Objective
The aim of this study was to determine if the fibrinogen/C-reactive protein (CRP) ratio could be used in obstetrics as a predictor for a disseminated intravascular coagulation.
Study Design
One hundred eleven patients with hemolysis, elevated liver enzymes, and low platelet count syndrome at the Department of Obstetrics and Fetomaternal Medicine (General Hospital, Vienna, Austria) were selected and divided into 2 groups (overt disseminated intravascular coagulation, no overt disseminated intravascular coagulation). The classical parameters and the fibrinogen/CRP ratio were compared. The analysis was carried out using IBM SPSS statistical package (SPSS, Inc, Cary, NC).
Results
The fibrinogen/CRP ratio showed significant differences. The receiver-operating characteristic analysis showed for the ratio (area under the curve, 0.74) significantly better discriminative power than for fibrinogen (area under curve, 0.59). The odds ratio for the fibrinogen/CRP ratio was 7.04. Finally, significant correlations between the ratio and the neonatal outcome were found.
Conclusion
We suggest the implementation of the fibrinogen/CRP ratio within patients with hemolysis, elevated liver enzymes, and low platelet count syndrome as a diagnostic and prognostic factor for the occurrence of disseminated intravascular coagulation.
Hemolysis, elevated liver enzymes, and low platelet count (HELLP) are alterations in laboratory tests that characteristically refer to a life-threatening obstetric complication considered by many to be a variant of preeclampsia (9-14%). It was in 1982 by Dr Louis Weinstein, who coined the acronym HELLP for this set of abnormalities. The HELLP syndrome is aggravated in nearly 4-21% with a further complication, the disseminated intravascular coagulation (DIC). For identifying this diagnosis the International Society of Thrombosis and Haemostasis (ISTH) Subcommittee on DIC has proposed laboratory criteria for calculating an overt DIC. This score, somehow seen as gold standard, includes the coagulation parameters platelet count, fibrinogen, prothrombin time (PT) as well as the D-dimer.
However, these ISTH criteria are faced critically by several recent studies. According to the ISTH criteria, hypofibrinogenemia is regarded to reflect the consumption of coagulation factor during intravascular fibrin formation, but mostly the DIC patients do not show decreased levels. Kor-anantakul and Lekhakula could demonstrate that only 33% of the patients with pregnancy-related conditions that were predisposed to DIC development fulfilled the ISTH criteria regarding the parameter fibrinogen.
So Kim et al revealed a new diagnostic issue: for the author’s study population 1056 patients who developed a DIC because of different underlying diseases were selected and investigated regarding coagulation parameters. It turned out that only 8-6% of these patients showed fibrinogen levels under the cutoff value of 100 mg/dL as proposed in the ISTH criteria. Considering fibrinogen as an acute phase reactant it seems to be clear getting elevated in patients with infection and inflammatory disease. A plasma level of fibrinogen that is not adjusted to this acute phase reaction, as in the ISTH criteria, cannot be considered as a useful marker of DIC. Instead the authors applied a new predictor and so they calculated the ratio of fibrinogen to another acute phase reactant, the C-reactive protein. Kim et al succeeded in demonstrating that this new fibrinogen/C-reactive protein (CRP) ratio is able to improve the sensitivity of the DIC score for identifying those patients with active coagulation factor consumption. In addition, they could also show the ratio′s prognostic use in regard of the 28-day mortality in the patients suspected of having DIC. Thus, Kim et al suggested replacing the parameter fibrinogen by their new fibrinogen/CRP ratio to gain both, more diagnostic and prognostic power for DIC.
The authors’ evaluation of the new fibrinogen/CRP ratio was among patients, who developed a DIC because of the following underlying diseases : sepsis/infection, trauma, organ destruction, solid tumor, hematologic malignancy, and vascular abnormalities. So the question arose whether the findings of Kim et al regarding the new fibrinogen/CRP ratio, could also be applicable in cases of HELLP syndrome.
The primary goal was to analyze the function of the fibrinogen/CRP ratio within obstetric patients, who developed a DIC because of an underlying HELLP syndrome. In detail, the fibrinogen/CRP ratio was compared with classical coagulation parameters for the purpose of a more reliable prediction of a DIC. Furthermore, the ratio′s predictive use regarding the neonatal outcome was of high interest and therefore the secondary goal of the study was if there is an association between the new fibrinogen/CRP ratio and the neonatal outcome.
Materials and Methods
Subjects
The medical records of pregnant women admitted for delivery at the Vienna General Hospital between Jan. 1, 2003, and Dec. 31, 2009, were carefully reviewed retrospectively if their pregnancies were complicated by a HELLP syndrome. The syndrome was determined antenatally by the presence of all 3 of the following criteria :
- –
hemolysis (characteristic appearance of peripheral blood smear and serum lactate dehydrogenase [LDH] level ≥600 U/L or serum total bilirubin level ≥1.2 mg/dL)
- –
elevated liver enzymes (serum aspartate aminotransferase [AST] concentration ≥70 U/L)
- –
low platelet count (<100.000 cells/μL)
Design
In regard to a DIC, these 111 pregnant women were divided into 2 groups. In a first step, the diagnostic criteria of DIC proposed by the Scientific Subcommittee on DIC of the ISTH, as shown in Table 1 , were used. Within this scoring system the parameters PT, platelet count, fibrinogen, and D-dimer are considered and evaluated with points between 0 and 3. A total count under 5 points represents a no overt DIC, whereas a score equal and higher than 5 refers to an overt 1. The normal reference ranges for the parameters fibrinogen and platelet count were given ( Table 1 ). The PT was declared as the International Normalized Ratio (INR), for which, according to the guidelines of the laboratory of the General Hospital of Vienna, a normal reference range of 75–140% was assumed. An INR between 74-50% was regarded as moderate decrease and an INR fewer than 49% as a severe reduction. In regard to Kim et al’s a D-dimer level <0.4 μg/mL was considered as no increase, a level between 0.4 and 4.0 μg/mL was considered as a moderate increase and a level >4.0 μg/mL was considered as a strong increase. However, recent studies showed that this DIC scoring system is insufficiently accurate to make or to reject a diagnosis of DIC.
| Variable | Reference ranges | Points |
|---|---|---|
| PT/INR | PT normal | 0 |
| PT prolonged 3–6 sec | 1 | |
| PT prolonged >6 sec | 2 | |
| Platelet count | >100/nL | 0 |
| 50–100/nL | 1 | |
| <50/nL | 2 | |
| D-dimer | Normal | 0 |
| Slightly elevated | 2 | |
| Strongly elevated | 3 | |
| Fibrinogen | >1 g/L | 0 |
| <1 g/L | 1 |
Consequently, the classification of the 111 patients in regard to a DIC was done with another criterion in a second step. They were separated into the 2 groups according to their D-dimer levels. It is known that physiologically this parameter can increase slightly during pregnancy and therefore only levels above 4.0 μg/mL were considered to indicate an overt DIC. In the following, for each woman, test results for DIC diagnosis from the hospital database were collected regarding the following classical parameters: blood platelet count (G/L), fibrinogen (mg/dL), antithrombin (%), thrombine time (sec), activated partial thromboplastin time (aPTT) (sec), and PT (sec). Peripheral blood was drawn into commercially available vacutainers and the tests were performed within 2 hours after the blood sampling.
In addition, the CRP was evaluated to calculate the new parameter fibrinogen/CRP ratio. Kim et al proposed that the optimal cutoff point of the ratio is 104.
For the analysis if the fibrinogen/CRP ratio can also be considered as parameter with prognostic power, the adverse neonatal outcome of the 111 patients was evaluated. Therefore, the following variables were chosen: Apgar scores (at 1, 5, and 10 minutes), umbilical cord artery pH, stillbirth, small for gestational age (SGA), transfer to the neonatal intensive care unit (NICU), and birthweight.
Statistical hypothesis
Regarding the primary goal of the study whether there are significant differences between the groups regarding the classical coagulation parameters and the fibrinogen/CRP ratio, the following hypotheses could be established. Because of the fact that the ratio got tested in an obstetric population for the first time, null hypotheses (H0) were assumed (μ 1 = μ 2 ). H0.1: patients with HELLP syndrome and overt DIC (according to the ISTH criteria) do not differ from patients with HELLP syndrome without overt DIC regarding the classical coagulation parameters (platelet count, fibrinogen, antithrombin, thrombin time, aPTT, PT). H0.2: patients with HELLP syndrome and DIC (according to D-dimer levels) do not differ from patients with HELLP syndrome without DIC regarding the classical coagulation parameters (platelet count, fibrinogen, antithrombin, thrombin time, aPTT, PT). H0.3: patients with HELLP syndrome and overt DIC (according to ISTH criteria) do not differ from patients with HELLP syndrome without overt DIC regarding the fibrinogen/CRP ratio. H0.4: patients with HELLP syndrome and DIC (according to D-dimer levels) do not differ from patients with HELLP syndrome without DIC regarding the fibrinogen/CRP ratio.
According to the secondary goal whether the fibrinogen/CRP ratio has also prognostic power regarding the adverse neonatal outcome, the following hypothesis could be created. However, based on Kim et al’s significant correlations were assumed. Thus alternative hypotheses were formulated (ρ ≠ 0):
There is a correlation between a low fibrinogen/CRP ratio and (H1.5) a low Apgar score, (H1.6) the umbilical cord artery pH, (H1.7) a low birthweight, (H1.8) the number of SGAs, (H1.9) the number of stillbirth, (H1.10) the transfers to the NICU.
Statistical analyses
Data were stored and analyzed using IBM SPSS (Statistical Package for Social Science release 19.0; SPSS, Inc, Cary, NC) in an IBM-compatible computer and are given as median values and ranges.
To analyze differences between the 2 groups regarding the classical parameters as well as the fibrinogen/CRP ratio Mann-Whitney U tests for independent samples were carried out. The receiver-operating characteristics (ROC) analysis was performed to predict the prognostic power of the variables fibrinogen, CRP, and fibrinogen/CRP ratio. In addition, the odds ratios (OR) were estimated using logistic regression analysis and the 95% confidence intervals were computed. For the secondary goal of this study, whether there is an association between the ratio and the neonatal outcome, Spearman’s tests as well as Mann-Whitney U tests were used. For all statistical tests, the level of significance was set at a 2-sided P < .05.
Considerations on the power of the Wilcoxon 2-sample test for an overall sample size of n = 111: As the asymptotic relative efficiency of the 2-sample Wilcoxon test to the 2-sample t test is 0.95, we assume an effective sample size of n = 105 if Wilcoxon’s test is used rather than a t test on approximately normally distributed data. Then a power of 0.85 will be achieved when testing an underlying difference in means of Cohen’s d = 0.6 (assuming a 2-sided alpha of .05). Thus, the given sample size of n = 111 should be sufficient to detect a shift of distributions in the magnitude of d = 0.6, located between a medium (d = 0.05) and a large (d = 0.8) effect according to Cohen. Power analysis was done by means of stplan (version 4.5, 2010, MDACC; Department of Biomathematics, Houston, TX).
Results
Descriptive statistics
In the case that the ISTH criteria were used, overt DIC could be diagnosed in 11.7% of the 111 pregnancies. Median DIC score of all the 111 patients was 3.00 (2-6). According to the classical coagulation parameters the levels of platelet, fibrinogen, antithrombin, and PT were lower in patients with overt DIC (n = 13) than in patients without overt DIC (n = 98). In contrast the thrombin time and the aPTT were increased in the patients with overt DIC ( Table 2 ). Based on the D-dimer, the patients with levels >4.0 μg/mL (n = 40) had lower platelet counts, fibrinogen, antithrombin, PT, and higher aPTT and thrombin time than those with levels ≤4.0 μg/mL (n = 38) ( Table 3 ).
| Variable | Reference category | All patients (n = 111) | No overt DIC (n = 98) | Overt DIC (n = 13) |
|---|---|---|---|---|
| Platelet count | 150-350 G/L | 99.00 (20-275) | 104.5 (28-275) | 41 (20-252) |
| Fibrinogen | 180-390 mg/dL | 449.00 (14.80-747) | 459.5 (14.80-747) | 419 (78-647) |
| Antithrombin | 70-120% | 79 (10-551) | 80 (33-551) | 65 (10-87) |
| Thrombin time | 14 sec | 14.4 (12.10-21.90) | 14.35 (12.10-21.90) | 14.60 (12.10-21.60) |
| aPTT | 28-40 sec | 33.50 (20.90-64) | 33.40 (20.90-64) | 36.30 (30.80-46.60) |
| PT/INR | 75-140% | 133 (29-150) | 135 (51-150) | 124 (29-150) |
| Variable | D-dimer ≤4 μg/mL (n = 38) | D-dimer >4 μg/mL (n = 40) |
|---|---|---|
| Platelet count | 115 (46-268) | 67.50 (20-275) |
| Fibrinogen | 456 (170-747) | 427.00 (78-686) |
| Antithrombin | 77 (41-120) | 70.00 (10-142) |
| Thrombin time | 14.20 (13.10-18.40) | 14.65 (12.10-21.90) |
| aPTT | 33.30 (20.90-64) | 34.30 (28.80-48.90) |
| PT/INR | 137.50 (82-150) | 127.50 (29-150) |
In both cases the fibrinogen/CRP ratio was markedly lower in patients with overt DIC (70.51, 44.73-4446.15) and with D-dimer levels >4 μg/mL (132.86, 41.92-4446.15) than in those with no overt DIC (292.51, 22.65-5410) and D-dimer levels ≤4 μg/mL (396.00, 61.16-4757.14).
Among the 111 patients with HELLP syndrome, the adverse neonatal outcome was associated with stillbirths (n = 96, 1.8%), SGAs (n = 111, 14.4%), and transfers to the NICU (n = 111, 45.9%). In addition, the parameters birthweight (n = 107), Apgar score (n = 96), and umbilical cord artery pH (n = 93) were evaluated and shown in Table 4 .
| Variable | Reference category | All patients (n = 111) | No overt DIC (n = 98) | Overt DIC (n = 13) |
|---|---|---|---|---|
| Umbilical cord artery pH | >7.3 | 7.26 (6.97-7.35) | 7.26 (6.97-7.35) | 7.28 (7.24-7.32) |
| Birthweight | 3500-3700g | 1948 (445-3920) | 1940 (445-3920) | 2230 (1065-3330) |
| Apgar score | ||||
| 1 min | 8 (0-10) | 8 (0-10) | 8 (0-9) | |
| 5 min | 9-10 | 9 (0-10) | 9 (0-10) | 9 (0-10) |
| 10 min | 10 (0-10) | 10 (0-10) | 9 (0-10) |
Analytical statistic
Fibrinogen/CRP ratio is a sensitive parameter for overt DIC
For the comparison of the classical parameters between the groups overt DIC and no overt DIC based on the ISTH criteria, Mann-Whitney U tests were carried out. This nonparametric method was chosen due to the fact that the variables showed no normal contribution (Kolmogorov-Smirnov <0.05). The mean ranks for the coagulation parameters are shown in Table 5 . Statistically significant differences between the 2 groups could be found according to platelet count, antithrombin, and aPTT. However, the differences regarding the parameters fibrinogen, thrombin time, and PT turned out to be not significant. According to D-dimer levels, the groups showed statistically significant only according to platelet count. For all the other classical coagulation parameters no significant differences could be found ( Table 6 ).
