ORIGINAL

ARTICLES

Family history, body mass index, selected dietary factors, menstrual history, and risk of moderate to severe acne in adolescents and young adults
Anna Di Landro, MD,a Simone Cazzaniga, PhD Math,a Fabio Parazzini, MD,b Vito Ingordo, MD,c Francesco Cusano, MD,d Laura Atzori, MD,e Francesco Tripodi Cutr , MD,f Maria Letizia Musumeci, MD,g a h Cornelia Zinetti, BS, Enrico Pezzarossa, MD, Vincenzo Bettoli, MD,i Marzia Caproni, MD,j Giovanni Lo Scocco, MD,k Angela Bonci, MD,l Pierluca Bencini, MD,m and Luigi Naldi, MD,a and the GISED Acne Study Group Bergamo, Milano, Taranto, Benevento, Cagliari, Napoli, Catania, Cremona, Ferrara, Firenze, Prato, and Reggio Emilia, Italy
Background: Genetic and environmental components may contribute to acne causation. Objective: We sought to assess the impact of family history, personal habits, dietary factors, and menstrual history on a new diagnosis of moderate to severe acne. Methods: We conducted a case-control study in dermatologic outpatient clinics in Italy. Cases (205) were consecutive those receiving a new diagnosis of moderate to severe acne. Control subjects (358) were people with no or mild acne, coming for a dermatologic consultation other than for acne. Results: Moderate to severe acne was strongly associated with a family history of acne in rst-degree relatives (odds ratio 3.41, 95% condence interval 2.31-5.05). The risk was reduced in people with lower body mass index with a more pronounced effect in male compared with female individuals. No association with smoking emerged. The risk increased with increased milk consumption (odds ratio 1.78, 95% condence interval 1.22-2.59) in those consuming more than 3 portions per week. The association was more marked for skim than for whole milk. Consumption of sh was associated with a protective effect (odds ratio 0.68, 95% condence interval 0.47-0.99). No association emerged between menstrual variables and acne risk. Limitations: Some degree of overmatching may arise from choosing dermatologic control subjects and from inclusion of mild acne in the control group. Conclusions: Family history, body mass index, and diet may inuence the risk of moderate to severe acne. The inuence of environmental and dietetic factors in acne should be further explored. ( J Am Acad Dermatol 2012;67:1129-35.) Key words: body mass index; contraceptive pill; diet; sh; menstrual history; milk; moderate to severe acne; risk factors; skim milk; smoking.

From the Centro Studi Gruppo Italiano Studi Epidemiologici in Dermatologia (GISED), Fondazione per la Ricerca Ospedale a Maggiore, Bergamoa; Clinica Ostetrico Ginecologica, Universit degli studi di Milano, Fondazione Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Ca Granda, Ospedale Maggiore, a Policlinico, Milanob; Centro Ospedaliero Militare, Tarantoc; Unit di Dermatologia, Presidio Ospedaliero Multizonale G. Rummo, a degli Beneventod; Clinica Dermatologica, Ospedale Universit Studi, Cagliarie; Clinica Dermatologica, Policlinico Universitario, a di Cataniag; Unit a di Napolif; Clinica Dermatologica, Universit Dermatologia, Azienda Ospedaliera, Cremonah; Clinica Dermatologica, Azienda Ospedaliera Universitaria, Ferrarai; Clinica Dermatologica, Ospedale S. Maria Nuova-Orbatello, Firenzej;

Unit a di Dermatologia, Ospedale Misericordia e Dolce, Pratok; Unit a di Dermatologia, Arcispedale S. Maria Nuova, Reggio Emilial; and Istituto di Chirurgia e Laserchirurgia in Dermatologia, Milano.m Funding sources: None. Conflicts of interest: None declared. Accepted for publication February 4, 2012. Reprint requests: Luigi Naldi, MD, Centro Studi GISED, Via Garibaldi 13, 24100 Bergamo, Italy. E-mail: luigi.naldi@gised.it. Published online March 2, 2012. 0190-9622/$36.00 2012 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2012.02.018

1129

1130 Di Landro et al

J AM ACAD DERMATOL

DECEMBER 2012

Genetic and environmental components contribfew comedones are present; (2) mild acne, where ute to the pathogenesis of acne.1 The prevalence of lesions include several noninflammatory comedones acne varies in different countries and over time, and with less than 10 inflammatory lesions; (3) moderate it has been postulated that different lifestyles may acne with many comedones, papules, pustules, but influence acne prevalence.2 Along this line, a few no nodules; and (4) severe acne with inflammatory studies reported that acne was virtually absent in nodules in addition to papules and pustules. non-Westernized countries in those living and eating A total of 205 patients aged 10 to 24 years (mean in their traditional manner, age, 17.2 6 3.1 years, 50.2% whereas acne began to apmale) and 358 control subCAPSULE SUMMARY pear when these populations jects (mean age 17.6 6 3.8 changed their eating habits years, 43.0% male) were inWesternized diet and high consumption to those more similar to terviewed. Lesions in paof milk and skim milk have been 3 Western populations. tients were present for at associated with acne. To analyze the role played least 6 months in 176 cases We confirmed that milk and skim milk by different factors, includ(85.8%). Areas involved were are associated with the risk of moderate ing family history, dietary the face in 201 cases (98.0%), to severe acne. We also documented an factors, smoking, and mentrunk in 104 cases (50.7%), increased risk of acne with increased strual history in acne develand other locations in 18 body mass index, and a protective effect opment, we conducted a cases (8.8%). Most patients of fish consumption. We ruled out an case-control study in young had already received topical increased risk for many dietetic factors, patients with a diagnosis of treatment for the disease at smoking, and menstrual history. moderate to severe acne verthe time of the examination. sus control subjects with no Reasons for dermatologic Diet should be an issue when discussing or mild acne at several derconsultation in control subacne management with patients. matologic outpatient clinics jects were assessment of nevi in Italy. in 106 cases (29.6%), wart treatment in 27 cases (7.5%), eczematous dermatitis METHODS in 21 cases (5.8%), and several other less prevalent This was a multicenter case-control study of risk conditions in the remaining control subjects. factors for acne in people aged 10 to 24 years Trained interviewers performed the interviews attending a dermatologic outpatient department in with both patients and control subjects using a 15 Italian cities (6 in Northern and 9 in Central and standardized questionnaire to collect information Southern Italy). The protocol was reviewed and on general sociodemographic characteristics (ie, approved by the Ethics Board of the Coordinating age, gender, marital status, education, and occupaCenter at Ospedali Riuniti di Bergamo. tion), personal habits (eg, smoking and alcohol Patients were those receiving a new diagnosis of consumption), menstrual history (limited to femoderate to severe acne by a dermatologist who males), and relevant medical history. A food frewere consecutively observed at the participating quency questionnaire was used to record centers during the study period. Those who could information on the intake of selected food items, not speak Italian were excluded. including the usual number of portions per week The control subjects were people with no or mild and the portions during the week before the interacne lesions without treatment for acne, coming for a view. Intakes of foods less than once a week but that dermatologic consultation other than for acne on were reported at least once a month were coded as randomly selected days at the same centers where 0.5 portions per week. cases were identied. They were matched with cases The food groups investigated included pasta and for age in quinquennia. Participants provided inbread consumption; beef, pork, chicken, ham/saformed written consent to enter the study. The study lami, sh, and eggs; whole, skim, or partially skim was approved by the ethical committees of the milk and other dairy products (cheese, goat cheese); participating hospitals. The participation rate was cakes, sweets, chocolate, and ice cream; and vegegreater than 95% for both patients and control tables and fruit. subjects. The classication of menstrual cycles was based Acne severity was dened according to a global on questions about the age at menarche, use of oral score,4 using photographs limited to the face to help contraceptive pills, average length of menstrual with judgment. Four categories were considered as cycles, and the number of periods during the past follows: (1) no or minimal acne lesions, where only a 12 months. Denitions of menstrual cycle patterns
d d d

J AM ACAD DERMATOL
VOLUME 67, NUMBER 6

Di Landro et al 1131

Abbreviations used: BMI: CI: OR: body mass index condence interval odds ratio

stepwise method. These factors are listed in the footnotes of the tables presenting results.

RESULTS
Table I shows the distribution of patients and control subjects, and corresponding ORs, according to age, gender, family history of the condition, and BMI. A family history of acne in first-degree relatives increases the risk of having the condition (OR 3.41, 95% CI 2.31-5.05). Compared with subjects with BMI less than 18.5 kg/m2, the OR of acne was 1.90 (95% CI 1.09-3.31) in those with a BMI of 18.5 to 23 kg/m2 and 1.94 (95% CI 1.02-3.68) in those with a BMI greater than 23 kg/m2. When stratifying by gender, the ORs were 2.25 (95% CI 0.89-5.68) in male and 1.67 (95% CI 0.83-3.37) in female individuals with BMI 18.5 to 23 kg/m2, whereas they were 3.29 (95% CI 1.16-9.28) in male and 1.30 (95% CI 0.54-3.11) in female individuals with BMI greater than 23 kg/m2. No association between smoking and acne risk emerged. The relationship between dietary factors and risk of acne is shown in Table II. The analysis showed a significant difference in the weekly consumption of milk between patients and control subjects. The association was more marked for skim milk than for whole milk. Consumption of fish was associated with a protective effect, with an OR of 0.68 (95% CI 0.47-0.99). No marked difference was documented in the estimated ORs when the analysis was considered separately in male and female individuals. No association between acne and other dietary items emerged. Table III shows the distribution of female patients and control subjects according to age at menarche, menstrual pattern, and use of oral contraceptives. No association emerged for these variables with acne risk.

were as follows: (1) premenarche was no menarche before 16 years of age; (2) primary amenorrhea was menarche not started after 16 years of age; (3) secondary amenorrhea was the absence of menstruation for 180 days or more; (4) oligomenorrhea was an average menstrual cycle of 42 to 180 days; (5) polymenorrhea was an average menstrual cycle of 21 days or less; (6) regular menstrual cycles were an average menstrual cycle of 22 to 41 days; (7) irregular menstrual cycles were average menstrual cycles of 22 to 41 days and two or more menstrual cycles with a length of less than 22 or more than 41 days during the past year; and (8) oral contraceptive use was the use of oral contraceptives for contraception, irregular menstrual cycles, dysmenorrhea, or acne.5 The average daily consumption of alcohol was computed assuming the following pure alcohol contents in each type of drink: 150 mL of wine = 330 ml of beer = 30 mL of spirits = 10 to 12 g of pure alcohol. To be classied as an ex-smoker or drinker, a patient was requested to have had a period of abstinence of at least 6 months before the date of diagnosis. Anthropometric measures, including height and weight, were also obtained. Body mass index (BMI) was calculated as weight (kg)/height (m2). BMI was categorized using extreme quintiles rounded to nearest meaningful values as cut-off points. Although there is no general consensus regarding BMI thresholds in young people, our categories could be interpreted as follows: less than 18.5 kg/ m2 underweight or at risk of underweight, 18.5 to 23 kg/m2 normal weight, and greater than 23 kg/m2 overweight or at risk of overweight.6 For each variable, we computed the odds ratios (OR) of acne as estimates of relative risks and the corresponding 95% condence intervals (CI). Variable categorization was based on clinically meaningful thresholds or tertiles of data distribution. Because of data grouping in consumption categories, food variables were categorized using the rst tertile of distribution as the cut-off point. To account for the effects of several potential confounding factors, simultaneously, conditional multiple logistic regression analysis with maximum likelihood tting was used. Included in the regression equations were terms in the data set signicantly associated with the risk of acne and selected by a forward

DISCUSSION
Before discussing our results, potential limitations of the study should be considered. The choice of dermatologic control subjects may have resulted in some degree of overmatching if risk factors were shared among skin conditions. We tried to include control subjects from several different diagnostic categories, and no marked differences in OR estimates were observed when the main diagnostic categories were considered separately. In addition, inclusion of patients with mild acne in the control group may have resulted in some attenuation of risk estimates. We acknowledge that this is a rather conservative bias, and actual risks may be higher than our estimates. Other sources of bias such as selection bias or confounding factors are unlikely to

1132 Di Landro et al

J AM ACAD DERMATOL

DECEMBER 2012

Table I. Distribution of patients and control subjects and corresponding odds ratios according to demographic variables, body mass index, family history, and smoking
Case, n (%) Control, n (%) OR (95% CI)*

Table II. Distribution of patients and control subjects and corresponding odds ratios according to intake of selected food items and risk of acne
Dietary items* Case, n (%) Control, n (%) OR (95% CI)y

Gender Female Male Age group, yy 10-14 15-19 20-24 Body mass index, kg/m2 \18.5 18.5-23 [23 Family history in firstdegree relatives No Yes Smoking habits Nonsmoker Smoker

102 (49.8) 204 (57.0) 103 (50.2) 154 (43.0) 1.14 (0.79-1.66) 42 (20.5) 88 (24.6) 117 (57.1) 138 (38.5) 46 (22.4) 132 (36.9)
z

23 (11.2) 69 (19.3) 136 (66.3) 213 (59.5) 1.90 (1.09-3.31)x 46 (22.4) 76 (21.2) 1.94 (1.02-3.68)x

107 (52.2) 280 (78.2) 98 (47.8) 78 (21.8) 3.41 (2.31-5.05) 180 (87.8) 305 (85.2) 25 (12.2) 53 (14.8) 1.00 (0.56-1.77)

Milk [3 dd/wk 125 (61.3) 173 (48.6) 1.78 (1.22-2.59) Whole 42 (67.7) 62 (57.9) 1.64 (0.81-3.33) Skim 75 (75.8) 95 (60.5) 2.20 (1.18-4.10) Cheese/yogurt 76 (37.2) 137 (38.4) 0.94 (0.64-1.37) [3 dd/wk Bread/pasta 194 (94.6) 326 (91.6) 1.39 (0.65-2.98) [3 dd/wk Cakes/sweets 94 (46.1) 164 (45.9) 0.98 (0.68-1.42) [3 dd/wk Chocolate 69 (33.7) 126 (35.3) 0.93 (0.63-1.37) [3 dd/wk Fruits/vegetables 117 (57.1) 220 (61.6) 0.84 (0.58-1.22) [3 dd/wk Fish $ 1 dd/wk 94 (45.9) 190 (53.4) 0.68 (0.47-0.99) Red meat 78 (38.1) 117 (32.8) 1.22 (0.83-1.79) [3 dd/wk Ham/salami 70 (34.2) 99 (27.8) 1.27 (0.86-1.88) [3 dd/wk
CI, Confidence interval; dd, average daily portion; OR, odds ratio. *Reference category is below reported threshold. y Multivariate estimates including terms for age, body mass index, family history of acne.

CI, Confidence interval; OR, odds ratio. *Multivariate estimates including terms for age, body mass index, family history of acne. y Matching variable. z Reference category. x 2 x Test for trend P = .085.

have produced marked effects, especially considering that patients and control subjects were interviewed at the same institution and that the rate of those agreeing to participate in the study was very high. Furthermore, the ndings were largely consistent when the analyses were run separately for each participating center or when the models of analysis included terms for center (data not shown). It is possible that prejudice concerning the role that dietetic factors may play in acne severity could have inuenced the collection of dietary history. It should be noted, however, that the specic association of acne with dairy products was not known to patients and control subjects, and that only a limited part of the questionnaire was related to dietary factors, so the attention of the interviewers and patients was not exclusively driven by dietetic aspects. A major strength of this study is that the agreement to participate in the study was very high at greater than 95%, and that we could adjust the estimates for several important confounders.

The general results of this study conrm that family history increases the risk of acne and suggest that BMI and a diet rich in milk and poor in sh may inuence the risk of the disease. No association emerged between intake of other food items and risk of acne. Likewise, no relationship emerged between menstrual characteristics or the use of oral contraceptives and the risk of moderate to severe acne. The results of this study should be considered in the framework of published studies on risk factors for acne. In our study, a family history of acne was the strongest predictor of moderate to severe acne. Familial and twin studies provide supporting evidence that genetic factors play a major role in determining susceptibility to acne.1 Genetic modeling using acne scoring in a large twin study, limited to women, conducted in the United Kingdom showed that 81% of the variance of acne was attributable to genetic effects and only the remaining 19% was attributed to unshared environmental factors.7 Results from other twin studies of acne reported heritability estimates between 50% and 90%.8,9 Limited and controversial data are available concerning the relationship between BMI and acne. In a study conducted in Taiwan in 3274 schoolchildren (aged 6-11 years), the mean BMI in students with acne was signicantly higher than that of unaffected

J AM ACAD DERMATOL
VOLUME 67, NUMBER 6

Di Landro et al 1133

Table III. Distribution of female patients and control subjects and corresponding odds ratios according to menstrual characteristics and risk of acne
Case, n (%) Control, n (%) OR (95% CI)*

Age at menarche, y # 12 [12 Menstrual cycle pattern Regular Irregular Use of oral contraceptives No Yes

52 (52.0) 48 (48.0)

y 99 (51.6) 93 (48.4) 1.08 (0.64-1.83)

68 (67.3) 147 (76.2) 33 (32.7) 46 (23.8) 1.56 (0.89-2.74)

37 (78.7) 10 (21.3)

32 (84.2) 6 (15.8) 0.74 (0.20-2.74)

CI, Confidence interval; OR, odds ratio. *Multivariate estimates including terms for age, body mass index, family history of acne. y Reference category.

students, with no gender difference. The association was stronger when inammatory lesions were considered.10 It was postulated that obesity may be accompanied by peripheral hyperandrogenism, which may be associated with increased sebum production and the development of severe acne. At variance with these results, a study conducted in the United Kingdom within the Glasgow Alumni Cohort failed to document any association between acne and BMI.11 However, information on acne was based on notes reported during the students visit to the university health services, and the prevalence of acne was remarkably low. A significant correlation between acne lesion counts and BMI in men aged 18 to 25 years but not in participants younger than 18 years was documented in a randomized study of low-glycemic-load diets to improve acne lesions.12 In a study of women with polycystic ovary syndrome conducted in Saudi Arabia, acne was associated with overweight and obesity.13 On the contrary, in another study from Taiwan investigating the impact of obesity on cutaneous manifestations of clinical hyperandrogenism, obese women presented with less acne than nonobese subjects, although BMI had a significant positive correlation with serum total testosterone.14 As pointed to by a meta-analysis, no data are available in women given the diagnosis of polycystic ovary syndrome concerning the impact of diet, exercise, or behavioral or combined treatments on the prevalence of acne.15 Our data indicate a main protective effect on the development of acne from lower BMI values and a lack of a clear dose-response relationship. The effect

was stronger in male than in female individuals. The relationship between BMI and acne is intriguing and cannot be simply explained by hormonal factors. Interestingly, android but not gynoid obesity is associated with hyperlipidemia, and high lipid levels have been associated with severe acne.16 Published studies on the association between diet and acne risk are somewhat inconsistent. Most of the studies published up to the end of the past century found no signicant role for diet in acne. In a review of the literature published in 2005, Magin et al17 concluded that there was poor evidence for any role for dietary factors in the management of acne. More recent analyses supported the association between some dietary factors such as a Western diet or highglycemic-load diet and acne.3,12 In our study, a diet high in milk but not in cheese was associated with the risk of acne. This result is consistent with those reported in two separate large cohort studies performed in teenagers in the United States. In these studies, Adebamowo et al18,19 found a positive association between milk intake during adolescence and a history of acne. This association was more marked for skim milk than for other forms of milk. In our study, the estimates for skim milk were slightly higher than those for whole milk, but did not reach statistical significance, possibly owing to limited statistical power. The stronger association with skim milk makes it unlikely that the risk is related to the fat content of milk. This is reinforced in our study by the lack of an association of acne with other sources of animal fat such as beef. On the other hand, the hormonal constituents of skim milk could be of a sufficient amount to cause biological effects in young consumers. At variance with a long-lasting belief, we found no association between acne and chocolate consumption in our study. Similarly, a study performed by Fulton et al20 showed no association between acne, sebum, and comedones formation and the consumption of chocolate bars. The chocolate bars used in that study had no milk component. Recently, a study showed an association between acne and a high-glycemic-load diet, with improvement of symptoms after changing to a higher protein, lower-glycemic-load diet.12 It was postulated that Western diets characterized by a high glycemic index could lead to hyperinsulinemia, triggering a hormonal response capable of promoting unregulated tissue growth and enhanced androgen synthesis. In our study, weekly consumption of cakes/sweets and chocolate was not associated with a higher risk of acne. An interesting result of our study is the inverse relationship between moderate to severe acne and

1134 Di Landro et al

J AM ACAD DERMATOL

DECEMBER 2012

the intake of sh. To our knowledge, such an association was previously only reported in a casecontrol study from Korea.21 In that study, fish intake, namely, white fish and green fish or blue tuna, was significantly higher in the control group than in the acne group (P = .03). The association will require confirmation in additional studies. A few contrasting reports pointed toward a relationship between acne and smoking. Smoking has been linked with adverse effects on the skin owing to alterations in the skin microcirculation, keratinocytes, and collagen and elastin synthesis. Smoking could decrease antioxidants, which could cause alterations in sebum composition. An Italian study showed a correlation between smoking and postpubertal acne, leading to its being dened as smokers acne.22 Another report, however, found a lower prevalence of severe acne in actively smoking male individuals than in nonsmokers.23 We did not document any association, positive or negative, of smoking habits with acne. Finally, hormonal factors have commonly been associated with acne, but only a few studies have analyzed the role of menstrual history on the risk of the condition. We found no relationship between age at menarche, menstrual cycle pattern, or use of oral contraceptives and acne. Oral contraceptives have been linked, although not consistently, with a reduced prevalence of acne.2 In agreement with our results, the large-scale twin study conducted in the United Kingdom mentioned above found no association between acne and reproductive or hormonal potential risk factors apart from the use of oral contraceptives. Rather than being causal, this association could be explained by the fact that the contraceptive pill is one of the therapeutic options for acne.7 In conclusion, our study conrms the important role of a family history on the risk of moderate to severe acne, and suggests that lower BMI values, especially in boys and men, may have a protective effect. In agreement with studies conducted in the United States, a diet rich in whole and skim milk inuenced the risk of moderate to severe acne irrespective of family history and BMI. Finally, our study points to a diet rich in sh as being protective against acne severity. The inuence of environmental and dietetic factors in acne that develops in adolescents should be further explored.
REFERENCES 1. Ballanger F, Baudry P, NGuyen JM, Khammari A, Dr eno B. Heredity: a prognostic factor for acne. Dermatology 2006;212: 145-9. 2. Jemec GB, Linneberg A, Nielsen NH, Frlund L, Madsen F, Jrgensen T. Have oral contraceptives reduced the prevalence

3.

4. 5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

of acne? A population-based study of acne vulgaris, tobacco smoking and oral contraceptives. Dermatology 2002;204: 179-84. Cordain L, Lindeberg S, Hurtado M, Hill K, Eaton SB, Brand-Miller J. Acne vulgaris: a disease of Western civilization. Arch Dermatol 2002;138:1584-90. Burton JL, Cunliffe WJ, Stafford I, Shuster S. The prevalence of acne in adolescence. Br J Dermatol 1971;85:119. van Hooff MHA, Voorhorst FJ, Kaptein MBM, Hirasing RA, Koppenaal C, Schoemaker J. Relationship of the menstrual cycle pattern in 14-17 year old adolescents with gynecological age, body mass index and historical parameters. Hum Reprod 1998;13:2252-60. World Health Organization. Physical status: the use and interpretation of anthropometry. Report of a WHO expert committee. Technical report series No. 854. Geneva: World Health Organization; 1995. Bataille V, Snieder H, MacGregor AJ, Sasieni P, Spector TD. The influence of genetics and environmental factors in the pathogenesis of acne: a twin study of acne in women. J Invest Dermatol 2002;119:1317-22. Friedman GD. Twin studies of disease heritability based on medical records: application to acne vulgaris. Acta Genet Med Gemellol (Roma) 1984;33:487-95. Kirk KM, Evans DM, Farthing B, Martin NG. Genetic and environmental influences on acne in adolescent twins. Twin Res 2001;4:190. Tsai MC, Chen W, Cheng YW, Wang CY, Chen GY, Hsu TJ. Higher body mass index is a significant risk factor for acne formation in schoolchildren. Eur J Dermatol 2006;16:251-3. Galobardes B, Davey Smith G, Jeffreys M, Kinra S, McCarron S. Acne in adolescence and cause-specific mortality: lower coronary heart disease but higher prostate cancer mortality; the Glasgow alumni cohort study. Am J Epidemiol 2005;161: 1094-101. Smith RN, Mann NJ, Braue A, M akel ainen H, Varigos GA. A low-glycemic-load diet improves symptoms in acne vulgaris patients: a randomized controlled trial. Am J Clin Nutr 2007;86: 107-15. Tamimi W, Siddiqui IA, Tamim H, AlEisa N, Adham M. Effect of body mass index on clinical manifestations in patients with polycystic ovary syndrome. Int J Gynaecol Obstet 2009;107: 54-7. Yang JH, Weng SL, Lee CY, Chou SY, Hsu CS, Hsu MI. A comparative study of cutaneous manifestations of hyperandrogenism in obese and non-obese Taiwanese women. Arch Gynecol Obstet 2010;282:327-33. Moran LJ, Hutchison SK, Norman RJ, Teede HJ. Lifestyle changes in women with polycystic ovary syndrome. Cochrane Database Syst Rev 2011;2:CD007506. Arora MK, Seth S, Dayal S. The relationship of lipid profile and menstrual cycle with acne vulgaris. Clin Biochem 2010;43: 1415-20. Magin P, Pond D, Smith W, Watson A. A systematic review of the evidence for myths and misconceptions in acne management: diet, face-washing and sunlight. Fam Pract 2005;22: 62-70. Adebamowo CA, Spiegelman D, Danby FW, Frazier AL, Willett WC, Holmes MD. High school dietary dairy intake and teenage acne. J Am Acad Dermatol 2005;52:207-14. Adebamowo CA, Spiegelman D, Berkey CS, Danby FW, Rockett HH, Colditz GA, et al. Milk consumption and acne in teenaged boys. J Am Acad Dermatol 2008;58:787-93. Fulton JE Jr, Plewig G, Kligman AM. Effect of chocolate on acne vulgaris. JAMA 1969;210:2071-4.

J AM ACAD DERMATOL
VOLUME 67, NUMBER 6

Di Landro et al 1135

21. Jung JY, Yoon MY, Min SU, Hong JS, Choi YS, Suh DH. The influence of dietary patterns on acne vulgaris in Koreans. Eur J Dermatol 2010;20:768-72. 22. Capitanio B, Sinagra JL, Ottaviani M, Bordignon V, Amantea A, Picardo M. Acne and smoking. Dermatoendocrinol 2009;1: 129-35. 23. Klaz I, Kochba I, Shohat T, Zarka S, Brenner S. Severe acne vulgaris and tobacco smoking in young men. J Invest Dermatol 2006;126:1749-52.

GISED ACNE STUDY GROUP

Benevento: Presidio Ospedaliero Multizonale G. RummoeDr Francesco Cusano, Dr Errico Assunta Bologna: Bellaria MaggioreeDr Anna Lanzoni Cagliari: Ospedale Universit a degli StudieDr Laura Atzori Catania: Azienda Ospedaliera Universit a Policlinico G. RodolicoeDr Maria Letizia Musumeci, Prof Giuseppe Micali

Cremona: Azienda OspedalieraeDr Enrico Pezzarossa Ferrara: Azienda Ospedaliera UniversitariaeDr Vincenzo Bettoli, Prof Annarosa Virgili Firenze: Ospedale S. Maria Nuova-Orbatello I ClinicaeDr Marzia Caproni, Dr Emiliano Antiga, Dr Walter Volpi, Prof Paolo Fabbri Milano: Universit a Scienze DermatologicheeDr Mauro Barbareschi Napoli: PoliclinicoeDr Francesco Tripodi Cutr Napoli: Universit a Studi Federico IIeDr Gabriella Fabbrocini Prato: Ospedale Misericordia e DolceeDr Giovanni Lo Scocco Reggio Emilia: Arcispedale S. Maria NuovaeDr Angela Bonci Taranto: Centro Ospedaliero MilitareeDr Vito Ingordo, Dr Nicola Licci