Change in body temperature, not acute-phase reaction, predict anti-Osteoporosis efficacy after the first administration of Zoledronic acid: a prospective observational cohort study

Abstract

Background

Acute-phase reactions (APRs) are common among people treated for the first time with zoledronate (ZOL). The current view is that both the APRs caused by ZOL and its efficacy are related to the mevalonic acid pathway. However, the relationship between APRs and ZOL efficacy remains unclear.

Methods

This was a prospective observational cohort study involving postmenopausal women with osteoporosis in Shanghai, China, for 1 year. A total of 108 patients with an average age of 67.4 ± 5.8 years were treated with 5 mg intravenous ZOL for the first time. Data on demographic characteristics, APRs, blood counts, bone turnover markers, including C-telopeptide collagen crosslinks (CTX) and N-terminal propeptide of type 1 collagen (PINP), and bone mineral density (BMD) were collected.

Results

(1) The results did not reveal a relationship between APRs and changes in bone turnover markers and BMD but showed that changes in body temperature (T) within 3 days after administration were positively correlated with changes in the BMD of the LS at Month 12 (β = 0.279 P = 0.034). (2) This effect was mediated mainly by changes in serum CTX (b = 0.046, 95% CI [0.0010–0.0091]). (3) The ROC curve revealed that when T increased by 1.95 °C, the sensitivity and specificity of identifying clinically important changes in LS BMD after 1 year were optimized.

Conclusions

In this study, we tested the hypothesis that people with elevated body T after initial ZOL treatment had greater improvements in BMD and better outcomes.

Trial registration

NCT, NCT03158246. Registered 18/05/2017.

Introduction

Osteoporosis has become a major public health problem with impacts on both quality of life and quantity of life. In China, the prevalence of osteoporosis among adults aged 40 years or older is 5.0% among men and 20.6% among women [1]. Bisphosphonates (BPSs) are the most commonly used therapy for osteoporosis [2].

Zoledronic acid (ZA), a third-generation nitrogen-containing BP, represents the most potent inhibitor of bone resorption and is long-acting in bone marrow, clearly improving the bone mineral density (BMD) and reducing the risk of vertebral, nonvertebral and hip fractures [3, 4]. ZA has been recommended in guidelines as a first-line treatment for extremely high fracture risk [5, 6]. Currently, ZOL is widely used in the clinic because of its significant efficacy, convenient annual administration, good patient compliance and few gastrointestinal reactions [7].

As a unique reaction after the administration of nitrogenous bisphosphonate (NBPS), acute-phase reactions (APRs) after the infusion of 5 mg of zoledronic acid for the first time are common. Fever and musculoskeletal pain occurred 24 to 36 h after NBP administration. Fever was noted to be associated with a decline in circulating lymphocyte number and increases in circulating γδ-T cells, IL-6 and tumour necrosis factor-alpha (TNF-a). These reactions are collectively referred as the acute phase response.

The main mechanism of ZOL is to inhibit the activity of farnesyl pyrophosphate (FPP) synthetase, the key enzyme of the mevalonate pathway, which can selectively suppress osteoclastic bone resorption. This mechanism may also be the basis for the occurrence of APRs [8]. ZOL inhibits FPP synthetase, resulting in the accumulation of isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP), the upstream intermediates of FPP synthetase of this metabolic pathway in monocytes, and then activates γδ- T cells via the release of proinflammatory cytokines, such as interleukin-6, interferon-γ and tumour necrosis factor β [9,10,11]. Therefore, APRs are more significant after ZOL treatment in those who are more sensitive to the mevalonate pathway. Consequently, both APRs caused by ZOL and its efficacy are likely related to the mevalonic acid pathway.

Approximately 30-54.9% of patients treated with a once yearly intravenous infusion of ZOL experienced an ARP, characterized by a transient flu-like syndrome with fever, chills, flushes, fatigue, malaise, and musculoskeletal and gastrointestinal symptoms following the first infusion [12, 13]. Therefore, we designed the present study, which lasted for 1 year after the first ZOL treatment infusion. In this study, we explored the relationships between APR and changes in bone turnover markers and bone mineral density after the first application of ZOL in postmenopausal women with osteoporosis to assess whether the therapeutic effect could be predicted by the occurrence of APR after the first ZOL administration.

Methods

Study design

This was a prospective observational cohort study involving postmenopausal women with osteoporosis in Shanghai, China, for 1 year. The study was registered at ClinicalTrials.gov via the Protocol Registration and Results System (PRS) on 18/05/2017 (ClinicalTrials. gov ID: NCT03158246, First posted date 18/05/2017) by Cttq. The Formula n = 2[(Z_α+Z_β)σ/δ]² was used to calculate the sample size. We calculated the sample size that was necessary to achieve a statistical power of 80% with a significance level of 0.05. Thus, α was set at a level of 0.05, whereas β was set at a level of 0.2. Therefore, Z_α was 1.96 (2-sided testing), whereas Z_β was 0.84. According to our previous research, σ, which was the standard deviation of the change in BMD between the two groups, was 0.022; δ, the difference in the mean BMD between the two groups, was 0.009. Therefore, the calculated sample size (n) was determined to be 95. Considering the loss to follow-up, the sample size was increased by 15%, resulting in a final sample size of 110. The flowchart of this study is shown in Fig. 1. A total of 110 patients were ultimately enrolled. A total of 108 patients with an average age of 67.4 ± 5.8 years and a mean body mass index of 22.3 ± 3.2 kg/m² who were treated with 5 mg intravenous ZOL for the first time between September 2017 and March 2018 completed all the visits and were included in the analysis. Two subjects were excluded from the study. The first subject withdrew her informed consent because she moved to another city 2 months after enrolling in the study. The second subject passed away in a car accident 7 days after enrolment. In our final analysis, we chose not to incorporate the data from these two patients because of an acceptable loss to follow-up rate of 1.8% and because their withdrawals were unrelated to the examination of drug side effects and therapeutic effects. A large amount of missing data can severely undermine the validity of inference and conclusions of a study. It is generally believed that if the proportion of missing values is less than 5%, samples containing missing values can be deleted, which will not have a significant effect on the overall analysis of the dataset. Otherwise, multiple imputation, weighting and maximum likelihood-based methods can be used to deal with incomplete data. Intravenous ZOL was administered at a dose of 5 mg over 15 min, and calcium (600 mg/d) and vitamin D (400 IU/d) were given as basic supplements following the infusion. To ensure compliance with treatment, calcium and vitamin D supplements were uniformly given by the investigators. A diary card was provided to request that the subjects record medication usage and adverse effects. Telephone interviews were conducted on Days 1 − 3 and at Months 3 and 9. Adverse events and calcium and vitamin D supplement consumption of the subjects were recorded. Clinical visits were conducted on Day 14 and at Months 6 and 12. The subjects were required to return the remaining calcium and vitamin D supplements and empty medicine bottles. Compliance with treatment was judged on the basis of the remaining amount of supplements. New calcium and vitamin D supplements were subsequently distributed for the next clinical visit. Oral T (T) at baseline and every morning for 3 days and on Day 14 after the infusion was recorded. Baseline and Day 14 blood samples were used to measure complete blood counts, liver and kidney function, and bone turnover marker (BTM) levels. Moreover, BTMs were also measured at Months 6 and 12. The BMDs of the lumbar spine (L2–4) and left total hip were measured at baseline and at 6 and 12 months after infusion. The investigators were blinded to the treatment allocation. The technicians involved in the measurement of BMD and BTM were blinded to any discernible clinical information regarding the participants, employing a rigorous blinding process. The study protocol and procedures were approved by the ethics committee of the hospital (No: 20170068).

Fig. 1

Flowchart of this study

Study populations

Postmenopausal women aged from 46 to 80 years with osteoporosis eligible for intravenous ZOL were eligible for inclusion if they had a bone mineral density T score of -2.5 or less at either the lumbar spine, hip, or femoral neck; a T score between − 1.0 and − 2.5; fragility fracture of the proximal humerus, pelvis, or distal forearm; or low-trauma spine or hip fracture regardless of BMD. Moreover, they had to be capable of monitoring and recording basic data on their body T and clinical symptoms.

The exclusion criteria were as follows: (1) patients who had previously used bisphosphonates; (2) patients who had used parathyroid hormone; (3) patients who had undergone surgical treatment in the past three months; (4) patients with a total serum calcium level of less than 2.1 mmol/L (8.4 mg/dL) or a serum-free calcium level of less than 0.95 mmol/L (3.8 mg/dL) or untreated hypocalcaemia; (5) patients with a Cockcroft calculated creatinine clearance of less than 35 mL/min; and (6) patients who refused to take ZOL, those with causes of osteoporosis other than postmenopausal osteoporosis, and patients with other diseases with an unstable status. Informed consent was obtained from all individual participants included in the study.

Study methods

Basic information collection and anthropometric measurements

Medical history, including fracture history and intake of antiosteoporosis drugs, was assessed, and a physical examination was performed. Height was measured using a stadiometer to the nearest 0.01 m. Body weight was measured by applying a standard balance beam scale to the nearest 0.1 kg. Body mass index (BMI) was calculated as the body weight divided by the square of the height (kg/m²).

Measurement of bone mineral density and the definition of osteoporosis

The BMDs of the lumbar spine (L2–4) and left total hip were measured with a dual-energy X-ray absorptiometry densitometer (Hologic Delphi A; Hologic Inc., Methuen, MA, USA). The precision error in our laboratory was 0.8% for the lumbar spine, 1.05% for the femur neck and 0.97% for the total hip. The densitometer was standardized with a standard phantom before each measurement.

Laboratory measurements

All blood samples were obtained in the morning after a 10 h overnight fast and were stored immediately at − 80 °C for subsequent assays. N-terminal propeptides of type 1 collagen (P1NP), C-telopeptide collagen crosslinks (CTXs) and 25OHD were measured via electrochemical luminescence (Roche Diagnostics, Boston, MA, USA), with intra- and interassay coefficients of variance (CVs) below 3.5% and 8.4% for CTX, below 2.6% and 4.1% for P1NP, and below 7.8% and 10.7% for 25OHD. Blood counts included white blood cell (WBC), neutrophil (N), lymphocyte (L), monocyte (M), eosinophil (E), basophil (B), red blood cell (RBC), and platelet (PLT) counts.

Definition of APR

Previous analyses of APRs have been performed in several studies, and we used a similar definition for our study [14]. The definition utilized adverse events, which were reported as occurring within 3 days of the first administration of ZOL. The adverse events were categorized according to the Medical Dictionary for Regulatory Activities (MedDRA) version 9 [15]. Preferred terms meeting the definition of APR were grouped into five symptom clusters: fever; musculoskeletal events (e.g., pain and joint swelling); gastrointestinal events (e.g., abdominal pain, vomiting, and diarrhoea); eye inflammation; and other events (including fatigue, nasopharyngitis, and oedema). Oral T was checked in each patient at 06:00 at baseline and on Days 1 to 3 and 14 after ZOL administration. If a patient’s body temperature exceeded 37.3 °C, it was remeasured every 4 h. Standardized digital thermometers were utilized for all temperature measurements. Body temperature data were centralized and stored in a single location rather than being dispersed across multiple sites or investigators. Thresholds were based on prior study recommendations [16].

Statistical analysis

SPSS v23 (SPSS Inc., Chicago, IL, USA) was used to analyse the data. The Shapiro‒Wilk test was performed to determine whether variables were normally distributed. Continuous variables are expressed as means with standard deviations. Skewed data are presented as the median with the interquartile range (25–75%). Classification variables are expressed as percentages. Data that were not normally distributed (all blood counts) were log transformed to normality before being used in the statistical analyses. After normality, independence, and equal variance (homoscedasticity) of the variables were checked, two-sample t tests or Mann‒Whitney U tests were used to analyse differences between groups. After the independence of variables (noncollinearity), normality of the residuals, independence between observations, and equal variances were checked, multivariate linear regression models were constructed to analyse the correlation between dependent variables, such as BMD, and independent continuous variables, such as the change in temperature, PINP, CTX, and blood count indices. The SPSS PROCESS macro programme was used to test the mediating effect of CTX on the relationship between T and BMD. The analysis facilitated estimation of the indirect effect with a normal theory approach and a bootstrap approach to obtain confidence intervals [17]. In PROCESS, Model 4 with 10,000 interactions was established to determine the mediation of the regression models, CTX with T and BMD. The regression coefficient of T to CTX was denoted as path a, as shown in Fig. 2. Second, the regression coefficient of T to BMD as the total effect was denoted path c. Third, the regression coefficients for both T and CTX to BMD as the direct effect were denoted paths b and c. Finally, the indirect effect was examined by 95% bootstrapped confidence intervals (CIs) using 10,000 bootstrapped samples. The receiver operating characteristic (ROC) curve and area under the curve (AUC) were used to estimate the utility, and the cut-off value, sensitivity and specificity were calculated simultaneously. Statistical significance was set at P < 0.05.

Fig. 2

Mediating effect of change in CTX between change in T and change in BMD. CTX: cross-linked C-telopeptide of type I collagen; T: temperature; BMD: bone mineral density AUC: area under the curve; BMD: bone mineral density; LS: lumbar spine

Results

Baseline and follow-up characteristics of the studied population

The baseline and follow-up characteristics of the subjects are shown in Table 1. The body T increased significantly within 3 days after the administration of ZOL and returned to the baseline level on Day 14. The BMD of the lumbar spine (LS) and femur neck (FN) increased significantly at 6 and 12 months after treatment. Compared with the baseline level, the BMD of the LS increased by 2.2% (Month 6) and 2.5% (Month 12), and the BMD of the FN increased by 1.7% (Month 6) and 2.1% (Month 12). In terms of bone metabolism turnover markers, the level of serum CTX decreased significantly after 2 weeks of treatment, with a decrease of 92.0%. The level of CTX increased slightly at 6 and 12 months after treatment, but it was still significantly lower than the baseline level, with decreases of 80.0% and 71.5%, respectively. P1NP did not change significantly 2 weeks after treatment but decreased at 6 and 12 months after treatment, with decrease rates of 62.9% (Month 6) and 56.0% (Month 12), respectively. The WBC count increased 2 weeks after treatment, the proportion of neutrophils increased substantially, and the percentages of lymphocytes and monocytes decreased.

Table 1 Baseline and follow-up characteristics of the studied population

Adverse events occurring within 3 days of infusion and treatment-related changes in BMD caused by APRs

As shown in Table S1, 75.9% of the subjects had APRs within 3 days after treatment, of which fever and chills were the most common, accounting for 64.8% of the subjects, followed by musculoskeletal pain, accounting for 38.9% of the subjects, and gastrointestinal, ophthalmic and other symptoms accounted for 13.0%, 3.7% and 29.6% of the subjects, respectively. The subjects were categorized into two groups by APR occurrence: APR+ (i.e., with APR) and APR- (i.e., without APR) (Table 2). The results revealed no significant difference between the two groups in BMD changes in the LS, FN or total hip (TH) at Months 6 and 12 (Table 2).

Table 2 Treatment-related changes in BMD by APR occurrence

Relationship between factors and change in BMD adjusted for age and BMI

Multivariate linear regression models were constructed to analyse the correlations between changes in body T within 3 days and 25OHD at baseline, and changes in BTM (baseline vs. 12 months) and changes in BMD (baseline vs. 12 months) are shown in Table 3. The results showed that after adjustment for age and BMI, changes in T (baseline vs. 3 days) were positively correlated with changes in LS BMD (baseline vs. 12 months) (β = 0.270 P = 0.034), and changes in CTX (baseline vs. 12 months) were negatively correlated with LS BMD (baseline vs. 12 months) (β=-0.486 P = 0.000). The change in PINP (baseline vs. 12 months) was negatively correlated with the change in LS BMD (baseline vs. 12 months) (β=-0.409 P = 0.001). The 25OHD at baseline was not correlated with the change in BMD. None of the factors were correlated with changes in FN BMD or TH BMD (baseline vs. 12 months).

Table 3 Relationships between factors and changes in BMD adjusted for age and BMI

Mediating effect of change in CTX (baseline vs. 12 months) between changes in T (baseline vs. 3 days) and changes in LS BMD (baseline vs. 12 months)

Table 4; Fig. 2 illustrate the results of the bootstrapped mediation analysis to determine whether CTX mediated the relationship between the change in T BMD and the change in LS BMD. With respect to the total effect, the change in T was positively linked to the change in LS BMD (b = 0.123, P = 0.00371). Furthermore, the change in T was negatively correlated with the change in the serum CTX concentration. The change in serum CTX was negatively correlated with the change in BMD. Moreover, CTX had a significant mediating effect on the relationship between the change in T and the change in the BMD of LS patients (b = 0.046, 95% CI [0.0010–0.0091]). However, there was no direct effect between the change in T and the change in BMD. The results suggested that the influence of the change in T on the change in the BMD of the LS was mainly mediated by the change in CTX.

Table 4 Mediating effect of change in CTX (baseline vs. 12 M) between change in T (baseline vs. 3D) and change in BMD (baseline vs. 12 M)

Change in T (baseline vs. 3 days) to predict the increase in LS BMD (baseline vs. 12 months)

According to the clinical consensus, a change in BMD by more than 2.77 times the precision error is defined as the least significant change (LSC) [18]. In this study, the precision errors were 0.80% for the spine, 1.05% for the femur neck and 0.97% for the total hip; thus, the LSCs were 2.2% for the spine, 2.90% for the femur neck and 2.67% for the total hip. Therefore, a change in LS BMD of more than 2.2% at 12 months was defined as a clinically significant change. The ROC curve was used to observe the sensitivity and specificity of the change in T to judge the clinically significant change in LS BMD. The results showed that when T increased by 1.95 °C, the sensitivity and specificity of judging the clinical significance change in LS BMD were optimized. The sensitivity and specificity were 68.2% and 78.1%, respectively, and the area under the curve was 0.724 (Fig. 3).

Fig. 3

Change in T to predict the increase in BMD in LS

Discussion

In this study, we investigated the hypothesis that the treatment effect in patients with APRs after the first administration of ZOL was better than that in patients without APRs. However, the results did not reveal a relationship between the APRs and changes in bone turnover markers and BMD but revealed that changes in T within 3 days after administration were positively correlated with changes in the BMD of the LS at month 12 (β = 0.270 P = 0.034). This effect was mediated mainly by changes in the serum CTX concentration (b = 0.046, 95% CI [0.0010–0.0091]). Moreover, the ROC curve revealed that when T increased by 1.95 °C, the sensitivity and specificity of judging the clinical importance of the change in LS BMD after 1 year were optimized.

APRs are common among people who are treated for the first time with ZOL. The incidence varies from 30–54.9%^{[12–13,19−20]}. APRs generally appear within 3 days following the first infusion, and common symptoms include fever, musculoskeletal pain, and gastrointestinal symptoms [19]. APRs are more common among younger subjects, those using nonsteroidal anti-inflammatory drugs (NSAIDs), and non-Japanese Asians, whereas they are less common in smokers, patients with diabetes, previous oral bisphosphonate users and Latin Americans [20]. Currently, research on APRs has focused mostly on drug safety. However, according to previous studies, the mevalonate pathway that causes APRs is also related to the efficacy of the medication. Therefore, we hypothesized that people with APRs after the first application of ZOL would experience better efficacy after treatment. At present, only 3 studies have been conducted to examine the relationship between APRs and treatment efficacy, and the results were inconsistent [21, 22, 19]. Subanalyses of the phase III ZONE study conducted in Japan demonstrated that patients with APRs presented significantly greater increases in total hip BMD at 6 and 12 months and greater suppression of BTMs compared with patients without APRs [22]. Another study conducted by Lu suggested a potential association between APRs occurrence and decreased refracture risk in patients with osteoporotic fractures and in patients undergoing orthopaedic surgery [19]. The HORIZON Pivotal Fracture Trial showed no significant difference in treatment-related changes in BMD with or without APRs; however, subjects starting ZOL who experienced APRs had a greater reduction in vertebral fracture risk [21]. In addition to racial differences, an important reason for this inconsistency lies in the choice of research methods. APRs are composed of a group of symptoms, including fever, musculoskeletal pain, gastrointestinal symptoms, and eye symptoms. In fact, most of these symptoms are self-reported by patients and are difficult to quantify and compare objectively, resulting in deviations in the results. Fever is the most common symptom in APRs. In studies in Japan and the United States, 75.7% and 47.5% of people with APRs, respectively, had fever symptoms [21, 22]. Therefore, in this study, body T, the most common and easily quantifiable indicator of APRs, was selected as the observation indicator to analyse its possible correlation with the efficacy of ZOL after treatment, including changes in BTM and BMD.

Our results revealed that when grouped by APRs, there were no significant differences in changes in BMD in the LS, FN, or TH between the two groups at Months 6 and 12, which is consistent with the results reported in the HORIZON-PFT study. However, when we used the change in body T within 3 days after the first infusion of ZOL as an observation indicator, the results showed that there wasrevealed a significant positive correlation between the changes in T (baseline vs. 3 days) and LS BMD (baseline vs. 12 months) (β = 0.270 P = 0.034). Moreover, further mediating effect analysis revealed that the influence of the change in T on the change in the BMD of LS patients was mediated mainly by the change in the serum CTX level (b = 0.046, 95% CI [0.0010–0.0091]).

Bone continuously undergoes modelling and remodelling to maintain bone metabolism and structural completion. This process of self-renewal, in which osteoclasts constantly absorb old bone and osteoblasts constantly form new bone, is known as bone turnover. Bone turnover markers (BTMs) are metabolic products or enzymes produced during bone turnover processes and are classified as bone formation or bone resorption markers. Serum CTX is a sensitive marker recommended by the IOF to reflect bone resorption [23]. Our results revealed a negative correlation between the changes in serum CTX levels at Week 2, Month 6, and Month 12 after infusion of ZOL and the changes in BMD in the LS and TH at Month 12. Although our study did not evaluate the occurrence of fractures because it was a one-year clinical trial, according to the reported studies, compared with those with the lowest quartile level of serum CTX, the risk of vertebral fracture increased by 1.4–2.2 times, and the risk of nonvertebral fracture increased by 1.8–2.5 times in the subjects with the highest quartile of serum CTX independent of BMD [24]. The greater the decrease in serum level of CTX, the greater is the decrease in fracture risk [25].

Notably, our results suggested that the impact of changes in body T on serum CTX after infusion of ZOL may be related to monocytes. The changes in T within 3 days after treatment were negatively correlated with the changes in serum CTX levels at Month 6 and Month 12. Moreover, the changes in T were also negatively correlated with the changes in monocytes in the peripheral blood at Week 2 (Table S2). Next, we subclassified the subjects into four groups according to the degree of monocyte reduction after treatment, and the changes in serum CTX were compared between the lowest quartile and highest quartile groups. The results revealed that the decrease in serum CTX was greater in subjects in the highest quartile group compared with those in the lowest quartile group, and the decrease in CTX at Month 12 was significantly different (− 78.2% vs. − 57.8%, P < 0.05) (Table S3).

Currently, both peripheral monocytes and γδ T cells are believed to be rapidly activated after treatment with ZOL, which ultimately affects the clinical severity of APR [20, 26, 27]. ZOL is taken up by monocytes, which thereby acquire the ability to ‘‘present’’ isoprenoid metabolites or related structures to γδ T cells [28]. Through mutual crosstalk, both monocytes and γδ T cells undergo a series of activation and differentiation steps that ultimately determine the severity of the APRs and replicate similar events to those that occur in acute infection [9]. Notably, the effects on circulating monocytes appear to be transient, lasting only 3 days; subsequently, rapid inactivation and extravasation occur [20]. Therefore, in our study, we observed a significant decrease in the number of monocytes at week 2 after treatment. On the one hand, the number of decreased monocytes reflects the severity of APRs in the body. On the other hand, it also reflects the sensitivity of osteoclasts to ZOL. Osteoclasts differentiate from monocyte-derived macrophage precursors and are responsible for bone absorption. Therefore, changes in the number of serum mononuclear cells after treatment can partially indicate the sensitivity of osteoclasts to ZOL. As shown in our study, the subjects with greater reductions in monocyte counts also presented greater decreases in serum CTX levels. The level of CTX reflects the activity of osteoclasts and the level of bone resorption.

Special attention should be given to a recent study conducted by Lu, which focused on elderly osteoporotic fracture patients undergoing orthopaedic surgery [19]. In that study, the long-term outcomes of APRs after initial ZOL administration were investigated. The results demonstrated that patients who experienced APR had a 73% lower rate of refracture than did those who did not, which indicated that patients with APRs experienced better drug efficacy. Our research yielded similar findings, showing that the increase in body T within 3 days after initial ZOL was related to changes in BTM and an increase in BMD after one year. However, surprisingly, Lu’s research also revealed that patients who had an APR had a 97% higher risk of mortality than patients who did not. In Lu’s study, all the subjects presented newly identified hip/morphological vertebral osteoporotic fractures and had undergone orthopaedic surgery in the hospital. Most of the subjects had received ZOL treatment for 7 − 14 days after the orthopaedic operation. In fact, acute fracture events and surgery comprise a special period for patients with osteoporosis. Research has shown that trauma itself can exacerbate the occurrence of ARPs [29]. Compared with those after nonsurgical interventions, the odds ratios of experiencing APRs after minimally invasive or open surgery were 3.54 and 5.71, respectively. Therefore, the results of Lu’s study can only represent the impact of ARPs on mortality and refracture after ZOL treatment of hospitalized osteoporotic fracture patients. However, the subjects in our study were recruited from the outpatient service, and none of them had undergone surgical treatment in the past three months. Therefore, our results are more universally applicable to patients with osteoporosis.

Both the APRs and efficacy of ZOL were related to the mevalonate pathway. Our results suggested that those patients who are sensitive to ZOL were more likely to have an increase in body T, a greater decrease in serum CTX and a greater increase in BMD after the first dose. A three-year HORIZEN-PFT study has shown that people with APRs are less likely to suffer vertebral fractures [21]. In our study, the degree of increase in body T after initial medication might indicate the sensitivity of patients to ZOL and potentially predict the treatment efficacy to some extent. Our study further revealed that an increase in body T of 1.95 °C or more within 3 days after the first administration of ZOL predicted a clinically significant increase in lumbar bone density at 1 year after the first administration of ZOL.

There are several limitations to our study. (1) While this study was conducted as a cohort study, it is worth noting that the follow-up period was limited to one year. The current view is that there is a significant change in BTM after 3 months of treatment with ZOL. A significant change in BMD was observed after 6 months of treatment with ZOL. Moreover, the fastest increase in BMD can be observed after the first year of treatment with ZOL [30, 31]. Therefore, osteoporosis efficacy, including changes in BTM and BMD, can be evaluated over one year of follow-up. However, the relatively short duration of the follow-up period might account for the observation limited to changes in BTM and BMD, with no occurrences of fractures recorded within our study cohort. Three existing studies have investigated the long-term effects of zoledronic acid in patients with and without APR. One of them is the HORIZON PFT study, which lasted for 3 years and focused on postmenopausal women. BMD and fracture occurrence were observed [21]. The results suggest that women initiating ZOL who experience an APR will have a greater reduction in vertebral fracture risk with ZOL. However, there was no significant difference in treatment-related changes in BMD. Another study conducted in Japanese patients treated with a once yearly intravenous infusion of ZOL 5 mg for 2 years investigated the relationship between APR and efficacy [22]. Patients with APRs presented significantly greater increases in total hip BMD at 6 and 12 months and greater reductions in BTMs than patients without APRs. However, fracture risk was not assessed in this study. The other is a retrospective study involving elderly osteoporotic fracture patients who underwent orthopaedic surgery, which has been discussed above [19]. (2) Despite meeting the requirements established by our power analysis, the sample size was still relatively small. (3) Given the higher prevalence of osteoporosis among postmenopausal women, the study participants in this exploratory study were limited to postmenopausal women residing in Shanghai. The results of this study present evidence that individuals exhibiting elevated body T levels following initial ZOL treatment may experience greater improvements in BMD. (4) In this study, certain external variables, such as lifestyle and nutritional status, were not collected, which might compromise the robustness of the findings. To gain a more comprehensive understanding of the long-term efficacy, safety, and sustained effects of ZOL, further studies with longer follow-up durations are needed. In future studies with longer follow-up durations (e.g., longer than 2 years), the change in occurrence of fractures can be better evaluated. In addition, in further research, external variables such as lifestyle, exercise and dairy intake will be collected to increase the validity of the results.

In summary, the results of this study indicated for the first time that (1) there was no significant correlation between the occurrence of ARPs within 3 days of the first administration of ZOL and the change in BMD after 1 year, whereas the increase in body T positively correlated with the increase in BMD, suggesting that the increase in body T predicted better treatment. (2) The correlation between the increase in body T and increase in BMD was mediated by the change in serum CTX. (3) An increase in body T of 1.95 °C or above predicted that the BMD in the LS would increase significantly 1 year after the first administration of ZOL.

In this study, we tested the hypothesis that people with elevated body T after initial ZOL treatment had greater improvements in BMD and better outcomes. The establishment of this hypothesis will be conducive to the prediction of the clinical curative effect and guide doctors in timely intensive treatment for people with poor curative effects. It is widely recommended in guidelines that the measurement of main serum bone turnover biomarkers, namely, CTX and PINP, should be performed over a time period ranging from 3 − 6 months posttreatment for assessment of drug efficacy. However, the use of a body temperature monitor allows for a timelier, cost-effective, and convenient assessment in comparison, particularly in regions where access to these laboratory tests may be limited. More importantly, since APRs are very common in patients with initial ZOL treatment, the establishment of this hypothesis will help alleviate patients’ fear and anxiety surrounding APRs, improve patients’ confidence and compliance with antiosteoporosis therapy, and increase humanistic care for patients.