This is a Letter to the Editor that the Long COVID Initiative submitted to Annals of Internal Medicine on 13 July 2022 in response to “A Longitudinal Study of COVID-19 Sequelae and Immunity: Baseline Findings.”
Laura C. Chambers, PhD, MPH1,2; Francesca L. Beaudoin, MD, PhD, MS1,2
1Department of Epidemiology, School of Public Health, Brown University, Providence, Rhode Island, USA
2Long COVID Initiative, School of Public Health, Brown University, Providence, Rhode Island, USA
Corresponding author: Laura C. Chambers, Brown University Department of Epidemiology, Box G-S121-2, 121 South Main Street, Providence, Rhode Island, 02912; Email: email@example.com; Phone: (734) 904-5279
Funding and conflicts of interest: The Long COVID Initiative at the Brown School of Public Health is supported by the Hassenfeld Family Foundation. All authors declare that they have no conflicts of interest.
Word count: 400
Reference count: 5
To the Editor:
Sneller and colleagues’ recent article1 received controversial coverage in social media for their conclusions that extensive diagnostic evaluation did not identify a cause of post-acute sequelae of COVID-19 (PASC) and people with a history of anxiety disorder are at increased risk of PASC. While we may be apt to dismiss such critiques in the age of misinformation, it is worth further examination lest we draw erroneous inference about anxiety and PASC.
The authors’ primary analyses included 189 participants with a history of COVID-19 (104 with and 85 without PASC). Ignoring power lost due to (appropriate) correction of the false discovery rate, the study sample was sufficiently powered to detect an odds ratio of 4.6 (assuming 5% prevalence of the characteristic among controls). An effect size of this magnitude is large and probably unreasonable given that PASC may represent multiple distinct conditions with different etiologies and risk factors. Exploratory immunologic and virologic evaluations included an even smaller subset of 100 participants further limiting power. The noted absence of associations may be true or simply due to limited power.
Importantly, the study measured prevalent PASC at study enrollment, which occurred about five months after COVID-19 onset. Given that some patients experience symptom resolution within five months,2 participants with resolved PASC at enrollment would have been misclassified as controls “without PASC,” biasing results towards the null. The design was also subject to recall and selection bias that would have been mitigated with a prospective design or different sampling scheme. It is not surprising that people with a history of anxiety who thought they had PASC might be more apt to volunteer for a study about PASC. The authors acknowledge the limitations of self-referral, but it is also important to highlight how different this sample is from the general population. For example, 48% of participants had an advanced degree versus 13% of the US population.3 Finally, analyses of risk factors for PASC were unadjusted, so associations may have been confounded by measured (e.g., female sex is associated with both PASC and anxiety disorders) and unmeasured (e.g., genetic) characteristics.4
This study contributes much needed data on PASC and highlights critical areas for future research. However, we must not draw broad conclusions from a small study with important limitations, and appropriate messaging around the science is critical. Patients with PASC already experience stigma and difficulty accessing care;5 our early work on PASC should not exacerbate this.
1. Sneller MC, Liang CJ, Marques AR, et al. A Longitudinal Study of COVID-19 Sequelae and Immunity: Baseline Findings. Ann Intern Med 2022; Epub ahead of print.
2. UpToDate. Type, Proportion, and Duration of Persistent COVID-19 Symptoms. Available at: https://www.uptodate.com/contents/image/print?imageKey=PULM%2F130356. Accessed: July 7, 2022.
3. US Census Bureau. Educational Attainment in the United States: 2018 (Table 2). Available at: https://www.census.gov/data/tables/2018/demo/education-attainment/cps-detailed-tables.html. Accessed: July 7, 2022.
4. DeMartini J, Patel G, Fancher TL. Generalized Anxiety Disorder. Ann Intern Med 2019; 170(7):ITC49-64.
5. Gorna R, MacDermott N, Rayner C, et al. Long COVID Guidelines Need to Reflect Lived Experience. Lancet 2021; 397(10273):455-457.