Osteoporosis (OP) is a chronic, progressive disease that manifests as low bone mass, degenerated bone micro-architecture, as well as enhanced bone fragility and fracture risk (1). Based on a 2018 Chinese epidemiological survey, 19.2% of OP patients were above 50 years old, and 32.0% of the Chinese population above 65 years old was diagnosed with OP (2). Among the numerous fragility fractures associated with OP, vertebral and hip fractures are the most correlated with enhanced morbidity and mortality (3). Based on the current statistics, around 30–50% females and 15–30% males will likely experience OP-related fracture in their lifetime (4, 5). Therefore, early detection of high-risk patients, and subsequent treatment, is most crucial to the health and quality of life of these patients.

Bisphosphonates (BPs) are widely used for treating OP, and they are reported to markedly reduce osteoclast-based bone resorption, thus diminishing the potential for vertebral and non-vertebral fractures (6). Zoledronate (ZOL) 5 mg is an intravenous BP, administered once a year, and it was approved for treating and preventing postmenopausal or glucocorticoid-induced OP, as well as enhance bone mass in OP males. Since it is infused once a year into patients, the patient adherence to ZOL is exceptionally high, and it avoids gastrointestinal absorption/irritation challenges commonly observed with oral BPs (7). Despite this, there are reports of transient acute phase reactions (APRs) like fever, myalgia, and flu-like symptoms during the first 3 days post first ZOL administration (8). To ensure enhanced patient adherence, it is critical to develop approaches that either prevent or better manage these APRs in patients receiving their first ZOL infusion.

Prior investigations identified certain APR risk factors, which can serve as indicators for APR development (9, 10). Among these factors, appropriate serum 25-hydroxy vitamin D (25[OH]D) content prior to ZOL infusion was demonstrated to markedly reduce APRs incidence (11–13). However, apart from clinical trials, the real world medical evidences on the independent correlation between serum 25(OH)D levels and APR risk in large patient populations are relatively scarce. In addition, there is no standard cut-off for appropriate 25(OH)D concentrations and ZOL-induced APR risk prior to ZOL infusion. Thus, the goal of our research was to further explore the correlation between serum 25(OH)D content and APR risk in the clinic.

This investigation was by far the largest of its kind in China, and it demonstrated a strong independent and negative correlation between serum 25(OH)D content and ZOL fever risk. Our research was based on the full use of real-world clinical data, and the inclusion of some new covariates that may influence fever risk during analysis, namely, serum β-CTX levels and calcitonin usage. In different adjusted models, the negative linear relationship remained stable. We also observed a second non-linear correlation between 25(OH)D content and ZOL fever risk, and the turning point of the adjusted smoothed curve was 35 ng/mL for serum 25(OH)D levels. These evidences indicated that serum 25(OH)D offered a certain level of protection against ZOL-induced APR fever in the daily clinical practice setting, and it became more significant after serum 25(OH)D levels reached 35 ng/mL.

Even though our findings did not provide a strong causal association between 25(OH)D content and APR, and a randomized, controlled trial (RCT) is needed to validate our findings, a potential relationship between the two is still of marked significance. Bertoldo et al. proposed this hypothesis in an earlier paper (11). ZOL is a new generation nitrogen-based bisphosphonate (N-BP), and several reports suggested that an intravenous administration of N-BPs are taken up by monocytes and dendritic cells, which, in turn, suppresses farnesyl pyrophosphate (FPP) synthase activity, thereby, accumulating intracellular metabolites like isopentenylpyrophosphate (IPP) and dimethylallyl diphosphate (DMAPP) upstream of the FPP synthase in the mevalonate network. These metabolites are robust gamma-delta T cell receptor agonists, which activate the gamma-delta T cells, and formation of proinflammatory cytokines like interleukin-6 (IL-6), tumor necrosis factor alpha (TNF- alpha), and interferon- gamma (IFN-gamma), which are crucial for APR pathogenesis by the same lymphocytes (23–25). Vitamin D is a steroid hormone that not only contributes to the calcium/phosphate metabolism, but also negatively regulates the adaptive immune system. Moreover, it promotes innate immunity through antimicrobial peptide synthesis (26, 27). A prior report revealed that the vitamin D receptor (VDR) expression is markedly enhanced in gamma-delta T cells. Particularly, BP ligands like IPP and DMAPP markedly enhances VDR levels in the gamma-delta T cells (28). Vitamin D is a potent regulator of gamma-delta T cell response, and it selectively downregulates its inflammatory properties, and inhibits the expression of proinflammatory cytokines following IPP-based activation (28).

Our analysis also revealed that the 25(OH)D content above 35 ng/mL is likely necessary to prevent ZOL-triggered fever. At present, the data on the optimal 25(OH)D level for immunomodulation is rather scarce. Bischoff-Ferrari et al. presented a combined result of multiple studies evaluating the serum 25(OH)D threshold with regards to bone mineral density (BMD), lower-extremity function, dental health, risk of falls, fractures, and colorectal cancer (29). In terms of all analyzed endpoints, the most optimal serum 25(OH)D concentration begins at 30 ng/mL, and the best range is between 36 and 40 ng/Ml (29). This result is consistent with our study, and corroborates with the data by Bertoldo et al., which indicated that a prime amount of 25(OH) (>40 ng/mL) must be obtained prior to N-BPs administration (11). However, the average serum 25(OH)D was 21.15 ng/mL in this study, and the value was even lower in the winter and spring (19.70 ng/mL), or before the first infusion of ZOL (20.47ng/mL). These values were well below the 35 ng/mL threshold that we recommend in this paper.

Alternately, there is much controversy regarding the suitable upper limit for serum 25(OH)D. In addition, the safety of elevated serum 25(OH)D levels among various patient populations remains inconclusive. A logical upper limit, computed from the concentration of 25(OH)D in sun-exposed healthy young adults, is 50 ng/mL, until more information is obtained (15). Given these evidences, we speculated that 35-50 ng/mL of serum 25(OH)D may be the target range needed to prevent ZOL-induced fever. In terms of the amount of vitamin D supplement, studies involving older individuals revealed that 25(OH)D levels can be enhanced by 20-26 ng/mL to a mean of 40 ng/mL using only 800 IU vitamin D per day (30, 31). Multiple scientific organizations suggested that adults 50 years and older must consume a minimum of 1,000 IU of vitamin D a day (15). The National Academy of Medicine suggested that 4,000 IU of vitamin D once a day is a safe upper limit for the general population (32, 33). Hence, we suggest a daily vitamin D dosage above 1,000 IU for OP patients prior to their first ZOL infusion.

Based on prior research, fever is less prevalent after the second and third ZOL infusions. In fact, fever incidences within three days post ZOL infusion in the aforementioned study ranged from 45.73, 13.65, and 5.26% after infusions 1 to 3, respectively. These incidences were slightly higher than the incidences reported in the ZONE study (38.8% after the first infusion and 7.4% after the second infusion), which evaluated Japanese postmenopausal women with OP (34), and were higher relative to the HORIZON PFT study (20% after the first, 4% after the second, and 1% after the third infusion) (8). These discrepancies may be the result of any of the following three factors: Firstly, a difference in fever assessment methodology can produce different results. The HORIZON PFT study recorded self-reported fever, whereas, we actively recorded patient body temperature for analysis in our research. This may have contributed to the reduced fever incidence in the PFT investigation, relative to our work. Secondly, the PFT subgroup analysis was stratified by various racial groups. Those most at risk were non-Japanese Asians (including Chinese) and Pacific Islanders, and they carried a univariate OR of 2.20 and 3.39, after adjustment of other variables in the model (8). The racial differences in APR prevalence potentially reflect the differing thresholds for the reporting and recording of adverse events, or it may represent race-based alterations in cytokine formation or activity. Thirdly, although the baseline 25(OH)D levels of the PFT study were not reported, the baseline levels were about 26 ng/mL in the ZONE study, which were higher than the serum 25(OH)D levels before the first (20.47 ng/mL) and second infusions (23.47 ng/mL) in our research. The baseline 25(OH)D levels prior to ZOL infusion may influence the APR severity to ZOL in patients with OP.

Owing to a suspected analgesic influence (35, 36), we provided a short-term calcitonin administration to OP patients receiving their first ZOL infusion. However, the increased risk of ZOL fever with previous use of calcitonin in our study was surprising, which is contrary to the results reported by Reid (8). Common complications of parenterally infused calcitonin are as follows: nausea, injection site inflammatory reactions, and vasomotor complications like sweating and flushing (15). Beyond these effects, we observed no published evidences on the relationships between elevated fever risk and a combined treatment of ZOL and calcitonin. In addition, uncovering a potential mechanism is very challenging. Since a combined treatment of calcitonin and BP can rapidly and effectively reduce blood calcium levels (37), we hypothesized that the lower calcium concentration induces a transient secondary elevation in parathyroid hormone (PTH) concentration. PTH, in turn, elevates IL-6 synthesis within osteoblasts, and acts as a bone resorbing agent (38, 39). Elevated serum IL-6 concentration is prevalent in individuals with primary or secondary hyperparathyroidism (40, 41). IL-6, as mentioned above, is associated with ZOL-induced APR pathogenesis (42–45). Thus, a combined treatment of ZOL and calcitonin can likely increase the potential for side effects. Of note, before our data is confirmed, the association between fever and a combined treatment of ZOL and calcitonin require further exploration in the near future.

We did not observe any NSAID-based protection against ZOL-induced fever, despite publications reporting a strong likelihood of APR reduction following ZOL infusion with NSAID usage (37, 46, 47). However, the NSAID usage in our present investigation was not only based on prophylactic therapy, but was instead used to treat other medical conditions related to persistent pain and/or inflammation, which may further complicate patient condition. To improve the reliability of the results, future research should use the prophylactic use of NSAIDs as a covariate. The lack of relationship between statin usage and APR risk was somewhat expected due to a similar report by Reid (8). Given that the present investigation included a relatively small quantity of statin users, additional examinations are warranted to determine its relation, if any, to the regulation of APR risk.

Our conclusions can greatly impact clinical practice. Identification of APR risk factors can potentially prevent APR occurrence and enhance tolerability of intravenous ZOL. Particularly, a “vitamin D supplemental period” is essential prior to ZOL infusion, particularly the first ZOL infusion, such that an appropriate 25(OH)D (probably>35 ng/mL) level is obtained to prevent ZOL-induced APR. Patients in high altitude or during winter and spring (the Northern Hemisphere) tend to have reduced baseline vitamin D levels (48, 49), and may require an extension of this supplemental period to increase the amount of vitamin D supplementation before ZOL infusion.

According to the findings of the univariate analysis, OP patients who have higher 25(OH)D, hemoglobin, lymphocyte counts, calcium, high-density lipoprotein levels, or have lower neutrophil counts, monocyte counts, alanine transaminase, β-CTX, P1NP, and ferritin levels, or who have had previous ZOL infusions, are female, with no fractures, with hypertension, and no calcitonin use, and if the season is autumn (October to November) may have less risk of ZOL-induced fever. These variables can be selected as candidate predictors of ZOL-induced fever. We have recently published an article predicting the risk of ZOL-induced fever (50). Knowledge of these predictors allows for early intervention in patients at high risk for fever. In addition, the risk/protective factors we have identified may provide clinical evidence for the future development of new prediction models.

Our research has a number of strengths. First, certain potential confounders, such as, infection and young OP patients (age<50 years) were excluded prior to analysis. Moreover, known confounding factors, namely, age, gender, order of ZOL infusion, main diagnosis, season of blood collection, year of blood collection, calcitonin usage, and β-CTX were corrected using multiple logistic regression analyses. Second, we generated a generalized linear model to assess the linear association between serum 25(OH)D levels and ZOL-induced fever risk, and employed GAM to identify a nonlinear association between the two variables. GAM holds strong benefits in terms of nonlinear associations, and it performs adequate non-parametric smoothing and fitting of data to a regression spline. Therefore, using GAM, we were able to better identify the true association between exposure and endpoint. Third, fever was assessed using quantifiable maximum axillary temperature or CRP, rather than imprecise patient self-report.

However, this study also has certain limitations. Firstly, this study only estimated the ZOL-induced fever risk and did not include the four remaining symptoms of APR. Additionally, in-depth data on the prophylactic use of NSAIDs was not collected. Secondly, our study was an analytical retrospective investigation, and, thus, any observed associations did not represent causation. Therefore, specific intervention trials are necessary to support our recommendations. Thirdly, our research employed a single-center design, with a relatively small population size. Therefore, our findings may not be applicable to other biographical ethnic groups, or for intravenous BP other than ZOL. Given these limitations, we recommend additional investigations, involving large-scale, multi-center RCTs, and using people of different races, to better ensure robustness of our study results.

In summary, this study demonstrated an independent negative correlation between serum 25(OH)D content and ZOL fever risk. According to our findings, 25(OH)D above 35 ng/mL may be more effective in preventing ZOL-induced APR. If this is confirmed, a “vitamin D supplemental period” is warranted prior to ZOL infusion, particularly the first ZOL infusion, to ensure appropriate 25(OH)D levels that protect against ZOL-induced fever.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Affiliated Kunshan Hospital of Jiangsu University. The ethics committee waived the requirement of written informed consent for participation.

Study design: KL and QS. Study conduct: CL and KL. Data collection: Y-QG. Data analysis: KL. Data interpretation: CL and QS. Drafting manuscript: KL. Revising manuscript content: QS and KL. Approving final version of manuscript: KL, CL, QS and Y-QG. KL and CL take responsibility for the integrity of the data analysis. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. All authors contributed to the article and approved the submitted version.

The study was supported by National Natural Science Foundation of China (CN) (82172441), Scientific Research Project of Gusu School of Nanjing Medical University (CN) (GSKY20210244), Clinical Medical Science and Technology Development Fund of Jiangsu University (CN) (JLY2021048) and Suzhou Key Clinical Diagnosis and Treatment Technology Project (CN) (LCZX202024).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.