CKD serum contains a mixture of PTH fragments that progressively accumulate as eGFR falls below 20 mL/min/1.73 m².[12] These fragments, along with post-transcriptional modifications of PTH can interfere with PTH function and measurement.[12]

The development of immunoassays has mitigated cross-reaction issues to some extent. The widely used 2nd generation assay, known as intact PTH (iPTH), improves upon the 1st generation by using two different antibodies: one against the C-terminus and another against the N-terminus (1-34) (Fig. 1). Although this method improves specificity for (1-84) PTH, it still cross-reacts with N-terminal truncated PTH fragments, believed to be the 7-84 fragment.[13] However, mass spectrometry has challenged this, suggesting other interfering fragments such as 34-84 or 34-77.[12]

The 3rd generation assay, known as whole PTH, targets the N-terminus (1-4), presumably capturing only (1-84) PTH. Measured concentrations are usually 30% to 50% lower than iPTH.[13,14] However, this assay can cross-react with amino-PTH (PTH with phosphorylation at serine 17) and may, in rare cases, cause higher whole PTH concentrations than iPTH.[15,16] Due to unclear evidence supporting whole PTH over iPTH in predicting bone turnover, KDIGO 2017 continues to recommend the 2nd generation PTH assay in clinical practice.[8,14]

Oxidized forms of PTH, which cause partial or complete loss of function, are proposed to form in CKD and cross-react with PTH assays.[17] However, current assays do not show additional benefits in differentiating bone turnover compared to whole PTH and even question their existence due to undetectability by mass spectrometry.[12,18,19]

Another issue with iPTH and whole PTH assays is high inter-assay variability.[13] This could restrict result comparison across assays and hampers defining diagnostic cut points for disease. The recent publication proposes using LC-MS to standardize 2nd and 3rd generation PTH by regression equation, potentially harmonizing PTH measurement in the future.[19]

Increasing serum PTH in CKD partly compensates for the skeletal system hyporesponsiveness to normalize serum calcium and bone turnover rate.[20,21] This condition is attributed to multiple mechanisms such as altered function or expression of PTH1R and CaSR.[22] Additionally, the variation of parathyroid responsiveness is influenced by race, primary renal disease, comorbidities, mineral metabolism and treatment.[23] For example, African American and Japanese populations tend to have a higher degree of resistance than Caucasian populations.[24,25] This could explain the broad range of diagnostic cut points when applying PTH as a diagnostic test, in particular discriminating low from non-low turnover ROD, as shown in Table 2.

The distribution of bone turnover subtypes changes as CKD progresses. A biopsy study of 56 patients in CKD G3-G4 showed that about 35% had low turnover ROD, with high bone turnover increasing from 4% in CKD G3 to 36% in CKD G4.[30] In dialysis patients, there has been a shift towards a higher prevalence of low bone turnover, partly due to changes in clinical practice over the last decade.[31] While mineralization defects have become rare, they still occur in up to 40% of the population in areas with high aluminum toxicity.[32]

PTH is widely available and commonly used to manage CKD patients. KDIGO suggests maintaining PTH levels at 2 to 9 times the upper normal limit (UNL) for the assay in CKD G5D, based on epidemiological data showing the lowest mortality rate.[33] However, this range is not necessarily a target for optimizing bone health, as a high proportion of patients within this target still have abnormal bone turnover ROD.[14]

For diagnostic purposes, iPTH is still a reasonable tool for distinguishing between high and non-high bone turnover ROD, with levels around 4 to 5 times above the UNL showing area under the curves (AUCs) of moderate accuracy (0.7-0.8).[14,34-36] Conversely, iPTH shows inconsistent performance in distinguishing low from non-low bone turnover ROD, with one study reporting AUCs of poor accuracy (0.56) with a cut point of 2.3 times above the UNL. [14,34-36] This variability can be attributed to differences in PTH responsiveness and assay noise.

BTMs have gained interest as alternative diagnostic tools. Recent studies on the diagnostic accuracy of BTMs and PTH in CKD are summarised in Table 2 and 3. For both low and high bone turnover, bone ALP and intact P1NP show promise, with AUCs ranging from 0.7-0.8 across various CKD populations. For example, Salam et al.[34] proposed a cut point of bone ALP <20 IU/L by IDS assay (equivalent to <30 IU/L by Quidel assay) with a negative predictive value at 96% to rule out low bone turnover ROD in CKD G4-5D.

In cases of suspected mineralization defects, such as underlying renal tubulopathy affecting serum calcium and phosphate, and untreated renal tubular acidosis, a bone biopsy remains necessary to confirm the diagnosis.[37] Elevated bone ALP and ALP levels can indicate bone mineralization defects. PTH can also increase in response to low calcitriol, irrespective of bone turnover status. Patients with aluminum toxicity can be symptomatic, including bone pain, pruritus, higher phosphate levels, and lower PTH compared to non-aluminum toxicity groups.[38] In regions without aluminum toxicity, dialysis patients with mineralization defects tend to have longer dialysis duration, increased bone ALP and increased PTH compared to unaffected groups with low/normal calcium and vitamin D levels.

Cross-sectional studies reported inconsistent results. Some found a positive relationship between prevalent fractures in CKD G2-5D and increased PTH, bone ALP, TRAP5b, and P1NP, while others did not.[39-41]

Regarding predicting fracture incidence, PTH and bone ALP are the most commonly studied markers. CKD stages G1-G5 with SHPT during follow-up show a 30% increased fracture risk compared to those with PTH in the normal range.[42] A study of 5100 CKD stages G3-G4 found a linear relationship between fractures and PTH levels.[43] The dialysis outcomes and practice patterns study (DOPPS) study, involving 28,888 hemodialysis patients from nine countries, found a weak linear relationship between normalized PTH levels (ratio of PTH to the reference) and fractures, with the lowest hazard ratio of fracture in normalized PTH <1.[23] The current management of secondary hyperparathyroidism, a multi-center observational study (COSMOS) cohort study of 6,797 hemodialysis patients in Europe found a 4% increase in fracture incidence for each 100 pg/mL increment of PTH. Baseline PTH >800 pg/mL compared to 300 to 800 pg/mL increased the hazard ratio of fracture by 60% and recurrent fractures doubled, retaining significance in time-dependent models.[6] Both DOPPS and COSMOS adjusted for multiple risk factors and the competing risk of death, but only COSMOS lost significance after adjustment.

Previous studies found that PTH <150 or 300 pg/mL was associated with an increased fracture risk, given the J curve association between PTH and fracture risk.[44-46] These associations failed to replicate in recent large studies, possibly due to different populations and treatment patterns, as studies before the year 2,000 might include cases with aluminum toxicity. Different adjustment factors and modeling methods could also contribute to result variations. DOPPS showed a J-curve association between PTH levels and mortalities.[23] Possibly, in the low PTH spectrum, the higher risk of death attenuates the fracture risk association after adjustment for competing mortality.

Serum ALP and its major isoform bone ALP can predict fracture incidence. Several studies in hemodialysis patients showed increased bone ALP at baseline and during follow-up associated with increased fracture incidence.[40,46,47] Larger population studies measuring serum ALP showed similar trajectories, with increasing baseline ALP in linear association with fracture.[23,48] Stratifying patients based on PTH and ALP, high serum ALP drives higher fracture event rates in the same PTH strata, producing a stronger association than high serum PTH, indicating the robustness of ALP in predicting fracture outcomes.[23,48]

Osteoporosis treatments are classified into antiresorptive and anabolic agents. Anti-resorptive agents, which include bisphosphonates and denosumab, slow down the bone turnover rate, allowing time for mineralization. Anabolic agents, including PTH and PTHrP peptides, form new bone through activated remodeling and modeling-based bone formation. Romosozumab has a dual effect of increasing bone formation and, to a lesser extent, decreasing bone resorption.

Most bisphosphonates are not licensed for use in patients with a GFR <35 mL/min/1.73 m², except for risedronate, which is extended for use until a GFR <30 mL/min/1.73 m². The restrictive use in CKD is due to concerns about accelerated CKD progression by 15% in patients with a GFR <45 mL/min/1.73 m².[66] This limitation makes the drug less appealing in CKD G3b-G5 despite some evidence suggesting BMD improvement and no acute kidney injury or hypocalcemia.[66,67]

In CKD G5D, some studies reported LS BMD improvement.[55,57,68] For example, a randomized study of CKD G5D patients receiving either denosumab or IV alendronate showed significant improvement in LS BMD by 5.6% at 12 months compared to baseline. The study also reported suppression of TRAP5P, bone ALP, and P1NP from 3 months through the end of the study at 12 months, except for a rebound of P1NP after 6 months. However, some studies report contradictory results, showing no BMD improvement in either the LS or hip regions.[53,64]

Another challenge of using bisphosphonates in CKD G5D is the unknown optimal regimen. Overaccumulation of drugs in the body is a concern for dialysis patients due to diminished kidney clearance, a tendency to develop low albumin levels, and body volume expansion, all of which can alter the total free drug in body fluids. Augmenting clearance through hemodialysis might be a possible solution for some dialysable bisphosphonates.[69,70] However, intensive pharmacokinetic and pharmacodynamic studies are required to determine the suitable regimen for both the drug and dialysis.

Denosumab, a monoclonal antibody against the receptor activator of nuclear factor κ-B ligand, is licensed for use in all GFR ranges without dose adjustment.[71] Its benefit in CKD G5D was confirmed in multiple small studies, showing BMD improvement of 2.5% to 4.5% at the hip and 5.6% to 5.7% at the LS over a 1-year treatment period.[56,57,72]

The decline of serum TRAP5b can be observed as early as two weeks after treatment, continuing until it reaches its nadir at 40% to 60% below pre-treatment levels at three months. This is unlike bone ALP, which gradually declines over 1 to 2 years of treatment by 10% to 40% of the initial level.[57,72-74] These effects could be sustained for up to four years while on treatment.[73]

One study suggested hip BMD gain could plateau after 3.5 years of treatment.[56] This early plateau compared to the general population warrants further exploration into the association between BTMs and BMD to understand different responses in this population. When prescribing denosumab, serum calcium should be closely monitored as severe hypocalcemia could occur in up to 10.5% of CKD G5D patients, which can lead to ventricular arrhythmia, seizures, and death.[75] Low baseline bone ALP is associated with hypocalcemia and high baseline TRAP5b at 670 unit mU/mL can predict the risk of hypocalcemia events with sensitivity and specificity >0.7.[72] These markers could alert physicians to high-risk groups.

Off-label use of teriparatide, a PTH peptide, at doses of 20 mcg/day or above improved hip and LS BMD in CKD G5D but not with reduced doses.[52,53,64] Over one year of treatment, two studies reported bone ALP rising four weeks after commencing treatment and maintaining elevated levels for up to six months to one year. TRAP5b showed persistent elevation during treatment in one study, while another reported a significant decrease after six months. [51,52]

As CKD progresses, adverse effects such as hyperuricemia and hypercalcemia tend to increase.[60] Additionally, high doses of teriparatide could cause transient hypotension and myalgia, leading to nearly half of the participants withdrawing from trials within one year.[51,52] Follow-up CAC scores over one year of treatment did not show progression compared to placebo, which is reassuring in terms of safety in this population.[64]

Romosozumab, a monoclonal antibody targeting sclerostin, is approved for treating postmenopausal osteoporosis. A small study in CKD G5D revealed an increased BMD of 9% at the LS and 2.5% at the total hip over one year of treatment and up to 14.9% and 5.4% with sequential denosumab for one year thereafter.[76] Achieving BMD at one year is about 2/3 and 1/3 of the figures observed in postmenopausal studies at the LS and total hip, respectively, corresponding to subtle BTM changes.[77,78] TRAP5b gradually decreased over 12 months. ALP increased by 30% at months 1 to 3 and total P1NP increased by 15% at 3 and 6 months before both returned to their baseline thereafter.

Despite impressive BMD figures, using this medication is not without risk. Calcium scores at the coronary artery and thoracic aorta significantly progressed after 1 year.[76] Although progressive vascular calcification was not unexpected, whether it is accelerated by romosozumab cannot be answered due to the lack of a control group. This raises concerns about increased vascular calcification from a pathophysiological viewpoint and the safety signal of cardiovascular mortality from postmenopausal osteoporosis trials, suggesting that this medication should be used cautiously in the CKD population.[77]