Acanthosis nigricans: A dermatologic marker of metabolic disease

Introduction

Acanthosis nigricans (AN), a skin lesion characterized by hyperpigmentation and papillary hypertrophy, is associated with several systemic illnesses. It is classified into 4 types (Curth et al): (a) Malignant AN (Type I) - a cutaneous paraneoplastic syndrome associated with adenocarcinoma (b) Pseudo - AN (Type III) - associated with several syndromes in which obesity and endocrinopathies especially the insulin resistant state coexist (c) True benign AN (Type II) - familial, present at birth or beginning in childhood (d) Druginduced AN (Type IV).

It is suggested that most, if not all patients with AN have either clinical or subclinical. Insulin Resistance (IR).[2] The IR state is associated with several manifestations e.g. IGT/DM, hypertension, dyslipidemia, disorders of blood coagulation, accelerated atherogenesis, ovarian thecal hyperplasia and functional hyperandrogenism.[3] Measures to counter IR may delay or prevent these manifestations.

We therefore undertook this study to estimate the Insulin Sensitivity (IS) of a group of patients with clinically suspected and biopsy proven AN.

Materials and Methods

Study population: Thirty - six patients (28 females, 8 males) referred by the Dermatology services with biopsy proven AN were included in this study after an informed consent. Eight normal subjects (5 females, 3 males) comprised the control population.

The subjects were evaluated in the Endocrinology Clinic of the institution. All of them were sedentary individuals and none had taken any medication in the past. They were questioned in detail regarding the sequelae of IR viz. diabetes mellitus, hypertension, hyperandrogenism and major familial illnesses. A detailed menstrual history was elicited from females. The controls had no personal or family history of any illness.

All subjects underwent a detailed physical examination including anthropometry (height, weight, waist and hip circumference, mid - arm circumference - MAC, skin fold thickness - SFT measured over the biceps, triceps, subcapsular folds). From these data - Body mass index (BMI), waist/hip ratio (WHR) and lean body mass (LBM) were calculated.[4]

Skin biopsy

Four mm. punch biopsies of the neck or axillary lesion were performed under local anesthesia with precautions. The biopsy was fixed in formalin, sectioned and stained with H & E. The sections were reported by dermatologist trained in dermatopathology.[5]

Dynamic tests

Plasma glucose and serum insulin levels were estimated after a 12 hour overnight fast basally and at 60, 90 and 120 min. after a 75 gm. glucose load (OGTT). Insulin Tolerance Test (ITT) with measurement of Glucose Disposal Rate (GDR): After a 10 hr overnight fast, 0.1 IU/Kg. body weight of Purified Porcine Insulin was administered and blood was collected for glucose estimation at - 10, 0, 10, 20, 30, 40, 50 and 60 min. after the injection.[6] Patients were monitored at the beside for symptomatic/asymptomatic hypoglycemia using a glucometer.

Laboratory methods

Plasma glucose was estimated by Glucose Oxidase Method (kit supplied by M/S. Impero Diagnostics). Insulin levels were estimated by RIA (Double antibody technique) using kits supplied by the BARC. The Intra - assay and Inter - assay coefficient of variation were < 10%.

Statistical analysis

The OGTT were classified as normal impaired and overtly diabetic by criteria laid down by WHO.[7]

The integrated insulin response during the OGTT was computed by Area Under the Curve (AUC-I) calculated by the trapezoidal method.[8] It was compared to the insulin response of controls by applying Student′s unpaired ′t′ test. The GDR - (Slope Kmax) was measured from the fall of blood glucose from 10′ to 40′ by the ITT. These timings were selected because it takes 10′ for the stabilization of the insulin action, which lasts for 30.′6

Log BG at 40′ - log BG at 10′

Slope Kmax = 0.5 hr

Slope in patients with differing extents of glucose tolerance was compared to slope in control using Student′s ′t′ test. The indices of adiposity were correlated to the insulin resistance by applying the linear regression analysis - coefficient of correlation - ′r′ value.

Results

This unselected study population comprised of 28 females and 8 males. Referrals for AN were predominantly made in the prepubertal age group and in the second and third decades [Table - 1].

The mother of one child (aged < 10 years) had gestational diabetes mellitus and AN and was overtly diabetic at the time of evaluation.

Skin sites

The neck (100%), axilla (80.6%) and groin (61.1 %) were the predominantly affected sites. AN also occurred at the axillae (13. 8%), the inframammary regions (10.3%), the nasal bridges and periorally. AN was evident clinically and was histologically significant in this study population.

Body Mass Index:

[Table - 1]

Six (3 males and 3 females) of eight control subjects had normal BMI (22.0+/-0.9) and two were obese (26.5 +/- 0.9 Kg/m2). Eleven patients (4 males and 7 females) with AN had normal BMI (21 - 27). One subject had a BMI < 20 (generalized lipoatrophy). Most (69.4%) of the patients were ′morbidly′ obese using WHO criteria.

Waist/Hip Ratio:

[Table - 1]

All control subjects had a W/H ratio < 1.0 except two obese subjects. 62.5% males with AN had W/H ratio upto 1.0 and 37.5% had ratio > 1.15; whereas 39.3% of the females had a W/H ratio < 0.85. 28.6% had a ratio 0.86 - 1.0 and 32.1 % had a ratio > 1.1. There was a significant positive correlation between BMI and W/H ratio in both, controls (r =0. 82; p< 0.01) and AN patients (r=0.57; p< 0.01).

Skin Fold Thickness (mm):

[Table - 2] a, b

SFT did not differ between AN patients and controls, though lean subjects had a lower SFT than the obese. Obese - AN patients had significantly high SFT compared to their non-obese counterparts.

Mid -Arm Circumference (mm):

[Table - 2] a, b

MAC was significantly higher in patients with AN (30.5 +/- 0.88 mm) as compared to controls (23.9+/- 0.76 mm), and in the obese compared to the non- obese population.

Lean Body Mass:

[Table - 2] a, b

LBM was significantly high in the AN compared to controls and in obese subjects with AN compared to non - obese.

Oral Glucose Tolerance Test:

All the controls and 20 with AN patients had normal GTT, The other patients had either IGT (8) or overt diabetes (8). However, all AN patients had higher glycemic responses (AUC - G) compared to controls (p< 0.0005). All AN patients irrespective of the glycemic responses, had highly significant hyperinsulinemia (AUC- I) compared to controls (p< 0.0005). It is Interesting to observe that the insulin response to glucose load in overtly diabetic AN patients was lower compared to normal and those with impaired GTT. The fasting, peak and 2 hr. glucose and insulin levels were significantly higher in AN patients compared to controls [Table - 3]a. No significant difference was observed between the glycemic (AUC - G) and insulinemic (AUC-I) responses of obese and non - obese AN patients [Table - 3] b.

Insulin Mediated Glucose Disposal Rate (GDR/ Slope K):

[Table - 3] a b

Slope K was significantly reduced in AN patients (-0.66+/- 0.07) compared to controls (-0.39+/- 0.08; p< 0.01) indicating a significant IR state in AN.

In obese AN patients, Slope K was reduced compared to obese controls (-0.21 +/0.14 in C & -0.71 +/-0.11 in AN; ns), whereas in non - obese, it was comparable in controls and AN patients (-0.46+/-0.08 in C & - 0. 59 +/0.13 in AN).

When the whole study population was considered, there was no correlation between adiposity and Slope K except with SIFT In these subjects insulin sensitivity was significantly reduced with high SFT (r= 0.5; t = 2.02; p < 0.05). However, slope K correlated positively with BMI (r= -0.71; t= 2.01;p< 0.01) and W/H ratio (r= -0.95; t= 5.1;p< 0.01) in subjects with IGT in our study.

Discussion

This study was carried out on an unselected population of patients with AN, who sought evaluation with the dermatologist for possibly cosmetic reasons and were diagnosed on biopsy to have AN. There were more females than males (28 v/s 8; 77.8% v/s 22.2%) in this population at an average of 26.3+/- 1.7 years. The overall prevalence of AN was higher in females possibly because it can be disfiguring and upsetting for the patients. The female preponderance in this study could be because obesity is another determinant of IR. In almost all populations, more women are obese and overweight than men.[9] Our population, was from the middle to the lower income group and yet, had subjects with morbid obesity. Even in the U.S.A., the prevalence of obesity in the overweight category was higher among women living below the poverty line than above it.[10]

Stuart et al[11],[12] have reported the prevalence of AN to range between 19-38% among native Americans. AN was present in 7.1 % of school children in Galveston, Texas. The incidence was 0.5% in whites, 5.5% in Hispanics and 13.3% in back children.

This population was selected at a clinic located in Western India and hence was divided between Maharashtrians (55.5%), Muslims (27.8%), Gujratis (8.3), Sindhis (5.6%) and Christians (2.8%). In studies in Western population, ethnic difference has been described in subjects with AN. [14] To the best of our knowledge, no population studies are available for Indians.

AN predominantly was seen on the neck, axillae, groins and perineum. No patient in this population manifested either generalized hyperpigmentation or mucosal involvement. It commonly involves the flexural areas, particularly the axillae, inguinal region, inframammary region and neck. In its most developed form, AN could involve almost the entire body, especially when it is a manifestation of internal malignancy especially the GI tract.[15]

AN could occur - (a) as genito - developmental or nevoid (b) due to friction and maceration as in obesity (c) epidermotrophic as in acromegaly and these factors could affect the area of distribution.

It may be difficult to diagnose clinically, except in its most evolved form.[16] In a histopathologic evaluation of women who manifested hyperandrogenism and PCOS, AN was demonstrable on biopsy although absent clinically.[17]

Patients with clinically evident AN, were almost always obese. Significantly 30.6% of our subjects were non - obese individuals. Again, the W/H ratio was significantly android in 60% obese and 50% of normal BMI individuals, especially in the female population. This implies that the distribution of body fat was abnormal in spite of a normal BMI. Subjects with AN had a higher LBM as well. The LBM has been shown to increase in parallel with fat mass in a predictable way in obese people.[18] This is possibly because of the anabolic effects of insulin.

Glucose tolerances ranged from normal (55.6%) to impaired (22.2%) to overtly diabetic (22.2%). All three groups were hyperinsulinemic as compared to controls. Overt diabetics possibly had exhaustion of the pancreatic B-cell secretion as evidenced by the lower insulin responses compared to those with normal or impaired glucose tolerance.[19] The glucose disposal rates were significantly lower in AN (obese or nonobese) as compared to controls, confirming that AN is associated with a significant IR. Actually, today It is being recognized that IR, whatever its cause, stimulates insulin secretion and that the increased insulin levels stimulate IGF-1 receptors on various tissues.In the skin, stimulation of the IGF - 1 receptors on keratinocytes leads to excessive epidermal growth. AN therefore today is considered as an epiphenomenon of the IR state.[20] However, when we correlated the insulin resistance indices (AUC - I and Slope K) to indices of adiposity, we found correlation only with skin fold thickness. Perhaps the small sample size is responsible for this discrepancy. However in subjects with impaired - GT, there was a positive correlation with BMI and W/H ratio as well.

Dunaif et al also showed that the severity of the histopathologically diagnosed AN correlated positively only with magnitude of peripheral insulin resistance (i.e. decreased insulin - mediated glucose disposal) rather than basal or glucose challenge - induced hyperinsulinemia.[17] It is also pertinent to note that all patients with congenital syndrome of severe IR have AN, but acquired forms of IR are less commonly associated with it.

AS is evident from our study, AN is a marker for insulin resistance. Obese acanthotics may be more severely IR than their nonacanthotic counterparts. Also, it is a rare condition, not every obese individual manifests AN. Acanthosis occurs in patients with hyperinsulinemia owing to stimulation of the insulin like growth factor - I and epidermal growth factor, which cause proliferation of the epidermal cells.

Life style and behavioural modification have proven baneficial in the prevention of several disorders associated with insulin resistance. Perhaps individuals with AN are another cohort that must be targeted for lifestyle modification, so that the serious consequences of IR may be avoided.

Acknowledgement

The authors express their gratitude to Dr. N.A. Kshirsagar, Dean for permitting us to publish this article. This study was carried out under the aegis of the′ Nair Golden Jubilee Research Fund′. Gratitude is also due to Mr. Venkat (Skin Dept) and residents of the Endocrinology Department.