Diet and Acne—Challenges of Translating Nutritional Epidemiologic Research Into Clinical Practice
John S. Barbieri, JAMA Dermatol. 2020;156(8):841-843.
There is growing evidence that diet may influence the incidence and severity of acne vulgaris, one of the most common skin conditions in the US.1 Much of the previous literature on the role of diet in acne has focused on the association of milk consumption and high glycemic-load diet with acne. For instance, milk consumption is thought to increase insulin and insulinlike-growth factor 1 levels, which can induce lipogenesis and proliferation of both keratinocytes and sebocytes.2,3 In addition, insulinlike-growth factor 1 can stimulate androgen synthesis and decrease production of sex hormone–binding globulin. Bovine insulinlike-growth factor 1 and androgens present in milk may also promote the development of acne.2,3
Observational studies, including a recent meta-analysis of 14 studies, have suggested an association between acne and milk intake.4 There is uncertainty whether milkfat percentage (ie, whole, low-fat, and skim) influences the strength of this association.
Whey protein, which represents 20% of the protein in cow’s milk, has also been suggested to contribute to the development of acne, and a small case series of 5 men found that discontinuation of whey protein supplement use resulted in apparent improvement of their acne.5
Similar to milk consumption, high glycemic-load diets have also been proposed to increase insulin and insulinlike-growth factor 1 levels, which could result in the development of acne.3 However, while observational studies and small trials have supported an association between high glycemic-load diet and acne,6,7 a Cochrane review on complementary acne therapies8 found insufficient evidence to support a low glycemic-load diet for the management of acne.
Although there have been several observational studies and small trials on the role of diet in acne, these studies have been limited by focusing on only a few domains of diet. In results from the NutriNet-Santé study published in this issue of JAMA Dermatology, Penso and colleagues9 report on the association between a broad range of dietary behaviors and adult acne. The NutriNet-Santé study is a prospective, observational cohort study that includes extensive dietary phenotyping via repeated online questionnaires. This cohort has been previously used to evaluate dietary associations with other inflammatory skin diseases, such as psoriasis.10 In November 2018, an acne questionnaire was distributed to this cohort to gather patient-reported data on acne history, which were used for this study. After adjusting for potential confounders, such as age, sex, total energy intake, smoking status, physical activity, and body mass index, consumption of milk (per glass: adjusted odds ratio [aOR], 1.12; 95% CI, 1.00-1.25), sugary beverages (per glass: aOR 1.18; 95% CI 1.01-1.38), and fatty and sugary products (per portion: aOR 1.53; 95% CI 1.09-2.16) were each found to be independently associated with current acne. Penso and colleagues9 also performed a principal components analysis in which a healthy dietary pattern was associated with decreased odds of current acne and a fatty and sugary dietary pattern was associated with increased odds of current acne. Similar patterns were observed in a subgroup analysis of women with acne.
These findings are largely in line with prior observational studies suggesting an association between milk consumption and high glycemic-load diet with acne.4,6,7 A small study in which college students were randomized to consume a milk chocolate bar or equivalent glycemic load of jelly beans reported increased acne activity in the chocolate bar group.11 Although an association between milk chocolate consumption and acne was noted in the univariate analysis, the present study by Penso and colleagues9 did not identify any association between milk or dark chocolate consumption and acne in their multivariate model.
The study by Penso and colleagues9 has several strengths, including its large sample size, attempts to control for potential confounders, and the comprehensive assessment of dietary habits of the participants. However, the study is not without limitations. Since the NutriNet-Santé cohort does not have data on milkfat percentage or whey protein consumption, the potential association between these factors and acne could not be assessed. In addition, given that many of the identified dietary associations in the multivariate analysis were only weakly significant (P = .01 to .05), there is a concern for spurious associations given the large number of dietary factors evaluated, although the similar findings in the principal coordinates analysis are reassuring.
As the authors discuss, an important challenge of nutritional epidemiologic research is the inability to determine whether there is a causal association between diet and the presence of acne. Unfortunately, associations identified in nutritional epidemiologic research are often unable to be replicated when evaluated in randomized trials.12 Given the complexity of diet and potential for confounding and selective reporting bias, some suggestions to improve the quality of nutritional epidemiologic research include improving transparency with prespecified analytic plans, increased data sharing, and independent analyses by other nutritional epidemiologists, including contrarians, to replicate study findings.13
In addition, effect sizes observed in this study were generally small (eg, ORs <1.2) and an association between diet and acne may not be clinically meaningful for many patients. Given these limitations, the findings of this study should be interpreted with caution with respect to clinical practice. There is a need for prospective, randomized trials to evaluate the influence of diet on acne. A small, 32-patient trial in Korea found that individuals randomized to a low glycemic-load diet had decreased sebaceous gland activity, greater reductions in acne lesions, and higher patient self-assessment scores at 10 weeks than those in the control diet group.6 Similar findings were found in an Australian trial involving 43 young adult men with acne.7 Future trials that assess both the mechanistic and clinical influence of diet on acne are needed to translate hypotheses generated from nutritional epidemiologic research into clinical practice.
Given the uncertainty surrounding both causal inference as well as the magnitude of the association between diet and acne, dietary interventions may be better thought of as adjunctive strategies rather than primary treatment options. Because some dietary interventions, such as avoiding dairy products, may have other relevant effects (eg, reduced calcium and vitamin D intake), they should be cautiously implemented. Nevertheless, given the potential overall health benefits of a healthy or low glycemic-load diet, and 2 small trials supporting its effectiveness in acne, a low glycemic-load diet is a reasonable recommendation for patients looking for dietary modifications that may improve their acne.
Overall, it is encouraging to see that the relatively comprehensive evaluation of the association between dietary factors and acne by Penso and colleagues9 is largely aligned with prior research identifying an association between high glycemic-load foods and milk consumption with acne. However, before we can confidently make dietary recommendations for patients with acne, trials are needed to evaluate whether dietary interventions can improve or prevent acne and how the effect size of such interventions compares with other standard treatment modalities.
