Bone fractures are a complication that greatly affects quality of life. In addition, immobility caused by fractures increases mortality [11]. KTx recipients have an increased risk of fractures. Recently, Evenepoel et al. observed in 518 KTx recipients within a 5.2 year follow-up a fracture incidence of 14.1 per 1000 person-years [12]. The authors found that low BMD was associated with fractures, independent of classic determinants, including history of fractures. In our larger study with a comparable length of follow-up, the incidence was 10 per 1000 person-years, but unlike the incidental fractures counted by Evenepoel et al., we excluded fractures related to adequate trauma or malignancy.
A recent Cochrane analysis addressed the question of the efficacy of different treatments for the prevention of MDB after KTx. The primary efficacy endpoint considered was bone fracture. It remained uncertain whether, apart from bisphosphonates, any other class of drugs reduced fractures. In the absence of trial data, the effect of PTX on fracture risk also could not be adequately assessed [13].
In chronic hemodialysis patients, Rudser et al. showed that parathyroidectomized patients have a reduced fracture risk. Several mechanisms are discussed as potential mechanisms. First, PTH excess is avoided and therefore high bone turnover lesions are mitigated. Second, PTX-induced hungry bone syndrome, which increases bone mineral uptake, may inherit long-term protective effects on fractures [14]. A protective role of PTX against fracture events has already been shown for primary HPT [8], although the mechanisms in both settings are not comparable. In 1994, Grotz et al. even reported in 100 KTx patients that pretransplant PTX increased the risk of post-transplant fractures [15]. However, this has not been confirmed in subsequent studies, and both drugs and surgical techniques and indications for PTX have changed since then [16]. In contrast, in the present era (of calcimimetics, which probably cannot reduce fracture risk in ESRD or KTx patients [17],[18][13] we herein generate evidence that PTX may act as an independent protective factor against pathologic fractures in KTx recipients with HPT. This association persists when adjusted for the well-known risk factors of age, sex and dialysis vintage, as well as for underlying renal disease, which was associated with fracture events in univariate analysis. Since eGFR at different time points was inconsistently identified as significant risk factor for fractures in univariate Cox proportional hazard regression analysis (after two and three years, but not after 1,4 and 5 years, Fig. 1A), we did not include this parameter in further analysis. Moreover, the addition of eGFR as reflection of renal function besides age to multivariable analysis seems problematic, because age is already considered in the eGFR estimation formula (CKD-EPI formula) (Levey et al. 2009). Further, age is known to be strongly associated with graft function and outcome in KTx recipients (Legendre et al. 2014). Reasons for this include the fact that living donations, which have a better graft outcome, are more common in younger recipients (Hart et al. 2019) and organs provided under the European senior program (ESP) have lower organ quality because they meet the expanded donor criteria (ECD) (Pascual et al. 2008).
Isaksson et al. matched a cohort of 590 PTX patients on dialysis or with functioning allograft for assessment of fracture risk [19]. It was observed that PTX reduced hip fracture risk, but only in female patients, compared with non-PTX patients. For this study, the authors distinguished only between KTx and non-KTx patients at the time of PTX. Whether patients received KTX after PTX was not considered. Nevertheless, these results point in the same direction as ours do.
Another study analyzed hypercalcemia control with subtotal PTX compared with cinacalcet use in KTx recipients. Cruzado et al. showed a superior effect of PTX for calcium and PTH normalization and increase of bone mineral density but due to the relatively short follow-up period of twelve months and small patient numbers, they were not able to sufficiently comment on fracture events [20].
Based on data suggesting that hypoparathyroidism after PTX in KTx recipients correlates with a significant decrease of renal function, PTX should be indicated with caution, and the initial treatment approach should be pharmacological with administration of calcimimetics [6, 21, 22]. This approach was largely followed in our center. Interestingly, Mathur et al. did not observe an association between treatment of sHPT and posttransplant delayed graft function, graft failure or death, but the proportion of PTX-treated patients in their cohort of 5094 KTx recipients was small (4.5%) [23]. However, our data do not show a significant difference in eGFR courses between patients with and without PTX (Table 3).
In the group of patients in whom a drug approach was followed to lower PTH, it seems trivial that the PTH value remains elevated (Fig. 3) [24]. This is paralleled by higher calcium levels and lower phosphate levels (Table 3). Nevertheless, there is also evidence to support a preference for PTX over the use of cinacalcet. In patients with tertiary hyperparathyroidism after KTx, cinacalcet can normalize serum calcium levels, but unlike PTX, it cannot normalize PTH levels[25] [26]. However, there is no consensus on which PTH value clearly defines post-transplant HPT [27], but PTH levels were found to be an important negative independent predictor of BMD, intriguingly, more deleterious than the cumulative dose of corticosteroids or inflammation [5]. In our cohort, one year after KTx PTH levels have decreased 2.2-fold compared to pre KTx. However, one year after KTx only 23.8% of the patients showed PTH levels within the by KIDGO recommended range for patients with CKD, 71.7% had values above and 4.5% below (see Results 3.2.4). [6]. This phenomenon has been observed in many studies [27]. Looking at both groups (PTX vs. non-PTX), there was no difference in PTH levels pointing to the fact that many patients had undergone subtotal PTX and controlled HPT in the other group. Considering that the group without fractures included 17.8% PTX patients (vs. 2.2% in the fracture group), it becomes clear that the average PTH level in the group without fractures was higher in the non-PTX patients and therefore is probably not the decisive influencing factor (Table 3). Nevertheless, there is evidence that persistent HPT could be a risk factor for fractures after KTx. Therefore, it could be important that parathyroid recovery from CKD-induced HPT was incomplete even 1 year after KTx in a large subset of patients (Fig. 1). While some studies link PTH-related stimulation of bone turnover to fractures [10], other sources describe (circulating) PTH levels as poorly predictive of underlying bone turnover [28]. In contrast to Perrin et al., who described a PTH above a cut-off of 130 pg/ml measured three months after KTx as a significant risk factor for fracture events in a cohort of 143 KTx recipients with a total of 22 fracture events [10], we could not confirm this observation in our data (Fig. 4). Moreover, according to our data, we could not support a clear cut-off for PTH being associated with an increased fracture risk (Suppl. Figure 1). In congruence with Perrin et al., we did not observe alkaline phosphatase as a marker for bone turnover elevated in the patients with fracture events. However, in contrast to Perrin et.al, we found alkaline phosphatase elevated in those patients without PTX. Nevertheless, it seems possible that bone abnormalities induced by HPT can be explained by the elimination of skeletal resistance to PTH occurring during CKD after renal transplantation.
VDRa (1,25(OH)2D3 (Calcitriol)) are more frequently applied in patients after PTX than in patients without PTX (Table 3). VDRa provide a well-studied protective effect on bone metabolism, whereas the efficacy of nutritional vitamin D preparations, frequently used in non-PTX KTx patients, has not been clearly established yet [29]. However, VDRa use did not differ in patients with and without fractures (Table 2).
In KTx recipients, immunosuppressive therapy is a specific contributing factor to MBD and osteoporosis. In addition to the osteoporotic effects of long-term steroid use [30], CNI medication may also affect bone metabolism. Luo et al. described increased bone resorption in patients with tacrolimus trough levels above 6 ng/ml [31]. Increased bone loss from CNI use has also been demonstrated in a rat model [32]. In contrast, we did not find an association between higher tacrolimus levels or tacrolimus metabolism, represented by the Tac C/D ratio [33], and pathologic fracture events in our cohort. Nevertheless, Tacrolimus may influence bone resorption to an extent that is clinically not apparent and is overshadowed by more relevant factors. Recently, the effect of steroids on BMD was prospectively analyzed in de novo KTx patients using steroid minimization protocols. The authors showed that a cumulative methylprednisolon dose of 2.5 ± 0.8 g (after 1 year) and 5.8 ± 3.3 g (after 5 years) caused only limited BMD changes and was predominantly related to remodelling activity rather than corticosteroid exposure [34]. Since treatment per protocol in our center results in exposure of approximately 3.4 g prednisone after 1 year and 10.7 g after 5 years (in an average patient on steroids without rejection), we cannot exclude a relevant influence of steroids on BMD in our cohort. However, steroid use and rejection rates did not differ in the groups.
Interestingly, polycystic kidney disease (ADPKD) as an underlying disease for ESRD was associated with the incidence of fractures after KTx in univariate analysis. This fits with the observations of Gitomer et al. who describe a bone metabolism defect in ADPKD patients with CKD stages 1 and 2, although ADPKD was not associated with an increased risk of fractures in ESRD patients [35]. Nevertheless, we could not confirm this association in the multivariate analysis.
This study has several limitations. Due to its retrospective nature, it is only hypothesis generating. Since we did not assess corticosteroid, VDRa, vitamin D doses or levels, we cannot comment on that and cannot exclude related effects. However, native vitamin D is routinely prescribed to all patients without hypercalcemia (weekly dose: 7.000–10.000 IE), because low dose prednisone therapy (~ 5 mg/d) is used in most patients in our center (data not shown). Moreover, additional information on FGF23 and bone density would have provided further valuable information on the bone homeostasis and morphology of patients after PTX compared to those without PTX.
In conclusion, our study demonstrates an association between PTX and fracture events after KTx, which is independent from elevated PTH levels at the time of KTX or afterwards. Therefore, keeping in mind that PTX surgeries were performed considerably prior to KTx, fracture events might be the result of a long-lasting sHPT in the foregoing history of patients without PTX to a significant extent. As PTX does not influence eGFR after KTx and offers protective effects on fracture risk in our study, our data supports a more generous indication for PTX in ESRD-patients with HPT waiting for KTx.