Bernard G Fikkers and Johannes G van der Hoeven
We thank Dr Dongelmans and coworkers for their interest in our study. They state that they are surprised by our high complication rate and that their complication rate is in accordance with those found in other series, referring to just two studies.
The first study, that by Polderman and coworkers , which employed the GWDF technique, found a major complication rate of 4.0%. The other study, that by Berrouschot and coworkers , in which the multiple dilator technique was employed, reported a 7.9% major perioperative complication rate, including one death (caused by tracheal laceration). Minor complications were not reported. It is difficult to believe that the patients in those two series suffered only from major complications and not any minor ones! We prospectively collected all our data and found major complication rates of 7.6% with GWDF and 5.3% with CBR. Because the difference between major and minor complications is important, we have decided for future research to categorize complications related to percutaneous tracheostomy as minor, intermediate and major (Fig. 1). Using these new definitions, the major complication rates in our series are 2.3% and 2.9%, respectively, because most major complications would be redefined as intermediate. Moreover, because we meticulously registered our perioperative complications, we are able to inform readers about all other complications they may encounter, although the majority is rarely clinically relevant.
We congratulate our colleagues from Amsterdam for their excellent results. We analyzed the available literature published up until 2002 and found that major complications varied from 0% to 14% (average 3.0%) in 28 studies (4066 patients) that used the multiple dilator technique; from 0% to 4.9% (average 3.0%) in six studies (461 patients) that used the GWDF technique; and from 1.3% to 5.0% (average 2.8%) in three studies (286 patients) using the CBR technique. We therefore feel that our results are completely in accordance with the existing literature. (For full details of our analysis and reference details, see Additional file 1.)