Incidence and predictors of long Covid-19 in hospitalized patients: A cohort study

INTRODUCTION

The long-term effects of severe acute respiratory syndrome corona virus 2 (SARS-CoV-2) are frequently denoted inter-changeably as long Covid and post-Covid syndrome. Clinical guidelines were proposed for the diagnosis and management of long Covid-19 in April 2021.¹ Most observational studies using patient health records and clinical data show that the prevalence of long Covid following acute SARS-CoV-2 infection is 10%–30%, with signs and symptoms occurring after hospital discharge. Chest pain, fatigue, dyspnoea, hair loss, depression, headache, muscle pain, cough, and joint pain are the most common reported long Covid symptoms.^1–3 Furthermore, long Covid results in a wide range of manifestations, affecting different organs and systems in the body. The presence of co-morbid conditions increases the risk of developing long Covid symptoms.⁴ Our aim was to determine the incidence and predictors of long Covid at 1-year follow-up among patients admitted with a diagnosis of Covid-19.

METHODS

In this prospective observational study, we collected clinical data of 393 patients admitted and diagnosed with Covid-19 disease at a tertiary cardiac care hospital from 27th March to 27th May 2021.

Patients diagnosed to have Covid-19 based on a positive reverse transcriptase polymerase chain reaction (RT-PCR) test, radiology, and laboratory findings, and admitted to our institution based on the guidelines of the Government of India, Ministry of Health and Family Welfare, were included in this study.⁵ Patients with missing data were excluded.

Their baseline characteristics, laboratory findings, treatment, and outcome data were recorded. We scheduled a follow-up, 1 year after diagnosis of Covid-19 disease at our hospital. Long Covid was defined as the persistence of symptoms beyond 4 weeks of SARS-CoV-2 infection, which are not explained by other causes.⁶ Classification of Covid as mild, moderate, and severe was done as per the Ministry of Health and Family Welfare, Government of India.⁵

At the time of follow-up, pulse, blood pressure, temperature, respiratory rate, and peripheral capillary oxygen saturation (SpO₂) were recorded. Electrocardiograms (ECGs) and 2D ECGs were performed on all patients. Additionally, blood was drawn for complete blood count, C-reactive protein, D-dimer, and HbA1c. Lipid profiles were measured using International Federation of Clinical Chemistry (IFCC)—approved enzymatic methods with a commercially available kit on an auto-analyser (ARCHITECH PLUS ci4100, Germany)— Lipid levels were classified according to the classification suggested by the National Cholesterol Education Program (NCEP) and the Adult Treatment Panel III (ATP III) guidelines.⁷ The institutional ethics committee approved this study (UNMICRC/Allied/2021/18).

RESULTS

We found the incidence of long Covid-19 to be 34.4% after hospital discharge at 1-year follow-up. The baseline characteristics and risk factors of the patients are given in Table 1.

At presentation, 37 (9.4%) patients had ground glass opacity, 16 (4.1%) had local patchy, and 77 (19.6%) patients had bilateral patchy consolidation on the chest X-ray. 77 (19.6%) patients had a ‘severe’ score on computed tomography (CT-SS) of the chest. During the illness, patients were managed in the hospital according to disease severity and many patients received Remdesivir (71.5%) and steroids (65.1%). Tocilizumab (5.1%) and immunosuppressant (3.1%) drugs were also used. At the time of admission, 174 (44.3%) patients required oxygen; 67 (17%) patients were on nonrebreather mask (NRBM) and 7 (1.8%) patients required high flow nasal cannula (HFNC). 35 (8.9%) patients needed bi-level positive airway pressure (BiPap) support.

At 1-year follow-up, 135 (34.4%) patients had post-Covid symptoms. Most patients reported joint pain (40%) and fatigue (33.3%; Table 2).

Other common symptoms were dyspnoea (22.9%), postural orthostatic tachycardia (POTS) (11.1%), hair loss (5.9%), chest pain (4.4%), weakness (3.7%), pedal oedema (2.2%), and mucormycosis (2.2%). New onset hypertension was detected in 32.3% of patients at 1-year follow-up.

Based on the presence of post-Covid-19 symptoms at 1-year follow-up, we classified patients into two groups: those with absence of long Covid (65.6%) and those with long Covid (34.4%) and compared their clinical features (Table 3).

At follow-up, most patients (86%) had a normal ECG (Table 4). 1.8% patients had T-wave inversion, 2.5% had left bundle branch block (LBBB), and 1.8% had right bundle branch block (RBBB). An echocardiogram (Echo) was also obtained. Baseline left ventricular ejection fraction (LVEF) was 55%–60% in the majority of patients (86%). We compared the baseline Echo of patients who had low LVEF: 26 patients (6.6%) had LVEF of 30%–50% at the 1-year examination and 24 (6.1%) had low LVEF even at first hospitalization. Only 2% patients had below 30% LVEF. 14.8% patients had diastolic dysfunction. 82.4% had pulmonary hypertension.

On regression analysis, older age (56.64 years), female gender (39.3%), in-hospital complications (13.3%), presence of diabetes (35.6%), high BMI (27.73 kg/m²) and severity of Covid-19 disease (14.1%) at the time of admission were significantly associated with the risk of developing long Covid-19. The use of steroids and duration of hospitalization were also higher in the long Covid group but were not significant (Table 5).

We determined the cut-off values for BMI and age from the receiver operating characteristic curve (ROC). Our cut-off value of BMI >25.9 kg/m² showed an area under the curve (AUC) of 0.61 (95% CI 0.56–0.66, p=0.0003) with 67.9% sensitivity and 54.4% specificity. At the age of >60 years, the AUC was 0.57 (95% CI 0.52–0.62, p=0.03), with 43% sensitivity and 67.8% specificity, among those who had Covid.

At 1-year follow-up, low haemoglobin was found in 7 (1.8%), raised CRP-Q (>5 mg/L) in 3 (0.8%), and raised Ddimer (>500 ng/L) in 67 (17%) patients. Among the lipid profile were increased level of triglyceride (TG) in 147 (37.4%), high density lipoprotein cholesterol/low density lipoprotein cholesterol ratio (LDL/HDL ratio) in 146 (37.2%), very low-density lipoprotein (VLDL) in 109 (27.7%), cholesterol in 101 (25.7%), total lipids (TL) in 108 (27.5%), low density lipoprotein cholesterol (LDL-C) in 96 (24.4%) and low high-density lipoprotein cholesterol (HDL-C) levels in 200 (50.9%) patients. According to the NCEP guidelines, 73% patients had dyslipidaemia. HbA1c was also increased in 87 (22.1%) patients. New onset of type 2 diabetes was seen in 3.85% of the patients. D-dimer levels were also increased in 67 (17%) patients.

DISCUSSION

This prospective observational study was designed to address post-recovery symptoms in patients who experienced acute symptoms of Covid-19. Many hypotheses have been proposed for the pathophysiology of long Covid, but the mechanism is still unclear.⁸

Long Covid generally manifests as a combination of disorders and complications. People recovering from Covid-19 sometimes show symptoms of a condition called postural orthostatic tachycardia syndrome (POTS). Typical arrhythmias can occur in individuals with Covid, and long-term Covid-19 has been associated with tachycardia. Common symptoms of Covid-19 POTS are tachycardia, headaches, and dyspnoea. A recently published review suggests that POTS is commonly associated with fatigue and connective tissue disorder. Our study found an 11.1% incidence of post-Covid-19 POTS; several studies have reported a higher prevalence of POTS.^9,10

A recent study reported joint pain as a generalized long-term symptom of Covid-19.¹¹ One north Indian study reported that 15.6% patients had joint pain post-Covid from three to six months, 20.8% at six weeks to three months, and 28.6% after 3 months. We also found that 13.7% patients had joint pain as a long Covid symptom post-Covid disease.¹²

Symptoms such as fatigue are common (33.3%) long-term complaints post-Covid-19 recovery. In studies conducted in Italy, the UK, and France, fatigue was found to be the most common post-Covid-19 manifestation.^13–15 César Fernándezde-las-Peñas et al. describe the incidence of persistent fatigue and dyspnoea in the largest multicentre study published with a long-term follow-up period in hospitalized Covid-19 survivors.¹⁶

There are many different types of hair loss and a multitude of factors that can contribute to it. One proposed theory for post-Covid-19 hair loss is that interleukin-6 (IL-6), a pro-inflammatory cytokine, may play a role.¹⁷ A Saudi Arabian study found 26% patients reporting hair loss at 3-months follow-up.¹⁸ We found that 5.9% patients had hair fall as a long-term Covid symptom.

A case series reported that 9 patients who recovered from Covid-19 completely developed pedal oedema.¹⁹ We found that only 2.2% patients had pedal oedema after Covid-19 recovery. Another Indian study reported that diabetes and widespread use of corticosteroids in the background of Covid-19 appears to increase mucormycosis.²⁰ We found 3 patients with mucormycosis, and all had diabetes and were treated with steroids.

We found new onset of hypertension in 32.3% and type 2 diabetes in 3.8% patients at 1-year follow-up. A recently published systematic review evaluated changes in blood pressure, sugar, and lipid profiles of Covid recovered patients at follow-up to identify new-onset of hypertension, diabetes, and dyslipidaemia.²¹ Disease prevalence and risk factors for long Covid tend to be higher in females, increasing age, obesity, and poor sociodemographic factors.^22–24

We found that BMI above 25.9 kg/m², and age >60 years were independent predictors of post-Covid symptoms. Thompson et al. reported that a high BMI increased the risk of long Covid.²³ Similarly, a study by Sudre et al. reported that patients with prolonged symptoms were more likely to be obese.²² Davis et al. reported that hospitalized Covid-19 patients presented with more symptoms post-Covid recovery, which is supported by our findings.¹⁰ Our study results show that lipid profile values remained high during follow-up. Wrona et al. and Li et al. reported that LDL-C, triglycerides, and TC were significantly higher at follow-up.^21,25

Limitations of our study include it being a single-centre experience, a small number of patients, and a lack of lipid profile data of the cohort at the time of admission.