Efficacy of moisturizers in paediatric atopic dermatitis: A systematic review and meta-analysis of randomised controlled trials

Search strategy

We searched Ovid Embase, Ovid Medline, Web of Science, Cochrane Database of Controlled Trials (CENTRAL), CINAHL, GREAT and DARE for relevant randomised controlled trials from inception to July 31, 2020. A cross-reference check was attempted to identify additional studies.

Eligibility criteria

Selected studies had to meet the following inclusion criteria: (1) participants age ≤15 years old with atopic dermatitis, diagnosed by a physician using U.K. Working Party’s diagnostic criteria for atopic dermatitis or Hanifin and Rajka criteria, who have received prior treatment and are flare-free at study initiation;^16,17 (2) intervention of any type of topical moisturizer, applied daily at any amount or duration and (3) a control group of no moisturizer treatment. Cointervention such as cleansing gel, bodywash, topical corticosteroid, topical calcineurin inhibitor, antibiotics or antihistamine were allowed if both groups were administered the same agent; (4) studies that investigated our pre-specified outcomes and (5) experimental studies that were either a randomised controlled trial or controlled clinical trial published in English.

The required exclusion criteria: (1) Any other forms of moisturizer apart from cream or lotion; (2) studies that compared moisturizer with active ingredient versus standard moisturizer without such ingredient and (3) studies that investigated adjunctive effect of moisturizer with another agent (such as topical corticosteroid, topical calcineurin inhibitor, antibiotics and antihistamines) compared to moisturizer alone.

Data collection

The titles and abstracts were independently reviewed by two authors against the criteria. Full texts were evaluated, when necessary, especially when abstracts did not clarify the participants’ age or presence of a no moisturizer control. Two researchers collected data independently using a data extraction form. Details of participants (age and gender, inclusion and exclusion criteria, baseline data, number randomised and number of dropouts), method (study design, blinding, randomisation, study duration, investigators and outcome assessors), intervention groups (details of treatment regimen, dosage, frequency, location of application and duration), results (types of outcomes collected and timing of assessment), funding source and declarations of interest were collected. Contact with the research authors was attempted when essential information was unclear. Bias was evaluated independently by two authors using the Cochrane ‘Risk of bias’ tool, in which the risk in each bias domain was graded as either ‘high,’ ‘unclear,’ or ‘low.’¹⁸ We also intended to assess publication bias using funnel plot analysis, given the known limitations of the method. Any disagreements between the authors were adjudicated by a third reviewer.

Study selection

The initial search retrieved a total of 3590 publications [Figure 1]. After de-duplication, we assessed the titles and abstracts of 1988 studies.We then screened the full text of 94 articles and found six that satisfied the inclusion criteria. Common reasons for exclusion of articles were the use of vehicle/placebo instead of a no treatment control,^20-24 and wrong comparisons of topical corticosteroids versus moisturizers.^25-28

Close

Figure 1:

Preferred Reporting Items for Systematic Reviews and Meta-Analysis flow diagram showing the methodology in selecting articles for final analyses

Export to PPT

Study characteristics

Table 1 illustrates the characteristics of included publications.^29-34 They were published between 2006 and 2017, were mostly multi-center and conducted in Europe.

Patients had mild to moderate severity (Scoring Atopic Dermatitis (SCORAD) of 5 to 35 or Investigator Global Assessment of 2 to 3), except for Grimalt et al. were done on moderate to severe disease (SCORAD 20–70).³¹ The number of participants ranged from 45 to 335 in each study, with a total of 748 subjects (intervention 436, control 312). Participants’ age varied from one month to 12 years. The study durations ranged from one to six months, with a mean length of 2.4 months.

Tiplica et al. had a three-arm design in which the moisturizer of interest, Dexeryl® (Pierre Fabre) and reference emollient, Atopiclair® (Sinclair Pharma) were compared against no treatment.³³ Other moisturizers were investigated included Exomega Milk® (Pierre Fabre).^30,31 Avène Xeracalm Balm®(Pierre Fabre),²⁹ Eucerin® Eczema Relief Body Crème (Beiersdorf)³⁴ and Cetaphil Restoraderm® (Galderma).³² While Tiplica et al. did not administer any cointervention, the other five studies gave a standardised cleanser or hygiene product to research arms.

Four separate publications have declared their source of funding as the pharmaceutical industry of the studied product.^29,31-33 Giordano-Labadie et al. and Weber et al. did not report their sponsorship, but the intervention was manufactured by the company that employed their researchers, making these parties the likely funding sources.^30,34

Risk of bias within studies

As summarised in Figure 2, all trials incorporated randomisation and allocation concealment, which limited selection bias. However and most strikingly, investigators and outcome assessments were not blinded in all of the studies.^29-34 Two of six publications had an unclear risk for attrition bias due to high dropout rates in the control or had missing data from questionnaires.^31,33 Bianchi et al.²⁹ and Tiplica et al.³³ demonstrate high risk for publication bias because they did not report results on individual signs and symptoms, although the data were collected as part of their Scoring atopic dermatitis assessments. Taken together, we appraised all included studies as showing a high overall risk for bias. We did not evaluate funnel plot symmetry to assess publication bias due to insufficient number of trials.

Close

Figure 2:

Risk of bias summary: review authors’ judgments about each risk of bias item for each included study

Export to PPT

Effect of moisturizer on outcomes

Extension of time to flare

Moisturizer treated participants experienced an extension of 13.52 days in time to flare compared to control [Figure 3a; 3 studies, n=231;95% confidence interval 0.05–26.99, I² 88%].^32-34 Apart from the wide confidence interval and limited sample size, the quality of evidence was downgraded to ‘very low’ due to risk of bias, indirectness and inconsistency [Table 2].^33,34

Close

Figure 3a:

Forest plot for differences in time to flare (days) in moisturizer user versus non-user

Export to PPT

Table 2: Summary of findings with assessment of evidence quality according to the Grading of Recommendations, Assessment, Development and Evaluation framework

Outcome	Number of participants (studies)	Certainty of the evidence (GRADE)	Relative effect (95% CI)	Anticipated absolute effects
				Risk with No Moisturizer	Risk difference with Moisturizer
Extension of time to flare (days)	231(3 randomized controlled trials)	⨁◯◯◯ VERY LOWa,b,c,d	-	The mean extension was 0 days	MD 13.52 higher(0.05 higher to 26.99 higher)
Risk of relapse after one month of latency	441(3 randomized controlled trials)	⨁⨁◯◯ LOWa,c	RR 0.47(0.31–0.72)	454 per 1,000	241 fewer per 1,000(313 fewer to 127 fewer)
Risk of relapse after two months of latency	441(3 randomized controlled trials)	⨁⨁◯◯ LOWa,c	RR 0.65(0.47–0.91)	589 per 1,000	206 fewer per 1,000(312 fewer to 53 fewer)
Risk of relapse after three months of latency	441(3 randomized controlled trials)	⨁⨁◯◯ LOWa,c	RR 0.63(0.47–0.83)	675 per 1,000	250 fewer per 1,000(358 fewer to 115 fewer)
Rate of remission	377(2 randomized controlled trials)	⨁◯◯◯ VERY LOWa,b,c,d,e	HR 2.68(1.18–6.09)	672 per 1,000	278 more per 1,000(60 more to 327 more)
Reduction in SCORing Atopic Dermatitis (Scoring atopic dermatitis) score	613(4 randomized controlled trials)	⨁◯◯◯ VERY LOWa,b,d,f	-	The mean reduction in Scoring atopic dermatitis was 0	MD 3.46 lower(6.05 lower to 0.87 lower)
Reduction in xerosis	235(3 randomized controlled trials)	⨁◯◯◯ VERY LOWa,e,g	RR 0.72(0.50–1.04)	509 per 1,000	143 fewer per 1,000(254 fewer to 20 more)
Reduction in redness, swelling, oozing/crusting, scratch marks and lichenification	224(2 randomized controlled trials)	⨁⨁◯◯ LOWa,e	-	Not pooled	Not pooled
Reduction in pruritus score	389(2 randomized controlled trials)	⨁⨁◯◯ LOWa,h	-	Not pooled	Not pooled
Reduction in sleep loss score	483(2 randomized controlled trials)	⨁⨁◯◯ LOWa,i	-	Not pooled	Not pooled
Improvement in quality of life	559(3 randomized controlled trials)	⨁◯◯◯ VERY LOWa,k,l	-	Not pooled	Not pooled
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval, MD: Mean difference, RR: Risk ratio, HR: Hazard ratio
Grading of Recommendations, Assessment, Development and Evaluation Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
Explanations a: Downgraded one level for performance, detection and publication bias b: Downgraded one level for inconsistency (serious heterogeneity: I2> 50%) c: Downgraded one level for indirectness (Tiplica et al. allowed the use of topical corticosteroid during remission)32 d: Downgraded one level for imprecision (wide confidence interval) e: Downgraded one level for imprecision (limited sample size) f: Downgraded one level for indirectness (Giordano-Labadie et al., Grimalt et al., and Tiplica et al., allowed the use of topical corticosteroid during remission)29-30, 32 g: Downgraded one level for indirectness (the proportion of patients with xerosis were compared instead of xerosis scores) h: Downgraded one level for indirectness (Bianchi et al. reported changes from baseline score, while Tiplica et al. reported post-study scores between groups)31,32 i: Downgraded one level for indirectness (Tiplica et al. reported post-treatment scores for patients, while Grimalt et al. reported scores for caregivers)29,32 j: Downgraded two levels for inconsistency (Giordano-Labadie et al. and Tiplica et al. reflected improvement in quality of life, while Grimalt et al. did not) 29-30, 32 k: Downgraded one level for indirectness (each included studies utilized different scoring system to assess quality of life)

Risk of relapse after each month of latency

Results from three randomised controlled trials (n=441) show that those who applied moisturizer experienced a lower risk of flare-up compared to control after the first month of latency [Figure 3b, pooled risk ratio 0.47, 95% confidence interval 0.31–0.72, I² 28%].^32-34 The protective effect was present throughout the second and third month but with smaller magnitudes of effect [Figure 3c, pooled risk ratio 0.65, 95% confidence interval 0.47–0.91, I² 35%; Figure 3d, pooled risk ratio 0.63, 95% confidence interval 0.47–0.83, I2 33%, respectively]. We regarded the certainty of evidence for these outcomes as ‘low’ because of lack of investigator, participant and outcome assessment blinding and indirectness of comparison [Table 2].

Close

Figure 3b:

Forest plot for risk of relapse at one month of latency in moisturizer user versus non-user

Export to PPT

Close

Figure 3c:

Forest plot for risk of relapse at two month of latency in moisturizer user versus non-user

Export to PPT

Close

Figure 3d:

Forest plot for risk of relapse at three month of latency in moisturizer user versus non-user

Export to PPT

Rate of remission

Pooled result from Tiplica et al. and Weber et al., with durations of three and six months, accordingly, demonstrated that moisturizer-treated patients were 2.68 times more likely to experience a remission [Figure 3e; n=378;95% confidence interval 1.18–6.09, I² 56%].Although our finding demonstrates a marked difference, Weber et al. has individually shown a rate of flare that is more than double of Tiplica et al. (Weber et al. 45 participants, hazard ratio 4.95, 95% confidence interval 1.62–15.1; Tiplica et al. 335 participants, hazard ratio 2.01 and 95% confidence interval 1.44–2.81, respectively).^33,34 Along with serious heterogeneity, risk of bias, indirectness and imprecision, the certainty of evidence was ranked as ‘very low’[Table 2].

Close

Figure 3e:

Forest plot for rate of remission in moisturizer user versus non-user

Export to PPT

Changes in global disease severity and individual signs and symptoms

When considering global changes in severity, the moisturizer group shows greater improvement in disease status, evidently from a –3.46 points difference in SCORAD from the control [Figure 3f; 4 randomised controlled trials, n=638; 95% confidence interval:6.05––0.87, I² 89%].^29-31,33 The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) certainty was ‘very low’ quality due to glaring risk of bias, inconsistency, indirectness and imprecision [Table 2].

Close

Figure 3f:

Forest plot for differences in SCORing Atopic Dermatitis (SCORAD) index in moisturizer user versus non-user

Export to PPT

Investigator-assessed signs showed no improvements. Pooled analysis reveals a non-significant reduction in risk of xerosis after product use [Figure 3g;3 randomised controlled trials, n=235; pooled risk ratio 0.72 and 95% confidence interval 0.50–1.04, I² 42%].^29,31,32 Grimalt et al. and Giordano-Labadie et al. have described negligible improvements in redness, swelling, oozing/crusting, scratch marks and lichenification, although numerical data were not reported.^29-31,33 These outcomes showed ‘very low’ and ‘low’ certainty of evidence, respectively [Table 2].

Close

Figure 3g:

Forest plot for risk of xerosis in moisturizer user versus non-user

Export to PPT

Moisturizer appeared to alleviate participant-assessed symptoms of pruritus and sleep loss. In Bianchi et al., the moisturizer group demonstrated greater pruritus score changes than the control (–74.6% and –35.9%, respectively), while in Tiplica et al., treatment resulted in lower post-study scores (Dexeryl® 0.42 ± 0.08, P<0.001; Atopiclair® 0.56 ± 0.08, P<0.001; no moisturizer 1.09 ± 0.08). Similarly, for sleep loss, the post-treatment scores were lower in patients and caregivers in Tiplica et al. and Grimalt et al., respectively (Tiplica et al.: Dexeryl® 0.27 ± 0.06, P=0.013; Atopiclair® 0.27 ± 0.06, P=0.014; no moisturizer 0.47 ± 0.06; Grimalt et al. 0.26 ± 0.48 vs. 0.53 ± 0.77, P=0.006).^31,33 The GRADE certainty for both outcomes were ‘low’ due to lack of investigator, participant and outcome assessment, blinding and indirectness of comparison, which prevented pooling of results [Table 2].

Taken together, our findings suggest moisturizer marginally improves atopic dermatitis severity by relieving patient’s subjective illness rather than the objective signs.

Quality of life

Giordano-Labadie et al. and Tiplica et al. have shown that moisturizer users experienced greater, but marginal, score changes in children’s dermatology life quality index and patient-oriented eczema measure compared to control (Giordano-Labadie et al.: –0.8 ± 0.4; P=0.001 and –0.4 ± 1.5; P=0.172, accordingly; and Tiplica et al.: Dexeryl® -3.23 ± 0.10; Atopiclair® -2.00 ± 0.10; no moisturizer 0.72 ± 0.11).^30,33 On the other hand, Grimalt et al. have shown no significant differences in Infant’s Dermatitis Quality of Life Index scores between treated and control groups (P=0.131).³¹ Similar to other outcomes, we assessed the GRADE quality as ‘very low’ due to high risk of bias, conflicting individual results and differences in scoring system used [Table 2].

Limitations

There is a dire need to conduct randomised controlled trials with more robust and standardised designs. We propose that future trials must: (1) include blinding of participant, investigator and outcome assessments; (2) recruit larger sample sizes with longer follow-up periods beyond six months because moisturizers are safe;¹¹(3) provide details on the level of care (i.e. community clinic and secondary care) and time of the year/season in which participants were recruited, as these crucial factors impact how moisturizers are prescribed and used, to ensure improved result generalizability;^38,39 (4) control the strength of topical corticosteroid used between trials or avoid them entirely if the main goal is to assess moisturizer efficacy; (5) utilize a universal scoring system for each outcome to ensure directness and allow inter-trial comparison and quantitative synthesis and (6) further assess minimal clinically important differences for each scoring system, as these values determines the applicability of the product in real-life practices. Not until high quality randomised controlled trials are available could we conclude the efficacy of moisturizer in paediatric atopic dermatitis.