Wu W, Robinson-Bostom L, Kokkotou E, Jung H, Kroumpouzos G. Dyslipidemia in Granuloma AnnulareA Case-Control Study. Arch Dermatol. 2012;148(10):1131–1136. doi:10.1001/archdermatol.2012.1381
Author Affiliations: Departments of Dermatology (Drs Robinson-Bostom and Kroumpouzos) and Health Services, Policy and Practice (Ms Jung), The Warren Alpert Medical School of Brown University, Providence, Rhode Island; Division of Gastroenterology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts (Dr Kokkotou); and South Shore Medical Center, Atrius Health, Norwell, Massachusetts (Dr Kroumpouzos). Mr Wu is a medical student at The Warren Alpert Medical School of Brown University.
Objectives To estimate the prevalence of dyslipidemia (DLP) among individuals with idiopathic granuloma annulare and to examine associations with disease characteristics and comorbidities, such as metabolic syndrome.
Design Case-control study using review of medical records from January 2002 through December 2010.
Setting A university hospital and a health care system.
Participants Adults consisting of 140 patients and 420 individuals serving as controls matched for age, sex, race/ethnicity, hypertension, type 2 diabetes mellitus, and hypothyroidism.
Main Outcome Measures Prevalence of DLP and its associated components (hypercholesterolemia, hypertriglyceridemia, elevated low-density lipoprotein cholesterol, and low high-density lipoprotein cholesterol) in idiopathic granuloma annulare, as well as odds ratio of DLP after adjusting for confounding variables.
Results The prevalence of DLP was 79.3% among granuloma annulare cases and 51.9% among controls (P < .001). Granuloma annulare was associated with DLP (odds ratio, 4.04; 95% CI, 2.53-6.46) after adjusting for confounding variables. Statistically significant differences between patient and control groups were found for levels of total cholesterol, triglycerides, and low-density lipoprotein cholesterol (P < .001 for all variables). A statistically significant association between the extent of disease and DLP (P = .02) was shown. Annular lesion morphology was associated with hypercholesterolemia (P = .02) and DLP (P = .01).
Conclusions This study indicates an association between granuloma annulare and DLP. Dyslipidemia is more common in generalized than in localized/disseminated disease, and the annular lesion morphology is associated with hypercholesterolemia and DLP. Physicians should be aware of these important associations and consider them in the management of granuloma annulare.
Granuloma annulare (GA) is a benign, usually self-limited, granulomatous skin disease of unknown etiopathogenesis. The characteristic eruption manifests as flesh-colored to erythematous or purple papules with a tendency toward annular grouping and a potential to involve any area of the skin.1- 3 The generalized form of the disease shows a more chronic, relapsing course; a later age of onset; rare spontaneous resolution; and poorer response to therapy compared with the localized form.1 Many precipitating factors have been reported, and various hypotheses as to the etiopathogenesis of GA have been put forward1,2; however, little is known about the initial triggers and associated comorbidities. Associations with systemic disease have been described but not consistently corroborated.3- 6 An association with diabetes mellitus1,6 has been debated.7 The prevalence of other comorbidities, such as hypertension, obesity, and thyroid, cardiovascular, and malignant disease, has not been systematically investigated. Data from uncontrolled studies1,3 and case series8 have not been validated by large-scale controlled studies.
Co-occurrence of GA with necrobiosis lipoidica diabeticorum has been reported,9- 11 especially in patients with diabetes mellitus, and these diseases share certain histopathologic features, including lipid deposition.10,12 Similarly, several cases of eruptive xanthoma associated with hypertriglyceridemia showed GA-like histopathologic characteristics.13,14 Interestingly, the T-cell repertoire (CD4+ T cells, helper T cell subtype 1 profile)/cytokine pattern (tumor necrosis factor[TNF], interferon γ, interleukin 2 [IL-2], and IL-4) in GA15,16 is similar to that of skin diseases that have been associated with dyslipidemia (DLP), such as lichen planus and psoriasis17,18; psoriasis also has been associated with metabolic syndrome.18 Sarcoidosis is another granulomatous process that shows a similar cytokine pattern19,20 and has been associated with low levels of high-density lipoprotein cholesterol (HDL-C) and hypertriglyceridemia.21 Chronic inflammation in these diseases could explain the link with DLP.22 Interestingly, an uncontrolled study of 100 patients with generalized GA by Dabski and Winkelmann3 reported percentages of hypercholesterolemia (19.6%) and hypertriglyceridemia (23.3%) in patients with documented lipid values. The study showed a higher prevalence of elevated serum lipid levels in the annular rather than nonannular GA subgroups. However, the study did not include the reference range for the lipid values and measurements of low-density lipoprotein cholesterol (LDL-C) and HDL-C levels, which would have allowed for a determination of DLP prevalence as well as comparison with subsequent studies. To our knowledge, no other studies have investigated the association between GA and components of DLP. In the past 15 years, there have been no studies on comorbidities in GA or clinical studies relevant to the etiology of the disease.
This lack of information prompted us to investigate the prevalence of DLP, metabolic syndrome, and other comorbidities (eg, hypertension, type 2 diabetes mellitus, and hypothyroidism) in patients with idiopathic GA. We also investigated whether an association exists between clinicopathologic features of GA (extent, morphologic characteristics, and histopathologic type) and DLP or other comorbidities.
This case-control study, consisting of a medical record review (from 2002 to 2010), was conducted at Rhode Island Hospital in Providence and South Shore Medical Center, Atrius Healthcare System, in Norwell, Massachusetts. Inclusion criteria were age older than 18 years, histopathologic confirmation of diagnosis, and adequate clinical data on the skin disease, comorbidities, lipid levels, and medications. Data on age, sex, race/ethnicity, socioeconomic status, age at onset of GA, location and clinical features of skin lesions, histopathologic diagnosis, lipid values, comorbidities (type 2 diabetes mellitus, hypertension, hypothyroidism, obesity, and metabolic syndrome), smoking status, and use of lipid-lowering medications and/or β-blockers were collected. Socioeconomic status was approximated by medical insurance status (scale of 1-3, with 1 being the lowest and 3 the highest). Lesions were characterized as localized, disseminated (>10 lesions), or generalized (involving the trunk and upper or lower extremities), as previously defined.23 Annular GA lesions were composed of individual coalescing papules arranged in a ringlike or circinate configuration.3 In accordance with previous studies,3 cases with predominantly annular GA lesions were classified in the annular subgroup. A thorough search revealed 140 cases with complete historic and clinical/laboratory data that were selected from a pool of 400 patients; 420 individuals serving as controls, with a ratio to cases of 3:1, were selected from a pool of 4000. Controls were selected from primary care clinics at Rhode Island Hospital and South Shore Medical Center and were confirmed to not have a history of GA and skin diseases associated with hyperlipidemia, such as necrobiosis lipoidica. The records of patients serving as controls were matched to cases on the following variables: age, sex, race/ethnicity, type 2 diabetes mellitus, hypothyroidism, and hypertension. Furthermore, controls closely approximated cases in demographics and socioeconomic status. The study was approved by the Lifespan institutional review board, the major teaching hospitals affiliated with the Warren Alpert Medical School of Brown University.
Hypertension was defined as blood pressure of 130/85 mm Hg or higher and obesity as a body mass index higher than 30 (calculated as weight in kilograms divided by height in meters squared). Hypothyroidism was defined as overt disease validated by elevated (>5.0 mIU/L) thyroid-stimulating hormone and low (<0.8 ng/dL; to convert to picomoles per liter, multiply by 12.871) free thyroxine levels and/or active pharmacologic treatment. Type 2 diabetes mellitus was defined as recurrent or persistent hyperglycemia diagnosed by demonstrating any one of the following: fasting plasma glucose level, 126 mg/dL or more (to convert to millimoles per liter, multiply by 0.0555); plasma glucose level, 200 mg/dL or more, 2 hours after a 75-g oral glucose ingestion as in a glucose tolerance test; symptoms of hyperglycemia and random plasma glucose level, 200 mg/dL or more; and hemoglobin A1c concentration, 6.5% or more (to convert to proportion of total hemoglobin, multiply by 0.01).24
Metabolic syndrome was defined by the criteria of the International Diabetes Federation.25 The definition includes central obesity (defined as a waist circumference with ethnicity-specific values or body mass index >30) and any 2 of the following: elevated triglyceride concentration (>150 mg/dL; to convert to millimoles per liter, multiply by 0.0113) or specific treatment for this lipid abnormality, reduced HDL-C level (<40 mg/dL in men and <50 mg/dL in women; to convert to millimoles per liter, multiply by 0.0259) or specific treatment for this lipid abnormality, elevated blood pressure (systolic ≥130 mm Hg and/or diastolic ≥85 mm Hg) or treatment of previously diagnosed hypertension, and elevated fasting plasma glucose level (≥100 mg/dL) or previously diagnosed type 2 diabetes.
Histopathologic subtypes of GA included palisaded, interstitial, mixed (palisaded and interstitial), and deep. Cases with histopathologic characteristics suggestive of a granulomatous drug reaction26 and those showing histopathologic features associated with systemic disease (active vasculopathy, especially in the setting of an atypical clinical presentation)4 were excluded from the study. The age of the control at the time lipid values were determined was matched to that of the patient at the onset of disease. If a patient or control was receiving a lipid-lowering medication, such as a statin or fibrate, the lipid values obtained before the medication was started were reported. Hypercholesterolemia was diagnosed when the total cholesterol level was higher than 200 mg/dL (to convert to millimoles per liter, multiply by 0.0259) and hypertriglyceridemia when the triglyceride concentration was higher than 150 mg/dL. Low levels of HDL-C were diagnosed as 40 mg/dL and high levels of LDL-C were diagnosed when higher than 130 mg/dL. Dyslipidemia was defined as the presence of 1 or more of its associated components: hypercholesterolemia, elevated LDL-C, low HDL-C, and hypertriglyceridemia.
Descriptive frequencies, means, and medians were used to describe the study and control populations. The proportions of patients with DLP were compared between cases and controls by univariate analyses, using χ2 tests to compare categorical variables between the groups and unpaired t tests for comparison of continuous variables. The median values of the groups were compared with the Mann-Whitney test. Analysis of variance testing was used when a variable had 3 or more groups. All P values are 2-sided, and significance was assessed at P ≤ .05. Given the binary nature of the dependent variable, we used multivariate logistic regression with robust standard errors to examine the association between GA and DLP. The model was adjusted for potential confounders, including age and sex, in addition to cardiovascular risk factors (obesity, metabolic syndrome, and smoking).27 Use of β-blockers was also adjusted for because the earlier generation of this class, nonvasodilating β-blockers, may cause DLP.28 Independent variables were chosen based on previous studies.27,29 Adjusted odds ratios (ORs) with 95% CIs and P values were calculated. Lipid levels and age were analyzed as continuous variables. Multicollinearity between independent variables was assessed using the variation inflation factor. All statistical analyses were performed using commercial software (Stata/SE, version 10; StataCorp).
The study included 140 patients with GA (age, 19-81 years) and 420 controls (age, 18-84 years) (Table 1). The female to male ratio (2.8) was in accordance with previous studies.1 Controls were matched for age, sex, race/ethnicity, type 2 diabetes mellitus, hypothyroidism, and hypertension. Dyslipidemia was diagnosed in 79.3% of the patients compared with 51.9% of the controls (P < .001). Statistically significant differences between the patient and control groups were shown for total cholesterol, triglycerides, and LDL-C (P < .001 for all variables). The breakdown of patients with 1, 2, 3, or 4 components of DLP was 25.2%, 43.2%, 21.6%, and 9.9%, respectively. The respective percentages for controls were 37.6%, 42.2%, 14.7%, and 5.5%.
The prevalence of generalized GA was 16.4%, which is between that reported by Dabski and Winkelmann (9.5%)3 and that reported by Studer et al (25%).1 The distribution of lesions was in accord with previous studies.1 The upper extremity was the most common site involved (67.1%), followed by the lower extremity (50.7%) (Table 2). The lower extremity was more involved in disseminated/generalized than in localized GA (P < .001). The annular morphologic characteristics were less common in generalized than localized/disseminated GA, but the difference did not reach statistical significance. Palisaded was the most common histopathologic type. There was no association between histopathologic type and GA morphologic characteristics.
A statistically significant association resulted between the extent of GA and DLP (P = .02), with generalized GA showing a higher prevalence of DLP than the disseminated and localized subgroups (100% vs 71.4% and 77.3%) (Table 3). In addition, annular GA morphology was associated with hypercholesterolemia (P = .02) and DLP (P = .01). Clinical subgroups were not significantly correlated with any other comorbidities. There were no associations between the histopathologic type and lipid abnormalities or other comorbidities.
Results of multivariate logistic regression are shown in Table 4. The regression model demonstrated that GA was associated with DLP (OR, 4.04; 95% CI, 2.53-6.46), even after controlling for confounders, including age, sex, smoking, obesity, metabolic syndrome, and β-blocker use.
The present study showed an association between GA and DLP. After controlling for confounding variables, patients with GA had 4 times the odds of developing DLP compared with the controls. The study also revealed statistically significant quantitative differences for total cholesterol, triglycerides, and LDL-C between the GA and control groups. Generalized GA showed a higher prevalence of DLP than did the localized and disseminated subgroups, and the annular lesion morphology was associated with hypercholesterolemia and DLP.
Although associations of GA with systemic disease have been reported,1,4- 6 comorbidities have not been systematically investigated. Preliminary observations that support an association between GA and DLP include co-occurrence with necrobiosis lipoidica,9- 11 microdroplet lipid staining in GA (45.5% of generalized and 37.5% of localized),12 and sharing histopathologic features with disorders associated with DLP, such as necrobiosis lipoidica9,10 and eruptive xanthoma.13,14 The study by Dabski and Winkelmann12 showed that microdroplet lipid staining was positive in 80% of annular and 39.3% of nonannular GA cases, and this observation is consistent with the results of our study, which showed an increased prevalence of DLP and hypercholesterolemia in annular vs nonannular GA subgroups. The lipid staining in the study12 was seen only inside the histiocytes (ie, not outside the cellular aggregates of the granuloma). Interestingly, other granulomatous processes, such as the granulomatous variant of chronic pigmented purpuric dermatosis, have been associated with hypercholesterolemia.30,31 The lesions in cases of granulomatous pigmented purpuric dermatosis were clinically similar to those of GA, and it was suggested that an inflammatory reaction of microvasculature induced by microdroplets or microdeposits associated with hyperlipidemia may trigger a granulomatous dermatosis.31
The association of GA with DLP may be related to the inflammation caused by the disease. Inflammation plays an important role in the development of DLP,32,33 and chronic inflammation in diseases with cytokine patterns similar to that of GA, such as lichen planus and psoriasis,17,18 may trigger DLP.22 Treatment of DLP has been associated with a clinical improvement in psoriasis,34 and TNF inhibitors, which are helpful in psoriasis, have been associated with a beneficial increase in HDL-C level35 or decrease in serum triglycerides concentration.36 Tumor necrosis factor has been associated with the maintenance stage of granulomatous processes such as GA,37 and GA has been treated with TNF inhibitors,38,39 which may be attributed not only to downregulation of the helper T-cell subtype 1 cytokine pattern but also improvement of DLP. Nonetheless, GA onset has been associated with TNF inhibitor treatment (uncontrolled),40 possibly via autoimmune mechanisms.
On the other hand, a granulomatous process may be secondary to promotion of inflammation and/or microvascular dysfunction, including decreased skin blood flowmotion,41 as a result of hypercholesterolemia.42 Although microangiopathy has been reported in GA,43 microvascular function has not been studied. Still, the microangiopathy in GA has been considered similar to that observed in necrobiosis lipoidica and diabetes, with high β-glucuronidase levels possibly as a defense mechanism involved in the degradation of mucopolysaccharides and glycoproteins around blood vessels.44 This concept is supported by the histopathologic finding of thick, multilayered, hyalinized basal lamina around capillaries in GA, which was most prominent in the generalized form.45 Increased levels of oxidized LDL-C have been shown in psoriasis lesions46 and were associated with the immune inflammatory events that result in progressive skin damage. Interestingly, oxidized LDL-C has been shown to increase IL-2 receptors in T cells,47 and IL-2 is dramatically increased in GA.15 Therefore, it needs to be investigated whether oxidized LDL-C–mediated tissue damage can be involved in the pathogenesis of GA.
Our study has strengths and limitations. To our knowledge, this is the largest study on patients with well-characterized GA and the first to include adequate documentation of lipid abnormalities. To minimize selection biases, we included patients from both hospital and community environments. Furthermore, the distribution of potentially confounding factors, such as age, sex, obesity, type 2 diabetes mellitus, hypothyroidism, and hypertension, was homogeneous in the GA and control groups. Drug therapy (ie, β-blockers) could not have biased the results because it was balanced between the 2 groups. A possible limitation is that socioeconomic status was approximated based on insurance data and did not include data on educational level and annual income. Furthermore, we documented smoking status as current or never and therefore could not determine whether history of smoking could have introduced potential confounding. Similarly, the effects of other lifestyle factors, such as exercise and alcohol use, were not measured, although other included variables (ie, low level of exercise is associated with obesity, and excessive alcohol use is associated with obesity, metabolic syndrome, diabetes, and hypertension)48 may partially account for these factors.
The results of our study have important clinical consequences. First, a diagnosis of GA warrants an evaluation for possible co-occurrence of DLP. Because of its serious complications, DLP should be considered a more severe comorbidity than GA and warrants prompt management. Furthermore, the presence of generalized GA and/or annular lesion morphology should trigger a high index of suspicion for DLP. Finally, the close relationship with DLP may be relevant when defining management options for refractory GA, such as chronic generalized GA. In this case, combined treatment options, including lipid-lowering medications and/or TNF inhibitors, may be beneficial.
In conclusion, this study indicates an association between GA and DLP. Generalized GA shows a significantly higher prevalence of DLP than the disseminated and localized subgroups, and annular lesion morphology is associated with hypercholesterolemia and DLP in this study. Clinicians should be aware of these associations and consider them in the management of GA. Additional studies are needed to validate that GA can serve as a new marker for DLP.
Correspondence: George Kroumpouzos, MD, PhD, Department of Dermatology, Rhode Island Hospital, APC 10, 593 Eddy St, Providence, RI 02903 (email@example.com).
Accepted for Publication: March 29, 2012.
Published Online: June 18, 2012. doi:10.1001 /archdermatol.2012.1381
Author Contributions: Mr Wu and Dr Kroumpouzos had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Wu and Kroumpouzos. Acquisition of data: Wu, Robinson-Bostom, and Kroumpouzos. Analysis and interpretation of data: Wu, Robinson-Bostom, Kokkotou, Jung, and Kroumpouzos. Drafting of the manuscript: Wu and Kroumpouzos. Critical revision of the manuscript for important intellectual content: Wu, Robinson-Bostom, Kokkotou, Jung, and Kroumpouzos. Statistical analysis: Kokkotou and Jung. Administrative, technical, or material support: Wu and Robinson-Bostom. Study supervision: Kroumpouzos.
Financial Disclosure: None reported.
Additional Contributions: Ann Sayles, BS, manager of the Biochemistry Department, Rhode Island Hospital, and Nina Gabel-Jorgensen, BA, data analyst at South Shore Medical Center, Atrius Health, provided invaluable help in collecting the data on controls for the study.