Background/Purpose: BMD is a strong predictor of fracture risk in untreated patients. Recent evidence suggests that BMD achieved during treatment also reflects fracture risk; thus, T-scores are being considered as a target to guide osteoporosis treatment. In the Active-Controlled Fracture Study in Postmenopausal Women with Osteoporosis at High Risk trial (ARCH [NCT01631214]), romosozumab (Romo), an investigational bone-forming agent with a dual effect of increasing bone formation and decreasing bone resorption, followed by alendronate (ALN) had greater efficacy in fracture risk reduction and BMD gains vs ALN alone (Saag NEJM 2017). Of note, a cardiovascular imbalance was observed with Romo in ARCH and a comprehensive assessment of these data is ongoing. Here we explored the relationship between T-scores achieved on-study after 1 year with Romo or ALN and subsequent fracture risk.

Methods: Postmenopausal women with osteoporosis and prior fragility fracture were randomized 1:1 to receive Romo 210mg SC QM or ALN 70mg PO QW for 12 months, followed by open-label (OL) ALN 70mg PO QW for ≥12 months, with an event-driven primary analysis. We examined change from baseline in BMD and T-scores at 12 months and the relationship between total hip T-scores at month 12 and subsequent nonvertebral (NVT) fracture rates. We also compared fractures in the OL period, including new vertebral (VT) fractures in year 2 (based on month 24 spine radiographs) and clinical, NVT, and hip fractures between arms in the full OL period.

Results: ARCH enrolled 4093 patients (2046 Romo, 2047 ALN); mean baseline T-scores were –2.96 at the lumbar spine and –2.80 at the total hip. 3465 patients (1739 Romo, 1726 ALN) received ≥1 OL ALN dose in the OL period (median 1.9 years follow-up). Mean total hip BMD increased by 6.2% for Romo and 2.8% for ALN in the first year, with increases in T-score of 0.31 and 0.15, respectively. At month 12, the achieved total hip T-score was associated with the 1-year NVT fracture rate observed in the OL period (Figure) and the relationship was independent of the drug received in the first year. During the OL period, when all patients were on ALN, patients who received Romo first had a 75% lower relative risk of new VT fracture (P<0.001), and had reductions in clinical (32%, P=0.001), NVT (19%, P=0.120), and hip (40%, P=0.041) fractures vs patients who received ALN first.

Conclusion: Higher absolute total hip T-scores achieved on therapy at month 12 resulted in subsequent lower fracture risk regardless of the treatment received, with ongoing benefits from building a BMD foundation. These data support the concept of a T-score target to improve outcomes in osteoporosis treatment.

Figure: Month 12 total hip T-score and nonvertebral fracture rate during the open-label period