Key Points
Question
What are the prevalence and incidence rates of pediatric alopecia areata (AA) in the US?
Findings
In this cohort study of 5801 children with AA, the PEDSnet database was used to evaluate the prevalence and incidence of pediatric AA across time, age, sex, race and ethnicity, and geographic region. The overall prevalence of pediatric AA was 0.11%; the 11-year overall incidence rate between 2009 and 2020 was 13.6 cases per 100 000 person-years, and incidence rates were significantly higher among children who self-identified as Asian, Black, and Hispanic.
Meaning
This study presents a large epidemiologic description of AA in children, and the results suggest an identification of high-risk demographic subgroups and that new therapeutics and trials should be directed toward these affected groups.
Abstract
Importance
Pediatric alopecia areata (AA) prevalence and incidence data are key to understanding the natural history of this medical disease.
Objective
To determine the prevalence and incidence of AA in a pediatric population across time, age, sex, race and ethnicity, and geographic areas within the US.
Design, Setting, and Participants
In this multicenter cohort study conducted among 5 children’s hospitals, data (January 2009 to November 2020) were collected from a standardized electronic health record (PEDSnet database, version 4.0) to evaluate the incidence and prevalence of pediatric AA. The study cohort included patients younger than 18 years with at least 2 physician visits during which a diagnosis code for AA was recorded, or 1 dermatologist specialty visit for which AA was recorded.
Main Outcomes and Measures
The prevalence denominator population comprised 5 409 919 patients. The incidence denominator population was 2 896 241. We identified 5801 children for inclusion in the AA cohort, and 2398 (41.3%) had 12 months or more of follow-up and were included in the incidence analysis.
Results
Of 5801 patients in the AA cohort, the mean (SD) age was 9.0 (4.5) years, 3259 (56.2%) were female, 359 (6.2) were Asian, 1094 (18.9%) were Black, 1348 (23.2%) were Hispanic, and 2362 (40.7%) were White. The overall prevalence of pediatric AA was 0.11%, and the participants in the AA cohort were more often older, female, and members of a racial and ethnic minority group than the full PEDSnet population. The 11-year overall incidence rate of pediatric AA between 2009 and 2020 was 13.6 cases per 100 000 person-years (95% CI, 13.1-14.2). The incidence rate by age was normally distributed and peaked at age 6 years. Rates were 22.8% higher in female patients than male patients (15.1 cases per 100 000 person-years for females vs 12.3 cases per 100 000 person-years for males). Additionally, incidence rates were highest among Hispanic children (31.5 cases per 100 000 person-years).
Conclusions and Relevance
This cohort study examined the prevalence and incidence rates of pediatric AA in the US across time, age, sex, race and ethnicity, and region from 2009 to 2020, finding a prevalence of 0.11% (doubling during the last decade) and incidence rate of 13.6 cases per 100 000 person-years. Additionally, the results identified Asian and Hispanic children as high-risk demographic subgroups who were shown to be 2 and 3 times more likely, respectively, to receive a diagnosis of AA.
This cohort study examines the prevalence and incidence of alopecia areata in a pediatric population across time, age, sex, race and ethnicity, and geographic areas within the US.
Introduction
Alopecia areata (AA) is an autoimmune disease that leads to nonscarring hair loss. Current epidemiologic descriptions of AA, mostly in adults, vary in estimations of population prevalence, ranging from 0.2% to 2%.1,2,3,4,5,6,7 To our knowledge, there have been few assessments of pediatric AA prevalence and incidence. Prior studies used data from a national registry and a single health system.8,9
It is unclear how pediatric AA prevalence and incidence have changed over time, or how age, race and ethnicity, sex, and geographic region are associated with disease rates. We leveraged PEDSnet (pedsnet.org, which was established to produce evidence by sharing data, expertise, and other resources across several of US children’s hospitals10,11) to produce what is to our knowledge the largest and most accurate epidemiological description of AA in children to date. This information can potentially assist clinicians in advocating for resources for children with AA and payers and health systems in planning for AA therapeutics. This article also underscores which populations have been missed in recent trials and studies.
Methods
Study Setting and Data Source
PEDSnet is a multi-institutional clinical research network containing data from approximately 7 million children from US children’s hospitals. The PEDSnet institutions participating in this study included the Children’s Hospital of Philadelphia (CHOP; Philadelphia, Pennsylvania), Children’s Hospital Colorado (Aurora, Colorado), Nationwide Children’s Hospital (Columbus, Ohio), Nemours Children’s Health System (Delaware only), and Seattle Children’s Hospital (Seattle, Washington). PEDSnet standardizes electronic health record (HER) data from each institution to the PEDSnet common data model, which is based on the Observational Medical Outcomes Partnership common data model.10
Study Cohort
Electronic health record data from January 2009 to November 2020 (PEDSnet database, version 4.0) were used for all analyses. The study cohort included patients with at least 2 physician visits in any clinical setting during which a diagnosis code for AA was recorded, or 1 dermatologist specialty visit for which AA was recorded. To assess the accuracy of AA diagnoses, we (L.C.S. and P.M.) reviewed 100 cases from 1 institution (CHOP) and found a 95% positive predictive value for the diagnosis code. The CHOP institutional review board approved the study, and informed consent was waived because of the retrospective nature of the study and use of deidentified data.
We defined 2 denominator data sets that were used to calculate prevalence and incidence, respectively. The requirement for the prevalence denominator was having 1 or more physician visit during the study period. To be included in the incidence denominator, which was a subset of the prevalence denominator, patients had to have at least 12 months of observation in the database.
Variable Construction
The follow-up period was computed as the time from the first visit coded for AA to the last physician visit coded for AA. The PEDSnet database lists race and ethnicity as distinct fields. Race and ethnicity was self-identified. They were combined in our analysis so that if the patient was Hispanic, the race and ethnicity was Hispanic; otherwise it was based on the race field.
Data Analysis
Prevalence was calculated as the number of AA cases in each category divided by the prevalence denominator. Incidence was calculated by taking the time of observation as the period from the first visit in the incidence cohort to either the end of the time in the study period or the first occurrence of AA after the first visit in the incidence cohort. Patient-days were calculated based on patient birth dates and the dates of visits at the beginning and end of their time in the incidence cohort. We used Stata, version 16 (StataCorp), to perform Shapiro-Wilk tests of normality to evaluate the distribution of our data. We used multivariable logistic regression to identify risk factors for development of AA. Odds ratios and risk ratios were almost identical for occurrence of AA.12
Results
Cohort Formation and Attrition
The prevalence denominator population comprised 5 409 919 patients, and the incidence denominator population was 2 896 241 (Table 1). We identified 5801 children with an AA diagnosis for inclusion in the AA cohort. Of those in the AA cohort, 2398 (41.3%) had 12 months or more of follow-up and were included in the incidence analysis.
Table 1. Characteristics of Prevalence, Incidence, and AA Populations.
| Category | Denominator | AA cohort (n = 5801) | |
|---|---|---|---|
| Prevalence (n = 5 569 768) | Incidence (n = 2 896 241) | ||
| Age, mean (SD), y | 7.0 (11.3) | 8.1 (6.6) | 9.0 (4.5) |
| Median (Q1-Q3) | 3.0 (0.3-10.1) | 6.6 (2.4-12.7) | 8.6 (5.3-12.6) |
| Follow-up, mean (SD), y | 4.7 (5.6) | 3.2 (3.2) | 5.3 (3.9) |
| Median (Q1-Q3) | 2.7 (0.2-7.8) | 2.2 (0.2-5.5) | 5.2 (1.4-8.7) |
| Sex, No. (%) | |||
| Male | 2 820 226 (50.6) | 1 410 294 (48.7) | 2542 (43.8) |
| Female | 2 748 501 (49.3) | 1 485 837 (51.3) | 3259 (56.2) |
| Unknown | 1041 | 110 | 0 |
| Race and ethnicity, No. (%) | |||
| Asian | 218 094 (3.9) | 104 103 (3.6) | 359 (6.2) |
| Black | 900 261 (16.2) | 563 554 (19.5) | 1094 (18.9) |
| Hispanic | 603 390 (10.8) | 319 579 (11.0) | 1348 (23.2) |
| White | 3 094 716 (55.6) | 1 598 707 (55.2) | 2362 (40.7) |
| Othera | 753 307 (13.5) | 310, 298 (10.7) | 638 (11.0) |
Abbreviations: AA, alopecia areata; Q, quarter
Participants self-identified as Other.
Prevalence of Pediatric AA
The overall prevalence of pediatric AA was 0.11% (Table 2). Prevalence by year increased steadily, with an overall 2-fold increase from 2009 to 2019 and then dropped slightly in 2020. A test of normality showed a significant negative skew of prevalence by age, with a higher prevalence among older children. Prevalence was higher in females than in males (0.12% vs 0.09%, respectively), and AA was most prevalent among those who identified as Hispanic (0.23%) and Asian (0.17%). Prevalence across the institutions ranged from 0.06% to 0.13% (Table 2).
Table 2. Prevalence of AA by Year, Age, Sex, Race and Ethnicity, and Institution.
| Category | No. (%) | |
|---|---|---|
| Prevalence, denominator population | AA cohort (prevalence, %) | |
| Overall | 5 409 919 | 5801 (0.11) |
| Calendar year | ||
| 2009 | 1 069 674 | 404 (0.04) |
| 2010 | 1 146 812 | 532 (0.05) |
| 2011 | 1 213 796 | 591 (0.05) |
| 2012 | 1 262 563 | 618 (0.05) |
| 2013 | 1 284 340 | 715 (0.06) |
| 2014 | 1 336 367 | 855 (0.06) |
| 2015 | 1 398 651 | 972 (0.07) |
| 2016 | 1 474 138 | 1081 (0.07) |
| 2017 | 1 521 833 | 1234 (0.08) |
| 2018 | 1 523 480 | 1251 (0.08) |
| 2019 | 1 525 325 | 1262 (0.08) |
| 2020 | 1 465 308 | 942 (0.06) |
| Age, y | ||
| 0 | 1 386 620 | 24 (<0.01) |
| 1 | 1 190 712 | 156 (0.01) |
| 2 | 1 039 495 | 365 (0.04) |
| 3 | 946 301 | 504 (0.05) |
| 4 | 903 014 | 612 (0.07) |
| 5 | 873 733 | 656 (0.08) |
| 6 | 821 140 | 751 (0.09) |
| 7 | 783 673 | 732 (0.09) |
| 8 | 758 231 | 717 (0.09) |
| 9 | 740 902 | 710 (0.10) |
| 10 | 730 241 | 676 (0.09) |
| 11 | 732 614 | 687 (0.09) |
| 12 | 723 896 | 660 (0.09) |
| 13 | 718 831 | 622 (0.09) |
| 14 | 718 058 | 643 (0.09) |
| 15 | 699 624 | 593 (0.08) |
| 16 | 667 465 | 552 (0.08) |
| 17 | 600 055 | 459 (0.08) |
| Sex | ||
| Male | 2 748 160 | 2542 (0.09) |
| Female | 2 660 917 | 3259 (0.12) |
| Race and ethnicity | ||
| Asian | 209 237 | 359 (0.17) |
| Black | 882 687 | 1094 (0.12) |
| Hispanic | 594 553 | 1348 (0.23) |
| White | 3 005 040 | 2362 (0.08) |
| Institution | ||
| CHOP | 1 297 281 | 1738 (0.13) |
| Colorado | 1 033 094 | 1367 (0.13) |
| Nationwide | 884 725 | 705 (0.08) |
| Nemours, Delaware | 583 034 | 581 (0.10) |
| Seattle | 646 369 | 822 (0.13) |
Abbreviations: AA, alopecia areata; CHOP, Children’s Hospital of Philadelphia, Colorado, Children’s Hospital Colorado; Nationwide, Nationwide Children’s Hospital; Nemours, Nemours Children’s Health System; Seattle, Seattle Children’s Hospital.
Incidence of Pediatric AA
The 11-year overall incidence rate of pediatric AA from 2009-2020 was 13.6 cases per 100 000 person-years (Table 3). The incidence rate by age was normally distributed and peaked at age 6 years. Rates were 22.8% higher in females than males. Additionally, rates were highest among Hispanic children, followed by Asian children, Black children, and White children (Table 3).
Table 3. Incidence of AA by Age, Sex, and Race and Ethnicity.
| Category | Incidence denominator population | AA incidence cohort | Patient-years | Incidence per 100 000 patient-years | Adjusted odds ratio (95% CI) | P value |
|---|---|---|---|---|---|---|
| Overall | 2 896 241 | 2398 | 17 576 550.4 | 13.6 | NA | NA |
| Age, y | ||||||
| 1 | 606 880 | 47 | 398 347.4 | 11.8 | 1 [Reference] | 1 [Reference] |
| 2 | 825 490 | 107 | 683 055.8 | 15.7 | NA | NA |
| 3 | 943 240 | 146 | 814 675.5 | 17.9 | ||
| 4 | 1 017 405 | 187 | 890 844.7 | 21.0 | ||
| 5 | 1 062 273 | 174 | 938 296.9 | 18.5 | ||
| 6 | 1 081 318 | 203 | 956 776.6 | 21.2 | ||
| 7 | 1 084 755 | 191 | 960 416.0 | 19.9 | ||
| 8 | 1 076 042 | 163 | 950 593.8 | 17.1 | ||
| 9 | 1 058 171 | 172 | 930 802.6 | 18.5 | ||
| 10 | 1 033 552 | 139 | 906 584.7 | 15.3 | ||
| 11 | 1 005 499 | 135 | 872 232.3 | 15.5 | ||
| 12 | 969 035 | 140 | 836 169.5 | 16.7 | ||
| 13 | 937 548 | 133 | 809 182.6 | 16.4 | ||
| 14 | 912 306 | 125 | 784 689.5 | 15.9 | ||
| 15 | 891 389 | 117 | 768 184.1 | 15.2 | ||
| 16 | 870 095 | 125 | 752 632.2 | 16.6 | ||
| 17 | 841 275 | 94 | 731 037.9 | 12.9 | ||
| Sex | ||||||
| Male | 1 485 837 | 1110 | 9 045 677.1 | 12.3 | 1 [Reference] | 1 [Reference] |
| Female | 1 410 294 | 1288 | 8 530 401.9 | 15.1 | 0.80 (0.74-0.87) | <.001 |
| Race and ethnicity | ||||||
| Asian | 104 103 | 125 | 542 320.0 | 23.1 | 2.02 (1.67-2.44) | <.001 |
| Black | 563 554 | 644 | 3 788 471.9 | 17.0 | 1.73 (1.55-1.92) | <.001 |
| Hispanic | 319 579 | 581 | 1 845 753.7 | 31.5 | 3.07 (2.76-3.42) | <.001 |
| White | 1 598 707 | 854 | 9 645 530.8 | 8.8 | [Reference] | 1 [Reference] |
| Other/unknowna | 310 298 | 194 | 1 754 474.1 | 11.1 | 1.17 (1.00-1.37) | .047 |
Abbreviation: AA, alopecia areata.
Participants self-identified as Other.
Logistic regression analysis to calculate adjusted odds ratios (aORs) showed general agreement with unadjusted incidence data. Males were significantly less likely to receive a diagnosis of AA compared with females (aOR, 0.80; 95% CI, 0.74-0.87; P < .001) (Table 3). Analysis across race and ethnicity revealed significantly increased rates among Asian, Black, and Hispanic populations compared with White children. Hispanic children had the greatest risk (aOR, 3.07; 95% CI, 2.76-3.42), followed by Asian children (aOR, 2.02; 95% CI, 1.67-2.44) and Black children (aOR, 1.73; 95% CI, 1.55-1.92).
Discussion
This cohort study examines the prevalence and incidence rates for pediatric AA in the US using EHR data from 5 large pediatric health systems that participated in PEDSnet, a national clinical research network. The overall prevalence of pediatric AA was 0.11%. The overall incidence rate of AA in children between 2009 and 2020 was 13.6 cases per 100 000 person-years. Prevalence of AA by age peaked at age 9 years and was negatively distributed, with higher prevalence found among older children. A prior study reported that AA prevalence by age followed a bell-shaped pattern with seemingly normal distribution among children, peaking at age 9 years and 12 years for females and males, respectively.9 In our study, the incidence rate of AA by age was normally distributed and peaked at age 6 years.
One of the most remarkable findings in this study was the much higher AA incidence rates in patients who self-identified by race and ethnicity as Hispanic and Asian compared with those who identified as White. While to our knowledge this has not been reported previously for AA, similar findings have been reported recently in systemic lupus erythematosus and autoimmune hepatitis, for which Asian and Hispanic populations had higher incidence and prevalence rates with worse outcomes.13,14,15 The higher incidence rates of AA among these populations in the present cohort study were found across all geographic locations, including in regions with smaller populations of Asian and Hispanic children. Recent and historical national registry data,4,8 as well as epidemiological data, have reported AA incidence and prevalence from a greater than 70% White population. With these new data, efforts should be placed on increasing education in diverse communities to expand awareness about AA and access to care for children living with this disease.
Limitations
There are limitations to this study associated with the use of EHR data. Most importantly, the study depends on the accuracy of the data generated during clinical care. Although the positive predictive value of AA diagnosis was found to be 95.0% among those reviewed, several patients with nonspecified alopecia or other diagnoses were captured by AA diagnostic codes. Second, using the population of children receiving care at PEDSnet member institutions, which were all large and urban pediatric academic centers, may have led to an overrepresentation of children with other medical conditions or complex disease.
Conclusions
To our knowledge, this cohort study represents one of the largest epidemiological descriptions of AA in children. Using PEDSnet, we reported the prevalence and incidence rates of pediatric AA across time, age, sex, race and ethnicity, and region from 2009 to 2020. We found an overall pediatric AA prevalence of 0.11% and incidence rate of 13.6 cases per 100 000 person-years, with prevalence doubling during the last decade. We also identified high-risk demographic subgroups, namely Asian and Hispanic children.
References
- 1.Safavi KH, Muller SA, Suman VJ, Moshell AN, Melton LJ III. Incidence of alopecia areata in Olmsted County, Minnesota, 1975 through 1989. Mayo Clin Proc. 1995;70(7):628-633. doi: 10.4065/70.7.628 [DOI] [PubMed] [Google Scholar]
- 2.Mirzoyev SA, Schrum AG, Davis MDP, Torgerson RR. Lifetime incidence risk of alopecia areata estimated at 2.1% by Rochester Epidemiology Project, 1990-2009. J Invest Dermatol. 2014;134(4):1141-1142. doi: 10.1038/jid.2013.464 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Soh BW, Kim SM, Kim YC, Choi GS, Choi JW. Increasing prevalence of alopecia areata in South Korea. J Dermatol. 2019;46(9):e331-e332. doi: 10.1111/1346-8138.14863 [DOI] [PubMed] [Google Scholar]
- 4.Lee HH, Gwillim E, Patel KR, et al. Epidemiology of alopecia areata, ophiasis, totalis, and universalis: a systematic review and meta-analysis. J Am Acad Dermatol. 2020;82(3):675-682. doi: 10.1016/j.jaad.2019.08.032 [DOI] [PubMed] [Google Scholar]
- 5.Villasante Fricke AC, Miteva M. Epidemiology and burden of alopecia areata: a systematic review. Clin Cosmet Investig Dermatol. 2015;8:397-403. doi: 10.2147/CCID.S53985 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Benigno M, Anastassopoulos KP, Mostaghimi A, et al. A large cross-sectional survey study of the prevalence of alopecia areata in the United States. Clin Cosmet Investig Dermatol. 2020;13:259-266. doi: 10.2147/CCID.S245649 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Alshahrani AA, Al-Tuwaijri R, Abuoliat ZA, Alyabsi M, AlJasser MI, Alkhodair R. Prevalence and clinical characteristics of alopecia areata at a tertiary care center in Saudi Arabia. Dermatol Res Pract. 2020;2020:7194270. doi: 10.1155/2020/7194270 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wohlmuth-Wieser I, Osei JS, Norris D, et al. Childhood alopecia areata—data from the National Alopecia Areata Registry. Pediatr Dermatol. 2018;35(2):164-169. doi: 10.1111/pde.13387 [DOI] [PubMed] [Google Scholar]
- 9.Caldwell CC, Saikaly SK, Dellavalle RP, Solomon JA. Prevalence of pediatric alopecia areata among 572,617 dermatology patients. J Am Acad Dermatol. 2017;77(5):980-981. doi: 10.1016/j.jaad.2017.06.035 [DOI] [PubMed] [Google Scholar]
- 10.Forrest CB, Margolis PA, Bailey LC, et al. PEDSnet: a national pediatric learning health system. J Am Med Inform Assoc. 2014;21(4):602-606. doi: 10.1136/amiajnl-2014-002743 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Forrest CB, Margolis P, Seid M, Colletti RB. PEDSnet: how a prototype pediatric learning health system is being expanded into a national network. Health Aff (Millwood). 2014;33(7):1171-1177. doi: 10.1377/hlthaff.2014.0127 [DOI] [PubMed] [Google Scholar]
- 12.Cummings P. The relative merits of risk ratios and odds ratios. Arch Pediatr Adolesc Med. 2009;163(5):438-445. doi: 10.1001/archpediatrics.2009.31 [DOI] [PubMed] [Google Scholar]
- 13.Izmirly PM, Wan I, Sahl S, et al. The incidence and prevalence of systemic lupus erythematosus in New York County (Manhattan), New York: the Manhattan Lupus Surveillance Program. Arthritis Rheumatol. 2017;69(10):2006-2017. doi: 10.1002/art.40192 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lee B, Holt EW, Wong RJ, et al. Race/ethnicity is an independent risk factor for autoimmune hepatitis among the San Francisco underserved. Autoimmunity. 2018;51(5):258-264. doi: 10.1080/08916934.2018.1482884 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wong RJ, Gish R, Frederick T, Bzowej N, Frenette C. The impact of race/ethnicity on the clinical epidemiology of autoimmune hepatitis. J Clin Gastroenterol. 2012;46(2):155-161. doi: 10.1097/MCG.0b013e318228b781 [DOI] [PubMed] [Google Scholar]