COVID-19: To treat or not to treat? April 26 update

Dr R. Phillip Dellinger, professor of medicine, senior critical care attending at Cooper University Health Care, past president of the Society of Critical Care Medicine, and lead author of the Surviving Sepsis Campaign guidelines (2004, 2008, 2012), talks with Dr Roman Jaeschke about some of the striking features of COVID-19, including thromboembolic events, and current patterns of practice.

R. Phillip Dellinger, MD, MSc: Roman, good to see you, my friend. I’m Phil Dellinger, senior critical care attending at Cooper University Hospital and professor of medicine at Cooper Medical School of Rowan University. I spent my first week with ventilated coronavirus disease 2019 (COVID-19) patients; it was the week before last. Anywhere from 9 to 13 patients on my service. I go back to the intensive care unit (ICU) tomorrow and we’ll have a similar number of patients.

How is this different for someone who’s spent a lifetime taking care of critically ill patients, with special interest in sepsis? I think the majority if not all of these patients are septic. I believe that it’s humbling when I move from one room to the next, it’s just the same thing again, and again, and again, and again.

Yes, there is some heterogeneity in these patients. We’ve all heard about the L phenotype, with relatively good compliance, sometimes near normal, with not many changes on computed tomography (CT) and with severe hypoxemia. We’ve also heard about the H phenotype, which looks like our traditional acute respiratory distress syndrome (ARDS). I think the big difference that we clearly see is this L phenotype, where the hypoxemia doesn’t make any sense when you look at the CT scans. It’s postulated that this is related to vasoplegia and loss of hypoxic vasoconstriction. It doesn’t respond as predictably to proning as the traditional ARDS H-type presentation. But it may, and we’re autoproning those patients. We’re actually having the patients prone themselves. We’re using selective inhaled vasodilators in this H group, with variable results.

If the progression goes to the H type, with the more traditional atelectatic-looking lung and consolidation, we treat it pretty much the same as any other ARDS that we’re used to treating. The only difference, it seems, is that some of those patients do have this lingering vasoplegia.

Roman Jaeschke, MD, MSc: Thank you for this introduction. Let me introduce you briefly to our audience. Doctor Phillip Dellinger is the person who ran the first 3 editions of the Surviving Sepsis Campaign guidelines.

Let’s talk about the management. You mentioned the L phenotype, which is good lung compliance and lack of good explanation. You mentioned that you are using pulmonary vasodilators?

R. Phillip Dellinger: Yes, inhaled epoprostenol, inhaled nitric oxide, with variable results. It seems to improve the oxygenation in some patients.

We’re also proning the ones that are intubated as well as the ones that are not. It’s the so-called “cooperative proning” or “autoproning,” where the patient is asked to prone themselves and to stay on their stomach for as long as they can tolerate. Some of them are able to sleep on their stomach.

Roman Jaeschke: One of my colleagues, Waleed Alhazzani, who’s now quite prominently involved in the Surviving Sepsis Campaign, is actually starting a trial of self-proning of nonintubated patients. Maybe we’ll know a little bit more.

Any comments about potential thromboembolic phenomena responsible for this hypoxia?

R. Phillip Dellinger: It’s clearly the most prothrombotic sepsis that we’ve seen, or that I’ve seen. I think we first noticed it when we were putting our first early COVID-19 patients on continuous renal replacement therapy (CRRT) and the machine just clotted. Even heparin didn’t handle the clot with your CRRT.

Then we saw patients with pulmonary embolism (PE) and with deep vein thrombosis (DVT). And then we began to read autopsy data about fibrin, microthrombi, and the pulmonary vasculature. Perhaps a reason why some of these patients are difficult to ventilate is because of the associated dead space.

I wouldn’t say it’s a policy—although it may soon be a policy—but the great majority of attendings will fully anticoagulate any patient with a D-dimer level of 3 mg/L or more; or some people would say 3000 ng/mL, depending on what your units are.

Roman Jaeschke: We are gravitating this way. It would be fascinating to learn what people are doing around the world.

Thank you for this start. Any other thoughts on how this situation differs from what you’ve experienced or expected?

R. Phillip Dellinger: Certainly the L phenotype, where you see a little bit of “ground glass,” no consolidation, no atelectasis, but yet severe hypoxemia thought to be related to vasoplegia. The hyperinflammatory response with the very high ferritin and high C-reactive protein (CRP) levels, high interleukin (IL) 6—we don’t measure it, but some people do. There’s this hyperinflammatory piece, the procoagulation piece, both macrocirculation clotting and suspected microcirculation clotting. It’s something.

I think the one area that we’re most conflicted on is the anti-IL-6 therapy. That hyperinflammatory response would make you want to use it, but yet we’ve been trying to quench that hyperinflammatory response for 20 years, unsuccessfully, in traditional sepsis, so why would it make a difference here? We’re also concerned about getting rid of the virus and what effect giving an anti-inflammatory drug would have on the body’s ability to clear the virus.

Roman Jaeschke: You started to talk about anti-IL-6. There’s been a huge controversy. I’ve been talking to different people, there are completely different policies and patterns of practice. What’s your pattern of practice? I’m not even asking the evidence behind it. What are you doing specifically?

R. Phillip Dellinger: I like it for septic shock where you don’t have a secondary bacterial cause. If I think the cardiovascular instability is related to COVID-19 and I see that really high ferritin, that’s a patient where—at least this week—I want to give it to them until somebody publishes evidence-based medicine. And I know how much you love evidence-based medicine.

Roman Jaeschke: We have to make decisions without good evidence anyway. So you would use some IL-6 antagonists. Anything else that you are or were routinely using?

R. Phillip Dellinger: Currently we’re tending not to use steroids. Pretty much all of our COVID-19 patients in the ICU were on Plaquenil (hydroxychloroquine), azithromycin, and lopinavir/ritonavir. I guess the World Health Organization (WHO) study that was briefly up on their website and then came down apparently showed that Plaquenil did not appear to make any difference and there’s concern about cardiac toxicity. So it may not be in the regimen this week.

Roman Jaeschke: But maybe it will be next week.

R. Phillip Dellinger: Could be. You never know.

Roman Jaeschke: By the way, if you have patients with this H type, more classic ARDS, do you tend to use steroids?

R. Phillip Dellinger: I tend not to. These data have been so controversial, as you know, for the last 10 or 15 years.

Roman Jaeschke: We have more meta-analyses than primary studies, that’s for sure.

Maybe the last question. You were the leader behind the first few iterations of the Surviving Sepsis Campaign. Effectively, we were all learning from them how to treat people with sepsis. You are experienced in practice guideline development. Do you have any advice for people who today are producing those guidelines for COVID-19?

R. Phillip Dellinger: I think my recommendation would be that in situations where there is no evidence to inform decision about recommendations, the best thing is to just say we can’t make a recommendation and wait for data instead of trying to force it like a round peg in a square hole. You just have to say we need evidence and let it be what it is.

Roman Jaeschke: There are 2 schools of thought. One is saying, “People are desperately ill. Let’s use something pending the evidence.” And the other one is saying, “We have no evidence, so we can’t use it.” Where do you stand on this? It’s a matter of opinion, or almost religion. Where are you?

R. Phillip Dellinger: An example I’ve been using for my colleagues and my trainees is with hydroxychloroquine, which is part of our ICU admission protocol—maybe not for long after the WHO study we’re hearing about. I would tell colleagues and trainees that if I had to bet half of my life savings on whether that drug works or not, I would say no. But if I was in the ICU with COVID-19, I would want the drug.

Roman Jaeschke: I won’t tell you my opinion, but it’s not that far yet. The obvious question is if there’s harm behind it.

R. Phillip Dellinger: I would insist they meticulously follow my QT interval. But it looks like that drug’s not going to work. We’ll see after the WHO study is peer reviewed.

Roman Jaeschke: Do you know where it is submitted?

R. Phillip Dellinger: I do not. But based on what we’ve been doing with peer review and what I’m seeing, I think peer review is still thorough but it’s rapid.

Roman Jaeschke: Correct me if I’m wrong, because I’m not 100% sure if it’s the same study I’m thinking about, but this was still an observational study, correct? Or was it a randomized controlled trial (RCT)?

R. Phillip Dellinger: I think it’s an RCT.

Roman Jaeschke: OK. We’ll wait for publication. I don’t think we’ve been ever discussing studies before they are even peer reviewed. We’ll see what happens.

R. Phillip Dellinger: It’s a new time.

Roman Jaeschke: It’s been a tremendous pleasure talking to you. Thank you.