Endogenous gas formation of carbon dioxide used for wound flooding--an experimental study with implications regarding gas microembolism during cardiopulmonary bypass.

BACKGROUND: Gas microembolisation is an identified risk in cardiac surgery. Flooding the wound with carbon dioxide is a method proposed to reduce this problem. The high solubility of carbon dioxide is beneficial, but may also cause problems. The gas solubility diminishes at warming and endogenous bubbles are formed when cold blood saturated with carbon dioxide is returned by cardiotomy suction. METHODS: The release of endogenous gas was measured at high resolution in an experimental digital model. A medium (water or blood) was incubated and equilibrated with gas (100% carbon dioxide or air) at a low temperature (10 °C or 23 °C). The temperature was increased to 37 °C and the gas release was measured, at rest and at fluid motion. RESULTS: The amount of carbon dioxide released at warming was substantial for both water and blood (both p=0.005). The effect was more pronounced when the temperature differential increased (p=0.005). However, blood and water differed in these terms: with water, the release of carbon-dioxide started instantly at warming; with blood, carbon dioxide remained dissolved and was released at fluid motion. When blood was warmed from 10 °C to 37 °C, the gas release corresponded to 44.4% (40.6/46.5) of the medium volume (median with quartile range). CONCLUSION: Gas dissolved in a medium becomes released at warming, as confirmed here. Blood exposed to carbon dioxide became heavily oversaturated at warming, with the gas instantly released at fluid motion. The amount of contained gas increased with a higher temperature differential. Our study has relevance to wound flushing, using carbon dioxide, in cardiac surgery. The clinical consequences of these findings remain to be answered.

Influence of pericardial suction blood retransfusion on memory function and release of protein S100B.

BACKGROUND: To study the influence of pericardial suction blood (PSB) on postoperative memory disturbances and release patterns of protein S100B during and after cardiopulmonary bypass (CPB). METHODS: Sixty male patients admitted for coronary artery bypass surgery were prospectively randomized to receive PSB either by using conventional cardiotomy suction retransfusion or after cell-saver processing. RESULTS: The concentration of S100B rose during the period of CPB from 0.065 +/- 0.004 to 0.24 +/- 0.001 microg/L (p < 0.001). PSB contained 18.0 +/- 1.7 microg/L of S100B. Direct retransfusion from the cardiotomy reservoir made the systemic level increase to 1.42 +/- 0.19 microg/L compared to 0.25 +/- 0.02 microg/L using a cell-saver. Signs of postoperative memory dysfunction ( > 1 SD) were discovered in one of three tests, but were unrelated to technique of retransfusion. No associations were found between serum concentrations of S100B and memory function. CONCLUSION: In this study, retransfusion of PSB during cardiac surgery appeared not to cause memory disturbances. PSB contained high concentrations of protein S100B making its use as a marker of cerebral injury unsuitable.

Inflammatory mediators in autotransfusion drain blood after knee arthroplasty, with and without leucocyte reduction.

BACKGROUND AND OBJECTIVES: The aim of this study was to evaluate whether leucocyte-reducing filters influenced complement activation and the formation of pro-inflammatory cytokines in autotransfusion drain blood after knee arthroplasty. MATERIALS AND METHODS: Twenty-three patients undergoing knee arthroplasty were divided into two groups. All patients were given salvage blood postoperatively. In Group A, a leucocyte filter was connected between the wound and the drain blood container. In Group B the drain blood was not leucocyte filtered. Complement split products and cytokines were analysed in circulating blood and in drain blood, together with blood-cellular differential counts. RESULTS: Drain blood showed activation vs. venous blood, with elevated concentrations of C3a, SC5b-9, interleukin (IL)-6, IL-8, polymorphonuclear (PMN) elastase and tumour necrosis factor-alpha (TNF-alpha) (P<0.05 to P<0.001). The leucocyte filter reduced TNF-alpha (P<0.01), but triggered complement activation (P<0.05). Room-temperature incubation increased the concentration of IL-8 (P<0.01), which was seen in both venous and drain blood. The leucocyte filter prevented formation of IL-8 (P<0.01). In drain blood at 24 h the inflammatory reactions accelerated (P<0.05-0.001), although the filter reduced the leucocyte counts and TNF-alpha concentrations. CONCLUSIONS: The leucocyte filter reduced IL-8 and TNF-alpha in drain blood, but at the same time triggered complement activation. Incubation affected the inflammatory spectrum of both drain blood and control venous blood, and the filtering reduced this activation.

Theoretical and experimental aspects of erythrocyte filterability testing; flow acceleration and systemic resistance.

Blood cell filterability is an established method in blood rheology. The dynamics at flow onset and its relevance to the data interpretation is, however, not fully known. This paper aims to investigate what controls the length and slope of flow acceleration as the medium accelerates to reach the steady state, and how this phenomenon may interfere with the data output. The acceleration time was not constant. With buffer the steady-state flow showed a logarithmic correlation (p<0.05) versus acceleration time and a linear correlation (p<0.001) versus acceleration slope. With 5% erythrocyte resuspension the steady-state flow instead demonstrated a linear relationship versus acceleration time (p<0.001) and no correlation versus acceleration slope. A cut-off timing of 0.6s is suggested to avoid artifacts associated with flow acceleration. The possible influence on data interpretation from the flow channel systemic resistance was also addressed, and found to significantly underestimate measurable changes in erythrocyte properties from unprocessed flow curves. This was despite the traditional correction for blank filtration flow. Both acceleration and effects from systemic resistance do probably have minor influence on the historic data interpretation but could perhaps be considered in the methodology to sharpen the data output.

Impaired contractility of left-over vein grafts used for CABG and the possible trauma caused by air exposure.

OBJECTIVE: The patency of vein grafts limits the long-term results after bypass surgery. Vein graft trauma may contribute to the clinical results. DESIGN: Left-over pieces of vein after routine coronary bypass surgery were compared with references of the same vein that were stored with maximum care. Blood perfusion (20 h, 37 degrees C) was also tested to evaluate possible recovery. Vessel rings were measured by isometric tone recordings. RESULTS: Left-over veins showed a contractile dysfunction compared to reference (p < 0.01-0.001, n = 13), and that did not recover after blood perfusion. Four trauma models were tested, using saphenous vein graft (SVG)-references (n = 5). Chemical exposures (0-90 min) to NaCl-papaverin or cardioplegic medium affected the baseline (p < 0.05-0.01). With mechanical distension, up to 450 mmHg leak testing, an increased norepinephrine (NE) reactivity appeared. Air contact, 20 min, reduced the NE reactivity to half maximum (p < 0.05). Longer exposures (40-60 min) completely abolished the contractility. CONCLUSIONS: The saphenous vein smooth muscle function appeared permanently traumatized following routine harvesting and handling. Air exposure gave dramatic deterioration and could in theory contribute to hamper the long-term results after bypass surgery.

D-glucose additive protects against osmotic-induced decrease in erythrocyte filterability.

Glucose has long-term effects on erythrocyte filterability owing to sorbitol accumulation and glycosylation of intracellular proteins, but glucose effects prior to these long-term metabolic consequences have not been studied to the same extent. D-glucose is osmotically inert in erythrocytes because of facilitated diffusion. Erythrocyte volume regulation, hemolysis and filterability were studied with reference to effects of glucose derivatives dissolved in water. Control situations were water alone or buffer. Salt-stock dilution by water, D-glucose, or 3-O-M-glucose in water (both 570 mmol/L additives) created volume increases of 4.0 +/- 0.2%, 3.4 +/- 0.5% and 3.3 +/- 0.2%, respectively, whereas L-glucose resulted in a 2.2 +/- 0.4% volume decrease (all p < 0.001 versus baseline, mean values +/- SEM). D-glucose at a final concentration 30 mmol/L did not display any osmotic properties. Despite erythrocyte swelling, electrolyte-free glucose did not induce any significant changes in filterability with either 3-microm or 5-microm filters. 3-O-M-glucose gave a 7.8 +/- 1.4% decrease in 3-microm filterability (mean value +/- SEM, p < 0.001), but not to the same extent as by water alone at the corresponding dilution rate (31.7 +/- 3.5%, mean value +/- SEM, p < 0.001). L-glucose caused a 2.0 +/- 0.8% decrease in filterability across the 5-microm (mean value +/- SEM, p < 0.05), but not with the 3-microm filters. Stressing the glucose transport system further, to simulate the water-like properties of electrolyte-free glucose during intravenous infusion, we found that glucose becomes osmotically active and prevents hemolysis at these extreme concentrations (> 114 mmol/L). We conclude that an in vivo concentration range of D-glucose protects the erythrocytes from a decreased filterability induced by osmotic swelling but without having any osmotic properties.

Narrow-pore flow behavior of blood cells in cardioplegic suspension.

Cardioplegia alters the ionic composition of the myocardium and also the blood in a way that may influence the cellular capillary flow behavior. We measured changes in RBC volume and narrow-pore flow resistance of blood cardioplegia versus crystalloid medium. Potassium, magnesium and sodium as osmotic control caused an expected cell shrinkage and reduced the flow resistance through 3 microm pores; however, stressing the osmosis further resulted in increased resistance. No major effects were seen with the 5 microm filters. Twenty percent blood cells in the cardioplegic medium caused a 360% increase in 5 microm pore resistance. There were no obvious additional filterability effects of the cardioplegic additives other than their osmotic patterns. There may be a theoretical advantage in having a cell-free medium in terms of flow resistance. Using blood cardioplegia, a limited hypertonicity may be beneficial in reducing the capillary flow resistance of RBC.

Response in shape and size of individual p31 cancer cells to cisplatin and ouabain: a computerized image analysis of cell halo characteristics during continuous perfusion.

BACKGROUND: Volume regulation is essential for cellular functions, including cell death, such as apoptosis. Flow cytometry is standard for nonadherent cells, such as blood cells. Our aim was to explore image analysis methods to study adherent cancer cells of a solid tumor. METHODS: P31 mesothelioma cells were perifused (40 min) and studied by phase-contrast microscopy. A noise reduction of the cell contour was tested to more accurately yield the cell shape factor (SF). The optical halo around the cell was analyzed for information about membrane blebbing. RESULTS: The projected cell area (PCA) slowly increased under control perfusion, the halo outside more than the halo inside. Cisplatin (apoptosis) caused an immediate increase in the PCA-halo outside (5.9 +/- 1.2 %, P < 0.01, 1-5 min) and the SF indicated decreased roundness (P < 0.05). The SF-halo inside became more irregular than the outside, which was different from the control cells. The morphology reflected instant blebbing, and the cell bodies showed fragmentation after about 20 min. Ouabain resulted in only small changes in PCA and SF, significantly different from both control and cisplatin conditions. CONCLUSIONS: Image analysis (PCA and SF) on perifused adherent cancer cells may serve as a tool to follow the sensitivity of cancer chemotherapy and to study cell death patterns.

Red blood cell trauma during cardiopulmonary bypass: narrow pore filterability versus free haemoglobin.

Ten patients admitted for coronary artery bypass grafting were investigated with respect to the influence of cardiopulmonary bypass (CPB) on red blood cell (RBC) trauma. Blood samples were collected prior to, at the start of, and at 30 and 60 min of CPB. RBC deformability was assessed by filtering re-suspended RBCs through a polycarbonate membrane using a computer-controlled filtrometer. Multiple regression analysis was employed to evaluate RBC flow-curve characteristics denoted by the initial filtration rate (IFR) and clogging slope (CS). Release of free haemoglobin was determined concomitantly. IFR was estimated at 90.39 microl/s and CS at -5.32 microl/s2 prior to CPB. During 60 min of CPB, neither IFR nor CS deviated significantly (p > 0.05) from these reference values. However, release of free haemoglobin increased significantly (p < 0.018) from the start of CPB to the 60-min determination. In conclusion, 60 min of CPB seems not to alter significantly RBC deformability in a 5 microm pore filtration model, despite a significant release of haemoglobin.

Microrheology of filtered autotransfusion drain blood with and without leukocyte reduction.

Autotransfusion of filtered knee drain blood (FKDB) is frequently practised in orthopaedic surgery, but questioned because it contains inflammatory cytokines, contaminants from lysed blood cells, debris and chemicals from the wound. We have studied the microrheology (5 microm pore filtration) of FKDB (n = 23) with versus without the addition of a leukocyte reducing filter (LRF) in line with the drain. As expected the whole blood clogging was reduced (p < 0.01) due to the lowered leukocyte number by the LRF. FKDB plasma contains clogging particles of unknown origin. With the LRF the increased plasma clogging was reduced (p approximately 0.05). With resuspended erythrocytes there was an increase in clogging rate in FKDB at 24 hours. This increase was abolished with the addition of the LRF, which may indicate that the erythrocyte trauma results from the incubation together with leukocytes in the drain container. There is a potential for further improvement of the filters in autotransfusion drains.

Serotoninergic modulation of cell volume response to estramustine: an image-analysis study on perifused individual glioma cells.

A technique of microscopy with computerised detection of early morphological changes during continuous perifusion was used to monitor the geometry changes of cultured glioma cells (MG-251) when exposed to 40 mg/L estramustine phosphate (EMP) alone or in combination with granisetron (0.1 mumol/L), ondansetron (0.1 mumol/L), or serotonin (1 mumol/L). When the cells were exposed to EMP, cell volume measured as projected cell area (PCA) rapidly increased. Serotonin and ondansetron, but not granisetron, prevented the acute EMP response (PCA). Serotonin, but none of the 5-HT3 receptor antagonists, protected against the cytotoxicity of EMP to the glioma cells as measured by a fluorometric microculture assay. Our results demonstrate hitherto unknown differences between selective 5-HT3 receptor antagonist on the cellular response to EMP and shows the necessity to study the receptor antagonists from viewpoint of interference with the antitumour drug effects on malignant cells. The perifusion technique could be used to study the effects of serotoninergic agonists and antagonists on cell volume regulation of cells exposed to anticancer drugs.

Treatment of myeloproliferative disorders with hydroxyurea: effects on red blood cell geometry and deformability.

Hydroxyurea (HU) is used in suppressing the bone marrow and producing fetal-like red blood cells (RBCs). These RBCs are large in size and may theoretically disturb the microcirculation. In five patients with myeloproliferative disorders (MPD), the RBC geometry and deformability were analyzed before and after 6 to 8 months of HU treatment. In untreated MPD, the RBC geometry and filterability was normal. After HU, the RBC membrane area increased 24% and the cell volume increased 39% (P <.005). This change resulted in a 12% increase in the minimum cylindrical diameter (MCD). From a static bending model of initial deformation, the RBC diametrical cross-section had a significantly increased section modulus. However, this increase in profile stiffness was compensated for by its larger projected cell area and, thus, pressure load on the RBC corpuscle. The resulting resistance to initial deformation therefore remained unchanged after HU. These findings were tested experimentally; with 3-microm filter membranes, HU treatment caused a significant increase in flow resistance (P <.02), in accordance with MCD. However, with 5-microm pores, no difference was seen, again in consonance with the theoretical findings of initial deformation. Because most capillaries are larger than 3 microm, we suggest that HU is acceptable from a perspective of cellular microrheology.

Filterability of autotransfusion blood cells and plasma after total knee arthroplasty.

Postoperatively filtered drain blood can be used for autotransfusion, but quality has been questioned. The blood contains activated or destroyed blood cells, debris, and chemicals from the wound. The ConstaVac autotransfusion system was used in 18 patients operated with a total knee prosthesis. Samples were taken from drain blood and compared with venous blood at different sampling times (to 24 h) and different incubation conditions. Micro-rheology was measured as pore filterability (5 microm) in whole blood, erythrocyte resuspension, and plasma. The filtration flow resistance was significantly lower in drain whole blood, but was unchanged in erythrocyte resuspension compared to venous blood. However, the rate of filter clogging was unchanged in drain whole blood, but was increased in the erythrocyte resuspension at 2 h. Drain plasma filterability indicated particle contamination. In total, the filterability of resuspended drain erythrocytes and cell-free plasma is affected, but this is not significant enough to have effects on drain whole blood filterability which is increased, possibly due to the leukocyte reduction. From these perspectives, we consider drain blood acceptable for autotransfusion.

Use of polyester leukocyte elimination filters in blood filterability research.

We tested a new routine to eliminate leukocytes for blood rheology measurements using commercial leukocyte absorbing filters (here PALL RC400). These filters were punched out and fitted in smaller chambers through which blood was filtered under controlled suction pressure (< 30 mm Hg). This technique resulted in a very effective leukocyte elimination to 0.0022% but also a platelet reduction to 0.2%. The process causes a small but significant hemolysis with free hemoglobin, of the order of 0.06% of the filtered erythrocytes. A small fraction of the erythrocytes were retained in the filter, versus plasma, to reduce the hematocrit on the order of 1.4%. The leukocyte filtration did not cause any detectable functional trauma to the erythrocytes, measured as micro-pore filterability of normal and glutaraldehyde (GA) hardened erythrocytes. However, when 10% of the erythrocytes were hardened with GA, which caused an increase in pore clogging slope (p < 0.05), the additional passage through the leukocyte elimination filter removed this measured change in clogging. This observation suggests that the leukocyte elimination filter may selectively remove, not only leukocytes and platelets, but also hardened erythrocytes. Reticulocyte counting did not reveal any selective removal of young erythrocytes. In general, we find the presented method reproducible, efficient and easy for eliminating leukocytes for blood rheology research although the risk of removing undeformable erythrocytes must be considered.

Autotransfusion after total knee arthroplasty. Effects on blood cells, plasma chemistry, and whole blood rheology.

Postoperative drain blood was collected and reinfused using the ConstaVac system (Stryker, Kalamazoo, MI) in 30 patients after total knee arthroplasty. Of the total 1.1-L volume of postoperative bleeding, 60% was reinfused. No clinical complications were observed. Differences between venous blood and drain blood and between venous blood and drain blood after separate incubation were studied with respect to acidic and inflammatory effects on blood cells, plasma chemistry, and whole blood rheology. In drain blood, leukocyte and platelet counts were reduced (P < .001), probably as a result of consumption in the wound. Acidic incubation occurs in the drain container because of production of lactate from glucose, with a minimum pH at 5 hours of 7.2. The low pH caused slight but significant erythrocyte swelling (P < .01). The complement C3d indicated leukocyte activation, although of modest magnitude. Despite incubation and complement activation, maximum erythrocyte hemolysis after 24 hours of incubation was less than 1%. Drain blood showed a lower resistance against micropore filtration than venous blood (P < .001), mainly because of the reduced number of leukocytes, and remained unchanged with incubation. Although the autotransfusion system can be improved with respect to blood quality, filtered drain blood should be considered acceptable for reinfusion.

Combined use of micropipette aspiration and perifusion for studying red blood cell volume regulation.

We combined micropipette aspiration with a new technique of perifusion to study volume behavior of single red blood cells (RBCs) during anisotonic conditions. Other techniques have suggested that RBC volume changes are complete within 1 s, but our pipette data indicate a slower volume response of 5-10 s. This difference appears to be due to the partial aspiration of the cell into the pipette, which in part prevents exposure of the membrane area to the anisotonic medium. However, we found that medium will pass the aspirated RBC (-20 mm H2O aspiration pressure) and enter the pipette at a volume rate of 0.3 micron3/s; this rate was measured by introducing a flow marker, a separate RBC or a cell fragment, into the pipette. The osmotic balance across the RBC also affected the flow: if the outside fluid was made hypertonic, the pipette flow decreased, but a hypotonic conditions, the flow increased and contributed to the exposure of the aspirated portions of the RBC. However, after the initial 10 s, the RBC geometry can be precisely monitored; at 200 mOsm, the RBC swelled by 41.0 +/- 0.9% and then demonstrated a small but significant (P < 0.001) regulatory volume decrease. We suggest that this technique is very precise in measuring the dynamics of geometric regulation, although acute changes are affected by the partial cell aspiration and pipette flow.

Computerized detection of morphological changes to glioma cells during estramustine and ion-channel blocker perifusion.

A perifusion technique for microscopy with computerized detection of early changes in cell morphology during continuous perifusion was used to show that the geometry of cultured glioma cells (MG-251) changes rapidly when they are exposed to estramustine phosphate (EMP). When the cells were exposed to 20 or 40 mg l(-1) EMP, cell volume projected cell area (PCA) rapidly increased. When the Na+,K+-ATPase blocker ouabain (100 micromol l(-1)) was added to the EMP (40 mg l(-1)) perifusion, the acute EMP response was eradicated. When the PCA curve for ouabain alone was subtracted from the curve of combined ouabain and EMP perifusion, the resulting curve showed that ouabain completely blocked the EMP-induced increase in PCA. When the Na+, K+, Cl- co-transport inhibitors bumetanide (10 micromol l(-1)), or furosemide (100 micromol l(-1)), were added to EMP (40 mg l(-1)), the acute increase in PCA seen for EMP alone was also completely blocked. This study shows that inhibitors of ion transmembrane transport can modify EMP-induced cell volume increases. This may be of particular importance since the blockers have been found to interfere also with the cytotoxic function of EMP during cell culture. Thus, it is possible that cell volume changes could serve as a rapid technique for predicting the cytotoxic activity of antineoplastic drugs.

Analysis of flow acceleration during erythrocyte filtration: dependence of hematocrit and cell rigidity.

At flow onset the blood filtration rate accelerates to a steady state, this may affect the interpretation of red blood cell (RBC) filterability. We studied the acceleration of flow while the pressure is built up across the filter to analyse effects of various hematocrits and RBC rigidity by glutaraldehyde (GA) hardening. This was analysed by a new filtration system with high time resolution and unlimited filtration volume. The system uses a digital balance that samples the accumulated weight (e.g., filtration rate through 5 microns Nucleopore membranes) with on-line computer communication. The filtration is computer controlled via a pneumatic valve. White blood cells (WBC) were removed prior to filtration by a WBC-eliminating filter to avoid clogging artifacts. When flow is initiated a steady state is reached at 0.3-0.4 s. This timing was also tested and confirmed by a video monitoring technique of filtration flow into a horizontal pipette. The digital balance has a mathematical function to reduce the effects of vibration noise; when this function was activated the apparent acceleration was retarded to 1.2 s. With any of these techniques the steady state timing did not vary with the hematocrit, however, the volume of filtered suspension during acceleration varied with both the hematocrit and the GA hardening (p < 0.001). Extrapolation to yield the initial filtration rate from the relative flow curve (RBC suspension divided by buffer flow) varied depending on if the acceleration phase was included or not. In the most unfavourable situation, with GA-hardened RBC, this difference was 340% (p < 0.01). The slope to calculate clogging rate was affected in a similar way. Moreover, with the most GA-hardened RBC a delay in flow onset was observed with this technique. The acceleration phenomenon may cause artifacts in systems employing volume-derived filtration kinetics because of fixed volumes of filtrated medium.

Thiocyanate induces vasoconstriction in rat-tail artery primed with norepinephrine.

The vasoregulating effects of thiocyanate (SCN-) were studied by means of a new rat tail artery perfusion model that uses constant driving pressure (1000 mm H2O). When a 30 mm long artery was perfused with 20 micromol/L verapamil, the flow rate increased 6%. Norepinephrine (10 to 1000 nmol/L) caused a dose-dependent flow inhibition. SCN- alone (0.05 to 5 mmol/L) had a slight, if any, effect on the arterial flow. However, when the artery was pretreated with norepinephrine (1000 nmol/L) for 10 minutes, followed by a basal-medium washout that left a 13% norepinephrine-induced flow inhibition, 0.05 to 5 mmol/L SCN- caused a marked flow rate reduction of between 20.4% +/- 9.8% and 28.0% +/- 21.8%, as calculated for the entire test perfusion. The decrease in flow rate correlated with the SCN- concentration (p < 0.05) and showed a clear reversibility. A similar SCN(-)-induced (0.05 mmol/L) vasoconstriction (p < 0.05) was seen when the artery was given a basal tone by continuous perfusion with low-dose norepinephrine, 10 nmol/L. Nifedipine (100 nmol/L) abolished the effect of 0.05 mmol/L SCN- but did not affect the norepinephrine priming. We conclude that SCN- amplifies the norepinephrine-induced vascular smooth muscle tone and that this may be caused by an altered calcium channel activity.

Effects of pressure on red blood cell geometry during micropipette aspiration.

Micropipette aspiration is a potentially useful and accurate technique to measure red blood cell (RBC) geometry. Individual RBCs are partially aspirated and from the resulting sphere diameter, total cell length, and pipette diameter, membrane area and cell volume can be calculated. In this study we have focused on possible shape artifacts associated with the aspirated portion of RBC. We observed that the apparent RBC geometry (calculated area and volume) changed markedly (P < 0.001) with the applied aspiration pressure; for normal human RBC the area increased by 5.6 +/- 0.6% and volume decreased by 4.7 +/- 0.6% when the aspiration pressure was increased from 20 to 100 mm water. The calculated membrane area dilation modulus was 7.4 dyn/ cm, which is far below the expected value, and microscopic observations revealed a membrane folding artifact as a possible artifact. These assumptions were strengthened by using a short-duration (3 s) pressure peak of 20-100-20 mm water. The folding then disappeared permanently, but a small (0.31 +/- 0.09%; P < 0.001) area decrease was detected which yields a realistic dilation modulus of 215 dyn/cm. We conclude that membrane folding can critically affect RBC micropipette measurements and that a transient pressure peak can unfold the RBC membrane, thus allowing accurate measurements of RBC geometry.