Associations of Serum 25(OH)D, PTH, and β-CTX Levels with All-Cause Mortality in Chinese Community-Dwelling Centenarians

This longitudinal cohort study explored the associations of 25-hydroxyvitamin D [25(OH)D], parathyroid hormone (PTH), and β-C-terminal telopeptide of type 1 collagen (β-CTX) levels with all-cause mortality in centenarians. The study included 952 centenarians (81.4% female). During a median follow-up of 32 months, 752 (78.9%) centenarians died. The estimated 1-year, 3-year, and 5-year survival rates were 80.0%, 45.7%, and 23.6%, respectively. The association of mortality with 25(OH)D was linear, whereas the associations with PTH and β-CTX were J-shaped, with a lower risk below the median levels. Compared with 25(OH)D of ≥30 ng/mL, 25(OH)D < 30 ng/mL was associated with increased mortality (HR 1.52, 95% CI 1.24–1.86, p < 0.001). Compared with PTH of ≤65 pg/mL, PTH > 65 pg/mL was associated with increased mortality (HR 1.30, 95% CI 1.08–1.56, p = 0.005). Compared with β-CTX of <0.55 ng/mL, β-CTX ≥ 0.55 ng/mL was associated with increased mortality (HR 1.30, 95% CI 1.10–1.54, p = 0.002). A higher β-CTX level (even in the clinical reference range of 0.55–1.01 ng/mL) was associated with increased mortality (HR 1.23, 95% CI 1.04–1.47, p = 0.018). Centenarians with 25(OH)D < 30 ng/mL, PTH > 65 pg/mL, and β-CTX ≥ 0.55 ng/mL had a 2.77-fold (95% CI 1.99–3.85, p < 0.001) increased risk of mortality when compared with those with 25(OH)D of >30 ng/mL, PTH < 65 pg/mL, and β-CTX < 0.55 ng/mL. Lower serum 25(OH)D and higher PTH and β-CTX were independently correlated with increased all-cause mortality in Chinese community-dwelling centenarians.


Introduction
One hundred years is considered to be close to the maximum human lifespan, and centenarians provide an ideal model for investigating traits of longevity among individuals in whom age-related diseases have been largely avoided or delayed [1]. Remodeling of bone, including resorption and formation, is a continuous process throughout life and continues even in extreme senescence. Bone remodeling is affected by 25-hydroxyvitamin D [25(OH)D] and parathyroid hormone (PTH) levels and is reflected by levels of bone

Baseline Characteristics and Follow-Up
In all, 46 of the 1002 centenarians in the CHCCS had missing values at baseline and 4 were lost during follow-up. The baseline characteristics of the study participants are summarized in Table 1. Of the 952 centenarians who were finally selected, 81.4% were female. The median age was 102 years (IQR, 100, 104). Approximately three-quarters (75.5%) had hypertension, and a minority had diabetes mellitus (9.9%) or a history of fracture (8.6%). The respective percentages of centenarians with abnormal serum calcium, phosphorus, ALP, PTH, P1NP, and osteocalcin levels were 27.1%, 14.8%, 7.2%, 24.3%, 30.5%, and 18.9%. In contrast, 80.8% of participants had an abnormal serum 25(OH)D level, and 75.1% had an abnormal β-CTX level (Figure 1).
During a median follow-up of 32 months (IQR, 15, 55), nearly four-fifths of participants died. The estimated 1-year, 3-year, and 5-year survival rates were 80.0%, 45.7%, and 23.6%, respectively. There was no significant difference in the serum P1NP between centenarians who remained alive and those who died during follow-up. There were statistically significant differences in serum calcium, phosphorus, ALP, and osteocalcin levels between centenarians who remained alive and those who died; however, the differences in serum calcium, phosphorus, ALP, and osteocalcin were relatively small (0.02 mmol/L or a 0.9% decrease, 0.04 mmol/L or a 3.7% increase, 7 U/L or a 9.2% increase, and 2.1 ng/mL or a 7.6% increase, respectively). Compared with centenarians who remained alive, those who died had a significantly lower median 25(OH)D level (21.15   During a median follow-up of 32 months (IQR, 15, 55), nearly four-fifths of participants died. The estimated 1-year, 3-year, and 5-year survival rates were 80.0%, 45.7%, and

Associations of 25(OH)D, PTH, and β-CTX with All-Cause Mortality
The Kaplan-Meier curves, Cox proportional hazards models, and RCS analyses confirmed that serum calcium, phosphorus, ALP, P1NP, and osteocalcin levels were not significantly associated with increased all-cause mortality. However, the Kaplan-Meier survival curves revealed a significant association of a lower 25(OH)D level with increased all-cause  (Figure 2A-D). Compared with the highest quartile, the multivariable-adjusted HRs for mortality in the lowest quartile, second quartile, and third quartile were 1.72 (95% CI 1.38-2.14), 1.53 (95% CI 1.24-1.91), and 1.46 (95% CI 1.18-1.81), respectively ( Table 2). Mortality was significantly higher in centenarians with a 25(OH)D level below the median (<21.5 ng/mL) than in those with a 25(OH)D level above the median (≥21.5 ng/mL) (1.32, 95% CI 1.13-1.53). Compared with a 25(OH)D level of ≥30 ng/mL, the multivariable-adjusted HRs for mortality in centenarians with 25(OH)D levels of <10 ng/mL and 10-30 ng/mL were 1.86 (95% CI 1.34-2.58) and 1.52 (95% CI 1.25-1.87), respectively. Compared with a 25(OH)D level of ≥30 ng/mL, the multivariable-adjusted HR for mortality in centenarians with 25(OH)D levels of <30 ng/mL (including <10 ng/mL and 10-30 ng/mL) was 1.52 (95% CI 1.24-1.86). The associations of 25(OH)D as a continuous variable with all-cause mortality were further investigated in univariate and multivariate Cox proportional hazard models using RCS analysis ( Figure 3). The relationship between 25(OH)D and mortality was linear in both unadjusted and multivariable-adjusted models (p nonlinear = 0.518 and p nonlinear = 0.526, respectively; Figure 3A,B).      The relationship between PTH and mortality was approximately J-shaped in both unadjusted (C) and multivariable-adjusted (D) models. For centenarians with a PTH level in the range of 15.56-43.89 pg/mL, the risk of all-cause mortality reached a nadir, with a significant positive association above but a non-significant negative association below this range. The relationship between the β-CTX level and mortality was nearly linear in the unadjusted model (E) but J-shaped in the multivariable-adjusted model (F). For centenarians with a β-CTX level of 0.227-0.406 ng/mL, the risk of allcause mortality reached a nadir, with a significant positive association above but a non-significant negative association below this range. It should be noted that centenarians with β-CTX levels in the clinical reference range (0.55-1.01 ng/mL) did have a significantly increased mortality. The data are adjusted for demographic factors (age, sex), body mass index, lifestyle factors (cigarette smoking, alcohol consumption, outdoor activities), comorbidities (incidence of hypertension, diabetes mellitus, cardiovascular disease), past medical history (fractures, surgery), and other possible confounding factors (serum creatinine, hemoglobin, C-reactive protein, homocysteine, and low-density lipoprotein cholesterol levels). 25(OH)D, 25-hydroxyvitamin D; β-CTX, β-C-terminal telopeptide of type 1 collagen; HR, hazard ratio; PTH, parathyroid hormone. Kaplan-Meier survival curves revealed significant associations of higher PTH levels with increased all-cause mortality ( Figure 2E-H). The highest quartile of PTH was associated with a 43% increase in risk of mortality compared with the second quartile (HR 1.43, 95% CI 1.16-1.76); however, the lowest and third quartiles were not associated with a significantly increased risk of mortality compared with the second quartile ( Table 2). The mortality rate was significantly higher in centenarians with a PTH level above the median value of >43.89 pg/mL than those with a PTH level below this value (HR 1.22, 95% CI 1.05-1.42). Compared with a PTH level in the range of 15-65 pg/mL, the multivariable-adjusted HRs for mortality in centenarians with a PTH level of <15 pg/mL and >65 pg/mL were 1.09 (95% CI 0.75-1.59) and 1.30 (95% CI 1.08-1.56), respectively. Compared with a PTH level of ≤65 pg/mL, the multivariable-adjusted HR for mortality when PTH was >65 pg/mL was 1.30 (95% CI 1.08-1.56). The relationship between PTH and mortality was approximately J-shaped in both unadjusted and multivariable-adjusted models (p nonlinear = 0.001 and p nonlinear = 0.010, respectively; Figure 3C,D). For centenarians with a PTH level in the range of 15.56-43.89 pg/mL, the risk of all-cause mortality reached a nadir, with a significant positive association above but a non-significant negative association below this range.
Kaplan-Meier survival curves revealed that a higher β-CTX level was significantly associated with increased all-cause mortality ( Figure 2I-L). The highest quartile of β-CTX was associated with a 60% increase in the risk of mortality in comparison with the second quartile (HR 1.60, 95% CI 1.29-1.97). However, the lowest and third quartiles were not associated with a significantly increased risk of mortality in comparison with the second quartile ( Table 2). Centenarians with a β-CTX level higher than the median value of ≥0.406 ng/mL had a significantly higher mortality risk than those with a β-CTX level below this value (HR 1.21, 95% CI 1.04-1.40). Compared with a β-CTX level of <0.55 ng/mL, the multivariableadjusted HRs for mortality at β-CTX levels of 0.55-1.01 ng/mL and >1.01 ng/mL were 1.23 (95% CI 1.04-1.47) and 2.03 (95% CI 1.41-2.92), respectively. Compared with a β-CTX level of <0.55 ng/mL, the multivariable-adjusted HR for mortality at β-CTX levels of ≥0.55 ng/mL (including 0.55-1.01 ng/mL and >1.01 ng/mL) was 1.30 (95% CI 1.10-1.54). The relationship between the β-CTX level and mortality was nearly linear in the unadjusted model but J-shaped in the multivariable-adjusted model (p nonlinear = 0.169 and p nonlinear = 0.043, respectively; Figure 3E,F). For centenarians with a β-CTX level of 0.227-0.406 ng/mL, the risk of all-cause mortality reached a nadir, with a significant positive association above but a non-significant negative association below this range. It should be noted that centenarians with β-CTX levels in the clinical reference range (0.55-1.01 ng/mL) did have a significantly increased mortality ( Figure 3F).

Discussion
The CHCCS included a large cohort of 1002 community-dwelling centenarians with a median follow-up of more than 2.5 years for all participants and of more than 5 years for those who remained alive. The loss to follow-up rate was 0.4%, and nearly 80% of centenarians reached the endpoint of death. Therefore, the CHCCS provides a good model for investigating aging and longevity in community-dwelling centenarians. Our previous work suggested that the serum 25(OH)D level might be associated with activities of daily living, functional dependence, and symptoms of depression at baseline in participants in the CHCCS [27][28][29]. In the present study, we further demonstrated that not only the 25(OH)D level but also the PTH and β-CTX levels were associated with all-cause mortality. Serum P1NP and osteocalcin levels were not significantly associated with mortality.
The association of 25(OH)D with all-cause mortality has been extensively examined in community-based individuals aged > 80 years. Several studies have found a significant association (HR 1.56-2.02) [10][11][12][13], while others have reported a negative association [14][15][16][17]. However, to date, the association between 25(OH)D and all-cause mortality has not been investigated in centenarians. Therefore, the present study provides the first evidence of a linear association between the serum 25(OH)D level and all-cause mortality in centenarians.
The association between PTH and all-cause mortality has rarely been investigated in the general population, especially in the cohort aged >80 years. Furthermore, the available evidence is conflicting. Sambrook et al. showed that PTH was associated with increased mortality in 842 older adults (mean age > 82 years) during a mean follow-up of 31 months [30]. The Helsinki Ageing Study reported a significant impact of elevated serum PTH on all-cause mortality in three age cohorts (75, 80, and 85 years) during a follow-up of 17 years [31]. However, the Malmo Osteoporosis Prospective Risk Assessment cohort, which included 1044 community-dwelling women who were aged >85 years and followed for 10 years, found that PTH was not independently associated with all-cause mortality [32]. However, until now, the association between PTH and all-cause mortality has not been investigated in centenarians. The present study provides the first evidence indicating that the risk of all-cause mortality is significantly increased in centenarians with an elevated PTH but unaffected in those with a PTH level below the normal range. The relationship between PTH and mortality was J-shaped. PTH could affect the calcium level directly or indirectly and has been correlated with vascular calcification and increased cardiovascular risk [33]. Interleukin-1 has been shown to participate in parathyroid cell function and secretion of PTH, and an interleukin-1 receptor antagonist to be able to regulate hypercalcemia [34,35].
Recent studies indicate that interleukin-1 receptor antagonism is a promising therapeutic strategy for cardiovascular disease [36].
The International Osteoporosis Foundation and the International Federation of Clinical Chemistry and Laboratory Medicine recommend P1NP as a marker of bone formation and β-CTX as a marker of bone resorption [2,3]. However, there is limited information on associations of P1NP and β-CTX with mortality. Bager et al. found a U-shaped association of β-CTX with all-cause mortality [37]. A high β-CTX level was also found to be an independent predictor of all-cause mortality in 1112 frail elderly subjects (mean age 86 years) [38]. However, the associations of all-cause mortality with P1NP and β-CTX have not previously been investigated in centenarians. The present study provides the first evidence that a higher β-CTX level (even in the clinical reference range of 0.55-1.01 ng/mL) is associated with a significantly increased risk of all-cause mortality in centenarians. The relationship between β-CTX and mortality was J-shaped. P1NP levels were not associated with all-cause mortality.
Several studies have demonstrated the effect of an interaction between 25(OH)D and PTH on all-cause mortality in the general population but not in centenarians. The Cardiovascular Health Study followed 2312 elderly individuals (aged ≥65 years) for 14 years and showed that all-cause mortality was increased by a further 16% in participants with a 25(OH)D < 15 ng/mL who had a PTH level of ≥65 pg/mL in comparison with their counterparts who had a PTH level of <65 pg/mL (HR 1.43 vs. 1.27) [39]. The CopD Study (a retrospective cohort study conducted in Copenhagen, Denmark) enrolled 34,996 adults (mean age 51 years) who were followed for a median of 3 years. All-cause mortality was significantly higher in participants with 25(OH)D < 25 nmol/L and PTH ≥ 7.6 nmol/L than in those with 25(OH)D < 25 nmol/L and PTH < 7.6 nmol/L (10.3% vs. 4.7%, p < 0.0001) [40]. The Health Aging and Body Composition study followed 2638 community-dwelling elderly individuals (aged 71-80 years) for 8.5 years and found that all-cause mortality was a further 60% higher in participants with 25(OH)D < 20 ng/mL if PTH was ≥70 pg/mL than if it was <70 pg/mL (HR 1.96 vs. 1.36) [41]. The Longitudinal Aging Study Amsterdam, which included 1317 elderly individuals (aged 65-85 years) who were followed for 18 years, reported that the relationship of 25(OH)D with all-cause mortality was partly mediated by PTH [42]. However, whether PTH and β-CTX could have an additional effect on the association between 25(OH)D and all-cause mortality in centenarians has not been studied. Our current study provides the first data suggesting that the risk of all-cause mortality is significantly higher in centenarians with at least two risk factors, namely, low 25(OH)D and high PTH and β-CTX, than in those with high 25(OH)D, low PTH and β-CTX.
This study had some limitations. First, we did not analyze the relationship between cause-specific mortality, such as cardiovascular-associated or cancer-associated mortality, because the causes of death were not documented. Second, 25(OH)D, PTH, and β-CTX levels were only measured at baseline. Therefore, we did not analyze any associations between changes in these parameters during follow-up and all-cause mortality.

Conclusions
To the best of our knowledge, the CHCCS is the largest study with the longest followup demonstrating the associations of 25(OH)D, PTH, and β-CTX levels with all-cause mortality in Chinese community-dwelling centenarians. Our findings highlight the importance of PTH and β-CTX levels in the relationship between 25(OH)D insufficiency and all-cause mortality in centenarians.