[HSF] how others will see us

[Tea Acuff]

As mentioned earlier in a thread about bias, evidence and aprotonin I have thoughts 
in referrence to http://stroke.ahajournals.org/cgi/content/37/11/2759 (In Stroke, 2006, "Off-Pump Surgery is associated with reduced occurence of stroke and other mordidity as compared to traditional coronary artery bypass grafting..." by Sedrekyan, Wu, ...Treasure)
I don't read much of our literature, but am occasionally shown articles by those trying to interact with us and learn a lot. I found this an interesting perspective for those that seek to know us. In general I believe the findings to agree with my past decade of beating heart experience. (I think "off pump " is a simplistic, yet commonly discussed, benchmark for patients even if I do that as my standard.) I try to do only what is required. I, perhaps like the authors (is that why it is in "Stroke"), have stoppped trying to change surgeon's minds, although there may be reason to reexplore these issues. The discussion does point out the fact that the RCTs tend (as do more so the cardiology papers) to pick the "better" patients of which I think there is general agrement will have less benefit with a particular technique, but let's just deal with the findings.
To wit:
1) 255 abstracts were narrowed to 41 RCTs by two authors for a meta analysis of 3996 patients, not all endpoints were so represented as patients available for mortality were 3996 but patients for reintervention were only 1660
2) significant differences favoring off pump were stroke 0.6% vs 1.6% (p=.03), afib, 16.9% vs 24.8% (p<.001), wound infection  3.8% vs 7.7% (p<.001), MI 2.6% vs 3.4% (p=.26) renal failure 1% vs 2%(p=.26)
3)favoring on pump: reintervention (9 trials only) 2.4% vs 1.2% (p=.08) cross over 7.9% vs 1.6% (duh!) (p=.08)
4) neutral: mortality 1.6% vs 1.6% (p=.89)
5) there were less grafts per patient in the off pump (-.26) but this varied inversely by study size of the number of patients per group with the "large" groups >50 (-.15) and small <25 (-.61). also studies with much higher on pump mortality (heterogeneity) than the norm were excluded.  
 Interesting the authors also did elsewhere a meta analysis of aprotonin as statistical analysis appears to be an interest of theirs.

My reflections:
This is a (very) good risk group: mortality = 1.6%
Surgeons are the least able to be objective about the data, and with good reason. If it takes 10 or more years to be good at something and you plan to do surgery anyway, it is stupid to change unless you think it is worth 10 more years of your life (or 5 if you a are a quick learner). It is not like changing types of stents that can be learned in a couple of weeks or even 50 cases. I personally think off pump skills take >500 cases to be similar to on pump (assuming that you have done 500 on).
If the main reason more surgeons don't adopt off pump techniques is for concern of poor anastomoses, why do on pump techniques have more MI's?
How many surgeons would prospectively trade 1% MI intraop for 1% reintervention postop? (I did not note that angina was LESS common in the postop off than on pump patient by the away which may be a futher bias...the "we did all we could"...)
One of the keys to understanding this data is to think about what I claim EBM (evidence based medicine) is about. It is the study of the behavior of doctors to therapy as much as the patients response to our behavior.
Here is how I logically think (I may be completely wrong, but it is my logic) about how off pump can have less MI, equal (? if we use these numbers as real world) mortality, and more reintervention. Surgeons over time become comfortable with their ability to do "necessary" anatomoses. All things being equal, which only is imagined to happen in a study not the real world, it is easier to do tinier and more diseased vessels with the heart still (intra surgeon not inter surgeon). More intra surgeon MI's come from stupid choices than stupid technique. (It may well be that surgeons with bad technique have more MI's , but this will not show up in randomization.) Just because you can sew it does not mean you should sew it. That is, surgeons adjust to their techniques ( if they have any brains at all) to their comfort which is in no (necessary) way similar to the needs of the patient. OFF and ON pump techniques are like bulls in a china shop. Just beacuse one is smaller and can go
 down a smaller isle doesn't mean it will do less damage automatically.
Reasons for less stroke are both simple and impossibly complex.
Reasons for less afib are so far and remain largely the newest conjecture.
Ultimately the reasons for less morbidity of any kind is that IF you can accomplish what you need, less is always a better technique than more. 
 If you can't and didn't (do what was needed), you may have to go back earlier  (reintervention) than you thought which is a judgement call.
If we don't stop reading papers for statistics and think about what they mean, we (and our patients) are dead in the water.

Tea