Can COVID-19 result in cognitive dysfunctions? The need for a multidisciplinary approach in rehabilitation for post-COVID-19 people

In 2020, only a small number of studies investigated the cognitive sequelae of COVID-19. However, the majority of those used screening test that lacked high level of sensitivity to adequately answer this question.^{1, 2} The exception was one study that analyzed, with a complete neuropsychological battery, 35 post-COVID-19 patients suggesting an impact of SARS-CoV-2 on memory, attention and executive functions.³ In our study, we provide further data that can be used to identify the rehabilitation needs for this clinical population. An observational prospective cohort study was carried on in a rehabilitation unit during the pandemic period. Of 56 hospitalized post-COVID-19 patients, 23 were excluded: three patients had visual or hearing impairment, two patients were not Italian native-speaking, one patient was illiterate, ten patients had premorbid neurological diseases or pre-COVID-19 critical illness, and seven patients showed, in the COVID-19 acute phase, symptoms of stroke, seizures or impaired consciousness.

Neuropsychological profiles were evaluated using standardized performance test that investigated core areas of cognitive functioning (memory, attention, executive functions, language and visual-constructive abilities): 1) Mini Mental State Examination (MMSE); 2) Trail Making Test (TMT; part A and B); 3) Digit span (forward and backward); 4) Short-Story Memory Test (immediate and delayed recall); 5) Rey Auditory Verbal Learning Test (RAVLT-immediate and delayed recall); 6) Rey-Osterrieth Complex Figure Test (ROCF-copy and delayed recall); 7) Colored progressive matrices (PM47); and 8) Phonemic verbal fluency (PVF). Considering the possible effects of age, gender and educational level on test performances, the row scores were adjusted for all three according to published normative data. Furthermore, since critical and non-critical illnesses seem to be both associated with a greater decline in cognitive functioning in older adults,⁴ we decided to cluster the sample by age (<65 years of age and >65 years of age). For quantitative variables, D’Agostino-Pearson Test was used to evaluate their normal distribution (assuming P>0.05 as normal distribution). All variables were described accordingly to their distribution, by mean and standard deviation or by median and interquartile range. Chi-square test was used to analyze qualitative variable described as frequency and percentages. The significance level was set to a P<0.05.

Of 23 patients that were analyzed sixteen were males and seven were females (mean age 64±12; mean educational level 9±5) and fourteen of them were over 65 years of age.

For all 23 patients, the mean time of COVID-19 hospitalization was 42±14 days (min 21, max 67); 14 (61%) patients received invasive artificial ventilation and 5 (22%) non-invasive ventilation (Continuous Positive Airway Pressure [CPAP]). All the clinical and demographic characteristics examined showed a normal distribution.

Pathological scores were detected in MMSE (3; 13%), Digit Span forward (4; 17%), Digit Span Backward (1; 4%); TMT A (4; 17%), TMT B (10; 43%); Short-story memory test-IR (2; 9%); Short-story memory test–DR (2; 9%), ROCF-DR (3; 13%), ROCF-COPY (14; 61%), RAVLT-IR (2; 9%), RAVLT-DR (3; 13%), PVF (2; 9%), PM47(1; 4%). On average, our population performed abnormally on the ROCF copy (Table I).

Table I. —Neuropsychological scores of the study population.

Neuropsychological variables	Scores
MMSE	28 (19-30)§
TMT-A	40.4 (10.8-218.4)§
TMT-B	215.03 (163.52)#
TMT B-A	158.96 (137.06)#
Digit span forward	5.24 (1.08)#
Digit span backward	3.83 (0.8)#
Short-story memory test-IR	5.58 (0-14.16)§
Short-story memory test-DR	5.58 (0-7.82)§
RAVLT-IR	39.22 (14.23)#
RAVLT-DR	8.13 (4.21)#
ROCF-DR	13.53 (4.08)#
ROCF-COPY	26.42 (5.53)# *
PVF	30.57 (8.93)#
PM47	26.32 (4.16)#

^#Mean (standard deviation) if P>0.05 at D’Agostino-Pearson Test; ^§median (range);*pathological score. IR: Immediate recall; DR: delayed recall.

Considering the performances on all the neuropsychological test, 78% of our sample showed pathological scores in one or more test. If we observe the two age subgroups, 64% of over 65 years-old patients showed pathological performances in 2 or more test (N.=9); this percentage dropped to 22% in patients under 65 years of age (N.=2). Whereas, if we consider performance test scores that showed no impairment, percentages are reversed (7% >65 y vs. 44% <65 y). Altered performances, in one test only, were found in 33% of under 65 years-old patients and in 29% of over 65 years-old patients.

The data is in line with a recent systematic review and confirm the possibility of finding cognitive sequelae among this clinical population.²

If we consider that a high percentage of our patients (82%) were treated with mechanical ventilation and came from medical ICUs, these findings are consistent with previous studies that reported neuropsychological impairment to be a common outcome among acute lung injured and mechanically ventilated patients post discharge.⁵

In our sample, the most substantial result involved visuo-constructive abilities, considering that 61% of patients obtained a pathological score in ROCF copy. Post-COVID-19 patients showed major difficulties in coordinating fine motor skills, understanding visuospatial relationships, as well as planning the reproduction of complex geometric figures. These findings are in line with other studies that administered the ROCF test to critically ill patients and detected impaired performances.⁵

Since the ROCF test simultaneously evaluates many different cognitive functions, it could than serve as a sensitive tool in detecting alterations among this clinical population.

As expected, patients over 65 years of age were proven to be at a greater risk in developing a multidomain impairment,⁴ particularly in visuo-constructive, attention, and executive areas. However, surprisingly, global cognition seems to be altered in a small percentage of participants (3; 13%) which could mean that, in most cases, post-COVID-19 patients acquired new cognitive alterations, but not a global loss of cognitive functions. Moreover, our data also showed altered performances in under 65 years-old patients. This implies, in agreement with recent literature, that young adults could also experience mild cognitive deficits that may warrant a deeper investigation.²

These findings have important implications for clinical practice for two main reasons. Firstly, test such as ROCF (altered in 61%) or TMT (altered in 43%) investigated higher cognitive functions which are essential in conducting activities of daily living such as driving, resulting in risks to both individual and public safety. It is worth noting that in our sample we only included patients who needed additional rehabilitation sessions and were therefore hospitalized in our rehabilitation unit, but most patients that were discharged from acute COVID-19 units returned home without a neuropsychological evaluation or a cognitive rehabilitation. Secondly, since literature highlights the possibility of finding persistent symptoms,² the identification of the neuropsychological profile of COVID-19 patients may allow the development of adequate rehabilitation strategies and, consequently, the mitigation of possible adverse, long-lasting sequelae. For these reasons, our data should be considered as a warning to all health care professionals involved in post-COVID-19 patient management.

The main limitations of the study were due to the emergency situation, which hindered the possibility of collecting more data (including neuroimaging), having a large sample size and planning subsequent follow-up. These limitations do not allow us to draw generalizable conclusions, but data can serve as a trigger to further studies.