Journal of Dermatology 2015; 42: 343–352
doi: 10.1111/1346-8138.12783
ORIGINAL ARTICLE
Eosinophilic pustular folliculitis: The transition in sex differences and interracial characteristics between 1965 and 2013 Takashi NOMURA,1,2 Mayumi KATOH,1 Yosuke YAMAMOTO,1,3 Kenji KABASHIMA,1 Yoshiki MIYACHI1 1
Department of Dermatology, Kyoto University Graduate School of Medicine, 2Department of Dermatology, Ijinkai Takeda General Hospital, 3Department of Healthcare Epidemiology, School of Public Health, Kyoto University Graduate School of Medicine, Kyoto, Japan
ABSTRACT Eosinophilic pustular folliculitis (EPF) is characterized by a non-infectious infiltration of eosinophils in the hair follicles. It has three variants: (i) classic EPF; (ii) immunosuppression-associated EPF, which herein is subdivided into HIV-associated (IS/HIV) and non-HIV-associated (IS/non-HIV); and (iii) infancy-associated EPF (I-EPF). The rarity of EPF has hindered our understanding of this entity. To examine the characteristics of EPF, with respect to age, sex, race, and chronology, published in case reports to date, we queried PubMed using the following terms: (“eosinophilic pustular folliculitis” [All Fields] OR “eosinophilic folliculitis” [All Fields]) AND (“1965/1/1” [PDAT]: “2013/12/31” [PDAT]). Additional Japanese cases were collected from Igaku Chuo Zasshi through Ichushi-Web, JDream III, and secondhand quotations from domestic periodicals published in Japan. Proceedings were excluded. The PubMed search produced 275 citations containing 358 cases of EPF (224 men, 132 women, and two of unspecified sex); these cases involved classic EPF (101 Japanese and 81 non-Japanese), IS/HIV (4 Japanese and 85 non-Japanese), IS/non-HIV (4 Japanese and 20 non-Japanese), and I-EPF (4 Japanese and 59 nonJapanese). Ichushi generated an additional 148 citations containing 207 cases of Japanese (148 men and 59 women), which included cases of classic EPF (181 cases), IS/HIV (14 cases), IS/non-HIV (9 cases), and I-EPF (3 cases). There was no sex difference in the classic EPF cases reported between 2003 and 2013, whereas IS/HIV, IS/non-HIV, and I-EPF were predominated by men. There is room for reconsideration of sex differences, particularly with regard to classic EPF. The rarity and specificity of I-EPF in Japan may reflect a state of uncertainty about this entity.
Key words:
chronology, eosinophilic pustular folliculitis, Ofuji’s disease.
INTRODUCTION Eosinophilic pustular folliculitis (EPF) is an inflammatory dermatosis that was originally described by Seiichi Ise and Shigeo Ofuji in 1965 as “a variant of superficial pustular dermatosis” in a female patient.1 Five years later, the term “EPF” was coined in a report describing three male patients.2 An EPF eruption consists of papulopustules that tend to form an annular plaque; the condition is characterized by a prominent eosinophilic infiltrate within and around the pilosebaceous units, often accompanied by microabscesses rich in eosinophils.2,3 This type of EPF is called Ofuji’s disease, classic EPF, or classical EPF.4,5 In this study, we call this type “classic EPF,” based on the classification by Nervi et al.5 Two additional variants of EPF were reported in the 1980s: immunosuppression-associated EPF (IS-EPF) and infancy-associated EPF (I-EPF).5 IS-EPF comprises HIV-associ-
ated eosinophilic folliculitis (HIV-EF) and a similar entity observed in immune-deficient patients without HIV infection. IEPF affects the scalp, unlike classic EPF. Although histological findings of I-EPF are similar to that of classic EPF, the entity of I-EPF is unclear.6 It is believed that all three variants of EPF affect men more frequently than women.5 The male-to-female ratio is approximately 5:1 among Japanese patients with classic EPF.7,8 However, a study from Singapore reported a ratio of 1:1.6 for classic EPF.9 Regardless, the epidemiology of EPF variants has not been studied because of the rarity of EPF. Therefore, the basic features of EPF remain obscure. This situation prompted us to survey all of the literature available on databases to revisit the characteristics of EPF. This study aimed to examine all of the EPF cases cited on PubMed and Ichushi, a Japanese database, between 1965 and 2013 with respect to age, sex, race, and
Correspondence: Takashi Nomura, M.D., Ph.D., Department of Dermatology, Kyoto University Graduate School of Medicine, 54 Shogoin Kawaharacho, Sakyoku, Kyoto 606-8507, Japan. Email: [email protected] Received 26 November 2014; accepted 16 December 2014.
© 2015 Japanese Dermatological Association
343
T. Nomura et al.
chronology. Our study revealed insignificant sex differences in classic EPF and a state of uncertainty in the entity of I-EPF.
Chronology of EPF Chronological changes were plotted for the cumulative number of cases for each sex. Cumulative sex ratios in a year were calculated as:
METHODS Proportion of male patients in a yearðMale%Þ
Subjects We collected publications in which EPF was mentioned that were released between 1965 and 2013. The query keywords employed were “eosinophilic pustular folliculitis,” “eosinophilic pustular dermatosis,” “eosinophilic folliculitis,” “Ofuji’s disease,” and “eosinophilic pustulosis.” We obtained a list of 273 citations by searching for (“eosinophilic pustular folliculitis” [All Fields] OR “eosinophilic folliculitis” [All Fields]) AND (“1965/1/1” [PDAT]: “2013/12/31” [PDAT]). To this list, we added the first case reported by Ise and Ofuji in 1965 and an electronically published article from 2013 that was omitted from a search conducted in early 2014 because it was labeled as “PubMed – in process”. We analyzed the contents of the 275 citations and counted the number of cases accompanying descriptions of clinical findings. Cases other than EPF, such as mycosis fungoides, follicular mucinosis, drug eruption, parasite infestation, or episodic eosinophilic dermatosis of the face, were excluded, bringing the total number of bona fide EPF cases on PubMed to 358. Similarly, we compiled a list of Japanese cases that were cited on Igaku Chuo Zasshi (Ichushi; Japana Centra Revuo Medicina), but not on PubMed. We used Ichushi-Web and the JDream III system to obtain data from between 1983 and 2013. Cases published between 1965 and 1982 were collected manually. We found 148 publications (excluding proceedings) to obtain 208 cases (207 Japanese and 1 Thai) of bona fide EPF.
¼ ðcumulative number of male patients in a yearÞ= ðcumulative number of male and female patients in a yearÞ 100:
Sex ratios over the previous decade were calculated as: Decadal proportion of male patients in a yearðMale%=10YsÞ ¼ ðsum of male patients in the previous decadeÞ= ðsum of male and female patients in the previous decadeÞ 100; where the result was omitted when no cases were reported in over a decade, which led to division by zero.
RESULTS Chronology of publications that mentioned EPF on PubMed The number of annual publications referring to EPF has increased since the first case of EPF was described by Ise and Ofuji in 1965 (Fig. 1).1 In the first decade following 1965, EPF was reported mainly in Japanese dermatological periodicals written in Japanese. Although none of them is included in PubMed, a glimpse of the situation in those days was reported in
Classification and analysis of the cases According to the classification by Nervi et al., EPF is classified into three subtypes: classic, IS-EPF, and I-EPF.5 Classic EPF presents with chronic and recurrent annular clusters of sterile follicular papules and pustules; these are often superimposed on plaques and have a tendency toward central clearing and peripheral expansion without underlying diseases. IS-EPF involves an underlying disease causing immunosuppression, such as HIV infection, hematological diseases, or internal malignancy. In this study, we divided IS-EPF into HIV-associated (IS/HIV) and non-HIV-associated (IS/non-HIV) according to the presence or absence of seropositivity for HIV, respectively. I-EPF is a variant of EPF with the least characterization.6 We specified I-EPF as EPF that caused sterile papulopustules in children—defined as under the age of 10 years in this study —and that was diagnosed histologically as having eosinophildominant infiltration. However, some cases of children were classified as having IS-EPF because of their underlying diseases. We reviewed each case in terms of age at first consultation or at onset of exanthema, sex, race, and chronology (i.e. year of publication).
344
Figure 1. Summary of published works of eosinophilic pustular folliculitis (EPF) on PubMed. There were a total of 275 publications between 1965 and 2013 on PubMed. Query words used were (“eosinophilic pustular folliculitis” [All Fields] OR “eosinophilic folliculitis” [All Fields]) AND (“1965/1/1” [PDAT]: “2013/12/ 31” [PDAT]). Since the first description of EPF by Ise and Ofuji in 1965, the number of publications referring to EPF has increased steadily. The identification of HIV-associated EPF and infancy-associated EPF in 1986 and 1984, respectively, boosted publication rates.
© 2015 Japanese Dermatological Association
Characteristics of EPF
English as, “the 18 cases described up till now . . . were all Japanese, of which, 15 were males and three females, 18– 61 years old.”10 The first non-Japanese case of EPF was reported in 1976 in Sweden,11 followed by the second in Spain in 1977.12 The first reports of I-EPF and HIV-EF appeared in the USA in 1984 and 1986, respectively. Since then, the publication rate of EPF has increased sharply.
Regional distribution of cases on PubMed In the 275 publications on PubMed published between 1965 and 2013, we found 358 cases of bona fide EPF (Table 1). The sum of the cases from Japan (n = 114), Europe (n = 112), and North America (n = 70) exceeded 80% of the total (n = 358). From Asia—other than Japan—52 cases (15% of the total) were reported. There were 182 cases of classic EPF reported from Japan (n = 101; 56% of total), Asia (n = 39; 21%), Europe (n = 31; 17%), North America (n = 9; 5%), and Oceania (n = 2; 1%). There were 89 cases of IS/HIV or HIV-EF reported from Europe (n = 37; 42%), North America (n = 32; 36%), Asia (n = 11; 12%), Japan (n = 5; 6%), Oceania (n = 3; 3%), and Africa (n = 1; 1%). There were 24 cases of IS/non-HIV reported from Europe (n = 17; 71%), Japan (n = 4; 17%), North America (n = 2; 8%), and Asia (n = 1; 4%). There were 63 cases of IEPF reported from Europe (n = 27; 43%), North America (n = 27; 43%), Japan (n = 4; 6%), Africa (n = 2; 3%), South America (n = 2; 3%), and Asia (n = 1; 2%).
Differences between the sexes in EPF cases on PubMed Among the 358 cases on PubMed, there were 224 male patients, 132 female patients, and 2 patients with unspecified sex (Table 2). The proportions of each subclass differed between the sexes. The 224 male cases comprised 97 classic EPF patients (43.3% of 224), 62 IS/HIV patients (27.7%), 15 IS/ non-HIV patients (6.7%), and 50 I-EPF patients (22.3%). The 132 female cases comprised 85 classic EPF patients (64.4% of 132), 26 IS/HIV patients (19.7%), 9 IS/non-HIV patients (6.8%), and 12 I-EPF patients (9.1%). The sex ratio also differed among the subtypes (Table 2). The sex difference was marginal in classic EPF (97 male patients and 85 female patients; 53% vs 47%). However, male patients were more frequently affected than female patients with regard to IS/HIV (62 male patients and 26 female patients;
70% vs 30%), IS/non-HIV (15 male patients and 9 female patients; 62% vs 38%), and I-EPF (50 male patients and 12 female patients; 81% vs 19%).
Differences between Japanese and non-Japanese populations There were 113 Japanese cases on PubMed (55 male and 58 female patients) that showed characteristic tendencies (Table 2). First, 101 cases were classic EPF, comprising 90% of the total. Second, the classic EPF cases showed no significant difference in the sex ratio (48 male and 53 female patients; 48% vs 52%), which is in sharp contrast to previous analyses of Japanese cases published in Japanese periodicals.7,13 Finally, there were only four cases of IS/HIV (four male patients), four cases of IS/non-HIV (two male and two female patients), and four cases of I-EPF (one male and three female patients), which was much fewer than the similar non-Japanese cases. There were 245 non-Japanese cases on PubMed (169 male patients, 74 female patients, and 2 individuals of unspecified sex). There were 81 cases of classic EPF (49 male and 32 female patients; 60% vs 40%), 85 IS/HIV (58 male patients, 26 female patients, and 1 individual of unspecified sex), 20 IS/non-HIV (13 male and 7 female patients), and 59 I-EPF (49 male patients, 9 female patients, and 1 individual of unspecified sex) (Table 2). To summarize, twothirds of non-Japanese EPF cases were either IS/HIV or IEPF. There were 207 Japanese cases on Ichushi (148 male and 59 female patients) that were described in original papers not cited on PubMed (Table 2). There were 181 cases of classic EPF (124 male and 57 female patients; 69% vs 31%), 14 IS/ HIV (14 male patients), 9 IS/non-HIV (8 male patients and 1 female patient), and 3 I-EPF (2 male patients and 1 female patient).
Distribution of age The distribution of age differed between the subtypes (Fig. 2). In classic EPF and IS/HIV, the age of the majority of the patients was between 20 and 50 years old. The age of IS/nonHIV patients ranged from the teens to nearly 80 years old, reflecting the distribution of hematological diseases, such as leukemia and lymphoma. Most I-EPF occurred in patient younger than 2 years of age.
Table 1. Regional variation of EPF cited on PubMed (1965–2013) Region/subtype
Japan†
Asia‡
Europe
Africa
North America
South America
Oceania
Global
Classic IS/HIV IS/non-HIV I-EPF Total
101 5 4 4 114
39 11 1 1 52
31 37 17 27 112
0 1 0 2 3
9 32 2 27 70
0 0 0 2 2
2 3 0 0 5
182 89 24 63 358
EPF, eosinophilic pustular folliculitis; I-EPF, infancy-associated EPF; IS/HIV, HIV-infected immunosuppression-associated EPF; IS/non-HIV, HIV-free immunosuppression-associated EPF. † IS/HIV cases from Japan included four Japanese men and one Filipino woman; the Filipino woman was counted as non-Japanese in the analysis. ‡ Asia includes Asian countries in the Near East, Middle East, South Asia, South-East Asia, and Far East Asia except Japan.
© 2015 Japanese Dermatological Association
345
T. Nomura et al.
Table 2. Published cases of EPF between 1965 and 2013 All†
Male patients
PubMed Total 358 (100%) 224 (100%) Classic 182 (50.8%) 97 (43.3%) IS/HIV 89 (24.9%) 62 (27.7%) IS/non-HIV 24 (6.7%) 15 (6.7%) I-EPF 63 (17.6%) 50 (22.3%) Japanese patients on PubMed Total 113 (100%) 55 (100%) Classic 101 (89.5%) 48 (87.3%) IS/HIV 4 (3.5%) 4 (7.3%) IS/non-HIV 4 (3.5%) 2 (3.6%) I-EPF 4 (3.5%) 1 (1.8%) Non-Japanese patients on PubMed Total 245 (100%) 169 (100%) Classic 81 (33.1%) 49 (29.0%) IS/HIV 85 (34.7%) 58 (34.3%) IS/non-HIV 20 (8.2%) 13 (7.7%) I-EPF 59 (24.0%) 49 (29.0%) Japanese patients on Ichushi Total 207 (100%) 148 (100%) Classic 181 (87.4%) 124 (83.8%) IS/HIV 14 (6.8%) 14 (9.5%) IS/non-HIV 9 (4.3%) 8 (5.4%) I-EPF 3 (1.5%) 2 (1.3%)
(a)
(b)
(c)
(d)
Female patients 132 85 26 9 12
(100%) (64.4%) (19.7%) (6.8%) (9.1%)
58 53 0 2 3
(100%) (91.4%) (0.0%) (3.4%) (5.2%)
74 32 26 7 9
(100%) (43.2%) (35.1%) (9.5%) (12.2%)
59 57 0 1 1
(100%) (96.6%) (0%) (1.7%) (1.7%)
Classic, classic EPF; EPF, eosinophilic pustular folliculitis; I-EPF, infancy-associated EPF; IS/HIV, HIV-infected immunosuppression-associated EPF; IS/non-HIV, HIV-free immunosuppression-associated EPF. † Sex was not specified for two non-Japanese cases on PubMed, one involving a 38-year-old IS/HIV patient and the other involving a 2-weekold I-EPF patient.
Figure 2. Age distribution of eosinophilic pustular folliculitis (EPF). (a) Points indicate the age of Japanese and non-Japanese (Non-J) classic EPF, (b) HIV-infected immunosuppression-associated EPF (IS/HIV), (c) HIV-free immunosuppressionassociated EPF (IS/non-HIV), and (d) infancy-associated EPF (I-EPF). An HIV-infected male patient aged 8 months (0.67 years) was classified as IS/HIV. The bottom and top of the box are the first and third quartiles, and the band inside the box indicates the median. The whiskers indicate the minimum and maximum of the data. †There were no Japanese women with IS/HIV. F, female; M, male.
The 182 patients with classic EPF cited on PubMed ranged in age from 11 to 74 years old, with a median/mean of 33.0/35.4 years. This included 97 male patients whose ages ranged from 12 to 74 years, with a median/mean of 33.0/ 37.0 years, and 85 female patients whose ages ranged from 11 to 72 years, with a median/mean of 32.0/33.5 years (Fig. 2a). There were 101 Japanese classic EPF patients whose ages ranged from 15 to 72 years, with a median/ mean of 32.0/35.9 years. This included 48 male patients whose ages ranged from 15 to 72 years, with a median/ mean of 32.5/37.2 years, and 53 female patients whose ages ranged from 15 to 72 years, with a median/mean of 32.0/ 34.7 years. There were 81 non-Japanese classic EPF patients whose ages ranged from 11 to 74 years, with a median/mean of 33.0/34.8 years. This included 49 male patients whose ages ranged from 12 to 74 years, with a median/mean of 34.0/36.9 years, and 32 female patients whose ages ranged from 11 to 58 years, with a median/ mean of 31.5/31.5 years. The 181 Japanese patients with classic EPF cited on Ichushi ranged in age from 16 and 75 years old, with a median/mean of 36.0/37.2 years. This included 124 male patients whose ages ranged from 16 to 75 years, with a median/mean of 37.5/38.2 years, and 57 female patients whose ages ranged from 17 to 75 years, with a median/mean of 33.0/35.1 years.
Of the 89 patients with IS/HIV cited on PubMed, the age range of 88 patients was from 0.7 (8 months) to 69 years old, with a median/mean of 38.0/38.5 years. This included 61 male patients whose ages ranged from 0.7 (8 months) to 69 years, with a median/mean of 37.0/38.7 years; 26 female patients whose ages ranged from 24 to 51 years, with a median/mean of 39.5/38.1 years; and a 38-year-old patient of unspecified sex (Fig. 2b). The age of a non-Japanese man with IS/HIV was not specified. The four Japanese patients with IS/HIV had ages ranging from 29 to 69 years old, with a median/ mean of 36.5/42.8 years; these men were aged 29, 35, 38, and 69 years old. Of the 85 non-Japanese patients, the age range of 84 patients was from 0.7 (8 months) to 67 years, with a median/mean of 38.0/38.3 years. This included 58 male patients, and the age range for 84 of them was from 0.7 (8 months) to 67 years old, with a median/mean of 37.0/ 38.5 years. There were 26 female with IS/HIV patients whose ages ranged from 24 to 51 years, with a median/mean of 39.5/38.1 years. There was also a 38-year-old of unspecified sex and a non-Japanese man with an unspecified age. The 14 Japanese patients with IS/HIV cited on Ichushi were exclusively men whose ages ranged from 19 to 58 years old, with a median/mean of 43.0/41.1 years old. The 24 patients with IS/non-HIV cited on PubMed ranged in age from 12 to 76 years old, with a median/mean of 45.0/
346
© 2015 Japanese Dermatological Association
Characteristics of EPF
44.7 years. This included 15 male patients whose ages ranged from 12 to 76 years, with a median/mean of 53.0/51.1 years, and 9 female patients whose ages ranged from 12 to 60 years, with a median/mean of 31.0/34.1 years (Fig. 2c). This included four Japanese patients, specifically two male patients aged 41 and 55 years old and two female patients aged 12 and 22 years. This also included 20 non-Japanese patients who ranged in age from 12 to 76 years, with a median/mean of 46.0/ 47.2 years. The 20 non-Japanese patients included 13 male patients whose ages ranged from 12 to 76 years, with a median/mean of 53.0/51.5 years, and 7 female patients whose ages ranged from 25 to 60 years, with a median/mean of 39.0/ 39.0 years old. The nine Japanese patients with IS/non-HIV cited on Ichushi ranged in age from 16 to 74 years old, with a median/mean of 53.0/47.0 years old; eight of these patients were men whose ages ranged from 16 to 74 years, with a median/mean of 55.5/46.9 years, and a 48-year-old woman. The 63 patients with I-EPF cited on PubMed ranged in age from a few hours to 9 years old, with a median/mean of 0.5/ 1.3 years. This included 50 male patients whose ages ranged from 24 h to 8 years old, with a median/mean of 0.5/1.2 years old; 12 female patients whose ages ranged from a few hours to 9 years old, with a median/mean of 0.5/1.8 years old; and a 2-week-old infant of unspecified sex (Fig. 2d). There were four Japanese I-EPF patients whose ages ranged from 2 months to 5 years old, with a median/mean of 2.1/2.4 years old; this included a 2-month-old male patient and three female patients ranging in age from 3 months to 5 years old, with a median/ mean of 4.0/3.1 years old. The 59 non-Japanese patients ranged in age from a few hours to 9 years old, with a median/ mean of 0.5/1.3 years old. This included 49 male patients whose ages ranged from 24 h to 8 years old, with a median/ mean of 0.6/1.3 years old; 9 female patients whose ages ranged from a few hours to 9 years old, with a median/mean of 0.4/1.4 years old; and a 2-week-old infant of unspecified sex. The three Japanese patients with I-EPF cited on Ichushi ranged in age from 4 to 9 years old, with the median/mean of 8.0/7.0 years old, and included an 8-year-old male patient, a 9-year-old male patient, and a 4-year-old female patient.
Chronology of classic EPF It was previously believed that male patients were more susceptible to classic EPF, especially within the Japanese population.5,7,8 Indeed, we found male dominance among Japanese patients with classic EPF in the literature, which was published mainly in Japanese periodicals between 1980 and 2010; of 91 cases of classic EPF identified, 66 involved male patients (73% of the total).13 Similarly, we found that the 181 Japanese cases of classic EPF on Ichushi that appeared between 1965 and 2013 included 124 male patients (69%) (Table 2). However, the 101 Japanese patients with classic EPF on PubMed during the same period included only 48 male patients (48% of 101) (Table 2). This discrepancy prompted us to study the chronological variations in the sex ratios of patients cited on PubMed and Ichushi. We made a set of time lines to assess chronologically the varying degrees of male dominance (Fig. 3). The first
© 2015 Japanese Dermatological Association
Japanese patient with classic EPF, a 42-year-old woman, appeared on PubMed in 1965; the first cases involving male patients, aged 18, 25, and 27 years old, appeared in 1970.1,2 The first non-Japanese patient, a 22-year-old man, was published in 1976.11 In 1986, a 58-year-old woman was reported as the first non-Japanese female patient with classic EPF on PubMed.14 The accumulation of female patients, both Japanese and non-Japanese, was slow until the end of the 20th century on PubMed (Fig. 3a, upper row of the left and middle columns). However, it accelerated in the beginning of the 21st century, resulting in proportions of Japanese and nonJapanese male patients that had declined to
doi: 10.1111/1346-8138.12783
ORIGINAL ARTICLE
Eosinophilic pustular folliculitis: The transition in sex differences and interracial characteristics between 1965 and 2013 Takashi NOMURA,1,2 Mayumi KATOH,1 Yosuke YAMAMOTO,1,3 Kenji KABASHIMA,1 Yoshiki MIYACHI1 1
Department of Dermatology, Kyoto University Graduate School of Medicine, 2Department of Dermatology, Ijinkai Takeda General Hospital, 3Department of Healthcare Epidemiology, School of Public Health, Kyoto University Graduate School of Medicine, Kyoto, Japan
ABSTRACT Eosinophilic pustular folliculitis (EPF) is characterized by a non-infectious infiltration of eosinophils in the hair follicles. It has three variants: (i) classic EPF; (ii) immunosuppression-associated EPF, which herein is subdivided into HIV-associated (IS/HIV) and non-HIV-associated (IS/non-HIV); and (iii) infancy-associated EPF (I-EPF). The rarity of EPF has hindered our understanding of this entity. To examine the characteristics of EPF, with respect to age, sex, race, and chronology, published in case reports to date, we queried PubMed using the following terms: (“eosinophilic pustular folliculitis” [All Fields] OR “eosinophilic folliculitis” [All Fields]) AND (“1965/1/1” [PDAT]: “2013/12/31” [PDAT]). Additional Japanese cases were collected from Igaku Chuo Zasshi through Ichushi-Web, JDream III, and secondhand quotations from domestic periodicals published in Japan. Proceedings were excluded. The PubMed search produced 275 citations containing 358 cases of EPF (224 men, 132 women, and two of unspecified sex); these cases involved classic EPF (101 Japanese and 81 non-Japanese), IS/HIV (4 Japanese and 85 non-Japanese), IS/non-HIV (4 Japanese and 20 non-Japanese), and I-EPF (4 Japanese and 59 nonJapanese). Ichushi generated an additional 148 citations containing 207 cases of Japanese (148 men and 59 women), which included cases of classic EPF (181 cases), IS/HIV (14 cases), IS/non-HIV (9 cases), and I-EPF (3 cases). There was no sex difference in the classic EPF cases reported between 2003 and 2013, whereas IS/HIV, IS/non-HIV, and I-EPF were predominated by men. There is room for reconsideration of sex differences, particularly with regard to classic EPF. The rarity and specificity of I-EPF in Japan may reflect a state of uncertainty about this entity.
Key words:
chronology, eosinophilic pustular folliculitis, Ofuji’s disease.
INTRODUCTION Eosinophilic pustular folliculitis (EPF) is an inflammatory dermatosis that was originally described by Seiichi Ise and Shigeo Ofuji in 1965 as “a variant of superficial pustular dermatosis” in a female patient.1 Five years later, the term “EPF” was coined in a report describing three male patients.2 An EPF eruption consists of papulopustules that tend to form an annular plaque; the condition is characterized by a prominent eosinophilic infiltrate within and around the pilosebaceous units, often accompanied by microabscesses rich in eosinophils.2,3 This type of EPF is called Ofuji’s disease, classic EPF, or classical EPF.4,5 In this study, we call this type “classic EPF,” based on the classification by Nervi et al.5 Two additional variants of EPF were reported in the 1980s: immunosuppression-associated EPF (IS-EPF) and infancy-associated EPF (I-EPF).5 IS-EPF comprises HIV-associ-
ated eosinophilic folliculitis (HIV-EF) and a similar entity observed in immune-deficient patients without HIV infection. IEPF affects the scalp, unlike classic EPF. Although histological findings of I-EPF are similar to that of classic EPF, the entity of I-EPF is unclear.6 It is believed that all three variants of EPF affect men more frequently than women.5 The male-to-female ratio is approximately 5:1 among Japanese patients with classic EPF.7,8 However, a study from Singapore reported a ratio of 1:1.6 for classic EPF.9 Regardless, the epidemiology of EPF variants has not been studied because of the rarity of EPF. Therefore, the basic features of EPF remain obscure. This situation prompted us to survey all of the literature available on databases to revisit the characteristics of EPF. This study aimed to examine all of the EPF cases cited on PubMed and Ichushi, a Japanese database, between 1965 and 2013 with respect to age, sex, race, and
Correspondence: Takashi Nomura, M.D., Ph.D., Department of Dermatology, Kyoto University Graduate School of Medicine, 54 Shogoin Kawaharacho, Sakyoku, Kyoto 606-8507, Japan. Email: [email protected] Received 26 November 2014; accepted 16 December 2014.
© 2015 Japanese Dermatological Association
343
T. Nomura et al.
chronology. Our study revealed insignificant sex differences in classic EPF and a state of uncertainty in the entity of I-EPF.
Chronology of EPF Chronological changes were plotted for the cumulative number of cases for each sex. Cumulative sex ratios in a year were calculated as:
METHODS Proportion of male patients in a yearðMale%Þ
Subjects We collected publications in which EPF was mentioned that were released between 1965 and 2013. The query keywords employed were “eosinophilic pustular folliculitis,” “eosinophilic pustular dermatosis,” “eosinophilic folliculitis,” “Ofuji’s disease,” and “eosinophilic pustulosis.” We obtained a list of 273 citations by searching for (“eosinophilic pustular folliculitis” [All Fields] OR “eosinophilic folliculitis” [All Fields]) AND (“1965/1/1” [PDAT]: “2013/12/31” [PDAT]). To this list, we added the first case reported by Ise and Ofuji in 1965 and an electronically published article from 2013 that was omitted from a search conducted in early 2014 because it was labeled as “PubMed – in process”. We analyzed the contents of the 275 citations and counted the number of cases accompanying descriptions of clinical findings. Cases other than EPF, such as mycosis fungoides, follicular mucinosis, drug eruption, parasite infestation, or episodic eosinophilic dermatosis of the face, were excluded, bringing the total number of bona fide EPF cases on PubMed to 358. Similarly, we compiled a list of Japanese cases that were cited on Igaku Chuo Zasshi (Ichushi; Japana Centra Revuo Medicina), but not on PubMed. We used Ichushi-Web and the JDream III system to obtain data from between 1983 and 2013. Cases published between 1965 and 1982 were collected manually. We found 148 publications (excluding proceedings) to obtain 208 cases (207 Japanese and 1 Thai) of bona fide EPF.
¼ ðcumulative number of male patients in a yearÞ= ðcumulative number of male and female patients in a yearÞ 100:
Sex ratios over the previous decade were calculated as: Decadal proportion of male patients in a yearðMale%=10YsÞ ¼ ðsum of male patients in the previous decadeÞ= ðsum of male and female patients in the previous decadeÞ 100; where the result was omitted when no cases were reported in over a decade, which led to division by zero.
RESULTS Chronology of publications that mentioned EPF on PubMed The number of annual publications referring to EPF has increased since the first case of EPF was described by Ise and Ofuji in 1965 (Fig. 1).1 In the first decade following 1965, EPF was reported mainly in Japanese dermatological periodicals written in Japanese. Although none of them is included in PubMed, a glimpse of the situation in those days was reported in
Classification and analysis of the cases According to the classification by Nervi et al., EPF is classified into three subtypes: classic, IS-EPF, and I-EPF.5 Classic EPF presents with chronic and recurrent annular clusters of sterile follicular papules and pustules; these are often superimposed on plaques and have a tendency toward central clearing and peripheral expansion without underlying diseases. IS-EPF involves an underlying disease causing immunosuppression, such as HIV infection, hematological diseases, or internal malignancy. In this study, we divided IS-EPF into HIV-associated (IS/HIV) and non-HIV-associated (IS/non-HIV) according to the presence or absence of seropositivity for HIV, respectively. I-EPF is a variant of EPF with the least characterization.6 We specified I-EPF as EPF that caused sterile papulopustules in children—defined as under the age of 10 years in this study —and that was diagnosed histologically as having eosinophildominant infiltration. However, some cases of children were classified as having IS-EPF because of their underlying diseases. We reviewed each case in terms of age at first consultation or at onset of exanthema, sex, race, and chronology (i.e. year of publication).
344
Figure 1. Summary of published works of eosinophilic pustular folliculitis (EPF) on PubMed. There were a total of 275 publications between 1965 and 2013 on PubMed. Query words used were (“eosinophilic pustular folliculitis” [All Fields] OR “eosinophilic folliculitis” [All Fields]) AND (“1965/1/1” [PDAT]: “2013/12/ 31” [PDAT]). Since the first description of EPF by Ise and Ofuji in 1965, the number of publications referring to EPF has increased steadily. The identification of HIV-associated EPF and infancy-associated EPF in 1986 and 1984, respectively, boosted publication rates.
© 2015 Japanese Dermatological Association
Characteristics of EPF
English as, “the 18 cases described up till now . . . were all Japanese, of which, 15 were males and three females, 18– 61 years old.”10 The first non-Japanese case of EPF was reported in 1976 in Sweden,11 followed by the second in Spain in 1977.12 The first reports of I-EPF and HIV-EF appeared in the USA in 1984 and 1986, respectively. Since then, the publication rate of EPF has increased sharply.
Regional distribution of cases on PubMed In the 275 publications on PubMed published between 1965 and 2013, we found 358 cases of bona fide EPF (Table 1). The sum of the cases from Japan (n = 114), Europe (n = 112), and North America (n = 70) exceeded 80% of the total (n = 358). From Asia—other than Japan—52 cases (15% of the total) were reported. There were 182 cases of classic EPF reported from Japan (n = 101; 56% of total), Asia (n = 39; 21%), Europe (n = 31; 17%), North America (n = 9; 5%), and Oceania (n = 2; 1%). There were 89 cases of IS/HIV or HIV-EF reported from Europe (n = 37; 42%), North America (n = 32; 36%), Asia (n = 11; 12%), Japan (n = 5; 6%), Oceania (n = 3; 3%), and Africa (n = 1; 1%). There were 24 cases of IS/non-HIV reported from Europe (n = 17; 71%), Japan (n = 4; 17%), North America (n = 2; 8%), and Asia (n = 1; 4%). There were 63 cases of IEPF reported from Europe (n = 27; 43%), North America (n = 27; 43%), Japan (n = 4; 6%), Africa (n = 2; 3%), South America (n = 2; 3%), and Asia (n = 1; 2%).
Differences between the sexes in EPF cases on PubMed Among the 358 cases on PubMed, there were 224 male patients, 132 female patients, and 2 patients with unspecified sex (Table 2). The proportions of each subclass differed between the sexes. The 224 male cases comprised 97 classic EPF patients (43.3% of 224), 62 IS/HIV patients (27.7%), 15 IS/ non-HIV patients (6.7%), and 50 I-EPF patients (22.3%). The 132 female cases comprised 85 classic EPF patients (64.4% of 132), 26 IS/HIV patients (19.7%), 9 IS/non-HIV patients (6.8%), and 12 I-EPF patients (9.1%). The sex ratio also differed among the subtypes (Table 2). The sex difference was marginal in classic EPF (97 male patients and 85 female patients; 53% vs 47%). However, male patients were more frequently affected than female patients with regard to IS/HIV (62 male patients and 26 female patients;
70% vs 30%), IS/non-HIV (15 male patients and 9 female patients; 62% vs 38%), and I-EPF (50 male patients and 12 female patients; 81% vs 19%).
Differences between Japanese and non-Japanese populations There were 113 Japanese cases on PubMed (55 male and 58 female patients) that showed characteristic tendencies (Table 2). First, 101 cases were classic EPF, comprising 90% of the total. Second, the classic EPF cases showed no significant difference in the sex ratio (48 male and 53 female patients; 48% vs 52%), which is in sharp contrast to previous analyses of Japanese cases published in Japanese periodicals.7,13 Finally, there were only four cases of IS/HIV (four male patients), four cases of IS/non-HIV (two male and two female patients), and four cases of I-EPF (one male and three female patients), which was much fewer than the similar non-Japanese cases. There were 245 non-Japanese cases on PubMed (169 male patients, 74 female patients, and 2 individuals of unspecified sex). There were 81 cases of classic EPF (49 male and 32 female patients; 60% vs 40%), 85 IS/HIV (58 male patients, 26 female patients, and 1 individual of unspecified sex), 20 IS/non-HIV (13 male and 7 female patients), and 59 I-EPF (49 male patients, 9 female patients, and 1 individual of unspecified sex) (Table 2). To summarize, twothirds of non-Japanese EPF cases were either IS/HIV or IEPF. There were 207 Japanese cases on Ichushi (148 male and 59 female patients) that were described in original papers not cited on PubMed (Table 2). There were 181 cases of classic EPF (124 male and 57 female patients; 69% vs 31%), 14 IS/ HIV (14 male patients), 9 IS/non-HIV (8 male patients and 1 female patient), and 3 I-EPF (2 male patients and 1 female patient).
Distribution of age The distribution of age differed between the subtypes (Fig. 2). In classic EPF and IS/HIV, the age of the majority of the patients was between 20 and 50 years old. The age of IS/nonHIV patients ranged from the teens to nearly 80 years old, reflecting the distribution of hematological diseases, such as leukemia and lymphoma. Most I-EPF occurred in patient younger than 2 years of age.
Table 1. Regional variation of EPF cited on PubMed (1965–2013) Region/subtype
Japan†
Asia‡
Europe
Africa
North America
South America
Oceania
Global
Classic IS/HIV IS/non-HIV I-EPF Total
101 5 4 4 114
39 11 1 1 52
31 37 17 27 112
0 1 0 2 3
9 32 2 27 70
0 0 0 2 2
2 3 0 0 5
182 89 24 63 358
EPF, eosinophilic pustular folliculitis; I-EPF, infancy-associated EPF; IS/HIV, HIV-infected immunosuppression-associated EPF; IS/non-HIV, HIV-free immunosuppression-associated EPF. † IS/HIV cases from Japan included four Japanese men and one Filipino woman; the Filipino woman was counted as non-Japanese in the analysis. ‡ Asia includes Asian countries in the Near East, Middle East, South Asia, South-East Asia, and Far East Asia except Japan.
© 2015 Japanese Dermatological Association
345
T. Nomura et al.
Table 2. Published cases of EPF between 1965 and 2013 All†
Male patients
PubMed Total 358 (100%) 224 (100%) Classic 182 (50.8%) 97 (43.3%) IS/HIV 89 (24.9%) 62 (27.7%) IS/non-HIV 24 (6.7%) 15 (6.7%) I-EPF 63 (17.6%) 50 (22.3%) Japanese patients on PubMed Total 113 (100%) 55 (100%) Classic 101 (89.5%) 48 (87.3%) IS/HIV 4 (3.5%) 4 (7.3%) IS/non-HIV 4 (3.5%) 2 (3.6%) I-EPF 4 (3.5%) 1 (1.8%) Non-Japanese patients on PubMed Total 245 (100%) 169 (100%) Classic 81 (33.1%) 49 (29.0%) IS/HIV 85 (34.7%) 58 (34.3%) IS/non-HIV 20 (8.2%) 13 (7.7%) I-EPF 59 (24.0%) 49 (29.0%) Japanese patients on Ichushi Total 207 (100%) 148 (100%) Classic 181 (87.4%) 124 (83.8%) IS/HIV 14 (6.8%) 14 (9.5%) IS/non-HIV 9 (4.3%) 8 (5.4%) I-EPF 3 (1.5%) 2 (1.3%)
(a)
(b)
(c)
(d)
Female patients 132 85 26 9 12
(100%) (64.4%) (19.7%) (6.8%) (9.1%)
58 53 0 2 3
(100%) (91.4%) (0.0%) (3.4%) (5.2%)
74 32 26 7 9
(100%) (43.2%) (35.1%) (9.5%) (12.2%)
59 57 0 1 1
(100%) (96.6%) (0%) (1.7%) (1.7%)
Classic, classic EPF; EPF, eosinophilic pustular folliculitis; I-EPF, infancy-associated EPF; IS/HIV, HIV-infected immunosuppression-associated EPF; IS/non-HIV, HIV-free immunosuppression-associated EPF. † Sex was not specified for two non-Japanese cases on PubMed, one involving a 38-year-old IS/HIV patient and the other involving a 2-weekold I-EPF patient.
Figure 2. Age distribution of eosinophilic pustular folliculitis (EPF). (a) Points indicate the age of Japanese and non-Japanese (Non-J) classic EPF, (b) HIV-infected immunosuppression-associated EPF (IS/HIV), (c) HIV-free immunosuppressionassociated EPF (IS/non-HIV), and (d) infancy-associated EPF (I-EPF). An HIV-infected male patient aged 8 months (0.67 years) was classified as IS/HIV. The bottom and top of the box are the first and third quartiles, and the band inside the box indicates the median. The whiskers indicate the minimum and maximum of the data. †There were no Japanese women with IS/HIV. F, female; M, male.
The 182 patients with classic EPF cited on PubMed ranged in age from 11 to 74 years old, with a median/mean of 33.0/35.4 years. This included 97 male patients whose ages ranged from 12 to 74 years, with a median/mean of 33.0/ 37.0 years, and 85 female patients whose ages ranged from 11 to 72 years, with a median/mean of 32.0/33.5 years (Fig. 2a). There were 101 Japanese classic EPF patients whose ages ranged from 15 to 72 years, with a median/ mean of 32.0/35.9 years. This included 48 male patients whose ages ranged from 15 to 72 years, with a median/ mean of 32.5/37.2 years, and 53 female patients whose ages ranged from 15 to 72 years, with a median/mean of 32.0/ 34.7 years. There were 81 non-Japanese classic EPF patients whose ages ranged from 11 to 74 years, with a median/mean of 33.0/34.8 years. This included 49 male patients whose ages ranged from 12 to 74 years, with a median/mean of 34.0/36.9 years, and 32 female patients whose ages ranged from 11 to 58 years, with a median/ mean of 31.5/31.5 years. The 181 Japanese patients with classic EPF cited on Ichushi ranged in age from 16 and 75 years old, with a median/mean of 36.0/37.2 years. This included 124 male patients whose ages ranged from 16 to 75 years, with a median/mean of 37.5/38.2 years, and 57 female patients whose ages ranged from 17 to 75 years, with a median/mean of 33.0/35.1 years.
Of the 89 patients with IS/HIV cited on PubMed, the age range of 88 patients was from 0.7 (8 months) to 69 years old, with a median/mean of 38.0/38.5 years. This included 61 male patients whose ages ranged from 0.7 (8 months) to 69 years, with a median/mean of 37.0/38.7 years; 26 female patients whose ages ranged from 24 to 51 years, with a median/mean of 39.5/38.1 years; and a 38-year-old patient of unspecified sex (Fig. 2b). The age of a non-Japanese man with IS/HIV was not specified. The four Japanese patients with IS/HIV had ages ranging from 29 to 69 years old, with a median/ mean of 36.5/42.8 years; these men were aged 29, 35, 38, and 69 years old. Of the 85 non-Japanese patients, the age range of 84 patients was from 0.7 (8 months) to 67 years, with a median/mean of 38.0/38.3 years. This included 58 male patients, and the age range for 84 of them was from 0.7 (8 months) to 67 years old, with a median/mean of 37.0/ 38.5 years. There were 26 female with IS/HIV patients whose ages ranged from 24 to 51 years, with a median/mean of 39.5/38.1 years. There was also a 38-year-old of unspecified sex and a non-Japanese man with an unspecified age. The 14 Japanese patients with IS/HIV cited on Ichushi were exclusively men whose ages ranged from 19 to 58 years old, with a median/mean of 43.0/41.1 years old. The 24 patients with IS/non-HIV cited on PubMed ranged in age from 12 to 76 years old, with a median/mean of 45.0/
346
© 2015 Japanese Dermatological Association
Characteristics of EPF
44.7 years. This included 15 male patients whose ages ranged from 12 to 76 years, with a median/mean of 53.0/51.1 years, and 9 female patients whose ages ranged from 12 to 60 years, with a median/mean of 31.0/34.1 years (Fig. 2c). This included four Japanese patients, specifically two male patients aged 41 and 55 years old and two female patients aged 12 and 22 years. This also included 20 non-Japanese patients who ranged in age from 12 to 76 years, with a median/mean of 46.0/ 47.2 years. The 20 non-Japanese patients included 13 male patients whose ages ranged from 12 to 76 years, with a median/mean of 53.0/51.5 years, and 7 female patients whose ages ranged from 25 to 60 years, with a median/mean of 39.0/ 39.0 years old. The nine Japanese patients with IS/non-HIV cited on Ichushi ranged in age from 16 to 74 years old, with a median/mean of 53.0/47.0 years old; eight of these patients were men whose ages ranged from 16 to 74 years, with a median/mean of 55.5/46.9 years, and a 48-year-old woman. The 63 patients with I-EPF cited on PubMed ranged in age from a few hours to 9 years old, with a median/mean of 0.5/ 1.3 years. This included 50 male patients whose ages ranged from 24 h to 8 years old, with a median/mean of 0.5/1.2 years old; 12 female patients whose ages ranged from a few hours to 9 years old, with a median/mean of 0.5/1.8 years old; and a 2-week-old infant of unspecified sex (Fig. 2d). There were four Japanese I-EPF patients whose ages ranged from 2 months to 5 years old, with a median/mean of 2.1/2.4 years old; this included a 2-month-old male patient and three female patients ranging in age from 3 months to 5 years old, with a median/ mean of 4.0/3.1 years old. The 59 non-Japanese patients ranged in age from a few hours to 9 years old, with a median/ mean of 0.5/1.3 years old. This included 49 male patients whose ages ranged from 24 h to 8 years old, with a median/ mean of 0.6/1.3 years old; 9 female patients whose ages ranged from a few hours to 9 years old, with a median/mean of 0.4/1.4 years old; and a 2-week-old infant of unspecified sex. The three Japanese patients with I-EPF cited on Ichushi ranged in age from 4 to 9 years old, with the median/mean of 8.0/7.0 years old, and included an 8-year-old male patient, a 9-year-old male patient, and a 4-year-old female patient.
Chronology of classic EPF It was previously believed that male patients were more susceptible to classic EPF, especially within the Japanese population.5,7,8 Indeed, we found male dominance among Japanese patients with classic EPF in the literature, which was published mainly in Japanese periodicals between 1980 and 2010; of 91 cases of classic EPF identified, 66 involved male patients (73% of the total).13 Similarly, we found that the 181 Japanese cases of classic EPF on Ichushi that appeared between 1965 and 2013 included 124 male patients (69%) (Table 2). However, the 101 Japanese patients with classic EPF on PubMed during the same period included only 48 male patients (48% of 101) (Table 2). This discrepancy prompted us to study the chronological variations in the sex ratios of patients cited on PubMed and Ichushi. We made a set of time lines to assess chronologically the varying degrees of male dominance (Fig. 3). The first
© 2015 Japanese Dermatological Association
Japanese patient with classic EPF, a 42-year-old woman, appeared on PubMed in 1965; the first cases involving male patients, aged 18, 25, and 27 years old, appeared in 1970.1,2 The first non-Japanese patient, a 22-year-old man, was published in 1976.11 In 1986, a 58-year-old woman was reported as the first non-Japanese female patient with classic EPF on PubMed.14 The accumulation of female patients, both Japanese and non-Japanese, was slow until the end of the 20th century on PubMed (Fig. 3a, upper row of the left and middle columns). However, it accelerated in the beginning of the 21st century, resulting in proportions of Japanese and nonJapanese male patients that had declined to
