Although the authors mention in their letter that our results are partly in contrast with their own findings using side-stream dark-field imaging during cardiac surgery, we suggest that both studies show rather similar alterations during cardiopulmonary bypass (CPB) for cardiac surgery. We found a transient 10% decrease of functional capillary density during CPB,1and den Uil et al. reported a reduction of microvascular perfusion index and an increased proportion of patients with impaired flow during CPB that normalizes within the first hours after surgery.2
The authors point out that difference in analysis routine exists: Dr. den Uil et al. used a semiquantitative analysis3to calculate the microvascular flow index, whereas we assessed the microcirculatory parameters of erythrocyte velocity, vessel diameter, and functional capillary density using Cap-image (Dr. Zeintl GmbH, Heidelberg, Germany). We do agree that automatic software to analyze routines must be well selected; however, Cap-image software is extremely well validated and has been frequently used in animal models using intravital microscopy. In the current study, all parameters were analyzed manually: diameter by drawing a vertical straight line from one vessel wall to the other, erythrocyte velocity with the use of a line-shift diagram, and functional capillary density by marking all visual perfused capillaries in a selected video sequence. Moreover, we recently found a good correlation between functional capillary density analyzed with this technique and the procedure described by De Backer et al. 3giving a correlation coefficient of 0.868 (our unpublished data, May 2007 comparison study of the analysis procedures as used by De Backer et al. and our analysis routine). Using this Cap-image diameter and erythrocyte velocity could not be analyzed in all microvessels because in this two-dimensional optical method, not all vessels in one region of interest are in focus. It has been discussed by Lindert et al. 4that using orthogonal polarization spectral imaging with a European phase alternating line (PAL) video standard, blood flow velocities can only be measured up to a maximum of approximately 1,000 μm/s, a fact that restricts its use in studies on arteriolar perfusion. In our experience, erythrocyte velocity in venules and capillaries (these are the type of vessels that can be mainly visualized with orthogonal polarization spectral imaging in sublingual tissue) is usually below 1,000 μm/s.
Regarding the patient selection and the inclusion criteria in our study, we indeed selected low-risk, elective patients and excluded any emergency or high-risk procedures. Therefore, postoperative morbidity and mortality were low; however, logistic European System of Cardiac Operative Risk Evaluation (EuroSCORE) was not assessed. We consider this reasonable to establish the normal range of microcirculatory changes during CPB in uncomplicated cases. Of course it would be most relevant to assess any possible relation between intraoperative microvascular hypoperfusion and postoperative outcome. However, it seems sensible to first characterize microcirculatory changes during uncomplicated operations. Although interesting, single case reports on patients with impaired microvascular perfusion during CPB and poor outcome do not substantially increase our knowledge regarding the role of optical methods in microcirculatory monitoring in cardiac surgery.
Finally, although De Backer et al. have defined a cutoff at 20 μm for the diameter of small and large microvessels, this will not enable a differentiation between venules and capillaries as suggested by Dr. den Uil et al. , because many smaller venules will certainly have diameters lower than 20 μm, especially taking into consideration that orthogonal polarization spectral imaging and the side-stream dark-field technique will contrast erythrocytes only, and this will underestimate vessel diameter by up to 5 μm.5
Further studies will have to show whether—as in septic patients—these optical methods are able to predict patients’ outcome and can be influenced by therapeutic approaches.
*Ludwig-Maximilian University Munich, Munich, Germany. firstname.lastname@example.org