Introduction

Atopic eczema (atopic dermatitis, hereafter “eczema”) is a chronic inflammatory skin condition characterized by red, itchy skin lesions [1]. There is a stepwise approach to the treatment of eczema, depending on disease severity and response [2]. Emollients are topical therapies which act to “soothe, smooth and hydrate the skin” [3]. They are the first-line treatment for eczema across all severities and should be used continually with and in larger quantities than other topical treatments [2].

Given the widespread and long-term use of emollients in treating eczema, an understanding of just how common adverse events are associated with emollients is important. However, the reporting of adverse events associated with topical eczema therapies has focused on the use of corticosteroids and calcineurin inhibitors, such as tacrolimus [4, 5]. A recent Cochrane review of emollients used to treat eczema only included adverse event data from randomized control trials (RCTs) [1].

The aim of this study was to conduct a restricted review [6] into adverse events associated with the use of emollients in eczema.

Methods

Literature Search

The Medline database (Ovid) was searched from inception (1946) to June 2018 for adverse events associated with emollient use in patients with eczema. A search strategy was developed using keywords from the electronic database, which combined search terms for eczema, emollients and adverse events (see Electronic Supplementary Material [ESM] Appendix 1). Duplicate publications were removed, and one of the authors (AB) assessed eligibility by screening titles and abstracts against the inclusion and exclusion criteria. All identified papers were then obtained and read in full. References of all included papers were screened for any additional eligible papers. When AB was uncertain about the inclusion or exclusion of a paper, its eligibility was discussed with a second author (MR) and agreement was reached. The reasons for exclusion were recorded.

Eligibility Criteria

Studies investigating “leave-on” emollients for the treatment of participants with atopic eczema/dermatitis of any age were included in this restricted review and included RCTs, cohort studies, case–control studies and case reports/series. Studies excluded were: those not in humans; those not published in English; those in which the emollient was not used for the treatment of eczema or was mixed with other topical therapies for eczema (e.g. topical corticosteroids); and those in which adverse events related to emollient use were not included in the title or abstract.

Data Extraction

One author (AB) developed and piloted a data extraction tool before applying it to all the included papers. The main outcome measure of interest was the frequency and nature of adverse events associated with emollient use. Expected adverse events related to emollient use included pruritus, rash, erythema, pain, burning and hypersensitivity. Type and name of emollient, study type, setting, patient population and disease diagnostic criteria (if any) applied were also recorded.

Compliance with Ethics Guidelines

This article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors, so ethical approval was not required.

Results

Database Search

A total of 897 papers were initially identified; following the removal of duplicate studies, 892 titles and abstracts were screened for eligibility. The full text of 102 papers were then screened, and 80 papers excluded. Two papers were added from the references of the included papers which resulted in a total of 24 papers included in the review (see Fig. 1).

Fig. 1
figure 1

Flowchart of study selection

Study Characteristics

The characteristics of the included papers are summarized in Table 1. Most studies were RCTs (17, 71%) [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23], followed by non-randomized interventional studies (5, 21%) [24,25,26,27,28]. There was one cohort [29] and one case–control study [30]. Most of the papers (13, 54%) did not specify the settings in which the studies were conducted [7, 11, 13, 14, 18,19,20, 22, 23, 25,26,27]; however, ten studies took place in a specialist setting (7 secondary care centers [9, 10, 12, 15, 24, 28, 30], 2 research centers [8, 21] and 1 mobile dermatological center [16]). One study took place in both a secondary care and a research centre [17]. No studies took place in a primary care/community setting (see Table 1).

Table 1 Characteristics of the papers included in the review

There were ten (42%) studies involving just children [10,11,12, 14, 15, 17, 18, 21, 27, 28], five (21%) with only adults [7, 8, 19, 22, 23] and nine (38%) involving both children and adults [9, 13, 16, 20, 24,25,26, 29, 30]. Eleven (46%) studies did not specify how the diagnosis of eczema was confirmed [7, 8, 10, 13, 15, 22, 23, 25, 26, 29, 30]. Six (25%) of the studies used Hanifin and Rajka criteria [9, 14, 18, 19, 21, 27] and four (17%) used the UK Working Party’s Group criteria [11, 12, 17, 24]. The eczema of participants in one (4%) study was diagnosed by a dermatologist [16], in one (4%) study diagnosis was based on Rajka and Langeland criteria [20] and in one (4%) study diagnosis was based on criteria of the Japanese Dermatological Association [23].

Emollient Type and Application

Across all 24 studies included in the review, 29 named emollients were evaluated (see Table 2). The most common types of emollient were creams (15, 52%) [9, 11, 12, 15, 16, 19, 21,22,23,24,25,26, 28, 29], lotions (3, 10%) [7, 17, 18] and ointments (4, 14%) [10, 12, 30, 20]. In one study the emollient was a gel [27], in one study it was a “jelly” [12] and one study explored emollients applied to a coated undershirt [14]. The formulation of four (14%) emollients could not be determined [7, 8, 12, 13].

Table 2 Studies according to emollient type and constituents

When grouping the emollients according to their provenance or humectants (Table 2), three (10%) contained urea [7, 22], five (17%) contained ceramide [8, 24, 25, 28, 29], four (14%) contained glycerol [11, 12, 22, 23] and four (14%) had a herbal basis [14, 16, 25, 30]. Thirteen (45%) were classified as “other” [9, 10, 12, 13, 15, 17, 18, 20, 21, 26, 27].

Adverse Events Associated With Emollients According to Constituents

No serious treatment-related adverse events were reported for any emollient. The proportion of participants experiencing treatment-related adverse events varied between 2.4 and 58.8% [9, 11, 12, 15, 17, 20, 24, 26, 27]. The most common adverse events were skin related and often mild. The range of participants experiencing non-treatment-related adverse events varied between 4.2 and 43.1% [11, 15, 17, 19, 20] (see Table 3).

Table 3 Summary of adverse events associated with emollients

Urea-Containing Emollients

Participants using Iso-Urea or 10% urea lotion [7] reported 21 adverse events not related to the study treatment, including common colds and headaches (Table A1, ESM Appendix 2). The five treatment-related adverse events reported were all related to the skin. There was one withdrawal due to adverse events with participants using Iso-Urea (erythema on eczema plaques) and two withdrawals due to adverse events with participants using 10% urea lotion (irritant contact dermatitis and pruritus). A study into urea cream [22] only reported on adverse skin reactions, which included smarting, stinging, itching and dryness/irritation. No serious adverse events were reported for any of the urea-containing emollients.

Ceramide-Containing Emollients

Of the five ceramide-containing emollients evaluated, four (Cetaphil Restoraderm Body Moisturizer [8], NeoCera [25], CeraVe Moisturizing Cream [29], TriCeram Cream [28]) were not associated with any no treatment-related adverse events (Table A2, ESM Appendix 2). Seven participants using Curel Moisture Cream [28] experienced pruritus and one experienced a warm sensation after application. There was one withdrawal due to adverse events from the use of Curel Moisture Cream (worsening of rashes and acneiform papules on the face). No serious adverse events were reported for any of the ceramide-containing emollients.

Glycerol-Containing Emollients

Of the four glycerol-containing emollients evaluated, treatment-related adverse events were only not reported by participants using N-acetyl-hydroxyproline cream (AHYP) [23] (Table A3, ESM Appendix 2). Three participants experienced treatment-related adverse events using Dexeryl [11] (mild to moderate erythema, burning, pruritus); one participant using glycerine/petroleum (and baby oil as a soap substitute) [12] experienced itching, although this was perceived to be with associated with the baby oil. Participants using glycerin cream [22] only reported on adverse skin reactions, which included smarting, stinging, itching and dryness/irritation. Thirty-one participants experienced non-treatment-related adverse events when using Dexeryl [11]. No serious adverse events were reported for any of the glycerol-containing emollients.

Herbal Emollients

Treatment-related or non-treatment-related adverse events were not reported for three of the four herbal emollients evaluated: Dermovix [30], borage oil coated on undershirts [14] and 10% chamomile hydro-alcoholic extract cream [16] (Table A4, ESM Appendix 2). In a study on Hamamelis distillate [19], five participants reported experiencing itching, erythema, stinging, lichenification or dry skin. Across the two treatment groups, there were three non-treatment-related adverse events (herpes simplex infection, cystitis and bronchitis). No serious adverse events were reported for any of the herbal emollients.

Other Emollients

Three of the 13 emollients classified as “other” (10% sodium cromoglycate ointment [10], emulsifying ointment [12], Eucerin Soothing Lotion 12% omega [18]) did not result in any treatment-related adverse events (Table A5, ESM Appendix 2). The range of treatment-related adverse events reported was from 5.0 to 58.8%. Three participants experienced transient pain after the application of a triclosan-containing emollient [9]. One participant using Cetomacrogrol (and baby oil as a soap substitute) [12] and one participant using petroleum jelly (and baby oil as a soap substitute) [12] experienced itching; however this was perceived to be associated with the baby oil. One participant withdrew due to a treatment-related adverse event (erythema after use) when using Atobarrier Cream [26]. Ten participants experienced mild post application skin dryness after the use of Atrapro Antipruritic HydroGel [27]. The only treatment-related adverse events reported after use of CHD-FA 3.5% (carbohydrate-derived fulvic acid) emollient was a short-lived burning sensation [13]. The use of Verum [15] resulted in two withdrawals due to adverse events (itch, erythema and spreading of the eczema). For Altoderm lotion [17] there were seven reports of treatment-related adverse events (erythema and pruritus and application site burning) and one report of an allergic reaction (stinging, itching, redness, skin evolved into eczema). Fifteen participants reported treatment-related adverse events (sensation of burning at the site of application, pruritus and erythema) when using an ointment base containing propylene carbonate [20]. Ten adverse events were reported by participants who had used MAS063DP (Atopiclair) Cream [21].

Participants using seven of the emollients (10% sodium cromoglycate ointment [10], Cetomacrogrol [12], Emulsifying ointment [12], petroleum jelly [12], Atrapro Antipruritic HydroGel [27], CHD-FA 3.5% [13], Eucerin Soothing Lotion 12% omega [18]) did not have any non-treatment-related adverse events. The range of non-treatment-related adverse events was between 6 and 43%. Three participants using Verum [15] reported non-treatment-related adverse events (bronchitis, worsened teething and concussion), and 22 participants using Altoderm Lotion reported non-treatment-related adverse events [17]. Seven participants reported a non-treatment-related adverse event (exacerbation of atopic dermatitis) when using an ointment base containing propylene carbonate [20]. Fifty non-treatment-related adverse events were reported by participants who used MAS063DP (Atopiclair) Cream [21].

No serious treatment-related adverse events were recorded in the “other” emollients. Boguniewicz et al. [21] reported two serious (as determined by the author) non-treatment-related adverse events (tonsillectomy and acute asthma exacerbation); however, it was not possible to determine whether these occurred in the treatment group (MAS063DP [Atopiclair] Cream) or the vehicle group.

Discussion

Summary of Main Findings

Across the 24 studies identified, 29 emollients were evaluated. Most of these 24 studies were RCTs, and there was a mix of studies involving children, adults or both. The location of many studies was unspecified, but the most common setting was specialist care. The criteria for diagnosing eczema was also not stated in many of the papers. The most common type of emollient was cream, with the formulation of some not described.

No serious treatment-related adverse events were reported for any of emollients. The proportion of participants experiencing treatment-related adverse events varied between 2.4 and 58.8% [9, 11, 12, 15, 17, 20, 24, 26, 27], and the most common adverse events were skin related and often mild. The range of participants experiencing non-treatment-related adverse events varied between 4.2 and 43.1% [11, 15, 17, 19, 20].

Strengths and Limitations

We believe this our review is the first review with the specific aim to quantify and compare the frequency of adverse events between different emollients across all types of published studies. We followed a “restricted review” approach [6] and were able to rapidly identify and summarize relevant international publications involving all types of studies. However, due to resource limitations we did not attempt to apply any criteria to assess the quality of the research that we identified. We found that few studies reported on adverse events and of those that did, data was generally of poor quality. It was difficult to interpret results and compare emollients due to missing data, confusion over whether adverse events referred to the treatment or control and inconsistency in reporting by subject or total numbers of the adverse event. Several studies compared the “active” emollient to a “vehicle,” but the extent to which such a comparator is “inert” is arguable. We adopted the descriptions of the emollients given by the authors of the included papers at face value and summarized these accordingly, such as, for example, “urea containing,” “herbal”. As a result, it is possible that some emollients may have been “misclassified”; however, reporting of the original studies means we were unable to better delineate similarities and differences between the emollients.

Limitations of this review itself are the searching of only one database, exclusion of papers not written in English and inability to retrieve the full text of some articles which potentially could have been included. Restrictions to the search criteria (“eczema” and “adverse events”) will have excluded a number of potentially relevant studies in which emollients were evaluated in people without eczema and in those where data on adverse events are reported in the manuscript but not the title or abstract. This limitation has been confirmed by the omission of some papers from the results of the search strategy which were included in the recent Cochrane Review [1]. Only one reviewer screened all the titles/abstracts and retrieved papers, but in cases of uncertainty, eligibility for inclusion was discussed with another author.

The wide variation observed in the proportion of participants experiencing an adverse event across the included studies may reflect a true difference between the different emollients, the population studied, how different researchers collected adverse event data and/or the number of participants in the study. In other words, we have presented adverse events as proportions of the total number of participants in each study, but the smaller the denominator, the greater the potential error around that estimate.

Comparison with Existing Literature

In common with previous systematic reviews that have included emollients for eczema, we found that adverse reporting in studies of emollients is poor and when reporting is present, the data suggest that in general these events are of low frequency and mild [1, 31]. Van Zuuren et al. [1] reported that the relative risks of an adverse event from all emollients was 24 per 100 (95% confidence interval 19–30) participants compared with 23 per 100 with vehicle, placebo or no moisturizer (10 RCTs, 1275 participants, follow-up range from 4 weeks to 6 months) but that there were more adverse events associated with urea-containing creams (65 per 100 participants in 1 RCT in which 129 participants were compared with placebo) and oat-containing moisturizers (9 per 100 participants in 1 RCT in which 173 participants were compared with no moisturizer).

Conclusions

Due to the limitations of both this review and the available literature, our findings should be interpreted with caution. Nevertheless, they provide a basis for further detailed research, with recommendations for a full systematic review of adverse effects associated with emollient use and better design and reporting of future studies that evaluate emollients, including adverse events.

Based on the findings of this restricted review, clinicians and patients can be reassured that the emollients studied appear to be generally safe to use. However, the emollients described in these studies are not in common use in many countries. Until further data become available, patients and clinicians should be encouraged to report adverse events via their national systems, such as the Yellow Card scheme [32].