Persistence of Long-COVID symptoms in a heterogenous prospective cohort.

Dear editor,

We read with interest the article by Fang and colleagues, showing the results of a multicenter prospective cohort study related to COVID-19 sequelae one year after acute illness, as well as potential risk factors.

We present our data collected from May 11, 2020 to September 24, 2021 at the “Luigi Sacco” University Hospital, Milan, where the ARCOVID (Ambulatorio Rivalutazione COVID) outpatients’ clinic began to follow “long haulers”, periodically controlling the persistence of physical and psychological symptoms, in order to assess their duration and the predictive factors associated with their resolution. Patients aged >18 years with confirmed COVID-19 (by PCR/antibody detection) were either referred by the physicians who had taken care of them in the acute phase or came voluntarily. After signing written informed consent they were enrolled in the AntiCROWN longitudinal study of anti-S1/S2 IgG response and clinical follow-up, approved by the “Comitato Etico Interaziendale Area 1″, n. 2020/ST/158. Throughout the first visit patients received a standardized clinical examination, serological sample to detect anti-S1/S2 IgG levels, 6-minute-walk test in case of dyspnea, and, if necessary, blood work and then sometimes were referred to other specialists. Moreover, they were asked about ongoing symptoms. The follow-up continued using questionnaires sent to each patient every 3 months.

Descriptive statistics included absolute frequencies and percentages for categorical variables and medians with interquartile range [IQR] for continuous variables. Patients were grouped according to the WHO classification of severity of SARS-CoV-2 infection.

Demographic and clinical characteristics of subjects in the four groups were compared using the χ2 test or Fisher's exact test where appropriate for categorical variables and the nonparametric Kruskal-Wallis test for continuous variables. With regard to the outcome of symptom resolution in the acute phase of infection, a survival analysis was conducted using Kaplan-Meier curves and univariable and multivariable Cox regression models to identify factors associated with symptom disappearance. R software v.3.6.2 was used for statistical analysis. A p-value <0.05 was considered statistically significant.

We enrolled 1168 patients, 49,7% females and 50,3% males, median age 57 years: 41% were mild, 26% moderate, 11% severe and 22% critical based on WHO COVID-19 disease severity classification. Of our patients 58,3% had been discharged from an hospital after acute COVID-19. Overweight patients were 59%, the median BMI being 25.47. Regarding comorbidities 37.8% were cardiovascular disease, 26,1% metabolic disorders, 11.4% pulmonary disease, 10,2% diabetes, 6,2% immunological diseases, 4,3% renal diseases, 3,7% autoimmune diseases, 2,7% cancer, 1,7% liver disease. Table 1 shows the distribution of the main Long-COVID symptoms among the different WHO severity groups.

Table 1

Long-COVID symptoms’ prevalence according to the WHO severity score.

Symptoms, n (%)	WHO COVID-19 disease severity score					p-value
	Overall	Mild	Moderate	Severe	Critical
Dyspnea,	428 (36.6)	118 (24.9)	132 (43.7)	57 (43.5)	121 (46.4)	<0.001
Telogen effluvium	117 (10.0)	50 (10.5)	34 (11.3)	14 (10.7)	19 (7.3)	0,407
Fatigue	624 (53.4)	224 (47.3)	181 (59.9)	79 (60.3)	140 (53.6)	0,002
Myalgia and arthralgia	426 (36.5)	161 (34.0)	112 (37.1)	50 (38.2)	103 (39.5)	0,474
Palpitations	147 (12.6)	67 (14.1)	46 (15.2)	15 (11.5)	19 (7.3)	0,02
Anosmia	261 (22.3)	141 (29.7)	63 (20.9)	24 (18.3)	33 (12.6)	<0.001
Ageusia	246 (21.1)	130 (27.4)	55 (18.2)	25 (19.1)	36 (13.8)	<0.001
Amnesia	168 (14.4)	62 (13.1)	54 (17.9)	14 (10.7)	38 (14.6)	0,164
Headache	113 (9.7)	57 (12.0)	31 (10.3)	7 (5.3)	18 (6.9)	0,041
Anxiety and panic attack	121 (10.4)	52 (11.0)	36 (11.9)	12 (9.2)	21 (8.0)	0,442
Insomnia	123 (10.5)	46 (9.7)	34 (11.3)	10 (7.6)	33 (12.6)	0,405

Fig. 1 shows the decay curves and estimated median time to resolution for the main symptoms with the multivariate analysis of predictive factors of persistence. Ageusia and anosmia displayed similar curves, steeper in the first 100 days, with a flattening of the slope thereafter and a median time to resolution of 155 and 200 days, respectively, 46% and 48% of still reporting persistent symptoms at 300 days. Baseline antibody production was associated with protracted ageusia.

Fig. 1

Decay curves and median decay rate of the main Long-COVID symptoms and multivariate analysis of associated risk factors.

Serological sample: anti-spike IgG (AU/mL).

Palpitations are disproportionate accelerations of the heartbeat during exercise or at rest that resolve slowly. Their estimated median time to resolution (95% CI) was 425 days (282–489). The estimated median time to resolution of anxiety and panic attacks was 391 days (313-NA). The median time to resolution of headache was estimated to be 379 days (264–505). For amnesia and insomnia the slow decay rate currently makes it impossible to foresee a median time of resolution. After one year less than 20% of our patients have resolved memory problems, and only about 15% have resolved insomnia. After median 35 days beyond the onset of COVID-19 phase of the disease 17.3% of the population reported telogen effluvium, the only truly post-COVID symptom, which had a median time to resolution of 299 days.

Researchers have initially approached Long-COVID by phone calls performed 60 days after discharge, which revealed the persistence of at least one symptom in 66–100% of subjects, according to disease severity, , with implications of job loss and mental health impact, A review of such short-term evaluations lead to a comprehensive description of the frequency of each Long-COVID symptom. Subsequently, clinical cohorts provided medical visits, physical examination and questionnaire-based follow-up. Huang et al., on a cohort of 1655 patients discharged from hospital reported the persistence at six months of fatigue in 63%, sleep difficulties in 26% and anxiety/depression in 23%. The severity of sequelae correlated to the severity of the acute phase. Liu et al. described in 594 patients discharged from Tongji Hospital, Wuhan, the persistence of at least one symptom in 28.4% at 12 months. Obstructive, restrictive, and mixed pulmonary function impairment persisted in 1.9%, 4.7%, 0.2% of the patients. Electrocardiogram abnormalities, including arrhythmia, ST-T change and conduction block remained in 242 patients (49.8%). Our study’ peculiarity is that the population is wider, including all grades of the WHO severity scale, and the observation period is longer than 12 months. Biases are the fact that symptoms were reported as present/absent, without severity scales. Specialists have gradually gathered around the project and we hope that more insight can be given in the future. Clinical visits detected sometimes a fluctuating course of the symptoms, leading to depression. The cause of fatigue has not been clarified yet, but it has been compared to post-infectious fatigue. A discrepancy between dyspnea, leading to a significant reduction in exercise endurance, and persistent lung damage is not uncommon. Anosmia and ageusia are so peculiar that pathophysiology is still unclear. Symptoms often evolve into altered smell and taste, which worsens the patients’ quality of life. The mechanism underlying the onset of COVID-related amnesia is still debated as the virus has shown some neurotropic and vasculotropic affinity, as well as the ability to stimulate the production of neurotoxic cytokines. What happens after COVID-19 disease cannot be summarily referred to a simple convalescence period in which symptoms gradually but linearly decrease.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.