12 weeks. Both inflammatory and non-inflammatory lesions were improved. The improvement in counts was on the order of 20 percent for both. Local reactions (roughness) were few and similar.

E.02.03-Isotretinoin 0.1 percent (topical) vs. Isotretinoin 0.05 percent (topical)

There was one medium trial, Trial 202. ¹²¹ Each concentration had a statistically significant effect on papule and comedo counts, and that effect was statistically equivalent between the arms. Side effects were limited to one patient dropping out due to erythema in the 0.05 percent group.

E.02.04-Isotretinoin (topical) vs. Tretinoin (topical)

There were two small trials, Trial 90 ¹²² and Trial 98, ¹²³ of patients with moderate acne in Trial 90, but not categorizable in Trial 98. Trial 90 had no losses to followup, but patients and providers were not blinded adequately, while Trial 98 had arms that were not completely comparable. Decrease in counts of comedones and inflammatory lesions at 4 and 12 weeks were large and comparable. Only Trial 90 provides data on adverse reactions. The rate of side effects (irritation) was about 50 percent in isotretinoin patients, and about 65 percent (stinging, erythema, and desquamation) in the tretinoin patients. Comparison P-values were not provided.

E.03.01-Motretinide (topical) vs. Vehicle (topical)

There are two trials: Trial 61, ¹²⁴ a small trial with major losses to followup, noncomparability of the arms, and the execution differing, thereby, from the design; and a large trial, Trial 60, ¹²⁵ by the same authors, of mild through moderate acne in a three-arm trial. Although motretinide led to statistically significant changes with respect to baseline in Trial 61, the sample size was so small that the decrease was not statistically significantly different from the vehicle effect. In the larger trial, the subjective rating of total lesions change was not statistically significantly different from vehicle. There was a high incidence, but no statistically significant differences, in side effects (erythema, desquamation, pruritus, burning).

E.03.02-Motretinide (topical) vs. Motretinide (topical)

There was one medium trial, Trial 253, ¹²⁶ of patients with moderate through severe acne. Motretinide twice daily was compared against itself as a once daily administration. Both led to statistically significant and meaningful decreases in papules and pustules, and in overall change, but neither was better than the other. Side effects were not reported.

E.03.03-Motretinide (topical) vs. Tretinoin (topical)

There were three trials, Trial 59, ¹²⁷ Trial 253, ¹²⁶ and Trial 60, ¹²⁵ of patients with the full range of acne severity. Trial 60, being the largest, had mild through moderate degrees of severity. Trial 59 had arms comparable on a number of characteristics beyond age and severity, but the trial was weakened by major losses to followup.

The improvement in total lesion count in each arm was statistically significant in Trial 59, but not statistically significantly different between arms. Papule and pustule counts were similarly effectively reduced in Trial 253, with no statistically significant difference between arms. In Trial 60, over half the patients in the two arms had a 50 percent or more improvement, and, again, there was no statistically significant difference between the two arms.

Side effects (pruritus, burning, erythema, and desquamation) occurred in most patients in the tretinoin group in Trial 60, and statistically significantly less so, but still very often, in the motretinide group. The absolute rate was lower in Trial 59, but tretinoin's rate was still statistically significantly greater than motretinide's.

E.04.01-Retinyl Beta-glucuronide (topical) vs. Vehicle (topical)

There was one small trial, Trial 145, ¹²⁸ a split-face trial of males with mild to moderate acne using stratified randomization and no losses to followup. The retinoid conferred a statistically significant decrease in non-inflammatory lesions at 18 weeks. Most other outcomes were equivalent numerically and statistically. Adverse reactions were equivalent.

E.04.02-Retinyl Beta-glucuronide (topical) vs. Tretinoin (topical)

There was one small trial, Trial 146, ¹²⁸ including patients with mild to moderate acne. Both treatments led to a statistically significant decrease in physicians' rating of overall change. No outcome differed in a statistically significant way between the treatments. Almost all tretinoin-treated patients had erythema, drying, peeling or cracking, while the retinyl group did not.

E.05.01-Tretinoin (topical) vs. Inert (topical)

The strengths of tretinoin examined ranged from 0.025 percent to 0.1 percent, and included both gel and cream. There were six trials: five three-arm trials, Trial 58, ¹²⁹ Trial 60, ¹²⁵ Trial 198, ¹³⁰ Trial 220, ¹³¹ and Trial 219, ¹³² and one two-arm trial, Trial 270, ¹³³ comprising 10 relevant comparisons. Trial 58, Trial 60, and Trial 198 all involved a comparison with 0.05 percent; the remaining comparison concerned the weaker 0.025 percent strength. Trial 58, Trial 220, and Trial 219 were multi-center trials. Most patients had mild to moderate acne, except for patients in Trial 58, where the acne was severe, and in Trial 198, where we could not categorize the severity.

Four trials assessed global, overall change. In three (Trial 60, Trial 198, and Trial 270), the change was large and statistically significant. In Trial 219, the change was large, but not statistically significant. Most trials measured lesion counts. Total lesion counts were statistically significantly improved only in one trial (Trial 219), where there was borderline statistical significance at 12 weeks. The magnitude was fairly large (about 40 percent), however. Tretinoin led to a greater decrease in non-inflammatory lesions in all trials, although statistically significantly in only three of the four trials. Inflammatory lesions decreased as well, but fewer trials achieved a statistically significant result compared with vehicle. Only Trial 59 and Trial 270 distinguished papules from pustules, and papules were statistically significantly improved only with the higher tretinoin dose in the first trial, but both pustules and papules were improved in the second.

Side effects were frequent. In two trials (Trial 58 and Trial 60), erythema and other local symptoms were almost universal early on, but then dropped off. Local symptoms included erythema, peeling, pruritus, burning, and irritation. The frequency was almost always statistically significantly greater than in the inert control. The frequency was not insignificant, however, in the control groups. One trial (Trial 220) reported on systemic symptoms, which occurred about one-third of the time, but not statistically significantly more often in the tretinoin group.

E.05.02-Tretinoin (topical) vs. Benzoyl Peroxide (topical)

There were three trials. Trial 149 ⁹⁷ was a large five-arm trial of benzoyl peroxide 5 percent and 10 percent alone and with tretinoin in Indonesian patients where acne severity cannot be categorized. There was one medium trial, Trial 38, ¹³⁵ and one large trial, Trial 221, ⁹⁴ of patients with moderate acne. The first trial used lower strength (0.025 percent) and the second, higher (0.1 percent).

In Trial 38, there was early increase in papules in tretinoin (data abstracted but not shown in the evidence table), but statistically significant improvement in comedones, papules, and pustules by 12 weeks. The decrease was statistically significantly greater in the tretinoin group for comedones and papules, but the differences were small. Both patients and clinicians rated the tretinoin group as more improved. In Trial 221, there was over a 30 percent decrease (after 8 weeks) in the tretinoin group for papules, pustules, and, most strongly, for comedones. The benzoyl peroxide groups had a greater effect on papules, but otherwise had a statistically significant equivalent effect. Qualitative assessment of severity change was the same in both groups. In Trial 149, the amount of time to 75 percent or more improvement was not statistically significantly greater in the benzoyl peroxide group. Overall change was highest in the tretinoin/benzoyl peroxide 5 percent group, and statistically significantly greater than the other groups.

Side effects (erythema, irritation, soreness) were few in Trial 38 and the same in the two arms. In Trial 221, the rate of these side effects was about 20 percent and not statistically significantly different for the two arms.

E.05.03-Tretinoin (topical) vs. Adapalene (topical)

See E.01.02.

E.05.04-Tretinoin (topical) vs. Azelaic acid (topical)

There were two large trials, Trial 188 ³² and Trial 316. ¹³⁶ Acne severity was mild in the second trial and could not be categorized in the first. Trial 188 used a higher strength of tretinoin, and treatment continued for 6 months.

Change in non-inflammatory lesions was comparable in the tretinoin arms of the two trials, although neither within- or between-arm P-values are supplied. Change in papules was greater and statistically significantly so in the azelaic group in Trial 316.

In Trial 188 (higher strength, longer duration of treatment), side effects of erythema and scaling occurred in about 30 percent of subjects, and were statistically significantly greater than in the azelaic (about 20 percent) group. In Trial 316, the same side effects occurred and statistically significantly more frequently in the tretinoin than in the azelaic group.

E.05.05-Tretinoin (topical) vs. Isotretinoin (topical)

See E.02.04..

E.05.06-Tretinoin (topical) vs. Retinyl Beta-glucuronide (topical)

See E.04.02..

E.05.07-Tretinoin (topical) vs. Tretinoin (topical)

There were eight trials. Three trials were large three-arm trials: Trial 58, ¹²⁹ Trial 220, ¹³¹ and Trial 219. ¹³² There were two medium three-arm trials, Trial 84 ¹³⁴ and Trial 198. ¹³⁰ There were three small multi-arm trials: Trial 119 ¹¹⁹ (seven arms), Trial 120 ¹¹⁹ (six arms), and Trial 302 ¹³⁷ (four arms). Trial 119, Trial 120, and Trial 302 were very small (less than 10 patients per arm) and very short (3-week) multi-arm trials. Trial 58 and Trial 220 were multi-center trials. One-quarter of the patients had severe acne (Trial 58); one-third had mild acne (Trial 220); one-quarter had mild to moderate acne (Trial 219); and for the remainder, severity could not be categorized. In fact, not all subjects in Trial 119 and Trial 120 had acne, since these were trials to evaluate tolerability and not effectiveness. Pharmaceutical employees were co-authors in these two trials. Trial 302 suffered from major protocol departures, although losses to followup were minor. Trial 84 reported comparable improvement, but no P-values, between the two strengths of tretinoin. The strengths of tretinoin examined ranged from 0.025 percent gel through 0.1 percent cream.

Two trials evaluated overall change. In Trial 198, both lower and higher dose tretinoin conferred improvement or marked improvement that was not statistically significant between the two groups. In Trial 219, the addition of polyolprepolymer-2 did not increase overall improvement. The three trials that reported total lesion count change (Trial 58, Trial 198, and Trial 220) showed no statistically significantly different change across tretinoin strengths. Four trials reported on non-inflammatory lesions specifically (Trial 58, Trial 198, Trial 220, and Trial 302). One trial (Trial 58) yielded a statistically greater decrease from the 0.05 percent preparation compared with the 0.025 percent, but the absolute difference was small. The same trials reported on inflammatory lesions. Only two trials demonstrated statistically significant differences in efficacy, and the magnitude of these differences was small.

All trials reported on side effects. There was a strength-response relationship demonstrated in Trial 58, in Trial 119, and in Trial 120. Side effects from 0.1 percent strength were so frequent and severe that most subjects stopped using the treatment. Side effects noted included erythema, peeling, and irritation. The addition of polyolprepolymer-2 in Trial 219 did not improve the side effect rate.

E.05.08-Tretinoin (topical) vs. Neomycin/Methylprednisolone (topical)

There was one medium multi-center, split-face trial, Trial 13, ¹³⁸ of patients with the full range of acne severity. Although 18 centers were involved, the authors reported no details regarding baseline characteristics, beyond gender and age; so it is difficult to assess the generalizability of their findings. The outcomes are hard to interpret in the absence of baseline data, but the actual change scores and comedo counts appear the same in the two groups. No comparison statistics were provided. Qualitative assessment (number of patients comedo-free, inter-treatment patient preference) was the same in both groups. Side effect rates of burning, stinging, dry skin, and erythema are equal in the two groups.

E.05.09-Tretinoin (topical) vs. Chloramphenicol/Hydrocortisone (topical)

There was one medium multi-center trial, Trial 37, ¹³⁹ of patients with moderate acne. Evidence of randomization is good, but there were major losses to followup, and the paper was co-authored by a pharmaceutical employee. Patients rated statistically significantly "marked improvement" after 12 weeks in both groups, with no statistical difference between the groups. Both groups had decreases in all lesion types (including cysts), but the baseline numbers were small. Qualitative assessment of improvement was the same in both group. The authors provided no data on side effects.

E.06.01-Vitamin A (topical) vs. Inert (topical)

There were two medium trials, Trial 35 ¹⁴⁰ and Trial 34, ¹⁴¹ of patients with mild to moderate acne. Trial 34 had no losses to followup. Trial 35 used a matched-control design.

In Trial 35, there was a marked difference numerically between the groups for comedones and papules, although P-values were not provided and cannot be calculated. Subjective assessment of overall change was statistically significant. There does not appear to be a difference with respect to pustules, but the absolute number of pustules is small, meaning a larger sample size would be needed to discern a difference. This is one of the few trials to stratify findings by sex. The effect of vitamin A appears to have been different for males and females: papules were reduced in males, but not in females, while comedo counts were reduced a similar proportion (60-80 percent). Assessments of overall change were statistically significant in both strata. In Trial 34, assessed response was better in the treated group; our calculated P-value demonstrated statistical significance.

In Trial 34, every patient treated with vitamin A had erythema, while the control group had none. Trial 35 provided no side effect data.

E.06.02-Vitamin A (topical) vs. Keratolytic (topical)

There was one large three-arm trial, Trial 272 ⁹² (N=189), with benzoyl peroxide and with sulfur/resorcinol as comparators. Percent change at 14 weeks was numerically markedly better in the vitamin A arm than in either comparator arm, although P-values were not provided nor could be calculated. No side effect data were provided.

E.06.03-Vitamin A (topical) vs. Oxytetracycline (oral)

There was one large multi-center trial, Trial 57 ¹⁴² at 2 university clinics and 13 private clinics, for a total of 238 patients 11 to 44 years old, where acne severity cannot be categorized. The enroller/treater was adequately blinded, and the arms were comparable by age, sex, skin color, skin type, and number and type of acne, but there were major losses to followup. The trial compared topical vitamin A, oral oxytetracycline, and the two together.

Across the three groups, improvement in inflammatory and non-inflammatory lesions was high (over half of patients achieving over 50 percent improvement), but no differences were statistically significant. The rate of side effects was high: up to 25 percent with erythema and desquamation in the vitamin A group and about 8 percent in the antibiotic group, but the between-arm P-value addresses all side effects together and is not statistically significant.

G.01.01-Tretinoin/Erythromycin (topical) vs. Erythromycin (topical) vs.Tretinoin (topical) vs. Vehicle

There was one small four-arm trial, Trial 238 ¹⁴³ of patients with moderate to severe acne. Physicians' assessment of overall change appeared greater in the combination group, less so in each component group, and least in the vehicle group. No comparison P-values were reported. No side effect data were reported.

G.02.01-Tretinoin/Clindamycin (topical) vs. Clindamycin (topical)

There were two small trials, two-arm Trial 46 ¹⁴⁴ and three-arm Trial 283, ¹⁴⁵ and two large two-arm trials, Trial 45 ¹⁴⁴ and Trial 47, ¹⁴⁴ all of adolescents and young adults with moderate to severe acne (some patients with mild acne in Trial 283). In Trials 45, 46, and 47 (performed by the same investigators), there were no losses to followup, but inconsistency in outcome ascertainment and noncomparable arms. In Trial 283, recruitment and treatment assignment were stratified by gender.

The combination treatment resulted in statistically significant, but small, improvements in non-inflammatory, but not inflammatory, acne in Trials 45 and 47. The overall severity score was also statistically significant, but small, in these two trials. Only Trial 283 reported on side effects, which were numerically fewer in the clindamycin group but statistically different across the three arms of that trial.

G.02.02-Tretinoin/Clindamycin (topical) vs. Tretinoin (topical)

There were two large trials, Trial 43 ¹⁴⁴ and Trial 44, ¹⁴⁴ and one medium trial, Trial 42, ¹⁴⁴ all two-arm trials of moderately to severely affected adolescents to young adults. All had no losses to followup but had noncomparable arms and inconsistent ascertainment across arms.

Of severity, inflammatory, and non-inflammatory changes, inflammatory change with respect to baseline was greater in the combination group in Trials 42 and 43, but not statistically greater in Trial 44. Severity change was statistically significantly greater in the combination group only in Trial 43, at 12 weeks. Side effects were not reported.

G.02.03-Tretinoin/Clindamycin (topical) vs. Tetracycline (topical)

There was one small three-arm trial, Trial 283, ¹⁴⁵ of adolescents and young adults, stratified by gender recruitment and treatment assignment. There were no statistically significant differences in effectiveness or in side effects.

G.03.01-Tretinoin/Benzoyl Peroxide (topical) vs. Benzoyl Peroxide (topical)

See D.01.09.

H.01.01-Clindamycin (oral) vs. Inert (oral)

There was one medium trial, Trial 56, ¹⁴⁶ of adolescents with moderate acne. Changes in non-inflammatory and inflammatory lesions were large (48 percent and 75 percent decrease, respectively) and statistically significant at 6 and 13 weeks. Adverse reactions, gastrointestinal in particular, were the same; there was numerically more mild diarrhea in the clindamycin group. P-values were not provided.

H.01.02-Clindamycin (oral) vs. Tetracycline (oral)

There were two small trials, Trial 265 ¹⁴⁷ and Trial 275, ¹⁴⁸ of moderately affected patients (adolescents exclusively in Trial 265). Trial 265 was afflicted with major losses to followup, while Trial 275 made explicit its education of patients for applying the medication.

In Trial 265, efficacy was numerically equivalent for overall change, for severity, and for lesion count, at 4 and at 8 weeks, but no P-values were provided or could be calculated. In both treatments, the effect was about a 50 percent decrease in lesion counts, using group means. In Trial 275, overall change, according to patients, were equivalent for the two groups, about two-third of patients having been "much improved."

In Trial 265, esophagitis or mild diarrhea/cramping was present in equal, small numbers in both groups. In Trial 275, two patients withdrew from the clindamycin group, none from the tetracycline group.

H.02.01-Demethylchlortetracycline (oral) vs. Placebo (oral)

There was one small trial, Trial 162, ¹⁴⁹ one medium trial, Trial 192, ¹⁵⁰ and three large trials, Trial 312, ¹⁵¹ Trial 193 ¹⁵⁰ and Trial 71, ¹⁵² the last two being cross-over trials. Acne severity was moderate to severe in one trial (Trial 162), mild through severe in another (Trial 312), and not categorizable in the others. Only one trial (Trial 71) provided the gender breakdown. Although it had a cross-over design, Trial 71 had enough errors in assignment that it had, effectively, non-random assignment. Trial 162, a parallel-design trial, had too many placebo patients receive the anti-bacterial (unplanned cross-over). In two trials, the dose was 150 mg once daily; in the other two, 150 mg four times daily.

Assessment of relative improvement was statistically significantly greater for demethylchlortetracycline only in Trial 162. The difference was not statistically significant in the other four trials.

In Trial 162 (higher dose), about one-third of patients reported photosensitivity; other patients report gastrointestinal (GI) disturbance and photo-onycholysis, none of which occurred in placebo-treated patients (P=0.01). In Trial 71 (higher dose), 10 percent of patients complained of photosensitivity (versus less than 1 percent in the placebo group).

H.02.02-Demethylchlortetracycline (oral) vs. Phenethicillin (oral)

There was one medium trial, Trial 312, ¹⁵¹ of patients with mild through severe acne. Physicians' assessment of overall change was numerically and statistically the same. Similarly, patients' assessment and photo-based severity assessment were both the same in the two treatment groups.

H.02.03-Demethylchlortetracycline (oral) vs. Tetracycline (oral)

There were two large trials, Trial 70 ¹⁵³ and Trial 71, ¹⁵² of patients with non-categorizable acne severity. Trial 70 had biased ascertainment because more severely affected patients finished their course of antibiotics earlier. Treatment assignment in Trial 71 was made incorrectly in about one-quarter of the patients.

In Trial 70 (lower dose), physicians' assessment of overall change was better in the tetracycline group (P=0.01), although there were eight (about 10 percent) relapses 15 weeks after therapy in the tetracycline group. In Trial 71 (higher dose), severity change was not statistically significantly different. The actual improvement rate was higher in Trial 70 than in Trial 71.

In Trial 71 (higher dose), 10 percent of demethylchlortetracycline assigned patients complained of photosensitivity. In Trial 70, there were numerically more side effects (all GI-based) in the tetracycline group, but no statistics are provided.

H.03.01-Doxycycline (oral) vs. Placebo

There was one large cross-over trial, Trial 273, ¹⁵⁴ of patients with mild acne. The trial was limited by major losses to followup. Decreases in inflammatory counts was statistically significant in the doxycycline arm in both phases. Assessment of improvement, however, was not different for comedones or for cysts. There were zero side effects in either group.

H.03.02-Doxycycline (oral) vs. Fusidic (oral)

There was one large trial, Trial 1 ¹⁵⁵ (N=110), of late adolescents with mild-moderate acne. Patients were not blinded, and a co-author was a pharmaceutical employee. Total lesion count, at 6 and 10 weeks, dropped significantly in both groups so that the effectiveness was statistically significant within each arm, but not statistically significant across arms. Assessments of overall change, by patient and by investigator, were also not statistically different. There were numerically more side effects in the fusidic acid group (4 vs.1), but no P-values were reported.

H.03.03-Doxycycline (oral) vs. Erythromycin Stearate (oral)

There was one medium two-site trial, Trial 25, ¹⁵⁶ of young adults with moderate acne. The trial had excellent methodological control. Severity decreased to mild in both groups; P-value was not provided. Change in inflammatory lesions was large and statistically significant within each group, but not statistically different across groups.

H.03.04-Doxycycline (oral) vs. Minocycline (oral)

There was one medium trial, Trial 257, ¹⁵⁷ and two small trials, Trial 311 ¹⁵⁸ and Trial 155, ¹⁵⁹ of patients with a full range of acne severity. In Trial 257, a co-author was a pharmaceutical employee. In Trial 155 there were major losses to followup.

In Trial 311, there was no difference in severity at 12 weeks. In Trial 257 and Trial 155, there were many lesion-based outcomes. In Trial 155, investigators attempted to obtain patients' preferences on a 100 cm visual analogue scale, but the responses were too inconsistent to be usable. Although no differences were statistically significant, both treatments were effective, decreasing counts by about 50 percent by 12 weeks for both inflammatory and non-inflammatory lesions; P-values were not supplied by the authors.

Side effects (dyspepsia, headache, nausea, vertigo) were few and comparable in the two groups.

H.04.01-Erythromycin (oral) vs. Doxycycline (oral)

See H.03.03.

H.04.02-Erythromycin Base (oral) vs. Erythromycin Stearate(oral)

There were two trials: one small trial, Trial 24, ¹⁶⁰ of patients with moderate acne over only 2 weeks, and one small seven-arm trial, Trial 340, ¹⁶¹ of patients with moderate to severe acne treated between 6 and 18 months. In Trial 340, preference over placebo was evaluated, but difficult to interpret. In Trial 24, only papule counts were reported. The counts were significantly (and statistically) decreased with respect to baseline in each arm, although the stearate did statistically significantly better. Side effects (gastrointestinal distress) appeared in both groups, but the difference was not statistically significant.

H.04.03-Erythromycin (oral) vs. Tetracycline (oral)

There were four trials: two small trials (Trial 6 ¹⁶² and Trial 36 ¹⁴⁰ ), one small seven-arm trial (Trial 340 ¹⁶¹ ), and one large multi-center trial (Trial 125 ¹⁶³ ). Severity ranged from mild to moderate (Trial 125) to moderate (Trial 36), to severe (Trial 340); baseline acne severity could not be categorized in Trial 6. Trial 36 included student health centers, in addition to dermatologists' offices and academic clinics, but had a pharmaceutical co-author. Doses varied: Trial 6 compared erythromycin 250 mg twice daily with tetracycline, 250 mg twice daily; Trial 36, erythromycin acistrate 200 mg once daily versus tetracycline 250 mg once daily; and Trial 125, erythromycin 500 mg once daily versus tetracycline of unknown dose.

Trial 36 reported that, on the face, back, and chest, severity improved at 12 and 24 weeks, with a within-group P-value of 0.05, but between groups, the between-arm P-value was not statistically significant. Trial 125 provided no P-values, but numerically, the effects were the same for both treatments for comedones, papules, and pustules. Patients' subjective assessments of change were numerically similar for the two treatments as well. Trial 6 also did not present P-values but presented numerically similar assessments of severity change in the two treatments. Trial 340 presented inter-treatment comparison. Although difficult to interpret (they are based on comparison to placebo), a few more patients preferred two different types of tetracycline (phosphate, citric acid, but not novobiocin) than preferred erythromycin.

Trial 6 had about 8 percent of patients with side effects, equal in the two groups. Trial 36 had up to 20 percent side effects, again equal in the two groups numerically. In Trial 125, there were again about 8 percent side effects. Numerically, more patients dropped out of erythromycin due to side effects, but these are probably statistically not significant.

H.04.04-Erythromycin (oral) vs. Triacetyl-oleandomycin (oral)

There was one small seven-arm trial, Trial 340, ¹⁶¹ of patients with moderate to severe acne treated between 6 and 18 months. Preference over placebo was evaluated, but difficult to interpret. More patients preferred triacetyl-oleandomycin over placebo than preferred erythromycin over placebo. No side effect data were reported.

H.04.05-Roxithromycin 150 mg (oral) vs. Roxithromycin 300 mg (oral)

There was one small trial, Trial 4, ¹⁶⁴ where acne severity cannot be categorized. A single outcome, severity change, was high for both groups, and not statistically significantly different. No side effect data were reported.

H.04.06-Oleandomycin (oral) vs. Ttriacetyl-oleandomycin (oral)

There is one large trial, Trial 247 ¹⁶⁵ (N=142), of patients where acne severity cannot be categorized. The authors reported only one outcome, physicians' perception of overall change, which was "excellent" in about one-quarter of cases in each arm, and not statistically significantly different. No data on side effects were reported.

H.05.01-Minocycline (oral) vs. Placebo (oral)

There was one medium cross-over trial, Trial 161, ¹⁶⁶ and one large cross-over trial, Trial 273. ¹⁵⁴ Acne severity was mild in Trial 273, and could not be categorized in Trial 161. The high-dose trial (Trial 273) was weakened by major losses to followup.

Acne severity in Trial 161 improved within the minocycline arm to a statistically significant extent and it improved, but not to a statistically significant extent in the placebo arm, in both phases of the cross over. However, a between-arm P-value was not provided. Inflammatory lesion counts in Trial 273 are improved by minocycline to a statistically significant extent in both phases. Non-inflammatory lesions, nodules, and cysts were improved, but not apparently more than by placebo; between-arm P-values were not provided.

There were two side effects (vertigo and urticaria) leading to patient withdrawals in Trial 161, and zero side effects reported in Trial 273.

H.05.02-Minocycline (oral) vs. Clindamycin (topical)

There were two medium trials, Trial 268 ¹⁶⁷ and Trial 307, ¹⁶⁸ of patients with moderate to severe acne. Trial 268 was a multi-center trial of student health centers that used balanced, block randomization. Trial 307 was a matched-control trial.

Overall assessment at 4 and 12 weeks was equivalent between the two groups in Trial 268, and was highly statistically significant within each group. Severity was not statistically significantly different between arms in Trial 307. Lesion-count decreases were equivalent between the two groups in both trials. Only Trial 268 documented statistically significant decreases within each group.

Side effects (six patients with allergy, diarrhea, impetiginization, dryness, and stinging) were few and similar between the groups.

H.05.03-Minocycline (oral) vs. Doxycycline (oral)

See H.03.04.

H.05.04-Minocycline (oral) vs. Tetracycline (oral)

There were two large two-arm trials, Trial 72 ¹⁶⁹ and Trial 171, ¹⁷⁰ and one medium trial, Trial 298, ¹⁷¹ of patients with moderate to severe acne. Trials 72 and 298 evidenced good blinding, and in Trial 298, agreement in assessment among patients, reviewers, and investigator was measured and was high (between 60 and 86 percent). Trial 171 had major losses to followup and inconsistent ascertainment across arms.

Severity change was not statistically significantly different between the two arms in Trial 298 or Trial 171. Lesion counts were improved by both treatments in Trial 72, having reduced comedones, papules, and pustules about 30 percent by 19 weeks; P-values were not provided.

Side effects (nine total in Trial 298) were not statistically significantly different between the arms, including GI and vaginal symptoms. There was one case of hives in the minocycline group in Trial 171.

H.06.01-Oxytetracycline (oral) vs. Placebo (oral)

There was one large cross-over trial, Trial 250 ¹⁷² (N=76), in patients with moderate to severe acne. There were major losses to followup and there were irregularities in execution. Severity change favored oxytetracycline in both phases of the cross-over (by our calculated chi-square test). Over half the patients had an "excellent" response in the oxytetracycline arm. No side effect data were reported.

H.06.02-Oxytetracycline (oral) vs. Benzoyl Peroxide (topical)

There was one small three-arm trial, Trial 255, ⁸⁴ of patients with moderate acne. Data were provided on many outcomes on the face, chest, and back. For oxytetracycline, many of these results were statistically significant within its arm. Although the benzoyl peroxide arm had no P-values listed, the numerical values were similar to those in the anti-bacterial group. The differences between the two arms were not statistically significant. There were six side effects reported (GI symptoms, severe facial reaction, vaginal candidiasis).

H.06.03-Oxytetracycline (oral) vs. Benzoyl Peroxide/Metronidazole (topical)

There was one medium trial, Trial 123, ¹⁰⁵ of young adults with moderate to severe acne. Assessments of overall change by both patient and physician were about the same in the two groups (about one-third "much improved") and not statistically significantly different. No side effect data were reported.

H.06.04-Oxytetracycline (oral) vs. Vitamin A (topical)

See E.06.03.

H.06.05-Oxytetracycline (oral) vs. Tetracycline (oral)

There was one small three-arm trial, Trial 255, ⁸⁴ of patients with moderate acne. Both treatments reduced face severity and comedonal count with statistical significance. Both also reduced the inflammatory lesion count, although only tetracycline's effect reached statistical significance, but the difference between the two groups was not statistically significant. There were four side effects (GI symptoms).

H.06.06-Oxytetracycline (oral) vs. Cotrimoxazole (oral)

There was one small matched-control trial, Trial 152, ¹⁷³ of only 16 patients with mild through severe acne. There was no statistically significantly different preference for one treatment over the other. There was one report of side effects (flatulence/abdominal pain).

H.06.07-Oxytetracycline (oral) vs. Trimethoprim (oral)

There was one small trial, Trial 128. ¹⁷⁴ Overall change was mostly "fair," but not statistically significantly different between arms. Lesion counts were statistically significantly reduced in both arms, with no statistically significant difference between the arms. Side effects (five patients) reported were not statistically significantly different between arms.

H.07.01-Cotrimoxazole (oral) vs. Vehicle (oral)

There were two medium cross-over trials, Trial 160 ¹⁷⁵ and Trial 223. ¹⁷⁶ Neither trial specified the patients' acne severity, and Trial 223 did not provide a sex and age breakdown, but did provide evidence of good blinding. In Trial 160, improvement in severity was large and statistically significant, whether treated first or second with the antibiotic. There was no improvement from the vehicle. No between-arm P-value was provided. In Trial 223, the way that the lesion data were reported precludes our making a comparison. However, that trial provided some side effect data (dizziness, headaches, red macular rash), which were too few to make inferences.

H.07.02-Sulfadimethoxine (oral) vs. Placebo (oral)

There was one large cross-over trial, Trial 44, ¹⁷⁷ of patients with severe acne. There were no losses to followup. The only outcome is physicians' preference across the two treatment phases, comparing treatments within patients. The P-value we calculated for the authors (based on chi-square) is 0.001. There were no data on side effects.

H.08.01-Tetracycline (oral) vs. Inert (oral)

There were nine trials of tetracycline 250 mg at different dosing schedules. Trials of twice daily administration were: Trial 69, ¹⁷⁸ Trial 201, ¹⁷⁹ Trial 10, ⁸⁰ Trial 33, ⁴⁴ Trial 23, ⁸¹ Trial 313, ⁸³ and Trial 136, ¹⁸⁰ the last five being three-arm trials. Trials of four-times-daily administration were Trial 71 ¹⁵² and Trial 349, ¹⁸¹ a four-arm trial. Six trials gave baseline severity information: Trial 33 studied mildly affected patients; Trial 10, Trial 23, Trial 249, and Trial 136 studied moderately affected subjects; Trial 313 patients had mild to moderate acne. Of the trials that provided ages, all were late adolescent/young adult. Trial 71 and Trial 313 were generalizable to college or local populations of young adults. Trial 23 and Trial 349 were properly randomized and Trial 201 had proper blinding. Trial 71 was a cross-over trial; but because of errors in randomization, almost all subjects received tetracycline first. Trial 23 had a pharmaceutical employee as a co-author.

For outcomes of relative improvement, and in all nine trials, tetracycline conferred statistically significantly greater improvement than placebo as early as 6 weeks and as late as 13. The magnitude of the improvement cannot be summarized from the data reported. Trial 33 provides some lesion count data: by 8 weeks, the decrease in inflammatory lesions was statistically significantly different from placebo.

Five trials (Trial 33, Trial 69, Trial 201, Trial 349, and Trial 136) reported on side effects; the total number of affected patients was 15.

H.08.02-Tetracycline (oral) vs. Clindamycin (topical)

There were five trials: Trial 29, ¹⁸² Trial 33, ⁴⁴ Trial 186, ¹⁸³ Trial 318 ¹⁸⁴ and Trial 136, ¹⁸⁰ of patients with mild, moderate, unspecified (two trials), and moderate to severe acne. Trial 136 was a three-arm trial. Trials 29 and 33 both used complementary placebos (topical and oral) to ensure patient blinding. Trial 186 suffered from major losses to followup.

Trial 186 resulted in equal numbers, and statistically nonsignificant differences, in many lesion-based outcomes, ranging from comedones to cysts. In Trial 136, papule and pustule counts were reduced equally, but physicians rated clindamycin as leading to statistically significantly greater improvement than tetracycline, while patients rated themselves as improved to the same degree. In Trial 33, there were statistically significant changes within each group. Clindamycin conferred a statistically significantly greater decrease in inflammatory lesions at 8 weeks, but conferred no advantage in Trial 29 at 4 weeks. Overall changes, as determined by patients or physicians, were better in the tetracycline group, although P-values were not provided. In Trial 318, there was no statistically significant difference in pustules counts.

Side effects included diarrhea in both groups (Trial 33) or erythema (Trial 186), for a total of four cases.

H.08.03-Tetracycline (oral) vs. Erythromycin (topical)

There was one medium trial, Trial 279, ¹⁸⁵ of patients with moderate to severe acne. Overall assessment was that patients were improved in both arms, with no statistically significant difference. Counts of many lesion types were presented. For many of them, within-arm improvement was statistically significant, but the difference between arms was not statistically significant.

Two patients withdrew because of adverse reactions.

H.08.04-Tetracycline (oral) vs. Tetracycline (topical)

There were two small trials and one medium trial, Trial 10, ⁸⁰ Trial 23, ⁸¹ and Trial 313, ⁸³ all three-arm trials of mild to moderately affected patients. Trial 23 reported evidence of randomization but had a pharmaceutical-company as co-author, and Trial 313 was generalizable to the local population.

Only relative improvement outcomes were provided. Patients improved in all arms, with a statistically significant difference between oral and topical treatment in only one trial, Trial 23.

No side effect data were reported.

H.08.05-Tetracycline (oral) vs. Azelaic Acid

There was one small trial, Trial 22, ¹⁸⁶ and two large trials, Trial 167 ¹⁸⁷ and Trial 166, ¹⁸⁷ of patients with mild through severe acne. Trials 167 and 166 were adequately randomized. Trial 22 suffered from major losses to followup.

Relative improvement was "excellent" or "good" in the majority of patients in both arms in Trial 167 and Trial 166. Lesion counts were reduced in both arms in Trial 22 and Trial 166. Only the former, the smallest trial, found a statistically significant difference between the treatments in one comparison of lesion counts, out of a total of nine count-based outcomes.

A single case of abdominal pain, in Trial 22, was reported.

H.08.06-Tetracycline (oral) vs. Meclocycline/Sulfosalicylate (topical)

There was one medium trial, Trial 164, ¹⁸⁸ of young adults with mild to moderate acne. Of the many outcomes assessed, both treatments reduced all lesions about 70 percent. Tetracycline was statistically significantly better in treating open comedones. Otherwise, the reduction was the same in the two groups. The authors report that there were no side effects, but they provide no specific data.

H.08.07-Tetracycline (oral) vs. Clindamycin (oral)

See H.01.02.

H.08.08-Tetracycline (oral) vs. Tetracycline (oral)

There were two medium trials that compared different regimens of tetracycline. Trial 212, ¹⁸⁹ where acne severity could not be discerned, compared tetracycline sustained release 250 mg once daily with tetracycline 250 mg twice daily in young adults. The trial was limited by major losses to followup. Trial 317, ¹⁹⁰ of severely affected patients, compared tetracycline 250 mg four times daily with a similar tetracyline regimen plus proteolytic enzymes in severely affected patients.

In Trial 212, two-thirds of patients were improved at 12 weeks, but there was no statistically significant difference between treatments. In Trial 317, patients were much improved in both groups at 12 weeks, but with no statistically significant difference between treatments.

Side effects were few (diarrhea [1] and sore mouth [1]) in Trial 212.

H.08.09-Tetracycline (oral) vs. Isotretinoin (oral)

See K.02.03.

H.08.10-Tetracycline (oral) vs. Cyproterone/Ethinyl Estradiol (oral)

See M.01.03.

H.08.11-Tetracycline (oral) vs. Gugulipid (oral)

See N.01.02.

H.08.12-Tetracycline (oral) vs. Ocimum Basilicum (topical)

See N.01.04.

H.08.13-Tetracycline (oral) vs. Zinc (oral)

See N.07.03.

H.08.14-Tetracycline (oral)/Povidone-iodine (topical) vs. Tetracycline (oral)/Vehicle (topical)

There was one small trial, Trial 236, ²⁹ of moderately to severely affected patients notable by the absence of treatment information regarding povidone-iodine and by major losses to followup. No comparison statistics were provided, but outcomes were comparable. There were zero side effects reported.

H.09.01-Dapsone (oral) vs. Placebo (oral)

Although dapsone is an anti-microbial, it does not have activity against P. acnes.

There was one small trial, Trial 293. ¹⁹² There were no losses to followup. There was a statistically significant difference in severity change favoring dapsone. This is one of the few trials to stratify results by gender. The results were about the same for each sex, although P-values were not reported. Side effects were numerically greater in dapsone, but not statistically significantly different.

H.09.02-Phenethicillin (oral) vs. Placebo (oral)

There was one medium trial, Trial 312, ¹⁵¹ of mixed acne severity, where patients were not blinded. Physicians' assessment of overall change was that most patients were judged to be unchanged, and overall change for the treatment arm was not statistically significantly different from control.

H.09.03-Phenethicillin (oral) vs. Demethylchlortetracycline (oral)

See H.02.02.

H.09.04-Sodium Fusidate (oral) vs. Sodium Fusidate (topical)

There was one medium trial, Trial 344, ¹⁹³ of all levels of acne severity. Treatment assignment was not properly randomized and patients and care provider were not blinded. There was a blinded outcome assessment made, however. The numbers in each improvement category were similar, although no between-arm P-values were provided. There was about a 10 percent side effect rate in the oral group and zero percent in the topical, but numbers were small.

H.09.05-Tetracycline/Novobiocin (oral) vs. Placebo (oral)

There were two trials: one small trial, Trial 279, ¹⁹¹ and one small seven-arm trial, Trial 340, ¹⁶¹ with treatment between 6 and 18 months; both trials had patients with moderate to severe acne. There were no losses to followup in Trial 279 and there was well-executed blinding. On physicians' evaluation, the antibiotic group improved statistically significantly more than the placebo group. Side effects were comparable but infrequent. In Trial 340, the outcome was inter-treatment comparison against placebo. The tetracycline-novobiocin arm had the fewest number of patients rating it as the preferred treatment.

I.01.01-Cotrimoxazole (oral)/Sulfur/Resorcinol vs. Cotrimoxazole (oral)/Placebo

There was one large multi-center, three-arm trial, Trial 126 (N=211). ¹⁹⁴ Methodological issues included 22 percent of enrolled patients not randomized, subjects' baseline acne severity not ascertainable, major losses to followup, and patients and outcome assessors were not blinded. There were many outcomes, including severity assessments and lesion counts, showing no statistically significant difference between the groups. Erythema appeared more frequent (about one-third) in the treatment group, but a P-value was not provided.

I.01.02.-Cotrimoxazole (oral)/Sulfur/Resorcinol vs. Cotrimoxazole (oral)/Tretinoin

There was one large multi-center, three-arm trial, Trial 126 (N=211). ¹⁹⁴ Methodological issues included 22 percent of enrolled patients not randomized, subjects' baseline acne severity not ascertainable, major losses to followup, and patients and outcome assessors not blinded. All efficacy differences (severity change and lesion counts) had P-values of 0.001. Desquamation and erythema appeared comparable, but P-values were not provided.

I.02.01-Tetracycline (oral)/Sulfur/Salicylic Acid/Polyethylene Scrub vs.Tretinoin/Benzoyl Peroxide/Water Avoidance

There was one medium three-arm trial, Trial 323, ¹⁹⁵ that included many patients of a narrow age range, but a wide range of acne severity. The report shows evidence of good randomization, blinding, and no losses to followup. There were many efficacy outcomes. The tetracycline/keratolytic combination achieved significant within-arm P-values for total lesion change, for comedo change, and for papule+pustule change. The tretinoin/keratolytic combination achieved a significant within-arm P-value for comedo change. The between-arm P-values were not significant, however. Ratings of side effects were low, but standard deviations were not provided to enable a comparison between the arms.

J.01.01-Cotrimoxazole (oral)/Tretinoin (topical) vs. Cotrimoxazole (oral)/Placebo (topical)

There was one large multi-center, three-arm trial, Trial 126 (N=211), ¹⁹⁴ where acne severity cannot be categorized. Only 78 percent of recruited patients were enrolled, patients were not blinded, outcome assessor was not blinded, and there were major losses to followup, as well as other major protocol departures. The outcomes were discretized improvement in pustule counts (statistically significant difference in favor of the combination therapy), in papule counts (not statistically significant), and in severity (statistically significant). Side effects of erythema and desquamation appear comparable in the two groups, although comparison statistics are not provided.

J.01.02-Oxytetracycline (oral)/Vitamin A (topical) vs. Oxytetracycline (oral)/Placebo (topical) vs. Vitamin A (topical)/Placebo (oral)

There was one large multi-center, three-arm trial, Trial 57, ¹⁹⁶ where the subjects in arms were similar along many dimensions, but where there were major losses to followup (N=238). Most patients were judged to have had 75-100 percent reduction in lesions, but P-values were not reported. No comparison between arms reached statistical significance. Over two-thirds of patients had local side effects in each group. The difference between arms is not statistically significant.

J.01.03-Tetracycline (oral)/Tretinoin (topical)/Emollient vs. Tetracycline (oral)/Sulfur/Salicylic Acid/Polyethylene Scrub (topical) vs. Tretinoin/Benzoyl Peroxide/Water

There was one three-arm trial, Trial 323, ¹⁹⁵ that included many patients of a narrow age range, but a wide range of acne severity. The report shows evidence of good randomization, blinding, and no losses to followup. All arms yielded changes in comedones and inflammatory lesions of about 40 percent. The tretinoin/benzoyl peroxide arm resulted in decreases that were statistically significantly larger. No side effects data were reported.

K.01.01-Alitretinoin (oral) vs. Isotretinoin (oral)

See K.02.04.

K.02.01-Isotretinoin (oral) vs. Placebo (oral)

There was one small trial, Trial 269, ¹⁹⁷ that compared isotretinoin to placebo in patients with severe acne. Patients in the isotretinoin arm had a decrease in the number of cysts while patients in the placebo arm had an increase in the number of cysts. This difference was statistically significant. Side effects of cheilitis (1), dermatitis (1), and arthralgia (1) were reported in the treatment arm.

K.02.02-Isotretinoin (oral) vs. Dapsone (oral)

There was one small trial, Trial 276, ¹⁹⁸ which evaluated males with severe acne. In the isotretinoin arm, there was statistically significant overall change as rated by the patient on a visual analogue scale beginning at 4 weeks of followup. Improvement in overall severity was statistically significant beginning at week 8. Statistically significant improvement in chest/back and face/neck were seen at week 16. All of these improvements persisted post-therapy. In the dapsone arm, there was some statistically significant improvement over baseline as rated by the patient in overall change and in face/neck count beginning at week 16. The isotretinoin arm had more improvement than the dapsone arm in the post-therapy face/neck count, though P-values for comparison between the treatment arms were not always provided. Other comparisons such as 16-week chest/back lesion count and total severity were not statistically significant. Cheilitis was reported in 17 patients in the isotretinoin arm but none in the dapsone arm.

K.02.03-Isotretinoin (oral) vs. Tetracycline Hydrochloride (oral)

There was one small trial, Trial 206, ¹⁹⁹ that evaluated patients with severe acne. This trial had major losses to followup. Patients in both treatment arms had improvement in the number of cysts, cyst diameter, and comedones and pustules present, though no P-values were provided. There were no statistically significant differences between the arms at 12 weeks; however, at post-therapy followup, the patients in the isotretinoin arm had statistically significantly greater improvement in cyst count, cyst diameter, and comedo and pustule count than patients in the tetracycline arm. More local side effects were reported in the isotretinoin arm (70 vs. 11) and more systemic side effects were reported in the isotretinoin arm (9 vs. 6), although no P-values were provided.

K.02.04-Isotretinoin (oral) vs. Alitretinoin (oral)

There was one small trial, Trial 127. ²⁰⁰ In this trial, males received either alitretinoin or isotretinoin. Patients in both treatment arms reported side effects including abdominal pain or diarrhea (2 vs. 0), dryness of various membranes (20 vs. 23), headache (5 vs. 1) and red face (6 vs. 10). Overall, similar numbers of side effects were reported, although no P-values were provided.

K.02.05-Isotretinoin vs. Etretinate

There is one medium trial, Trial 130, ²⁰¹ of patients with severe acne. While patients in both arms showed improvement, isotretinoin was statistically significantly more effective in reducing inflammatory nodules at the end of treatment and 8 weeks of post-therapy than was etretinate. Improvement for both treatments was notable on the face more than on the chest or back. Side effects of cheilitis were noted in all patients, and conjunctival injection was noted in 25 percent of patients in each treatment group.

K.02.06-Isotretinoin (oral) vs. Isotretinoin (oral)

There were six trials. Trial 5 ²⁰² was a small trial of patients with severe acne, while Trial 185 ²⁰³ was a medium trial of patients with moderate acne, Trial 332 ²⁰⁴ was a medium trial of patients with severe acne, and Trial 320 ²⁰⁵ was a large multi-center trial of patients with severe acne. Trial 106 ²⁰⁶ was a small three-arm trial as was Trial 190. ²⁰⁷ Examined together, these trials evaluate isotretinoin at different doses, comparing 0.1mg/kg to 0.3, 0.5 and 1.0 mg/kg, and comparing 0.3 to 0.5 and 0.5 to 1.0 mg/kg. In Trials 5, 185, and 332, the higher dose of isotretinoin was associated with either more patients rated as "very good" or "good" improvement, or with greater percent change in non-inflammatory lesions. However, in Trials 320 and 332, there was statistically significant improvement in all treatment groups, with no significant differences in improvement between the different dosage groups.

In Trials 5, 106, 185, and 332, the higher doses were also associated with more side effects, with cheilitis, facial dermatitis, desquamation and epistaxis being the most common complaints. Taken together, these side effects were reported at least 14 times per arm in Trial 5. In Trial 320, there were minor differences in the side effect profiles at different doses, but 42 percent of patients given the lowest dose of 0.1 mg/kg/day required treatment. Similarly, a dose response for side effects was suggested by Trial 106, but P-values were not provided. More systemic side effects were noted at the 1.0 mg/kg dose. No P-values were presented.

K.03.01-Vitamin A (oral) vs. Placebo (oral)

There was one medium trial, Trial 12, ²⁰⁸ comparing vitamin A to placebo. Similar numbers of patients were rated "improved" (34 vs. 33) and "no change" (23 vs. 17), with more patients being identified as "worse" in the placebo arm (3 vs. 22). No P-values were presented. No side effect data were presented.

K.03.02-Vitamin A/Zinc Sulfate (oral)

There was one small four-arm trial, Trial 233, ²⁰⁹ that evaluated vitamin A palmitate (oral), zinc sulfate (oral), their combination, and placebo in patients with moderate to severe acne. Severity change was numerically similar in the vehicle and vitamin A groups, and greater (and equal) in the zinc and combination group. No comparison statistics were provided. There were no patients with dry skin in the vehicle and vitamin A group, and four each in the other two groups. No comparison P-values were available.

M.01.01-Cyproterone (topical) vs. Vehicle (topical)

There was one small trial, Trial 141. ²¹⁰ Women with mild and moderate acne were evaluated. Neither the patient nor the outcome assessor was blinded. At 12 weeks, there was improvement in the face severity, total lesion count and total lesion percent change compared to baseline but no within-arm P-values were provided. There were statistically significant differences between the treatment arm and the vehicle arm in terms of severity change and total lesions percent change. No side effect data were provided.

M.01.02-Cyproterone/Ethinyl Estradiol (oral) vs. Minocycline Hydrochloride (oral)

There was one medium trial, Trial 245, ²¹¹ that evaluated women with mild or moderate acne. This trial was not blinded and had major losses to followup. Both treatments demonstrated statistically significant reductions in the papule, pustule, and comedo mean counts at week 24 compared to baseline. No P-values for between group comparisons were provided. Side effects of nausea, vomiting, giddiness, bleeding problems, and headaches were reported in both arms. In addition, diarrhea, heartburn, and depression were reported in the minocycline arm. Differences in the number of patients reporting side effects were not statistically significant.

M.01.03-Cyproterone/Ethinyl Estradiol (oral) vs. Tetracycline/Placebo (oral) vs.Cyproterone/Ethinyl Estradiol/Tetracycline (oral)

There was one medium trial of female patients, Trial 137, ²¹² a three-arm trial of cyproterone/ethinyl estradiol with tetracycline or the tetracycline placebo versus tetracycline (with placebo for birth control pills). This trial had major losses to followup. In all arms, there were statistically significant improvements in the inflammatory lesion change, total severity change, and face severity change at all time points. The improvement in non-inflammatory lesions was numerically different, but not statistically significantly different. In the tetracycline arm, there was statistically significant improvement in all of the outcome measures over baseline. At week 24, there was no statistically significant difference between the arms in terms of inflammatory lesion change, but the cyproterone/ethinyl estradiol arm shows statistically significantly greater improvement in total severity. Side effects of hypertension (1) and dysmenorrhea (1) were reported in the cyproterone/ethinyl estradiol arm. Side effects of chest pain (1), headache (1), menstrual irregularity (1), and breast tenderness (1) were reported in the cyproterone/ethinyl estradiol/tetracycline arm. No side effects were reported in the tetracycline arm.

M.01.04-Cyproterone/Ethinyl Estradiol (0.035mg) (oral) vs. Cyproterone(topical)

There was one small trial, Trial 141. ²¹⁰ Women with mild and moderate acne were evaluated. Neither the patient nor the outcome assessor was blinded. At week 12, both arms showed improvement in face severity, total lesion count and percent change, although no P-values were provided for comparison to baseline. There were statistically significant differences between the topical cyproterone arm and the cyproterone/ethinyl estradiol arm in terms of total severity change and Leeds severity score, with the topical cyproterone showing greater improvement. However, there were no statistically significant differences between the arms in total lesion count. No side effect data were provided.

M.01.05-Cyproterone/Ethinyl Estradiol (0.05mg) (oral) vs. Cyproterone/Ethinyl Estradiol (0.035mg) (oral)

There were three medium and large trials which evaluated the combination of cyproterone/ethinyl estradiol in various strengths in women: Trial 48, ²¹³ Trial 66, ²¹⁴ and Trial 117. ²¹⁵ The studies evaluated the combination of 2 mg cyproterone with 0.035 mg vs. 0.050 mg of ethinyl estradiol in women with moderate to severe acne. The trials demonstrated statistically significant improvement in severity within both arms with respect to baseline. No statistically significant difference between the arms was found. Systemic side effects were noted once in 22 patients in each arm, and more than once in at least 10 patients in each arm, although no P-values were provided.

M.01.06-Cyproterone (2mg)/Ethinyl Estradiol (oral) vs. Cyproterone (0.05mg)/Ethinyl Estradiol (oral)

There was one medium trial, Trial 237, ²¹⁶ of female patients with mild to moderate acne. The trial demonstrated statistically significant improvement in face papule count, face severity, and total severity by week 26. There was no numerical difference between the two arms in papule or nodule count, but there was a numerical difference in face severity, with the high cyproterone dose demonstrating greater improvement. P-values between the two arms, however, were not available and could not be calculated.

M.01.07-Cyproterone/Ethinyl Estradiol (oral) vs. Desogestrel/Ethinyl Estradiol (oral)

There was one small trial, Trial 55, ²¹⁷ and two large trials, Trial 87 ²¹⁸ and Trial 103 ²¹⁹ (all women), some with mild acne (Trial 55) and the remainder unspecified, that compared cyproterone/ethinyl estradiol to desogestrel/ethinyl estradiol at different doses. In Trial 55, there was no blinding. The trial was administered at two sites. Trial 87 had major losses to followup. In both trials, both treatment arms demonstrate some improvement in severity of acne. By long term followup, this improvement is statistically significant for both cyproterone sites in the cyproterone arm, but only for one site in the desogestrel arm. P-values were not provided in Trial 87. Comparing the two treatment arms, there was statistically greater improvement in the cyproterone arm at one site, but no difference between the arms at the other site or in Trial 87. No side effect data were presented.

M.01.08-Cyproterone/Ethinyl Estradiol (oral) vs. Levonorgestrel/Ethinyl Estradiol (oral)

There was one large trial of female patients with unspecified acne severity, Trial 48 ²¹³ (N=133). It compared cyproterone with low-dose and high-dose ethinyl estradiol to levonorgestrel/ethinyl estradiol. Results show that both cyproterone/ethinyl estradiol and levonorgestrel/ethinyl estradiol improved total lesion percent change, but no P-values were given in the comparison to baseline. The cyproterone/ethinyl estradiol arm had a statistically significantly greater decrease in total lesion change at 16 weeks than the levonorgestrel/ethinyl estradiol arm, but there was no statistically significant difference between the arms at 24 weeks. No side effect data were presented.

M.01.09-Cyproterone/Ethinyl Estradiol (oral) vs. Norethisterone/Ethinyl Estradiol (oral)

There was one medium trial of women with moderate to severe acne, Trial 237. ²¹⁶ Comparison of cyproterone/ethinyl estradiol to norethisterone/ethinyl estradiol demonstrated statistically significant reduction of face severity and total severity in both arms by week 26. Papule count was only statistically significantly reduced in the cyproterone/ethinyl estradiol arm. There were no statistically significant differences between the arms in nodule count or papule count. There was statistically significantly more improvement in the norethisterone arm in total severity and face severity. No side effect data were presented.

M.02.01-Desogestrel/Ethinyl Estradiol (oral) vs. Gestodene/Ethinyl Estradiol(oral)

There was one trial, Trial 225, ²²⁰ a small trial of patients with mild and moderate acne. This trial had limitations including lack of blinding and major losses to followup. Both treatments showed statistically significant improvement in acne severity grade at 36 weeks compared to baseline. The patients that received desogestrel/ethinyl estradiol had more improvement than those in the gestodene/ethinyl estradiol arm, although no P-values were provided. Similar numbers of patients reported systemic side effects in the desogestrel/ethinyl estradiol arm (6) and in the gestodene/ethinyl estradiol arm (4).

M.02.02-Desogestrel/Ethinyl Estradiol (oral) vs. Levonorgestrel/EthinylEstradiol (oral)

There were two trials. Trial 264 ²²¹ was a small trial of women with mild and moderate acne. Trial 263 ²²² was a medium trial. Trial 264 demonstrated statistically significant improvement in acne severity over baseline in both treatment arms, with the desogestrel arm having statistically significantly greater improvement than the levonorgestrel arm. Similarly, in Trial 263 there were patients in both arms who were rated "improved" at 24 weeks, but there were statistically significantly more improved patients in the desogestrel arm. Side effects were reported in Trial 263. Two patients in the desogestrel arm and three patients in the levonorgestrel arm reported irregular bleeding. Two patients in the desogestrel arm and four in the levonorgestrel arm reported that their acne was worse and dropped out. While these numbers were similar, no P-value was provided.

M.03.01-Spironolactone (oral) vs. Placebo (oral)

There were three trials. One was a small four-arm trial, Trial 156, ²²³ of patients with moderate to severe acne, limited by the absence of blinding. Results were stratified by gender. Women seemed to have lower severity scores than men, but no comparison statistics were reported. The second was a large cross-over trial, Trial 249, ²²⁴ of 200 mg daily in women with mild-to-moderate acne (N=58 in each arm). The third was a very small five-arm trial, Trial 132, ²²⁵ of four doses of spironolactone tested (50, 100, 150, and 200 mg daily) in a mixed-gender pool of patients with severe acne. Trial 249 was weakened by major losses to followup.

In Trial 249, spironolactone provided greater improved lesion counts (75 percent vs. 20 percent), greater improvement in patients' perceived severity (90 percent vs. 10 percent) and in photography-based severity assessment (55 percent vs. 20 percent). There were side effects deemed "positive" by the authors (decreased oily skin and breast enlargement) as well. In Trial 132, the higher doses of spironolactone may have led to higher assessments of overall change, but the numbers were very small.

There was a high rate of menstrual irregularity in Trial 249 and about a 10 percent rate of severe nausea, bad enough to lead to patient withdrawal; but the actual numbers are small. Trial 132 did not report any side effects.

M.04.01-Delta-0.2 percent Chlormadione (topical) vs. Vehicle (topical)

There was one small trial, Trial 2. ²²⁶ There were no patients who had any change in their acne reported by either physician or patient, although no P-values were provided. No data on side effects were reported.

M.04.02-Delta-5 percent Chlormadione (topical) vs. Vehicle (topical)

There was one small trial, Trial 2. ²²⁶ There were no patients who had any change in their acne reported by either physician or patient, although no P-values were provided. No data on side effects were reported.

M.04.03-Delta-5 percent Chlormadione (topical) vs. Delta-0.2 percentChlormadione (topical)

There was one small trial, Trial 2. ²²⁶ There were no patients who had any change in their acne reported by either physician or patient, although no P-values were provided. No data on side effects were reported.

M.04.04-16-Epiestriol-3-Allyl Ether (oral) vs. Placebo (oral)

There was one large trial (N=134). Trial 19 ²²⁷ compared the synthetic estrogen 16-epiestriol-3-allyl ether to placebo. After 12 weeks of treatment, there was no statistically significant difference in acne severity between the treatment arm and the placebo arm, with similar numbers of patients rated as "no change" (31 vs. 34), "improved" (19 vs. 26) and "marked improvement" (16 vs. 8). Similar number of patients also reported improvement in oiliness in both groups.

M.04.05-Epiestriol (topical) vs. Vehicle (topical)

There was one small split-face trial, Trial 295. ²²⁸ After 12 weeks, more patients were termed "better" in terms of severity in the placebo arm than in the treatment arm, though no P-values were provided. Both arms had the same number of patients rated as "worse" (7), "unchanged" (18), and "improved" (19) in terms of overall change as rated by the physician at 12 weeks. No side effect data were provided.

M.04.06-Inocoterone Acetate (topical) vs. Vehicle (topical)

There was one large trial (N=153), Trial 213, ²²⁹ that evaluated male patients with moderate acne. While there was no statistically significant difference in the inflammatory lesion count or the comedo count between the treatment arm and the placebo arm at 4 weeks, there was a statistically significantly greater decrease in the inflammatory lesion count at weeks 12 and 16 in the treatment arm. There remained no statistically significant difference between the two arms in terms of comedo count at 12 or 16 weeks. The number of inflammatory lesions decreased steadily from baseline through week 16 in the treatment arm, but no P-values were provided for within group comparison. Local reactions in the treatment arm included burning (32), stinging (18), itching (10), redness (19), and peeling (27). There were no statistically significant differences between the two arms with respect to the number of patients reporting local adverse reactions.

M.04.07-Levonorgestrel/Ethinyl Estradiol (oral) vs. Cyproterone/EthinylEstradiol (oral)

See M.01.08.

M.04.08-Norethisterone/Placebo/Ethinyl Estradiol (oral) vs.Cyproterone/Ethinyl Estradiol (oral)

See M.01.09.

M.04.09-Norgestimate/Ethinyl Estradiol (oral) vs. Placebo (oral)

There were two large trials of women, Trial 218 ²³⁰ (moderate acne) and Trial 280 (mild to moderate acne). ²³¹ While Trial 218 had major losses to followup, it did demonstrate decreases in total lesions, comedo count, and inflammatory lesion count at 26 weeks compared to baseline although no P-values were provided. The differences were statistically significant compared to placebo. There was no statistically significant improvement in the number of nodules compared to placebo. Similarly, Trial 280 demonstrated change in total lesion count, inflammatory lesions, and comedones, although no P-value was provided for comparison to baseline. The percent change, both in inflammatory lesions and in total lesions, was statistically significant compared to placebo. Both trials demonstrate statistically significant overall change reported by both investigator and patient. Trial 280 reported similar numbers of adverse events reported between treatment arm and placebo arm, with 13 patients discontinuing therapy and 113 patients reporting events in the treatments arm, and 5 patients discontinuing therapy and 119 other events reported in the placebo arm. No P-values were presented.

M.04.10-Quinestradol (oral) vs. Placebo (oral)

There was one medium trial of women with mild to moderate acne, Trial 248, ²³² with major losses to followup. After 8 weeks, patients in the treatment arm demonstrated a statistically significant decrease in the comedo count and total lesion count compared to baseline. However, there were no statistically significant differences demonstrated between the treatment arm and the placebo arm in terms of either outcome. No side effect data were presented.

M.04.11-Serum Gonadotropin (oral) vs. Placebo (oral)

There was one small trial, Trial 271, ²³³ a trial of patients with mild, moderate, and severe acne. In addition to serum gonadotropin (with unclear proposition of luteinizing hormone and/or follicle stimulating hormone) or placebo, patients received ancillary treatments including ultraviolet light and antibiotics. Similar numbers of patients were rated as "improved" (8 vs. 9), "no change" (9 vs. 9), and "worse" (2 vs. 3) in the treatment and placebo arms after 12 weeks of treatment, although no P-values were provided. No side effect data were presented.

N.06.045-Ibuprofen/Tetracylcine (oral) vs. Iinert (oral)

There was one small two-arm trial, Trial 349, ¹⁸¹ of young adults with mild to moderate acne. The arms were ibuprofen/tetracycline, tetracycline, ibuprofen, placebo. (Each arm had appropriate placebo to make two medications for each patient.) Patient blinding was documented, but physicians were not. At 8 weeks, total change was statistically significantly greater for the combination over the other three arms. While the change in single-treatment arms was greater than in placebo arm, P-values were not provided for those comparisons. Side effects data were not provided.