Self-reported cognitive and affective complaints associated with olfactory loss in an online survey of individuals with COVID-19

Abstract

Purpose: The symptomatology associated with COVID-19 is very diverse, ranging from flu-like symptoms to those affecting olfaction, cognition or mood for long periods. The present study explored the associations between self-reported olfactory deficits and cognitive and emotional complaints in a large-scale online survey conducted among individuals who had COVID-19.

Methods: Two complementary online studies were set up, one in France and the other in Mexico, involving 3108 and 364 volunteers respectively, to investigate the link between olfactory loss in COVID-19 and self-reported cognitive and emotional changes. Cognitive and affective complaints were assessed using simple yes/no items inspired by previously published studies, but not based on standardized clinical questionnaires.

Results: A first result was that cognitive difficulties are more frequent in COVID-19 individuals with long-standing olfactory disorders than in patients who have recently developed olfactory disorders. In addition, we also showed that the prevalence of cognitive difficulties is higher in COVID-19 patients with olfactory disorders than in those without. Furthermore, cognitive difficulties in patients with long-term olfactory disorders are more strongly associated with memory difficulties than with attention difficulties. Finally, mood disorders were more frequent in COVID-19 participants with olfactory loss than in those without.

Conclusion: Taken together, these data suggest that in COVID-19, the duration of olfactory loss is a key factor, strongly associated with cognitive and affective impairment. These data should provide us with further guidance on the management of people affected, which should not simply be unimodal, targeting just one category of symptoms.

Keywords: Anosmia; Cognition; Mood; Olfaction; Symptoms; long-COVID.

Fig. 1

(A) Pie chart representing the percentage of participants reporting at least 1 cognitive difficulty (29%, red portion) or no cognitive difficulty (71%, blue portion). The other panels illustrate the significant variables affecting the probability to report at least 1 cognitive difficulty. (B) Odds-ratios (OR) and 99% confidence intervals of the significant variables. Note that for continuous variables, OR are given for each standard deviation of the corresponding variable. (C) Effect of OD duration. In C, the black points represent raw data, while for illustrative purposes the red points represent the probability for participants binned per week; the size of black and red points vary according to their sample size

Fig. 2

Significant variables related to the number of reported cognitive difficulties. (A) Model estimate (for 1 standard deviation) and (B) Effects of OD duration on the number of reported cognitive difficulties. In B, circle size is proportional to the number of participants, the blue line and the grey shaded area display the prediction from the model and its associated 95% confidence interval, respectively

Fig. 3

Blue and red lines correspond to items dealing with memories and with attention, respectively. Plain and dashed lines correspond, respectively, to a significant and a non-significant effect of OD duration on the probability (estimated from a logistic regression) to report a cognitive difficulty. Shaded areas around the lines represent their associated 95% confidence intervals

Fig. 4

Pie chart presenting the proportion of participants who reported at least 1 cognitive difficulty (red portion) or no cognitive difficulty (blue portion) as a function of virus strains. In the top chart, participants had one of the variants of SARS-CoV-2 with associated OD. In the chart on the left, participants had the original strain of SARS-CoV-2 and at the same time period than for the top chart but were much rarer. In the chart on the right, participants had the original strain of the SARS-CoV-2 with associated OD. They were matched by propensity score matching to the characteristics (according to age, gender, OD type, OD onset speed, OD constancy, OD duration, month of COVID-19 onset, smoking status and BMI) of the participants in the top chart

Fig. 5

Effect of the presence of OD (left, pink bars: with OD; right, blue bars: without OD) on the simultaneous presence of either mood disorder or cognitive difficulty