Research Center Inserm, U897, Bordeaux 2 University, France.
Int Psychogeriatr. 2011 Nov;23(9):1405-12. doi: 10.1017/S1041610211001165. Epub 2011 Jul 22.
The objective of this study was to investigate the relationship between social network and cognitive decline, taking into account the potential bias of reverse causality.
The study sample comprised 2055 elderly participants without dementia. We assessed baseline social functioning across four variables: size of social network, satisfaction with relationships, perception of being understood, and participation in social activities. A neuropsychological battery was proposed at baseline and repeated throughout follow-up. Linear mixed models were used to investigate the relationship between social network and baseline cognitive performances and cognitive decline during the 20-year follow-up.
When controlling for the reverse causality bias by excluding participants who developed dementia during the study follow-up and after adjusting for covariates, the results showed that better social functioning at baseline was associated with better initial performances in the Isaacs Set Test and the Wechsler Paired Associate Test. However, there was no significant association with further cognitive decline. By contrast, when the bias of reverse causality was not controlled for (i.e. no exclusion of participants who developed dementia), the association between social network and global cognitive decline measured by MMSE was found to be statistically significant.
With the opportunity to exclude participants who developed dementia, and the particularly long follow-up of participants, we were able to investigate the relationship between social networks and age-related cognitive decline with a minimization of reverse causality bias. The results suggest that even though higher social functioning is concomitantly associated with better cognitive performances, it may not prevent subsequent decline.
本研究旨在探讨社会网络与认知衰退之间的关系，同时考虑到反向因果关系的潜在偏差。
研究样本包括 2055 名无痴呆的老年人。我们评估了社会功能的四个方面：社交网络的大小、人际关系的满意度、被理解的感知度和参与社会活动的程度。在基线时提出了一个神经心理学测试套件，并在整个随访过程中重复测试。线性混合模型用于研究社会网络与基线认知表现之间的关系，并在 20 年的随访期间研究认知衰退。
通过排除在研究随访期间和之后发生痴呆的参与者来控制反向因果关系的偏差，并调整了协变量后，结果表明，基线时更好的社会功能与 Isaacs 集合测试和韦氏成对关联测试中的初始表现更好相关。然而，与进一步的认知衰退没有显著关联。相比之下，当不控制反向因果关系的偏差时（即不排除发生痴呆的参与者），发现社会网络与 MMSE 测量的整体认知衰退之间的关联具有统计学意义。
通过有机会排除发生痴呆的参与者，以及参与者的特别长的随访时间，我们能够以最小化反向因果关系偏差的方式研究社会网络与年龄相关的认知衰退之间的关系。结果表明，尽管更高的社会功能与更好的认知表现同时相关，但它可能无法预防随后的衰退。
