D-Dimer: Predictor of Postpartum Hemorrhage among Pre-Eclampsia at Kilimanjaro Christian Medical Centre ()
1. Introduction
Postpartum hemorrhage is responsible for around 25% of maternal deaths worldwide, reaching 60% in developing countries [1], and is the major leading cause of maternal deaths in Tanzania [2]. Pre-eclampsia is seen to be a second major contributor to maternal mortality reaching 14% worldwide, 12.9% in developed regions, and 16% in sub-Saharan Africa [3]. Women with pre-eclampsia are more likely to suffer postpartum hemorrhage compared to non-pre-eclamptic women. A study done in Australia showed 7.4% of women with pre-eclampsia developed postpartum hemorrhage [4] while that in Pakistan was 27.6% [5], 22% in Kenya [6], and 24% in Songea Tanzania [7].
Postpartum hemorrhage was found to be associated with pregnancy complications due to coagulation defects caused by Pre-eclampsia, such complications are ICU admissions, Shock, Acute kidney injury, and loss of fertility due to hysterectomy [8]. D-dimer tends to rise in women with pre-eclampsia [9], this is due to exacerbation of the hypercoagulable state in pre-eclampsia, platelet activation, and overproduction of thrombin [10]. D-dimer is the ultimate degradation product of a fibrin clot cross-linked by factor XIII [11].
Few studies that have shown the association between D-dimer and Postpartum hemorrhage reported its value is controversial and differs according to ethnicity and lifestyle [11] [12]. Studies have shown higher odds of association between D-dimer and PPH, with an elevated D-dimer level of (≥2.02-µg/ml) among women with pre-eclampsia with severe features in developing postpartum hemorrhage [11]. The cut-off D-dimer value of 1.555 μg/ml has shown a sensitivity of 94.1% and a specificity of 58.6% in predicting PPH among women with pre-eclampsia with severe features [11].
Several studies have used fibrinogen as a marker for a coagulable state in predicting postpartum hemorrhage, which is not available in our setting [13] [14] [15]. From this study, we assessed the association, sensitivity, and specificity of D-dimer as a laboratory predictor of postpartum hemorrhage among women with pre-eclampsia at KCMC hospital.
2. Methods
2.1. Study Design and Area
This was a hospital-based analytical cross-sectional study conducted at Kilimanjaro Christian Medical Center (KCMC) from September 2022 to March 2023. KCMC is a tertiary referral hospital located in the Kilimanjaro region in the Northern part of Tanzania, serving patients from the Kilimanjaro region and its neighboring regions in the Northern zone such as Arusha, Manyara, Tanga, and even other parts of the United Republic of Tanzania and sometimes serves people from neighboring countries like Kenya. Kilimanjaro is one of the regions in the Northern part of Tanzania comprised of seven districts, covering 13250 Km2. The region was projected by the year 2020 to have an estimated population of 1,951,252 people and an annual growth rate of 2.3% [16].
2.2. Study Population
This study included pre-eclamptic women who were in their third trimester, either in the latent phase of labor or those who were prepared for cesarean section and consented to participate in the study, and women with deep venous thrombosis and pulmonary embolism were excluded. A sample size of 195 participants was enrolled in the study by using the Keshlies formula with a prevalence of 10% proportion of PPH attributed to pre-eclampsia [17].
2.3. Study Settings
Identification of women with pre-eclampsia was performed in the obstetrics and gynecology ward. Consent from the patients was obtained and blood samples for plasma D-dimer were taken during the latent phase of labor or when prepared for cesarean section. Clinical data such as Age, Area of residency, Marital status, Gravidity, Gestation age, Body mass index, the severity of pre-eclampsia, number of fetuses, number of previous scars, blood pressure before and after delivery, hemoglobin, platelet levels before delivery, Hematocrit levels before delivery, birth weight, the score of the baby, estimated blood loss, and other causes of PPH, if occurred such as retained placenta, perineal and cervical tear, abruption placenta and placenta previa, was extracted from the patient’s file in the Electronic file system. Furthermore, 12 - 18 hours post vaginal delivery or Cesarean section (C/S), blood for hematocrit was taken again. Hematocrit results were compared pre and post-vaginal delivery or C/S. A fall of hematocrit by 10% was considered as post-partum hemorrhage (PPH) together with clinical assessment of the patient [18].
2.4. Statistical Methods
Descriptive statistics were summarized using medians (Inter quartile range) for continuous variables, and categorical variables were summarized by frequency and proportions and presented by table and narrations. A proportion of post-partum hemorrhage was computed as a 10% change in hematocrit before and after delivery. Normality for continuous variables was checked by plotting histograms and assessing skewness. The Mann-Whitney U-test compared the median d-dimer level between post-partum categories.
Classical logistic regression was used to determine the association between d-dimer level and post-partum hemorrhage adjusting for other factors. Crude and Adjusted Odds Ratio (OR) with their respective 95% confidence interval were reported to estimate the magnitude of the association. Furthermore, the receiver operating characteristic curve (ROC) was taken to evaluate their predictive values in pre-eclampsia complicated with postpartum hemorrhage. A statistical significance was considered for a p-value of less than 0.05.
Cleaned datasets were analyzed using Stata version 15.0 (Stata Corp, College Station, TX).
2.5. Ethical Consideration
Ethical approval to conduct the study was obtained from KCMU College and Research Ethics Review Committee with clearance no; PG 148/2022. Permission was also obtained from the department’s Head of Obstetrics and gynecology. Written and informed consent was obtained from study participants, with prior information about the study procedure before they decided to participate in the study. Participants were provided with identity numbers to preserve anonymity as one of the methods of maintaining confidentiality. Women who sustained postpartum hemorrhage were managed according to postpartum hemorrhage protocol.
3. Results
3.1. The Enrolment of Study Participants
195 women with pre-eclampsia in their third trimester who signed the consent form were enrolled in the study. Participants were divided among pre-eclampsia with severe features and pre-eclampsia without severe features. Among those two groups were further divided into those who had PPH and those who did not have PPH (Figure 1).
Figure 1. The enrollment of participants and distribution of cases among pre-eclampsia at KCMC.
3.2. Social Demographic and Obstetrics Characteristics
A total of 195 women were enrolled in this study. More than half of the participants 134 (68.7%) were aged between 20 - 34 years and 22.1% between the age of 35 - 39 years, with a mean age of 29.9 (±6.07). The majority of women 149 (76.4%) were either married or cohabiting and only 5 (2.6%) were either separated or divorced. Most of the participants 128 (65.6%) delivered by Cesarean section and more than half were referrals 114 (58.5%). On the other hand, there was almost an equal distribution between women who had Pre-eclampsia without severe feature 100 (51.3%) and Pre-eclampsia with severe feature 95 (48.7%) (Table 1).
3.3. The Median D-Dimer Level among Women with/without Severe Features of Pre-Eclampsia at KCMC
The median D-dimer level among all pre-eclamptic women was 0.623 µg/ml with an interquartile range of 0.85 µg/ml to 1.35 µg/ml. The median D-dimer level in pre-eclamptic women with severe features was 0.75 µg/ml which was higher compared to that in pre-eclamptic women without severe features which was 0.5505 µg/ml with a significant p-value of 0.011 (Figure 2).
3.4. The Median D-Dimer Level in Postpartum Hemorrhage among Pre-Eclampsia at KCMC
There was a high median D-dimer level of 1.13 µg/ml in women who had PPH compared to the median D-dimer level in women without PPH 0.591 µg/ml, among pre-eclampsia with a significant p-value of 0.001 (Figure 3).
3.5. The Association of D-Dimer and Postpartum Hemorrhage among Women with Pre-Eclampsia at KCMC
3.5.1. Association of D-Dimer among Participants Who Had Pre-Eclampsia with Severe Features against PPH
For every unit increase of µg/ml in the d-dimer level, there is a 47% significant
Table 1. Sociodemographic and obstetric characteristics of women who delivered at KCMC with pre-eclapsia (n = 195).
Characteristics |
N |
% |
Age group (years) |
|
|
15 - 19 |
9 |
4.6 |
20 - 34 |
134 |
68.7 |
35 - 39 |
43 |
22.1 |
40+ |
9 |
4.6 |
Mean (SD) |
29.9 ± 6.07 |
|
Residence |
|
|
Arusha |
9 |
4.7 |
Manyara |
2 |
1.0 |
Moshi |
182 |
93.3 |
Tanga |
1 |
0.5 |
Mwanza |
1 |
0.5 |
Referrals |
|
|
Referrals |
114 |
58.5 |
Not referrals |
81 |
41.5 |
Marital Status |
|
|
Married/Cohabiting |
149 |
76.4 |
Separated/Divorced |
5 |
2.6 |
Single |
41 |
21 |
Mode of delivery |
|
|
SVD |
67 |
34.4 |
CS |
128 |
65.5 |
Gestation age at delivery (weeks + days) |
|
|
Early pre-term (28 - 32) |
27 |
13.9 |
Late pre-term (32 + 6 - 36 + 6) |
81 |
41.5 |
Early term (37- 38 + 6) |
54 |
44.6 |
Late term (39 - 41 + 6) |
33 |
16.9 |
Gravidity |
|
|
1 |
59 |
30.3 |
2 - 4 |
109 |
55.9 |
5+ |
27 |
13.8 |
Parity |
|
|
1 |
67 |
34.4 |
2 - 4 |
111 |
56.9 |
5+ |
17 |
8.7 |
Pre-eclampsia |
|
|
Pre-eclampsia without severe features |
100 |
51.3 |
Pre-eclampsia with severe features |
95 |
48.7 |
Number of fetuses |
|
|
Single |
186 |
95.4 |
Twin |
9 |
4.6 |
Number of previous scars |
|
|
None |
138 |
70.8 |
One |
32 |
16.4 |
Two or more |
25 |
12.8 |
Birth weight (grams) |
|
|
<2500 |
93 |
47.7 |
2500 - 3999 |
100 |
51.3 |
>4000 |
2 |
1.03 |
Median (IQR) |
2500 (1800 - 3000) |
|
![]()
Figure 2. The Median D-dimer level among women with/without severe features of pre-eclampsia at KCMC.
Figure 3. The median D-dimer level in PPH among pre-eclampsia at KCMC.
increase in the odds of having postpartum hemorrhage among women with pre-eclampsia with severe features (cOR: 1.47; 95% CI: 1.19 - 1.83). After controlling for confounders there was a 45% significant increase in the odds of having postpartum hemorrhage. (aOR: 1.43; 95% CI: 1.11 - 1.84) (Table 2).
Table 2. Bivariate and Multivariable logistic regression showing an association between D-dimer and PPH among pre-eclampsia at KCMC.
Characteristics |
COR (95% CI) |
p-Value |
AOR (95% CI) |
p-Value |
With severe features (n = 95) |
|
|
|
D-dimer level |
1.47 (1.19 - 1.83) |
<0.001 |
1.45 (1.09 - 1.92) |
0.011 |
Gestation age (in weeks) |
0.80 (0.64 - 1.00) |
0.055 |
0.79 (0.59 - 1.05) |
0.112 |
Maternal Age (in years) |
1.04 (0.96 - 1.14) |
0.322 |
1.10 (0.94 - 1.29) |
0.224 |
Parity |
1.19 (0.78 - 1.81) |
0.415 |
0.81 (0.36 - 1.79) |
0.601 |
Number of fetuses |
|
|
|
|
Singleton |
1.00 |
|
|
|
Twin |
3.68(0.30 - 14.62) |
0.306 |
31.89 (1.27 - 798) |
0.035 |
Previous scars |
0.53 (0.11 - 2.50) |
0.426 |
0.84 (0.15 - 4.69) |
0.842 |
Mode of delivery |
|
|
|
|
SVD |
1.00 |
|
1.00 |
|
CS |
1.63 (0.32 - 8.29) |
0.558 |
3.83 (0.57 - 25.71) |
0.168 |
Other factors contributing to PPH (previa, abruption, tears, retained products) |
No |
1.00 |
|
1.00 |
|
Yes |
27.3 (2.53 - 294.76) |
0.006 |
25.73 (3.16 - 209.33) |
0.002 |
Without severe features (n = 100) |
|
|
|
D-dimer lever |
1.13 (0.91 - 1.41) |
0.277 |
1.14 (0.68 - 1.92) |
0.612 |
Gestation age (in weeks) |
0.95 (0.84 - 1.07) |
0.412 |
0.92 (0.74 - 1.14) |
0.426 |
Maternal Age (in years) |
0.96 (0.88 - 1.05) |
0.403 |
1.04 (0.89 - 1.23) |
0.562 |
Parity |
0.65 (0.40 - 1.06) |
0.084 |
0.35 (0.14 - 0.89) |
0.028 |
Number of fetuses |
|
|
|
|
Singleton |
1.00 |
|
|
|
Twin |
11.73 (1.57 - 87.88) |
0.017 |
27.36 (1.26 - 592.08) |
0.035 |
Previous scars |
0.70 (0.29 - 1.68) |
0.425 |
1.02 (0.92 - 1.14) |
0.671 |
Mode of delivery |
|
|
|
|
SVD |
1.00 |
|
1.00 |
|
CS |
1.67 (0.30 - 9.14) |
0.556 |
0.80 (0.12 - 5.59) |
0.826 |
Other factors contributing to PPH (previa, abruption, tears, retained products) |
No |
1.00 |
|
1.00 |
|
Yes |
30.33 (2.37 - 389.0) |
0.009 |
41.97 (4.41 - 753.53) |
0.005 |
Key: OR: Odds ratio, CI: Confidence Interval, SVD: Spontaneous Vaginal delivery. AOR was adjusted for age, parity, number of fetuses, previous scar, and mode of delivery.
3.5.2. Association of D-Dimer among Participants Who Had Pre-Eclampsia without Severe Features against PPH
On average, for every unit increase of µg/ml in the d-dimer level, there is 13% increase in odds of having postpartum hemorrhage among women who had pre-eclampsia without severe features (cOR: 1.13; 95% CI: 0.91 - 1.41) and after controlling for confounders there was a 14% significant increase in the odds of having postpartum hemorrhage among women who had pre-eclampsia without severe features (aOR: 1.14; 95% CI: 0.68 - 1.92) (Table 2).
3.6. D-Dimer Cut-Off Point in Predicting Postpartum Hemorrhage among Pre-Eclampsia at KCMC
The cutoff point of the D-dimer level in predicting PPH among women with pre-eclampsia was 0.66 µg/ml with sensitivity of 75% and specificity of 55%.
Pre-eclamptic women with severe features: Cuttoff point was 3.58 µg/ml, sensitivity 50% and specificity of 96%, AUC 0.73, 95% CI: 0.51 - 0.89.
For women who had pre-eclampsia without severe features, the cutoff point was 0.53 µg/ml with a sensitivity of 95% and specificity of 53%, AUC 0.77, 95% CI: 0.61 - 0.87 (Table 3), (Figure 4).
Table 3. ROC curve of D-dimer for evaluation with PPH.
Variable |
Cutoff point |
Sensitivity |
Specificity |
AUC |
95% CI |
Lower |
Upper |
D-dimer |
0.66 |
75.0% |
55.0% |
0.72 |
0.60 |
0.84 |
Key: AUC: Area under the curve. CI: Confidence Interval.
Figure 4. ROC curve for the sensitivity and specificity of the cut-off point of the d-dimer level.
4. Discussion
Findings from this study there was a higher D-dimer level of 0.75 µg/ml among pre-eclampsia with severe features than pre-eclampsia without severe features 0.55 µg/ml. Also, there was a highly significant median D-dimer level of 1.13 µg/ml in women with PPH compared to the median D-dimer level in those who had no PPH of 0.591 µg/ml. A true association was seen after adjusting for other factors between D-dimer and PPH. That is for every unit increase of µg/ml of D-dimer level among pre-eclamptic women with severe features, there was a significant increase of 45% in the odds of having PPH, and a significant increase of 14% among pre-eclampsia without severe features. Furthermore, the cut-off point of a D-dimer level of 0.66 µg/ml significantly predicts PPH with a sensitivity of 75% and specificity of 55% and for those who had severe features cutoff point was 3.58 µg/ml predicts PPH with sensitivity of 50% and specificity of 96%. For those without severe features, the cutoff point was 0.53 µg/ml, sensitivity of 95%, and specificity of 53%.
Median D-dimer levels that were seen in women who had PPH were higher at 1.13 µg/ml compared to the median D-dimer level in those who had no PPH at 0.591 µg/ml among pre-eclamptic women with/without severe features (Figure 3). Shao reported that D-dimer levels were elevated in the PPH group with severe features of pre-eclampsia (2.02 µg/ml) compared to those who had no PPH (1.37 µg/ml) [11]. High values of D-dimer levels in our study seen in the PPH group compared to the non-PPH group correlated significantly to Shao’s study, but the values of D-dimer levels in our study were seen to be lower than Shao’s. This may be due to the selected group in Shao’s study being more sick/severe cases only while the selected group in our study were both with and without severe features of pre-eclampsia. This was supported also by Figure 2 which showed a higher D-dimer level of 0.75 µg/ml among pre-eclampsia with severe features than pre-eclampsia without severe features 0.55 µg/ml due to aggressiveness in the severe form of pre-eclampsia. This high value of D-dimer above 1 explains the hypercoagulable state in pre-eclampsia and more aggressive hyperconsumption of fibrin hence high D-dimer is expected to be seen in severe form. Thus these women are more prone to bleeding. This indicates that D-dimer might be associated with postpartum hemorrhage among pre-eclamptic women. A range above 1 should be used as a value for diagnosis, as few studies showed this value in pre-eclampsia as a predictor for PPH.
After adjusting for other factors there was an association seen between D-dimer and postpartum hemorrhage among pregnant women with pre-eclampsia in both groups, i.e.; those who had severe features and those who had no severe features. These factors are maternal age, gestational age, parity, number of fetuses, previous scar, mode of delivery, and other factors contributing to postpartum hemorrhage such as abruption placenta, placenta previa, retained products and cervical tear. Thus for every unit increase of µg/ml in the d-dimer level, there was a 14% significant increase in the odds of having postpartum hemorrhage among women who had pre-eclampsia without severe features (aOR: 1.14; 95% CI: 0.68 - 1.92), and for every unit increase of µg/ml in the d-dimer level, there was a 45% significant increase in the odds of having postpartum hemorrhage among women with pre-eclampsia with severe features (aOR: 1.45; 95% CI: 1.09 - 1.92) (Table 2). These findings support Shao’s study as D-dimer significantly associated with PPH among women who had pre-eclampsia with severe features [11]. This indicated that in pre-eclampsia, there is hyper-coagulopathy thus the formation of fibrin clots and hence D-dimer formation. This shows that D-dimer can be used as a predictive tool, as there is an association seen between D-dimer and Postpartum hemorrhage among pre-eclampsia. Higher odds of 1.45 of D-dimer in sustaining PPH among those who have severe features and 1.14 higher odds of D-dimer in sustaining PPH among those with pre-eclampsia without severe features.
The cut-off point of D-dimer in predicting postpartum hemorrhage was 0.66 µg/ml which had a sensitivity of 75% and specificity of 55% with a narrow confidence interval (Table 3). The area under the curve (AUC) of 0.72 shows a good prediction ability to discriminate between individuals with and those without postpartum hemorrhage (Figure 4). This means D-dimer has 72% predictability against postpartum hemorrhage. Hence patients who are at risk of developing postpartum hemorrhage can be predicted early. A good prediction was also seen among pre-eclamptic women without severe features with a D-dimer cut-off point of 0.53 µg/ml together with a sensitivity of 95% and specificity of 53%. The area under the curve was 0.77, which showed a 77% good prediction ability of D-dimer against postpartum hemorrhage among pre-eclamptic women without severe features. Those who had Pre-eclampsia with severe features had a higher cut-off point of 3.58 µg/ml with a sensitivity of 50% and specificity of 96%. This shows that for those who had severe features the cutoff point is more specific with 73% predictability against postpartum hemorrhage. Shao et al. reported that the cut-off D-dimer value was 1.555 μg/ml in predicting PPH with a sensitivity of 94.1% and a specificity of 58.6% among women with pre-eclampsia with severe features [11]. The difference might be due to disparities in testing methods, sample size and differences in race. This shows that D-dimer is more sensitive and more specific in predicting PPH in women with pre-eclampsia especially those with severe features and can be applied in clinical services to save maternal morbidity and mortality. The cut-off point of above 1 can be used as a predictive value as is sensitive in picking women who are at risk in sustaining postpartum hemorrhage as the value is supported by our study and other studies as well. Thus D-dimer can be used as a screening value in women who have pre-eclampsia with severe features.
5. Strength
This study used Receiver operating characteristics (ROC) in predicting the cut-off point which provided us with a precise and valid estimate that can be used for clinical implementation.
Also, this study specifically looked into high-risk groups, and an association of factors was done to get the true value. Thirdly, this is one of the first few studies that tried to predict the association between d-dimer level and postpartum hemorrhage in Africa. Lastly, the Estimation of PPH was by checking hematocrit thus, it did not affect those who delivered by c/s as they received fluid post-delivery.
6. Limitation
This study did not check for the D-dimer levels in the first and second trimesters. Also, this study design did not allow checking for the D-dimer levels post-delivery for follow-up. Lastly, there are limited studies that could be used to compare with our study in the discussion part.
7. Conclusion
D-dimer can be used to predict postpartum hemorrhage among pre-eclampsia, especially for those who have severe features. This shows that D-dimer has high specificity in predicting PPH among women who have pre-eclmpsia with severe features and can be applied in clinical services to save women’s maternal morbidity and mortality. Blood products such as fresh frozen plasma, platelets, and whole blood together with tranexamic acid should be readily available for women with pre-eclampsia especially those with severe features with a D-dimer level of 3.58 µg/ml and above during delivery. We wish to generalize the results as it is one hospital with one pathological disease ie pre-eclampsia, disease physiology and laboratory maker seem to apply to other patients.
8. Recommendations
More studies with a longer follow-up period and larger sample size should be considered to explain or confirm our findings.
Women with pre-eclampsia should be checked for D-dimer levels antenatally, those who tend to have a rise in D-dimer values should be prepared early as they approach term, as D-dimer increases with gestation age [19] [20] [21] hence they are at high risk of developing postpartum hemorrhage.
Blood products such as fresh frozen plasma, platelets, and whole blood together with tranexamic acid should be readily available for women with pre-eclampsia with a raised D-dimer level especially those who had pre-eclampsia with severe features having 3.58 µg/ml and above during delivery.
Consent for Publication
Not applicable.
Authors’ Contributions
All authors contributed to the discussion, conclusion, and review of this work. However, PSS and MJM advised on the analysis and statistic part. All authors contributed to this study’s design, drafting, and manuscript preparation.
Acknowledgements
We would like to thank our Almighty God, families, friends, and all study participants who participated and contributed to the accomplishment of this study.
Abbreviations
C/SCesarean Section
DICDisseminated Intravascular Coagulopathy
GAGestation Age
ICUIntensive Care Unit
KCMCKilimanjaro Christian Medical Center
OBGYNObstetrics and Gynecology
PPHPostpartum Hemorrhage
SVDSpontaneous Vaginal Delivery
WHOWorld Health Organization