Settings and Study Design

This cross-sectional study enrolled elderly participants (age ≥ 60 years) between 2020 and 2022 at the All India Institute of Medical Sciences, a tertiary care public hospital in New Delhi (North India). The institute's ethics committee approved the study (Ref. No. IEC-485/02.08.2019, AA-6/04.10.2019), and participants gave written informed consent. We performed the study according to the ethical standards of the Declaration of Helsinki and its subsequent amendments.

Objectives

The primary objective was to evaluate the test characteristics (sensitivity, specificity, positive and negative likelihood ratios, positive and negative predictive values, and diagnostic accuracy) of MoCA as a screening test for cognitive impairment and validate against a detailed regional neuropsychological battery in people with T2D.

Inclusion and Exclusion Criteria

We included participants with T2D (age ≥ 60 years) who were followed up at the Neurology and Endocrinology outpatient departments of our institute. We excluded those who were unwilling to participate, were known cases of major depressive disorder or psychosis, those with overt hypothyroidism [thyroid stimulating hormone (TSH) > 10 with subnormal T4], established dementia, disabling stroke, active delirium, significant visual, language or hearing disability interfering with neuropsychological testing, and impaired instrumental activities of daily living (cognitive disability index score ≥ 16, on the Scale for the Instrumental Activity of Daily Living in the Elderly (IADL-EDR) [15]). We also included a control group (age ≥ 60 years, without T2D) for comparison and for deriving the cut-offs to define cognitive impairment. These participants were either people seeking care (enrolled from the same clinics which provided cases) or were spouses of the cases. For the control group, we used the same exclusion criteria as cases, with the additional exclusion of those with malignancy, history of any stroke, Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) score ≥ 3.27. IQCODE is a structured questionnaire that captures the history of cognitive decline over the last 10 years as elicited from the informant, and a score ≥ 3.27 indicates a possible presence of dementia [16, 17].

MeasurementsClinical Interview

All participants underwent a clinical interview to obtain information on sociodemographic parameters, medical history (duration of diabetes, history of hypertension, coronary artery disease, stroke, transient ischemic attack, smoking, alcohol consumption, details on medications, diabetes-related micro and macrovascular complications), depression [Patient health Questionnaire-9 (PHQ-9)] followed by anthropometric and biochemical measurements [18].

Psychological Testing

The Montreal Cognitive Assessment Scale (MoCA) We used the Hindi or English version of the MoCA. It assesses the following cognitive domains: orientation (time), visuospatial (cube copying, clock drawing), naming (three-line drawings of animals), language (sentence repetition and verbal fluency), memory (delayed recall of a five-word list) and attention/executive [forward (five digits) and backward (three digits) digit span, tapping to the letter 'A', serial subtraction (100–7), alternating trail making (five-point spread of alternating numbers and alphabets), abstraction (similarity between items), phonemic fluency ka] functions. It takes 10 min to administer. The maximum score is 30. If the patient has 12 or fewer years of education, one point is added to his final score [10].

Neuropsychological Battery The detailed neuropsychological battery (Indian Council of Medical Research Neurocognitive toolbox or ICMR-NCTB) included the following tests—Trail Making A & B, Phonemic fluency ka, ma, pa (Attention/Working memory/Executive functions); Verbal learning test—total learning, learning over trials, delayed recall, delayed recognition (verbal memory); Modified Taylor Complex Figure (MTCF) test—immediate and delayed recall (visual memory); MTCF copy and line bisection test (Visuo-perceptual functions); and Picture naming test, Category fluency for animals, vegetables, food (language) [14]. It is a culturally validated tool. We used a total of 17 tests under five cognitive domains. The IQCODE and IADL-EDR are also part of the toolbox. Certified psychologists administered the above battery, including MoCA. We obtained written permission to use ICMR-NCTB from Indian Council of Medical Research.

Definition of Mild Cognitive Impairment

MCI was defined using actuarial criteria based on neuropsychological testing [19]. As per the actuarial criteria, impairment in a cognitive domain was defined as performance ≥ 1.0 SD below the control group on at least two tests representing a cognitive domain under evaluation. To differentiate MCI from mild dementia, instrumental activities of daily living (iADLs) were supposed to be normal in MCI cases (all participants had normal iADLs, as impairment in the same was an exclusion criteria).

Test Characteristics of MoCA for MCI, with ICMR-NCTB as the Reference Standard

The sensitivity, specificity, positive and negative likelihood ratios, positive predictive value (PPV), negative predictive value (NPV) and percent correctly diagnosed, were calculated to determine the discriminant validity of MoCA for detecting cognitive impairment (on neuropsychological testing) using an online calculator from MedCalc. We evaluated the original cut-off of 26 proposed by Nasreddine et al. and a lower cut-off (cut-off 23) proposed by a recent meta-analysis [10, 12]. We further evaluated the test characteristics at other cut-offs to provide the readers and researchers with the desired information for their interpretation for clinical or research purpose.

Statistical Analysis

We analyzed data using Stata 15.0 (StataCorp, College Station, TX, USA) and presented data as number (%), mean ± SD, or median (Q25–Q75) as appropriate. Pearson chi-square test or Fisher's exact test were used (as appropriate) for qualitative variables. For quantitative variables, normality was assessed using the Shapiro–Wilk test. Student's t test for independent samples was used to compare the difference in means for the normally distributed quantitative variables. Wilcoxon rank-sum test was used for quantitative variables without normal distribution. A p value of < 0.05 was taken significant.

For each of the 17 psychological tests, a z-score was calculated [z-score = (individual participant raw score—control group mean)/control group SD] for each participant. The mean and standard deviation were derived by pooling data from all the controls. For all the tests, a higher z-score indicated a better performance except for trail making test (TMT) —A & B. The signs of these two tests were inverted for a uniform interpretation. For a given test, impairment was defined as a z-score ≥ 1.0. For a given domain, impairment was defined as any two tests (representing that domain) with z-scores ≥ 1.0. The frequency of overall MCI was calculated as the number of participants with cognitive impairment in any of the five domains [performance ≥ 1.0 SD below the control group (z-scores ≥ 1.0) on at least two tests of any one cognitive domain]. The test characteristics of MoCA for MCI using the ICMR detailed neuropsychological battery as reference were evaluated using online calculator from MedCalc.