Advertisement

Comparative retrospective study on the validity of point-of-care testing device for massive obstetrical hemorrhage: dry hematology vs thromboelastography

Open AccessPublished:October 19, 2022DOI:https://doi.org/10.1016/j.ajogmf.2022.100778

      BACKGROUND

      Early recognition of hypofibrinogenemia and prompt initiation of transfusion therapy in patients with massive obstetrical hemorrhage can improve prognosis. There are reports on the usefulness of point-of-care testing, which provides quicker test results compared with fibrinogen measurements using the conventional Clauss method.

      OBJECTIVE

      This study aimed to compare and investigate the diagnostic accuracy of dry hematology and thromboelastography in point-of-care testing for the diagnosis of hypofibrinogenemia.

      STUDY DESIGN

      A single-center, retrospective study of 126 massive obstetrical hemorrhage cases with point-of-care testing before treatment was initiated. The correlation of fibrinogen values with the Clauss method and the diagnostic accuracy for hypofibrinogenemia were compared between dry hematology and thromboelastography.

      RESULTS

      Fibrinogen value in dry hematology showed a strong positive correlation with values measured by the Clauss method, and the diagnostic accuracy for hypofibrinogenemia was high, but there were many residuals above 100 mg/dL, and the distribution of these residuals was not uniform. Although thromboelastography cannot be used to directly measure fibrinogen values, maximum amplitude citrated functional fibrinogen, amplitude-10 citrated rapid thromboelastography, and amplitude-10 citrated functional fibrinogen showed a strong positive correlation with fibrinogen values using the Clauss method, and no significant difference in correlation or diagnostic accuracy was observed relative to dry hematology.

      CONCLUSION

      Dry hematology and thromboelastography were equally accurate in diagnosing hypofibrinogenemia, with results correlating well with fibrinogen values measured by the Clauss method.

      Key words

      Why was this study conducted?

      The number of reports on the diagnostic accuracy of point-of-care testing (POCT) devices in massive obstetrical hemorrhage (MOH) is limited. Although various types of POCT devices are currently used in clinical practice, few comparisons have been done between them.

      Key findings

      POCT devices were highly accurate in diagnosing hypofibrinogenemia, and there was no difference in accuracy among POCT devices.

      What does this add to what is known?

      We compared the diagnostic accuracy for hypofibrinogenemia in MOH between devices and found no substantial differences.

      Introduction

      Massive obstetrical hemorrhage (MOH) is a life-threatening condition and one of the leading causes of maternal death.
      • Say L
      • Chou D
      • Gemmill A
      • et al.
      Global causes of maternal death: a WHO systematic analysis.
      MOH often presents with coagulopathy, especially when fibrinogen initially falls below the hemostatic threshold and requires high-dose coagulation factor replacement. Because blood fibrinogen levels correlate with the severity of MOH,
      • Era S
      • Matsunaga S
      • Matsumura H
      • Murayama Y
      • Takai Y
      • Seki H.
      Usefulness of shock indicators for determining the need for blood transfusion after massive obstetric hemorrhage.
      it is important to diagnose hypofibrinogenemia early and initiate appropriate coagulation factor replacement. In recent years, point-of-care testing (POCT), a rapid and simple measurement of blood coagulation function, has attracted attention, and there have been many reports on its clinical effectiveness in the field of obstetrics.
      • Collins PW
      • Cannings-John R
      • Bruynseels D
      • et al.
      Viscoelastometry guided fresh frozen plasma infusion for postpartum haemorrhage: OBS2, an observational study.
      ,
      • Spasiano A
      • Matellon C
      • Orso D
      • et al.
      Functional fibrinogen (FLEV-TEG) versus the Clauss method in an obstetric population: a comparative study.
      The CG02N whole blood coagulation analyzer (A&T Corporation, Kanagawa, Japan) can rapidly and quantitatively measure fibrinogen levels, whereas TEG 6s (Haemonetics Corporation, Braintree, MA) does not directly measure fibrinogen levels but performs a comprehensive evaluation of coagulation and hemostatic function, including the influence of platelets, using whole blood. TEG 6s can measure the viscoelasticity of blood clots using various reagents simultaneously, resonating the clots and expressing their amplitudes graphically. ROTEM (Pentapharm GmbH, Munich, Germany), which is based on the principle of thromboelastometry, is also widely used in daily clinical practice and its use has been reported in many cases.
      • Collins PW
      • Cannings-John R
      • Bruynseels D
      • et al.
      Viscoelastometry guided fresh frozen plasma infusion for postpartum haemorrhage: OBS2, an observational study.
      The measurement principles of thromboelastography and thromboelastometry are generally the same, with the only difference being the pins and cups of the testing equipment. Our medical institution uses the CP3000 (Sekisui Medical Co, Ltd, Tokyo, Japan), which measures fibrinogen using the Clauss method for the definitive diagnosis of hypofibrinogenemia, but it has drawbacks such as a long examination time, large size, and high cost of the device for installation in a primary medical institution. The aforementioned POCT equipment enables testing in a short time and at low cost, and test results are reported to correlate well with fibrinogen levels in conventional blood testing.
      • Hayakawa M
      • Gando S
      • Ono Y
      • et al.
      Rapid evaluation of fibrinogen levels using the CG02N whole blood coagulation analyzer.
      Because POCT devices can quickly and easily assess blood coagulation activity, there have been many reports of their use in the field of emergency medicine, such as trauma
      • Da Luz LT
      • Nascimento B
      • Shankarakutty AK
      • Rizoli S
      • Adhikari NK.
      Effect of thromboelastography (TEG®) and rotational thromboelastometry (ROTEM®) on diagnosis of coagulopathy, transfusion guidance and mortality in trauma: descriptive systematic review.
      and cardiac surgery,
      • Bolliger D
      • Tanaka KA.
      Point-of-care coagulation testing in cardiac surgery.
      and some reports of use for MOH.
      • Roberts TCD
      • De Lloyd L
      • Bell SF
      • et al.
      Utility of viscoelastography with TEG 6s to direct management of haemostasis during obstetric haemorrhage: a prospective observational study.
      However, studies comparing the usefulness of each POCT instrument in treating MOH are scarce, and no studies have compared the usefulness of dry hematology and thromboelastography. Therefore, we retrospectively examined the diagnostic accuracy of POCT (dry-hematology, thromboelastography, or both) in MOH cases seen at our institution in correlation with fibrinogen levels measured by the Clauss method to detect hypofibrinogenemia (≤150 mg/dL, ≤200 mg/dL), which is particularly important in MOH. In addition, Bland–Altman plots were used to measure the residuals between the POCT device measurements and the fibrinogen values in the Clauss method.

      Materials and Methods

      Study setting

      Figure 1 shows the research targets of this study. This was a retrospective study of 320 cases of MOH seen at Saitama Medical University, Saitama Medical Center, between April 2016 and March 2019. A retrospective analysis was performed on 126 patients who underwent simultaneous blood testing with POCT (CG02N only, TEG 6s only, or both) and fibrinogen measurement with CP3000 using the conventional Clauss method during the first blood draw before transfusion, after being diagnosed with MOH. Data were collected retrospectively from medical records. Because of the difference in the timing of the introduction of CG02N and TEG 6s at our institution, the results were categorized into cases measured only with CG02N (Group 1), cases measured only with TEG 6s (Group 2), and cases measured with both (Group 3).
      Figure 1
      Figure 1Study flowchart
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      POCT, point-of-care testing.

      About point-of-care testing devices

      In this study, CG02N based on dry hematology and TEG 6s based on thromboelastography were used as POCT instruments for obstetrical hemorrhage cases. The measurement mechanism of dry hematology is that fibrinogen can be measured directly by dry methods without using reagents, and the results are available within tens of seconds. Thromboelastography is a wet method that uses reagents to evaluate the coagulation and fibrinolytic systems of the blood, and takes approximately 30 minutes to obtain all results. The details of the POCT devices are provided in Supplemental File 1.

      Various definitions of hypofibrinogenemia

      Reported cutoff values for hypofibrinogenemia vary.
      • Matsunaga S
      • Takai Y
      • Seki H.
      Fibrinogen for the management of critical obstetric hemorrhage.
      Fibrinogen ≤200 mg/dL has been reported to be a risk factor for postpartum hemorrhage,
      • Cortet M
      • Deneux-Tharaux C
      • Dupont C
      • et al.
      Association between fibrinogen level and severity of postpartum haemorrhage: secondary analysis of a prospective trial.
      ,
      • Charbit B
      • Mandelbrot L
      • Samain E
      • et al.
      The decrease of fibrinogen is an early predictor of the severity of postpartum hemorrhage.
      and fibrinogen levels at ≤130 to 155 mg/dL can lead to high risk of massive blood transfusions
      • Era S
      • Matsunaga S
      • Matsumura H
      • Murayama Y
      • Takai Y
      • Seki H.
      Usefulness of shock indicators for determining the need for blood transfusion after massive obstetric hemorrhage.
      ,
      • Wang L
      • Matsunaga S
      • Mikami Y
      • Takai Y
      • Terui K
      • Seki H.
      Pre-delivery fibrinogen predicts adverse maternal or neonatal outcomes in patients with placental abruption.
      in MOH. In this study, we measured the diagnostic accuracy of POCT for detecting fibrinogen levels ≤150 mg/dL and ≤200 mg/dL.

      Inclusion criteria

      This study included patients who came to our institution and were diagnosed with MOH (total blood loss ≥1000 mL

      WHO guidelines approved by the guidelines review committee. WHO recommendations for the prevention and treatment of postpartum haemorrhage, World Health Organization. https://apps.who.int/iris/bitstream/handle/10665/75411/9789241548502_eng.pdf. (website last accessed on November 9, 2022)

      ) postpartum, or those who were diagnosed with MOH after delivery in another institution and were transferred to our institution.

      Exclusion criteria

      Cases with a total blood loss of <1000 mL, cases in which endpoints could not be evaluated retrospectively (eg, incomplete medical records), and cases in which blood transfusion or tranexamic acid was administered before POCT measurement were excluded.

      Method of analysis for Group 1

      The correlation between fibrinogen levels from CG02N and those from CP3000 measured by the Clauss method was examined in all 65 patients who were tested only with CG02N. We also estimated the diagnostic accuracy of CG02N for hypofibrinogenemia (≤150 mg/dL and ≤200 mg/dL) using the Clauss method.

      Method of analysis for Group 2

      In Group 2, only 29 patients who were measured using TEG 6s were included. Because TEG 6s does not directly measure fibrinogen values, we examined the correlation between those indexes that can be directly measured and fibrinogen values measured using the conventional Clauss method. We also estimated the diagnostic accuracy of each TEG 6s index against hypofibrinogenemia using the Clauss method.

      Method of analysis for Group 3

      A total of 32 patients who were simultaneously tested using CG02N and TEG 6s were included in the study. The correlation between fibrinogen values measured using both CG02N and TEG 6s and those measured using the Clauss method was determined. The diagnostic accuracy of the 2 methods for hypofibrinogenemia was then compared.

      Ethics approval and participant consent

      This study was approved by the Ethics Committee of the Saitama Medical Center, Saitama Medical University.

      Characteristics of the medical institution where the study was conducted

      This study was conducted at Saitama Medical Center, Saitama Medical University, a general perinatal medical center located in Saitama Prefecture. Our institutions accept all cases of MOH that occur in primary and secondary medical facilities in Saitama Prefecture. Therefore, we often encounter patients who have lost >1000 mL of blood but are transferred to our institution without receiving blood transfusions or tranexamic acid.
      The basic policy for MOH patients at our institution is described below. When a patient with MOH is transported to our institution, blood analysis is performed immediately on arrival. Simultaneously with the measurement of coagulation activity by POCT, fibrinogen measurement by the Clauss method is initiated. Blood transfusion is then started on the basis of the patient's vital signs, total blood loss at the time of transport, and results from POCT. For hypofibrinogenemia, fibrinogen concentrate should be administered first, with the red cell concentrate (RCC) and fresh frozen plasma (FFP) being approximately 1:1. After transfusion, frequent blood tests are performed to check for vital signs and to minimize complications from excessive transfusion.

      Statistical analysis

      For the comparison between dry hematology and thromboelastography in Group 3, the Fisher exact ratio test was used for nominal variables, the Shapiro–Wilk test for normality, and t-tests were used for continuous variables. For comparisons of patient background among Groups 1, 2, and 3, the Fisher exact ratio test was used for nominal variables, and 1-way analysis of variance for continuous variables. The area under the curve (AUC) in the receiver operating characteristic (ROC) curve was used to determine the accuracy of the hypofibrinogenemia diagnosis. JMP version 14 (SAS Institute, Cary, NC) was used for statistical analysis.

      Results

      Comparison between groups

      Table 1 shows the comparison of obstetrical backgrounds among the groups. Group 2 had significantly fewer gestational weeks at delivery compared with Groups 1 and 3 (35.5±3.9 vs 37.9±4.2 vs 37.9±4.2 weeks), higher body mass index (24.0±3.7 vs 22.6±3.6 vs 21.8 ±3.8), and higher cesarean delivery rates (82.8% vs 36.9% vs 59.4%), respectively. There were no significant differences in maternal age, primiparity rates, or intrauterine fetal death rates among the 3 groups.
      Table 1Obstetrical characteristics of each group
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      CharacteristicGroup 1: analyzed with CG02N

      (n=65)
      Group 2: analyzed with TEG 6s

      (n=29)
      Group 3: analyzed with both CG02N and TEG 6s

      (n=32)
      P value
      Maternal age (y)33.8±5.735.4±5.434.9±5.9.394
      Gestational age at delivery (wk)37.9±4.2
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      35.5±3.9
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      37.9±3.0
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .015
      There were statistically significant differences among the three groups.
      Primiparity40 cases (61.5%)18 cases (62.1%)16 cases (50%).522
      Intrauterine fetal death4 cases (6.2%)0 case (0%)3 cases (9.4%).2671
      Body mass index22.6±3.624.0±3.721.8±3.8.0676
      There were statistically significant differences among the three groups.
      Cesarean delivery24 cases (36.9%)
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      24 cases (82.8%)
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      19 cases (59.4%)
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      <.0001
      There were statistically significant differences among the three groups.
      The Fisher exact ratio test was used to compare nominal variables and 1-way analysis of variance was used to compare continuous variables.
      a,b Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      c There were statistically significant differences among the three groups.
      Table 2 lists the results of a blood test from the hospital laboratory for each group, and the amount of blood loss, transfusion volume, and causative disease of MOH. Group 2 had less preexamination and total bleeding than Groups 1 and 3, significantly lower doses of RCC, FFP, platelet concentrate, and fibrinogen concentrate, and significantly higher hemoglobin and fibrinogen levels on blood tests.
      Table 2Hematological data of each group
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      Hematological dataGroup 1: analyzed with CG02N (n=65)Group 2: analyzed with TEG 6s (n=29)Group 3: analyzed with both CG02N and TEG 6s (n=32)P value
      Preexamination blood loss (mL)2044±1029
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      1393±906
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      1992±1164
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0175
      Total blood loss (mL)3510±2386
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      2360±1759
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      3212±2467
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0367
      There were statistically significant differences among the three groups.
      Results of a blood test from hospital laboratoryHemoglobin (g/dL)7.5±2.0
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      8.7±2.0
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      7.8±2.1
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0437
      There were statistically significant differences among the three groups.
      Platelet count (104/μL)15.6±6.814.5±5.713.4±6.2.2922
      APTT (s)42.5±35.937.6±16.834.8±18.8.3817
      PT% (%)79.6±25.589.2±21.082.7±27.2.176
      FDP (μg/mL)174.4±293.3
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      62.8±44.6
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      196.3±309.0
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0031
      There were statistically significant differences among the three groups.
      Fibrinogen (mg/dL)202.1±95.9
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      262.3±98.8
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      192.3±99.4
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0129
      There were statistically significant differences among the three groups.
      Total blood transfusion volumeRed cell concentrate (mL)1620±1566
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      705±1065
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      1299±1500
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0062
      There were statistically significant differences among the three groups.
      Fresh frozen plasma (mL)1303±1595
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      621±993
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      1178±1530
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0218
      There were statistically significant differences among the three groups.
      Platelet concentrate (mL)112±243
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      34±94
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      144±245
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0163
      There were statistically significant differences among the three groups.
      Fibrinogen concentrate (g)2.4±2.8
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      1.0±2.0
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      2.7±3.4
      Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      .0163
      There were statistically significant differences among the three groups.
      Cause of PPHAtonic hemorrhage29 cases (44.6%)21 cases (72.4%)15 cases (46.9%)
      Perineal laceration15 cases (23.1%)3 case (10.3%)4 cases (12.5%)
      Retained products of conception, placenta accreta9 cases (13.8%)3 cases (10.3%)2 cases (6.3%)
      Placental abruption8 cases (12.3%)2 cases (6.9%)7 cases (21.9%)
      Amniotic fluid embolism3 cases (4.6%)0 case (0%)1 case (3.1%)
      Others1 case (1.5%)0 case (0%)3 case (9.3%)
      The Fisher exact ratio test was used to compare nominal variables and 1-way analysis of variance was used to compare continuous variables.
      APTT, activated partial thromboplastin time; FDP, fibrinogen degradation product; PPH, postpartum hemorrhage; PT, prothrombin time.
      a,b Multiple comparisons were performed for variables that showed statistically significant differences among the three groups. Compared to a, b showed a statistically significant difference
      c There were statistically significant differences among the three groups.

      Analysis for Group 1 and Group 2

      Figure 2 shows a scatter plot of fibrinogen values measured using the Clauss method with CP3000 and fibrinogen values measured with CG02N. The correlation coefficient was 0.7944 (P<.001), indicating a significant correlation between the 2 variables.
      Figure 2
      Figure 2Scatter plots for fibrinogen measured by Clauss method and CG02N
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      The solid red line represents the regression line and the dashed red line represents the 95% confidence interval; regression line: y=49.905871+0.6937446*x; correlation coefficient: 0.7944; coefficient of determination (R2): 0.631071.
      Table 3 lists the diagnostic accuracy of each POCT device for hypofibrinogenemia using the Clauss method. The AUCs calculated from ROC curves for fibrinogen <150 mg/dL and <200 mg/dL for the respective definitions of hypofibrinogemia were 0.969 and 0.881, respectively.
      Table 3Area under the curve values using receiver operating characteristic curves for hypofibrinogenemia of ≤150 mg/dL and ≤200 mg/dL
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      GroupAUC value for fibrinogen ≤150 mg/dL95% CIAUC value for fibrinogen ≤200 mg/dL95% CICorrelation coefficient for fibrinogen values by Clauss method
      Analysis 1 (measured by CG02N)Fibrinogen measured by CG02N0.969(0.925–1.000)0.881(0.798–0.963)0.794
      Analysis 2 (measured by TEG 6S)MA-CFF (mm)0.936(0.837–1.000)0.940(0.851–1.000)0.74
      A-10 CRT (mm)0.987(0.952–1.000)0.920(0.822–1.000)0.57
      A-10 CFF (mm)0.962(0.891–1.000)0.946(0.838–1.000)0.67
      Analysis 1 and Analysis 2 show correlation coefficients between fibrinogen values determined by the Clauss method and each item in TEG 6s.
      A-10 CFF, amplitude-10 citrated functional fibrinogen; A-10 CRT, amplitude-10 citrated rapid thromboelastography; AUC, area under the curve; CI, confidence interval; MA-CFF, maximum amplitude citrated functional fibrinogen.
      Table 3 also lists the AUC for each TEG 6s value using the ROC curve for hypofibrinogenemia and the correlation coefficient with the fibrinogen value using the Clauss method (for TEG 6s, only the 3 parameters with high AUC values are listed; other parameters are included in Supplemental File 2). The diagnostic accuracy of each of the TEG 6s parameters was high for levels of hypofibrinogenemia at ≤150 mg/dL and ≤200 mg/dL, with high AUC values for maximum amplitude citrated functional fibrinogen (MA-CFF), amplitude-10 citrated rapid thromboelastography (A-10 CRT), and amplitude-10 citrated functional fibrinogen (A-10 CFF), and other parameters. Significant positive correlations with fibrinogen values were also observed using the Clauss method.
      Figure 3 shows a Bland–Altman plot of fibrinogen values measured with CG02N vs those measured using the Clauss method. The mean difference in fibrinogen values for CG02N vs the Clauss method was 17.3 mg/dL (95% confidence interval, 0.61–33.9). The plots did not show a consistent distribution, and there were a certain number of measured residuals that showed residuals of 100 mg/dL or more, regardless of fibrinogen value.
      Figure 3
      Figure 3Bland-Altman plot of fibrinogen of CG02N vs Clauss method
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      The vertical axis represents the difference between the fibrinogen values of the Clauss method and CG02N, and the horizontal axis represents the mean of the 2 values. The solid red line indicates the mean of the difference, and the dashed red line indicates its 95% confidence interval.

      Analysis for Group 3

      Table 4 lists the AUC values for the fibrinogen values from the CG02N and each of the TEG 6s parameters using the ROC curve for hypofibrinogenemia, the P values of the ROC comparison between CG02N and each of the TEG 6s parameters, and the correlation coefficient of the Clauss method for fibrinogen values (for TEG 6s, only the top 3 parameters with the highest AUC values are listed; other parameters are presented in Supplemental File 3). Both CG02N and TEG 6s had high AUC values for the diagnosis of hypofibrinogenemia and high diagnostic accuracy for both definitions of hypofibrinogenemia. There was no significant difference in the AUC values of CG02N and TEG 6s on the ROC curve for diagnostic accuracy for hypofibrinogenemia. A strong positive correlation was shown between the Clauss method and fibrinogen levels in CG02N (as in Analysis 1), and TEG 6s also showed strong positive correlations for MA-CFF, A-10 CRT, and A-10 CFF.
      Table 4Area under the curve values using receiver operating characteristic curve for hypofibrinogenemia at ≤150 mg/dL and ≤200 mg/dL
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      GroupAUC value for fibrinogen ≤150 mg/dL95% CIP value
      P value of t-test comparing CG02N against AUC values for hypofibrinogenemia.
      AUC value for fibrinogen ≤200 mg/dL95% CIP value
      P value of t-test comparing CG02N against AUC values for hypofibrinogenemia.
      Correlation coefficient for fibrinogen values by Clauss method
      Fibrinogen measured by: CG02N (mg/dL)0.961(0.882–1.000)0.949(0.878–1.000)0.91
      TEG 6sMA-CFF (mm)0.952(0.885– 1.000).7580.933(0.835–1.000).7770.87
      A-10 CRT (mm)0.942(0.854–1.000).3250.961(0.903–1.000).7260.87
      A-10 CFF (mm)0.961(0.882–1.000)10.957(0.897–1.000).7950.91
      Analysis 3 and comparison with AUC values of CG02N for hypofibrinogenemia were done using Clauss method correlation coefficients between fibrinogen values and each item.
      A-10 CFF, amplitude-10 citrated functional fibrinogen; A-10 CRT, amplitude-10 citrated rapid thromboelastography; AUC, area under the curve; CI, confidence interval; MA-CFF, maximum amplitude citrated functional fibrinogen.
      a P value of t-test comparing CG02N against AUC values for hypofibrinogenemia.
      Figure 4 shows a scatter plot of fibrinogen values from CG02N and each parameter of TEG 6s against fibrinogen values using the Clauss method. Only the 3 parameters having the highest correlation coefficients with fibrinogen values measured by the Clauss method among all the TEG 6s parameters are shown. The nonlinear model using logistic regression is indicated by the solid line in the figure. Fibrinogen values were generally linearly correlated across the entire range for CG02N and MA-CFF and A-10 CFF in TEG 6s. A-10 CRT tended to give higher results when fibrinogen using the Clauss method was low and lower results when fibrinogen was high.
      Figure 4
      Figure 4Scatter plots for fibrinogen measured by each POCT and Clauss method
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      The nonlinear model using logistic regression is indicated by solid lines.
      A-10 CFF, amplitude-10 citrated functional fibrinogen; A-10 CRT, amplitude-10 citrated rapid thromboelastography; MA-CFF, maximum amplitude citrated functional fibrinogen.
      Figure 5 shows a Bland–Altman plot for the fibrinogen values measured by the Clauss method in Analysis 3. The fibrinogen values in CG02N showed a certain number of plots with large residuals >100 mg/dL regardless of the distribution of fibrinogen values in the Clauss method, similar to the results of Analysis 1. For the other parameters in the TEG 6s (MA-CFF, A-10 CRT, A-10 CFF), the correlations were strong, but the variation tended to increase as the fibrinogen values increased.
      Figure 5
      Figure 5Bland-Altman plot of fibrinogen of TEG6s vs Clauss method
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      The vertical axis shows the difference between the fibrinogen values of the Clauss method and each TEG 6s parameter, and the horizontal axis shows the mean of the 2 values. The solid red line indicates the mean difference and the dashed red line the 95% confidence interval.
      A-10 CFF, amplitude-10 citrated functional fibrinogen; A-10 CRT, amplitude-10 citrated rapid thromboelastography; MA-CFF, maximum amplitude citrated functional fibrinogen.
      Table 5 uses ROC curves to indicate the comparison of the A-10 CFF and TEG 6s parameters that correlate best with fibrinogen values using the Clauss method for the diagnostic accuracy for hypofibrinogenemia. The A-10 CFF cutoff for the diagnosis of hypofibrinogenemia at ≤150 mg/dL was 4.0 mm (sensitivity, 0.909; specificity, 1.000), and a comparison of the AUC with CG02N showed no significant difference between the two. For the diagnosis of hypofibrinogenemia at ≤200 mg/dL, the A-10 CFF cutoff was 5.0 mm (sensitivity, 0.800; specificity, 1.000), and similarly, there was no significant difference between the two.
      Table 5Comparison of diagnostic accuracy of amplitude-10 citrated functional fibrinogen and dry hematology for hypofibrinogenemia measured by the Clauss method
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      GroupFibrinogen measured by CG02NA-10 CFF value measured by TEG 6sP value comparing the AUC of the two
      Hypofibrinogenemia (≤150 mg/dL) measured by Clauss methodCutoff value148.1 (mg/dL)4.0 (mm)
      AUC0.9610.9611.000
      Hypofibrinogenemia (≤200 mg/dL) measured by Clauss methodCutoff value211.0 (mg/dL)5.0 (mm)
      AUC0.9490.957.795
      The diagnostic accuracy of dry hematology and A-10 CFF for hypofibrinogenemia by the Clauss method were compared, respectively, in Analysis 3. AUC values in receiver operating characteristic curves were calculated for both methods.
      A-10 CFF, amplitude-10 citrated functional fibrinogen; AUC, area under the curve.

      Discussion

      Principal findings

      In this study, we compared the diagnostic accuracy for hypofibrinogenemia between CG02N using dry hematology and TEG 6s using thromboelastography. The diagnostic accuracies of the 2 devices were equivalent. CG02N showed high diagnostic accuracy compared with CP3000 at fibrinogen levels of approximately 150 mg/dL and a large residual difference in measurements of ≥300 mg/dL. TEG 6s also showed high diagnostic accuracy in MA-CFF, A-10 CRT, and A-10 CFF parameters for hypofibrinogenemia, defined as fibrinogen ≤150 mg/dL and ≤200 mg/dL, respectively, and significant correlation with fibrinogen levels.
      Dry hematology provides test results faster than thromboelastography. In the clinical situation of MOH, where immediate response is required, rapid diagnosis of hypofibrinogenemia by dry hematology is extremely important for decision-making in the clinical situation. This study demonstrated the noninferiority of dry hematology compared with thromboelastography, and its faster examination time.

      Results

      It has been reported that fibrinogen values measured by dry hematology correlate well with fibrinogen values measured by the Clauss method.
      • Ogawa S
      • Tanaka KA
      • Nakajima Y
      • et al.
      Fibrinogen measurements in plasma and whole blood: a performance evaluation study of the dry-hematology system.
      This study also revealed good correlation for CG02N with existing fibrinogen values, as in previous literature, but it also examined its diagnostic accuracy for hypofibrinogenemia and compared it with that of TEG 6s. In this study, the analysis focused on blood fibrinogen level as the final outcome. Fibrinogen levels constitute the most important blood coagulation index in MOH, and although many coagulation factors are required for hemostasis, fibrinogen requires the highest blood concentration.
      • Hiippala ST
      • Myllylä GJ
      • Vahtera EM.
      Hemostatic factors and replacement of major blood loss with plasma-poor red cell concentrates.
      In MOH, it has been noted that the total dosage of RCC and FFP may be reduced and prognosis improved if hypofibrinogenemia is identified early and fibrinogen concentrate is administered.
      • Matsunaga S
      • Takai Y
      • Nakamura E
      • et al.
      The clinical efficacy of fibrinogen concentrate in massive obstetric haemorrhage with hypofibrinogenaemia.
      Early identification of hypofibrinogenemia using POCT devices and development of transfusion protocols based on the results may improve maternal prognosis in MOH and save medical resources, including transfusion products, as has been mentioned in a few similar reports.
      • Snegovskikh D
      • Souza D
      • Walton Z
      • et al.
      Point-of-care viscoelastic testing improves the outcome of pregnancies complicated by severe postpartum hemorrhage.
      Although the study did not have a protocol in place that dictated the transfusion dose according to the POCT results, it did provide cutoff values for hypofibrinogenemia with CG02N and TEG 6s (especially for A-10 CFF); therefore, prospective studies or randomized controlled studies using POCT instruments will be considered in the future.
      As mentioned above, there have been numerous reports on the usefulness of POCT devices in the fields of cardiovascular surgery and emergency medicine; however, reports from the field of obstetrics are scarce. Pregnancy has special hemodynamic and coagulability characteristics that make it difficult to treat hemostatic coagulation in the same manner as in other adults. Compared with nonpregnant women, circulating blood volume increases by up to 50% by approximately 34 weeks of gestation in normal pregnancies,
      • Hill CC
      • Pickinpaugh J.
      Physiologic changes in pregnancy.
      and all coagulation factor activities are increased in pregnant women, except for factor 13, which is relatively suppressed in the fibrinolytic system.
      • Hui C
      • Lili M
      • Libin C
      • et al.
      Changes in coagulation and hemodynamics during pregnancy: a prospective longitudinal study of 58 cases.
      Although CG02N is generally considered to show no measurement residuals in patients with general bleeding disorders,
      • Ogawa S
      • Tanaka KA
      • Nakajima Y
      • et al.
      Fibrinogen measurements in plasma and whole blood: a performance evaluation study of the dry-hematology system.
      a certain amount of residuals were observed regardless of fibrinogen level. This may have been because of increased coagulation factor activity in the pregnant women described above. Fibrinogen is also thought to have some effect because obstetrical disseminated intravascular coagulation (DIC) patients undergo a dramatic change in blood properties from hypercoagulable to impaired coagulation at ≤200 mg/dL.
      • Cortet M
      • Deneux-Tharaux C
      • Dupont C
      • et al.
      Association between fibrinogen level and severity of postpartum haemorrhage: secondary analysis of a prospective trial.
      Although dry hematology is affected by hematocrit values, pregnant women develop pseudoanemia because of hemodilution; therefore, the hematocrit calculation method for the general population may not accurately determine coagulability in pregnant women.
      Fibrinogen requires the highest concentration of any coagulation factor during hemostasis
      • Hiippala ST
      • Myllylä GJ
      • Vahtera EM.
      Hemostatic factors and replacement of major blood loss with plasma-poor red cell concentrates.
      ; therefore, when infusion is used to maintain vital signs such as blood pressure in the event of massive bleeding, there is a risk of dilutional coagulopathy.
      • Levy JH
      • Szlam F
      • Tanaka KA
      • Sniecienski RM.
      Fibrinogen and hemostasis: a primary hemostatic target for the management of acquired bleeding.
      Furthermore, MOH is associated with the risk of consumptive coagulopathy, in which a large amount of coagulation factor is consumed to stop bleeding from the placental abruption surface, and fibrinogen is also consumed first, resulting in more severe bleeding.
      • Thachil J
      • Toh CH.
      Disseminated intravascular coagulation in obstetric disorders and its acute haematological management.
      In conclusion, POCT is useful in MOH because it is thought to improve prognosis by recognizing hypofibrinogenemia earlier and helping to provide appropriate transfusion therapy, compared with other forms of massive hemorrhage.
      The TEG 6s parameters that showed high diagnostic accuracy for hypofibrinogenemia with the Clauss method were MA-CFF, A-10 CRT, and A-10 CFF. Figure 6 shows the actual TEG 6s output screen and a description of the parameters. The difference between A-10 CRT and A-10 CFF is that A-10 CRT measures the maximum coagulation capacity, whereas A-10 CFF measures platelet binding to fibrinogen by adding a GIIb/IIIa receptor antagonist that inhibits platelet fibrinogen binding to measure the strength of clots. It has been reported that MA in TEG 6s can be measured within 10 minutes and that MA values (especially MA-CFF) correlate well with fibrinogen values in cardiac bypass surgery,
      • Ortmann E
      • Rubino A
      • Altemimi B
      • Collier T
      • Besser MW
      • Klein AA.
      Validation of viscoelastic coagulation tests during cardiopulmonary bypass.
      making it suitable for intraoperative management and for determining whether to administer fibrinogen concentrate.
      • Tamura T.
      Predicting results of fibrinogen and platelet levels by TEG6s during cardiopulmonary bypass: a pilot study.
      ,
      • Yamamoto Y
      • Sato Y
      • Takahashi M
      • Yamamoto H
      • Echizen M
      • Uchida T.
      TEG6s Platelet Mapping assay for the estimation of plasma fibrinogen concentration during cardiovascular surgery: a single-center prospective observational study.
      In this study, A-10 was the only factor with strong correlation to fibrinogen values, but A-10 represents the amplitude height 10 minutes after MA, and it takes approximately 15 to 20 minutes to obtain results. The time required to measure fibrinogen by the Clauss method is generally 30 to 50 minutes, depending on the report.
      • Toulon P
      • Ozier Y
      • Ankri A
      • Fléron MH
      • Leroux G
      • Samama CM.
      Point-of-care versus central laboratory coagulation testing during haemorrhagic surgery. A multicenter study.
      We consider that TEG 6s MA and A-10 are instruments capable of providing results of coagulation capacity that correlate well with fibrinogen levels in the blood, more quickly than conventional testing instruments such as CP3000.
      Figure 6
      Figure 6Explanation of TEG 6s output screen and each parameter
      Nakamura. Comparison of diagnostic accuracy of point-of-care testing devices in massive obstetrical hemorrhage. Am J Obstet Gynecol MFM 2022.
      The actual output screen of TEG 6s, with processing.

      Clinical implications

      Dry hematology and thromboelastography can rapidly diagnose hypofibrinogenemia in patients with MOH, and both have demonstrated high diagnostic accuracy. Early detection of hypofibrinogen may not only reduce mortality and total blood loss, but also be cost-effective in reducing the total amount of blood transfusions required. However, to prove their effectiveness, it is necessary to conduct a randomized study with a POCT-using group and a POCT-nonusing group.

      Research implications

      Very few studies have demonstrated the usefulness of POCT for MOH. In addition to the need for high-quality studies such as randomized controlled trials in the future, active use of POCT in regions and countries with limited medical resources may be encouraged if cost-effectiveness can be demonstrated.

      Strengths and limitations

      This study has several limitations. Although it focused solely on blood fibrinogen levels and their correlation, it is important to note that MOH is not only caused by a decrease in blood fibrinogen levels. In particular, MOH with consumptive coagulopathy may present as DIC with severe hyperfibrinolysis, and further research is needed to incorporate the evaluation of the fibrinolytic system, which is possible with TEG 6s, into the management of MOH.
      In this study, Group 2 (analyzed only with TEG 6s) had significantly less background bleeding and milder coagulopathy compared with the other groups. This higher prevalence of patients with mild disease in Group 2 indicates selection bias. We consider this to be because of the fact that the proportion of consumptive coagulopathies (placental abruption or amniotic fluid embolization), which tend to cause rapid coagulopathy, was smaller in the MOH group than in the other groups. Although it would have been desirable to unify the backgrounds among the groups, we did not perform background-adjusted matching because the number of cases tested with TEG 6s alone was small (29 in total), and we were concerned that such matching would be highly confounded. However, we did not consider it appropriate to interpret the results only by comparing the diagnostic accuracy for hypofibrinogenemia between Group 1 (dry hematology only) and Group 2 (thromboelastography only), which were heterogeneous, and thus we added an analysis for Group 3, in which the same patients were measured with 2 point-of-care tests at the same time. This should resolve the difficulty in interpreting the results owing to heterogeneity between groups.

      Conclusion

      Dry hematology and thromboelastography (especially MA-CFF, A-10 CRT, and A-10 CFF) showed equivalent diagnostic accuracy for hypofibrinogenemia, with results correlating well with fibrinogen measured by the Clauss method.

      Acknowledgments

      We would like to thank Editage (www.editage.com) for the English-language editing.

      Appendix. Supplementary materials

      References

        • Say L
        • Chou D
        • Gemmill A
        • et al.
        Global causes of maternal death: a WHO systematic analysis.
        Lancet Glob Health. 2014; 2: e323-e333
        • Era S
        • Matsunaga S
        • Matsumura H
        • Murayama Y
        • Takai Y
        • Seki H.
        Usefulness of shock indicators for determining the need for blood transfusion after massive obstetric hemorrhage.
        J Obstet Gynaecol Res. 2015; 41: 39-43
        • Collins PW
        • Cannings-John R
        • Bruynseels D
        • et al.
        Viscoelastometry guided fresh frozen plasma infusion for postpartum haemorrhage: OBS2, an observational study.
        Br J Anaesth. 2017; 119: 422-434
        • Spasiano A
        • Matellon C
        • Orso D
        • et al.
        Functional fibrinogen (FLEV-TEG) versus the Clauss method in an obstetric population: a comparative study.
        BMC Anesthesiol. 2019; 19: 90
        • Hayakawa M
        • Gando S
        • Ono Y
        • et al.
        Rapid evaluation of fibrinogen levels using the CG02N whole blood coagulation analyzer.
        Semin Thromb Hemost. 2015; 41: 267-271
        • Da Luz LT
        • Nascimento B
        • Shankarakutty AK
        • Rizoli S
        • Adhikari NK.
        Effect of thromboelastography (TEG®) and rotational thromboelastometry (ROTEM®) on diagnosis of coagulopathy, transfusion guidance and mortality in trauma: descriptive systematic review.
        Crit Care. 2014; 18: 518
        • Bolliger D
        • Tanaka KA.
        Point-of-care coagulation testing in cardiac surgery.
        Semin Thromb Hemost. 2017; 43: 386-396
        • Roberts TCD
        • De Lloyd L
        • Bell SF
        • et al.
        Utility of viscoelastography with TEG 6s to direct management of haemostasis during obstetric haemorrhage: a prospective observational study.
        Int J Obstet Anesth. 2021; 47103192
        • Matsunaga S
        • Takai Y
        • Seki H.
        Fibrinogen for the management of critical obstetric hemorrhage.
        J Obstet Gynaecol Res. 2019; 45: 13-21
        • Cortet M
        • Deneux-Tharaux C
        • Dupont C
        • et al.
        Association between fibrinogen level and severity of postpartum haemorrhage: secondary analysis of a prospective trial.
        Br J Anaesth. 2012; 108: 984-989
        • Charbit B
        • Mandelbrot L
        • Samain E
        • et al.
        The decrease of fibrinogen is an early predictor of the severity of postpartum hemorrhage.
        J Thromb Haemost. 2007; 5: 266-273
        • Wang L
        • Matsunaga S
        • Mikami Y
        • Takai Y
        • Terui K
        • Seki H.
        Pre-delivery fibrinogen predicts adverse maternal or neonatal outcomes in patients with placental abruption.
        J Obstet Gynaecol Res. 2016; 42: 796-802
      1. WHO guidelines approved by the guidelines review committee. WHO recommendations for the prevention and treatment of postpartum haemorrhage, World Health Organization. https://apps.who.int/iris/bitstream/handle/10665/75411/9789241548502_eng.pdf. (website last accessed on November 9, 2022)

        • Ogawa S
        • Tanaka KA
        • Nakajima Y
        • et al.
        Fibrinogen measurements in plasma and whole blood: a performance evaluation study of the dry-hematology system.
        Anesth Analg. 2015; 120: 18-25
        • Hiippala ST
        • Myllylä GJ
        • Vahtera EM.
        Hemostatic factors and replacement of major blood loss with plasma-poor red cell concentrates.
        Anesth Analg. 1995; 81: 360-365
        • Matsunaga S
        • Takai Y
        • Nakamura E
        • et al.
        The clinical efficacy of fibrinogen concentrate in massive obstetric haemorrhage with hypofibrinogenaemia.
        Sci Rep. 2017; 7: 46749
        • Snegovskikh D
        • Souza D
        • Walton Z
        • et al.
        Point-of-care viscoelastic testing improves the outcome of pregnancies complicated by severe postpartum hemorrhage.
        J Clin Anesth. 2018; 44: 50-56
        • Hill CC
        • Pickinpaugh J.
        Physiologic changes in pregnancy.
        Surg Clin North Am. 2008; 88: 391-401
        • Hui C
        • Lili M
        • Libin C
        • et al.
        Changes in coagulation and hemodynamics during pregnancy: a prospective longitudinal study of 58 cases.
        Arch Gynecol Obstet. 2012; 285: 1231-1236
        • Levy JH
        • Szlam F
        • Tanaka KA
        • Sniecienski RM.
        Fibrinogen and hemostasis: a primary hemostatic target for the management of acquired bleeding.
        Anesth Analg. 2012; 114: 261-274
        • Thachil J
        • Toh CH.
        Disseminated intravascular coagulation in obstetric disorders and its acute haematological management.
        Blood Rev. 2009; 23: 167-176
        • Ortmann E
        • Rubino A
        • Altemimi B
        • Collier T
        • Besser MW
        • Klein AA.
        Validation of viscoelastic coagulation tests during cardiopulmonary bypass.
        J Thromb Haemost. 2015; 13: 1207-1216
        • Tamura T.
        Predicting results of fibrinogen and platelet levels by TEG6s during cardiopulmonary bypass: a pilot study.
        J Clin Anesth. 2019; 58: 59-60
        • Yamamoto Y
        • Sato Y
        • Takahashi M
        • Yamamoto H
        • Echizen M
        • Uchida T.
        TEG6s Platelet Mapping assay for the estimation of plasma fibrinogen concentration during cardiovascular surgery: a single-center prospective observational study.
        J Anesth. 2022; 36: 79-88
        • Toulon P
        • Ozier Y
        • Ankri A
        • Fléron MH
        • Leroux G
        • Samama CM.
        Point-of-care versus central laboratory coagulation testing during haemorrhagic surgery. A multicenter study.
        Thromb Haemost. 2009; 101: 394-401