Participants are from the IDCD study, which investigates the effects of long-term type 2 diabetes-related characteristics on cognitive decline. The IDCD study design has been previously described in detail18. Briefly, the IDCD recruited community-dwelling elderly individuals with type 2 diabetes (65+ years old) living in central Israel, from approximately 11,000 clients enrolled in the diabetes registry of the Maccabi Healthcare Services (MHS). MHS is the second largest health maintenance organization (HMO), treating a representative cross-section of 2 million Israeli citizens. The MHS diabetes registry was established in 1998 to facilitate diabetes management and to improve treatment. Any of the following criteria were sufficient for enrolment into the registry: (1) HbA1c >7.25% (55.7 mmol/mol); (2) glucose >11.10 mmol/l on two exams more than 3 months apart; (3) purchase of diabetic medication twice within 3 months supported by an HbA1c >6.5% (47.5 mmol/mol) or glucose >6.94 mmol/l within half a year; (4) diagnosis of type 2 diabetes (ICD-9 code19 by a general practitioner, internist, endocrinologist, ophthalmologist or type 2 diabetes advisor, supported by an HbA1c >6.5% (47.5 mmol/ mol) or glucose >6.94 mmol/l within half a year. These criteria have been validated by 20 physicians in the MHS against their own practice records20. IDCD inclusion criteria were having type 2 diabetes; normal cognition at entry; being free of any neurological (e.g., Parkinson’s disease, stroke), psychiatric (e.g., schizophrenia) or other diseases (e.g., alcohol or drug abuse) that might affect cognition; Hebrew fluency, and having an informant. Participants were assessed by a physician experienced in assessment and diagnosis of dementia, and by a neuropsychologist, who administered a neuropsychological battery. Carotid atherosclerosis was assessed during the IDCD 36 months follow up. Four hundred seventy-two IDCD participants underwent the carotid US assessments and had complete data on Hp and sociodemographic and cardiovascular variables.
Data on risk factors and possible confounders were obtained using two methods: the Maccabi Diabetes Registry and data collection during the baseline visit of the IDCD cohort. Variables available through the Diabetes Registry were computed as the average of all the measurements done in Maccabi since entry to the Diabetes Registry and included time in the T2D registry (a good proxy of duration of T2D)21, BMI, HbA1c, fasting glucose, total, HDL and LDL cholesterol, triglycerides, vitamin B-12, vitamin D, and folic acid levels. Systolic and diastolic blood pressure was measured during the carotid US examination. During the baseline visit of the IDCD study blood samples were obtained to evaluate CRP, IL-6 and the phenotype of APOE and haptoglobin.
Ultrasound assessment procedures
All examinations were performed at the Department of Neurology, Sheba Medical Center by one of 2 qualified and experienced ultrasound technicians, after obtaining informed consent. Subjects were placed in a supine position and rested for 5 minutes prior to assessing their vital signs. Carotid Ultrasound Doppler was performed using the premium EPIQ 7 US system (Philips, Netherlands). The following indices of carotid stiffness and atherosclerosis were assessed:
Carotid intima-media thickness (cIMT): IMT is defined as the distance between the media–adventitia interface and the lumen–intima interface. Measurements were performed bilaterally at the far wall of the common carotid artery (CCA) 1.0 cm proximal to the carotid bifurcation. The mean value of computer-based points was used. For each individual, cIMT was determined as the average of 3 measurements for each artery, as was automatically computed by the QLAB software (Philips, Netherlands).
3-D Carotid plaque volume: Patients with detectable plaques were assessed for plaque volume. In standard optimized mode, using the mechanic volumetric VL13-5 broadband linear array transducer, 3D plaque scanning volume data were obtained automatically. For each volume approximately 250 single transverse images (frames) were obtained with an interval of 0.15 mm. Plaque volume was automatically calculated using the Vascular Plaque Quantification (VPQ) module (QLAB software), after selecting the beginning and ending frame and selecting at least one key frame within the plaque region.
Carotid Distensibility: Following static B-mode real-time imaging from a longitudinal section of the CCA after 5 minutes of rest, dynamic CINE looped M-mode images of consecutive cardiac cycles were stored for later offline analysis. Distensibility was assessed using the distension of both CCAs, measuring the change in diameter in systole relative to diastolic during the cardiac cycle. The vessel lumen diameter was assessed from the from the near wall to the far wall of the CCA. The maximal systolic lumen diameter was determined visually and from the R-wave of the ECG-recording and the minimal lumen diameter was used for the diastolic diameter. The end-diastolic diameter (Dd), the absolute stroke change in diameter during systole (∆A), and the relative stroke change in diameter (∆A/Dd) were computed as the mean of 10 cardiac cycles of one successive recording. Blood pressure was measured before and after the measurement session and pulse pressure (∆P) was defined as the difference between the systolic and diastolic blood pressure. The cross-sectional arterial wall distensibility coefficient was calculated according to the following equation:
Cross sectional distensibility coefficient (DC) =
Carotid Elastography: In the B-mode display, a midsection of straight CCA in longitudinal plane is chosen. The shear wave elastographic mode was activated to show paired images of B-mode and elastography at the same time. The probe was handed with standardized pressure in the 2nd – 4th quintile of the linear pressure scale, as seen using standardized real-time measurement displayed on a linear scale. Elastographic images are displayed with different color mapping for the softest, intermediate and hardest components, according to the different levels of strain. On a representative static image, the relative strain ratio (SR), between blood to carotid arterial wall were measured. The first region of interest (ROI) for the arterial wall strain was manually placed at the midpoint of posterior wall of displayed carotid artery. The second ROI for the blood strain was placed at the center of arterial lumen. SR was calculated automatically by dividing strain value of the blood by that of carotid arterial wall, using the QLAB software. Measurement were performed during 10 heart beats and an average of 3 images, preferably consecutive, were used as the elastography index.
Blood serum samples for Hp phenotype were drawn from subjects during the baseline visit. Genotyping was determined using polyacrylamide gel electrophoresis method according to established methods22.
All variables were reviewed for abnormal values, to assess skewedness and outliers. Characteristics of study participants were compared using independent-samples T-test, Wilcoxon rank-test and ANOVA, as appropriate for continuous variables, and χ2 test for categorical variables.
Total burden of carotid plaque volume was computed as the sum of plaques volume in the right and left carotid arteries. Due to a non-normal distribution of carotid plaque volume, it was a-priori categorized into 4 groups: no plaque group, and tertiles of the plaque volume: small plaque (volume ≤122 mm3) medium plaque (volume 122.1-271 mm3) and large plaque (volume >271 mm3). Impaired cIMT, carotid distensibility and elastography were defined as the worst quartile. cIMT worst quartile was determined as higher or equal to 0.9 mm, carotid distensibility was determined as the lower or equal to 13.03 [kPa-3] and carotid elastography was determined as higher or equal to 0.925 SR.
Haptoglobin phenotype was also used as an ordinal variable where 1-1 phenotype was the reference group for the 2-1 and 2-2 phenotypes. Mantel-Haenszel linear by linear association was used for the trend analysis between carotid plaque volume and haptoglobin phenotype. ANCOVA was used to estimate the association of cIMT, carotid distensibility and elastography with Hp phenotype. Ordinal logistic regression was used to estimate the odds ratio (OR) and 95% confidence interval (CI) for the association between carotid plaque volume with Hp phenotype. Binary logistic regression was used to estimate the odds ratio (OR) and 95% confidence interval (CI) for the association between impaired cIMT, carotid distensibility and elastography with Hp phenotype. Primary covariates in all analyses were age and gender as they are strongly associated with both predictors and outcomes. Secondary covariates were LDL cholesterol, triglycerides, Il-6, eGFR, diabetes treatment (none, oral hypoglycemic, insulin, oral hypoglycemic+insulin), duration of diabetes and smoking. Statistical analysis was performed using SPSS software version 24.