Portal de Periódicos da CAPES

Diet and acne: a review of the evidence

2009; Wiley; Volume: 48; Issue: 4 Linguagem: Inglês

10.1111/j.1365-4632.2009.04002.x

1365-4632

Elsa H. Spencer, Hope Ferdowsian, Neal D. Barnard,

Contact Dermatitis and Allergies

Acne vulgaris is the most common dermatologic condition in the USA, affecting more than 17 million Americans of all ages, although it is especially common in adolescents.1 Moreover, approximately 80–90% of American adolescents experience acne.2 Severe acne is associated with low self-esteem, poor body image, social withdrawal, and depression.3 Pharmaceutical acne treatments are costly and have potentially severe side-effects. Adolescent acne is typically the result of clogged, infected, pilosebaceous follicles. Adults may experience fewer comedones and more inflammatory lesions.1 Normally, sebum travels up the follicle to the skin surface. Hormones may increase sebum production and cause follicular cells to hyperproliferate and block the follicular opening, forming a comedo.4 Complete follicle blockage results in closed comedones (i.e. "whiteheads"), whereas incomplete blockage results in open comedones ("blackheads"). Comedo formation typically occurs over the course of 2–3 weeks. Acne may manifest in the form of noninflammatory comedones, superficial inflammatory lesions (papules, pustules), and/or deeper inflammatory lesions (nodules, cysts). Inflammatory lesion formation occurs most commonly when Propionibacterium acnes colonizes the pilosebaceous unit, triggering follicular rupture and a neutrophil cascade.5 Rarely, acne may have nonbacterial causes.6 In studies of diverse populations, individuals with acne commonly attribute the condition3, 7-9 or its exacerbation3, 7, 8, 10 to diet. Chocolate and oily or fatty foods are commonly implicated;7, 10-12 however, reviews prior to 2007 have concluded that diet plays no important role in acne and that the condition is primarily attributable to genetic predisposition and hormonal influences.13-15 Two large twin studies16, 17 have reported on the heritability of acne. Estimated heritability (genetic variance/phenotypic variance) ratios for acne risk and severity range from 0.5 to 0.9 among adolescent16 and adult17 pairs of monozygous and dizygous twins. Walton et al.18 reported that sebum excretion is influenced by genetic factors, but that the development of clinical disease is mediated by environmental factors. These studies suggest that genetic factors alone do not fully account for the acne risk. Despite the genetic regulation of sebum excretion and other determinants of acne, environmental influences, such as diet, may act as modifiers of gene expression. Recently, well-designed, controlled, prospective studies have supported the association between specific dietary factors and acne. We therefore critically examined the quality and strength of the published literature examining the association between diet and acne. We conducted a review of the relationship between diet and acne using the following keywords: "acne,""acne vulgaris,""diet,""nutrition,""food,""food allergy,""vitamin," and "chocolate." The following databases and periods were included: Medline since 1949; Embase-Medicine & Embase-Psychology (EMBASE) since 1980; the Cochrane Central Register of Controlled Trials (Cochrane) since 1898; Database of Abstracts on Reviews and Effectiveness (DARE) since 1990; PsycInfo since 1967; and the Cumulative Index to Nursing and Allied Health Literature (CINAHL) since 1982. Articles were also obtained by bibliography review. Articles published in languages other than English were reviewed if English translations were available. Observational and interventional human studies with participants of any age, sex, or health status were included. Articles were excluded if they reported only on the potential effects of topical, herbal, or vitamin preparations, or on the diagnosis, treatment, or pathogenesis of acne. A small sample size, a lack of a control group, and unclear statistical methods were not reasons for exclusion. The primary author (EHS) reviewed the titles and abstracts of all potentially relevant articles to determine whether they met the eligibility criteria. Of the 59 abstracts and full articles retrieved, 31 articles were excluded for reporting only on the potential effects of topical, herbal, or vitamin preparations, or on the diagnosis, treatment, or pathogenesis of acne. The remaining 28 articles were included. We excluded one study that did not report on diet. Of the 27 relevant articles, 21 were observational studies and six were clinical trials (Table 1). We found 15 cross-sectional,3, 7-12, 19-26 two case–control,27, 28 and four prospective cohort29-32 studies, as summarized in Table 1. Population-based studies suggest that acne prevalence is lower in rural societies than in industrialized populations. Cordain et al.19 studied the Kitavan islanders of Papua New Guinea (n = 1200) and the Aché hunter-gatherers of Paraguay (n = 115). The islanders subsisted mainly on root vegetables, fruit, fish, and coconut. Their intake of dairy products, coffee, alcohol, cereals, oils, sugar, and salt was minimal. An estimated two-thirds of the Aché hunter-gatherer diet consisted of sweet manioc, peanuts, maize, and rice. Approximately one-quarter of their diet consisted of flour, sugar, and meat. No cases of acne were detected in either population. The authors suggested that the low fat intake and the absence of high-glycemic-index foods may explain the low prevalence of acne in these populations. Freyre et al.20 compared acne prevalence in three Peruvian populations, including indigenous and white populations (n = 2214). Among 12–18-year-olds, the indigenous population showed a significantly (P < 0.001) lower acne prevalence (28%) than the white population (45%) or those of mixed ancestry (43%). Each adolescent group had a lower prevalence of acne than that reported in 12–18-year-old Americans.20 Bechelli et al.21 assessed the prevalence of acne in 9955 Brazilian schoolchildren: 8980 were impoverished urban children, whereas 975 were from rural areas. Less than 3% of the combined population (2.7%) demonstrated evidence of acne. Two reports have suggested that acne prevalence increases as populations adopt a Western diet through migration or cultural change. Reports of northern Canadian Inuits made no mention of acne until acculturation with their southern neighbors and subsequent increases in soda, beef, dairy products, and processed foods, after which the acne prevalence increased.22 Pre-World War II Okinawans, who traditionally followed a diet of sweet potatoes, rice, and vegetables, together with some soybeans, but little meat, reported an increase in acne prevalence after adopting a diet high in animal products.23 Seven studies have assessed the perceptions of factors believed to affect acne.3, 7-12 In a 2007 study, Rigopoulos et al.7 assessed the beliefs about acne among 13–18-year-old Greek students with and without acne. Self-reported acne was present among 59% of students. Among 316 students with and without acne, 62% cited diet as a causal factor, and 66% believed chocolate was an exacerbating factor. In 2006, El-Akawi et al.10 reported that, of 166 Jordanian male and female untreated clinic patients with acne, participants believed that their acne was aggravated by nuts (89%), chocolate (85%), cakes/biscuits (57%), oily food (53%), fried food (52%), eggs (42%), or milk, yogurt, and cheese (23%). Nearly one-fifth (19%) believed that consuming fruits and vegetables improved their acne. A survey administered by Ikaraoha et al.11 to 174 Nigerian students aged 18–32 years demonstrated that 75% of participants believed that an oily or fatty diet contributed to their acne. Of 130 male and female Saudi Arabian patients attending an acne clinic, more than one-quarter of participants believed that diet caused (26%) or exacerbated (32%) their acne.8 A 2003 study9 of Saudi Arabian high school and college students (n = 517), aged 15–29 years, demonstrated that 72% of students with and without acne (and 79% of 217 students with acne) believed that diet contributed to acne. In a 2001 study in the Journal of the American Academy of Dermatology, Tan et al.3 reported that acne was believed to be caused by diet less frequently than by hormonal or genetic factors. Although 32% of acne patients believed that diet caused their acne, 64% believed that hormones and 38% believed that genetics were responsible. In addition, 44% of participants believed that diet aggravated their acne. In 2001, Green and Sinclair12 reported that almost half of 215 sixth-year Australian medical students believed that diet aggravated their patients' acne, citing chocolate and oily or fatty foods as the most common dietary factors. Kaymak et al.24 examined the association between acne and the glycemic index and glycemic load of the daily diet, insulin sensitivity, and insulin-like growth factor (IGF) levels in 91 university students (n = 49 acne patients, n = 42 control patients). Participants completed a food frequency questionnaire from which the authors calculated the glycemic index using published reports. Physicians assessed acne as well as insulin resistance through the calculation of the homeostatic model assessment (HOMA) index: [fasting insulin (microU/mL) × fasting blood glucose (mmol/L)/22.5]. There were no significant differences in fasting glucose or insulin levels, and none of the participants had insulin resistance. In patients with acne, levels of IGF-1 were higher and levels of IGF-binding protein-3 were significantly lower than those of controls. Participants with acne of more than 2 years' duration ate a diet with a significantly higher glycemic index than did participants with acne of less than 2 years' duration. Khanna et al.25 studied energy, carbohydrate, protein, and fat intake among 200 students in India with and without acne. They compared the diets of those with severe, moderate, mild, and no acne, and reported no dietary differences (P > 0.05) among the four groups; however, the researchers used a t-test, intended for the comparison of two groups, rather than the appropriate analysis of variance statistical test to compare the four groups. It is unclear whether the results from appropriate statistical tests would have differed from those reported. In a 1956 study published in the British Medical Journal, Bourne26 found that adult British soldiers, aged 20–40 years, with acne were significantly heavier than those without acne (73.2 kg vs. 67.5 kg; P = 0.013). Adolescent British soldiers aged 15–19 years with acne weighed more than those without acne (62.7 kg vs. 60.5 kg), although the differences were not significant. In a 2005 study, Adebamowo et al.27 tested the hypothesis that milk (whole, powdered, low-fat, and skimmed) intake was associated with a risk of teenage acne. More than 47,000 nurses were questioned about their adolescent diets and whether they had experienced "physician-diagnosed severe acne" during their teenage years. Prevalence ratios (PRs), comparing acne prevalence at the highest (more than three servings per day) to lowest (one serving or less per week) intake categories, were computed. In multivariate models adjusted for energy, present age, age of menarche, and body mass index, total milk intake was associated with severe acne [PR = 1.22; 95% confidence interval (CI), 1.03–1.44]. Severe acne prevalence increased as the milk fat content decreased: PR = 1.12 (CI, 1.00–1.25) for whole milk; PR = 1.16 (CI, 1.01–1.34) for low-fat milk; PR = 1.44 (CI, 1.21–1.72) for skimmed milk. Trend tests were significant for total milk [P(trend) = 0.002] and skimmed milk [P(trend) = 0.003]. The only nutrients significantly and positively associated with acne were vitamin D supplementation and total energy intake. High saturated fat intake was inversely associated with acne risk (PR = 0.88; CI, 0.80–0.94). There were no associations between acne and soda, French fries, pizza, or chocolate. The authors hypothesized that the hormones found in milk products were responsible for milk's association with acne. Bett et al.28 tested the hypothesis that acne patients consumed more sugar than age- and sex-matched controls. They compared the sugar consumption in 16 patients with acne with that in 16 patient controls with warts and 16 healthy age- and sex-matched office and factory worker controls. Diet was assessed by food frequency questionnaire. There were no significant differences in sugar consumption among the groups (121 g/day for acne patients vs. 111 g/day for wart patients and 120 g/day for healthy controls; P > 0.05). Four prospective cohort studies have evaluated the associations between diet and acne.29-32 Adebamowo et al. followed 4273 boys29 and 6094 girls,30 aged 9–15 years, in 1996. Information on dietary intake was collected between 1996 and 1998, and acne prevalence and severity were assessed in 1999. The authors investigated the association between self-reported acne severity and cow's milk (whole/2%, 1%, skimmed, chocolate) intake. PRs, comparing acne prevalence at the highest (two or more servings per day) to lowest (one serving or less per week) intake categories, were computed. After adjustment for baseline age, height, and energy intake, acne severity in girls30 was significantly associated with the intake of all categories of cow's milk: PR = 1.20 [CI, 1.09–1.31; P(trend) < 0.001] for total milk; PR = 1.19 [CI, 1.06–1.32; P(trend) < 0.001] for whole milk; PR = 1.17 [CI, 1.04–1.31; P(trend) = 0.002] for low-fat milk; PR = 1.19 [CI, 1.08–1.31; P(trend) < 0.001] for skimmed milk; PR = 1.29 [CI, 1.08–1.53; P(trend) = 0.02] for chocolate milk. In boys,29 the association between milk intake and acne was significant for total milk (PR = 1.16; CI, 1.01–1.34) and skimmed milk (PR = 1.19; CI, 1.01–1.40) intake. The test for trend was significant only for skimmed milk intake [P(trend) = 0.02]. Among both boys and girls, there were no significant associations between acne and the intake of nonmilk dairy foods, French fries, pizza, or chocolate. In a 2003 study, 22 university students were followed for one semester by Chiu et al.31"Dietary quality" was determined by the number of meals eaten per day and subjective self-ranking of "diet quality" on a four-point scale. Descriptive dietary data were not provided. Perceived dietary quality was inversely associated with acne exacerbation and severity (r = –0.48, P = 0.02). During the course of a study on radiation treatment for acne in 2083 patients between 1925 and 1949, Robinson32 collected 1–2 weeks of dietary records in a subset of patients. The number of dietary records was not reported. Milk products were the most frequently cited acne-causing food. Patients were subsequently advised to follow a low-fat, dairy-free diet; however, postintervention follow-up results were not published. Table 1 summarizes the results of six dietary intervention studies.33-38 Three33, 34, 36 of the six studies included control groups. Smith et al.34 assessed the effect of a low-glycemic-load diet (25% energy from protein and 45% energy from low-glycemic-index carbohydrates) on acne and insulin sensitivity. Participants (n = 43, all male, aged 15–25 years) were randomly assigned in a parallel design to the dietary intervention or control group urged to regularly include carbohydrates without receiving information on the glycemic index. Participants were followed for 12 weeks. Blind dermatologists assessed the number of acne lesions every four weeks, starting at baseline. Relative to those on the control diet, participants on the low-glycemic-load diet experienced greater reductions in counts of all lesions (51% vs. 31%; P = 0.03) and inflammatory lesions (45% vs. 23%; P = 0.02). Participants in the intervention group experienced a significant improvement (i.e. increase) in insulin sensitivity and significant changes in androgen levels, compared with participants in the control group.39 A positive correlation was observed between the change in total lesion counts and the change in insulin sensitivity as measured by the HOMA index (r = 0.38, P = 0.01). A change in sex hormone-binding globulin (SHBG) levels also correlated negatively with a change in lesion counts (r = −0.38, P = 0.01). Sebum sampling was completed by 31 participants (n = 16 intervention participants, n = 15 control participants).33 At baseline and 12 weeks, follicular sebum outflow and the composition of skin triglycerides were assessed. Follicular sebum outflow and the proportions of sebum fatty acids did not differ between the groups. Compared with baseline levels (P = 0.007), however, participants in the intervention (but not the control) group demonstrated an increased ratio of saturated to monounsaturated sebum fatty acids. The change in ratio correlated negatively and significantly with the change in acne lesion counts (r = −0.39, P = 0.03). The study authors concluded that the desaturation of sebum fatty acids may play a role in acne development. Anderson35 tested the effect of four foods on acne. For 1 week, medical students were asked to consume daily servings of a self-selected test food [6 small (39-g) chocolate bars, 0.95 L of milk, 113 g of roasted (iodized) salted peanuts, or 0.71 L of cola]. Nine participants developed new acne lesions in 10 days of the tests. Nine participants developed no new lesions. The results were not grouped by test foods, and the study was neither controlled nor blind. The authors reported that the self-selected foods were not associated with acne. Two studies specifically tested the acne-causing ability of milk chocolate. In an uncontrolled intervention in eight university students, Grant and Anderson38 fed participants a 9.75-oz milk chocolate bar and instructed them to eat an identical second bar the next day. By day five, four participants had developed new lesions, but four had not. Statistical analyses were not presented. The investigators concluded that large amounts of chocolate did not aggravate acne. In a single-blind crossover trial36 designed to test the effect of chocolate on acne, 65 participants were asked to consume one 112-g dairy-free chocolate bar every day for 4 weeks. The bar was enriched 10-fold with cocoa solids and cocoa butter. During the control phase, subjects were instructed to consume a similar sized chocolate bar enriched with 28% partially hydrogenated vegetable (i.e. "trans") fat in place of the cocoa paste and cocoa butter. A 30% increase or decrease in acne lesions was considered to be clinically significant. Differences between experimental and control groups did not reach the designated clinical threshold and were not specifically reported. The authors concluded that large amounts of chocolate did not significantly affect acne, sebum composition, or sebum secretion. In 1965, Gaul37 prescribed a 3-month low-salt diet for 30 acne patients with baseline urinary chloride excretion of at least 12–20 g/L (the salt intake was not reported). In four cases, salt restriction reduced the number and severity of pustules and cysts. Quantitative analyses and data for the 26 other study participants were not reported. Population-based studies have suggested that, as diets Westernize, acne prevalence increases. Observational studies, including one case–control study27 and two large and well-controlled prospective cohort studies,29, 30 have demonstrated an association between cow's milk intake and acne prevalence and severity (Table 2). One prospective cohort study24 demonstrated an association between high-glycemic-index foods and longer acne duration, whereas two randomized controlled trials33, 34 demonstrated that a low-glycemic-index diet reduced acne risk. Studies have been inconclusive with respect to the association between acne and chocolate (perhaps because of methodological limitations), and have failed to demonstrate an association between acne and salt or iodine intake. Methodologic issues may have limited the conclusions that could be drawn from the literature before 2005. Prior to that date, two prospective studies24, 32 failed to explicitly state the follow-up duration, only one intervention36 utilized a control group, and another intervention38 failed to clearly define the changes in acne. Published studies suffered from a small sample size, lack of appropriate controls, and incomplete reporting of results. Recently published prospective studies24, 29, 30, 33, 34 have utilized clearly defined interventions, outcomes, and appropriate statistical analyses. Previous studies have provided the theoretical basis for an association between diet and acne. Acne typically results from excess sebum production, which causes follicular cells to hyperproliferate and block the follicle opening. Bacteria can then hypercolonize the follicle, and an immune response results in inflammation. Sebum production may be influenced by androgens and hormonal mediators, such as SHBG and IGF-1, all of which may be influenced by dietary factors, as described below. Three large studies reported a positive association between milk intake and acne. The studies of Adebamowo et al.29, 30 demonstrated that higher levels of milk consumption were associated with acne risk in both boys and girls. In the Nurses' Health Study II,27 women who consumed more milk as adolescents (based on later recall) showed a greater prevalence of severe acne than those with less frequent consumption. These findings are supported by previous population-based studies,19, 22, 23 in which areas with minimal consumption of dairy products had a very low acne prevalence. In recent observational studies, skimmed milk was consistently associated with acne, suggesting that the fat content of milk does not appear to affect its acne-causing ability. Some authors have reported that the hormones in milk, such as IGF-1, 5α-reduced steroids, and α-lactalbumin, may survive milk processing and affect the pilosebaceous unit.40 Milk consumption also increases IGF-1 production, which has been associated with ovarian androgen production in premenarchal girls and acne in adult women.41, 42 Findings from the studies of Smith et al.33, 34, 39 have illustrated the various interactions between glycemic load, insulin sensitivity, hormonal mediators, and acne. Regular consumption of foods with a high glycemic index elevates serum insulin concentrations, which may stimulate sebocyte proliferation and sebum production,43 suppress SHBG concentrations and raise androgen concentrations,44 and contribute to acne. Conversely, low-glycemic-index foods have been shown to increase SHBG and reduce androgen levels; higher SHBG levels have been associated with lower acne severity. Intervention participants in Smith's study34 also lost weight, and the low-glycemic-load diet was higher in polyunsaturated fat and fiber and lower in saturated fat, compared with the comparison diet. These differences may be expected to influence inflammation and acne-related hormones. A 1997 study of 871, 10–11-year-old girls followed prospectively for 5 years45 found that those with severe (vs. mild or moderate) comedonal acne had significantly higher androgen levels and significantly earlier menarche. Although diets high in saturated fat increase the concentration of IGF-1, low-fat, high-fiber diets tend to decrease the concentrations of IGF-146 and androgens,47 and increase the concentration of SHBG. The ratio of omega-6 to omega-3 fatty acids in Western diets is commonly at least 10 : 1, compared with ratios of 4 : 1 in Japan and 2 : 1 in historic, nonindustrialized populations.48 Studies have suggested that inflammatory markers increase as this ratio increases. Omega-6 fats are precursors to proinflammatory mediators and have been associated with the development of inflammatory acne.6 In contrast, high levels of omega-3 fatty acids have been shown to decrease inflammatory factors,49 and may reduce acne risk by decreasing IGF-1 levels and preventing hyperkeratinization of sebaceous follicles. Grant and Anderson38 and Anderson35 tested the acne-causing ability of milk chocolate bars. Grant and Anderson38 did not provide a control group, and neither study reported quantitative statistical analyses. Fulton et al.36 provided a control group and quantified their results. Moreover, they specifically tested cocoa rather than nonfat milk solids, milkfat, or sugar; however, the fat and sugar contents of treatment and control bars were nearly identical, minimizing the potential to detect the effects of fats or sugars on acne; the study findings did not provide conclusive evidence to establish whether cocoa solids influence acne formation. Evidence suggests that components of Western diets, particularly dairy products, may be associated with acne. The hormonal effects of dietary components, such as glycemic index levels or fat or fiber intake, may mediate the effect of diet on acne risk. Until 2005, cross-sectional, case–control, cohort, and clinical intervention studies designed to address the relationship between diet and acne typically failed to incorporate adequate controls, objective measures, and appropriate statistical analyses. Well-designed prospective studies published since 2005 have elucidated the mechanisms whereby particular foods and dietary constituents may influence acne risk and severity. In order to test the efficacy of dietary interventions, prospective, randomized trials, including controls for environmental stressors, acne medications, age, pubertal stage, and age at menarche, are essential.