Risk of fractures in vitiligo patients treated with phototherapy—A retrospective population-based cohort study

G Model DESC 2965 No. of Pages 7

Journal of Dermatological Science xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Journal of Dermatological Science journal homepage: www.jdsjournal.com

Risk of fractures in vitiligo patients treated with phototherapy—A retrospective population-based cohort study Po-Ching Loa, 1,2 , Cheng-Yuan Lib,c,d,1,3 , Weng-Foung Huanga,2,* , Yi-Wen Tsaia,2 , Han-Nan Liub,d,e,3 , Yun-Ting Changb,d,3 , Hsien-Yi Chiuf,4 , Tsen-Fang Tsaig,h,4,** a

Institute of Health and Welfare Policy, National Yang Ming University, Taipei, Taiwan Department of Dermatology, National Yang Ming University, Taipei, Taiwan Graduate Institute of Clinical Medicine, National Taiwan University College of Medicine, Taipei, Taiwan d Department of Dermatology, Taipei Veterans General Hospital, Taipei, Taiwan e National Defense Medical Center, Taipei, Taiwan f Department of Dermatology, National Taiwan University Hospital Hsin-Chu Branch, Hsinchu, Taiwan g Department of Dermatology, National Taiwan University Hospital,Taipei, Taiwan h Department of Dermatology, National Taiwan University College of Medicine, Taipei, Taiwan b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 26 September 2015 Received in revised form 12 February 2016 Accepted 26 February 2016

Background: Phototherapy might increase bone mineral density. However, it is unknown whether phototherapy can reduce the risk of fractures in patients with vitiligo. Objectives: To investigate the effect of phototherapy on fracture risks in vitiligo patients aged 40 or older. Methods: This population-based cohort study used the 2000–2010 Taiwan National Health Insurance Research Database (NHIRD) to identify 3863 patients newly diagnosed with vitiligo between 2003 and 2009 at age 40 years. Study subjects were classified into three cohorts: (1) frequent phototherapy; (2) infrequent phototherapy; and (3) no phototherapy. Patients were followed until the first hip or vertebral fracture or 31 December 2010. Data were analysed using Cox regression models and also stratified by age and gender. Results: Frequent phototherapy decreased the fracture risks (adjusted hazard ratio (aHR) = 0.32, p = 0.009) in vitiligo patients. Stratification by age and gender confirmed the fracture prevention effect of frequent phototherapy in patients aged 40–64 years (aHR = 0.14, p = 0.016) and in female patients (aHR = 0.31, p = 0.024). Conclusions: This study provides the first evidence that frequent phototherapy can reduce the risk of fractures among middle-aged and among female vitiligo patients. ã 2016 Published by Elsevier Ireland Ltd on behalf of Japanese Society for Investigative Dermatology.

Keywords: Fracture Phototherapy Vitiligo Gender National Health Insurance Research Database

1. Introduction

Abbreviations: aHR, adjusted hazard ratio; CI, confidence interval; DDD, defined daily dose; HR, hazard ratio; ICD-9-CM, International Classification of Disease, Ninth Revision, Clinical Modification; NHI, National Health Insurance; NHIRD, National Health Insurance Research Database; SD, standard deviation. * Corresponding author at: Institute of Health and Welfare Policy, National Yang Ming University, No. 155, Sec. 2, Linong St., Beitou District, Taipei, Taiwan. ** Corresponding author at: Department of Dermatology, National Taiwan University Hospital, No.7 Chung San South Road, Taipei, Taiwan. E-mail addresses: [email protected], [email protected] (W.-F. Huang), [email protected] (T.-F. Tsai). 1 These authors contributed equally to this work. 2 No. 155, Sec. 2, Linong Street, Taipei, Taiwan. 3 No. 201, Sec. 2, Shipai Road, Taipei, Taiwan. 4 No. 7 Chung San South Road, Taipei, Taiwan.

Vitiligo is caused by selective loss of epidermal melanocytes that results in patches of white skin. Worldwide prevalence of vitiligo is 1–2% overall, but this varies between different populations. Those affected can suffer devastating psychological distress [1]. Vitiligo affects women younger and more prevalent than men, which may reflect an aetiologic association with underlying autoimmune dysregulation to which females are more susceptible, or, indirectly, greater concerns to the negative effect on the appearance [1]. Diverse mechanisms that may contribute to the aetiology of vitiligo have been proposed; these include autoimmune, genetic, biochemical, oxidative stress, neural, and viral factors [2]. Vitiligo has also been associated with other autoimmune diseases, for example, autoimmune thyroid disease, rheumatoid arthritis,

http://dx.doi.org/10.1016/j.jdermsci.2016.02.010 0923-1811/ ã 2016 Published by Elsevier Ireland Ltd on behalf of Japanese Society for Investigative Dermatology.

Please cite this article in press as: P.-C. Lo, et al., Risk of fractures in vitiligo patients treated with phototherapy—A retrospective populationbased cohort study, J Dermatol Sci (2016), http://dx.doi.org/10.1016/j.jdermsci.2016.02.010

G Model DESC 2965 No. of Pages 7

2

P.-C. Lo et al. / Journal of Dermatological Science xxx (2015) xxx–xxx

diabetes mellitus, psoriasis, and lupus erythematosus, among others [3,4]. Notably some such vitiligo-associated diseases, for instance, rheumatoid arthritis, are also potential risk factors for osteoporosis and fractures [5–7]. Various treatment options that have been used include topical corticosteroids, calcineurin inhibitors, vitamin D3 analogues (calcipotriene), systemic steroid, anti-oxidants, phototherapy, and surgery [2]. Topical calcineurin inhibitors and topical corticosteroids are the common first-line therapy, with phototherapy, particularly narrowband ultraviolet light B, reserved as second-line therapy for patients with extensive vitiliginous lesions or lesions on exposed areas after first-line therapy fails [8]. Some studies have found vitiligo patients to have lower serum vitamin D levels; however, this remains controversial [9,10]. Others have reported increased serum vitamin D in vitiligo patients treated with phototherapy [11]. Since vitamin D plays a significant role in calcium and bone mineral metabolism [12], it is rational to hypothesize that vitiligo patients treated with phototherapy may have a reduced risk of fractures, consequent to elevated vitamin D levels. Although previous studies have reported increased serum 25-hydroxyvitamin D3 synthesis and bone mineral density in elderly psoriatic women treated with phototherapy [13,14], none have investigated whether phototherapy could actually reduce fracture risks. Therefore, we conducted a retrospective populationbased cohort study to analyse the incidence of primary hip or vertebral fracture among vitiligo patients aged 40 years or older and to determine whether phototherapy can reduce their risk of fractures. 2. Materials and methods 2.1. Study design We classified eligible patients into three cohorts, based on the frequency of phototherapy (Fig. 1), and follow up their outcome. We defined frequency of phototherapy as the accumulated times of phototherapy divided by follow-up period. This “average”

frequency of phototherapy was used because the patients had heterogeneous observational period, from one year to eight years. A frequent phototherapy cohort comprised patients who received phototherapy (national reimbursement codes, phototherapy: 51019B) at least 24 times per year on average; those who received phototherapy between 1 and 23 times per year on average comprised the infrequent phototherapy cohort; the control cohort received no phototherapy. We defined receiving at least 24 times of phototherapy per year on average as frequent phototherapy since Bogh et al.’s study indicated that ultraviolet B exposure of 1 standard erythema dose every second week is sufficient for maintaining summer vitamin D levels during the winter [15]. This operational definition resulted in a frequent phototherapy cohort receiving more phototherapies than Bogh et al. suggested (a ultraviolet B dose every second week), theoretically. The reimbursement code of phototherapy (51019B) in the NHIRD dataset covers mainly narrowband UVB, and does not cover PUVA, lower level laser, and excimer laser therapy. The entry date for each cohort was the first date of vitiligo diagnosis. Subjects were followed-up until an incident of (new) hip or vertebral fracture, withdrawal from the NHI program, or 31 December 2010, whichever came first. Fig. 2 shows the study design. 2.2. Data source All data were obtained from the Taiwan National Health Insurance (NHI) Research Database (NHIRD) from 2000 to 2010. The Taiwan NHI program covers 99% of the Taiwan population and offers comprehensive medical care to beneficiaries – the NHIRD, which has been described in detail elsewhere [16–18], stores their medical records and demographic data. This research used a specific NHIRD subject dataset that contains all patients in Taiwan diagnosed with vitiligo from 2000 to 2010. This study was exempt from full review by an Institutional Review Board because the NHIRD consists of de-identified secondary data made available for research.

Fig. 1. Flowchart of study cohort selection.

Please cite this article in press as: P.-C. Lo, et al., Risk of fractures in vitiligo patients treated with phototherapy—A retrospective populationbased cohort study, J Dermatol Sci (2016), http://dx.doi.org/10.1016/j.jdermsci.2016.02.010

G Model DESC 2965 No. of Pages 7

P.-C. Lo et al. / Journal of Dermatological Science xxx (2015) xxx–xxx

3

Fig. 2. Study design.

2.3. Study subjects Fig. 1 shows the study cohort selection process. Since bone loss begins in middle life in women and mainly after age 70 in men [19,20], we selected patients newly diagnosed with vitiligo (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] code 709.01) between 2003 and 2009 at an age of 40 years or older. Using years 2000–2002 as the run-in, we identified patients with new-onset vitiligo from 1 January 2003–31 December 2009 at age 40 years or older. Vitiligo cases were defined as patients having made at least two visits to a dermatologist for vitiligo during the study period. 2.4. Inclusion and exclusion criteria Patients with mild vitiligo are less likely to receive phototherapy, which causes potential selection bias. However, vitiligo severity cannot be directly ascertained from NHIRD data. In attempt to balance vitiligo severity between the comparison cohorts, we used frequency of visits for vitiligo within 1 year of diagnosis as a surrogate for severity, and further selected those who visited for vitiligo more than three times within 1 year from diagnosis. Patients who visited fewer than three times were considered having mild vitiligo and excluded from the study. We also excluded patients with medical history of osteoporosis (ICD-9-CM code 733.0), hip fractures, (ICD-9-CM code 820), or vertebral fractures (ICD-9-CM codes 805, 806) before being diagnosed with vitiligo, since these diseases are strongly associated with the principal study outcome, namely fractures. To ensure a consistent entry date for comparison cohorts, we also excluded those patients who began phototherapy beyond 180 days from the date of diagnosis. 2.5. Outcome and covariates The study outcome was the first outpatient visit or hospitalisation for a hip or vertebral fracture. Fractures caused by traumatic injuries (e.g., from traffic accidents) were excluded.

The statistical model was controlled for variables considered as potential confounding factors; these included: patients’ demographic data (i.e., age and gender), geographical region (northern, central, southern, eastern Taiwan) and occupation (four types according to attribute of insurant’s institution: public, labor, farmers and fishermen, and others); history of rheumatoid arthritis, overweight and obesity, alcohol dependence syndrome, tobacco use disorder, secondary osteoporosis, type 2 diabetes mellitus, chronic kidney disease, stroke, psoriasis, hypertension, hyperlipidemia, coronary heart disease, Parkinson’s diseases, glucocorticoids exposure, and use of any potent and superpotent topical corticosteroids, anti-osteoporosis medications (e.g., calcitonin, alendronate, etc.), calcitriol therapy, calcium supplement; and number of visits for vitiligo after study entry. Secondary osteoporosis indicates certain diseases that potentially cause osteoporosis, which include type 1 diabetes mellitus, hyperthyroidism, hypogonadism, chronic liver disease, and viral hepatitis [5]. Glucocorticoids exposure was defined as exposure to oral or parenteral glucocorticoids for more than 90 defined daily doses during the study period. 2.6. Statistical analysis Patient characteristics and baseline variables were presented as numbers and percentages or means with standard deviation. Data were analysed using ANOVA, chi-squared, or Fisher exact test, as appropriate. A two-tailed P-value <0.05 was considered statistically significant. Crude incidence rates for the fractures were presented as events per 100,000 person-years. Adjusted risks of fractures for the phototherapy cohorts compared to no phototherapy were analysed using Cox proportional hazards regression models. Age-stratified and gender-stratified analyses were conducted to investigate the pre-specified association. The age-stratified analyses were carried out for subjects aged 40–64 years and those aged 65. The gender-stratified analyses were carried out for male and female patients.

Please cite this article in press as: P.-C. Lo, et al., Risk of fractures in vitiligo patients treated with phototherapy—A retrospective populationbased cohort study, J Dermatol Sci (2016), http://dx.doi.org/10.1016/j.jdermsci.2016.02.010

G Model DESC 2965 No. of Pages 7

4

P.-C. Lo et al. / Journal of Dermatological Science xxx (2015) xxx–xxx

An analysis of dose-response relationship between the frequency of phototherapy and the risks of fractures were further conducted to examine the pre-specified association. Based on the frequency of phototherapy received, the patients were classified into four groups: (1) less than 24 times; (2) 24–47 times; (3) 48–71 times; and (4) equal or greater than 72 times phototherapy per year on average. 3. Results

kidney disease, psoriasis, or Parkinson’s diseases. However, the no phototherapy cohort had proportionally higher likelihood of a history of hypertension, type 2 diabetes mellitus, and coronary heart disease. There were no between cohort differences in the proportions of patients taking anti-osteoporosis medications or calcitriol therapy after study entry. However, a significantly higher proportion of patients in the no phototherapy cohort than in either phototherapy cohort took potent and superpotent topical corticosteroids and calcium supplements.

3.1. Baseline characteristics 3.2. Fracture risks We enrolled 3863 patients with new-onset vitiligo at age 40 years. Table 1 summarises their demographic characteristics, medical history, and use of treatments after the diagnosis of vitiligo. The mean age of all study subjects was 57 years, and 56% were female. Following the diagnosis of vitiligo, 1261 (32.6%) received phototherapy: 10.5% frequently and 22.1%, infrequently. The other 2602 patients (67.4%) received no phototherapy. There were no between cohort differences in history of hyperlipidemia, stroke, rheumatoid arthritis, secondary osteoporosis, glucocorticoids exposure, malignant neoplasm, overweight and obesity, alcohol dependence syndrome, tobacco use disorder, chronic

Mean durations of follow-up were 3.77 years for the no phototherapy cohort, 3.57 years for frequent phototherapy, and 4.38 years for infrequent phototherapy. Table 2 presents the crude incidence rates (105 person-years) and hazard ratios for fractures. 3.3. Frequent phototherapy cohort Multivariate analysis of the whole study population showed that the frequent phototherapy cohort had significant lower risk of fractures than the no phototherapy cohort (Table 2). Stratified

Table 1 Summary of the characteristics of three study cohorts from 2003 to 2009 (n = 3863). Variables

No phototherapy

Frequent phototherapy

Infrequent phototherapy

Number of subjects Demographics Age, years (mean, SD) Female (n, %) Geographical region Northern (n, %) Central (n, %) Southern (n, %) Eastern (n, %) Occupation Public (n, %) Labor (n, %) Farmers and fishermen (n, %) Others (n, %) Comorbidities Hypertension (n, %) Hyperlipidemia (n, %) Stroke (n, %) Rheumatoid arthritis (n, %) Secondary osteoporosis (n, %) Type 2 diabetes mellitus (n, %) Glucocorticoids exposurea (n, %) Malignant neoplasm (n, %) Overweight and obesity (n, %) Alcohol dependence syndrome (n, %) Tobacco use disorder (n, %) Chronic kidney disease (n, %) Psoriasis (n, %) Coronary heart disease (n, %) Parkinson’s diseases (n, %) Therapy Vitiligo visits within 1 year (mean, SD) Phototherapy (n, %) Total cumulative counts (mean, SD) Frequency (counts per year on average; mean, SD) Potent and superpotent topical corticosteroids (n, %) mg per day (mean, SD) Anti-osteoporosis medications (n, %) DDD (mean, SD) Calcium supplement (n, %) DDD (mean, SD) Calcitriol therapy (n, %) DDD (mean, SD)

2602

405

856

p-value

<0.001 0.056 <0.001

58.71 1481

12.19 56.92%

54.20 205

10.06 50.62%

55.23 474

11.16 55.37%

1387 495 671 35

53.31% 19.02% 25.79% 1.35%

257 47 93 5

63.46% 11.60% 22.96% 1.23%

412 220 211 11

48.13% 25.70% 24.65% 1.29%

96 909 361 1226

3.69% 34.93% 13.87% 47.12%

41 175 37 149

10.12% 43.21% 9.14% 36.79%

41 375 108 330

4.79% 43.81% 12.62% 38.55%

772 495 18 32 482 310 245 209 13 4 14 40 33 135 20

29.67% 19.02% 0.69% 1.23% 18.52% 11.91% 9.42% 8.03% 0.50% 0.15% 0.54% 1.54% 1.27% 5.19% 0.77%

82 73 0 3 69 35 36 33 3 0 2 3 11 8 0

20.25% 18.02% 0.00% 0.74% 17.04% 8.64% 8.89% 8.15% 0.74% 0.00% 0.49% 0.74% 2.72% 1.98% 0.00%

183 132 5 17 129 74 82 69 6 0 3 10 12 31 7

21.38% 15.42% 0.58% 1.99% 15.07% 8.64% 9.58% 8.06% 0.70% 0.00% 0.35% 1.17% 1.40% 3.62% 0.82%

<0.001 0.060 0.275 0.155 0.067 0.009 0.924 0.997 0.559 0.729 0.878 0.415 0.076 0.006 0.186

5.68 – – – 2479 0.11 37 2.06 258 14.09 11 0.00

3.86 – – – 95.27% 0.25 1.42% 35.70 9.92% 132.51 0.42% 0.06

5.94 100.00% 58.44 6.61 91.82% 0.14 0.93% 61.69 7.48% 172.20 0.47% 0.07

<0.001 – – – <0.001 <0.001 0.546 0.646 <0.001 0.254 0.927 0.814

<0.001

17.27 405 202.29 57.99 373 0.15 5 0.21 14 2.57 1 0.00

7.17 100.00% 138.45 31.24 92.10% 0.24 1.23% 2.41 3.46% 37.91 0.25% 0.05

8.84 856 46.74 8.19 786 0.08 8 2.47 64 15.08 4 0.00

SD—standard deviation; DDD—defined daily dose. a Exposed to oral or parenteral glucocorticoids for more than 90 DDDs.

Please cite this article in press as: P.-C. Lo, et al., Risk of fractures in vitiligo patients treated with phototherapy—A retrospective populationbased cohort study, J Dermatol Sci (2016), http://dx.doi.org/10.1016/j.jdermsci.2016.02.010

G Model DESC 2965 No. of Pages 7

P.-C. Lo et al. / Journal of Dermatological Science xxx (2015) xxx–xxx

5

Table 2 Cox proportional hazards models for fractures in no phototherapy, infrequent and frequent phototherapy cohort. Variable Base case analysis No phototherapy Infrequent phototherapy Frequent phototherapy Stratified analysis By age Age 40  64 No phototherapy Infrequent phototherapy Frequent phototherapy Age  65 No phototherapy Infrequent phototherapy Frequent phototherapy By gender Female No phototherapy Infrequent phototherapy Frequent phototherapy Male No phototherapy Infrequent phototherapy Frequent phototherapy

No. of subjects

No. of fractures

PYs

2602 856 405

72 18 8

9803.69 3748.85 1443.85

1787 680 339

24 10 2

815 176 66

Rate/ 105 PYs

Unadjusted HR (95%CI)

P-value

Adjusted HR (95%CI)a

P-value

734.42 480.15 554.08

1 0.64 0.76

reference (0.38, 1.07) (0.36, 1.57)

– 0.086 0.457

1 0.87 0.32

reference (0.51, 1.50) (0.13, 0.75)

– 0.628 0.009

6770.96 3039.38 1229.65

354.46 329.01 162.65

1 0.86 0.45

reference (0.41, 1.81) (0.11, 1.92)

– 0.694 0.283

1 1.09 0.14

reference (0.50, 2.35) (0.03, 0.69)

– 0.836 0.016

48 8 6

3032.73 709.46 214.20

1582.73 1127.61 2801.12

1 0.71 1.78

reference (0.34, 1.50) (0.76, 4.15)

– 0.369 0.185

1 0.80 0.59

reference (0.36, 1.75) (0.20, 1.74)

– 0.572 0.337

1481 474 205

49 12 6

5557.38 2080.97 738.80

881.71 576.65 812.13

1 0.63 0.92

reference (0.34, 1.19) (0.39, 2.14)

– 0.157 0.841

1 1.00 0.31

reference (0.52, 1.95) (0.11, 0.86)

– 0.991 0.024

1121 382 200

23 6 2

4246.31 1667.88 705.05

541.65 359.74 283.67

1 0.64 0.53

reference (0.26, 1.58) (0.13, 2.26)

– 0.336 0.392

1 0.75 0.29

reference (0.28, 2.00) (0.05, 1.61)

– 0.564 0.156

HR, hazard ratio; CI, confidence interval; PYs, Person-years. a Adjusted for age, gender (gender was not adjusted in each gender-specific analysis), geographical region, occupation, comorbidities of rheumatoid arthritis, overweight and obesity, alcohol dependence syndrome, tobacco use disorder, secondary osteoporosis, type 2 diabetes mellitus, chronic kidney disease, stroke, psoriasis, hypertension, hyperlipidemia, coronary heart disease, Parkinson’s diseases, systemic glucocorticoids exposure, potent and superpotent topical corticosteroids, calcium supplement, and number of visits for vitiligo after the entry date.

analyses confirmed the fracture prevention effect of frequent phototherapy in the 40–64 years stratum and in female patients. Among patients aged 65 and among male patients, the frequent phototherapy cohort did not have significantly lower risk of fractures as compared with the no phototherapy cohort.

significant decrease of fracture risks was observed in the most frequent phototherapy group (72 times phototherapy per year on average; adjusted hazard ratio = 0.21, p = 0.039).

3.4. Infrequent phototherapy cohort

Although previous studies have shown increased serum vitamin D level and bone mineral density after phototherapy in psoriasis [13,14], there are no prior studies to show the effect of phototherapy on fracture risks. Serum vitamin D level and bone mineral density are only surrogate markers of fracture risks. To our knowledge, this is the first report on fracture risks in vitiligo patients treated with phototherapy. Our findings provide early evidence that frequent phototherapy might decrease the risk of fractures in vitiligo patients, especially among females and those aged 40–64 years. Phototherapy might have a promising beneficial effect on fracture reduction; however, the risk of photoaging and photocarcinogenicity should also be considered. Nonetheless, nonCaucasian skin is less prone to photoaging, and the risks have been reported to be further reduced in vitiliginous skin [21]. In addition,

Multivariate analysis showed that the infrequent phototherapy cohort had a risk of fractures comparable to the no phototherapy cohort (Table 2). Likewise, all stratified analyses, either based on age or gender, showed that the infrequent phototherapy cohort had a similar risk of fractures as compared with the no phototherapy cohort. 3.5. Dose response relationship The analysis of dose-response relationship showed a possible relationship between the frequencies of phototherapy and the risks of fractures (Table 3). As the frequency of the phototherapy increased, a decreasing trend in the fracture risks was observed. A

4. Discussion

Table 3 The Cox proportional hazards model for investigating the dose-response relationship between phototherapy frequency and the risk of fractures. Frequency of phototherapy (times per year on average)

Unadjusted HR (95% CI)

P-value

Adjusted HR (95%CI)a

P-value

<24 24–47

1 1.01

– 0.987

1 0.46

– 0.126

48–71

0.42

0.388

0.26

reference (0.41, 2.48)

(0.17, 1.24)

(0.06, 3.01) 72

0.94

0.181 (0.03, 1.88)

0.927 (0.23, 3.80)

reference

0.21

0.039 (0.05, 0.92)

HR, hazard ratio. a Adjusted for age, gender, geographical region, occupation, comorbidities of rheumatoid arthritis, overweight and obesity, alcohol dependence syndrome, tobacco use disorder, secondary osteoporosis, type 2 diabetes mellitus, chronic kidney disease, stroke, psoriasis, hypertension, hyperlipidemia, coronary heart disease, Parkinson’s diseases, systemic glucocorticoids exposure, potent and superpotent topical corticosteroids, calcium supplement, and number of visits for vitiligo after the entry date.

Please cite this article in press as: P.-C. Lo, et al., Risk of fractures in vitiligo patients treated with phototherapy—A retrospective populationbased cohort study, J Dermatol Sci (2016), http://dx.doi.org/10.1016/j.jdermsci.2016.02.010

G Model DESC 2965 No. of Pages 7

6

P.-C. Lo et al. / Journal of Dermatological Science xxx (2015) xxx–xxx

the therapeutic dose of UVB is usually lower in vitiligo compared with other dermatoses [22,23]. Despite the lower UVB dose used in vitiligo, our study still observed a lower fracture risk in patients treated with frequent phototherapy, which suggests a strong link toward phototherapy and reduced fracture risks. The prevalence of osteoporosis rises steadily with advancing age [24]; after 35 years of age bone loss increases gradually as part of the natural ageing process [25]. By age of 75 years, approximately half of the population will have osteoporosis. Osteoporosis is asymptomatic until the first fracture occurs, whereafter significant mortality and morbidity incur considerable social costs that include direct medical costs and indirect costs resulting from reduced quality of life, disability, and even death [26]. We defined fractures as our study outcome because this is a more important clinical outcome than osteoporosis per se. Furthermore, osteoporosis in Taiwan is under-diagnosed [27,28], and is at best only a surrogate marker of fracture risks. We think that the statistically significant decrease in fracture was observed in the patients aged 40–64 years for two reasons. First, since bone loss in women starts to increase after 35 years of age, preventive strategies such as phototherapy, might be more beneficial when we target patients with early bone loss who still have sufficient remaining bone density to prevent further fractures. In other words, despite receiving phototherapy, patients older than 65 may already have bone loss that is too severe and bone structure that is too fragile to prevent further fractures. Second, vitiligo is an early-onset disease, with more than half of patients developing vitiligo before their twenties [29]. Consequently, the number of patients aged over 65 years in our study might be too small to detect any statistical significance. The mechanism by which frequent phototherapy decreases the fracture risks in vitiligo patients may relate to increased levels of serum vitamin D. Vitiligo patients treated with phototherapy are known to have increased serum vitamin D levels, with moderate correlation between vitamin D levels and the degree of pigmentation [11]. Since vitamin D plays a significant role in calcium and bone mineral metabolism [12], the elevated vitamin D level in vitiligo patients treated with phototherapy might further influence osteo-homeostasis and enhance osteogenesis, with consequently decreased risk of fractures. Our observation that phototherapy had no significant effect on fracture prevention in the infrequent phototherapy cohort suggests that there may be a dosage threshold or response relationship between phototherapy and fracture prevention. Our study showed that the patients might need frequent phototherapy (more than 72 times annually) to accumulate or maintain sufficient serum vitamin D to prevent further fractures. Due to the lack of melanin and melanocytes in the vitiliginous skin, a lower starting as well as a maintenance UVB dosage was often used for the treatment of vitiligo to prevent sunburn. Thus, the requirement of frequency of phototherapy (72 instead of 24 times per year [15]) to achieve a significant reduction of fracture incidence might be due to either the small number of fractures sufferers in the study or the use of UVB dose which is lower than minimal erythema dose in vitiligo treatment, or both. A previous study found that serum vitamin D plateaus after 3 weeks of narrowband ultraviolet B phototherapy [30], which suggests that there might also be a plateau in the fracture prevention effect of phototherapy. Further studies will be needed to test this hypothesis. Both men and women lose bone mass during ageing, and sex hormones also have important roles in bone loss. In women, decreased estrogen due to the menopause causes bone resorption to increase by 90% while bone formation increases by only 45%, resulting in overall bone loss. However, men have a more gradual decrease in sex hormone levels with ageing, which may account for the less severe decrease in bone strength [31]. Given this gender

difference in rates of bone loss, it is possible that gender factor may also influence fracture risk prevention in phototherapy. The frequent phototherapy had a fracture prevention effect among our female patients. As bone loss in women begins at an earlier age and more abruptly compared with men [19,20], the fracture risks in middle-aged vitiligo patients may be much higher in females than in males, thus phototherapy may be more beneficial to female than male vitiligo patients. We concerned the influences of baseline comorbidities and therapy between cohorts. In terms of comorbidities, rheumatoid arthritis and secondary osteoporosis are well-confirmed risk factors for fractures [5], and their proportions between three cohorts were not significantly different. On the other hand, although the proportions of hypertension, type 2 diabetes mellitus and coronary heart disease were significantly different between three cohorts, their influences are limited because they are not strong predictors for fractures. Furthermore, the bias caused by coronary heart disease might be limited due to its low prevalence. To minimize possible bias, we have adjusted all obtainable confounders in our statistical models. With regard to the differences in therapy between cohorts, the proportions of patients using calcium supplement (a protector) was the lowest among the frequent phototherapy cohorts. Meanwhile, the patients in the frequent phototherapy cohort consumed the highest dose of potent and superpotent topical corticosteroids (a risk factor for fractures). These imply that the protective effect of frequent phototherapy could be underestimated if neglecting these variables in the model. Thus, we also adjusted these variables in our models. Our study had some limitations. Firstly, the phototherapy cohorts had proportionately lower rates of several comorbidities, and perhaps were more health conscious than the no phototherapy cohort. Our statistical models had adjusted the influence of comorbidities on fractures but not the health conscious which was not available in the NHIRD dataset. It might have led us to overestimate the fracture protection effect of frequent phototherapy. Secondly, some other confounding factors for fractures, including parental history of fractures, tobacco use, alcohol intake, physical activity, body weight or body-mass index, and sun exposure history again were not available in the NHIRD dataset. Some related indicators, for example, geographic region and occupation, were selected as proxy to reflect the sun exposure history. Certainly, the degree to which sun exposure may be represented by these proxy variables are unclear. Finally, although we used a national cohort, limited numbers of patients received phototherapy, especially above age 65 years, which may have underpowered the analysis to demonstrate statistical significance. This may explain why we found no statistically significant differences between either phototherapy cohort and the no phototherapy cohort. 5. Conclusions This study provides the first evidence that frequent phototherapy can decrease fracture risks in middle-aged and female vitiligo patients. Further studies are warranted to elucidate the possible mechanisms involved and to find out whether the benefit of frequent phototherapy may extrapolate to other patient populations at increased risk of osteoporosis. IRB status This study was exempt from full review by an Institutional Review Board because the Taiwan National Health Insurance Research Database consists of de-identified secondary data made available for research.

Please cite this article in press as: P.-C. Lo, et al., Risk of fractures in vitiligo patients treated with phototherapy—A retrospective populationbased cohort study, J Dermatol Sci (2016), http://dx.doi.org/10.1016/j.jdermsci.2016.02.010

G Model DESC 2965 No. of Pages 7

P.-C. Lo et al. / Journal of Dermatological Science xxx (2015) xxx–xxx

Funding sources This article has no funding source. Conflicts of interest All authors have no conflicts of interest to declare. Statement on any prior presentation This study has not been published previously in whole or part, except in abstract form. Acknowledgements This study is based on data from the Taiwan National Health Insurance Research Database, which is provided by the National Health Insurance Administration (NHIA), Ministry of Health and Welfare and managed by National Health Research Institutes (NHRI). The NHIA and NHRI, however, played no part in designing the study or analyzing and interpreting the data. The interpretation and conclusions contained herein do not represent any official position of the NHIA, or the NHRI. References [1] A. Alikhan, L.M. Felsten, M. Daly, V. Petronic-Rosic, Vitiligo: a comprehensive overview Part I. Introduction, epidemiology, quality of life, diagnosis, differential diagnosis, associations, histopathology, etiology, and work-up, J. Am. Acad. Dermatol. 65 (2011) 473–491. [2] P.E. Grimes, I. Hamzavi, M. Lebwohl, J.P. Ortonne, H.W. Lim, The efficacy of afamelanotide and narrowband UV-B phototherapy for repigmentation of vitiligo, JAMA Dermatol. 149 (2013) 68–73. [3] G. Laberge, C.M. Mailloux, K. Gowan, P. Holland, D.C. Bennett, P.R. Fain, et al., Early disease onset and increased risk of other autoimmune diseases in familial generalized vitiligo, Pigment Cell Res. 18 (2005) 300–305. [4] J.I. Silverberg, N.B. Silverberg, Clinical features of vitiligo associated with comorbid autoimmune disease: a prospective survey, J. Am. Acad. Dermatol. 69 (2013) 824–826. [5] J.A. Kanis on behalf of the World Health Organization Scientific Group, Assessment of osteoporosis at the primary health-care level, Technical Report, World Health Organization Collaborating Centre for Metabolic Bone Diseases, University of Sheffield, UK, 2007. [6] B. Abrahamsen, D. Brask-Lindemann, K.H. Rubin, P. Schwarz, A review of lifestyle, smoking and other modifiable risk factors for osteoporotic fractures, Bonekey Rep. 3 (2014) 574. [7] Z.Z. Lin, J.J. Wang, C.R. Chung, P.C. Huang, B.A. Su, K.C. Cheng, et al., Epidemiology and mortality of hip fracture among patients on dialysis: Taiwan National Cohort Study, Bone 64 (2014) 235–239. [8] L.M. Felsten, A. Alikhan, V. Petronic-Rosic, Vitiligo: a comprehensive overview Part II: treatment options and approach to treatment, J. Am. Acad. Dermatol. 65 (2011) 493–514. [9] A. Beheshti, H. Ghadami, A. Barikani, F. Haj Manouchehri, Assessment of vitamin D plasma levels in patients with vitiligo vulgaris, Acta Med. Iran. 52 (2014) 601–606. [10] I. Ustun, G. Seraslan, C. Gokce, S. Motor, Y. Can, M. Ugur Inan, et al., Investigation of vitamin d levels in patients with vitiligo vulgaris, Acta Dermatovenerol. Croat. 22 (2014) 110–113. [11] M. Sehrawat, T.C. Arora, A. Chauhan, H.K. Kar, A. Poonia, V. Jairath, Correlation of vitamin d levels with pigmentation in vitiligo patients treated with NBUVB therapy, ISRN Dermatol. (2014) 493213.

7

[12] M. Premkumar, T. Sable, D. Dhanwal, R. Dewan, Vitamin D homeostasis, bone mineral metabolism, and seasonal affective disorder during 1 year of Antarctic residence, Arch. Osteoporos. 8 (2013) 129. [13] A. Osmancevic, K. Landin-Wilhelmsen, O. Larko, D. Mellstrom, A.M. Wennberg, L. Hulthen, et al., UVB therapy increases 25(OH) vitamin D syntheses in postmenopausal women with psoriasis, Photodermatol. Photoimmunol. Photomed. 23 (2007) 172–178. [14] A. Osmancevic, K. Landin-Wilhelmsen, O. Larko, D. Mellstrom, A.M. Wennberg, L. Hulthen, et al., Risk factors for osteoporosis and bone status in postmenopausal women with psoriasis treated with UVB therapy, Acta Dermatol. Venereol. 88 (2008) 240–246. [15] M.K. Bogh, A.V. Schmedes, P.A. Philipsen, E. Thieden, H.C. Wulf, A small suberythemal ultraviolet B dose every second week is sufficient to maintain summer vitamin D levels: a randomized controlled trial, Br. J. Dermatol. 166 (2012) 430–433. [16] F.Y. Hsiao, C.L. Yang, Y.T. Huang, W.F. Huang, Using Taiwan national health insurance research databases for pharmacoepidemiology research, J. Food Drug Anal. 15 (2007) 99–108. [17] C.L. Cheng, Y.H. Kao, S.J. Lin, C.H. Lee, M.L. Lai, Validation of the National Health Insurance Research Database with ischemic stroke cases in Taiwan, Pharmacoepidemiol. Drug Saf. 20 (2011) 236–242. [18] The National Health Research Institutes, National Health Insurance Research Database, The National Health Research Institutes, Taiwan, 2015. [19] S. Khosla, Pathogenesis of age-related bone loss in humans, J. Gerontol. A. Biol. Sci. Med. Sci. 68 (2013) 1226–1235. [20] B.L. Riggs, S. Khosla, L.J. Melton 3rd, Sex steroids and the construction and conservation of the adult skeleton, Endocrinol. Rev. 23 (2002) 279–302. [21] A. Paradisi, S. Tabolli, B. Didona, L. Sobrino, N. Russo, D. Abeni, Markedly reduced incidence of melanoma and nonmelanoma skin cancer in a nonconcurrent cohort of 10,040 patients with vitiligo, J. Am. Acad. Dermatol. 71 (2014) 1110–1116. [22] M.M. Kleinpenning, T. Smits, J. Boezeman, P.C. van de Kerkhof, A.W. Evers, M.J. Gerritsen, Narrowband ultraviolet B therapy in psoriasis: randomized doubleblind comparison of high-dose and low-dose irradiation regimens, Br. J. Dermatol. 161 (2009) 1351–1356. [23] Z. Hallaji, M. Ghiasi, A. Eisazadeh, M.R. Damavandi, Evaluation of the effect of disease duration in generalized vitiligo on its clinical response to narrowband ultraviolet B phototherapy, Photodermatol. Photoimmunol. Photomed. 28 (2012) 115–119. [24] J.A. Kanis, Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report, WHO Study Group, Osteoporos. Int. 4 (1994) 368–381. [25] J. Burch, S. Rice, H. Yang, A. Neilson, L. Stirk, R. Francis, et al., Systematic review of the use of bone turnover markers for monitoring the response to osteoporosis treatment: the secondary prevention of fractures, and primary prevention of fractures in high-risk groups, Health Technol. Assess. 18 (2014) 1–180. [26] F. Bleibler, A. Konnopka, P. Benzinger, K. Rapp, H.H. Konig, The health burden and costs of incident fractures attributable to osteoporosis from 2010 to 2050 in Germany–a demographic simulation model, Osteoporos. Int. 24 (2013) 835–847. [27] N.P. Yang, C.Y. Deng, Y.J. Chou, P.Q. Chen, C.H. Lin, P. Chou, et al., Estimated prevalence of osteoporosis from a Nationwide Health Insurance database in Taiwan, Health Policy 75 (2006) 329–337. [28] J.S. Hwang, D.C. Chan, J.F. Chen, T.T. Cheng, C.H. Wu, Y.K. Soong, et al., Clinical practice guidelines for the prevention and treatment of osteoporosis in Taiwan: summary, J. Bone Miner. Metab. 32 (2014) 10–16. [29] R. Yaghoobi, M. Omidian, N. Bagherani, Vitiligo: a review of the published work, J. Dermatol. 38 (2011) 419–431. [30] L. Feldmeyer, G. Shojaati, K.S. Spanaus, A. Navarini, B. Theler, D. Donghi, et al., Phototherapy with UVB narrowband, UVA/UVBnb, and UVA1 differentially impacts serum 25-hydroxyvitamin-D3, J. Am. Acad. Dermatol. 69 (2013) 530–536. [31] T. Willson, S.D. Nelson, J. Newbold, R.E. Nelson, J. LaFleur, The clinical epidemiology of male osteoporosis: a review of the recent literature, Clin. Epidemiol. 7 (2015) 65–76.

Please cite this article in press as: P.-C. Lo, et al., Risk of fractures in vitiligo patients treated with phototherapy—A retrospective populationbased cohort study, J Dermatol Sci (2016), http://dx.doi.org/10.1016/j.jdermsci.2016.02.010