Postprandial Elevation of Activated Factor VII in Young Adults
The Northwick Park Heart Study found that factor VII (FVII) activity was a risk factor for ischemic heart disease, and other studies based on indirect assays of activated factor VII (FVIIa) found an elevation of FVIIa postprandially. We hypothesized that postprandial elevation of FVIIa would produce intermittent activation of factor X to Xa and, subsequently, prothrombin to thrombin. We chose to study postprandial activation of coagulation with a new assay specific for FVIIa that uses soluble tissue factor and with a prothrombin fragment 1+2 (F1+2) assay to detect the activation of prothrombin by factor Xa. We fed a high-fat breakfast (30 g/m2) to 30 healthy volunteer subjects (30.8±9.8 years; range, 20 to 49 years) on no medication. Fasting blood samples were collected for FVIIa, FVII antigen (FVIIag), and F1+2 as well as triglycerides and total and HDL cholesterol. A significant difference was found between fasting (2.82±1.49 ng/mL, mean±SD) and 6-hour postprandial (3.45±2.08 ng/mL) FVIIa levels (P<.004); FVIIag did not change significantly (mean, 0.89 U/mL fasting and 0.90 U/mL at 6 hours). In contrast, F1+2 levels were slightly lower 6 hours postprandially than fasting (median, 0.39 versus 0.44 nmol/L, P<.02). Four-hour postprandial triglyceride levels correlated significantly (ρ=0.51, P<.02) with 6-hour postprandial FVIIag but not with 6-hour postprandial FVIIa. Postprandial F1+2 levels (at 6 hours) correlated significantly (ρ=0.39, P<.04) with fasting FVIIag levels but not with 6-hour postprandial FVIIa levels. Thus, the basal FVIIag level, in the fasting state, may be involved in control of the generation of F1+2. We found a postprandial increase in FVIIa levels after a dietary fat load but did not find a concomitant postprandial burst of activation of factor X and prothrombin as measured by F1+2. Further studies are to test whether postprandial FVIIa generation leads to enhanced activation of coagulation.
Presented in part at the Annual Meeting of the American Society of Hematology, Nashville, Tenn, December 4, 1994, and published in abstract form in Blood 1994;84:191. Preliminary results of the effect of anticoagulants on F1+2 levels were presented in part at the XIVth Congress of the International Society on Thrombosis and Haemostasis, New York, NY, July 5, 1993, and published in abstract form in Thromb Haemost 1993;69:625.
- Received December 4, 1995.
- Revision received July 25, 1996.
The Northwick Park Heart Study identified elevated levels of FVII coagulant activity as a risk factor for ischemic heart disease and cardiovascular death.1 2 FVII is proteolytically cleaved by FIXa, FXa, FXIIa, or thrombin into FVIIa.3 FVIIa is a serine protease that activates FX in the presence of its cofactor, TF. FVIIa has a long plasma half-life (relative to other procoagulant serine proteases) of about 2.5 hours.4 In the normal state a trace level of FVIIa is found in plasma, perhaps to allow for priming of the clotting cascade.5 Elevated plasma FVIIa levels may, therefore, be a useful marker of a hypercoagulable state.
Studies that have measured FVIIa by using indirect methods6 7 8 have found elevated levels of postprandial FVIIa in subjects given standardized test meals. Recent reports have also demonstrated positive associations between FVII coagulant activity and F1+2,9 10 the polypeptide that is released from prothrombin when it is converted to thrombin by FXa.
We studied 30 healthy young subjects given a high-fat meal in order to measure possible postprandial FVII and prothrombin activation. We hypothesized that the basal FVII zymogen level, measured as FVIIag, may control the generation of FXa and subsequent prothrombin activation. We chose to measure prothrombin activation via the F1+2 assay and tested the correlation of fasting FVIIag with postprandial F1+2. Because TG levels correlate with FVIIag levels11 and with indirect measurements of FVIIa,6 11 we also tested the correlation of postprandial TG levels with postprandial FVIIag and FVIIa.
We hypothesized that postprandial elevation of FVIIa would also produce intermittent activation of FX to FXa, and, subsequently, prothrombin to thrombin. We chose to study postprandial activation of FVII by using a new assay specific for FVIIa5 12 that employs a mixture of phospholipid and a truncated STF apoprotein that retains cofactor function toward FVIIa but not FVII.13 We hypothesized that both FVIIa and F1+2 levels would increase after subjects ingested a high-fat meal, and we tested the correlation of postprandial FVIIa levels with postprandial F1+2 levels. We chose to measure FVIIa levels before and 6 hours after ingesting a high-fat meal because previous studies6 14 have suggested that FVII activation is detectable 2½ hours after the postprandial TG peak at 3 to 4 hours. We also measured HDL-C and TC levels in our subjects.
A sample size estimate of 30 subjects was determined by calculating a sample size for paired samples. We estimated a mean of 3.5 ng/mL FVIIa at baseline,5 and using an α of .05 and a power of 80%,15 we hypothesized a 15% mean increase in FVIIa postprandially. We also hypothesized that postprandial FVIIa levels would be significantly correlated with postprandial TG6 and postprandial F1+2 levels and that fasting FVIIag levels would be correlated with postprandial F1+2 levels. All samples were measured in duplicate.
The protocol for obtaining blood samples was approved by the Committee on Research Involving Human Subjects at the University at Stony Brook following the principles of the Declaration of Helsinki. Written informed consent was obtained from all study participants. Both men and women, ages 20 to 49 years, were recruited on a volunteer basis. Volunteers taking any medications, including aspirin and oral contraceptives, were not eligible to participate since our goal was to study normal, healthy subjects. Data were collected by questionnaire on each subject regarding age, gender, height, and risk factors for ischemic heart disease. Weight was measured and used to calculate surface area, which was used to determine the amount of fat to be given to each subject. Twenty-two men and eight women were recruited (age, 30.8±9.8 years). Three subjects were found to be at high risk for ischemic heart disease based on family history of premature myocardial infarction (defined as ≤60 years of age for women and ≤50 years of age for men) in a first-degree relative. The remaining 27 low-risk subjects were nonsmokers with no family history of premature ischemic heart disease.
Each subject was instructed to fast overnight for 12 hours and report to the study the following morning. Consent forms and data sheets were then completed, and a baseline blood sample was collected. The subjects were then fed a serving of high-fat premium ice cream (Häagen-Dazs Co) based on their surface area (30 g fat/m2; range, 42 to 72 g). The subjects were asked to return at 4 and 6 hours for two more blood samples. The 4-hour time was optional, but strongly encouraged, in order to accrue a subset for measuring peak postprandial serum TGs. Twenty-two of the 30 subjects returned for the 4-hour draw. No more than four subjects were tested per day due to the time limitations of the assays.
Citrated blood samples were collected by using a double-syringe technique. The first 5 mL were drawn into a sterile evacuated tube (Becton Dickinson Systems) and allowed to clot; serum was prepared by centrifugation at 3000g for 18 minutes for determination of serum TC, TG, and HDL-C levels. A second syringe (10-mL) was drawn, 5 mL of which was added immediately to a 1:10 volume of 3.8% sodium citrate in a polypropylene tube for FVII assays. The remaining 5 mL was added to an evacuated tube containing EDTA/PPACK thrombin inhibitor and aprotinin (SCAT-I tubes, Haematologic Technologies) for F1+2 analysis. Plasma was separated by centrifugation at 3000g for 18 minutes at room temperature within ½ hour of blood collection. All samples were kept at room temperature to avoid cold activation and were handled with polypropylene tubes, pipet tips, and Fibrometer cups (Becton Dickinson) to prevent contact activation. All samples for the FVIIa clotting assay and serum lipid levels were tested on fresh samples.
One 5-mL sample was drawn into a sterile evacuated tube for testing TG levels. Blood was collected from the same arm as for the fasting sample.
The same method was followed as in the fasting sample except that blood was drawn from the other arm in most cases. Aliquots of each plasma sample were made. One aliquot was used for the FVIIa assay, and the remaining aliquots were frozen at −80°C for future studies. FVIIag and F1+2 were assayed in all subjects from fasting and 6-hour samples after all samples were obtained (<1 month).
Determination of TC, HDL-C, and TGs
Serum TC and TGs were assayed on a Kodak Ektachem DT60 analyzer (Eastman Kodak Co) by using Ektachem reagents according to the manufacturer's instructions. Results were calculated in milligrams per deciliter. Samples for HDL-C determination were prepared by dextran sulfate/magnesium precipitation of the VLDL and LDL cholesterol fractions by using Ektachem reagents.
Determination of FVIIa
Plasma samples were diluted in 0.15 mol/L NaCl (Fisher)/0.05 mol/L Tris (Sigma)/0.1% (wt/vol) bovine serum albumin (fraction V, fatty acid free, Sigma) (TBS-BSA). CaCl2 stock was from BDH. FVII-deficient plasma was purchased from George King Biomedical. RBC (Sigma) was prepared fresh on the day of each assay in 0.85% NaCl according to the manufacturer's instructions. This concentration of RBC was 10-fold greater than the final concentration recommended by the manufacturer for that used in the partial thromboplastin time test. STF (STF1-218),12 a generous gift from Dr Yale Nemerson (Mt Sinai School of Medicine, New York, NY), was diluted to a concentration of 1 μmol/L in TBS-BSA and 30 mmol/L CaCl2, divided into aliquots, placed in 5-mL polypropylene tubes, and stored at −80°C. On the day of the assay, an equal volume of the 1 μmol/L STF preparation was added to the RBC preparation and kept at 37°C. This STF/RBC preparation was vortexed repeatedly to maintain a homogeneous suspension.
Plasma from a subject on oral contraceptive therapy was cold-activated in glass test tubes for 24 hours at 4°C, divided into 1-mL aliquots, and stored at −80°C to serve as the FVIIa standard. It was calibrated against a known quantity of recombinant human FVIIa (a gift of Novo Nordisk). One normal donor's plasma was assayed twice daily to serve as an internal control.
The FVIIa clotting assay, performed on a Fibrometer (BBL, Becton Dickinson), was similar to the assay developed by Morrissey et al.5 Diluted sample (100 μL) was added to 0.1 mL FVII-deficient plasma in a Fibrometer cup at 37°C; after 5 minutes, 0.1 mL of 37°C STF/RBC mixture was added, and the clotting time was determined. Diluted samples were assayed in duplicate. The standard curve was prepared by assaying cold-activated plasma at dilutions of 1:50, 1:100, 1:200, 1:400, and 1:1000. Control and volunteer subject samples were assayed at 1:2 and 1:4 dilutions and were kept at room temperature to avoid cold activation. Subject samples were assayed as soon as possible after each specimen was collected. One dilution of the standard was reassayed at the end of each batch of samples to verify the stability of the reagents. The interassay and intra-assay CVs were 13% and 15.5% based on six and 11 determinations, respectively, of a single normal donor's plasma.
FVII Enzyme Immunoassay
FVIIag was measured by using an FVII enzyme immunoassay kit (Diagnostica Stago).16 The anti-FVIIag–peroxidase conjugate was used at the concentration recommended by Diagnostica Stago. This assay was performed on samples thawed within 1 month of collection and freezing. The interassay CV was 4.2% based on seven determinations of a single normal donor's plasma.17 For this study, the intra-assay CV was 7% based on 11 determinations of a single normal donor's plasma.
Determination of F1+2
F1+2 was measured by using a microELISA method (Thrombonostika F1.2, Organon Teknika) according to the manufacturer's directions except that blood samples were placed in a mixture of EDTA/PPACK/aprotinin. The interassay CV ranged from 8.6% to 10.4%, and the intra-assay CV ranged from 3.3% to 10.9% (for both CVs, n=20 and 27, respectively); according to the manufacturer, a minimum of 0.1 nmol/L plasma F1+2 can be detected. Although the manufacturer recommends the use of heparinized plasma samples, Bauer et al18 report that asymptomatic patients with congenital antithrombin III deficiency have substantially higher plasma F1+2 levels in samples placed in a mixture of citric acid/sodium citrate/dextrose/EDTA/adenosine/heparin than in those placed in a mixture of EDTA/PPACK inhibitor/aprotinin. To confirm the superiority of this anticoagulant mixture on the F1+2 assay, we performed a preliminary study of 12 normal healthy volunteers (6 men and 6 women, aged 33.5±9.4 years) on no medication and 7 subjects (4 men and 3 women, aged 47.4±18.8 years) with possible hypercoagulable states (2 patients had type II diabetes and 1 each had premature peripheral vascular disease, primary pulmonary hypertension, pulmonary embolus, carotid artery occlusion, and type III hyperlipidemia). We compared all samples drawn into 4.0 U/mL sodium heparin (Sigma) versus those collected into an inhibitor mix of 10 mmol/L EDTA (Sigma), 0.1 μmol/L PPACK (Bachem), and 10 μg/mL aprotinin (Trasylol, FBA Pharmaceuticals) (final concentrations). All samples were obtained by clean venipuncture via a two-syringe technique.16 Plasma was prepared by centrifugation at 3000g for 18 minutes at room temperature within 1/2 hour of blood collection. A 450-μL plasma sample was added to a polypropylene tube; 50 μL sample treatment reagent (Organon Teknika) was then added, and the sample was mixed thoroughly and stored at −80°C until assay (<1 month). The mean values and 2-SD range for the 12 normal subjects were virtually identical on heparinized plasma (0.6±0.3 nmol/L) and inhibitor mix plasma (0.7±0.3 nmol/L). All samples had detectable F1+2. However, 2 normal subjects whose sample collection was flawed (one because of a difficult venipuncture and one because of a 1-hour lag in sample processing) had discordantly high results on the heparinized sample (10.9 and 7.1 nmol/L) compared with the inhibitor mix sample (1.6 and 0.48 nmol/L). In addition, 2 of the 7 patients with suspected hypercoagulable states had discordantly high results on heparinized plasma (8.4 and 1.67 nmol/L) compared with inhibitor mix plasma (4.4 and 0.33 nmol/L). Therefore, the inhibitor mix was chosen as the preferred anticoagulant for the present study.
To determine whether postprandial lipids affected the in vitro assays, 0.02 mg/mL phosphatidylserine/phosphatidylcholine vessicles (30%/70%; Sigma) were added to fasting plasma samples and assayed for FVII. The addition of phosphatidylserine/phosphatidylcholine did not significantly affect either the FVIIag or FVIIa assays. According to the manufacturer, adding up to 500 mg/dL of lipid does not affect the F1+2 assay.
Results of the FVIIa coagulation assay data were analyzed on an IBM personal computer by log transformation of the data in the Parlin program19 for parallel-line analysis of bioassays. All other analyses were performed by using CSS:Statistica software (Statsoft). Observed distributions of all clotting factor and lipid assays were analyzed for departure from normality by using χ2 and Kolmogorov-Smirnov d statistics. Correlations of FVIIa and F1+2 results with TG levels were analyzed by using Spearman's ρ statistic. We chose prospectively to test only four correlations: 4-hour TG with 6-hour FVIIa or FVIIag levels and 6-hour F1+2 with fasting FVIIag or 6-hour FVIIa levels. This strategy was chosen to perform only those correlations that would test our hypotheses and to avoid indiscriminate data probing. Student's two-tailed t test for dependent samples was used to determine significant differences in fasting versus 6-hour FVIIa levels and fasting versus 6-hour FVIIag levels. A value of P<.05 was considered significant. Differences in fasting versus 6-hour F1+2 levels and between fasting versus 4-hour TG levels were analyzed by using the Wilcoxon matched-pairs test. Values are mean±SD unless otherwise noted.
Effect of a High-Fat Meal on FVIIa and F1+2
There was a significant difference (P<.004) between fasting (2.82±1.49 ng/mL) and 6-hour postprandial (3.45±2.08 ng/mL) FVIIa levels, representing a 26% mean increase in postprandial FVIIa (Table 1⇓ and Fig 1⇓) in the 30 subjects. FVIIa levels were higher postprandially in 21 of the 30 subjects, unchanged in 5, and lower in 4 (Fig 1⇓). In contrast, FVIIag levels did not change (fasting and 6-hour postprandial, 0.89±0.20 and 0.90±0.20 U/mL, respectively) (Table 1⇓). Median TG levels increased significantly (P<.00005) 4 hours postprandially (241 mg/dL) compared with fasting (101 mg/dL) (Table 1⇓). Median F1+2 levels were slightly lower (P<.02) 6 hours postprandially (0.39 nmol/L) than fasting (0.44 nmol/L) (Table 1⇓ and Fig 2⇓). The lower median F1+2 level postprandially is not readily explained by the insensitivity of the assay below 0.1 nmol/L, since only 2 subjects (Nos. 1 and 17) had <0.1 nmol/L F1+2 in the postprandial sample only (Fig 2⇓). Of the 29 subjects, 6 showed no change in postprandial F1+2, 5 showed an increase in F1+2, and 18 showed a decrease (Fig 2⇓).
Correlation of FVIIa With Other Variables
Correlations of FVIIa, FVIIag, F1+2, and TG levels are summarized in Table 2⇓ and Figs 3⇓ and 4. Six-hour postprandial FVIIag levels correlated significantly with 4-hour postprandial TG levels (ρ=0.51, P<.02), but 6-hour postprandial FVIIa levels did not. Six-hour postprandial F1+2 levels correlated significantly with fasting FVIIag levels (ρ=0.39, P<.04) but not with 6-hour postprandial FVIIa levels. Although it was not one of the original goals of the study, we subsequently tested the correlation of fasting FVIIag with fasting F1+2 levels; we found no significant correlation (ρ=0.28, P<.14), nor was there any significant correlation between fasting FVIIa and fasting F1+2 (ρ=0.04, P=.89).
We found a significant 26% increase in postprandial compared with fasting FVIIa levels after a single high-fat meal. This mean increase reflected an increase in 21 of the 30 subjects and was not due to skewed results in a limited number of subjects. Only 4 of the 30 subjects had a lower FVIIa level postprandially. Our postprandial mean value of FVIIa (3.45±2.08 ng/mL) agrees well with the mean values of 3.73±1.44 ng/mL (n=87) of Morrissey et al5 and 4.34±1.57 ng/mL (n=20) of Wildgoose et al,12 whose studies used clotting assays similar to ours on plasma from nonfasting healthy subjects. In addition, Kario et al,20 using a fluorogenic assay, found a mean FVIIa value of 2.5 ng/mL in samples from 110 fasting older Japanese subjects, which was similar to our mean value of 2.82 ng/mL on plasma from fasting subjects. The assays of all the above-mentioned investigators used a mutant STF that retained cofactor function toward FVIIa but not FVII. Thus, all assays directly measured plasma FVIIa. Using a mutant STF assay for FVIIa, Miller et al21 report that FVIIa rose postprandially after a high-fat meal compared with baseline in subjects with FXI or FXII deficiency and normal control subjects.
We found no correlation between postprandial FVIIa and postprandial TG levels. Although earlier reports found a significant correlation between postprandial FVII activity and postprandial TG levels,6 14 these studies used indirect measurements of FVIIa that may reflect levels of both zymogen FVII and FVIIa.22 Investigators20 23 who have used assays specific for FVIIa report no correlations between fasting FVIIa and fasting TG levels. The preponderance of evidence based on the specific FVIIa assay therefore supports the conclusion that lipoprotein metabolism, as reflected by TG levels, is not important in controlling FVIIa generation and/or clearance. The rise in postprandial FVIIa may, however, be influenced by circadian fluctuation14 or genetic polymorphism.23 24
We have shown25 26 that fasting FVIIag levels are significantly correlated with fasting TG levels, in agreement with others.6 11 20 23 27 28 We now report that postprandial FVIIag levels are highly correlated with postprandial TG levels. As one might expect from these data, there is also evidence of a positive correlation between pre- and postprandial TG levels; compared with age-matched control subjects, subjects with hypertriglyceridemia have elevated pre- and postprandial TG levels after ingesting a single high-fat meal.29
We found a slight but statistically significant postprandial decrease in F1+2 levels in our subjects given a high-fat meal, whereas FVIIa levels significantly increased. Because the plasma half-life of F1+2 is ≈90 minutes,30 it is unlikely that we would find an increase in postprandial FVIIa but miss detecting a concomitant increase in F1+2 levels. Therefore, our hypothesis that F1+2 would increase postprandially in conjunction with increased FVIIa can be rejected with confidence.
Our unexpected finding of a slight decrease in postprandial F1+2 levels is intriguing; whether it reflects a significant biological change is uncertain. Eichinger et al31 found no significant increase in F1+2 after hemophilic subjects were infused with either purified FVIII or FIX concentrates, whereas infusion of recombinant FVIIa in FVII-deficient patients led to an increase in F1+2 levels that was within the normal range for this assay. The sudden infusion of FVIIa into the circulation may not precisely mimic the more gradual postprandial in vivo generation of FVIIa. Additionally, Eichinger et al31 found no significant increase in F1+2 after infusion of FIX concentrate into four patients with hemophilia B, despite the fact that initially low levels of FVIIa normalized after the infusion. We speculate that control mechanisms, such as an increase in TF pathway inhibitor, may be elicited to dampen activation of prothrombin in the postprandial state after the in vivo generation of FVIIa and may therefore actually cause a slight decrease in F1+2.
If the postprandial F1+2 level had increased, we had hypothesized that it would correlate significantly with the postprandial FVIIa levels; however, we found no significant correlation, a finding adding further support to the lack of a biological relation between the postprandial FVIIa increase and F1+2 levels. The postprandial F1+2 level was significantly correlated with the fasting FVIIag level; since the latter is an indication of a chronic elevation in plasma FVII mass, this significant positive correlation suggests the possibility that the basal level of plasma FVII mass influences the postprandial coagulation response, even though no overall rise in F1+2 was detected. On the other hand, this positive correlation may not be biologically meaningful, since the statistical significance was dependent on a single patient with very high levels of F1+2 and FVIIag (Fig 4⇓). Further studies are needed to clarify the postprandial response of F1+2. On the basis of our data, it is premature to accept the intuitively attractive hypothesis that the postprandial increase in plasma FVIIa will lead to enhanced activation of coagulation. Further studies are needed to clarify the possible physiological role of increased postprandial FVIIa levels.
Selected Abbreviations and Acronyms
|CV||=||coefficient of variation|
|F1+2||=||prothrombin fragment 1+2|
|FVII, FVIII, FIX, FX, FXI, or FXII||=||factor VII, VIII, IX, X, XI, or XII|
|FVIIa, FIXa, FXa, or XIIa||=||activated factor VII, IX, X, or XII|
|FVIIag||=||factor VII antigen|
|PPACK||=||d-phenyl-alanyl-prolyl-arginine chloralmethyl ketone|
|RBC||=||rabbit brain cephalin|
|STF||=||soluble tissue factor|
This study was supported by Vascular Diseases Academic Award HL-02821 from the National Institutes of Health. We thank Jolyon Jesty, DPhil, for the Parlin program, Arlene Neuroth, RN, for help with blood sample collection, and Helen Giles for outstanding secretarial assistance.
Hagen FS, Gray CL, O'Hara P, Grant FJ, Saari GC, Woodbury RG, Hart CE, Insley M, Kisiel W, Kurachi K, Davie EW. Characterization of a cDNA coding for human factor VII. Proc Natl Acad Sci U S A. 1986;83:2412-2416.
Seligsohn U, Kasper CK, Østerud B, Rapaport SI. Activated factor VII: presence in factor IX concentrates and persistence in the circulation after infusion. Blood. 1979;53:828-837.
Morrissey JH, Macik BG, Neuenschwander PF, Comp PC. Quantitation of activated factor VII levels in plasma using a tissue factor mutant selectively deficient in promoting factor VII activation. Blood. 1993;81:734-744.
Silveira A, Karpe F, Blombäck M, Steiner G, Walldius G, Hamsten A. Activation of coagulation factor VII during alimentary lipemia. Arterioscler Thromb. 1994;14:60-69.
Wildgoose P, Nemerson Y, Hansen LL, Nielsen FE, Glazer S, Hedner U. Measurement of basal levels of factor VIIa in hemophilia A and B patients. Blood. 1992;80:25-28.
Neuenschwander PF, Morrissey JH. Deletion of the membrane anchoring region of tissue factor abolishes autoactivation of factor VII but not cofactor function: analysis of a mutant with a selective deficiency in activity. J Biol Chem. 1992;267:14477-14482.
Snedecor GW, Cochran WG. Statistical Methods. 6th ed. Ames, Iowa: Iowa State University Press; 1967:111-114.
Kario K, Miyata T, Sakata T, Matsuo T, Kato H. Fluorogenic assay of activated factor VII: plasma factor VIIa levels in relation to arterial cardiovascular diseases in Japanese. Arterioscler Thromb. 1994;14:265-274.
Miller GJ, Martin JC, Mitropoulos KA, Esnouf MP, Cooper JA, Morrissey JH, Howarth DJ, Tuddenham EGD. Activation of factor VII during alimentary lipemia occurs in healthy adults and patients with congenital factor XII or factor XI deficiency, but not in patients with factor IX deficiency. Blood. 1996;87:4187-4196.
Morrissey JH. Tissue factor interactions with factor VII: measurement and clinical significance of factor VIIa in plasma. Blood Coagul Fibrinolysis. 1995;6:S14-S19.
Moor E, Silveira A, van't Hooft F, Suontaka AM, Eriksson P, Blombäck M, Hamsten A. Coagulation factor VII mass and activity in young men with myocardial infarction at a young age: role of plasma lipoproteins and factor VII genotype. Arterioscler Thromb Vasc Biol. 1995;15:655-664.
Bernardi F, Marchetti G, Pinotti M, Arcieri P, Baroncini C, Papacchini M, Zepponi E, Ursicino N, Chiarotti F, Mariani G. Factor VII gene polymorphisms contribute about one third of the factor VII level variation in plasma. Arterioscler Thromb Vasc Biol. 1996;16:72-76.
Hoffman CJ, Miller RH, Hultin MB. Correlation of factor VII activity and antigen with cholesterol and triglycerides in healthy young adults. Arterioscler Thromb. 1992;12:267-270.
Hoffman CJ, Lawson WE, Miller RH, Hultin MB. Correlation of vitamin K–dependent clotting factors with cholesterol and triglycerides in healthy young adults. Arterioscler Thromb. 1994;14:1737-1740.
Bauer KA, Goodman TL, Kass BL, Rosenberg RD. Elevated factor Xa activity in the blood of asymptomatic patients with congenital antithrombin deficiency. J Clin Invest. 1985;76:826-836.
Eichinger S, Mannucci PM, Tradati F, Arbini AA, Rosenberg RD, Bauer KA. Determinants of plasma factor VIIa levels in humans. Blood. 1995;86:3021-3025.