An Unscripted Debate at WoCoVa 2016

I just returned from WoCoVa 2016 in Lisbon. What a meeting! Representatives from over 60 countries, 1000 participants and so much interest in improving all aspects of device safety - from selection, insertion, care and maintenance. My heartiest congratulations to the organizers for arranging such a fantastic scientific session.

On a personal note - it was also nice to finally meet many of the people I have corresponded with via email, putting faces to names and shaking hands after all these years. I also appreciated the many different aspects of vascular access in the world. It's fascinating to me how the same device is often inserted by different people, cared for in different ways, and viewed upon variably based upon these and other factors. For those of you that have never been to WoCoVa, I highly encourage it -- it is refreshing and opens your eyes to important differences in our practice.

One of the highlights of the meeting for me was debates with key leaders regarding the evidence. At many sessions that I attended, there was  a palpable anxiety regarding the evidence base and whether or not it applies to current practice. Most notably, I heard the revered Drs. Pittiruti and Lamberti (both of whom I have tremendous respect for), speak about systematic reviews and meta-analysis and the limitations of current data. While I understood their concerns, I also was vexed regarding their interpretation of these studies that have been published in top-tier journals. In particular, they discussed our recent meta-analysis regarding PICC DVT published in the Lancet. I had the opportunity to challenge some of their assertions in an unscripted mini-debate. Here, in summary, are the key issues brought forward and my responses:

1. Meta-analyses of observational data is not appropriate or valid.
Response: This is simply not accurate. There is a robust science regarding how to use data from observational studies in meta-analyses, including guidelines and a checklist of best practices. Meta-analyses of observational data are responsible for key learnings, including those of global climate change and it's effect on species, the association between tobacco use and cancer, as well as the recent and Zika virus outbreak by the World Health Organization. These analyses are also extremely hard to do - much harder than analyses of randomized trials. And it's not as if RCTs are the silver bullet to solve all of our problems -- in fact, quite the opposite, as so nicely outlined by this author. They are therefore KEY to understanding and improving our knowledge about key scientific questions. Top tier journals, including the Lancet, understand this and have very stringent requirements to ensure such studies are well done and adhere to the highest standards of practice. In fact, our review was independently reviewed and met stringent scientific criteria to be included by DARE, an organization that reviews the quality of systematic reviews.

2. The studies put together are extremely heterogeneous,  do not reflect current practice.
Response: It is true that the studies included in our analysis span over 50 years and practice with PICCs has changed considerably. For instance, Drs. Pittiruti and Lamberti cite use of ultrasound as a key advancement that influences safety. While I  understand these concerns, there are important caveats. First, the key to a well-done systematic review is the use of an explicit protocol that includes all, not some of the evidence. Including only a few studies or focusing on those that (in the readers assertion) are the most representative (which is what they chose to do) is the key to introducing selection bias, the very thing systematic reviews were designed to counteract. Second, it is important to remember that systematic reviews cannot, by definition, improve upon the quality of the evidence. They are a representation of what is out there - and well done reviews are an important clue to a "signal" as to what current evidence suggests. Finally, it is always important to look at evidence holistically, not just from a single studies point of view. In this context, it is worth noting that the signal that PICCs might be more thrombogenic than other devices was not simply suggested by our review - but also by those of Saber and colleagues, Pikwer and colleagues, and also by Johanssen and colleagues. Is it possible that all of these are flawed. Possibly. But I doubt it. 

3. The data do not support greater risk of thrombosis associated with PICCs than central venous catheters.
Response: This assertion, unfortunately, occurred with some misrepresentation of the data from our studies and those of others. First, only certain studies from our analysis of 62 were picked and chosen to fit the bill for this argument based on ultrasound use, a modern era of PICC placement, etc. Asides from the huge concerns regarding selection bias here, I think it is wrong to pick studies only because insertion practices were better. I realize this is very important when it comes to thrombosis, but ultrasound represents but a fraction of  time for a PICC. What about the rest of the life of the device? We know that PICCs are more likely to develop malposition than other devices (this was conceded by the speakers as well), and since tip malposition is a key factor for thrombosis - how does an ultrasound at insertion at the time prevent this in the future? Second,  none of the numbers presented had confidence intervals around them. Confidence intervals are the key to understanding effect sizes -- in fact, in meta-analysis, they are FAR more important than the actual point estimate. When you actually look at the confidence intervals of the data points used by the speakers, they completely overlap those of the conclusion of the study. Third, concerns regarding including asymptomatic and symptomatic thrombosis and inclusion of both of these data were not correct -- we understood these are separate entities and did separate them in subgroup analysis, only to come up with the same conclusion. I will add that there are several new studies that suggest PICCs might be more thrombogenic, including a recently published RCT, a large study of DVT in medical patients, and a study from Italy by Dr. Bertoglio who also spoke at the conference. Again, remembering the totality of the evidence -- is it possible that these are all wrong? Maybe, but again - I doubt it. 

Take Aways:
1. Please note that I am obviously conflicted being the writer of this document and of the Lancet paper.  I am also an empiricist, not a rationalist (more on this topic here). I believe good science requires data and that all decisions must be made taking this data into consideration. This was not what was happening at this session, prompting me to take on two leaders in our field whom I have learned much from.  I think it is important to respect the evidence we have. In a field that struggles to find any type of evidence, we simply cannot afford to throw out the baby with the bathwater. There are patients in all of these studies that might have been hurt by the choices we providers made when placing PICCs. It is simply not right to assume all of the data is inaccurate given this important pretext. Absence of evidence is not evidence of absence. 

2. I do not want you to walk away from this seeing this post as more rhetoric and verbal jousting. Rather, regardless of our differences,  I share with Drs. Pittiruti and Lamperti the need for better data in helping make this assertion. Given the effect sizes and numbers involved, only a large, likely international, multi-site trial can give us the answers we need when it comes to thrombosis risk between these devices.  I see our role in leading just such a study, not arguing over the data. I was happy that we agreed on this -- and that we will work towards such a study in the years to come.

3. In the end, we agreed to disagree. But everyone reading this should know that this was a huge win for our field. This type of debate - by people who are passionate, knowledgeable and able to communicate in a non-defensive, conflict-free way - is precisely what vascular access needs. At the end of the day - the specialty won. And that, my friends, is what is most important. 

I look forward to your thoughts regarding the session (if you were there) or this post.