Abstract
Objective
Cognitive impairment in type 2 diabetes is associated with cerebral glucose hypometabolism. Providing a glucose substitute such as ketone bodies might restore metabolic balance in glucose-compromised neurones and improve cognitive performance. We aimed to investigate if β-hydroxybutyrate (ketone body) infusion acutely affects cognitive performance, measured by a neuropsychological test battery, in patients with type 2 diabetes.
Design
Randomised, placebo-controlled, double-blind cross-over trial.
Methods
Eighteen patients with type 2 diabetes received i.v. ketone body (β-hydroxybutyrate) and placebo (saline) infusion in a randomised order on two separate occasions. On both days of examination, blood glucose was clamped at 7.5 mmol/L and a neuropsychological test battery was used to assess global cognitive performance (primary outcome) and specialized cognitive measures of verbal memory, working memory, executive function, psychomotor speed, and sustained attention.
Results
During neurocognitive testing, β-hydroxybutyrate concentrations were 2.4 vs 0.1 mmol/L. Working memory assessed by Wechsler Adult Intelligence Scale letter-number-sequencing significantly improved by 1.6 points (95% CI: 0.7, 2.4; non-adjusted P < 0.001) corresponding to a 17% increase in performance during ketone infusion compared to placebo. There was no change for global cognitive performance or any other cognitive measure after adjusting for multiple comparisons. Blood concentrations of β-hydroxybutyrate and glycaemic status did not associate with test performance; however, insulin resistance measured by HOMA was related to improved working memory performance during ketone infusion (β = 4%; 95% CI: 1.1, 7.7; P = 0.012).
Conclusions
Ketone infusion specifically improved working memory performance in patients with type 2 diabetes in the absence of changes in global cognition.
Introduction
Type 2 diabetes (T2D) is associated with a more rapid cognitive decline in mid- and late-life (1, 2, 3) and increased risk of mild cognitive impairment and dementia disorders (4, 5). Cognitive impairment in T2D has been observed for memory, executive function, attention, and processing speed (3, 6, 7, 8, 9) and is likely to affect abilities of daily living and diabetes self-management (10, 11). The underlying mechanisms for T2D-associated cognitive impairment are largely unknown. However, converging evidence suggests that cerebral hypometabolism may be a contributing factor (12). Brain regions crucial for cognition have reduced glucose metabolism with increasing insulin resistance and T2D which is associated with deterioration of memory and executive functions (13, 14, 15, 16, 17). In addition, cerebral glucose uptake is reduced in obese and individuals with T2D (18). This indicates that increased availability of alternative substrates, such as ketone bodies, could be a treatment strategy for cognitive impairment in T2D.
Ketone bodies are small lipid-derived molecules that serve as alternative energy substrates during fasting and dietary carbohydrate restriction (19). During experimentally induced hypoglycaemia, ketone administration attenuates and delays cognitive impairment and hormonal stress responses (20, 21) suggesting that ketone bodies uphold cerebral metabolism even during acute glucose deficiency. Unlike glucose, cerebral ketone metabolism seems to be less affected by peripheral insulin resistance (22) indicating that ketone bodies might be an attractive treatment in T2D-associated cognitive impairment.
Treatment of cognitive impairment with ketogenic agents or dietary ketosis has been investigated in patients with mild cognitive impairment and Alzheimer’s disease, both characterized by cerebral hypometabolism (23), demonstrating beneficial effects on cognition (24, 25, 26, 27, 28, 29). No trial has previously evaluated the potential effects of ketone bodies on cognitive performance in patients with T2D.
Our aim was to explore for the first time if hyperketonaemia would affect cognitive performance in T2D patients with no known cognitive or neurological disease. To test this, we evaluated participants’ cognitive performance on neuropsychological tests for attention, working memory, executive functions and verbal memory as well as a global measure of cognition during i.v. infusion of the ketone body β-hydroxybutyrate (β-OHB). We have previously demonstrated that i.v. administration of β-OHB acutely decreases cerebral glucose combustion with a maintained oxygen consumption indicating a metabolic shift in the brain from glucose to ketone oxidation (30).
Subjects and methods
Study design and participants
This randomised, double-blinded, cross-over trial was designed to compare cognitive performance in response to acute hyperketonaemia with placebo in patients with T2D.
Eighteen participants, clinically diagnosed with T2D, were enrolled in the study. Inclusion criteria were age 35–70 years, BMI 23–35 kg/m2, HbA1c below 80 mmol/mol (9.5%), and treatment with diet or any glucose-lowering medication except insulin and sodium-glucose cotransporter (SGLT) 2 inhibitors; the latter was allowed for up to 3 weeks before screening. Key exclusion criteria were recent ischemic cardiac disease, severe diabetes complications, or other major illnesses. The full list of inclusion and exclusion criteria is included in the supplementary material (Supplementary Table 1, see section on supplementary materials given at the end of this article).
Cognitive performance was assessed during ketone and saline infusion on two separate occasions. A wash-out period of 2–6 weeks was scheduled between each visit to minimize learning effects. During both experimental visits, blood glucose levels were clamped to 7.5 mmol/L to minimize inter-visit glucose fluctuations and endogenous ketogenesis. Both participants and cognitive test administrators were blinded to the intervention.
This study was approved by the local ethics committee and Danish Data agency and was carried out in accordance with the Helsinki Declaration of Good Clinical Practice. Written informed consent was obtained from all participants. The study is registered at ClinicalTrials.gov (NCT03657537).
Experimental procedures
Participants arrived at the Research Department after an overnight fast, having avoided strenuous physical activity and alcohol 3 days before. Blood samples were drawn from an antecubital vein; this arm was kept heated with a heating pad throughout the examination to obtain vasodilation as a proxy for arterialisation. Two other catheters were inserted into the contralateral arm; one to stabilize blood glucose levels with continuous infusion of human soluble insulin (Actrapid, Novo Nordisk) and glucose (20% w/v), the second for continuous infusion of either ketone solution or saline (0.9% w/v). The ketone solution was prepared by the Regional Pharmacy, providing a pH-adjusted sterile solution of Na-DL-3-hydroxybutyrate (Gold Biotechnology, St Louis, MO, USA) dissolved in sterile water (7.5% w/v).
After a 15-min baseline period, a continuous insulin infusion was initiated with a 3-min priming regime (200 mU × m−2 × min−1) and thereafter a constant infusion rate of 75 mU × m−2 × min−1 which was kept throughout the neurocognitive testing. Simultaneously, a continuous infusion of either ketone solution or saline was initiated at a constant infusion rate (0.22 g × kg−1 × h−1), as previously reported (30). Glucose infusions were adjusted to maintain target plasma glucose concentration of 7.5 ± 0.2 mmol/L. After 120 min, neurocognitive testing was initiated and lasted for approximately 45 min (Fig. 1).
Experimental design. Cognitive assessment and blood sampling for hormone responses are shown. β-OHB, β-hydroxybutyrate; PG, plasma glucose. A full colour version of this figure is available at https://doi.org/10.1530/EJE-19-0710.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Experimental design. Cognitive assessment and blood sampling for hormone responses are shown. β-OHB, β-hydroxybutyrate; PG, plasma glucose. A full colour version of this figure is available at https://doi.org/10.1530/EJE-19-0710.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Experimental design. Cognitive assessment and blood sampling for hormone responses are shown. β-OHB, β-hydroxybutyrate; PG, plasma glucose. A full colour version of this figure is available at https://doi.org/10.1530/EJE-19-0710.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Neurocognitive test battery
A comprehensive neuropsychological test battery was designed to evaluate overall cognitive performance. Seven neuropsychological tests within four different cognitive domains (‘verbal memory’, ‘working memory and executive functions’, ‘psychomotor speed’, and ‘sustained attention’) were selected. The following tests were included: Rey Auditory Verbal Learning Test (RAVLT) (31) (verbal memory), Trail Making Test (TMT) (32) part A and B (psychomotor speed and executive function), Symbol Digit Modalities Test (SDMT) (33) (psychomotor speed), Wechsler Adult Intelligence Scale III Letter-Number Sequencing test (WAIS-LNS) (34) (working memory), Verbal Fluency test (letters S and D) (35) (executive function), and Rapid Visual Processing (RVP) test from the Cambridge Neuropsychological Test Automated Battery (CANTAB) using A’ (RVP-A) and mean latency for correct responses (sustained attention).
To minimize learning effects from first to second visit, alternative versions of RAVLT and SDMT were used. Verbal IQ was estimated with the Danish Adult Reading Test (DART) equivalent to the National Adult Reading Test (NART).
The primary endpoint was a global cognitive composite score, estimated by the average of the four domains. The secondary endpoint was SDMT, a test of psychomotor speed and complex attention previously demonstrated to be heavily affected in patients with T2D (8).
Neurocognitive testing was conducted by fully trained research assistants from the NEAD Group, Psychiatric Center Copenhagen (www.neadgroup.org). The research assistants were blinded to the intervention.
Blood samples
Plasma glucose was measured every 5 min to ensure target glucose concentration using the glucose oxidase method (YSI 2300 STAT PLUS, Yellow Springs, OH, USA). Ketone concentrations of β-OHB was measured every 15 min using the FreeStyle Precision Neo ketone monitoring system (Abbott Diabetes Care Inc.). Additional blood samples were measured at baseline (time = −15 min), test start (10 min prior to neurocognitive testing), and test end (immediately after neurocognitive testing), evaluating hormonal responses to β-OHB infusion. Serum insulin, C-peptide, non-esterified fatty acids, growth hormone and cortisol were determined using the IMMULITE 2000 immunoassay system (Siemens Healthcare). Plasma samples of adrenaline and noradrenaline were extracted by solid phase extraction and the eluate was evaporated and reconstituted before analysis by liquid chromatography-tandem mass spectrometry. Plasma glucagon was measured by a commercially available ELISA kit (Mercodia) according to the manufacturer’s procedure.
To evaluate metabolic response, a venous blood gas sample was acquired measuring lactate, pH, and potassium levels at baseline and test start.
Statistical analysis
The primary outcome (global cognition) and four cognitive domains (verbal memory, working memory and executive functions, psychomotor speed, and sustained attention) were constructed from the average of standardized test scores. Standardization was accomplished by using the residuals from separate linear mixed models with visit as fixed effect and participant as random effect, hereby minimizing the influence of potential minor learning effects present during the second experimental visit. Scores where larger values represent worse performance were inversed in cognitive domains and composite. All cognitive outcomes (primary outcome, cognitive domains, and individual test scores) were analysed by linear mixed model with a treatment-visit interaction as fixed effect and participant as random effect.
Bonferroni correction for multiple testing was applied providing a significance level of P = 0.003 (0.05/16 cognitive outcomes).
Univariable regression analyses were performed to explore associations between changes in cognitive performance (ketone vs placebo) and variables of diabetes status: Homeostatic Model Assessment of Insulin Resistance (HOMA-IR), c-peptide, fasting glucose, glycated haemoglobin (HbA1c), and diabetes duration and variables of cognitive reserve: DART and years of education. Additionally, regression analyses were used to evaluate the associations between changes in cognitive performance and β-OHB concentrations.
Model assumptions were evaluated by visual inspection of residual and normal probability.
Data are presented as mean ± s.d. or median (IQR) unless otherwise indicated.
Sample size calculation was performed for primary outcome by a paired t-test, with 80% power and 5% significance level. We estimated a mean difference of 0.43 and a standard deviation of 0.56 based on a previous trial examining the effect of hyperketonaemia on cognitive performance in elderly participants using a comparable test battery (25). Assuming a 10% drop out rate, a sample of n = 18 participants had to be enrolled in the study.
Randomization was carried out by an independent project-nurse. Participants were block randomized using the blockrand and set.seed function in R.
All statistical analyses and graphical illustrations were performed using R (version 3.2.3).
Results
Twenty-one patients were screened between August 2018 and January 2019. Eighteen matched inclusion and exclusion criteria and were enrolled in the study, all completing both experimental visits. Participants (five women) were 65 ± 4 years old, had a BMI of 30.3 ± 4.5 kg/m2, and HbA1c of 50.5 ± 8.1 mmol/mol (6.8 ± 0.7%) with a mean diagnose duration of 10.0 ± 5.2 years by the time of inclusion. Detailed participant characteristics are presented in Table 1.
Baseline characteristics (n = 18).
| Age (years) | 65 ± 4 |
| BMI (kg/m2) | 30.3 ± 4.5 |
| Sex (m/f) | 13/5 |
| Race, of European descent (%) | 100 |
| HbA1c (mmol/mol) | 50.5 ± 8.1 |
| HbA1c (%) | 6.8 ± 0.7 |
| Diabetes duration (years) | 10.0 ± 5.2 |
| Number of diabetic medications | 1.4 ± 0.9 |
| Number of medications | 4.6 ± 1.9 |
| Systolic blood pressure (mmHg) | 134 ± 12 |
| Diastolic blood pressure (mmHg) | 78 ± 8 |
| Heart rate (beat/min) | 73 ± 15 |
| Educational level (years) | 14.3 ± 3.5 |
| DART score | 41.5 (34, 43) |
BMI, body mass index; DART, Danish adult reading test; HbA1c, glycated haemoglobin.
There were no adverse reactions to the ketone infusion, though mild symptoms such as light headache and light-headedness were reported during intervention.
Throughout neurocognitive assessment (time: 120–165 min), systemic β-OHB concentrations were significantly increased at the ketone infusion day compared to the placebo day with mean levels of 2.4 ± 0.6 mmol/L and 0.1 ± 0.0 mmol/L, respectively (P < 0.001; Fig. 2A). Plasma glucose levels were stabilized during the neurocognitive assessment (time 120–165 min), though mean glucose concentrations were significantly lower during placebo infusions (7.3 ± 0.4 vs 7.5 ± 0.6 mmol/L; P < 0.001; Fig. 2B).
Plasma glucose (A) and β-hydroxybutyrate (B) concentrations during ketone (open circles dashed line) and saline (closed circles solid line) infusion. Data is presented as mean ± s.e. β-OHB, β-hydroxybutyrate.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Plasma glucose (A) and β-hydroxybutyrate (B) concentrations during ketone (open circles dashed line) and saline (closed circles solid line) infusion. Data is presented as mean ± s.e. β-OHB, β-hydroxybutyrate.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Plasma glucose (A) and β-hydroxybutyrate (B) concentrations during ketone (open circles dashed line) and saline (closed circles solid line) infusion. Data is presented as mean ± s.e. β-OHB, β-hydroxybutyrate.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Neurocognitive assessment
Ketone infusion did not affect the primary outcome (global cognitive composite) compared to saline; the estimated treatment difference (ETD) was 0.1 (95% CI −0.1, 0.4; P = 0.292). Performance of WAIS-LNS (working memory) improved significantly from a mean score of 9.2 ± 2.1 during placebo to a mean score of 10.8 ± 2.4 during ketone infusion (ETD of 1.6; 95% CI 0.7, 2.4; P < 0.001) equivalent to a mean performance improvement of 17%. No effect was found for the remaining test scores or cognitive domains after adjusting for multiple comparisons. If Bonferroni corrections were left out, there was an improved domain score for working memory and executive functions (ETD of 0.4; 95% CI: 0.1, 0.7; P = 0.004) and a tendency towards improved verbal memory (ETD of 0.3; 95% CI: −0.0, 0.7; P = 0.059) during ketone infusion. Further, a modest worsening of secondary outcome (SDMT) performance was detected during ketone infusion with a mean test score of 51.6 ± 10.6 compared to 54.3 ± 11.7 (ETD of −2.8; 95% CI: −5.2, −0.3; P = 0.027) resulting in a deterioration of psychomotor speed (ETD of −0.4; 95% CI: −0.8, −0.1; P = 0.026) (Table 2).
Cognitive assessment during ketone and saline infusion. Mean raw scores from the neurocognitive tests. Estimated differences (ketone – placebo) were analysed by linear mixed effect models.
| Placebo | Ketone | ETH | |||
|---|---|---|---|---|---|
| Mean ± s.d. | Mean ± s.d. | E | 95 % CI | P | |
| Verbal memory | |||||
| Total recall I–V | 40.3 ± 9.7 | 43.4 ± 9.9 | 1.7 | −0.2, 6.4 | 0.066 |
| Immediate recall | 8.1 ± 2.1 | 8.4 ± 2.5 | 0.3 | −0.4, 1.1 | 0.359 |
| Delayed recall | 7.7 ± 2.4 | 8.1 ± 2.7 | 0.4 | −0.6, 1.4 | 0.419 |
| Recognition | 12.8 ± 2.1 | 13.0 ± 2.0 | 0.4 | −0.6, 1.3 | 0.463 |
| Composite score | 0.3 | −0.0, 0.7 | 0.059 | ||
| Working memory and executive functions | |||||
| TMT-B, s* | 87.0 ± 19.5 | 86.6 ± 26.3 | 0.7 | −11.6, 13.1 | 0.908 |
| WAIS-lns†,** | 9.2 ± 2.1 | 10.8 ± 2.4 | 1.6 | 0.7, 2.4 | <0.001 |
| Verbal fluency | 21.3 ± 9.4 | 22.6 ± 9.3 | 1.1 | −0.6, 2.9 | 0.201 |
| Composite score | 0.4 | 0.1, 0.7 | 0.004 | ||
| Psychomotor speed | |||||
| SDMT | 54.3 ± 11.7 | 51.6 ± 10.6 | −2.8 | −5.2, −0.3 | 0.027 |
| TMT-A, s* | 38.3 ± 10.3 | 39.6 ± 10.5 | 1.9 | −3.6, 7.5 | 0.495 |
| Composite score | −0.4 | −0.8, −0.1 | 0.026 | ||
| Sustained attention | |||||
| RVP A′ | 0.90 ± 0.05 | 0.90 ± 0.06 | 0.01 | −0.01, 0.02 | 0.542 |
| Mean latency, ms* | 454.7 ± 85.3 | 455.9 ± 133.6 | 2.4 | −41.2, 46.1 | 0.914 |
| Composite score | 0.1 | −0.3, 0.5 | 0.687 | ||
| Global cognition | 0.1 | −0.1, 0.4 | 0.292 | ||
*Higher scores reflect worse performance. †One data point from the ketone infusion day was excluded from the WAIS-lns analysis, due to invalid assessment. **Significant after Bonferroni correction (P < 0.003).
E, model estimate; ETD, estimated treatment difference.
Neither parameters for diabetes status (HOMA-IR, c-peptide, fasting glucose, HbA1c and diabetes duration) nor cognitive reserve (DART and years of education) was associated with change in global cognition or any of the cognitive domains (P > 0.05). Further regression analyses of individual tests showed that higher HOMA-IR was associated with a greater relative improvement in WAIS-LNS performance during ketone infusion (β = 4%; 95% CI: 1.10, 7.65; P = 0.012) and that longer diabetes duration was negatively associated with relative change in SDMT performance (β = −0.9%; 95% CI: −1.6, −0.2; P = 0.017). β-OHB levels during cognitive testing were not associated with change in any of the cognitive measures (P > 0.05).
Hormone and metabolic responses
Glucagon concentrations were reduced by the clamp procedure, but there was considerably less reduction in glucagon concentrations during ketone infusion compared to saline. Adrenaline responses were modestly inhibited by ketone infusion. In contrast, noradrenaline responses were modestly increased at test start (10 min prior to neuropsychological testing). Neither insulin, c-peptide, non-esterified free fatty acids, cortisol, nor growth hormone differed between the two intervention days (Fig. 3A, B, C, D, E, F, G and H).
Hormone responses (A-H) during ketone (open circles dashed line) and saline (closed circles solid line) infusion at baseline, test start (10 min before cognitive assessment) and test end (immediately after cognitive assessment). Data are presented as mean ± s.e. *P < 0.05; ***P < 0.001 for ketone vs saline infusion. NEFA, non-esterified free fatty acid.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Hormone responses (A-H) during ketone (open circles dashed line) and saline (closed circles solid line) infusion at baseline, test start (10 min before cognitive assessment) and test end (immediately after cognitive assessment). Data are presented as mean ± s.e. *P < 0.05; ***P < 0.001 for ketone vs saline infusion. NEFA, non-esterified free fatty acid.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Hormone responses (A-H) during ketone (open circles dashed line) and saline (closed circles solid line) infusion at baseline, test start (10 min before cognitive assessment) and test end (immediately after cognitive assessment). Data are presented as mean ± s.e. *P < 0.05; ***P < 0.001 for ketone vs saline infusion. NEFA, non-esterified free fatty acid.
Citation: European Journal of Endocrinology 182, 2; 10.1530/EJE-19-0710
Ketone infusion induced a condition of mild metabolic alkalosis increasing mean pH levels from 7.41 ± 0.02 at baseline to 7.48 ± 0.02 after 2-h infusion, which was a significant increase compared to placebo (P = 0.002). Despite the increased pH levels, we found that lactate concentrations increased significantly more during ketone infusion (P < 0.001) from 1.19 ± 0.38 to 1.91 ± 0.38 mmol/L vs 1.17 ± 0.38 to 1.41 ± 0.28 mmol/L during placebo. Additionally, plasma concentrations of potassium ions decreased during both occasions (placebo: 4.08 ± 0.32 to 3.74 ± 0.05 mmol/L vs ketone: 4.13 ± 0.39 to 3.3 ± 0.43 mmol/L), thus more pronounced during ketone infusion (P = 0.002) (Table 3).
Metabolic response during ketone and saline infusion. Data are presented as mean ± SD raw concentrations. Difference between mean change in concentrations during ketone – placebo was analysed by linear mixed model.
| Placebo | Ketone | ETH | |||||
|---|---|---|---|---|---|---|---|
| Baseline | Clamp | Baseline | Clamp | E | 95% CI | P | |
| Lactate (mmol/L) | 1.17 ± 0.38 | 1.41 ± 0.28 | 1.19 ± 0.38 | 1.91 ± 0.38 | 0.48 | 0.18, 0.77 | 0.002 |
| K+ (mmol/L) | 4.08 ± 0.32 | 3.74 ± 0.50 | 4.13 ± 0.39 | 3.3 ± 0.43 | −0.49 | −0.80, −0.17 | 0.002 |
| pH | 7.41 ± 0.03 | 7.41 ± 0.03 | 7.41 ± 0.02 | 7.48 ± 0.02 | 0.07 | 0.05, 0.09 | <0.001 |
ETD, estimated treatment difference; E, model estimate.
Discussion
This study aimed to examine if an acute increase in circulating ketone body levels would affect cognitive performance and hormonal responses in individuals with T2D. When β-OHB was infused, ketone body levels were moderately increased to a mean of 2.4 mmol/L during cognitive assessment. Similar infusions of β-OHB have previously been shown to markedly increase cerebral blood flow and drastically change cerebral metabolism from glucose towards ketone oxidation (30, 36). Yet, global cognition and most cognitive tests were unaffected by the intervention, suggesting that glucose and ketone bodies are both efficient substrates for most cognitive functions in patients with T2D. In response to acutely elevated β-OHB levels working memory (measured by WAIS-LNS) was significantly improved by an average of 17% (this effect survived adjustment for multiple comparisons across the 16 cognitive measures). No other cognitive measure significantly differed between the two examination days after adjustment for multiple comparisons.
Even though ketone levels were moderately increased by ketone infusion, there was no indication of ketoacidosis which suggests that treatments inducing higher ketone levels in patients with type 2 diabetes are metabolically safe and well tolerated. The glucose clamp required fewer corrections during the ketone infusion than during saline infusion and resulted in a slightly higher (0.2 mmol/L) glucose level during the cognitive tests; however, we consider this to be of no clinical significance. The significant differences in glucagon levels (higher during ketone body infusion compared to saline) are in line with previous findings (37) and may be of importance since glucagon-related peptides may have direct effects on the nervous system (38).
Previous studies have reported enhanced global cognition and improvements in tests assessing memory and executive functions such as paragraph recall, working memory, and cognitive flexibility after a single ketogenic meal (25, 26). In contrast, selective attention, visual memory, and visual search speed showed no improvements in these studies, suggesting that mainly verbal memory and some executive functions are sensitive to acute elevated ketones. Indeed, when analysing a composite score only comprising measures of verbal memory and executive functions, we found a significant improvement in the ketone group (P = 0.003). Despite this, we found no improved global cognition when including all tests.
Working memory was the only executive function which was significantly improved by ketone infusion. Working memory is important for higher-order executive functions and covers the capacity to temporarily hold information in mind while mentally working with it, which is important for reasoning, problem solving, and planning (39). A larger working memory capacity may therefore be relevant for daily activities. Patients with T2D generally perform worse in working memory tasks (8, 40) and exhibit glucose hypometabolism in related brain regions (14). The improved WAIS-LNS score found during ketone infusion indicates a normalization in test performance (41) and corresponds to a Cohen’s d effect size of 0.9 (calculated by the mean score difference divided by the residual s.d.), which is considered a clinically relevant effect size (42). Normalization of test performance may reflect a restored net metabolism in affected brain regions. In line with this theory, HOMA-IR was positively associated with the relative increase in WAIS-LNS performance. Since cerebral hypometabolism is more pronounced with increasing HOMA-IR (13, 14) this might reflect a larger potential for improving cerebral metabolism among individuals who are more insulin resistant. Measures of glycaemic status and other variables related to diabetes status (HOMA-IR, c-peptide, fasting glucose, HbA1c, and diabetes duration) were not associated with the relative change in WAIS-LNS score. The study was, however, underpowered for such analyses and interpretations should be made with caution.
Examining our crude data, our study showed a modest decrease in SDMT scores reflecting psychomotor speed and complex attention during ketone infusion. In a recent study, we have shown that SDMT performance is severely affected when glucose availability is reduced by acute non-severe hypoglycaemia (43). This suggests that optimal SDMT performance is susceptible to decreased cerebral glucose utilization and may not be rescued by higher levels of ketones, which further indicates that some cognitive functions may be more dependent on glucose than others; however, this effect did not survive adjustment for multiple comparisons. Future studies should include tests for a wide range of cognitive functions when examining the effects of ketones in relation to cognition.
Slightly elevated ketone body concentrations (0.0–1.2 mmol/L) have previously been shown to correlate with enhanced cognitive performance (24, 25, 26) supporting the hypothesis that ketone bodies function as alternative energy supporting undernourished neurons. In contrast, we were unable to detect this association in our study population at higher ketone concentrations (1.5–3.0 mmol/L). In healthy individuals, moderate ketone increments, as seen here, reduce cerebral glucose metabolism but maintains net metabolism by substituting glucose with ketones (30, 36, 44). At less elevated ketone concentrations, glucose utilization may not be inhibited (45). In support of this, a recent study suggested that long-term ketogenic treatment inducing only slight ketosis preserves cerebral glucose metabolism and increases net metabolism in patients with mild cognitive impairment (29). Whether net metabolic rate is increased by moderate ketone levels exceeding normal rates in individuals with T2D or other patient groups with cerebral hypometabolism remains to be explored. Thus, we cannot exclude that lower concentrations of ketones may be more efficient in supporting net metabolism and thereby treating cognitive impairment.
In the clinical setting, SGLT2 inhibitors are known to induce ketosis to a similar extent as in trials with ketogenic supplementations (46). Interestingly, recent epidemiological data suggest a lower risk for developing dementia in patients with T2D treated with SGLT2 inhibitors compared to other diabetes agents (47). This could potentially mean that an already approved drug with a known and safe side effect profile could have the potential for treating cognitive dysfunctions.
The focus of this study was to establish if ketone bodies influence cognitive functions in patients with T2D. Therefore, our study design did not include any metabolic data making it impossible to explore if any changes in cognitive outcomes were related to cerebral metabolism. From this acute setting, it is unsure to say how longer treatment of ketone supplementation will affect cognition in individuals with T2D. Our assessment of working memory was based on WAIS-LNS only and therefore does not provide a full evaluation of this cognitive domain; consequently this finding should be confirmed in future studies examining this domain by more than a single test. Our study was also limited to participants without dementia or mild cognitive impairment and with good glycaemic control as well as absence of any severe diabetic complications. However, we have no reason to believe that individuals with poorer glycaemic control or cognitive disease would have less effect from the ketone treatment, in fact we would expect the opposite. It is notable that we administered a racemic mixture of D- and L-β-OHB, but only measured the D-isoform. The L-isoform of β-OHB is metabolised at much lower rates and is not expected to contribute to the metabolism within the experimental timeframe (48).
In conclusion, this study demonstrates that levels of ketone bodies known to severely affect cerebral energy combustion, significantly improves performance on a working memory test. Future studies should consistently investigate several aspects of cognitive functioning to elucidate both positive and negative findings of treatments that induce ketosis. We propose to further study whether ketosis induced by ketogenic diets, carbohydrate restriction, or SGLT2 inhibitors offer an alleviation of cognitive impairment in patients with T2D.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/EJE-19-0710.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by the Danish Alzheimer Foundation; Skibsreder Per Henriksen, R. og hustrus Foundation; and The Jascha Foundation.
Acknowledgements
The authors thank L Schmidt and H Wodshow (University of Copenhagen, Denmark) for valuable assistance during experimental visits. Research assistants from the NEAD Group (Psychiatric Center Copenhagen, Denmark) are gratefully acknowledged for their assistance in neurocognitive assessments. The authors acknowledge J Nymann (Department of Clinical Biochemistry, Hvidovre Hospital, Denmark) for performing the serum analyses and I Brandslund and M Bergmann (Department of Clinical Biochemistry, Hospital Lillebaelt, Vejle, Denmark) for performing the LC-MS analyses of catecholamines. The authors particularly want to thank the patients who participated in this study.
References
- 1↑
Rawlings AM, Sharrett AR, Schneider AL, Coresh J, Albert M, Couper D, Griswold M, Gottesman RF, Wagenknecht LE & Windham BG et al.Diabetes in midlife and cognitive change over 20 years: a cohort study. Annals of Internal Medicine 2014 785–793. (https://doi.org/10.7326/M14-0737)
- 2↑
Cukierman T, Gerstein HC, Williamson JD. Cognitive decline and dementia in diabetes – systematic overview of prospective observational studies. Diabetologia 2005 2460–2469. (https://doi.org/10.1007/s00125-005-0023-4)
- 3↑
Reijmer YD, van den Berg E, Ruis C, Kappelle LJ, Biessels GJ. Cognitive dysfunction in patients with type 2 diabetes. Diabetes/Metabolism Research and Reviews 2010 507–519. (https://doi.org/10.1002/dmrr.1112)
- 4↑
Lu FP, Lin KP, Kuo HK. Diabetes and the risk of multi-system aging phenotypes: a systematic review and meta-analysis. PLoS ONE 2009 e4144. (https://doi.org/10.1371/journal.pone.0004144)
- 5↑
Cheng G, Huang C, Deng H, Wang H. Diabetes as a risk factor for dementia and mild cognitive impairment: a meta-analysis of longitudinal studies. Internal Medicine Journal 2012 484–491. (https://doi.org/10.1111/j.1445-5994.2012.02758.x)
- 6↑
Pelimanni E, Jehkonen M. Type 2 diabetes and cognitive functions in middle age: a meta-analysis. Journal of the International Neuropsychological Society 2019 25 215–229. (https://doi.org/10.1017/S1355617718001042)
- 7↑
Palta P, Schneider AL, Biessels GJ, Touradji P, Hill-Briggs F. Magnitude of cognitive dysfunction in adults with type 2 diabetes: a meta-analysis of six cognitive domains and the most frequently reported neuropsychological tests within domains. Journal of the International Neuropsychological Society 2014 278–291. (https://doi.org/10.1017/S1355617713001483)
- 8↑
Wong RH, Scholey A, Howe PR. Assessing premorbid cognitive ability in adults with type 2 diabetes mellitus – a review with implications for future intervention studies. Current Diabetes Reports 2014 547. (https://doi.org/10.1007/s11892-014-0547-4)
- 9↑
McCrimmon RJ, Ryan CM, Frier BM. Diabetes and cognitive dysfunction. Lancet 2012 2291–2299. (https://doi.org/10.1016/S0140-6736(12)60360-2)
- 10↑
Sinclair AJ, Girling AJ, Bayer AJ. Cognitive dysfunction in older subjects with diabetes mellitus: impact on diabetes self-management and use of care services. All Wales Research into Elderly (AWARE) Study. Diabetes Research and Clinical Practice 2000 203–212. (https://doi.org/10.1016/s0168-8227(00)00195-9)
- 11↑
Feil DG, Zhu CW, Sultzer DL. The relationship between cognitive impairment and diabetes self-management in a population-based community sample of older adults with Type 2 diabetes. Journal of Behavioral Medicine 2012 190–199. (https://doi.org/10.1007/s10865-011-9344-6)
- 12↑
Arnold SE, Arvanitakis Z, Macauley-Rambach SL, Koenig AM, Wang HY, Ahima RS, Craft S, Gandy S, Buettner C & Stoeckel LE et al.Brain insulin resistance in type 2 diabetes and Alzheimer disease: concepts and conundrums. Nature Reviews: Neurology 2018 168–181. (https://doi.org/10.1038/nrneurol.2017.185)
- 13↑
Willette AA, Bendlin BB, Starks EJ, Birdsill AC, Johnson SC, Christian BT, Okonkwo OC, La Rue A, Hermann BP & Koscik RL et al.Association of insulin resistance with cerebral glucose uptake in late middle-aged adults at risk for Alzheimer disease. JAMA Neurology 2015 1013–1020. (https://doi.org/10.1001/jamaneurol.2015.0613)
- 14↑
Baker LD, Cross DJ, Minoshima S, Belongia D, Watson GS, Craft S. Insulin resistance and Alzheimer-like reductions in regional cerebral glucose metabolism for cognitively normal adults with prediabetes or early type 2 diabetes. Archives of Neurology 2011 51–57. (https://doi.org/10.1001/archneurol.2010.225)
- 15↑
Li W, Risacher SL, Huang E, Saykin AJ & Alzheimer’s Disease Neuroimaging Initiative. Type 2 diabetes mellitus is associated with brain atrophy and hypometabolism in the ADNI cohort. Neurology 2016 595–600. (https://doi.org/10.1212/WNL.0000000000002950)
- 16↑
Roberts RO, Knopman DS, Cha RH, Mielke MM, Pankratz VS, Boeve BF, Kantarci K, Geda YE, Jack CR Jr & Petersen RC et al.Diabetes and elevated hemoglobin A1c levels are associated with brain hypometabolism but not amyloid accumulation. Journal of Nuclear Medicine 2014 759–764. (https://doi.org/10.2967/jnumed.113.132647)
- 17↑
Garcia-Casares N, Jorge RE, Garcia-Arnes JA, Acion L, Berthier ML, Gonzalez-Alegre P, Nabrozidis A, Gutierrez A, Ariza MJ & Rioja J et al.Cognitive dysfunctions in middle-aged type 2 diabetic patients and neuroimaging correlations: a cross-sectional study. Journal of Alzheimer’s Disease 2014 1337–1346. (https://doi.org/10.3233/JAD-140702)
- 18↑
Hwang JJ, Jiang L, Hamza M, Sanchez Rangel E, Dai F, Belfort-DeAguiar R, Parikh L, Koo BB, Rothman DL & Mason G et al.Blunted rise in brain glucose levels during hyperglycemia in adults with obesity and T2DM. JCI Insight 2017 95913. (https://doi.org/10.1172/jci.insight.95913)
- 19↑
Owen OE, Morgan AP, Kemp HG, Sullivan JM, Herrera MG, Cahill GF Jr. Brain metabolism during fasting. Journal of Clinical Investigation 1967 1589–1595. (https://doi.org/10.1172/JCI105650)
- 20↑
Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J. Effect of hyperketonemia and hyperlacticacidemia on symptoms, cognitive dysfunction, and counterregulatory hormone responses during hypoglycemia in normal humans. Diabetes 1994 1311–1317. (https://doi.org/10.2337/diab.43.11.1311)
- 21↑
Page KA, Williamson A, Yu N, McNay EC, Dzuira J, McCrimmon RJ, Sherwin RS. Medium-chain fatty acids improve cognitive function in intensively treated type 1 diabetic patients and support in vitro synaptic transmission during acute hypoglycemia. Diabetes 2009 1237–1244. (https://doi.org/10.2337/db08-1557)
- 22↑
Castellano CA, Hudon C, Croteau E, Fortier M, St-Pierre V, Vandenberghe C, Nugent S, Tremblay S, Paquet N & Lepage M et al.Links Between metabolic and structural changes in the brain of cognitively normal older adults: a 4-year longitudinal follow-up. Frontiers in Aging Neuroscience 2019 15. (https://doi.org/10.3389/fnagi.2019.00015)
- 23↑
Croteau E, Castellano CA, Fortier M, Bocti C, Fulop T, Paquet N, Cunnane SC. A cross-sectional comparison of brain glucose and ketone metabolism in cognitively healthy older adults, mild cognitive impairment and early Alzheimer’s disease. Experimental Gerontology 2018 18–26. (https://doi.org/10.1016/j.exger.2017.07.004)
- 24↑
Henderson ST, Vogel JL, Barr LJ, Garvin F, Jones JJ, Costantini LC. Study of the ketogenic agent AC-1202 in mild to moderate Alzheimer’s disease: a randomized, double-blind, placebo-controlled, multicenter trial. Nutrition and Metabolism 2009 31. (https://doi.org/10.1186/1743-7075-6-31)
- 25↑
Ota M, Matsuo J, Ishida I, Hattori K, Teraishi T, Tonouchi H, Ashida K, Takahashi T, Kunugi H. Effect of a ketogenic meal on cognitive function in elderly adults: potential for cognitive enhancement. Psychopharmacology 2016 3797–3802. (https://doi.org/10.1007/s00213-016-4414-7)
- 26↑
Reger MA, Henderson ST, Hale C, Cholerton B, Baker LD, Watson GS, Hyde K, Chapman D, Craft S. Effects of beta-hydroxybutyrate on cognition in memory-impaired adults. Neurobiology of Aging 2004 311–314. (https://doi.org/10.1016/S0197-4580(03)00087-3)
- 27↑
Taylor MK, Sullivan DK, Mahnken JD, Burns JM, Swerdlow RH. Feasibility and efficacy data from a ketogenic diet intervention in Alzheimer’s disease. Alzheimer’s and Dementia 2018 28–36. (https://doi.org/10.1016/j.trci.2017.11.002)
- 28↑
Newport MT, VanItallie TB, Kashiwaya Y, King MT, Veech RL. A new way to produce hyperketonemia: use of ketone ester in a case of Alzheimer’s disease. Alzheimer’s and Dementia 2015 99–103. (https://doi.org/10.1016/j.jalz.2014.01.006)
- 29↑
Fortier M, Castellano CA, Croteau E, Langlois F, Bocti C, St-Pierre V, Vandenberghe C, Bernier M, Roy M & Descoteaux M et al.A ketogenic drink improves brain energy and some measures of cognition in mild cognitive impairment. Alzheimer’s and Dementia 2019 625–634. (https://doi.org/10.1016/j.jalz.2018.12.017)
- 30↑
Svart M, Gormsen LC, Hansen J, Zeidler D, Gejl M, Vang K, Aanerud J, Moeller N. Regional cerebral effects of ketone body infusion with 3-hydroxybutyrate in humans: reduced glucose uptake, unchanged oxygen consumption and increased blood flow by positron emission tomography. A randomized, controlled trial. PLoS ONE 2018 e0190556. (https://doi.org/10.1371/journal.pone.0190556)
- 31↑
Schmidt M. Rey Auditory Verbal Learning Test – A Handbook. Los Angeles, CA: Western Psychological Services, 2004.
- 32↑
Army Individual Test Battery. Manual of Directions and Scoring. Washington, DC: War Department, Adjutant General’s Office, 1944.
- 34↑
Wechsler D. Wechsler Adult Intelligence Scale III. San Antonio: The Psychological Corporation, 1997.
- 35↑
Borkowski JG, Benton AL, Spreen O. Word fluency and brain damage. Neuropsychologia 1967 135–140. (https://doi.org/10.1016/0028-3932(67)90015-2)
- 36↑
Hasselbalch SG, Madsen PL, Hageman LP, Olsen KS, Justesen N, Holm S, Paulson OB. Changes in cerebral blood flow and carbohydrate metabolism during acute hyperketonemia. American Journal of Physiology 1996 E746–E751. (https://doi.org/10.1152/ajpendo.1996.270.5.E746)
- 37↑
Henry RR, Brechtel G, Lim KH. Effects of ketone bodies on carbohydrate metabolism in non-insulin-dependent (type II) diabetes mellitus. Metabolism: Clinical and Experimental 1990 853–858. (https://doi.org/10.1016/0026-0495(90)90132-v)
- 38↑
Nilsson M, Gjedde A, Brock B, Gejl M, Rungby J. The effects of incretin hormones on cerebral glucose metabolism in health and disease. Neuropharmacology 2018 243–250. (https://doi.org/10.1016/j.neuropharm.2017.12.031)
- 39↑
Diamond A. Executive functions. Annual Review of Psychology 2013 135–168. (https://doi.org/10.1146/annurev-psych-113011-143750)
- 40↑
Takeuchi A, Matsushima E, Kato M, Konishi M, Izumiyama H, Murata Y, Hirata Y. Characteristics of neuropsychological functions in inpatients with poorly-controlled type 2 diabetes mellitus. Journal of Diabetes Investigation 2012 325–330. (https://doi.org/10.1111/j.2040-1124.2011.00170.x)
- 41↑
Jørgensen KF & Selskabet Danske Neuropsykologer. Danish Norms for Neuropsychological Tests (in Danish). Kbh: Dansk Psykologisk Forlag, 2012. 149 sider, illustreret + 1 cd-rom p.
- 42↑
Miskowiak KW, Burdick KE, Martinez-Aran A, Bonnin CM, Bowie CR, Carvalho AF, Gallagher P, Lafer B, Lopez-Jaramillo C & Sumiyoshi T et al.Methodological recommendations for cognition trials in bipolar disorder by the International Society for Bipolar Disorders Targeting Cognition Task Force. Bipolar Disorders 2017 614–626. (https://doi.org/10.1111/bdi.12534)
- 43↑
Nilsson M, Jensen N, Gejl M, Bergmann ML, Storgaard H, Zander M, Miskowiak K, Rungby J. Experimental non-severe hypoglycaemia substantially impairs cognitive function in type 2 diabetes: a randomised crossover trial. Diabetologia 2019 1948–1958. (https://doi.org/10.1007/s00125-019-4964-4)
- 44
Courchesne-Loyer A, Croteau E, Castellano CA, St-Pierre V, Hennebelle M, Cunnane SC. Inverse relationship between brain glucose and ketone metabolism in adults during short-term moderate dietary ketosis: a dual tracer quantitative positron emission tomography study. Journal of Cerebral Blood Flow and Metabolism 2017 2485–2493. (https://doi.org/10.1177/0271678X16669366)
- 45↑
Mikkelsen KH, Seifert T, Secher NH, Grondal T, van Hall G. Systemic, cerebral and skeletal muscle ketone body and energy metabolism during acute hyper-D-beta-hydroxybutyratemia in post-absorptive healthy males. Journal of Clinical Endocrinology and Metabolism 2015 636–643. (https://doi.org/10.1210/jc.2014-2608)
- 46↑
Ferrannini E, Baldi S, Frascerra S, Astiarraga B, Heise T, Bizzotto R, Mari A, Pieber TR, Muscelli E. Shift to fatty substrate utilization in response to sodium-glucose cotransporter 2 inhibition in subjects Without diabetes and patients With Type 2 diabetes. Diabetes 2016 1190–1195. (https://doi.org/10.2337/db15-1356)
- 47↑
Wium-Andersen IK, Osler M, Jorgensen MB, Rungby J, Wium-Andersen MKK. Antidiabetic medication and risk of dementia in patients with type 2 diabetes. A nested case-control study. European Journal of Endocrinology 2019 499–507. (https://doi.org/10.1530/EJE-19-0259)
- 48↑
Stubbs BJ, Cox PJ, Evans RD, Santer P, Miller JJ, Faull OK, Magor-Elliott S, Hiyama S, Stirling M, Clarke K. On the metabolism of exogenous ketones in humans. Frontiers in Physiology 2017 848. (https://doi.org/10.3389/fphys.2017.00848)
