Introduction

Duchenne muscular dystrophy (DMD) is an X-linked disorder characterized by progressive muscle weakness, leading to loss of independent ambulation by age 12 years, and mortality from cardiorespiratory failure by early adulthood. The use of long-term glucocorticoid (GC) therapy prolongs ambulation, improves cardiac and pulmonary function, delays onset of scoliosis and contractures, and preserves upper extremity function [1,2,3]. However, long-term GC therapy has significant adverse effects, including osteoporosis, delayed puberty/hypogonadism, increased weight gain, and growth failure [1, 3, 4].

Osteoporosis is a common and important comorbidity in patients with DMD [4, 5]. Vertebral and long bone fractures are frequent following the introduction of GC therapy [2]. Apart from direct toxicity of GC on bone, delayed puberty/hypogonadism and disruption of the bone-muscle unit that results from muscle weakness are also detrimental to bone health in patients with DMD [2]. Unlike other groups of patients with GC-induced osteoporosis (GIO), such as those with acute leukemia or rheumatologic disorders, spontaneous recovery of vertebral fractures and vertebral reshaping are not typically observed in patients with DMD, probably due to the ongoing need for GC therapy and limited bone growth [6]. Pharmacologic treatment for GIO in this population is therefore indicated to prevent osteoporosis and its complications, which include premature loss of ambulation (a psychologically devastating milestone), pain (further impacting quality of life), or fat embolism (a potentially life-threatening event).

The off-label use of bisphosphonates (BP) has become standard practice for the treatment of GIO in patients with DMD, although there is no clear evidence on which to base consensus as to type, dose, and duration of therapy in this population [1, 7]. Both oral and intravenous BP have been reported to be associated with lower fracture frequency and increased bone mineral density (BMD) in patients with DMD [8,9,10]. However, even with the combination of vitamin D and calcium, BP therapy is frequently inadequate to prevent fractures [2]. In adult (non-DMD) patients with GIO, current treatment recommendations include supplementation with vitamin D and calcium, in addition to other therapeutic agents, including BP and parathyroid hormone (PTH) [11,12,13]. There is an urgent need to find therapeutic agents that are effective to treat or prevent osteoporosis in pediatric/adolescent patients with DMD.

Recombinant human PTH (rhPTH) (1–34), teriparatide, is an anabolic agent that promotes the formation of bone by increasing the number of osteoblasts, activating pre-existing osteoblasts, and decreasing apoptosis in osteoblasts [14]. It intuitively better addresses the underlying pathophysiology of GIO and has been shown to be superior to anti-resorptive agents such as BP to increase BMD and reduce fracture risk [15]. The predominant effects of teriparatide therapy are seen in areas of high bone turnover, mainly cancellous bone (such as the spine), which are preferentially affected by GC, with improvement in skeletal microarchitecture [16]. Several pivotal studies of teriparatide therapy in adults with osteoporosis, including GIO, have shown it to be safe and effective, with significant improvement in BMD and reduction of fracture risk [14, 17,18,19]. Teriparatide has been shown to be safe and effective when used as replacement therapy for children with hypoparathyroidism [20,21,22]. However, it has neither been used nor studied in children and adolescents for the treatment of osteoporosis because of concern for the potential for osteosarcoma based on observations in animal studies [23, 24]. There has been a single case report of teriparatide therapy for osteoporosis in a 20-year-old man with DMD, in whom 18 months of treatment was associated with decreased bone pain, increased BMD, and improved quality of life [25].

We used teriparatide to treat 6 adolescent/young adult patients with DMD and severe osteoporosis, on long-term GC, and who continued to sustain fragility fractures despite treatment with BP, or who had an intolerance to BP. The aims of this study were to evaluate the safety and efficacy of teriparatide for treatment of severe osteoporosis in patients with DMD.

Materials and methods

Study participants

Patients were considered eligible for the study by the following inclusion criteria: (i) a diagnosis of DMD, confirmed by genetic mutation analysis and/ or muscle biopsy; (ii) long-term and concurrent GC therapy of at least 1-year duration; (iii) diagnosis of osteoporosis, with a vertebral compression fracture, or height-adjusted lumbar spine (LS) BMD Z-score < − 2.0 and pathological long bone fractures [26], with progression or continuing to fracture despite being on BP therapy, or intolerance to BP; and (iv) estimated glomerular filtration rate (eGFR) > 49 mL/min, as measured by cystatin C [27]. Patients were excluded by the following criteria: (i) hypercalcemia, defined by serum calcium concentration (corrected for serum albumin) exceeding 11 mg/dL; (ii) nephrocalcinosis or nephrolithiasis (confirmed by ultrasound), with hypercalciuria, as defined by 24-h urine calcium excretion exceeding 4 mg/kg (as teriparatide treatment for osteoporosis could cause hypercalcemia/hypercalciuria); (iii) concurrent digoxin, hydrochlorothiazide, or furosemide treatment; or (iv) inability to follow up or comply with recommended teriparatide treatment.

Eligible patients were referred to Endocrinology by the Comprehensive Neuromuscular Care Center at Cincinnati Children’s Hospital Medical Center (CCHMC). Patient information, laboratory, and radiographic data were accessed by review of the electronic medical records. Informed consent and assent were obtained at study enrollment. The study was approved by the CCHMC Institutional Review Board.

Treatment protocol

Bisphosphonate therapy was discontinued at the time of teriparatide start, as the literature suggested that the simultaneous combination of BP and teriparatide may be detrimental to bone health [28]. Teriparatide was administered as a once-daily subcutaneous injection of 20 mcg, with the intent to treat for 2 years. Patients were seen at teriparatide initiation and at 6-monthly intervals over 2 years. Patients were assessed to ensure their calcium intake by diet and supplementation met required daily amounts for age, and those with known hypercalciuria discontinued or received lower supplementation to meet their needs but prevent hypercalcemia and hypercalciuria [16]. Vitamin D supplementation was continued, to maintain serum 25-hydroxyvitamin D concentrations of 30–60 ng/mL. Other therapies that could have affected bone health, including GC, and testosterone, remained constant, unless determined by clinical need.

Bone health monitoring

Fractures and vertebral morphometry

Fracture history was reviewed and confirmed radiographically during each clinic visit. Long bone fracture rates were calculated pre- and on 2 years of teriparatide therapy. The group fracture rate was assessed by the total number of long bone fractures divided by the sum of the times from the first fracture to the start of teriparatide for each patient. All patients had radiographic assessment of thoracic and lumbar vertebrae at baseline and every 6 months during teriparatide therapy, and if symptomatic. Morphometric measurement of vertebrae on lateral spine radiographs using 6-point vertebral morphometry, as previously described [29], was performed by J.C.T. and P.H. Baseline and follow-up anterior, middle, and posterior vertebral body heights were compared to document response to therapy. Vertebral height ratios were assessed by the Genant semi-quantitative method: anterior-to-posterior heights, middle-to-posterior heights, and posterior-to-upper and lower adjacent posterior vertebral heights [30]. Vertebral height ratio losses were graded as follows: grade 0 (normal), < 15%; grade 0.5 fracture (uncertain or questionable), ≥ 15–< 20%; grade 1 (mild) fracture, ≥ 20–< 25%; grade 2 (moderate) fracture, ≥ 25–< 40%; grade 3 (severe) fracture, ≥ 40%.

Bone mineral density and content

All patients had dual X-ray absorptiometry (DXA) scans at baseline and at 6- to 12-month intervals to assess areal bone mineral density (aBMD) and bone mineral content (BMC) of the LS, whole body (WB), and lateral distal femur (LDF). DXA scans were performed using a Hologic densitometer that had software with the same bone detection algorithms (version 12.3) for the analysis of all scans. The coefficient of variation at our center was < 1% at LS and WB sites. For the LDF site, three regions of interest (R1, primarily trabecular bone; R2, a transitional zone; and R3, primarily cortical bone) were identified [31]. Bone mineral apparent density (BMAD) of the lumbar spine was calculated using methods previously described [32, 33]. The aBMD Z-scores were calculated from pediatric reference data [34, 35]. As accurate height measurements were not possible in the majority of these patients, aBMD Z-scores were not corrected for height.

Bone formation markers

Bone-specific alkaline phosphatase and procollagen type 1 amino-terminal propeptide (P1NP) were obtained as morning fasting samples at the start of teriparatide treatment, and at 6, 12, 18 and 24 months of therapy.

Safety monitoring

Laboratory testing was performed at the start of treatment, and at 1, 3, 6, 12, 18, and 24 months after initiation of therapy, and included intact PTH, renal panel, 25-hydroxyvitamin D, eGFR as measured by cystatin C, 24-h urine Litholink studies (Litholink Corporation, Chicago, IL, USA), urine calcium, and urine osmolality.

Statistical analysis

Data were analyzed using SAS®, version 9.4 (SAS Institute, Cary, NC). Continuous data were presented as means ± SD and ranges (minimum-maximum), and categorical data were presented as percentages. Due to the small sample size, statistical analyses were not performed on changes over time for the laboratory, vertebral morphometry, or DXA data. Inter- and intra-observer agreement of vertebral height measurements was assessed using an intraclass correlation estimated from the general linear mixed model. Landis and Koch cut-points were used to define the intraclass correlations, with values between 0.81 and 1.0 being excellent agreement [36].

Results

Baseline evaluation

Six patients were eligible and enrolled in the study (Table 1). All patients were treated with long-term GC (mean 28.3 ± 5.4 mg per day of prednisone-equivalent) and had a history of long bone and vertebral fractures, and none reported bone pain or back pain at baseline. All patients had been previously started on oral BP therapy when they had first developed early signs of bone fragility, such as mild vertebral height loss or compression deformities, in accordance with our center’s bone health protocol. All had continued to sustain fragility fractures despite treatment with vitamin D, calcium supplements, and oral BP, except patient 2 who was not on oral BP due to intolerance to treatment (swelling of lips). All patients had sufficient 25-hydroxyvitamin D concentrations (30–60 ng/mL), normal serum calcium concentrations, and renal function (Table 2). Patients 3 and 6 had hypercalciuria (urine calcium/osmolality ratio > 0.25 and/or 24-h urine calcium excretion > 4 mg/kg), but no ultrasonographic evidence of nephrocalcinosis or nephrolithiasis at baseline. Their vitamin D and calcium supplements were decreased or discontinued prior to the start of teriparatide. All patients who had delayed puberty/hypogonadism were treated with testosterone therapy, and all had appropriate serum testosterone concentrations for age during the study period.

Table 1 Baseline characteristics of 6 patients with DMD
Full size table
Table 2 Laboratory safety indices, bone formation markers, and urine studies at baseline and on teriparatide therapy
Full size table

All patients completed a 2-year course of teriparatide treatment, except for patient 4, who died following a cardiac arrest at 1 year post-initiation of treatment. The cardiac event was attributed to congestive heart failure, unrelated to teriparatide, as the patient had baseline severe cardiomyopathy, a common comorbidity of DMD.

Bone health

Long bone fractures

The 6 patients had a total of 17 long bone fractures, over a 20.3 patient-year period prior to the initiation of teriparatide therapy (0.84 fractures per year). The time from the last fracture to teriparatide start ranged from 0.5 to 3.5 years (Table 1). The lower extremities were the most common sites of fractures (82%). Patient 3, who was close to losing ambulation, sustained another lower extremity long bone fracture following a fall injury at 6 months of therapy, resulting in permanent loss of ambulation. Overall, on teriparatide therapy, the long bone fracture rate appeared to decrease to 0.09 fractures per year (1 fracture over an 11.0 patient-year period).

Vertebral morphometry and fractures

The intraclass correlation coefficients for intra-observer measurements of vertebral heights were 0.88–0.93, and those for inter-observer measurements were 0.81–0.88, namely excellent. Lumbar vertebral heights were measured in all 6 patients. Due to severe thoracic kyphoscoliosis being prevalent in some of our patients, as well as severe osteoporosis, accurate measurement of some thoracic vertebral heights was not possible. There appeared to be no major changes in median lumbar vertebral heights over 2 years of teriparatide treatment (Fig. 1). Genant semi-quantitative assessment showed stable or improved fractures in most T9-L4 vertebrae during teriparatide therapy (Fig. 2). No bone or back pain was reported at baseline or during treatment.

Fig. 1
figure 1

Box and whisker plots of the anterior, middle, and posterior lumbar vertebral heights (L1–L4) at baseline and 1 and 2 years on teriparatide

Full size image
Fig. 2
figure 2

Changes in vertebral fractures (T9–L4) assessed by the Genant semi-quantitative method at baseline and 2 years* on teriparatide therapy. *Comparison of baseline and 1 year of teriparatide therapy for patient 4

Full size image

Bone density

The absolute aBMD and BMC values by DXA at all skeletal sites and LS BMAD appeared to remain stable, while aBMD Z-scores appeared to continue to decline over time (Fig. 3).

Fig. 3
figure 3

BMC, BMD, BMD Z-score, and BMAD trajectories of individual patients on teriparatide therapy

Full size image

Bone formation markers

While serum bone-specific alkaline phosphatase remained relatively stable, there were modest increases in P1NP concentrations at 2 years of teriparatide treatment in all patients (Table 2).

Safety indices

Mean serum PTH, calcium, phosphorus, magnesium, and cystatin C eGFR remained within normal/acceptable ranges throughout the course of treatment (Table 2). All patients had 24-h urine studies at baseline, but only patients 1 and 3 completed these at 2 years of teriparatide therapy. Mean 24-h urine calcium excretion and urine calcium/osmolality ratios remained stable in all patients. The 24-h urine oxalate and urine supersaturation-calcium oxalate was stable in patient 1 but increased over time in patient 3. Patient 6, who had hypercalciuria at baseline, developed a non-obstructive renal calculus at 22 months of teriparatide therapy.

Discontinuation of teriparatide

All except patient 4 completed a 2-year course of teriparatide therapy. Following discontinuation of teriparatide, patients 3, 5, and 6 transitioned to treatment with an oral or intravenous BP. Patient 1 switched therapy to denosumab. Patient 2 was continued on testosterone therapy only by his non-CCHMC endocrinologist.

Discussion

The key findings from our study suggest that teriparatide therapy for the treatment of GIO in adolescent and young adult patients with DMD and severe osteoporosis is associated with stable bone health and no major safety concerns. We evaluated bone health in a comprehensive manner, which included the assessment of vertebral morphometry by 6-point measurement (novel methodology in this patient population) and Genant semi-quantitative grading, in addition to long bone fracture rate and DXA indices of multiple skeletal sites. We found that vertebral morphometry and fractures were relatively stable, in keeping with lack of new vertebral fractures described in a prior case report of an adult patient with DMD treated with teriparatide [25]. Long bone fracture rate appeared to decrease over the 2 years of teriparatide treatment, and we observed a modest increase in P1NP.

We did not observe an increase in BMD while on teriparatide treatment, in contrast to gains seen in previous studies in adults with GIO [19, 37]. In our cohort, absolute BMD values at all skeletal sites appeared to remain unchanged. While there was a continuing (unchanged) trend of declining BMD Z-scores, this was not surprising as our patients with DMD were compared with standards for healthy males who undergo puberty with accelerated growth, and who continue to accrue bone mineral during adolescence. Adolescent patients with DMD are subject to GC-induced growth retardation and delayed puberty/hypogonadism, which could account for the observed decline of DXA indices that have been standardized for age and sex, but not height or pubertal stage.

In adults with GIO, teriparatide has been shown to be superior to BP with regard to effects on BMD, bone microarchitecture, and vertebral fracture outcomes [19, 37]. There are several possible reasons why we may not have observed similar gains. GC doses and durations reported in these studies (prednisone 7.5–8.8 mg daily for 2.3–6.4 years) were much lower compared with the treatment regimens in our patients with DMD. The adult patient with DMD (who was BP-naïve) treated with teriparatide in the previously cited case report was also on a lower GC dose (prednisone 5 mg daily) when compared with our patients [25]. Thus, differences in GC regimens and cumulative dosing could have been responsible for an attenuated response to teriparatide therapy in our patients, who remained on high doses of GC. Another possible contributing factor could have been prior long-term BP exposure that may have blunted the extent of bone remodeling from teriparatide [38, 39], although there are conflicting data on this theory. Several studies have shown the suppressive effects of BP can be overcome during teriparatide therapy, suggesting that a washout period is not needed [40,41,42], and newer studies have reported combined BP and teriparatide therapy may be beneficial [43,44,45]. Additional factors could be the severity of osteoporosis in our cohort (who represented the more severe extreme of the osteoporosis spectrum of our DMD patient population as a whole), and the presence and persistence of other risk factors (namely immobility, lack of weight-bearing ability, and severe muscle weakness due to advanced muscular dystrophy). These factors may also explain the smaller increases in bone formation markers in our cohort, in contrast to more significant increases in P1NP, osteocalcin, and bone-specific alkaline phosphatase in adults with GIO who were treated with teriparatide [19]. Lastly, teriparatide dose may also play a role in BMD response. Although the teriparatide dose of 20 mcg daily that we used is considered standard dosing, recent studies in postmenopausal women suggested that high-dose teriparatide (40 mcg daily) was associated with greater BMD gain, but without benefit for fracture outcomes [46, 47]. Further studies are needed to evaluate whether patients with high-risk osteoporosis, particularly those with DMD, would benefit from a higher dose of teriparatide.

We found that teriparatide was safe and well tolerated in adolescent and young adult patients with DMD. Hypercalciuria is a commonly reported side effect of teriparatide therapy for treatment of osteoporosis in adults and hypoparathyroidism in children, as a result of transient hypercalcemia following intermittent administration [22, 48]. We observed stable serum calcium concentrations and urinary calcium excretion (24-h urine calcium and urine calcium/osmolality ratio) during treatment with teriparatide. An increase in 24-h urine oxalate and urine supersaturation-calcium oxalate, despite stable urine calcium, was observed in one patient. This may have been due to a decrease in his dietary and supplemental calcium intake, as recommended after the finding of hypercalciuria prior to teriparatide start. A decrease in dietary calcium to bind with oxalate could lead to more free intestinal oxalate being absorbed and subsequently excreted in the urine. Finally, although one patient developed nephrolithiasis, hypercalciuria was already present at baseline despite appropriate calcium and vitamin D intake. Hypercalciuria is common in patients with DMD, conferring risk for nephrocalcinosis even without teriparatide treatment [7].

Another common concern for use of teriparatide treatment in children is the theoretical risk for the development of osteosarcoma, based on studies in rodents [49]. However, teriparatide doses and durations in these studies were much greater relative to those in humans. No patients in our cohort were known to develop osteosarcoma on short-term teriparatide therapy, including in 4 patients who had open epiphyses at baseline (8 patient-years of follow-up), in line with post-marketing surveillance studies that reported no apparent connection between osteosarcoma and teriparatide use in adults [50, 51]. Findings from long-term studies of children with hypoparathyroidism also support the safe use of teriparatide in children [22].

Strengths of our study include novel treatment and comprehensive bone health assessment, including vertebral morphometry, in this patient population. Vertebral reshaping is typically overlooked when monitoring treatment for osteoporosis in children [7]. We evaluated vertebral morphometry by using both vertebral heights and the more commonly used Genant semi-quantitative method. Vertebral height measurements allow for more in-depth comparisons of continuous variables, compared with categorical comparisons using the Genant method. However, none of these measures assesses bone quality or microarchitecture. Vertebral morphometry evaluation may also be subjective. Nevertheless, our intra- and inter-observer correlation coefficients were excellent, all above 0.81. Another point meriting consideration is the limitation of BMD use in the DMD population itself. BMD Z-scores presented in this study were not corrected for height (due to most patients being non-ambulatory) or puberty, and presence of lumbar vertebral fractures may falsely increase LS BMD. However, trajectories for absolute BMD values during treatment still serve as a useful tool for monitoring treatment response, as long as these caveats are taken into consideration when interpreting the data. Our study was also limited by the small number of patients in this cohort. Lastly, compliance with teriparatide treatment relied on patients’ reports.

In conclusion, in our 2-year clinical experience of teriparatide treatment of 6 patients with DMD, we observed stable bone health, with modest increases in P1NP, and no safety concerns. There is a rationale for the use of anabolic agents such as teriparatide to treat GIO, and our findings will help inform the next steps in the quest to find better therapies to treat this challenging patient population. Larger intervention studies in patients with DMD not previously treated with long-term BP, studies using higher doses of teriparatide, and/or controlled trials comparing teriparatide with conventional BP treatment or newer anabolic agents, or combination therapies, are needed to better understand the efficacy of teriparatide for treatment of GIO in youth with DMD.