Objective
To describe GP consultation rates for post-COVID-19 sequelae in non-hospitalised and hospitalised COVID-19 patients, and among non-hospitalised individuals describe how GP consultation rates for post-COVID-19 sequelae change over time, as well as following COVID-19 vaccination.
Design
Population-based cohort study.
Setting
1,392 general practices in England contributing to the Clinical Practice Research Datalink Aurum database.
Participants
456,002 COVID-19 cases diagnosed between 1st August 2020 to 14th February2021 (44.7% male; median age 61 years), either hospitalised within two weeks of diagnosis or not hospitalised and followed-up for a maximum of 9.2 months. A negative control group of individuals without COVID-19 (N=38,511) and patients with influenza before the pandemic (N=21,803) were used to contextualise findings.
Main outcome measures
Comparison of GP consultation rates for new symptoms, diseases, prescriptions and healthcare utilisation (HCU) in hospitalised and non-hospitalised individuals separately pre and post-COVID-19 infection using Cox regression and negative binomial regression for HCU. This was repeated for the negative control and influenza cohorts. In non-hospitalised individuals, outcomes were additionally described over time following COVID-19 diagnosis and compared pre and post-COVID-19 vaccine in individuals who were symptomatic post-COVID-19, using negative binomial regression.
Results
Relative to the negative control and influenza cohorts non-hospitalised patients (N=437,943) had significantly higher GP consultation rates for multiple sequelae, the most common being loss of smell/taste (HRadj 5.28 [95%CI 3.89 to 7.17]; p<0.001), venous thromboembolism (VTE) (3.35 [2.87 to 3.90]; p<0.001), lung fibrosis (2.41 [1.37 to 4.25]; p=0.002), and muscle pain (1.89 [1.6 3to 2.20]; p<0.001), as well as HCU post COVID-19 diagnosis compared with 1-year prior. In terms of absolute proportions, the most common outcomes ≥ 4-weeks post-COVID-19 diagnosis in non-hospitalised patients were joint pain (2.5%), anxiety (1.2%), and non-steroidal anti-inflammatory drugs (NSAID) prescriptions (1.2%). Hospitalised patients (N=18,059) also had significantly higher GP consultation rates for multiple sequelae, the most common being VTE (16.21 [11.28 to 23.31]; p<0.001), nausea (4.65 [2.23 to 9.21]; p<0.001),) paracetamol prescriptions (3.68 [2.86 to 4.74]; p<0.001), renal failure (3.42 [2.67 to 4.38]; p<0.001), as well as HCU after COVID-19 diagnosis compared with 1-year prior. In terms of absolute proportions, the most common outcomes ≥4-weeks post-COVID-19 diagnosis in hospitalised patients were VTE (3.5%), joint pain (2.7%), and breathlessness (2.8%). In non-hospitalised patients, anxiety and depression, abdominal pain, diarrhoea, general pain, nausea, chest tightness, and tinnitus persisted throughout follow-up. GP consultation rates for chest tightness, anorexia, smell/taste loss, tinnitus, chest pain, ischaemic heart disease, asthma, gastroesophageal reflux, prescriptions and HCU were lower in non-hospitalised patients post first COVID-19 vaccination relative to pre-vaccination.
Conclusions
GP consultation rates for post-COVID-19 sequelae differ between hospitalised and non-hospitalised individuals. In non-hospitalised individuals, whilst some rates of sequelae decreased over time, others such as anxiety and depression remained persistent. Rates of some outcomes did decrease after vaccination in this group.