Association between Parathyroid Hormone Levels and Inflammatory Markers among US Adults
Background and Aims. High levels of parathyroid hormone (PTH) appear to be associated with an increased mortality. Previous studies concerning the relationship of inflammatory markers with hyperparathyroidism have yielded inconsistent results. This study investigated whether serum PTH concentrations were independently associated with several inflammatory markers among the US adults. Materials and Methods. Using data from the National Health and Nutrition Examination Survey, we examined the relation between serum PTH and C-reactive protein (CRP), red cell distribution width (RDW), and platelet-to-lymphocyte ratio (PLR) levels with weighted linear regression. Additionally, we examined the relation with increased modified Glasgow Prognostic Score (mGPS) by using weighted logistic regression. Results. CRP, RDW, and PLR values increased with increasing serum PTH concentration. After extensively adjusting for covariates, CRP and RDW increased linearly and across PTH categories (all ), while PLR marginally increased ( and using PTH as a categorical and continuous variable, resp.). The odds ratio of increased mGPS was 1.11 and 1.31 across PTH categories and with increasing PTH levels continuously. Conclusion. These nationally representative data indicate that serum PTH levels are independently associated with several inflammatory markers in the US population. The casual relationship between PTH levels and inflammation remains to be elucidated.
Parathyroid hormone (PTH) is the principle regulator of calcium and phosphorus homeostasis. PTH modulates osteoblast activity and osteoclast resorption, increases renal tubular reabsorption of calcium, and stimulates conversion of 25-hydroxyvitamin D to 1,25-dihydroxyvitamin D by 1-hydroxylase in the kidney. In addition, PTH acts on bone cells to increase expression of fibroblast growth factor-23 (FGF-23) . Higher PTH concentrations are associated with an increased mortality risk among general older populations [2, 3]. In the presence of hypovitaminosis D, absence of secondary hyperparathyroidism is characterized by lower rates of bone turnover and reduced mortality when compared with counterparts who manifest a physiological PTH elevation in response to vitamin D deficiency . PTH has been considered to be a complementary biomarker in heart failure .
PTH stimulates interleukin-6 (IL-6) production by osteoblasts and liver cells [6, 7]. In turn, IL-6 may modulate acute-phase protein synthesis in the liver [8, 9]. It has been proposed to administrate vitamin D in the elderly to reduce serum levels of IL-6 and C-reactive protein (CRP) and, possibly, to decrease the risk of thromboembolic vascular events . In this context, patients with hyperparathyroidism will, theoretically, have higher levels of IL-6, CRP, or tumor necrosis factor- (TNF-). However, previous studies have had conflicting results. Some showed elevated levels of these inflammatory markers in hyperparathyroidism [11–14], whereas others found that levels of CRP, IL-6, and leukocytes were similar in patients and controls [15, 16]. Furthermore, results for the effects of parathyroidectomy on subclinical inflammation were also inconsistent. Studies have showed decreased , increased [16–18], or unchanged [13, 19] levels of inflammatory markers after parathyroidectomy.
Such association studies may be underpowered to detect meaningful differences due to small sample sizes. Serum concentrations of 25-hydroxyvitamin D and PTH were found to independently associate with blood pressure and hypertension in the National Health and Nutrition Examination Survey (NHANES) in 2003 through 2006 . The study did not specifically evaluate the relationship between PTH and CRP levels. In addition, other parameters such as IL-6 and TNF- were not available in the NHANES database. Recently, we have validated the significance of another inflammatory marker, neutrophil-to-lymphocyte ratio, in differentiated thyroid cancer . In the present study, we hypothesized that PTH levels might be associated with inflammatory status in the general US population. To test this hypothesis, we examined the cross-sectional associations between demographic, lifestyle, and serum factors and intact PTH using the large-scale data set released by the US NHANES 2003-2004 and 2005-2006.
2. Materials and Methods
2.1. Study Design and Population
NHANES provides nationally representative cross-sectional data for the health status of the US civilian noninstitutionalized population. In this analysis, we used data from the 2003-2004 and 2005-2006 NHANES that obtained serum PTH levels. The design and operation are available on the Centers for Disease Control and Prevention website [22, 23]. This analysis was limited to participants 20 years of age and older. Ethical approval for the study was obtained from the Research Ethics Review Board of National Center for Health Statistics, and all participants gave written informed consent.
2.2. Data Collection
Data were collected at all study sites by trained personnel using standardized procedures. Sociodemographic information such as age, gender, and race/ethnicity was recorded during the interview. Current smokers were defined as participants who were currently smoking and had smoked ≥100 cigarettes in their life. Blood pressure was measured at the mobile examination centers by physicians with mercury sphygmomanometers using a standard protocol. Up to four readings of systolic and diastolic blood pressure were averaged.
2.3. Serum Measurements
The complete blood count and leukocyte differential count were measured with the Beckman Coulter MAXM analyzer. Red cell distribution width (RDW) was derived from the coefficient of variation of the red cell volume distribution histogram. The platelet-to-lymphocyte ratio (PLR) was calculated. Serum creatinine concentration was measured by kinetic Jaffe assays with picrate (Beckman LX20). Estimated glomerular filtration rate (eGFR) was determined using the chronic kidney disease epidemiology collaboration (CKD-EPI) equation . Serum total calcium was measured by an indirect ion-selective electrode method (Beckman LX20). When serum albumin level was 4.0 g/dL, serum calcium levels were corrected using the following formula: corrected calcium (mg/dL) measured total calcium (mg/dL) + 0.8 [4.0 − serum albumin (g/dL)]. Serum 25-hydroxyvitamin D was measured using a Diasorin (formerly Incstar) 25(OH)D assay. Serum PTH was measured by electrochemiluminescence immunoassay on an Elecsys 1010 autoanalyzer (Roche Diagnostics). CRP was measured by latex-enhanced nephelometry on a Behring nephelometer.
The modified Glasgow Prognostic Score (mGPS) was calculated as previously described . Participants with an elevated C-reactive protein concentration (>1 mg/dL) and a decreased albumin concentration (<3.5 g/dL) were assigned score 2. Those with an elevated C-reactive protein concentration (>1 mg/dL) were assigned score 1, and patients with a C-reactive protein concentration of ≤1 mg/dL and any albumin concentration were assigned score 0.
2.4. Statistical Analysis
All statistical analyses were computed by using survey commands of STATA (STATA Corporation) to incorporate sample weights and to adjust for clusters and strata of the complex sample design. Where distributions appeared nonnormal, we used natural log transformations to normalize their right-skewed distributions (white blood cell count, triglycerides, CRP, and PTH).
Serum PTH levels were stratified into clinically relevant categories (6–39, 40–59, 60–99, and ≥100). Potential confounding factors were chosen on the basis of previous studies or of their biologic plausibility. Tests of trend were calculated across PTH groups. Variance estimates were calculated using Taylor series linearization. Sample weights, which account for the differential probabilities of selection, nonresponse, and noncoverage, were incorporated into the variance estimation process. Univariate modeling was performed to determine whether a directional trend existed between variables and three inflammatory markers (CRP, RDW, and PLR). We examined whether the observed associations persisted within the subgroups stratified by eGFR and PTH. Subsequently, we constructed full multivariable linear regression models to test associations between inflammatory markers and serum PTH. Multivariate models were adjusted for age, sex, race/ethnicity, smoking status, glycohemoglobin, albumin, high-density lipoprotein, (logged) triglycerides, corrected total calcium, 25-hydroxyvitamin D, and eGFR. Trends across PTH categories were also assessed in linear regression models by using continuous (logged) PTH values. In addition, we performed logistic regression with a dichotomous outcome of increased mGPS (>0), adjusting simultaneously for the same covariates. All statistical tests were considered significant for .
Amongst the 8948 participants who formed our main analysis sample, the weighted mean age was 49.4 years. The study sample consisted of 4322 men and 4626 women. Weighted mean PTH was 46.3 pg/mL (SE 0.5). The characteristics of the study according to PTH levels are outlined in Table 1. The results indicate that with increasing PTH levels, age, body mass index, systolic and diastolic pressure, glycohemoglobin, and serum levels of creatinine tended to increase, while eGFR and serum levels of albumin, calcium, phosphorus, and vitamin D tended to decrease with increasing PTH levels. In all, eGFR level was a major determinant of PTH levels.
Three inflammatory markers (CRP, RDW, and PLR) were examined in our study. The weighted mean (95% CI) CRP, RDW, and PLR was 0.49 (0.47–0.51) mg/dL, 12.87% (12.84%–12.91%), and 139.6 (137.7–141.5), respectively. The unadjusted eGFR-stratified levels of CRP, RDW, and PLR by PTH categories are displayed in Figure 1. CRP and RDW values increased at higher PTH categories independent of eGFR tertile (all values <0.01). The unadjusted PLR tended to increase across PTH categories, but the statistically significant difference was observed only among participants in the lowest tertile of eGFR ().
Results of the multivariate linear regression predicting levels of three inflammatory markers are presented in Table 2. After adjusting for age and sex (Model 1), the linear relationships of PTH with CRP and RDW were statistically significant (). These relationships persisted after further adjusting for covariates including race/ethnicity, smoking status, glycohemoglobin levels (Model 2), and other potential confounders (Model 3). After further adjusting for PTH determinants including calcium, vitamin D, and eGFR (Model 4), CRP and RDW remained significantly associated with PTH categories (). Moreover, when (logged) PTH was entered as a continuous variable in linear regression models, it was independently associated with CRP and RDW (both ). PLR also showed increasing trends. PLR was marginally associated with PTH after multivariate adjustment for covariables, assessed either by categories of or by continuous PTH ( and , resp.).
Additionally, systemic inflammation was examined as a dichotomous variable (mGPS > 0). Excluding subjects with missing data of CRP () and albumin (), 1083 (12%) participants had increased mGPS. We found significant correlations between mGPS and other inflammatory markers, namely, CRP, RDW, and PLR (all ). Participants with increased mGPS had higher PTH concentrations (weighted mean 51.9 and 45.5, resp.; ). As shown in Figure 2, the multivariate odds ratios for increased mGPS according to increasing PTH categories were 1, 1.05, 1.29, and 1.27 (odds ratio for trend 1.11). The trend remained significant when (logged) PTH levels were included in the logistic regression as a continuous variable ( values for Model 1 to Model 4 were <0.001, <0.001, <0.001, and 0.013, resp.). The odds ratio was 1.39, 1.42, 1.46, and 1.31 from Model 1 to Model 4, respectively. This is consistent with the observation that serum albumin levels were inversely associated with PTH concentrations.
PTH is considered as a uremic toxin which may inflict damage on multiple organs . Recently, we have shown that in dialysis patients with secondary hyperparathyroidism, the symptom burden had a negative impact on patients’ quality of life, and parathyroidectomy significantly improved symptoms and quality of life . Epidemiological studies have consistently reported higher total and cardiovascular mortality associated with PTH levels [2, 3, 28, 29]. In addition, associations with cancer mortality and noncancer, noncardiovascular mortality have been reported . Pleiotropic effects of PTH are also reflected by numerous nonskeletal, nontraditional manifestations of primary hyperparathyroidism .
Accumulating evidence suggests that higher PTH levels may be associated low-grade inflammation. Adipose tissue from patients with primary hyperparathyroidism showed upregulation of inflammatory genes . Dietary-induced hyperparathyroidism in rodents led to increased serum proinflammatory cytokine production . Furthermore, in obese adolescents, PTH levels were closely correlated with CRP levels and triglycerides: high-density lipoprotein ratio . The PTH level, but not the vitamin D level, is an independent predictor of metabolic syndrome . However, a paradoxical increase in CRP levels after parathyroidectomy was observed in some studies [16, 18]. Therefore, the cause-and-effect relationship between PTH and inflammation remains unclear.
Inflammatory markers vary considerably across studies. The most extensively studied biomarkers of inflammation in cardiovascular disease are CRP and IL-6. The CRP level is consistently associated with the risk of cardiovascular disease . Interestingly, elevated levels of CRP are associated with increased risk of several types of cancer . In this study, we found a positive association between PTH and CRP, after adjustment for multiple potential confounders. Our results are in agreement with findings from epidemiological studies showing that PTH correlates with cardiovascular and even cancer mortality . RDW is a quantitative measure of variability in the size of circulating erythrocytes, with higher values reflecting greater heterogeneity in cell sizes (i.e., anisocytosis). RDW is an easy, inexpensive, routinely reported parameter as a part of the complete blood count test. Recent studies exhibit a strong correlation between RDW and CRP [37–39]. Of note, data from NHANES III showed that RDW was associated with cardiovascular disease, cancer, and chronic lower respiratory tract disease . RDW was shown to be a powerful predictor of mortality in community-dwelling older adults . As a surrogate marker for the inflammatory state, we found that RDW levels strongly correlated with PTH and CRP levels. These findings further corroborate the association between PTH and inflammation.
Chronic inflammation is often associated with reactive thrombocytosis. Accompanied with systemic inflammation, release of various immunological mediators (including IL-6) increases circulating platelet counts as a result of megakaryocyte proliferation . Platelets release the thromboxanes and other mediators, and consequently, patients with higher platelets may have increased inflammation. Smith and colleagues found that PLR is a prognostic marker in patients with periampullary cancer . Subsequent study showed that PLR is an independent predictor of long-term mortality after non-ST segment elevation myocardial infarction . In dialysis patients, PLR was positively correlated with IL-6 levels, and PLR was superior to neutrophil-to-lymphocyte ratio in terms of inflammation prediction . Therefore, we used PLR as another inflammatory marker in the present study. Our analysis showed the trend of association between PLR and PTH levels. However, the association was no longer statistically significant in the fully adjusted model (i.e., Model 4). The nonsignificant results may be due to direct effects of PTH on platelets and lymphocytes [45–47]. Alternatively, there may be a nonlinear relationship between PTH levels and blood cell derangements induced by the acute phase reaction.
The combination of serum CRP and albumin has been used to derive the mGPS . Hepatic albumin biosynthesis is downregulated by inflammatory stimuli as part of a negative acute phase reaction. Previous studies suggest that mGPS is superior to the original GPS and has greater consistency and is of more use . The observation was based on the results that a low albumin concentration alone was uncommon and was not significantly associated with cancer-specific survival in many cancers. In this study, increased mGPS was analyzed as a dichotomous variable (>0 versus 0) and taken as another marker of inflammation. Although mGPS is usually used to predict outcome in cancer patients, our results indicate that serum PTH levels were strongly associated with increased mGPS. Overall, our consistent findings buttress the validity of the association found between PTH levels and inflammation.
A major strength of our study is that it was based on nationwide, population-based sampling survey data. However, some limitations should be acknowledged. First, our analyses point to an association between PTH levels and inflammation but do not indicate a direction for the causal relationship because of the observational and cross-sectional study design. Second, other potential confounding variables could influence the findings. For example, in the NHANES database, there were no data on the FGF-23/Klotho axis. Higher FGF-23 levels were reported to be independently associated with higher levels of inflammatory markers in patients with chronic kidney disease . Moreover, both systemic and local inflammation may decrease kidney Klotho expression, and Klotho downregulates inflammation . Our finding may be a reflection of the association between low Klotho levels and inflammation. Third, no repeat measurements of PTH or other inflammatory markers were available, and therefore, the data set was limited to measures collected at a single time point for each participant. Finally, caution should be taken in interpreting results in relation to populations other than US adults.
In conclusion, using the NHANES data from the US adult population, we found an association between higher serum concentrations of PTH and several inflammatory markers (CRP, RDW, PLR, and mGPS). Studies have shown that vitamin D supplementation has a suppressive effect on PTH levels , but it is still unclear whether the suppression has any impact on clinical outcomes. Given that inflammation is involved in the pathophysiology of cardiovascular disease, cancer, aging, and other conditions, our results provide clues on which to base further investigations of the mechanistic aspects of the PTH-inflammatory link.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
This study was supported by Grants from National Science Council of Taiwan (NSC-100-2314-B-195-001-MY3 and NSC-102-2314-B-195-018) and Mackay Memorial Hospital (MMH-10303 and MMH-10316).
J. Silver, M. Rodriguez, and E. Slatopolsky, “FGF23 and PTH—double agents at the heart of CKD,” Nephrology Dialysis Transplantation, vol. 27, pp. 1715–1720, 2012.View at: Google Scholar
S. B. Kritchevsky, J. A. Tooze, R. H. Neiberg et al., “Health ABC Study. 25-Hydroxyvitamin D, parathyroid hormone, and mortality in black and white older adults: the health ABC study,” Journal of Clinical Endocrinology & Metabolism, vol. 97, pp. 4156–4165, 2012.View at: Google Scholar
A. J. van Ballegooijen, I. Reinders, M. Visser et al., “Serum parathyroid hormone in relation to all-cause and cardiovascular mortality: the Hoorn study,” Journal of Clinical Endocrinology & Metabolism, vol. 98, pp. E638–E645, 2013.View at: Google Scholar
S. Gandhi and R. B. Myers, “Can parathyroid hormone be used as a biomarker for heart failure?” Heart Failure Reviews, vol. 18, pp. 465–473, 2013.View at: Google Scholar
C. W. G. M. Lowik, G. van der Pluijm, H. Bloys et al., “Parathyroid hormone (PTH) and PTH-like protein (PLP) stimulate interleukin-6 production by osteogenic cells: a possible role of interleukin-6 in osteoclastogenesis,” Biochemical and Biophysical Research Communications, vol. 162, no. 3, pp. 1546–1552, 1989.View at: Google Scholar
M. A. Mitnick, A. Grey, U. Masiukiewicz et al., “Parathyroid hormone induces hepatic production of bioactive interleukin-6 and its soluble receptor,” American Journal of Physiology: Endocrinology and Metabolism, vol. 280, no. 3, pp. E405–E412, 2001.View at: Google Scholar
A. Grey, M.-A. Mitnick, S. Shapses, A. Ellison, C. Gundberg, and K. Insogna, “Circulating levels of interleukin-6 and tumor necrosis factor-α are elevated in primary hyperparathyroidism and correlate with markers of bone resorption—a clinical research center study,” Journal of Clinical Endocrinology and Metabolism, vol. 81, no. 10, pp. 3450–3454, 1996.View at: Publisher Site | Google Scholar
C.-Y. Guo, P. A. Holland, B. F. Jackson et al., “Immediate changes in biochemical markers of bone turnover and circulating interleukin-6 after parathyroidectomy for primary hyperparathyroidism,” European Journal of Endocrinology, vol. 142, no. 5, pp. 451–459, 2000.View at: Google Scholar
C. G. Øgard, M. D. Engelmann, C. Kistorp, S. L. Nielsen, and H. Vestergaard, “Increased plasma N-terminal pro-B-type natriuretic peptide and markers of inflammation related to atherosclerosis in patients with primary hyperparathyroidism,” Clinical Endocrinology, vol. 63, no. 5, pp. 493–498, 2005.View at: Publisher Site | Google Scholar
E. Chertok-Shacham, A. Ishay, I. Lavi, and R. Luboshitzky, “Biomarkers of hypercoagulability and inflammation in primary hyperparathyroidism,” Medical Science Monitor, vol. 14, no. 12, pp. CR628–CR632, 2008.View at: Google Scholar
A. Halabe and B. Shohat, “Effect of parathyroid adenoma excision on interleukin-6 (IL-6) and IL-2 receptor levels,” Metabolism: Clinical and Experimental, vol. 49, no. 2, pp. 192–194, 2000.View at: Google Scholar
E. G. Almqvist, A.-G. Bondeson, L. Bondeson, and J. Svensson, “Increased markers of inflammation and endothelial dysfunction in patients with mild primary hyperparathyroidism,” Scandinavian Journal of Clinical and Laboratory Investigation, vol. 71, no. 2, pp. 139–144, 2011.View at: Publisher Site | Google Scholar
G. Zhao, E. S. Ford, C. Li, P. M. Kris-Etherton, T. D. Etherton, and L. S. Balluz, “Independent associations of serum concentrations of 25-hydroxyvitamin D and parathyroid hormone with blood pressure among US adults,” Journal of Hypertension, vol. 28, no. 9, pp. 1821–1828, 2010.View at: Publisher Site | Google Scholar
C. L. Liu, J. J. Lee, T. P. Liu, Y. C. Chang, Y. C. Hsu, and S. P. Cheng, “Blood neutrophil-to-lymphocyte ratio correlates with tumor size in patients with differentiated thyroid cancer,” Journal of Surgical Oncology, vol. 107, pp. 493–497, 2013.View at: Google Scholar
“Centers for Disease Control and Prevention,” National Health and Nutrition Examination Survey 2003-2004, http://wwwn.cdc.gov/nchs/nhanes/search/nhanes03_04.aspx.View at: Google Scholar
Centers for Disease Control and Prevention, National Health and Nutrition Examination Survey 2005-2006, http://wwwn.cdc.gov/nchs/nhanes/search/nhanes05_06.aspx.
A. S. Levey, L. A. Stevens, C. H. Schmid et al., “CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate,” Annals of Internal Medicine, vol. 150, pp. 604–612, 2009.View at: Google Scholar
D. C. McMillan, “The systemic inflammation-based Glasgow Prognostic Score: a decade of experience in patients with cancer,” Cancer Treatment Reviews, vol. 39, pp. 534–540, 2013.View at: Google Scholar
S. P. Cheng, J. J. Lee, T. P. Liu et al., “Parathyroidectomy improves symptomatology and quality of life in patients with secondary hyperparathyroidism,” Surgery, vol. 155, pp. 320–328, 2014.View at: Google Scholar
R. Alemzadeh and J. Kichler, “Parathyroid hormone is associated with biomarkers of insulin resistance and inflammation, independent of vitamin D status, in obese adolescents,” Metabolic Syndrome and Related Disorders, vol. 10, pp. 422–429, 2012.View at: Google Scholar
O. Yousuf, B. D. Mohanty, S. S. Martin et al., “High-sensitivity C-reactive protein and cardiovascular disease: a resolute belief or an elusive link?” Journal of the American College of Cardiology, vol. 62, pp. 397–408, 2013.View at: Google Scholar
G. Lippi, G. Targher, M. Montagnana, G. L. Salvagno, G. Zoppini, and G. C. Guidi, “Relation between red blood cell distribution width and inflammatory biomarkers in a large cohort of unselected outpatients,” Archives of Pathology & Laboratory Medicine, vol. 133, pp. 628–632, 2009.View at: Google Scholar
Z. Förhécz, T. Gombos, G. Borgulya, Z. Pozsonyi, Z. Prohászka, and L. Jánoskuti, “Red cell distribution width in heart failure: prediction of clinical events and relationship with markers of ineffective erythropoiesis, inflammation, renal function, and nutritional state,” American Heart Journal, vol. 158, no. 4, pp. 659–666, 2009.View at: Publisher Site | Google Scholar
R. A. Smith, L. Bosonnet, P. Ghaneh et al., “The platelet-lymphocyte ratio improves the predictive value of serum CA19-9 levels in determining patient selection for staging laparoscopy in suspected periampullary cancer,” Surgery, vol. 143, no. 5, pp. 658–666, 2008.View at: Publisher Site | Google Scholar
K. Turkmen, F. M. Erdur, F. Ozcicek et al., “Platelet-to-lymphocyte ratio better predicts inflammation than neutrophil-to-lymphocyte ratio in end-stage renal disease patients,” Hemodialysis International, vol. 17, pp. 391–396, 2013.View at: Google Scholar
A. Benigni, M. Livio, and P. Dodesini, “Inhibition of human platelet aggregation by parathyroid hormone. Is cyclic AMP implicated?” American Journal of Nephrology, vol. 5, no. 4, pp. 243–247, 1985.View at: Google Scholar
J. Munoz Mendoza, T. Isakova, A. C. Ricardo et al., “Chronic renal insufficiency cohort. Fibroblast growth factor 23 and Inflammation in CKD,” Clinical Journal of the American Society of Nephrology, vol. 7, pp. 1155–1162, 2012.View at: Google Scholar
M. C. Izquierdo, M. V. Perez-Gomez, M. D. Sanchez-Nino et al., “Klotho, phosphate and inflammation/ageing in chronic kidney disease,” Nephrology Dialysis Transplantation, vol. 27, supplement 4, pp. iv6–i10, 2012.View at: Google Scholar