Renal complications in patients with chronic hypoparathyroidism on conventional therapy: a systematic literature review

A systematic literature review was performed to summarize the frequency and nature of renal complications in patients with chronic hypoparathyroidism managed with conventional therapy. Methodology was consistent with the recommendations outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement. Peer-reviewed journal articles with specified medical subject heading terms were identified using the PubMed, EMBASE, and Cochrane databases. Data were extracted from eligible articles based on prespecified parameters for clinical outcomes of renal calcifications and disease. Because of the heterogeneity of the data, a meta-analysis could not be conducted. From 1200 potentially relevant articles, data were extracted from 13 manuscripts that reported data for ≥1 of the 19 predefined renal outcomes for ≥10 adult patients (n = 11 manuscripts) or pediatric patients (n = 2 manuscripts). The collective data provide evidence that adult and pediatric patients with chronic hypoparathyroidism and treated with conventional therapy (oral calcium and active vitamin D) had an increased risk of renal complications. The reported rate of nephrolithiasis was up to 36%, with the lowest rates in studies reporting shorter duration of disease. The rate of nephrocalcinosis was up to 38%. Some studies reported a combined nephrolithiasis/nephrocalcinosis outcome of 19% to 31%. Data for renal disease that encompassed a range of renal insufficiency to chronic kidney disease were reported in 10 articles; the reported rates ranged from 2.5% to 41%. In patients who receive long-term treatment with oral calcium and active vitamin D, chronic hypoparathyroidism may be associated with an increased risk of renal complications compared with the general population.


Introduction
Hypoparathyroidism is a rare endocrine disorder caused by absent or inappropriately low levels of parathyroid hormone (PTH) [1]. Mineral homeostasis cannot be maintained because of the loss of the PTH-controlled pathways involving bone, kidney and the gastrointestinal (GI) tract. The absorption of calcium in the GI tract is greatly decreased because of the loss of activation of 25-hydroxyvitamin D (25[OH]D) to 1,25dihydroxyvitamin D (1,25 [OH] 2 D), which stimulates absorption of both calcium and phosphate [2]. The skeleton ceases to be a ready source of calcium because of exceedingly low bone turnover [3]. There is reduced calcium reabsorption and urinary phosphate excretion in the kidney because of the loss of the effect of PTH [2,4]. The results of these pathophysiological factors in hypoparathyroidism are hypocalcemia and hyperphosphatemia [1,2,5].
Conventional treatment in patients with chronic hypoparathyroidism is oral calcium and active vitamin D (eg, calcitriol), as well as parenteral forms of vitamin D and thiazide diuretics as needed [1,4,5]. Over the lifetime of an individual, chronic hypoparathyroidism and long-term therapy with oral calcium and active vitamin D appear to be associated with an increase in the risk of renal complications based on a number of retrospective studies in adult or pediatric patients with chronic hypoparathyroidism [6][7][8][9][10]. A case-controlled retrospective study found that increased calcium-phosphate product (ie, [serum calcium concentration] × [serum phosphate concentration]) was associated with increased risk of renal disease in patients with hypoparathyroidism [8]. Mitchell et al. suggested that conventional treatment may increase the risk of hypercalciuria, itself a risk factor for nephrolithiasis, nephrocalcinosis, and impaired renal function [9].
The objective of this systematic literature review is to summarize the reported frequency and nature of renal complications in patients with chronic hypoparathyroidism managed conventionally with calcium and active vitamin D. The specific renal outcomes investigated were nephrolithiasis/kidney stones, nephrocalcinosis, and chronic kidney disease (CKD). In addition, estimated glomerular filtration rate (eGFR) levels were also investigated. Any reported associations between each of the renal outcomes and relevant biochemical or disease parameters in the published selected articles are included in this review.

Data sources
Methodology was consistent with the recommendations outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [11]. A methodology protocol specified the process. Primary eligibility criteria for the inclusion of peer-reviewed journal articles are listed in Table 1. Searches were conducted in PubMed ® / MEDLINE ® for English-language abstract-containing articles published in peer-reviewed journals from database inception to 15 November 2018. Additional peer-reviewed journal articles not in PubMed ® /MEDLINE ® were identified by similar searches conducted in EMBASE ® and Cochrane ® databases for the same timeframe.

Search strategy
The strategy employed a database search string composed of free text and controlled vocabulary terms (ie, medical subject headings [MeSH] for PubMed ® /MEDLINE ® ). Selected MeSH terms included nephrocalcinosis, nephrolithiasis, kidney calculi, and all related terms for kidney stones, renal insufficiency, and chronic kidney/renal disease. This approach was broad based but also included precise terminology to capture publications that would potentially have data values for the predefined relevant clinical outcomes ( Table 2). The following is the search string that was used, formatted for PubMed ®

Screening and data extraction process
Articles with abstracts were reviewed according to the Table 1 eligibility criteria by two independent reviewers; a third reviewer resolved any nonconsensus selections. Articles excluded were assigned a reason for rejection. Eligible articles that remained after abstract screening underwent a full article review by each of the two independent reviewers to extract all data reported for the 19 predefined relevant clinical outcomes ( Table 2). Extracted data were reviewed, and a subset of articles containing data for the most relevant renalrelated outcomes (ie, nephrolithiasis/kidney stones, nephrocalcinosis, and CKD) and eGFR levels was selected. One reviewer conducted a second round of extraction, not specified in the protocol, to capture data reported for associations between renal outcomes and predefined biochemical-related outcomes. One reviewer conducted a third round of extraction, not specified in the protocol, to capture available data for thiazide use, blood pressure in the context of reported hypertension, and diabetes mellitus. The individual eligible articles reported data that were collected using differing heterogeneous methodologies that precluded any aggregation of the extracted data and a meta-analysis.

Articles selected
The process of the peer-review database search for data of interest yielded 1200 articles (Fig. 1). Following screening and assessment for eligibility, 74 of the 1200 articles had data that reported one or more of the 19 predefined clinical outcomes listed in Table 2. Of the 74 articles with data for one or more of the 19 predefined clinical outcomes, 21 reported data for nephrolithiasis/kidney stones, nephrocalcinosis, CKD, or eGFR levels. Of these 21 papers, 13 articles were the ultimate focus of this review because they reported data for nephrolithiasis/kidney stones, nephrocalcinosis, CKD, or eGFR levels from studies of ≥10 adult (n = 11) or pediatric patients (n = 2) with chronic hypoparathyroidism.

Renal calcifications
Treatment of hypoparathyroidism with conventional therapies may result in hypercalciuria, which is a risk factor for nephrolithiasis and nephrocalcinosis [9]. Nephrolithiasis is defined by the appearance of solid calcium-containing stones in the collecting system of the kidney [12][13][14].
Abstracts identified through the protocol-defined PubMed ® /MEDLINE ® / Embase ® and Cochrane ® database searches N=1200 Full-text articles screened n=100 No. of records excluded based on title or abstract n=1100 No. of records excluded with reasons n=26 Reasons for exclusions: Data for the predefined outcomes not reported (12) Patients not on conventional therapy (7) Review or guideline article (4) Transient hypoparathyroidism (1) Not hypoparathyroidism (1) Unrecovered article (1) Second-round data extraction n=21 Full-text articles excluded for not reporting data for nephrolithiasis/kidney stones, nephrocalcinosis, CKD or eGFR levels. n=53 Screening Identification Included for extraction #1

Included for extraction #2
Data extracted from full-text articles n=74 Full-text articles with ≥10 patients n=13 Full-text articles with <10 patients n=8 Included for analysis Nephrocalcinosis refers to the parenchymal deposition of calcium salt crystals within the interstitium of the kidney and usually involves the renal medulla [12,[15][16][17]. Nephrolithiasis and nephrocalcinosis are conditions that may coexist [12,16]. Nephrolithiasis is commonly diagnosed by abdominal computed tomography (CT) or ultrasonography [18]. Nephrocalcinosis is typically detected using ultrasound imaging as increased bilateral, symmetrical echogenicity within renal pyramids, or by abdominal CT [12,16].

Nephrolithiasis/kidney stones
Among the six articles with data on the percentage of patients with nephrolithiasis/kidney stones, rates varied from 0% [10] to 35.5% [19] ( Table 3, Fig. 2) [6,7,10,[19][20][21]. Differences in study designs, patient populations, and overall size of the studies may help explain some of the variation in the reported rates of this renal complication. Furthermore, the considerably different assessment methods used in the six studies, ranging from diagnostic codes [6,7], ultrasound [10,20,21], and patient self-reporting [19], likely contributed substantially to the heterogeneity in the reported rates. In the pediatric study of Levy et al., the lowest outcome rate of nephrolithiasis was reported (0%) based on ultrasound [10]. In adult studies, rates were reported using diagnostic codes (1%-2%) [6,7], ultrasound (8%-30%) [20,21], and patient self-reporting (35.5%) [19]. These findings suggest that current understanding regarding the frequency of this complication in patients with hypoparathyroidism is likely affected by the choice of diagnostic coding or ultrasound. Meola et al. evaluated a crosssection of patients with chronic hypoparathyroidism on conventional treatment to determine the proportion who met biochemical parameter targets defined by the European Society of Endocrinology (ESE) treatment guidelines [22] in order to meet the ESE treatment goal to relieve symptoms of hypocalcemia and maintain calcium levels in the low or slightly below normal reference range. As part of that study, Meola et al. reported that 30% of the study population had nephrolithiasis detected by renal ultrasound and that most were asymptomatic [21]. Meola et al. used a study design with age-and gendermatched healthy normative controls and determined that the rate and risk of nephrolithiasis/kidney stones was significantly higher in patients with postsurgical chronic hypoparathyroidism versus controls (30% vs 5%, P < 0.0001; odds ratio, 8.2).
In the two studies that used age-and gender-matched controls, there was an increased risk of nephrolithiasis/kidney stones in patients with postsurgical chronic hypoparathyroidism, but not in patients with nonsurgical chronic hypoparathyroidism (hazard ratios, 4.82 and 0.80, respectively; Table 3) [6,7]. Underbjerg et al. suggest that for renal outcomes there may be an interaction between the age of the individual and duration of disease, with increased risk in older patients [7].
A separate study by Bohrer et al. used a new Likert-scale questionnaire to capture patient self-reported impairment reported as distress level scores [23]. Patients with higher total symptom or distress level scores had more illness manifestations, including kidney stones (plus bone changes, basal ganglia calcification, and cataracts), and had lower serum calcium levels than patients with lower total symptom or distress level scores (both P < 0.05).

Nephrocalcinosis
Among the four articles reporting specifically the percentage of patients with nephrocalcinosis, rates varied from 0% to 38% (Table 4, Fig. 2) [10,[19][20][21]. Similar to the nephrolithiasis/kidney stones outcome, between-study differences in design, methodology and population size may explain the large discrepancies in the reported rates of nephrocalcinosis. In order of lowest to highest, the reported occurrences in the three adult studies used ultrasound (0%) and patient self-reporting (28%); none of the studies reported diagnostic codes [19][20][21]. Hadker et al. used a patient selfreporting cross-sectional survey of adult patients and found that those who indicated severe disease on the questionnaire (no definition or description of severity grades was provided) reported a significantly higher occurrence of nephrocalcinosis versus patients who reported having mild disease (22% vs 6%; P ≤ 0.05) [19].
Levy et al. pediatric study that used renal ultrasound reported 38% of patients had nephrocalcinosis, in contrast to the finding that no patients had nephrolithiasis/ kidney stones [10]. This study also enabled evaluation of changes in nephrocalcinosis using a staging system developed by Boyce et al., in which stage 0 was no echogenicity; stage I, mild echogenicity around medullary pyramid borders; stage II, moderate echogenicity around and inside pyramids; and stage III, severe echogenicity of entire pyramids [12]. Of the 11 patients with nephrocalcinosis after the initial ultrasound, the nephrocalcinosis resolved in two patients (18%), remained in early stage I in three patients (27%), and progressed from stage I to III in six patients (55%). In the two patients with resolved nephrocalcinosis, both had DiGeorge syndrome, and calcium concentrations were more frequently within the target range versus patients in whom nephrocalcinosis persisted (81% ± 7.6% vs 56% ± 8.5%; P = 0.01).

Combined data for nephrolithiasis and/or nephrocalcinosis
Four articles reported data on the percentage of patients with nephrolithiasis and/or nephrocalcinosis as a combined outcome; all studies used ultrasound or CT scans [9,[24][25][26]. Among the four studies, the rates were similar to those reported in the studies using separated nephrolithiasis and nephrocalcinosis outcomes (19% to 31%; Table 5). In a study by Leidig-Bruckner et al. of 33 patients with postsurgical hypoparathyroidism (and medullary thyroid carcinoma), there were two cases noted for hospitalization for symptomatic nephrolithiasis [26]. The article also provided details about the nine patients (27%) with documented renal calcifications; five patients had initially received high cholecalciferol dosages, and two patients had received dihydrotachysterol. Also, three of the nine patients receiving cholecalciferol/dihydrotachysterol experienced transient renal failure.

Chronic kidney disease
Data for CKD, renal insufficiency, and eGFR levels were reported in 10 articles [6-10, 19, 21, 25-27]. Eight adult studies reported the percentage of patients with CKD based on standard methods of eGFR <60 mL/min/1.73 m 2 or ≥ stage 3 classification [28], or renal insufficiency international classification of diseases (ICD) ICD-8 and ICD-10 codes. One survey reported CKD based on adult patients self-reporting for chronic kidney failure [19]. The pediatric study of Levy et al. used the revised Schwartz estimating equation for nonchronic kidney disease populations. The methods used by each study are detailed in Table 6. The rates of CKD varied among studies from 2.5% to 41% ( Fig. 3 and Table 6). In a survey by Hadker et al., 2.5% of patients with milder hypoparathyroidism symptoms and 19% of patients with severe hypoparathyroidism symptoms self-reported having CKD [19]. In the two studies that used age-and gendermatched controls, there was an increased risk of renal insufficiency in both patients with postsurgical or nonsurgical chronic hypoparathyroidism (hazard ratios, 4.95 and 6.01, respectively; Table 6) [6,7].
There were five adult studies that reported eGFR data using either the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, Cockcroft-Gault (eCrCl) or Modification of Diet in Renal Disease (MDRD) formula [8,9,21,26,27]. Four articles reported similar percentages of patients with eGFR <60 mL/min/1.73 m 2 , ranging from 12% to 23% of patients; the overall populations had a mean duration of disease of 9 to 15.9 years [8,21,26,27]. In a study by Mitchell et al., 41% of 120 patients had eGFR <60 mL/min/ 1.73 m 2 ; the overall population had a mean duration of disease of 17 years [9].

Thiazide use, hypertension, and diabetes mellitus
Renal outcomes may be affected by thiazide use, blood pressure in the context of reported hypertension, and diabetes mellitus. However, an examination of the 13 articles in this review found that very few reported any data on the following indices: thiazide use (n = 2) [9,21], blood pressure in the context of reported hypertension (n = 2) [8,21], and diabetes mellitus (n = 3) [6,8,21]. In the two studies that reported thiazide data, the percentages of patients prescribed this medication were 2% and 20%; the authors of the paper with the higher percentage noted they were unable to determine whether the patients were prescribed thiazides for hypertension, hypercalciuria, or both [9,21]. In the two studies that reported the percentage of patients with hypertension or diabetes mellitus type 1 or 2, ranges were 3.5% to 18% and 1% to 8.4%, respectively [8,21]. The authors of these four articles did not infer any relationship with renal outcomes based on the limited data. These conditions may have been underreported in the articles because they were not the primary focus of the publications and these diagnoses were based on hospital records. The collation of the predefined renal outcomes and diseaserelevant biochemical parameters affords the opportunity to reveal relationships not previously described. However, it was recognized that there are limitations to this approach in that an individual study with single biochemical measures (24-h urine collection) may not characterize the longitudinal status of that parameter. The association of nephrolithiasis/kidney stones with a number of biochemical or clinically relevant parameters was only reported by one study of 90 patients [21]. No significant correlation (P = 0.98) was seen between the presence of kidney stones and the duration of hypoparathyroidism, 24-h urinary calcium excretion, total albumin-corrected serum calcium, or vitamin D status (Table 3).
Association data with nephrocalcinosis and a number of biochemical or clinically relevant parameters were reported by the Levy et al. pediatric study (n = 29) [10]. In a multivariate analysis, the most significant predictors for nephrocalcinosis were the degree of relative hypercalcemia (area under the curve [AUC] of total calcium concentrations >9.6 mg/dL) and the degree of hyperphosphatemia (AUC above age-adjusted phosphate concentrations; R 2 = 0.47, P < 0.01; Table 4). Odds ratios for the association between nephrocalcinosis and degree of hypercalcemia and degree of hyperphosphatemia were 1.027 (95% CI, 1.003-1.052) and 1.004 (95% CI, 1.001-1.008), respectively. Compared with 18 patients in the study without nephrocalcinosis, the nine patients with unresolved nephrocalcinosis had a greater degree of hypercalcemia (AUC of total serum calcium concentrations >9.6 mg/dL; P = 0.005), hyperphosphatemia (AUC above age-adjusted phosphate; P = 0.01), and hypocalcemia (AUC of total serum calcium concentrations <8.0 mg/dL; P = 0.004) and a greater duration of hypocalcemia (percentage of time with total calcium <8.0 mg/dL; P = 0.003).
Association data with the combined nephrolithiasis and/or nephrocalcinosis outcome and biochemical parameters were only reported in a study by Lopes et al. of 55 patients [25]. Weight-adjusted 24-h urinary calcium was higher in patients with renal calcification versus those without (3.3 vs 1.8 mg/kg/day, respectively; P < 0.05; Table 5). However, there was no correlation between serum and urinary levels of calcium and the presence of calcification.
Mitchell et al. in their study of 120 patients examined correlations between CKD and a number of biochemical or clinically relevant parameters; eGFR <60 mL/min/1.73 m 2 was compared with age-matched normative controls (Table 6) [9]. Univariate analyses found that eGFR levels were negatively correlated with age (P < 0.001), duration of disease (P < 0.001), average time-weighted serum calcium (P < 0.001), and estimated proportion of time with serum calcium higher than 9.5 mg/dL (P < 0.001). Average timeweighted serum phosphate and average calcium-phosphate product were not correlated with eGFR levels. Multivariate regression analyses of the predictors from the univariate analyses found that eGFR levels remained significantly associated with age (P < 0.001), duration of disease (P = 0.032), and proportion of time with relative hypercalcemia (P = 0.005). No association was seen between eGFR levels and either 24-h urine calcium values or presence of renal calcification. In a univariate analysis of the pediatric study, lower eGFR was associated with higher calcium concentrations (r = −0.42, P = 0.02) and a greater proportion of time with relative hypercalcemia (r = −0.41, P = 0.03) [10].
In a case-controlled retrospective study of 431 patients with national health registry data of long-term complications, Underbjerg et al. applied a composite endpoint of renal stones (defined by ICD codes) and renal insufficiency (defined by eGFR <60 mL/min/1.73 m 2 ) to describe renal disease in patients with hypoparathyroidism [8]. This study showed that a decreased risk of any renal disease was associated with a higher dose of alfacalcidol supplementation (>1 vs ≤1 μg/day; P = 0.03). An increased risk of any renal disease was associated with an increased serum calciumphosphate product (>2.80 mmol 2 /L 2 ), increased number of hypercalcemic episodes, and long duration of disease. Predictors of any incidence of renal disease were disease duration (≥12.7 vs <12.7 years; P < 0.01) and increased calcium-phosphate product (≤2.80 vs >2.80 mmol 2 /L 2 ; P < 0.01). Although the articles used divergent outcome measures and methodologies to assess biochemical parameters, they nevertheless revealed valuable insights into factors associated with renal outcomes in patients with chronic hypoparathyroidism.

Limitations
There was a significant heterogeneity in the data and in the methods of reporting data for each of the renal outcomes within published articles of clinical data studies of adult and pediatric patients with chronic hypoparathyroidism. Given the relatively low prevalence of hypoparathyroidism, it is not surprising that there are large gaps in the reporting of the key disease-related biochemical parameters studies of patients with chronic hypoparathyroidism; prospective studies are needed to address these knowledge gaps. The methodology for the collection of CKD information based on eGFR data was heterogeneous and often unclear. Only one study explicitly reported a collection method according to the CKD definition (ie, low eGFR levels on ≥2 occasions with an interval of ≥3 months) [26]. However, the low eGFR rate reported in the majority of the articles suggests a common comorbidity in patients with chronic hypoparathyroidism. There are  Of the 9 pts with renal calcifications, 2 were treated with calcitriol from the beginning of treatment, 5 initially received high cholecalciferol doses, and 2 received dihydrotachysterol [see Table 6 for the reported eGFR data and renal calcifications]   The eGFR was negatively correlated with the duration of hypoparathyroidism (r = −0.62; P = 0.0001).
This correlation remained significant after adjusting for chronological age (partial correlation, adjusted for age r = −0.35, P = 0.04).
The correlation between eGFR and duration of hypoparathyroidism was independent from the degree of hypoparathyroidism (partial or complete) and also independent from the radiological presence of calcification More pts with calcifications had eGFR <60 mL/min/1.73 m 2 (ie, CKD) 22% (2/9 pts) than those without calcifications 8% (2/24 pts); differences were not significant At last visit, eGFR was lower in pts with calcifications (9/33 pts) than in those without calcifications ( Table 6. Patient numbers (N) refer to the number of patients with hypoparathyroidism in the study. a Patients self-reporting in a crosssectional survey. eGFR, estimated glomerular filtration rate; ICD, international classification of diseases and related health problems eGFR <60 mL/min/1.73m 2 ) [8]. The serum calcium-phosphate product level was within generally recommended reference ranges but was relatively high, leading Underbjerg et al. to suggest that treating physicians should target the lower part of the reference range. A similar point was made by the authors for target serum phosphate levels based on their association findings with increased risk for complications and mortality. Other studies in the general population reported that high serum phosphate was associated with harmful effects on renal function [32] and impairment of microvascular function in individuals with normal renal function [33]. In adult patients with chronic hypoparathyroidism, an increased number of hypercalcemic episodes and a longer duration of illness were associated with an increased risk of any incidence of renal disease [8]. In pediatric patients, a greater degree of hypocalcemia or repeated episodes of hypocalcemia were unexpectedly associated with the development of nephrocalcinosis [10], which could be explained by the need for higher doses of oral calcium to maintain normal serum calcium and/or concomitant hyperphosphatemia. However, these assertions require further investigation. The pediatric study provided definitive information about the presence of early mild renal impairment in children with chronic hypoparathyroidism. The authors noted that early renal impairment in childhood aligns with longitudinal studies in adults, and may progress to CKD in adulthood [12].

Concluding remarks
Renal complications and an increased risk of adverse renal events in patients with chronic hypoparathyroidism who receive conventional therapy were observed consistently in a systematic literature review. There is an unmet need for additional large-scale studies, including more studies with standardized CKD definitions methodology, to better establish the factors that increase the risk of renal complications in patients with hypoparathyroidism.