Diabetes is associated with risk of postoperative cognitive dysfunction: A meta-analysis.

Diabetes Metabolism Research and Reviews 2017;33:e2284 doi:10.1002/dmrr.2884

Presented by: Dr C Beynon

Background

Post operative cognitive dysfunction (POCD) is common, especially in the elderly and affects 10-38% during the first 2-3 months following surgery. This paper looks to quantify the risk of POCD associated with known diabetes, chronic hyperglycaemia and those with a history of hypoglycaemia.

Design & Setting

  • This is a meta-analysis study.
  • Pubmed and Cochrane databases were searched for longitudinal studies.
  • The criteria for study inclusion were:
  1. Adult patients (>18years) undergoing surgery, where the association between POCD and diabetes, glycaemic levels and hypoglycaemic episodes was included as relative risk or odds ratios.
  2. Cognitive function pre and post surgery had to be measured using standardised, performance based, neuropsychological assessments.
  3. Glycaemic level was defined by blood glucose or HbA1c levels.
  4. Diabetic status or a history of hypoglycaemia was ascertained from hospital records and self reporting.
  5. The papers had to be in the English language

Results of systematic search

CB 5th Dec pic.png

Outcomes

  • 14 studies out of the 246 identified in the initial search were used for data extraction this included 2642 patients.
  • Fields extracted included proportion of men, mean age, surgical procedure and type of anaesthesia, definition of POCD, exposure to diabetes, hyperglycaemia and history of hypoglycaemia, associations of exposure with POCD and covariates.
  • Data analysis included use of computer programmes to estimate summary relative risk ratios, fixed and random effect modelling, I2 index calculated to look for statistical heterogeneity and sub group analysis.
  • Sensitivity analysis was used to assess the influence of individual studies on the pooled sub group analysis.

Results

  • A total of 2642 patients over 14 studies were analysed. Patients with diabetes has a 1.26 fold increased risk of POCD compared to non diabetics.
  • One study demonstrated a higher HbA1c level associated with increased risk of POCD (RR per percent higher HbA1c 2.0)

Conclusions

  • Diabetic patients have a higher risk of POCD than non diabetics. In middle aged to older adults that risk is increased by 26% compared to diabetic free individuals.
  • Those with poorer glycaemic control are at further risk (relative risk per percent higher 2.0)

Strengths

  • The paper set out clear inclusion and exclusion criteria for their meta-analysis paper and asked a specific question.
  • The patients underwent formal assessment for POCD.
  • Their methods of review included STROBE quality scores.
  • The explanation of statistical tests and modelling applied to the studies included in the paper appear to be comprehensive.

Weaknesses

  • Of the studies included 8/14 contained patients undergoing cardiac surgery with cardio-pulmonary bypass. 3/14 contained carotid endarterectomy cases under GA. Only 1/14 contained patients undergoing major general or orthopaedic surgery.
  • The majority of studies included had small numbers of patients (4/14 <100, 6/14 100-200, 4/14>200 (the largest contained 585).
  • The papers were assessed by one person, ideally and to avoid bias this should be performed by 2 people.

Implications

  • Perioperative glycaemic control is important and should be optimised prior to surgery if possible to reduce to additional risk of POCD associated with hyperglycamia.
  • Diabetic patients are at higher risk of dementia and POCD and consideration should be given to anaesthetic technique to minimise risk due to anaesthetic if possible.
  • The predominance of cardiac and carotid surgery within the studies included may have overestimated the risk of POCD, although other studies have found that at 3 months post surgery the type of surgery or anaesthetic has no influence on the incidence of POCD.

Potential for impact

This paper aims to quantify the increased risk of hyperglycaemia on POCD and I believe it has done that. It provides yet more detail on the importance of good glycaemic control pre and perioperatively.