What is the infectious risk of arterial catheters?

Welcome back to Depth of Anesthesia. This is a podcast that critically explores our clinical practices. I'm David Hao and I'm a fellow at Brigham and Women's Hospital. Our first guest today is Dr. Benjamin Weaver. Dr. Weaver is a PGY1 resident in the MGH Anesthesiology Residency Program and a graduate of the Boston University School of Medicine.

I'm so excited to announce that this is actually the first episode we're recording that was produced through the new podcast selective at MGH. Dr. Weaver did a masterful job combing through the literature, and I know our listeners are going to be very intrigued by the findings. Dr. Weaver, welcome to the show. Thanks for having me, David. Our second guest today is Dr. Alyssa Walsh.

Our listeners may recall that Dr. Walsh joined us for episode 10. in which we explored whether a subjective assessment of functional capacity was predictive of perioperative complications. Doctor Walsh is a current fellow in the surgical ICU at MGH She's also a former chief resident and graduate of the MGH Anesthesia Residency Program. Dr. Welsh, welcome back to the show. Glad to join the show again. So our case today is that of a 45-year-old woman who's admitted to the SICU after unexpected blood loss during a laparoscopic hysterectomy.

On initial evaluation, the team determines that she will require an arterial line for hemodynamic monitoring while she's in the intensive care unit. The resident starts preparing for the procedure and asks their attending why it's not customary to use full barrier precautions with arterial line placement as compared to central line placement. The attending responds that arterial lines have lower rates of infection and thus do not require the same level of precautions. Dr. Weaver, what are some of the claims in this case? I think there are a few here.

The first one being pretty obvious, and that being that arterial lines have lower rates of infections compared to central lines. The others are that barrier precautions are an effective means for reducing the incidence of Line infections. And then lastly, that full barrier precautions do not reduce arterial line associated infections when they could. So here now are two questions for our listeners to think about. What's the level of your agreement with the primary claim that arterial lines have lower rates of infection compared to central lines?

And what is the level of evidence for what you believe? Dr. Weaver, how'd you get interested in exploring the subject? Well, um coming into the elective, we discussed a few things to talk about on the podcast, but when we went over the subject, I remembered throughout medical school and then early in residency. really just seeing a stark difference in the approaches to placing central lines and arterial lines. All my experience has been with central lines being placed in a very regimented, very sterile full uh barrier precaution manner, but very frequently I would see arterial lines placed in a rather cavalier manner, you know, quickly with only gloves and no real other barrier precautions. So I thought this would be an interesting topic to explore and see what the differences were.

Yeah, and that's pretty consistent with my experiences as well. I've had many attendings claim to me that the air is a high flow system and thus unlikely to get infected. And oftentimes when I've noticed a breach in sterility, the overall attitude is to proceed. Doctor Walsh, I wonder if that's been your experience as well. Totally been the same thing.

I've gotten the same uh sort of explanation. Oh, it's a high pressure system. There can't possibly be an infection that gets in except in very rare circumstances Even today I was in the ICU when we were talking about a patient with bacteremia of an unknown source and the responding clinician made the statement like, oh, of course it can't be the arterial line, like they don't get infected So I think that's definitely an attitude that I've seen uh pretty pervasively. Well I think that speaks to the importance of exploring the evidence on this topic. Funny enough, just about a week or two ago I actually saw a post on Reddit exploring common dogmas in anesthesia and one of the dogmas listed was that arterial lines must be done with sterile technique.

In fact, one of the comments, which actually got 17 upvotes, says, quote, I never got the arterial line one. Who's ever seen an infected arterial line? The arteries are a high pressure, high flow system So interestingly, we actually have some formal survey data that's explored antiseptic technique and the perceived infectious risk of arterial lines. Dr. Weaver, what can you tell us about this study? Yeah, this is so this is mostly a survey as opposed to some sort of retrospective or you know prospective study, but this was done in 2015 in the uh Journal of Critical Care Medicine.

And they pulled their subscribers to try and determine some of the feelings and beliefs towards arterial line and their infections. Essentially, they sent an anonymous twenty-two question survey to over eleven thousand clinicians, including doctors nurse practitioners, respiratory therapists, and nurses. And in this study, they had some pretty interesting findings. They found that only 44 % of respondents reported compliance with current CDC guidelines for arterial line insertion. And those are minimal barrier precautions where the entire patient is not gowned as we do in central lens and the insert the uh the individual inserting the arterial line was not fully draped or gowned, I should say.

They also found that there's a pretty significant underestimation of arterial line infection rate the mean and median infection uh estimates were 0.3 and 0.1 per 1,000 catheter days when uh and as we'll talk about this later when the CBC reference rate for this is actually 2.2 per catheter days. So the average respondent was almost a order of magnitude off. And then um only 15% reported using Full barrier precautions when placing arterial lines, thirty-nine percent supported mandatory full barrier precautions for arterial insertion And then lastly, just as a very specific anesthesia thing, anesthesia respondents were the least in favor of mandatory full barrier precautions. Well looking at this data, I'm not surprised that only fifteen percent reported use of full barrier precautions, but I am pretty surprised that only forty-four percent of respondents reported compliance with CDC guidelines, which, as you discussed, includes sterile gloves, surgical cap, surgical mask, and a small sterile drape.

Dr. Walsh, as a fellow in the ICU, how have you thought about this risk and how have you seen some of the beliefs around the space evolve over time? Yeah, I think, you know, I share the same surprise with the survey data. I certainly know, as we've already alluded to, there definitely is a more laissez-faire attitude with the arterial lines. And I've almost never seen full barrier precautions used, except outside of arterial access for mechanical circulatory support. However, I was surprised it was only 44 % for the typical CDC guidelines Although when I think back to it, I do think that as you alluded to before as well, David, I do see, you know, sometimes there's a small contamination or an emergent placement.

And that may result in lower adherence or contamination, which may be what we're seeing reported in this survey. I often think that this is justified by the fact that we often don't really the arterial line infections, which I think really ultimately represents either a lack of follow-up, such as from anesthetic care. or from lack of identification of the arterial line as a potential source of infection, such as in the case that I discussed earlier that we saw today. That makes a lot of sense, and I think that's one of the pervasive challenges with our practices in anesthesia is that a lot of times we don't see the downstream implications of our intraoperative or perioperative decision making And I think one of the core issues that I really want to start looking at the evidence on is whether or not arterial lines are a significant source of infection, because I think that really underlines this entire conversation.

As we've already touched on a bit in the survey data, the general perception is that arterial lines are not a significant source of infection. But I wonder, Dr. Weaver, what did you find in the literature? So well I guess there there does seem to be quite a bit of data that with regards to the incidence of arterial line infections, but I ended up kind of anchoring on this study by O'Horo et al. in 2014 that was published in the Critical Care Medicine Journal The study looked at uh meta-analysis of many different studies, including eleven hundred and fifty-three studies that met the eventual inclusion criteria. but they were assessing infection rates in arterial catheters used in both post-surgical care and critical illness.

Notably, they did exclude studies that were not in those settings, such as those with ECMO. And if I'm understanding this study correctly, the studies were only included if they had a clear definition of arterial catheter-related bloodstream infection. that included correlation of a catheter tip culture to a separate blood culture in a patient with signs and symptoms compatible with sepsis and no other source of sepsis had to have been identified. Is that right? That's correct, yeah. Got it.

So what did the authors find? Most notably, they found that the pooled incidence of catheter-related bloodstream infections was 3.4 per 1,000 catheters or 0.96 per 1,000 catheter days. They found that there was an increased risk of infection for femoral uh arterial lines with a relative risk of 1.93 compared to other insertion sites. And then they found notably when they tried to assess the sterile practices throughout the different studies that uh many different sterile practices were uh were used throughout. There was one where there were non-sterile procedures all around.

Ten studies were only with sterile gloves. And then there were another 23 that had only down, cap, mask, and a full barrier and a full barrier drape. So that's to say actually full barrier precautions in those 23 studies. Okay, so a pooled incidence of 3.4 per 1000 catheters or 0.96 per 1,000 catheter days with, as you mentioned, a fair amount of variability in terms of sterile practices across studies. Dr. Walsh, how do we put these numbers into context?

So the literature seems to suggest that centraline infection risk is somewhere between 0.9 to 3.4 catheter-related infections per 1,000 catheter days So given that this meta-analysis demonstrates a rate of 0.96 per thousand catheter days, this suggests that arterial line infection rates fall within the same spectrum as central line infections And thus we should consider this as a significant issue. Absolutely. And I think this is data that I wasn't familiar with prior to Dr. Weaver's research into the subject, but it's definitely eye-opening. And Dr. Weaver, one of the things you observed is that many of the studies used catheter line colonization as a surrogate for infectious risk. This was specifically identified in the limitations section of the meta-analysis, in which the authors observe, quote, the practice of drawing blood cultures from arterial catheters makes differentiating colonization from catheter related bloodstream infection difficult and limited the number of studies which we can include I I think it's important to establish whether this is indeed a reasonable strategy.

I think this is a limitation in some ways, particularly with regards to formulating some meta-analysis assessing arterial line infectious risk. But in other ways I think this is actually a powerful tool to use when assessing the risk of these, the the the absolute risk of these lines. And this was studied in a paper by Rinders et al. in 2002. where they looked to see if there was an association between catheter tip colonization as a surrogate and endpoint for catheter-related bloodstream infections. In this study, they looked at 29 different studies for a total of 60 study groups. with 1,281 catheter colonized catheters and 259 catheter-related bloodstream infections for a total of 94,595 catheter days In the study, the mean catheter colonization and catheter-related bloodstream infection incidences were 13.5 per 1,000 catheter days and 2.7 per 1,000 catheter days, respectively Um and what the study found was that there was a really good correlation between catheter colonization and catheter related bloodstream infections with an R value of 0.69 and an R squared of 0.48 for a p-value in their study of less than 0.001.

And I'm not uh I don't claim to be statistically savvy enough to really parse through this, but it did seem to be a pretty well done study And I think it does lend credence to the fact that we can actually use this as a proxy measurement for what the actual risk of arterial lime infections is Okay, so in this study that aggregated a total of sixty study groups, there was good correlation between catheter colonization and catheter related bloodstream infection, which gives us some evidence that colonization is a reasonable surrogate. This methodology is definitely repeated in future studies, so I'm glad that we were able to find this study. Now, taking all of this into consideration, let's head into some of the literature that actually compares central line and arterial line infections head to head.

Yeah, so as we mentioned before, it does appear that uh from the meta-analysis that arterial line infections are comparable uh with regards to risk of a line associated infection when compared to central lines Helpfully, there was a study by Lucid et al. in 2010 in the Critical Care Medicine Journal, which directly compared the rate of infections between these two different kinds of lines. The study was initially designed to assess the effects of dressing changes on lime colonization and infections, comparing dressing changes every three or seven days, and then comparing dressing changes with or without chlorhexine and impregnated sponges. In this study, maximum barrier precautions were used in the placement of both arterial and central venous lines, and there were a total of 3,532 catheters for a total of 27,541 catheter days In total, there were about uh actually there were exactly six uh 1,617 arterial lines and 1,915 central venous lines Doctor Weaver, I just want to clarify one aspect.

It sounds like the original study was designed to look at the effects of dressing changes on Lyme colonization and infection. The subsequent study we're discussing here was a post hoc analysis. Is that right? Yeah, absolutely. So maybe not the best study to assess this, but it did seem to be one of the only ones I found that was directly comparing the risk of infection in these two lines with similar placement techniques Okay, perfect.

What were some of the findings based on this study? I guess we can break this down by um catheter type. So in arterial lines, they found that there were uh 11.4 colonizations per 1,000 catheter days They found that there were 0.99 major catheter related infections per 1,000 catheter days. They found that there was a daily hazard rate for colonization of 1.3 % on day five and two point four percent on day ten. 3% on day 15 for uh for all these arterial lines. In the central venous group, they found that there were 11.1 calmizations per 1,000 catheter days. 1.09 major catheter-related infections per 1,000 catheter days, and then a central venous catheter daily hazard ratio for colonization of 1.2 on day 5, 1.6 on day 10, and 1.4 on day 15.

Yeah, so in summary, there doesn't seem to be any difference in the rates of colonization or major catheter-related infections, and then there doesn't really seem to be any difference in the absolute number of infection c infectious cases either Excellent. So my takeaway from this is that based on this post hoc analysis of a study that originally randomized to different dressing change intervals and use or lack thereof of chlorhexidine impregnated sponges, we find no difference in terms of the risk of colonization in catheter related infection between arterial lines and central venous catheters Dr. Walsh, any other thoughts about this study? So this study was likely highly generalizable to the ICU setting, given that it was a large number of patients with a mixed medical surgical population.

And as we talked about in the prior study, catheter-line colonization has been shown to be a reasonable surrogate for infectious risk. So what we're seeing from this study is that the colonization as well as catheter-related infection ratio does seem to be similar between arterial and central lines. Of course, one important factor for this study is that all patients did receive maximum barrier precautions, which does bring us to a more pragmatic study. Yeah, I would also say interestingly, I did come across another study that actually looked at prospectively to see what the difference was between arterial line and central venous catheter infection rates This was a study also published in the Critical Care Medicine Journal by Co. et al. in 2008.

And this was a 24-month prospective study in Australia at one of the at an Australian teaching hospital The population study was a 95% adult population, which excluded post-cardiothoracic surgical patients and acute neurosurgical cases but otherwise was uh open to most surgical uh most surgical cases. In this study they actually looked at um all short-term intravascular devices including arterial lines and central venous lines inserted in the ICU as well as in the operating theaters and in the emergency department at this hospital. Arterial lines in this study were placed under, quote, strict aseptic technique. end quote, um, which included large sterile drapes, sterile gloves, and chlorohexidine for skin preparation. Central venous catheters, on the other hand, were inserted with maximal barrier c precautions, including sterile gloves, gown large drapes, mask, and cap.

So there is some discrepancy in the insertion techniques here, but I think they do better mirror the actual insertion practices in the United States and for how we do at our current hospital Dr. Weaver, how did this study approach the issue of sampling catheters and what were some of the definitions that they ended up applying? I think they did what ends up being probably the most uh elegant solution and probably the also the most uh labor-intensive one for sampling these catheters. And in this study, what they did is they removed their catheters under sterile technique and immediately transported them for uh quantitative tip culture sampling. And then to assess for Lyne infections, they would culture the tips and then do blood cultures and try to cross-reference those two to make sure that they isolate from the tip culture and the blood cultures were the same Got it.

So what did they end up finding? The best way to look at the study results, I think, are the same way looked at the prior and differentiating the different results, bad catheter, type. In this study, they ended up having 321 arterial lines cited in 252 patients for an average length of catheter of 3.4 days and a total of 1,082 catheter days For the arterial lines, they found that 5.3 % of them ended up having a colonized catheter. And they found that there was a colonization rate of 15.71 and a bloodstream infection rate of 0.92 per 1,000 catheter days. They also found that there was an insignificant trend towards higher colonization rates in femoral lines when compared to all others.

And they also found that the colonization rate of arterial lines was higher when inserted in the operating room or in the Department of Medicine when compared to the ICU. With the rights of the central lines, there were a total of 618 central lines cited in 410 patients for an average line length of 6.5 days per central line. There was a total of 4,040 catheter days. On this arm of the trial, they found that 11% of the catheters were colonized at the time of removal, and they found that there was a colonization rate of 16.83 and a bloodstream infection rate of 2.23 per 1000 catheter days. Alright, so in terms of takeaways, we have a prospective study that gives us data suggesting that the colonization and catheter related bloodstream infection rates for arterial catheters and central venous lines are relatively similar.

Colonization was fifteen point seven one for arterial lines compared to 16.83 for central lines and the catheter related bloodstream infection rate was 0.92 in arterial catheters and 2.23 in central lines per 1000 catheter days. Doctor Walsh, any thoughts about these findings? Yeah, so I think in our prior study we were looking at central lines and arterial lines that were placed both under maximal barrier precautions, and we saw no difference in either bloodstream infections or colonization rates And now we have this study where the central lines and arterial lines were both placed under their typical manner in the United States with maximal barrier precautions for the central lines and the strictly septic technique. Arterial lines, and we still are not seeing a difference in the rates of either colonization or bloodstream infection.

It is important to consider the Hawthorne effect as a modulator for infectious risk because people do know that they were being evaluated for infectious outcomes. Which probably meant that they had a better adherence to sterile techniques, particularly when you think about our earlier survey data regarding practices around arterial line placement I do think it's also interesting to think about other factors that govern rate of infection, such as post-insertion, catheter, management, and maintenance. Overall, I thought that this is a very interesting study that specifically addressed the question of infectious risk under more typical practices for placement. Those are definitely some important points to consider. And I think on the basis of the findings of these two studies, we do have to ask a fundamental question, which takes us back to our second claim that barrier precautions are an effective means of reducing the incidence of Lyon infection.

Doctor Weaver, do we have evidence that increased barrier precautions are necessary or Put another way, would additional precautions reduce the infectious risk of arteriolines? Yeah, so to my knowledge, actually, there's really only one study that looks at this with any clarity. And that's a randomized control trial that was done in 2003. This was also completed by Renders et al. And in this study they prospectively randomized patients in the in a medical ICU at a large university hospital uh in a one-to-one ratio to have arterial lines placed uh with sterile barrier precautions including sterile gloves, a mask, a cap, a sterile gown, and a large sterile sheet And then to a standard of care group where only hand washing, sterile gloves, and skin disinfection with chlorohexidine and 70% alcohol was used The end point for the study were arterial line tip cultures in which strains were cultured and identified only when there was greater than 1,000 colony forming units per catheter, and they matched the positive blood cultures that also resulted.

Well, first off, I'm fairly impressed the study exists, so props to the authors for conducting it. I'm also somewhat surprised this isn't a study we talk about more. I think the standard of care group is actually fairly pragmatic. The only difference that I can think of with our practice at MGH is that we use a small sterile drape or towels on top of everything else they mentioned. Doctor Weaver, what were some of the findings from the study?

So I found the results of the study pretty interesting also, and I think they kind of get to the the question that we led the podcast with with regards to barrier precautions actually being effective. In the end, they had 272 total arterial line tip cultures for the analysis and a total of 2,343 cat third days. There were 129 patients in the sterile barrier precaution group and 143 patients in the standard of care group. The study was stopped after the results of only one interim analysis that were performed. Notably, there was a relatively large dropout rate that was noted for 101 of the randomized patients with catheters were not available for the final analysis In the end, they found that no differences were actually seen in the incidence of arterial line colonization when compared between the full barrier precautions and the standard of care group.

So definitely some interesting results and I do want to point out that the study only included peripheral arterial catheters, specifically those in the dorsalis, pedus, and radial artery. And I'm not aware of any study to date that has examined this for other arterial line locations. Obviously this is a small study, but it does seem to address the fundamental question we've had about barrier precautions. Dr. Walsh, I do want to get your thoughts about the author's decision to discontinue the study at the interim analysis. Yeah, so it was kind of an interesting aspect of the paper.

So um As we all know, the statistical power calculation tells us how many patients are required to correctly reject the null hypothesis, or in other words, to correctly identify an effect that's actually there In this paper, they identified that 3,650 catheter days were required to identify a difference between the arterial line and central line cohorts. And they had planned an interim analysis. When they did that interim analysis, what they showed was that basically the chance of finding a 50% or greater difference between those cohorts would have been vanishingly small, like under 1.5 % if the trial had been continued until the predefined recruitment numbers were attained. And thus, because there was such a low chance that they were going to actually find the effect, that was why they discontinued the study.

Another interesting point that the authors bring up in their discussion section is about whether a trial with sufficient power to demonstrate 50% reduction In catheter-related bloodstream infections would even be feasible. Such a study would require two groups of 1,100 people for a total of 9,000 catheter days. With all this in mind, I want to pivot briefly to explore the origins of full barrier precautions for central line insertion I think this could definitely be an entire episode on its own, but it ties in nicely with what we're discussing today, and I think it's going to be interesting to see the evidence on which such a strong standard of practice is built. Dr. Weaver, tell us about your findings.

What I found really kind of stemmed from looking at where the actual guidelines for full barrier precautions are most clearly written. And those are in the 2011 CDC Guidelines for the Prevention of Intrabascular Caphetor-related infections. In those guidelines, they advocate for the use of full body drapes sterile gloves and gowns, masks, surgical caps, and full skin preparation to prevent uh central venous catheter infections. Um and to support these guidelines, they cite four studies The first of these studies is a non-randomized pre-post-observational trial in which third-year medical students and PGY-1 residents were given a one-day course on infection and control practices. They found that there was increased perceived need for full sterile drapes, an increased use of full-size drapes, and a decrease in infections per 1,000 catheter days from 4.151 to 2.92.

The second study that they cite in these guidelines was a review article about the infectious complications of swan gain's pulmonary artery catheters, which I could not find access to. The third study was a small randomized controlled study in Minerva Anestologica, which is an Italian journal, where 82 patients were admitted to a mixed medical surgical ICU. and had central venous catheters placed with either maximal barrier precautions or control precautions, which were a mask, cap, sterile gloves, and a small drape. In this study, they actually weren't looking at in bloodstream infections or colonization rates, but rather just the rates of skin colonization. In the abstracts, the results showed an increased risk of skin colonization in the control precautions with an odds ratio of 3.4.

The last study is really the only randomized control trial that they cite, and that actually assessed for a catheter colonization and rates of bloodstream infections. So in this last study, this does seem to be the fundamental paper which uh advocates for central venous catheter insertions with full bearing precautions. In this study, 402 adult patients in an outpatient oncology setting at a cancer clinic were randomized into groups where either central lines were placed with full barrier precautions or minimal barrier precautions like we do. with arterial line placement. In this study, 12 catheter colonizations were found in the control group and four were found in the increased barrier group. There were 0.5 per 1,000 catheter days in the control group and 0.08 in the increase barrier group.

So overall, this paper does paint a picture in which Full barrier precautions do limit the number of bloodstream infections caused by catheters, but I would say that this is kind of limited in the general inpatient medicine setting given that these were typically in patients who were receiving immunosuppressive therapy and with catheters in place for a much longer time than the average study you would assess in a in a normal inpatient population. I think all this together paints a picture that I think is far less convincing than I would have expected given the rigid adherings that we actually have to but full barrier precautions in the insertion of central venous catheters, particularly in opposition to the arterial lens Really, really interesting stuff.

And I'll make sure these studies from the CDC guidelines are included in our show notes and on the website for anyone who might be interested in diving into this a bit more. It's always interesting to me when we look back into the historical context from which some of these practices originated and oftentimes we find that the evidence is quite fragile and these papers and evidence then seem to be propagated indefinitely. So I'm glad that you actually took the time to look into the evidence on which these CDC guidelines were based. I really enjoyed working through some of these interesting articles today. Dr. Walsh and Dr. Weaver, can you leave our listeners with some key takeaways?

So the infectious potential of arterial lines is often neglected, and I think from all the literature that we've gathered here, we've seen seen that arterial lines by a number of metrics have an infection rate that's likely on par with that of central line catheters Yeah, I definitely agree with Dr. Walsh. It seems as though, based on the published evidence, we are pretty much systematically underestimating the infectious risk attributable to arterial lines. That being said, I'm not actually in that the evidence surrounding barrier precautions, full or otherwise, is as convincing as I would have assumed Awesome. So heading back to the claims that we discussed at the top of the show, here now are two questions for our listeners to think about.

What is the level of your agreement with the claim that arterial lines have lower rates of infection compared to central lines, and that barrier precautions are an effective means of reducing the incidence of line infection? And what is the level of evidence for what you believe? Dr. Weaver and Dr. Walsh, thank you so much for joining me on the episode today. I thought this was a wonderful topic to tackle that really has immense clinical relevance. Thank you so much for having me.

It was great to be on the podcast and take the elective with you. Thank you, David. Glad to be back. Thank you to our listeners for joining me on the show today. If you enjoy the work that we do here, please share this podcast with your friends and colleagues.

Our mission is to challenge clinical dogma for the sake of improving patient care. And we hope that you'll join us on this mission. As always, please send us your questions and comments to depth of anesthesia at gmail. com. I'm also super excited to announce that Dan Saddawi-Konefka and I are going to be presenting a live episode at the upcoming IARS meeting. It is fully virtual now, so if you're planning on attending, please stop by our session.

Stay hungry and keep asking questions. So the first thing is about it, supposed to be split, the worst things by the worst things were.