The chloride/phosphate ratio combined with alkaline phosphatase as a valuable predictive marker for primary hyperparathyroidism in Chinese individuals

The chloride/phosphate ratio (Cl/PO4) has been suggested to have a role in primary hyperparathyroidism (PHPT), but the associations between Cl/PO4 combined with ALP level and PHPT has not been well-studied. Our aim was to investigate the predictive value of combination Cl/PO4 with ALP for PHPT. A cross-sectional retrospective analysis was conducted to examine 172 patients diagnosed with PHPT categorized into two groups: normocalcaemic primary hyperparathyroidism (NPHPT) group and hypercalcaemia PHPT group. We found that Cl/PO4 levels and ALP levels in the NPHPT and hypercalcaemia PHPT group were both significantly higher than normal controls. Cl/PO4 and ALP levels were an independent risk factor for PHPT. Cl/PO4 combined with ALP increased the receiver-operating characteristic curves (ROC-AUC) and the diagnostic value in NPHPT and hypercalcaemia PHPT group (0.913; 95% CI, 0.744–1.000 and 0.932; 95% CI, 0.897–0.966, respectively), specificity of 92.8% and sensitivity of 98%. In conclusion, combination Cl/PO4 with ALP might be a low-cost, simple, available predictive marker of PHPT in Chinese individuals, particularly Chinese remote region where the method used to measure PTH cannot be done. Moreover, due to serum calcium level in NPHPT, Cl/PO4 combined with ALP level measurement have great potential to predict significant occurrence of NPHPT.

Anthropometric evaluation and biochemical measurements. General information on the 222 participants, such as gender and age, was obtained through inquiry. Anthropometric parameters such as height, body weight, and blood pressure were measured and the Body Mass Index (BMI) was calculated as weight/height2. Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured enzymatically on a Hitachi 7600 analyser (Hitachi, Tokyo, Japan). Serum sodium, potassium, alkaline phosphatase (ALP), phosphorus, chloride, and calcium were measured using automated techniques before surgery. ALP was detected by AMP buffer method of the ALP kit (KH-G-C-012) obtaining from Shanghai Kehua Bio-engineering Co., Ltd (Shanghai, China). The normal range for ALP is from 14 to 112 U/L. We calculated the chloride/phosphate ratio (Cl/PO4) as chloride (mEq/L)/phosphate (mg/100 mL). The total serum calcium levels used in our research was adjusted for albumin. The formula to use is: corrected calcium = measured total serum calcium in mg/dL + 0.8*(4.0-patient's serum albumin concentration in g/dL). Serum PTH was measured by using an electrochemiluminescence immunoassay (ECLIA) on an Elecsys autoanalyser (E170; Roche Diagnostic, Mannheim, Germany). All biochemical values were measured from the same blood sample collected after an overnight fast. Related indicators were detected by Department of Medical Laboratory in Shanghai Jiao Tong University Affiliated Sixth People's Hospital. And the detection methods of related indicators had not changed.
Statistical analyses. Statistical analysis was performed using SPSS for Windows software (ver. 17.0; SPSS Inc., Chicago, IL, USA). Normally distributed data are presented as means ± standard deviation (SD). Non-normally distributed data determined using the Kolmogorov-Smirnov test are expressed as medians and interquartile range (25-75%). Two groups were compared using Student's unpaired t-test. The composition ratio of categorical variables was calculated using the chi-square test. Pearson's correlation was used to evaluate the association among the Cl/PO4 ratio, ALP and other general biochemical parameters. To analyse whether Cl/ PO4 and ALP might be related to PHPT independently, we applied logistic regression models to adjust for other clinical and biochemical variables. The predictive performance of each model was assessed using area the under the receiver operating characteristic (ROC-AUC) curve with 95% confidence intervals (CI). All P-values were two-tailed and statistical significance was considered as P < 0.05.

Results
Subject characteristics and Serum Cl/PO4 or ALP Levels at Baseline. Table 1 summarises the clinical characteristics of the study subjects with primary hyperparathyroidism (PHPT) and normal controls. PHPT was subdivided into normocalcaemic primary hyperparathyroidism (NPHPT) and hypercalcaemia groups according to the serum calcium level. In this cross-sectional analysis of NPHPT and normal controls, no significant differences in the proportion of males, BMI, diastolic blood pressure (DBP), ALT, AST, or potassium were observed. However, patients with NPHPT were older, and had higher systolic blood pressure (SBP), ALP [107.0 (74.0-269.0) vs. 61. 5  3), P = 0.002], and a lower haemoglobin (Hb) level (P < 0.001), than controls. After adjusting for age and sex in the NPHPT group, SBP (P = 0.118) and calcium (P = 0.094) lost significance. The hypercalcaemia group of PHPT was older (P < 0.01) than the normal controls, 3)] were higher in the hypercalcaemic PHPT group than in the normal controls (all P < 0.001). ALT (P = 0.757) did not differ in the hypercalcaemic PHPT group after adjusting for age and sex. There were no differences in DBP, AST, or potassium levels among the groups at baseline. Tables 2 and 3, correlation analysis revealed that the Cl/PO4 and ALP levels correlated with PTH and calcium in all of the subjects. Cl/PO4 levels had positive associations with age (r = 0.134, P = 0.047), ALP (r = 0.357, P < 0.001), PTH (r = 0.463, P < 0.001), sodium (r = 0.202, P = 0.003), and calcium (r = 0.586, P < 0.001) and negative associations with BMI (r = −0.251, P < 0.001) and Hb (r = −0.235, P < 0.001). Simultaneously, the ALP level was positively associated with Cl/PO4 (r = 0.357, p < 0.001), PTH (r = 0.647, p < 0.001), calcium (r = 0.58, p < 0.001), and negatively associated with BMI (r = −0.141, p = 0.036) and Hb (r = −0.307, p < 0.001). To minimise confounding effects in different groups, additional analyses were performed. Cl/PO4 and ALP levels were still significantly positively associated with PTH, calcium in both groups.

Baseline Cl/PO4 and ALP Level Are Predictors of PHPT.
The stepwise multiple regression model was refined to assess the factors that were independently associated with PHPT (Table 4 and Table 5). Cl/PO4 levels were an independent risk factor for hypercalcaemia group of PHPT (odds ratio (OR): 0.832; 95% CI, 0.769-0.900; P < 0.001), after adjusting for age, sex, BMI, sodium and Hb. To determine if calcium had a confounding effect, the same stepwise analysis was performed in the NPHPT subgroup. Cl/PO4 levels were an independent risk factor for this group (odds ratio (OR): 0.797; 95% CI, 0.667-0.951; P = 0.012), after adjusting for some factors ( Table 4). The ALP model also yielded results similar to the Cl/PO4 model (Table 5). This result shows that the Cl/ PO4 and ALP levels were predictors of diagnosing PHPT.
Predictive value of the Cl/PO4 plus ALP model. The receiver-operating characteristic curves shown in Fig. 1 represent the predictive value of different models for PHPT events in both groups. In the NPHPT group, the ROC-AUC was 0.743 (95% CI, 0.587-0.899) for Cl/PO4 alone. However, Cl/PO4 combined with the ALP model significantly increased the ROC-AUC (0.913; 95% CI, 0.744-1.000). Cl/PO4 >33.07 (specificity = 80%, sensitivity = 73.7%), ALP >71.5 (specificity = 66%, sensitivity = 94.7%), the combination of Cl/PO4 and ALP (specificity = 83%, sensitivity = 94%, positive predictive value = 53.1% and negative predictive value = 73.9%) can be used to predict diagnosing PHPT. In the hypercalcaemia group of PHPT, the ROC-AUC of Cl/PO4 alone was 0.897 (95% CI, 0.851-0.942), and increased significantly to 0.932 (95% CI, 0.897-0.966) when combined with ALP ( Fig. 2). A high specificity of 80% and sensitivity of 88.9% were achieved at a Cl/PO4 threshold of 33.05. A high specificity of 92% and sensitivity of 67.3% were achieved at a ALP threshold of 90.0. When Cl/PO4 combined with ALP, specificity of 92.8%, sensitivity of 98%, positive predictive value of 51.4% and negative predictive   addition to PTH 7,10 . The predictive value of Cl/PO4 < 33 for excluding the diagnosis was greater than 99%. Some studies 13,24 have reported that serum ALP is positively associated with PTH in PHPT, and PTH significantly influenced the bone isoenzyme of ALP. Nasir et al. 25 suggested ALP as a readily available marker to detect metabolic bone disorders. ALP levels were shown to be increased in PHPT patients, and serum ALP is a potential indicator for the occurrence of postoperative hypocalcemia in PHPT patients 9,23 . Our findings which Cl/PO4 threshold >33.07 and ALP >71.5 in the NPHPT group, Cl/PO4 >33.05 and ALP threshold of 90.0 in the hypercalcaemia group of PHPT were consistent with theirs, although there were some differences due to the different populations. We found that both Cl/PO4 and ALP might be helpful biochemical predictors of PHPT. Moreover, the combination model significantly increased the predicted value of PHPT in both groups. Therefore, our finding suggests that Cl/PO4 combined with ALP is an inexpensive screening tool for predicting PHPT without the costly PTH assays. These findings provide important clues regarding the pathophysiological mechanism underlying the association between Cl/PO4 and PHPT. One possible explanation is that PTH decreases the tubular reabsorption of phosphate in the proximal renal tubule, which increases urinary phosphate and decreases serum phosphate 25,26 . The reabsorption of bicarbonate from the proximal renal tubule is decreased by the action of PTH, causing the reabsorption of chloride and increasing the secretion of bicarbonate from the kidney 27 . These findings may partly account for the observed association between serum Cl/PO4 and PHPT.
Previous studies 25,[28][29][30] suggested that elevated ALP levels existed in the patients of primary hyperparathyroidism. Serum alkaline phosphatase originated principally from bone. Elevated level of ALP is a marker of  osteoblast activity increase. Bone metabolism is a dynamic balance of the process by common regulation of bone resorption and bone formation. Although osteoclasts are the primary target organ for PTH to accelerate bone resorption and destruction, functional PTH receptors were expressed in both osteoblasts and osteoclasts. When PTH combined with its receptors, they could indirectly stimulate effect mediated by osteoblasts. Due to bone formation and bone resorption processes coupling in bone remodeling units, osteoblast activity also increased in PHPT with the increase in osteoclast activity. Osteoblasts responded to PTH with an increase in ALP as bone morphogenetic marker. Additionally, as the research was observing this trend of ALP slight elevation currently, the further studies including more samples should be performed. So in clinical work, when we found that ALP was close to upper limit of the normal range or higher than normal range, we need to further detect other laboratory indicators of PHPT so that they could increase the predicted value of PHPT. The above findings may partly account for the observed association among serum Cl/PO4, ALP and PHPT. Simultaneously, in the study, we found significant difference of Hb and sodium between patients with hyperparathyroidism and normal, but there was no significant difference after adjusting for Hb and sodium. On the one hand, the pathophysiological mechanism of elevated Hb in PHPT patients has not yet known. On the other hand, as for serum sodium level was different in two groups, one possible pathophysiological mechanism was that patients with hyperparathyroidism have higher aldosterone levels, and elevated aldosterone levels have positive correlation with sodium intake 30 . Another possible pathophysiological mechanism was that the reabsorption of sodium was induced by PTH 25,26 . Further research is needed to clarify the potential pathophysiological mechanism. This study had several limitations. First, the age and sex of control group were not very match up to the PHPT group, so we performed the analysis adjusting for age and sex in each step to eliminate the bias from age and sex. Second, a small sample size from a single centre cannot fully reflect the entire Chinese population. The study was based on parathyroid adenoma of PHPT. Further prospective studies that include parathyroid carcinoma and parathyroid hyperplasia cases should be performed. Additionally, we have analysed the patients of PTH-dependent hypercalcemia in the study. Whether the chloride/phosphate ratio combined with alkaline phosphatase as a valuable predictive marker for PTH-independent hypercalcemia, further study is urged to for certification. Third, given the cross-sectional design of the study, the most important limitation is a lack of causation. Therefore, follow-up studies should be prospective and include a large number of cases to confirm our findings.
In conclusion, we revealed an association among Cl/PO4, ALP and PHPT, especially NPHPT, in Chinese individuals. Due to its low cost and availability, Cl/PO4 could increase the rate of diagnosis of PHPT, particularly NPHPT, when combined with ALP. The clinical significance of this model should be corroborated in larger studies.