6-month consequences of COVID-19 in patients discharged from hospital: a cohort study

The study addresses the largely unknown long-term health consequences of COVID-19 after hospital discharge. As of early January 2021, COVID-19 had caused over 83 million cases and 1.8 million deaths globally. While acute epidemiological and clinical characteristics are well described, evidence on persistent symptoms and organ dysfunction beyond three months post-discharge is limited and based on small cohorts. The purpose was to characterise a broad spectrum of 6-month outcomes after COVID-19 hospitalisation and to identify risk factors, particularly the role of acute disease severity, for adverse long-term consequences.

Evidence before this study: A PubMed search up to Nov 5, 2020, identified limited follow-up studies of discharged COVID-19 patients, generally with small sample sizes and follow-up up to approximately 3 months. Reported issues included persistent symptoms (fatigue, dyspnoea), impaired pulmonary function, chest imaging abnormalities, and reduced quality of life. However, representativeness and detail were insufficient, and risk factors for long-term health consequences were unknown. This study adds the largest cohort (n=1733) with 6-month follow-up, quantifying symptoms, psychological outcomes, pulmonary function, imaging findings, extrapulmonary organ function, and antibody dynamics, and identifying disease severity and sex as risk factors.

Design and setting: Ambidirectional cohort study at Jin Yin-tan Hospital, Wuhan, China, including all laboratory-confirmed COVID-19 patients discharged between Jan 7 and May 29, 2020. Eligibility and exclusions: Excluded were patients who died before follow-up; those with difficulty in follow-up due to psychotic disorder, dementia, or readmission for underlying diseases; those unable to ambulate due to osteoarthropathy or immobilised pre/post-discharge (e.g., stroke, pulmonary embolism); those who declined; were unreachable; or lived outside Wuhan or in nursing/welfare homes. All discharges met national criteria (afebrile ≥3 days, respiratory symptom improvement, imaging resolution of acute lesions, and two negative SARS-CoV-2 tests ≥24 h apart). Ethics approval KY-2020-78.01; written informed consent obtained. Follow-up procedures: Appointments scheduled via telephone; missed visits could be rescheduled twice. In-person outpatient assessments included: - Questionnaires: self-reported symptoms (new/persistent/worsened vs pre-COVID), mMRC dyspnoea scale, EQ-5D-5L, EQ-VAS, and cardiovascular/cerebrovascular events registration. - Physical examination and 6-minute walk test (6MWT). - Blood tests: complete blood count, serum creatinine, haemoglobin, and HbA1c. - Antibody testing: For participants enrolled in the LOTUS trial, paired plasma from acute phase (median 23 days from onset) and follow-up were tested for IgM/IgA/IgG against nucleoprotein, spike, and RBD by ELISA, and neutralising antibodies by microneutralisation on Vero cells. Ancillary testing (stratified sampling): Disproportionate random sampling by peak seven-category severity scale (3: no oxygen; 4: oxygen; 5–6: HFNC/NIV/IMV) for pulmonary function tests (PFTs), chest HRCT, and ultrasonography of lower limb veins and abdomen. All with scale ≥5 were invited for all three tests; scales 3 and 4 sampled at a 1:2 ratio. Measurements and definitions: - Disease severity: highest seven-category scale during hospitalisation (1–7; 7=death). - PFTs: performed per American Thoracic Society standards using Master Screen PFT; included spirometry (FEV1, FVC, FEV1/FVC), lung volumes (TLC, FRC, RV), and carbon monoxide diffusion capacity (DLCO; not corrected for haemoglobin). Predicted values per Enright and Sherrill. - Chest HRCT: SIEMENS 64-slice scanner, 1-mm slices. AI software quantified lesion volumes per lobe; semiquantitative CT score (0–25) summed over lobes based on percentage lobe involvement. - Renal function: eGFR via CKD-EPI equation; AKI per standard criteria (appendix). HbA1c used to identify diabetes. - Outcomes: Primary—symptoms (fatigue/muscle weakness, sleep difficulties, hair loss, smell disorder), exercise capacity (6MWD), health-related quality of life (EQ-5D-5L domains, EQ-VAS), lung function (including DLCO), and chest CT patterns. Secondary—extrapulmonary organ function (eGFR, HbA1c, venous thrombosis, abdominal organ ultrasound) and antibody titres/seropositivity. Statistical analysis: Continuous variables as median (IQR), categorical as counts (%). Comparisons by severity scale (3, 4, 5–6) and by sex using Mann–Whitney U test, chi-square, or Fisher’s exact as appropriate. Multivariable logistic regression for categorical outcomes and linear regression for continuous outcomes evaluated associations with disease severity, adjusting for age, sex, smoking status, education, comorbidities (hypertension, diabetes, cardiovascular, cerebrovascular, malignancy, COPD, CKD), and treatments (corticosteroids, antivirals: lopinavir-ritonavir, arbidol, chloroquine phosphate, hydroxychloroquine; IVIG). Antibody acute vs follow-up comparisons used paired t tests (titres) and McNemar tests (seropositivity). Additional models explored risk factors for diffusion impairment, anxiety/depression, fatigue/muscle weakness, and percentage change in CT score from acute to follow-up. Two-sided p<0.05 considered significant. Analyses in SAS 9.4.

Cohort and follow-up: - 2469 discharged; 1733 enrolled after 736 excluded. Median age 57 years; 52% male. Follow-up June 16–Sept 3, 2020; median 186 days from symptom onset; 153 days from discharge. Symptoms and quality of life at 6 months: - 68% (1119/1650) reported at least one symptom; higher in females. - Most common: fatigue or muscle weakness 52% (855/1654); sleep difficulties 26% (437/1655). - Anxiety or depression: 23% (367/1616), more common in females. Exercise capacity: - 6MWD below lower limit of normal: 17% (74/428) in severity 3; 13% (144/1147) in severity 4; 28% (32/114) in severity 5–6. Pulmonary function and imaging (subset tested): - Lung diffusion impairment (DLCO <80% predicted): 22% (18/83) severity 3; 29% (48/165) severity 4; 56% (48/86) severity 5–6. - After multivariable adjustment vs severity 3: severity 5–6 had higher odds of diffusion impairment (OR 4.60, 95% CI 1.85–11.48). - Chest HRCT median CT scores: severity 3 = 3.0 (IQR 2.0–5.0); severity 4 = 4.0 (3.0–5.0); severity 5–6 = 5.0 (4.0–6.0); significant difference between 3 and 5–6. Common patterns: ground-glass opacities and irregular lines; consolidations largely resolved by ~5 months. - Lesion volumes were greater in severity 5–6 vs 3 (e.g., lung lesion volume β ≈ +34 cm³, p<0.05). Adjusted associations (multivariable): - Severity 5–6 vs 3: higher odds of anxiety/depression (OR 1.76, 95% CI 1.05–2.96) and fatigue/muscle weakness (OR 2.75, 95% CI 1.61–4.69). Differences for severity 4 vs 3 were generally not significant. - Female sex associated with higher odds of diffusion impairment (OR 2.22, 95% CI 1.24–3.98) and anxiety/depression (OR 1.80, 95% CI 1.38–2.34). - Per 10-year age increase: higher odds of diffusion impairment (OR 1.27, 95% CI 1.02–1.60) and fatigue/muscle weakness (OR 1.25, 95% CI 1.15–1.37); smaller percentage improvement in CT score from acute to follow-up. Antibody dynamics (n=94 with paired samples): - Neutralising antibody seropositivity declined from 96.2% (acute) to 58.5% (follow-up); median titre declined from 19.0 to 10.0. - Seropositivity of N-IgM, RBD-IgM, S-IgM, N-IgA, RBD-IgA, S-IgA, and RBD-IgG decreased significantly; N-IgG and S-IgG seropositivity remained high (>90%). Overall titres of all isotypes and neutralising antibodies waned over time, with heterogeneous IgG responses (a minority showed >20% increases). Extrapulmonary findings: - At follow-up, 35% (487/1393) had eGFR <90 mL/min/1.73 m²; among those without AKI and with normal eGFR during acute phase, 13% (107/822) developed eGFR <90 at follow-up. - Newly diagnosed diabetes: 58 patients without prior history. - Lymphocytopenia resolved in most: among 488 with acute-phase lymphocytopenia, 97% had lymphocytes ≥0.8×10^9/L at follow-up. - No lower-limb deep venous thrombosis detected in 390 ultrasounds; abdominal ultrasounds were normal in those assessed.

This large, 6-month follow-up cohort demonstrates that most previously hospitalised COVID-19 patients continue to experience symptoms—particularly fatigue or muscle weakness, sleep difficulties, and psychological distress. Greater acute disease severity is linked to higher risk of impaired pulmonary diffusion, greater residual chest imaging abnormalities, and more fatigue and anxiety/depression, underscoring the need for targeted post-discharge rehabilitation and mental health support for severely ill survivors. Pulmonary findings (persistent DLCO impairment and interstitial-type HRCT changes) align with long-term sequelae reported after SARS and influenza, suggesting ongoing interstitial remodeling and possible fibroblast activation during convalescence. While corticosteroids benefit acute severe COVID-19, this study did not find evidence that they accelerate recovery of lung function or radiographic abnormalities. The significant waning of neutralising antibodies from acute illness to 6 months raises concern for susceptibility to reinfection, although IgG seropositivity (N- and S-IgG) remained high. Heterogeneity in humoral trajectories suggests individual variability in immune durability, supporting ongoing surveillance and possibly vaccination strategies. Extrapulmonary consequences include new or persistent renal dysfunction and incident diabetes, indicating potential ongoing organ involvement beyond the lungs and the importance of monitoring multi-organ health post-COVID-19. Overall, the findings address the study question by defining the spectrum and risk factors of long-term consequences and highlight populations (those with severe acute illness and females) who may benefit most from follow-up and interventions.

At approximately 6 months after symptom onset, many COVID-19 survivors report persistent symptoms, most commonly fatigue or muscle weakness and sleep difficulties, with notable rates of anxiety or depression. Those with greater acute-phase severity show more impaired pulmonary diffusion and residual imaging abnormalities and are key targets for post-discharge care. Neutralising antibodies decline substantially over time, warranting vigilance for possible reinfection. Continued, longer-term follow-up in larger and more diverse populations is needed to determine the trajectory of recovery, resolution of pulmonary abnormalities, durability of immunity, and long-term extrapulmonary outcomes.

- Lack of pre-COVID baseline pulmonary function and 6-minute walk data limits attribution of impairments solely to COVID-19, though chronic pulmonary/heart disease prevalence was low (self-reported, potential underestimation). - Smoking status and some comorbidities collected during early pandemic response may be underreported. - New symptoms were not stratified into persistent vs worsened vs de novo post-discharge. - Mild cases managed in Fangcang shelter hospitals were not included, limiting generalisability beyond hospitalised populations. - Limited sample size for paired antibody analyses reduces precision of immune durability estimates. - Ongoing follow-up suggests results are interim; further data may refine risk estimates.