2013 Volume 36 Issue 10 Pages 1622-1626
Denosumab, a fully human monoclonal antibody that inhibits the receptor activator of nuclear factor-κB ligand, inhibits the activation of osteoclasts. Some clinical trials have shown that denosumab suppresses bone resorption in patients with advanced cancer, but hypocalcemia has been reported as a serious adverse effect after the administration of denosumab. It is difficult to predict hypocalcemia in such cases because the risk factors for denosumab-induced hypocalcemia have not been reported. Accordingly, the aim of the present study was to identify the risk factors for hypocalcemia induced by denosumab. We retrospectively reviewed the records of patients who had received denosumab at Tokushima University Hospital between April 2012 and May 2013. Fifty-three patients were analyzed and eleven patients had hypocalcemia after administration of denosumab. Univariate logistic regression analysis revealed that the patients who had not been administered zoledronic acid before receiving denosumab or had lower creatinine clearance (CCr) appeared to have a higher risk of hypocalcemia (p<0.05). The cut off value of CCr was 50.4 mL/min calculated by receiver-operator characteristics curves. Moreover, multivariate logistic regression analysis revealed that non-administration of zoledronic acid (odds ratio 10.43, p<0.05) and CCr less than 50.0 mL/min (odds ratio 5.90, p<0.05) were independent risk factors for denosumab-induced hypocalcemia. These findings provide useful information regarding the monitoring of hypocalcemia in patients receiving denosumab.
Bone metastases are common in advanced cancer.1) Bone metastases lead to fractures and spinal compression due to the born resorption induced by activated osteoclasts.2) As a result, the quality of life (QOL) in patients with advanced cancer is decreased significantly. Therefore, prevention of bone resorption is important for maintaining QOL. Zoledronic acid, which inhibits farnesyl pyrophosphate synthetase in osteoclasts, has been used for preventing skeletal-related events caused by metastatic bone disease.3–5)
Recently, many studies have reported that the interaction between the receptor activator of nuclear factor-κB (RANK), which is expressed in osteoclasts, and the RANK ligand (RANKL), which is expressed in osteoblasts, is associated with the activation of osteoclasts.6) RANKL is the essential mediator of osteoclast differentiation, activation, and survival, and thereby a key contributor to bone resorption. In a murine model of established bone metastases, inhibition of RANKL was shown to prevent osteoclast-mediated bone resorption.7) Denosumab, a fully human monoclonal RANKL antibody, was shown to prevent the activation of osteoclasts via the inhibition of RANKL in a human model.8) Moreover, some clinical trials have shown that denosumab suppresses bone resorption more effectively than zoledronic acid in patients with advanced cancer.9–13) Therefore, denosumab can be considered a new drug which prevents skeletal-related events via a mechanism different from that of zoledronic acid.
Despite the promise of denosumab, hypocalcemia has been reported as a serious adverse effect of the drug. Indeed, studies have shown that the incidence of hypocalcemia induced by denosumab is higher than that induced by zoledronic acid.11–14) To prevent hypocalcemia, prophylactic administration of calcium and vitamin D is recommended concomitant with the administration of denosumab.15–17) However, it is difficult to predict the induction of hypocalcemia by denosumab because hypocalcemia is likely to occur even if denosumab is administered prophylactically. Therefore, identification of the risk factors related to this side effect would be clinically useful. In this study, therefore, we attempted to retrospectively identify the risk factors for hypocalcemia induced by denosumab.
Between April 2012 and May 2013, 68 patients received initial administration of denosumab at Tokushima University Hospital. Patients who had not undergone testing for serum calcium levels both before and after administration of denosumab were excluded, leaving 53 patients who were eligible for this study. This study was reviewed and approved by the Ethics Committee of Tokushima University Hospital.
Assessment of the Risk FactorsWe collected the data on gender, age, body weight, cancer type, serum creatinine level, serum albumin level, serum calcium level before and after administration of denosumab, the history of prophylactic administration of calcium and/or activated vitamin D3 and the history of administration of zoledronic acid. If the patients had been administered calcium and/or vitamin D as supplements before administration of denosumab, they were included in the prophylactic administration group. If the serum albumin level was less than 4.0 mg/dL, the adjusted serum calcium level was calculated according to the following equation: adjusted serum calcium level=serum calcium level (mg/dL) – serum albumin level (mg/dL)+4.0. Creatinine clearance (CCr) was calculated as described by Cockcroft and Gault.18) Hypocalcemia was defined as an adjusted serum level of less than 8.8 mg/dL, which is the lower limit of the standard at our facility. If the patients were administered denosumab at several times within the study period, the changes in the serum calcium level were observed at the first administration of denosumab to decide whether or not they had hypocalcemia. Based on this measure, the patients were divided into a hypocalcemia group (n=11) and a non-hypocalcemia group (n=42), and we compared the factors between the two groups to identify those associated with denosumab-induced hypocalcemia.
Statistical AnalysesFisher’s exact probability test, Mann–Whitney U-test and Student’s t-test were used to assess differences between the hypocalcemia group and non-hypocalcemia group. To investigate the risk factors for denosumab-induced hypocalcemia, univariate and multivariate logistic regression analyses were performed. We calculated the area under receiver-operator characteristics (ROC) curves to estimate sensitivity, specificity and cut off values for factors obtained by univariate logistic regression analysis. In the multivariate logistic regression analysis, the forced entry method was employed using the factors that were significantly different in the univariate logistic regression analysis. All analyses were done using excel (Microsoft). All recorded p values were two sided, and values of less than 0.05 were considered statistically significant.
Table 1 shows the patient characteristics in this study. Among the 53 patients, 18 patients had lung cancer, 10 patients had breast cancer, 8 patients had prostate cancer, 6 patients had kidney cancer and 11 patients had other cancers. Thirty-four patients had received the prophylactic administration of calcium and/or activated vitamin D3 and 1 patient had received calcium and vitamin D as supplement. Twenty-one patients had received the administration of zoledronic acid before receiving denosumab. Of these patients, 19 patients had received denosumab within 3–4 weeks and 2 patients had received within 1–2 months after zoledronic acid therapy. Thirty-nine patients had normal renal function (≧50 mL/min), 10 patients had moderate renal failure (30 mL/min≦CCr<50 mL/min), and 4 patients had severe renal failure (<30 mL/min). No patients had hypocalcemia before denosumab treatment, and 11 patients had hypocalcemia after receiving denosumab.
| Characteristics | Means±S.D. or number of patients |
|---|---|
| Total number of patients | 53 |
| Gender | |
| Male | 34 |
| Female | 19 |
| Age | 66.2±10.6 |
| (range) | (39–86) |
| Body weight | 55.1±8.7 |
| Serum albumin | 3.4±0.7 |
| Cancer type | |
| Lung cancer | 18 |
| Breast cancer | 10 |
| Prostate cancer | 8 |
| Kidney cancer | 6 |
| Others | 11 |
| Prophylactic administration | |
| Ca and vitamin D3 | 26 |
| Only Ca | 5 |
| Only vitamin D3 | 3 |
| Supplement | 1 |
| Pretreatment with zoledronic acid | 21 |
| CCr | 69.0±27.0 |
| ≧50 mL/min | 39 |
| 30≦CCr<50 | 10 |
| <30 | 4 |
| Hypocalcemia | 11 |
In Table 2, the patient characteristics are compared between the hypocalcemia group and non-hypocalcemia group. The percentage of patients in the hypocalcemia group who had received zoledronic acid before administration of denosumab was significantly smaller than that in the non-hypocalcemia group (p<0.05). In addition, there was a significant difference in CCr between the two groups (p<0.05). No significant differences were found between the two groups with regard to the other factors investigated. Moreover, univariate logistic regression analysis revealed that there was a significant correlation between the incidence of hypocalcemia and the percentage of patients who had received zoledronic acid before administration of denosumab and CCr (Table 3). To estimate the prediction ability of denosumab-induced hypocalcemia, the area under the ROC curves (AUC) obtained by univariate logistic regression analysis of CCr was calculated. The AUC was 0.72, which was evaluated as having moderate accuracy for detecting some events (Fig. 1). The cut off value of CCr calculated by the ROC curve was 50.4 mL/min. When using this cut off value, the sensitivity and specificity of CCr to detect the incidence of hypocalcemia after denosumab administration were 63.6% and 81.0%, respectively.
| Factor | Non-hypocalcemia | Hypocalcemia | p Value |
|---|---|---|---|
| Means±S.D. or number of patients | Means±S.D. or number of patients | ||
| Male/Female | 29/13 | 5/6 | 0.17a) |
| Age | 65.2±10.4 | 69.8±11.1 | 0.13b) |
| Body weight | 56.0±8.8 | 51.7±6.9 | 0.15c) |
| Serum albumin | 3.3±0.7 | 3.6±0.5 | 0.13c) |
| Cancer type | |||
| Lung cancer | 16 | 2 | 0.29a) |
| Breast cancer | 8 | 2 | 1.00a) |
| Prostate cancer | 6 | 2 | 0.67a) |
| Kidney cancer | 5 | 1 | 1.00a) |
| other cancer | 7 | 4 | 0.21a) |
| Serum calcium level before administration of denosumab | 9.9±1.1 | 9.6±1.2 | 0.41c) |
| Prophylactic administration | 28 | 7 | 1.00a) |
| Pretreatment with zoledronic acid | 20 | 1 | 0.04a) |
| CCr | 73.2±28.1 | 53.3±15.8 | 0.03c) |
a) Fisher’s exact probability test. b) Mann–Whitney U-test. c) Student’s t-test.
| Factor | Odds ratio | 95% CI | p Value |
|---|---|---|---|
| Male | 0.37 | 0.096–1.449 | 0.15 |
| Age | 1.05 | 0.976–1.121 | 0.21 |
| Body weight | 0.94 | 0.859–1.023 | 0.15 |
| Serum albumin | 2.49 | 0.751–8.236 | 0.14 |
| Cancer type | |||
| Lung cancer | 0.36 | 0.069–1.888 | 0.23 |
| Breast cancer | 0.81 | 0.149–4.461 | 0.81 |
| Prostate cancer | 1.33 | 0.230–7.743 | 0.75 |
| Kidney cancer | 0.74 | 0.077–7.076 | 0.79 |
| Other cancer | 2.86 | 0.655–12.455 | 0.16 |
| Serum calcium level before administration of denosumab | 0.70 | 0.298–1.649 | 0.16 |
| Prophylactic administration | 0.88 | 0.219–3.500 | 0.85 |
| No pretreatment with zoledronic acid | 9.09 | 1.066–77.497 | 0.04* |
| CCr (per 1 mL/min decrease) | 1.04 | 1.002–1.061 | 0.04* |
* Included in multivariate logistic regression analysis
We carried out a multivariate logistic regression analysis using the factors of the history of administration of zoledronic acid before receiving denosumab and CCr less than 50.0 mL/min. The results of the multivariate logistic regression analysis are shown in Table 4. Multivariate logistic regression analysis suggested that the patients who had not received zoledronic acid before administration of denosumab (odds ratio 10.43, 95% confidence interval 1.12–97.26; p<0.05) or who had CCr less than 50 mL/min (odds ratio 5.90, 95% confidence interval 1.24–27.99; p<0.05) had a higher risk of hypocalcemia induced by denosumab. Therefore, we chose the non-administration of zoledronic acid and CCr less than 50 mL/min as independent risk factors for hypocalcemia induced by denosumab.
| Factor | Odds ratio | 95% CI | p Value |
|---|---|---|---|
| No pretreatment of zoledronic acid | 10.43 | 1.118–97.259 | 0.039 |
| CCr less than 50 mL/min | 5.90 | 1.243–27.990 | 0.026 |
In our study, 11 patients had hypocalcemia induced by denosumab. The incidence of hypocalcemia reported in this study (20.8%) seems to be much higher than those reported earlier.9–13) This may be due to apparently higher lower limit of normal serum calcium level at our facility (8.8 mg/dL) than those (8.4–8.5 mg/dL) reported elsewhere to define hypocalcemia. It is clear, however, that hypocalcemia was induced by denosumab in our study because no patients had hypocalcemia before denosumab treatment. We confirmed only 1 patient developed grade 3 hypocalcemia and the others developed grade 1 in our study. Further analyses with more patients who had grade 3 or 4 hypocalcemia will be needed to identify the risk factors associated with much severe hypocalcemia induced by denosumab.
To identify the risk factors for hypocalcemia, we first compared the differences between the hypocalcemia group and non-hypocalcemia group. The percentages of patients who were administered zoledronic acid before administration of denosumab and the CCr values were significantly different between these two groups, though no significant differences were found in regard to the other factors investigated. It has been reported that the prophylactic administration of calcium (≧500 mg) and vitamin D (≧400 IU) decreased the risk of hypocalcemia induction by denosumab.15–17) The average dose of calcium was 197 mg in the hypocalcemia group and 293 mg in the non-hypocalcaemia group in this study and there was no significant difference between the two groups (Student’s t-test, p=0.19). Underdosing of calcium may be the reason why no significant differences were observed between the two groups in the percentage of patients who received the prophylactic administration of calcium and/or activated vitamin D3.
Multivariate logistic regression analysis revealed that the patients who had not received zoledronic acid before administration of denosumab appeared to have a higher risk of hypocalcemia induced by denosumab. This may have been related to the secretion of parathyroid hormone (PTH). PTH increases the concentration of serum calcium by promoting the activation of vitamin D and secretion of inorganic phosphate from the kidneys.19) It is known that the secretion of PTH is increased after administration of zoledronic acid and the increases of PTH concentration are continued for about 1–2 months.3,20,21) In this study, denosumab was administered within 2 months after administration of zoledronic acid in the zoledronic acid pretreatment group. Therefore, we considered that the PTH concentration in the patients of the zoledronic acid pretreatment group were likely increased. Such increases in the PTH concentration could have prevented the denosumab-induced decreases in the serum calcium level. However, we could not verify this hypothesis because we could not obtain laboratory data on the PTH concentration, as this test was not routinely performed in the study subjects. A prospective study will be needed to elucidate the relationship between the incidence of hypocalcemia induced by denosumab and the PTH concentration.
Multivariate logistic regression analysis also revealed that the patients who had CCr less than 50 mL/min appeared to have a higher risk of denosumab-induced hypoclcemia. There was a tendency for the serum calcium level of the patients with renal failure to be decreased, because the absorption of calcium in the enteron was reduced due to the inactivation of vitamin D.19) We should therefor pay special attention to hypocalcemia when denosumab is administered to zoledronic acid-naive patients with renal failure.
Although we did not clarify the other factors associated with hypocalcemia induced by denosumab, factors affecting the secretion of PTH may be risk factors, since PTH is an important hormone for maintaining serum calcium levels. For example, radiation to the thyroid gland or hypoparathyroidism may be risk factors of hypocalcemia induced by denosumab. Moreover, it is known that glucocorticoid increases the loss of calcium from the kidneys. Drugs that affect the serum calcium level may thus be related to the induction of hypocalcemia by denosumab.
In conclusion, the present study showed that the non-administration of zoledronic acid and CCr less than 50 mL/min were independent risk factors for the induction of hypocalcemia by denosumab. Though further prospective analyses with more patients will be needed to collect further evidence, the present work should provide useful information regarding the monitoring of hypocalcemia in patients receiving denosumab.
