Long-term cognitive effects of COVID-19 studied with repeated neuropsychological testing

Abstract

COVID-19 has a wide range of clinical manifestations. Apart from respiratory symptoms, neurological manifestations and cognitive impairment are commonly seen. Due to lack of a clear consensus on the definition of long COVID a pragmatic approach of assessing change from neurological baseline has become prevalent. We present the case of a high functioning woman in her late 40s who had a mild course of acute illness but developed cognitive difficulties. No organic causes for her cognitive difficulties were found. Neuropsychological testing showed impairment in right frontal lobe functioning. The patient showed a response to empirical treatment with Attention deficit hyperactivity disorder (ADHD) medications. Repeat neuropsychological testing 1 year later showed continued dysfunction in executive functioning. We present the details of her test findings and discuss the proposed pathophysiology behind neurocognitive effects of long COVID.

Background

COVID-19 has a wide range of clinical manifestations. Apart from respiratory symptoms, neurological manifestations and cognitive impairment are commonly seen with greater risk in hospitalised and severely ill patients, but often seen in mild disease or initially asymptomatic individuals.¹ The affected domains most described in the studies are executive functions, attention, episodic memory,² processing speed, category fluency, memory encoding, and recall,³ phonemic fluency, semantic verbal fluency, set shifting, inhibitory control, and working memory.⁴

The pathological mechanisms hypothesised range from the direct effects of cellular damage due to viral invasion, secondary inflammatory responses, decreased angiotensin-converting enzyme 2 (ACE2) activity that regulates neuroprotective and neuro-immunomodulatory functions; oxidative stress; and hypoxia, sepsis and/or multiorgan damage related to severe COVID-19.⁴ Vascular involvement may also underlie the cognitive changes with studies reporting elevated d-dimer levels.⁵ Studies suggest some patients may be genetically predisposed to a COVID-19-triggered autoimmune response.⁶ This association is significant since the same allele confers a higher risk of sporadic Alzheimer’s disease.⁷

The term ‘long COVID’ is an initial social media description with the literature since using many other terms including ‘long-haul’, ‘post-COVID’, ‘post-acute COVID syndrome (PACS)’ and ‘post-acute sequelae of SARS-CoV-2 (PASC)’ to refer to persistent symptoms and/or delayed or long-term complications beyond acute COVID-19.⁸ The WHO suggested the following clinical case definition ‘post COVID-19 occurs in individuals with a history of probable or confirmed SARS-CoV-2 infection, usually 3 months from the onset of COVID-19 with symptoms that last for at least 2 months and cannot be explained by an alternative diagnosis’.⁹ The International Classification of Diseases, Tenth Revision refers to it under the ‘Post-COVID conditions, unspecified’ code.¹⁰

Case presentation

The patient is a female in her late 40s with a high performing job who contracted COVID-19 infection with a mild course of the initial acute illness which was diagnosed after a positive PCR test. Her primary symptoms included non-specific aches and pains. She did not have a fever, respiratory symptoms nor did she require hospitalisation and recovered with at home supportive care.

Post recovery from the infection she reported having a reduced sense of taste and smell, fatigue, decreased attention, decreased ability to multitask and memory problems such as retrieval difficulties. She complained of forgetfulness and difficulty with focus. She developed anxiety about these symptoms and noticed that these symptoms made it harder to perform her usual tasks. She also noticed changes in her language like substituting irrelevant words, having trouble thinking of a word she wants to say and finding that her thoughts go blank. She described this state as being ‘foggy all the time.’

Physical examination at 3-month follow-up with her primary care physician did not show any abnormalities and a diagnosis of post COVID-19 syndrome was considered. She has a history of taking citalopram 10 mg daily for 5–6 years on which her depression has been well controlled. She has no history of cognitive, neurological or relevant medical illness. There is no family history of early-onset cognitive disease. Laboratory studies to rule out reversible causes like thyroid dysfunction or vitamin deficiency were normal. MRI brain did not show any relevant abnormalities and evaluation by a neurologist ruled out any neurological disorders. In view of her significant dysfunction, she was referred for neuropsychological testing and a complete psychodiagnostic interview was completed 4 months post infection.

Investigations

Immediate neuropsychological testing results

Patient’s full-scale intellectual functioning score was in the average range. She performed in the borderline or moderately impaired range on the Boston Naming Test, a test of confrontational naming. The patient performed in the low average range on the New Learning trial of the Rey Auditory Verbal Learning Test. She performed in the very poor range of functioning on the 5-minute Verbal Recall (short-term memory for words) and in the low average range of functioning at 30 min (long-term memory for words). She was given the Rey Complex Figure Test, a test of visuospatial, executive functioning and graphomotor praxis. She was able to perform in the average range on the copy, 5-min and 30-min visual recall and recognition. The patient performed in the average range on measures of visual integrative skills and visual problem-solving. She denied anxiety and depression at the time of this testing.

Of particular interest was her performance on measures of executive functioning: Trail Making Test, Verbal Fluency Test, Design Fluency Test and Color Word Interference. The four measures are part of a battery of tests: Delis-Kaplan Executive Functioning System (DKEFS). Each of the four tests had subtests. There was statistically significant scatter in the scores within each test. Trail Making Test scores had statistically significant scatter with Visual Scanning, Number Sequencing and Letter Sequencing being statistically stronger than Number Letter Switching and Motor Speed. Verbal Fluency Test scores had statistically significant scatter with Category Switching Total Correct Responses and Category Switching Total Switching Accuracy being statistically stronger than Letter Fluency and Category Fluency. Design Fluency Test had statistically significant scatter with Switching Total Correct being statistically stronger than Filled Dots and Empty Dots Only. Color Word Interference Test scores also had statistically significant scatter with Inhibition Switching being statistically stronger than Color Naming and Inhibition. When we compare Word Reading to Inhibition the Word Reading is statistically stronger.

Immediate testing analysis

Post COVID-19 symptoms expressed by the patient were exhibited on neuropsychological assessment also. Testing was remarkable for fluctuations in executive functioning skills. The aetiology of the change in functioning is likely manifested in the frontal lobe, particularly the right frontal lobe.

Differential diagnosis

The possibility of a depressive episode or anxiety disorders was considered but the patient did not meet criteria for any psychiatric disorders on evaluation. She did mention an increase in anxiety related to her cognitive issues postinfection, but they did not meet severity for a clinical diagnosis. Scores on Beck Depression Inventory and Zung Anxiety Self-Assessment at initial visit and 1-year follow-up were within normal limits. The patient did not endorse pain symptoms, malaise or fatigue ruling out disorders like chronic fatigue syndrome and fibromyalgia which may be associated with attentional difficulties.

Treatment

The patient was started on lisdexamfetamine, a prodrug of the dextroamphetamine, to help with the cognitive difficulties by her primary care provider. Amphetamines, which are stimulant medications used commonly in the treatment of ADHD, are thought to regulate the reuptake and release of norepinephrine and dopamine into the extraneuronal space and help with attention and focus.¹¹ Lisdexamfetamine was used off-label to help with these cognitive symptoms initially at a dose of 30 mg daily which was then titrated up to 40 mg daily taken as needed on days when sustained focus was needed. The medication was started after the initial neuropsychological testing and she had not taken lisdexamfetamine on either day of testing and is thereby unlikely to affect the test results. She reported subjective improvement in attention and concentration when taking the medication. She took the medication only on the days that she needed to do tasks requiring sustained attention or on days that she anticipated having difficulty coping with the tasks at hand. She herself created a chart system for cognitive rehabilitation (for short term memory), where she used a stenographer notebook to write down daily tasks/goals. She then follows up directly across the column and turns the page over to make a list of tasks to do for the next day. She also began a workout programme and increased her daily fluid intake.

Outcome and follow-up

One-year postinfection the patient continued to have forgetfulness, but she reported some improvement. She continued to have difficulties doing more than one task at a time.

The patient continued to be employed in the same profession at the time of re-evaluation 1 year later. She was transferred to a new unit in the company. She had to learn about a new product in the company which she was able to do using her chart system as a coping strategy. She was eventually promoted to a supervisory role in her new unit despite ongoing cognitive difficulties.

Follow-up neuropsychological testing results 1 year later

The performance on the DKEFS subtests of Number Sequencing, Motor Speed, Letter Fluency, Category Fluency and all Design Fluency subtests had statistically significant changes from initial testing to follow-up. Number Sequencing decreased by 1 SD (13–10). This is a measure of speed of processing and digit recognition as well as sequencing. However, Motor Speed, with no other distractor, improved by 5 points. She has more difficulty with performing a task that requires two stimuli rather than one. Letter Fluency and Category Fluency increased (12–17 and 14–19). It appears as though she is able to generate words with more accuracy at 1-year follow-up. This is typically referred to as a function of the left frontal temporal region of the brain. Of particular concern, were the decreases on all subtest scores in Design Fluency (see online supplemental material for table of test scores). Design Fluency is typically thought of as a function of the right frontal lobe and showed statistically significant decreases across all subtests and the composite.

The performance on Color Word Interference was consistent between the 2 years. There were statistically significant differences in scores on the follow-up subtests of the DKEFS. With respect to tests of dementia, 3 scores improved: Boston Naming Test, Rey Auditory Verbal Learning New Learning and 5-minute Verbal Recall (short-term verbal memory). All other scores were relatively the same as in initial assessment results.

Implications of follow-up testing results

Design fluency measures the following skills: motor speed, visual attention, constructional skills, visual-perceptual skills, executive functioning skills including fluency in generating visual patterns, initiation of problem solving, creativity in drawing new designs, and a combined effort of observing the rules of the current task while drawing a novel or new design each time. Each subtest includes a new demand including inhibition of responses and simultaneously engaging in cognitive shifting.

The statistically significant decreased performances on all four subtests from initial to 1-year testing indicates that the patient’s performance had decreased in skills such as basic visual skills, motor speed, visual perceptual skills, constructional skills along with executive functioning skills in the areas of initiation of problem solving, generating visual patterns, creativity, combined behaviours of drawing designs while following a new set of rules on each of the four tasks; including the skills of initiation and cognitive shifting.

The decrease in her ability to fluently produce designs under time constraints with specific rules could be relevant to her need to rely on physical reminders (using a journal, lists and calendar). She is not able to function as efficiently as in the past without compensatory devices. Therefore, she used problem solving and created her own rehabilitative tool to perform more efficiently at her job. She also trained staff in using her techniques, per the request of her supervisor.

Since the patient was not taking lisdexamfetamine at initial testing, and taking it only on needed basis and had not taken it on the day of follow-up testing, we can conclude that it did not affect the test results. Biologically, we would hypothesise that the inflammatory process has centred more in the right frontal lobe as reflected by her performance from initial to 1-year follow-up testing. The decrease in scores might reflect that long-term effects of COVID-19 might progress over time.

Discussion

The biology of ‘long COVID’ is not well characterised. Hyposmia, mood disorders, cognitive impairment, sleep disorders and dysautonomia have been linked to dysfunction of ‘ACE2-rich’ brain areas.¹² Neuroimaging studies using 18F-FDG (fluorodeoxyglucose) brain positron emission tomography in patients with ‘long- COVID’ show hypermetabolism in the orbital gyrus, olfactory gyrus, temporal lobe (including hippocampus and amygdala), hypothalamus, thalamus, brainstem and cerebellum,^{13 14} and bear similarities with disorders like chronic fatigue syndrome and the postural orthostatic tachycardia syndrome.⁸ Kanberg and colleagues showed that the changes caused by CNS injury from COVID-19 attenuate at 6 months despite continuing neurocognitive symptoms which appear to outlast the injury¹⁵ which is similar to the presentation in this case.

There are no reliable data to facilitate identification of high-risk patients for long COVID in clinical practice. Discordant findings are reported with some studies reporting pre-existing neurological morbidity, increasing age and female gender as risk factors while other studies reporting the opposite.⁸ Just as earlier published case reports have added to the understanding of long COVID syndrome and its mental and physical aspects,^{16 17} this case adds to the growing literature on the cognitive impairments of long COVID. Reports on the neurocognitive effects have underlined memory deficits in non-hospitalised patients,¹⁸ but there are no reports to our understanding where a patient was followed up with repeat full neuropsychological testing as seen in the current report. Our case also highlights the involvement of the frontal lobe and impairment of executive functioning which is be additionally distressing to individuals in their occupational setting.

Some case reports have shown initial indication that different classes of medications used to treat ADHD might be beneficial in patients with long COVID,^{19 20} which is in keeping with the response to attention deficits seen in our patient. However, further studies are needed to determine the efficacy of these treatment options. This patient showed benefit from using lisdexamfetamine in addition to using coping strategies. Regular exercise, dietary changes, and therapy also have a role in mitigating symptoms and help recovery.

In summary, cognitive deficits were seen even in the case of mild COVID-19 illness and deficits persisted up to a year of follow-up. These deficits are evident on neuropsychological testing and caused significant functional dysfunction. Treatment with stimulant medications, and use of coping strategies helped mitigate some of the deficits and improve functioning.

Learning points.

• Even mild COVID-19 illness may be associated with dysfunction in the frontal lobe, particularly right frontal lobe, evident immediately after COVID-19 and persisting at retesting 1 year later.

• The lasting nature of neurocognitive deficits post COVID-19 and subtle changes in functioning might be missed or attributed to other causes.

• Providers need to consider this possibility when evaluating patients presenting with cognitive difficulties, as history of COVID-19 infection in the past may not be readily offered or considered relevant by the patient especially with mild illness.

• Individuals may need to learn new coping strategies to deal with the change or require appropriate accommodations.

• Psychostimulant medications like those used to treat ADHD may be beneficial in mitigating some of the symptoms.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).