Denosumab in Patients With Fibrous Dysplasia Previously Treated With Bisphosphonates

Bas C. J. Majoor; Socrates E. Papapoulos; P. D. Sander Dijkstra; Marta Fiocco; Neveen A. T. Hamdy; Natasha M. Appelman-Dijkstra


J Clin Endocrinol Metab. 2019;104(12):6069-6078. 

In This Article


Seven patients had polyostotic FD, four had MAS, and one patient had severe monostotic craniofacial disease (Table 1). Endocrine functions were stable for years in the four patients with MAS: two patients had precocious puberty and were not receiving any hormonal treatment at adult age; two had growth hormone excess and were receiving octreotide and had normal values of serum IGF1, one of whom also had hyperthyroidism and was receiving methimazole titration therapy and had normal serum TSH. All women included in the study were premenopausal.

All patients had been treated previously with bisphosphonates, predominantly olpadronate, for a median period of 8.8 years (range, 0.7 to 22.1). Other bisphosphonates used were pamidronate, zoledronate, alendronate, and risedronate (Table 2). Median SBS was 20.8 (0.3 to 64.7). Previous long-term treatment with bisphosphonates had led to temporary normalization of serum ALP activity or serum P1NP values in three patients, and temporary normalization of serum CTX levels in all patients.[12] Bisphosphonate treatment had also resulted in intermittent responses of pain complaints in all patients, but in none of the patients did this treatment result in disappearance of pain at any point during follow-up.

Patients were treated for at least one year with denosumab with a median follow-up of 15.5 months (range, 12 to 19 months). Median time between the last bisphosphonate treatment and first denosumab injection was 4.0 months (range, 0 to 103 months).

At the start of denosumab treatment, all 12 patients reported having pain at the site of their bone lesions. Prior to start of treatment with denosumab, 8 patients used a variety of analgesics, including paracetamol (n=8), nonsteroidal anti-inflammatory drugs (NSAIDs; n = 4), morphine derivatives (e.g., oxycodone; n = 5), and 1 patient used pregabalin. The 4 patients who did not use any analgesics did however have pain, but previous attempts to relieve the pain with analgesics were unsuccessful and they had stopped taking them. Serum values of biochemical markers of bone turnover varied markedly between patients at the time of starting treatment with denosumab (Table 1). Serum ALP activity was increased in 10 of 12 patients (median, 191 IU/L; range, 75 to 538 IU/L), serum P1NP in 12 of 12 patients (median, 241 ng/mL; range, 73 to 1235 ng/mL), and serum CTX in 3 of 12 patients (median, 375 pg/mL; range, 114 to 999 pg/mL). However, blood was collected in the nonfasting state, which can affect measured values of CTX, so that serum CTX levels should be interpreted with caution in this case series.[28] Serum calcium, adjusted for albumin, phosphate, creatinine and 25-OHD vitamin D concentrations, and plasma PTH levels were within their respective normal laboratory reference ranges in all studied patients. In contrast, serum FGF-23 values were increased in 9 of the 12 patients (Table 1).

Biochemical Response to Treatment

Initially, a 60 mg dose of denosumab was administered subcutaneously at six-month intervals in six patients but failed to maintain the decrease in serum BTMs observed at three months at the six-month evaluation time point (Figure 1). For example, serum ALP activity, although decreased at three months after the injection, returned to baseline values after six months, in contrast to the additional decrease observed in the patients who received a second denosumab injection after three months. Changes in serum P1NP levels were similar (Figure 1). Linear mixed model analysis (Table 3 and Table 4) showed a substantial difference over time for ALP (P = 0.007) and P1NP (P = 0.025), but not for CTX (P = 0.162), between patients who received denosumab at three-month intervals compared with those who received the agent at six-month intervals. The response of ALP and P1NP values did not differ between patients with FD/MAS and those with FD.

Figure 1.

(A–D) Percentage changes in serum ALP and P1NP levels with subcutaneous injections of 60 mg of denosumab (black arrows) in 12 patients with FD/MAS. Normal ranges: ALP 40 to 98 IU/L and P1NP <59 ng/mL. Percentage changes consistently show the different trend in changes in BTMs with denosumab given at 6-mo intervals compared with 3-mo intervals. Patients with 6-mo schedules clearly show an initial decrease in bone turnover after 3 mo, with values of ALP and P1NP returning to baseline after 6 mo. Patients with 3-mo schedules show a similar decrease after 3 mo, but further decrease of ALP and P1NP levels after 6-mo of treatment. Numbers in boxes correspond to individual patient numbers shown in Table 1.

These results indicated that the 6-month denosumab regimen used in the treatment of osteoporosis is inadequate for inducing a sustained reduction of biochemical markers of bone turnover in patients with FD and all patients were subsequently treated with a 3-month schedule of 60 mg of denosumab. In one patient (patient 7, Figure 2) with severe disease, serum ALP activity decreased from 538 IU/L reaching a plateau of 321 IU/L after 18 months of treatment with denosumab given at 6-month and later 3-month intervals. Continuation of treatment at a dose of 120 mg every 3 months resulted in a further decrease of serum ALP activity to 100 IU/L, a near-normal value. This patient was not included in the further analysis of the pharmacodynamic responses to denosumab treatment.

Figure 2.

Sequential changes in serum ALP levels in a patient with polyostotic FD (patient 7) during treatment with bisphosphonates (BPs, green boxes) and denosumab (DMab, blue arrows). The gray bar represents the normal laboratory reference range of serum ALP (40 to 98 IU/L). This patient was treated initially with denosumab at a 6-mo interval followed by three 3-mo injections and three 3-mo doses of 120 mg resulting in normalization of bone turnover and reduction in pain complaints. Note the improved response with subcutaneous injections of 60 mg of denosumab and the normalization of serum ALP levels with subsequent injections of 120 mg of denosumab.

In the 11 patients who received 60 mg of denosumab, serum ALP activity decreased from 212 ± 39 IU/L to 79 ± 6 IU/L (P = 0.004) at the end of the observation period (median, 15.5 months). Similarly, serum P1NP values decreased from 346 ± 111 ng/mL to 56 ± 17 ng/mL (P = 0.023) (Figure 3). Serum CTX decreased from 412 ± 251 pg/mL to 315 ± 318 pg/mL, a non-significant decrease (P = 0.403).

Figure 3.

Box plots of median and range of serum ALP (left panel, normal range 40 to 98 IU/L) and serum P1NP (right panel, normal range <59 ng/mL) levels of 11 patients with FD/MAS before and at the end of treatment with denosumab. The gray bars indicate normal ranges for serum ALP and P1NP values. *P = 0.004, **P = 0.023.

At the end of the observation period, of the 10 patients with increased serum ALP activity at baseline, values normalized in 7 (70%) (range final serum ALP: 48 to 113 IU/L), whereas serum P1NP levels that were increased in all patients reached the normal range in 9 of 12 (75%) (range final serum P1NP: 21 to 183 ng/mL). There were no statistically significant differences in response between patients with FD and those with FD/MAS, but numbers were small.

There was no correlation between pretreatment SBS and changes in biochemical parameters of bone turnover under treatment with denosumab. There was, however, a significant correlation between pretreatment and final values of serum ALP activity (Spearman ρ = 0.71, P = 0.015).

Clinical Response to Treatment

Ten of the 12 patients treated with denosumab reported improvement of pain, 6 of whom reporting complete disappearance of pain. In 3 of the 10 responders, pain decreased early after the first denosumab injection, in 4 it did not change after 3 months, and in 3 patients pain increased transiently (not requiring additional analgesics) at 3 months. Thereafter, pain progressively decreased in all. In 5 of the 6 patients who became pain-free this response was observed 6 to 12 months after the start of treatment. In the 1 patient (patient 7, Figure 2) although pain decreased substantially with treatment, it disappeared completely after 730 days concurrently with the administration of the highest denosumab dose. These results suggest that longer treatment periods may be needed for optimal clinical benefit in the majority of the patients. Of 8 patients who were using analgesics prior to start of denosumab treatment (paracetamol, n = 5; NSAIDs, n = 3; morphine, n = 2), 5 patients used a lesser dose and/or less frequent and 3 stopped their use. In these patients, reduction of pain symptoms was associated with increased mobility and function, although this was not quantified with validated questionnaires. In 1 patient with severe disease of the right lower limb, in addition to the reported reduction in pain, local tenderness and skin temperature were also reduced over the affected femur. In the patient with craniofacial FD (patient 5) pain levels remained unchanged after the first 3 months and increased after decompression surgery of the optic nerve. Although bone turnover normalized after treatment with denosumab in this patient, postsurgical headaches persisted. No other patient had a surgical procedure during the period of observation.

Adverse Events

Treatment was well tolerated and none of the patients developed symptomatic or asymptomatic hypocalcemia. One patient (patient 1) discontinued treatment after one year because of a skin rash thought to be possibly related to denosumab. She was started on bisphosphonates that could not, however, maintain the decreased level of BTMs achieved with denosumab although these did not increase above pretreatment levels nor were there any changes in serum calcium. In two patients with MAS (patients 10 and 12), treatment was associated with transient, asymptomatic, mild decrease in serum phosphate concentration to nadir values of 0.63 mmol/L (normal range, 0.8 to 1.5mmol/L), associated with an increase in intact PTH concentrations to 11.2 and 20.7 pmol/L, respectively, in the absence of changes in serum calcium. Both patients had a decreased tubular phosphate reabsorption in relation to glomerular filtration rate, 0.53 and 0.7, respectively, and increased FGF-23 levels prior to treatment with denosumab. Both were using active vitamin D metabolites. Serum phosphate and PTH concentrations normalized by increasing the dose of vitamin D supplementation and no phosphate supplements were required. There were no cases of osteonecrosis of the jaw or atypical femoral fractures during the period of observation.