Edited by: Ferdinando Franzoni, University of Pisa, Italy
Reviewed by: Andrea Natali, University of Pisa, Italy; Carlo A. Pruneti, University of Parma, Italy
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Members of the aging population who undergo surgery are at risk of postoperative cognitive dysfunction (POCD). Exploring an effective and reliable early predictor of POCD is essential to the identification of high-risk patients and to making prospective decisions. The purpose of this study was to examine whether preoperative insulin resistance is an independent predictor of POCD.
A total of 124 patients aged 60 years and older and who were scheduled for gastrointestinal surgery were enrolled in a prospective observational clinical study. All participants completed a battery of neuropsychological tests before surgery and 7 days later. POCD was defined as a decline of at least 1.5 SD on two or more of neuropsychological tests. Plasma concentration of the tumor necrosis factor α (TNF-α), C-reactive protein (CRP), and S-100β protein were measured. The status of insulin resistance was assessed by Homeostasis Model Assessment–Insulin Resistance (HOMA-IR). The relationship between HOMA-IR and POCD was assessed by Multivariable logistic regression models and the receiver operating characteristic (ROC) curve.
Fifty one patients (41.1%) were diagnosed with POCD at 7 days after surgery. Preoperative HOMA-IR values of the POCD group were significantly higher than the non-POCD group. Furthermore, CRP and TNF-α levels of the POCD group were significantly higher at each postoperative time point (
Preoperative insulin resistance is an effective predictor for the occurrence of POCD. Targeted prevention and treatment strategies of insulin resistance may be effective interventions of patients at risk for POCD.
Postoperative cognitive dysfunction (POCD) is a common neuropsychological complication after surgery, with influences on various aspects of cognitive functioning, such as learning, memory, attention, and executive function, particularly in the elderly. Currently, the evaluation of POCD is based on a battery of neuropsychological test differences between pre- and postoperative performance and is often delayed. POCD has been confirmed to be associated with multiple poor outcomes, including increased surgical complications, prolonged hospitalization, impaired quality of life, increased risk of disability, and mortality (
Recently, there has been an increase in awareness on the contribution of metabolic risk factors such as obesity, diabetes, dyslipidemia to the occurrence of POCD (
Thus, we aimed to evaluate the prospective association between insulin resistance and the subsequent risk of POCD, with a focus on whether insulin resistance is an independent predictor of POCD.
This was a prospective observational clinical trial. All procedures of study were approved by the ethical committee of the Third Xiangya Hospital of Central South University, China (K18178). The study was registered on the Chinese Clinical Trial Registry (ChiCTR1800019768). All study participants provided written informed consent before enrollment.
Eligible patients were older than 60 years, classified under American Society of Anesthesiologists physical status I–III, and scheduled for elective gastrointestinal surgery lasting at least 2 h under general anesthesia. Additional inclusion criteria were that the participants must have the ability to communicate with the interviewer and complete a battery of neuropsychological testing. Exclusion criteria were any patients who had a preexisting psychiatric or neurological disease (e.g., intracranial tumors), had undergone surgery within the past 12 months, and registered a baseline Mini-Mental State Examination (MMSE) score < 23.
On admission, all patients who were accepted provided detailed history records such as self-reports of a diagnosis of hypertension, diabetes mellitus, heart disease, and previous stroke, a detailed list of medication and were administered a clinical examination. Clinical examination includes the measurement of weight, height, waist circumference and blood pressure, using an average of 3 seated blood pressure measurements.
The metabolic syndrome was defined using a modified version of the definition recommended by the American Heart Association (
The homeostasis model assessment index (HOMA-IR) was used to assess the degree of insulin resistance (
After admission, venous blood was obtained fasting in the morning for each patient and analyzed for biochemical parameters (triglycerides, HDL-C, fasting glucose and fasting Insulin levels) using standard laboratory techniques which is already a clinical routine in our hospital. At baseline and on postoperative days (PODs) 1, 3, and 7, blood samples were collected, processed by centrifugation, then were frozen at −70°C until assayed. Neuronal injury S100β protein and 2 biomarkers of inflammation CRP and TNF-α were measured in the central laboratory of our institution, using enzyme linked immunosorbent assay (ELISA) or turbidimetric inhibition immune assay.
Subjects were first screened with SDS, MMSE to exclude subjects with serious depression or cognitive impairment. Those enrolled participants completed a battery of neuropsychological tests conducted by a trained interviewer before surgery and 7 days after surgery. This cognitive test battery consisted of Verbal Learning and fluency Test, Visuospatial Memory and Delayed Recall Test, Benton Judgment of Line Orientation, Trail Making Test Parts A and B, Digit Span and Digit Symbol Substitution Test, which primarily focus on memory, attention, and executive function. Thedetailed content of these cognitive tests have been described previously (
A postoperative neuropsychological disorder in a test has been defined as a deterioration of one standard deviation (SD) compared to the preoperative test results (namely “the 1 SD criterion”) (
In addition to the above neuropsychological tests, the following tests were also administered: (1) MMSE was used to exclude pre-existing cognitive impairment, (2) the Self-rating Depression scale (SDS) and the Self-rating Anxiety scale (SAS) were used to assess anxiety and depression at baseline, (3) the Visual analog scales 0–10 were used to assess postoperative pain and the presence of mood disorder.
During the entire perioperative period, all clinical management followed recognized clinical practice. No premedication was administered before surgery. All patients received general anesthesia. To ensure that all patients were under similar levels of anesthesia, the depth of anesthesia was monitored by the bispectral index, which is required to be maintained between 40 and 60. Except for this requirement, anesthetic agents and other aspects of management (blood pressure targets, use of vasoactive drugs) were at the attending anesthesiologist’s discretion. Generally, midazolam, sufentanil, neuromuscular blockade, and propofol were used for anesthesia induction, followed by volatile (sevoflurane) or/and intravenous (propofol) anesthesia. Endotracheal intubation was the most common form of airway maintenance. Postoperative pain therapy was most commonly treated with patient controlled analgesia device using sufentanil in the initial period. The surgery was carried out by skilled surgeons in a routine laparoscopic excision of gastrointestinal cancer. All aspects of the patient’s condition were documented in case report forms.
Descriptive statistics of variables were examined in patients with and without POCD. Quantitative data were expressed as the mean ± standard deviation. Categorical data were expressed as the number and percentage. Statistical differences between two groups were investigated using
A total of 124 patients were included. The trial details are shown in
Enrollment and follow-up of study participants.
The demographic and clinical characteristics of the study participants.
Age (y) | 69.8 ± 5.4 | 71.1 ± 5.2 | 68.8 ± 5.3 | 0.027∗ |
Female, n (%) | 60 (48.4%) | 28 (54.9%) | 32 (43.8%) | 0.225 |
Height (cm) | 160.7 ± 7.5 | 160 ± 8.1 | 161 ± 7 | 0.6 |
Weight (kg) | 58.8 ± 10.1 | 60.9 ± 9.8 | 57.4 ± 10.2 | 0.06 |
Education (y) | 6.7 ± 2.8 | 6.2 ± 3 | 7.1 ± 2.6 | 0.068 |
Smoking, n (%) | 52 (41.9%) | 22 (43.1%) | 30 (41.0%) | 0.821 |
Alcohol, n (%) | 41 (31.1%) | 18 (35.3%) | 23 (32.4%) | 0.186 |
Inactivity | 57 (45.9%) | 25 (49%) | 32 (43.8%) | 0.681 |
1–2 times/week | 25 (20.2%) | 11 (21.6%) | 14 (19.2%) | |
≥3 times/week | 42 (33.9%) | 15 (29.4%) | 27 (37%) | |
Type 2 diabetes mellitus | 43 (34.7%) | 23 (45.1%) | 20 (27.4%) | 0.042∗ |
Metabolic syndrome | 44 (35.5%) | 26 (50.9%) | 18 (24.7%) | 0.003∗ |
Self-reported cardiac disease | 23 (18.5%) | 11 (21.6%) | 12 (16.4%) | 0.47 |
Self-reported transient ischemic attack | 11 (8.9%) | 5 (9.8%) | 6 (8.2%) | 0.76 |
Waist circumference (cm) | 83.7 ± 9.0 | 87.0 ± 9.7 | 81.4 ± 7.7 | 0.002∗∗ |
Systolic blood pressure (mmHg) | 128.0 ± 17.0 | 129.3 ± 15.0 | 129 ± 15.4 | 0.918 |
Diastolic blood pressure (mmHg) | 76.9 ± 10.2 | 78.0 ± 10.2 | 76.0 ± 10.1 | 0.277 |
Fasting plasma glucose (mmol/L) | 6.0 ± 1.8 | 6.7 ± 2.3 | 5.6 ± 1.1 | < 0.001∗ |
High density lipoprotein (mmol/L) | 1.18 ± 0.3 | 1.1 ± 0.3 | 1.2 ± 0.3 | 0.019∗ |
Triglycerides (mmol/L) | 1.6 ± 1.9 | 1.8 ± 1.1 | 1.5 ± 0.9 | 0.083∗ |
Insulin (mmol/L) | 8.7 ± 4.0 | 11.3 ± 3.7 | 7.0 ± 3.2 | < 0.001∗ |
HOMA-IR | 2.5 ± 1.7 | 3.5 ± 1.7 | 1.9 ± 1.7 | < 0.001∗ |
Duration of surgery (min) | 223.9 ± 76.8 | 221.5 ± 68.6 | 222.2 ± 75.8 | 0.96 |
Blood loss (ml) | 217.7 ± 205.9 | 256.7 ± 242.4 | 185.8.0 ± 140.6 | 0.079 |
Hospital length of stay (d) | 19.5 ± 6.0 | 20.4 ± 7.0 | 18.8 ± 5.2 | 0.113 |
Pneumonia | 32 (25.8%) | 19 (37.3%) | 13 (16.9%) | 0.015∗ |
Surgery-related complications | 16 (12.9%) | 7 (13.7%) | 9 (12.3%) | 0.884 |
Baseline SAS | 24.3 ± 2.8 | 24.6 ± 2.7 | 24.1 ± 2.9 | 0.198 |
Baseline SDS | 25 ± 3.2 | 24.7 ± 3 | 25.4 ± 3.5 | 0.313 |
Baseline MMSE | 25.9 ± 2.5 | 25.4 ± 2.7 | 26.2 ± 2.3 | 0.085 |
VAS 1 day after surgery | 4.40 ± 0.98 | 4.4 ± 0.99 | 4.3 ± 1.01 | 0.503 |
Preoperative SAS and SDS showed no difference between the two groups (
The results of neuropsychological testing of participants at baseline and at 7 days after surgery were listed in
Neuropsychological test results at baseline and 7 days after surgery.
Hopkins verbal learning test-revised | 13.4 ± 3.3 | 14.2 ± 2.5 | 0.182 | 10.6 ± 3.5 | 12.3 ± 2.7 | 0.003∗∗ |
Brief visuospatial memory test-revised | 6.2 ± 2.2 | 6.6 ± 2.0 | 0.291 | 4.3 ± 2.4 | 5.4 ± 2.3 | 0.011∗ |
Trail-making test (Parts A and B)# | 310.5 ± 81.1 | 287.6 ± 74.4 | 0.104 | 395.9 ± 113.1 | 336.7 ± 87.7 | 0.002∗∗ |
Benton judgment of line orientation | 15.9 ± 2.8 | 15.5 ± 2.5 | 0.334 | 12.6 ± 3.0 | 14.0 ± 2.5 | 0.01∗ |
Digit span test | 16.1 ± 2.9 | 15.8 ± 3.2 | 0.696 | 13.5 ± 3.4 | 15.1 ± 3.2 | 0.06 |
Symbol-digit modalities test | 17.9 ± 4.9 | 19.1 ± 4.7 | 0.167 | 15.7 ± 5.4 | 17.7 ± 4.6 | 0.03∗ |
HVLT-R delayed recall test | 3.7 ± 1.4 | 4.0 ± 1.2 | 0.99 | 2.7 ± 1.9 | 3.1 ± 1.3 | 0.152 |
HVLT-R discrimination index | 22.3 ± 1.3 | 22.1 ± 1.8 | 0.549 | 21.3 ± 2.3 | 21.8 ± 1.5 | 0.183 |
BVMT-R delayed recall test | 2.8 ± 1.1 | 3.0 ± 1.3 | 0.365 | 2.0 ± 1.8 | 2.2 ± 1.3 | 0.535 |
BVMT-R discrimination index | 11.2 ± 4.1 | 11 ± 1.1 | 0.74 | 10.2 ± 4.3 | 10.6 ± 1.3 | 0.52 |
Verbal fluency test | 38.8 ± 8.8 | 42.4 ± 8.7 | 0.034∗ | 35.2 ± 9.4 | 38.1 ± 8.5 | 0.105 |
Associations between metabolic risk factors, insulin resistance and POCD.
Fasting plasma glucose | 1.638 | 1.235–2.174 | 0.01∗ | 0.984 | 0.679–1.424 | 0.93 |
Waist circumference (cm) | 1.08 | 1.032–1.132 | 0.001∗∗ | 1.052 | 0.991–1.116 | 0.095 |
HDL | 0.213 | 0.054–0.830 | 0.026∗ | 0.502 | 0.106–2.367 | 0.383 |
TG | 1.379 | 0.95–2.003 | 0.091 | NA | ||
5 component risk factors | 3.385 | 1.310–8.752 | 0.012∗ | 3.13 | 1.21–8.14 | 0.019∗ |
Metabolic syndrome | 3.437 | 1.599–7.389 | 0.002∗∗ | 0.991 | 0.332–2.961 | 0.987 |
Insulin resistance (HOMA-IR) | ||||||
Continuous | 2.277 | 1.64–3.17 | < 0.001∗∗ | 2.069 | 1.266–3.382 | 0.004∗∗ |
Dichotomized (HOMA-IR > 2.6) | 6.89 | 3.05–15.58 | < 0.001∗∗ | 3.26 | 1.07–9.91 | 0.037∗ |
Associations between Insulin Resistance and POCD in subgroup analysis based on metabolic members.
No | 81 | 28 | 2.169 | 1.603–4.280 | < 0.001∗ | 2.476 | 1.460–4.198 | 0.001∗ |
Yes | 43 | 23 | 1.851 | 1.163–2.946 | 0.009∗ | 2.384 | 1.260–4.512 | 0.008∗ |
No | 70 | 30 | 2.016 | 1.292–3.147 | 0.002∗ | 1.8 | 1.007–3.010 | 0.025∗ |
Yes | 54 | 21 | 2.817 | 1.611–4.924 | < 0.001∗ | 3.461 | 1.605–7.462 | 0.002∗ |
No | 74 | 24 | 1.995 | 1.344–2.962 | 0.001∗ | 2.43 | 1.466–4.028 | 0.001∗ |
Yes | 50 | 27 | 2.122 | 1.252–3.596 | 0.005∗ | 2.007 | 1.015–3.968 | 0.045∗ |
No | 72 | 27 | 2.459 | 1.538–3.932 | < 0.001∗ | 2.699 | 1.486–4.904 | 0.001∗ |
Yes | 52 | 24 | 1.939 | 1.245–3.020 | 0.003∗ | 2.074 | 1.153–3.730 | 0.015∗ |
No | 80 | 25 | 1.868 | 1.271–2.744 | 0.001∗ | 1.863 | 1.208–2.875 | 0.005∗ |
Yes | 40 | 26 | 2.494 | 1.326–4.690 | 0.005∗ | 3.579 | 1.521–8.424 | 0.003∗ |
The ROC curves (
Receiver operating characteristic analysis of the preoperative HOMA-IR value. ROC, Receiver operating characteristic; HOMA-IR, Homeostasis Model Assessment–Insulin Resistance.
In addition, most metabolic risk factors showed a weak-to-moderate relationship with incident POCD after adjustment for age and postoperative pneumonia. However, after adjustment for all covariates and the other metabolic risk factors, associations were substantially meaningless. Only individuals with the worst metabolic condition, holding five risk factors for metabolic syndrome have a three-times larger odds (OR 3.13, 95% CI, 1.21–8.14) of developing POCD when compared to individuals with no metabolic risk factors. However, metabolic syndrome was not associated with an increased risk of incident POCD at follow-up.
The levels of plasma biomarkers of brain injury and systemic inflammation are listed in
Plasma Biomarker Levels in Patients with and without Postoperative Cognitive Dysfunction (POCD).
D0 | 34.6 ± 6.9 | 36.5 ± 8.0 | 33.2 ± 5.6 | 0.023∗ |
D1 | 47.0 ± 11.6 | 52.1 ± 11.9 | 43.5 ± 9.9 | 0.001∗∗ |
D3 | 43.5 ± 12.2 | 47.9 ± 12.6 | 40.4 ± 11 | 0.001∗∗ |
D7 | 33.5 ± 9.6 | 37.2 ± 11.5 | 30.9 ± 7 | 0.003∗∗ |
D0 | 6.2 ± 12.6 | 7.4 ± 15.3 | 5.3 ± 10.4 | 0.411 |
D1 | 58.9 ± 26.2 | 66.2 ± 29.1 | 53.7 ± 22.9 | 0.011∗∗ |
D3 | 114.7 ± 38.5 | 134.6 ± 40.2 | 100.7 ± 30.6 | 0.000∗∗ |
D7 | 39.8 ± 26.5 | 49.0 ± 31.8 | 33.28 ± 19.9 | 0.008∗∗ |
D0 | 303.4 ± 68.5 | 314.2 ± 68.6 | 295.2 ± 67.8 | 0.110 |
D1 | 364.8.1 ± 64.6 | 377.0 ± 73.4 | 356.4 ± 56.7 | 0.088 |
D3 | 335.1 ± 79.6 | 346.4 ± 95.2 | 327.3 ± 66.2 | 0.309 |
D7 | 306.7 ± 75.7 | 329.0 ± 92.9 | 294.0 ± 54 | 0.436 |
Plasma biomarker levels in two groups at different times. TNF-α
The correlation of HOMA-IR and systemic inflammation. HOMA-IR and TNF-α concentration at baseline
In this study, we found a high prevalence of preexisting insulin resistance in older patients undergoing gastrointestinal surgery, and most importantly, preexisting insulin resistance was associated with subsequent incident POCD. This association was independent of age, postoperative pneumonia, as well as all other metabolic derangement components. These findings suggest that preexisting insulin resistance associated with risk of POCD independently.
Insulin resistance is not only a shared hallmark characteristic of metabolic disease, but also a shared neuropathological process underlying cognition aging and AD.
Many studies have drawn a strong link between insulin resistance and cognitive decline (
Currently, many animal studies have indicated that insulin and insulin resistance play a prominent role in amyloid beta metabolism and influence AD pathology (
Age has been shown to be an independent risk factor of POCD repeatedly (
Many elderly individuals suffer from multiple metabolic diseases, such as diabetes, hypertension and abdominal obesity (
As noted in previous work, both clinical and preclinical studies indicate that inflammatory reactions could be responsible for incident POCD (
Our study has limitations. One limitation is the lack of data on the ratio of CSF to serum insulin level which is closely associated with brain insulin resistance, because the influence of insulin resistance on cognitive function is mainly on the brain insulin signaling pathways. In addition, we did not evaluate amyloid beta accumulation in plasma or in CSF. This could represent a further investigation of the relationship between insulin resistance and cognitive impairment in terms of amyloid brain deposition. Additional studies with more complete data are necessary. Investigations should include POCD at 3 months and activities of daily living (ADLs). All of the above may be in favor to determine the exact contribution of insulin resistance to POCD.
In conclusion, insulin resistance is associated with an increased risk of POCD in the elderly. This observation of a link between insulin resistance and POCD suggests that insulin resistance is an effective predictor of incident POCD. Thus, consideration of insulin resistance status may help clinicians and patients to make prospective decisions. The findings warrant further direction of research, particularly with respect to the underlying mechanisms and possible treatment strategies of POCD.
The data for this manuscript will be made available by the authors to qualified researchers upon reasonable request. Requests to access the data should be directed to the corresponding author.
All participants provided a signed written informed consent before enrolment in the study. All procedures described in this study were carried out in accordance with the declaration of Chinese Clinical Trial Registry (ChiCTR1800019768).
WO and XH designed the clinical experiment. WO directed the research group in all aspects. XH was the main investigator in this research. CQ contributed to sample selection and provision. BZ and JC were responsible for data acquisition. XH, GL, and WO carried out the data analysis and interpretation, and drafted the manuscript. All authors revised the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The Supplementary Material for this article can be found online at: