文献速递认知障碍与痴呆

巴西未确诊痴呆的患病率与特征:基于ELSII-Brazil的5年全国代表性分析

Prevalence and Profile of Undiagnosed Dementia in Brazil: A 5-Year Nationally Representative Analysis From the ELSI-Brazil.
PubMed 文献 · Neurology · 2026-11-06

基于巴西老龄化纵向研究(ELSI-Brazil)两波全国代表性样本,比较2015-2016年与2019-2021年60岁及以上社区居住成人的未确诊痴呆患病率;未确诊痴呆患病率由5.53%降至3.86%(-1.66个百分点,p=0.013),痴呆患者中未确诊比例由89.66%降至77.17%(p=0.002),总痴呆患病率两波间无显著差异(6.21% vs 5.02%;p=0.104)。

要点

  • 未确诊痴呆患病率5年间由5.53%降至3.86%,下降1.66个百分点(p=0.013)。
  • 痴呆人群中未确诊比例由89.66%降至77.17%(p=0.002)。
  • 总痴呆患病率两波间无显著差异(6.21% vs 5.02%;p=0.104)。

核心结论

  • 巴西未确诊痴呆患病率仍高,但5年间显著下降,尤其在女性、非白人、受教育≤4年者和东北部居民等弱势群体中。
  • 作者认为这些发现提示初级保健的扩展与延续可能开始覆盖历史上服务不足的人群。
  • 作者指出地区不平等仍显著,强调需要有针对性的公共卫生政策和主动病例发现以缩小偏远地区的诊断差距。

研究设计与样本

  • 研究分析巴西老龄化纵向研究(ELSI-Brazil)两波数据,该研究为全国代表性的社区居住成人研究。
  • 比较第1波(2015-2016年)与第2波(2019-2021年)60岁及以上样本:第1波5,432人(平均年龄70.3±7.97岁,女性60.0%),第2波6,929人(平均年龄71.0±8.28岁,女性59.5%)。
  • 痴呆由结合认知测试、功能状态和知情者报告的计算方法定义;未确诊痴呆为计算方法阳性但既往无医生阿尔茨海默病诊断。
  • 估计调查加权的人群患病率及基于设计的波内与波间差异(以百分点pp表示),并采用Benjamini-Hochberg错误发现率(FDR)校正。

主要发现

  • 未确诊痴呆患病率由第1波5.53%(95%CI 4.46-6.59)降至第2波3.86%(95%CI 3.11-4.61),变化-1.66个百分点(p=0.013)。
  • 总痴呆患病率两波间无显著差异(6.21% vs 5.02%;p=0.104)。
  • 痴呆人群中未确诊比例由89.66%降至77.17%(p=0.002)。
  • 下降集中于历史上服务不足的群体:女性-2.57个百分点、非白人-2.72个百分点、受教育≤4年者-2.71个百分点、东北部居民-4.68个百分点(所有FDR q=0.010)。

关键数值与推荐

  • 未确诊痴呆患病率:第1波5.53%(95%CI 4.46-6.59)vs 第2波3.86%(95%CI 3.11-4.61),第2波较第1波降低1.66个百分点(p=0.013)。
  • 总痴呆患病率:第1波6.21% vs 第2波5.02%,两波间无显著差异(p=0.104)。
  • 痴呆人群中未确诊比例:第1波89.66% vs 第2波77.17%(p=0.002)。
  • 亚组下降幅度(百分点):女性-2.57、非白人-2.72、受教育≤4年者-2.71、东北部居民-4.68(所有FDR q=0.010)。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND AND OBJECTIVES: Dementia underdiagnosis is a major barrier to public health strategies, particularly in low- and middle-income countries. Although Brazil has a high prevalence of dementia, data on underdiagnosis remain scarce. We aimed to investigate the prevalence and sociodemographic profile of individuals with undiagnosed dementia in Brazil over 5 years. METHODS: We analyzed 2 waves of the Brazilian Longitudinal Study of Aging, a nationally representative study of community-dwelling adults, comparing samples from wave 1 (W1; 2015-2016) and wave 2 (W2; 2019-2021) aged 60 years or older. Dementia was defined by an algorithm combining cognitive testing, functional status, and informant report; undiagnosed dementia was algorithm-positive dementia without a prior physician diagnosis of Alzheimer disease. We estimated survey-weighted population prevalences and design-based within-wave and wave-to-wave differences (in percentage points, pp), applying the Benjamini-Hochberg false discovery rate (FDR) correction. RESULTS: The samples included 5,432 participants (mean age 70.3 ± 7.97 years, 60.0% female) in W1 and 6,929 (mean age 71.0 ± 8.28, 59.5% female) in W2. The prevalence of undiagnosed dementia decreased from 5.53% (95% CI 4.46-6.59) in W1 to 3.86% (3.11-4.61) in W2 (-1.66 pp; p = 0.013), whereas total dementia prevalence did not differ significantly between waves (6.21% vs 5.02%; p = 0.104). Among people with dementia, the undiagnosed proportion fell from 89.66% to 77.17% (p = 0.002). Declines were concentrated among historically underserved groups: women (-2.57 pp), non-White individuals (-2.72 pp), adults with ≤4 years of education (-2.71 pp), and residents of the Northeast (-4.68 pp; all FDR q = 0.010). DISCUSSION: Although undiagnosed dementia in Brazil remains high, its prevalence decreased significantly over the 5 years, particularly among vulnerable groups, individuals with low education, women, non-White populations, and in the Northeast. These findings suggest that the expansion and continuity of primary care are beginning to reach historically underserved populations. However, substantial regional inequalities persist, highlighting the need for targeted public health policies and active case finding to bridge diagnostic gaps in isolated territories.

来源:Neurology,查看原文。版权归原作者所有。

引用

  1. de Bittencourt MM, da Silva TM, Barbosa BJAP, et al. Prevalence and Profile of Undiagnosed Dementia in Brazil: A 5-Year Nationally Representative Analysis From the ELSI-Brazil. Neurology. 2026. PMID: 42993913. doi:10.1212/WNL.0000000000218735

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。