5: PCR and antibody tests should not be used as inclusion criteria for trial participants
Only an estimated 1%-3% of cases in the first wave of COVID were documented, and the CDC estimates that only 25% of cases through September 2021 were documented. Similarly, antibody tests are unreliable to determine past infection, as roughly a third of patients don’t seroconvert, and a similar proportion serorevert within a few months. Using polymerase chain reaction (PCR) and antibody testing to determine who should be included in clinical trials limits who is eligible to participate in research, particularly those who have been ill for longer. Additionally, the majority of those who serorevert are women, so using antibody tests for inclusion introduces a selection bias and may miss mechanisms of immune system functioning that are part of long COVID.
PCR tests also have high false-negative rates and requiring them in research excludes people with lower viral loads with long COVID, which would confound findings.
These issues with testing also lead to COVID-infected people accidentally being included in control groups, which ruins the credibility of the research findings completely.
6: Include comparator groups
There are several common diagnoses that occur in people with long COVID, including ME/CFS, postural orthostatic tachycardia syndrome, small-fiber neuropathy, mast cell activation syndrome, and Ehlers-Danlos syndrome.
Identifying people with these conditions within the trial cohort improves research across all fields, benefiting all groups, and helps clarify what types of patients benefit most from certain medications.
7: Identify the right endpoints; avoid the wrong ones
Even though our understanding of the pathophysiology of long COVID is still evolving, it’s still possible to do clinical trials by identifying strong endpoints and outcome measures.
Several tools have been designed for viral-onset conditions and should be used alongside other endpoints. Postexertional malaise and autonomic symptoms, which are some of the most common symptoms of long COVID, can be measured with the validated DSQ-PEM and COMPASS-31, respectively. Tools for cognitive dysfunction trials should capture specific and common types of impairment, like processing speed.
Endpoints should be high-impact and aim for large improvements that have clinical significance over small improvements that do not have clinical significance.
Objective tests should be incorporated where possible; some to consider include natural killer cell functioning, cerebral blood flow, T-cell functioning, levels of reactivated herpesviruses, blood lactate levels, and microclots, as testing becomes available.
Mental health outcomes shouldn’t be primary endpoints, except where a trial is targeting a specific mental health condition because of COVID (for example, premenstrual dysphoric disorder).
If mental health conditions are tracked secondarily, it’s vital not to use questionnaires that include physical symptoms like fatigue, difficulty concentrating, difficulty sleeping, or palpitations, as these artificially increase depression and anxiety scores in chronically ill respondents. Tools that include physical symptoms (Patient Health Questionnaire–9, Beck Anxiety Inventory, Beck Depression Inventory) can be replaced with scales like the PHQ-2, General Anxiety Disorder–7, Hospital Anxiety and Depression Scale, or PROMIS-29 subscales.
Because certain cytokines and other inflammatory markers may naturally decrease over time without corresponding improvement in the ME/CFS subtype, caution should be taken when using cytokines as endpoints.