Bone turnover markers to explain changes in lumbar spine BMD with abaloparatide and teriparatide: results from ACTIVE

Abaloparatide [1] is a synthetic peptide analogue of the human parathyroid hormone-related protein used in the treatment of postmenopausal osteoporosis in women at high risk for fracture. Abaloparatide binds to the parathyroid hormone 1 receptor, similar to teriparatide or PTH 1-34. Recent in vitro studies indicate that there are at least two stable, high-affinity PTHR1 conformations, R⁰ and RG. Binding to R⁰ results in prolonged signalling responses in cells and prolonged calcemic responses in animals, whereas selective binding to RG results in more transient responses. Abaloparatide has greater selectivity for the RG conformation than teriparatide and is thought to promote more transient signalling which results in an anabolic effect with less resorption [2].

Bone turnover markers (BTMs) can be used to monitor the effects of osteoporosis treatment. The International Osteoporosis Foundation has proposed that the reference marker for bone formation should be procollagen type I N propeptide (s-PINP) and serum carboxy-terminal cross-linking telopeptide of type I collagen (s-CTX) for bone resorption [3].

s-PINP has proven to be the most sensitive marker for detecting the response of bone turnover to teriparatide [4]. S-PINP increases within 3 days of starting this treatment [5] and reaches a peak after 3 to 6 months of treatment [6]. The early change in s-PINP correlates with the later change in BMD whether the treatment was with teriparatide [7, 8] or PTH 1-84 [9]. The relationship between early changes in biochemical markers of bone turnover and subsequent BMD response to abaloparatide has not been examined.

The ACTIVE Study compared abaloparatide with blinded placebo and with open-label teriparatide over 18 months [10]. In ACTIVE, the increase in lumbar spine BMD was similar for abaloparatide compared to teriparatide and the increase in total hip BMD was greater for abaloparatide, yet the increase in the bone formation marker s-PINP was lower with abaloparatide than teriparatide. The increase in the bone resorption marker s-CTX was also lower, raising the question of whether the relative increase of formation and resorption might differ and explain the differences in the pattern of BMD changes. The relative changes in bone resorption and formation can be estimated mathematically by calculating the mean and standard deviation of the relevant markers and subtracting resorption from formation [11]. This approach has been used during therapy of glucocorticoid-induced osteoporosis with teriparatide [12].

If the PINP is increased more by teriparatide than by abaloparatide, then why is the increase in BMD at the spine at 18 months similar for the two drugs and the increase in BMD at the total hip at 18 months different? Perhaps this could be explained by studying the relationship between the change in bone formation and bone resorption markers quantitatively using an ‘uncoupling index’.

Among the 600 randomly selected patients in the BTM subset, 184 received placebo, 189 received abaloparatide, and 227 received teriparatide. The baseline demographics were similar between the abaloparatide and the teriparatide groups (Table 1). They were also similar to the patients treated in the main trial with abaloparatide (n = 824) and teriparatide (n = 818) [10].

The serum concentrations of s-PINP and s-CTX during treatment and the uncoupling index (UI) are shown in Fig. 1. Both abaloparatide and teriparatide increased s-PINP and s-CTX concentrations as reported by Miller [10]. The pattern of increase was similar for the two agents during the first 3 months of treatment, but thereafter, s-PINP and s-CTX concentrations were greater for teriparatide than abaloparatide. Despite these differences, the UI was generally similar for abaloparatide and teriparatide during 18 months of treatment. The mean UI was moderately greater for abaloparatide than teriparatide at 1 month (1.743 vs 1.493, respectively; P = 0.03), but not at 3 months (Fig. 1, Table 2) or at later time points.

For patients included in this BTM analysis, BMD increased at the lumbar spine, total hip, and femoral neck with abaloparatide and teriparatide similar to the overall cohort [10]. After 18 months of treatment, the percentage change in lumbar spine BMD was similar between agents, but the percentage increase in total hip and femoral neck BMD was significantly greater for abaloparatide as compared to teriparatide (Fig. 2).

There was a significant correlation between early change in s-PINP and subsequent increases in BMD for both therapies. The strongest correlations between s-PINP and BMD were observed between 3 and 6 months; the 3-month time point was considered most clinically useful for providing early feedback and is described in more detail. The relationship between the log ratio of s-PINP at 3 months and percentage change in lumbar spine at 18 months for abaloparatide was r = 0.561 (P < 0.0001) and r = 0.198 (P = 0.003) for teriparatide; the correlation coefficient for abaloparatide was significantly larger than for teriparatide (P < 0.0001) (Table 3, Fig. 3). The removal of an outlying point (ratio of about 20) only had a small effect on this correlation for the teriparatide group (it changed from an r value of 0.198 to 0.215). The slope of the regression line between log ratio of s-PINP at 3 months and increase in lumbar spine BMD at 18 months was greater for abaloparatide (slope = 5.96) than for teriparatide (slope = 2.39) (Fig. 3; P = 0.0005) indicating that a similar early increase in s-PINP was associated with a greater increase in subsequent lumbar spine BMD. A significant correlation between the 3-month log ratio of s-PINP and total hip BMD was also observed for abaloparatide (r = 0.273; P < 0.001) but not for teriparatide (Table 3).

There was also a significant correlation between early change in s-CTX and subsequent increases in BMD. The relationship between change in the log ratio of s-CTX at 3 months and percentage change in lumbar spine BMD at 18 months for abaloparatide was r = 0.451 (P < 0.0001) and r = 0.042 (P = 0.53) for teriparatide; the correlation coefficient for abaloparatide was significantly larger than for teriparatide (P < 0.0001) (Table 3, Fig. 3). A significant correlation between the 3-month log ratio of s-CTX and total hip BMD was also observed for abaloparatide (r = 0.166; P = 0.023), but not for teriparatide (Table 3).

We also examined the relationship between the UI and subsequent increases in BMD. There was a significant correlation between the 3-month UI for abaloparatide and lumbar spine BMD after 18 months (r = 0.247; P < 0.001), but not for teriparatide (r = 0.115; P = NS; P = NS for the comparison between UI for abaloparatide versus teriparatide) (Table 3). A similar pattern was observed with total hip BMD (Table 3).

The changes in BTM could be influenced by changes in 1,25(OH)₂D. However, the mean increases in 1,25(OH)₂D at 1, 6, and 18 months for treatment with abaloparatide were 26, 16, and 13% and for teriparatide were 27, 18, and 15%.

In this study, we have tried to explain why two anabolic drugs differ in the increase in the bone formation marker s-PINP, yet had similar lumbar spine BMD increase. We think this might be explained by an alteration in the relationship between change in s-PINP and BMD and by the similarity in the increase in s-PINP relative to s-CTX.

We concluded this because we had examined the relationship between changes in BTMs and changes in BMD and found that this relationship differed for abaloparatide and teriparatide, with the slope being steeper for abaloparatide. It is unknown as to why the relationship is different for these two drugs. It may be that the relative effects on modelling and remodelling differ, but this would require a comparative study of bone histomorphometry.

An alternative, yet speculative, explanation is that regions of the skeleton differ in their responses to the treatments. BTMs reflect metabolism occurring within the entire skeleton; hence, if there were particularly potent effects of teriparatide on the skull [13], this would result in greater increases in BTMs, but not greater anabolic effects on BMD at the lumbar spine or total hip.

We hypothesised that the change in balance between formation and resorption early in treatment would be more predictive of 18-month change in lumbar spine BMD than a single marker of either formation or resorption alone. Thus, we calculated the UI, which increased in response to both anabolic agents. Such a finding has been reported previously for teriparatide in glucocorticoid-induced osteoporosis, with a value of about 2 at 3 months [12]. However, the UI was calculated differently in the current study as we addressed the skewed distribution of the BTMs by logarithmic transformation. We had used a similar approach recently to demonstrate an increase in bone turnover associated with bone loss in postmenopausal women [14]. The early difference in UI between treatments at 1 month, which was no longer apparent at 3 months or later time points, could help explain why lumbar spine BMD increased more quickly with abaloparatide but was similar to teriparatide after 18 months; however, this does not fully explain why there was a greater increase in proximal femur BMD with abaloparatide.

Bone histology has been reported for the ACTIVE study [15]. It would be helpful in the future to reconstruct the bone remodelling cycle in order to calculate ‘bone balance from resorption depth and mean wall thickness’ as this is the histological equivalent of ‘uncoupling index’.

The strengths of our study are the large sample size given that we are studying surrogates, the continuous outcomes, and the inclusion of both teriparatide and abaloparatide in the same study, with patients randomly assigned to treatment. Also, the BTM and BMD were measured using state-of-the-art techniques, and we analysed the data while taking account of the skewed distribution of BTMs. The weaknesses are that teriparatide was given as an open-label treatment and there is an insufficient number of fractures to compare the relationship between changes in bone turnover markers and fracture.

In conclusion, early changes in s-PINP correlate with the percentage change in lumbar spine BMD after 18 months of treatment with both abaloparatide and teriparatide, although the correlation with abaloparatide was greater. The UI was generally similar between abaloparatide and teriparatide, supporting that while absolute levels of s-PINP and s-CTX were lower with abaloparatide than teriparatide, the balance between the formation and resorption marker was similar.