Results from our historic cohort study showed that after 9-15 months follow-up, surgical cure of pHPT is associated with a decline in kidney function compared to surveillance without surgical intervention. Subgroup analyses showed that this especially applies to those with an initial (near-) normal kidney function, whereas no decline in eGFR was observed if kidney function was impaired at baseline.
Our results are in accordance with the findings from two previous retrospectives studies by Tassone et al (5) and Garcia-Martin et al (17). In both studies, kidney function was assessed prior to and following PTX. Our study provides further support for these findings by including a group of patients not treated by PTX, allowing for comparisons between PTX and non-PTX treated patients. Furthermore, our findings extend previous findings by including a much larger sample of patients, allowing for a more detailed assessment of the importance of kidney function at baseline. Whereas previous studies showed that PTX is associated with a decline in kidney function only if eGFR at baseline is above 60 mL/min, but not if eGFR at baseline is below 60 mL/min, our analyses suggest that the decline in kidney function following PTX only applies to those with a baseline eGFR above 80 mL/min.
PHPT is associated with an increased risk of kidney impairment through several of different mechanisms which include hypercalciuria, development of kidney calcifications and higher rate of hypertension (21, 22).
Similar to previous studies, our study showed a high prevalence of kidney impairment as 26.2 % of our included patients had an eGFR below 60 mL/min at baseline (23-25). However, during one year of surveillance without surgery, eGFR remained stable in the non-PTX group. This is consistent with several previous studies showing that kidney function remains largely stable during different follow-up times (11, 12, 26, 27).
Thus, no rapid decline in kidney function occurs shortly after diagnosis in patients managed conservatively. pHPT is known as a disease that often remains silent for several years before diagnosed (28). Several of the patients included in our study had presumably been suffering from pHPT for a long period prior to diagnosis, which may have resulted in kidney impairment. Of importance, the decline in kidney function following PTX was only of small magnitude in absolute terms and probably without clinical importance. Accordingly, despite a small decrease in eGFR following PTX, it seems appropriate to shorten the duration of the disease by surgical cure and thereby presumably lower the risk of further (long-term) kidney complications. Moreover, successful PTX is associated with additional beneficial effects such as improved BMD with reduced risk of fracture and a better quality of life (27, 29-36).
The mechanism causing a decreased eGFR following PTX needs to be further elucidated. It may be speculated whether it is due to the well-known effect of PTH as a vasodilator (37) or the diuretic actions of hypercalciuria. Thus, activation by calcium of the ductal apical calcium-sensing receptor may inhibit arginine vasopressin (AVP) mediated increases in intracellular cAMP and increase water excretion (38). Concerning PTH as a vasodilator, a successful PTX could result in relative kidney vasoconstriction, followed by a decrease in eGFR. A similar effect might also be attributable to the lowering of aldosterone levels in response to PTX. PHPT has been suggested to increase plasma levels of aldosterone, with a reduction following PTX (39, 40). The reduced aldosterone levels will lower fluid volume and thereby increase plasma creatinine levels and decrease eGFR. This mechanism has been well described following adrenalectomy in patients treated for primary hyperaldosteronism (41, 42). Moreover, muscle function is known to be impaired in patients with pHPT. Muscle function may increase following cure for pHPT which may contribute to increased levels of creatinine (43).
Of interest, in our mutually adjusted multiple regression analyses, PTH levels at baseline were inversely associated with percentage change in eGFR. In other words, the higher PTH levels at baseline, the larger was the decrease in eGFR. As no effect was observed in patients with impaired kidney function, it may be speculated whether PTH only causes kidney vasodilation if kidney function is well preserved. Of interest, PTH therapy of hypoparathyroidism for eight years has been associated with a stable kidney function, whereas a decline in kidney function often occurs if patients are treated with conventional therapy in terms of active vitamin D and calcium supplements (44). Further studies should aim to investigate if this potential renoprotective effect of PTH replacement therapy in hypoparathyroidism only applies to patients with an initial normal kidney function or also may be beneficial to patients with impaired kidney function. Even if endogenous PTH acts as a mild kidney vasodilator in pHPT, any protective effects would come at the cost of an increased risk of kidney calculi and it is important to stress that patients in conservative treatment would be monitored for the development of stone disease or nephrocalcinosis and that developing this complication would be a strong indication for fast-tracking them for parathyroidectomy.
Strength and limitations:
The strength of this study is firstly the large cohort which includes both a group that has received PTX and a control group, opposite the recent studies only investigating kidney function prior to and after PTX. However, our study design using a historic cohort makes it impossible to draw causal conclusions from our findings. As shown in Table 1, characteristics between PTX-treated and non-PTX treated patients differ on several parameters. Certainly, conservatively managed patients were older, had lower serum calcium levels and poorer kidney function at baseline. There is some obvious selection bias because the groups are not randomized, and there are reasons causing patients to be operated or not. Guidelines strongly favour surgery in patients with impaired kidney function, but young age is an additional factor than strongly speaks for PTX in endocrine practice. In addition to the question of surgery referral bias, we do not have information on whether patients have received medical treatment (e.g., bisphosphonates, cinacalcet etc.). Such medical treatment could influence kidney function.
Another strength of our study is that our data is based on the Danish NPR, which is a well-validated database (18). Data on biochemistry were available from 1977, the year when NPR was founded. It was only possible to validate the pHPT diagnosis by biochemistry in 1/3 of the patients. Of importance, our initial search for patients in the Danish NPR was very broad in order to aim at identifying as many patients with pHPT as possible. Many of the patients initially identified, were suffering from other causes of hyperparathyroidism. Thus, the fact that many of the patients initially identified could not be confirmed as suffering from pHPT should not be interpreted as a lack of specificity of the register but as a consequence of a broad search strategy with deliberately high sensitivity. Furthermore, we were only able to search for biochemical findings in databases from the major current laboratories. As our study includes a time span of more than 40 years, some laboratories do no longer exist or have not maintained their database on biochemical findings for the entire time period. On the other hand, we feel very confident that patients identified were suffering from pHPT. There is no reason to assume a connection between, whether biochemistry was available for a patient and the change in kidney function after PTX/non-PTX-treatment, why we do not think that this is a cause of selection bias. Furthermore, follow-up after 9-15 months was only registered in half of the patients with a confirmed diagnosis. This is probably due to the retrospective design of this study, where patients have been followed in the clinic and not according to a predefined protocol.
Biochemistry was performed by various laboratories and we have no information on the exact measurement methods. We tried to compensate for this by using local reference ranges used by the specific laboratories in the relevant periods. In addition, to minimize differences caused by the different measurement methods, a percentage difference between baseline and follow-up levels of creatinine and eGFR was calculated, as most patients had these measurements performed in the same laboratory.
Comparing our study with similar studies, other studies used additional exclusion criteria for the study population, like in the study from Tassone et al, thyroid and liver disease patients were excluded (5). We did not include information on comorbidities, although this could have added information on studied patients. Nevertheless, we excluded patients with known kidney impairment >1 year before the time of diagnosis, to eliminate other reasons of kidney impairment as well as tertiary hyperparathyroidism as a reason for the hyperparathyroid hypercalcemia.
We cannot exclude that the observed effect on kidney function may be due to regression towards the mean. Prospective studies are needed to get a more profound understanding of this.
Lastly, as pHPT is considered a slowly progressing disease, our follow-up time (9-15 months) was relatively short. In further studies, it could be of interest to study the long-term effects of PTX on kidney function.