Serum 25-hydroxyvitamin D levels modulate the acute-phase response associated with the first nitrogen-containing bisphosphonate infusion

Abstract

The acute-phase response (APR) is the most frequent side effect after the first dose of intravenous nitrogen-containing bisphosphonates (N-BPs). It has been demonstrated in vitro that N-BPs stimulate γδ T-cell proliferation and production of cytokines and that vitamin D is able to modulate them. Therefore, we have studied the relationship between bone metabolism parameters, particularly for 25-hydroxyvitamin D [25(OH)D], and APR in patients treated with 5 mg zoledronic acid intravenously. Ninety N-BP-naive osteoporotic women (63.7 ± 10.6 years of age) were stratified for the occurrence of APR (APR⁺) or not (APR^–) and quantified by body temperature and C-reactive protein (CRP). The APR⁺ women had significantly lower 25(OH)D levels than the APR^– women. Levels of 25(OH)D were normal (>30 ng/mL) in 31% of APR⁺ women and in 76% of APR^– women. The odds ratio (OR) to have APR in 25(OH)D-depleted women was 5.8 [95% confidence interval (CI) 5.30–6.29; p < .0002] unadjusted and 2.38 (95% CI 1.85–2.81; p < .028) after multiple adjustments (for age, body mass index, CRP, calcium, parathyroid hormone, and C-telopeptide of type I collagen). Levels of 25(OH)D were negatively correlated with postdose body temperature (r = −0.64, p < .0001) and CRP (r = −0.79, p < .001). An exponential increase in fever and CRP has been found with 25(OH)D levels lower than 30 ng/mL and body temperature less than 37 °C, whereas normal CRP was associated with 25(OH)D levels above 40 ng/mL. The association between post-N-BPs APR and 25(OH)D suggests an interesting interplay among N-BPs, 25(OH)D, and the immune system, but a causal role of 25(OH)D in APR has to be proven by a randomized, controlled trial. However, if confirmed, it should have some practical implications in preventing APR. © 2010 American Society for Bone and Mineral Research.

Introduction

Amino-bisphosphonates (N-BPs) are currently considered the most important class of drugs used for the inhibition of osteoclast activity in common metabolic bone diseases, such as postmenopausal osteoporosis, Paget disease, bone metastases, multiple myeloma, and hypercalcemia. Paget disease, metastatic bone disease, and hypercalcemia are usually treated by administering N-BPs intravenously, whereas the oral administration of N-BPs is commonly used for the treatment of osteoporosis. Recently, intravenous zoledronic acid and ibandronate have been licensed for the treatment of osteoporosis.

The major adverse event of intravenously administered N-BPs is the development of an acute-phase response (APR), which is a nonspecific physiologic immune-driven reaction to systemic challenge (i.e., infection, malignant tumors, or tissue damage). The reported incidence of APR ranges between 10% and 50% in patients treated for the first time with N-BPs, even if the real incidence may be underestimated because in some studies patients were instructed to take acetaminophen after the infusion.1–4 APR is characterized by a transient mild flu-like syndrome with fever, fatigue, myalgia, and malaise.2, 5 The APR usually develops within 24 to 36 hours of the first infusion and resolves spontaneously within 2 to 3 days.2 The symptoms are due to an increased circulating level of interleukin 6 (IL-6), tumor necrosis factor β (TNF-β), and interferon-γ (IFN-γ).6 This cytokine cascade results in inflammation and causes the hepatic production of a variety of proteins, collectively referred to as acute-phase proteins. The two major acute-phase proteins in humans are C-reactive protein (CRP) and serum amyloid A, both of which can increase up to 1000 times over normal levels during an APR.7 The pathogenic mechanism of APR has been clarified only recently. It has been suggested that intravenous N-BPs are internalized by endocytic cells, probably monocytes or dendritic cells, causing the inhibition of farnesyl phyrophosphate (FPP) synthase. This inhibition leads to a lack of farnesylation and geranylgeranylation of translated small GTPase interfering with a variety of essential cellular functions for survival and activity. Moreover, the FPP synthase inhibition leads to intracellular accumulation of isoprenyl pyrophosphate metabolites upstream of FPP synthase in the mevalonate pathway.8 In particular, this leads to the accumulation of isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) (alkylamine isoprenyl pyrophosphate monoesters), which are potent agonists of the γδ T-cell receptor. They are naturally recognized by γδ T cells with subsequent activation and release of TNF-β, IL-6, and IFN-γ, which are the proinflammatory cytokines involved in the development of APR.9 It has been demonstrated in vitro that the stimulatory effect of N-BPs on γδ T-cell proliferation was consistently abrogated by simultaneous treatment with mevastatin, which inhibits the mevalonate pathway upstream of FPP synthase. This inhibition prevents the accumulation of IPP and DMAPP and therefore prevents the stimulatory effect on γδ T cells.10, 11 Interestingly, it has been demonstrated in vitro that activation of human γδ T cells by IPP leads to the upregulation of the vitamin D receptor and that vitamin D negatively regulates in a dose-dependent manner the phospholigand-induced γδ T-cell expansion and IFN-γ production.12

Although the molecular mechanisms of APR related to N-BPs have been elucidated, it is still unclear why only a minority of patients suffers from APR after infusion of N-BPs, as well as which are the predisposing factors to development of APR and the basis of the high interindividual variability of APR severity. Excluding the previous use of N-BPs and the difference among the N-BPs with regard to the induction of symptomatic APR,13, 14 we conducted a study to evaluate the relationship between the common parameters of bone metabolism and APR in osteoporotic patients, with particular regard for vitamin D.

Materials and Methods

Subjects

Ninety female subjects (age 63.7 ± 10.6 years of age) were consecutively enrolled in the study between April and November 2008 in a single bone metabolic unit. All patients needed treatment with N-BPs for osteoporosis. Subjects with a history of cancer, autoimmune disease or immune deficit, alcohol abuse, recent or chronic corticosteroid therapy, statins and immune therapy, or other drugs that potentially could interfere with the immune system were not included in the study. Furthermore, subjects previously treated with intravenous or oral N-BPs were excluded. A questionnaire was used to evaluate other factors, such as smoking and concomitant use of calcium and vitamin D supplementation.

Study design and procedures

For all subjects, body weight and height (Harpender stadiometer, Wunder SA.BI, Bridgend, UK) were assessed, and body mass index (BMI, kg/m²) was calculated. All the women were postmenopausal. All subjects included in the study were treated with zoledronic acid 5 mg diluted in 0.9% saline and infused over 15 minutes. Forty-four percent of the subjects (40/90) were taking vitamin D supplementation as cholecalciferol 400 IU/day before starting zoledronic acid (mean period time of 1.5 ± 1.0 years) and continued after the infusion. All other subjects were supplemented with cholecalciferol 400 IU/day and calcium 500 to 1000 mg/day when necessary starting the day of the infusion. The biochemical parameters of bone metabolism analyzed in this study were serum calcium [Roche/Hitachi coefficient of variation (CV) < 10%; Roche Diagnostics, Mannheim, Germany], serum parathyroid hormone (PTH; DiaSorin Liaison N-tact PTH, CV < 10%; DiaSorin, Inc., Stillwater, FL, USA), serum 25-hydroxyvitamin D [25(OH)D; DiaSorin Liaison 25(OH) vitamin D, total CV < 15%; Diasorin), C-telopeptide of type I collagen (CTX; Elecsys B-CrossLaps/serum assay, CV < 20%; Roche Diagnostics), and CRP (Roche/Hitachi CV < 10%; Roche Diagnostics).

The blood samples were taken at baseline before N-BP infusion and on the first and third day after the infusion. The patients were invited to record in a diary the axillary body temperature, other symptoms referred to APR (i.e., flu-like syndrome, fatigue, malaise, bone/muscle pain, and headache), and the use of acetaminophen (500 mg only when necessary) for the 3 days immediately after the infusion.

The occurrence of an acute-phase response (APR⁺) was categorized as the presence of fever (body temperature > 37 °C) or the presence of at least one other APR symptom. To quantify the severity of the APR, the highest postdose body temperature and the highest CRP value over the 3 days after the infusion were considered.

Patients with a baseline level of 25(OH)D above 30 ng/mL were considered vitamin D repleted, with a level between 20 and 29 ng/mL were considered vitamin D insufficient, and with levels below 20 ng/mL were considered deficient.15 Biochemical parameters were analyzed independently by laboratory technicians who were unaware of the clinical data.

Statistical analysis

Baseline measures were analyzed descriptively. The demographic characteristics and biochemical parameters of bone turnover were compared according to the APR group (APR⁺ when APR occurred, APR^– when APR did not occur) with analysis of variance (ANOVA) and the Wilcoxon rank test for continuous variables. Categorical variables and proportions were analyzed using the Chi-square test. Analysis of covariance (ANCOVA) was used for adjusting the values for any confounding factors.

Associations between continuous variables were examined using Pearson's correlation coefficients and multivariate linear regression. Logistic regression analysis was performed to study the associations between 25(OH)D levels and APR, adjusted for potential confounding factors. The odds ratio (OR) included 95% confidence intervals (95% CIs). Differences were considered significant at p < .05. All statistical procedures were carried out using a computer program (Statgraphics Centurion, Version XV, Herndon, VA, USA).

Results

The demographic, clinical, and biochemical characteristics of all subjects entirely and of the subjects stratified for APR⁺ or APR^– are reported in Table 1. Of the subjects treated with 5 mg intravenous zoledronic acid, 67.7% (61/90) developed an APR. Forty-five percent of the study population (41/90) had normal 25(OH)D levels (49.6 ± 17.4 ng/mL), 25% (22/90) had 25(OH)D levels between 21 and 29 ng/mL, and 30% (27/90) had 25(OH)D levels lower than 20 ng/mL. Among APR⁺ subjects, only two reported body temperature below 37 °C. In APR⁺ subjects, there was a not significantly higher baseline CRP than in APR^– subjects (p < .070). The APR^– and APR⁺ levels were not significantly different for age, BMI, bone mineral density (BMD), or bone turnover marker (CTX). APR⁺ subjects had significantly lower 25(OH)D levels (25.40 ± 14.22 ng/mL versus 47.18 ± 22.95 ng/mL; p < .001), lower calcium levels (9.36 ± 0.34 mg/dL versus 9.52 ± 0.32 mg/dL; p < .042), and higher PTH levels (66.47 ± 32.20 pg/mL versus 53.50 ± 24.10 pg/mL; p < .041) than APR^– subjects.

Table 1. Demographic, Clinical, and Biochemical Characteristics of Study Population According to Acute Phase Response Development (APR⁺/APR^–)

	All patients	APR^–	APR^+	p
The values are expressed as means ± standard deviation. Differences were considered significant at p < .05.
Age (years)	61.1 ± 11.07	65.33 ± 10.42	60.24 ± 11.19	.220
BMI (kg/cm^2)	25.55 ± 3.49	24.96 ± 1.83	25.94 ± 4.14	.560
Height (cm)	156.7 ± 5.76	156.4 ± 5.68	156.6 ± 6.12	.930
Weight (kg)	63.21 ± 10.74	61.57 ± 2.57	64.06 ± 13.27	.630
Ca (mg/dL)	9.47 ± 0.34	9.52 ± 0.32	9.36 ± 0.34	.042
CTX (ng/mL)	0.59 ± 0.31	0.54 ± 0.33	0.61 ± 0.29	.290
25(OH)D (ng/mL)	32.3 ± 20.09	47.18 ± 22.95	25.40 ± 14.22	.000
CRP (mg/L)	5.91 ± 5.79	4.2 ± 1.54	8.2 ± 4.39	.070
PTH (pg/mL)	64.25 ± 29.75	53.5 ± 24.1	66.47 ± 32.2	.041
T-score neck (DS)	−1.88 ± 0.6	−2.01 ± 0.68	−1.76 ± 0.56	.410
T-score spine (DS)	−2.52 ± 1.26	−2.88 ± 1.13	−2.28 ± 1.31	.310

In the APR⁺ group, 41% of subjects had deficient levels of 25(OH)D, 28% had insufficient levels, and 31% were normal. Conversely, in the APR^– group, 76% of subjects had normal 25(OH)D levels, 17% had insufficient levels, and 7% had deficient levels (Fig. 1).

Figure 1. Percent distribution of patients with vitamin D deficiency, insufficiency, or sufficiency between APR⁺ and APR^– groups.

The OR of having APR in subjects low in 25(OH)D (<30 ng/mL) compared with normal 25(OH)D subjects was 5.80 (95% CI 5.30–6.29; p < .0002) unadjusted, 3.40 (95% CI 3.08–3.72; p < 0.003) after adjustment for baseline CRP, calcium, CTX, and PTH and 2.38 (95% CI 1.85–2.81; p > .028) after multiple adjustments (for age, BMI, baseline CRP, calcium, CTX, and PTH) (Fig. 2).

Figure 2. Relative risk and 95% confidence interval of developing APR having basal serum levels of 25(OH)D below 30 ng/mL. Square: Adjusted for age, BMI, serum calcium, serum CTX, PTH, basal CRP. Circle: Adjusted for serum calcium, serum CTX, PTH, basal CRP. Diamond: Unadjusted.

Table 2 lists the matrix for simple and multiple regression coefficients among age, BMI, baseline CRP, serum calcium, PTH, CTX, 25(OH)D, body temperature, and CRP postdose. BMI and PTH were positively related and, conversely, 25(OH)D was negatively related to fever and CRP postdose. Baseline CRP was negatively related to 25(OH)D and positively related to CRP postdose. However, 25(OH)D remained negatively related to postdose body temperature and CRP when adjusted for baseline PTH, CRP, and BMI.

Table 2. Simple and Multiple Correlation Coefficients Between Continuous Variables in Osteoporosis Patients

	CRP 0	BMI	Age	CTX	PTH	25(OH)D	Ca	25(OH)D (adjusted for PTH, BMI, and CRP 0)
^*p < .05; ^**p < .001, ^***p < .000). CRP 0 = basal C-reactive protein, CRP 3 = C-reactive protein on day 3.
Fever	0.331	0.476^*	0.355	0.299	0.190	−0.678^***	−0.248	−0.518^**
CRP 3	0.320	0.567^**	0.211	−0.009	0.512^*	−0.622^***	−0.140	−0.527^*

The negative correlation between body temperature (t, °C) (r = −0.64, R² = 41.38, p < .000) and CRP (mg/L) (r = −0.79, R² = 62.48, p < .000) and baseline 25(OH)D level is shown in Fig. 3.

Figure 3. The relationship between baseline serum 25(OH)D levels and temperature (A) (r = −0.64, R² = 41.38, p < .000) and CRP (B) (r = −0.79, R² = 62.48, p < .000) after the first zoledronic acid infusion.

After zoledronic acid infusion, acute changes (within 3 days) in bone turnover occurred. CTX and serum calcium decreased significantly from baseline on the third day (−93 ± 3.4%, p < .01, and −6.3 ± 5.1%, p < .0034, respectively). Only in one subject did serum calcium decrease below the low-normal limit (7.35 mg/dL). Percent calcium decrease was negatively related to PTH increase on the third day (r = −0.68, p < .0001). Finally, 25(OH)D decreased significantly on the third day (−17.3 ± 9.6%, p < .02). The level of 25(OH)D on the third day became 18.8 ± 7.9 ng/mL in subjects with low baseline levels of 25(OH)D and 35.1 ± 9.06 ng/mL in subjects with normal 25(OH)D at baseline. The percent variation in PTH on the third day postinfusion was significantly related to the percent decrease in 25(OH)D (r = −0.42, p < .045; data not shown).

Among the 19 APR⁺ subjects with normal 25(OH)D baseline levels (31%), 15 patients showed an increase in PTH of 74.2 ± 48.4% and a decrease in 25(OH)D of −30.1 ± 8.1% on the third postdose day. Their 25(OH)D absolute levels fell from 42.8 ± 8.1 ng/mL (range 31 to 67 ng/mL) to 27.2 ± 1.7 ng/mL (range 24 to 29 ng/mL) (Fig. 4).

Figure 4. (A) Individual data on serum 25(OH)D levels before and 3 days after zoledronic acid infusion in the 19 subjects with normal baseline levels of 25(OH)D who developed APR. (B) PTH levels before and 3 days after the infusion in the same cohort of patients.

Discussion

To the best of our knowledge, this is the first investigation that describes the association between 25(OH)D serum levels and the development of APR after the first infusion of N-BPs. About 70% of the patients with APR⁺ had 25(OH)D levels below 30 ng/mL, and in the majority of patients (76%) in whom APR did not occur, the 25(OH)D level was above 30 ng/mL. Levels below 30 ng/mL induced a threefold increased risk of APR associated with zoledronic acid. Furthermore, we found that 25(OH)D levels are significantly and negatively related in a dose-dependent manner to the severity of APR expressed as both intensity of fever and CRP levels. The relationship between CRP and 25(OH)D resulted in a stronger response than that with fever. CRP is the major acute-phase protein in humans, and it can increase up to 1000 times over normal during an APR.7 Therefore, CRP probably captures APR more sensitively, probably also detecting subclinical APR. Furthermore, since the subjects in our study were instructed to take acetaminophen when intolerable fever or other symptoms occurred, temperature likely could underestimate the real entity of APR and is encumbered by a higher interindividual variability than CRP, which is not influenced by acetaminophen.

In the literature, the frequency of APR ranges between 30% and 55% of the patients receiving an initial dose of an intravenous N-BP.13 In our study, the incidence of APR was quite a bit higher (about 67%) than that reported in the Horizon Study, where the incidence of APR (considering the five most frequent symptoms) after the first infusion was about 31%.16 The difference could be sustained by unknown predisposing factors. Particularly in our study population, the prevalence of a low 25(OH)D level was high (55% of the subjects) but in agreement with European and national epidemiologic data.17, 18 The baseline levels of 25(OH)D in the randomized, controlled trials evaluating zoledronic acid efficacy in osteoporosis are not mentioned.16, 19

Baseline CRP, PTH, and BMI also were related to both APR parameters (fever or CRP) and 25(OH)D levels. The negative relationship between PTH and 25(OH)D is well known, deficiency of 25(OH)D causing secondary hyperparathyroidism.20 All indices of adiposity are negatively associated with serum 25(OH)D, and it is well known that subjects with normal BMI are more likely than overweight and obese women to have high serum 25(OH)D levels.21–23 CRP and 25(OH)D have been found associated in many studies.24–33 However, the contribution of baseline CRP, PTH, and BMI in the increase in CRP and fever after zoledronic acid infusion seems to be minor because the regression coefficients between postdose fever and CRP with 25(OH)D remained statistically significant after baseline PTH, CRP, and BMI adjustment (see Table 2).

We did not found in the literature any studies aimed at finding the predisposing factors to the development of APR after first-dose infusion of N-BPs or investigations aimed at assessing the causes of the extreme interindividual variability of severity of APR. Clearly, the identification of risk factors of APR could help to prevent it and to avoid the most frequent and disabling problem in the tolerability of intravenous N-BPs.

Although the findings of our study clearly did not prove a causal relationship between 25(OH)D levels and APR and a randomized, controlled trial is indicated to clarify the role of 25(OH)D, the association could generate an interesting hypothesis on the basis of a strong rationale. Vitamin D is a potent immunomodulator34, 35 exerting an inhibitory action on the adaptive immune system and promoting innate immunity through the production of antimicrobial peptides such as cathelicidin.35 The most well-established functions of vitamin D within the adaptive immune system concern the ability of 1,25(OH)₂D to modulate T-lymphocyte production and function, suppressing proliferation and immunoglobulin production and retarding the differentiation of B-cell precursors into plasma cells. In addition, it inhibits T-cell proliferation, in particular, the Th1 cells producing IFN-γ and IL-2 and activating macrophages. At least in part these actions on T-cell proliferation and differentiation stem from actions of 1,25(OH)₂D on dendritic cells to reduce their antigen-presenting capability.35 There are data suggesting that vitamin D may specifically modulate the APR. It has been demonstrated in vitro that the vitamin D receptor (VDR) is upregulated in a subpopulation of T cells mainly involved in the APR, the γδ T cells. In particular, expression of the VDR can be upregulated in γδ T cells after activation by phosphate ligands such as IPP and DMAPP.12 Interestingly, vitamin D can strongly modulate the γδ T-cell response, selectively downregulating the inflammatory properties of γδ T cells and inhibiting the expression of IFN-γ and TNF-β after IPP stimulation.12. Also, there are in vivo data suggesting that poor vitamin D status may promote the APR. Vitamin D levels drop during winter in people with scarce sun exposition and in the elderly. Interestingly, both fibrinogen and CRP have been reported to show a seasonal fluctuation that reaches a peak during the winter,24, 25 tends to increase with aging,26, 27 and is lowered by increasing sun exposure.28–32 It has been reported that 500 IU of cholecalciferol daily is able to lower elevated CRP levels in patients with critical illnesses and poor 25(OH)D.32 IL-6 is believed to be the principal physiologic stimulant of so-called type II acute phase proteins by human hepatocites, including fibrinogen and CRP.36, 37 PTH can trigger IL-6 production in osteoclasts and perhaps in other tissues as well; in particular, it has been reported to enhance hepatic secretion of IL-6.38 In our study population, baseline levels of CRP were slightly but not significantly higher in subjects with low 25(OH)D levels than in subjects with normal 25(OH)D levels, and it is likely that the secondary hyperparathyroidism in these subjects could be associated with an increase in the serum levels of acute-phase proteins.

The curve representing the relationship between vitamin D levels and fever or CRP clearly indicates a dose-response relationship. It suggests that clinically significant differences in the severity of APR among subjects could be ascribed to deficiency of vitamin D, with an impressive and exponential increase in CRP and fever with levels of 25(OH)D below 30 ng/mL.

The curve also suggests that levels of 25(OH)D greater than 40 ng/mL are necessary to avoid fever and perhaps greater than 50 ng/mL to maintain CRP lower than 10 ng/L. Although there is no consensus, the common definition of normal for 25(OH)D is 30 ng/mL or greater.15, 39 Nowadays, there is evidence for nonclassical actions of vitamin D.40 Vitamin D receptors are found in most tissues, and many of these tissues also contain the enzyme CYP27B1, which is capable of producing 1,25(OH)₂D from the circulating form of vitamin D.15 The most advantageous serum 25(OH)D concentration for several nonskeletal outcomes such as colon cancer, oral health, falls, lower extremity function, insulin resistance, hypertension, and multiple sclerosis has been estimated to be greater than 30 ng/mL, the best being between 36 and 40 ng/mL.41 There is no published evidence concerning optimal levels of 25(OH)D for immunomodulating effects. Our data suggest that the serum levels of 25(OH)D to prevent an APR induced by zoledronic acid probably should be greater than those reported for other nonskeletal effects. Zoledronic acid is by far the most potent inhibitor of the mevalonate pathway, and data from pivotal trials in osteoporosis treatment indicate that it has the highest incidence of APR with respect to the other N-BPs given intravenously.3, 8 It is likely that the rank order of inhibitory potency of N-BPs on FPP synthase matches not only with the antiresorptive potency14 but also the immunomodulating effects on γδ T cells. Therefore the “protective” levels of 25(OH)D on γδ T cells to prevent APR could differ among N-BPs on the basis of the potency of inhibition of FPP synthase and consequently the intensity of γδ T-cell stimulation.

In our study we observed 19 patients who experienced APR despite normal baseline 25(OH)D levels. Interestingly, in 79% of these subjects, 25(OH)D decreased acutely (within 3 days) after the zoledronic acid infusion, reaching levels lower than 30 ng/mL. Zoledronic acid infusion suppresses bone turnover more acutely and more deeply than oral BPs.1, 42 In our study, the acute changes in bone turnover are similar to those found by Generali and colleagues in cancer patients.43 Zoledronic acid, by reducing bone resorption acutely, induces a positive bone remodeling imbalance within a few days. The decrease in calcium levels cause a transient secondary increase of PTH that, in turn, leads to an acute decrease of about 18% in 25(OH)D levels, likely through enhancing metabolic clearance of 25(OH)D in 1,25(OH)₂D. Although under basal conditions the estimated half-life of 25(OH)D is about 15 days, there is much evidence that an acute endogenous PTH increase or PTH(1-34) infusion causes a significant and rapid conversion of 25(OH)D into 1,25(OH)₂D within 18 to 24 hours, necessarily lowering 25(OH)D levels.44–48 In our study, the increase in PTH worsened the 25(OH)D deficiency in subjects with low 25(OH)D levels at the baseline but, more interestingly, shifted acutely a number of subjects having 25(OH)D levels just above the normal level before infusion to 25(OH)D levels below normal immediately afterward. We suppose that in these cases 1,25(OH)₂D levels increased, especially in subjects with sufficient 25(OH)D levels, even if we did not measure it. However, the important vitamin D substrate for several nonclassical actions of vitamin D, including the immunomodulating effect, is 25(OH)D and not the active 1,25(OH)₂D.15 Immune cells such as macrophages contain the enzyme CYP27B1, which is responsible for converting circulating 25(OH)D into 1,25(OH)₂D, and the intracrine, autocrine, or paracrine synthesis of 1,25(OH)₂D seems to be important for innate and adaptive immunity.15, 49 Therefore, from our data we could suggest that a basal level of 25(OH)D above 40 ng/mL would be adequate to prevent APR by ensuring the maintenance of an adequate substrate level of 25(OH)D.

It is well known that APR is usually limited to the first infusion of zoledronic acid, with a significant attenuation after subsequent infusions. This phenomenon is poorly explored and discussed in literature. We studied the association of 25(OH)D and APR after only the first infusion, and we could only speculate about the subsequent attenuation of APR.

A plausible hypothesis could be drawn, considering that 1 year after the initial infusion the serum CTX is still at the lower limit of normal,16 and it is likely that because of vitamin D supplementation, the 25(OH)D serum level will be better than baseline. Therefore, it is probable that a subsequent infusion produces minimal stimulation of PTH and a minimal drop in 25(OH)D, reducing the risk of developing an APR.

A further suitable hypothesis is linked to the immunomodulating mechanism of zoledronic acid. Zoledronic acid in vivo expands a subset of γδ T cells, the effector γδ T cells, while strongly and persistently decreasing naive and memory γδ T-cell subsets.50 After stimulation with IPP, the production of IFN-γ, inducing an APR, is typical of the effector cells, whether or not the proliferation activity of γδ T cells is a property of naive and memory cells.50 Furthermore, a decrease of effector γδ T cells in the peripheral blood was observed after 56 and 90 days after the first zoledronic acid infusion, perhaps because of their migration into peripheral tissues, where they can perform terminal functions.51 The changes in γδ T-cell subsets are long-lasting, and it is likely that subsequent infusions of zoledronate will not induce an APR because of the poor number of residual naive T cells and because of the distribution into peripheral tissues of effector T cells. On this basis, we could hypothesize that attenuation of the APR after the first infusion is probably independent of 25(OH)D serum levels.

Our study might have some direct implications for clinical practice. APR associated with N-BPs currently is managed empirically with acetaminophen.16, 19, 42, 52 An attempt to modulate the APR associated with intravenous N-BPs using statins, as was suggested based on in vitro studies,11 turned out not to be effective.53 Our data, if confirmed, could suggest a simple and safe way to prevent the APR. Usually, vitamin D supplements are given at the start of BP therapy so that at the beginning of therapy patients still have low 25(OH)D levels. However, achievement of adequate levels of 25(OH)D (probably > 40 ng/mL) should be reached before the first infusion of N-BPs. Finally, since 25(OH)D is the only useful substrate for many nonskeletal effects, including APR, only cholecalciferol should be used for supplementation and not the active metabolites.41

Our study have some limitations. We used a chemiluminescent method to measure 25(OH)D serum levels. The results obtained with different assays may be very different, and the comparability of 25(OH)D assays has important implications for the discussion on the minimum required serum 25(OH)D level.54–56 The data, particularly the 25(OH)D levels useful to prevent APR, should be confirmed with more sensitive methods. Further investigations should be carried out to better clarify the role of 25(OH)D in the prevention of APRs, particularly randomized, controlled trials with and without supplementation with cholecalciferol before infusion of zoledronic acid in order to determine the preventive effect of 25(OH)D and the optimal dosage of vitamin D necessary to minimize APR symptoms.

In conclusion, this study shows for the first time the association between low levels of vitamin D and the APR induced by the infusion of N-BPs, suggesting an interesting interplay among N-BPs, vitamin D, and the immune system. If it is confirmed, it would have some practical implications, such as suggesting a feasible and safe approach to management of APR symptoms.

Disclosures

The authors state that they have no conflicts of interest.