作为一个图书馆，NLM 提供对科学文献的访问。NLM 数据库的收录并不意味着 NLM 或美国国立卫生研究院认可或同意其内容。
Learn more: PMC Disclaimer | PMC Copyright Notice
了解更多信息：PMC 免责声明 | PMC 版权声明
营养素。2022 年 11 月；14（21）：4629.
pmcid: pmc9655488if：5.9 Q1
Vitamin D Status and Risk of All-Cause and Cause-Specific Mortality in Osteoarthritis Patients: Results from NHANES III and NHANES 2001–2018
骨关节炎患者的维生素 D 状态与全因和特定原因死亡率风险：NHANES III 和 NHANES 2001-2018 的结果
学术编辑 Connie Weaver
作者信息 文章注释 版权和许可信息 PMC 免责声明
Abstract 摘要
Objectives: The role of Vitamin D (VD) in calcium balance and bone health makes VD a vital factor in osteoarthritis (OA). Studies that have evaluated the effect of VD on OA patients have mainly been performed on a short-term basis. In this analysis, we aimed to evaluate whether VD was associated with mortality, a long-term outcome, in OA patients. Methods: Participants with self-reported OA from NHANES III and NHANES 2001–2018 were included. Associations of 25(OH)D concentrations with mortality risk were assessed continuously using restricted cubic splines and by categories (i.e., <25.0, 25.0–49.9, 50.0–74.9, and ≥75.0 nmol/L) using the Cox regression model. Sensitivity and stratified analyses were performed to evaluate the robustness of the results. Results: A total of 4570 patients were included, of which 1388 died by 31 December 2019. An L-shaped association was observed between 25(OH)D concentrations and all-cause mortality, whereas an inverse association was found for cardiovascular disease (CVD) mortality. The adjusted hazard ratios (95% confidence intervals) across four categories were 1.00 (reference), 0.49 (0.31, 0.75), 0.45 (0.29, 0.68), and 0.43 (0.27, 0.69) for all-cause mortality and 1.00 (reference), 0.28 (0.14, 0.59), 0.25 (0.12, 0.51), and 0.24 (0.11, 0.49) for CVD-specific mortality; no significant associations were found for cancer-specific mortality. Similar results were observed when stratified and sensitivity analyses were performed. Conclusions: Compared with patients with insufficient or deficient serum 25(OH)D, those with sufficient 25(OH)D concentrations had a lower risk of all-cause and CVD mortality, supporting a beneficial role of VD on a long-term basis.
目的：维生素 D（VD）在钙平衡和骨骼健康方面的作用使其成为骨关节炎（OA）的一个重要因素。评估维生素 D 对 OA 患者影响的研究主要是短期研究。在这项分析中，我们旨在评估 VD 是否与 OA 患者的死亡率（一种长期结果）相关。研究方法纳入 NHANES III 和 NHANES 2001-2018 中自我报告有 OA 的参与者。使用限制性三次样条连续评估25（OH）D浓度与死亡风险的关系，并使用Cox回归模型按类别（即<25.0、25.0-49.9、50.0-74.9和≥75.0 nmol/L）评估25（OH）D浓度与死亡风险的关系。为评估结果的稳健性，进行了敏感性分析和分层分析。结果共纳入 4570 名患者，其中 1388 人在 2019 年 12 月 31 日前死亡。观察发现，25（OH）D浓度与全因死亡率呈L型关系，而与心血管疾病（CVD）死亡率呈反向关系。四个类别的调整后危险比（95% 置信区间）分别为：全因死亡率 1.00（参考值）、0.49（0.31，0.75）、0.45（0.29，0.68）和 0.43（0.27，0.69）；心血管疾病死亡率 1.在心血管疾病特异性死亡率方面，分别为 1.00（参考值）、0.28（0.14，0.59）、0.25（0.12，0.51）和 0.24（0.11，0.49）；在癌症特异性死亡率方面未发现显著关联。在进行分层分析和敏感性分析时也观察到了类似的结果。结论与血清25(OH)D不足或缺乏的患者相比，25(OH)D浓度充足的患者全因死亡和心血管疾病死亡的风险较低，支持VD的长期有益作用。
关键词维生素 D 骨关节炎 死亡率 NHANES
1. Introduction 1.导言
Osteoarthritis (OA) is a common musculoskeletal disorder that affected approximately 527.8 million people worldwide in 2019 [1]. In the US, more than 32.5 million adults have OA, and it has become a dominant cause of pain and disability [2]. In addition, OA can impact both the physical and mental health of individuals, and such a prevalent disease has manifested in tremendous societal and personal expenses [2,3]. The pathogenesis and progression of OA are multifactorial, including the joint structures, non-modifiable factors (age, gender, etc.), and modifiable factors such as obesity and Vitamin D status [4].
骨关节炎（OA）是一种常见的肌肉骨骼疾病，2019 年全球约有 5.278 亿人受到影响[ 1]。在美国，超过 3250 万成年人患有 OA，它已成为导致疼痛和残疾的主要原因[ 2]。此外，OA 还会影响个人的身体和心理健康，这种流行性疾病造成了巨大的社会和个人支出[2, 3]。OA 的发病机制和进展是多因素的，包括关节结构、不可改变因素（年龄、性别等）以及肥胖和维生素 D 状态等可改变因素[4]。
As a steroidal hormone, Vitamin D produces a variety of biological effects on many target tissues [5]. The well-established roles of Vitamin D in the calcium metabolism of bone make Vitamin D a highly influential factor that can be used for the prevention and management of OA [5]. Although a conclusive result has not been reached, several randomized control trials indicated that Vitamin D supplementation was possible to alleviate pain and improve joint function in knee OA patients, especially in those with a low serum Vitamin D level (<50 nmol/L) [6,7,8]. Such benefits might also be relevant to increasing outdoor activity and mental health in OA patients, thus contributing to better survival in terms of quality and quantity [9]. Additionally, Vitamin D has been shown to have benefits in many non-skeletal diseases, such as cardiovascular disease (CVD), diabetes, and cancers, via its effects of reducing inflammation, anti-proliferative activity, anti-oxidative activity, etc. [10]. In a general population, a study including 365,530 subjects from UK Biobank demonstrated that a higher level of serum 25(OH)D (a stable indicator of the Vitamin D status) was nonlinearly related with a decrease in the risk of all-cause, CVD-specific, and cancer-specific mortality [11]. A similar association was reported for all-cause and CVD-specific mortality but not for cancer-specific mortality when the population was restricted to diabetic patients [12].
作为一种类固醇激素，维生素 D 可对许多靶组织产生多种生物效应[5]。维生素 D 在骨钙代谢中的作用已得到证实，因此维生素 D 是一种极具影响力的因素，可用于预防和治疗 OA [ 5]。尽管尚未得出定论，但几项随机对照试验表明，补充维生素 D 有可能减轻膝关节 OA 患者的疼痛并改善其关节功能，尤其是那些血清维生素 D 水平较低（<50 nmol/L）的患者[6, 7, 8]。这些益处也可能与增加 OA 患者的户外活动和心理健康有关，从而有助于提高生存质量和数量[9]。此外，维生素 D 还通过其减少炎症、抗增殖活性、抗氧化活性等作用，对许多非骨骼疾病，如心血管疾病（CVD）、糖尿病和癌症有益处[10]。[ 10].在普通人群中，一项包括英国生物库 365,530 名受试者的研究表明，血清 25（OH）D（维生素 D 状态的稳定指标）水平越高，全因死亡率、心血管疾病特异性死亡率和癌症特异性死亡率的风险就越低[11]。当研究对象仅限于糖尿病患者时，全因死亡率和心血管疾病特异性死亡率也有类似的关系，但癌症特异性死亡率则没有[12]。
Although the benefit of Vitamin D in mortality reduction has been reported in the general population and diabetic patients, there have only been a few studies that have focused on such a relationship for OA patients so far. Therefore, in this study, we aimed to evaluate whether serum 25(OH)D concentrations were associated with all-cause and cause-specific mortality in OA patients from NHANES III and the newly released NHANES 2001–2018 database.
虽然维生素 D 有助于降低普通人群和糖尿病患者的死亡率，但迄今为止只有少数研究关注 OA 患者的这种关系。因此，在本研究中，我们旨在评估血清 25(OH)D 浓度是否与 NHANES III 和新发布的 NHANES 2001-2018 数据库中 OA 患者的全因死亡率和特定原因死亡率相关。
2. Materials and Methods 2.材料和方法
2.1. Study Population 2.1.研究对象
National Health and Nutrition Examination Survey (NHANES), sponsored by National Center for Health Statistics (NCHS), is a large-scale and periodic program that harvests nationally representative health-related data of non-institutionalized citizens in the United States. Details of survey design and data files are publicly available on https://www.cdc.gov/nchs/nhanes/ (accessed on 31 August 2022). The ethics protocol was approved by the Research Ethics Review Board of NCHS, and informed consent was signed by all recruited participants.
全国健康与营养调查（NHANES）由美国国家卫生统计中心（NCHS）发起，是一项大规模的定期计划，收集美国非住院公民的全国代表性健康相关数据。有关调查设计和数据文件的详细信息可在 https://www.cdc.gov/nchs/nhanes/（访问日期：2022 年 8 月 31 日）上查阅。伦理协议已获得美国国家卫生与健康委员会研究伦理审查委员会的批准，所有受访者均签署了知情同意书。
Here, the data from NHANES III (from 1988 to 1994) and NHANES 2001–2018 were used in our analysis, and only respondents (aged > 18 years) with self-reported OA were included in our analysis. OA patients were identified by answering “Yes” to the question “Has a doctor ever told you that you had arthritis?” and then selecting “Osteoarthritis” to the question “Which type of arthritis was it?”. As a common chronic disease, self-reported information was considered to be reliable in most instances; an 85% concordance between self-reported OA and clinically well-defined OA was exhibited in a previous study [13]. At first, 5735 OA patients were identified. After excluding those who did not have 25(OH)D concentrations measurements (647), mortality (31), or other relevant covariates (487), 4570 individuals were eventually included in our study (Figure 1).
在此，我们使用了 NHANES III（1988-1994 年）和 NHANES 2001-2018 年的数据进行分析，并且只将自我报告有 OA 的受访者（年龄大于 18 岁）纳入分析范围。对于 "医生是否曾告诉您患有关节炎？"这一问题，回答 "是"，然后在 "是哪种类型的关节炎？"这一问题中选择 "骨关节炎"，即可识别出 OA 患者。作为一种常见的慢性疾病，自我报告的信息在大多数情况下被认为是可靠的；在之前的一项研究中，自我报告的 OA 与临床明确定义的 OA 之间的一致性达到了 85%[13]。研究首先确定了 5735 名 OA 患者。在排除那些没有测量 25（OH）D 浓度（647 人）、死亡率（31 人）或其他相关协变量（487 人）的患者后，最终有 4570 人被纳入我们的研究（图 1）。
2.2. Measurement of Serum 25(OH)D
2.2.测量血清 25(OH)D
Note that there exist methodological discrepancies in measuring serum 25(OH)D between survey cycles in NHANES. More specifically, the DiaSorin RIA kit (Stillwater, MN, USA) was applied in NHANES III and NHANES 2001–2006, while a standardized liquid chromatography–tandem mass spectrometry (LC-MS/MS) method became an alternative starting from the 2007–08 cycle. We performed standardization using the regression method, referring to the Analysis Guidance Manual provided by NHANES, to unify all serum 25(OH)D data to be LC-MS/MS equivalent for subsequent analyses. The specific conversion method can be accessed at https://wwwn.cdc.gov/nchs/nhanes/vitamind/analyticalnote.aspx (accessed on 31 August 2022). Moreover, for analytes with serum 25(OH)D concentrations below the lower limit of detection (LLOD), the imputed fill value, LLOD/sqrt(2), was placed in the analytic result field.
需要注意的是，NHANES 不同调查周期之间测量血清 25(OH)D 的方法存在差异。更具体地说，在 NHANES III 和 NHANES 2001-2006 中使用的是 DiaSorin RIA 试剂盒（Stillwater, MN, USA），而从 2007-08 周期开始，标准化的液相色谱-串联质谱（LC-MS/MS）方法成为替代方法。我们参照 NHANES 提供的《分析指导手册》，采用回归法进行标准化，将所有血清 25(OH)D 数据统一为等效的 LC-MS/MS，以便进行后续分析。具体转换方法可访问 https://wwwn.cdc.gov/nchs/nhanes/vitamind/analyticalnote.aspx（访问日期：2022 年 8 月 31 日）。此外，对于血清 25（OH）D 浓度低于检测下限（LLOD）的分析物，在分析结果字段中放置了推算的填充值 LLOD/sqrt(2)。
2.3. Mortality Ascertainment
2.3.死亡率的确定
The corresponding mortality information for each participant was acquired by a linkage to the National Death Index (NDI) up to 31 December 2019. The ICD-10 was applied to determine the underlying causes of death [14]. The primary mortality outcomes considered in our study included all-cause mortality, CVD-specific mortality (codes I00–I09, I11, I13, and I20–I51), cancer-specific mortality (codes C00–C97), as well as other-cause mortality. More information regarding linkage methodology and analytic considerations was elucidated elsewhere [15].
通过与截至 2019 年 12 月 31 日的国家死亡指数（NDI）进行链接，获得了每位参与者的相应死亡信息。采用 ICD-10 确定基本死因[14]。我们的研究考虑的主要死亡结果包括全因死亡率、心血管疾病特异性死亡率（代码 I00-I09、I11、I13 和 I20-I51）、癌症特异性死亡率（代码 C00-C97）以及其他原因死亡率。有关链接方法和分析注意事项的更多信息已在其他地方阐明[15]。
2.4. Covariates 2.4.变量
A number of variables were identified as potential confounders in our analysis in the following main aspects: (1) Demographic information: age (years); gender (male, female); race/ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, others); education level (below, equivalent, and above high school); family poverty income ratio (PIR; ≤1, 1< to ≤3, >3). (2) Health status: the body mass index (BMI) was classified as normal weight (<25 kg/m2), overweight (25 to <30 kg/m2), and obesity (≥30 kg/m2); history of diabetes mellitus (fasting glucose level ≥126 mg/dL), hypertension (resting blood pressure ≥140/90 mmHg), CVD, cancer, chronic lung disease, and renal disease (urine albumin-to-creatinine ratio ≥ 30 mg/g) were ascertained using the self-reported doctor’s diagnosis, medication use, and/or laboratory-measured biochemical indicators. (3) Lifestyle factors: physical activity (inactive (has no leisure-time physical activity), insufficient (1~5 times/week for moderate activity with METs ranging from 3~6 or 1~3 times/week for vigorous activity with METs > 6), as well as active (has more leisure-time moderate or vigorous activity than above mentioned)); smoking status (individuals who reported smoking <100 cigarettes over lifetime were coded as never smokers, those who had smoked ≥100 cigarettes but quit smoking as past smokers, and those who still smoked now as current smokers); drinking status (never drinker, abstainer (defined as >12 drinks in any given year in life but without alcohol consumption in the past 12 months), and current drinker (defined as having as least 1 drinks in past 12 months)). In addition, whole-blood biochemical markers, measured in NHANES Laboratory, such as lead and C-reactive protein (CRP), were also considered in the sensitivity analysis.
在我们的分析中，一些变量被确定为潜在的混杂因素，主要包括以下几个方面：（1）人口信息：年龄（岁）；性别（男、女）；种族/民族（非西班牙裔白人、非西班牙裔黑人、墨西哥裔美国人、其他）；教育程度（高中以下、同等学历、高中以上）；家庭贫困收入比（PIR；≤1，1<至≤3，>3）。(2) 健康状况：体重指数（BMI）分为正常体重（<25 kg/m 2 ）、超重（25 至 <30 kg/m 2 ）和肥胖（≥30 kg/m 2 ）；糖尿病史（空腹血糖水平≥126 mg/dL）、高血压史（静息血压≥140/90 mmHg）、心血管疾病史、癌症史、慢性肺部疾病史和肾脏疾病史（尿白蛋白与肌酐比值≥30 mg/g）。(3) 生活方式因素：体力活动（不活跃（没有业余体力活动）、不足（每周进行 1~5 次中等强度活动，METs 在 3~6 之间，或每周进行 1~3 次剧烈活动，METs > 6）以及活跃（业余时间进行比上述更多的中等强度或剧烈活动））；吸烟状况（一生中吸烟少于 100 支者为从不吸烟者，吸烟≥100 支但已戒烟者为过去吸烟者，现在仍吸烟者为现在吸烟者）；饮酒状况（从不饮酒者、戒酒者（定义为一生中任何一年饮酒>12 支，但在过去 12 个月中没有饮酒）和现在饮酒者（定义为在过去 12 个月中至少饮酒 1 次））。 此外，敏感性分析还考虑了 NHANES 实验室测量的全血生化指标，如铅和 C 反应蛋白 (CRP)。
2.5. Statistical Analysis
2.5.统计分析
Appropriate sampling weights were taken into considerations in all analyses for the complex, multistage sample design implemented by NHANES. As recommended by the Endocrine Society Clinical Practice Guidelines, all included OA patients were classified into four categories, that is, serum 25(OH)D < 25.0 nmol/L was regarded as severe deficiency, 25.0–49.9 nmol/L as moderate deficiency, 50.0–74.9 nmol/L as insufficient, and ≥75.0 nmol/L as sufficient [16]. A descriptive analysis was followed to evaluate between-group differences in baseline characteristics, and χ2 tests were used for categorical variables, whereas generalized linear regressions were used for continuous variables.
由于 NHANES 采用了复杂的多阶段抽样设计，因此所有分析都考虑了适当的抽样权重。根据《内分泌学会临床实践指南》的建议，所有纳入的 OA 患者被分为四类，即血清 25（OH）D < 25.0 nmol/L 为严重缺乏，25.0-49.9 nmol/L 为中度缺乏，50.0-74.9 nmol/L 为不足，≥75.0 nmol/L 为充足[16]。采用描述性分析评估基线特征的组间差异，分类变量采用χ 2 检验，连续变量采用广义线性回归。
Before fitting the Cox regression model, tests based on Schoenfeld residuals were initially incorporated for the proportional hazards (PHs) assumption. The results suggested that two predictors (i.e., physical activity and PIR) did not satisfy the PH assumption in several models. In those cases, we employed the stratified Cox model instead. Estimated effects size was presented as hazard ratios (HRs) and corresponding 95% confidence intervals (CIs). Serum 25(OH)D concentration, as the independent variable, was modeled categorically (four categories) and continuously (log-transformed value), respectively. In Model 1, age, gender, and race/ethnicity were adjusted; Model 2 further adjusted (from Model 1) for education level, BMI, PIR (when applicable), physical activity (when applicable), smoking status, and drinking status. In addition, Model 3 further adjusted (from Model 2) for diabetes mellitus, hypertension, CVD, cancer, chronic lung disease, and renal disease. Furthermore, the non-linear association of interest (on a continuous scale) was examined by the restricted cubic spline regression models with full adjustment for confounders.
在拟合 Cox 回归模型之前，首先对基于 Schoenfeld 残差的比例危险（PHs）假设进行了检验。结果表明，在几个模型中，两个预测因子（即体力活动和 PIR）不符合 PH 假设。在这种情况下，我们采用了分层 Cox 模型。估计效应大小以危险比（HRs）和相应的 95% 置信区间（CIs）表示。血清 25(OH)D 浓度作为自变量，分别以分类（四个类别）和连续（对数变换值）的形式进行建模。在模型 1 中，对年龄、性别和种族/民族进行了调整；模型 2 对教育水平、体重指数、PIR（如适用）、体力活动（如适用）、吸烟状况和饮酒状况进行了进一步调整（在模型 1 的基础上）。此外，模型 3 还对（模型 2 的）糖尿病、高血压、心血管疾病、癌症、慢性肺病和肾病进行了进一步调整。此外，在对混杂因素进行充分调整的情况下，通过限制性三次样条回归模型对相关的非线性关联（连续尺度）进行了检验。
Stratified analyses, only for all-cause mortality with a sufficient number of deaths, were also performed to explore underlying effect modification, including age (≤60 years, >60 years), gender, race/ethnicity (non-Hispanic white, others), BMI (<30 kg/m2, ≥30 kg/m2), smoking status (never smoker, current smoker, past smoker), drinking status (never drinker, current drinker, abstainer), physical activity (inactive, insufficient, or active), chronic lung disease (yes, no), and renal disease (yes, no). The potential interactions between serum 25(OH)D concentrations and these stratifying variables were detected by adding their cross-product terms into the model accordingly.
此外，还对有足够死亡人数的全因死亡率进行了分层分析，以探讨潜在的效应修正，包括年龄（≤60 岁，>60 岁）、性别、种族/民族（非西班牙裔白人，其他）、体重指数（<30 kg/m 2 、≥30千克/米 2 ）、吸烟状况（从不吸烟、目前吸烟、过去吸烟）、饮酒状况（从不饮酒、目前饮酒、戒酒）、体力活动（不活跃、不足或活跃）、慢性肺部疾病（是、否）和肾脏疾病（是、否）。血清 25（OH）D 浓度与这些分层变量之间的潜在交互作用是通过在模型中加入相应的交叉项来检测的。
In addition, to check the robustness of the association, a series of sensitivity analyses were performed as follows: (1) considering that cancer often leads to worse survival, we excluded subjects with a history of cancer to avoid a situation in which cancer patients were unevenly distributed across comparison groups; (2) likewise, analyses were conducted after the exclusion of OA patients with a history of CVD; (3) we also removed those whose deaths occurred during the first two years of follow-up to avoid potential reverse causality as much as possible; (4) history of fracture, lipid profiles (i.e., the ratio of total cholesterol to HDL) [17], CRP [18], and blood lead [19], thought to likely influence the observed relationships according to suggestive biological links, were separately adjusted in the model.
此外，为了检验这种关联的稳健性，我们还进行了以下一系列敏感性分析：(1）考虑到癌症通常会导致生存率降低，我们剔除了有癌症病史的受试者，以避免出现癌症患者在对比组中分布不均的情况；（2）同样，在剔除有心血管疾病病史的 OA 患者后进行了分析；（3）我们还剔除了在随访前两年死亡的患者，以尽可能避免潜在的反向因果关系；（4）骨折病史、血脂情况（即总胆固醇与高密度脂蛋白的比值[ 17]、CRP [ 18]和血铅、17]、CRP[18]和血铅[19]，这些指标根据提示性生物学联系被认为可能会影响观察到的关系，在模型中分别进行了调整。
All analyses were carried out in Stata/SE version 15.1 and R version 4.0.5.
所有分析均在 Stata/SE 15.1 版和 R 4.0.5 版中进行。
3. Results 3.成果
3.1. Baseline Characteristics
3.1.基线特征
There were 4570 OA patients included in this analysis, of which approximately 65.53% were female. The mean age of all included patients was 61.82 ± 12.94 years old. The geometric mean (95% CI) of serum 25(OH)D concentration was 63.90 (95% CI: 63.08–64.73) nmol/L; a total of 16.54% of the patients were Vitamin D deficient (<50 nmol/L), and 48.15% were Vitamin D insufficient (<75 nmol/L). By the census day of 31 December 2019, 1388 OA patients died (median follow-up: 7.6 (interquartile range: 4.2–12.3) years), including 427 (30.8%) from CVD, 268 (19.3%) from cancer, and 693 (49.9%) from other causes.
本次分析共纳入 4570 名 OA 患者，其中约 65.53% 为女性。所有纳入患者的平均年龄为（61.82 ± 12.94）岁。血清 25(OH)D 浓度的几何平均数（95% CI）为 63.90（95% CI：63.08-64.73）毫摩尔/升；共有 16.54% 的患者缺乏维生素 D（<50 毫摩尔/升），48.15% 的患者维生素 D 不足（<75 毫摩尔/升）。截至2019年12月31日普查日，共有1388名OA患者死亡（中位数随访时间：7.6年（四分位间范围：4.2-12.3年）），其中427人（30.8%）死于心血管疾病，268人（19.3%）死于癌症，693人（49.9%）死于其他原因。
Table 1 presents the baseline information of OA patients by serum 25(OH)D status. OA patients with higher levels of serum 25(OH)D tended to be older, non-Hispanic whites, as well as well educated. They were less likely to be overweight/obese, be current smokers, or have comorbidities such as diabetes, cancer, and renal disease. Regarding biochemical markers, patients with higher serum 25(OH)D concentrations tended to have higher HDL and CRP. We did not observe significant differences in LDLs and triglycerides.
表 1 按血清 25（OH）D 状态列出了 OA 患者的基线信息。血清 25（OH）D 水平较高的 OA 患者多为老年人、非西班牙裔白人和受过良好教育的人。他们超重/肥胖、吸烟或患有糖尿病、癌症和肾病等并发症的可能性较小。在生化指标方面，血清 25(OH)D 浓度较高的患者往往具有较高的高密度脂蛋白和 CRP。我们没有观察到低密度脂蛋白和甘油三酯的明显差异。
Table 1 表 1
Basic characteristics of participants with OA according to serum 25(OH)D concentrations in NHANES III and NHANES 2001–2018.
根据 NHANES III 和 NHANES 2001-2018 中血清 25(OH)D 浓度得出的 OA 患者基本特征。
| Characteristic | Serum 25(OH)D Concentrations (nmol/L) a 血清 25（OH）D 浓度（毫摩尔/升） a | ||||
|---|---|---|---|---|---|
| <25.0 (n = 136) | 25.0–49.9 (n = 1054) 25.0-49.9 (n = 1054) | 50.0–74.9 (n = 1579) 50.0-74.9 (n = 1579) | ≥75.0 (n = 1801) | p | |
| Age (years, mean ± SD) 年龄（岁，平均 ± SD） | 60.89 ± 13.54 | 59.50 ± 13.96 | 59.89 ± 13.49 | 63.70 ± 11.95 | <0.001 |
| Gender, n (%) 性别，n (%) | |||||
| Male | 31 (15.40) | 355 (32.96) | 642 (40.57) | 591 (31.78) | <0.001 |
| Female | 105 (84.60) | 699 (67.04) | 937 (59.43) | 1210 (68.22) | |
| Race/ethnicity, n (%) 种族/民族，n (%) | |||||
| Non-Hispanic white 非西班牙裔白人 | 45 (53.21) | 558 (71.70) | 1068 (81.44) | 1392 (88.78) | <0.001 |
| Non-Hispanic black 非西班牙裔黑人 | 57 (26.47) | 255 (13.15) | 190 (5.48) | 145 (3.29) | |
| Mexican American 墨西哥裔美国人 | 22 (11.86) | 137 (5.68) | 121 (2.69) | 89 (1.51) | |
| Others | 12 (8.45) | 104 (9.47) | 200 (10.38) | 175 (6.41) | |
| Education level, n (%) 教育程度，n (%) | |||||
| Below high school 高中以下 | 49 (27.09) | 325 (19.84) | 359 (13.54) | 310 (10.32) | <0.001 |
| High school or equivalent 高中或同等学历 | 35 (22.50) | 270 (28.13) | 391 (24.78) | 420 (22.70) | |
| Above high school 高中以上 | 52 (50.40) | 459 (52.03) | 829 (61.68) | 1071 (66.98) | |
| Family poverty income ratio, n (%) 家庭贫困收入比率，n (%) | |||||
| ≤1 | 33 (18.19) | 192 (15.49) | 218 (9.40) | 194 (7.59) | <0.001 |
| 1< to ≤3 | 58 (45.14) | 465 (39.33) | 645 (34.74) | 725 (31.37) | |
| >3 | 45 (36.68) | 397 (45.19) | 716 (55.85) | 882 (61.04) | |
| Body mass index, n (%) 体重指数，n (%) | |||||
| Normal | 18 (10.72) | 177 (15.42) | 340 (20.93) | 475 (22.74) | <0.001 |
| Overweight | 36 (27.67) | 326 (30.33) | 495 (29.01) | 609 (33.86) | |
| Obese | 82 (61.62) | 551 (54.26) | 744 (50.06) | 717 (43.40) | |
| Physical activity, n (%) 体育活动，n (%) | |||||
| Inactive | 112 (84.55) | 675 (59.63) | 816 (46.96) | 962 (48.79) | <0.001 |
| Insufficient | 7 (3.90) | 157 (15.05) | 296 (19.64) | 259 (15.84) | |
| Active | 17 (11.55) | 222 (25.32) | 467 (33.39) | 580 (35.37) | |
| Smoking status, n (%) 吸烟状况，n (%) | |||||
| Never smoker | 66 (37.58) | 506 (45.41) | 730 (44.55) | 890 (51.50) | <0.001 |
| Current smoker 目前吸烟 | 35 (30.94) | 216 (24.13) | 242 (15.35) | 220 (13.40) | |
| Past smoker | 35 (31.49) | 332 (30.46) | 607 (40.10) | 691 (35.09) | |
| Drinking status, n (%) 酗酒情况，n（%） | |||||
| Never drinker 从不喝酒 | 30 (18.71) | 188 (13.48) | 226 (10.23) | 221 (8.37) | <0.001 |
| Current drinker 当前饮酒者 | 74 (63.13) | 574 (59.82) | 1065 (72.22) | 1229 (73.97) | |
| Abstainer | 32 (18.16) | 292 (26.70) | 288 (17.55) | 351 (17.67) | |
| Comorbidities, n (%) 合并症，n (%) | |||||
| Diabetes mellitus 糖尿病 | 44 (24.23) | 312 (26.38) | 363 (20.01) | 373 (19.44) | <0.001 |
| Hypertension | 88 (67.56) | 693 (61.84) | 1006 (58.90) | 1229 (63.56) | 0.049 |
| Cardiovascular disease 心血管疾病 | 29 (24.87) | 226 (18.75) | 293 (18.16) | 386 (18.87) | 0.154 |
| Cancer | 15 (12.44) | 178 (19.57) | 299 (19.69) | 438 (25.67) | <0.001 |
| Chronic lung disease 慢性肺病 | 31 (33.28) | 147 (12.93) | 207 (13.79) | 219 (10.11) | 0.004 |
| Renal disease 肾病 | 38 (28.21) | 224 (18.18) | 271 (13.05) | 277 (13.10) | <0.001 |
| Whole-blood biochemical markers (GM, 95% CI) 全血生化指标（GM，95% CI） | |||||
| Total cholesterol (n = 4529), mmol/L 总胆固醇（n = 4529），毫摩尔/升 | 5.130 (1.941, 5.325) | 5.168 (5.096, 5.240) | 5.083 (5.030, 5.137) | 5.003 (4.952, 5.053) | <0.001 |
| HDL (n = 4526), mmol/L 高密度脂蛋白（n = 4526），毫摩尔/升 | 1.349 (1.283, 1.417) | 1.297 (1.274, 1.320) | 1.325 (1.305, 1.344) | 1.419 (1.399, 1.439) | <0.001 |
| LDL (n = 2149), mmol/L 低密度脂蛋白（n = 2149），毫摩尔/升 | 2.651 (2.431, 2.891) | 2.980 (2.893, 3.071) | 2.925 (2.856, 2.997) | 2.772 (2.711, 2.835) | 0.002 |
| Triglyceride (n = 4549), mmol/L 甘油三酯（n = 4549），毫摩尔/升 | 1.306 (1.144, 1.491) | 1.436 (1.377, 1.497) | 1.391 (1.341, 1.443) | 1.320 (1.275, 1.367) | 0.016 |
| CRP (n = 3440), mg/dL CRP （n = 3440），毫克/分升 | 0.423 (0.341, 0.526) | 0.334 (0.311, 0.359) | 0.269 (0.253, 0.287) | 0.230 (0.215, 0.245) | <0.001 |
| Lead (n = 4023), umol/L 铅（n = 4023），umol/L | 0.105 (0.090, 0.122) | 0.095 (0.090, 0.099) | 0.080 (0.077, 0.083) | 0.068 (0.066, 0.071) | <0.001 |
NHANES, National Health and Nutrition Examination Survey; 25(OH)D, 25-Hydroxyvitamin D; OA, osteoarthritis; HDL, high-density lipoprotein; LDL, low-density lipoprotein; CRP, C-reactive protein; GM, geometric mean; CI, confidence interval; n, number of subjects; SD, standard deviation; %, weighted percentage. a Vitamin D status was categorized into four groups, severe deficiency (<25.0 nmol/L), moderate deficiency (25.0–49.9 nmol/L), insufficient (50.0–74.9 nmol/L), and sufficient (≥75.0 nmol/L), according to the Endocrine Society Clinical Practice Guidelines. All estimates accounted for complex survey designs.
25（OH）D，25-羟基维生素 D；OA，骨关节炎；HDL，高密度脂蛋白；LDL，低密度脂蛋白；CRP，C 反应蛋白；GM，几何平均数；CI，置信区间；n，受试者人数；SD，标准差；%，加权百分比。根据《内分泌学会临床实践指南》，维生素 D 状态分为四类：严重缺乏（<25.0 nmol/L）、中度缺乏（25.0-49.9 nmol/L）、不足（50.0-74.9 nmol/L）和充足（≥75.0 nmol/L）。所有估计值均考虑了复杂的调查设计。
3.2. Vitamin D and Mortality
3.2.维生素 D 与死亡率
The associations of serum 25(OH)D concentrations (on a continuous scale) with mortality risk can be visualized in Figure 2. An L-shaped association was found for all-cause mortality, and the concentration of serum 25(OH)D related to the lowest all-cause mortality risk was approximately 84.50 nmol/L. An inverse association was found for CVD-specific mortality.
血清 25（OH）D 浓度（连续标度）与死亡风险的关系见图 2。全因死亡率呈 "L "形关联，与最低全因死亡风险相关的血清 25（OH）D 浓度约为 84.50 nmol/L。心血管疾病特异性死亡率呈反比关系。
Spline analyses of all-cause (A), CVD-specific (B), cancer-specific (C), and other-cause (D) mortality according to serum 25(OH)D concentrations, with the frequency distribution histogram in the background. Solid red lines represent multivariable-adjusted hazard ratios, with light-red areas showing the 95% confidence intervals derived from restricted cubic spline regressions (spline analyses were adjusted for age, gender, race/ethnicity, education, body mass index, physical activity, smoking status, drinking status, diabetes mellitus, hypertension, cardiovascular disease, cancer, chronic lung disease, and renal disease); arrows indicate the concentration of 25(OH)D with the lowest risk of mortality; over 99% of the participants were located to the left of the red dotted lines.
根据血清 25(OH)D 浓度对全因死亡率（A）、心血管疾病特异性死亡率（B）、癌症特异性死亡率（C）和其他原因死亡率（D）进行的样条分析，背景为频率分布直方图。红色实线表示多变量调整后的危险比，浅红色区域表示限制性三次样条回归得出的 95% 置信区间（样条分析已对年龄、性别、种族/民族、教育程度、体重指数、体力活动、吸烟状况、饮酒状况、糖尿病、高血压、心血管疾病、癌症、慢性肺病和肾病进行调整）；箭头表示死亡风险最低的 25(OH)D 浓度；超过 99% 的参与者位于红色虚线左侧。
The results based on serum 25(OH)D concentrations categories are summarized in Table 2. After multivariable adjustment, categories with higher levels of 25(OH)D were significantly related to low all-cause mortality and CVD mortality but not to cancer-specific mortality. The multivariable-adjusted HRs (95% CIs) across four categories of serum 25(OH)D concentrations were 1.00 (reference), 0.49 (0.31, 0.75), 0.45 (0.29, 0.68), and 0.43 (0.27, 0.69) for all-cause mortality and 1.00 (reference), 0.28 (0.14, 0.59), 0.25 (0.12, 0.51), and 0.24 (0.11, 0.49) for CVD-specific mortality.
表 2 总结了根据血清 25(OH)D 浓度分类的结果。经多变量调整后，25（OH）D 水平较高的类别与全因死亡率和心血管疾病死亡率较低显著相关，但与癌症特异性死亡率无关。四类血清 25（OH）D 浓度的多变量调整 HRs（95% CIs）分别为 1.00（参考值）、0.49（0.31，0.75）、0.45（0.29，0.68）和 0.43（0.27，0.69）；心血管疾病特异性死亡率分别为 1.00（参考值）、0.28（0.14，0.59）、0.25（0.12，0.51）和 0.24（0.11，0.49）。
Table 2 表 2
HRs (95% CIs) for all-cause and cause-specific mortality according to serum 25(OH)D concentrations among OA patients in NHANES III and NHANES 2001–2018.
根据 NHANES III 和 NHANES 2001-2018 中 OA 患者的血清 25(OH)D 浓度得出的全因死亡率和特定病因死亡率的 HRs（95% CI）。
| Serum 25(OH)D Concentrations (nmol/L) 血清 25（OH）D 浓度（毫摩尔/升） | Per One-Unit Increment in 每增加一个单位 Natural Log-Transformed 25(OH)D 自然对数转换的 25（OH）D | ||||
|---|---|---|---|---|---|
| <25.0 | 25.0–49.9 | 50.0–74.9 | ≥75.0 | ||
| All-cause mortality 全因死亡率 | |||||
| Number of deaths/total 死亡人数/总数 | 58/136 | 429/1054 | 510/1579 | 391/1801 | |
| Model 1 a | Ref. | 0.43 (0.27, 0.67) | 0.36 (0.23, 0.56) | 0.32 (0.20, 0.53) | 0.66 (0.53, 0.83) |
| Model 2 b | Ref. | 0.46 (0.30, 0.71) | 0.40 (0.26, 0.62) | 0.39 (0.24, 0.63) | 0.76 (0.61, 0.94) |
| Model 3 c | Ref. | 0.49 (0.31, 0.75) | 0.45 (0.29, 0.68) | 0.43 (0.27, 0.69) | 0.81 (0.65, 1.00) |
| CVD mortality 心血管疾病死亡率 | |||||
| Number of deaths 死亡人数 | 22 | 138 | 156 | 111 | |
| Model 1 a | Ref. | 0.27 (0.12, 0.57) | 0.20 (0.09, 0.41) | 0.17 (0.08, 0.38) | 0.47 (0.30, 0.73) |
| Model 2 b | Ref. | 0.27 (0.13, 0.57) | 0.23 (0.11, 0.47) | 0.22 (0.10, 0.48) | 0.59 (0.37, 0.95) |
| Model 3 c | Ref. | 0.28 (0.14, 0.59) | 0.25 (0.12, 0.51) | 0.24 (0.11, 0.49) | 0.61 (0.40, 0.94) |
| Cancer mortality 癌症死亡率 | |||||
| Number of deaths 死亡人数 | 8 | 75 | 96 | 89 | |
| Model 1 a | Ref. | 0.70 (0.29, 1.73) | 0.56 (0.23, 1.37) | 0.65 (0.27, 1.59) | 0.85 (0.58, 1.26) |
| Model 2 b | Ref. | 0.77 (0.31, 1.91) | 0.69 (0.27, 1.73) | 0.83 (0.34, 2.03) | 1.00 (0.70, 1.43) |
| Model 3 c | Ref. | 0.74 (0.29, 1.87) | 0.68 (0.26, 1.73) | 0.81 (0.32, 2.04) | 1.00 (0.71, 1.42) |
| Other mortality 其他死亡率 | |||||
| Number of deaths 死亡人数 | 28 | 216 | 258 | 191 | |
| Model 1 a | Ref. | 0.52 (0.30, 0.90) | 0.47 (0.28, 0.78) | 0.39 (0.23, 0.68) | 0.73 (0.57, 0.94) |
| Model 2 b | Ref. | 0.56 (0.32, 0.98) | 0.51 (0.31, 0.86) | 0.45 (0.26, 0.78) | 0.80 (0.63, 1.02) |
| Model 3 c | Ref. | 0.64 (0.35, 1.15) | 0.61 (0.36, 1.03) | 0.54 (0.30, 0.96) | 0.88 (0.68, 1.14) |
NHANES, National Health and Nutrition Examination Survey; 25(OH)D, 25-Hydroxyvitamin D; OA, osteoarthritis; CVD, cardiovascular disease; HR, hazard ratio; CI, confidence interval; Ref., reference. a Model 1 adjusted for age, gender, and race/ethnicity; b Model 2 further adjusted (from Model 1) for education, body mass index, physical activity, smoking status, and drinking status; c Model 3 further adjusted (from Model 2) for diabetes mellitus, hypertension, cardiovascular disease, cancer, chronic lung disease, and renal disease.
NHANES，美国国家健康与营养调查；25（OH）D，25-羟维生素 D；OA，骨关节炎；CVD，心血管疾病；HR，危险比；CI，置信区间；Ref.，参考文献。 a 模型 1 对年龄、性别和种族/人种进行了调整； b 模型 2（在模型 1 的基础上）对教育程度、体重指数、体力活动、吸烟状况和饮酒状况进行了进一步调整； c 模型 3（在模型 2 的基础上）对糖尿病、高血压、心血管疾病、癌症、慢性肺部疾病和肾脏疾病进行了进一步调整。
3.3. Stratified and Sensitivity Analyses
3.3.分层分析和敏感性分析
The results of the stratified analysis are shown in Table 3. Such inverse associations remained when analyses were stratified by age (≤60 or >60 years), gender (male or female), race/ethnicity (non-Hispanic white or other), BMI (<30 or ≥30 kg/m2), smoking status (never, current, or past), drinking status (never, current, or abstainer), physical activity (inactive or insufficient/active), chronic lung disease (yes or no), renal disease (yes or no), and any type of comorbidities (yes or no), even though in some strata, the associations became insignificant. Of note, no significant interactions (p interaction > 0.05) were found between serum 25(OH)D and stratum variables except for the variable of any type of comorbidities (yes or no).
分层分析结果见表 3。按年龄（≤60 岁或大于 60 岁）、性别（男性或女性）、种族/民族（非西班牙裔白人或其他）、体重指数（＜30 或≥30 kg/m 2 ）、吸烟状况（从未、目前或过去）、饮酒状况（从未、目前或禁酒）、体力活动（不活跃或不充分/活跃）、慢性肺部疾病（是或否）、肾脏疾病（是或否）进行分层分析时，上述逆相关性依然存在、是或否）、肾脏疾病（是或否）以及任何类型的合并症（是或否），尽管在某些分层中，相关性变得不显著。值得注意的是，除了任何类型的合并症（是或否）这一变量外，血清 25（OH）D 与分层变量之间没有发现明显的交互作用（p interaction > 0.05）。
Table 3 表 3
Stratified analyses of the associations (HRs, 95% CIs) between serum 25(OH)D concentrations and all-cause mortality among OA patients in NHANES III and NHANES 2001–2018.
对 NHANES III 和 NHANES 2001-2018 中 OA 患者血清 25(OH)D 浓度与全因死亡率之间的关系（HRs、95% CIs）进行分层分析。
| Serum 25(OH)D Concentrations (nmol/L) 血清 25（OH）D 浓度（毫摩尔/升） | p interaction b | ||||
|---|---|---|---|---|---|
| <25.0 a | 25.0–49.9 a | 50.0–74.9 a | ≥75.0 a | ||
| Age (years) | 0.46 | ||||
| ≤60 (n = 1624) | Ref. | 0.13 (0.07, 0.25) | 0.15 (0.08, 0.26) | 0.19 (0.08, 0.43) | |
| >60 (n = 2946) | Ref. | 0.71 (0.40, 1.24) | 0.58 (0.33, 1.03) | 0.57 (0.31, 1.04) | |
| Gender | 0.83 | ||||
| Male (n = 1619) 男性（n = 1619） | Ref. | 0.22 (0.09, 0.54) | 0.20 (0.08, 0.50) | 0.21 (0.08, 0.52) | |
| Female (n = 2951) 女性（n = 2951） | Ref. | 0.58 (0.33, 1.02) | 0.53 (0.31, 0.91) | 0.51 (0.28, 0.92) | |
| Race/ethnicity 种族/民族 | 0.32 | ||||
| Non-Hispanic white (n = 3063) 非西班牙裔白人（n = 3063） | Ref. | 0.46 (0.26, 0.82) | 0.44 (0.26, 0.75) | 0.44 (0.24, 0.79) | |
| Others (n = 1507) 其他（n = 1507） | Ref. | 0.56 (0.31, 1.01) | 0.44 (0.23, 0.83) | 0.28 (0.14, 0.56) | |
| Body mass index 体重指数 | 0.63 | ||||
| <30 (n = 2476) | Ref. | 0.67 (0.36, 1.26) | 0.71 (0.38, 1.32) | 0.61 (0.32, 1.17) | |
| ≥30 (n = 2094) | Ref. | 0.35 (0.20, 0.61) | 0.28 (0.15, 0.50) | 0.33 (0.17, 0.62) | |
| Smoking status 吸烟情况 | 0.12 | ||||
| Never smoker (n = 2192) 从不吸烟（n = 2192） | Ref. | 0.66 (0.38, 1.14) | 0.68 (0.40, 1.17) | 0.68 (0.39, 1.21) | |
| Current smoker (n = 713) 当前吸烟者（n = 713） | Ref. | 0.18 (0.07, 0.45) | 0.19 (0.07, 0.54) | 0.16 (0.06, 0.43) | |
| Past smoker (n = 1665) 过去吸烟者（n = 1665） | Ref. | 0.57 (0.23, 1.46) | 0.43 (0.17, 1.08) | 0.42 (0.17, 1.03) | |
| Drinking status 饮酒状况 | 0.27 | ||||
| Never drinker (n = 665) 从不饮酒者（n = 665） | Ref. | 0.70 (0.29, 1.68) | 0.70 (0.30, 1.64) | 0.41 (0.17, 1.00) | |
| Current drinker (n = 2942) 当前饮酒者（n = 2942） | Ref. | 0.30 (0.14, 0.62) | 0.27 (0.13, 0.57) | 0.29 (0.13, 0.66) | |
| Abstainer (n = 963) 禁欲者（n = 963） | Ref. | 0.83 (0.41, 1.70) | 0.68 (0.31, 1.47) | 0.64 (0.29, 1.40) | |
| Physical activity 体育活动 | 0.98 | ||||
| Inactive (n = 2565) 非活动（n = 2565） | Ref. | 0.53 (0.30, 0.93) | 0.50 (0.29, 0.87) | 0.52 (0.28, 0.98) | |
| Insufficient or active (n = 2005) 不足或活跃（n = 2005） | Ref. | 0.38 (0.18, 0.79) | 0.32 (0.15, 0.68) | 0.29 (0.13, 0.63) | |
| Chronic lung disease 慢性肺病 | 0.73 | ||||
| Yes (n = 604) 是（n = 604） | Ref. | 0.45 (0.15, 1.34) | 0.33 (0.10, 1.09) | 0.28 (0.08, 0.99) | |
| No (n = 3966) 否 (n = 3966) | Ref. | 0.48 (0.31, 0.75) | 0.46 (0.29, 0.72) | 0.44 (0.27, 0.71) | |
| Renal disease 肾病 | 0.18 | ||||
| Yes (n = 810) 是 (n = 810) | Ref. | 0.71 (0.32, 1.58) | 0.78 (0.35, 1.70) | 0.68 (0.31, 1.49) | |
| No (n = 3760) 否 (n = 3760) | Ref. | 0.42 (0.24, 0.73) | 0.37 (0.21, 0.65) | 0.37 (0.19, 0.71) | |
| Any type of comorbidities 任何类型的合并症 | <0.001 | ||||
| Yes (n = 3704) 是 (n = 3704) | Ref. | 0.53 (0.32, 0.88) | 0.47 (0.29, 0.77) | 0.44 (0.25, 0.77) | |
| No (n = 866) | Ref. | 0.08 (0.03, 0.24) | 0.11 (0.04, 0.30) | 0.11 (0.03, 0.34) | |
NHANES, National Health and Nutrition Examination Survey; 25(OH)D, 25-Hydroxyvitamin D; OA, osteoarthritis; HR, hazard ratio; CI, confidence interval; n, number of subjects; Ref., reference. a Model adjusted for age, gender, race/ethnicity, education, body mass index, physical activity, smoking status, drinking status, diabetes mellitus, hypertension, cardiovascular disease, cancer, chronic lung disease, and renal disease. b Wald test was performed to examine the interaction between continuous serum 25(OH)D and stratification variables.
NHANES，美国国家健康与营养调查；25(OH)D，25-羟维生素 D；OA，骨关节炎；HR，危险比；CI，置信区间；n，受试者人数；Ref. a 模型调整了年龄、性别、种族/民族、教育程度、体重指数、体力活动、吸烟状况、饮酒状况、糖尿病、高血压、心血管疾病、癌症、慢性肺病和肾病。对连续血清 25（OH）D 与分层变量之间的交互作用进行了 Wald 检验。
To further test the robustness of our findings, several sensitivity analyses were performed. Sensitivity analyses excluding OA patients with a history of cancers (n = 930), patients with a history of CVD (n = 934), and patients who died within two years of follow-up (n = 182) showed similar results (Supplementary Tables S1–S3). When further adjusted for fractures (yes or no, n = 3480), lipid profiles (n = 4525), CRP (n = 3440), or lead (n = 4023), the results did not change largely (Supplementary Table S4).
为了进一步检验我们研究结果的稳健性，我们进行了多项敏感性分析。敏感性分析排除了有癌症病史的 OA 患者（n = 930）、有心血管疾病病史的患者（n = 934）和在随访两年内死亡的患者（n = 182），结果显示相似（补充表 S1-S3）。如果对骨折（是或否，n = 3480）、血脂概况（n = 4525）、CRP（n = 3440）或铅（n = 4023）进行进一步调整，结果基本没有变化（补充表 S4）。
4. Discussion 4.讨论
In this study of 4570 OA patients from the NHANES III and NHANES 2001–2018 cohorts, after a median follow-up of 7.6 (interquartile range: 4.2–12.3) years, an L-shaped association was found for all-cause mortality, and the concentration of serum 25(OH)D related with the lowest all-cause mortality risk was approximately 84.50 nmol/L. An inverse association was found for CVD-specific mortality. Additionally, the results based on categories of 25(OH)D concentrations indicated that OA patients with sufficient serum 25(OH)D were associated with a decreased risk of all-cause and CVD mortality compared with those with lower 25(OH)D levels; however, such associations were not significant for cancer-specific mortality. Similar results were found when the analyses were stratified by factors including age, race/ethnicity, BMI, smoking status, drinking status, physical activity, and comorbidities. Intriguingly, when the analysis was stratified by gender, relatively lower HRs were observed in men than in women. No profound changes were observed in the sensitivity analyses, suggesting that our results were robust.
在这项对 4570 名来自 NHANES III 和 NHANES 2001-2018 年队列的 OA 患者进行的研究中，在中位随访 7.6 年（四分位间范围：4.2-12.3）后，发现全因死亡率呈 L 型关联，与最低全因死亡风险相关的血清 25(OH)D 浓度约为 84.50 nmol/L。心血管疾病特异性死亡率呈反比关系。此外，根据 25（OH）D 浓度分类得出的结果表明，与 25（OH）D 水平较低的患者相比，血清中 25（OH）D 充足的 OA 患者的全因死亡和心血管疾病死亡风险较低；但是，这种关联在癌症特异性死亡率方面并不显著。如果按照年龄、种族/人种、体重指数、吸烟状况、饮酒状况、体力活动和合并症等因素进行分层分析，也会发现类似的结果。耐人寻味的是，当按性别进行分层分析时，男性的 HR 值相对低于女性。在敏感性分析中没有观察到深刻的变化，这表明我们的结果是稳健的。
Our findings elucidated a protective role of Vitamin D in reducing mortality among OA patients, which is in line with previous studies concentrating on the general population and diabetic patients [11,12]. Although Vitamin D is considered as a highly influential factor in OA, only a few studies found a protective effect on OA onset and progression [8,20,21]. Of note, studies that did not support a preventive role of Vitamin D were conducted mainly among the population with higher 25(OH)D status (≥50 nmol/L) [6,7,22]; it is, therefore, possible that the use of an inappropriate population might conceal the true effect of Vitamin D [23]. Our study, however, found a beneficial effect of using Vitamin D on the long-term outcome, i.e., mortality reduction. In addition, by making rough comparisons, we found that the HRs reported in our analysis were slightly smaller than those among diabetic patients when similar covariates were adjusted. This finding indicated that OA patients might benefit more from higher Vitamin D levels than diabetic patients, thus highlighting the significance of sufficient Vitamin D intake in OA patients.
我们的研究结果阐明了维生素 D 在降低 OA 患者死亡率方面的保护作用，这与之前针对普通人群和糖尿病患者的研究结果一致[11, 12]。虽然维生素 D 被认为是对 OA 有很大影响的因素，但只有少数研究发现维生素 D 对 OA 的发病和进展有保护作用[8, 20, 21]。值得注意的是，不支持维生素 D 预防作用的研究主要是在 25（OH）D 状态较高（≥50 nmol/L）的人群中进行的[6, 7, 22]；因此，使用不适当的人群可能会掩盖维生素 D 的真正作用[23]。然而，我们的研究发现，使用维生素 D 对长期结果（即降低死亡率）有益。此外，通过粗略比较，我们发现在调整了类似的协变量后，我们分析报告的 HRs 略小于糖尿病患者的 HRs。这一发现表明，与糖尿病患者相比，OA 患者可能会从更高的维生素 D 水平中获益更多，从而凸显了摄入充足的维生素 D 对 OA 患者的重要意义。
Bone diseases are a major cause of death in the elderly [24], and it is widely known that vitamin D helps the body to absorb and use calcium, protecting the bones [25]. Serum vitamin D is first converted into hormone calcitriol (known as “active Vitamin D”), which can promote the absorption of calcium, for example, by acting on the nuclei of small intestinal mucosal cells [26]. Calcium is subsequently combined with other minerals to form hard crystals, giving bones strength and structure. Additionally, calcium also has benefits in other body systems, such as reducing blood pressure and cholesterol levels, preventing tooth loss, and etc. [27]. The benefits mentioned above, whether relative to the bones, or endocrine or dental, have an essential impact on the reduction in mortality in later life. On the contrary, in the situation of Vitamin D deficiency, the body has to take calcium from its stores in the skeleton. Weakened existing bone leads to a less favorable survival [24].
骨骼疾病是老年人死亡的主要原因[24]，众所周知，维生素 D 有助于人体吸收和利用钙，保护骨骼[25]。血清维生素 D 首先转化为激素钙三醇（称为 "活性维生素 D"），它可以促进钙的吸收，例如通过作用于小肠粘膜细胞核[26]。钙随后与其他矿物质结合，形成坚硬的晶体，赋予骨骼强度和结构。此外，钙还对身体其他系统有益，如降低血压和胆固醇水平、防止牙齿脱落等[ 27]。[ 27].上述益处，无论是对骨骼、内分泌还是牙齿，都对降低晚年死亡率有着至关重要的影响。相反，在缺乏维生素 D 的情况下，人体不得不从骨骼中的储存中摄取钙。现有骨骼的减弱会导致存活率降低[24]。
Despite the benefit of Vitamin D in mortality reduction, a ceiling effect was identified via the restricted cubic spline analysis. An L-shaped association was found between serum 25(OH)D concentrations and all-cause mortality, and such a finding was consistent with that of Fan et. al., where the study population was the general population (regardless of disease history) from a large cohort, UK Biobank [11]. Nevertheless, they reported that the concentration of 25(OH)D associated with the lowest risk of all-cause mortality was 60 nmol/L [11], which was lower than that in our study (84 nmol/L). A possible reason could be the different populations used (general population vs. OA patients). Regarding the association with CVD mortality, an inverse trend was found in our analysis. Related studies focused on OA patients are scarce; however, several studies focused on the general population or prediabetic patients showed a similar decreasing trend without stating a ceiling effect, which is in line with our finding [28,29]. However, the conclusion on this issue is inclusive. For instance, a reverse J-shaped association was observed in CopD Study, and the 25(OH)D level of 70 nmol/L was related with the lowest CVD mortality risk [30]. Moreover, Xiao et al. recently published a study that also focused on the NHANES database [31]; they observed an L-shaped association between serum 25(OH)D concentrations and CVD and all-cause mortality, and the concentration of 54.40 nmol/L was related with the lowest all-cause mortality risk. The disparities between their findings and our ours could be partly explained by the follow-up period and population sample size, as our analysis was based on newly released data, NHANES 2001–2018, which included more OA patients (4570 vs. 2556) and had a longer follow-up time.
尽管维生素 D 有降低死亡率的作用，但通过限制性三次样条分析发现了上限效应。研究发现，血清 25（OH）D 浓度与全因死亡率之间存在 L 型关联，这一发现与 Fan 等人的研究结果一致，他们的研究对象是英国生物库（UK Biobank）大型队列中的普通人群（不考虑疾病史）[11]。不过，他们报告称，与全因死亡风险最低相关的 25（OH）D 浓度为 60 毫摩尔/升[11]，低于我们的研究（84 毫摩尔/升）。可能的原因是使用的人群不同（普通人群与 OA 患者）。关于与心血管疾病死亡率的关系，我们的分析发现呈反向趋势。以 OA 患者为研究对象的相关研究很少；不过，一些以普通人群或糖尿病前期患者为研究对象的研究显示了类似的下降趋势，但没有说明上限效应，这与我们的发现一致[28, 29]。然而，关于这个问题的结论是包容性的。例如，CopD 研究观察到一种反向 J 型关联，25（OH）D 水平为 70 nmol/L 时心血管疾病死亡风险最低[30]。此外，Xiao 等人最近发表了一项同样以 NHANES 数据库为对象的研究[31]；他们观察到血清 25（OH）D 浓度与心血管疾病和全因死亡率之间呈 L 型关联，浓度为 54.40 nmol/L 时，全因死亡风险最低。他们的研究结果与我们的研究结果之间存在差异的部分原因可能是随访时间和人口样本量，因为我们的分析是基于新发布的数据，即 NHANES 2001-2018，其中包括了更多的 OA 患者（4570 对 4570）。 2556），随访时间更长。
Although the associations in several subgroups did not reach statistical significance due to limited sample size, in general, the results of stratified analyses were in line with that of our main analysis. Of note, when the analysis was stratified by gender, relatively lower HRs were observed in men than in women, which suggested men might benefit more from higher Vitamin D status than women in terms of mortality reduction. Such findings can be possibly explained by gender differences in the pathogenesis and treatment of OA. Studies showed that women tended to experience a higher prevalence and more severe OA (such as decreased cartilage volume, experiencing more clinical pain, and physical difficulty) than men [32]. More importantly, most female patients included in our study were postmenopausal women that suffered hormonal changes, especially decreasing estrogen levels [33]. In such a population, higher Vitamin D exclusively might not be able to achieve benefits comparable to those in men, and more gender-specific factors such as an adequate estrogen level might also of importance [34,35]. The identification of gender-based differences in mortality reduction would be beneficial for the development of gender-personalized therapeutic strategies for OA patients, especially for postmenopausal women.
虽然由于样本量有限，几个亚组的相关性未达到统计学意义，但总体而言，分层分析的结果与我们的主要分析结果一致。值得注意的是，当按性别进行分层分析时，男性的 HR 值相对低于女性，这表明在降低死亡率方面，男性可能比女性更受益于较高的维生素 D 水平。这些研究结果可能与 OA 的发病机制和治疗方法的性别差异有关。研究表明，与男性相比，女性的 OA 患病率更高，病情也更严重（如软骨量减少、临床疼痛加剧、体力不支等）[32]。更重要的是，我们研究中的大多数女性患者都是绝经后妇女，她们的荷尔蒙发生了变化，尤其是雌激素水平下降[33]。在这样的人群中，只摄入较多的维生素 D 可能无法获得与男性相媲美的益处，更多的性别特异性因素（如充足的雌激素水平）可能也很重要[34, 35]。确定在降低死亡率方面存在的性别差异，将有利于为 OA 患者，尤其是绝经后妇女，制定针对不同性别的治疗策略。
One of the major strengths in the current study is that a nationally representative sample of OA patients in the U.S. with long follow-up was used. Second, information on the cause of death was acquired via the linkage to National Death Index records. Moreover, the comprehensive data from NHANES allowed us to adjust a wide range of confounders such as lifestyle factors, socioeconomic status, race/ethnicity, and comorbidities.
本研究的主要优势之一是采用了具有全国代表性的美国 OA 患者样本，并进行了长期随访。其次，我们通过与国家死亡指数记录的链接获得了死因信息。此外，NHANES的综合数据使我们能够调整生活方式因素、社会经济状况、种族/民族和合并症等多种混杂因素。
Several limitations of our study should be noted. Firstly, due to the lack of repeated Vitamin D measurements, we could not ascertain the associations of dynamic Vitamin D status with mortality. Secondly, the inclusion of OA patients was based on a simple question, “Doctors ever told you had OA”, without further confirmation, using participants’ medical records. However, we did not anticipate that such bias could influence our results significantly. Another limitation in our study was the limited statistical power for detecting the associations between 25(OH)D and cancer-specific mortality, given the low frequency of cancer deaths. Additionally, studies showed that the role of Vitamin D in cancer incidence and survival and in different sites of cancer might be different [36,37,38]. Therefore, when feasible, future studies that evaluate the effect of Vitamin D on different sites of cancer are warranted. Finally, due to the study design, we could not directly deduce the causality of Vitamin D with mortality in OA patients. Theoretically, this question can be addressed with the utilization of randomized controlled trials and Mendelian randomization analyses, which unfortunately we were not able to perform in this study. Future studies that evaluate the potential casual associations are of vital use in elucidating the effect of Vitamin D on a long-term basis (such as mortality reduction) on OA patients.
我们的研究存在一些局限性。首先，由于缺乏对维生素 D 的重复测量，我们无法确定动态维生素 D 状态与死亡率之间的关系。其次，纳入 OA 患者的依据是一个简单的问题，即 "医生是否曾告诉您患有 OA"，而没有使用参与者的医疗记录进行进一步确认。然而，我们并没有预料到这种偏差会对我们的结果产生重大影响。我们研究的另一个局限性是，由于癌症死亡的频率较低，检测 25（OH）D 与癌症特异性死亡率之间关系的统计能力有限。此外，研究表明，维生素 D 在癌症发病率和存活率中的作用以及在不同部位癌症中的作用可能不同[36, 37, 38]。因此，在可行的情况下，今后有必要开展研究，评估维生素 D 对不同部位癌症的影响。最后，由于研究设计的原因，我们无法直接推断维生素 D 与 OA 患者死亡率的因果关系。从理论上讲，这个问题可以通过随机对照试验和孟德尔随机分析来解决，但遗憾的是，我们在这项研究中未能做到这一点。未来评估潜在偶然关联的研究对于阐明维生素 D 对 OA 患者的长期影响（如降低死亡率）至关重要。
5. Conclusions 5.结论
This study elucidated that OA patients with sufficient serum 25(OH)D were significantly associated with a decreased risk of all-cause and CVD mortality than those at lower 25(OH)D levels, which suggested a beneficial role of Vitamin D on a long-term basis. In addition, relatively lower HRs were observed in men than in women, highlighting the need of developing gender-personalized therapeutic strategies for OA patients, especially postmenopausal women.
这项研究表明，与 25（OH）D 水平较低的患者相比，血清中 25（OH）D 水平充足的 OA 患者的全因死亡和心血管疾病死亡风险明显降低，这表明维生素 D 具有长期的有益作用。此外，与女性相比，男性的HRs相对较低，这表明有必要为OA患者，尤其是绝经后女性制定针对不同性别的治疗策略。
Acknowledgments 致谢
We thank for all generous voluntary participation of the U.S. residents who have given their personal time to make the NHANES surveys possible.
我们感谢所有自愿参与 NHANES 调查的美国居民，感谢他们为调查付出的个人时间。
Abbreviations 缩略语
BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; CVD, cardiovascular disease; HDL, high-density lipoprotein; HR, hazard ratio; ICD, the codes of International Statistical Classification of Diseases and Related Health Problems; LC-MS/MS, liquid chromatography–tandem mass spectrometry; LDL, low-density lipoprotein; LLOD, lower limit of detection; MET, metabolic equivalent; NCHS, National Center for Health Statistics; NDI, National Death Index; NHANES, National Health and Nutrition Examination Survey; OA, osteoarthritis; PHs, proportional hazards; PIR, family poverty income ratio; VD, Vitamin D.
BMI（体重指数）；CI（置信区间）；CRP（C反应蛋白）；CVD（心血管疾病）；HDL（高密度脂蛋白）；HR（危险比）；ICD（疾病和相关健康问题国际统计分类代码）；LC-MS/MS（液相色谱-串联质谱）；LDL，低密度脂蛋白；LLOD，检测下限；MET，代谢当量；NCHS，国家卫生统计中心；NDI，国家死亡指数；NHANES，国家健康与营养检查调查；OA，骨关节炎；PHs，比例危害；PIR，家庭贫困收入比；VD，维生素 D。
Supplementary Materials 补充材料
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu14214629/s1, Table S1 HR (95% CIs) for all-cause and cause-specific mortality according to serum 25(OH)D concentrations among OA patients after excluding participants with history of cancers (n = 3640); Table S2 HR (95% CIs) for all-cause and cause-specific mortality according to serum 25(OH)D concentrations among OA patients after excluding participants with history of CVD (n = 3636); Table S3 HR (95% CIs) for all-cause and cause-specific mortality according to serum 25(OH)D concentrations among OA patients after excluding participants who died within two years of follow-up (n = 4388); Table S4 HR (95% CIs) for all-cause and cause-specific mortality according to serum 25(OH)D concentrations among OA patients with further adjustment of fracture, lipid profiles, CRP, or lead.
以下辅助信息可从以下网址下载：https://www.mdpi.com/article/10.3390/nu14214629/s1，表 S1 排除癌症病史参与者（n = 3640）后，根据 OA 患者血清 25（OH）D 浓度得出的全因和特因死亡率 HR（95% CIs）；表 S2 排除心血管疾病病史参与者（n = 3636）后，根据 OA 患者血清 25（OH）D 浓度得出的全因和特因死亡率 HR（95% CIs）；表 S3 排除随访两年内死亡的参与者（n = 4388）后，根据 OA 患者血清 25（OH）D 浓度得出的全因和特定原因死亡率的 HR（95% CIs）；表 S4 进一步调整骨折、血脂、CRP 或铅后，根据 OA 患者血清 25（OH）D 浓度得出的全因和特定原因死亡率的 HR（95% CIs）。
点击此处查看更多数据文件。 (120K, zip)
Author Contributions 作者供稿
Conceptualization, J.W., F.H. and Y.L.; Data curation, J.F., Y.Y., S.M., D.C. and L.S.; Methodology, J.W., J.F. and Y.L.; Supervision, F.H. and Y.L.; Validation, S.M., D.C., L.S. and F.H.; Visualization, Y.Y.; Writing—original draft, J.W. and J.F.; Writing—review and editing, J.W., J.F., Y.Y., S.M., D.C.,L.S., F.H. and Y.L. All authors have read and agreed to the published version of the manuscript.
概念化，J.W.、F.H. 和 Y.L.；数据整理，J.F.、Y.Y.、S.M.、D.C. 和 L.S.；方法学，J.W.、J.F. 和 Y.L.；监督，F.H. 和 Y.L.；验证，S.M.、D.C、撰写-原稿，J.W.和J.F.；撰写-审阅和编辑，J.W.、J.F.、Y.Y.、S.M.、D.C.、L.S.、F.H.和Y.L. 所有作者均已阅读并同意手稿的出版版本。
Institutional Review Board Statement
机构审查委员会声明
The ethics protocol was approved by the Research Ethics Review Board of National Center for Health Statistics (https://www.cdc.gov/nchs/nhanes/irba98.htm (accessed on 31 August 2022)), and informed consent was signed by all recruited participants.
该伦理方案已获得国家卫生统计中心研究伦理审查委员会（https://www.cdc.gov/nchs/nhanes/irba98.htm（2022年8月31日访问））的批准，所有被招募的参与者都签署了知情同意书。
Data Availability Statement
数据可用性声明
The data used in this study are freely available for download by the public at: https://wwwn.cdc.gov/nchs/nhanes/Default.aspx (accessed on 31 August 2022).
本研究使用的数据可在以下网址免费下载：https://wwwn.cdc.gov/nchs/nhanes/Default.aspx（访问日期：2022 年 8 月 31 日）。
Conflicts of Interest 利益冲突
The authors have no relevant financial nor non-financial interests to disclose.
作者没有相关的经济或非经济利益需要披露。
Footnotes 脚注
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
出版商注：MDPI 对出版地图和机构隶属关系中的管辖权主张保持中立。
References 参考资料
1.Diseases G.B.D., Injuries C. Global burden of 369 diseases and injuries in 204 countries and territories, 1990--2019: A systematic analysis for the Global Burden of Disease Study 2019.Lancet.2020;396:1204-1222.[ PMC free article] [ PubMed] [ Google Scholar].
2.倡议 USBaJ .BMUS; Rosemont, IL, USA: 2018.[谷歌学者］
3.Zhao T.、Ahmad H.、Winzenberg T.、Aitken D.、de Graaff B.、Jones G.、Palmer A.J.骨关节炎患者与非骨关节炎患者健康相关生活质量的横断面和时间差异：一项为期 10 年的前瞻性研究。doi: 10.1093/rheumatology/keaa787IF：5.5 Q1 .[ PubMed] [ CrossRef] [ Google Scholar].
4.Allen K.D., Golightly Y.M. Epidemiology of osteoarthritis：State of the evidence.Curr.Opin.Rheumatol.2015;27:276-283. Doi: 10.1097/BOR.0000000000000161IF：5.1 Q2 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
5.Mabey T., Honsawek S.维生素 D 在骨关节炎中的作用：分子、细胞和临床观点。Int.J. Endocrinol.2015;2015:383918. doi: 10.1155/2015/383918IF: 2.8 Q3 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
6.Jin X.、Jones G.、Cicuttini F.、Wluka A.、Zhu Z.、Han W.、Antony B.S.E.、Wang X.、Winzenberg T.、Blizzard L.等：《补充维生素 D 对有症状的膝骨关节炎患者胫骨软骨体积和膝关节疼痛的影响》。美国医学会杂志》。2016;315:1005. DOI: 10.1001/JAMA.2016.1961IF：120.7 Q1 .[ PubMed] [ CrossRef] [ Google Scholar].
7.McAlindon T.、LaValley M.、Schneider E.、Nuite M.、Lee J.Y.、Price L.L.、Lo G.、Dawson-Hughes B.《补充维生素 D 对有症状骨关节炎患者膝关节疼痛和软骨量减少的影响》。美国医学会杂志》。2013;309:155. doi: 10.1001/jama.2012.164487IF：120.7 Q1 .[ PMC 免费文章] [ PubMed] [ CrossRef] [ Google Scholar].
8.Zheng S.、Jin X.、Cicuttini F.、Wang X.、Zhu Z.、Wluka A.、Han W.、Winzenberg T.、Antony B.、Aitken D.等：《保持维生素 D 充足与改善膝骨关节炎的结构和症状相关》。Am.J. Med.2017;130:1211-1218. doi: 10.1016/j.amjmed.2017.04.038IF：5.9 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] 。
9.Manoy P.、Yuktanandana P.、Tanavalee A.、Anomasiri W.、Ngarmukos S.、Tanpowpong T.、Honsawek S.《补充维生素 D 可提高骨关节炎患者的生活质量和身体表现》。Nutrients.2017;9:799. doi: 10.3390/nu9080799IF：5.9 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
10.Sassi F., Tamone C., D'Amelio P. 《维生素 D：营养素、激素和免疫调节剂》。Nutrients.2018;10:1656. Doi: 10.3390/nu10111656IF：5.9 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
11.Fan X., Wang J., Song M., Giovannucci E.L., Ma H., Jin G., Hu Z., Shen H., Hang D. Vitamin D Status and Risk of All-Cause and Cause-Specific Mortality in a Large Cohort: Results from the UK Biobank.J. Clin.内分泌。Metab.DOI: 10.1210/clinem/dgaa432IF：5.8 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] 。
12.Wan Z.、Guo J.、Pan A.、Chen C.、Liu L.、Liu G. 《糖尿病患者血清 25-羟维生素 D 浓度与全因和特定原因死亡率的关系》。糖尿病护理》。DOI: 10.2337/dc20-1485IF：16.2 Q1 .[ PubMed] [ CrossRef] [ Google Scholar].
13.March L.M.、Schwarz J.M.、Carfrae B.H.、Bagge E. 自我报告骨关节炎的临床验证。Osteoarthr.软骨。1998; 6:87-93. Doi: 10.1053/joca.1997.0098IF：7.0 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] [ PubMed] [ CrossRef] [ Google Scholar].
14.国际疾病和相关健康问题统计分类》。第5版。世界卫生组织；瑞士日内瓦：2016 年。第 10 次修订。[谷歌学者］
15.国家卫生统计中心调查数据与国家死亡指数-2019 年关联死亡率档案（LMF）的关联：Linkage Methodology and Analytic Considerations.国家卫生统计中心；美国马里兰州海茨维尔：2021 年。[谷歌学者］
16.Holick M.F.、Binkley N.C.、Bischoff-Ferrari H.A.、Gordon C.M.、Hanley D.A.、Heaney R.P.、Murad M.H.、Weaver C.M.，内分泌学会《维生素 D 缺乏症的评估、治疗和预防》：内分泌学会临床实践指南》。J. Clin.内分泌。Metab.2011;96:1911-1930. DOI: 10.1210/JC.2011-0385IF：5.8 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] 。
17.Dibaba D.T. 补充维生素 D 对血清脂质的影响：系统回顾和荟萃分析。Nutr.Dibaba D.T. Effect of vitamin D supplementation on sera lipid profiles: A systematic review and meta-analysis. Nutr：6.1 Q1 .[ PubMed] [ CrossRef] [ Google Scholar].
18.Amirkhizi F.、Asoudeh F.、Hamedi-Shahraki S.、Asghari S. 《维生素 D 状态与膝关节骨关节炎患者的炎症生物标志物和临床症状相关》。膝关节。DOI: 10.1016/j.knee.2021.12.006IF：1.9 Q3 .[ PubMed] [ CrossRef] [ Google Scholar] 。
19.Nelson A., A Shi X., A Schwartz T., Chen J.-C., Renner J.B., Caldwell K.L., Helmick C.G., Jordan J.M. Whole blood lead levels are associated with radiographic and symptomatic knee osteoarthritis：约翰斯顿县骨关节炎项目的横断面分析。Arthritis Res.2011;13:R37. doi: 10.1186/ar3270IF：4.9 Q2 .[ PMC 免费文章] [ PubMed] [ CrossRef] [ Google Scholar].
20.Laslett L.L.、Quinn S.、Burgess J.R.、Parameswaran V.、Winzenberg T.M.、Jones G.、Ding C. 中度维生素 D 缺乏与老年人膝关节和髋关节疼痛的变化有关：一项为期 5 年的纵向研究。Ann.Rheum.Dis.2014;73:697-703. DOI: 10.1136/annrheumdis-2012-202831IF: 27.4 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] 。
21.Wang X.、Cicuttini F.、Jin X.、Wluka A.E.、Han W.、Zhu Z.、Blizzard L.、Antony B.S.E.、Winzenberg T.、Jones G.等：膝关节骨关节炎患者膝关节积液-滑膜炎体积测量及补充维生素 D 的效果。Osteoarthr.Cartil.2017;25:1304-1312. DOI: 10.1016/j.joca.2017.02.804IF：7.0 Q1 .[ PubMed] [ CrossRef] [ Google Scholar].
22.Arden N.K.、Cro S.、Sheard S.、Doré C.J.、Bara A.、Tebbs S.A.、Hunter D.J.、James S.、Cooper C.、O'Neill T.等：《补充维生素 D 对膝骨关节炎的影响，VIDEO 研究》：随机对照试验。Osteoarthr.软骨。2016;24:1858-1866. doi: 10.1016/j.joca.2016.05.020IF：7.0 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
23.Park C.Y. Vitamin D in the Prevention and Treatment of Osteoarthritis：从临床干预到细胞证据。Nutrients.2019;11:243. Doi: 10.3390/nu11020243IF：5.9 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
24.Ravindrarajah R.、Hazra N.C.、Charlton J.、Jackson S.H.D.、Dregan A.、Gulliford M.C.《按 80 岁以上虚弱程度分列的骨折发生率和死亡率》：使用英国电子健康记录的队列研究。BMJ Open.2018;8:e018836. DOI: 10.1136/bmjopen-2017-018836IF: 2.9 Q2 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
25.Bouillon R., Marcocci C., Carmeliet G., Bikle D., White J.H., Dawson-Hughes B., Lips P., Munns C.F., Lazaretti-Castro M., Giustina A., et al. Skeletal and Extraskeletal Actions of Vitamin D: Current Evidence and Outstanding Questions.内分泌。DOI: 10.1210/ER.2018-00126IF: 20.3 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
26.胃酸、钙吸收及其对骨骼健康的影响。Physiol.2013;93:189-268. DOI: 10.1152/physrev.00015.2012IF：33.6 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] 。
27.Cormick G., Belizan J.M. Calcium Intake and Health.Nutrients.DOI: 10.3390/nu11071606IF：5.9 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
28.Daraghmeh A.H.、Bertoia M.L.、Al-Qadi M.O.、Abdulbaki A.M.、Roberts M.B.、Eaton C.B.维生素 D 假设的证据：NHANES III 扩展死亡率随访。动脉粥样硬化。2016;255:96–101. doi: 10.1016/j.atherosclerosis.2016.04.007IF:5.3 Q2 .[ PubMed] [ CrossRef] [ Google Scholar] 。
29.Lu Q.、Wan Z.、Guo J.、Liu L.、Pan A.、Liu G. 《糖尿病前期成人血清 25- 羟维生素 D 浓度与死亡率的关系》。J. Clin.内分泌。Metab.DOI: 10.1210/clinem/dgab402IF：5.8 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] 。
30.Durup D.、Jørgensen H.L.、Christensen J.、Tjonneland A.、Olsen A.、Halkjær J.、Lind B.、Heegaard A.-M.、Schwarz P. 《血清 25-羟基维生素 D 与心血管疾病死亡率之间的反向 J 型关系》：CopD 研究。J. Clin.内分泌。Metab.2015;100:2339-2346. doi: 10.1210/jc.2014-4551IF：5.8 Q1 .[ PubMed] [ CrossRef] [ Google Scholar] 。
31.Xiao Q.、Bin Cai B.、Yin A.、Huo H.、Lan K.、Zhou G.、Shen L.、Ben He B. 骨关节炎患者血清 25- 羟维生素 D 浓度与心血管和全因死亡率的 L 型关系：NHANES数据库前瞻性队列研究的结果。BMC Med.DOI: 10.1186/s12916-022-02510-1IF：9.3 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
32.Tschon M.、Contartese D.、Pagani S.、Borsari V.、Fini M.性别是骨关节炎的关键决定因素，而不仅仅是混杂变量。临床数据的系统回顾。J. Clin.Med.Doi: 10.3390/jcm10143178IF：3.9 Q2 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
33.Fenton A., Panay N. Estrogen, menopause and joints.Climacteric.2016;19:107-108. doi: 10.3109/13697137.2016.1151151IF: 2.8 Q2 .[ PubMed] [ CrossRef] [ Google Scholar].
34.Xiao Y.P.、Tian F.M.、Dai M.W.、Wang W.Y.、Shao L.T.、Zhang L. 《雌激素相关药物是治疗骨关节炎的新选择吗？Arthritis Res. Ther.2016;18:151. doi: 10.1186/s13075-016-1045-7IF：4.9 Q2 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
35.Jung J.H.、Bang C.H.、Song G.G.、Kim C.、Kim J.H.、Choi S.J.绝经后妇女的膝关节骨关节炎与绝经激素治疗：一项全国性横断面研究。Menopause.2018;26:598-602. Doi: 10.1097/GME.0000000000001280IF: 2.7 Q2 .[ PubMed] [ CrossRef] [ Google Scholar].
36.Jeon S.M.，Shin E.A.《探索维生素 D 在癌症中的代谢和功能》。Exp.Mol. Med.Med.2018;50:1-14. Doi: 10.1038/s12276-018-0038-9IF：12.8 Q1 .[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
37.Heath A.K.、Kim I.Y.、Hodge A.M.、English D.R.、Muller D.C.《维生素 D 状态与死亡率》：观察性研究的系统回顾。Int.环境。Res. Public Health.2019;16:383. DOI: 10.3390/IJERPH16030383.[ PMC free article] [ PubMed] [ CrossRef] [ Google Scholar].
38.Freedman D.M.、Looker A.C.、Chang S.C.、Graubard B.I.《美国血清维生素 D 与癌症死亡率的前瞻性研究》。10.3 Q1 ：10.3 Q1 .[ PubMed] [ CrossRef] [ Google Scholar].
营养素》的文章由多学科数字出版研究所 (MDPI) 提供。