General discussion 9 277 The RECoVERED cohort aimed to ensure approximately equal enrolment of participants across the disease spectrum, rather than be representative of the pyramid of COVID-19 severity in the Dutch population. To gain insight into the prevalence of long COVID in the general population, we must therefore look to population-based studies. The national Nivel Surveillance evaluated long COVID among adults visiting primary care services 3-12 months after verified SARS-CoV-2 infection during the first wave of COVID-19 in the Netherlands[53]. Among 10,313 electronic patient records of non-hospitalised individuals who visited their primary care physician for ongoing symptoms, the authors reported a long COVID proportion of 23% when using a conservative definition and 37% when using a less stringent definition. These initial analyses, however, did not evaluate long COVID among hospitalised individuals, although the authors aim to link their dataset with hospital databases in the future[54]. Moreover, it remains unclear to what extent ongoing symptoms were attributable to SARS-CoV-2 infection, as like our study, the Nivel Surveillance study did not enrol a COVID-19-negative control group. In contrast, a population-based study conducted by the RIVM recruited test-negative controls in addition to individuals with COVID-19. Three months after infection (for cases) or start of symptoms (for symptomatic test-negative controls), almost half of SARS-CoV-2-infected individuals reported at least one symptom (48.5%) whilst one-third (29.8%) of controls also reported ongoing complaints. However, the study was unable to evaluate to what extent participants’ symptoms may result from any pre-existing conditions. In the Lifelines Cohort – a prospective, population-based cohort in the north of the Netherlands[55], in which 5.0% of male participants and 2.5% of female participants were hospitalised due to COVID-19 – researchers were able to conduct powerful analysis which may more closely represent the occurrence of long COVID in the Dutch population. The authors used both historical symptom data (serving as pre-pandemic controls for later infected individuals) as well as contemporary, non-infected controls that were matched to SARSCoV-2-positive individuals on various confounding factors. The study concluded that 13% of persistent symptoms reported at 90-150 days after infection could be attributed to COVID-19 in this cohort. A caveat of the Lifelines Cohort remains, however, that 98% of study participants were ethnically white, limiting the external validity of findings. In summary, estimating the occurrence of long COVID is challenging, and depends greatly on the choice of study design, population and analysis method. For public health professionals willing to evaluate the potential disease burden of long COVID in the Dutch population, it may be of greatest interest to evaluate how frequently long COVID occurs among the largest group of COVID-19 cases: those with initially mild disease[56].
RkJQdWJsaXNoZXIy MTk4NDMw