BritishJournal of Haernatology, 1979.41,95-103.
Serial Changes in Coagulation and Viscosity during Sickle-Cell Crisis S. G. N. RICHARDSON, K. B. M A ~ H E W J.SSTUART, , A. M. GEDDES A N D R. M. WILCOX
Department of Huematology, Queen Elizabeth Hospital, Birmingham, and Department of Communicable and Tropical Diseases, East Birmingham Hospital (Received 23 February 1978; accepted for publication
24
April 1978)
SUMMARY. Coagulation activity and whole-blood viscosity were measured in the steady state, and serially during painful crisis, in eight patients with sickle-cell anaemia. Platelet and coagulation activation occurred in the steady state and became more pronounced early in crisis. Whole-blood viscosity increased during crisis in parallel with plasma fibrinogen. Similar changes were found in a parallel study of 20 patients with localized bacterial or viral infection who did not have sickle-cell anaemia. Reports of platelet activation, hypercoagulability, and hyperviscosity during painful crisis therefore reflect secondary changes arising from vascular stasis, precipitating infection, and an acute-phase protein reaction. Although secondary, these changes may contribute to vascular occlusion by an additive effect in vessels already partially occluded by sickled cells. The importance of coagulation activation in the pathogenesis of vaso-occlusive (painful) crisis in sickle-cell anaemia is not known. Histological studies have shown that blood vessels adjacent to infarcted areas are occluded by fibrin as well as by sickled cells (Diggs, 1965) and various coagdation changes during crisis have also been described. Plasma fibrinogen may be elevated in the presence (Green et al, 1970) or absence (Gordon et al, 1974; Richardson et a l , 1976) of precipitating infection. Factor VIII coagulation activity (FVIII: C) may be elevated both in crisis and in the steady state (Abildgaard et al, 1967; Green et al, 1970). Crisis may also be accompanied by a fall in platelet count (Gordon et al, 1974), shortened platelet survival (Haut et al, 1973) and evidence of platelet activation (Barnhart et al, 1973; Stuart et al, 1974; Gruppo et a l , 1977). Coagulation activation of this type may contribute to vascular sludging, and also promote further sickling, by causing an increase in whole-blood viscosity. While viscosity is already slightly raised in the steady state of sickle-cell disease, in part due to irreversibly sickled cells (Chien et al, 1970), there is a further significant increase early in crisis in parallel with a rise in plasma fibrinogen (Richardson et al, 1976). Fibrinogen is a major determinant of whole-blood viscosity, particularly a t low shear-rates (Chien et al, 1967), but the formation of large molecular weight complexes of fibrin monomer with intact fibrinogen causes an even greater increase in viscosity than fibrinogen alone (Copley et al, 1966). These complexes may therefore Correspondence: DrJohn Stuart, Consultant Haematologist, Queen Elizabeth Hospital, Birmingham B I s zTH. 0007-1048/79/0100-0095$02.00 011979 Blackwell Scientific Publications
95
96
S. G. N . Richardson et a1
contribute substantially to hyperviscosity and sludging, and Alkjaersig et a1 (1976) have demonstrated the presence of these complexes in the plasma a t the onset of, and possibly preceding, crisis. It remains uncertain, however, whether these changes are of aetiological significance or whether they occur as a secondary response to sickling and local stasis. The present study was therefore designed to investigate serial changes in platelet activation, thrombin generation, blood coagulation, and blood viscosity during crisis in comparison with steady state values. Many coagulation proteins are, however, acute-phase reactants (proteins which increase in plasma as a non-specific response within 24-48 h of tissue damage or inflammation) and since the onset of crisis is often associated with localized bacterial or viral infection, a parallel study was made in patients with infection but without sickle-cell anaemia, to differentiate the acute-phase reaction arising from infection. PATIENTS AND METHODS
Patients with Sickle-Cell Anaemia Seven patients with homozygous sickle-cell anaemia (SS), and one with sickle-cell/P-thalassaemia (SIP-thal), of mean age 16.3 years (range I 1-25 years) were each studied on an average of two occasions in the steady state. The mean steady state value for each coagulation parameter was then determined. Nine episodes of painful crisis requiring hospital admission occurred in these patients and were preceded o r accompanied by bacterial or viral infection. Three episodes were associated with pneumococcal or H.influenzae chest infection while, in the remaining six, crisis was preceded by a sore throat or 'flu-like illness with purulent sputum from which no organism was isolated. The mean duration of symptoms ofinfection prior to treatment was 3.3 d (range 1-7 d) and the mean duration of fever after admission was 7.2 d (range 4-10 d). All patients were treated with oral or intravenous hydration, antibiotics and analgesics. Crisis pain was present on average for 24 h prior to admission and lasted for 5-8 d (mean 6.4 d) after admission. Patients with Infections A total of 20 patients (mean 28.6 years, range 16-48 years), who did not have sickle-cell disease but who were admitted to an Infectious Diseases unit with bacterial ( I I ) or viral (nine) infection, was studied (Table I). The mean duration of symptoms of infection prior to treatment was 2.8 d (range 1-9 d), all were febrile on admission, and the mean duration of fever after treatment commenced was 2.6 d (range 1-5 d). Control Group A control group of 22 healthy asymptomatic white subjects (mean age 28.7 years, range 18-45 years) was studied to establish reference values for the tests employed. Coagulation Tests Blood was obtained with minimal venous occlusion; 9 ml were collected into I ml of 3 . 1 3 YO sodium citrate on melting ice and the following coagulation tests were performed on the same day: factor V assay (Wolf, 1953); FVIII: C two-stage assay (modification of Poole & Robinson,
Coagulation-Viscosity in Sickle-Cell Crisis
97
TABLE I. Clinical details of zo patients without sickle-cell anaemia admitted to an Infectious Diseases unit Type of infection Bacterial Strep. pyogenes Tonsillitis Sinusitis Strep. p y o p n e s Cellulitis Strep. pyogenes Erysipelas Strep. pyogenes Bronchopneumonia H. injhenzae Lobar pneumonia Strep. pneumoniae Meningitis N. meningitidis Urinary tract infection Esch. coli Paratyphoid Salm. paratyphi Viral Influenza1 illness Upper respiratory tract infection Herpes zoster Viral stomatitis Glandular fever Total
No. ofpatients
4 I I I
2 20
1959); and serial-dilution protamine sulphate test for fibrin monomer (Gurewich & Hutchinson, 1971). The heparin-thrombin clotting time (HTCT) was measured by the method of O’Brien (1974) and the plasma heparin neutralizing activity was expressed as the ratio of the heparin-thrombin clotting time of pooled normal plasma divided by that of the test plasma (HTCT ratio). Fibrinogen and cryofibrinogen. An additional 4.5 ml of blood were collected into 0.5 ml of 3. I 3 YOsodium citrate and kept at room temperature for measurement of thrombin-clottable fibrinogen (Ellis & Stransky, 1961) and cryofibrinogen (Howie et al, 1971). Factor V I I I R :Ag. Citrated plasma was stored at -zo°C and factor VIII antigenic activity (FVIIIR:Ag) assayed in batches (Laurell, 1966) using Behringwerke antiserum. Platelet counts were performed using an improved Neubauer counting chamber (Brecher & Cronkite, 1950), and irreversibly sickle cell (ISC) counts by the method of Serjeant et a1 (1969). Serum antithrombin III activity was measured by a modification of von Kaulla’s method (Peterson et al, 1970) and blood, with added epsilon aminocaproic acid (10 mglml), was tested for fibrin degradation products (FDP) using the Thrombo-Wellcotest latex kit method (Garvey & Black, I 972).
Whole-Blood Viscosity Blood was collected into solid EDTA (1.5 mg/ml) for measurement of whole-blood viscosity a t 37°C a t shear rates of 23 s-l and 230 5 - l using a Wells-Brookfield viscometer as described previously (Richardson et al, 1 ~ 7 6and ) corrected to a standard haematocrit of 0.45. Statistical Analysis Statistical significance was determined by Student’s t test.
98
S . G. N.Richardson
et at
RESULTS In Table 11, the mean levels of coagulation parameters and whole-blood viscosity in eight patients with sickle-cell anaemia in the steady state, and also five white subjects who had previously undergone splenectomy, are compared with the mean levels obtained in 22 healthy control subjects. The sickle-cell anaemia patients, compared with the healthy control group, showed significantly higher levels of FVIII: C, FVIIIR: Ag, antithrombin 111, and corrected whole-blood viscosity at both shear-rates (23 s-l and 230 s-l), together with lower levels of factor V and higher heparin-thrombin clotting time ratio. T h e splenectomized patients showed no significant coagulation abnormality compared with the healthy control subjects. Serial coagulation and whole-blood viscosity changes in sickle-cell painful crisis are shown in Table 111. Most patients were studied daily for the first 3 d and subsequently less frequently; the mean levels obtained on each of the first 3 d, and on days 4-5,6-8 and 9-14 are compared with the mean steady state values. Significant changes were observed by the third day in plasma fibrinogen, cryofibrinogen, heparin-thrombin clotting time ratio, antithrombin 111, FVIIIR : Ag/FVIII :C ratio, and whole-blood viscosity at both shear rates. Fibrin degradation products (FDP) were detected at low levels throughout crisis, but not in the steady state, and the serial-dilution protamine sulphate test for fibrin monomer was negative in all cases throughout crisis. There was no significant change in the percentage of ISC in crisis compared with the steady state. In Table IV, the mean coagulation and whole-blood viscosity measurements, taken serially a t similar intervals during bacterial or viral infection in 20 patients without sickle-cell anaemia, are compared with the mean values obtained in the healthy control group. Significant changes were observed in fibrinogen, cryofibrinogen, FVIII : C, FVIIIR :Ag, platelet count, heparinthrombin clotting time ratio, antithrombin I11 level, FVIIIR: Ag/FVIII :C ratio, and wholeblood viscosity. Serum FDP were detected throughout the period of illness a t levels comparable to those obtained during painful crisis. No fibrin monomer was demonstrated by the serial-dilution protamine sulphate test in any of the infected patients.
DISCUSSION O u r eight patients with sickle-cell anaemia in the steady state showed a significant elevation of FVIII :C, FVIIIR :Ag, plasma heparin-neutralizing activity, antithrombin 111, and blood viscosity, and a reduced level of factor V activity, compared with 22 healthy white controls. The changes in factor V and FVIII:C are similar to the findings of Leslie et al (197s) who compared steady state patients with normal, age-matched black controls. Raised factor VIII levels have been reported in other haemolytic anaemias (Egeberg, 1962; Mannucci et al, r969), and may reflect increased reticulo-endothelial cell activity. Leslie et a1 (1975) also found a significantly elevated mean platelet count in the steady state and this difference in statistical significance between the two studies may reflect the younger mean age of our patient group (16.3 years t, 26.8 years) which included some young patients in whom autosplenectomy from repeated microinfarction is often incomplete. Leslie et al (1975) failed to show any rise in antithrombin I11 but they used an immunoassay in contrast to our bioassay.
0.j1f0.24 0.25+0.19
NS NS
3.57k0.37
3.8720.70
NS NS
SS+S/thalsteadystate(X)
V3
P
Serial Changes in Coagulation and Viscosity during Sickle-Cell Crisis S. G. N. RICHARDSON, K. B. M A ~ H E W J.SSTUART, , A. M. GEDDES A N D R. M. WILCOX
Department of Huematology, Queen Elizabeth Hospital, Birmingham, and Department of Communicable and Tropical Diseases, East Birmingham Hospital (Received 23 February 1978; accepted for publication
24
April 1978)
SUMMARY. Coagulation activity and whole-blood viscosity were measured in the steady state, and serially during painful crisis, in eight patients with sickle-cell anaemia. Platelet and coagulation activation occurred in the steady state and became more pronounced early in crisis. Whole-blood viscosity increased during crisis in parallel with plasma fibrinogen. Similar changes were found in a parallel study of 20 patients with localized bacterial or viral infection who did not have sickle-cell anaemia. Reports of platelet activation, hypercoagulability, and hyperviscosity during painful crisis therefore reflect secondary changes arising from vascular stasis, precipitating infection, and an acute-phase protein reaction. Although secondary, these changes may contribute to vascular occlusion by an additive effect in vessels already partially occluded by sickled cells. The importance of coagulation activation in the pathogenesis of vaso-occlusive (painful) crisis in sickle-cell anaemia is not known. Histological studies have shown that blood vessels adjacent to infarcted areas are occluded by fibrin as well as by sickled cells (Diggs, 1965) and various coagdation changes during crisis have also been described. Plasma fibrinogen may be elevated in the presence (Green et al, 1970) or absence (Gordon et al, 1974; Richardson et a l , 1976) of precipitating infection. Factor VIII coagulation activity (FVIII: C) may be elevated both in crisis and in the steady state (Abildgaard et al, 1967; Green et al, 1970). Crisis may also be accompanied by a fall in platelet count (Gordon et al, 1974), shortened platelet survival (Haut et al, 1973) and evidence of platelet activation (Barnhart et al, 1973; Stuart et al, 1974; Gruppo et a l , 1977). Coagulation activation of this type may contribute to vascular sludging, and also promote further sickling, by causing an increase in whole-blood viscosity. While viscosity is already slightly raised in the steady state of sickle-cell disease, in part due to irreversibly sickled cells (Chien et al, 1970), there is a further significant increase early in crisis in parallel with a rise in plasma fibrinogen (Richardson et al, 1976). Fibrinogen is a major determinant of whole-blood viscosity, particularly a t low shear-rates (Chien et al, 1967), but the formation of large molecular weight complexes of fibrin monomer with intact fibrinogen causes an even greater increase in viscosity than fibrinogen alone (Copley et al, 1966). These complexes may therefore Correspondence: DrJohn Stuart, Consultant Haematologist, Queen Elizabeth Hospital, Birmingham B I s zTH. 0007-1048/79/0100-0095$02.00 011979 Blackwell Scientific Publications
95
96
S. G. N . Richardson et a1
contribute substantially to hyperviscosity and sludging, and Alkjaersig et a1 (1976) have demonstrated the presence of these complexes in the plasma a t the onset of, and possibly preceding, crisis. It remains uncertain, however, whether these changes are of aetiological significance or whether they occur as a secondary response to sickling and local stasis. The present study was therefore designed to investigate serial changes in platelet activation, thrombin generation, blood coagulation, and blood viscosity during crisis in comparison with steady state values. Many coagulation proteins are, however, acute-phase reactants (proteins which increase in plasma as a non-specific response within 24-48 h of tissue damage or inflammation) and since the onset of crisis is often associated with localized bacterial or viral infection, a parallel study was made in patients with infection but without sickle-cell anaemia, to differentiate the acute-phase reaction arising from infection. PATIENTS AND METHODS
Patients with Sickle-Cell Anaemia Seven patients with homozygous sickle-cell anaemia (SS), and one with sickle-cell/P-thalassaemia (SIP-thal), of mean age 16.3 years (range I 1-25 years) were each studied on an average of two occasions in the steady state. The mean steady state value for each coagulation parameter was then determined. Nine episodes of painful crisis requiring hospital admission occurred in these patients and were preceded o r accompanied by bacterial or viral infection. Three episodes were associated with pneumococcal or H.influenzae chest infection while, in the remaining six, crisis was preceded by a sore throat or 'flu-like illness with purulent sputum from which no organism was isolated. The mean duration of symptoms ofinfection prior to treatment was 3.3 d (range 1-7 d) and the mean duration of fever after admission was 7.2 d (range 4-10 d). All patients were treated with oral or intravenous hydration, antibiotics and analgesics. Crisis pain was present on average for 24 h prior to admission and lasted for 5-8 d (mean 6.4 d) after admission. Patients with Infections A total of 20 patients (mean 28.6 years, range 16-48 years), who did not have sickle-cell disease but who were admitted to an Infectious Diseases unit with bacterial ( I I ) or viral (nine) infection, was studied (Table I). The mean duration of symptoms of infection prior to treatment was 2.8 d (range 1-9 d), all were febrile on admission, and the mean duration of fever after treatment commenced was 2.6 d (range 1-5 d). Control Group A control group of 22 healthy asymptomatic white subjects (mean age 28.7 years, range 18-45 years) was studied to establish reference values for the tests employed. Coagulation Tests Blood was obtained with minimal venous occlusion; 9 ml were collected into I ml of 3 . 1 3 YO sodium citrate on melting ice and the following coagulation tests were performed on the same day: factor V assay (Wolf, 1953); FVIII: C two-stage assay (modification of Poole & Robinson,
Coagulation-Viscosity in Sickle-Cell Crisis
97
TABLE I. Clinical details of zo patients without sickle-cell anaemia admitted to an Infectious Diseases unit Type of infection Bacterial Strep. pyogenes Tonsillitis Sinusitis Strep. p y o p n e s Cellulitis Strep. pyogenes Erysipelas Strep. pyogenes Bronchopneumonia H. injhenzae Lobar pneumonia Strep. pneumoniae Meningitis N. meningitidis Urinary tract infection Esch. coli Paratyphoid Salm. paratyphi Viral Influenza1 illness Upper respiratory tract infection Herpes zoster Viral stomatitis Glandular fever Total
No. ofpatients
4 I I I
2 20
1959); and serial-dilution protamine sulphate test for fibrin monomer (Gurewich & Hutchinson, 1971). The heparin-thrombin clotting time (HTCT) was measured by the method of O’Brien (1974) and the plasma heparin neutralizing activity was expressed as the ratio of the heparin-thrombin clotting time of pooled normal plasma divided by that of the test plasma (HTCT ratio). Fibrinogen and cryofibrinogen. An additional 4.5 ml of blood were collected into 0.5 ml of 3. I 3 YOsodium citrate and kept at room temperature for measurement of thrombin-clottable fibrinogen (Ellis & Stransky, 1961) and cryofibrinogen (Howie et al, 1971). Factor V I I I R :Ag. Citrated plasma was stored at -zo°C and factor VIII antigenic activity (FVIIIR:Ag) assayed in batches (Laurell, 1966) using Behringwerke antiserum. Platelet counts were performed using an improved Neubauer counting chamber (Brecher & Cronkite, 1950), and irreversibly sickle cell (ISC) counts by the method of Serjeant et a1 (1969). Serum antithrombin III activity was measured by a modification of von Kaulla’s method (Peterson et al, 1970) and blood, with added epsilon aminocaproic acid (10 mglml), was tested for fibrin degradation products (FDP) using the Thrombo-Wellcotest latex kit method (Garvey & Black, I 972).
Whole-Blood Viscosity Blood was collected into solid EDTA (1.5 mg/ml) for measurement of whole-blood viscosity a t 37°C a t shear rates of 23 s-l and 230 5 - l using a Wells-Brookfield viscometer as described previously (Richardson et al, 1 ~ 7 6and ) corrected to a standard haematocrit of 0.45. Statistical Analysis Statistical significance was determined by Student’s t test.
98
S . G. N.Richardson
et at
RESULTS In Table 11, the mean levels of coagulation parameters and whole-blood viscosity in eight patients with sickle-cell anaemia in the steady state, and also five white subjects who had previously undergone splenectomy, are compared with the mean levels obtained in 22 healthy control subjects. The sickle-cell anaemia patients, compared with the healthy control group, showed significantly higher levels of FVIII: C, FVIIIR: Ag, antithrombin 111, and corrected whole-blood viscosity at both shear-rates (23 s-l and 230 s-l), together with lower levels of factor V and higher heparin-thrombin clotting time ratio. T h e splenectomized patients showed no significant coagulation abnormality compared with the healthy control subjects. Serial coagulation and whole-blood viscosity changes in sickle-cell painful crisis are shown in Table 111. Most patients were studied daily for the first 3 d and subsequently less frequently; the mean levels obtained on each of the first 3 d, and on days 4-5,6-8 and 9-14 are compared with the mean steady state values. Significant changes were observed by the third day in plasma fibrinogen, cryofibrinogen, heparin-thrombin clotting time ratio, antithrombin 111, FVIIIR : Ag/FVIII :C ratio, and whole-blood viscosity at both shear rates. Fibrin degradation products (FDP) were detected at low levels throughout crisis, but not in the steady state, and the serial-dilution protamine sulphate test for fibrin monomer was negative in all cases throughout crisis. There was no significant change in the percentage of ISC in crisis compared with the steady state. In Table IV, the mean coagulation and whole-blood viscosity measurements, taken serially a t similar intervals during bacterial or viral infection in 20 patients without sickle-cell anaemia, are compared with the mean values obtained in the healthy control group. Significant changes were observed in fibrinogen, cryofibrinogen, FVIII : C, FVIIIR :Ag, platelet count, heparinthrombin clotting time ratio, antithrombin I11 level, FVIIIR: Ag/FVIII :C ratio, and wholeblood viscosity. Serum FDP were detected throughout the period of illness a t levels comparable to those obtained during painful crisis. No fibrin monomer was demonstrated by the serial-dilution protamine sulphate test in any of the infected patients.
DISCUSSION O u r eight patients with sickle-cell anaemia in the steady state showed a significant elevation of FVIII :C, FVIIIR :Ag, plasma heparin-neutralizing activity, antithrombin 111, and blood viscosity, and a reduced level of factor V activity, compared with 22 healthy white controls. The changes in factor V and FVIII:C are similar to the findings of Leslie et al (197s) who compared steady state patients with normal, age-matched black controls. Raised factor VIII levels have been reported in other haemolytic anaemias (Egeberg, 1962; Mannucci et al, r969), and may reflect increased reticulo-endothelial cell activity. Leslie et a1 (1975) also found a significantly elevated mean platelet count in the steady state and this difference in statistical significance between the two studies may reflect the younger mean age of our patient group (16.3 years t, 26.8 years) which included some young patients in whom autosplenectomy from repeated microinfarction is often incomplete. Leslie et al (1975) failed to show any rise in antithrombin I11 but they used an immunoassay in contrast to our bioassay.
0.j1f0.24 0.25+0.19
NS NS
3.57k0.37
3.8720.70
NS NS
SS+S/thalsteadystate(X)
V3
P
