Targets for parathyroid hormone in secondary hyperparathyroidism: is a “one-size-fits-all” approach appropriate? A prospective incident cohort study

In the 70’s, secondary hyperparathyroidism (SHPT) was described as a severe bone disease occurring in young end-stage renal disease (ESRD) patients with significant duration of dialysis. When parathyroid hormone (PTH) was very high, up to 1000 ng/L, associated with hypercalcemia, the only treatment was subtotal parathyroïdectomy [1, 2]. In the 90’s, access to kidney transplantation for the young and dialysis for the old led to a rapid ageing of dialyzed population [3]. During the first decade of the millennium, a new paradigm emerged. First, SHPT was considered not only as a bone disease, but also as a vascular disease [4]. Second, SHPT turned out to be a biological rather than a clinical disease at bedside [5]. A “one-size-fits-all” approach was recommended: PTH under 300ng/L from 2003 to 2009 according to K-DOQI [6]. Due to variability in PTH measurement, the target was modified in 2009: “maintaining PTH levels in the range of approximately two to nine times the upper normal limit for the assay” [7, 8]. Third, cinacalcet tends to be seen by the clinicians as the most appropriate solution for the treatment of SHPT due to its mechanism of action, when conventional therapy is not effective enough [6, 9]. But the randomized controlled EVOLVE study, published in 2012, failed to demonstrate efficacy of cinacalcet to reduce the risk of death or major cardiovascular events [10]. Today, the exact importance of PTH is still debated [11].

Large observational cross-sectional studies about SHPT have recently been published [12‐15]. An incidence/prevalence bias may have hampered a precise description of SHPT phenotypes [16]. In order to capture the phenotypes of SHPT at bedside, we meticulously enrolled all dialysis patients of the REIN registry - Region of Lorraine with newly marked PTH elevation in a prospective observational study from December 2009 to May 2012. At inclusion, we delivered a validated questionnaire to measure clinical symptoms. With an original statistical analysis, we demonstrated that high PTH levels matched with 4 very different phenotype profiles, suggesting that a “one-size-fits-all” target approach for SHPT was not appropriate.

The pharmacoepidemiological EPHEYL (Étude PHarmacoÉpidémiologique de l’hYperparathyroïdie secondaire en Lorraine) study is a 2-year, open-cohort, prospective, observational study on incident SHPT, i.e. newly diagnosed, with a 2-year follow-up, set in the 12 dialysis units located in the French region of Lorraine (public or private).

EPHEYL is a well-characterized cohort of patients with incident severe SHPT diagnosis described not only on the basis of initiation of cinacalcet but also a cut-off value for PTH [6]. An incident population helps to accurately describe diseases, avoiding bias related to incidence and prevalence mix [16]. We used an appropriate methodology, MCA and ascendant hierarchical clustering, to identify homogeneous subgroups of cases with a high statistical level validity [22]. Our four clustered subgroups consisted of homogeneous patients with same medical history, same prior therapy, and probably similar characteristics concerning mineral bone diseases and cardiovascular co-morbidities.

SHPT symptoms are difficult to assess due to the lack of specificity. The self-administered questionnaire developed by Pasieka et al. was used in several studies on primary and secondary hyperparathyroidism to quantify severity of symptoms using median values [18‐20]. In the EPHEYL study, one out of two patients suffered from at least one symptom. But the most frequent symptoms (thirst, weakness, fatigue, and pain of joints) were not specific. As the questionnaire was developed in the context of parathyroidectomy, its validity is questionable at early stage of SHPT.

The PTH cut-off value of 500 ng/L was chosen at the time of 2003 K-DOQI [6]. Its enabled to focus on SHPT patients without adynamic bone disease [8, 23]. Furthermore, no patient had hypercalcemia, suggesting that there was no tertiary or autonomized SHPT. This result is consistent with the incident type of our cohort, as tertiary SHPT were found in previous studies including prevalent SHPT patients [6, 24]. Nevertheless, we know that PTH is subject to many simultaneous types of variability [7, 11]. Our study points out obstacles with the use of PTH to precisely diagnose SHPT. The distribution of PTH at a cut-off value of 500, according to the new recommendation: “maintaining PTH levels in the range of approximately two to nine times the upper normal limit for the assay” was wide (Figure 2). Jean et al. have suggested that PTH should be replaced with specific biochemical markers of bone such as bone ALP and beta cross-laps to follow-up SHPT [24]. These measurements, however, are too costly to be recommended in routine clinical practice [8]. Finally, in the context of quite vague recommendations, clinicians should be aware that a binary approach for SHPT diagnosis, i.e. absence/presence, is not adequate. There is definitely a grey zone for diagnosis which limits are not easily defined. We should recommend an observation period before acting strongly.

In this grey zone, our study identified four statistically distinct subgroups of patients. Our description of each group reflected a clinical reality, and was therefore clinically appropriate. Noteworthy, at bedside, these distinct phenotypes should be distinguished by doctor rather by biological cut-offs. This pleads for patient-doctor contact. A recent publication has demonstrated a positive association between patient-doctor contact and outcomes [25]. Last but not least, our study reinforces the recent publication by Levin that has recommended acknowledging the heterogeneity of chronic kidney disease populations and appropriately characterizing populations for studies [26].

The group of “elderly patients with a few cardiovascular comorbidities”, in majority with normocalcemia and normophosphatemia, had a PTH which, at first, should impressed clinicians. In another hand, normal serum phosphorus could not be explained by malnutrition; despite their old age, nutritional markers (such as albumin and phosphatemia) were not statistically different from those in the other groups. PTH seemed to be associated with a good clinical condition and a low prevalence of comorbidities. These results are consistent with those from previous studies in showing that, particularly in elderly, PTH is inversely correlated with score of comorbidities [12, 27]. At the time of diagnosis of SHPT, this subgroup of patients did not require presumably an intensive therapeutic management.

The group of “younger patients with severe cardiovascular comorbidities”, by contrast, consisted of a majority of patients (66%) with PTH within targets [8]. However, the mean PTH was similar to the one found in the previous group. They seemed to be most likely to have bone and cardiovascular complications, as in previous cohorts, possibly linked to hyperphosphatemia, with a high proportion of cardiovascular diseases, diabetes and obesity [4, 13]. As a result, they should require an intensive care management. Of note, most of them had hyperphosphatemia and hypocalcemia despite multiple treatments. It is obvious that this phenotype is characterized by very low adherence to first-line strategies for SHPT, e.g. calcium, vitamin D and phosphorus binders. For this treatment category, a recent meta-analysis has pointed out poor adherence in a majority of patients [28].

The group of “intermediate patients” seemed to show intermediate characteristics between those with both previous groups. Most of them had hyperphosphatemia without hypocalcemia and the highest PTH. They seemed to be more likely to be at high risk for cardiovascular events due to uncontrolled hyperphosphatemia; as a result, they should be cautiously monitored [21].Finally, the group of patients “who initiated cinacalcet” was very different from the others (Figure 3). Most of patients had hypocalcemia and hyperphosphatemia. Patients were given already multiple conventional treatments before cinacalcet initiation. At the time of cinacalcet initiation, an overwhelming majority of patients (86%) had PTH within KDIGO™ targets which were applicable during the study period. Our study has confirmed that cinacalcet was prescribed for broadened indications in real life, and highlighted that the benefit-risk balance of cinacalcet was not favorable in patients with low PTH. This phenotype “cinacalcet user” including such patients should not exist. Before such therapeutic agents were marketed, indications for surgical parathyroidectomy were limited to symptomatic SHPT patients with very high PTH (>1000 ng/L). Initially presented as an alternative to parathyroidectomy, cinacalcet is now prescribed in pre-symptomatic patients with PTH > 300 ng/L, perhaps due to previous KDOQI recommendations.

Its prescription is based on studies suggesting that cinacalcet should have a protective effect on cardiovascular disease outcomes and reduce risk of fractures [29]. Recently, the prospective EVOLVE study failed to demonstrate the role of cinacalcet in reducing cardiovascular mortality and events [10]. These negative results can be explained, in particular, by the fact that 23% of patients in the treated group and 11% in the placebo group received the trade formulation of cinacalcet, which skewed the results. The EPHEYL study has shown that the use of cinacalcet was likely to be physician-dependent rather than characteristic of patients, and decision-making to prescribe it was insufficiently detailed. Thus in the group of patients who initiated cinacalcet, 43% of them who should have been treated with cinacalcet were not taking treatment 3 months later. Considering that low PTH levels characterize adynamic bone disease and/or a high prevalence of comorbidities, we should currently recommend to prescribe cinacalcet only after conducting a rigorous investigation on its benefit/risks [23, 27]. The strength of our study was to identify a subgroup of incident SHPT patients treated with cinacalcet despite low PTH in a real life setting. The matter whether cinacalcet should be contraindicated in patients with low PTH is of great interest.

Our study has some limitations. Its observational nature which may seem like a drawback is strength actually. Pharmacoepidemiological study, as EPHEYL, reflects routine clinical practices with various therapeutic strategies which are not always consistent with previous randomized trials [10]. Second, we have taken into account only data at the inclusion, making sense as regards to the incident nature of a disease. The 2-year follow-up of the EPHEYL cohort should reinforce the relevance of the classification. Third, most of patients were included with a PTH cut-off value of 500ng/L, while the latest KDIGO™ guidelines have recommended maintaining PTH in ranges, at approximately 2 to 9 times the upper normal limit, rather than absolute values. KDIGO™ recommendations were not implemented when we wrote our article. As a result, data on PTH have been detailed in this report. Fourth, the methodology allowed identification of distinct profiles, not completely disjunctive. It is therefore possible that some patients of distinct subgroups should have partially similar characteristics at junctions. Notwithstanding differences between groups are relevant. Fifth, clinical symptoms assessed by PAS questionnaire need to be interpreted with caution, due to the difficulty met to collect data from self-administered questionnaire.

In conclusion, four significantly distinct profiles of dialysis patients with a recent severe SHPT diagnosis were identified on the basis of clinical, biological and therapeutic data routinely available. Our well-characterized incident cohort, coupled with an original methodological approach, highlights a contemporary picture of daily clinical practice. Our study reinforces that the benefit-risk balance of cinacalcet is not positive in patients with low PTH, and raises the matter whether cinacalcet should be contraindicated in such patients. A “one-size-fits-all” target for PTH approach is probably not appropriate. Therapeutic management needs to be adjusted to the four different phenotypes.