This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Orr-Walker, B. J. Articles by Reid, I. R. Search for Related Content PubMed PubMed Citation Articles by Orr-Walker, B. J. Articles by Reid, I. R.

Premature Hair Graying and Bone Mineral Density¹

Department of Medicine, University of Auckland, Auckland, New Zealand

Address all correspondence and requests for reprints to: Brandon J. Orr-Walker, Department of Medicine, University of Auckland, Private Bag, 92019, Auckland, New Zealand.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
In a recent case-control study, premature hair graying was found to be associated with osteopenia, suggesting that this might be a clinically useful risk factor for osteoporosis. We report a reexamination of this possibility in 293 healthy postmenopausal women. Subjects experiencing onset of hair graying in their 20s tended to have lower bone mineral density throughout the skeleton (adjusted for age and weight) than those with onset of graying later in life. The same was true for those in whom the majority of their hair was gray by the age of 40 yr (n = 16), in whom bone density was reduced by 7% in the femoral neck, 8% in the femoral trochanter, and 4% in the total body (P < 0.05) when compared with those not prematurely gray. Bone density at the lumbar spine and Ward’s triangle showed similar trends that were not significant. However, premature hair graying explained only 0.6–1.3% of the variance in bone mineral density within the population. We conclude that premature hair graying is associated with low bone density, but that its infrequency in the normal postmenopausal population leads to its accounting for only a tiny fraction of the variance of bone density.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
THE existence of clinical markers of osteoporosis is an attractive concept to help identify those people at particular risk and may also give some insight into the pathogenesis of the condition. Recently, premature hair graying was reported to be associated with osteopenia (1). In this case-control study, premature hair graying was associated with a greater than 4-fold increased risk of osteopenia [defined as a lumbar spine bone mineral density (BMD) T-score < -1]. This observation was made in subjects in whom there were no other identifiable risk factors for osteopenia nor conditions that might result in hair depigmentation.

This important observation was subject to a number of methodological limitations however. First, the subjects were drawn from three different sources, specifically patients referred by their doctors, self-referrals, and volunteer subjects in clinical trials. It is not clear, therefore, that the cases and controls came from the same reference population, nor whether the proportion of cases and controls from each source were different. Second, although information on a large number of subjects was available for analysis, rigorous exclusion of potentially confounding conditions resulted in only 15% of those subjects with osteoporosis being included in the analysis. However 64% of those with premature hair graying were included, suggesting that the exclusion criteria affected cases and controls differently. Third, the decision to define premature hair graying as the majority of hair gray before age 40 yr was not justified and therefore might represent a post hoc analysis of the data. Supporting this possibility, Rosen et al. (1) were unable to show a significant correlation between age of hair graying and raw or age-adjusted BMD. Furthermore, defining osteopenia as a dichotomous variable is arbitrary and does not reflect the continuous relationship between BMD and fracture risk. In addition, BMD at sites other than the lumbar spine were not reported. Finally, inclusion of both male and female subjects in the same analysis may have introduced bias if there was gender disproportion in the contingency tables. The applicability of this interesting association to an otherwise healthy population at risk of osteoporosis (e.g. normal postmenopausal women) is, therefore, uncertain. We readdressed this question by examining the association between premature hair graying and BMD throughout the skeleton in 293 healthy postmenopausal women.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Subjects

The study population comprised 404 normal postmenopausal women involved in studies of osteoporosis prevention in our department. Exclusion criteria included disorders of calcium metabolism; renal, thyroid, or hepatic dysfunction; other major systemic illness; current use of anticonvulsant or glucocorticoid drugs; past use of hormone replacement therapy for >6 months; and use of sodium fluoride, calcitonin, anabolic steroids, or bisphosphonates.

Study protocol

All women were sent a written questionnaire asking them to recall when their hair began to turn gray, and when the majority of their hair (>70%) turned gray. The responses were classified by decade of age or as ’not yet gray.’ Subjects were also asked whether they had pernicious anemia or vitiligo, conditions associated with premature hair graying. Data from subjects with either condition were excluded from the analysis. Premature hair graying was defined a priori as the majority hair graying before age 40 yr.

Measurements

BMD was assessed using a Lunar DPX-L dual-energy x-ray absorptiometer (Lunar, Madison, WI). Scans of the whole body, lumbar spine (L2-L4), and proximal femur were performed.

Statistical analysis

All analyses used the programs of SAS version 6.04 (SAS Institute, Cary, NC). One-way ANOVA was performed to test the hypothesis that the mean BMD at each site did not change across the decades of hair graying. To assess whether there was a critical age of hair graying that was associated with osteopenia, BMDs of subjects grouped in consecutive decades were compared by Student’s t test. Comparison of the BMDs of those with premature hair graying with those of the remainder of the cohort was made using Student’s t test. Comparisons of weight- and age-corrected BMDs were made using the general linear models procedure of SAS. Because all comparisons were made a priori, no adjustment of (0.05) was needed or made. All tests were two-tailed. Data are given as the mean ± SEM unless indicated otherwise.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Subjects

Questionnaires were sent to 404 women, and completed replies received from 334 (83%). Four subjects reported pernicious anemia and 37 had vitiligo. Therefore, data from 293 subjects (73%) were available for analysis. Subjects’ ages ranged from 45–88 yr (mean ± SD, 60 ± 5.8 yr)

BMD and decade of onset of hair graying

The mean BMD of subjects, grouped by the decade of commencement of hair graying, is summarized in Table 1. ANOVA showed differences in the age- and weight-adjusted BMD at all skeletal sites except the femoral trochanter. When data were analyzed by comparison of successive decades, there were differences in age- and weight-adjusted BMD between those who had onset of hair graying in their 20s vs. those with onset in their 30s in the total body and at the femoral trochanter, femoral neck, and Ward’s triangle, with lower BMD in those with earlier graying. In addition, subjects who reported commencement of hair graying in their 50s had lower age- and weight-adjusted BMDs in the lumbar spine and Ward’s triangle than those with commencement of hair graying in their 40s.

The mean BMD of subjects grouped by the decade during which the majority of their hair became gray is summarized in Table 2. ANOVA revealed no differences between groups at any site except for unadjusted BMD in the total body (P = 0.05) and a trend at the same site after adjustment for age and weight (P = 0.07). When data were analyzed by comparison of successive decades, age- and weight-adjusted BMD was significantly lower in those with the majority hair graying occurring during their 30s compared with those in whom it occurred in their 40s atn the lumbar spine, femoral neck, femoral trochanter, and in the total body.

The relationship between premature hair graying and BMD is summarized in Table 3. There were significant differences between the two groups in age- and weight-adjusted BMD in the total body and at the femoral neck and femoral trochanter, with lower BMD in those with premature hair graying. Nonsignificant trends towards lower BMD were noted at the other two sites. Current smoking (n = 25) or any history of smoking (n = 121) were unrelated to premature hair graying (², P = 0.57 and P = 0.87, respectively) and had no effect on BMD.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The decision to define premature hair graying as the majority of hair graying before 40 yr of age was made a priori in the present study, based on the findings of Rosen et al. (1). However, in our subject population it also appears to be the most sensitive cut off, because it marked the only significant difference of age- and weight-adjusted BMDs between decades during which the majority of hair turned gray. A similar BMD step was present in the age-of-onset of graying data (a decade earlier), representing essentially the same subjects and further justifying this cut off as the definition of premature graying.

Although significant differences in unadjusted BMD were apparent between those with the majority of hair graying in their 50s and those in their 60s, when adjusted for age and weight, significant differences only remained between those with the majority hair graying in their 30s and those in whom it developed in their 40s. The differences in unadjusted BMD between those with the majority of hair graying in their 50s vs. those in their 60s are likely to be because of the greater current age and lower weight in the latter group. BMD differences according to the reported age of onset of graying were apparent between the 20s and 30s and also the 40s and 50s. This is essentially the same group as those just discussed, though why the differences are now significant following appropriate adjustment is not clear. This may simply be a chance finding resulting from the number of comparisons we have carried out.

It is appropriate to consider the extent to which the results of this study agree with those of Rosen et al. (1). Our data are expressed in terms of the difference in bone density between women with and without premature hair graying, and also in terms of the percent of variance of BMD accounted for by premature hair graying; whereas Rosen calculated the odds ratio for a prematurely gray subject having osteopenia. Our data can also be analyzed in that way. The odds ratio of a prematurely gray woman having a bone density more than one SD below the age- and weight-adjusted mean was 1.81 (95% confidence interval, 0.56–5.79, P = 0.32) at the lumbar spine, 3.32 (95% confidence interval, 1.06–10.43, P = 0.04) at the femoral neck, 2.64 (95% confidence interval, 0.73–9.56, P = 0.14) at the trochanter, 1.36 (95% confidence interval, 0.42–4.37, P = 0.61) at Ward’s triangle, and 1.48 (95% confidence interval,0.41–5.43, P = 0.55) in the total body. The 95% confidence intervals for all these measures overlap with that found by Rosen et al. at the lumbar spine, indicating that the two studies are broadly consistent in their findings.

It is interesting to speculate as to why hair graying may be linked to low BMD. Apart from potentially confounding associations with autoimmune conditions such as vitiligo, Addison’s disease, Grave’s disease, premature hypogo-nadism, and Werner’s syndrome (each excluded from both our own and Rosen’s studies), premature graying has been shown to be inherited in an autosomal dominant pattern (2) and to be less frequent in racial groups with higher BMD (e.g. blacks) (3). This suggests that it may be a linked to genetic factors that influence BMD. Alternatively, the processes that lead to loss of scalp melanin production might also impact on bone turnover.

	Acknowledgments

 
We thank G. Gamble for his expert statistical assistance.

	Footnotes

 
¹ This work was supported by the Health Research Council of New Zealand.

Received March 14, 1997.

Revised July 9, 1997.

Accepted July 21, 1997.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Rosen CJ, Holick MF, Millard PS. 1994 Premature graying of hair is a risk marker for osteopenia. J Clin Endocrinol Metab. 79:854–857.[Abstract]
2. Mosher DB, Fitzpatrick TB, Ontonne J-P. 1979 Abnormalities of pigmentation. In: Fitzpatrick TB, Eisen AZ, Wolff K, Freedberg IM, Austen KF, eds. Dermatology in general medicine. New York: McGraw-Hill; 582–590.
3. Arnold HL, Odom RB, James MD, eds. 1990 Andrew’s diseases of the skin, 8th ed. Philadelphia: WB Saunders; 891.

This article has been cited by other articles:

				D. J. Morton, D. Kritz-Silverstein, D. J. Riley, E. L. Barrett-Connor, and D. L. Wingard Premature Graying, Balding, and Low Bone Mineral Density in Older Women and Men: The Rancho Bernardo Study J Aging Health, April 1, 2007; 19(2): 275 - 285. [Abstract] [PDF]

				P. C. Arck, R. Overall, K. Spatz, C. Liezman, B. Handjiski, B. F. Klapp, M. A. Birch-Machin, and E. M. J. Peters Towards a "free radical theory of graying": melanocyte apoptosis in the aging human hair follicle is an indicator of oxidative stress induced tissue damage FASEB J, July 1, 2006; 20(9): 1567 - 1569. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Orr-Walker, B. J. Articles by Reid, I. R. Search for Related Content PubMed PubMed Citation Articles by Orr-Walker, B. J. Articles by Reid, I. R.