© 1991 S. Karger AG. Basel 0301-0147/91 /0211-0030S2.75/0
Haemostasis 1991;21:30-36
Haemorrhagic Effects of Unfractionated and Two Low Molecular Weight Heparins, Enoxaparin and Fragmin, in Rats Christen J. Bang, Arnold Berslad, Ingebrigt Talstad Sections of Gastroenterology and Haematology, Medical Department, Haukeland University Hospital, Bergen, Norway
Key Words. Primary haemostasis • Low molecular weight heparin • Unfractionated heparin
Introduction The haemorrhagic and antithrombotic ef fects of heparin are both clearly dose related. Doses of heparin, high enough to virtually abolish thrombosis, carry unacceptable risks of bleeding [1-3]. This general rule seems to be valid for the new low molecular weight heparins as well, although experimental evi dence suggests a reduced bleeding tendency [4, 5], The potential for inducing haemor rhage is, however, not easily assessed. All experimental bleeding models are relatively inaccurate, and the primary haemostasis is usually more dependent on platelet function
than on fibrin formation and polymerization [6, 7], In the rat and several other animal spe cies, however, haemostasis in the gastric mu cosa seems largely independent of platelets [8], The arrest of bleeding, then, is mainly dependent on coagulation. Logically, under such conditions, haemostasis is easily ham pered by the administration of an anticoagu lant, like heparin, and the gastric mucosa may constitute a unique model to test the haemorrhagic effects of heparins. Using this sensitive model, we found that the low mo lecular weight heparin, enoxaparin, had to be administered at doses which were five to
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Abstract. The effects on primary haemostasis of unfractionated heparin and of the two low molecular weight heparins, enoxaparin and fragmin, were compared in two rat models, one employing the gastric mucosa and the other the tail skin. All three heparin preparations prolonged the bleeding time and increased the blood loss dose dependently. The prolonga tion of the bleeding time per unit dose caused by unfractionated heparin was significantly greater than the prolongation caused by either one of the two low molecular weight heparins. In the gastric mucosa, but not in the tail skin, enoxaparin prolonged the bleeding time significantly less than fragmin (p < 0.05).
31
Heparin and Bleeding
Materials and Methods Male Wistar rats weighing 200-300 g were anaes thetized with pentobarbital sodium (60 mg/kg). Tra cheostomy was performed and a PE 240 catheter in serted into the trachea to maintain an open airway. The vena jugularis was cannulated with a PE 50 cath eter which was kept open by continuous infusion of saline at a rate of 0.6 ml/h (Harvard infusion pump). Laparotomy was performed via a midline incision. The stomach was lifted out and opened along the greater curvature, ensuring that the vascular connec tions were intact. After removal of the gastric content, the stomach was clamped in the bottom of a chamber, as described by Mersereau and Hinchey [10]. The body temperature was kept constant at 37 °C by an animal blanket control unit. Unfractionated heparin (Heparin; Nycomed, Nor way) was administered intravenously at doses of 10, 50," 150, 200. 500. and 1,000 U/kg (U = anti-Xa units). Fragmin (Fragmin; Kabi, Sweden) and enoxa parin (Klexane; Rhone-Poulenc, France) were admin istered similarily at doses of 50, 200, 500, and 1,000 U/kg. Batches of different concentrations obtained by dilution with saline, which were coded and stored in the refrigerator for further use. were then adminstered blindly and in random order, both with respect to drug and dose. Each dose level of each drug was investigated in 6 rats. Ten minutes after administra
tion of the drug, a 2-mm cut was made with a pair of iris scissors in the glandular part of the gastric muco sa. The bleeding was monitored as described by Whit tle et al. [8] by perfusing the chamber with isotonic saline (37°C) at a rate of 3 ml/min (Ismatec Reglo 100), taking care that the perfusate was directed over the mucosal surface close to, but not onto, the lesion. The chamber was emptied at 1-min intervals during a 20-min period. The amount of blood in the perfusate was quantified by reading the optical density of each 1-min sample at 540 nm after having haemolyzed the erythrocytes by freezing and thawing [11]. One refer ence standard for the haemoglobin concentration was prepared for each rat by adding 10 pi of the rat’s arterial blood to 3 ml saline. Simultaneously, a 5 mm long and 1 mm deep free hand cut was made with a Simplate device in the mid dle of the rat tail at its dorsal side, taking care to avoid the large veins. The tail was then submersed vertically into a beaker containing isotonic saline at a tempera ture of 37 °C [12], The bleeding was monitored for a maximum of 20 min, and the cessation of bleeding determined visually. The following variables were determined: (1) the gastric mucosal bleeding time, i.e., the time period during which the bleeding rate was > 1 pl/min - this limit was chosen to reduce disturbing factors like leakage of blood from the chamber ring or refluxed material from the duodenum; (2) the total mucosal blood loss during the 20-min period, and (3) the rat tail bleeding time. The statistical analyses are based on regression analyses of (1) log-transformed bleeding time over dose of the drugs given, excluding dose levels which gave maximal median bleeding time, and (2) logtransformed blood loss over dose of the drugs given. The slopes of the regression lines were compared by t tests. P values < 0 .0 5 (two-sided) were considered statistically significant. Wilcoxons’ rank-sum test was used to compare bleeding times at different dose levels.
Results There were considerable variations in the bleeding response to a given dose of heparin (table 1; fig. 1,2), both in the gastric mucosa and in the tail skin. The median bleeding
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
ten times higher than the doses of unfrac tionated heparin in order to achieve a simi lar prolongation of the bleeding time [9]. The low molecular weight heparins repre sent different mixtures of small heparin mol ecules and may thus influence haemostasis differently. In the present study, employing two bleeding models, the gastric mucosa and the tail skin in rats, we have studied three heparin preparations, the two low molecular weight heparins fragmin and enoxaparin and conventional unfractionated heparin, com paring their potentials for inducing bleeding. The design of the study was blind as to the administered preparation and dose.
32
Bang/Berstad/T alstad
times were, however, dose dependently af fected by all three heparin preparations, and the maximal prolongation of the median bleeding time, 20 min, was achieved by high doses of fragmin and unfractionated hepa rin, but not by enoxaparin. This was the case both in the gastric mucosa and in the tail skin. Gastric Mucosal Bleeding A significant prolongation of the median bleeding time, as compared with controls, was seen at 200 anti-Xa U/kg of unfraction ated heparin (p = 0.001) and fragmin (p < 0.05) and at 500 anti-Xa U/kg (p = 0.001) of enoxaparin. The dose-response curves for the two low molecular weight heparins were
displaced to the right, i.e., towards higher doses, of that for unfractionated heparin (fig. 1). The bleeding time per unit dose, as expressed by the slope of the regression line for log response over dose (table 2), was sig nificantly shorter for both fragmin (p < 0.05) and enoxaparin (p < 0.001) than for unfractionated heparin. Enoxaparin pro longed the bleeding times significantly less than fragmin (p < = 0.05). A significant increase in blood loss was seen at 150 anti-Xa U/kg of unfractionated heparin (p = 0.001), at 200 anti-Xa U/kg of fragmin (p < 0.05), and at 500 anti-Xa U/kg of enoxaparin (p < 0.01) when compared with controls. The blood loss per unit dose, as expressed by the slope of the regression
Table 1. Gastric mucosal and tail skin bleeding in response to graded doses of unfractionated heparin and two low molecular weight heparins (fragmin and enoxaparin)
Saline Heparin
Gastric bleeding time, min
Gastric blood loss pi
Tail bleeding time s
median
median
median
3 10 50 150 200 500 1,000
4 6 15 19 20 13
Fragmin
50 200 500 1,000
Enoxaparin
50 200 500 1,000
range 1-5
17
range 5-22
150
range 135-160
2-10 1-16 4-20 6-20 17-20 13-20
17 23 63 131 125 314
12-118 6-200 32-193 31-1,381 47-931 89-590
180 225 415 1,200 1,200 1,200
105-340 120-350 260-795 540-1,200 1,200-1,200 1,200-1,200
4 13 20 19
1-16 1-20 7-20 7-20
15 50 99 99
5-374 10-986 34-365 18-1,815
175 245 585 1,200
95-300 120-840 255-1,200 580-1,200
4 3 11 18
1-8 2-20 6-14 7-20
19 19 36 92
5-138 8-273 19-312 44-204
115 175 400 495
105-170 120-400 195-570 225-1,200
Each dose level tested in 6 rats.
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Drugs anti-Xa U/kg
Heparin and Bleeding
33
GASTRIC MUCOSAL BLEEDING TIME (MIN)
Fig. 1. Dose-response curves for the bleeding time from induced gastric mu cosal lesions after intravenous adminis tration of unfractionated heparin, fragmin. and enoxaparin. Median values of 6 rats at each dose level. TAIL BLEEDING TIME (SECONDS)
Fig. 2. Dose-response curves for the rat tail bleeding time after intravenous administration of unfractionated hepa rin, fragmin, and enoxaparin. Median values of 6 rats at each dose level.
Gastric bleeding time, min Gastric blood loss, pi Tail bleeding time, s
a b a b a b
Heparin
Fragmin
Enoxaparin
1.1165 0.00831 2.8528 0.01018 4.9733 0.00892
1.0996 0.003470 2.9426 0.00381 5.0859 0.00198
1.0411 0.00172 2.8489 0.00175 4.9574 0.00133
Significance of the differences see Results.
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Table 2. Linear transformation of the dose-response curves according to the formula log y = bx + a, where x = dose in anti-Xa units at submaximal dose levels and y = bleeding time or blood loss (pi)
34
line for log response over dose (table 2), was significantly smaller for enoxaparin (p < 0.001) than for unfractionated heparin. The difference in blood loss between unfraction ated heparin and fragmin was not signifi cant. Rat Tail Bleeding The median bleeding times in response to the low molecular weight heparins were dis placed to the right of that for unfractionated heparin (fig. 2). The maximal median bleed ing time was reached for unfractionated hep arin at > 200 anti-Xa U/kg, for fragmin at 1,000, while enoxaparin never reached max imal bleeding time, not even at 1,000 antiXa U/kg. As compared with controls, in creased bleeding occurred at 50 anti-Xa U/kg of unfractionated heparin (p < 0.05), at 200 anti-Xa U/kg of fragmin (p < 0.05), and at 500 anti-Xa U/kg of enoxaparin. The bleeding time per unit dose, when applying the low molecular weight heparins, was sig nificantly shorter than when applying un fractionated heparin (p < 0.001). There was no significant difference between the two low molecular weight heparins.
Discussion In the present study, the haemorrhagic effects of three heparin preparations were compared in two different bleeding models: the platelet-dependent primary bleeding time in the rat tail and the platelet-indepen dent bleeding time in the gastric mucosa. Both models demonstrate a weaker haemor rhagic effect of the low molecular weight heparins, fragmin and enoxaparin, than of unfractionated heparin, at equal anti-Xa ac tivity.
Whereas the primary bleeding time of the tail skin was visually determined, estimation of the gastric bleeding time was based on the time with a blood loss of more than 1 pl/min. Contrary to the primary haemostasis of the skin, the gastric bleeding is characterized byepisodes of rebleeding [11], and the gastric bleeding time may, therefore, comprise epi sodes of rebleeding during the 20-min obser vation period. The dependency of the gastric haemostasis on coagulation of fibrin and the episodes of rebleeding may explain the dif ferences in the results achieved using the two bleeding models. In the gastric mucosa, but not in the tail skin, enoxaparin prolonged the bleeding time significantly less than did frag min. The difference between the two low molecular weight heparins was not signifi cant when comparing gastric blood loss. The initial blood loss from an induced lesion is critically dependent on number and size of the blood vessels injured. Due to anatomical variations, the vessel injury, and hence the blood loss, is difficult to standardize, even in uniform lesions. The greater variation in blood loss than in bleeding time in repeated experiments may explain that the effects of the two low molecular weight heparins on bleeding time differed significantly, whereas their effects on blood loss did not. In the present models, therefore, the bleeding time seems to be the most sensitive response vari able for evaluation of the haemorrhagic ef fect of the heparins. Low molecular weight heparins are more readily absorbed after subcutaneous injec tion and have a longer half-life than unfrac tionated heparin [13]; the effect of the differ ent heparins on bleeding, therefore, is time dependent [14]. We administered the drugs intravenously and induced the bleeding shortly (10 min) thereafter. Differences in
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Bang/Berstad/Talstad
Heparin and Bleeding
side effect, not only of the traditional unfrac tionated heparin, but also of the new, low molecular weight heparins [ 19]. Whether the antithrombotic/haemorrhagic effect ratio in man is improved by the low molecular weight heparins is so far not clearly estab lished.
Acknowledgements The authors are grateful to Prof. Jon Lekven, Head of the Surgical Research Institute, Haukeland University Hospital, and his assistants for providing excellent laboratory facilities.
References 1 Salzman EW: Low-molecular-weight heparin. Is small beautiful? N Engl J Med 1986;315:957-959. 2 Levine MN, Planes A, Hirsh J, et al: The relation ship between anti-factor Xa level and clinical out come in patients receiving enoxaparine low mo lecular weight heparin to prevent deep vein thrombosis after hip replacement. Thromb Haemost 1989;62:940-944. 3 Salzmann EW, Dcykin D, Mayer Shapiro R, et al: Managment of heparin therapy. Controlled pro spective trial. N Engl J Med 1975:292:1046— 1050. 4 Cade JF, Buchanan MR, Boneu B, et al: A com parison of the antithrombotic and haemorrhagic effect of low molecular weight heparin fractions: The influence of the method of preparation. Thromb Res 1984;35:627-636. 5 Holmer E, Mattson C, Nilsson S: Anticoagulant and antithrombotic effects of heparin and low molecular weight heparin fragments in rabbits. Thromb Res 1982;25:475-485. 6 Borchgrewink CF: The mechanism of the pro longed bleeding time provoked by dextran sul phate or by excessive doses of heparin. Acta Med Scand 1961;170:365-373. 7 Mielke CH, Kaneshiro MM, Maher IA, et al: The standardized normal Ivy bleeding time and its prolongation by aspirin. Blood 1969;34:204215.
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
drug absorption and metabolism were thus avoided. The reduced bleeding previously demon strated in response to low molecular weight heparins as compared with unfractionated heparin in platelet-dependent bleeding mod els has been ascribed to the different effects of these heparins on platelets [4, 15], Platelet interference could possibly explain the present results in the tail skin, but not in the gastric mucosa. Here the haemostasis seems largely independent of the platelets, because potent inhibitors of platelets, such as prosta cyclin or thromboxane receptor antagonists, were found not to prolong the bleeding time in this bleeding model [8], It is, therefore, unlikely that the different haemorrhagic ef fects of the heparins are due to their different effects on the platelets. Our results confirm that the antihaemo static effect of a particular heparin prepara tion is not fully expressed by its anti-Xa activity. It is also recognized that the haem orrhagic effects of various heparins are not strongly connected with their different ef fects on global assays of anticoagulation such as the activated partial prothrombin time [16], However, a significant correlation has been found between lipase release and bleed ing time in response to heparin and other sulphated polysaccharides [17], and recently it has been demonstrated that some basic disaccharide units of heparin are ten-fold more potent inhibitors of haemostasis than heparin itself [ 18], These various results sup port the notion that the precise mechanism by which heparin induces bleeding is not known. The clinical relevance of our findings re mains to be shown. No appropriate experi mental human model seems to exist. Clinical experience indicates that bleeding is a major
35
8 Whittle BJR, Kauffman GL, Moncada S: Hemo static mechanisms, independent of platelet aggre gation, arrest gastric mucosal bleeding. Proc Natl Acad Sci USA 1986;83:5683-5687. 9 Bang C, Berstad A, Talstad I: Gastric mucosal bleeding after unfractionated and low molecular weight heparin in rats. Scand J Gastroenterol 1990;25:379-382. 10 Mersereau WA, Hinchey EJ; Effect of gastric acid ity on gastric ulceration induced by hemorrhage in the rat, utilizing a gastric chamber technique. Gas troenterology 1973:64:1130-1135. 11 Berstad A, Almodovar K, Berstad K, et al: Effect of tranexamic acid on gastric bleeding in rats. Scand J Gastroenterol 1988;23:402-406. 12 Dejana E, Villa S, de Gaetano G: Bleeding time in rats: A comparison of different experimental con ditions. Thromb Haemost 1982;48:108-111. 13 Dawes J, Bara L, Billaud E, et al: Relationship between biological activity and concentration of a low molecular weight heparin (PK10196) and un fractionated heparin after intravenous and subcu taneous administration. Haemostasis 1986; 16: 116-122. 14 Pangrazzi J, Abbadini M, Zameta M, et al: Anti thrombotic and bleeding effects of a low molecu lar weight heparin fraction. Biochem Pharmacol 1985;34:3305-3308. 15 Fernandez F, N’Guyen P, Van Ryn J, et al: Hem orrhagic doses of heparin and other glycosaminoglycans induce a platelet defect. Thromb Res 1986;43:491-495.
Bang/Berstad/T alstad
16 Buchanan MR, Ofosu FA, Fernandez F, et al: Lack of relationship between enhanced bleeding induced by heparin and other sulfated polysaccha rides and enhanced catalysis of thrombin inhibi tion. Semin Thromb Hemost 1986;12:324-327. 17 Barrcwcliffe TW, Merton RE, Gray E, et al: Hep arin and bleeding: An association with lipase re lease. Thromb Haemost 1988;60:434-436. 18 Nader HB, Tersariol ILS, Dietrich CP: Structural requirements of heparin disaccharides responsible for hemorrhage: Reversion of the antihemostatic effect of ATP. Faseb J 1989;3:2420-2424. 19 Bergqvist D, Burmark US, Frisell J, et al: Prospec tive double-blind comparison between fragmin and conventional low-dose heparin: Thromboprophylactic effect and bleeding complications. Hae mostasis 1986; 16(suppl 2): 11 —18.
Received: August 17, 1990 Accepted in revised form: October 4, 1990 Christen J. Bang, MD Section of Gastroenterology Medical Department Haukeland University Hospital N-5021 Bergen (Norway)
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
36
Haemostasis 1991;21:30-36
Haemorrhagic Effects of Unfractionated and Two Low Molecular Weight Heparins, Enoxaparin and Fragmin, in Rats Christen J. Bang, Arnold Berslad, Ingebrigt Talstad Sections of Gastroenterology and Haematology, Medical Department, Haukeland University Hospital, Bergen, Norway
Key Words. Primary haemostasis • Low molecular weight heparin • Unfractionated heparin
Introduction The haemorrhagic and antithrombotic ef fects of heparin are both clearly dose related. Doses of heparin, high enough to virtually abolish thrombosis, carry unacceptable risks of bleeding [1-3]. This general rule seems to be valid for the new low molecular weight heparins as well, although experimental evi dence suggests a reduced bleeding tendency [4, 5], The potential for inducing haemor rhage is, however, not easily assessed. All experimental bleeding models are relatively inaccurate, and the primary haemostasis is usually more dependent on platelet function
than on fibrin formation and polymerization [6, 7], In the rat and several other animal spe cies, however, haemostasis in the gastric mu cosa seems largely independent of platelets [8], The arrest of bleeding, then, is mainly dependent on coagulation. Logically, under such conditions, haemostasis is easily ham pered by the administration of an anticoagu lant, like heparin, and the gastric mucosa may constitute a unique model to test the haemorrhagic effects of heparins. Using this sensitive model, we found that the low mo lecular weight heparin, enoxaparin, had to be administered at doses which were five to
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Abstract. The effects on primary haemostasis of unfractionated heparin and of the two low molecular weight heparins, enoxaparin and fragmin, were compared in two rat models, one employing the gastric mucosa and the other the tail skin. All three heparin preparations prolonged the bleeding time and increased the blood loss dose dependently. The prolonga tion of the bleeding time per unit dose caused by unfractionated heparin was significantly greater than the prolongation caused by either one of the two low molecular weight heparins. In the gastric mucosa, but not in the tail skin, enoxaparin prolonged the bleeding time significantly less than fragmin (p < 0.05).
31
Heparin and Bleeding
Materials and Methods Male Wistar rats weighing 200-300 g were anaes thetized with pentobarbital sodium (60 mg/kg). Tra cheostomy was performed and a PE 240 catheter in serted into the trachea to maintain an open airway. The vena jugularis was cannulated with a PE 50 cath eter which was kept open by continuous infusion of saline at a rate of 0.6 ml/h (Harvard infusion pump). Laparotomy was performed via a midline incision. The stomach was lifted out and opened along the greater curvature, ensuring that the vascular connec tions were intact. After removal of the gastric content, the stomach was clamped in the bottom of a chamber, as described by Mersereau and Hinchey [10]. The body temperature was kept constant at 37 °C by an animal blanket control unit. Unfractionated heparin (Heparin; Nycomed, Nor way) was administered intravenously at doses of 10, 50," 150, 200. 500. and 1,000 U/kg (U = anti-Xa units). Fragmin (Fragmin; Kabi, Sweden) and enoxa parin (Klexane; Rhone-Poulenc, France) were admin istered similarily at doses of 50, 200, 500, and 1,000 U/kg. Batches of different concentrations obtained by dilution with saline, which were coded and stored in the refrigerator for further use. were then adminstered blindly and in random order, both with respect to drug and dose. Each dose level of each drug was investigated in 6 rats. Ten minutes after administra
tion of the drug, a 2-mm cut was made with a pair of iris scissors in the glandular part of the gastric muco sa. The bleeding was monitored as described by Whit tle et al. [8] by perfusing the chamber with isotonic saline (37°C) at a rate of 3 ml/min (Ismatec Reglo 100), taking care that the perfusate was directed over the mucosal surface close to, but not onto, the lesion. The chamber was emptied at 1-min intervals during a 20-min period. The amount of blood in the perfusate was quantified by reading the optical density of each 1-min sample at 540 nm after having haemolyzed the erythrocytes by freezing and thawing [11]. One refer ence standard for the haemoglobin concentration was prepared for each rat by adding 10 pi of the rat’s arterial blood to 3 ml saline. Simultaneously, a 5 mm long and 1 mm deep free hand cut was made with a Simplate device in the mid dle of the rat tail at its dorsal side, taking care to avoid the large veins. The tail was then submersed vertically into a beaker containing isotonic saline at a tempera ture of 37 °C [12], The bleeding was monitored for a maximum of 20 min, and the cessation of bleeding determined visually. The following variables were determined: (1) the gastric mucosal bleeding time, i.e., the time period during which the bleeding rate was > 1 pl/min - this limit was chosen to reduce disturbing factors like leakage of blood from the chamber ring or refluxed material from the duodenum; (2) the total mucosal blood loss during the 20-min period, and (3) the rat tail bleeding time. The statistical analyses are based on regression analyses of (1) log-transformed bleeding time over dose of the drugs given, excluding dose levels which gave maximal median bleeding time, and (2) logtransformed blood loss over dose of the drugs given. The slopes of the regression lines were compared by t tests. P values < 0 .0 5 (two-sided) were considered statistically significant. Wilcoxons’ rank-sum test was used to compare bleeding times at different dose levels.
Results There were considerable variations in the bleeding response to a given dose of heparin (table 1; fig. 1,2), both in the gastric mucosa and in the tail skin. The median bleeding
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
ten times higher than the doses of unfrac tionated heparin in order to achieve a simi lar prolongation of the bleeding time [9]. The low molecular weight heparins repre sent different mixtures of small heparin mol ecules and may thus influence haemostasis differently. In the present study, employing two bleeding models, the gastric mucosa and the tail skin in rats, we have studied three heparin preparations, the two low molecular weight heparins fragmin and enoxaparin and conventional unfractionated heparin, com paring their potentials for inducing bleeding. The design of the study was blind as to the administered preparation and dose.
32
Bang/Berstad/T alstad
times were, however, dose dependently af fected by all three heparin preparations, and the maximal prolongation of the median bleeding time, 20 min, was achieved by high doses of fragmin and unfractionated hepa rin, but not by enoxaparin. This was the case both in the gastric mucosa and in the tail skin. Gastric Mucosal Bleeding A significant prolongation of the median bleeding time, as compared with controls, was seen at 200 anti-Xa U/kg of unfraction ated heparin (p = 0.001) and fragmin (p < 0.05) and at 500 anti-Xa U/kg (p = 0.001) of enoxaparin. The dose-response curves for the two low molecular weight heparins were
displaced to the right, i.e., towards higher doses, of that for unfractionated heparin (fig. 1). The bleeding time per unit dose, as expressed by the slope of the regression line for log response over dose (table 2), was sig nificantly shorter for both fragmin (p < 0.05) and enoxaparin (p < 0.001) than for unfractionated heparin. Enoxaparin pro longed the bleeding times significantly less than fragmin (p < = 0.05). A significant increase in blood loss was seen at 150 anti-Xa U/kg of unfractionated heparin (p = 0.001), at 200 anti-Xa U/kg of fragmin (p < 0.05), and at 500 anti-Xa U/kg of enoxaparin (p < 0.01) when compared with controls. The blood loss per unit dose, as expressed by the slope of the regression
Table 1. Gastric mucosal and tail skin bleeding in response to graded doses of unfractionated heparin and two low molecular weight heparins (fragmin and enoxaparin)
Saline Heparin
Gastric bleeding time, min
Gastric blood loss pi
Tail bleeding time s
median
median
median
3 10 50 150 200 500 1,000
4 6 15 19 20 13
Fragmin
50 200 500 1,000
Enoxaparin
50 200 500 1,000
range 1-5
17
range 5-22
150
range 135-160
2-10 1-16 4-20 6-20 17-20 13-20
17 23 63 131 125 314
12-118 6-200 32-193 31-1,381 47-931 89-590
180 225 415 1,200 1,200 1,200
105-340 120-350 260-795 540-1,200 1,200-1,200 1,200-1,200
4 13 20 19
1-16 1-20 7-20 7-20
15 50 99 99
5-374 10-986 34-365 18-1,815
175 245 585 1,200
95-300 120-840 255-1,200 580-1,200
4 3 11 18
1-8 2-20 6-14 7-20
19 19 36 92
5-138 8-273 19-312 44-204
115 175 400 495
105-170 120-400 195-570 225-1,200
Each dose level tested in 6 rats.
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Drugs anti-Xa U/kg
Heparin and Bleeding
33
GASTRIC MUCOSAL BLEEDING TIME (MIN)
Fig. 1. Dose-response curves for the bleeding time from induced gastric mu cosal lesions after intravenous adminis tration of unfractionated heparin, fragmin. and enoxaparin. Median values of 6 rats at each dose level. TAIL BLEEDING TIME (SECONDS)
Fig. 2. Dose-response curves for the rat tail bleeding time after intravenous administration of unfractionated hepa rin, fragmin, and enoxaparin. Median values of 6 rats at each dose level.
Gastric bleeding time, min Gastric blood loss, pi Tail bleeding time, s
a b a b a b
Heparin
Fragmin
Enoxaparin
1.1165 0.00831 2.8528 0.01018 4.9733 0.00892
1.0996 0.003470 2.9426 0.00381 5.0859 0.00198
1.0411 0.00172 2.8489 0.00175 4.9574 0.00133
Significance of the differences see Results.
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Table 2. Linear transformation of the dose-response curves according to the formula log y = bx + a, where x = dose in anti-Xa units at submaximal dose levels and y = bleeding time or blood loss (pi)
34
line for log response over dose (table 2), was significantly smaller for enoxaparin (p < 0.001) than for unfractionated heparin. The difference in blood loss between unfraction ated heparin and fragmin was not signifi cant. Rat Tail Bleeding The median bleeding times in response to the low molecular weight heparins were dis placed to the right of that for unfractionated heparin (fig. 2). The maximal median bleed ing time was reached for unfractionated hep arin at > 200 anti-Xa U/kg, for fragmin at 1,000, while enoxaparin never reached max imal bleeding time, not even at 1,000 antiXa U/kg. As compared with controls, in creased bleeding occurred at 50 anti-Xa U/kg of unfractionated heparin (p < 0.05), at 200 anti-Xa U/kg of fragmin (p < 0.05), and at 500 anti-Xa U/kg of enoxaparin. The bleeding time per unit dose, when applying the low molecular weight heparins, was sig nificantly shorter than when applying un fractionated heparin (p < 0.001). There was no significant difference between the two low molecular weight heparins.
Discussion In the present study, the haemorrhagic effects of three heparin preparations were compared in two different bleeding models: the platelet-dependent primary bleeding time in the rat tail and the platelet-indepen dent bleeding time in the gastric mucosa. Both models demonstrate a weaker haemor rhagic effect of the low molecular weight heparins, fragmin and enoxaparin, than of unfractionated heparin, at equal anti-Xa ac tivity.
Whereas the primary bleeding time of the tail skin was visually determined, estimation of the gastric bleeding time was based on the time with a blood loss of more than 1 pl/min. Contrary to the primary haemostasis of the skin, the gastric bleeding is characterized byepisodes of rebleeding [11], and the gastric bleeding time may, therefore, comprise epi sodes of rebleeding during the 20-min obser vation period. The dependency of the gastric haemostasis on coagulation of fibrin and the episodes of rebleeding may explain the dif ferences in the results achieved using the two bleeding models. In the gastric mucosa, but not in the tail skin, enoxaparin prolonged the bleeding time significantly less than did frag min. The difference between the two low molecular weight heparins was not signifi cant when comparing gastric blood loss. The initial blood loss from an induced lesion is critically dependent on number and size of the blood vessels injured. Due to anatomical variations, the vessel injury, and hence the blood loss, is difficult to standardize, even in uniform lesions. The greater variation in blood loss than in bleeding time in repeated experiments may explain that the effects of the two low molecular weight heparins on bleeding time differed significantly, whereas their effects on blood loss did not. In the present models, therefore, the bleeding time seems to be the most sensitive response vari able for evaluation of the haemorrhagic ef fect of the heparins. Low molecular weight heparins are more readily absorbed after subcutaneous injec tion and have a longer half-life than unfrac tionated heparin [13]; the effect of the differ ent heparins on bleeding, therefore, is time dependent [14]. We administered the drugs intravenously and induced the bleeding shortly (10 min) thereafter. Differences in
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
Bang/Berstad/Talstad
Heparin and Bleeding
side effect, not only of the traditional unfrac tionated heparin, but also of the new, low molecular weight heparins [ 19]. Whether the antithrombotic/haemorrhagic effect ratio in man is improved by the low molecular weight heparins is so far not clearly estab lished.
Acknowledgements The authors are grateful to Prof. Jon Lekven, Head of the Surgical Research Institute, Haukeland University Hospital, and his assistants for providing excellent laboratory facilities.
References 1 Salzman EW: Low-molecular-weight heparin. Is small beautiful? N Engl J Med 1986;315:957-959. 2 Levine MN, Planes A, Hirsh J, et al: The relation ship between anti-factor Xa level and clinical out come in patients receiving enoxaparine low mo lecular weight heparin to prevent deep vein thrombosis after hip replacement. Thromb Haemost 1989;62:940-944. 3 Salzmann EW, Dcykin D, Mayer Shapiro R, et al: Managment of heparin therapy. Controlled pro spective trial. N Engl J Med 1975:292:1046— 1050. 4 Cade JF, Buchanan MR, Boneu B, et al: A com parison of the antithrombotic and haemorrhagic effect of low molecular weight heparin fractions: The influence of the method of preparation. Thromb Res 1984;35:627-636. 5 Holmer E, Mattson C, Nilsson S: Anticoagulant and antithrombotic effects of heparin and low molecular weight heparin fragments in rabbits. Thromb Res 1982;25:475-485. 6 Borchgrewink CF: The mechanism of the pro longed bleeding time provoked by dextran sul phate or by excessive doses of heparin. Acta Med Scand 1961;170:365-373. 7 Mielke CH, Kaneshiro MM, Maher IA, et al: The standardized normal Ivy bleeding time and its prolongation by aspirin. Blood 1969;34:204215.
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
drug absorption and metabolism were thus avoided. The reduced bleeding previously demon strated in response to low molecular weight heparins as compared with unfractionated heparin in platelet-dependent bleeding mod els has been ascribed to the different effects of these heparins on platelets [4, 15], Platelet interference could possibly explain the present results in the tail skin, but not in the gastric mucosa. Here the haemostasis seems largely independent of the platelets, because potent inhibitors of platelets, such as prosta cyclin or thromboxane receptor antagonists, were found not to prolong the bleeding time in this bleeding model [8], It is, therefore, unlikely that the different haemorrhagic ef fects of the heparins are due to their different effects on the platelets. Our results confirm that the antihaemo static effect of a particular heparin prepara tion is not fully expressed by its anti-Xa activity. It is also recognized that the haem orrhagic effects of various heparins are not strongly connected with their different ef fects on global assays of anticoagulation such as the activated partial prothrombin time [16], However, a significant correlation has been found between lipase release and bleed ing time in response to heparin and other sulphated polysaccharides [17], and recently it has been demonstrated that some basic disaccharide units of heparin are ten-fold more potent inhibitors of haemostasis than heparin itself [ 18], These various results sup port the notion that the precise mechanism by which heparin induces bleeding is not known. The clinical relevance of our findings re mains to be shown. No appropriate experi mental human model seems to exist. Clinical experience indicates that bleeding is a major
35
8 Whittle BJR, Kauffman GL, Moncada S: Hemo static mechanisms, independent of platelet aggre gation, arrest gastric mucosal bleeding. Proc Natl Acad Sci USA 1986;83:5683-5687. 9 Bang C, Berstad A, Talstad I: Gastric mucosal bleeding after unfractionated and low molecular weight heparin in rats. Scand J Gastroenterol 1990;25:379-382. 10 Mersereau WA, Hinchey EJ; Effect of gastric acid ity on gastric ulceration induced by hemorrhage in the rat, utilizing a gastric chamber technique. Gas troenterology 1973:64:1130-1135. 11 Berstad A, Almodovar K, Berstad K, et al: Effect of tranexamic acid on gastric bleeding in rats. Scand J Gastroenterol 1988;23:402-406. 12 Dejana E, Villa S, de Gaetano G: Bleeding time in rats: A comparison of different experimental con ditions. Thromb Haemost 1982;48:108-111. 13 Dawes J, Bara L, Billaud E, et al: Relationship between biological activity and concentration of a low molecular weight heparin (PK10196) and un fractionated heparin after intravenous and subcu taneous administration. Haemostasis 1986; 16: 116-122. 14 Pangrazzi J, Abbadini M, Zameta M, et al: Anti thrombotic and bleeding effects of a low molecu lar weight heparin fraction. Biochem Pharmacol 1985;34:3305-3308. 15 Fernandez F, N’Guyen P, Van Ryn J, et al: Hem orrhagic doses of heparin and other glycosaminoglycans induce a platelet defect. Thromb Res 1986;43:491-495.
Bang/Berstad/T alstad
16 Buchanan MR, Ofosu FA, Fernandez F, et al: Lack of relationship between enhanced bleeding induced by heparin and other sulfated polysaccha rides and enhanced catalysis of thrombin inhibi tion. Semin Thromb Hemost 1986;12:324-327. 17 Barrcwcliffe TW, Merton RE, Gray E, et al: Hep arin and bleeding: An association with lipase re lease. Thromb Haemost 1988;60:434-436. 18 Nader HB, Tersariol ILS, Dietrich CP: Structural requirements of heparin disaccharides responsible for hemorrhage: Reversion of the antihemostatic effect of ATP. Faseb J 1989;3:2420-2424. 19 Bergqvist D, Burmark US, Frisell J, et al: Prospec tive double-blind comparison between fragmin and conventional low-dose heparin: Thromboprophylactic effect and bleeding complications. Hae mostasis 1986; 16(suppl 2): 11 —18.
Received: August 17, 1990 Accepted in revised form: October 4, 1990 Christen J. Bang, MD Section of Gastroenterology Medical Department Haukeland University Hospital N-5021 Bergen (Norway)
Downloaded by: King's College London 137.73.144.138 - 5/22/2018 12:11:13 PM
36
