Rapidly fatal pneumonia due to COVID-19 confronts us with an expanding pandemic for which we have no vaccination and no direct treatment. Thousands are dead, and many more would be lost without heroic ongoing efforts to support patients through respiratory failure and allow time for recovery.

The rapidly mounting tragedies strike not only the infirm and elderly but an increasing number of healthy younger individuals, and they plead for immediately available specifically effective therapies. In response, many would short-circuit the tedious process of drug discovery by jumping immediately from anecdotes of benefit to large scale treatments.

Just over 100 years ago, Sir Alexander Fleming published a paper in the British Journal of Surgery meticulously evaluating antiseptic agents used to treat battle field wounds in World War I. His startling discovery was that successful disinfectants of inanimate surfaces not only failed to prevent or treat wound infections but amplified bacterial growth, defeated natural defenses and increased damage. Some antiseptics embedded in pastes dramatically accelerated the development of gangrene.

Even after Fleming published, some antiseptics remained in use for wound care because of strongly held but wrong beliefs. In retrospect, anecdotes followed by non-critical thinking led to more deaths due to untested therapies than to the battlefield wounds and infections themselves.

Between World Wars — on Sept. 3, 1928, to be precise — Fleming noticed that a fungus contaminating his culture-plates killed Staphylococcus species, the original bacterial target of his studies. This anecdote led to carefully done experiments reported in 1929 in the British Journal of Experimental Pathology, and to additional careful studies by other British scientists.

These studies grew increasingly difficult to perform as resources were siphoned away because of armed conflicts in Europe and Asia in the 1930s. Nevertheless, after numerous laboratory investigations, followed by animal and human studies to demonstrate safety and efficacy, penicillin was successfully used in 1941 but could not be produced in large quantities. Collaboration between British and American scientists and pharmaceutical companies led to sufficient supply to treat just 10 patients by 1942. Increasing and finally ample supplies by 1944 saved many lives otherwise doomed by wounds and infections during the remainder of World War II.

In the current crisis, multiple anecdotes seemingly reveal great benefits of various potential treatments for COVID-19 infections. Some advocate rapid deployments of these untested medications including the re-purposing of hydoxychloroquine. But these stories and accompanying advocacy seem like those from World War I that led to enormous numbers of deaths due to treatments with unproven agents.

We face a choice. We can treat COVID-19 infections on a grand scale with untested medications without testing or knowledge of potential benefits or harms. With hydroxychloroquine, we have already observed deaths from self-medication and diversion to untested uses may harm people who depend on it for other diseases where benefits are known.

Instead, we can let clinician-scientists design and perform rigorous trials as rapidly as possible to discover benefits, if any, or to confirm whether already demonstrated harms would predominate, in an echo of World War I antiseptic treatments for wound infections. People are dying from COVID-19, and more will die as we investigate hydroxychloroquine and other drugs that may have a direct benefit. But the alternative risks amplifying harms and deaths by unguided decisions.

Physicians have been taught to follow the principal of primum non nocere, first do no harm, for millennia. In the middle of a pandemic, when it is most difficult, we must follow this ancient advice and test treatments before we use them for the sake of our loved ones, friends, neighbors and even our enemies.

Theodore G. Liou

Theodore G. Liou is director of the Adult Cystic Fibrosis Center and professor of internal medicine in the Division of Respiratory, Critical Care and Occupational Pulmonary Medicine at the University of Utah.