Abstract：
BACKGROUND: When using biological variation (BV) data, BV estimates need to be robust and representative. High-endurance athletes represent a population under special physiological conditions, which could influence BV estimates. Our study aimed to estimate BV in athletes for metabolism and growth-related biomarkers involved in the Athlete Biological Passport (ABP), by 2 different statistical models.
METHODS: Thirty triathletes were sampled monthly for 11 months. The samples were analyzed for human growth hormone (hGH), insulin-like growth factor-1 (IGF-1), insulin-like growth factor binding protein 3 (IGFBP-3), insulin, and N-terminal propeptide of type III procollagen (P-III-NP) by immunoassay. Bayesian and ANOVA methods were applied to estimate within-subject (CVI) and between-subject BV.
RESULTS: CVI estimates ranged from 7.8% for IGFBP-3 to 27.0% for insulin, when derived by the Bayesian method. The 2 models gave similar results, except for P-III-NP. Data were heterogeneously distributed for P-III-NP for the overall population and in females for IGF-1 and IGFBP-3. BV components were not estimated for hGH due to lack of steady state. The index of individuality was below 0.6 for all measurands, except for insulin.
CONCLUSIONS: In an athlete population, to apply a common CVI for insulin would be appropriate, but for IGF-1 and IGFBP-3 gender-specific estimates should be applied. P-III-NP data were heterogeneously distributed and using a mean CVI may not be representative for the population. The high degree of individuality for IGF-1, IGFBP-3, and P-III-NP makes them good candidates to be interpreted through reference change values and the ABP.
METHODS: Thirty triathletes were sampled monthly for 11 months. The samples were analyzed for human growth hormone (hGH), insulin-like growth factor-1 (IGF-1), insulin-like growth factor binding protein 3 (IGFBP-3), insulin, and N-terminal propeptide of type III procollagen (P-III-NP) by immunoassay. Bayesian and ANOVA methods were applied to estimate within-subject (CVI) and between-subject BV.
RESULTS: CVI estimates ranged from 7.8% for IGFBP-3 to 27.0% for insulin, when derived by the Bayesian method. The 2 models gave similar results, except for P-III-NP. Data were heterogeneously distributed for P-III-NP for the overall population and in females for IGF-1 and IGFBP-3. BV components were not estimated for hGH due to lack of steady state. The index of individuality was below 0.6 for all measurands, except for insulin.
CONCLUSIONS: In an athlete population, to apply a common CVI for insulin would be appropriate, but for IGF-1 and IGFBP-3 gender-specific estimates should be applied. P-III-NP data were heterogeneously distributed and using a mean CVI may not be representative for the population. The high degree of individuality for IGF-1, IGFBP-3, and P-III-NP makes them good candidates to be interpreted through reference change values and the ABP.
摘要：
背景：使用生物变异（BV）数据时，BV估计需要稳健且具有代表性。高耐力运动员代表着特殊生理条件下的人群,这可能会影响BV估计。我们的研究旨在评估运动员生物护照（ABP）中涉及的代谢和生长相关生物标志物的BV，通过两种不同的统计模型。
方法：每月抽取30名铁人三项运动员，共11个月。分析样品的人生长激素（hGH），胰岛素样生长因子-1（IGF-1），胰岛素样生长因子结合蛋白3（IGFBP-3），胰岛素，免疫测定III型前胶原的N端前肽（P-III-NP）。贝叶斯和ANOVA方法用于估计受试者内(CVI)和受试者间BV。
结果：CVI估计范围为IGFBP-3的7.8％至胰岛素的27.0％，当通过贝叶斯方法得出时。这两个模型给出了类似的结果，除了P-III-NP。P-III-NP的数据在总体人群中以及在女性中的IGF-1和IGFBP-3中分布不均。由于缺乏稳态，未估计hGH的BV成分。所有被测量的个性指数都低于0.6，除了胰岛素.
结论：在运动员群体中，对胰岛素应用普通CVI是合适的，但对于IGF-1和IGFBP-3，应采用性别特异性估计。P-III-NP数据是异质分布的，使用平均CVI可能不代表人群。IGF-1，IGFBP-3和P-III-NP的高度个性化使其成为通过参考变化值和ABP进行解释的良好候选者。
方法：每月抽取30名铁人三项运动员，共11个月。分析样品的人生长激素（hGH），胰岛素样生长因子-1（IGF-1），胰岛素样生长因子结合蛋白3（IGFBP-3），胰岛素，免疫测定III型前胶原的N端前肽（P-III-NP）。贝叶斯和ANOVA方法用于估计受试者内(CVI)和受试者间BV。
结果：CVI估计范围为IGFBP-3的7.8％至胰岛素的27.0％，当通过贝叶斯方法得出时。这两个模型给出了类似的结果，除了P-III-NP。P-III-NP的数据在总体人群中以及在女性中的IGF-1和IGFBP-3中分布不均。由于缺乏稳态，未估计hGH的BV成分。所有被测量的个性指数都低于0.6，除了胰岛素.
结论：在运动员群体中，对胰岛素应用普通CVI是合适的，但对于IGF-1和IGFBP-3，应采用性别特异性估计。P-III-NP数据是异质分布的，使用平均CVI可能不代表人群。IGF-1，IGFBP-3和P-III-NP的高度个性化使其成为通过参考变化值和ABP进行解释的良好候选者。
