Estimates of Within-Subject Biological Variation of Protein C, Antithrombin, Protein S Free, Protein S Activity, and Activated Protein C Resistance in Pregnant Women.
Measurement of thrombophilia markers may be indicated as part of risk assessment in pregnancy, but may also be monitored (e.g., antithrombin) as part of a diagnostic work-up of pregnancy complications with suspected development of disseminated intravascular coagulation (DIC) (4, 5). Correct interpretation of the results is important for evaluating the severity of disease and further management of the patient. Reference intervals for coagulation parameters in nonpregnancy cannot be used in pregnancy because of the physiological changes in pregnancy (1, 6-9). Knowledge of the course of the coagulation parameters in pregnancy is therefore essential for interpretation. There are some studies describing the course of coagulation parameters during pregnancy and postpartum (1, 6-9). However, the results have been inconsistent, especially concerning the results during the last trimester of pregnancy (7, 10). In addition, no studies have estimated the within-subject biological variation ([CV.sub.I]) in pregnant women, although nonpregnant [CV.sub.I] exists for some coagulation parameters (11-18).
The aim of this study was to describe the course of five different coagulation parameters in the same women during pregnancy, shortly after delivery, and in the postpartum period, to estimate the [CV.sub.I] of these parameters in pregnancy by using a tool for transformation into a steady-state situation, and to compare the results with similar results from healthy, nonpregnant women.
Materials and Methods
Twenty healthy pregnant women were included early in pregnancy. The median age for the pregnant women was 31 years (range 22-38 years), median duration of gestation at inclusion was 8 weeks (range 5-13 weeks), and median gestational week at delivery was 41 (range 37-42). The pregnancies were largely uncomplicated, but we included 3 women with Caesarean deliveries (subjects 2, 8, and 16), and 1 woman (subject 4) with more than 500 mL blood loss after delivery. In addition, one woman was heterozygous (subject 1) and one woman was homozygous (subject 5) for factor V Leiden. The median age for the 19 nonpregnant women was 33.5 years (range 23-40 years). No participants had a personal or family history of bleeding or thromboembolic disease, and all had been healthy during earlier pregnancies, except for 1 nonpregnant woman who had preeclampsia in 1 pregnancy. At each blood sampling, information of health status and medication was collected. Especially, symptoms or signs of VTE were not experienced, and diagnostic imaging for VTE was not undertaken. Study participants were recruited by an advertisement in the local newspaper and on the Web page of Haukeland University Hospital. The study was approved by the Regional Committee of Medical Ethics of Western Norway, and the women gave informed consent to participate in the study in a form accepted by the Norwegian Social Science Data Services.
Blood samples were drawn by venipuncture between 8 AM and 1:30 PM by experienced phlebotomists, with study participants in the sitting position. The samples were collected in 0.129 mmol/L (3.8%) sodium citrate tubes and centrifuged at 2500g for 15 min. Plasma was transferred to microtubes (1 mL), frozen within 2 h of blood sampling and stored at minus 80 [degrees]C until analysis. Blood samples were collected every fourth week in pregnancy and at 1-3 days and 2 and 6 weeks after delivery. The gestational age was calculated retrospectively according to the date of term determined by routine ultrasound in the 18th week of pregnancy. The time of sampling was allocated into nine 4-week time windows during pregnancy at gestational weeks 5-9, 9-13, etc. In nonpregnant women, blood samples were collected every 4th week with 10 samples from each woman, apart from 3 women who became pregnant after 5, 5, and 7 blood samples, respectively, had been collected.
Frozen plasma from each participant was rapidly thawed in a 37 [degrees]C water bath for 5 min, thoroughly mixed, and analyzed in duplicate in the same analytical run. The samples were analyzed on a STA-R instrument (Stago), with STA-Stachrom Protein C, STA-Stachrom Antithrombin, STA Liatest Free Protein S, STA-Staclot Protein S from Stago, and Coatest APC Resistance VS (with added factor V--deficient plasma) from Chromogenix IL. The APCR result is based upon the ratio of an activated partial thromboplastin time reagent (APTT1) with added activated protein C and one APTT reagent without added activated protein C (APTT2). Internal quality control materials [STA Liatest Control-N and Control-P (Stago), Scandinorm and Scandipath (Stago), and APC1 and APC2 (Chromogenix)] were supplied by the manufacturers. More than one lot of both controls and reagents were used for some of the coagulation parameters, but no systematic changes in the concentrations of the controls during the period of analysis were detected. Total analytical CVs (within- and between-run variation) calculated for the internal controls and within-run analytical variation ([CV.sub.A]) calculated for duplicate participants' samples were [less than or equal to] 10% and <4%, respectively (see Tables 1 and 2 in the Data Supplement that accompanies the online version of this article at http://www. clinchem.org/content/vol63/issue4).
TRANSFORMATION OF ORIGINAL RESULTS INTO MoM AND lnMoM
To be able to create a steady-state situation of the coagulation parameters, independent of the systematic changes in some of the parameters during pregnancy and postpartum, the results were transformed into multiples of the median (MoM), and then the natural logarithm of MoM (lnMoM) was calculated (19). Thereby, an unsteady-state situation is transformed into a steady-state condition. Transformation into MoM is done by calculating the median for each parameter for each sampling time (Table 1) and then dividing each individual pregnant woman's result for the corresponding sampling time with the calculated median for this time (19). Finally, ln-transformation (lnMoM) was used to make the distribution of the pooled residuals gaussian (20). The Shapiro-Wilk test was used to test for normality of the pooled residuals. The measured coagulation test results and MoM values are performed on a ratio scale, and CV can be calculated even if the distributions are skewed, whereas only SD can be calculated for lnMoM on an interval scale. These logarithmic SDs are excellent estimates of CV for MoM results and can also be calculated by the formula [CV.sub.mom] = [square root]([e.sup.var(lnMoM)]-1) (21), where var(lnMoM) is the variance of lnMoM ([SD.sup.2]). The differences between normally distributed variables are normally distributed, while the differences in percentages are not. Thus, the reference change value (RCV) calculation will not be correct when using CVs (20). By using the ln-transformation, it is possible to work with SD, and the correct RCV for percent can then be estimated when the RCVs are transformed back (21).
For nonpregnant women, the activities were transformed to MoM and lnMoM, as for the pregnant women, for the purpose of comparing the transformed values, but transformation should in principle not be necessary given the steady-state situation in nonpregnant women.
To examine for outliers in the duplicates in the samples for each woman and in the mean of each subject, the method of Burnett was used (22). In the pregnant women, subject 17 for PC, subject 4 for AT, subjects 1 and 14 for PS free, subject 13 and samples 1 and 2 for subject 14 for PS activity, and subjects 1 and 5 for APCR were excluded based upon outlying subjects or samples. In addition, sample 11 for subject 11 was excluded both for PS free and PS activity because of a result out of measurement range (>150%). In the nonpregnant women, only subject 33 was excluded for PS free analysis. There was no known reason for the outlying results in these women, except for subject 14, in whom pregnancy was achieved by in vitro fertilization (hormonal treatment given in the beginning of the pregnancy), probably contributing to the low PS results in the first 2 samples; and subjects 1 and 5, who were heterozygous and homozygous for factor V Leiden. After the exclusions, results from 18-20 pregnant women and from 18-19 nonpregnant women were included in the study, depending upon the parameter examined.
TEST OF VARIANCE HOMOGENEITY
A presupposition for the practical use of the pooled [CV.sub.I] to be representative for all individuals is that there is variance homogeneity among the individual [CV.sub.I]s. The individual variation for each person--i.e., the within-person SD ([SD.sub.P]) and within-person CV ([CV.sub.P]) (which includes analytical variation)--was estimated for each woman for the absolute results, as well as for the transformed results of MoM and lnMoM to test for variance homogeneity. Testing for homogeneity of these [SD.sub.P] and [CV.sub.P] values was performed by the Bartlett test (23). Variance homogeneity means that all individual [SD.sub.P] are distributed as if they were all random samples taken from the same gaussian-pooled [SD.sub.P] distribution, which implies that they must follow a curve around pooled [SD.sub.P] described by the formula pooled [SD.sub.P] x [square root] [chi-square]/df, where df is the degrees of freedom. This is illustrated on a rankit scale (24) (same as probability scale (23) or normal scale (25)), where the cumulated ranked individual SDP were plotted as fractions on the right ordinate as function of calculated [SD.sub.P] (or [CV.sub.P]). This was compared with the theoretical [chi square]-function for homogenous distribution of SD-values: pooled [SD.sub.P] [square root] [chi square]/df shown on the same rankit scale using the mean df for series of results from each woman (19, 23, 26). The closer the cumulated fractions are to the theoretical curve, the better the homogeneity of variances. Owing to the high number of tests for variance homogeneity, the P value for the Bartlett test was set to 0.01.
BIOLOGICAL AND ANALYTICAL VARIATION
The biological variations, separated into the within-subject ([CV.sub.I]) and between-subject ([CV.sub.G]) biological variations, and within-series analytical variation ([CV.sub.A]), were estimated from analysis of variance with the statistical model for repeated subsampling (nested design) (26). To compare the different steps in transformation for pregnant women, CVs were calculated for the non-transformed absolute results, their MoM values; and finally, the SDs [[SD.sub.I], [SD.sub.G], and [SD.sub.A], with [SD.sub.G] being "between-subject SD (lnMoM)," SDI, "within-subject SD (lnMoM)," [SD.sub.A], "within-series analytical SD (lnMoM)"] were calculated for their lnMoM-transformed values (21). Only the lnMoM had homogenous [SD.sub.P]-values and gaussian-distributed residuals for most of the parameters. The SDs ([SD.sub.I], [SD.sub.G], and SDa) calculated from the lnMoM results were transformed back into CVs (as described previously in this article) (21). For nonpregnant women, biological and analytical variations were calculated as for the pregnant women for comparison. 95% CIs were calculated according to Burdick and Graybill (27, 28).
REFERENCE CHANGE VALUE
For the within-subject variation ([CV.sub.I]) and RCV to be useful and representative for the studied population, variance homogeneity has to be present. Gaussian distribution of the residuals for the lnMoM-transformed data is necessary for a reliable RCV (20, 26, 29). RCV for SD (lnMoM) values is calculated as RCV = z X [square root of 2] X [square root] ([SD.sub.A.sup.2] + [SD.sub.I.sup.2]), or for the comparable CVs, RCV = z X [square root or 2] X [square root]([CV.sub.A.sup.2] + [CV.sub.I.sup.2]).
A unidirectional test with 95% probability (z = 1.64) was chosen because only a decrease of the coagulation parameters was considered. A difference between two lnMoM values has to be judged according to the RCV on a logarithmic level. "RCV" refers to the probability that a change within the RCV limits can be explained by analytical and within-subject variation with a certain probability (i.e., that the patient is in a stable situation), and not the counterhypothesis, the probability that a true change has occurred (30).
THE COURSE OF COAGULATION PARAMETERS DURING PREGNANCY
The course of the different coagulation parameters for each of the pregnant women as a function of time are shown in Fig. 1A-E (lines in color), while the results for nonpregnant women are shown as black lines with shadows. The medians for each sampling time are shown in Table 1. An increasing spread (increasing variation) of PC results were seen throughout pregnancy compared to nonpregnant women (Fig. 1A). AT levels were similar or slightly lower during pregnancy compared to nonpregnancy, and decreased slightly right after delivery (Fig. 1B). Results for PS free and PS activity were lower than in nonpregnancy, already in the first sample (weeks 5-9), and decreased further throughout pregnancy, with PS free decreasing less than PS activity (Fig. 1, C and D and Table 1). At 6 weeks postpartum, 30% and 40% of the results from pregnant women were still not within the reference interval for nonpregnant women for PS free and PS activity, respectively (Fig. 1, C and D). Both APTT tests (APTT1 and APTT2), from which the APCR ratio is calculated, decreased steadily during pregnancy (see online Supplemental Fig. 1, A and B), but the ratios between them (APCR) were nearly constant in the whole period, and also for the 2 women who were factor V Leiden heterozygous and homozygous (subjects 1 and 5, respectively) (Fig. 1E). PC, AT, PS free, and PS activity levels tended to reach a maximum (increase) at week 2 postpartum, and then decreased again 6 weeks postpartum (Fig. 1A-D).
TRANSFORMATION INTO MoM AND lnMoM
When the results of the coagulation parameters in pregnancy were transformed to MoM, and thereafter to lnMoM, the values were normalized around 1 and 0 (shown for PS free and PS activity in Fig. 2, A and B (MoM) and Fig. 2, C and D (lnMoM), respectively). After exclusion of outliers, the lnMoM values show gaussian-distributed residuals except for PS free (see online Supplemental Table 3). The results of the coagulation parameters in nonpregnant women were "stable" throughout the study period (Fig. 1, black lines and shaded area), but again, gaussian-distributed residuals were not found for PS free (data not shown).
TEST FOR VARIANCE HOMOGENEITY
Variance homogeneity was found for lnMoM results in pregnant and nonpregnant women for all coagulation parameters after excluding outliers, except for PS free [Table 2, Fig. 3A-E (pregnant), and online Supplemental Fig. 2A-E (nonpregnant)].
WITHIN-SUBJECT BIOLOGICAL VARIATION
The original data (Fig. 1A-D) had to be transformed into MoM values to obtain a steady-state situation in pregnancy, and then transformed to lnMoM to be able to calculate RCVs (see online Supplemental Table 3 and Supplemental Fig. 3). In pregnancy (9 samples), the transformation from absolute results to MoM and lnMoM reduced the within-subject variation ([CV.sub.I]) for PS free and PS activity, whereas the between-subject variation ([CV.sub.G]) was not changed (see online Supplemental Table 4). In the nonpregnant women, transformation into MoM and lnMoM had no effect on the estimates of within- and between-subject variation (see online Supplemental Table 4).
When comparing pregnant and nonpregnant women, only the [CV.sub.I] (lnMoM) for PS free and PS activity were significantly higher in pregnant women, and there was no difference in the [CV.sub.G] (lnMoM) for any of the parameters (Table 2). Comparing [CV.sub.I] (lnMoM) and [CV.sub.G] (lnMoM) calculated only in pregnancy (9 samples) to results calculated in pregnancy and postpartum (12 samples), neither [CV.sub.I] nor [CV.sub.G] were different (see online Supplemental Table 4). Except for [CV.sub.G] for APCR, [CV.sub.I] and [CV.sub.G] were not different if the excluded subjects/samples were included (see online Supplemental Table 5), but variance homogeneity was not present if they were included.
REFERENCE CHANGE VALUE
The RCVs for lnMoM were not significantly different in pregnant and nonpregnant women, except for PS free, whereas the RCV in the pregnant women were slightly higher than in the nonpregnant women (Table 2).
The main new finding in the present study was that the within-subject variation of PC, AT, PS free and PS activity, and APCR in healthy pregnant women, after being "corrected" for the physiological changes in these parameters during pregnancy, was comparable to that found in nonpregnant healthy women. Thus, the within-subject variation is comparable in pregnant and non-pregnant women, but the homeostatic set point varies through pregnancy.
COURSE OF COAGULATION PARAMETERS DURING PREGNANCY
There are discrepant findings in the literature concerning the course of PC and AT activity during pregnancy. Some studies conclude that PC and AT are unaffected (1, 7, 31), while other studies indicate an increase in PC (32-34) or a decrease in AT (8, 10). The larger spread in PC during pregnancy has not been shown before. The reason for the PC increase in some of the pregnant women is not clear, but it has been speculated that this is a physiological reaction to counterbalance the decrease in PS (32). The decrease in AT shown 1-3 days after delivery in the present study confirms the finding of a recent study (10). The reason for an AT decrease right after delivery may be an increased consumption of AT in conjunction with delivery, in which the hypercoagulable state of pregnancy reaches its peak (7, 19).
PS activity, and to a lesser extent PS free, decrease throughout pregnancy (Table 1 and Fig. 1, C and D), which is in line with earlier studies of these parameters in pregnancy (1, 7, 31). The very early reduction in PS levels, already at 5-9 weeks (Fig. 1, C and D), confirms the findings in another study (32). The decrease of PS in pregnancy and other situations with increased estrogen concentrations (35) is probably caused by downregulation of the protein S (alpha) (PROS1) gene expression by 17[beta]-estradiol via the estrogen receptor [alpha] (6). However, it was surprising that PS free and PS activity were normalized in only two-thirds of the women during the 6 weeks postpartum (Fig. 1, C and D). The decrease of APTT1 and APTT2, used as part of the APCR test, has not been described before. The reduction is most probably caused by the increase in coagulation factors like fibrinogen, factor VIII, and factor IX (1, 7). Since APTT1 and APTT2 decrease to a similar degree, resulting in largely unchanged APCR ratios during pregnancy, the results confirm that APCR with added factor V-depleted plasma may be used in pregnancy to detect factor V Leiden mutation (9, 36).
The slight increase of PC, AT, PS free, and PS activity at 2 weeks postpartum and then decrease at 6 weeks postpartum (Fig. 1A-D) have been shown in a few studies, but not consistently for all parameters (9, 10, 37) and not emphasized. An increased production ("overshooting") may occur as a response to the increased coagulation activity and consumption of these inhibitors right after delivery.
TRANSFORMATION INTO MoM AND lnMoM
Since PS results decreased gradually during pregnancy (Table 1 and Fig. 1, C and D), normalization was necessary to obtain a "steady state" of the results, making it possible to calculate the within-subject variation ([CV.sub.I]). Since changes also were seen both for AT and PC, although not as systematically as for PS, all components were transformed into MoM and lnMoM for both normalization and conformance with a gaussian distribution. A similar transformation to lnMoM was also performed for D-dimer in pregnancy in a previous study (19).
TEST FOR VARIANCE HOMOGENEITY AND CALCULATION OF [CV.sub.I]
The validity of transformation to lnMoM during pregnancy is the establishment of a kind of "steady state," with gaussian distributions and with variance homogeneity of the parameters. This may improve the interpretation of a complicated system of changes in the ordinary course during the pregnancy. Variance homogeneity and gaussian distribution was achieved for all parameters except for PS free (Fig. 3 and online Supplemental Fig. 2). [CV.sub.I] and RCV for PS free should therefore be interpreted with caution both for pregnant and nonpregnant women since it will not be representative for all women. Generally, the [CV.sub.I] was lower for nonpregnant than for pregnant women, but this was only statistically significant for PS free and PS activity.
RCV AND INTERPRETATION OF CHANGES IN PREGNANCY
In general, the five coagulation parameters studied are mostly used for diagnostic purposes (i.e., hereditary thrombophilia). Testing before pregnancy is preferred to avoid false diagnoses of hereditary thrombophilia caused by the physiological changes. Although gestation-specific reference intervals are suggested (7), it is uncertain if they can be reliably used for diagnostic purposes in pregnancy.
In a monitoring situation, the difference between (two) test results is of importance. The RCV defines the lower and upper limits of expected results for the difference between two consecutive measurements that can be explained by analytical and biological variation (bidirectional test). Thus, if the second result is outside these limits, there is a higher probability that it is caused by a pathological event (disease). Comparing the size of a decrease in a coagulation parameter with the RCV for that parameter could be helpful when suspecting consumption of coagulation factors in pregnancy (e.g., DIC). As an example, RCV SD (unidirectional) can be used to evaluate if a decrease in AT in a pregnant woman with a suspected complication is "as expected," or if it is larger than expected. First, the two AT results that are compared are transformed into lnMoM (e.g., AT decrease of 103% to 85% from week 36 to week 38 is a change in lnMoM AT from -0.01 to -0.15). The decrease in lnMoM AT is then compared to the calculated lnMoM RCV SD (unidirectional), shown to be -0.09 (Table 2). The decrease in the example (-0.15-0.01 = -0.14) is numerically larger than expected (see online Supplemental Table 6). Whether the change is caused by a pathological process (e.g., DIC caused by preeclampsia) will have to be judged by the clinician in each case.
The transformation of concentrations to lnMoM is a tool for changing an unsteady-state situation during pregnancy to a steady-state position, which allows estimation of [CV.sub.I]. After this transformation, the [CV.sub.I] of coagulation parameters were comparable in pregnant and nonpregnant women, although the pregnant women had a set value that varied throughout pregnancy. Information about the within-subject variation in pregnant women can be a supplemental tool in the monitoring of women with suspected pregnancy complications if the physiological changes of coagulation parameters make single measurements difficult to interpret. This method, however, needs to be further validated.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: Reagents free of cost from Stago; reagents at reduced cost from Alfa-lab. A.H. Kristoffersen, PhD and post-doctoral scholarship from Western Regional Health Authorities.
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, and final approval of manuscript.
Acknowledgments: We thank Bente Asbjornsen and Solveig Vannes for excellent technical assistance. Thanks also to Stago for supporting the study with reagents free of cost and Alfa-laboratory for APCR reagents at reduced costs. Additional thanks to the Western Regional Health Authorities for supporting A.H. Kristoffersen with PhD and postdoctoral scholarships.
(1.) Brenner B. Haemostatic changes in pregnancy. Thromb Res 2004;114:409-14.
(2.) Jacobsen AF, Skjeldestad FE, Sandset PM. Incidence and risk patterns of venous thromboembolism in pregnancy and puerperium-a register-based case-control study. Am J Obstet Gynecol 2008;198:233.e1-233.e7.
(3.) De Stefano V, Rossi E. Testing for inherited thrombophilia and consequences for antithrombotic prophylaxis in patients with venous thromboembolism and their relatives. A review of the guidelines from scientific societies and working groups. Thromb Haemost 2013; 110:697-705.
(4.) Li WJ, Sha M, Ma W, Zhang ZP, Wu YJ, Shi DM. Efficacy evaluation of D-dimer and modified criteria in overt and nonovert disseminated intravascular coagulation diagnosis. Int J Lab Hematol 2016;38:151-9.
(5.) Wada H, Matsumoto T, Yamashita Y, Hatada T. Disseminated intravascular coagulation: testing and diagnosis. Clin Chim Acta 2014;436:130-4.
(6.) Suzuki A, Sanda N, Miyawaki Y, Fujimori Y, Yamada T, Takagi A, et al. Down-regulation of PROS1 gene expression by 17beta-estradiol via estrogen receptor alpha (ERalpha)-Sp1 interaction recruiting receptor-interacting protein 140 and the corepressor-HDAC3 complex. J Biol Chem 2010;285:13444-53.
(7.) Szecsi PB, Jorgensen M, Klajnbard A, Andersen MR, Colov NP, Stender S. Haemostatic reference intervals in pregnancy. Thromb Haemost 2010;103:718-27.
(8.) Wickstrom K, Edelstam G, Lowbeer CH, Hansson LO, Siegbahn A. Reference intervals for plasma levels of fibronectin, von Willebrand factor, free protein S and antithrombin during third-trimester pregnancy. Scand J Clin Lab Invest 2004;64:31-40.
(9.) Kjellberg U, Andersson NE, Rosen S, Tengborn L, Hellgren M.APC resistance and other haemostatic variables during pregnancy and puerperium. Thromb Haemost 1999;81:527-31.
(10.) James AH, Rhee E, Thames B, Philipp CS. Characterization of antithrombin levels in pregnancy. Thromb Res 2014;134:648-51.
(11.) Costongs GM, Bas BM, Janson PC, Hermans J, Brombacher PJ, van Wersch JW. Short-term and long-term intra-individual variations and critical differences of coagulation parameters. J Clin Chem Clin Biochem 1985; 23:405-10.
(12.) Wada Y, Kurihara M, Toyofuku M, Kawamura M, Iida H, Kayamori Y, et al. Analytical goals for coagulation tests based on biological variation. Clin Chem Lab Med 2004;42:79-83.
(13.) Thompson SG, Martin JC, Meade TW. Sources of variability in coagulation factor assays. Thromb Haemost 1987;58:1073-7.
(14.) Nguyen ND, Ghaddar H, Stinson V, Chambless LE, Wu KK. ARIC hemostasis study-IV. Intraindividual variability and reliability of hemostatic factors. The Atherosclerosis Risk in Communities (ARIC). Thromb Haemost 1995;73:256-60.
(15.) Melzi d'Eril G, Anesi A, Rizzo V, Trotti R. Biological variation in protein C, protein S and antithrombin concentrations in plasma of healthy subjects. Eur J Clin Chem Clin Biochem 1997;35:257-60.
(16.) Horne MK 3rd, McCloskey DJ, Cullinane AM, Merryman PK, Rick ME, Hortin GL, et al. Parameters of coagulant and fibrinolytic capacity and activity in postmenopausal women: within-subject variability. Thromb Res 2002;107:229-33.
(17.) Chambless LE, McMahon R, Wu K, Folsom A, Finch A, Shen YL. Short-term intraindividual variability in hemostasis factors. The ARIC study. Atherosclerosis Risk in Communities Intraindividual Variability Study. Ann Epidemiol 1992;2:723-33.
(18.) Banfi G, Del Fabbro M. Biological variation in tests of hemostasis. Semin Thromb Hemost 2009;35:119-26.
(19.) Kristoffersen AH, Petersen PH, Sandberg S. A model for calculating the within-subject biological variation and likelihood ratios for analytes with a time-dependent change in concentrations; exemplified with the use of D-dimer in suspected venous thromboembolism in healthy pregnant women. Ann Clin Biochem 2012;49: 561-9.
(20.) Roraas T, Stove B, Petersen PH, Sandberg S. Biological variation: the effect of different distributions on estimated within-person variation and reference change values. Clin Chem 2016;62:725-36.
(21.) Fokkema MR, Herrmann Z, Muskiet FA, Moecks J. Reference change values for brain natriuretic peptides revisited. Clin Chem 2006;52:1602-3.
(22.) Burnett RW. Accurate estimation of standard deviations for quantitative methods used in clinical chemistry. Clin Chem 1975;21:1935-8.
(23.) Bliss CI. Statistics in biology. Statistical methods for research in the natural sciences. New York: McGraw-Hill Book Company; 1967. Vol. I.
(24.) Hyltoft Petersen P, Blaabjerg O, Andersen M, Jorgensen LG, Schousboe K, Jensen E. Graphical interpretation of confidence curves in rankit plots. Clin Chem Lab Med 2004;42:715-24.
(25.) Altman DG. Practical statistics for medical research. London: Chapman & Hall; 1994.
(26.) Fraser CG, Harris EK. Generation and application of data on biological variation in clinical chemistry. Crit Rev Clin Lab Sci 1989;27:409-37.
(27.) Burdick RK, Graybill FA. Confidence intervals on variance components. New York: Marcel Dekker; 1992. Chapter 5.4, The unbalanced two-fold nested random model; p 96-108.
(28.) Roraas T, Petersen PH, Sandberg S. Confidence intervals and power calculations for within-person biological variation: effect of analytical imprecision, number of replicates, number of samples, and number of individuals. Clin Chem 2012;58:1306-13.
(29.) Harris EK, Yasaka T. On the calculation of a "reference change" for comparing two consecutive measurements. Clin Chem 1983;29:25-30.
(30.) Petersen PH, Sandberg S, Iglesias N, Soletormos G, Aarsand AK, Brandslund I, Jorgensen LG. 'Likelihood-ratio' and 'odds' applied to monitoring of patients as a supplement to 'reference change value' (RCV). Clin Chem Lab Med 2008;46:157-64.
(31.) Franchini M. Haemostasis and pregnancy. Thromb Haemost 2006;95:401-13.
(32.) Said JM, Ignjatovic V, Monagle PT, Walker SP, Higgins JR, Brennecke SP. Altered reference ranges for protein C and protein S during early pregnancy: implications for the diagnosis of protein C and protein S deficiency during pregnancy. Thromb Haemost 2010;103: 984-8.
(33.) Cerneca F, Ricci G, Simeone R, Malisano M, Alberico S, Guaschino S. Coagulation and fibrinolysis changes in normal pregnancy. Increased levels of procoagulants and reduced levels of inhibitors during pregnancy induce a hypercoagulable state, combined with a reactive fibrinolysis. Eur J Obstet Gynecol Reprod Biol 1997;73: 31-6.
(34.) Bremme K, Ostlund E, Almqvist I, Heinonen K, Blomback M. Enhanced thrombin generation and fibrinolytic activity in normal pregnancy and the puerperium. Obstet Gynecol 1992;80:132-7.
(35.) van Ommen CH, Fijnvandraat K, Vulsma T, Delemarre Van De Waal HA, Peters M. Acquired protein S deficiency caused by estrogen treatment of tall stature. J Pediatr 1999;135:477-81.
(36.) Kadauke S, Khor B, Van Cott EM. Activated protein C resistance testing for factor V Leiden. Am J Hematol 2014;89:1147-50.
(37.) Saha P, Stott D, Atalla R. Haemostatic changes in the puerperium '6 weeks postpartum' (HIP Study) implication for maternal thromboembolism. BJOG 2009;116:1602-12.
Ann H. Kristoffersen, [1, 2] * Per H. Petersen,  Thomas Roraas,  and Sverre Sandberg [1, 2, 3]
 Laboratory of Clinical Biochemistry, Haukeland University Hospital, Bergen, Norway;  Norwegian Quality Improvement of Primary Health Care Laboratories (Noklus), Haraldsplass Deaconess Hospital, Bergen, Norway;  Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway.
* Address correspondence to this author at: Laboratory of Clinical Biochemistry, Haukeland University Hospital, NO-5021 Bergen, Norway. Fax +47-55973115; e-mail firstname.lastname@example.org.
Received August 31, 2016; accepted November 22, 2016.
Previously published online at DOI: 10.1373/clinchem.2016.265900
 Nonstandard abbreviations: VTE, venous thromboembolism; PC, protein C; AT, antithrombin; PS, Protein S; DIC, disseminated intravascular coagulation; [CV.sub.I], within subject CV; [CV.sub.P], within-person CV (individual CV)-includes both [CV.sub.A] and [CV.sub.I]; APCR, activated protein C resistance; APTT, activated partial thromboplastin time; [CV.sub.A], within-series analytical CV; MoM, multiples of the median; lnMoM, natural logarithm of MoM; RCV, reference change value; [CV.sub.G], between-subject CV; [SD.sub.G], between-subject SD (lnMoM); [SD.sub.I], within-subject SD (lnMoM); [SD.sub.A], within-series analytical SD (lnMoM).
Caption: Fig. 1. Variation in PC(A), AT(B), PS free (C), PS activity (D), and APCR ratio (E) as a function of time for 20 healthy pregnant women during pregnancy (from week 5 until week 40) and the postpartum period (from 1-3 days until 6 weeks after delivery), and 19 nonpregnant women during 40 weeks. Each colored line represents the trajectory of the concentration for individual pregnant women, and each black line does the same for nonpregnant women. The shaded area encloses the trajectory for nonpregnant women. All results are included in the graphs except PS free and PS activity for sample 11 (2 weeks), subject 11, because of values above measurement limit (>150%). Weeks (as in "Weeks 5-9") = weeks of blood sampling in pregnancy; days (as in "1-3 days") = days after delivery; weeks (as in "2 weeks") = weeks after delivery.
Caption: Fig. 2. Variation in MoM and lnMoM PS free (A and C) and PS activity (Band D) as a function of time for 18-20 healthy pregnant women during pregnancy and the postpartum period. Each line represents the trajectory of the MoM and lnMoM values for the individual pregnant women. All results are included in the graphs, except PS free and PS activity for sample 11 (2 weeks), subject 11 because of values above the measurement limit (>150%). Weeks (as in "Weeks 5-9") = weeks of blood sampling in pregnancy; days (as in "1-3 days") = days after delivery; weeks (as in "2 weeks") = weeks after delivery.
Caption: Fig. 3. Distribution of cumulated fractions of [SD.sub.P] as function of [SD.sub.P] values for lnMoM PC (A), lnMoM AT (B), lnMoM PS free (C), lnMoM PS activity (D), and lnMoM APCR ratio (E) in 18-20 pregnant women during pregnancy. The smooth curve represents the expected distribution of the [SD.sub.P] values [around the pooled SD according to ([chi square]/df)]. For lnMoM, the distributions of [SD.sub.P] values are homogeneous for all (P > 0.01), except PS free (P = 0.008; Bartlett test).
Table 1. Medians for PC, AT, PS free, and PS activity and APCR calculated for each sampling time in pregnancy and postpartum for healthy pregnant women. (a) Gestational week in pregnancy Sample 1 2 3 4 number Sampling 5-9 9-13 13-17 17-21 Parameter time Weeks Weeks Weeks Weeks PC (a) (%), n = 19 93 103 113 119 AT (b) (%), n = 19 106 100 101 101 PS free (c) (%), n = 18 57 52 52 50 PS activity (e) (%), n = 19 61 56 54 47 APCR (ratio), (f) n = 18 2.4 2.4 2.4 2.4 Gestational week in pregnancy Sample 5 6 7 8 number Sampling 21-25 25-29 29-33 33-37 Parameter time Weeks Weeks Weeks Weeks PC (a) (%), n = 19 118 116 116 106 AT (b) (%), n = 19 104 107 107 104 PS free (c) (%), n = 18 44 42 40 43 PS activity (e) (%), n = 19 44 42 40 37 APCR (ratio), (f) n = 18 2.4 2.4 2.4 2.3 Postpartum Sample 9 10 11 12 number Sampling 37-40 1-3 2 6 Parameter time Weeks Days Weeks Weeks PC (a) (%), n = 19 112 115 119 114 AT (b) (%), n = 19 99 98 113 111 PS free (c) (%), n = 18 42 50 83 (d) 72 PS activity (e) (%), n = 19 36 35 71 (d) 69 APCR (ratio), (f) n = 18 2.3 2.4 2.4 2.4 (a-f) Excluded outlying subjects or samples (listed below). (a) Subject 17. (b) Subject 4. (c) Subjects 1 and 14. (d) Sample 11 of subject 11 was excluded from PS free and PS activity for all calculations because of values above the measurement range (>150%). (e) Subject 13 and samples 1 and 2 of subject 14. (f) Subjects 1 and 5. Table 2. [CV.sub.I], [CV.sub.G], [CV.sub.A], and unidirectional RCV calculated for lnMoM transformed to CVs and for SDs for different coagulation parameters in healthy pregnant (P) and nonpregnant (NP) women. (a) lnMoM lnMoM [CV.sub.I], % [CV.sub.G], % Parameter (95% CI) (95% CI) PC (c) (P), n = 19 8.4(7.5-9.6) 13.6(10.0-20.3) PC (NP), n = 19 6.7(6.1-7.6) 14.6(10.7-21.4) AT (d) (P), n = 19 3.8(3.3-4.3) 7.1 (5.3-10.6) AT(NP), n = 19 3.4 (3.0-3.9) 5.1 (4.0-8.0) PS free (e) (P), n = 18 11.5(10.2-13.1) 13.5(9.8-20.9) PS free (f) (NP), n = 18 8.7 (7.8-9.8) 17.7(12.5-25.4) PS activity (g) (P), n = 19 9.3(8.2-10.6) 16.2 (11.9-24.3) PS activity (NP), n = 19 7.1 (6.4-8.1) 18.8(14.0-28.1) APCR (h) (P), n = 18 0.5(0-1.2) 3.4 (2.5-5.1) APCR (NP), n = 19 1.3(0.7-1.7) 5.4 (4.1-8.1) lnMoM lnMoM [CV.sub.A], % RCV% Parameter (95% CI) decrease (b) PC (c) (P), n = 19 1.4(1.3-1.6) -18.0 PC (NP), n = 19 1.0 (0.9-1.1) -14.6 AT (d) (P), n = 19 1.5(1.4-1.7) -9.0 AT(NP), n = 19 1.7(1.5-1.9) -8.5 PS free (e) (P), n = 18 2.3 (2.1-2.6) -23.8 PS free (f) (NP), n = 18 1.9(1.7-2.1) -18.7 PS activity (g) (P), n = 19 3.4 (3.0-3.8) -20.5 PS activity (NP), n = 19 3.0 (2.7-3.3) -16.5 APCR (h) (P), n = 18 2.6 (2.3-2.9) -6.0 APCR (NP), n = 19 2.5 (2.2-2.8) -6.2 lnMoM Variance RCV SD homogeneity Parameter decrease (b) PC (c) (P), n = 19 -0.20 P = 0.11 PC (NP), n = 19 -0.16 P = 0.17 AT (d) (P), n = 19 -0.09 P = 0.09 AT(NP), n = 19 -0.09 P = 0.68 PS free (e) (P), n = 18 -0.27 P <0.01 PS free (f) (NP), n = 18 -0.21 P <0.01 PS activity (g) (P), n = 19 -0.23 P = 0.25 PS activity (NP), n = 19 -0.18 P = 0.27 APCR (h) (P), n = 18 -0.06 P = 0.03 APCR (NP), n = 19 -0.07 P = 0.36 (a) Only the 9 samples in pregnancy were included (the 3 samples after delivery were excluded). (b) Unidirectional, 95% probability (z = 1.64). (c-h) Excluded outlying subjects or samples (listed below). (c) Subject 17. (d) Subject 4. (e) Subjects 1 and 14. (f) Subject 33. (g) Subject 13 and samples 1 and 2 of subject 14. (h) Subjects 1 and 5.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Hemostasis and Thrombosis|
|Author:||Kristoffersen, Ann H.; Petersen, Per H.; Roraas, Thomas; Sandberg, Sverre|
|Date:||Apr 1, 2017|
|Previous Article:||Estimated Glomerular Filtration Rate and Albuminuria Are Associated with Biomarkers of Cardiac Injury in a Population-Based Cohort Study: The...|
|Next Article:||Characterization of a Blood Spot Creatine Kinase Skeletal Muscle Isoform Immunoassay for High-Throughput Newborn Screening of Duchenne Muscular...|