Acute Kidney Injury Is Associated With Subtle but Quantifiable Neurocognitive Impairments

Jessica A. Vanderlinden; Joanna S. Semrau; Samuel A. Silver; Rachel M. Holden; Stephen H. Scott; J. Gordon Boyd

Disclosures

Nephrol Dial Transplant. 2022;37(2):285-297.

Abstract and Introduction

Abstract

Graphical Abstract

(Enlarge Image)

Background: Acute kidney injury (AKI) is associated with long-term morbidity and mortality. The effects of AKI on neurocognitive functioning remain unknown. Our objective was to quantify neurocognitive impairment after an episode of AKI.

Methods: Survivors of AKI were compared with age-matched controls, as well as a convenience sample of patients matched for cardiovascular risk factors with normal kidney function (active control group). Patients with AKI completed two assessments, while the active control group completed one assessment. The assessment included a standardized test: the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), and a robotic assessment: Kinarm.

Results: The cohort consisted of 21 patients with AKI, 16 of whom completed both assessments, and 21 active control patients. The majority of patients with AKI had Kidney Disease: Improving Global Outcomes Stage 3 AKI (86%), 57% received dialysis and 43% recovered to ≤25% of their baseline serum creatinine by their first assessment. Compared with the RBANS, which detected little impairment, the Kinarm categorized patients as impaired in visuomotor (10/21, 48%), attention (10/20, 50%) and executive tasks (11/21, 52%) compared with healthy controls. Additionally, patients with AKI performed significantly worse in attention and visuomotor domains when compared with the active controls. Neurocognitive performance was generally not impacted by the need for dialysis or whether kidney function recovered.

Conclusions: Robotic technology identified quantifiable neurocognitive impairment in survivors of AKI. Deficits were noted particularly in attention, visuomotor and executive domains. Further investigation into the downstream health consequences of these neurocognitive impairments is warranted.

Introduction

Acute kidney injury (AKI) defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria affects one in five hospitalized adults.^[1] AKI is particularly prevalent in patients who undergo cardiovascular surgery,^[2] or who are critically ill.^[3] Regardless of the cause of AKI, patients who experience an episode of AKI have an increased risk of morbidity and mortality.^[2–4]

AKI often occurs in the context of multi-organ dysfunction, either due to the inciting injury affecting multiple organs simultaneously, or as an indirect result of kidney dysfunction.^[5] The brain may be particularly vulnerable to dysfunction in the context of AKI. It has been hypothesized that inflammation, uremia, cytotoxicity, osmolality disturbances and alterations in blood–brain barrier permeability are possible mechanisms underlying cerebral dysfunction in patients with AKI.^[6] These pathophysiological processes may lead to impaired neurocognitive performance in individuals with AKI.

The neurocognitive outcomes for survivors of AKI have not been extensively studied. Using administrative data, two retrospective studies have demonstrated that a diagnosis of AKI was an independent risk factor for dementia.^[7,8] More recently, AKI has been shown to be associated with long-term neurocognitive impairment in children with severe malaria.^[9] In patients with chronic kidney disease (CKD), the high burden of neurocognitive impairment is well-described, and affects the domains of attention, memory and executive function.^[10] We also recently demonstrated that robotic technology can precisely quantify impairments in visuomotor function in patients with CKD.^[11] In contrast, there are no prospective data on adult neurocognitive impairment after AKI.

The objective of this prospective observational study was to use robotic technology (Kinarm) to quantify neurocognitive impairment in patients who recently experienced an episode of AKI. In this study, neurocognitive performance was compared with both healthy controls, as well as an active control group: patients matched for cardiovascular risk factors (but without any history of kidney disease). We hypothesize that Kinarm will be able to identify impairments in patients with AKI that are not detected using a traditional clinical assessment tool, and that patients with AKI will perform worse than patients without kidney disease.

1 2 3 4

Next Section

Table 1. Patient characteristics
Patient characteristics	AKI (n = 21)	Active control (n = 21)	P-value
Age, mean (range), years	71.4 (57–84)	63.9 (53–79)	<0.01
Sex, n (%)
Male	16 (76.2)	19 (90.5)	>0.05
Handedness, n (%)
Right	20 (95.2)	20 (95.2)	>0.05
Ethnicity, n (%)
Caucasian	21 (100)	21 (100)	>0.05
Highest reported education, n (%)
High school	7 (33.3)	9 (42.9)	>0.05
College or university	13 (61.9)	8 (38.1)	–
Masters	1 (4.8)	3 (14.3)	–
Unknown	0 (0)	1 (4.8)	–
Comorbidities, n
Diabetes	9	5	>0.05
Hypertension	8	12	>0.05
Cardiovascular^a	21	21	–
Cancer	7	2	–
Other	26	22	–
Cause of AKI, n (%) of totalN
Cardiovascular^f	4 (19.0)	–	–
Infection^g	6 (28.6)	–	–
ATN	3 (14.3)	–	–
Rhabdomyolysis	4 (19.0)	–	–
Obstruction^h	3 (14.3)	–	–
Otherⁱ	8 (38.1)	–	–
KDIGO AKI stage, n (%)
Stage 1	1 (4.8)	–	–
Stage 2	2 (9.5)	–	–
Stage 3	18 (85.7)	–	–
Recovered ≤25% of kidney function, n (%)
Yes	9 (42.9)	–	–
No	12 (57.1)	–	–
Baseline parameters before incident AKI
Baseline Cr, mean (range), μmol/L^j	104.1 (67–198)	–	–
Existing kidney disease, n (%)
Yes	6 (28.6)	0 (0)	<0.01
No	14 (66.6)	21 (100)	–
Unknown	1 (4.8)	0 (0)	–
Hospital AKI data
Highest hospital Cr, mean (range), μmol/L	715.5 (271–1762)	–	–
Required dialysis
Yes	12 (57.1)	–	–
No	9 (42.9)	–	–
First cognitive testing (n = 21)
Months between AKI incident and test 1, mean (range)	6.8 (2–15)	–	–
Cr, mean (range), μmol/L	139.2 (68–361)	81.3 (65–111)	<0.001
eGFR, mean (range), mL/min/1.73 m²	47.7 (10–98)	84.9 (59–104)	<0.0001
Second cognitive testing (n = 16)
Months between tests 1 and 2, mean (range)^j	4.6 (4–14)	–	–
Cr, mean (range), μmol/L^j	151.5 (76–338)	–	–
eGFR, mean (range), mL/min/1.73 m^{2 j}	41.4 (11–93)	–	–

Summary of baseline characteristics.
^aAKI cardiovascular comorbidities: coronary artery disease (n = 7), atrial fibrillation (n = 4), peripheral vascular disease (n = 3), pacemaker (n = 2), heart failure (n = 2), deep vein thrombosis (n = 1), transcatheter aortic valve implantation (n = 1) and aortic valve replacement (n = 1).
^bAKI cohort cancer diagnoses: ovarian (n = 2), prostate (n = 2), colon (n = 2), gastric (n = 1) and lung (n = 1).
^cCardiac cohort cancer diagnoses: ovarian (n = 1) and prostate (n = 1).
^dAKI cohort other comorbidities: sleep apnea (n = 6), dyslipidemia (n = 4), chronic obstructive lung disease (COPD) (n = 4), gout (n = 3), hypothyroidism (n = 3), osteoarthritis (n = 3), gastroesophageal reflux disease (GERD) (n = 2) and achondroplasia (n = 1).
^eCardiac cohort other comorbidities: dyslipidemia (n = 7), hypothyroidism (n = 3), sleep apnea (n = 3), osteoarthritis (n = 3), GERD (N = 2), COPD (n = 2), kidney stones (n = 1) and Gilbert Syndrome (n = 1).
^fCardiovascular causes of AKI: cardiac surgery (n = 3) and heart failure (n = 1).
^gInfection causes of AKI: sepsis (n = 4), viral gastritis (n = 1) and infection unspecified (n = 1).
^hObstruction cause of AKI: obstructive neuropathy (n = 1), bladder obstruction (n = 1) and obstruction unspecified (n = 1).
ⁱOther cause of AKI: pneumonia (n = 4), contracted induced (n = 2), hepatorenal (n = 1) and hyperkalemia (n = 1).
^jIndicates excluded values that were unknown or not available at the time of assessment, as one patient's pre-AKI Cr was unknown at the time of the AKI incident and five patients did not complete their second assessment. Significant differences between the AKI and cardiac cohort are in bold type. Note the cause of AKI and patient comorbidities were multifactorial for the majority of patients. Therefore if AKI was caused by more than one attribute (e.g. infection and obstruction), or the patient had more than one comorbidity in each section (e.g. dyslipidemia and gout), it was accounted for in both sections or in the same section depending on the classification.

Table 2. Neurocognitive battery tasks, task components and descriptions, domains assessed and test performance
Neurocognitive test	Task/domain	Task components/task description	Domain assessed	AKI test 1 (no. impaired, %)	AKI, no. completed test 1 (max n = 21)	AKI test 2 (no. impaired, %)	AKI, no. completed test 2 (max n = 16)	Active control (no. impaired, %)
RBANS	Attention	Digit span, symbol coding	Attention	2 (10.0)	20	0 (0)	16	0 (0)
	Immediate memory	List learning, story memory	Memory	1 (4.8)	21	0 (0)	16	1 (4.8)
	Delayed memory	List recall, list recognition, story recall, figure recall	Memory	2 (10.0)	20	1 (6.3)	16	1 (4.8)
	Visuospatial	Figure copy, line orientation	Visuospatial	3 (15.0)	20	2 (12.5)	16	2 (9.5)
	Language	Picture naming, semantic fluency	Language	0 (0)	20	0 (0)	16	0 (0)
	Total scale score	A summary score based on the performance on all the RBANS domains, representative of global performance on the RBANS	Composite performance score	1 (5.0)	20	0 (0)	16	0 (0)
Kinarm	Arm position matching	Patients are asked to mirror match the arm position the robot has moved one arm to with their other arm	Proprioception, Somatosensory	1 (5.6)	18	1 (7.7)	13	1 (4.8)
	Ball on bar	A bar connects the patients two arms with a ball on top of it. Patients are then instructed to place this ball in various targets on the screen. As the task progresses, the ball becomes less fixed to the bar increasing the task's difficulty	Visuomotor skill, motor range and coordination	3 (13.0)	13	2 (16.7)	12	2 (10%)^a
	Object hit (OH)	The patient's hands become paddles and they are instructed to hit as many falling balls away from them as possible. As the task progresses the quantity and speed at which the balls fall increases	Attention, visuomotor	4 (25.0)	16	4 (30.8)	13	1 (4.8)
	OH and avoid	Similar to OH, but this time the patients must only hit two specific shapes while avoiding all other shapes. As the task progresses the quantity and speed at which the targets fall increases as in the OH task	Attention, visuomotor and executive Function	4 (23.5)	17	3 (25.0)	12	1 (4.8)
	Visually guided reaching (VGR)	The patient's hand is represented by a dot. They are then instructed to move this dot as quickly and accurately as possible into various targets on the screen	Visuomotor	10 (47.6)	21	5 (31.3)	16	2 (9.5)
	Reverse VGR	Similar toVGR, but the light representing the hand now moves in the opposite direction of the arm movement	Visuomotor, executive function	11 (52.4)	21	8 (50.0)	16	5 (23.8)
	Spatial span	Patients are shown a sequence in a 3 × 4 box matrix and asked to replicate it. The length of the sequence depends on if previous response was correct or not	Working memory	1 (6.7)	15	0 (0)	12	1 (4.8)
	Trail making test A (TMT-A)	The patient is asked to connect the numbers 1–25 in order as quickly as possible	Attention, visuomotor	10 (50.0)	20	6 (40.0)	15	2 (9.5)
	Trail making test B	Similar to TMT-A, but the patient is now asked to alternate between number and letters (1-A-2-B-3) as quickly as possible	Attention, visuomotor, Executive function	5 (25.0)	20	6 (40.0)	15	1 (4.8)

The neurocognitive battery performed by patients, which comprised: the RBANS and Kinarm. The table also depicts the domains assessed and tasks that make up these domains for the RBANS and the Kinarm task descriptions. Test performance and number of patients who completed the task at each time point are also included, along with the active control group. Patients did not complete tasks mainly due to time constraints. The number of impaired was determined by a score outside the 95% cut off healthy normative data (RBANS: <75.25, Kinarm: >1.96).
^aOne patient did not complete the ball on bar task in the active control group, making the n value for this task 20. All other task for the active control group had an n value of 21.

processing....

Acute Kidney Injury Is Associated With Subtle but Quantifiable Neurocognitive Impairments

Abstract and Introduction

Abstract

Introduction

Tables

Tables

References

Authors and Disclosures

Authors and Disclosures

Sidebar

Sidebar

Key Learning Points

Most Popular Articles

Email This

Feedback