Quality of life in atopic dermatitis in Asian countries: a systematic review

Atopic dermatitis (AD) is a common chronic inflammatory skin condition which impacts psychological wellbeing and social relationships. There have been studies of AD’s impact on quality of life (QoL) in Western countries, but these findings cannot be directly extrapolated to Asian populations with genetic, environmental and cultural differences. Therefore, we aimed to systematically review the literature pertaining to QoL impairment in AD in East and Southeast Asia to characterize the impact of AD on patients and their families, and to identify the factors affecting the degree of QoL impairment. A search of English language papers was conducted on MEDLINE, EMBASE, PSYCInfo, Global Health and Web of Science. Observational studies measuring QoL using single or multi-item instruments in people with self-reported or physician diagnosed atopic dermatitis were included. 27 studies from 29 articles were included and synthesized. There is data documenting QoL impairment in AD sufferers and their families, across a wide range of Asian countries, healthcare settings and ages. Aspects of QoL impacted to a greater extent included symptoms of itch, feelings of embarrassment, and sleep disturbance. Severity of disease affects the degree of impairment of QoL, but there is no apparent link between QoL impairment and patient demographic factors, or other medical factors such as age at diagnosis or duration of illness. Our findings also highlighted the need for clinicians to actively explore the impact of patient’s symptoms, especially in an Asian context where healthcare communications are traditionally doctor-centric.


Introduction
Atopic dermatitis (AD) is a common type of chronic inflammatory skin condition, with a particular subtype associated with a heightened immune response to common allergens [1]. It is characterized by itchy lesions most commonly on the flexural surfaces and affects mainly children.
Quality of life (QoL) is a multidimensional construct encompassing one's mental, physical and social wellbeing. While non-life threatening, atopic dermatitis is associated with QoL impairment; the itching may affect mood and sleep hygiene, and the lesions may cause embarrassment, thus impacting on psychological wellbeing and social relationships [1]. The wellbeing of carers may also be affected, as they often must modify their lifestyle to provide care. In view of its importance, QoL has been included as one of the four core outcome domains to ideally be measured in all atopic dermatitis clinical trials [2]. While QoL of AD patients has been widely studied in Western countries [3][4][5], these findings cannot automatically be extrapolated to the Asian context as genetic, environmental and cultural factors may affect clinical manifestation and prevalence of AD between races [6,7]. It is has been reported that filaggrin null mutations which vary between people of different ethnicities may be positively correlated with the severity of AD [8]. For example, the Filaggrin null mutation c.3321delA is not found in western populations but has been reported in East Asian populations such as Japan, China, Korea, Taiwan and Singapore [6]. Differences in the cultural context may also lead to differences in individual's perception of their disease. In addition the greater involvement of family with the Confucian principle of family-centered care giving [9], may impact on family QoL.
With these differences, it was pertinent to summarize what is known about the impact of AD specifically in the Asian context. Hence, this systematic review aimed to qualitatively synthesize and critique the published literature with two research questions in mind. First, does atopic dermatitis impact QoL in Asian countries, and if so, how and to what extent? Second, what are the determinants of QoL in atopic dermatitis among the Asian population?

Methods
The protocol was first developed using the PRISMA Statement [10] as a guideline. Details of the protocol were registered on PROSPERO and can be accessed at http:// www. crd. york. ac. uk/ PROSP ERO/ displ ay_ record. php? ID= CRD42 01810 6613.
A search strategy was developed with MeSH headings and keywords relevant to the population (people with atopic dermatitis), study design (observational studies), context (people from Asian countries), and outcome (QoL). The search for relevant studies was conducted in Aug 2018 on electronic databases MEDLINE, EMBASE, PSYCInfo, Global Health, and Web of Science (from inception to 21 Aug 2018).
The review process was conducted independently by two of our authors (Huang and Choo). Titles and abstracts of studies were retrieved and screened for their relevance to the research question using the inclusion/exclusion criteria in Table 1. Potentially eligible articles were then accessed, and their full text assessed for eligibility. Full-text articles failing to fulfil the selection criteria were excluded and the reasons documented.
Data extraction were then conducted on full text articles which were eligible, using a pre-piloted form. Data extracted included general information, sample characteristics, study methodology, outcome measurements, and any other significant results or factors not in our pre-determined categories. The form was continuously reviewed and modified to reflect more accurately the information included in studies.

Data analysis
A qualitative synthesis of the findings was carried out and reported as a narrative summary.
Additional analysis was conducted on the Children's Dermatology Life Quality Index (CDLQI) and Dermatology Life Quality Index (DLQI) scores. The CDLQI and DLQI scores were interpreted using validated severity stratifications by Waters et al. [11] and Hongbo et al. [12], respectively, to determine the degree of QoL impairment. Scores for each question in these questionnaires were grouped under six headings, and an aggregate score for each heading was calculated as according to the CDLQI and DLQI instructions for detailed analysis. The aspects of QoL with the three highest scores from each study were then highlighted to identify trends.

Risk of bias assessment
Included studies were assessed by both reviewers for risk of selection bias and information bias using an adapted tool (Appendix) based on Dodoo-Schittko et al. [13], and modified from the Newcastle-Ottawa Scale.

Results
The search yielded 4396 articles after removal of the duplicates. After screening and assessing full text for eligibility, 27 studies from 29 articles were included, as outlined in the PRISMA flow diagram (Fig. 1). The extracted data from the included studies are represented below (Table 2). Of the 27 studies, the countries represented included Singapore (four studies), Malaysia (two studies), Japan (five studies), Korea (six studies), Hong Kong (nine studies) and Taiwan (one study). Most studies were performed on patients under specialist care (dermatology clinics or hospitals), only one study was conducted in a primary care setting [14]. Four studies were population surveys using self-reported diagnoses to identify people with AD [15][16][17][18][19]

QoL impairment in people with atopic dermatitis
QoL was impaired in Asian people with AD. Mean IDQoL scores ranged from 6.8 to 7.7, but there is no validated interpretation for the absolute value of IDQoL [20]. CDLQI and DLQI scores ranged for 4.8-15.2 and 4.8-12.0, respectively, with most studies describing a "moderate" or "very large effect" on QoL as according to the interpretation by Waters et al. [11] and Hongbo et al. [12]. The Short Form Health Survey indicated that AD patients have a statistically significant QoL impairment in both Physical Component Scores (PCS) and Mental Component Scores (MCS) compared to healthy controls (HCs) [15,21], except for one study [22] which showed only impairment in MCS and not PCS. Impairment of QoL was also statistically significant in AD patients compared to healthy controls in two studies using the EQ5D scale [16,18,19], with one study [23] reporting a 30-35% decrease in QoL by AD. Some studies showed a statistically significant QoL impairment in AD patients using Skindex-16 [24] or WHOQoL-26 [25].  Aspects of QoL most affected were "symptoms and feelings" and "daily activities". "Treatment" was least affected

Aspects of QoL affected are in the domains of symptoms and feelings, and sleep
The aspects of QoL affected were specified numerically in only four studies using CDLQI and four studies using DLQI (Table 3), while other studies using these QoL instruments only described the various aspects of QoL affected qualitatively. Symptoms of itch and feelings of embarrassment were the most pertinent aspects of QoL. Furthermore, sleep appeared to be an important aspect of QoL [22,[26][27][28][29], though one study [18,19] showed no difference in sleep duration between AD patients and healthy controls. The findings from the analysis of CDLQI and DLQI was corroborated by Higaki et al. [24], which found that "symptoms" and "feelings" were more impacted than "functioning". One study [30] using Short Form (SF)-36 and one [18,19] using EQ5D also showed similar results, with statistically significant impact in aspects of QoL such as symptoms of "pain" and "discomfort", and mental wellbeing.

QoL impairment in family members of people with AD
AD also impacts the QoL of family members. The Dermatitis Family Impact (DFI) scores ranged from 4.8 to 9.4 [21,28,29,31]. Applying the interpretation proposed by Ricci et al. [3], three studies [21,28,29,32] showed a minor impact, while one study [31] showed no impact in QoL.

Determinants of QoL in people with AD
Of the 15 studies investigating the relationship between severity and QoL in AD, 14 showed a statistically significant correlation between severity and QoL impairment [14, 21-24, 26-29, 31-37]. Only one study observed that moderate/severe had similar levels of QoL impairment to mild AD patients [15]. However, this study acknowledged that their sample size was small and may have had insufficient power. Furthermore, their assessment of severity was self-rated, and had a different severity distribution from a previous study where severity of a similar sample population was measured based on clinical examination [38].
Beyond severity, there were no other clear links between general QoL impairment and patient demographics, such as age and gender, nor other medical factors, such as presence of concomitant atopic conditions, age at diagnosis, duration of illness and family history of atopy. However, a few studies found correlations between QoL and biological measurements, such as skin hydration [35,36], trans-epidermal Both total work productivity impairment (TWPI) and total activity impairment (TAI) scores were significantly correlated with the severity and QoL water loss [36] and several biomarkers [39]. Parent negativity [31,40] was also shown to be associated with a lower QoL. Lastly, steroid phobia, whether from patients [25] or parents [36] was linked to a decreased QoL of AD patients.

Discussion
QoL was impaired in people with atopic dermatitis in Asian countries with average CDLQI and DLQI scores ranging from 4.8 to 15.2 (CDLQI mean of means = 9.1) and 4.8-12.0 (DLQI mean of means = 9.1) respectively, similar to other studies in Denmark (CDLQI = 8, DLQI = 5) [4] and the United States (CDLQI = 5.8, DLQI = 6.6) [5]. A review article also quoted that DLQI values of AD patients ranged from 4.5 to 21.4, with the mean of means being 12.2 [41]. While sociocultural differences between these populations may provide some answers as to differences in QoL impairment between countries, it would be hasty to draw major generalizations regarding specific differences between Asian and non-Asian populations, due to the lack of a head to head comparison between the various aspects of QoL affected in these two populations. Different cultures also vary in the way they interpret questionnaires [42], limiting the comparability between countries even with the same QoL instrument. Aspects of QoL affected were in the domains of symptoms and feelings. In our analysis of the CDLQI, we also found that sleep was particularly affected among children. Indeed AD, like many inflammatory skin conditions such as psoriasis and urticaria, is known to be a highly pruritic disease [43]. This pruritus is enhanced at night, when transepidermal water loss (TEWL) in AD patients is greatest [44], affecting sleep quality through nocturnal awakenings [45]. Thus, clinicians should ensure that these aspects are adequately addressed in the care of their patients.
The nocturnal awakenings by AD patients also disrupted sleep for parents who often engage in co-sleeping (sleeping in the same bed) [46], or when they had to get up to attend to the child, leading to parents having their sleep reduced by a median of 45 and 39 min/night, respectively [47], consequently impacting family QoL.
There is a clear relationship between severity and degree of QoL impairment. The SCORAD and NESS used in included studies have subjective symptoms like itch as a category in its own right [48]. Itching also affected sleep quality and consequently QoL. Erythema was a feature in SCORAD and EASI [48], and a higher amount of erythema could be linked to greater impact in physical appearance and consequently social embarrassment.
Besides severity, there was no clear link between general QoL impairment and other medical or demographic factors. However, there were only a few studies investigating the effect of each factor on the QoL of AD patients. Therefore, it may be premature to conclude that these factors do not affect QoL. There were also other variables, such as "involvement of visible areas of the body", that have been reported to be determinants of QoL impairment in AD, but were not investigated in the studies we identified.
Our findings highlight the need for clinicians to actively explore the impact of patient's symptoms on QoL and consider using self-reported QoL questionnaires in their routine monitoring AD patients. This refinement of practice is especially important in the Asian context where patient-doctor communications are traditionally doctor-centered [9,49], with a focus on symptomatology rather than socio-emotional matters [50].
A limitation of this review is that countries included in this study were not entirely representative of SEA and East Asia; there was a lack of papers from China due to our literature search being confined to English language papers and because of resource constraints not using any Chinese bibliographic databases. Studies from developing countries in Asia were few, and it would be inappropriate to generalize the findings from developed counterparts given the differences in health care, illness behaviors and psychosocial characteristics. There remains a need for research to explore the impact of AD on QoL in these other countries. Furthermore, many papers were unclear about the study's inclusion and exclusion criteria, rendering it difficult to assess selection bias (Table 4). Studies should also be transparent about their sampling methods.
The studies included had many differences in their methods, whether in terms of diagnosis or severity scoring or outcome measures. This made comparisons between studies difficult. The interpretation by Ricci et al. [3] was not validated, which may affect the interpretation of DFI scores. Future research should be geared towards the validation of outcome measures and their interpretations and forming a consensus on the instruments used to measure QoL of AD patients.

Conclusion
QoL is impaired for both Asian AD sufferers and their family. Sufferers are most affected by the itch, sleep disturbance and embarrassment associated with AD. Severity of disease affects the degree of impairment of QoL on AD patients. Greater attention needs to be given to validation of instruments and consistency of their use, and future research should extend to the investigation of QoL on AD patients in other developing countries in Asia.

Description of study population
Inclusion and exclusion criteria should be adequately specific. They should include: 1. Clear definition of atopic dermatitis: papers should state if they used a criteria for diagnosis (e.g. UK Working Party criteria), patient diagnosis or if atopic dermatitis was self reported. 2. Demographics of study population: There should be further inclusion and exclusion criteria on the demographics (e.g. Age) of study population. 3. Setting: Studies should mention the setting in which people with atopic dermatitis were obtained (whether clinical or population based samples).

Categories:
• Clear: All three components (diagnosis, demographics and setting) were specified. • Unclear: two components were specified. • No inclusion and/or exclusion criteria: Only one or no component was specified.

Sampling method
This category assesses if the method of sampling from the study population (as specified by diagnosis, demographics and setting) was truly random and without bias. Categories: • Random: the sample was drawn randomly from study population/study specified or that there was complete enumeration from study population. • Not random: the sample was not drawn randomly. This includes choosing patients based on convenience or any other extrinsic factor. • Unclear: study did not state how the sample was drawn from study population.

Classification of exposure
The exposure in this instance refers to the diagnosis of atopic dermatitis. The method in which the diagnosis of atopic dermatitis was established was assessed. Categories: • Secure records: study should specify that diagnoses were made by physician or from secured medical records. Physician diagnoses can be established through the use of diagnostic criteria or clinical judgment. • Self reported: diagnoses of atopic dermatitis were reported by participants of the study, with no verification with secured records. • Unclear: study did not state how atopic dermatitis was established.

Classification of outcomes
The outcome in this instance refers to quality of life (QoL) of the study sample. The method of measuring QoL was assessed. Categories.
• Validated psychometric instruments: instruments which have been evaluated to have adequate validity and reliability.
• Non-validated psychometric instruments: instruments which have not been shown to have adequate validity and reliability. This includes translation of validated psychometric instruments which in themselves have not been validated. • Self-reported QoL: studies in which QoL was asked directly and self-reported.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in  [15] A, B, C Random (−) Self reported Validated instrument Aziah et al. [32] A, B, C Random (+) follow up > 80% Secure records Validated instrument Bae et al. [52] A, C Not random (−) Secure records Validated instrument Chen et al. [30] A, C Not random (−) Secure records Validated instrument Cheok et al. [16] A, B, C Random (−) Secure records Validated instrument Chuh and Chan et al. [14] A, B, C Not random (−) Secure records Validated instrument Ghani et al. [28,29] A, B, C Random (−) Secure records Validated instrument Higaki et al. [24] A, C Random (+) follow-up > 80% Secure records Validated instrument Ho et al. [21] A, B, C Random (−) Secure records Validated instrument Hon et al. [53] B, C Not random (−) Secure records Validated instrument Hon et al. [39] A, B, C Random (−) Secure records Validated instrument Hon et al. [54] A, C Unclear (−) Secure records Validated instrument Hon et al. [36] A, B, C Random (−) Secure records Validated instrument Hon et al. [55] A, B,C Random (−) Secure records Validated instrument Hon et al. [34] A, B, C Unclear (−) Secure records Validated instrument Hon et al. [35] A, B, C Unclear (−) Secure records Validated instrument Itakura et al. [17] A, B, C Random (−) Self reported Validated instrument Jang et al. [31] A, B, C Unclear (−) Self reported Validated instrument Kawashima et al. [25] A, B, C Unclear (−) Secure records Validated instrument Kim C et al. [23] A, C Unclear Unclear Secure records Validated instrument Kim DH et al. [26] A, C Unclear (−) Secure records Validated instrument Kwak et al. [18], Lee SH et al. [19] A, C Random (−) Self reported Validated instrument Lam et al. [22] A, B, C Not random (−) Secure records Validated instrument Ng et al. [27] A, B, C Unclear (−) Secure records Validated instrument Oh et al. [40] A Unclear (−) Unclear Validated instrument Yano et al. [37] C Unclear (−) Secure records Validated instrument the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.