In the early days of a pandemic caused by a virus with no existing treatments, many different compounds are often considered and tried in an attempt to help patients.
It all relates back to a profound question: How do we know what we know?
Many of these treatments fall by the wayside as evidence accumulates regarding actual efficacy. At that point, other treatments become standard of care once their benefit is proven in rigorously designed trials.
However, about seven months into the pandemic, we're still seeing political resurrection of a treatment that has been systematically studied and demonstrated in well-designed randomized controlled trials to not have benefit.
The hydroxychloroquine (and by extension chloroquine) story is a complicated one that was difficult to follow even before it became infused with politics. It is a simple fact that these drugs, long approved by the Food and Drug Administration (FDA), work in Petri dishes against various viruses including coronaviruses. This set of facts provided biological plausibility to support formally studying their use in the clinical treatment and prevention of COVID-19. As evidence from these studies accumulates, it is a cognitive requirement to integrate that knowledge and not to evade it. This also means evaluating the rigor of the studies.
In recent days we have seen groups yet again promoting the use of hydroxychloroquine in, what is to me, a baffling disregard of the multiple recent studies that have shown no benefit. Indeed, though FDA-approved for other indications like autoimmune conditions and preventing malaria, the emergency use authorization for COVID-19 has been rescinded (which means the government cannot stockpile it). Still, however, many patients continue to ask for the drug, compelled by political commentary, viral videos, and anecdotal data. Yet most doctors (like myself) are refusing to write the prescriptions outside of a clinical trial – a position endorsed by professional medical organizations such as the American College of Physicians and the Infectious Diseases Society of America. Why this disconnect?
It all relates back to a profound question: How do we know what we know? In science, we use the scientific method – the process of observing reality, coming up with a hypothesis about what might be true, and testing that hypothesis as thoroughly as possible until we discover the objective truth.
The confusion we're seeing now stems from an inability to distinguish between anecdotes reported by physicians (observational data) and an actual evidence base. This is understandable among the general public but when done by a healthcare professional, it reveals a disdain for reason, logic, and the scientific method.
The Difference Between Observational Data and Randomized Controlled Trials
The power of informal observation is crucial. It is part of the scientific method but primarily as a basis for generating hypotheses that we can test. How do we conduct medical tests? The gold standard is the double-blind, randomized, placebo-controlled trial. This means that neither the researchers nor the volunteers know who is getting a drug and who is getting a sugar pill. Then both groups of the trial, called arms, can be compared to determine whether the people who got the drug fared better. This study design prevents biases and the placebo effect from confounding the data and undermining the veracity of the results.
For example, a seemingly beneficial effect might be seen in an observational study with no blinding and no control group. In such a case, all patients are openly given the drug and their doctors observe how they do. A prime example is the 36-patient single-arm study from France that generated a tremendous amount of interest after President Trump tweeted about it. But this kind of a study by its nature cannot answer the critical question: Was the positive effect because of hydroxychloroquine or just the natural course of the illness? In other words, would someone have recovered in a similar fashion regardless of the drug? What is the role of the placebo effect?
These are reasons why it is crucial to give a placebo to a control group that is as similar in every respect as possible to those receiving the intervention. Then we attempt to find out by comparing the two groups: What is the side effect profile of the drug? Are the groups large enough to detect a relatively rare safety concern? How long were the patients followed for? Was something else responsible for making the patients get better, such as the use of steroids (as likely was the case in the Henry Ford study)?
Looking at the two major hydroxychloroquine trials, it is apparent that, when studied using the best tools of clinical trials, no benefit is likely to occur.
All of these considerations amount to just a fraction of the questions that can be answered more definitively in a well-designed large randomized controlled trial than in observational studies. Indeed, an observational study from New York failed to show any benefit in hospitalized patients, showing how unclear and disparate the results can be with these types of studies. A New York retrospective study (which examined patient outcomes after they were already treated) had similar results and included the use of azithromycin.
When evaluating a study, it is also important to note whether conflicts of interest exist, as well as the quality of the peer review and the data itself. In the case of the French study, for example, the paper was published in a journal in which one of the authors was editor-in-chief, and it was accepted for publication after 24 hours. Patients who fared poorly on hydroxychloroquine were also left out of the study altogether, skewing the results.
What Randomized Controlled Trials Have Shown
Looking at the two major hydroxychloroquine trials, it is apparent that, when studied using the best tools of clinical trials, no benefit is likely to occur. The most important of these studies to announce results was part of the Recovery trial, which was designed to test multiple interventions in the treatment of COVID-19. This trial, which has yet to be formally published, was a randomized controlled trial that involved over 1500 hospitalized patients being administered hydroxychloroquine compared to over 3000 who did not receive the medication. Clinical testing requires large numbers of patients to have the power to demonstrate statistical significance -- the threshold at which any apparent benefit is more than you would expect by random chance alone.
In this study, hydroxychloroquine provided no mortality benefit or even a benefit in hospital length of stay. In fact, the opposite occurred. Hydroxychloroquine patients were more likely to stay in the hospital longer and were more likely to require mechanical ventilation. Additionally, smaller randomized trials conducted in China have not shown benefit either.
Another major study involved the use of hydroxychloroquine to prevent illness in people who were exposed to COVID-19. These results, published in The New England Journal of Medicine, included over 800 patients who were studied in a randomized double-blind controlled trial and also failed to show any benefit.
But what about adding the antibiotic azithromycin in conjunction with hydroxychloroquine? A three-arm randomized controlled study involving over 500 patients hospitalized with mild to moderate COVID-19 was conducted. Its results, also published in The New England Journal of Medicine, failed to show any benefit – with or without azithromycin – and demonstrated evidence of harm. Those who received these treatments had elevations of their liver function tests and heart rhythm abnormalities. These findings hold despite the retraction of an observational study showing similar results.
Additionally, when used in combination with remdesivir – an experimental antiviral – hydroxychloroquine has been shown to be associated with worse outcomes and more side effects.
But what about in mildly ill patients not requiring hospitalization? There was no benefit found in a randomized double-blind placebo-controlled trial of 400 patients, the majority of whom were given the drug within one day of symptoms.
Some randomized controlled studies have yet to report their findings on hydroxychloroquine in non-hospitalized patients, with the use of zinc (which has some evidence in the treatment of the common cold, another ailment that can be caused by coronaviruses). And studies have yet to come out regarding whether hydroxychloroquine can prevent people from getting sick before they are even exposed. But the preponderance of the evidence from studies designed specifically to find benefit for treating COVID-19 does not support its use outside of a research setting.
Today – even with some studies (including those with zinc) still ongoing – if a patient asked me to prescribe them hydroxychloroquine for any severity or stage of illness, with or without zinc, with or without azithromycin, I would refrain. I would explain that, based on the evidence from clinical trials that has been amassed, there is no reason to believe that it will alter the course of illness for the better.
Failing to recognize the reality of the situation runs the risk of crowding out other more promising treatments and creating animosity where none should exist.
What has been occurring is a continual shifting of goalposts with each negative hydroxychloroquine study. Those in favor of the drug protest that a trial did not include azithromycin or zinc or wasn't given at the right time to the right patients. While there may be biological plausibility to treating illness early or combining treatments with zinc, it can only be definitively shown in a randomized, controlled prospective study.
The bottom line: A study that only looks at past outcomes in one group of patients – even when well conducted – is at most hypothesis generating and cannot be used as the sole basis for a new treatment paradigm.
Some may argue that there is no time to wait for definitive studies, but no treatment is benign. The risk/benefit ratio is not the same for every possible use of the drug. For example, hydroxychloroquine has a long record of use in rheumatoid arthritis and systemic lupus (whose patients are facing shortages because of COVID-19 related demand). But the risk of side effects for many of these patients is worth taking because of the substantial benefit the drug provides in treating those conditions.
In COVID-19, however, the disease apparently causes cardiac abnormalities in a great deal of many mild cases, a situation that should prompt caution when using any drugs that have known effects on the cardiac system -- drugs like hydroxychloroquine and azithromycin.
My Own Experience
It is not the case that every physician was biased against this drug from the start. Indeed, most of us wanted it to be shown to be beneficial, as it was a generic drug that was widely available and very familiar. In fact, early in the pandemic I prescribed it to hospitalized patients on two occasions per a hospital protocol. However, it is impossible for me as a sole clinician to know whether it worked, was neutral, or was harmful. In recent days, however, I have found the hydroxychloroquine talk to have polluted the atmosphere. One recent patient was initially refusing remdesivir, a drug proven in large randomized trials to have effectiveness, because he had confused it with hydroxychloroquine.
Moving On to Other COVID Treatments: What a Treatment Should Do
The story of hydroxychloroquine illustrates a fruitless search for what we are actually looking for in a COVID-19 treatment. In short, we are looking for a medication that can decrease symptoms, decrease complications, hasten recovery, decrease hospitalizations, decrease contagiousness, decrease deaths, and prevent infection. While it is unlikely to find a single antiviral that can accomplish all of these, fulfilling even just one is important.
For example, remdesivir hastens recovery and dexamethasone decreases mortality. Definitive results of the use of convalescent plasma and immunomodulating drugs such as siltuxamab, baricitinib, and anakinra (for use in the cytokine storms characteristic of severe disease) are still pending, as are the trials with monoclonal antibodies.
While it was crucial that the medical and scientific community definitively answer the questions surrounding the use of chloroquine and hydroxychloroquine in the treatment of COVID-19, it is time to face the facts and accept that its use for the treatment of this disease is not likely to be beneficial. Failing to recognize the reality of the situation runs the risk of crowding out other more promising treatments and creating animosity where none should exist.
The refocusing of medical research to COVID-19 is unprecedented in human history. Seven months ago, we barely were aware that the virus existed, and now a torrent of new information greets us each day online.
There are many unanswered questions about COVID-19, but perhaps the most fascinating is whether we even need to directly go after the virus itself.
Clinicaltrials.gov, the most commonly used registry for worldwide medical research, listed 1358 clinical trials on the disease, including using scores of different potential drugs and multiple combinations, when I first wrote this sentence. The following day that number of trials had increased to 1409. Laboratory work to prepare for trials presents an even broader and untabulated scope of activity.
Most trials will fail or not be as good as what has been discovered in the interim, but the hope is that a handful of them will yield vaccines for prevention and treatments to attenuate and ultimately cure the deadly infection.
The first impulse is to grab whatever drugs are on the shelf and see if any work against the new foe. We know their safety profiles and they have passed some regulatory hurdles. Remdesivir is the first to register some success against SARS-CoV-2, the virus behind the disease. The FDA has granted it expedited-use status, pending presentation of data that may lead to full approval of the drug.
Most observers see it as a treatment that might help, but not one that by itself is likely to break the back of the pandemic. Part of that is because it is delivered though IV infusion, which requires hospitalization, and as with most antiviral drugs, appears to be most beneficial when started early in disease. "The most effective products are going to be that ones that are developed by actually understanding more about this coronavirus," says Margaret "Peggy" Hamburg, who once led the New York City public health department and later the U.S. Food and Drug Administration.
Combination therapy that uses different drugs to hit a virus at different places in its life cycle have proven to work best in treating HIV and hepatitis C, and likely will be needed with this virus as well. Most viruses are simply too facile at evolving resistance to a single drug, and so require multiple hits to keep them down.
Laboratory work suggests that other drugs, both off-the-shelf and in development, particularly those to treat HIV and hepatitis, might also be of some benefit against SARS-CoV-2. But the number of possible drug combinations is mind-bogglingly large and the capacity to test them all right now is limited.
Viruses are simple quasi-life forms. Effective treatments are more likely to be specific to a given virus, or at best its close relatives. That is unlike bacteria, where broad-spectrum antibiotics often can be used against common elements like the bacterial cell wall, or can disrupt quorum sensing signals that bacteria use to function as biofilms.
More than a decade ago, virologist Benhur Lee's lab at UCLA (now at Mt. Sinai in New York City) stumbled upon a broad-spectrum antiviral approach that seemed to work against all enveloped viruses they tested. The list ranged from the common flu to HIV to Ebola.
Other researchers grabbed this lead to develop a compound that worked quite well in cell cultures, but when they tried it in animals, a frustrating snag emerged; the compound needed to be activated by light. As the greatest medical need is to counter viruses deep inside the body, the research was put on the shelf. So Lee was surprised to learn recently that a company has inquired about rights to develop the compound not as a treatment but as a possible disinfectant. The tale illustrates both the unanticipated difficulties of drug development and that one never knows how knowledge ultimately might be put to use.
Remdesivir is a failed drug for Ebola that has found new life with SARS-CoV-2. It targets polymerase, an enzyme that the virus produces to use host cell machinery to replicate itself, and since the genetic sequence of polymerase is very similar among all of the different coronaviruses, scientists hope that the drug might be useful against known members of the family and others that might emerge in the future.
But nature isn't always that simple. Viral RNA is not a two-dimensional assemblage of genes in a flat line on a table; rather it is a three-dimensional matrix of twists and turns where a single atom change within the polymerase gene or another gene close by might change the orientation of the RNA or a molecular arm within it and block a drug from accessing the targeted binding site on the virus. One drug might need to bind to a large flat surface, while another might be able to slip a dagger-like molecular arm through a space in the matrix to reach its binding target.
That is why a broad-spectrum antiviral is so hard to develop, and why researchers continue to work on a wide variety of compounds that target polymerase as a binding site.
Additionally, it has taken us decades to begin to recognize the unintended consequences of broad-spectrum rather than narrowly targeted antibiotics on the gut microbiome and our overall health. Will a similar issue potentially arise in using a broad-spectrum antiviral?
"Off-target side effects are always of concern with drugs, and antivirals are no exception," says Yale University microbiologist Ben Chen. He believes that "most" bacteriophages, the viruses that infect bacteria and likely help to maintain stability in the gut microbial ecosystem, will shrug off such a drug. However, a few families of phages share polymerases that are similar to those found in coronaviruses. While the immediate need for treatment is great, we will have to keep a sharp eye out for unanticipated activity in the body's ecosystem from new drugs.
Is an Antiviral Needed?
There are many unanswered questions about COVID-19, but perhaps the most fascinating is whether we even need to directly go after the virus itself. Mounting evidence indicates that up to half the people who contract the infection don't seem to experience significant symptoms and their immune system seems to clear the virus.
The most severe cases of COVID-19 appear to result from an overactive immune response that damages surrounding tissue. Perhaps downregulating that response will be sufficient to reduce the disease burden. Several studies are underway using approved antibodies that modulate an overly active immune response.
One of the most surprising findings to date involves the monoclonal antibody leronlimab. It was originally developed to treat HIV infection and works modestly well there, but other drugs are better and its future likely will be mainly to treat patients who have developed resistance to those other drugs.
The response has been amazingly different in patients in the U.S. with COVID-19 who were given emergency access to leronlimab – two injections a week apart, though the company believes that four might be better. The immune response and inflammatory cytokines declined significantly, T cell counts were maintained, and surprisingly the amount of virus in the blood declined too. Data from the first ten patients is available in a preprint while the paper undergoes peer review for publication. Data from an additional fifty patients will be added.
"We got lucky and hit the bulls' eye from a mile away," says Jay Lalezari, the chief science officer of Cytodyn, the company behind leronlimab. Dr. Jay, as he is widely known in San Francisco, built an adoring fan base running many of the early-phase drug studies for treating HIV. While touting leronlimab, Lalezari suspects it might best be used as part of a combination therapy.
The small, under-capitalized firm is struggling for attention in the vast pool of therapies proposed to treat COVID-19. It faces the added challenge of gaining acceptance because it is based on a different approach and mechanism of action, which involves a signaling molecule important to immune cell migration, than what most researchers and the FDA anticipate as being relevant to counter SARS-CoV-2.
All of the therapeutics under development will face some common sets of issues. One is the pressure to have results yesterday, because people are dying. The rush to disseminate information "make me worry that certain things will become entrenched as truth, even in the scientific community, without the actual scientific documentation that ordinarily scientists would demand," says Hamburg.
"It is becoming increasingly clear that the biggest problem for drug and vaccine makers is not which therapeutics or vaccine platform to pursue."
Lack of standardization in assays and laboratory operations makes it difficult to compare results between labs studying SARS-CoV-2. In the long run, this will slow down the iterative process of research that builds upon what has gone before. And the shut down of supply chains, from chemicals to cell lines to animals to air shipment, has the potential to further hobble research.
Almost all researchers consult with the FDA in putting together their clinical trials. But the agency is overwhelmed with the surge of activity in the field, and is even less capable of handling novel approaches that fall outside of its standard guidance.
"It is becoming increasingly clear that the biggest problem for drug and vaccine makers is not which therapeutics or vaccine platform to pursue. It is that conventional clinical development paths are far too lengthy and cumbersome to address the current public health threat," John Hodgson wrote in Nature Biotechnology.
Another complicating factor with this virus is the broad range of organ and tissue types it can infect. That has implications for potential therapies, which often vary in their ability to enter different tissues. At a minimum, it complicates the drug development process.
Remdesivir has become the de facto standard of care. Ideally, clinical trials are conducted using the existing standard of care rather than a placebo as the control group. But shortages of the drug make that difficult and further inhibit learning what is the best treatment regimen for regular clinical care.
"Understandably, we all really want to respond to COVID-19 in a much, much more accelerated fashion," says Hamburg. But ultimately that depends upon "the reality of understanding the nature of the disease. And that is going to take a bit more time than we might like or wish."
[This article was originally published on June 8th, 2020 as part of a standalone magazine called GOOD10: The Pandemic Issue. Produced as a partnership among LeapsMag, The Aspen Institute, and GOOD, the magazine is available for free online.]
In October 2006, Craig Mello received a strange phone call from Sweden at 4:30 a.m. The voice at the other end of the line told him to get dressed and that his life was about to change.
"We think this could be effective in [the early] phase, helping the body clear the virus and preventing progression to that severe hyperimmune response which occurs in some patients."
Shortly afterwards, he was informed that along with his colleague Andrew Fire, he had won the Nobel Prize in Physiology or Medicine.
Eight years earlier, biologists Fire and Mello had made a landmark discovery in the history of genetics. In a series of experiments conducted in worms, they had revealed an ancient evolutionary mechanism present in all animals that allows RNA – the structures within our cells that take genetic information from DNA and use it to make proteins – to selectively switch off genes.
At the time, scientists heralded the dawn of a new field of medical research utilizing this mechanism, known as RNA interference or RNAi, to tackle rare genetic diseases and deactivate viruses. Now, 14 years later, the pharmaceutical company Alnylam — which has pioneered the development of RNAi-based treatments over the past decade — is looking to use it to develop a groundbreaking drug for the virus that causes COVID-19.
"We can design small interfering RNAs to target regions of the viral genome and bind to them," said Akin Akinc, who manages several of Alnylam's drug development programs. "What we're learning about COVID-19 is that there's an early phase where there's lots of viral replication and a high viral load. We think this could be effective in that phase, helping the body clear the virus and preventing progression to that severe hyperimmune response which occurs in some patients."
Called ALN-COV, Alnylam's treatment hypothetically works by switching off a key gene in the virus, inhibiting its ability to replicate itself. In order to deliver it to the epithelial cells deep in the lung tissue, where the virus resides, patients will inhale a fine mist containing the RNAi molecules mixed in a saline solution, using a nebulizer.
But before human trials of the drug can begin, the company needs to convince regulators that it is both safe and effective in a series of preclinical trials. While early results appear promising - when mixed with the virus in a test tube, the drug displayed a 95 percent inhibition rate – experts are reserving judgment until it performs in clinical trials.
"If successful this could be a very important milestone in the development of RNAi therapies, but virus infections are very complicated and it can be hard to predict whether a given level of inhibition in cell culture will be sufficient to have a significant impact on the course of the infection," said Si-Ping Han, who researches RNAi therapeutics at California Institute of Technology and is not involved in the development of this drug.
So far, Alnylam has had success in using RNAi to treat rare genetic diseases. It currently has treatments licensed for Hereditary ATTR Amyloidosis and Acute Hepatic Porphyria. Another treatment, for Primary Hyperoxaluria Type 1, is currently under regulatory review. But its only previous attempt to use RNAi to tackle a respiratory infection was a failed effort to develop a drug for respiratory syncytial virus (RSV) almost a decade ago.
However, the technology has advanced considerably since then. "Back then, RNAi drugs had no chemical modifications whatsoever, so they were readily degraded by the body, and they could also result in unintended immune stimulation," said Akinc. "Since then, we've learned how to chemically modify our RNAi's to make them immunosilent and give them improved potency, stability, and duration of action."
"It would be a very important milestone in the development of RNAi therapies."
But one key challenge the company will face is the sheer speed at which viruses evolve, meaning they can become drug-resistant very quickly. Scientists predict that Alnylam will ultimately have to develop a series of RNAi drugs for the coronavirus that work together.
"There's been considerable interest in using RNAi to treat viral infections, as RNA therapies can be developed more rapidly than protein therapies like monoclonal antibodies, since one only needs to know the viral genome sequence to begin to design them," said David Schaffer, professor of bioengineering at University of California, Berkeley. "But viruses can evolve their sequences rapidly around single drugs so it is likely that a combinatorial RNAi therapy may be needed."
In the meantime, Alnylam is conducting further preclinical trials over the summer and fall, with the aim of launching testing in human volunteers by the end of this year -- an ambitious aim that would represent a breakneck pace for a drug development program.
If the approach does ultimately succeed, it would represent a major breakthrough for the field as a whole, potentially opening the door to a whole new wave of RNAi treatments for different lung infections and diseases.
"It would be a very important milestone in the development of RNAi therapies," said Han, the Caltech researcher. "It would be both the first time that an RNAi drug has been successfully used to treat a respiratory infection and as far as I know, the first time that one has been successful in treating any disease in the lungs. RNAi is a platform that can be reconfigured to hit different targets, and so once the first drug has been developed, we can expect a rapid flow of variants targeting other respiratory infections or other lung diseases."