Sex Hormones, Metabolic Status, and Obesity in Female Patients with Acne Vulgaris Along with Clinical Correlation: An Observational Cross-Sectional Study.

BACKGROUND: Acne vulgaris is a common dermatological disorder. Several hormones are suspected to play a role in its etiopathogenesis. AIMS AND
OBJECTIVES: The aim of this study was to analyze the role of sex-hormones, metabolic status, and obesity in acne vulgaris and correlate with its severity and symptom load.
MATERIALS AND METHODS: This cross-sectional observational study included 89 female patients with acne vulgaris and certain phenotypic markers such as prepubertal onset, late-onset, persistent course, hirsutism, acanthosis nigricans, acrochordons, premenstrual flare, and diminished response to isotretinoin; suggestive of an underlying hormonal pathology. All patients were subjected to physical examination to rule out obesity and metabolic syndrome along with serum biochemistry to detect sex hormones (testosterone, progesterone, estrogen), serum insulin and insulin resistance (HOMA-IR) and lipid profile.
RESULTS: Among 89 patients (mean age 21.3 ± 5.3 years), 34.8% presented with late-onset/persistent/pre-pubertal acne, 33.7% presented with premenstrual flare and 28.2% presented with hirsutism. Hormonal analysis revealed elevated testosterone and progesterone with low estrogen across all categories of patients. Testosterone was significantly elevated even in mild acne. Serum lipid profile was altered significantly only in hirsute females. In total, 36% and 20.2% patients presented with metabolic syndrome and obesity, respectively; however, neither was associated with severity of acne.
CONCLUSION: Sex-hormones, serum lipids, metabolic status, and body mass index are altered in acne vulgaris. All acne patients with endocrine markers should be evaluated for sex-hormones irrespective of severity and symptom load, whereas hirsutism may be regarded as clinical marker of lipid abnormalities. Metabolic syndrome and obesity do not seem to be directly correlated with acne severity. Thus, anti-androgens may be considered as adjuvant therapy in these patients, not responding to conventional therapy. Copyright:

Introduction

Acne vulgaris is one of the most common dermatological disorders globally, with a prevalence of about 22% in the Indian subcontinent.[1] It has a predilection for adolescent females, but adult variant is being increasingly reported lately. Adult acne (acne tarda) can either develop after 35 years age (late-onset acne) or continue beyond 25 years of age (persistent acne). Late-onset acne can be further classified into chin acne (distributed periorally or along the mandible) and sporadic acne.

The etiopathogenesis involves inflammation of pilosebaceous unit characterized by hyperseborrhea, abnormal follicular keratinization and colonization of pilosebaceous duct with propionibacterium acne.[2] Several contributory factors have been highlighted such as genetic predisposition, endocrine factors, environmental factors, stress and food habits;[3] the most important being endocrine factors characterized by the involvement of sex hormones especially androgens.[4]

Skin is being increasingly recognized as an endocrine organ, with sebaceous gland being the principal site for synthesis of diverse hormones and expression of associated receptors, especially androgens.[5] Sebocytes are the main site for synthesis and metabolism of sex-hormones from cholesterol to dihydrotestosterone (DHT) and estradiol by 5-α-reductase and aromatase, respectively.[4] Androgens and their precursor substrates increase the size of sebaceous glands, stimulate sebum production and enhance ductal keratinocyte differentiation and plugging, thus resulting in acne.[6] Several studies are exploring the etiologic role of androgens in acne, a classic example of androgen-mediated skin disease; however, data are lacking in the current setup. In addition, the relationship between severity of acne and circulating androgens is yet to be established.

Role of progesterone in the pathogenesis of acne is controversial. It is a competitive inhibitor of 5α reductase and might play some role to reduce sebaceous gland activity but in humans its sebum suppression effect is minimal.[7] However, according to Kanda et al.[8] progesterone aggravates acne by stimulating sebum secretion and keratinocyte proliferation. Therefore, there is lack of consensus regarding the relationship between acne vulgaris and serum progesterone levels in women.

Estrogen is hypothesized to exert a protective role in acne vulgaris, but scarcity of literature regarding plasma level of estrogen in these patients suggest a fertile area to be explored.[9]

Several authors have highlighted the association between acne and metabolic syndrome, especially in adolescents under the influence of a Western diet. It has been postulated that such a diet overstimulates one of the key conductors of metabolism, the nutrient- and growth factor-sensitive kinase mTORC1 in lesional skin and sebocytes. Increased activity of the later may lead to increased incidence of obesity, metabolic syndrome, type 2 diabetes mellitus and neurodegenerative disorders. Thus, acne may be considered a mTOR-mediated dermatoses and regarded as MetS of the pilosebaceous follicle.[10]

However, several authors have refuted this association in young males.[11] Thus, there is lack of consensus regarding the association of MetS and acne vulgaris, especially in Indian females.

Studies from several regions have proposed a positive correlation between acne and obesity in adolescent males with inconsistent data in females.[12] Obesity may result in reduced plasma level of sex hormone-binding globulin (SHBG), thus increasing the free testosterone level and inflammatory cytokines leading to acne. Hyperinsulinemia and insulin resistance may be additional factors. Thus, there is need of further studies to establish this relationship, especially in females.

We have explored the role of sex hormones, metabolic status and obesity in females with acne vulgaris and correlated with its clinical severity and symptom load as data is scarce in the present setup. Other relevant clinical characteristics such as hirsutism and menstrual disturbance have also been assessed which may indicate a hormonal pathology. We have also reviewed current literature to determine the prevalence of these associations. The authors hope these findings may improve treatment options to obtain better patient outcomes for this common disorder.

Materials and Methods

We conducted a cross-sectional, observational study involving 89 female patients of reproductive age presenting with acne vulgaris along with certain clinical features suggestive of an underlying hormonal pathology (e.g., hirsutism, acanthosis nigricans, acrocordons/skin tag and pre-menstrual flare) [Figures 1 and 2]. The study was carried out at the departments of Dermatology and Biochemistry of a tertiary care center in eastern India between January 2019 and June 2019 after obtaining approval from the Institutional ethical committee.

Figure 1

Acne vulgaris with hirsutism

Figure 2

Acne vulgaris with hirsutism and mandibular distribution, usually associated with premenstrual flare

The inclusion criteria included patients with acne (irrespective of severity) having one or more of the following characteristics: not responding to conventional therapy, especially those who failed to respond to isotretnoin therapy or relapsed shortly following discontinuation of isotretinoin or oral contraceptives (OC); late onset (adult acne, after 35 years of age), or sudden-onset, or severe, unresponsive or persistent acne (continues after 25 years of age) and prepubertal acne. Patients with history of premenstrual flare and/or showing clinical signs of hyperandrogenism and metabolic syndrome such as hirsutism, mid-truncal obesity, weight gain, acanthosis nigricans, acrochordons (skin tag), patterned hair loss, menstrual disturbances and elevated blood pressure were also included.

Patients aged less than 10 years and more than 45 years were excluded from the study. Other exclusion criteria included pregnancy and lactation, smoking, concomitant use of OC pills or any other hormonal/anti-androgen medication, for example, spironolactone or any drug affecting lipid metabolism, any systemic illness such as hypothyroidism or lipid metabolic disorder and/or infertility treatment.

All the study patients were recruited after fulfilling the inclusion and exclusion criteria and receiving their informed consent (or guardian) in accordance with the Declaration of Helsinki Principles of medical research.

Detailed history of each patient was taken with special emphasis on demographic profile, duration of symptoms, menstrual history and drug history. A single dermatologist clinically examined the skin of all patients to detect acne lesions, hirsutism, acanthosis nigricans and acrocordons/skin tag. Grading of acne vulgaris was done according to global acne grading system [Table 1].[13]

Table 1

The global acne grading system[22]

Location	Factor
Location	Score
Forehead	2
Left cheek	2
Chin	1
Right cheek	2
Nose	1
Chest and upper back	3

Interpretation: Each type of lesion is given a value depending on severity: no lesions = 0, comedones = 1, papules = 2, pustules = 3 and nodules = 4. The score for each area (local score) is calculated using the formula: Local score=Factor × grade (0-4). The global score is the sum of local scores and acne severity was graded using the global score. A score of 1-18 is considered mild, 19–30, moderate; 31-38, severe; and >39, very severe

Anthropometric measurements

Anthropometric measurements of the participants, such as height (cm) and weight (kg) were recorded using a stadiometer (SECA 213 Stadiometer) and a digital weighing machine (SECA 874 U digital scale). Blood pressure (mm Hg) was measured at the left brachial artery using a standard syphgmomanometer cuff. All evaluations were conducted in accordance with the standard guidelines, three times in each subject and mean values were obtained. Body mass index (BMI) was calculated as weight in kilograms divided by height in meter squared (kg/m2). The modified classification of BMI for Asian Indian populations was used in this study to define overweight (23–24.99 kg/m2) and obesity (>25 kg/m2).[14]

Biochemical analysis

Venous blood was collected through venepuncture of left anterior cubital vein. Twelve-hour fasting blood was taken for estimation of glucose, triglyceride (TG), total cholesterol (TC), high-density lipoprotein-cholesterol (HDL-C), and low-density lipoprotein-cholesterol (LDL-C). Concentrations were determined with kits by Erba Mannheim XL System Pack, EM 360 Transasia Autoanalyzer (Germany), in the Department of Biochemistry, which is included in Grade “A” category by Biorad Quality Assurance System (USA).

Hormonal parameters including luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol (E2), total testosterone (TT), and progesterone concentrations were estimated with kits by elecsys cobas e analyzer, Cobas e 411 autoanalyzer (USA). All samples were taken during the follicular phase of the menstrual cycle.

Reference ranges of the measured parameters are tabulated in Table 2.

Table 2

Reference ranges of measured hormones (luteal phase) and lipids

Hormone (follicular phase)	Reference interval
LH (mIu/mL)	2.4-12.5
FSH (mIu/mL)	3.5-12.5
TT (ng/mL)	0.2-0.8
E2 (pg/mL)	30-100
Prog (ng/mL)	0.15-0.7
Lipid (mg/dL)
T-CHO	<200
TG	<150
HDL-C	40-59
LDL-C	<100

LH=Leutinising hormone, FSH=Follicle-stimulating hormone, TT=Total testosterone, E2=Estradiol, Prog=Progesterone, HDL-C=High-density lipoprotein cholesterol, LDL-C=Low density lipoprotein cholesterol, E2=Estradiol, Prog=Progesterone, TC=Total Cholesterol, TG=Triglycerides

Statistical analysis

Data were entered in a Microsoft Excel spreadsheet. All the entries were double entered checking consistency. All the data obtained have been statistically analyzed using MedCalc® v12.5.0 and preserved for future reference. Descriptive statistics as mean and standard deviation was used. Normal distribution of numerical variables was determined using the Shapiro–Wilk test. Chi-square test was used for nonparametric data, whereas ANOVA and Kruskel–Walis test were applied for parametric data. A value of P < 0.05 was been taken as significant.

Results and analysis

In total, 89 female patients completed the study, their age ranging from 13 to 38 years, with mean being 21.3 ± 5.3 years. Almost 75% (n = 67) belonged to 15–25 years age group, whereas six patients (6.7%) presented with adult acne (acne onset >35 years). The mean duration of symptoms was 24.05 ± 31.9 months, almost 50% reported within 6 months of onset of acne. Approximately one-third of the patients presented with hirsutism (28.2%), acanthosis nigricans/acrocordons (31.5%) and metabolic syndrome (35.9%), clinical indicators of hormonal acne. However, menstrual disturbance was noted in only 9 (10.1%) patients. Most of the patients presented with mild acne (82% [n = 73]), followed by moderate and severe acne in 14.6% and 3.4% cases, respectively (GAGS scoring). The clinico-demographic details are tabulated in Table 3.

Table 3

Clinico-demographic profile of patients

Clinico-demographic characteristic	Total number of cases (%) [n=89]
Age group
<15 years	7 (7.9%)
16-24 years	67 (75.3%)
25-34 years	9 (10.1%)
≥35 years	6 (6.7%)
Duration of acne vulgaris
1-6 months	44 (49.4%)
7-24 months	17 (19.1%)
25-59 months	11 (12.4%)
≥60 months	17 (19.1%)
Hirsutism	25 (28.2%)
Menstrual disturbances (current or previous)	9 (10.1%)
Metabolic syndrome	32 (35.9%)
Acanthosis nigricans/acrochordon	28 (31.5%)
Poor response to Isotretinoin	3 (3.4%)

All the study subjects (89 female patients) were grouped into four different categories. Category1 ( = 31): Who developed late onset (adult acne, after 35 years of age), or pre-pubertal or sudden-onset, or severe unresponsive and persistent acne. Category 2 ( = 30): Who had all the characteristic of Category 1, but in addition had H/O premenstrual flare. Category 3 ( = 25): Who had all the characteristics of Category 2 but in addition developed hirsutism. Category 4 ( = 3): Who had all the characteristics of Category 3 but in addition showed poor response to isotretinoin therapy.

Hormonal analyses revealed elevated serum total testosterone and progesterone level in all categories compared to the reference interval (TT = 0.2–0.8 ng/mL), (PRO = 0.15–0.7 ng/mL); however, neither elevation was statistically significant with respect to severity of acne (P > 0.05, Chi-square test for trend). There was no significant difference across different categories of our patients [Table 4]. Serum testosterone was elevated in almost 50% of patients with mean being 0.89 ± 0.52 ng/mL, whereas serum progesterone was elevated in 61 (68.5%) subjects [mean: 3.08 ± 4.73 pg/mL]. Interestingly, inter group analysis showed a statistically significant association between elevated testosterone and mild acne (P = 0.02, Chi-square test); thus, the authors recommend analyzing serum testosterone level in all female acne patients with additional phenotypic manifestations such as premenstrual flare, hirsutism, adult-onset acne and persistent acne irrespective of its severity.

Table 4

Distribution of serum levels of hormones according to variety of symptoms among the acne patients (n=89)

Category	1 (n=31)	2 (n=30)	3 (n=25)	4 (n=3)	K-W test (P)
LH (mIU/mL) mean±sd	8.69±4.92	7.16±6.11	11.04±8.23	10.58±9.82	4.264 (0.119)
FSH (mIU/mL) mean±sd	3.63±1.12	3.65±2.17	4.12±1.5	4.00±1.69	5.263 (0.072)
Estradiol (pg/mL) mean±sd	29.41±15.48	30.95±21.07	40.23±23.12	42.52±27.51	0.086 (0.958)
Progesterone (ng/mL) mean±sd	2.20±2.15	2.83±3.83	5.89±4.32	5.59±9.14	2.63 (0.268)
Total testosterone (ng/mL) mean±sd	0.96±0.45	0.84±0.57	0.99±0.32	0.96±0.40	2.441 (0.295)

Serum estradiol concentration was low in 65 (73%) subjects and normal in the rest (EST = 30–100 pg/mL). Serum LH, FSH ratio were altered >2:1 across all categories.

Serum biochemistry revealed elevated LDL cholesterol, elevated fasting blood glucose (FBS) and elevated total cholesterol (CHO) in 21.3%, 12.4%, and 2.2% patients, respectively (normal ref range) [Table 2]. However, none had a statistically significant association with the severity of acne (P > 0.05, Chi-square test for trend). All the parameters were raised in third category (maximum symptom load including hirsutism), whereas in the other categories values were within normal range. Interestingly, inter-group analyses showed significant differences in serum CHO (F = 3.43, P < 0.05), TG (F = 3.93, P < 0.012) and LDL-C (F = 4.17, P < 0.009) among different categories of acne (ANOVA test) [Table 5].

Table 5

Distribution of serum lipid among the acne patients of different categories (n=89)

Category	1 (n=31)	2 (n=30)	3 (n=25)	4 (n=3)	ANOVA
TG (mg/dL) Mean±sd	88.86±28.57	103.21±48.02	233±45.78	114.31±50.48	F=3.931, P=0.012
CHO (mg/dL) Mean±sd	142.76±22.13	147.88±20.57	211±28.29	145.75±20.46	F=3.44, P=0.021
HDL (mg/dL) Mean±sd	50.24±10.94	48.60±10.78	141±15.23	47.06±6.51	F=0.307, P=0.82
LDL Mean±sd	85.48±16.38	89.76±15.02	141±20.12	91.06±15.89	F=4.175, P=0.009
BMI Mean±sd	22.19±3.24	22.81±3.39	33.63±3.27	22.73±4.58	F=3.16, P=0.029

Thirty-two patients (36%) presented with features of metabolic syndrome (MetS) according to the NCEP-ATP III criteria,[15] not significantly associated with the severity of acne or symptom load (category of patient) [P > 0.05, Chi-square test for trend]. However, interestingly presence of metabolic syndrome was statistically significant in subjects with mild acne (P = 0.04, Chi-square test) [Table 6]. Thus, all patients should be evaluated for metabolic syndrome irrespective of severity in the presence of additional clinical indicators such as hirsutism, premenstrual flare, adult or pre-pubertal acne. Our patients showed a normal mean serum insulin (0.78 ± 1.66 μIU/mL), whereas none of them had insulin resistance (mean HOMA-IR 0.17 ± 0.36).

Table 6

Metabolic syndrome (MetS) and its relation with the severity of acne (n=89)

Acne severity (GAGS score)	Metabolic syndrome n (%)			Level of association (P)
	Absent	Present	Total
Mild	48 (53.9)	25 (28.1)	73 (82)	0.04
Moderate	8 (8.9)	5 (5.6)	13 (14.6)	0.75
Severe	1 (1.1)	2 (2.2)	3 (3.4)	0.48
Very severe	0 (0)	0 (0)	0 (0)	0
Total	57 (64.04)	32 (35.9)	89 (100)	0.63

Obesity was recorded in 20.2% (n = 18) of our patients, whereas 20 patients (22.5%) were overweight according to the revised consensus BMI indices for Asian Indians.[14] However, its association with severity of acne was not statistically significant (P = 0.3, Chi-square test for trend). However, mean body mass index (BMI) showed a statistically significant difference across the categories of our patients (F = 3.16, P = 0.03, ANOVA), with hirsute patients (category 3) showing the maximum mean value (33.63 ± 3.27 kg/m2).

Discussion

In our study, the mean age was about 21 years, whereas almost 75% belonged to 15–25 year age group and six patients (6.7%) presented with adult-onset acne (onset >35 years). These data are consistent with that reported by Iftikhar et al.[1] and Akdogan et al.[16] This age group (15–25 years) coincides with maximum sebaceous activity under androgenic influence. The mean duration of our patients was similar to that of other studies.[116] In our study hirsutism was observed in 28.2% patients, much lower than that reported by Ambalal et al.[17] in 80.5% patients. This difference may have occurred due to a much smaller sample size in the later study (33 patients). Majority of our patients (90%) had regular menstruation in accordance with most other studies.[31617]

Most of our patients presented with mild acne [73 (82%)] with the median GAGS score being 14.95 ± 4.23. Maximum prevalence of mild acne has also been reported by other similar studies.[11819] This may be attributed to the selection of cases presenting for the first time.

In our study, we categorized the patients into four groups based on their symptom load, all potentially related to androgen excess. 33.7% of our patients presented with premenstrual flare, much lower than George et al.[3] (78.7%), probably because the later study was conducted exclusively on a cohort of adult-acne patients. Chin acne is usually associated with premenstrual flare,[17] our study also supported this observation.

Serum testosterone was elevated in almost half of our patients (49.4%), across all categories This finding is similar to that of most other authors.[12021] However, our finding is lower when compared with a similar study from Egypt (70%)[6] and much higher than a study from Bangladesh (10%)[22] and India (13.8%).[17] This variation may be explained by the heterogeneous patient profiles in these studies. In our study, elevated serum testosterone was not associated with age of patient, menstrual history or severity of acne. Rather, females with mild acne showed a significant elevation of testosterone. Clinically, all our patients presented with symptoms of androgen excess such as hirsutism, acanthosis nigricans and acrochordons and/or atypical acne (late-onset or persistent or pre-pubertal or isotretinoin-resistant or having pre-menstrual flare). Thus, serum testosterone estimation is recommended in all acne patients presenting with such symptoms irrespective of age, menstrual history or severity of lesions, as it is a cost-effective laboratory test to detect androgenic excess secondary to occult ovarian or adrenal dysfunction and/or defective receptor metabolism. In such patients, anti-androgen treatment should be considered along with treatment of primary pathology for better therapeutic response.

The role of estrogen in AV is debatable; however, elevated estrogen level is hypothesized to be protective by reducing endogenous androgen secretion resulting in sebo-suppression, the cornerstone of acne pathogenesis.[23] We detected low serum estrogen (estradiol) in 73% patients, thus favoring acne formation. Our observation is consistent with that of Akdogan et al.[16] and Bakry et al.[6] However, the role of estrogen rich foods, drugs containing high amounts of estrogen such as pills, skin patches or topical agents for treatment of acne remain inconclusive and questionable.

Serum progesterone level was elevated in 68.5% of our patients, irrespective of clinical categories. Several other authors have also documented increased serum progesterone level in majority of their patients.[2425] Our data regarding increased progesterone also corroborate the possible androgenic effect of progesterone in stimulating seborrhoea and keratinocyte proliferation.

The relationship between common blood lipids (TC, TG, LDL-C, HDL-C) and acne vulgaris remains controversial till date. We compared lipids among our patients and correlated it with their symptom load by categorizing the patients. In all categories except third category (maximum symptom load) we observed normal serum levels of TC, TG, LDL-C, and HDL-C, whereas in the 3rd category all components of lipid profile exceeded the normal reference range. This inter-group difference was statistically significant [Table 5]. However, there was no significant association with the severity of acne. Our study is well supported by Akawi et al.[26] and Cunha et al.[27] Thus, the authors recommend serum lipid assessment once females develop hirsutism/other features of virilization, irrespective of the severity of acne lesions.

Metabolic syndrome was detected in 36% of our patients, whereas Nagpal et al.[11] recorded it in 17% patients. None of our patients showed Insulin resistance but Nagpal et al.[11] recorded it in 22% subjects. This difference may be explained by the sole inclusion of male subjects in the later study. None of the studies including the current one showed any significant association with the severity of lesions. Thus, we feel every acne patient should be subjected to screening for metabolic syndrome irrespective of its severity or symptom load to rule out any occult systemic disorder such as diabetes or cardiovascular pathology. Female patients may be less prone to develop insulin resistance than their male counterparts; however, this observation needs to be validated by further large scale studies.

Eighteen (20.2%) patients were declared obese (BMI ≥25 kg/m2) in this study, whereas Akdogan et al.[16] reported it in 14.7% patients in Turkey. The slight difference may be attributed to differences in ethnicity and food habits between these two countries. Tsai et al.[28] proposed higher BMI to be a significant risk factor for acne in children. Maximum BMI was noted in patients presenting with hirsutism; thus, the latter may be considered a clinical marker of obesity. Secretion of adipokines in obese individuals might be the possible link, which reduce SHBG levels, thus increasing free testosterone level and worsening acne and hirsutism.[4] However, no link could be established between BMI and acne in this study.

Limitations

The major limitations of our study are a cross-sectional design with limited sample size and exclusive inclusion of acne vulgaris patients with visible signs of hormonal pathology such as premenstrual flare, adult-onset acne, and hirsutism, who are more prone to develop endocrine (sex-hormone) abnormalities, thus resulting in a selection bias. So, the authors recommend case-control studies with a larger sample size in the future to corroborate these findings.

Conclusion

Acne vulgaris is one of the most common adolescent dermatological disorders. Although several treatment modalities exist, some patients show inadequate response to conventional therapy. Additionally, these patients may present with characteristic phenotypic manifestations such as hirsutism, acanthosis nigricans and acrochordons or have a skewed age distribution ranging from pre-puberty to adulthood (>35 years), suggestive of hormonal involvement. We conducted this study on a cohort of such patients to detect several hormonal abnormalities ranging from raised androgen and progesterone to sub-normal estrogen levels, and also correlated them with the symptom load and severity of the disease. Recently, acne is being recognized as a systemic disorder rather than a transient pubertal condition. We also detected metabolic syndrome and obesity in several of our patients to support this view.

Thus, the authors recommend hormonal analysis of patients presenting with atypical symptoms or in a different age group, irrespective of severity, to detect any hormonal abnormality and initiate targeted anti-hormonal therapy to provide a favorable treatment outcome. Thus, hormonal therapy may be considered as an adjuvant in patients not responding to conventional therapy. Also, screening for metabolic syndrome needs to be done, as skin reflects the internal health, thus enabling the dermatologist to provide lifestyle modifications to their acne patients at an early age to prevent or delay the onset of serious systemic disorders.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.