Prevalence and predictors of post-COVID-19 symptoms in general … – BMC Infectious Diseases

In this Norwegian nationwide registry-based study of primary care patients we found that the risk of fatigue or shortness of breath were more than doubled 3 to 12 months after COVID-19 compared to a control group of unexposed persons. The likelihood for memory disturbance, other respiratory symptoms, pain in abdomen, chest and musculoskeletal system and headache were also significantly higher in the COVID-19 group.

The risk of fatigue, memory disturbance and shortness of breath varied with sex, age, education and country of origin among patients after COVID-19. These risk factors generally acted in the same way in the unexposed control group, but pre-pandemic registration of comorbidities was less associated with shortness of breath after COVID-19 compared to unexposed. Having had the same symptom before the pandemic increased the risk for fatigue, memory disturbance and shortness of breath after COVID-19, but the associations were weaker in the exposed than in the unexposed group. Yet, a higher prevalence of outcome symptoms among the exposed support the hypothesis that such symptoms may be attributed to COVID-19, also in a low-risk primary care population. Post-COVID-19 symptoms were more frequent among those who were hospitalised during the acute infection compared to COVID-19 patients who were not hospitalised.

This is a nationwide registry-based study including all persons with a positive PCR test for SARS-CoV-2 and all contacts in a GP list patient system for the whole population, ruling out potential bias from selection or loss to follow up. The diagnoses were recorded by GPs in a business as usual situation, which strengthens the generalizability of the findings.

Another strength was the comprehensive data of the chosen outcome symptoms, comorbidities, and GP utilisation before the pandemic.

Most of the included COVID-19 patients in this study were unvaccinated since vaccination started in Norway at the turn of the year 2020/2021. At the end of our inclusion period (20 February 2021), only 7.5% of the Norwegian population had had a first dose and 3.7% of the population had two doses vaccinated [21] of whom many were nursing homes residents not included in our study population, 85 years of age or older, or health care workers. Consequently, the present study can be seen as a baseline study regarding post-COVID-19 symptoms in a predominantly unvaccinated population.

A limitation of our study is that for GPs, it is common to register only one or two ICPC-2 diagnosis codes, while patients may have presented several problems in one consultation. This is likely to limit the number of symptoms that we identified from registered codes, resulting in underestimation of the symptom burden. Since the data are from GP consultations which do not capture all symptoms in the population, the present study cannot be interpreted as prevalence study of post-COVID-19 symptoms in the general population.

Also, under-registration of COVID-19 is likely, as not all infected persons were tested. This was particular the case in the first 23 months of the pandemic due to a shortage of test kits. However, during the first phase in 2020 the incidence of COVID-19 in Norway was low due to lockdowns. This is supported by a study among 27 700 randomly selected persons at the end of 2020 detecting antibodies against SARS-CoV-2 only among 0.9% of the sample, which is close to the prevalence in the population with a positive PCR test at that time at 0.8% [22]. Therefore, in the whole population without positive PCR test the undiagnosed cases represent a very low share and should have neglectable effect on the results in the current study.

Having had COVID-19 may alter the utilisation of healthcare, but according to a Norwegian study the increased use of GP services related to COVID-19 gradually decreased to a normal level during the first three months after COVID-19 [23].

The lack of a specific diagnostic code for post-COVID-19 makes it difficult to assess the epidemiology of post-COVID-19 health issues. Walker et al. [24] found little use of codes for post-COVID-19 introduced for British GPs and recommended more awareness on coding of post-COVID-19 symptoms to increase possibilities for research and care planning. The current study is based on diagnoses irrespectively of the GPs interpretation of a possible connection to the prior COVID-19.

Still, those who have undergone COVID-19 may be more alert to symptoms reported as post-COVID-19 and present them to their GP. GPs may also have altered their coding practice, being more attentive to and record symptoms that could be related to prior COVID-19. This could result in confirmation and detection bias in our study. If so, the differences between exposed and unexposed may be overestimated in the current study. However, this is a challenge in all studies on registered symptoms that may be related to a certain disease.

The number of individuals with positive PCR test for SARS-CoV-2 was not very high during the exposure period used in this study. This leads to a small number of individuals in some of the risk factor strata within the COVID-19 group, and this is also reflected in the wide confidence intervals for some RRs in Fig.3. The findings regarding these risk factors should be interpreted cautiously.

Determining the prevalence of post-COVID-19 symptoms in the population is methodologically challenging, including confirmation of infection (self-report, health care reports or registries) and approaches to define post-COVID-19.

The prevalence of symptoms reported in meta-analyses published early during the pandemic were generally much higher than our findings [12, 18, 19]. Not all patients seek a GP for symptoms and the GP reported diagnoses do not reflect all symptoms experienced by the patients. When an early meta-analysis reports a prevalence of fatigue of 23 to 60%, compared to 6% in the present study, this may reflect different patient populations. A strength of our study is that it is based in primary care and therefore closer to true population rates than hospital-based studies. A later study has indicated lower prevalence, with 6% reporting post-COVID-19 symptoms after three months and about 1% after 12 months, which is more line with our data [6]. Further, a Norwegian cohort study among young adults found no association between serological signs of COVID-19 and symptoms 6 months later [25], using the broad WHO definition for post-COVID [8]. However, they showed a trend for increased prevalence of postinfectious fatigue, in line with our findings of a highest frequency and HR for this outcome symptom.

Unlike previous studies, particularly those relying on self-reported symptoms [18, 19], we found a lower likelihood for psychological symptoms in the post-COVID-19 period. This could indicate that these symptoms were frequently mild and not perceived as a reason to consult a GP, or it is possible that GPs chose to code other symptoms during the visit. On the other hand, as shown by others, we found an increase in prevalence of anxiety and depression also unexposed, possibly related to a more general effect of the pandemic [26]. It is worth noting that registry-based studies tend to report a lower prevalence of mild mental health problems compared to self-reported surveys [27].

A Dutch study, using data from a large population-based cohort study initiated before the pandemic, was able to establish a control group like we did [28]. The diagnostic categories are not quite comparable to the GP diagnostic codes used in our study, but the findings are similar to ours. Also, a study on non-hospitalised patients from UK showed hazard ratio for fatigue, shortness of breath and chest pain in the post-COVID-19 period in line with our findings [4]. In a recent study from Norway based on SARS-CoV-2 positive cases at the turn of year 2021/2022 there was marked lower HR for most symptoms [3], with HR for fatigue of 1.24 and 1.29 for Delta and Omicron respectively compared to the HR of 2.1 in our study. That study reported HRs of 1.29 (Delta) and 1.69 (Omicron) for shortness of breath compared to 2.8 in our study. These differences can be explained by a shorter observation period, maximum 4 months as compared to 12, and different covariates, as we included pre-pandemic health problems. However, this may also indicate a decrease in risk for post-COVID-19 symptom later in the pandemic, and an effect of vaccination of most of the population at that stage.

Recently, based on 9764 tested persons in the US, an attempt was made to develop a definition on post-acute sequelae of SARS-CoV-2 infection (PASC) [29]. However, only around 1300 were unexposed and symptoms were self-reported. It was acknowledged that developing a better definition would require analysis of prospectively and uniformly collected data from diverse uninfected and infected individuals. Our study may add knowledge to this field using a large sample of both exposed and unexposed.

In general, after various infections a minority of patients experience long term post-acute health problems [10]. Such symptoms are common across different infectious diseases, suggesting some common underlying mechanisms that, however, are poorly understood. In light of this knowledge, a post-acute syndrome following COVID-19 is not surprising nor exceptional.

Female sex, higher age, belonging to an ethnic minority group and a high disease burden prior to infection are documented risk factors for post-COVID syndrome [4, 5, 13, 30]. We found that the risk factors for fatigue, memory disturbance and shortness of breath were rather similar in exposed and unexposed. However, some minor divergences were found, as belonging to the age group 25 to 60 years was a stronger risk factor for fatigue in the exposed compared to the unexposed group.

In the literature, comorbid conditions are found to increase the risk for post-COVID-19 symptoms [4, 5]. However, using a score for comorbidities, we found no increase in relative risk for fatigue or shortness of breath with increasing number of comorbid conditions among exposed and a slight increase in risk for memory disturbance with three or more comorbidities. For shortness of breath, we found a lower relative risk with increasing numbers of comorbidities after COVID-19 compared to unexposed. This could indicate shortness of breath to be part of post-COVID-19 syndrome unrelated to previous diseases since the infection affects the respiratory system, whereas among the unexposed the risk increased with increasing comorbidity score. On the other hand, in our study the diagnosis shortness of breath might not have been registered if other diagnosed conditions were considered to explain the symptom by the GP, and this may more often be the case with lung symptom compared to fatigue and memory disturbance.

Outcome symptoms such as fatigue, memory disturbance and shortness of breath are common in the general population, and having had the same symptom before the pandemic markedly increased the risk for having an outcome symptom in the observation period. In clinical practice it is difficult or impossible to determine in each case whether a symptom actually was present prior to the infection, whether it is caused or aggravated by the infection or whether it appears by chance. This underlines the challenge using post-COVID-19 as diagnosis in primary care, as shown by low-frequent use of such a code when introduced in the UK [24] and also indicate that a composite explanatory model is necessary to understand the mechanisms underlying post-COVID-19 symptoms [10, 25].

Go here to read the rest:

Prevalence and predictors of post-COVID-19 symptoms in general ... - BMC Infectious Diseases