The Importance of Including Long COVID Outcomes When Developing Novel Treatments for Acute COVID-19

Abstract

Amid efforts to develop effective treatments for acute COVID-19, there is growing recognition of the need to address long COVID as a key outcome measure. We argue there are 7 compelling reasons to include long COVID measurements in clincal trials investigating acute COVID-19 treatments: (1) Long COVID is not rare. (2) Long COVID is debilitating to individuals and has a high societal cost. (3) Those at high risk of severe COVID-19 are also at higher risk of developing long COVID if they are infected with COVID-19. (4) Treatments for acute COVID-19 may reduce the risk of long COVID. (5) Measures exist to track long COVID. (6) Long COVID considerations are potentially important for acute COVID-19 treatment decision making. (7) Deaths and hospitalizations due to COVID-19 are increasingly rare. While not every trial needs to include assessments of long COVID, it is worth the research burden to include assessments where possible, as this could facilitate the uptake of acute COVID-19 treatments that lessen the societal burden of long COVID.

In January of this year we were in the midst of a major COVID-19 surge, with some analyses suggesting that the magnitude is second only to the Omicron wave of early 2022 [1, 2]. In the last week of 2023, almost 35 000 people were hospitalized from the virus in the United States [3], and during the surge, about 1000 people in the United States are dying each week from COVID-19 [1]. The Centers for Disease Control and Prevention no longer tracks cases; however, data from circulating COVID-19 in wastewater crudely suggested that >1 million Americans were newly infected each day during the JN.1 variant surge [1]. While these rates are lower than they were at the peak of the pandemic, these are still troubling numbers. COVID-19 remains a clear and present danger. At the same time, we have made notable advances in the treatment of acute COVID-19. Note that acute COVID-19 refers to the period of infection where an individual has SARS-CoV-2 and would test positive via a polymerase chain reaction or antigen test. In addition, we have learned more about long COVID—that is, the development or continuation of symptoms beyond 12 weeks after acute COVID-19 infection—and that it affects about 11% of all patients with COVID-19 [4]. To further our knowledge about currently available treatments for acute COVID-19 and the exploration of and investment in novel treatments for the same, we should incorporate the prevention of long COVID as a key outcome measure. We argue that the research community is missing an opportunity to learn more about how to prevent long COVID by failing to systematically include measures of long COVID in clinical trials of COVID-19 treatments. While there are interventions that decrease the infection and transmission of COVID-19 and prevent acute COVID-19 and thereby long COVID, our focus is on the prevention of long COVID once someone has COVID-19.

To determine whether long COVID measurements are currently included in clinical trials for developing acute COVID-19 treatments, we conducted a systematic search of current or planned US clinical trials and found only 1 trial measuring long COVID. The goal of this article is to review and consider the current state of knowledge about treating acute COVID-19 and to highlight the opportunity to leverage these efforts to advance our knowledge about long COVID.

Thus far, treatments for COVID-19 infections have been important tools in the management of the pandemic and have substantially decreased health care burden through the reduction of hospitalization and death [5–7]. COVID-19 treatments are currently recommended only for patients infected with SARS-CoV-2 who are hospitalized or who are not hospitalized but have risk factors for severe COVID-19, as these patients have been shown to benefit from intervention [8]. The patient-specific risk factors for severe COVID-19 are presented in Table 1.

The current standard of care for individuals at risk for severe COVID-19 is to initiate treatment even with currently mild symptoms [8]. Considering the current population of the United States, approximately 87.5% of adults are at risk for severe COVID-19 and are thus eligible for treatment [10]. Note that 87.5% was the best estimate that we could locate. It was not the primary outcome of the study that reported the percentage, and the estimate has several limitations: 87.5% may be an underestimate because hospitalized patients with the risk factors were not included, or it may be an overestimate because those who receive a health system administered SARS-CoV-2 test may not be representative of the population. As displayed in Figure 1, the focus of this article is nonhospitalized patients with SARS-CoV-2 infection and risk factors for severe COVID-19, as depicted in the gray box.

Ritonavir-boosted nirmatrelvir (Paxlovid) and remdesivir are approved by the Food and Drug Administration for the treatment of acute COVID-19 in this group [11]. Additionally, molnupiravir is available under an emergency use authorization of the Food and Drug Administration [11]. National Institutes of Health (NIH) treatment guidelines recommend (in order of preference) Paxlovid, remdesivir, or molnupiravir for nonhospitalized patients with SARS-CoV-2 infection and risk factors for severe COVID-19 [11].

Unfortunately, these treatments come with significant limitations for those eligible to receive them. Paxlovid comes with a substantial risk of “rebound,” where users experience a recurrence of COVID-19 symptoms or a positive test result after having the disease and testing negative [12]. In addition, Paxlovid is contraindicated for use with many drugs commonly used to treat comorbidities among individuals with risk factors for severe COVID-19 [13], thereby reducing its utility. In addition, while Paxlovid is still our best treatment for nonhospitalized patients with SARS-CoV-2 infection and risk factors for severe COVID-19, its effectiveness has waned considerably since initial studies were completed: from 88% effective to 37% effective at preventing hospitalization or death [14, 15]. Remdesivir must be administered intravenously for 3 days, which presents a logistical challenge and limits availability [11, 16]. Molnupiravir is considered a second-line treatment due to its limited effectiveness in reducing hospitalization and death since the Omicron variant [11, 16].

Thus, there is a strong interest in developing new treatments for nonhospitalized patients with SARS-CoV-2 infection and risk factors for severe COVID-19, with several potential treatments under study [17, 18]. Early trials of treatments for acute COVID-19 focused on mitigating hospitalization and death [19]. Limiting hospitalization and death are important goals due to their practical importance and the health care burden that they pose [17, 20]. But today, there is an opportunity to expand the list of outcomes to other useful clinical end points, including long COVID.

Long COVID, broadly defined, is a set of signs, symptoms, and conditions that continue or develop beyond 12 weeks after acute COVID-19 infection [21, 22], and it is increasingly being recognized as debilitating and deserving of study and prevention [23]. More than 200 symptoms have been identified, affecting every organ system [24]. The severity of long COVID ranges widely but can profoundly compromise the ability of those with the condition to perform daily activities [25]. Studies have shown that people with long COVID often take substantial time off work, lose their jobs, and face financial difficulties as a result [25]. According to the NIH, long COVID has cost society $386 billion in the United States alone as of January 2023 [26], and the cost has continued to climb in the year since this estimation. The risk of long COVID grows after each infection, and reinfection can exacerbate, reintroduce, or introduce symptoms among patients previously afflicted with long COVID [27, 28].

The Centers for Disease Control and Prevention’s estimate on the prevalence of long COVID among those with confirmed COVID-19 infection is approximately 11%, but estimates vary significantly [4]. In patients with SARS-CoV-2 infection and risk factors for severe COVID-19, the prevalence is likely higher, as there is increasing evidence that many at high risk of severe COVID-19 are also at increased risk of long COVID [29, 30]. Failure to assess long COVID symptoms among those enrolled in acute COVID-19 treatment trials leaves clinicians having to guess which treatment may be best in preventing long COVID. This contributes to uncertainty about long-term outcomes for clinicians and patients.

Factors that put individuals at risk of severe COVID-19 are the same risk factors that place those who get COVID-19 at increased risk of developing long COVID. For example, risk factors for severe acute COVID-19 and long COVID include conditions outlined in Figure 1 as well as undervaccination [31, 32]. For individuals with preexisting medical conditions, long COVID adds to their list of medical conditions and exacerbates multimorbidity (the presence of multiple chronic conditions) and its associated multiplicative negative effects [33]. Individuals with risk factors for severe COVID-19 are therefore at higher risk of not only long COVID but additional consequences of the disease. Thus, it is especially important to reduce the risk of long COVID in this population.

We believe there are at least 7 compelling arguments in favor of exploring long COVID as an outcome in studies designed to treat COVID-19 among those at high risk of severe COVID-19 (Table 2). The first 3 points can be summarized from the data just described: long COVID is not rare; long COVID is debilitating to those affected and results in a high social cost; and those at high risk of severe COVID-19 are also at higher risk of developing long COVID. In what follows, we address the remaining reasons.

COVID-19 TREATMENT MAY REDUCE THE RISK OF LONG COVID

First, there is increasing evidence that current and novel COVID-19 treatments can lessen the risk of long COVID [23]. COVID-19 treatments may indirectly reduce the risk of long COVID by reducing the severity of COVID-19, as more severe infections increase the risk of long COVID [34]. In addition to this, there is increasing evidence that COVID-19 treatments may directly reduce the risk of long COVID [35–39].

A 2023 observational study involving >280 000 individuals at high risk for severe COVID-19 infected with COVID-19 found that Paxlovid was associated with a 26% reduction in risk for long COVID [35]. This finding was consistent across all subgroups and was unrelated to vaccination history or prior infection. According to the authors, this amounted to 4.51 fewer cases of long COVID for every 100 persons treated with Paxlovid [35]. A separate study with 647 patients also found that Paxlovid reduced the risk of long COVID 6 months postinfection [36].

The potential for reducing long COVID is not limited to Paxlovid. There is evidence that molnupiravir and remdesivir may also reduce the risk of long COVID [39, 40]. In addition, other treatments under study, such as dexamethasone and metformin, have shown potential to reduce the risk of long COVID [39]. There are plausible but unexplored potential biological mechanisms that may explain how different acute COVID-19 treatments may modulate the risk of long COVID [39].

Despite potential promise, current evidence on the ability of COVID-19 treatments to directly prevent long COVID has recently been appraised by the NIH as “limited” [16]. While a number of studies indicate a protective effect of existing treatments, there are other studies that demonstrate mixed results for the prevention of long COVID [39, 41–43]. In addition, efficacy comparisons of different acute COVID-19 treatments are challenging due to a number of confounding factors, such as differences in study populations, dosages, circulating variants, follow-up times, long COVID measures or proxies, and study designs [44]. In addition, current evidence is not informed by randomized controlled trials of COVID-19 treatments, as previous clinical trials did not include long COVID as an outcome. This was due in part to a lag between the identification of long COVID and the development of a clinical and research definition as well as validated symptom questionnaires.

MEASURES EXIST TO TRACK LONG COVID

There are now several measures that aid in the diagnosis of long COVID [44–46]. The questionnaires range from focus on symptom burden to the overall impact of long COVID on quality of life and were informed by extensive patient feedback [44, 46]. Some of the measures were developed with clinical trials in mind [44], and others are used in a patient registry on ClinicalTrials.gov [46]. Long COVID questionnaires include the following: the Post–COVID-19 Symptom Scale [47], the COVID-19 Yorkshire Rehabilitation Screening Scale [48], Long COVID Symptom and Impact Tools [45], the Symptom Burden Questionnaire for Long COVID [44], the Post-Acute (Long) COVID-19 Quality of Life [46], and the Post–COVID-19 Functional Status Scale [49]. In addition, a core outcome measurement set aims to provide a comprehensive evaluation for 12 core outcomes related to long COVID [50].

While each of these measures has its merits, we agree with calls to measure long COVID uniformly across studies to better facilitate comparison [50]. The relevant experts in the field ought to develop a standard measure that maximizes the likelihood of gaining the scientific knowledge needed and minimizes research participant/patient burden. In addition to self-reported symptoms, there is increasing evidence that biomarkers of long COVID exist and may be used in its diagnosis in the future [51]. However, given that biomarkers are not currently used in diagnosis and that symptoms play a pivotal role in the disorder [52], symptom questionnaires are the most accurate and clinically useful tools for diagnosing the disorder.

To explore whether current trials for acute COVID-19 among those at risk of severe COVID-19 include long COVID measurements, we reviewed open and ongoing trials designed to assess current and novel treatments for acute COVID-19 among those at risk of severe COVID-19. We found 19 such trials listed in ClinicalTrials.gov. Our selection criteria were as follows:

• Currently recruiting or future phase 2, 3, or 4 interventional clinical trials in the United States, published on ClinicalTrials.gov as of 27 December 2023

• Interventional clinical trials focused on the treatment of acute COVID-19 in nonhospitalized adults and children

• Does not specifically exclude all people with risk factors for severe COVID-19

We limited our search to currently recruiting or not yet recruiting trials to exclude those that are no longer recruiting participants and are near completion. We determined whether long COVID was measured by reviewing primary and secondary outcome measures listed in the ClinicalTrials.gov trial entry. The titles of patient questionnaires to be completed ≥3 months postinfection in the ClinicalTrials.gov entry were screened for references to long COVID symptoms. When this was unclear by title alone, we found and reviewed the questionnaire listed. Questionnaires that specifically named long COVID (or a synonym) as an outcome were considered long COVID related. Clinical trials that utilized questionnaires with questions about long COVID symptoms but were not specific to long COVID were categorized as trials with some data collection about long COVID symptoms. Clinical trials that did not use symptom assessments beyond 3 months postinfection were categorized as having no data collection related to long COVID. Because there is significant heterogeneity among those for whom specific treatments may be counterindicated due to comorbidities or drug-drug interactions, we did not require the inclusion of all people with risk factors for severe COVID-19 in each study. Fortunately, the studies that excluded certain people with risk factors for severe COVID-19 excluded different subgroups, which reduces the risk of certain people with risk factors for severe COVID-19 being ineligible for all studies and potential treatments. Our search strategy (Supplementary material) resulted in 109 results, which were narrowed to 19 studies that met our selection criteria.

Per Figure 2, we found only 1 study explicitly collecting data on long COVID via a validated questionnaire. About a quarter of investigators are collecting data about long COVID symptomology through non–long COVID measures. These investigators may have intended to use questionnaires with long COVID symptoms as a proxy to measure long COVID; however, this is of questionable validity and makes cross-study comparison especially difficult. It is understandable that previous trials did not collect these data due to a lack of ability to meaningfully study long COVID prior to the validation of questionnaires and the urgency in creating a treatment that reduces death, but given the variety of validated measures available today and the relatively low participant burden for completing them [44, 46], these measures should be included. More than two-thirds of current studies are not measuring anything related to long COVID. These studies have lost the opportunity to collect valuable data that would be relatively easy to capture [44, 46].

Future trials of novel treatments present an opportunity to systematically utilize measures of long COVID to accurately measure whether such treatments reduce the risk of long COVID among nonhospitalized patients with SARS-CoV-2 infection and risk factors for severe COVID-19 and by how much. Treatment trials should be evaluating the risk of long COVID in all trial arms, ideally through the use of validated and comparable long COVID measurements. Measuring long COVID would require follow-up to last at least 3 months postinfection for participants who develop long COVID to meet the established World Health Organization criteria [22]. Measuring long COVID once at the 3-month mark would provide valuable evidence on incidence. Measuring long COVID beyond this point is less crucial for the purpose of a COVID-19 treatment trial focused on a number of outcomes but could enrich studies by providing information on whether the duration or severity of long COVID differs among trial arms.

RISK OF LONG COVID RELEVANT TO TREATMENT DECISION MAKING

Measuring long COVID presents a practical opportunity to work on improving acute and chronic clinical outcomes in parallel. Treatments for acute COVID-19 are currently underutilized among high-risk individuals [10, 23, 53, 54]. If demonstrated, clinical trial evidence that acute COVID-19 treatments lower the risk of long COVID may present an additional incentive for those infected with COVID-19 to seek treatment, which could increase appropriate utilization of treatments.

There are barriers to the inclusion of long COVID–related measures, most notably increased cost. For clinical trials not already following participants beyond 3 months, extending the length of the trial may require considerable resources, such as additional participants to account for increased attrition over time. For clinical trials already following participants ≥3 months, the inclusion of long COVID measurements is much more feasible, and the additional burden on the participants could be minimized. We are not arguing that every trial include long COVID measures, but determining which and how many trials should will require coordination among sponsors and investigators based on burden, cost, and other logistical considerations. Coordination is important to ensure that enough research is incorporating long COVID measurements for each potential COVID-19 treatment.

DEATHS AND HOSPITALIZATIONS DUE TO COVID-19 ARE INCREASINGLY RARE

Adapting or coordinating trials to capture information on long COVID is a worthy investment. Knowledge about a treatment's ability to reduce the potential long-term consequences of long COVID has scientific and social value. This is especially true given the reduction in the numbers of hospitalizations and death due to COVID-19 [19], which are increasingly rare given the high levels of vaccine- and infection-mediated immunity [55]. COVID-19 experts are actively considering alternate end points to use in clinical trials [19] and must not overlook long COVID.

CONCLUSION

We have outlined 7 reasons why we should be measuring long COVID in clinical trials for acute COVID-19 treatments (Table 2) and highlighted that the majority of clinical trials are missing this opportunity to use such a measure (Figure 2). Now is the time to expand the scope of COVID-19 research. Otherwise, we risk delaying the uptake of COVID-19 treatments that improve acute COVID-19 and long-term sequalae. We face the potential drawback of evaluating treatments with incomplete information on meaningful long-term outcomes. This poses a risk of improperly weighing the value of different treatments and neglecting treatments that may be more effective at preventing long COVID as compared with others. The additional costs associated with longer follow-up should not be minimized. However, an up-front investment now could facilitate the uptake of treatments that lessen the substantial societal burden of long COVID.

Clinicians treating patients diagnosed with acute COVID-19 should assess long COVID symptoms at 3 months postinfection and at follow-up visits. Investigators conducting trials of treatment for acute COVID-19 should incorporate assessments of long COVID symptoms into their study protocols. For trials that have already started and will follow participants for at least 3 months, investigators should consider including an assessment of long COVID symptoms as soon as feasible. Clinicians and investigators with experience and expertise in the symptoms of long COVID should seek consensus on a standard assessment for long COVID symptoms. Professional societies, research sponsors, and health policy makers should support efforts to develop a standard assessment for long COVID.

In conclusion, given the need for new treatments and the lack of trial data on whether current COVID-19 treatments reduce the risk of long COVID, there is a need for considerations of long COVID to be highlighted in acute COVID-19 treatment clinical trials.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Financial support. This work was supported by the Intramural Research Program of the Clinical Center, National Institutes of Health.

References

Author notes