A1c and Blood Pressure Fluctuations Predict Fracture Risk
Two studies appearing in the April 2022 issue of the Journal of Clinical Endocrinology & Metabolism move our knowledge of risk from the obvious — age, estrogen deficiency, and frailty — to common data from office encounters, namely glycemic control and blood pressure (BP). These studies don't project far enough into the future to register as fracture susceptibility, however. Can someone share more information ?
@jovis-johny The first report is a multicenter US study that looked at correlations of A1c values as a risk-stratifying measurement. Of note, the authors are primarily computer scientists from a prominent engineering school, with a few physicians included.
Using a cohort of patients enrolled in a regional Medicare Advantage program, some with preexisting diabetes, and spanning a variety of medical conditions and medication prescriptions that physicians would commonly encounter in their senior populations, the investigators targeted red flags that preexisting or newly acquired diabetes might create for new fractures. They assessed how many new fractures occurred for different levels of glycemic control over 2 years since the A1c was last measured. They also determined which commonly used hypoglycemic agents seemed to protect from these new fractures and which increased the risk or identified people who might sustain falls from their side effects.
Starting with 4 million diabetic enrollees, they winnowed the population down to about 150,000 above a prespecified age who had been enrolled in the program longer than 3 years and had A1c data available. Among these, they found 18,000 who had sustained a fracture. The A1c taken 1-2 years before the fracture correlated with the likelihood of sustaining a fracture. For every 1% rise in A1c, the fracture risk increased about 8%. For those with A1c greater than 9% compared with those under better control, the added risk was about 30%.
Medication choice had its correlates as well, with those on metformin and dipeptidyl peptidase 4 inhibitors having relative protection while those on thiazolidinediones and insulin having higher fracture rates. Of interest, those who are on bisphosphonates for fracture prevention also had a 10% higher fracture rate. This was believed to be related to preferential prescriptions provided to those with the greatest fracture risk according to accepted criteria.
BP Variability and Fractures
As cardiovascular therapy becomes more successful, those benefiting from longer life spans also experience orthostasis or other BP fluctuations or may take medications that add to fracture susceptibility. Technology has enabled us to extract BP measurements over a series of office visits and match variations with a variety of health outcomes, including subsequent cardiac events, cognitive impairment, and even mortality. The second study was a retrospective review of large amounts of accumulated patient data with computerized sorting. Korea maintains a centralized health system with a database of encounters, covering 97% of their population. This enables endless explorations of medical situations and results of treatments. Korean investigators assessed the likelihood that BP variations over serial visits would predict subsequent fractures in those older than 50 years.
First, they had to define what "variability independent of the mean" was, resulting in a mathematical formula probably a little more intricate than one that would appear on the SATs. They then looked at fracture occurrences of many types with attention to expected confounders such as age, absolute BP levels independent of variability, socioeconomic divisions of the population, and medical care that would affect BP.
Those with the greatest BP variations tended to be older, female, and have other treated medical conditions. About 10% of the 3 million individuals studied sustained a fracture during the 7 years of review; about half of these were vertebral. Although those with the most variability sustained the most fractures, the peak risk was a modest 7% in the quartile that had both systolic and diastolic variability. The use of antihypertensive agents did not seem to add to the cumulative risk. When assessed with other comorbidities, BP variation remained a small but independent risk factor.
Many medical conditions such as fractures, myocardial infarctions, and many cancers have long silent phases with sudden dramatic presentations. Managing risk factors to minimize future morbidity has yielded gratifying results for many decades. Those risks must be part of the clinician's awareness and there needs to be effective intervention. A1c is one of those parameters that always registers as favorable or unfavorable when we see it. As our patients become older and more prone to injury, the first study suggests that some medical options for dealing with glycemia may have an osseous advantage.
Variability of BP may also be predictable but is less of a part of our exam room awareness. As its utility becomes better defined and the capacity of our electronic records generate that as data, it may also have a future role in our periodic assessments of the older people whom we guide through their later years.