Patient‐reported olfactory recovery after SARS‐CoV‐2 infection: A 6‐month follow‐up study

Abstract

Long-term results about smell recovery after SARS-CoV-2 infection are missing and it remains unclear how many COVID-19+ patients may suffer from permanent smell impairment. The aim of this study was to assess olfactory dysfunction (OD) in a cohort of COVID-19+ patients to characterize long-term patterns of olfactory recovery and identify predictors of poor olfactory restoration. An electronic platform was created in February 2020 by our institutional infectious diseases unit to collect demographic and clinical information on COVID-19+ patients, including the OD onset. In September 2020, 126 COVID-19+ patients with OD were identified and sent via e-mail a modified version of the American Academy of Otolaryngology–Head and Neck Surgery COVID-19 Anosmia Reporting Tool for Clinicians.1 This observational, retrospective study was conducted in accordance with the World Medical Association Declaration of Helsinki and approved by the ethics committee at our institution. Statistical analysis was performed using SSPS version 19.0 for Windows (IBM Corp, Armonk, NY). The association between anosmia/hyposmia recovery and other variables was explored using the Student t test or Mann-Whitney U test for continuous variables and the chi-square test or Fisher's exact test for discrete variables. A multivariate analysis using a binary logistic regression model was performed, which included the variables found to be statistically significant in the univariate analysis. p < 0.05 was considered statistically significant. Confidence intervals were set at 95%. There were 110 patients included in the study. Their age (mean ± standard deviation) was 41.4 ± 12.3 years, male-to-female ratio was 0.57 (70 females and 40 males), and follow-up time was 6.1 ± 1.1 months. A complete recovery (CR) from anosmia/hyposmia was registered in 70 patients (63%; median recovery time, 7-14 days; range, 4-90 days) and a partial recovery (PR) in 24 patients (22%; median recovery time, 1-3 months; range, 15-180 months). The difference between median recovery time in the CR and PR groups was statistically significant (p < 0.0001). In 16 patients (14.5%) anosmia/hyposmia was unchanged. Overall, persistent OD was self-reported in 40 patients (36.5%). In 99 patients (90%), OD occurred after other symptoms (mainly malaise, fever, and cough). On univariate analysis, fever and headache occurring after OD onset and treatment with hydroxychloroquine were significantly associated with a CR pattern (p = 0.043, p = 0.025, and p = 0.042, respectively). Detailed survey responses and univariate analysis data are presented in Table 1. The multivariate analysis showed that the only factor significantly associated with anosmia/hyposmia recovery was cigarette smoking (p = 0.044; odds ratio confidence interval, 1.1-61.8). The post hoc sample calculation for cigarette smoking obtained a power of 72.9%, with the alpha set at 0.05. Table 2 presents results of the multivariate analysis. Sudden anosmia/hyposmia constitutes a symptom that, even alone, can drive towards self-isolation and testing for severe acute respiratory syndrome‒coronavirus-2 (SARS-CoV-2) infection.2, 3 The early spread of the virus in Italy allowed us to conduct our study with the longest follow-up available, as verified at the time of data analysis. Our results suggest that 2 possible smell restoration patterns exist after SARS-CoV-2 infection: full recovery may occur after few days, whereas a slow partial recovery may intervene after 1 to 3 months. Most of the available reports with shorter follow-up periods likely captured the first and fastest resolution pattern. Before our study, only the study by Konstantinidis et al demonstrated a similar dual-resolution pattern, albeit in a smaller sample and with a shorter follow-up.4 These trends may be related to different pathophysiologic mechanisms, which still need to be clarified.5 Since COVID-19‒related OD mostly affects young people, who have long life expectancy, data about smell recovery are required to support the management and counseling of upcoming infected patients. It is noteworthy that 14.5% of patients had no change in OD at long-term follow-up. Other studies employing validated psychophysical olfactory testing reported lower percentages of patients with persistent OD (ie, 6%), which highlights a possible discrepancy between assessment methods.6 Among the predictors identified on multivariate analysis, cigarette smoking was independently associated with persistent OD after 6 months. However, our sample size was insufficient to draw univocal conclusions on the role of smoking in smell recovery. Smoking is known to increase angiotensin-coverting enzyme-2 expression of the nasal epithelium and olfactory bulb, thus facilitating virus propagation at these sites.5, 7, 8 This mechanism may explain a more severe smell dysfunction and/or a poorer chance of recovery, but the present results do not allow for confirmation of this association. One of the strengths of this study is the length of the follow-up, which at the time of the data analysis was the longest among the existing literature.4, 9, 10 Moreover, a uniform and accessible self-reporting scale was used, allowing comparison with other ongoing and future studies. Among the main limitations of this study is the low statistical power of the multivariate analysis, which did not allow for a definitive characterization of the association between cigarette smoking and smell recovery. Other limitations include the retrospective design and the absence of objective olfactory assessment. At the time of the creation of the electronic platform, it was not possible to perform objective smell assessments, as the infection was active. Therefore, because patients were initially classified as anosmic/hyposmic on a subjective basis, the long-term follow-up was performed using a self-report questionnaire due to the absence of a baseline for comparison. This limitation is shared by most studies of OD in COVID-19+ patients. It is uncertain whether the recovery trend reached a plateau after 6 months, and further studies with longer follow-up times and larger sample sizes are needed for clarification.