Individuals with confirmed SARS CoV-2 infection were at increased risk of reporting a wide range of symptoms at ≥12 weeks post-infection, compared to propensity score matched uninfected patients with no record of suspected or confirmed SARS CoV-2 infection, after accounting for both sociodemographic and clinical characteristics, and the reporting of symptoms prior to infection. The symptoms most associated with SARS CoV-2 infection included some that are already recognized in previous studies,23 such as anosmia, shortness of breath, chest pain and fever, but also included a range of other symptoms that have previously not been widely reported such as hair loss and sexual dysfunction. Prior SARS CoV-2 infection was independently associated with the reporting to primary care of 20 of the 33 symptoms included in the WHO case definition, and an additional 42 symptoms, beyond 12 weeks from infection. SARS CoV-2 infection was associated with a 26% relative increase in risk of reporting at least one of the symptoms included in the WHO case definition for Long COVID.
Among those with a history of confirmed SARS CoV-2 infection, several risk factors were associated with reporting symptoms 12 weeks or more post-infection. Female sex, younger age, belonging to a black, mixed race or other ethnic minority group, socioeconomic deprivation, smoking, high BMI, and presence of a wide range of comorbidities were associated both with symptoms included in the WHO definition of Long COVID, and with symptoms statistically associated with SARS CoV-2 infection reported 12 weeks or more post-infection.
Strengths & limitations
The strengths of the study include the large sample size, which included 486,149 individuals with a confirmed diagnosis of SARS CoV-2 infection and 1.9 million propensity score matched patients with no prior record of infection. The large sample size provided adequate statistical power to assess differences in the reporting of a wide range of symptoms between the infected and uninfected cohorts and estimation of the association between reporting of symptoms and important sociodemographic and clinical risk factors with a high level of precision.
A key strength of this study is the inclusion of an uninfected comparator group that did not have either suspected or confirmed COVID-19 and had been propensity score matched for sociodemographic factors, previously reported symptoms, and over 80 comorbidities. This enabled us to assess the independent association between exposure to SARS CoV-2 and the reporting of symptoms ≥12 weeks post-infection, after accounting for many important confounders.
Another strength is the large number of symptoms included in the analysis, which was based on a previous systematic review of the literature, 10 a scoping review of Long COVID questionnaires, and an extensive consultation with patients and clinicians.22 Symptom code lists were developed rigorously with systematic searches for relevant SNOMED-CT codes with extensive clinical input. We also assessed the outcome of Long COVID using the WHO case definition as well as a novel definition that incorporated symptoms that were statistically associated with a history of SARS CoV-2 infection.
A key limitation of the study is the use of routinely coded healthcare data. Coded symptom data in primary care records is likely to underrepresent the true symptom burden experienced by individuals with Long COVID.24 This could be due to reduced access to primary care (especially during the first wave of the pandemic), patients not consulting their GP about symptoms, or patients underreporting the full extent of their symptoms. In addition, much of a patient’s clinical history, in terms of the symptoms reported, are recorded as free text, rather than as SNOMED-CT codes. The symptom data we used for the study thus cannot be used to make inferences about the absolute prevalence of these symptoms. However, since this underrepresentation would be expected to affect both the infected and uninfected patients equally, the data used in the present analysis can still be used to examine relative differences in the reporting of symptoms between infected and uninfected patients. Conversely, with the evolving awareness of Long COVID, it is possible that those with a history of COVID-19 may have been more likely than uninfected patients to access primary care and alert clinicians of their symptoms, which could potentially lead to an inflation of the observed effect sizes.
Another limitation of the study is potential misclassification bias. Community testing for SARS CoV-2 was very limited during the first wave of the pandemic, and many non-hospitalised individuals with COVID-19 were not tested. It is therefore possible that some members of our uninfected cohort had been infected with SARS CoV-2 but had simply not been tested. We attempted to partially account for this bias by excluding individuals from the uninfected pool if they had a coded diagnosis of suspected COVID-19. However, this is unlikely to be 100% sensitive in identifying individuals with unverified COVID-19 from the uninfected pool, which would potentially have the effect of attenuating the observed effect sizes.
Relationship to other studies
Our findings support the results from our previous systematic review and meta-analysis on Long COVID symptoms.10 That review found the most prevalent symptoms to be fatigue, shortness of breath, muscle pain, joint pain, headache, cough, chest pain, altered sense of smell, altered taste and diarrhoea. Our current analysis was not able to assess symptom prevalence, but rather the relative difference in symptoms between individuals with a history of SARS CoV-2 and closely matched uninfected patients at ≥12 weeks post-infection. We similarly identified that anosmia, shortness of breath, fatigue, and chest pain to be symptoms significantly associated with SARS CoV-2 infection. By contrast, we also identified novel symptoms such as hair loss, sneezing, symptoms of sexual dysfunction (difficulties ejaculating and reduced libido), hoarse voice and fever to also be significantly associated. Also, like our review, we found that female sex and the presence of a range of comorbidities were associated with an increased risk of developing persistent symptoms. However, in contrast to our review, the present analysis found that younger age was associated with a higher risk of reporting symptoms at ≥12 weeks post-infection. This could partly be due to the adjustment for an extensive range of comorbidities or the differences in the populations studied. Most studies included in our review were based on hospitalised cohorts whereas our present study excluded hospitalised patients. Older patients with COVID-19 were more likely to be hospitalised than younger patients, and therefore be excluded from our study. Older non-hospitalised patients may therefore potentially have had milder disease with lower symptom burden.
An international online cohort study of people with confirmed and suspected long COVID found that respondents reported an average of 56 symptoms across an average of nine organ systems.8 This was a comprehensive review of symptom burden in these individuals, but the study lacked a control group and was therefore unable to make strong inferences about the relative contribution of SARS CoV-2 infection to these symptoms over and above pre-existing health conditions or psychosocial effects related to the pandemic. However, like this study, we also found that individuals with a history of confirmed SARS CoV-2 reported a broad range of symptoms, with a total of 62 symptoms being associated at 12 or more weeks post-infection. Importantly, we were able to control for potential confounders, including whether the symptoms of interest were reported prior to infection.
The COVID Symptom Study provided data on self-reported symptoms among participants enrolled on a app.12 Among those with symptoms persisting 28 days or longer post-infection, key symptoms included fatigue, headache, dyspnoea and anosmia, which were all also significantly associated at ≥12 weeks in our cohort. The COVID Symptom Study also found that Long COVID was associated with increasing BMI and female sex, which is in keeping with our findings. However, the study also found that the risk of reporting Long COVID symptoms increased with age, whereas our study observed the opposite trend after adjustment for a comprehensive range of potential confounders. Although the COVID Symptom Study is community-based, it includes individuals with a history of hospitalised and non-hospitalised COVID-19 so the reasons for the discrepant age trend may be due to the same reasons described above.
One of the largest population-based surveys on COVID-19 and Long COVID is the UK Office for National Statistics COVID Infection Survey.25 This estimated that as of 6th December 2021, 1.3 million people living in private households in the UK (2.0% of the population) were experiencing symptoms persisting beyond four weeks from SARS CoV-2 infection, and with 70% experiencing symptoms beyond 12 weeks. Fatigue, shortness of breath, anosmia and difficulty concentrating were the main symptoms reported. The prevalence was greatest in females, those from more socioeconomically deprived areas, people working in health and social care and individuals living with health conditions and disabilities. Our analysis showed similar symptoms, including cognitive effects, as well as similar risk factors. However, we were unable to assess the association between occupational status and reporting of symptoms due to a lack of occupational data in UK primary care records.
Most recently, Whittaker and colleagues undertook an analysis of 456,002 patients with COVID-19 in England using the CPRD Aurum database to determine the rates of GP consultations for post-COVID-19 sequalae.26 This included both hospitalised and non-hospitalised patients and two control groups consisting of patients without COVID-19 and those with influenza before the pandemic. Patients with COVID-19 managed in the community were significantly more likely to consult for loss of taste or smell and other symptoms such as joint pain, anxiety, depression, abdominal pain, and diarrhoea at ≥ 4 weeks post-infection compared to 12 months prior to infection. They also found that patients’ GP consultation rates for symptoms, prescriptions and healthcare use were mostly reduced in those who were managed in the community after the first COVID-19 vaccination dose.
We were unable to estimate the effect of vaccination and infection year on Long COVID symptoms in our study due to the very short follow-up period among those vaccinated and infected in the year 2021 [Median 8 (IQR 4-14) and 12 (7-16) days, respectively] compared to those unvaccinated and infected in the year 2022 [33 (16-77) and 64 (31-90) days, respectively]. Furthermore, the majority (81%) of patients vaccinated prior to infection in our cohort were infected with SARS CoV-2 within two weeks of vaccination, which would be before acquiring immunity from vaccination, thus restricting the validity of our data to assess the effects of vaccination on Long COVID.
Implications for practice, policy, and research
Further research is needed to estimate the prevalence of persistent symptoms associated with SARS CoV-2 infection among patients presenting to primary care. Much of the symptom data in primary care records is held in free text entries rather than as clinically coded data. Natural language processing could be used to leverage these textual data to gain more accurate estimates of the prevalence of these symptoms.
62 symptoms reported more than 12 weeks post-SARS CoV-2 infection were found to be associated with COVID-19. It is likely that these symptoms cluster differentially amongst the population, and clustering methods could potentially be used to contribute to the discussion about Long COVID phenotypes. This can allow for assessment of whether clinical outcomes and the underlying pathophysiology differs between these subgroups and potentially develop targeted therapies for the different phenotypic subgroups.
There is also a need to obtain patient-reported data on symptoms and assess the association between symptom burden, quality of life and work capability to ascertain which symptoms have the greatest impact on individuals. Finally, there is a need to understand the natural history of Long COVID by assessing symptom burden serially over time in a population-representative cohort with a history of COVID-19 alongside a matched control population.