Introduction:
Treatment-induced hemolysis is a rare complication in present-day hemodialysis, however when it does occur the results are often extremely serious. It has been recently speculated whether hemolysis would have characteristic effects on relative blood volume changes measured by different on-line measuring techniques. It was also noted that ultrasonic measurements such as those used in the BVM (Blood Volume Monitor, Fresenius
Medical Care, Bad Homburg, Germany) would be insensitive to primary effects of hemolysis but that characteristic effects would be expected with optical techniques such as those used in the CritLine-III (HemaMetrics, Kaysville, UT). Therefore, the aim of this study was to address the characteristics and the magnitude of free hemoglobin interfering with optical measurements in a series of in-vitro studies.

Material and methods:
In our in-vitro studies we used bovine blood at an O2-saturation of more than 98%. The blood was anticoagulated with EDTA-K2. The composition of blood circulating through a small blood circuit at approximately 200 or 400 mL/min was constantly measured by optical means using the CritLine-III device (HemaMetrics, Kaysville, Utah) with its sensor clip attached to the specified blood chamber. Actual blood flows were measured by timed blood volume collections. Each experiment was done with approximately 500 mL of blood using different baseline hematocrit levels obtained by mixing red blood cell concentrate and plasma. To this blood, 50 mL of hemolyzate were added in steps of 10 mL to reach maximfree hemoglobin concentrations in the range of 2 to 3 g/dL. CritLine readings were recorded after complete mixing.

Results:
Nine dilution series using whole bovine blood, covering a baseline hematocrit (H0) from 30 to 60% were done. Final optical measurements were overestimated (hematocrit) or underestimated (relative blood volume) in a clear linear relationship of the same magnitude. Similar linear relationship was observed when measurements
were done at high blood flows.

Conclusions:
The results obtained in this study confirm the expectation that optical measurements of hematocrit would be affected by hemolysis. However, taking a closer look, it becomes clear that the magnitude of the relative hematocrit error is small for free hemoglobin concentrations in the range of 0.1 to 0.2 g/dL which are usually associated with the onset of clinical symptoms.