NeuroVoices: Jennifer Frontera, MD, on New COVID-19 Data

Jennifer Frontera, MD

The results from a recently published study demonstrated that patients with COVID-19 may be at risk for developing new neurologic disorders, in addition to associations with an increased risk of in-hospital mortality and decreased likelihood of being discharged. During the study timeframe, 13.5% of patients hospitalized developed a new neurologic disorder in a median of 2 days from COVID-19 symptom onset.

Led by Jennifer Frontera, MD, the most commonly observed neurologic disorders were toxic/metabolic encephalopathy (6.8%), seizure (1.6%), stroke (1.9%), and hypoxic/ischemic injury (1.4%). Among patients with seizure, 34 of 74 (46%) had no prior history of seizure or epilepsy. Since the beginning of the pandemic, there has been a growing amount of research that has shown an association between the virus and neurologic disorders, but the uncertainties about a direct correlation between them, or if the conditions were previously underlying, remain.

For now, the published has at least given clinicians confirmation that there are these associations, while still learning more about the how the virus may impact neurologic disease pathology. Frontera, a professor of neurology at the NYU Langone Grossman School of Medicine, feels as though the data is on par with what has been observed. As part of NeurologyLive’s NeuroVoices series, she discussed her findings in full detail, surprising data points, and the challenges associated with differentiating death from COVID-19 or other underlying causes.

NeurologyLive: Were you shocked by any of your findings? Did you expect your results to reveal a greater percentage of patients developing new neurologic disorders?

Jennifer Frontera, MD: I think we knew that we were probably going to have a conservative estimate of neurologic disorders simply because we couldn’t hold sedation or paralysis on a lot of the very sickest patients to do a neurological exam. Those are the ones with the very severe acute respiratory distress syndrome (ARDS). The most surprising finding to us was the degree of hypoxic ischemic brain injury. Most of the literature has focused on patients who’ve had cardiac arrest. In fact, there’s literature that suggest that just hypoxemia without hypotension is a relatively benign phenomenon. But there is a large amount of literature within ARDS that shows that cognitive deficit after having ARDS is prevalent and common in those who survive it. It’s probably mediated in part by some hypoxic brain injury. We had patients who had really profound MRI changes, particularly in vulnerable areas of the brain. Among them included certain cortical layers, 356, medial temporal, CA1 region, Purkinje cells in the cerebellum, and some striatal neurons as well.

We saw these really surprising, profound MRI changes that are primarily due to hypoxemia. When we looked at these patients, a lot of them did have episodes of hypotension, but the median number of episodes of blood pressure dropped below a MAP of 65 was in 1. Much of their encephalopathy was probably driven primarily by hypoxemia, with a second insult of episodic hypotension. Toxic metabolic encephalopathy was the most common neurological disorder that we saw in the context of COVID. We know that that's common in general and hospitalized patients, but the severity of it was somewhat surprising to us, I would say.

How does your data compare to what has previously been observed?

We found that over there was 1.9% of the patients had a combination of ischemic and hemorrhagic stroke, and 1.4% for ischemic, which compared to a global paper that was put out which showed a 0.5% incidence of stroke globally among ischemic stroke patients with COVID-19. Not all the centers included in that global study had access to imaging modalities like we did. So our rates are probably representative. A lot of patients weren’t coming in with stroke during the time when New York was the epicenter of the pandemic. There are articles in the New England Journal, etc., that showed that patients weren’t coming in at that time, including a dip in April in terms of ischemic stroke admissions, which eventually rebounded to some extent. There’s probably more stroke generally happening that we did not see. How much of it is related to COVID-19 is hard to tell. We certainly saw patients that were sick enough to come in with their stroke. It’s hard to estimate but stroke related to COVID-19 is probably a low thing, maybe 1% or less.

Now the intracranial hemorrhages we saw were interesting because they were more often lumbar than deep. They’re highly predicted by being on anticoagulation. At the time in April, we were anticoagulating based on D-dimer levels. If your D-dimer level was less than 2000, anticoagulation was not suggested. If it was between 2000 and 10,000, we suggested it as an option. For patients over 10,000, it was suggested that you anticoagulant with Heparin. What happened is that we had patients where we didn’t know they had hypoxic ischemic injury, and so we anticoagulated them and they would bleed into right ventricular brain, which is not unexpected unfortunately.

We saw a proportion of patients that had really interesting multitude of microbleeds on gradient echo. There's one autopsy paper that I saw that did find possibly viral invasion at these sort of micro hemorrhage sites, but of course you’re always concerned for blood contamination in pathological samples that see viral particles. I think the direct viral invasion question has gotten much more interesting since we last spoke, because we have more pathologic date.

There was a paper published in Lancet by a German group that looked at 43 autopsy specimens. It was a nice paper because they looked at both the immunohistochemistry for the nucleocapsid, spike proteins, but also looked at polymerase chain reaction (PCR) for RNA. It's interesting that they did detect both within a variety of different tissues, particularly in the brainstem, and cranial nerves. But what's really interesting about this paper is that where they detected the virus did not co-localize with the most neuro pathological injury.

There’s another paper that found similar results, where they detected that the virus was not necessarily where the neuropathy issues stemmed, and that it’s possible the virus was an innocent bystander coming along after the brain is already injured by hypoxic injury or some other event. In fact, that group hypothesized that if you have hypoxic injury or blood brain barrier breakdown, the virus is able to transgress into this injured tissue, and exacerbate the hypoxic injury that’s already present, and worsen the situation. The primacy of neurotropism being the actual cause of neurologic injury that is still the question. Can the virus get in the brain? It appears it can, but is it just entering tissue that’s already damaged? I don’t know. Even the RNA levels I saw in the German paper were low, something like 4500 copies, which sometimes might suggest contamination by blood products.

I think we’ve advanced our understanding on that, but not a ton. In our data, we didn’t really have any patients that had meningitis or encephalitis referable to the virus itself. If it does occur, it’s probably very uncommon. The group in Liverpool has developed this causal relationship criteria between COVID-19 and a variety of different neurological injuries to try to establish some format or guidelines for how to indicate causality with COVID-19. That’s really important so that when people more forward because if someone has meningitis for a cause not related to COVID-19, you shouldn’t ascribe the condition immediately with the virus. It’s important that we’re careful about how we look at this.

Could you explain some of the challenges associated with telling whether a patient has died from COVID-19 or a symptom associated, or something that may have been underlying in their body before?

A paper just came out in JAMA about the excess mortality with COVID-19, where they looked at the expected mortality rates during the timeframe of a surge, and then the excess rates. They went by state too. For example, Rhode Island had 22% excess mortality thought to be related to COVID-19, compared to 65% for New York. The normal number of people that die and the number of people that died during the COVID surge were certainly different. When you can see that someone died of ARDS related to COVID, it’s easier to ascribe it to that factor.

We’re also in a very interesting political time. The Great Barrington Memorandum came out not too long ago and said that we should be lax with preventative measures like masking and social distance with a less vulnerable population and try to obtain herd immunity. I think what we saw in terms of mortality and excess mortality in New York is a really important lesson. Eric Liotta, MD, from Northwestern wrote a commentary on our paper that I thought was really insightful. He had pointed out that during the same timeframe in Chicago, they had an 8% mortality rate for all COVID emissions, while New York had a 21% mortality rate. We had similar rates of mechanical ventilation, which is sort of a surrogate marker for severity of illness. Additionally, the rates we reported at NYU Langone were not different from what Northwell and Columbia reported as well. They’re very high and definitely impacted by the disaster medicine that we had to practice with during a surge. This includes limited resources, limited personnel, people coming in from other parts, nurses, doctors, and patients being treated outside in tents. It certainly was a form of medical practice that is not within our routine. We saw an excess mortality in part to that, and compared to Chicago, they did not have a surge where they were short on resources or ventilators.

If you were to apply something like the Great Barrington memorandum idea of not masking or social distancing, you allow that surge to happen again. We weren’t doing anything and 21% of those COVID-19 patients died. We have a 12% excess mortality compared to a comparable group of Chicago. That’s a lot of people that you’re willing to sacrifice, and we still don’t have herd immunity. That’s a hypnotical, it’s not scientifically based method of preventive care. No matter what you level of fatigue, if someone said to you a year ago that all you have to do is wear a piece of clothing to save someone’s life, would you do it? Maybe it’s a little inconvenient of uncomfortable, but it doesn’t hurt you or really interfere with what you do.I think you’d be hard pressed to find somebody that would say “no, I’m not going to do it.” I think New York did pretty well in terms of masking after a high surge. We can thank our state government for having strict regulations. If you look at Japan or Taiwan, where they have a culture of masking anyway and started masking very early on, they’ve kept their levels extremely low. We know those tactics work, and those people are pretty much back to their normal pre-COVID lifestyles. It’s possible, but we really need to have better management and more consistent policies across the country.

Transcript edited for clarity.