In vivo effect of haemodilution with saline on coagulation
2002; Elsevier BV; Volume: 89; Issue: 6 Linguagem: Inglês
10.1093/bja/aef574
ISSN1471-6771
AutoresPetra Innerhofer, D Friès, Antón Klingler, Werner Streif,
Tópico(s)Hemodynamic Monitoring and Therapy
ResumoEditor—As the influence of i.v. fluids on the coagulation system is of importance and also sustains the discussion on optimal fluid therapy, we would like to comment on the results reported by Ng and colleagues1Ng KFJ Lam CCK Chan LC In vivo effect of haemodilution with saline on coagulation: a randomized trial.Br J Anaesth. 2002; 88: 475-480Crossref PubMed Scopus (151) Google Scholar and the suggestions presented in the accompanying editorial.2Ruttmann TG Haemodilution enhances coagulation.Br J Anaesth. 2002; 88: 470-472Crossref PubMed Scopus (38) Google Scholar The study design chosen by Ng and colleagues1Ng KFJ Lam CCK Chan LC In vivo effect of haemodilution with saline on coagulation: a randomized trial.Br J Anaesth. 2002; 88: 475-480Crossref PubMed Scopus (151) Google Scholar to investigate coagulation and haemodilution might eliminate the influence of tissue trauma provoked by surgery, but is not ‘without the effect of any confounding variables’.2Ruttmann TG Haemodilution enhances coagulation.Br J Anaesth. 2002; 88: 470-472Crossref PubMed Scopus (38) Google Scholar First, the ratio of blood withdrawn:replacement with normal saline was 1:2, which suggests that the haemodiluted patients were hypovolaemic rather than haemodiluted. Second, all study subjects were patients with malignant disease, who are known to exhibit some state of hypercoagulability because of their underlying disease. Third, the shortening of the reaction time (r-time) by 30% and the coagulation time (k-time) by 36–45% was statistically different from the controls, but they also showed prolongation of the r- and k-times, although no intervention occurred. The question arises as to whether such changes are relevant in vivo, as they might be explained as an in vitro phenomenon attributable to sedimentation of red cells in the thrombelastograph (TEG) cup. Uraemic patients have been shown to have an increased bleeding time, but also exhibit thrombelastographic signs of hypercoagulability measured by shorter r- and k-times, increased α-angle and maximum amplitude compared to normal controls.3Holloway DS Vagher JP Caprini JA Simon NM Mockros LF Thrombelastography of blood from subjects with chronic renal failure.Thromb Res. 1987; 45: 817-825Abstract Full Text PDF PubMed Scopus (32) Google Scholar Interestingly, the latter study showed a significant negative correlation between haematocrit and signs of hypercoagulability. Moreover, isolated reduction of the haematocrit from 40 to 10% also resulted in thrombelastographic signs of hypercoagulability.4Iselin BM William PFX Seifert B et al.Isolated reduction of haematocrit does not compromise in vitro blood coagulation.Br J Anaesth. 2001; 87: 246-249Crossref PubMed Scopus (67) Google Scholar Reports of accelerated coagulation during haemodilution might therefore be influenced by changes in red cell sedimentation occurring with changing haematocrit. It is likely that this effect also depends on the physicochemical properties of the dilution fluid used and the time needed to process the coagulation measurements. Our in vitro and in vivo data on the influence of colloids and crystalloids on coagulation show an accelerated initiation of coagulation during moderate haemodilution with Ringer's lactate and gelatin for the slow-reacting intrinsic TEG measurements only.5Fries D Innerhofer P Klingler A et al.The effect of the combined administration of colloids and lactated Ringer's solution on the coagulation system: an in vitro study using Thrombelastograph® Coagulation Analysis (ROTEG®).Anesth Analg. 2002; 94: 1280-1287Crossref PubMed Scopus (127) Google Scholar 6Innerhofer P Fries D Margreiter J et al.Effects of perioperatively administered colloids and crystalloids on primary platelet-mediated hemostasis and clot formation.Anesth Analg. 2002; 95: 858-865Crossref PubMed Scopus (158) Google Scholar In contrast to the data of Ng1Ng KFJ Lam CCK Chan LC In vivo effect of haemodilution with saline on coagulation: a randomized trial.Br J Anaesth. 2002; 88: 475-480Crossref PubMed Scopus (151) Google Scholar and Ruttmann,2Ruttmann TG Haemodilution enhances coagulation.Br J Anaesth. 2002; 88: 470-472Crossref PubMed Scopus (38) Google Scholar our results for accelerated coagulation were always accompanied by a reduction in clot firmness. Furthermore, we were surprised that, in the study by Ng and colleagues, the α-angle increased by about 71% in the haemodiluted group at 30 min, although platelets and fibrinogen decreased markedly. Given a normal value of 60°, this would mean that the angle would have reached values well above 90°. Was this really observed? Analysis of our in vivo data for the α-angle showed no significant difference in the intraoperative response profile (area under the curve minus baseline, AUCA-D) of extrinsic TEG measurements between different fluid regimens (hydroxyethyl starch, modified gelatin 4% combined with a basal infusion of Ringer's lactate, or Ringer's lactate), and a significant difference in intrinsic TEG tracings between the two colloid groups only (Table 1).Table 1Extrinsically- and intrinsically-activated TEG measurements during knee replacement surgery in groups randomized to receive hydroxyethyl starch 6% (HES) 200/0.5, modified gelatin 4% with Ringer's lactate (GEL), or lactated Ringer's solution (RL). Differences between groups were analysed by comparing the perioperative response profile (area under the curve, AUCA-D). Data are median (interquartile range): A, baseline; B, before tourniquet release; C, end of surgery; D, 2 h later. Statistically significant difference (AUCA-D P<0.05): *GEL vs HES. No other statistically significant differences between HES, GEL and RL groups. Analysis of repeated measures ANOVA (all groups pooled) showed no significant time effect.α-angleRLGELHES(°)(n=20)(n=20)(n=20)Extrinsic TEGA71 (64/76)71 (66/72)67 (64/71)B72 (66/74)71 (63/72)69 (64/73)C71 (66/77)69 (62/71)65 (60/69)D72 (61/76)70 (66/73)62 (58/67)AUCA-D2.3 (–7.5/11.5)–4.5 (–10.5/–0.5)–5.8 (–17.3/2.8)Intrinsic TEGA71 (68/74)68 (64/72)70 (64/73)B69 (66/74)70 (68/74)67 (64/71)C71 (63/75)69 (66/71)67 (62/71)D67 (61/75)70 (65/73)66 (63/68)AUCA-D2.5 (–11/7.5)2.5 (–6/7.5)*–9.3 (–12.5/–0.5) Open table in a new tab Another question arises concerning measurements of concentrations of thrombin–antithrombin complexes (TAT). Why were concentrations of TAT much higher in the control group, and why did they also show a large standard deviation not observed in the haemodiluted group? Despite these inter-group discrepancies, the authors assumed that thrombin generation was increased in their haemodiluted patients as concentrations of TAT did not change during haemodilution, but showed an insignificant decrease in the controls. A clinical study investigated coagulation after haemodilution with gelatin (in addition to a basal infusion of Ringer's lactate), which has also been associated with signs of hypercoagulability.7Hobisch-Hagen P Wirleitner B Mair J et al.Consequences of acute normovolaemic haemodilution on haemostasis during major orthopaedic surgery.Br J Anaesth. 1999; 82: 503-509Crossref PubMed Scopus (29) Google Scholar In this study, concentrations of TAT did not change significantly over several hours in haemodiluted or control patients. Yet 3 h after surgery, a significant increase was observed, but only in control patients. Furthermore, all measurements of activated coagulation and fibrinolysis showed no differences between groups. In summary, there is very little evidence that administration of i.v. fluids leads to clinically detectable hypercoagulability. Data from one study conducted in 1980, investigating only 30 ‘wet’ and 30 ‘dry’ patients during abdominal surgery without perioperative thrombosis prophylaxis, are insufficient to associate i.v. fluid therapy with development of deep venous thrombosis.2Ruttmann TG Haemodilution enhances coagulation.Br J Anaesth. 2002; 88: 470-472Crossref PubMed Scopus (38) Google Scholar 8Janvrin SB Davies G Greenhalgh RM Postoperative deep vein thrombosis caused by intravenous fluids during surgery.Br J Surg. 1980; 67: 690-693Crossref PubMed Scopus (89) Google Scholar Some data indicate that the impairment of coagulation is the least when predominantly crystalloids are used.5Fries D Innerhofer P Klingler A et al.The effect of the combined administration of colloids and lactated Ringer's solution on the coagulation system: an in vitro study using Thrombelastograph® Coagulation Analysis (ROTEG®).Anesth Analg. 2002; 94: 1280-1287Crossref PubMed Scopus (127) Google Scholar 6Innerhofer P Fries D Margreiter J et al.Effects of perioperatively administered colloids and crystalloids on primary platelet-mediated hemostasis and clot formation.Anesth Analg. 2002; 95: 858-865Crossref PubMed Scopus (158) Google Scholar These effects are seen as an advantage for surgical patients with respect to preventing further blood loss and limiting the need for transfusion of blood products. Moreover, reports on disseminated intravascular coagulopathy after administration of crystalloids are lacking in the literature. In vivo, the response of the coagulation system is mainly governed by the extrinsic pathway. Therefore, the impact of TEG measurements which are obtained without extrinsic activation, thus depending on the reactions provoked by contact to foreign surfaces, remains to be determined. Until it is proven clearly that hypercoagulability measured by TEG is not only an in vitro phenomenon but also relevant in vivo, it seems dangerous to recommend withholding i.v. fluids. Normovolaemia is essential for maintaining sufficient oxygen delivery to the tissues, and thus fluid therapy, not only influences coagulation, but also preserves organ function. Editor—Thank you for the opportunity to reply to the letter by Dr Innerhofer and colleagues commenting on our paper.1Ng KFJ Lam CCK Chan LC In vivo effect of haemodilution with saline on coagulation: a randomized trial.Br J Anaesth. 2002; 88: 475-480Crossref PubMed Scopus (151) Google Scholar We acknowledge Dr Innerhofer's point that the clinical significance of the hypercoagulability caused by crystalloid infusion remains to be evaluated. However, we believe that, given the prospective, randomized and controlled design, it is safe to state that our study is without the effect of any confounding variables. We cannot agree with Dr Innerhofer that, because replacement with saline was 1:2, patients in the haemodiluted group in our study were hypovolaemic. The usual 25–30% retention rule is oversimplified and refers to an equilibrium state only. The percentage retention of infused crystalloid in the intravascular compartment is dynamic. It depends upon the volume infused, the rate of infusion, and on when one is making the measurement. This has been discussed elegantly by Hahn and colleagues.9Hahn RG Drobin D Stahle L Volume kinetics of Ringer's solution in female volunteers.Br J Anaesth. 1997; 78: 144-148Crossref PubMed Scopus (79) Google Scholar As we were allowing only 10 min for the infused fluid to distribute in our study procedure, 1:2 is the ratio administered to achieve normovolaemia. In fact, if we assume normovolaemia then the drop in haemoglobin (Hb) concentration after removal of 10, 20 and 30% of blood volume can be predicted by the exponential equations:10Brecher ME Rosenfeld M Mathematical and computer modeling of acute normovolemic hemodilution.Transfusion. 1994; 34: 176-179Crossref PubMed Scopus (125) Google Scholar Hb10%=Hb0e−0.1 Hb20%=Hb0e−0.2 Hb30%=Hb0e−0.3 where Hb10%, 20%, 30% refers to the Hb concentration after removal of 10, 20 and 30% of blood volume and concomitant isovolaemic haemodilution. Hb0 is the initial Hb concentration. In other words, the Hb concentration after removal of 10, 20 and 30% of blood volume should be 0.90 Hb0, 0.82 Hb0 and 0.74 Hb0 respectively. In our study, the values were 0.90, 0.80 and 0.74 respectively, which confirmed almost exact isovolaemia. Dr Innerhofer's suggestion that the shortening of r- and k-times might be an artefact related to changes in red cell sedimentation is interesting, as they reported the effect may disappear when the TEG measurement was made more quickly. However, to accelerate measurement, the differences one would have to introduce are more than just the cell sedimentation rate. Certain components of the coagulation process would invariably be suppressed or enhanced. The differences in TEG variables one observes may not be attributable to differences in measurement time alone. The fact that an isolated reduction of haematocrit resulted in hypercoagulability4Iselin BM William PFX Seifert B et al.Isolated reduction of haematocrit does not compromise in vitro blood coagulation.Br J Anaesth. 2001; 87: 246-249Crossref PubMed Scopus (67) Google Scholar only proved that haemodilution-related hypercoagulability could develop in the absence of changes in concentrations of natural procoagulants or anti coagulants. It cannot be interpreted that the hypercoagulability is because of changes in red cell sedimentation during measurement, nor can the results exclude the possible contribution of changes in concentrations of natural procoagulants or anticoagulants. We are surprised that Dr Innerhofer quoted the normal value of α as 60°, which, as far as we know, is usually quoted between 29 and 43° for native TEG.11Haemoscope Corporation Thrombelastograph coagulation analyzer: Operational Manual 43. Skokie, IL, USA1990Google Scholar, 12Miller BE Bailey JM Mancuso TJ et al.Functional maturity of the coagulation system in children: an evaluation using thrombelastography.Anesth Analg. 1997; 84: 745-748Crossref PubMed Google Scholar, 13Caprini JA Zuckerman L Cohen E et al.Comparison of thrombo-elastography with common coagulation tests.Thromb Haemost. 1981; 46: 362Google Scholar In our study, the pre-dilution value of α was 26° in the haemodiluted group, increasing to 43° at 15 min, a 70% increase. The higher mean and variance in the TAT concentrations of the control group were because of two patients with high TAT concentrations in this group. Thus, the two groups may not be comparable directly with respect to TAT concentrations and, therefore, this was not attempted. What we have done is to follow the trend of TAT concentration in each group over time. Our suggestion that thrombin generation may be increased in the haemodiluted group was based on the lack of decline in TAT concentration despite haemodilution and a proportional decline in other haemostatic and cellular markers within this group. We fully agree with Dr Innerhofer that the native TEG is mainly measuring the contact activation (intrinsic) pathway of coagulation. Therefore the study, which demonstrated TEG features of hypercoagulability and yet prolonged bleeding time,3Holloway DS Vagher JP Caprini JA Simon NM Mockros LF Thrombelastography of blood from subjects with chronic renal failure.Thromb Res. 1987; 45: 817-825Abstract Full Text PDF PubMed Scopus (32) Google Scholar could not be simply discarded as reporting an artefact in the TEG measurement. The TEG is not a monitor of platelet function. Bleeding time measures adhesive properties of platelets (platelet–vWF–endothelium interaction) which is the major haemostatic defect observed in uraemic patients.14Weigert AL Schafer AI Uremic bleeding: pathogenesis and therapy.Am J Med Sci. 1998; 316: 94-104Crossref PubMed Scopus (145) Google Scholar Although TEG variables are affected by thrombocytopenia, interestingly, the addition of platelets that are totally incapable of adhesion or aggregation improves the TEG, and the possible mechanisms for this have been reported.15McNulty SE Sasso P Vesci J Schieren H Platelet concentrate effects on thromboelastography.J Cardiothorac Vasc Anesth. 1997; 11: 828-830Abstract Full Text PDF PubMed Scopus (11) Google Scholar Similarly, because TEG measures mainly the contact activation pathway, it is true that the clinical significance of our observation remains to be evaluated. We also agree that the patients we have studied probably had some pre-existing hypercoagulability. However, because of the parallel design of the study, this effect should be roughly similar in the haemodiluted and control groups. Being aware of these limitations, and with our total agreement with Dr Innerhofer on the importance of adequate circulating volume, it has never been our intention to advocate withholding i.v. crystalloids from surgical patients. We shall need properly designed outcome studies to address this issue. K. F. J. Ng C. C. K. Lam L. C. Chan Hong Kong, China
Referência(s)