Aliberti Márlon Juliano Romero, Avelino-Silva Thiago Junqueira, Covinsky Kenneth E, Bertola Laiss, Magaldi Regina Miksian, Boyd Cynthia M, Carpenter Christopher R, Yasuda Mônica Sanches, Suemoto Claudia Kimie
Laboratorio de Investigacao Medica Em Envelhecimento (LIM-66), Servico de Geriatria, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, Brazil.
Research Institute, Hospital Sirio-Libanes, Sao Paulo, Brazil.
J Am Geriatr Soc. 2025 Aug;73(8):2524-2530. doi: 10.1111/jgs.19494. Epub 2025 May 2.
Dementia often goes undetected in hospital settings, where cognitive assessments are challenging due to illness severity and delirium. This study aimed to (1) evaluate the accuracy of the Clinical Dementia Rating (CDR), based solely on knowledgeable informant reports, for detecting preexisting dementia and cognitive impairment in hospitalized patients compared to a gold-standard diagnosis, (2) and compare its performance to the 16-item Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE-16).
This cross-sectional study assessed consecutive patients aged ≥ 65 admitted to inpatient units from five hospitals in Brazil. The informant-based CDR was administered to knowledgeable informants within 48 h of admission, capturing patients' cognitive status 3 months before hospitalization to avoid the influence of acute cognitive impairments. Blinded experts provided a gold standard clinical diagnosis of dementia or cognitive impairment no dementia (CIND) based on a 90-min comprehensive assessment, including a standardized neuropsychology battery. Areas under the curve (AUC) examined diagnostic accuracy.
Of 65 participants (mean age = 79.4 years; women = 54%), 34% had dementia and 32% had CIND. Compared to the gold standard, the informant-based CDR showed excellent diagnostic accuracy for detecting dementia (AUC = 0.92; 95% confidence interval [CI] = 0.86-0.98) and cognitive impairment (AUC = 0.93; 95% CI = 0.88-0.98), with a cutoff of ≥ 1 showing 98% specificity for dementia and a cutoff of ≥ 0.5 showing 98% sensitivity for cognitive impairment. Compared to the IQCODE-16, the informant-based CDR had similar performance in detecting dementia and a nonsignificant slight advantage in identifying cognitive impairment (AUC = 0.93 vs. 0.84, p = 0.069), reducing unrecognized cognitive impairment based on medical record documentation from 70% to 2%.
Informant-based CDR is a valid and efficient tool for detecting dementia and cognitive impairment in hospitalized older adults, supporting early diagnosis and guiding multidisciplinary interventions in acute care. Its use in routine hospitals may help clinicians reduce undetected cognitive impairment, enhance decision making, and improve patient care.
在医院环境中，痴呆症常常未被发现，因为在这种环境下，由于疾病严重程度和谵妄，认知评估具有挑战性。本研究旨在：（1）与金标准诊断相比，评估仅基于知情者报告的临床痴呆评定量表（CDR）在检测住院患者中已存在的痴呆症和认知障碍方面的准确性；（2）并将其性能与16项老年人认知能力下降知情者问卷（IQCODE - 16）进行比较。
这项横断面研究评估了巴西五家医院住院病房中连续收治的年龄≥65岁的患者。在入院后48小时内，对知情者进行基于知情者的CDR评估，以获取患者住院前3个月的认知状态，以避免急性认知障碍的影响。不知情的专家基于90分钟的综合评估，包括标准化神经心理学测试组合，对痴呆症或无痴呆的认知障碍（CIND）进行金标准临床诊断。曲线下面积（AUC）用于检验诊断准确性。
65名参与者（平均年龄 = 79.4岁；女性 = 54%）中，34%患有痴呆症，32%患有CIND。与金标准相比，基于知情者的CDR在检测痴呆症（AUC = 0.92；95%置信区间[CI] = 0.86 - 0.98）和认知障碍（AUC = 0.93；95% CI = 0.88 - 0.98）方面显示出优异的诊断准确性，截断值≥1时对痴呆症的特异性为98%，截断值≥0.5时对认知障碍的敏感性为98%。与IQCODE - 16相比，基于知情者的CDR在检测痴呆症方面表现相似，在识别认知障碍方面有不显著的轻微优势（AUC = 0.93对0.84，p = 0.069），将基于病历记录的未被识别的认知障碍从70%降至2%。
基于知情者的CDR是检测住院老年人痴呆症和认知障碍的有效且高效的工具，有助于早期诊断并指导急性护理中的多学科干预。在常规医院中使用它可能有助于临床医生减少未被发现的认知障碍，增强决策制定，并改善患者护理。
