Evidence supporting PINP as a biological response marker during teriparatide treatment
The teriparatide Phase 3 Fracture Prevention Trial (NCT00670501) was a randomized, placebo-controlled trial of postmenopausal women with osteoporosis and vertebral fractures [5]. The prospectively collected biological markers of bone formation included carboxy-terminal extension peptide of procollagen type 1 (PICP) and serum bone-specific alkaline phosphatase (bone ALP). In the teriparatide group, PICP increased dynamically at 1 month but returned toward baseline by 3 months [24]. The transience of PICP elevation during teriparatide treatment suggests it is not an optimal biological response marker. In contrast, bone ALP gradually increased between baseline and 12 months in the teriparatide group [24]. Yet, because bone ALP does not show dynamic increases early during teriparatide treatment, it also may not be optimal as an early biological response marker in clinical practice.
Following completion of this trial, intact PINP was used to measure serum PINP concentrations from samples collected in a subset of patients at baseline and after 3 months on study drug [24]. These samples had been stored for at least 4 years and previously thawed at least twice prior to PINP assessment [17]. In a subsequent analysis of these trial results, the PINP response to teriparatide (signal) was found to be large compared to the variability (noise, or fluctuation, with placebo) of the test, as shown in Fig. 2 [17]. A relatively high signal-to-noise ratio of PINP relative to other prospectively defined markers of bone turnover was observed in the trial [17]. Accordingly, PINP may be particularly useful as a biological response marker in patients treated with teriparatide.
Subsequent teriparatide clinical trials have included prospectively collected PINP serum samples within their study design. For example, in a global, randomized, double-blind comparator trial of teriparatide vs. alendronate in 203 postmenopausal women with osteoporosis, intact PINP was assessed at baseline, 1, 3, 6, and 12 months of treatment [35]. Figure 3 shows changes in PINP concentration and urinary N-terminal cross-linking telopeptide of type I collagen corrected for creatinine (NTX-I; biochemical marker of bone resorption) in patients treated with teriparatide [35]. Following initiation of teriparatide, PINP concentration increased rapidly, with an increase significantly greater than baseline, and then further increased to remain significantly above baseline at 6 months (p < 0.010) [35]. Yet, the resorption marker NTX-I did not increase at 1 month but did increase after 3 months of teriparatide therapy [35]. In contrast, PINP and NTX-I concentrations decreased markedly during alendronate treatment [35]. Similarly, in another active comparator study of 44 postmenopausal women, PINP increased in patients treated with teriparatide and decreased in patients treated with risedronate, as measured by total PINP assay [36]. In addition, another active comparator study of 79 postmenopausal women demonstrated marked increases in total PINP concentration from baseline in women treated with teriparatide, with slight decreases in PINP from baseline during strontium ranelate therapy (NCT00239629) [37]. These observed changes in serum PINP concentration are consistent with the anabolic mechanism of action of teriparatide.
Changes in PINP concentration were evaluated during a Phase 3 study of teriparatide conducted in Japanese men and postmenopausal women with osteoporosis at high risk for fracture (NCT00433160) [38]. This double-blind, placebo-controlled trial randomized patients 2:1 to teriparatide (N = 137) or placebo (N = 70) and included prospectively defined intact PINP assessments at baseline, 1, 3, 6, and 12 months of treatment. Fig. 4 shows the effects of teriparatide vs. placebo on median PINP concentrations in this study population [39]. These changes in serum PINP concentration were significantly different at 1, 3, 6, and 12 months (p < 0.001) [39]. The sharpest rise in PINP concentration occurred within the first month of teriparatide treatment [39].
Patients treated with teriparatide may have been treated previously with antiresorptive drugs and switched to teriparatide. In the EUROFORS study, 758 postmenopausal women with osteoporosis who were either treatment-naive, previously treated with an antiresorptive, or who had failed antiresorptive treatments were all given teriparatide (NCT00191425) [40, 41]. Samples for PINP assessment were prospectively collected at baseline, 1, and 6 months and analyzed by total PINP [41]. As expected, baseline PINP was significantly lower in the previous antiresorptive groups than the treatment-naive group. PINP increased from baseline in all treatment groups, with significantly greater increases noted in the treatment-naive group than the previous antiresorptive groups at 1 month. However, there were no significant differences in PINP concentrations between the treatment-naive and the previous antiresorptive groups at 6 months [40, 41]. Overall, teriparatide increased biochemical markers of bone formation in postmenopausal women with established osteoporosis, regardless of previous long-term exposure to antiresorptive therapies [40, 41].
PINP was also evaluated in a prospective, open-label, nonrandomized study of 59 postmenopausal women with osteoporosis previously treated with raloxifene or with alendronate for at least 18 months, and then switched to teriparatide [42]. Intact PINP assessments were prospectively collected at baseline, 1, 3, 6, and 12 months of teriparatide treatment. Consistent with alendronate being a more potent inhibitor of bone turnover than raloxifene, baseline PINP concentrations were lower in the alendronate group compared with the raloxifene group. During the treatment with teriparatide, statistically significant increases in PINP concentration from baseline occurred in both groups, but with a significantly smaller increase in the first month observed in the alendronate group than in the raloxifene group. No statistically significant difference between treatment groups occurred at subsequent time points [42]. This study included serial osteocalcin assessments. During teriparatide treatment, this marker shows changes similar to those of PINP.
In another study of patients with prior bisphosphonate exposure (alendronate, risedronate, etidronate, or pamidronate), total PINP was lower at baseline (p = 0.010), at 3 months, and at 6 months during teriparatide treatment in 38 patients with prior bisphosphonate exposure compared with 14 patients without prior bisphosphonate exposure [43]. These findings suggest that previous treatment with the potent antiresorptive drug alendronate appears to delay the PINP response to teriparatide, but with continued teriparatide treatment, PINP increases are observed [42, 43].
PINP response to teriparatide may be impacted differently by different bisphosphonates. For example, in the OPTAMIZE study, patients treated with alendronate or risedronate for at least 24 months were switched to teriparatide for 12 months of treatment. In the prior risedronate group, the PINP increase was significantly greater after 3 months of teriparatide than in the prior alendronate group (NCT00130403) [44].
An open-label study of 137 osteoporosis drug-naive postmenopausal women with osteoporosis, randomized to treatment with either teriparatide alone or to teriparatide plus raloxifene, evaluated intact PINP at baseline, 1, 3, and 6 months of treatment to determine the effect of the concomitant use of teriparatide with raloxifene on serum PINP concentration (NCT00046137) [45]. PINP significantly increased in both treatment groups similarly, demonstrating concomitant use of teriparatide with raloxifene resulted in a typical anabolic PINP response [45].
In regard to concomitant use of teriparatide with alendronate or with raloxifene, a prospective, open-label study of 198 postmenopausal women, with osteoporosis who were previously treated with raloxifene or alendronate for at least 18 months, randomized patients to add teriparatide and continue their antiresorptive drugs or to switch to teriparatide alone (“add” vs. “switch”) (NCT00079924) [46]. Total PINP assessments were prospectively collected at baseline, 1, 3, 6, 12, and 18 months of treatment. In all groups, PINP increased significantly from baseline during teriparatide treatment. However, the increases were smaller in the groups that added teriparatide to ongoing antiresorptive therapy compared with the groups that switched to teriparatide. In the alendronate stratum at 6 months, median PINP increases from baseline were 64 % vs. 401 % (p < 0.001) in the add vs. switch groups, respectively. In the raloxifene stratum at 6 months, median PINP increases from baseline were 131 % vs. 259 % (p < 0.001) in the add vs. switch groups, respectively [46]. These differences in PINP for add vs. switch groups demonstrate relatively greater bone formation with the switch approach, although the switch approach was also associated with greater bone resorption [46].
PINP monitoring may also find application in the management of patients with glucocorticoid-induced osteoporosis. For patients treated with chronic systemic glucocorticoids, the skeletal defect at the tissue level is impaired bone formation [47]. In a randomized, double-blind, active comparator trial of 428 men and women with glucocorticoid-induced osteoporosis (prednisone equivalent 5 mg/day or more for at least 3 months and BMD T-score at spine or hip of −2 or less, or −1 plus a history of a fragility fracture), intact PINP assessments were prospectively collected at baseline, 1, 6, and 18 months of treatment with teriparatide or alendronate (NCT00051558) [48–52]. In the teriparatide group, PINP concentration was significantly increased (p < 0.001) by 1 month and peaked at 6 months, with an observed median 69.8 % increase [48]. Notably, PINP concentrations remained above baseline during the 36 months of treatment [49].
From the same glucocorticoid-induced osteoporosis trial, PINP data for 83 male patients have been reported. Intact PINP serum concentration increased significantly during teriparatide treatment, peaked at 1 to 6 months, and remained above baseline at 18 months [51]. In another study of glucocorticoid-induced osteoporosis in 92 men, the typical expected increase in intact PINP during teriparatide treatment was observed, which was statistically significant as compared with risedronate (p < 0.001) [53]. These results confirm an increase in PINP can be expected in males with glucocorticoid-induced osteoporosis treated with teriparatide (NCT00503399) [53].
Also from the glucocorticoid-induced osteoporosis trial described above, PINP data for 67 premenopausal women have been reported (NCT00051558) [53]. During treatment with teriparatide, PINP increased from baseline at all time points, with peak increases observed at 1 to 6 months [53]. Similarly, increases in PINP were also noted in a study of 21 premenopausal women with unexplained fragility fractures or low BMD treated with teriparatide (NCT01440803) [54]. Serum PINP increased at 1 month, was 150 % above baseline at 6 months, and declined toward baseline at 18- and 24-month assessments [54]. These results confirm an increase in PINP can be expected in premenopausal females treated with teriparatide.
The early onset of PINP response to teriparatide was assessed in 15 osteopenic postmenopausal women treated with daily SC teriparatide 20 mcg/day [55]. Intact PINP concentrations increased by 8.2 % after 2 days of treatment and by 111 % after 28 days of teriparatide treatment, with all observed PINP increases significant compared with baseline (p < 0.0001) [55]. This study showed approximately similar increases in PINP, PICP, and osteocalcin during 28 days of teriparatide treatment, but showed a smaller increase in bone ALP [55]. Following cessation of treatment at 28 days, concentrations of bone formation markers decreased to within 20 % of baseline values by day 56 [55].
Relationship of PINP concentration to bone formation during teriparatide treatment
A Phase 4 clinical trial of ten postmenopausal women with osteoporosis treated with teriparatide showed a typical increase in intact PINP (NCT00259298) [56]. Increase from baseline in serum PINP concentrations at 3 and 18 months of treatment were correlated with an increase from baseline in skeletal uptake of radiopharmaceutical technetium-99 m methylene diphosphonate on a bone scan at the same visit (r = 0.60 at 3 months, r = 0.78 at 18 months) [56].
In an active comparator study of 69 postmenopausal women with osteoporosis, those patients treated with teriparatide again showed a typical increase in intact PINP (NCT00927186) [57]. Additionally, the correlation between 6-month PINP concentration and bone formation, defined by the mineralizing surface of trans-iliac bone biopsies, was high (r = 0.85), confirming PINP concentration is related to bone formation at the bone tissue level [57]. These data demonstrate PINP concentration is strongly related to skeletal bone formation as assessed by bone scan and by trans-iliac bone biopsy.
Relationship of early change in PINP concentration to other outcomes during teriparatide treatment
Several teriparatide studies have shown statistically significant correlations between an early change in PINP concentration and a later increase in BMD during teriparatide treatment [22, 23, 35, 39, 40, 57, 58]. In general, the best correlations were obtained between change in PINP concentration at 1 month and later percent increase in lumbar spine BMD. For example, the relationship between the 1-month change in PINP vs. 12-month percent change in lumbar spine BMD is shown in Fig. 5 [39]. Note that nearly complete separation of patients treated with teriparatide from patients treated with placebo was achieved with these biomarkers. Using Spearman correlation analysis, the highest correlation coefficient value between the bone turnover marker and BMD response was observed between the change in PINP concentration at 1-month and the 12-month lumbar spine BMD percent change (r = 0.76; p < 0.010) [39]. These data support early increases in bone formation during teriparatide treatment, reflected by increases in PINP concentration, predict later increases in BMD.
Teriparatide studies have shown statistically significant correlations between early PINP change and later percent increases in bone strength at the spine as determined by finite element analysis. In the teriparatide vs. alendronate active comparator study described above, postmenopausal women with osteoporosis were randomized to teriparatide 20 mcg/day or alendronate 10 mg/day in a double-blind fashion [35]. Quantitative computed tomography scans with finite element modeling of the L3 vertebra were performed in a subset of 21 teriparatide patients at baseline and 18 months [4]. Pearson correlation coefficients were calculated for log-transformed changes from baseline in biochemical markers of bone turnover at various times and compared with changes from baseline in vertebral strength parameters at 18 months. The best predictor of 18-month vertebral strength increases was PINP increments at 1 month in teriparatide-treated patients. The correlation coefficients for vertebral compressive stiffness and volumetric BMD were 0.45 and 0.51, respectively, with p values <0.05 [4, 23]. In another study, 1- and 3-month total PINP increases vs. baseline were similarly correlated with 18-month percent increase in spine strength as assessed by finite element analysis in four different groups (NCT 00079924) [46]. Additional analysis of the correlation between increases in PINP and bone strength was reported in males with glucocorticoid-induced osteoporosis [52]. In this study, changes in intact PINP at 3 months were correlated with change in finite element analysis strength increments in the spine to anterior bending (r = 0.422), axial compression (r = 0.516), and axial torsion (r = 0.496) assessed at 6 and 18 months of treatment [52]. These findings support a correlation between early changes in serum PINP and later increases in spine BMD and strength in patients treated with teriparatide.
Data to fully assess the relationship of early PINP change with fracture risk reduction are not available, since the large Phase 3 teriparatide fracture trial included PINP assessments in only a subset of the patients [22, 23]. However, the significant relationship between early change in PINP and later increase in BMD described above, along with the significant relationship between increase in BMD and fracture risk reduction during teriparatide treatment [24], provide indirect evidence for a relationship between PINP change and fracture risk reduction.
In the randomized, double-blind, active comparator trial of 428 men and women with glucocorticoid-induced osteoporosis described above, there was a significant reduction in vertebral fracture in the teriparatide vs. alendronate group [48, 49]. A subset of patients from both treatment groups in the trial had PINP assessments. The incidence of vertebral fractures in the subset of patients with PINP assessments was similar to the incidence of vertebral fractures in the overall study population (Fig 6). Because most patients in the teriparatide group had increases in PINP > 10 mcg/L and most patients in the alendronate group did not, the incidence of vertebral fractures in the subgroup of patients with vs. without PINP increase from baseline >10 mcg/L was also similar to the overall incidence (Fig. 6).
PINP response in individual patients
The intact PINP assay label references a least significant change (LSC) for PINP of 21 % [19, 59]. However, this 21 % definition of LSC for PINP during teriparatide treatment is problematic because many patients treated with teriparatide have been previously treated with potent antiresorptive drugs such as alendronate, and as noted above, these patients often have a relatively low baseline PINP concentration. For example, a typical patient treated with alendronate might have a baseline intact serum PINP concentration of 19 mcg/L. In such a patient, use of the 21 % cut point (LSC) would lead to the conclusion that an increase in PINP to 23 mcg/L (a 21 % increase from baseline) represents a significant increase in PINP even though this is obviously a small increment and below what would be considered a meaningful change. To overcome this difficulty, Eastell and colleagues established an absolute change cut point by using the typical mean PINP concentration of 48 mcg/L in untreated postmenopausal women [19] and multiplying the value by 0.21 (LSC) to arrive at a LSC of >10 mcg/L [17]. Absolute change in PINP is simple to calculate and is not subject to the percentage difficulties that arise in patients with very low baseline PINP concentrations [17]. Eastell and colleagues [17] proposed an algorithm for using the observed change in PINP serum concentration to monitor patients treated with teriparatide. This algorithm was tested and slightly modified by Krege and colleagues [60] and then evaluated in detail once again by Tsujimoto and colleagues [39]. An illustration of this simplified approach to using serum PINP concentration to monitor patients treated with teriparatide is provided in Fig. 7.
A prospective Phase 3, double-blind, placebo-controlled clinical trial included prospectively defined intact PINP analyses at baseline, 1, 3, 6, and 12 months to assess response rates in individuals treated with placebo or teriparatide [38, 39] (NCT00433160). In the placebo group, 2 out of 66 patients had PINP increases >10 mcg/L at 1 or 3 months, indicating a false-positive rate of 3 % when employing the monitoring algorithm. Among the 136 patients in the teriparatide 20 mcg/day group, 129 patients had PINP increases > 10 mcg/L at 1 or 3 months, and 7 patients did not. Among these seven patients, three were noncompliant and four were compliant. The absence of an increase in PINP > 10 mcg/L at 1 or 3 months in 4 out of 136 compliant patients indicates a false-negative rate of 3 % when using the monitoring algorithm [39]. Table 2 summarizes the findings from Tsujimoto and colleagues [39], defining a significant increase in PINP as an increase > 10 mcg/L at 1 or 3 months and a significant increase in lumbar spine BMD as an increase > 3 % at 12 months. These results demonstrate that clinically useful complementary information may potentially be gained from monitoring both PINP and BMD during osteoanabolic therapy.
Table 2 Response of PINP and BMD during osteoanabolic therapy [39]
Defining significant responses in this same manner, Table 3 presents potential patient observations and health care professional responses illustrating the potential use of PINP monitoring to enhance teriparatide therapy for individual patients.
Table 3 Hypothetical clinical scenarios, with potential clinical responses, and relevant data from Tsujimoto and colleagues [39]
PINP response during teriparatide therapy has been evaluated in additional clinical trials in which PINP monitoring was prospectively defined [7, 17, 56]. In drug-naive postmenopausal women with osteoporosis, the proportion of patients with increases in intact PINP > 10 mcg/L was 85 % at 1 month and 87 % at 3 months [17]. In postmenopausal women with osteoporosis previously treated with raloxifene for at least 18 months and then switched to teriparatide, the proportion of patients with increases in intact PINP > 10 mcg/L was 96 % at 1 month and was 92 % at 3 months [17]. Yet, in postmenopausal women with osteoporosis previously treated with alendronate for at least 18 months and then switched to teriparatide, the proportion of patients with increases in intact PINP > 10 mcg/L was 79 % at 1 month and was 97 % at 3 months [17]. The observed delay in the rise of serum PINP concentration for the teriparatide group previously treated with alendronate may be a result of the prior potent antiresorptive therapy.
PINP response to teriparatide following previous therapies, with concurrent therapies, or with concomitant illnesses
Clinical study data provide insight into the effect of prior osteoporosis drug treatment on PINP response during teriparatide treatment. For example, the EUROFORS study included three osteoporosis treatment cohorts: (1) treatment-naive patients, (2) patients pretreated with antiresorptive drugs, and (3) patients pretreated with antiresorptive drugs with inadequate clinical outcome [40, 41]. Total PINP responses > 10 mcg/L were similar in these groups [60]. Overall, the percentage of patients achieving a PINP increase > 10 mcg/L in the EUROFORS study was 83 % (613/736) at 1 month and 91 % (630/689) at 6 months [60]. Among the 110 patients with 1-month PINP increases ≤10 mcg/L, 88 (80 %) had 6-month PINP increases > 10 mcg/l [60].
PINP response to teriparatide in patients treated with other concomitant osteoporosis medications was evaluated in a recent clinical trial including patients previously treated for at least 18 months with alendronate or raloxifene (NCT000799924) [61]. These patients were randomized either to switch to teriparatide or to add teriparatide to ongoing alendronate or raloxifene [61]. Defining a PINP response as an intact PINP increment > 10 mcg/L at 1 or 3 months, the response rates (previously unpublished data) are shown in Table 4 [61].
Table 4 PINP response by treatment group in Cosman and colleagues [61]
A study of the combination of teriparatide plus denosumab has recently been published [62]. With this treatment combination, intact PINP does not appear to increase, but rather PINP shows a significant decrease from baseline. This study suggests that an increase in PINP should not be anticipated to occur with combination teriparatide plus denosumab treatment [62].
PINP response to teriparatide in patients with other chronic medical illnesses has been evaluated. PINP was collected in a randomized, double-blind trial of men and women with glucocorticoid-induced osteoporosis (NCT00051558) [47–49]. These individuals had various baseline conditions for which they were treated with glucocorticoids including rheumatoid arthritis, systemic lupus erythematosis, polymyalgia rheumatica, vasculitis, respiratory disorders, and inflammatory bowel disease. In these medically complex individuals treated with teriparatide, glucocorticoids, and a variety of other concomitant medications, the proportion of individuals with increases in intact PINP > 10 mcg/L at 1 or 6 months was 88 % [63].