Do pulmonary artery catheters improve outcomes in cardiac surgery?

Welcome back to Depth of Anesthesia. This is a podcast that critically explores our clinical practices. I'm David Hao and I'm an anesthesiologist at the Massachusetts General Hospital. Joining us today is Dr. Taylor Thomas, a dual obstetric and adult cardiothoracic anesthesia fellow at MGH and recent graduate of our anesthesia residency program. Taylor completed the Depth of Anesthesia Podcast Elective and is a graduate of the Icahn School of Medicine at Mount Sinai.

Taylor, welcome to the show. Thanks so much, David. It's awesome to be here. I've actually learned a lot from this podcast over the years, so it's actually pretty exciting to be on this side of the mic and to have a chance to dive into one of these topics together. Absolutely.

I'm looking forward to it. And our second guest today is Dr. Matt Vanneman, a clinical associate professor in the Department of Anesthesiology, Perioperative and Pain Medicine at Stanford. Matt completed his anesthesia residency and cardiothoracic fellowship at MGH, where he also served as chief resident. Some of you might actually remember that Matt was our very first guest on the show. He joined us for an episode on the safety of administering cefazolin in patients with penicillin allergies.

Matt, it's great to have you back. Welcome. It's good to be back. It's really impressive what you've built here and decided to dive into this topic today. So our case today is that of a sixty-four-year-old male with a history of coronary artery disease.

Hypertension and diabetes who's presenting for elective coronary artery bypass grafting with cardiopulmonary bypass His preoperative evaluation was largely unremarkable, with an ejection fraction of 55 % and no significant valvular disease, no data on echocardiography. The surgical team has planned for a standard on pump cabbage with placement of three graphs. As part of the anesthetic plan, the residents suggest placement of an arterial line and a pulmonary artery catheter for hemodynamic monitoring. Taylor, what are some of the claims in this case? Yeah, so one of the central claims here is that using a pulmonary artery catheter during cardiac surgery can actually improve outcomes by giving us more precise human dynamic data to use to guide our intraop management.

So here now are two questions for our listeners to think about. What is the level of your agreement with the claim that routine use of a pulmonary artery catheter during cardiac surgery improves outcomes by giving us more precise hemodynamic data to guide management? And what is the level of evidence for what you believe? Taylor, how'd you actually get interested in exploring this specific subject? Well, David, during my training I noticed that pulmonary artery catheter use and cardiac surgery really varied.

Some of the anesthesia attendings use PA catheters routinely in select cases. Others almost never Some kind of said that it was a product of where they trained. They said that it was institution-dependent. But what really left me wondering was with so much variation in practice, what's the actual evidence guiding our practice? So interestingly, all the way back in October of 2013, over a decade ago, now the ASA released a list of quote specific tests or procedures commonly ordered, but not always necessary in anesthesiology.

And this was part of the ABIM Foundation's Choosing Wisely initiative. And so among these recommendations, the third stated: don't use pulmonary artery catheters routinely for cardiac surgery. in patients at low risk of hemodynamic complications, especially when alternative diagnostic tools like TEE are available. And to support this guidance, the ASA referenced a prospective observational study published in Anesthesia and Analgesia in 2011. It seems fitting that we begin our discussion here. So Taylor, could you walk us through the details of this study?

Yeah, of course. So the study titled Lack of Effectiveness of the Pulmonary Artery Catheter in Cardiac Surgery by Schwann and colleagues. It was published in Anesthesia and Analgesia in 2011 So this prospective observational study included a little over 5,000 patients undergoing coronary artery bypass grafting or cabbage with cardiopulmonary bypass. It spanned 70 institutions across 16 different countries between November 1996 and June 2000. So the patients that were included, they had to be at least 18, and they were scheduled for a cabbage with cardiopulmonary bypass.

The main exclusion criteria was pretty simple. It was just the presence of intraoperative transesophageal echocardiography. And PAC use was entirely discretionary based on clinician judgment. So of the total, 2,546 patients were analyzed after propensity score matching. And so they were divided into two cohorts of approximately 1,270-ish each, one with the PAC use and one without.

The predefined composite endpoint was the occurrence of any of the following. So death by any cause Cardiac dysfunction, so like myocardial infarction or congestive heart failure. Cerebral dysfunction, which included stroke or encephalopathy. Renal dysfunction, which included both renal dysfunction and renal failure, or pulmonary dysfunction, which was acute respiratory distress syndrome. They also looked at some secondary endpoints.

This included stuff like endotrope use, influid administration, as well as post-op intubation duration, as well as ICU length of stay All right, so after enrolling over 5,000 patients across 70 centers worldwide, the researchers, as you mentioned, split them into two match groups, one with PAC monitoring and one without. As you mentioned, it's important to note that PAC use along with all aspects of patient management was left entirely up to the medical teams. So no standardized protocols were in place. They used propensity score matching to make sure both groups were balanced and baseline characteristics. And with these groups lined up, the stage was then set to look at what actual impact PAC use might have.

That's exactly right. So the study found that PAC use was associated with worse outcomes across several important clinical endpoints. So major morbidity mortality was actually higher in the PAC group. at over 20% compared to about 15% in the non-PAC group. That meant that patients with PACs had almost a 70% higher likelihood of experiencing major complications or death. And among this, if you look at in-hospital mortality, it was actually doubled at like 3.5% versus 1.7% in patients with P.

A. catheter. Cardiac events, which included MI, congestive heart failure, or death from cardiac causes, were also more common in the PAC group. Cerebral events like stroke and cephalopathy were significantly higher as well. I think it was like an adjusted odds ratio a little greater than two. And renal events like renal dysfunction or failure also higher.

The adjusted odds rates was like 2.47. So PC patients didn't fare quite as well, but they also had longer ICU stays. I want to say it was like 14.5 % stayed more than four days compared to 10.1% in the non-PAC group. Their exturbation time was also prolonged. Then in terms of secondary endpoints PC patients received more inotropes on the day of surgery.

They also received more fluids and had a higher positive fluid balance. So taken together, these findings really suggest that While PAC monitoring provides detailed hemodynamic data, its use in the setting was not associated with improved outcomes. Matt, any thoughts about this particular study? Absolutely. So I find these findings to be pretty interesting.

So a few things. First, this is a large prospective cohort study. It's very large and international. which suggests that it's probably generalizable. The purpose of flupensity score matching is to try to reduce the risk of residual confounding By matching similar patients to one another. So in theory, the statistical methodology reduces the risk of residual compounding However, it's very important to remember that propensity score matching is only as good as the variables that go into the propensity score That means if there are unmeasured or unobserved variables that aren't included in the perspensive score and are confounding the association between homoyor decapitals and the outcome , Those wouldn't be controlled for in this type of analysis.

So it's important to remember that that type of problem or bias may remain even despite the author's best efforts. That being said, clearly there is no strong signal for help in this study or suggestion that a polymer capture is improving outcomes, which I think is notable. Another interesting aspect of the study is that there's no standardized hemodynamic management protocol. So everything was left to the patient discretion as it normally would have been. So that could mean that using the pulmonary artery catheter could lead to more frequent or different or more aggressive interventions, things like load administration or ox trop administration.

It seems unlikely that the catheter itself is resulting in these types of negative outcomes. But it must be, because it's a monitor, how the commissions are using that monitor and responding to the data that I've provided. So subtle changes in hemodynamics that might otherwise have gone unnoticed and probably been completely tolerated may have prompted interventions that instead could have tipped balance towards unfavorable blood. An additional strain to this study is it's a large-scale study where trans -esophageal echocardiography was specifically excluded. Now that's really important because David, as you mentioned, in the Choosing Wisley guidelines, transesophagyocardiography could in theory supplant pulmonary artery catheter use in many patients, which means this specific population, at least in theory, would have the greatest likelihood of benefit from a pulmonary artery catheter because transesophage wasn't used.

So, because of that exclusion, we can see specifically how the coronary calculator monitoring is associated with outcomes in isolation without the use of TE. So this should be the perfect setup to test the pulmonary artery catheter standing out by itself. But importantly, it doesn't really reflect contemporary trends where T is used almost routinely. Fantastic. So I think stepping back for a second, this does bring us to a bigger question, which is what does the broader evidence actually show about pulmonary artery catheter use in cardiac surgery?

And to answer that, we're going to move forward and look at one of the more contemporary analyses. This is a study by Brown and colleagues in 2022, which certainly adds to the evidence we've discussed and gives us some updated insights into this issue. Taylor, would you mind walking us through this study? Of course. So this study, like you mentioned by Brown and colleagues in 2022.

It included over 11,800 adult cardiac surgery patients treated at the University of Pittsburgh Medical Center between 2010 and 2018. So these were a broad mix of surgeries from isolated cabbage and valve surgeries to combined procedures like cabbage with mitral or aortic valve replacements. They also included both elective and urgent or emergent surgeries. They did exclude patients undergoing more complex operations like LVAD implantation, cardiac transplantation, aortic surgery for aneurysm or dissection, pulmonary endorterectomy, and other highly specialized procedures. This was an observational cohort study using a large, prospectively maintained database.

PC use importantly was not mandated by any institutional protocol, but was instead determined by the care team. And it was generally guided by factors like low cardiac output states, significant valvular disease, or reoperations. After identifying patients with and without PACs, the researchers performed propensity score matching to balance baseline characteristics between the groups. They also use multivariable logistic regression models to adjust for intraoperative variables when evaluating primary and secondary outcomes So to quickly summarize, this was a robust observational analysis with a large sample size and some thoughtful use of matching and regression to control for confounders. And the real strength here certainly is how it accounts for a broad real-world patient population and variable institutional practices.

Exactly. So after matching, they looked at over 7,000 patients, and there were about 3,500 with PACs and about 3,500 without. Intraoperative red blood cell transfusions were significantly more common in the PAC group, about 26% versus 23. Post-operative transfusions were also higher in that group at 40% compared to 35. And then ICU stays were also notably longer in the PAC group with a median duration of 48 hours versus 39 hours Despite these differences though in resource use, mortality was identical at 2.4% in both groups.

There were really no differences too in major complications like stroke, sepsis, or renal failure. And even when they adjusted for intraoperative variables like transfusions and bypass time, PHC use was not independently associated with operative mortality. whether they looked at the entire cohort or just at higher risk subgroups like patients with recent heart failure, mitral valve disease, or significant tricuspid regurge. Thanks, Taylor. I find this study interesting and how it sheds light on the broader implications of pulmonary artery capitals. And while it's reassuring that there is no increase ischemic time or risk of mortality There was a consistent finding that the polarity catheter was associated with increased intraoperative and post-operative red blood cell transfusion as well as longer ICU length of stay.

Now, the median ICU length of stay was saying by nine hours in the primary catheter group, which is of uncertain operational importance but is potentially notable because the polarity capital group may be receiving uh different vasoactive infusions and have different meaning strategies. Now, similar to the last study, these authors are using a propensity score matching strategy to try to reduce confounding, and then the second step of using multivariate regression It's important to remember that these regressions and propensity score strategies are only as good as the variables that are included in them, which means that it is still possible residual confounding may explain these results What stands out to me is how these findings suggest the bowernack artery catheters detailed hemodynamic data may be driving some intrinsic differences in management, or certainly at least be associated with them.

And it's not just the anesthesiologist interpreting numbers that the ICU teams and surgeons are also using these types of data to make clinical decisions. So these cascades of interpretation across multiple teams may be associated with more inventions, even if the data isn't necessarily resulting in better outcomes. I think that's a really important Matt. And it does take us to the next question, which is whether there are certain patient groups that might actually benefit from the added hemodynamic information a pulmonary artery catheter provides. After all, if the overall data doesn't necessarily support routine use, some clinicians might certainly argue that there could be value in more complicated or higher risk subgroups.

Exactly. And that's what this next study set out to explore. So in this study published in 2023 The investigators used almost 970,000 weighted records from the National Inpatient Sample spanning from 1999 to 2019. So the National Inpatient Sample, in case you're not familiar, it's the largest publicly available all-payer inpatient database in the U. S.

It captures five to eight million hospital discharges each year from a stratified sample of community hospitals. And just for the record, academic centers like ours is considered a community hospital per this definition. So what they did is they compared patients who received a pulmonary artery catheter with those who did not, adjusting for, again, baseline characteristics. And data were extracted on multiple outcomes, including in-hospital mortality and length of stay. So overall Adjusted analyses suggested that there was really no significant difference in mortality between the PAC and non-PAC group.

And interestingly, the PC group had shorter length of stay overall. It was like shorter by like a little less than half a day. But, you know, when you zoom into the subgroups, the picture changes a little bit. So for patients with congestive heart failure, PAC use was actually associated with higher mortality and longer length of stay. But by contrast, patients with pulmonary hypertension, mitral or tricuspid valve disease, or those undergoing combined surgeries They didn't show worse outcomes and in some cases they actually had shorter hospitalizations.

Yeah, so this is a really strong study. It is a very large, very widely generalizable, publicly available database. So there is nearly a million hospitalizations that are analyzed. This type of scale provides some element of strong generalizability. This study, similar to our other studies, did not find that the pulmonary catheter was associated with strong grouping.

And it's important to note that this study did suggest there may be a signal for harm simply in patients who had heart failure. I would, however, point out that a major drawback with a national inpatient sample is it's very difficult to grade the degree of severity of these different diagnoses. So it's uncertain if there was some residual confounding that may have been associated with both pulmonary catheters and and with the outcomes that may have explained this type of climate Nonetheless, it suggests that a pulmonary artery catheter may at least be associated with potentially negative outcomes in this small subgroup, although I would say further analysis would be required. We have to be careful with these conclusions, however.

Administrative databases really can't capture nuances like things like frailty, ejection fractions, and how clutents are reacting to polynomial recap data. Because the pulmonary artery catheter doesn't improve outcomes itself, it's important to understand how that information is being used, and these databases can't get at that type of mission activity. So, while the studies suggest that pulmonary R catheter use as routine use is not associated with an improvement in benefit, there is this potential signal that there's an association of harm with theStic heart failure patients Which then should make us ask, is this association of harm because we're selecting sicker patients to receive pulmonary catheters and therefore their baseline underlying illness? Is what is associated with the outcome Or is that there's something actually related to pulmonary catheter that's driving management, and that management actually may be harmful.

And that distinction is important to understand if this is just a finding of residual confounding or a true association. To the author's points, they themselves emphasize that prospective trials are needed. And that call for higher quality evidence is exactly where this recent systematic review and meta-analysis by wrong and colleagues come in. So they synthesize data across multiple studies, including some that we've already discussed, to evaluate PAC use in cardiac surgery more comprehensively in the hopes of addressing some of these concerns that we've raised. Yeah, that's right.

So they pulled together seven studies totaling 25,853 patients, the vast majority of whom were undergoing cabbage. Now, importantly, for our discussion, this meta-analysis includes some of the individual studies that we've already highlighted this session. So the primary outcome was really in hospital mortality. Across studies, PAC use was associated with a significantly increased risk of death with a pooled odds ratio of 1.57. Now, what's interesting is that this finding was consistent across multiple analyses.

It remained the same when each study was removed one at a time and Also, when the analysis was limited to patients with similar baseline characteristics matched between both like the pulmonary artery catheter group and the non-pulmonary artery catheter group For secondary outcomes, PAC use correlated with more intensive periop management and resource utilization Patients with PACs were more than twice as likely to receive intraoperative inotropes, and they experienced longer ICU stays, prolonged intupation times, and also higher hospitalization costs However, pulmonary artery catheter use was not associated with higher rates of a kidney injury, myocardial infarction, stroke, or infection, suggesting that the excess risk observed was not actually driven by these complications. Yeah, thanks, Taylor. So a meta-analysis is one of the finest quality ways to try to synthesize data together and to generate a single understanding from multiple studies.

And having this type of association and this type of effect size is like very strong and powerful signature. It's worth pointing out here that this is like multiple studies across multiple cohorts, which suggests that this is a reproducible finite. This association appears to be bringing many different places and probably is a generalizable association. But that's the key word. It's an association.

Six of the studies included in this meta-analysis are observational, which means there's always a risk of residual confounding if all the appropriate confounders are included in the main models. It's so possible that these patients received pulmonary artery catheters were just sicker in terms of their underlying baseline characteristics in unmeasured ways, and as a result, had more negative outcomes. Which may not really be related to the pulmonary artery catheter itself, but is actually just from the baseline sacro commodities. The one included a randomized trial was small and had higher crossover rates. So it's really hard to interpret that type of study if there's such a high crossover rate as make these data less reliable.

Most of the patients included in this man-analysis also were lower risk, they were elective cabbage patients, which means that these types of patients in this cohort will probably the least likely to benefit from aggressive hemogenamic monitoring chemical artery capital. And that context is fairly important when you consider how this association is going out. That's exactly right. And the authors do touch on that point because they suggest that one explanation for the increased morbidity and mortality is not the pulmonary artery catheter itself, but again, what clinicians are doing with the information it provides. So, for example, pulmonary artery catheter patients were much more likely to receive inotropes, and there's a body of evidence suggesting that excessive inotrope use can worsen outcomes in cardiac surgery.

Misinterpretation, incomplete use of that data again might have pushed management towards interventions that did more harm than good As a final note, it's worth mentioning a related study by Kadosh and colleagues from the Critical Care Cardiology Trials Network. This was a large multi-center analysis of over 13,000 admissions to contemporary cardiac ICUs, including nearly 4,000 patients with shock. So while the prior studies we've discussed have raised concerns about associations with harm, this study actually found that pulmonary artery catheter use in the cardiac ICU setting was associated with lower in-hospital mortality among shock patients, even after adjusting for patient and institutional factors with an adjusted odds ratio of 0.79. To my knowledge, this is the only study demonstrating a potential benefit, so certainly something worth spending some time on.

Matt, any thoughts on how to think about these findings? Yeah, so I have a couple of thoughts. The first one is the most important one, which you just highlighted, which is that to my knowledge, this is the only study that demonstrates an association of pulmonary R capital use and improvement blood So what's interesting about this study is that it's changing the focus of the patients from those that are originating from the operating room instead to those who are in the cardiac-intensive period. And the study assesses how the coronary artery catheters are being deployed in modern cardiac crater. So across 13,000 emissions at 34 centers There were about 3,800 patients which were presented in shock, and most of them were in cardiogenic shock.

The exact type of patient that may benefit from a pulmonary artery calculator to derive measurements that could guide clinical management of this shock. So not surprisingly, our captor use was the most common to patients with heart failure or those on mechanical circulatory support, which makes sense because those are the sickest. Most humodynamically complicated and the ones where the pulmonary counter -generated information might actually change clinical management that could hopefully improve the patient's outcome. What stands out here is there is a fairly strong signal that the hormonary artery catheter was associated with lower mortality. So among patients with shock, those managed with a pulmonary artery catheter are lower in hospital mortality of around 28% compared to those about 35% without a pulmonary artery capture.

And that difference remained even after adjusting for how likely each patient was to receive a pulmonary artery cathode in the first place. Or this is a version of propensity score matching, estimating the likelihood a patient would receive the pulmonary artery cathode treatment. So this suggests, at least in the cardiac intensive period unit setting, in patients with cardiogenic shock, the pulmonary artery capture may be associated with improved survival Which is a fairly strengthening contact to our surgical literature, where we really didn't see a strong association at that point. A critical point here is to note pulmonary artery cavities was widely variable across different institutions. While overall, 46% of the patients in the cohort who had shock received a pulmonary artery capture, that range widely.

From as low as 8% in some centers to 70% in others So it's hard to say for sure if it was a pulmonary catheter itself that was associated with this benefit, or if there was some other unmeasured variable for example, that were intrinsic to each center that had high PA catheter usage that may have helped the patient have a better outcome, that may have been independent upon the area. And so it's important to remember that residual confounding in these types of retrospective studies still is in play and certainly can modulate the association results. Absolutely. Super important points and a very important study to consider in the context of our discussion today. So we've reviewed a number of studies.

We've gone through our meta-analysis and systematic review. And before we wrap up, I do want to pull together some of the big picture takeaways from this episode. So Taylor, any closing thoughts for our audience? Sure, David. I think that the main theme across all these studies is really that routine use of pulmonary artery catheters and cardiac surgery, especially for those like lower risk populations who are growing things like isolated cabbages, really it doesn't appear to improve major clinical outcomes like mortality, stroke, renal failure.

And this was really shown across multiple well-designed observational studies and even when they used propensity matching to try to balance out patient characteristics I do think that the meta-analysis from 2024 really reinforces this point because it's showing that not only in association with mortality, but also greater use of inotropes, longer ICU stays. prolonged intubation times and higher costs associated with PACs, while finding really no reduction in complications like AKI, MI, or stroke So interestingly though, we do see an association with pulmonary artery catheter use and a consistent increase in resource utilization So things like more transfusions, longer ICU stays, and increased use of inotropes and fluids. And I definitely think that this can reflect a more vigilant approach to hemodynamics.

But at the end of the day, the data just seems to suggest that it doesn't translate to meaningful clinical benefits. So overall, while I I think that pulmonary artery catheters can provide detailed hemodynamic data The available evidence really suggests that in at least routine cardiac surgery and particularly for like isolated cabbages The additional information derived from these pulmonary artery catheters doesn't really lead to better outcomes. So this really underscores the importance of using pulmonary artery catheters selectively. Rather than as sort of a routine tool for every case. Fantastic summary, Taylor.

Any final thoughts, Matt? Yes, I think it's important to highlight that most of the surgical data that we've talked about today, as well as most that's in the literature that I'm aware of, really does not show an association of pulmonary artery catheter use with patient BEP And there is this small potential signal for possible harm, although it's hard to know if that's related to residual confounding compared to the pulmonary artery capture itself. It's important to note that there's a different story in these medical populations that are presenting in cardiogenic shock to the cardiac intensive curum A recent study from the Critical Care Cardiology Trials Network did find a potential association for pulmonary artery catheter use and Improved survival amongst patients with cardiogenic shock.

That said, this interpretation does require caution. And so a local practice and culture may have a stronger influence than specific standardized criteria that are being used from the culminary Catholic. So while this potential associative signal of pulmonary artery can have to be used for patient benefit in carriaging shock is intriguing, it is uncertain if that's a true causal finding or if that's just due to other factors I'd be interested to see if future studies looking at this in a randomized fashion might find any difference, and I hope that those studies are done in the future Thank you so much, Matt. I just want to really say a very sincere thank you to Taylor for all the work that she's done in preparing this been participating in the Depth of Anesthesia podcast elective.

Taylor, thanks so much for joining the show. Oh well thanks for having me David. As we mentioned at the top of the show, Matt was our first guest on the show. And I remember at the time we were talking about cefazolin and penicillin allergies. And at that time It was extremely common for clinicians in our institution to be giving alternative antibiotics.

And now fast forward maybe five years or so. I talk to the residents nowadays and it's just standard of care for them to give cefazolin, irrespective of penicillin allergies. And I just think it's so interesting to look at how time shifts clinical practices. And I just want to say that I think this kind of work continues to drive these things forward. So to our listeners, thank you for listening.

And to Matt, thanks so much for joining the show again and providing Providing your valuable insights. Thanks so much. We're having you, David. It's always a really good time. To our listeners, thanks so much for joining the show today.

As always, stay hungry and keep asking questions.