Summary
This study from the WHI cohort aimed to assess whether a second bone mineral density (BMD) measurement approximately three years after the initial assessment would improve the ability to predict the risk of fracture beyond the baseline BMD measurement alone [1]. The data of 7,419 women (mean age of 66 years at baseline) who underwent both baseline and year 3 BMD measurements were analyzed. Throughout the 9-year follow-up, 139 women (1.9%) experienced hip fractures, and 732 women (9.9%) suffered major osteoporotic fracture (MOF) including hip, spine, radius, ulna, wrist, upper arm, or shoulder fracture. Absolute change in total hip BMD was significantly associated with the risk of hip fracture (HR: 1.29; 95% CI, 1.08-1.54 per 1 SD decrease in BMD; p = 0.004) after adjustment for baseline BMD, age, race/ethnicity, history of fracture, physical activity level, BMI, physical function level, frequency of falls and hormone us. The relationship was weaker for the risk of MOF (1.11; 95% CI, 1.03-1.20 for 1 SD decrease in total hip BMD). The discriminative value for hip fracture as evaluated by the area under the receiver operating characteristic curve (ROC-AUC) was comparable for baseline total hip BMD, change in total hip BMD or the combination of baseline total hip BMD and change in total hip BMD. The femoral neck and lumbar spine BMD values had similar discrimination for hip fractures. The authors concluded that a second BMD measurement did not improve the capability to discriminate women who experienced a subsequent hip fracture or MOF from women who did not beyond the baseline BMD value alone and therefore that repeat BMD measurement would not be of sufficient interest to be advised in routine clinical practice for fracture risk assessment.
Commentary
The issue of repeated BMD measurements in postmenopausal women to assess fracture risk is still widely discussed. Most of the studies published so far [2-4] have reported a lack of benefit of repeated BMD assessment (performed 4 to 8 years later) beyond baseline BMD alone in discriminating postmenopausal women who experienced subsequent osteoporotic fractures (mainly hip fractures) from those who did not. The main limitation of these studies was that they were restricted to older subjects and thus did not allow to draw conclusions for younger postmenopausal women. In this current study, the authors indicate that 44% of the study population was younger than 65 years, although they did not give the mean age in this subgroup nor the percentage of women within the first 5-10 years after menopause.
Even though their data were consistent with previous studies, several methodological flaws need to be discussed since these are likely to limit the scope of their conclusions. At least three of them of particular importance must be acknowledged:
- The study population was very heterogeneous in terms of age, with 56% of subjects older than 65 years, body weight (35% of the population had a BMI > 30 kg/m2), ethnicity (15% were African-American women) and current treatment (47% were receiving MHT and 30%, a medication known to interfere with bone mass/loss). Even though the results were adjusted for most of these confounding factors, it is very questionable that neither age, BMI or MHT were found to interact with the rate of bone loss.
- The absolute BMD changes were minimal and only negative at the femoral neck (-0,028% per year). Very surprisingly, the mean total hip BMD was reported to increase over three years (+0,197 % per year) as well as lumbar spine BMD. However, the latter is likely explained by increasing spinal degenerative-changes with age.
- The comparisons of ROC-AUCs do not allow for a detailed assessment of the additional value of a given parameter. Another feature of the usefulness of a predictive model is the risk reclassification [5], and it would have been more appropriate to calculate the net reclassification improvement, namely the percentage of women who would have been reclassified with regards to their risk of fracture after taking into account the rate of bone loss. This would be of particular interest in women with a so-called “intermediate” fracture risk. In any case, it must be recalled that most of the WHI population was at low if not at very low risk of fracture. The mean total hip and femoral neck T-scores were -0.655 and -1.115, respectively (for a mean population age of 66 years) with only 3%, 8% and 12% of the women with a T-score ≤ -2.5 at the total hip, femoral neck and lumbar spine, respectively.
Finally and most importantly, it is evident that serial BMD measurement is only of interest in clinical situations where BMD is likely to vary rapidly and differently among patients, which is typically the case in early menopause. The impact of the rate of postmenopausal bone loss on subsequent fracture risk has been well established since the pioneering work of Claus Christiansen [6]. This new analysis of the WHI does not in any way question the value of bone densitometry monitoring at the beginning of menopause under certain conditions (for instance, vertebral T-score between -2 and -2.5, high bone remodelling, a clinical risk factor for fracture) [7,8]. Further studies are thus needed before addressing clinical decision-making regarding BMD monitoring in early postmenopausal women.
Florence Trémollieres, MD, PhD
Head of the Menopause and Metabolic Bone Disease Center, Hôpital Paule de Viguier, University Hospital of Toulouse, 330 avenue de Grande-Bretagne, TSA 700354, 31059 Toulouse, France
References :
- Crandall CJ, Larson J, Wright NC, et al. Serial bone density measurement and incident fracture risk discrimination in postmenopausal women. JAMA Intern Med 2020; doi:10.1001/jamainternmed.2020.2986.
https://pubmed.ncbi.nlm.nih.gov/32730575/ - Nguyen TV, Center JR, Eisman JA. Femoral neck bone loss predicts fracture risk independent of baseline BMD. J Bone Miner Res 2005; 20:1195-201.
https://pubmed.ncbi.nlm.nih.gov/15940372/ - Hillier TA, Stone KL, Bauer DC, et al. Evaluating the value of repeat bone mineral density measurement and prediction of fractures in older women: the study of osteoporotic fractures. Arch Intern Med 2007; 167:155-60.
https://pubmed.ncbi.nlm.nih.gov/17242316/ - Leslie WD, Morin SN, Lix LM. Manitoba Bone Density Program. Rate of bone density change does not enhance fracture prediction in routine clinical practice. J Clin Endocrinol Metab 2012; 97:1211-18.
https://pubmed.ncbi.nlm.nih.gov/22278427/ - Schousboe JT, Langsetmo L, Taylor BC et al. Fracture Risk Prediction Modeling and Statistics: What Should Clinical Researchers, Journal Reviewers, and Clinicians Know? J Clin Densitom 2017; 20:280-90.
https://pubmed.ncbi.nlm.nih.gov/28712982/ - Riis BJ, Hansen MA, Jensen AM, Overgaard K, et al. Low bone mass and fast rate of bone loss at menopause: equal risk factors for future fracture: a 15-year follow-up study. Bone 1996; 19:9-12.
https://pubmed.ncbi.nlm.nih.gov/16234965/ - Sornay-Rendu E, Munoz F, Duboeuf F et al. Rate of forearm bone loss is associated with an increased risk of fracture independently of bone mass in postmenopausal women: The OFELY study. J Bone Miner Res 2005; 20:1929-35.
https://pubmed.ncbi.nlm.nih.gov/16234965/ - Berger C, Langsetmo L, Joseph L et al. Association between change in BMD and fragility fracture in women and men. J Bone Miner Res 2009; 24: 361-9.
https://pubmed.ncbi.nlm.nih.gov/18847328/
