Diabetes, but not hypertension and obesity, is associated with postoperative cognitive dysfunction

Gunnar Lachmann, Insa Feinkohl, Friedrich Borchers, Thomas H. Ottens, Hendrik M. Nathoe, Anne Mette Sauer, Jan M. Dieleman, Finn M. Radtke, Diederik Van Dijk, Claudia Spies, Tobias Pischon

Research output: Contribution to journalJournal articleResearchpeer-review


Background/Aims: Older people undergoing surgery are at risk of developing postoperative cognitive dysfunction (POCD), but little is known of risk factors predisposing patients to POCD. Our objective was to estimate the risk of POCD associated with exposure to preoperative diabetes, hypertension, and obesity. Methods: Original data from 3 randomised controlled trials (OCTOPUS, DECS, SuDoCo) were obtained for secondary analysis on diabetes, hypertension, baseline blood pressure, obesity (BMI ≥30 kg/m2), and BMI as risk factors for POCD in multiple logistic regression models. Risk estimates were pooled across the 3 studies. Results: Analyses totalled 1,034 patients. POCD occurred in 5.2% of patients in DECS, in 9.4% in SuDoCo, and in 32.1% of patients in OCTOPUS. After adjustment for age, sex, surgery type, randomisation, obesity, and hypertension, diabetes was associated with a 1.84-fold increased risk of POCD (OR 1.84; 95% CI 1.14, 2.97; p = 0.01). Obesity, BMI, hypertension, and baseline blood pressure were each not associated with POCD in fully adjusted models (all p > 0.05). Conclusion: Diabetes, but not obesity or hypertension, is associated with increased POCD risk. Consideration of diabetes status may be helpful for risk assessment of surgical patients.

Original languageEnglish
JournalDementia and Geriatric Cognitive Disorders
Issue number3-4
Pages (from-to)193-206
Publication statusPublished - Nov 2018


  • Diabetes
  • Hypertension
  • Obesity
  • Postoperative cognitive dysfunction


Dive into the research topics of 'Diabetes, but not hypertension and obesity, is associated with postoperative cognitive dysfunction'. Together they form a unique fingerprint.

Cite this