EVIDENCE-BASED DERMATOLOGY: ORIGINAL CONTRIBUTION An Evidence-Based Review of the Efficacy of Antihistamines in Relieving Pruritus in Atopic Dermatitis Peter A. Klein, MD, MPA; Richard A. F. Clark, MD Objective: To critically review the body of clinical tri- use of antihistamines in relieving pruritus. One grade B als that refute or support the efficacy of antihistamines trial supported the efficacy of antihistamines in reliev- in relieving pruritus in patients with atopic dermatitis. ing pruritus. All remaining trials (grade C) lacked pla- cebo controls or randomization, or contained fewer than Design: Review of MEDLINE from 1966 through March 20 patients in each treatment group. 1999, the Cochrane Database of Systematic Reviews, and Best Evidence to identify therapeutic trials of antihista- Conclusions: Although antihistamines are often used mines in patients with atopic dermatitis. in the treatment of atopic dermatitis, little objective evi- dence exists to demonstrate relief of pruritus. The ma- Main Outcome Measures: All randomized con- jority of trials are flawed in terms of the sample size or trolled trials or clinical trials of antihistamines used in study design. Based on the literature alone, the efficacy the treatment of atopic dermatitis. We found 16 studies of antihistamines remains to be adequately investi- throughout the literature. gated. Anecdotally, sedating antihistamines have some- times been useful by virtue of their soporific effect and Results: Large, randomized, double-blind, placebo- bedtime use may be warranted. There is no evidence to controlled clinical trials with definitive conclusions (grade support the effectiveness of expensive nonsedating agents. A trials) have not been performed. Two grade B trials (small, rigorous, randomized trials with uncertain re- sults due to moderate to high a or b error) refuted the Arch Dermatol. 1999;135:1522-1525 TOPIC dermatitis is a com- receptor antihistamines in managing pru- mon chronic or relapsing ritus in atopic dermatitis, few random- eczematous dermatitis ized, double-blind, placebo-controlled characterized by intense clinical studies have evaluated efficacy. In pruritus and occurring pri- addition to older sedating formulations, ex- Amarily in infants and children with a per- pensive nonsedating agents are some- sonal or family history of atopy. Nine per- times used in the absence of any clinical cent to 12% of all children are affected with evidence that demonstrates relief of pru- atopic dermatitis,1,2 and 60% to 70% of ritus. those with mild to severe dermatitis will continue to experience symptoms into RESULTS adulthood.3,4 A significant proportion of patients who have outgrown the typical Sixteen studies have evaluated the effi- manifestations of the disease develop ir- cacy of antihistamines in relieving pruri- ritant dermatitis, which may be chronic tus in patients with atopic dermatitis. and work-disabling, especially in the con- Large, randomized, double-blind, placebo- text of wet, dirty, or caustic conditions.5 controlled clinical trials with definitive Pruritus is one of the most common conclusions have not been performed symptoms of atopic dermatitis. The itch- (grade A trials). Two grade B trials refute scratch cycle exacerbates damage to the the use of antihistamines in relieving pru- epidermal barrier thereby increasing wa- ritus and 1 grade B trial supports the ef- ter loss and drying, which creates a suit- ficacy of antihistamines. Four grade C tri- 7-22 From the Department of able environment for skin pathogens to als refute and 9 grade C trials support Dermatology, State University cause infection and flaring of symptoms. the antipruritic effects of antihistamines. of New York at Stony Brook. Despite the frequent use of histamine H1 These results are summarized in the ARCH DERMATOL / VOL 135, DEC 1999 WWW.ARCHDERMATOL.COM 1522 ©1999 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 cision of the study. Because of the small number of sub- jects and the imprecision of the point estimate of pruri- METHODS tus relief, a treatment effect cannot be ruled out. Wahlgren et al8 performed a randomized, double- We searched the MEDLINE database for trials of an- blind, placebo-controlled, crossover study of terfena- tihistamines published from 1966 through March dine, 60 mg twice daily, and clemastine fumarate, 2 mg 1999. The search used the keywords dermatitis, atopic/ twice daily for 3 days with intervening 4-day washout therapy or eczema/therapy and histamine H1 antago- periods. The study was randomized with a double- nists/therapy with publication types randomized con- dummy protocol, with each patient receiving 3 courses trolled trial or clinical trial. The Cochrane Database in random order: (1) active terfenadine and placebo clem- of Systematic Reviews and Best Evidence databases astine fumarate; (2) placebo terfenadine and active clem- were also searched using the keywords atopic der- astine fumarate; and (3) placebo terfenadine and pla- matitis and eczema. The quality of each trial was 6 cebo clemastine fumarate. Twenty-five adults fulfilling ranked by applying a modified version of Sackett’s 23 criteria for clinical evidence. Grade A trials are large, the criteria of Hanifin and Rajka were enrolled in the randomized, double-blind, placebo-controlled stud- study. Age, sex, disease duration, and history of atopy ies with low false-positive (a) and low false- were provided. Patients were permitted to use topical hy- negative (b) errors. Grade A trials lead to definitive drocortisone and emollients during the study. The quan- conclusions with precise point estimates of treat- tity of topical hydrocortisone used during the 4-day wash- ment effect. Grade B studies are also randomized, out period after each treatment course was comparable double-blind, placebo-controlled studies, but in- between groups. Patients recorded pruritus using visual clude a small number of patients, thereby increas- analog scales and a computerized method for self- ing the likelihood of high false-positive and/or false- recording. The authors found no significant difference negative errors. Grade B results may be statistically in the intensity of itch between the 3 treatment groups significant, for example, yet provide an imprecise point estimate of treatment effects. Grade C trials lack 1 using either assessment method. The sedative effect of or more of the following criteria: randomization, pla- clemastine fumarate was significantly greater than ter- cebo control, or blinding. Trials with fewer than 20 fenadine or placebo, yet its antipruritic effect did not patients in each treatment group are also grade C tri- differ. als. Case reports and case series are categorized as Hannuksela et al9 conducted a randomized, double- grade C trials. blind, placebo-controlled, parallel study of cetirizine hy- drochloride, a nonsedating antihistamine, at 10 mg, 20 mg, or 40 mg taken once daily for 4 weeks. The study included 178 adults, patient characteristics, such as mean Table. The grade B studies present the best evidence avail- age, sex, weight, and prior use of emollients and hydro- able and are discussed below. cortisone were provided. All groups were comparable at Berth-Jones and Graham-Brown7 conducted a ran- baseline. The criteria of Hanifin and Rajka23 were not used domized, double-blind, placebo-controlled, crossover in patient selection. Of the 178 patients, 51 were ex- study of 28 subjects given 120 mg of terfenadine twice cluded from the statistical analysis of efficacy: adverse daily for 1 week. The mean age and fulfillment of the di- reactions (20), noncompliance (19), protocol viola- agnostic criteria of Hanifin and Rajka23 were noted, but tions (5), use of a forbidden drug (5), lost to follow-up the characteristics of the study subjects and their dis- (1), and did not fulfill the inclusion criteria (1). The au- ease severity were not provided. Of the 28 subjects, 4 were thors neither revealed the characteristics and extent of excluded from the analysis due to failure to comply with disease in these dropout patients nor did they address the protocol either by altering their topical steroid treat- the potential effects of their exclusion. The active group ment or by failing to take the trial medications as di- included 26, 34, and 35 patients who received 10 mg, rected. The degree to which these dropout patients dif- 20 mg, and 40 mg of cetirizine hydrochloride daily for 4 fered in terms of disease severity from the remaining weeks, respectively. The placebo group had 32 subjects. subjects was not provided. The subjects were permitted Patients used emollients and hydrocortisone cream dur- to use topical steroids and emollients to maintain their ing the study and the containers were weighed before and dermatitis in a stable condition. The type of topical ste- after the trial. Pruritus was assessed using visual analog roid and the quantity used by each subject were not noted. scales by both investigators and patients. At the end of The first 3 days of the study were a washout period, and the trial, the investigators’ pruritus scores of the active assessment was performed only during the last 4 days of and placebo groups at all doses were significantly de- each treatment. Subjects used visual analog scales to re- creased compared with baseline values. However, none cord the severity of pruritus. Subjects were also exam- of the active groups showed a statistically significant im- ined by investigators at the end of each treatment pe- provement compared with the placebo group. The pa- riod and the severity of excoriation was evaluated. The tients’ pruritus scores at the final visit showed that the mean (±SE) visual analog scores for terfenadine and pla- active and placebo groups had a statistically significant cebo were 23.95 (±4.9) and 25.13 (±5.1), respectively. improvement from baseline values (P #.001). Only the The authors concluded that terfenadine was not effec- group receiving the 40-mg dose showed a statistically sig- tive in relieving pruritus.