Primary hyperparathyroidism (PHPT) is a common cause of hypercalcaemia among outpatients [1] . It is estimated that 0.3% of the general population have PHPT and 1% - 3% of postmenopausal women suffer from it [2] . Surgical treatment is clearly indicated in PHPT patients who present with severe hypercalcaemia, renal stone, osteoporosis and reduced renal function [3] . Nonetheless, studies have reported that 80% of patients who present with PHPT are asymptomatic (serum calcium < 3 mmol/L and no renal stones) [4] . It has been suggested that conservative approach may be more reasonable than surgical intervention in the management of stable asymptomatic PHPT [5] .

Bone remodelling involves the removal of mineralized bone by osteoclasts (bone resorption) followed by formation of bone matrix by osteoblasts (bone formation) that is subsequently mineralised. Osteoporosis is characterised by an imbalance between osteoclast and osteoblast activities (bone uncoupling), resulting in impaired bone remodelling. Raised parathyroid hormone (PTH) in PHPT increases bone loss by stimulating the osteoclastic bone resorption [6] . Increased rate of bone turnover is usually determined by measuring the serum concentration of C-terminal telopeptide (CTx), which is a cross-link peptide sequence of type I bone collagen. Biochemically, raised PTH, due to its direct effect on bone, increases not only CTx but also calcium both in blood and urine. Clinically, this results in reduced BMD, osteoporosis, impairment in renal functions and renal stones.

Studies have shown that anti-resorptive treatment in the form of bisphosphonates is as effective as surgical treatment in improving BMD in patients with PHPT [7] . Bisphosphonates are analogues of inorganic pyrophosphates. Bisphosphonates attach to hydroxyapatite binding sites on bony surfaces, especially surfaces undergoing active bone resorption. When osteoclasts begin to resorb bone that is impregnated with bisphosphonate, the bisphosphonate released during resorption impairs the ability of the osteoclasts to form the ruffled border, to adhere to the bony surface, and to produce the protons necessary for continued bone resorption. In addition to this, bisphosphonates causes apoptosis of osteoclasts by inhibiting farnesyldiphosphate synthase in osteoclast cells, resulting in disruption of a number of pathways including cytoskeletal organisation, cell proliferation, and survival [8] and thus reduces osteoclast activity. Bone formation is often reduced by bisphosphonates, which is probably an indirect effect of inhibition of bone resorption. Bisphosphonates that contain nitrogen (such as alendronate, risedronate, ibandronate, and zoledronic acid) have the most potent antiresorptive properties and are the most commonly used drugs in the treatment of osteoporosis. Large randomized controlled trials in post menopausal osteoporosis women have shown significant reduction in vertebral, non-vertebral and hip fractures [9] with bisphosphonates. Both the National Institute for Health and Clinical Excellence (NICE) UK and the UK National Osteoporosis Guideline Group recommend bisphosphonate as the first line treatment option for primary or secondary prevention of fracture in postmenopausal women [10] [11] .

The objective of this retrospective observational study was to primarily assess the effect of long-term bis- phosphonate therapy on BMD in patients with PHPT. Besides this, we also examined the effect of long-term bisphosphonate therapy on bone fragility fractures and biochemical markers of bone turnover in PHPT.

We observed a significant increase in lumbar spine BMD in patients with PHPT, who received bisphosphonate

IQR: Interquantile range, ^*p value < 0.05 significant.

PTH: parathyroid hormone; CTx: C-terminal telopeptide; SD: standard deviation.

therapy over an average period of 5 years. Anti-resorptive treatments such as bisphosphonates have shown a comparable increase in BMD as observed after surgical intervention in PHPT or in eucalcaemic population [7] . Since PHPT confers an increased risk of bone loss, bisphosphonates may be a reasonable alternative to surgery in patients with PHPT. PTH is preferentially known to affect the cortical bone more compared to trabecular bone [14] . Paradoxically, our study patients had lower bone density at lumbar spine (trabecular bone) compared to left femoral neck (cortical bone) before they were commenced on bisphosphonate therapy. Studies have shown that this effect of PTH may be due to its long-term biphasic action, which causes bone loss throughout the skeleton [7] . Bisphosphonates, by inhibiting the bone resorption, increase BMD in PHPT patients. We observed this affect in our study patients who received bisphosphonates for an average period of 5 years. Bisphosphonates are known to increase BMD significantly both at lumbar spine and left femoral neck density [15] [16] in PHPT. Nevertheless, we observed a significant increase in lumbar spine density and no significant change in left femoral neck density in our PHPT patients. The reason for this observed effect may be that studies so far have reported the short-term effect of bisphosphonates (2 years) on BMD in PHPT [7] [16] where as we studied the long-term effect of bisphosphonates on BMD. Besides bisphosphonates, hormone replacement therapies such as estrogens have also been reported to increase BMD in PHPT [17] [18] .

The fracture risk in PHPT increases by 1.5 to 2 fold compared to eucalcaemic population [19] . The PHPT patients in our study presented with fractures involving both trabecular and cortical bone (Table 3). This is contrary to the action of PTH on skeleton, which predominantly results in greater cortical bone loss [14] . However there is no clear evidence that in PHPT, fracture risk will be higher in skeleton predominantly rich in cortical bone [19] . We did not observe any significant increase in fracture rate in our cohort of study patients after receiving bisphosphonate therapy over an average period of 5 years. This was reflected by a non-significant change in bone resorption marker CTx at the end of treatment. Bisphosphonates are well known to reduce the fracture risk in eucalcaemic osteoporosis patients [20] . Thus, it has been presumed that bisphosphonates will have a similar action on bone fragility fractures in PHPT patients. Before starting bisphosphonate treatment 14 PHPT patients (28%) in our study had bone fragility fracture (baseline). In this cohort, only one patient had a recurrent fracture after 4 years of bisphosphonate treatment. Thus, bisphosphonates were able to prevent recurrent bone fragility fractures in all 14 patients except one, which corresponded with the significant increase in BMD. During the bisphosphonate therapy, 6 PHPT patients had new bone fragility fractures after an average period of 6 years of post therapy (Table 3). In total, there were 7 patients (including one who had recurrent fracture) patients (14%) who had fragility fractures when they were on bisphosphonate treatment. The mean serum 25(OH) vitamin D levels in these seven patients was 27.9 nmol/L (25(OH) vitamin D below 30 nmol/L suggest deficiency), suggesting the association of low vitamin D status and bone fragility fracture [21] . Therefore, we presume that vitamin D deficiency may have contributed to the bone fragility fractures in our PHPT patients in spite of being on bisphosphonate treatment. Recent meta-analysis suggests that optimal vitamin D levels (25(OH) vitamin D above 60 nmol/L) are most beneficial in reducing the risk of hip and non-vertebral fractures in eucalcaemic elderly population [22] . There is no evidence that correcting low vitamin D status will reduce the bone fragility fractures in PHPT patients. Nevertheless, observational studies show an inverse association of size of parathyroid adenoma, bone turnover, levels of serum calcium, PTH and trabecular BMD with 25(OH) vitamin D [23] [24] . Thus it has been recommended to correct the low vitamin D status and maintaining 25(OH)D above 50 nmol/L in PHPT [5] . Although there was a reduction in fracture rate in our study patients, this failed to achieve significance presumably due to a small sample size. None of the previous studies have reported the effect of anti-resorptive treatment on fracture risk in PHPT patients.

We observed a significant reduction in calcium both in blood and urine following bisphosphonate therapy suggesting that bisphosphonates are able to ameliorate the effect of raised PTH on bone resorption. Previous studies have shown a similar finding [25] [26] emphasising the use of bisphosphonates in the management of hypercalcaemia in PHPT.

Additionally, there was no deterioration in renal function with bisphosphonate therapy in our study patients (Table 4). Usually, hypercalcaemia in PHPT results into deterioration in renal functions as it induces nephrogenic diabetes inspidus [27] . Since there was a significant improvement in serum corrected calcium levels following bisphosphonate therapy in our cohort, we did not observe deterioration in renal function. At the same time, several studies have reported no decline in renal functions in PHPT [28] .

Our study suggests that long-term bisphosphonate therapy in PHPT improves lumbar BMD, prevents increase in rate of bone fragility fractures and improves hypercalcaemia. We recommend consideration of long-term bis- phosphonate therapy in patients with PHPT. Randomized controlled trials are required to ascertain effect of vitamin D supplementation on fragility fracture rate in PHPT.

Our study does have some limitations. Firstly it is an observational retrospective study and cause-effect relationship cannot be ascertained in this design. However, the findings of our study do concur with previous randomised control trials, which have shown increase in BMD with bisphosphonates. Secondly, it is likely that differences were not observed in some variables due to the size of the cohort. The sample size calculation was not indicated as we had included all patients with PHPT who had their records available.

Further prospective randomised controlled trials are required to assess the effect of bisphosphonate therapy and vitamin D supplementation on bone fragility fractures in PHPT patients.