Clinical examination for hyperlinear palms to determine filaggrin genotype: A diagnostic test accuracy study

Palmar hyperlinearity is a feature of ichthyosis vulgaris, the monogenic skin disorder caused by FLG loss‐of‐function mutations.


| INTRODUC TI ON
Palmoplantar hyperlinearity, keratosis pilaris and ichthyosis are features of ichthyosis vulgaris (MIM #146700) caused by loss-of-function (null) mutations in the gene encoding filaggrin (FLG). 1 FLG null mutations are semi-dominant, meaning individuals with one mutation have a mild phenotype and individuals with two null mutations have more severe ichthyosis. Palmar hyperlinearity was the clinical feature most strongly associated with FLG null genotype in a population-based study of children aged 7-9 years (heterozygote odds ratio 19.3 (95% confidence interval 11.7-31.7)) 2 but hyperlinearity can also occur in FLG wild-type individuals (those with no FLG mutations). 2 In addition to causing the monogenic dry skin condition of ichthyosis vulgaris, FLG null mutations also increase risk of the common complex trait atopic eczema. 3 The strongest and most highly significant effect of FLG null genotype on eczema risk is present in earlyonset, persistent and severe disease 4 associated with multiple other atopic conditions, including asthma and food allergy. [5][6][7] FLG haploinsufficiency is believed to contribute to the pathophysiology of atopic disease by multiple mechanisms in the biochemical, physical and microbial components of skin barrier formation and function. 8 The mechanisms by which FLG null genotype leads to palmoplantar hyperlinearity remain unknown.
Filaggrin deficiency has been targeted therapeutically using emollients containing filaggrin's constituent amino acids 9 and observational studies have reported differences in patient response to immunosuppressive treatment based on FLG genotype. 10 Knowledge of an eczema patient's FLG genotype could therefore be used for current and future personalized medicine strategies, but genotyping is not yet available in routine clinical practice.
We sought to test the hypothesis that examination for the clinical feature of hyperlinear palms (HLP) can be used as a proxy for FLG null genotype (having one or two loss-of-function mutations) in children with eczema or at high risk of atopic eczema.

| ME THODS
This study was conducted using all available data from three paediatric eczema studies: the CLOTHES trial of silk clothing which recruited 300 children aged 1-5 years with moderate-severe atopic eczema 11 ; the BEEP study, which studied 1394 infants at high risk of atopic eczema based on family history, up to 2 years of age 12 ; and a UK-Irish case collection which recruited 4053 children aged 0-16 years with doctor-diagnosed atopic eczema from secondary and tertiary care in the Republic of Ireland, Scotland, England and Northern Ireland. 5,13,14 The presence or absence of HLP or 'unsure' was recorded by research nurses in the CLOTHES and BEEP studies and medical doctors in the UK-Irish case collection. Research nurses received training in the observation of HLP using clinical photographs of the palmar aspects of children's hands, including paediatric ichthyosis vulgaris cases. The teaching material is shown G R A P H I C A L A B S T R A C T Palmar hyperlinearity is associated with FLG loss-of-function mutations. This diagnostic test accuracy study used data previously collected as part of three paediatric cohorts, including a total of 3656 children. We aimed to investigate whether the presence or absence of hyperlinear palms (HLP) could be used to detect FLG genotype in children. Thirty-two percent of participants (1159/3656) had FLG null mutation(s) and 37% (1347/3656) had HLP. The presence of HLP was not a reliable clinical sign for the detection of FLG mutations.

Key Messages
• Examination for palmar hyperlinearity is not a sensitive or specific way to detect filaggrin mutations • However, the absence of HLP can be used to exclude The clinicians and trained observers recording HLP were unaware of the participants' genotypes and similarly the laboratory staff carrying out genetic analysis did not have access to the phenotypic data.
Individuals had been genotyped for the four most prevalent FLG null mutations (R501X, 2282del4, R2447X and S3247X) as part of the previous studies. 5,[11][12][13][14] Individuals with one or two FLG null mutations (heterozygotes, homozygotes and compound heterozygotes) were considered as one group for this analysis, to be compared with the group of individuals with FLG wild-type genotype (having no mutations). Characteristics of the study participants are shown in Table 1. Individuals of white European ethnicity were selected because of knowledge about the prevalent FLG mutations in this population group. 15 All the participants in this study (or their parents or guardians) had given written informed consent as part of the original study consent process for their data and DNA from blood or saliva to be used for future research. 1,11,12 The utility of HLP (assessed as yes or no) as a proxy for FLG null genotype was reviewed using cross-tabulation and investigated by calculation of sensitivity, specificity, likelihood ratios, diagnostic odds ratios, positive and negative predictive values for each study and presented with 95% confidence intervals. For the BEEP study, this was repeated for the subset of children with eczema at 24 months old. The cases where HLP were assessed as unsure were included in a sensitivity analysis as no HLP. We have reported this diagnostic accuracy study, in which the presence/absence of HLP is the index test and FLG genotype is the reference standard, using STARD criteria. 16 These analyses were not pre-specified in the original study protocols.  Table 2). HLP was recorded as 'unsure' or not recorded in 13% of the total combined study population, but 15% in the UK-Irish collection (395/2623) ( Table 2).

TA B L E 1 Characteristics of children included in the FLG and HLP analysis
Cross-tabulation of FLG genotype by HLP in Table 3 shows that HLP are observed in children with and without FLG null mutations in each of the three studies. mild, moderate and severe in the UK-Irish collection). In these studies, the sensitivity and specificity of HLP for FLG null genotype are estimated to be 67% and 72% sensitivity, 75% and 60% specificity respectively (Table 4). In contrast, in the BEEP study, which comprised young children at high risk for atopic eczema, the sensitivity of HLP was only ~46% but the specificity was ~89%. Similar sensitivity and specificity were observed in BEEP in the subset of children who had developed eczema by 24 months of age. Figure 3 displays the sensitivity and specificity for each study.
The positive likelihood ratios in Table 4   Note: 95% confidence intervals are shown in parentheses; note positive and negative predictive values depend on the prevalence of FLG null mutations. Children whose HLP status could not be determined were excluded from analysis. a diagnosis based on UK working party criteria at 24 months of age.  (Tables 5 and 6) which limit the interpretation of these associations.

| Clinical implications
Genetic analysis is likely to increase in availability in the future, but

ACK N OWLED G EM ENTS
We are very grateful to the children and families who took part in these three studies and the clinicians and research teams who recruited patients and participants.

CO N FLI C T O F I NTE R E S T
The funding bodies have had no influence over study design, conduct, analysis or reporting of this study.