Treatment of Coronavirus Disease 2019—Evidence-based or Personalized Medicine?

(See the Brief Reports by Garcia-Vidal et al on pages 127–32 and Larson et al on pages 133–5; Viewpoints by Hall et al on pages 144–8.)

The speed and ferocity of the coronavirus disease 2019 (COVID-19) pandemic seem to have thrown the principles of sober 21st century scientific medicine out the window. Guidelines are recommending drug treatments on the basis of little more than anecdotal evidence, clinicians are treating patients on the basis of non–peer-reviewed preprints, and prestigious journals are accused of rushing underpowered and even uncontrolled observational studies into print. As editors, we are keenly aware of the difficult balance between the need for rapid communication and the desire for robust data. Cooler heads caution that evidence-based medicine must not be abandoned and that unproven therapies should not be used outside of proper randomized, controlled trials [1, 2]. However, this advice is widely unheeded, as clinicians understandably are making the calculation that their patient is deteriorating and may not live long enough for the results of carefully performed and analyzed studies to be available—and so they decide to take a chance. Neither side is necessarily wrong.

Although COVID-19 is as amenable to scientific study as any other infection, its enormous variability in clinical course and severity pose unique challenges for evidence-based medicine. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) initially evades interferon-dependent antiviral responses, allowing the virus to gain a foothold in the host [3]. Most individuals eventually mount a protective immune response and recover uneventfully. However, others, particularly older patients and those with comorbid conditions such as hypertension, obesity, and diabetes, experience more serious complications [4]. Patients with progressive COVID-19 develop lower respiratory tract involvement, which may cause hypoxia with few or no respiratory symptoms [5], a dangerous situation that can be initially overlooked. Worsening COVID-19 is typically accompanied by a dysregulated inflammatory response [6], which is triggered by the SARS-CoV-2 virus and associated with marked elevations of interleukin (IL)-6 and several other cytokines, and may continue and escalate even as the viral burden diminishes [7]. Acute respiratory distress is a frequent complication. However, in particularly severe cases, a hypercoagulable state and endothelial injury combine to create microthrombi in the lungs and other organs [8], often with lethal consequences.

This dynamic sequence of overlapping events with considerable heterogeneity among patients creates an enormous challenge for clinicians and clinical researchers alike. In the initial stages of infection, administration of interferon might be helpful to boost respiratory antiviral responses [9], but most patients require no specific intervention. Those with more significant infections require careful monitoring to ensure that they are not developing severe hypoxemia, requiring supplemental oxygen [10]. Antiviral agents are most likely to be useful early in the course of infection when viral loads are highest and before irreversible damage has occurred [11], but it is not a simple matter to predict the likelihood of clinical progression during the earliest stages of illness. A constellation of laboratory abnormalities are useful as biomarkers to monitor clinical progression: worsening lymphopenia and elevations of the neutrophil-to-lymphocyte ratio, C-reactive protein, LDH (lactate dehydrogenase), ferritin, and IL-6 are predictive of further deterioration [12–14]. Elevations of D-dimer may be particularly ominous as an indicator of intensifying coagulopathy and potentially lethal macro- or microvascular thrombotic events [15]. Elevations of troponin T or creatinine signal secondary cardiac or kidney injury [16, 17]. The clinician who attempts to help a patient navigate safely through the Scylla and Charybdis of virus and host response must ascertain the patient’s clinical trajectory and how best to intervene, usually without a robust body of evidence for guidance. For a clinical investigator, the challenge is to define the subsets of patients who are most likely to benefit from a specific intervention, whether an antiviral, an immunomodulator, an anticoagulant, a nonpharmacological intervention, or some combination of the above. The failure to precisely define the target population and the timing of an intervention may make effective therapies appear ineffective [18]. Similarly, the failure to properly control for patient heterogeneity and the natural history of COVID-19 may cause ineffective therapies to appear effective [19].

This issue of Clinical Infectious Diseases contains several articles that illustrate different aspects of the challenges clinicians who are caring for patients with COVID-19 are facing. Larson et al limited treatment to evidence-based interventions in 135 patients with COVID-19 [20], and none of their patients required mechanical ventilatory support or died, even though the cohort included patients with comorbidities associated with an increased risk of illness severity. This is a reminder that most COVID-19 infections are self-limiting with a benign clinical course, which makes it imperative to develop rigorous criteria to identify those patients who might benefit from specific interventions. Garcia-Vidal et al adopted a more personalized approach to hospitalized patients with COVID-19 in Barcelona, dividing them into groups based on whether they exhibited physiology consistent with hyperinflammation, bacterial coinfection, or hypercoagulability [21]. Lower rates of progression and mortality were observed in the patients who received treatment interventions that targeted their specific physiological pattern compared with control patients who received standard care. Along similar lines, Hall et al argue in a Viewpoints article that the variable immune responses exhibited by different patients with SARS-CoV-2 infections require a personalized approach with regard to the use of immunomodulatory therapy that is based on the patient’s specific immunophenotype [22].

Evidence-based medicine has revolutionized and brought greater rationality to medical practice [23]. However, evidence-based studies require patients who are sufficiently alike because of a shared condition, so that their responses to a clinical intervention will be generalizable. This seems to place evidence-based medicine at odds with personalized medicine, which emphasizes the aspects of each patient that make them a unique individual [24]. Patients with COVID-19 are not only not all alike, they can be profoundly different from one another with regard to both severity and pathophysiology. The heterogeneity of COVID-19 requires us to apply the principles of both evidence-based medicine and personalized approaches to make rational treatment decisions but also to ensure that we are not being misled by compelling anecdotes. For now, clinicians must carefully assess the physiological status of an individual patient and their clinical course while weighing the potential benefits and adverse effects of a treatment intervention, often on the basis of an inadequate evidence base. Clinical investigators who wish to improve this evidence base must be similarly aware of the complex dynamics of COVID-19 and take care to focus on patient populations that are most likely to benefit from a specific intervention, as well as to be wary of confounding by COVID-19’s myriad effects. Evidence-based and personalized approaches each have much to offer the practice of medicine, and we are going to need both to combat COVID-19.

Note

Potential conflicts of interest. F. C. F. reports personal fees and nonfinancial support from BioFire and Cepheid and personal fees from the Infectious Diseases Society of America outside the submitted work. R. T. S. reports grants from Gilead Sciences and personal fees from SEMPRA Energy, AbbVie, and CytoDyn outside the submitted work. In addition, R. T. S. has a patent pending in the area of long-acting anti-coronavirus compound discovery. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References