Oral H1 antihistamines as ‘add‐on’ therapy to topical treatment for eczema
The symptoms of eczema can lead to sleeplessness and fatigue and may have a substantial impact on quality of life. Use of oral H1 antihistamines (H1 AH) as adjuvant therapy alongside topical agents is based on the idea that combining the anti‐inflammatory effects of topical treatments with the blocking action of histamine on its receptors in the skin by H1 AH (to reduce the principal symptom of itch) might magnify or intensify the effect of treatment. Also, it would be unethical to compare oral H1 AH alone versus no treatment, as topical treatment is the standard management for this condition.
To assess the effects of oral H1 antihistamines as 'add‐on' therapy to topical treatment in adults and children with eczema.
We searched the following databases up to May 2018: the Cochrane Skin Group Specialised Register, CENTRAL, MEDLINE, Embase, and the GREAT database (Global Resource of EczemA Trials; from inception). We searched five trials registers and checked the reference lists of included and excluded studies for further references to relevant randomised controlled trials (RCTs). We also searched the abstracts of four conference proceedings held between 2000 and 2018.
We sought RCTs assessing oral H1 AH as 'add‐on' therapy to topical treatment for people with eczema compared with topical treatment plus placebo or no additional treatment as add‐on therapy.
Data collection and analysis
We used standard Cochrane methodological procedures. Primary outcome measures were 'Mean change in patient‐assessed symptoms of eczema' and 'Proportion of participants reporting adverse effects and serious adverse events'. Secondary outcomes were 'Mean change in physician‐assessed clinical signs', 'Mean change in quality of life', and 'Number of eczema flares'.
We included 25 studies (3285 randomised participants). Seventeen studies included 1344 adults, and eight studies included 1941 children. Most studies failed to report eczema severity at baseline, but they were conducted in secondary care settings, so it is likely that they recruited patients with more severe cases of eczema. Trial duration was between three days and 18 months. Researchers studied 13 different H1 AH treatments. We could not undertake pooling because of the high level of diversity across studies in terms of duration and dose of intervention, concomitant topical therapy, and outcome assessment. Risk of bias was generally unclear, but five studies had high risk of bias in one domain (attrition, selection, or reporting bias). Only one study measured quality of life, but these results were insufficient for statistical analysis.
Although this review assessed 17 comparisons, we summarise here the results of three key comparisons in this review.
Cetirizine versus placebo
One study compared cetirizine 0.5 mg/kg/d against placebo over 18 months in 795 children. Study authors did not report patient‐assessed symptoms of eczema separately for pruritus. Cetirizine is probably associated with fewer adverse events (mainly mild) (risk ratio (RR) 0.68, 95% confidence interval (CI) 0.46 to 1.01) and the need for slightly less additional H1 AH use as an indication of eczema flare rate (P = 0.035; no further numerical data given). Physician‐assessed clinical signs (SCORing Atopic Dermatitis index (SCORAD)) were reduced in both groups, but the difference between groups was reported as non‐significant (no P value given). Evidence for this comparison was of moderate quality.
One study assessed cetirizine 10 mg/d against placebo over four weeks in 84 adults. Results show no evidence of differences between groups in patient‐assessed symptoms of eczema (pruritus measured as part of SCORAD; no numerical data given), numbers of adverse events (RR 1.11, 95% CI 0.50 to 2.45; mainly sedation, other skin‐related problems, respiratory symptoms, or headache), or physician‐assessed changes in clinical signs, amount of local rescue therapy required, or number of applications as an indicator of eczema flares (no numerical data reported). Evidence for this comparison was of low quality.
Fexofenadine versus placebo
Compared with placebo, fexofenadine 120 mg/d taken in adults over one week (one study) probably leads to a small reduction in patient‐assessed symptoms of pruritus on a scale of 0 to 8 (mean difference (MD) ‐0.25, 95% CI ‐0.43 to ‐0.07; n = 400) and a greater reduction in the ratio of physician‐assessed pruritus area to whole body surface area (P = 0.007; no further numerical data given); however, these reductions may not be clinically meaningful. Results suggest probably little or no difference in adverse events (mostly somnolence and headache) (RR 1.05, 95% CI 0.74 to 1.50; n = 411) nor in the amount of 0.1% hydrocortisone butyrate used (co‐intervention in both groups) as an indicator of eczema flare, but no numerical data were given. Evidence for this comparison was of moderate quality.
Loratadine versus placebo
A study of 28 adults compared loratadine 10 mg/d taken over 4 weeks versus placebo. Researchers found no evidence of differences between groups in patient‐assessed pruritus, measured by a 100‐point visual analogue scale (MD ‐2.30, 95% CI ‐20.27 to 15.67); reduction in physician‐assessed clinical signs (SCORAD) (MD ‐4.10, 95% CI ‐13.22 to 5.02); or adverse events. Study authors reported only one side effect (folliculitis with placebo) (RR 0.25, 95% CI 0.01 to 5.76). Evidence for this comparison was of low quality. Number of eczema flares was not measured for this comparison.
Based on the main comparisons, we did not find consistent evidence that H1 AH treatments are effective as 'add‐on' therapy for eczema when compared to placebo; evidence for this comparison was of low and moderate quality. However, fexofenadine probably leads to a small improvement in patient‐assessed pruritus, with probably no significant difference in the amount of treatment used to prevent eczema flares. Cetirizine was no better than placebo in terms of physician‐assessed clinical signs nor patient‐assessed symptoms, and we found no evidence that loratadine was more beneficial than placebo, although all interventions seem safe.
The quality of evidence was limited because of poor study design and imprecise results. Future researchers should clearly define the condition (course and severity) and clearly report their methods, especially participant selection and randomisation; baseline characteristics; and outcomes (based on the Harmonising Outcome Measures in Eczema initiative).
Uwe Matterne, Merle Margarete Böhmer, Elke Weisshaar, Aldrin Jupiter, Ben Carter, Christian J Apfelbacher
Plain language summary
Do H1 antihistamine tablets or liquids help improve eczema symptoms in people who are already using creams or ointments for their eczema?
Are H1 antihistamines (which inhibit the action of chemicals released as part of an allergic reaction; known as 'histamines'), taken as tablets or liquid, effective and safe in people of any age with diagnosed eczema, if given in addition to creams and ointments, compared to treatment with an inactive substance (placebo) or nothing added to creams and ointments?
Eczema (also known as 'atopic eczema/dermatitis') is a skin disorder frequently affecting both children and adults. In developed countries, 10% to 20% of all people are affected by eczema during their lifetime. The main symptom is itch, which results in scratching and, together with skin inflammation, frequently produces reddening of the skin. The symptoms of eczema can lead to sleeplessness and fatigue, lowering quality of life. Antihistamines are frequently given for itch (specifically H1 antihistamines taken by mouth), and they may alleviate the symptoms of eczema when given in addition to conventional treatments directly applied to the skin (e.g. emollients, moisturisers, steroid creams), although they are not thought to cure it. Many antihistamines are available without prescription, for instance, cetirizine or loratadine. Although H1 antihistamines are frequently prescribed for treating eczema, we do not know whether they are effective and safe.
We searched for relevant studies up to May 2018. We included 25 randomised controlled trials with 3285 participants of all ages with diagnosed eczema. Eight studies included children or adolescents, and 17 included adults. The gender of participants and the severity of symptoms often were not reported. All studies were conducted in secondary care settings, including hospital clinics, research clinics, dermatology centres, and surgery centres, meaning that participants were likely to have more severe eczema than if recruitment occurred from first point of contact settings (i.e. primary care). All but one study compared H1 antihistamine versus placebo. Researchers studied 13 different H1 antihistamines, most of which were less sedating H1 antihistamines (known as 'second‐generation antihistamines'). Studies lasted between three days and 18 months. Seven trials received funding from pharmaceutical companies; they are the largest trials included in this review.
We found no convincing evidence that H1 antihistamines help patients with eczema.
One study compared cetirizine 0.5 mg/kg/d versus placebo (in children over a period of 18 months). No data were provided on patient‐assessed itch symptoms of eczema. Cetirizine is probably associated with fewer (mainly mild) adverse events and the need for slightly less additional H1 antihistamine to prevent flares. Even though physician‐assessed clinical signs were reduced in both groups, results show no differences between groups (all moderate‐quality evidence).
When compared with placebo, we found no evidence that an increased dose of cetirizine 10 mg/d over four weeks makes a difference in terms of patient‐assessed itch, number of side effects, physician‐assessed signs, or number of eczema flares as measured by the amount of treatment used (all low‐quality evidence). Side effects reported in both groups included drowsiness, skin‐related problems, breathing issues, and headaches.
Compared with placebo, fexofenadine 120 mg/d given to adults for one week probably slightly improves patient‐assessed itch, as well as producing a greater reduction in the area of itch, as assessed by a physician, and it probably makes little or no difference in the number of participants experiencing side effects (mostly drowsiness and headaches) or in the extent of treatment required as an indicator of the number of eczema flares (all moderate‐quality evidence).
We found no evidence of a difference between placebo and loratadine 10 mg/d given to adults for four weeks in patient‐assessed itch, occurrence of side effects, or physician‐assessed signs of eczema (all low‐quality evidence). This study did not measure the number of eczema flares. Study authors reported only one side effect (folliculitis), which occurred with placebo.
Only one study measured quality of life, but results could not be analysed.
Quality of the evidence
For all outcomes across key comparisons, evidence was of low to moderate certainty. Reasons for lowering the quality of evidence include concerns over how studies were carried out and inclusion of too few participants, leading to less accurate results.
Uwe Matterne, Merle Margarete Böhmer, Elke Weisshaar, Aldrin Jupiter, Ben Carter, Christian J Apfelbacher
Implications for practice
We found little conclusive evidence to establish the effectiveness of H1 AH as add‐on therapy to topical treatment for eczema. We derived this conclusion from careful consideration of the quality of included trials and their reporting, as well as the wide variation in comparisons and the fact that we found no opportunities to combine results in meta‐analyses.
Details of the specific dose and regimen of the interventions described in this section are found in Effects of interventions. All interventions in this section were compared against placebo.
With regard to the primary outcome ‘Patient‐assessed symptoms’, which we narrow here specifically to participant‐assessed pruritus, we found no evidence that cetirizine 10 mg/d or loratadine makes a difference in this outcome (both low‐quality evidence). Fexofenadine probably slightly improves this outcome (moderate‐quality evidence), but the clinical meaning of this slight improvement is uncertain. Trials assessing cetirizine 0.5 mg/kg/d did not report this outcome.
Fexofenadine probably makes little or no difference in the occurrence of adverse effects, and cetirizine 0.5 mg/kg/d is probably associated with fewer adverse events (both moderate‐quality evidence). We found no evidence that cetirizine 10 mg or loratadine makes a difference in this outcome (both low‐quality evidence).
In terms of physician‐assessed clinical signs, we found no evidence that loratadine or cetirizine 10 mg/d or 0.5 mg/kg/d makes a difference (low‐quality evidence). Fexofenadine probably improves this outcome (moderate‐quality evidence), but the clinical meaning of this improvement is uncertain.
Fexofenadine probably makes little or no difference in terms of the number of eczema flares (measured by ‘escalation of treatment’ or ‘use of topical anti‐inflammatory medications’) (moderate‐quality evidence). We found no evidence that cetirizine 10 mg makes a difference in the amount of treatment required (low‐quality evidence), but less cetirizine 0.5 mg/kg/d is probably needed, hence reducing the number of eczema flares (moderate‐quality evidence). Trials assessing loratadine did not report this outcome.
We could draw no conclusions about effects on quality of life as only one study assessed this outcome. The data provided by this single study were insufficient for statistical analysis.
Implications for research
One general implication of this review is that some uncertainty remains with regard to the question of whether H1 AH are effective in the treatment of eczema. Further, review findings are not robust. In light of these limitations, little evidence from the 25 included trials suggests that H1 AH have a clinically significant impact on eczema.
As the quality of evidence was often low and our risk of bias assessment arrived frequently at a judgement of unclear, we would like to recommend several methodological requirements for future studies, which should be randomised controlled trials. First, research would benefit from a clear definition of the condition examined, including its course and severity. In intervention studies, it is important to know who exactly was studied, which is why a standardised assessment should be carried out and appropriately reported. Many of the reviewed studies failed to provide this important information. Other relevant baseline data should be obtained and reported. For instance, a clear sociodemographic description should be provided. In our review, most studies failed to do so.
Second, we often noted insufficient detail on how randomisation was implemented, how allocation concealment was achieved, and how incomplete outcome data (attrition) were dealt with. Adherence to guidelines, such as the CONSORT (Consolidated Standards of Reporting Trials) statement (Moher 1998), would help in ensuring complete reporting. For example, to safeguard against incomplete outcome data, investigators need to routinely anticipate attrition and incorporate various measures to minimise dropouts during trial design (Hui 2013). Study participants should be well informed, and expectations with respect to study procedures should be realistic (Matsui 2012). Should substantial attrition occur, it is of paramount importance to assess whether attrition may have affected any effect estimates (Hui 2013). To ensure that intervention and placebo groups are balanced with regard to important clinical and demographic variables, investigators should employ and report adequate randomisation procedures and should check the distribution of these variables between groups.
Third, future research would benefit from a standardised assessment of outcomes. Guidance for choice of reliable and valid outcome measures may be obtained from the HOME initiative (Schmitt 2014). Future availability of more trials using the same standardised outcome measurement, similar intervention periods, and comparable dosages may allow pooling of data ‐ something we could not do. We also recommend that future studies should be conducted in a multi‐arm fashion with regard to dosages, such that several different dosages of the same H1 AH are compared against placebo. Further, concomitant treatment should be applied in a standardised way. For instance, studies should mimic clinical practice in accordance with existing guidelines, consisting of moisturisers and/or emollients in combination with either corticosteroids or other immunomodulatory agents (such as tacrolimus or pimecrolimus) (Werfel 2016). Additionally, we suggest that H1 AH interventions should be studied in patients with eczema with allergic comorbidity as assessed by allergic sensitisations (known as 'atopic' eczema).
Several of the included trials studied small or very small samples. As a consequence, we downgraded the quality of the evidence from such trials. Future studies should state what effect is considered clinically useful and should base the sample size calculation on this effect to achieve adequate power. This would greatly enhance confidence in effect estimates.
An important issue pertains to the generalisability of reported findings. Most studies appear to have based their selection rationale on easy availability of potential participants in clinical settings. In other words, the samples can be said to be convenience samples ‐ a fact that greatly limits the extent to which findings can be generalised to the population of patients with eczema on the whole. Authors of high‐quality intervention studies should design their trials to maximise the generalisability of study findings.
Moreover, although the duration of the intervention and of follow‐up was as long as 18 months for cetirizine and levocetirizine, up to 52 weeks for ketotifen, or eight weeks for olopatadine, only short‐ or medium‐term interventions and follow‐ups were available for acrivastine, azelastine, chlorpheniramine, fexofenadine, hydroxyzine, loratadine, terfenadine, or tazifylline. Future research should ensure a sufficient duration of the intervention to assess impact on eczema severity.
However, as current guidelines recommend the use of H1 AH as add‐on to conventional topical treatment during acute flares only (Werfel 2016), future studies could greatly benefit from comparing H1 add‐on treatment in groups of patients with acute flares that are comparable.
Only one study assessed quality of life, but the data it provided were insufficient for statistical analysis. It must be borne in mind though that many studies were conducted at a time when health‐related patient‐reported outcomes were given minor credence compared to today. It is recommended that future trials should include one such measure, in line with the consented core outcome set of the HOME initiative (Heinl 2016; Heinl 2017; Schmitt 2012).
There is scope for replication studies. For instance, olopatadine and fexofenadine were found to have positive effects, albeit of little clinical significance. Further confirmatory trials may be informative. There is also scope for investigating other H1 AH as oral add‐on interventions for eczema. For instance, other H1 AH, such as ebastine, rupatadine, desloratadine, diphenhydramine, dimetindene, clemastine, cyproheptadine, mizolastine, and promethazine, are used in clinical practice for the treatment of eczema. However, the efficacy and safety of these agents have not yet been studied in clinical trials.
Should future studies establish more robust evidence for specific H1 AH treatment effects on eczema and adequate safety, subsequent trials could try to compare H1 AH among each other.
In light of recent findings that not only histamine‐related pathways lead to mediation of pruritus (Ständer 2002; Tominaga 2016; Yosipovitch 2008), it may be worthwhile to study combination treatments in the future.