The Impact of Obesity on the Presentation of Primary Hyperparathyroidism

Abstract

Context:

Obesity has been associated with elevated serum PTH (sPTH) in the general population. Obesity may also alter the clinical presentation in patients with primary hyperparathyroidism (PHPT).

Objectives:

The objectives of the study were to compare the clinical presentation of obese (OB) vs nonobese (NO) PHPT patients and to assess the impact of obesity on the presentation of PHPT independent of serum calcium and PTH.

Patients:

Consecutive PHPT patients who underwent parathyroidectomy between 2003 and 2012 by a single surgical group participated in the study.

Setting:

The study was conducted at an academic medical center.

Design:

Cross-sectional review of records of preoperative demographic, historical, laboratory, and densitometry findings and intraoperative pathological findings were compared in OB vs NO patients.

Main Outcome Measures:

The prevalence of nephrolithiasis and osteoporosis was measured.

Results:

Two hundred forty-seven PHPT patients were included in this analysis. Fifty percent were OB and 79% were women. Mean body mass index was 25.3 ± 3.3 and 36.0 ± 5.2 kg/m² in the NO and OB groups, respectively. Age, gender, and race distribution was similar between the two groups. Serum calcium was similar between the groups (11.0 ± 0.7 mg/dL in NO vs 11.1 ± 0.9 mg/dL in OB, P = .13), whereas sPTH was higher in OB (151 ± 70 vs 136 ± 69 pg/mL, P = .03). The OB group exhibited higher prevalence of hypercalciuria (urine calcium > 400 mg per 24 h) (41% vs 23% in NO, P = .01) and nephrolithiasis (36% vs 21% in NO, P = .03). Despite higher sPTH, OB patients showed higher bone mineral density and a lower rate of osteoporosis (21% vs 35%, P = .05). Differences in the prevalence of hypercalciuria and osteoporosis between the groups persisted after adjustment for age, race, estimated glomerular filtration rate, gender, sPTH, and calcium.

Conclusions:

In PHPT patients, obesity is a risk factor for hypercalciuria and nephrolithiasis and is protective against osteoporosis. The impact of parathyroidectomy on the clinical features of obese PHPT patients merits further evaluation.

Primary hyperparathyroidism (PHPT) is the third most common endocrine disorder, with an incidence rate of 1 in 1000 in the United States (1). Patients with PHPT are on average more obese than the normal population (2). Furthermore, in healthy individuals without PHPT, greater body weight is associated with a higher serum PTH concentration (3). This positive correlation is thought to be due to a resistance to PTH in the obese (4, 5), but its impact in PHPT patients is not entirely known.

Compared with individuals with normal body mass index (BMI), obese individuals in the general (non-PHPT) population are known to have a lower prevalence of osteoporosis (6) and a higher prevalence of hypercalciuria and kidney stones (7–9). However, there is limited information on the impact of obesity on these clinical features in PHPT. Two studies have shown that obese PHPT patients display more severe disease manifestations in terms of symptoms of reflux, musculoskeletal pain, and depression and higher parathyroid gland weight (10, 11). However, these studies did not evaluate the impact of obesity on urinary or bone mineral density (BMD) parameters and did not control for the severity of hypercalcemia and the degree of PTH elevation when assessing the clinical findings of obese PHPT patients.

In this review of 247 consecutive PHPT patients who underwent parathyroidectomy, we investigated the impact of obesity on the clinical presentation and features of PHPT while controlling for disease severity.

Patients and Methods

We performed a retrospective review of 247 patients with PHPT and known BMI who underwent parathyroidectomy at University of Texas Southwestern Medical Center and Parkland Hospital between January 20, 2003, and January 4, 2012. Patients were excluded if they had secondary hyperparathyroidism, tertiary hyperparathyroidism, multiple endocrine neoplasia, chronic kidney disease stage 4 or greater or if serum calcium failed to normalize within 6 months after parathyroidectomy. Medical records were reviewed for symptoms of PHPT including nephrolithiasis, gastrointestinal reflux, or peptic ulcer disease, and a history of osteoporosis and fractures. The diagnosis of PHPT was confirmed upon review of the clinical features in the medical records as well as laboratory and postoperative histopathological findings. Surgery was performed on patients manifesting PHPT symptoms or in asymptomatic PHPT patients with one or more of the criteria for surgery defined by the National Institutes of Health consensus conference guidelines (12, 13). All parathyroidectomies were performed by one of five endocrine surgeons. BMI was calculated using body weight (kilograms) divided by height squared (square meters).

Patients were divided into two groups defined by BMI less than 30 kg/m² [nonobese (NO)[ and BMI ≥ 30 kg/m² or greater [obese (OB)]. Single-gland disease was defined by an intraoperative drop in serum PTH by greater than 50% and to within the reference range (<65 pg/mL) after resection of a single parathyroid gland. Other cases were defined as multigland disease. BMD was measured preoperatively in 161 patients, and 24-hour urine calcium was measured preoperatively in 162 patients. Serum 25-hydroxyvitamin D (25-OHD) concentration was measured preoperatively in 123 patients by HPLC/tandem mass spectrometry at the Mayo Medical Laboratories (Rochester, Minnesota). Serum PTH concentration was assessed by the Roche Cobas chemoluminescent assay (Roche Diagnostics Corp).

Statistical methods

The NO and OB groups were compared with the Fisher exact test and Wilcoxon rank sum test for categorical and continuous variables, respectively. Spearman correlation coefficients were used to assess the association between PTH and other variables. Slopes were compared between the NO and OB groups with linear regression models. PTH was log transformed prior to parametric analysis. Unadjusted and adjusted odds ratios were analyzed with logistic regression models. Results are expressed as mean and SD unless otherwise specified. A two-sided value of P < .05 was considered statistically significant. Statistical analysis was performed with SAS version 9.2 (SAS Institute).

Results

Table 1 shows the baseline characteristics of the 247 patients studied. The mean BMI for the NO group was 25.3 kg/m² as compared with 36.0 kg/m² in the OB group (P < .0001). There was no difference in age, gender, or race distribution between the two groups. Serum PTH was significantly higher in the OB group (151 ± 70 vs 136 ± 69 pg/mL, P = .03), whereas serum calcium, phosphorus, alkaline phosphatase, and estimated glomerular filtration rate (eGFR) were not different between the two groups (Table 1). The OB group had on average a higher 24-hour urinary calcium excretion (346 ± 164 vs 262 ± 167 mg per 24 h, P = .0005), although the 24-hour urine calcium to creatinine ratio (U_Ca/Cr) was not different between the two groups (257 ± 153 vs 248 ± 126 mg calcium per gram creatinine, P = .79). Hypercalciuria (defined as 24 h urine calcium > 400 mg/d) was significantly more common in the OB group (41% vs 23%, P = .01) that also had a higher prevalence of kidney stones (36% vs 21%, P = .03) (Figure 1). The prevalence of osteoporosis was lower in the OB group (21% vs 35%, P = .05) (Figure 1). Histologically, there was a similar proportion of multigland involvement between the groups (∼8%-9%), although the OB group had a higher average gland weight than the nonobese group (Table 1).

Table 1.

Demographic Characteristics and Clinical Features of Study Cohort

	NO	OB	P Value
Number of subjects	124	123
BMI, kg/m2	25.3 ± 3.3	36.0 ± 5.2	<.001
Age, y	57 ± 13	57 ± 10	.42
Female, %, n	81% (101)	79% (97)	.64
African American, %, n	20% (25)	23% (28)	.64
Caucasian, %, n	54% (67)	46% (56)	.20
Hispanic, %, n	17% (21)	28% (35)	.03
Other race, %, n	9% (11)	3% (4)	.07
Serum PTH, pg/mL	135.5 ± 69.2	150.6 ± 69.9	.03
Serum calcium, mg/dL	11.0 ± 0.7	11.1 ± 0.9	.13
Serum phosphorus, mg/dL	2.6 ± 0.5	2.5 ± 0.5	.09
Serum alkaline phosphatase, U/L	85.5 ± 32.5	97.3 ± 52.7	.29
Serum 25-OHD, ng/mL	23.6 ± 15.9	23.6 ± 10.2	.47
eGFR, mL/min per 1.73 m2	90 ± 30	84 ± 24	.27
24-Hour urine calcium, mg per 24 h	262 ± 167	346 ± 164	.0005
24-Hour UCa/Cr, mg/mg	257 ± 153	248 ± 126	.79
Fractures, %, n	6% (7)	5% (6)	.79
Diabetes, %, n	3% (4)	9% (11)	.067
GERD, %, n	27% (33)	38% (47)	.11
Single-gland disease, %	91.9% (114)	91.0% (112)	.82
Average gland weight, g	1.0 ± 1.3	1.3 ± 1.4	.02

Abbreviation: GERD, gastroesophageal reflux disease. Data are shown as mean ± SD except where otherwise specified.

Figure 1.

Prevalence of complications of primary hyperparathyroidism based on BMI. *, P < .05 between the obese and nonobese groups. Hypercalciuria was defined as 24-hour urine calcium greater than 400 mg/d.

Similar trends of higher serum PTH, 24-hour urine calcium, and BMD with increasing BMI were observed when patients were divided into four categories of BMI (normal BMI, overweight, obese, severely obese; Supplemental Table 1). Subjects with missing 24-hour urine and bone density results had similar mean age, gender, BMI, and serum calcium as those with nonmissing data (Supplemental Table 2).

There was a weak but significant positive correlation with BMI and PTH (Spearman R = 0.17, P = .007) (Figure 2A). Statistically, there was a nearly statistically significant interaction between serum PTH and BMI in determining serum calcium (P for interaction = .09). Consequently, subjects were grouped by tertiles of serum PTH and BMI quartiles (Figure 2B), and the relationship of serum calcium with serum PTH was somewhat different at lower vs higher BMI: in the lowest BMI quartile, average serum calcium increased progressively with increasing serum PTH. At higher BMI (BMI quartiles 2–4), the relationship between serum calcium and PTH is not as linear, and the average serum calcium in the middle PTH tertile is lower than that in the lowest and highest PTH tertiles. Furthermore, in the lowest PTH tertile, average serum calcium is higher with an increasing BMI quartile, whereas at higher serum PTH (PTH tertiles 2 and 3), the average serum calcium is relatively constant across BMI quartiles.

Figure 2.

Relationship between serum PTH and various metabolic findings in patients with primary hyperparathyroidism A, Moderate but significant positive relationship between serum PTH and BMI. Spearman correlation was 0.17; P = .01 B, Three-dimensional representation of the relationship between serum calcium, serum PTH, and BMI. Data are grouped by quartiles of BMI and tertiles of serum PTH, with the average serum calcium shown on the z-axis. C and D, Obese primary hyperparathyroidism patients (●) have significantly higher 24-hour urine calcium excretion than nonobese (□) patients with primary hyperparathyroidism when expressed in milligrams per day (C), but the 24-hour U_Ca/Cr ratio is not different between the two groups (D). Solid line represents regression line for obese group, whereas dotted line represents regression line for nonobese. E and F, Obese primary hyperparathyroidism patients (●) have significantly higher BMD than nonobese (□) patients with primary hyperparathyroidism at the lumbar spine (E) and the one third radius (F) at any given serum PTH. Solid line represents regression line for obese group, whereas the dotted line represents the regression line for the nonobese.

Twenty-four-hour urine calcium excretion was higher in the OB group, reflected in the significantly higher intercept when graphing urinary calcium vs serum PTH (Figure 2C, P < .001); however, the slopes of the two lines were not statistically different (P = .92). When urinary calcium excretion was expressed as U_Ca/Cr, there was no difference between OB and NO groups at any given serum PTH (Figure 2D). At any serum PTH value, OB subjects had a significantly higher BMD at both lumbar spine and the one third radius (P = .01 and P = .02, respectively) (Figure 2, E and F).

Overall, the odds ratio (OR) for osteoporosis in OB vs NO PHPT patients was 0.52 [95% confidence intervals (CIs) 0.29–0.91, P = .02] (Table 2). The lower risk of osteoporosis in the OB persisted after adjusting for age, race, gender, eGFR, serum PTH, serum calcium, history of diabetes, and history of thiazide use (OR 0.45, 95% CI 0.23–0.89, P = .02). OB patients also had a 2.2-fold increase risk for bone pain after multivariate adjustment (P = .02). The OR for hypercalciuria in the OB was 2.32 (P = .003) and 2.40 (P = .007) in unadjusted and multivariate adjusted analyses. There was a 2-fold increase in the rate of kidney stones in the OB (P = .01), which remained elevated, although was no longer statistically significant after multivariable adjustment (P = .13) (Table 2). For each 5-kg/m² increase in BMI, there was a significant decrease in the risk of osteoporosis by 25% (P = .01), an increase in the risk of bone pain by 28% (P = .02), and an increased risk of hypercalciuria by 28% (P = .01) and of kidney stones by 26% (P = .03) in unadjusted analyses (Table 2). After multivariable adjustment, similar trends were observed, although the OR for kidney stones was no longer significant (OR 1.25, 95% CI 0.99–1.57, P = .06) (Table 2).

Table 2.

Impact of Obesity and Increasing BMI on Clinical Presentation of Patients with PHPT

	OR (95% CI, P value) of complication in OB vs NO		OR (95% CI, P value) of complication with 5 kg/m2 rise in BMI
	Unadjusted	Multivariable adjusteda	Unadjusted	Multivariable adjusteda
Osteoporosis	0.52 (0.29–0.91, P = .02)	0.45 (0.23–0.89, P = .022)	0.75 (0.61–0.94, P = .013)	0.70 (0.53–0.92, P = .012)
Bone pain	1.86 (1.04–3.35, P = .038)	2.21 (1.13–4.32, P = .020)	1.28 (1.04–1.58, P = .018)	1.32 (1.06–1.66, P = .015)
Hypercalciuriab	2.32 (1.34–4.02, P = .0027)	2.40 (1.27–4.54, P = .007)	1.28 (1.05–1.55, P = .014)	1.26 (1.01–1.57, P = .035)
Kidney stones	2.10 (1.19–3.71, P = .011)	1.66 (0.85–3.21, P = .13)	1.26 (1.03–1.53, P = .026)	1.25 (0.99–1.57, P = .06)

^aAdjusted for age, gender, race, eGFR, serum PTH, serum calcium, use of thiazide diuretics, and presence of diabetes mellitus.

^bHypercalciuria was defined as 24-hour urine calcium greater than 400 mg/d.

There were only 123 patients with available preoperative serum 25-OHD measurement. In this subgroup, adjusting for serum 25-OHD did not alter the OR for kidney stones and osteoporosis with obesity.

Discussion

In the general (non-PHPT) population, obesity has been associated with a lower prevalence of osteoporosis, an increased risk of nephrolithiasis, and a higher serum PTH concentration (3–5). Although PHPT patients have been previously shown to be heavier than eucalcemic controls (4), the impact of obesity on the clinical presentation of PHPT has not been well studied. In this report, we found that obese patients with PHPT are more likely to present with hypercalciuria and nephrolithiasis and less likely to present with osteoporosis when compared with nonobese PHPT patients. We also showed for the first time that the impact of obesity on these complications is independent of serum calcium and PTH concentration. The underlying mechanism(s) are not entirely clear at this time and merit further exploration.

Biochemically, obese patients in our study population had a higher mean serum PTH than nonobese individuals. However, this did not correlate to a difference in serum calcium, phosphorus, or alkaline phosphatase. These findings are similar to those from the two published studies that examined the clinical features of obese PHPT patients (10, 11): in a report of 776 PHPT patients, morbid obesity (BMI ≥ 35 kg/m²) was associated with a higher preoperative serum PTH but a similar serum calcium as leaner PHPT patients (11). The second study from 196 PHPT patients also reported a higher preoperative serum PTH but similar serum calcium concentration in obese vs overweight and lean patients (10). In our population, serum 25-OHD (available in half of the patients) was also not different in the groups of patients, suggesting that vitamin D deficiency may not explain the higher serum PTH in the obese. The similar serum calcium concentration despite higher serum PTH in obese PHPT patients has been ascribed to difference in responsiveness of serum calcium to serum PTH elevation (ie, resistance to PTH), as had previously been reported in obese healthy (non-PHPT) subjects (4, 5). In our cross-sectional analysis, the degree of hypercalcemia with increasing serum PTH varied across BMI quartiles (Figure 2B). This finding is somewhat compatible with the resistance to PTH mechanism previously proposed, although dynamic studies (varying serum calcium and measuring PTH response in individual patients) would be needed to definitively confirm or refute this hypothesis in obese vs. nonobese PHPT patients.

With respect to complications of PHPT, obese patients in our study had a higher prevalence of hypercalciuria, and kidney stones and a lower prevalence of osteoporosis than the nonobese population. Even after adjusting for potential confounders, including age, gender, ethnicity, serum PTH, serum calcium, hydrochlorothiazide use, and history of diabetes, these associations remained essentially unchanged, except for prevalence of nephrolithiasis, which was no longer significant. These results suggest that the impact of obesity on the clinical presentation of PHPT is independent of the biochemical severity of the disease (ie, degree of hypercalcemia or hyperparathyroidism). These results are compatible with the known impact of obesity on the prevalence of hypercalciuria, nephrolithiasis, and osteoporosis in the general (non-PHPT) population (14–16).

The greater prevalence of hypercalciuria and nephrolithiasis with obesity in the non-PHPT population have ascribed to differences in dietary intake of salt and protein as well as to changes in insulin sensitivity that can in turn alter urine pH, a key biochemical factor in the process of kidney stone formation (17). In our cohort, 24-hour urinary calcium was higher in the obese compared with the nonobese, and the greater prevalence of hypercalciuria persisted after controlling for the presence of diabetes (a downstream consequence of insulin resistance). However, measurements of other risk factors that could link obesity to stone formation (such as urine sodium, pH, oxalate, citrate, and sulfate) were not obtained in our patient population. The lower prevalence of osteoporosis despite greater urinary calcium excretion in OB could be ascribed to either greater starting BMD in OB (prior to onset of PHPT) or to the fact that excreted calcium in PHPT may be primarily of intestinal origin (ie, derived from greater absorbed calcium) rather than skeletal origin (bone resorption). Still, despite the higher average BMD among OB PHPT patients, fracture prevalence was not significantly different between the OB and NO groups, possibly due to the relatively young age of the study population (average age of 57 y).

In terms of histological findings, we found that the weight of resected parathyroid glands was higher in obese patients, in accordance with prior studies. One previous report linked the higher parathyroid gland weight in obesity to greater fat content (10). Rates of multiglandular disease were not higher with obesity.

When evaluating PHPT patients, clinicians recommend parathyroidectomy for patients with history of kidney stones or in the presence of osteoporosis or severe hypercalciuria. In our analysis, these complications appear to be modulated by obesity itself and may not be related to severity of PHPT presentation alone. Thus, compared with nonobese PHPT patients, obese patients with PHPT may be referred to surgery more often due to a history of kidney stones and less frequently due to osteoporosis.

Finally, some investigators have found that parathyroidectomy in PHPT patients help to improve some cardiovascular (CV) risk indices (18, 19), especially in those patients with preexisting CV disease (19). However, other investigators have not found the same association (20). The CV benefits of parathyroidectomy in obese PHPT patients who are more likely to exhibit CV risk factors merits further evaluation.

Some limitations of this study are noted. Our cohort consisted of PHPT patients who have undergone parathyroidectomy, and the findings may not necessarily apply to all patients with mild and/or asymptomatic PHPT. Furthermore, BMD testing and 24-hour urine collection results were available on only approximately two thirds of the cohort because some patients underwent parathyroidectomy without these tests having been performed (due to presence of symptoms such nephrolithiasis or known diagnosis of osteoporosis obviating the need for additional testing prior to parathyroidectomy). Patients with missing tests were clinically similar to those with nonmissing results in terms of age, gender, BMI, and serum calcium distribution (Supplemental Table 2), making it less likely that our results are biased due to missing data. Finally, given the cross-sectional nature of this study, causal inferences between obesity and different complications of PHPT cannot be assumed.

Conclusions

In patients with PHPT, obesity itself is an independent risk factor for hypercalciuria and nephrolithiasis and is protective against osteoporosis. Although obese patients have a tendency toward higher serum PTH, there does not appear to be a resistance to PTH in the obese based on the similar serum calcium-PTH curve between obese and nonobese PHPT patients. Further studies in obese PHPT patients are necessary to evaluate the impact of parathyroidectomy on the clinical features of PHPT and to better understand the dynamic relationship between serum calcium and PTH.