HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Menicanti, L. A. Search for Related Content PubMed Articles by Menicanti, L. A.

Eur J Cardiothorac Surg 2006;29:63-64
© 2006 Elsevier Science NL

Editorial comment

Divisione di Cardiochirurgia, Policlinico San Donato, Via Morandi 30, 2007 San Donato, Milano, Italy

^_* Tel.: +39 02 5277 4521; fax: +39 02 5277 4327. (Email: menicanti{at}libero.it).

The article entitled ‘The Italian CABG Outcome Study: short-term outcomes in patients with coronary artery bypass graft surgery’ deals with the first Italian National Survey of CABG outcome and represents a very important step in the management of the public health in our country. The authors applied an elegant statistical analysis on outcome, but unfortunately they published the results too early because the data generating the statistical model are not completely realistic. There are such biases in the analysis that make the elegant statistical model totally useless. I wish to make the following comments:

(A) The study was voluntary and the participating centres were asked to present data covering at least 6 months of observation. We know from the literature that a good and reliable representative period of observation should last more than 2 years!
The period of observation was established from January 2002 through 30th September 2004, and each centre was free to choose the period, according to the preference of participants, therefore the range of the observational period was extremely wide (6–29 months). This first bias should be sufficient to consider the study as unreliable according to all guidelines for observational study.

(B) The definition of risk parameter was not well detailed and the reported rate of risk factors is, therefore, too high in respect to the normal biological variation. For example:

Cardiogenic shock is reported from 0.2% to 7.2%. This is completely unrealistic according to that reported in the literature (1%). It is evident that the definition of cardiogenic shock differed from centre to centre!

Diabetes: some centres reported an incidence of 46% in their population! This is completely out of the normal frequency (15–25%) in a population with coronary artery disease undergoing CABG.

Pulmonary disease: the reported rate of COPD ranges from 2.4% to 35% (normal rate is from 4% to 7%).

Obstructive peripheral arterial disease: some centres reported a rate as high as 40% (normal rate from the Euroscore database is 11%).

Unstable angina: this condition varies from 3.5% to 62% in the reported data! In the most important international database the incidence of unstable angina in patients submitted to CABG is around 12%.

Emergency: last but not least, this condition varies from 0.2% to 14%! This is one of the most altered risk factor reported in the study. In the STS and Euroscore databases, emergency is reported to vary between 1% and 4%.

It is very clear that mortality cannot be adjusted for risk factors that have such a high variation!

Moreover, the clinical monitoring was not as rigorous as it should have been in a well-designed trial. Very few records were reviewed by a monitor, and some centres, those activated in the last period, did not even receive one single visit. No centre received a second visit.

This study produced a ranking of centres which is unique according to all published data and the authors should have stressed it in the conclusions. In fact, the best performing centres (n = 8) are those with a low volume of work with a mean number of procedures per unit of 273 per year and only one centre is performing 425 cases per year! More interesting is the fact that all the centres performing more than 500 cases per year (n = 7) are in the low rank; two of them are in the lowest rank. It comes out from the study that the best performing centre has an activity volume of 146 cases per year; this centre does not have an Emergency Department and is not working during the summer! These data are in contrast to all published data.

In the conclusion, the authors observe that the adjusted mortality dropped from 2.8% in 2002 to 2.4% in 2003–2004. The difference is not statistically significant but nevertheless they consider it as a relevant improvement towards a mortality of 0.9% (mortality obtained by the eight centres classified as low-outliers) considered as the gold standard for the Italian population.

We did a simulation in our patient population: if we refuse surgery to the patients with a logistic Euroscore >40 we reduce our mortality by 50%! Thus, the best way to achieve the totally arbitrary ‘gold standard mortality’ of 0.9% is to refuse surgery to the sickest patients.

Finally, the way these results were published is astonishing: the media reported the rank of the adjusted mortality of the participating centres without any consideration or comment or a word of caution from the Institution conducting the survey about any possible error in ranking. The observed mortality was not reported! We know the incredible importance of statistics but we also know that statistics is a unique science, in that it has the ability of resuscitating patients or vice versa; for example, in the group of patients of Centre no. 23 there are 16 patients in good shape that are dead for statistics and there are six patients from Centre no. 15 who are dead but for statistics they are in perfect shape!

I think I have sufficiently explained the reasons why this paper should be read with extreme caution. The readers of the journal should be aware of the fact that the study suffers from too many biases and that the results are therefore totally unreliable.

This article has been cited by other articles:

				G. G.L. Biondi-Zoccai, M. Fusaro, and L. Inglese Potentials and pitfalls of clinical outcome research studies in cardiac surgery. Eur. J. Cardiothorac. Surg., May 1, 2006; 29(5): 855 - 856. [Full Text] [PDF]

				F. Seccareccia, C. A. Perucci, D. Fusco, and P. D'Errigo Reply to hekmat et Al. Eur. J. Cardiothorac. Surg., May 1, 2006; 29(5): 857 - 858. [Full Text] [PDF]

				L. A. Menicanti Reply to seccareccia et Al. Eur. J. Cardiothorac. Surg., May 1, 2006; 29(5): 859 - 860. [Full Text] [PDF]

				K. Hekmat, U. Mehlhorn, and T. Wahlers Do we need the new Italian risk stratification model for CABG patients? Eur. J. Cardiothorac. Surg., May 1, 2006; 29(5): 856 - 857. [Full Text] [PDF]

				F. Seccareccia, C. A. Perucci, P. D'Errigo, and D. Fusco Concerning the Editorial comment by Dr Menicanti. Eur. J. Cardiothorac. Surg., May 1, 2006; 29(5): 858 - 859. [Full Text] [PDF]

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Menicanti, L. A. Search for Related Content PubMed Articles by Menicanti, L. A.