The much anticipated revised guidelines for chamber quantification have been released and feature in the January publication of JASE. This is a massive undertaking and I want to start my review by acknowledging the efforts of the writing group – I can appreciate how difficult the task is! In today’s post I will aim to draw attention to some notable changes from the 2005 paper and share some of my opinions on these changes – it is not aimed to be a complete analysis of the document.
The 2005 paper is arguably the most significant echo paper published and the revised edition overall has improved significantly on this document. There has been a large growth in the amount of normal data available for the authors to base guidelines and the revised paper reflects the technological advances which have become routine clinical practice in many centres (i.e. real-time 3D and Strain). As an educator, the goal of a guideline paper is to reduce variability in how to measure and interpret findings. It represents the “go-to” document that summarises the various different techniques and papers available. This guideline revision has improved significantly on the previous guidelines, however I feel that several new challenges have arisen.
The most obvious change is a departure away from partitioned values for many of the measurements. The majority of measurments are reported as a mean+-SD and merely classified as either normal or abnormal. The further grading of mild, moderate, or severe is left to the individual institutions (or individual sonographers) to develop their own reference ranges. The authors go to great lengths to explain the justifications for this philosophical shift, citing statistical inaccuracies which arise from a non-Gaussian distribution of the data in physiological systems as a major limitation to most methods. This makes sense, as does the argument that a prognostic-significance approach to classifying degree of dilatation is equally flawed (we saw that when the LA volume data was published in the 2005 guidelines causing widespread confusion amongst many echocardiographers). The method for determining cutoff which makes the most sense is to draw on the vast knowledge of the writing group to arbitrarily draw some lines in the sand. The authors state that they didn’t see this as a scientifically rigorous method, however to me it is the one that makes the most sense. The writing group doesn’t seem to be suggesting that we change our reporting to a binary system of normal/abnormal, but they have given a free reign for labs to make up their own ranges. This seems a far less scientifically rigorous approach then at least laying down some expert opinions to use as a guide. After all that is the whole point of guidelines. If I compare two labs, one a tertiary referral hospital, and the other a small outpatient clinic performing predominately normal scans. The scale of mild, mod, severe is likely to be skewed to the right in the tertiary hospital and to the left in the outpatient clinic. Our biases based on our sample population will skew the qualitative grading, rather than having an objective set of numbers to compare with (if you see 100 ventricles all measuring less than 5cm, then a 5.5cm LV walks in, you are going to think it is huge in comparison, where as the tertiary centre that routinely sees 7cm+ LVs is not going to blink at a 5.5cm. I will be interested to see how this tracks over time – certainly an area for future research…
LA dimensions. Thankfully the new guidelines have been a lot stronger in discouraging linear AP dimensions as the sole measurement used for assessment of LA size. Planimetry of the LA area from the A4C view alone is also discouraged, in favor of biplane LA volume. This was always going to be the case, and I think the sooner we all move to measuring the same thing the better. I am not clear why the authors chose to include partitions for indexed LA volume, when the shift away from this in other measurements had been so obvious (and this wasn’t explained in the paper). It was great to see the cutoffs for normal shifting from 28 to 34 ml/m² – I think that alone will make a lot of people happy. I am also happy to see a deliberate ommision of gender specific ranges, with the authors citing that although there is a recognised difference in LA volume between males and females, the majority of this is corrected by indexing to BSA.
Curiously there is still discordance when comparing cutoffs for LV volumes to LV size. A cutoff of 120 ml is a very small LV volume. Using a Teicholz formula this would equate to an LV linear dimension of less than 5cm. Studies looking at LV opacification with contrast or looking at LV with other techniques (ct and MRI) all show the LV being a lot bigger than this in normal populations. A very interesting study for an enthusiastic student would be to look at how LV size is classified in patients using the different techniques…I suspect if we were using solely the LVEDV (indexed or absolute) from properly measured Simpsons traces, there would be a significantly higher proportion of abnormal classifications then if we use any other method.
Aortic measurement technique! This was the section I had the biggest hopes for. Nearly all of the measurements in echo have moved to a blood-tissue interface technique as it has the best reproducibility. It also makes sense that we measure the internal dimension of a chamber rather than including a variable space in the leading tissue interface. The authors acknowledge this, and given that it would correlate better with MRI/CT they had hoped to move to an inner edge to inner edge technique. The concern was based around whether this would lead to some patients with a dilated aorta not falling into prognostically significant cutoffs which may prevent them from having surgery. I consider this quite disappointing. Surely the error of our measurements is almost as great as the difference proposed (remember we have an axial resolution of 1-3 mm). This concern didn’t seem to hold back the authors of the aortic stenosis guidelines dropping the cutoff of severe AS from 50-40mmHg…
The aortic measurements section was not all disappointment. The authors did a really nice job of illustrating incorrect measurement technique for the aortic annulus. They also acknowledge the two different measurements we can obtain of the aortic root – an area we are starting to appreciate a lot more.
The revised guidelines have included strain and the use of 3d and contrast imaging as useful tools for the assessment of chamber quantification. This a big advancement from the 2005 paper. I also really like the overall layout of the document, particularly the approach of explainging the various techniques, providing a justification of the writing groups thoughts and then proving recommendations. This is where the expert opinion is able to be employed, and I would love to see more of this. The paper is easy to read and easy to find measurements thanks to a small table of contents on the front page.
It is easy for me to sit back and be critical of the paper, and I truly hope that my thoughts aren’t misinterpreted as a revolt of the guidelines. Quite the opposite. I sincerely acknowledge the difficulties the writing group face with this task, and on the whole, I think the document is fantastic. I hope that the next document in another 10 years sees a revision of LV volume values, and measurement techniques for measuring the aorta. I also hope that we see a return of partioned ranges for measurements. There is an unbelievable amount of experience/skill/knowledge in the collective writing group and I would love to see some further standardisation in how we report – the authors should take comfort in the knowledge that an arbitrary value arrived at by experts which forms the basis of our reporting is a lot more scientifically rigorous than just “leaving it up to the labs”.
So here is the call to action: WE need to drive the future direction of our profession. We can do this by writing quality papers on normal ranges and partitioned values. It may not sound like “sexy” research, but it forms the basis of ALL our scans. I would argue that it is THE sexiest research as it has the biggest impact clinically. We need to answer questions such as, “Does a shift away from partitioned values impact on reported qualifiers?” and “What impact would measuring inner edge to inner edge really have on patient outcomes?”. There are countless research questions coming to mind from this paper, it is up to us to answer them. I know I will be!
Subscribe to be updated on new Echo.Guru posts
Image courtesy of Stuart Miles at FreeDigitalPhotos.net