Correlation Between Serum Levels of Insulin-like Growth Factor 1, Dehydroepiandrosterone Sulfate, and Dihydrotestosterone and Acne Lesion Counts in Adult Women

Objectives  To determine if insulin-like growth factor 1 (IGF-1) and androgen levels (1) correlate with the presence and severity of acne in adult men and women, and (2) correlate directly with each other and interact in affecting acne.

Design  Case-control study and single-center examination of hormone levels in a cohort of volunteers.

Setting  Academic referral center.

Patients  Thirty-four subjects (8 women and 8 men with clinical acne, 10 women and 8 men without clinical acne). Clinical acne is defined by a history of persistent acne (acne present on most days for several years), recent acne treatment, and the presence of 10 or more inflammatory acne lesions and 15 or more comedones.

Interventions  Single visit for serum sampling.

Main Outcome Measures  Serum levels of IGF-1 and androgens were determined, adjusted for age, and compared based on the presence or absence of clinical acne using an analysis of covariance. Correlations between hormone levels and acne lesion counts were calculated within each subgroup. Correlations were also calculated between serum levels of IGF-1 and androgens. Further statistical testing was conducted to determine whether IGF-1 or androgens had a greater effect on acne lesion counts.

Results  Dehydroepiandrosterone (DHEAS), dihydrotestosterone (DHT), and IGF-1 correlated positively with acne lesion counts in women. Androstenedione and DHEAS correlated with acne lesion counts in men. Although the age-adjusted mean serum levels of IGF-1 were higher in women with clinical acne than in women without clinical acne, this difference did not achieve statistical significance. No difference in IGF-1 level was noted in men based on the presence of clinical acne. In women with clinical acne, IGF-1 correlated with DHT. In men with clinical acne, IGF-1 correlated with DHEAS and androstenedione. In men and women with clinical acne, the effects of androgens on increased acne lesion counts were dependent on the influence of IGF-1.

Conclusions  Increased IGF-1 levels in addition to androgens may influence acne in adult men and women. While IGF-1 appears to have a stronger effect on acne in women, androgens may play a greater role in acne for men. However, in both men and women these hormones are interrelated, possibly owing to reciprocal effects on hormone production.

Sebum production is one of the key factors in the development of acne. Maximum sebum production begins during puberty, which coincides with the peaking levels of growth hormone and insulin-like growth factor 1 (IGF-1) that occur in midpuberty.^1,2 Accordingly, elevated levels of growth hormone and IGF-1 may play a role in the development of acne. Since growth hormone is released in intermittent secretory bursts and is thus impractical to assess by random serum sampling, increased levels have not been directly demonstrated in patients with acne. On the other hand, IGF-1 is relatively stable and primarily reflects cumulative secretion of growth hormone, which makes it more suitable for serum testing. One study noted elevated serum levels of IGF-1 in 82 women with acne aged 20 to 25 years when compared with 31 age-matched controls.³

Similar to that of growth hormone and IGF-1, the level of dehydroepiandrosterone (DHEAS), the major adrenal androgen precursor, progressively increases during puberty. It appears to be a crucial factor in the initiation of sebum secretion in the prepubertal period and perhaps beyond.⁴ Since both IGF-1 and DHEAS follow a similar chronological trend, there may be a relationship between these 2 hormones. Correlations between DHEAS and IGF-1 have been found among other groups of patients, including obese women and normal prepubertal girls.^5,6 In addition to noting elevated IGF-1 levels, Aizawa and Niimura³ noted elevated levels of DHEAS in postadolescent women with acne who were aged 20 to 25 years. However, DHEAS levels did not correlate with those of IGF-1.

Several studies have shown that levels of androgens are higher in adult women with acne than in those without acne, yet there is little information available about the correlation of androgen levels with acne lesion counts in adults. In women in their 20s with acne, free testosterone levels correlated with the number of pustules, but no correlations were noted between levels of androgens and acne lesions in men.⁷ In preadolescent and adolescent girls, elevated DHEAS levels have also been correlated with acne severity but not with lesion counts.^8,9

The goal of the present study is to test the following hypotheses in adults with acne: (1) that IGF-1 levels are elevated in patients with clinical acne; (2) that IGF-1 correlates with androgen levels; and (3) that the extent of acne (lesion counts) correlates with levels of IGF-1 and androgens.

This study demonstrates that the number of total acne lesions, inflammatory lesions, and comedones each correlated with serum IGF-1 levels in women with clinical acne. In this group, serum IGF-1 levels also correlated with serum dihydrotestosterone (DHT) levels. However, the correlation of IGF-1 level with number of comedones and inflammatory lesions was not independent of the effects of DHT. In adult women, DHEAS and DHT also correlated with acne lesion counts (total, inflammatory, and comedone). Although the age-adjusted mean serum IGF-1 level was higher in women with clinical acne, this difference only approached statistical significance (P = .07).

The institutional review board of The Pennsylvania State University College of Medicine approved this study. Men and women aged 18 to 45 years with and without clinical acne were recruited for this study. Participants were questioned about severity, duration, and prior treatment for acne. Participants were classified as having clinical acne based on a positive history of persistent acne (acne present on most days for several years) and recent acne treatment and objective physical findings of a minimum of 15 comedones and 10 inflammatory papules on the face. One participant with a history of persistent acne and extensive comedonal acne (45 comedones) but only 8 inflammatory papules was also included in the clinical acne group. Participants were excluded from the study if they had received oral antibiotics within 1 month of the study; used topical acne medications such as antibiotics, benzoyl peroxide, or tretinoin within 2 weeks prior to the study; received prior treatment with oral retinoids; had known endocrine disease such as congenital adrenal hyperplasia, polycystic ovary syndrome, or adrenal or ovarian tumor; or were currently using any type of hormonal contraceptive, estrogens, corticosteroids, or finasteride. Women were observed in the luteal phase of their menstrual cycle. Data on levels of serum androgens and tissue androgen-metabolizing enzymes (5α-reductase and 17β-hydroxysteroid dehydrogenase) were previously reported in these subjects.¹⁰

Based on clinical examination, the numbers of comedones and inflammatory lesions were recorded, and total acne lesions were determined.

Levels of IGF-1 were determined by acid ethanol extraction and double antibody radioimmunoassay with intra-assay and interassay variance of 3.0% and 6.0%, respectively.¹¹ Since IGF circulates as a complex with binding proteins that inhibit its interaction with IGF-1 receptors, IGF binding protein 3 (IGFBP-3) was also determined from banked serum from the female subjects. Levels of IGFBP-3 were also measured via double antibody radioimmunoassay with an intra-assay variance of 3.4% to 8.0% and an interassay variance of 5.3% to 6.3%.¹² Androstenedione, testosterone, DHEAS, and DHT levels were determined by means of radioimmunoassay methods as previously reported.¹⁰

Statistical Analysis System version 8 (SAS Institute Inc, Cary, NC) was used to analyze data from the following subgroups: all women; women with clinical acne; women without clinical acne; all men; men with clinical acne; and men without clinical acne. An analysis of covariance was used to compare the age-adjusted serum hormone levels between women with clinical acne and women without clinical acne. The same analysis was performed for men with and without clinical acne. Serum levels of IGFBP-3 were also compared between women with and without clinical acne using analysis of covariance. For all 6 of the subgroups, correlation coefficients were calculated between the various hormone levels (IGF-1, DHEAS, androstenedione, testosterone, and DHT) and the acne lesion counts (total lesions, comedones, and inflammatory lesions).

In addition, correlation coefficients were calculated between the levels of various androgens and IGF-1 to determine the presence of hormone-hormone associations. When such associations were found, a multivariable regression analysis was used to determine if either IGF-1 or androgens independently correlated with acne lesion counts. Partial correlation coefficients were calculated between each independent variable (IGF-1 and androgen) and the dependent variable (acne lesion count). If either hormone had a partial correlation coefficient with a statistically significant P value, it was considered that this hormone exerted effects independent of the other hormone on the acne lesion count. On the other hand, if the correlation coefficient P value was not statistically significant, this indicated that the hormone’s effect on acne lesion count was not independent of the other hormone. Any t test finding, correlation coefficient, or multivariable regression analysis that produced a P value of .05 or lower was considered to demonstrate a significant association.

A total of 34 subjects participated in the study: 8 women with clinical acne (mean ± SD age, 31 ± 5 years), 10 women without clinical acne (age 33 ± 7 years), 8 men with clinical acne (age 27 ± 7 years), and 8 men without clinical acne (age 34 ± 7 years). Table 1 lists the range of the various acne lesion counts in the patient subgroups and details a significant difference in lesion counts between subjects with and without clinical acne.

The age-adjusted mean hormone levels were determined in each subgroup based on sex and the presence or absence of clinical acne (Table 2). While women with clinical acne had elevated IGF-1 and DHT levels compared with women without clinical acne, this difference only approached statistical significance (P = .08 and P = .06, respectively). No statistically significant difference in IGF-1 level was noted in men. No differences were noted in IGFBP-3 levels based on the presence of clinical acne in women.

Androstenedione, testosterone, and DHEAS levels were significantly greater in the women with clinical acne than in women without clinical acne, as previously reported.¹⁰ On the other hand, men did not have a significant difference in androgen levels based on the presence of clinical acne (Table 2). These data are reiterated herein to provide a framework for demonstrating the correlation between serum hormone levels and acne lesion counts.

In the all-women group, IGF-1, DHEAS, and DHT levels correlated with total acne lesions, comedones, and inflammatory lesions. Androstenedione and testosterone levels correlated with inflammatory lesions in this group as well. In women with clinical acne, IGF-1 levels correlated with total acne lesions, comedones, and inflammatory lesions (Table 3).

In the all-men group, DHEAS and androstenedione levels correlated with total acne lesions and comedones, with the exception of a lack of correlation between androstenedione and comedones. In men with clinical acne, DHEAS and androstenedione correlated with total acne lesions and comedones. No correlation between IGF-1 and acne lesions was noted in men (Table 3).

Statistically significant correlations were found between IGF-1 and certain androgens in the all-women group, women with clinical acne, and men with clinical acne (Table 4). In the all-women group, IGF-1 correlated withDHEAS (r = 0.54; P<.05) and DHT (r = 0.75; P<.001). In women with clinical acne, only DHT correlated with serum IGF-1 (r = 0.76; P<.05). In the all-men group, neither DHEAS nor androstenedione correlated with IGF-1. On the other hand, in men with clinical acne, DHEAS and androstenedione levels correlated with IGF-1 (r = 0.86; P<.01 and r = 0.82; P<.01, respectively) (Table 4).

The data in Table 3 demonstrate that IGF-1 and androgens each correlated with various acne lesion counts in the subgroups. Because of the potential for a joint influence of IGF-1 with an androgen, a multivariable regression analysis was performed when a significant correlation was noted between IGF-1 and a particular androgen as seen in Table 4. In the all-women group, IGF-1 correlated with DHEAS and DHT. A multivariable regression analysis demonstrated that the partial correlations of DHEAS and DHT with total lesions and with inflammatory lesions were each independent of IGF-1 (Table 5). However, the partial correlations of DHEAS and DHT with comedones were dependent on the effects of IGF-1. In women with clinical acne, IGF-1 correlated with DHT. The multivariable regression analysis showed that the partial correlation of IGF-1 with comedones and the partial correlation of IGF-1 with inflammatory lesions were each dependent on the correlation of IGF-1 with DHT. In contrast, IGF-1 correlated with total lesions independently of its correlation with DHT. In the all-men group, no correlation of IGF-1 with serum androgens was noted. In men with clinical acne, IGF-1 correlated with DHEAS and androstenedione. The multivariable regression analysis demonstrated that the partial correlations of DHEAS and androstenedione with total lesions and comedones were dependent on the correlation of these androgens with IGF-1 (Table 5).

Sebum production is a major factor in the pathogenesis of acne. Growth hormone and IGF-1 may play roles in sebaceous gland physiology as evidenced by the expression of receptors for growth hormone and IGF-1 on human sebaceous glands and in sebocytes.^13,14 Studies of the expression of growth hormone and IGF-1 receptors in human sebaceous glands and analysis in rat preputial glands suggest that these factors stimulate sebocyte differentiation and proliferation.^13,15 Clinically, in teenaged to middle-aged patients with acromegaly, sebum excretion rate correlates positively with the output of growth hormone by the pituitary.^16,17 Aizawa and Niimura³ reported that IGF-1 levels were significantly higher in 20- to 25-year-old women with acne than in controls: 6 women with acne had IGF-1 levels higher than the normal range. Since growth hormone stimulates IGF-1 production, it was hypothesized that these findings reflected an increased growth hormone secretion in women with acne.

Our study, performed in older women (mean ± SD age, 33 ± 6 years) supports the findings that IGF-1 levels are higher in adult women with acne than in those without acne. However, these differences did not achieve statistical significance, perhaps owing to our small sample size. As in the case of serum androgens that were previously reported in these patients, the mean IGF-1 levels were within the reference range in both groups with the exception of 1 woman (age 42 years) in the clinical acne group (IGF-1 level, 437 ng/mL).¹⁰ Furthermore, in the present study, IGF-1 level correlated with total acne count, comedone count, and inflammatory lesion count in adult women. This correlation between IGF-1 and total acne lesion count was especially strong in women with clinical acne (r = 0.88; P = .004). In men, however, there were no significant differences in serum levels of IGF-1 based on the presence of clinical acne and no correlation of IGF-1 with acne lesion counts (Table 3).

These and previous data imply that IGF-1 or another factor with a direct relationship to IGF-1 (such as growth hormone) plays a role in acne in women. It appears that even normal serum IGF-1 levels may contribute to acne in women. Based on the strong correlation of IGF-1 levels and acne lesion counts in women with clinical acne, it is possible that these women are more sensitive to variations in serum IGF-1 levels, whereas the putative effects of IGF-1 on sebaceous glands in men may be masked by the overriding effects of higher serum androgen levels.

Serum DHEAS levels correlate with the presence of acne in prepubertal children and in adolescent girls aged 14 to 17 years.^9,18 Levels of DHEAS were significantly elevated when compared with controls in women in their 20s with acne but not in men.⁷ In the present study, DHEAS levels were significantly higher in an older group of adult women with acne than in those without acne, but again not in men (Table 2). Since adrenal androgens (such as DHEAS) are the major contributors to circulating androgens in women and girls, it seems logical that the presence of acne relates to serum DHEAS levels more in women than in men. Since androgen levels are much lower in women, other hormones such as IGF-1 or growth hormone could have a greater effect on acne in women than in men.

In both men and women, DHEAS correlated with all acne lesion counts (total, comedo, and inflammatory). Although there was no significant difference in the age-adjusted mean serum levels of DHEAS in men with and without clinical acne, a significant correlation of DHEAS with acne lesion counts was observed within the all-men group and men with clinical acne, which suggests that the extent of acne (as indicated by lesion counts) directly correlates with serum DHEAS level in both men and women (Table 3).

Numerous studies point to an interrelationship between IGF-1, insulin, and androgens. These relationships are complex and not completely understood. They are best exemplified in women with polycystic ovary syndrome, which is characterized by insulin resistance leading to hyperinsulinemia, hyperandrogenemia, and increased acne. Insulin at high levels, as found in insulin-resistant subjects, can interact with the IGF-1 receptor. In addition, some studies have demonstrated elevated IGF-1 levels in women with polycystic ovary syndrome.¹⁹

Increasing levels of serum androgens appear to correlate with a rise in IGF-1 level. For example, when DHEAS was administered to postmenopausal women, IGF-1 level also increased.²⁰ Similarly, when oophorectomized women were given testosterone, a linear increase occurred in the IGF-1–IGFBP-3 ratio but not a decrease in IGFBP-3 levels.²¹ These findings suggest that increasing serum androgens may directly stimulate production of IGF-1.

Conversely, IGF-1 may also stimulate DHEAS production by the adrenal gland or even within the skin itself. Insulin-like growth factor 1 has been shown to play a major role in promoting and maintaining the expression of steroidogenic enzymes that are responsible for converting cholesterol into steroid precursors for the synthesis of DHEAS and androgens.²² Steroidogenic enzymes are expressed in human sebaceous glands where they may play a role in local androgen production.²³ Also, IGF-1 can induce 5α-reductase in human skin fibroblasts, leading to an increased conversion of testosterone to DHT.²⁴ In the present study, levels of DHEAS and androstenedione correlate with those of IGF-1 in men with clinical acne, which supports the hypothesis that IGF-1 may play a role in stimulating the activity of steroidogenic enzymes. In women, DHT correlates with acne lesions and with IGF-1. A high correlation exists between DHT and IGF-1 in women with clinical acne, suggesting that IGF-1 may influence DHT through stimulation of 5α-reductase (Table 4).

Despite these correlations, the data from our study demonstrate that IGF-1 and androgens can act independently or dependently on acne lesion counts, depending on the patient group or subgroup. For example, androgens exert independent effects of IGF-1 in the all- women and all-men groups. However, in men and women with clinical acne, IGF-1 appears to play a greater role. For example, in men with clinical acne, neither androgens nor IGF-1 exerted effects independent of one another. Moreover, in women with clinical acne, IGF-1 actually acted independently of androgens (Table 5).

The effects of hormones on the skin are complex and not well understood. Our study suggests that in adult women, serum levels of IGF-1 in addition to androgens may influence acne. Despite there being a correlation of acne with elevated IGF-1 and DHEAS, it would be difficult to use these hormones as laboratory markers of adult acne in the clinical setting because levels are usually within the normal range. Undoubtedly, these hormones have important effects in other organ systems in the adult in addition to effects in the skin, and therefore nonspecific inhibition of hormone synthesis could have undesirable effects. However, our data provide a rationale for the further study of androgen, IGF-1, and growth hormone receptor activity in the sebaceous gland. In the future, it may be possible to down-regulate or block receptor signaling in a tissue-specific manner to decrease the propensity for acne.

Correspondence: Diane Thiboutot, MD, Department of Dermatology, HU 14, The Pennsylvania State University College of Medicine, 500 University Dr, Hershey, PA 17033 (dthiboutot@psu.edu).

Accepted for Publication: June 7, 2004.

Funding/Support: This study was supported by National Institutes of Health (NIH) grant NIH K08 AR046766 (Dr Thiboutot) and NIH General Clinical Research Center grants M01 RR010732 and C06 RR016499 to The Pennsylvania State University College of Medicine.

Acknowledgment: We thank Jan Light, LPN, Jolene Sponhower, MA, and Chris Hamilton, BS, for assistance with patient recruitment and the technical aspects of the study.

Financial Disclosure: None.