Serum Phosphate: A Neglected Test in the Clinical Management of Primary Hyperparathyroidism

Abstract

Background

Although the inverse correlation between serum PTH and phosphate (P) levels in patients with primary hyperparathyroidism (PHPT) is well known, the relationship between P levels and the clinical picture of the disease has not been well investigated. This was thus the aim of this paper.

Patients

A total of 472 consecutive patients with PHPT attending our center were retrospectively evaluated at diagnosis.

Results

P levels lower than 2.5 mg/dL (HypoP) were found in 198/472 patients (41.9%). HypoP was mild (2-2.5 mg/dL), moderate (1-1.9 mg/dL), and severe (<1 mg/dL) in 168 (84.9%), 30 (15.1%), and 0 cases, respectively. P levels were lower in males than females. Patients with more severe bone density impairment at the radial (but not the vertebral or femoral) site had P levels significantly lower than other patients. PHPT severity was worse in HypoP patients, both clinically (higher prevalence of renal stones, but not of osteoporosis) and biochemically (higher serum calcium and PTH levels). All patients in the moderate HypoP group were either symptomatic or asymptomatic reaching surgical indication according to the latest guidelines.

Conclusions

We observed a relationship between P levels and biochemical and clinical features of PHPT severity. In asymptomatic PHPT patients, even moderate HypoP is predictive of surgical indication, regardless of age and hypercalcemia severity.

Classic primary hyperparathyroidism (PHPT) is characterized by inappropriately high concentrations of the PTH associated with high serum and urinary calcium, and low serum phosphate (P) concentrations (1, 2).

Measuring the levels of serum calcium (Ca) is essential in the diagnosis of PHPT and is included along with 24-hour urinary calcium (UCa), among the surgical criteria recommended by latest international guidelines for asymptomatic PHPT (aPHPT) management (3). On the other hand, serum and urinary P levels are not relevant when diagnosing PHPT, nor are they among the criteria for performing surgery in PHPT patients.

Current guidelines for PHPT management do not mention any cutoff for P levels and no recommendation is based on the level of this parameter (3-5).

It is widely accepted that PTH and P levels are inversely correlated, and the higher PTH levels, the lower the P levels (6, 7). Furthermore, in PHPT patients, serum P levels are reportedly lower in stone-formers (8, 9) and in males (10).

To the best of our knowledge, there are no published reports specifically exploring the relationship between P and the clinical manifestations of PHPT. We thus retrospectively evaluated the impact of P levels on the clinical manifestations of PHPT in a large, single-center, unselected series.

Patients and Methods

Design

A retrospective survey was conducted on the medical records of all patients diagnosed with PHPT who presented at our department between January 1997 and December 2020.

The study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional review board and the ethics committee of our institution. No informed consent was required for this study because we only retrospectively accessed a deidentified database for analysis purposes. All data had originally been collected as part of routine clinical and psychological procedures.

Patients

The patients were referred by general practitioners, primary care clinics, and subspecialty clinics.

PHPT diagnosis had been established by the presence of hypercalcemia and concomitant inappropriately high levels of serum PTH on at least 2 separate occasions (reference range 8.4-10.2 mg/dL and <65 ng/L for Ca and PTH, respectively; see Methods section).

Patients diagnosed with normocalcemic PHPT, multiple endocrine neoplasia, hyperparathyroidism-jaw tumor syndrome, familial hypocalciuric hypercalcemia, and parathyroid carcinoma were excluded.

No patients had been taking calcium or vitamin D supplements, estrogens or testosterone or selective estrogen receptor modulators, or bone-active medications for at least 6 months. Patients taking proton pump inhibitors or antacid medications at diagnosis were excluded from this series.

In agreement with Bilezikian et al (3), patients were classified as aPHPT on the basis of a lack of radiological signs of bone involvement, nephrolithiasis, and symptoms of hypercalcemia. Regarding bone involvement, all patients had routinely undergone dual X-ray absorptiometry and a radiographic evaluation of the skull and hands to check for signs of excess PTH effects on the bones, such as osteitis fibrosa cystica, subperiosteal resorption in the fingers, salt and pepper mottling of the skull, or brown tumors.

In terms of kidney involvement, patients were classified as symptomatic either if they had a recorded positive history for renal stones (ultrasound examination, urography, plain radiography, history of passing stones, or their endoscopic or surgical removal) or if renal stones (or calcinosis) had been diagnosed by routinely performed ultrasound in either asymptomatic or symptomatic patients.

The aPHPT patients not meeting the surgical criteria set out by the latest international guidelines (3) and retrospectively applied to all patients were considered as “mild asymptomatic” patients (11).

Methods

All blood samples were collected after overnight fasting and rest.

Serum total Ca, P, and creatinine as well as UCa were analyzed by a standard autoanalyzer using colorimetric and enzymatic methods, whereas ionized serum Ca was analyzed by a specific probe after correction for pH.

The minimum detectable level of serum P was 0.3 mg/dL, with inter- and intra-assay variation coefficients of 2% and 2.7%, respectively. Hypophosphatemia (HypoP) was defined as a confirmed serum level below 2.5 mg/dL, and it was considered mild, moderate, or severe if serum P was 2 to 2.5 mg/dL, 1 to 1.9 mg/dL, and <1 mg/dL, respectively (12).

The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration formula (13):

where SCr represents serum creatinine (in mg/dL), k is 0.7 for women and 0.9 for men, α is −0.329 for women and −0.411 for men, min represents the minimum of SCr/k or 1, and max represents the maximum of SCr/k or 1.

Serum intact PTH concentrations were measured up to 2012 using a 2-site immunochemiluminometric assay (Immulite 2000; DPC, Los Angeles, CA) with inter- and intra- assay variation coefficients of 6.3% to 8.8% and 4.2% to 5.7%, respectively. After 2012, serum intact PTH concentrations were measured using a new second-generation immunochemiluminometric assay (Cobas e411, Roche Diagnostics) with inter- and intra-assay variation coefficients of 3.1% to 6.5% and 1.4% to 3.2%, respectively. Normal ranges are 20 to 65 ng/L and 15 to 65 ng/L, respectively.

Serum 25-hydroxy-vitamin D (25OHD) levels were measured by a radioimmunoassay (DIAsource 25OHVit. D3-Ria-CT Kit - DIAsource Immuno Assays S.A., Nivelles, Belgium), with a detection limit of 0.6 µg/L (1.5 nmol/L) and inter- and intra-assay variation coefficient of 5.3% and 4.7%, respectively. Our laboratory periodically conducts quality control tests on kits used with materials provided by the manufacturer. Our laboratory is also a member of the External Quality Assessment Scheme for the estimation of 25OHD, conducted by the QualiMedLab-CNR (Pisa, Italy), as a means of determining the accuracy of results. A level <20 µg/L was considered the cutoff for deficiency.

Bone mineral density was measured at the lumbar spine (L2-L4), proximal femur, and distal third of the nondominant radius using the same instrument (DXA QDR-4500, Hologic, Bedford, MA) throughout the study period. Minor upgrades to the bone mineral density instrument, above all in the reporting and duration of the procedure, did not significantly affect the results. Data were analyzed as absolute measurements (in grams per square centimeter) and reported as T scores.

All patients underwent standard renal ultrasound using a 2- to 5-MHz-wide band convex transducer. For a definitive diagnosis of stones, which enables patients to be classified as positive or negative for nephrolithiasis, radiologists looked for hyperechogenic spots that were more than 2 mm in diameter, with a multiplanar evaluation of specific signs such as echogenicity, posterior acoustic shadowing, or a positive twinkle sign.

Statistical Analysis

Variables were preliminarily tested for normal distribution with the Shapiro-Wilks W test and data were expressed as mean ± SD when normally distributed, and as median and interquartile range when not normally distributed.

Continuous variables with non-normal and normal distribution were analyzed by the Mann-Whitney U test and t test for unpaired samples, respectively, as appropriate. Differences in categorical variables were analyzed by χ ² or Fisher test, as appropriate.

Two separate sets of linear regressions were conducted, with the P value set as the dependent variable. The first set was univariate, and the second included all the variables whose β-coefficient was significant in the univariate analysis. In the case of multiple independent variables, a multicollinearity test was performed. Variables were rejected from the analysis if there was a variance inflation factor > 2.

The level of statistical significance was set at P ≤ 0.05. All calculations were performed using SPSS (IBM SPSS Statistics, v. 21).

Results

Data were analyzed from 472 consecutive patients diagnosed with PHPT, with a female (F):male (M) ratio of 3.2:1 and a mean age of 61.2 ± 13 years.

Table 1 compares P levels in patients subdivided according to their demographic and clinical characteristics. Briefly, P levels were significantly lower in M patients and those whose 25OHD levels were lower than 20 µg/L. The gender difference remained significant also when comparing M with F patients divided according to menopausal state (pre-F and post-F, Table 2). Moreover, P levels were significantly lower in PHPT patients whose worst T-score was at the radial level than in patients whose worst T-score was at the lumbar or femoral site.

HypoP was found in 198 of the 472 patients (41.9%). Mean serum P in this group was 2.1 ± 0.26 mg/dL (range 1.1-2.5). HypoP was severe in nil and moderate in 30/198 (15.1%). Among these patients with moderate HypoP, none was classified as “mild” asymptomatic PHPT.

In patients with moderate HypoP, we then investigated the presence of the 2 criteria for surgery recommended by the latest guidelines evaluable without further tests (ie, serum Ca higher than 1 mg/dL over the upper limit of normal range and age younger than 50 years), showing that 8/30 (26.6%) had total Ca lower than 11.2 mg/dL, and 5/30 (16.7%) were also older than 50 years. Specifically, among these 5 patients, 2 were classified as symptomatic PHPT, whereas 3 were classified as aPHPT meeting surgical criteria. The surgical indication was met because of T score <-2.5 at the radial site and eGFR <60 mL/min in 2 and 1 patient, respectively.

Table 3 compares patients with and without HypoP. HypoP patients were significantly more frequently M, with a symptomatic disease, higher levels of PTH and serum Ca (both total and ionized), and lower levels of 25OHD.

In the whole series, at the linear regression analysis, P levels were negatively correlated with PTH, serum total, and ionized Ca, and positively correlated with 25OHD levels (Table 4).

As serum total and ionized Ca showed a variance inflation factor greater than 4, the variable total Ca was removed from the multiple linear regression analysis to avoid inflating the SE of the regression coefficients. In the multivariate regression analysis, P levels were significantly associated with ionized Ca (ß = -0.264, P < 0.001) and PTH levels (ß = -0.174, P = 0.006).

Finally, we compared P levels in asymptomatic patients, subdivided according to the presence or absence of the criteria recommended by the latest international guidelines for surgical intervention: P levels were significantly lower in patients with serum Ca 1 mg/dL higher than the normal limit than in patients with Ca <11.2 mg/dL (2.39 ± 0.58 vs 2.79 ± 0.54 mg/dL, respectively; P < 0.001) and in patients whose UCa levels were higher than 400 mg/day than in those with lower levels (2.24 ± 0.66 vs 2.77 ± 0.55 mg/dL, respectively; P = 0.002).

Discussion

The rate of HypoP in our unselected consecutive series of 472 PHPT patients was 41.7%, and was moderate in 15.1% of them. Serum P levels were significantly correlated with PTH, serum Ca, and 25OHD levels. Moreover, P levels were significantly lower in M than in F, and in patients whose worst T-score was at the radial level than in the remaining patients.

PHPT patients with HypoP were more frequently M with a higher rate of symptomatic forms of PHPT and renal stones. All patients with aPHPT and moderate HypoP met at least 1 criterion for surgery established by the latest international guidelines (3).

To date, P levels are not relevant in establishing the diagnosis of PHPT, nor are they among the criteria for performing surgery in aPHPT patients. Current guidelines for PHPT management do not mention any cutoff for P levels; in fact, no recommendations are based on this parameter (3-5). The reported prevalence of HypoP in PHPT is widely variable, ranging from 10% to 20% (14) to >50% (15). However, the serum Ca/P ratio has recently been proposed as an accurate index to diagnose PHPT, especially in patients with asymptomatic or nonclassical phenotypes of PHPT (16). The good diagnostic value of Ca/P ratio in detecting PHPT is driven by the known divergent pattern, in a mirror image, of serum Ca and P in PHPT (17, 18).

Although serum P alone is universally considered unreliable for diagnosing PHPT (16), its relationship with clinical manifestations of PHPT has been little investigated (19-21).

The results of our series confirm the generally accepted inverse relationship between PTH and serum P levels (6, 7) and the direct correlation between serum P and vitamin D levels, whose deficit is associated with some features reflecting a more severe biochemical and clinical phenotype of PHPT (22, 23).

When comparing symptomatic, asymptomatic meeting surgical criteria, and “mild” asymptomatic forms of PHPT, we also found a nonsignificant increase in P levels.

In line with previous reports (9, 10), we found lower P levels in M than in F patients. After splitting the F group according to menopausal state, the mean serum P level was still different between M and post-F, whereas it was similar between M and pre-F. This could be explained by the influence of the gonadal hormones on the hypophosphatemic effect of PHPT, as reported in an experimental hyperparathyroidism model (24). However, in the general population, serum P levels have been reported to be significantly higher in women than in men and, among women, in those not receiving estrogen replacement than in those on estrogen treatment (25).

As for renal involvement, some authors have reported lower P levels in stone-formers than in other PHPT patients (8, 9), and not univocal findings have been reported concerning the comparison between silent and symptomatic renal stones (26, 27). In this regard, in our series renal stones were significantly more frequent in PHPT patients with HypoP. Because renal stones are generally more frequent in M with PHPT (10), a gender-dragging effect cannot be ruled out.

Regarding bone involvement, no difference in P levels was recently reported between patients with predominant skeletal symptoms compared with patients with predominant renal manifestation in a small series of inpatients undergoing parathyroidectomy (19). In our series, no significant differences were found in the mean P levels of PHPT patients with or without radiological findings of osteitis fibrosa cystica. However, we found significantly lower P levels in PHPT patients with prevalent densitometric impairment at the radial site than at the vertebral or femoral site. P levels thus seem to be related to a preferential impairment of cortical bone in PHPT, as evaluated at the distal third of the radius (1).

PHPT patients with HypoP had significantly higher levels of PTH and serum Ca, and significantly lower levels of 25OHD. Moreover, they were more frequently M with a higher rate of symptomatic forms of PHPT and renal stones.

The level of HypoP was moderate in 15.1% of cases: all the patients in this group were either symptomatic or asymptomatic reaching surgical indication according to the latest guidelines (3).

In a small, but not negligible, group of patients with moderate HypoP, neither hypercalcemia was higher than 1 mg/dL over the normal range nor was age below 50 years. This finding could be of particular interest for the clinical management of aPHPT patients in which it may be difficult or impossible to carry out the systematic evaluation of all parameters recommended by the current guidelines. Given that in the West, PHPT is a largely asymptomatic disease (1), we believe that moderate HypoP should be considered as a supportive, inexpensive, and easily available means to identify aPHPT patients who may require surgery (16, 28).

Our study has several limitations. Despite the large cohort studied, this is a retrospective single-institution study, which may have been affected by selection bias. These findings cannot therefore be generalized indiscriminately to PHPT patients in other countries and ethnic groups. However, unlike previous studies, this study evaluated a large consecutive series of patients and thus better reflects real-life clinical practice.

Second, we had no information about dietary P intake, potentially affecting serum P levels. The dietary P in the general population is heterogeneous and not easy to estimate accurately. In any case, a recent paper evaluating the Italian-Mediterranean diet in adults found a sufficient P intake (29).

Finally, we did not use the gold standard “liquid chromatography aligned to mass spectrometry” to assess 25OHD levels, and our radioimmunoassay may have slightly overestimated 25OHD levels (in line with several radioimmunoassays). However, all measurements were performed in the same laboratory, thereby ensuring a good quality of data.

In conclusion, P levels in PHPT are related to biochemical and clinical features of disease severity, largely mirroring serum Ca levels. However, in aPHPT patients, moderate HypoP indicates that surgery may be required, regardless of age and hypercalcemia severity. Further larger controlled studies are needed to confirm our findings.

Abbreviations

Abbreviations
 
• 25OHD

  25-hydroxy-vitamin D

 
• aPHPT

  asymptomatic primary hyperparathyroidism

 
• BMD

  bone mineral density

 
• Ca

  calcium

 
• eGFR

  estimated glomerular filtration rate

 
• F

  female

 
• HypoP

  hypophosphatemia

 
• M

  male

 
• P

  phosphate

 
• PHPT

  primary hyperparathyroidism

 
• Scr

  serum creatinine

 
• UCa

  urinary calcium

Additional Information

Disclosures: Nothing to declare.

Data Availability

Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

References