Hair and Sweat Glands in Families With Hypohidrotic Ectodermal Dysplasia: Further Characterization

Objectives  To gather and compare clinical and histologic information from individuals affected by hypohidrotic ectodermal dysplasia (HED) and unaffected controlsubjects and to assess the value of these data in the diagnosis of HED.

Design  Volunteer subjects attending the 20th Annual Family Conference of the National Foundation for Ectodermal Dysplasia answered a questionnaire andperformed a starch-iodide sweat-function test. A subset of the subjects also donated samples of hair and 4-mm punch biopsy specimens of palmar and scalpskin. Specimens from each of these tests were assessed in a blinded fashion. Analysis was performed comparing affected and control subjects for each ofthe following parameters: quantification of eccrine structures in the skin biopsy specimens, analysis of hair sample trichograms for hair shaft defects,and qualitative classification of starch-iodide palm-print sweat-function test results.

Setting  An international conference for families and individuals with ectodermal dysplasias.

Subjects  A total of 40 subjects were included in the final analysis: 15 unaffected control subjects and 25 subjects with HED. Nine affected subjects and 9 unaffectedsubjects donated skin biopsy specimens.

Main Outcome Measure  This study was designed to assess the value of 4 simple tests in supporting the diagnosis of HED.

Results  Investigators were blinded to group during analysis of the test samples. Trichogram examination identified 3 hair shaft abnormalities, with a slightlyhigher prevalence in the affected group: variable shaft thickness, trichorrhexis nodosa, and pili torti. The sensitivity and specificity for each of thesefindings was less than 40%. Starch-iodide paper palm imprints identified a higher likelihood of diminished or absent sweat in the affected group, butthis test had a low sensitivity (44%) and an imperfect specificity (93%). Examination of horizontally sectioned skin biopsy specimens from the palmwere devoid of eccrine structures in a minority of affected subjects (sensitivity, 30%; specificity, 100%). In contrast, scalp biopsy specimens lacked eccrinestructures in the majority of affected subjects (sensitivity, 67%; specificity, 100%). Separate analysis excluding the subjects without apparent eccrine apparatusyielded comparable numbers of eccrine duct s from control and affected groups.

Conclusions  We have defined the value of simple, easily performed tests in the morphological diagnosis of HED. Noninvasive trichogram and sweat testing results can supportthe diagnosis of HED, but they are not sensitive or highly specific; horizontally sectioned 4-mm punch biopsy specimens of the scalp or palms that lack eccrinestructures are diagnostic of HED; scalp biopsy is much more sensitive than palmar biopsy; and a scalp biopsy specimen with detectable eccrine structuressuggests that a patient does not have HED.

Hypohidrotic ectodermal dysplasia (HED) is a rare genetic disorder that is characterized by fine, sparse hair; few, often pointed teeth; and diminishedor absent eccrine function. The inability to sweat puts affected infants at risk for life-threatening and brain-damaging episodes of hyperthermia.¹ Previous studies have used magnification of fingertipsor silicone palm impressions to highlight eccrine ostia and have found a qualitative diminution.² Case reports have also documentedabsent or hypoplastic eccrine glands in vertically sectioned biopsy specimens from the palms and forearms.³ These data havebeen frequently misinterpreted as a total absence of eccrine glands in persons with HED.^4,5 More accurate quantificationof eccrine structures in horizontally sectioned skin biopsy specimens has never been reported in these patients, nor have the results of histologicanalysis of hair shafts, hair follicles, or eccrine structures in the scalp.

Hypohidrotic ectodermal dysplasia is most commonly transmitted in an X-linked recessive fashion; autosomal recessive and dominant transmissionalso occur. All the candidate genes encode proteins in a single signaling pathway. Ectodysplasin-A,⁶ a member of thetumor necrosis factor ligand superfamily, has been identified as the affected protein coded in the Xq12-q13.1 region. Ectodysplasin-A is expressed in normalfetal and adult skin and hair and in adult teeth.⁷ The autosomal recessive form, localized to 2q11-q13, has been associated withdefects in the receptor for ectodysplasin-A, EDAR, a transmembrane protein.⁸ A third genotype was recently identified in families affected by HED and immunodeficiency. This X-linked defect is localized tothe IKKg (NEMO) gene, encoding a protein that is a likely constituent of the ectodysplasin signaling pathway.⁹

Despite recent advances in the genetic basis of this disorder, the diagnosis is still established clinically in the majority of patients. One diagnosticdilemma occurs when a fair-haired infant with a negative family history presents with frequent fevers of unknown origin. Another is when a healthy child presentswith a chief complaint of anhidrosis. Genetic analysis is not routinely available in these cases.

This study was designed to investigate the structure and function of eccrine sweat glands as well as hair shafts and follicles in individuals withHED. The primary goals were to gather clinical information from a relatively large cohort of affected families, to compare the affected subjects with asample of unaffected individuals, and to determine a clinicopathologic correlation that may be helpful in establishing the diagnosis.

Approval for this study was obtained from the institutional review board of Saint Louis University, St Louis, Mo. Subjects were enrolled at the 20thAnnual Family Conference of the National Foundation for Ectodermal Dysplasia, which was held July 20 and 21, 2001, in Collinsville, Ill. All subjects gavesigned, informed consent; were assigned a random number; and completed a questionnaire. The questionnaire included medical history, family history, and specific informationabout relevant symptoms (frequent fevers, overheating, early male pattern baldness, slow childhood hair growth, abnormal tooth development, and cloggednasal and ear secretions). This information was tabulated to determine the status of each subject: affected, carrier/unknown, or control (Table 1). The affected group included subjects who reported a previousspecific diagnosis of HED, a history of frequent fevers and/or overheating, and abnormal hair and/or tooth development. The control subjects were thosewho had no criteria of the affected group and who were unrelated volunteers, relatives with negative results on previous genetic testing, or the fathersof families with affected sons and unaffected daughters. (The results of genetic testing were reported by the subjects. The testing was performed on an investigationalbasis in the laboratory of Jonathan Zonana, MD, University of Oregon Health and Sciences University, Portland.) Subjects who did not fit in either groupwere classified as a carrier/unknown and were excluded from final analysis. A subset of subjects donated hair samples, participated in a noninvasive starch-iodidepaper test to determine sweat function, and provided 4-mm punch biopsy specimensfrom their scalp and palms.

Hair samples were obtained by pull test to bias the sample toward inclusion of more fragile shafts. Scissors-snipped samples were obtained from subjectswhose hair did not break easily. Hair samples were qualitatively examined by light microscopy (trichogram) and digital photomicrography. The starch-iodide paper test was performed to evaluate eccrine function. It was administered by exposing a vinyl-gloved hand to an infrared light source for 5 minutes.The glove was then removed, and the palm was immediately placed on a piece of specially prepared starch-iodide paper. Patterned, purple discolorationmarked sweat production. A commercially available 4-mm punch tool was used to obtain skin biopsy specimens from the hypothenar area of the nondominantpalm and the parietal area of the scalp after injection of local anesthesia (0.1 mL of 1% lidocaine with epinephrine). Hemostasis was achieved with 20%aluminum chloride (Drysol). A stitch using 4-0 nylon suture was placed and secured by slipknots.¹⁰ The wound was dressedwith antibiotic ointment and a clear, waterproof cover. Written wound care instructions were provided. The specimens were placed in formalin. After fixation,they were horizontally bisected at the level of the deep reticular dermis, serially sectioned in horizontal fashion, and processed using standard hematoxylin-eosinstaining. They were then viewed microscopically to quantify the number of eccrine sweat glands and hair follicles and to identify structural malformations.Digital photomicrographs of each specimen were obtained. JPEG files were processed by digitally marking each eccrine and follicular unit using commercially availablesoftware (Adobe PhotoDeluxe; Adobe, San Jose, Calif) to facilitate a more reliable quantification (Figure 1).All data collection was performed in a blinded fashion.

Forty subjects were evaluated. Of these, 15 were unaffected friends or relatives and 25 had HED. In the affected group, there were 17 male subjects,7 female subjects, and 1 person of unknown gender. In the control group, there were 10 male subjects and 5 female subjects. The control group included 8unrelated volunteers, 2 relatives with negative results on previous genetictesting, and 5 fathers of families with affected sons and unaffected daughters.Forty subjects underwent starch-iodide paper sweat test ing, 35 donated hair samples for trichogram, 18 underwent a scalp biopsy, and 20 underwent a palmarbiopsy. Of the 18 scalp biopsy specimens, 9 were from subjects who were previously diagnosed with HED and 9 were from unaffected subjects. Two of the 9 biopsyspecimens from unaffected subjects could not be evaluated because of processing problems; 7 were available for final analysis. Of the 20 palmar biopsy specimens,10 were from affected individuals with HED and 10 were from unaffected individuals.

There were 22 subjects affected with HED and 13 unaffected subjects who donated hair samples. Of the 22 affected subjects, 14 were male, 7 werefemale, and 1 was of unknown gender. Of the 13 unaffected subjects, 9 were male and 4 were female. Trichograms revealed numerous hair shaft abnormalitiesin the subjects with HED as well as in the control subjects. Of the 13 unaffectedsubjects, 3 had variable shaft thicknesses (the diameter of individual hairshafts varied by as much as twice the size of other hair shafts), 1 had pili bifurcati (longitudinal splitting), 1 had trichorrhexis nodosa, and 3 hadpili torti (twisting of the hair shaft). Of the 22 subjects affected withHED, 7 had variable shaft thickness, 1 had pili bifurcati, 3 had trichorrhexisnodosa, and 8 had pili torti. Longitudinal splitting (pili bifurcati) occurred equally in both groups, so statistical analysis was performed in the other3 categories: (1) identification of variable shaft thickness had a sensitivity of 32%, a specificity of 23%, a positive predictive value of 70%, and a negativepredictive value of 40%; (2) identification of trichorrhexis nodosa had a sensitivity of 14%, a specificity of 8%, a positive predictive value of 75%,and a negative predictive value of 39%; and (3) identification of pili torti had a sensitivity of 36%, a specificity of 23%, a positive predictive valueof 73%, and a negative predictive value of 42%.

There were 25 subjects affected with HED and 15 unaffected subjects who performed the starch-iodide paper sweat test. Of the 25 affected subjects,17 were male, 7 were female, and 1 was of unknown gender. Of the 15 unaffected subjects, 10 were male and 5 were female. The test was graded based on 4 subjectivelevels: intense, moderate, minimal, and no sweating (Figure 2). Of the 15 control subjects, 10 had an intense sweating response, 4 had a moderate sweating response, and 1 had a minimal sweatingresponse. Of the 25 subjects with HED, 9 had an intense response, 5 had a moderate response, 10 had a minimal response, and 1 had no sweating. The sensitivity,specificity, and positive/negative predictive values were determined based on the following definitions: a sweat test that was positive for HED was definedas the combination of minimal and no sweating; a negative sweat test result was defined as the combination of intense and moderate sweating (Figure 3). In regard to diagnosing HED, thestarch-iodide paper sweat test has a sensitivity of 44%, a specificity of93%, a positive predictive value of 92%, and a negative predictive value of50%.

To determine the value of the starch-iodide test in predicting the number of eccrine ducts in a palmar biopsy section, the average number of eccrineducts was calculated for the 4 qualitative groups of sweat test results (intense, moderate, minimal, and no sweating). The mean ± SD for subjects withintense sweating was 48 ± 16 (n = 22), for those with moderate sweating it was 36 ± 21 (n = 7), for those with minimal sweating it was 27 ±31 (n = 4), and for those with no sweating it was 0 ± 0 (n = 1) (Figure 4). The wide standard deviations, especially in the groups with moderate or minimal sweating, reflect the poorcorrelation and predictive value of the starch-iodide paper sweat test. The poor predictive value was also demonstrated by a negative sweat test resultin a volunteer-physician control subject who did not donate a palmar biopsyspecimen.

To determine the variability of eccrine duct expression, the average number of ducts was calculated for both the control and affected groups. Thenumber of eccrine ducts in a 4-mm palmar skin sample from 10 unaffected subjects was 42 ± 14 (median, 41; range, 17-63). The number of eccrine ductsin a 4-mm palmar skin sample from 10 affected subjects was 29 ± 26 (median, 28; range, 0-74); excluding the 3 subjects lacking eccrine ducts,the number was 41 ± 20 (Figure 5).

There were 10 subjects with HED and 10 unaffected subjects who performedthe starch-iodide paper sweat test. Of the 10 affected subjects, 6 were maleand 4 were female. Of the 10 unaffected subjects, 7 were male and 3 were female. Palmar biopsy specimens from 3 of 10 affected subjects lacked eccrine structures(all 3 subjects were male), while all 10 of the palmar biopsy specimens fromcontrol subjects showed some eccrine structures (Figure 6). The sensitivity, specificity, and positive/negative predictivevalues were determined based on the following definitions: a palmar biopsyspecimen positive for HED was defined as one with a complete lack of eccrinestructures; a negative palmar biopsy specimen was defined as one with eccrine structures. In regard to diagnosing HED, the palmar biopsy eccrine test hada sensitivity of 30%, a specificity of 100%, a positive predictive value of100%, and a negative predictive value of 41%.

There were 9 subjects with HED and 7 unaffected subjects who performedthe starch-iodide paper sweat test. Of the 9 affected subjects, 6 were maleand 3 were female. Of the 7 unaffected subjects, 7 were male and none were female. Scalp biopsy specimens from 6 (5 male, 1 female) of 9 affected subjectslacked eccrine structures, while all 7 of the scalp biopsy specimens fromcontrol subjects showed some eccrine structures (Figure 7). The sensitivity, specificity, and positive/negative predictivevalues were determined as in the palmar biopsy specimens. In regard to diagnosingHED, the scalp biopsy eccrine test had a sensitivity of 67%, a specificityof 100%, a positive predictive value of 100%, and a negative predictive valueof 70%.

Medical literature suggests that people with HED uniformly do not sweat and have diminished or absent eccrine structures. This study is the firstto demonstrate widely varying sweat function test results and eccrine densities in those affected with HED. We have also defined the value of some simpleclinical screening tests that can be used to support the diagnosis of HED.

Trichogram analysis is a less valuable diagnostic tool than the starch-iodidepaper sweat test, but as a simple and noninvasive examination, a trichogrammay support the diagnosis of HED. Identification of variable shaft thickness,trichorrhexis nodosa, or pili torti will support the diagnosis of HED in 70%to 75% of cases.

A starch-iodide paper sweat test that shows minimal or no sweat production is likely among those with HED. However, the majority of the affected subjectsin the present study exhibited moderate or intense sweating, confirming that demonstrable sweat production does not rule out the condition.

In contrast to widely held belief, quantification of eccrine ducts ina 4-mm palmar skin biopsy specimen is only a valuable diagnostic tool whenidentifiable eccrine structures are completely absent. This finding occurred in 3 of 10 subjects with HED, all 3 of whom were male. Among the affectedsubjects whose palmar biopsy specimens included eccrine ducts, the average number of ducts did not differ from that in control biopsy specimens, andthe SDs in the 2 groups were nearly equal. It is possible that individuals affected with HED simply have an extreme variability in the distribution ofeccrine ducts. However, these results suggest that there are at least 2 subsets of individuals affected with HED: those completely lacking eccrine structuresand those with a normal distribution of them. Recent studies have identified 3 unique loci for HED: 2 X-linked and 1 autosomal recessive.^6,8,9,11 Additionalstudies are warranted to identify a genetic correlation for our observation.

Compared with the palmar biopsy findings, identification of any eccrine structures in scalp biopsy specimens is a more valuable predictor of HED,with a specificity and positive predictive value of 100% and a sensitivity and negative predictive value of 70%. We identified no false-positives, andwe identified 70% true-negatives. Therefore, a scalp biopsy specimen that is bisected at the appropriate level of dermis and that lacks eccrine ductsor glands is highly suggestive of HED. Furthermore, if clinical suspicion is high for HED, and a 4-mm scalp biopsy specimen reveals identifiable eccrinestructures, then there is a 70% likelihood that the individual is not affected.

In the present study, the number of palmar eccrine structures was surprisingly unrelated to the degree of palmar sweating. Of 3 subjects who lacked eccrineducts in a single 4-mm horizonally sectioned palmar biopsy specimen, only 1 had no detectable sweating; the other 2 had minimal sweating. One of thecontrol subjects had minimal sweating and a biopsy specimen that revealed 55 eccrine ducts. Perhaps this poor correlation simply reflects the smallnumber of subjects in the minimal/no sweating groups, or technical difficulties with the starch-iodide test. However, it is likely that many individuals withHED have a normal amount of poorly functioning eccrine structures.

One limitation of the present study is that all of the subjects who donated biopsy specimens were adolescents or adults. Establishing objectivedata for early diagnosis of HED is most valuable for infants and young children, who are at highest risk from the adverse consequences of hyperthermia. However,the eccrine apparatus is fully formed by the third trimester and eccrine density diminishes with age, so we would expect scalp biopsy findings to be equallyreliable in this age group.

In conclusion, we have defined the value of a few simple tests in thediagnosis of HED (Table 2). Noninvasivetrichograms and sweat testing can only support the diagnosis of HED as they are not sensitive or highly specific. The results of quantification of eccrineducts in a 4-mm palmar skin biopsy specimen are similar in affected and controlgroups. A scalp or palmar biopsy specimen that lacks eccrine ducts or glandsis diagnostic of HED, but the findings of a scalp biopsy are much more sensitivethan those of a palmar biopsy. A scalp biopsy specimen that reveals eccrineducts or glands suggests that the patient does not have HED. Furthermore, the scalp is a technically easier site from which to obtain a biopsy specimenand carries a lower risk of problematic scarring. For these reasons, the scalp biopsy should be the preferred test in individuals suspected of having HED.Many recent molecular studies have confirmed the interaction of ectodysplasin and its receptor protein.^8,11 Inthe future, we hope to relate the expression of ectodysplasin to the structure and function of these affected cutaneous appendages.

Correspondence: Elaine Siegfried, MD, Department of Dermatology,Saint Louis University, 1034 S Brentwood Blvd, Suite 600, St Louis, MO 63117(Siegfried@centralderm.com).

Accepted for publication July 30, 2003.

This project was supported by a grant from the National Foundation forEctodermal Dysplasias, Mascoutah, Ill.

We would like to thank Diane Wainwright for her administrative support and Tracie Bryson, MD, Angel Allen, MD, Parul Shah, MD, Amy Cole, MD, AlannaBree, MD, and Sarah Jensen, MD, for their participation at the 20th AnnualFamily Conference of the National Foundation for Ectodermal Dysplasias, Collinsville,Ill.