Gender-Specific Association of Androgenetic Alopecia With Metabolic Syndrome in a Middle-aged Korean Population

S.M. Yi; S.W. Son; K.G. Lee; S.H. Kim; S.K. Lee; E.R. Cho; I.-H. Kim; C. Shin


The British Journal of Dermatology. 2012;167(2):306-313. 

In This Article


Over the past few decades, various authors have investigated the relationship between male AGA and various cardiovascular risk factors.[4,6,7,9,28–33] At the same time, many controversies exist regarding the presence of an association between AGA and those cardiovascular risk factors.[14–16,34] Several studies showed a relationship between coronary heart disease and AGA in men.[4,28–32] Early onset of AGA (before the age of 36 years) is a risk factor for early onset of severe coronary heart disease.[29] Other studies showed that vertex pattern baldness appears to be a marker for increased risk of myocardial infarction[32] and coronary artery disease[31] in men. However, other cross-sectional studies did not show this relationship.[15,34] There are also many studies that show the relationship between AGA and insulin-resistance-associated disorders, such as hyperinsulinaemia, obesity, hypertension, and dyslipidaemia.[6,7,9,33] However, Nabaie et al.[14] and Abdel Fattah and Darwish[16] did not find an association between AGA and insulin resistance.

Recently, two studies[20,21] have focused on the association between AGA and MetS, which comprise a combination of medical disorders that in concert increase the risk of developing cardiovascular disease and diabetes. Arias-Santiago et al.[20] conducted a case–control study to analyse the presence of cardiovascular risk factors included in the ATPIII criteria for MetS.[25] The study included 154 participants, 77 with early-onset AGA (40 male and 37 female) and 77 healthy control subjects (40 male and 37 female). They found a higher prevalence of MetS in the male and female AGA groups than in their respective control groups. Su and Chen[21] reported a significant association between AGA and MetS in 670 Taiwanese men in a population-based cross-sectional survey with adjustment for age, family history and smoking.

However, our study found significant differences from previous studies.[20,21] In the group of 1563 men, no statistically significant association was found between AGA and the presence of MetS. On the contrary, we found an association between AGA and MetS in the group of 1321 women after adjusting for significant risk factors, including age, family history of AGA and smoking status. Because the sample size of the female group in our study, especially Ludwig II and III, was too small to achieve statistical power, we compared a population of men with more advanced stages of hair loss with a population of women with mild as well as advanced hair loss. However, female patients with mild AGA (Ludwig class I) may be a less certain diagnosis and induce heterogeneity into the study population. In future, a study with larger numbers of patients is needed to confirm our findings. In addition, this study also found that a higher number of fulfilled components of MetS is associated with a greater risk of AGA. Among each of the components of MetS, there was no statistically significant association with AGA in either the male and female groups. There was no previous study that reported different results between men and women regarding the association of MetS with AGA. It is unclear whether that difference resulted from the different mechanism of hair loss between men and women or the failure to control for other unknown compounding factors in the population group. In addition, the difference in severity of hair loss between the male and female populations might be an important factor to induce gender-specific difference of AGA with MetS.

Up to now, various mechanisms of AGA have been suggested, but the exact mechanism is not well known. Several factors, including genetic predisposition, presence of sufficient androgens, androgen receptors and androgen receptor coactivators, are known to be associated with the development of AGA. Currently, AGA is thought to be a polygenetic condition with a wide spectrum of severity, age of onset and scalp location of hair loss. In addition, the fact that AGA has a polygenic mode of inheritance and wide phenotype range is likely a result of the actions of multiple genes that collectively contribute to the manifested phenotype. More recently, identification of new susceptibility genes on chromosomes 3q26 and 20p11 suggest that non androgen-dependent pathways may exist.[35,36]

Phenotypic differences between male and female AGA suggest the possibility of different mechanisms between male and female AGA. In men, hair loss frequently occurs in a characteristic 'horseshoe' pattern, in which hair is lost on the front and top of the head but not on the back and sides. Although AGA is as frequent in women as it is in men, the hair loss is usually less severe, does not occur in the characteristic pattern, and is manifested as an overall thinning. Some studies have suggested that male AGA and female AGA are different entities because of those differences in the pattern of hair loss and the age of onset. Therefore, the term, 'female pattern hair loss (FPHL)', has been commonly used rather than its synonym, female AGA. The role of androgens in FPHL has not been clearly established.

The possibility of the existence of different mechanisms between male and female AGA has been supported by many case reports and studies. Some investigators have failed to find evidence of raised androgen levels in women with FPHL,[37,38] and in all studies, there is a variable proportion of women with hair loss who do not show clinical or biochemical signs of androgen excess. Finasteride, a 5α-reductase inhibitor of proven efficacy in male pattern balding, failed to halt the progression of hair loss in a large controlled trial in postmenopausal women who had normal androgen levels.[39] The androgen receptor blocker cyproterone acetate similarly failed to prevent the progression of hair loss in a trial comparing cyproterone acetate to topical minoxidil in premenopausal women with hair loss.[40] A report of female pattern hair loss in a patient with 17α-hydroxylase deficiency supports the idea that hyperandrogenaemia does not necessarily lead to FPHL and that FPHL can develop in individuals with extremely low levels of circulating androgens.[41] Orme et al.[42] reported a young woman with hypopituitarism presenting with the clinical and histological features of FPHL in the absence of detectable levels of circulating androgens. Female pattern hair loss occurring in a patient with complete androgen insensitivity syndrome also suggests that mechanisms other than direct androgen action contribute to this common form of hair loss in women.[43]

When each component of MetS was considered individually, associations for all five components (waist circumference, TG, blood HDL-C, blood sugar and BP) were not statistically significant in our study. The pathophysiological link between MetS and AGA is not yet well understood. Insulin resistance associated with AGA has previously been reported and might contribute to this association.[6] However, as a raised fasting glucose level was not associated with AGA in our study, our finding cannot confirm the relationship between insulin or insulin resistance and AGA. Recently, two studies[16,21] reported no association between AGA and insulin resistance, concordant with our study.

In our study, smoking status was not a statistically significant factor responsible for AGA after controlling for age and family history. The association between smoking status and AGA has been addressed in five studies with inconsistent results.[12,13,17,18,21] Three studies[12,13,21] showed a positive association, one study[17] failed to demonstrate a statistically significant positive association, another study[18] showed opposite findings, albeit statistically nonsignificant. A harmful effect on the microvasculature of the dermal hair papilla, smoke genotoxicants-induced damage to DNA of the hair follicle, smoking-induced imbalance in the follicular protease or antiprotease systems, a relative hypoestrogenic state by inducing increased hydroxylation of oestradiol, and inhibition of aromatase are proposed mechanisms by which smoking influences AGA,[1] but its association with AGA is still vague.

The prevalence of AGA in our study was lower than the prevalences found in studies on Caucasians but was slightly higher than those in previous studies targeting Asian people.[44–46] The prevalences of AGA in the male and female groups (Norwood III or above in men, Ludwig I or above in women) were 41·2% and 16·7%, respectively. However, previous studies[44–46] targeted all age groups, but our study targeted only people above 40 years, and the mean age of participants in our study was around 55 years in both the male and the female groups.

In conclusion, our analysis of AGA and the prevalence of MetS in 2884 healthy male and female participants in the population-based Korean Health and Genome Study demonstrated quite different features compared with previous reports after making adjustments for other significant confounding factors. When multiple regression was used to adjust for age, family history and smoking, there was no significant association between the prevalence of MetS and moderate to severe androgenetic alopecia in the male group. On the contrary, a statistically significant positive association was noted between the prevalence of MetS and AGA in the female group. The different results between the male group and female group suggest that there may be quite different mechanisms that are yet to be defined between male and female AGA.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: