He thought he was over COVID-19, then had a stroke. Doctors are studying how to prevent more cases like his.

Troy Randle’s COVID-19 symptoms were difficult yet bearable. After recovering from fever, cough, and headache that started in late March, he was cleared to return to work as a cardiologist in the Virtua Health network in South Jersey.

But after Randle was back on the job for two days in mid-April, his head began to ache again, and it was different. He felt as if it were being squeezed in a vise.

An MRI revealed that a blood clot had blocked an artery in Randle’s brain. He was suffering a stroke.

Physicians worldwide were starting to report the same thing in some of the sickest patients with COVID, as well as a mysterious increase in smaller blood clots elsewhere in the body. In the lungs, the liver, the kidneys — even the toes, as in those purplish “COVID toes” depicted in many a social-media feed months ago. Many hospitals began treating COVID patients with high doses of blood thinners as a preventive measure.

Three international studies of whether that was the right call, including one overseen by researchers at the University of Pittsburgh, are starting to yield results. For patients in intensive care, to the surprise of some physicians involved in the trials, the answer seems to be no.

On Dec. 22, the National Institutes of Health announced that researchers in all three trials would stop enrolling new critically ill patients, because those who had received high doses of blood thinners were just as likely to need ventilators or other types of organ support as those who were given low doses. What’s more, the early results suggested that intensive-care patients on high doses of blood thinners were more likely to suffer adverse events such as bleeding.

» READ MORE: Vaccinations are ramping up in a ‘glimmer of hope,’ Dr. Anthony Fauci says

Yet the trials are continuing to enroll hospitalized patients with moderate COVID, along with others who are not in the hospital, as researchers remain optimistic that blood thinners will prove beneficial for them. Among dozens of participating locations worldwide is the Hospital of the University of Pennsylvania.

A detailed analysis of the lackluster results in intensive-care cases is underway, but the early thinking is that by the time patients are that sick, they may be in too fragile a state to get drugs that tip the delicate balance from clotting to bleeding, said Allyson Pishko, a Penn hematologist.

“Timing is a big question mark that’s still out there,” she said.

Randle, who was able to resume work a month after his stroke, still feels its lingering effects, sometimes struggling to hold a pen. He agrees that the blood thinners in the studies, called anticoagulants, may turn out to be a plausible preventive measure for some patients. Another option may be aspirin, which prevents blood clots in a different way and also reduces inflammation, he said.

On Wednesday, Randle got a different type of prevention: a vaccine. Though he may still have some lingering immunity from his infection in March, the official guidance dictates that recovered patients should receive the shots. At a Virtua clinic in Voorhees, he rolled up the left sleeve of his black T-shirt for the first of two injections of the vaccine made by Pfizer and BioNTech.

» READ MORE: They were lifelong best friends for 80 years. Then one died of COVID-19.

“If nothing else, it’s kind of like a booster,” he said.

While the science on the vaccines is well-established, the way in which COVID leads to blood clots remains something of a puzzle, said Matthew D. Neal, a UPMC trauma surgeon and critical-care specialist who is coleading the blood-thinner trial for Pittsburgh.

Any patients in the hospital are at risk of clotting simply from being confined to their beds, as lack of movement allows the blood to stagnate. But patients hospitalized with COVID have suffered clotting complications at a much higher rate than usual — in as many as 25% of cases, compared with fewer than 10% of those admitted for other conditions.

The coronavirus infects cells in the nasal passages and deeper in the airways by latching onto proteins called receptors. The same kind of receptor also is found on cells in the lining of arteries, leading some researchers to propose early on that the abnormal blood clots in some COVID patients were the result of the virus directly infecting the walls of blood vessels.

But evidence now suggests the main culprit behind the harmful clotting is the body’s own response to the infection. An overactive immune response to the virus can prompt a cascade of harmful inflammation, triggering the blood vessels to react as if they have been injured, Neal said.

» READ MORE: After seeing patients and taking risks all year, Pa.’s private doctors have been left out of the vaccination process

“This is viewed as being a respiratory illness, but the reality is these clotting complications can occur everywhere,” he said.

Large blood clots are typically diagnosed with CAT scans. Yet early in the pandemic, when hospitals in New Jersey and New York were overrun with patients, physicians started to look for reasonable shortcuts.

If patients’ blood samples contained high levels of D-dimer — a protein fragment that can result from the breakdown of blood clots — their doctors presumed that more clots were present and ordered blood thinners, said Rick Pescatore, an emergency physician who volunteered at hospitals in northern New Jersey for a month during the first wave of cases.

“The CAT scanners had to be shut down for two hours” after being used on a COVID patient, said Pescatore, now the chief physician and associate state medical director at the Delaware Division of Public Health. “We just sort of empirically started putting people on anticoagulants.”

Many physicians around the world were doing the same. Yet, at the same time, they were learning that COVID could be grimly unpredictable. Neal, the Pitt physician, said that one day several months ago, he had one patient die as a result of harmful blood clots, while another was lost to uncontrolled bleeding.

Neither patient was enrolled in the trial, but the outcomes were a sobering reminder that answers remain elusive.

“In the same day, I have seen death as the result of both ends of the spectrum,” he said.

Further analysis of the trial results may reveal that certain patients, such as those with higher D-dimer levels, might benefit from a particular drug combination of anticlotting or anti-inflammatory agents, while others might not, Neal said. With the continued surge in COVID patients, providing fresh subjects for the ongoing studies, the Pitt physician predicts that answers will come soon.