An evaluation of screening tests for defects in the protein C pathway: commercial kits lack sensitivity and specificity.

A comparative evaluation of four commercial coagulation test kits for screening the protein C pathway kits was performed at two centres. The tests were Acticlot V-OUT (V-OUT), the PCA test (PCA), the GradiThrom PCP test (PCP) and ProC Global (ProC). Reference ranges were established in 40 normal subjects and, with the exception of V-OUT and ProC, significant differences were observed between males and females. Consequently, sex-specific normal cut-off values (fifth percentile) were used that led to greatly improved sensitivity when compared with the manufacturers' recommended cut-off values. Plasma from patients with factor V Leiden (n = 23), congenital protein S deficiency (n = 19), congenital protein C deficiency (n = 11), lupus anticoagulant (n = 17) and thrombophilia with no demonstrable protein C pathway defect (n = 20) were tested. All kits showed 100% sensitivity to factor V Leiden, but sensitivity was variable for protein C deficiency (27-73%), and poor for protein S deficiency (29-35%). The lupus anticoagulant affected all kits to some degree, with 29-35% giving values below the fifth percentile of normal, whereas all kits gave 1/20 unexpected abnormal results in the thrombophilia group, with the same sample accounting for the abnormal results in three of the four kits. Overall sensitivity and specificity, respectively, for defects in the protein C pathway were: V-OUT, 60 and 91%; PCA, 70 and 86%; PCP 75 and 94%; and ProC, 66 and 88%. We conclude that all four kits lack the sensitivity and specificity required for routine laboratory screening for defects in the protein C pathway and cannot replace assays for the individual proteins of this system.

The importance of locally derived reference ranges and standardized calculation of dilute Russell's viper venom time results in screening for lupus anticoagulant.

Three commercial dilute Russell's viper venom time (DRVVT) kits were evaluated at four UK centres experienced at performing lupus anticoagulant (LA) tests. Each centre established a normal reference range for the DRVVT ratio calculated against local pooled normal plasma from 20 healthy normal subjects. Plasma from LA-positive patients and LA-negative thrombophilia patients was also tested. DRVVT ratios and the degree of correction were assessed in a variety of ways to reflect not only the UK national Guidelines, but also the manufacturers' recommendations. The reference range data showed a normal distribution in each case, but considerable variation in the mean and SD between the centres and reagents, with the mean +2SD value ranging from 1.06 to 1.19. The use of an arbitrary DRVVT ratio of < 1.1 as the cut-off value for normality, which is applied in many laboratories, is therefore inappropriate. Although no single kit had a clear overall advantage in terms of sensitivity and specificity, the way in which the screen and confirmation data were analysed had a major impact on the interpretation of the results. A data analysis method employing a mean plus two standard deviations (SDs) cut-off for normality, and judgement regarding confirmation of LA based on a percentage correction of DRVVT ratio, was the simplest and most consistent, with overall sensitivity and specificity values of 81% and 94%, respectively, for uncomplicated LA-positive and LA-negative thrombophilia samples. We conclude that the 1991 BSCH Guidelines are in need of revision, each laboratory should establish its own normal reference range for the DRVVT ratio and a common method should be used for calculating the degree of correction with confirmation reagents, so that LA results can be correctly interpreted between laboratories. Standardizing DRVVT interpretation in this way should improve the consistency of LA detection.

Late-form hemorrhagic disease of the newborn: a fatal case report with illustration of investigations that may assist in avoiding the mistaken diagnosis of child abuse.

Hemorrhagic disease of the newborn (HDN) is usually a self-limiting hemorrhagic disorder of childhood that occurs as a result of vitamin K deficiency. It may be defined as early or late form depending on the time of onset related to birth. HDN is recognized as one of several bleeding disorders that can mimic the findings of nonaccidental head injury and may lead to a mistaken diagnosis of child abuse. We present a single fatal case of late-onset HDN with illustration of hematologic assays that can be performed to assist the pathologist in making the correct diagnosis of HDN.

Recombinant and tissue extract thromboplastins for determination of international normalised ratio in over-anticoagulated patients.

The international normalised ratio (INR)/international sensitivity index (ISI) system is established for calibration of thromboplastins for laboratory monitoring of oral anticoagulant therapy. The calibration procedure employs patients stabilised on oral anticoagulants, and is therefore validated for patients within the therapeutic range. For practical reasons, the system is used for patients at all levels of therapy, including over-anticoagulated patients with particularly low levels of factors II, VII and X. We studied patients within and above the therapeutic range, using a thromboplastin containing recombinant human tissue factor (Innovin) and two tissue extract thromboplastins. In samples with INRs from 2.0 to 4.0, there was good agreement between results obtained with the three systems (mean INRs within 4% of each other). In patients with INRs > 4.0, results with a human placental extract reagent (Thromborel S) were similar to those obtained with a rabbit brain thromboplastin (IL PT Fib Hs Plus); mean INRs were 6.30 and 6.32 respectively (not significant). Results with Innovin (mean INR: 7.67) were significantly (P < 0.001) greater (on average by 22%) than those obtained with the other two materials. The discrepancy between results with different reagents negatively correlated with factor VII levels. Thus, the lower the factor VII level, the greater was the discrepancy between INRs. Unexpectedly, there was a positive correlation between factor V level and the difference between INRs with different reagents. Thus, the higher the factor V level, the greater was the discrepancy between INRs. The effect of these differences at higher INRs on patient management is unknown, but the recently revised UK guidelines recommend that management of these patients should be influenced by clinical factors, reducing the relative importance of discrepancies between results obtained with different systems.

Fibrinogen and von Willebrand factor in IDDM: relationships to lipid vascular risk factors, blood pressure, glycaemic control and urinary albumin excretion rate: the EURODIAB IDDM Complications Study.

The interrelationships between fibrinogen, von Willebrand factor, a marker of vascular endothelial cell damage, and serum lipids were explored in well-characterised subjects with insulin-dependent diabetes mellitus. The 2091 subjects were enrolled into a cross-sectional, clinic-based study of complications, from 16 European countries: the EURODIAB IDDM Complications study. The anticipated significant relationships between both plasma fibrinogen and plasma von Willebrand factor concentrations and age and glycaemic control, and between fibrinogen and body mass index, were noted. Fibrinogen, adjusted for age and glycated haemoglobin concentration, was also related to smoking habits and was higher in the quartiles with highest systolic and diastolic blood pressures. There was a clustering of vascular risk factors, with a positive relationship between plasma fibrinogen and serum triglyceride concentrations in both genders and between fibrinogen and total cholesterol in males. An inverse relationship between fibrinogen and high density lipoprotein cholesterol was also apparent in males. A prominent feature was a positive relationship between both fibrinogen and von Willebrand factor and albumin excretion rate (p < 0.001 and p < 0.003 respectively) in those with retinopathy but not in those without this complication. In view of previous observations on blood pressure and albuminuria in these subjects the findings are consistent with the hypothesis that microalbuminuria and increased plasma von Willebrand factor are due to endothelial cell perturbation in response to mildly raised blood pressure in subjects with retinopathy. Fibrinogen may also contribute to microvascular disease and its relationships to lipid vascular risk factors suggest a possible pathogenic role in arterial disease in diabetes.

Activated protein C resistance in homozygous sickle cell disease.

The activated protein C resistance (APC-R) ratios in 50 patients with steady state homozygous sickle cell (SS) disease and 59 healthy AA controls was measured. There was a significant reduction in median APC-R ratio in sickle cell disease compared to controls. This reduction in APC-R ratio was not explained by (1) the presence of the factor V Leiden, found in only one of 165 patients with SS disease including those tested for APC-R, or (2) the presence of lupus anticoagulants. However, the raised levels of factor VIIIC in SS patients in this study may be contributing to increased resistance to APC, which in turn may contribute to the vaso-occlusive complications of SS disease.

Characterization and specificity of anti-endothelial cell membrane antibodies and their relationship to thrombosis in primary antiphospholipid syndrome (APS).

Immunoblotting was used to detect antibodies reacting with membrane and cytosol preparations of human umbilical vein endothelial cells (HUVEC), fibroblasts and a T lymphoma line HUT78 in 18 patients with anticardiolipin antibodies (ACA) (14 of whom had had a thrombotic event), 11 patients with a recent myocardial infarction and 17 controls. Multiple membrane-specific antibodies to HUVEC were found in 10 of the patients with ACA (28 bands) and in nine of the patients with thromboses (27 bands) in contrast to only three of the patients with myocardial infarction (four bands) and one control (one band). The most frequently recognized HUVEC membrane epitopes were at 33 kD (four sera), 61-63 kD (five sera) and 76-79 kD (four sera). Although cross-reactivity with fibroblast and/or HUT78 membranes was seen at 33 kD, binding at 61-63 kD and 76-79 kD was specific for endothelial membranes. Although no correlations with the presence and titre of ACA were seen, HUVEC membrane-specific antibodies showed a correlation with venous thrombotic events.

The development of antiphospholipid antibodies in haemophilia is linked to infection with hepatitis C.

In haemophilia the presence of antibodies to antiphospholipid has been linked with HIV infection, but other possibilities have not been fully explored and the specificity for various phospholipids not established. In order to investigate further the pathogenesis and clinical significance of these antibodies, we have determined IgG and IgM antibodies to a variety of phospholipids, including cardiolipin, in the serum of 52 haemophiliacs, and related our findings to the presence of antibodies to HIV and hepatitis C virus (HCV), as well as to clotting factor concentrate usage and blood markers of liver biochemistry. Our results demonstrate that the presence of infection with hepatitis C virus is strongly associated with raised serum levels of antiphospholipid antibody even in the absence of HIV infection. They suggest that earlier conclusions on the relationships of antiphospholipid to HIV infection in haemophilia should be revised and that chronic infection with the hepatitis C virus should be added to the list of infective causes for the development of antiphospholipid antibody.

Bilirubin, ferritin, D-dimers and erythrophages in the cerebrospinal fluid of patients with suspected subarachnoid haemorrhage but negative computed tomography scans.

AIM: To assess the diagnostic value of cerebrospinal fluid (CSF) spectrophotometry, cytology, ferritin, and D-dimer measurements in the investigation of suspected subarachnoid haemorrhage in patients with negative or equivocal computed tomography (CT) scans. METHODS: CSF specimens submitted for assessment of xanthochromia were examined for erythrophages using a cytospin preparation stained with Wright's stain, for ferritin using the Ciba-Corning Magic IRMA assay, D-dimers using the Dimertest 2 latex agglutination slide test, and for bilirubin by scanning spectrophotometry. The patients were divided into three groups for data analysis and the results compared with the existing methods, CT, and angiogram results. Final diagnoses were reviewed by a consultant neurologist. RESULTS: Thirty six patients were recruited. In those patients with confirmed subarachnoid haemorrhage CSF cytology had a low sensitivity and there were false negative results with both the D-dimer and ferritin assays. Eleven patients with a negative or equivocal CT scan underwent angiography, but only one aneurysm and no arterio-venous malformations or bleeding points were identified. In the patient with the aneurysm there was no laboratory evidence of subarachnoid haemorrhage. Six patients had CSF abnormalities detected by the special tests only and in none of these cases was subarachnoid haemorrhage confirmed. All results were normal in four out of five cases of traumatic tap. CONCLUSIONS: This is a small study, but it shows that, depending on the timing of the lumbar puncture, false negative results can occur with both ferritin and D-dimer measurements. It suggests that neither of these tests adds significantly to the information provided by CT, visualisation of CSF, and spectrophotometry and confirms that, despite the use of spectrophotometry, D-dimer and ferritin assays in selecting patients for angiography, the proportion of patients with negative CT scans and colourless CSF with demonstrable vascular lesions remains low.

A comparison of methods for the measurement of activated factor VII.

A number of different methods are available for the measurement of factor VIIa. Almost all of these employ ratios of two different measurements of factor VII. In order to determine which is the most sensitive to activated factor VII we have compared currently available methods in the following groups: two patients with haemophilia A following treatment with activated recombinant factor VII (rVIIa); 6 normal plasmas during cold promoted activation of factor VII; normal individuals (n = 23); and patients with unequivocal disseminated intravascular coagulation (DIC, n = 19). Factor VII was measured in an amidolytic assay (VII:Amid) and an antigen assay (VII:Ag). Clotting activity was measured using rabbit (VII:C Rab), human (VII:C Hum) and bovine (VII:C Bov) thromboplastin. Of the clotting assays the most sensitive to the presence of factor VIIa was that which utilised bovine thromboplastin. Amidolytic and immunological measurements were unaffected by the activity state of factor VII. The ratios VII:C Rab/VII:Ag and VII:C Rab/VII:Amid were insensitive to activated factor VII. The ratios most sensitive to the presence of factor VIIa were VII:C Bov/VII:Amid and VII:C Bov/VII:Ag. The ratios VII:C Bov/VII:C Rab and VII:C Bov/VII:C Hum are less sensitive but have the advantage for epidemiological studies of narrower reference ranges.

Screening for lupus anticoagulant and anticardiolipin antibodies in women with fetal loss.

Sixty six women with first or second trimester fetal loss were investigated for the presence of lupus anticoagulant by routine coagulation tests and the dilute Russell's viper venom time with a platelet neutralisation procedure, and for raised anticardiolipin antibodies by an enzyme linked immunosorbent assay. Of 35 women with recurrent fetal loss, seven were positive for lupus anticoagulant and six had increased IgG anticardiolipin antibodies, while of 31 women with only one or two episodes of fetal loss, one had lupus anticoagulant and none increased IgG anticardiolipin antibodies. These findings were significantly different. There was no difference in the incidence of increased IgM anticardiolipin antibodies between the two groups (three and two cases, respectively). A further 11 women with intrauterine death in the third trimester were studied and lupus anticoagulant and raised IgM anticardiolipin antibodies were found in one case. No woman was known to have systemic lupus erythematosus. It is concluded that lupus anticoagulant and increased IgG anticardiolipin antibodies are independently associated with recurrent first and second trimester fetal loss and that such cases should be investigated, even in the presence of otherwise good health, by a comprehensive methodological approach.

Inhibition of activated protein C and its cofactor protein S by antiphospholipid antibodies.

We have investigated the effects of purified IgG fractions from plasma containing the lupus anticoagulant (LAC) and/or IgG anticardiolipin antibody (ACA) on the degradation of factor Va by an activated protein C-protein S complex. Plasma samples from 10 patients were studied. LAC was detected by a Russell's Viper venom technique. ACA was determined by ELISA. IgG fractions were obtained from each plasma sample by protein A-Sepharose fractionation. This fraction was shown to exhibit ACA/LAC activity. Using purified activated protein C (APC), protein S and phosphatidylserine/phosphatidylcholine, factor Va degradation was assessed in the presence and absence of IgG fractions from LAC/ACA containing plasmas. After 2 min incubation the mean factor Va degradation by APC and protein S in the presence of IgG LAC/ACA fractions was 14% compared with 52% with normal IgG. A similar effect was seen when phospholipid was substituted by washed freeze-thawed platelets. Experiments employing varying concentrations of protein S and phospholipid revealed marked differences in respect of the inhibitory specificity of the different antiphospholipid antibodies. These results indicate that antiphospholipid antibodies have an inhibitory effect on the activated protein C/protein S complex and provide some explanation for a relationship between antiphospholipid antibodies and thrombosis.

Antibody-dependent cellular cytotoxicity of vascular endothelium: characterization and pathogenic associations in systemic sclerosis.

Ten sera from 48 patients with systemic sclerosis were found to be capable of producing cytotoxicity of human umbilical venous and arterial endothelium when co-cultured with peripheral blood mononuclear cells. Fractionation of sera on Ultrogel and the preparation of monomeric IgG by ion exchange and affinity chromatography suggested that the cytotoxicity was mediated by anti-endothelial antibodies capable of pre-sensitizing target cells in a mechanism that resembled antibody-dependent cellular cytotoxicity. These anti-endothelial antibodies together with C1q-binding immune complexes and anti-cardiolipin antibodies were found in 18 of 28 patients so investigated, suggesting that multiple immunological mechanisms may be involved in the pathogenesis of the vascular lesion of systemic sclerosis.

Anticardiolipin antibodies in systemic sclerosis: immunological and clinical associations.

Anticardiolipin antibodies of IgG/IgM class were detected in seven of 28 patients with systemic sclerosis including five of 16 patients severely affected by extensive visceral disease. This severely affected sub-group also showed significant elevations of plasma levels of von Willebrand factor antigen in 10 cases and serum C1q binding activity in seven cases respectively. This triple association raises the possibility that multiple immunological mechanisms are involved in the pathogenesis of systemic sclerosis and its vascular lesions.

Elevated von Willebrand factor antigen in systemic sclerosis: relationship to visceral disease.

Plasma levels of the factor VIII complex (von Willebrand factor antigen, factor VIII coagulant and ristocetin co-factor) were measured in 28 patients with systemic sclerosis. Elevated von Willebrand factor antigen was found in 12 patients overall and in 10 of 16 patients characterized by severe extensive visceral disease, with a resulting positive correlation between the extent of visceral involvement and the plasma level of von Willebrand factor antigen (r = 0.60, p less than 0.001). Factor VIII coagulant and ristocetin co-factor levels, however, frequently failed to parallel the increases of von Willebrand factor antigen, supporting the view that these increases were due to in vivo endothelial damage. The findings suggest that vascular damage is an important aspect of the visceral lesions of systemic sclerosis.

A method for the determination of activated factor VII using bovine and rabbit brain thromboplastins: demonstration of increased levels in disseminated intravascular coagulation.

Factor VII levels as measured in one stage clotting assays are dependent on the degree of activation of factor VII and on the type of thromboplastin used. Bovine thromboplastin (BT) is known to be more sensitive to factor VIIa than human brain thromboplastin and the different sensitivities of these reagents to VIIa have formed the basis of a method for determining VIIa in plasma (16). Since human thromboplastin is no longer widely available, we have re-evaluated this method using bovine and two rabbit thromboplastins (Manchester Reagent, RT MR and Diagen activated, RT Diagen). Activated factor VII was generated in normal plasma by cold activation for 24 hours. During this period, factor VII assays using bovine thromboplastin increased markedly with time whereas the assays using rabbit thromboplastins showed considerably less change. The ratio of factor VII (bovine thromboplastin)/factor VII (rabbit thromboplastin) is a sensitive index of activated factor VII in plasma. The mean ratio in normal plasmas (using BT and RT MR) was found to be 1.02 (+/- s.d., range 0.80-1.24). The ratio was measured in 29 samples from patients with disseminated intravascular coagulation (D.I.C.) and was found to be increased in 24 (mean 1.71, range 0.93-3.38).

Platelet function and coagulation in patients with hepatobiliary disorders receiving cefotetan prophylaxis.

Eighteen patients with hepatobiliary disorders undergoing invasive biliary tract investigations were given the cephalosporin cefotetan and platelet function and coagulation factors were monitored. No significant changes were observed in the group as a whole, although marked alterations were seen in three patients in association with clinical complications unrelated to the antibiotic. Although cefotetan is structurally similar to cephalosporins which have been associated with bleeding disturbances, it does not appear to induce such abnormalities in a high risk group of patients with normal renal function when given for five days intravenously at a dose of 1 g twice daily.