Abstract
Objective
Maternal hemorrhage is a major contributor to pregnancy morbidity and mortality worldwide. Identifying risk factors for hemorrhage is critical to optimizing patient safety and preventing patient morbidity. Prior studies examining the potential association between mild thrombocytopenia and the risk of postpartum hemorrhage have had inconsistent findings. In this study, we seek to investigate whether mild thrombocytopenia increases the risk of clinically significant obstetrical hemorrhage in a large, modern cohort.
Methods
This is a retrospective cohort study of patients delivering ≥ 24 weeks singleton gestations within a single hospital system between 2013 and 2023. Data were extracted from the hospital electronic medical record (EMR) data warehouse using standardized definitions and diagnosis codes. Patients were categorized as having normal platelets (≥ 150 × 1000/µL) or mild thrombocytopenia (100–149 × 1000/µL) at delivery admission. The primary outcome was a bleeding composite inclusive of estimated blood loss at delivery ≥1000 mL, International Classification of Diseases (ICD) diagnosis of postpartum hemorrhage, or receipt of a red blood cell transfusion. Secondary outcomes included (1) additional uterotonic use after delivery other than synthetic oxytocin; (2) tranexamic acid use; (3) need for a hemorrhage procedure; (4) intensive care unit admission. Multivariate logistic regression was performed for the two statistically significant outcomes adjusting for potential confounders.
Results
Among 85,047 eligible subjects, 7194 (8.5%) had mild thrombocytopenia noted upon admission for delivery, while 77,853 (91.5%) had a normal platelet count. Patients with mild thrombocytopenia were more likely to be advanced maternal age, multiparous, have preeclampsia with severe features, have a neonate > 4000 g, and have an admission hemoglobin < 11 g/dL. Mild thrombocytopenia was associated with a higher risk of the bleeding composite (7.6% vs. 6.7%, aOR 1.22, 95% CI 1.11–1.34) and uterotonic use (13.4% vs. 11.7%, aOR 1.17, 95% CI 1.09–1.26) compared to normal platelet count group.
Conclusion
Patients with mild thrombocytopenia at the time of delivery hospitalization had increased odds of clinically significant postpartum bleeding. This patient population is not traditionally considered to be at increased risk of clinically significant hemorrhage; however, this variable should be considered in risk assessment tools to better identify patients at risk for hemorrhage.
Keywords: aspirin, bleeding, delivery, delivery admission, hemorrhage, mild thrombocytopenia, postpartum hemorrhage, uterotonic use
1. INTRODUCTION
Maternal hemorrhage is a major contributor to pregnancy morbidity and mortality worldwide [1, 2, 3]. Postpartum hemorrhage is typically defined as a cumulative blood loss of ≥1000 mL at the time of delivery or blood loss that is accompanied by signs or symptoms of hypovolemia within 24 h of birth irrespective of mode of delivery [4]. The recognition and identification of risk factors that may lead to hemorrhage is critical in optimizing patient safety and preventing patient morbidity [5, 6]. There are many postpartum hemorrhage risk factor assessment tools available to help with risk stratification; however, maternal thrombocytopenia is not typically included in these models [6]. While significant thrombocytopenia is associated with increased hemorrhage risk, this association is less clear for a mild decrease in platelet count.
Thrombocytopenia in pregnancy is traditionally defined as a platelet count of < 150 × 1000/µL and is relatively common, affecting approximately 12% of pregnancies [7]. During normal gestation, there is a mild decrease in platelet count attributed to hemodilution and increased clearance [8]. According to a systematic review, the mean platelet count at delivery is 237 × 1000/µL [9]. Prior studies have demonstrated a linear association between decreased platelet count and increased risk of postpartum hemorrhage [10]. The platelet cutoffs that are associated with clinically significant hemorrhage, however, are poorly understood [11]. Mild thrombocytopenia is defined as a platelet count between 100 × 1000 and 149 × 1000/µL according to the American College of Obstetricians and Gynecologists (ACOG). Mild thrombocytopenia in patients without a history of bleeding is usually attributed to gestational thrombocytopenia and is not thought to be a risk factor for clinically significant bleeding [12].
Some studies report a clinically significant increased risk of obstetrical hemorrhage even in patients with mild thrombocytopenia [13, 14, 15, 16, 17]. Many of these studies, however, are small and restricted to very specific patient populations. In this study, we seek to investigate whether mild thrombocytopenia increases the risk of clinically significant obstetrical hemorrhage in a large, contemporary cohort. We hypothesized that patients with mild thrombocytopenia at time of delivery were more likely to experience hemorrhage related complications than patients with a normal platelet count within our hospital system.
2. METHODS
We performed a retrospective cohort study of all patients who delivered a singleton gestation ≥24 weeks within our hospital system (including one tertiary referral hospital and three community hospitals) from October 2013, when our electronic medical record (EMR) was introduced, through January 2023. Using our hospital's EMR data warehouse, we compared bleeding outcomes for patients with mild thrombocytopenia (platelet count 100–149 × 1000/µL) at time of delivery hospitalization to patients with a normal platelet count (≥150 × 1000/µL). Patients were dichotomized based on their very first platelet count upon admission for delivery prior to birth as it is routine in our hospital to have admission bloodwork performed and this is the time when hemorrhage risk assessment is typically first performed. We excluded all patients with a platelet count < 100 × 1000/µL as these patients were more likely to have a significant underlying pathology leading to their thrombocytopenia and are at known increased risk for hemorrhage. Patients with other major risk factors for hemorrhage including multifetal gestations, suspected placenta accreta spectrum, placenta previa, bleeding or clotting disorders, anticoagulation use during pregnancy, and hemoglobinopathies and aplastic anemia were excluded from this analysis. We additionally excluded patients with significant laboratory outliers (hemoglobin < 7 g/dL or > 16 g/dL) and patients who delivered prior to obtaining admission bloodwork.
The primary outcome was a bleeding composite inclusive of an estimated blood loss at delivery ≥1000 mL, International Classification of Diseases (ICD) diagnosis of postpartum hemorrhage (Table S1), or receipt of a red blood cell transfusion during the delivery admission encounter. Of note, during the course of the study, the American College of Obstetricians and Gynecologists (ACOG) changed the definition of postpartum hemorrhage after a vaginal birth to reflect a cumulative blood loss of 1000 mL (as opposed to previous criterium of 500 mL). Since this was a retrospective analysis, we used the more recent definition. Secondary outcomes were (1) additional uterotonic use after delivery other than synthetic oxytocin inclusive of methylergonovine, carboprost tromethamine and/or misoprostol; (2) tranexamic acid use; (3) need for a hemorrhage procedure inclusive of dilation and curettage, hysterectomy, uterine artery embolization, B‐lynch suture, O'Leary suture, and intrauterine balloon; (4) intensive care unit (ICU) admission. Medication administration and surgical procedures were directly abstracted from within the medical record. Procedures were identified using appropriate current procedural terminology (CPT) codes and searches of operative notes for specific keywords and finally confirmed by hand abstraction.
Maternal characteristics, including age, delivery admission, body mass index (BMI), parity, delivery hospital, insurance type, preterm delivery, mode of delivery, neonatal birthweight, and first hemoglobin at delivery admission, were extracted directly from the EMR. Nulliparity was defined as no previous births ≥20 weeks of gestation. Race, ethnicity, marital status, smoking status, and primary language were obtained through patient self‐report as documented within the EMR. Preterm delivery was defined as delivery prior to 37 weeks of gestation. Hypertensive disorder of pregnancy was defined as ICD‐9/ICD‐10 diagnosis of gestational hypertension; preeclampsia with or without severe features; hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome; and eclampsia (Table S1). Chronic hypertension was defined by diagnostic codes (Table S1). Aspirin use and receipt of intravenous iron were based on documentation within the patient's medication list.
Bivariate analyses were performed using chi‐square for categorical variables based on non‐parametric distributions. Unadjusted logistic regression and multivariable logistic regression modeling of the primary outcomes (bleeding composite and uterotonic use) were performed, generating a crude and adjusted odds ratio (cOR and aOR, respectively) and 95% confidence intervals (CIs). Variables that were significant at p < 0.05 for both mild thrombocytopenia and the dependent variable (bleeding composite or additional uterotonic use) or the dependent variable only were used for model adjustment. All analyses were performed using SAS 9.4 (SAS Institute). This study was approved by the Institutional Review Board with waiver of informed consent.
3. RESULTS
Of the 110,690 pregnancies within our health system's research database, 85,047 patients met inclusion criteria. Of these, 7194 (8.5%) had mild thrombocytopenia noted upon admission for their delivery hospitalization, while 77,853 (91.5%) had a normal platelet count.
There were several important differences in baseline characteristics between patients who had mild thrombocytopenia compared to those with normal platelet counts at delivery admission (Table 1). Patients with mild thrombocytopenia were more likely to be of advanced maternal age, be multiparous, have severe preeclampsia, have a neonate > 4000 g, and have a delivery admission hemoglobin < 11 g/dL. Patients with mild thrombocytopenia had a different racial/ethnic profile and were less likely to be unmarried, have a BMI ≥30 kg/m2, nulliparous, smoke during pregnancy, have non‐commercial insurance, undergo induction of labor, have a preterm delivery, and use aspirin during pregnancy. There was no difference in primary language, intravenous iron use, or mode of delivery.
TABLE 1.
Association between mild thrombocytopenia, patient characteristics, and outcomes.
| Mild thrombocytopenia | Normal platelets | ||
|---|---|---|---|
| n = 7194 | n = 77,853 | p value | |
| A. Patient Characteristics | |||
| Age ≥35 years at delivery | 2074 (28.8%) | 20,710 (26.6%) | <0.01 |
| Race/ethnicity | <0.01 | ||
| Hispanic | 1322 (18.4%) | 18,810 (24.2%) | |
| Non‐Hispanic Black | 776 (10.8%) | 10,818 (13.9%) | |
| White, non‐Hispanic | 4112 (57.2%) | 39,030 (50.1%) | |
| Asian, non‐Hispanic | 403 (5.6%) | 4142 (5.3%) | |
| Other, non‐Hispanic | 581 (8.1%) | 5053 (6.5%) | |
| Marital status | <0.01 | ||
| Single/divorced/other | 1715 (23.8%) | 24,980 (32.1%) | |
| Married/living as married | 5479 (76.2%) | 52,873 (67.9%) | |
| Body mass index at delivery admission | <0.01 | ||
| < 30 kg/m2 | 3978 (55.3%) | 33,354 (42.8%) | |
| ≥30 kg/m2 | 3088 (43.7%) | 43,276 (56.5%) | |
| Missing data | 128 (1.8%) | 1223 (1.6%) | |
| Parity | <0.01 | ||
| Nulliparous | 3039 (43.0%) | 34,941 (44.9%) | |
| Multiparous | 4027 (56.0%) | 41,877 (53.8%) | |
| Missing data | 72 (1.0%) | 1035 (1.3%) | |
| Smoking status | <0.01 | ||
| Current smoker/quit in pregnancy | 261 (3.6%) | 4533 (5.8%) | |
| Non‐smoker/former smoker | 6933 (96.4%) | 73,320 (94.2%) | |
| Delivery hospital a | <0.01 | ||
| A | 1440 (20.0%) | 16,989 (21.8%) | |
| B | 2830 (39.3%) | 21,331 (27.4%) | |
| C | 539 (7.5%) | 6394 (8.2%) | |
| D | 2385 (33.2%) | 33,139 (42.6%) | |
| Insurance Status | <0.01 | ||
| Public/other | 2,088 (29.0%) | 29,076 (37.4%) | |
| Commercial | 5106 (71.0%) | 48,777 (62.7%) | |
| Primary language | 0.94 | ||
| Non‐English | 808 (11.2%) | 8766 (11.3%) | |
| English | 6386 (88.8%) | 69,087 (88.7%) | |
| Intravenous iron in pregnancy | 0.16 | ||
| None | 7091(98.6%) | 76,537 (98.3%) | |
| 1–3 doses | 39 (0.5%) | 567 (0.7%) | |
| 4 or more doses | 64 (0.9%) | 749 (1.0%) | |
| Hypertensive disorder of pregnancy b | <0.01 | ||
| No HDP | 6055 (84.2%) | 65,400 (84.0%) | |
| HDP without magnesium | 873 (12.1%) | 10,328 (13.3%) | |
| HDP with magnesium | 266 (3.7%) | 2125 (2.7%) | |
| Induction of labor | 2070 (28.8%) | 23,395 (30.1%) | 0.02 |
| Preterm delivery c | 407 (5.7%) | 5976 (7.7%) | <0.01 |
| Aspirin use d | 364 (5.1%) | 5496 (7.1%) | <0.01 |
| Mode of delivery | 0.09 | ||
| Vaginal | 4939 (68.7%) | 53,419 (68.6%) | |
| Cesarean | 2192 (30.5%) | 23,922 (30.7%) | |
| Missing | 63 (0.9%) | 512 (0.7%) | |
| Birthweight > 4000 g | 801 (11.1%) | 6777 (8.7%) | <0.01 |
| Delivery admission hemoglobin < 11 g/dL | 14,971 (19.2%) | 981 (13.6%) | <0.01 |
| B. Primary Outcomes | |||
| Bleeding composite e | 546 (7.6%) | 5234 (6.7%) | <0.01 |
| C. Secondary outcomes | |||
| Additional uterotonic use f | 967 (13.4%) | 9122 (11.7%) | <0.01 |
| Tranexamic acid use | 55 (0.8%) | 771 (1.0%) | 0.06 |
| Hemorrhage procedure g | 67 (0.9%) | 694 (0.9%) | 0.73 |
| Intensive care unit admission | 14 (0.2%) | 200 (0.3%) | 0.31 |
A, B, C represent community hospitals; D represents tertiary care hospital.
Inclusive of chronic hypertension; gestational hypertension; preeclampsia with or without severe features; and hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome.
Preterm delivery defined as < 37 weeks of gestation.
Aspirin use as documented in patient's medication list within electronic medical record.
Inclusive of estimated blood loss at delivery ≥1000 mL, International Classification of Diseases (ICD) diagnosis of postpartum hemorrhage, and/or receipt of a red blood cell transfusion.
Inclusive of methylergonovine, carboprost tromethamine, and/or misoprostol.
Hemorrhage procedure inclusive of dilation and curettage, hysterectomy, uterine artery embolization, B‐lynch suture, O'Leary suture, intrauterine balloon.
Patient characteristics, significantly associated with the two outcomes of interest (bleeding composite and additonal uterotonic use), were presented in Table 2.
TABLE 2.
Patient characteristics and association with bleeding composite and additional uterotonic use.
| Bleeding composite | Additional uterotonic use | |||||
|---|---|---|---|---|---|---|
| Yes | No | p value | Yes | No | p value | |
| Age ≥35 years at delivery | 1611 (27.9%) | 21,173 (26.7%) | 0.05 | 2708 (26.8%) | 20,076 (26.8%) | 0.9 |
| Race/Ethnicity | < 0.01 | < 0.01 | ||||
| Hispanic | 1704 (29.5%) | 18,428 (23.3%) | 2640 (26.2%) | 17,492 (23.3%) | ||
| Non‐Hispanic Black | 960 (16.6%) | 10,634 (13.4%) | 1270 (12.6%) | 10,324 (13.8%) | ||
| White, non‐Hispanic | 2514 (43.5%) | 40,628 (51.3%) | 4941 (49.0%) | 38,201 (51.0%) | ||
| Asian, non‐Hispanic | 296 (5.1%) | 4249 (5.4%) | 620 (6.2%) | 3925 (5.2%) | ||
| Other, non‐Hispanic | 306 (5.3%) | 5328 (6.7%) | 618 (6.1%) | 5016 (6.7%) | ||
| Marital status | < 0.01 | 0.36 | ||||
| Single/divorced/other | 2106 (36.4%) | 24,589 (31.0%) | 3207 (31.8%) | 23,488 (31.3%) | ||
| Married | 3674 (63.6%) | 54,678 (69.0%) | 6882 (68.2%) | 51,470 (68.7%) | ||
| Body mass index at delivery admission | < 0.01 | < 0.01 | ||||
| < 30 kg/m2 | 2113 (36.6%) | 35,219 (44.4%) | 4055 (40.2%) | 33,277 (44.4%) | ||
| ≥30 kg/m2 | 3569 (61.8%) | 42,795 (54.0%) | 5905 (58.5%) | 40,459 (54.0%) | ||
| Missing | 98 (1.7%) | 1253 (1.6%) | 129 (1.3%) | 1222 (1.6%) | ||
| Parity | < 0.01 | < 0.01 | ||||
| Nulliparous | 2734 (47.3%) | 35,302 (44.5%) | 5221 (51.8%) | 32,815 (43.8%) | ||
| Multiparous | 2983 (51.6%) | 42,921 (54.2%) | 4771 (47.3%) | 41,133 (54.9%) | ||
| Missing data | 63 (1.1%) | 1,044 (1.3%) | 97 (1.0%) | 1010 (1.4%) | ||
| Smoking status | 0.67 | 0.17 | ||||
| Current smoker/quit in pregnancy | 333 (5.8%) | 4,461 (5.6%) | 539 (5.3%) | 4255 (5.7%) | ||
| Non‐smoker/former smoker | 5447 (94.2%) | 74,806 (94.4%) | 9550 (94.7%) | 70,703 (94.3%) | ||
| Delivery hospital a | < 0.01 | < 0.01 | ||||
| A | 1622 (28.1%) | 16,807 (21.2%) | 1353 (13.4%) | 17,076 (22.8%) | ||
| B | 1261 (21.8%) | 22,900 (28.9%) | 3175 (31.5%) | 20,986 (28.0%) | ||
| C | 347 (6.0%) | 6586 (8.3%) | 865 (8.6%) | 6068 (8.1%) | ||
| D | 2,550 (44.1%) | 32,974 (41.6%) | 4696 (46.6%) | 30,828 (41.1%) | ||
| Insurance status | < 0.01 | 0.88 | ||||
| Public/other | 2574 (44.5%) | 28,590 (36.1%) | 3,704 (36.7%) | 27,460 (36.6%) | ||
| Commercial | 3206 (55.5%) | 50,677 (63.9%) | 6385 (63.3%) | 47,498 (63.4%) | ||
| Primary language | < 0.01 | < 0.01 | ||||
| Non‐English | 872 (15.1%) | 8702 (11.0%) | 1295 (12.8%) | 8279 (11.0%) | ||
| English | 4908 (84.9%) | 70,565 (89.0%) | 8794 (87.2%) | 66,679 (89.0%) | ||
| Intravenous iron in pregnancy | < 0.01 | < 0.01 | ||||
| None | 5650 (97.8%) | 77,978 (98.4%) | 9,882 (98.0%) | 73,746 (98.4%) | ||
| 1–3 doses | 56 (1.0%) | 550 (0.7%) | 77 (0.8%) | 529 (0.7%) | ||
| 4 or more doses | 74 (1.3%) | 739 (0.9%) | 130 (1.3%) | 683 (0.9%) | ||
| Hypertensive disorder of pregnancy b | < 0.01 | < 0.01 | ||||
| No HDP | 4361 (75.5%) | 67,094 (84.6%) | 8138 (80.7%) | 63,317 (84.5%) | ||
| HDP without magnesium | 1039 (18.0%) | 10,162 (12.8%) | 1,445 (14.3%) | 9756 (13.0%) | ||
| HDP with magnesium | 380 (6.6%) | 2011 (2.5%) | 506 (5.0%) | 1885 (2.5%) | ||
| Induction of labor | 1937 (33.5%) | 23,528 (29.7%) | < 0.01 | 3923 (38.9%) | 21,542 (28.7%) | < 0.01 |
| Preterm delivery c | 571 (9.9%) | 5,812 (7.3%) | < 0.01 | 759 (7.5%) | 5624 (7.5%) | 0.94 |
| Aspirin use d | 566 (9.8%) | 5,294 (6.7%) | < 0.01 | 887 (8.8%) | 4973 (6.6%) | < 0.01 |
| Mode of delivery | < 0.01 | < 0.01 | ||||
| Vaginal | 2972 (51.4%) | 55,386 (69.9%) | 6319 (62.6%) | 52,039 (69.4%) | ||
| Cesarean | 2764 (47.8%) | 23,350 (29.5%) | 3693 (36.6%) | 22,421 (29.9%) | ||
| Missing data | 44 (0.8%) | 531 (0.7%) | 77 (0.8%) | 498 (0.7%) | ||
| Birthweight > 4000 g | 755 (13.1%) | 6823 (8.6%) | < 0.01 | 1448 (14.4%) | 6130 (8.2%) | < 0.01 |
| Delivery admission hemoglobin < 11 g/dL | 1396 (24.2%) | 14,556 (18.4%) | < 0.01 | 2058 (20.4%) | 13,894 (18.5%) | < 0.01 |
A, B, and C represent community hospitals; D represents tertiary care hospital.
Inclusive of chronic hypertension; gestational hypertension; preeclampsia with or without severe features; and hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome.
Preterm delivery defined as < 37 weeks of gestation.
Aspirin use as documented in patient's medication list within electronic medical record.
The bleeding composite outcome occurred in 546 (7.6%) patients with mild thrombocytopenia and 5234 (6.7%) patients with normal platelet counts (cOR 1.14, 95% CI 1.04–1.25). This increased odds of experiencing the bleeding composite remained significant after adjusting for potential confounders including race/ethnicity, marital status, BMI, parity, delivery hospital, insurance status, primary language, intravenous iron use, hypertensive disorder of pregnancy, induction of labor, preterm delivery, aspirin use, mode of delivery, birthweight > 4000 g, delivery admission hemoglobin < 11 g/dL (aOR 1.22, 95% CI 1.11–1.34) (Table 3).
TABLE 3.
Mild thrombocytopenia and patient outcomes.
| Mild thrombocytopenia n (%) | Normal platelets n (%) | Crude OR | 95% CI | p value | aOR a | 95% CI | p value | |
|---|---|---|---|---|---|---|---|---|
| Bleeding composite b | 546 (7.6) | 5234 (6.7) | 1.14 | 1.04–1.25 | < 0.001 | 1.22 | 1.11–1.34 | < 0.001 |
| Additional uterotonic use c | 967 (13.3) | 9122 (11.7) | 1.17 | 1.09–1.26 | < 0.001 | 1.17 | 1.09–1.26 | < 0.001 |
Selection of the covariates were as follows: Meeting p < 0.05 for both mild thrombocytopenia and the dependent variable (bleeding composite or additional uterotonic use) or the dependent variable only. For the bleeding composite, this included age, race/ethnicity, marital status, body mass index, parity, delivery hospital, insurance status, primary language, intravenous iron use, hypertensive disorder of pregnancy, induction of labor, preterm delivery, aspirin use, mode of delivery, birthweight > 4000 g, delivery admission hemoglobin < 11 g/dL. For additional uterotonic use, this included race/ethnicity, body mass index, parity, delivery hospital, primary language, intravenous iron use, hypertensive disorder of pregnancy, induction of labor, aspirin use, mode of delivery, birthweight > 4000 g, delivery admission hemoglobin < 11 g/dL.
Inclusive of estimated blood loss at delivery ≥1000 mL, International Classification of Diseases (ICD) diagnosis of postpartum hemorrhage, or receipt of a red blood cell transfusion.
Inclusive of methylergonovine, carboprost tromethamine, and/or misoprostol.
Additional uterotonic was used in 967 (13.3%) of patients with mild thrombocytopenia and 9122 (11.7%) of patients with a normal platelet count (cOR 1.17, 95% CI 1.09–1.26). This association persisted after accounting for potential confounders including race/ethnicity, BMI, parity, delivery hospital, primary language, intravenous iron use, hypertensive disorder of pregnancy, induction of labor, aspirin use, mode of delivery, birthweight > 4000 g, delivery admission hemoglobin < 11 g/dL (aOR 1.17, 95% CI 1.09–1.26).
Patients with mild thrombocytopenia had no differences in rates of secondary outcomes compared to those with normal platelet counts (Table 1). Specifically, there was no difference in need for additional hemorrhage procedures occurring in 67 (0.9%) patients with mild thrombocytopenia and 694 (0.9%) with a normal platelet count (p = 0.73). There was additionally no difference in tranexamic acid use with 55 (0.8%) of patients with mild thrombocytopenia and 771 (1.0%) of patients with a normal platelet count receiving the medication (p = 0.06). Finally, there was no differences in rates of ICU admission occurring in 14 (0.2%) patients with mild thrombocytopenia and 200 (0.3%) of patients with a normal platelet count (p = 0.31).
4. DISCUSSION
Based on the results of this study, patients with mild thrombocytopenia at the time of hospital admission for delivery, defined as platelet count between 100 × 1000/µL and 149 × 1000/µL, were more likely to experience the bleeding composite and more likely to require uterotonics compared to patients with a normal platelet count. Mild thrombocytopenia is not traditionally considered a risk factor for postpartum hemorrhage; however, data from this large, modern cohort add to the growing body of literature that it should be considered a clinically significant risk factor for bleeding morbidity.
Prior research has demonstrated that thrombocytopenia, irrespective of its severity, is associated with increased risk of postpartum hemorrhage [11]. This association appears to be linear, with risk of hemorrhage increasing with increasing degree of thrombocytopenia [11]. The precise platelet count cutoff that confers clinically significant hemorrhage, however, is poorly studied. Traditionally recognized thresholds requiring intervention due to clinically significant hemorrhage vary by situation and include platelet count < 30 × 1000/µL for a vaginal delivery, < 50 × 1000/µL for a cesarean delivery, and < 70 × 1000/µL for neuraxial anesthesia [12, 18]. The significance of lesser degrees of thrombocytopenia, however, is less clear.
Some studies have begun to show a link between mild thrombocytopenia and obstetrical hemorrhage in select patient populations including in nulliparous, term, singleton, vertex pregnancies undergoing labor [16], in nulliparous pregnancies with any mode of delivery [17], in twin gestations [13], and in elective cesarean deliveries [14, 19]. Another study, however, found no correlation between mild thrombocytopenia at the time of cesarean delivery and risk of postpartum hemorrhage [15]. These studies were mostly conducted in smaller patient populations or with very specific inclusion criteria regarding parity and mode of delivery.
Rates of transfusions and hemorrhage‐related morbidity are continuing to increase across the United States [20]. There are multiple risk assessment tools available for obstetrical hemorrhage; however, the sensitivities and specificities vary widely [5, 21, 22]. These risk assessment tools do not traditionally include the presence of mild thrombocytopenia. This study demonstrated that mild thrombocytopenia was an independent risk factor for clinically significant obstetrical hemorrhage. Given that anticipation of hemorrhage can improve clinical outcomes [23], continuing to understand risk factors for obstetrical hemorrhage can make these risk assessment tools more robust and effective in preventing significant blood loss and thus, maternal morbidity.
This study has several important strengths. First, to our knowledge, our study is one of the largest cohorts that examines obstetrical hemorrhage risk in patients with mild thrombocytopenia at the time of delivery hospitalization. Second, we utilized robust methodology and rigorous statistical analysis to minimize potential confounding. In our analysis, we adjusted for variables that were significant for both mild thrombocytopenia and the dependent variable (bleeding composite or additional uterotonic use) or the dependent variable alone. This study, however, is not without limitation. One limitation is the use of our hospital's EMR data warehouse, which relies on accurate documentation within the medical record. While this may impact some of the demographic information and some other variables (e.g. aspirin use), the independent variables and primary outcomes were abstracted directly from the medical record minimizing the potential for coding errors. Another potential limitation is the use of estimated blood loss (EBL) rather than quantitative blood loss (QBL). QBL has been shown to be more accurate in quantifying blood loss and would have been preferred [24, 25]; however, given that EBL was consistently used, we believe that this measure can still be used to compare relative blood loss between the two patient groups. Further due to the format of the study, the team was not able to evaluate some of the clinical attributes that could have affected the risk of bleeding (e.g., length of labor, intraamniotic infection, total oxytocin dose, prior hysterotomy, etc.). Finally, there is a possibility for overestimating the number of postpartum hemorrhage cases prior to the change in the definition of postpartum hemorrhage after a vaginal birth; however, this overestimation should have occurred equally in both patients with and without mild thrombocytopenia.
Mild thrombocytopenia appears to be an independent risk factor for clinically significant obstetrical hemorrhage. Further research should be focused on two key areas. First, future studies should evaluate whether adding mild thrombocytopenia to risk assessment tools improves prediction of hemorrhage and consequent patient outcomes. Second, future research should focus on potential interventions that can decrease hemorrhage risk in patients with known mild thrombocytopenia. One potential area for exploration involves studying hemorrhage risk in patients with mild thrombocytopenia who are using daily low dose aspirin.
In conclusion, we found that patients with mild thrombocytopenia at the time of delivery hospitalization had increased odds of clinically significant obstetrical hemorrhage. This patient population is not traditionally considered to be at increased risk of clinically significant hemorrhage; however, this variable should be added to risk assessment tools to better identify patients at risk for hemorrhage.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Supporting information
Supporting Information
REFERENCES
- 1. Say, L. , Chou D., Gemmill A., Tunçalp Ö., Moller A.‐B., Daniels J., Alkema L., et al. 2014. “Global Causes of Maternal Death: A WHO Systematic Analysis.” Lancet Global Health 2(6): e323–33. 10.1016/S2214-109X(14)70227-X. [DOI] [PubMed] [Google Scholar]
- 2. Creanga, A. A. , Berg C. J., Ko J. Y., Farr S. L., Tong V. T., Bruce F. C., and Callaghan W. M.. 2014. “Maternal Mortality and Morbidity in the United States: Where Are We Now?.” Journal of Women's Health 23(1): 3–9. 10.1089/jwh.2013.4617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Committee on Practice B‐O . 2017. “Practice Bulletin No. 183: Postpartum Hemorrhage.” Obstetrics and Gynecology 130(4): e168–86. 10.1097/AOG.0000000000002351. [DOI] [PubMed] [Google Scholar]
- 4. Menard, M. K. , Main E. K., and Currigan S. M.. 2014. “Executive Summary of the reVITALize Initiative: Standardizing Obstetric Data Definitions.” Obstetrics and Gynecology 124(1): 150–3. 10.1097/AOG.0000000000000322. [DOI] [PubMed] [Google Scholar]
- 5. Main, E. K. , Goffman D., Scavone B. M., Low L. K., Bingham D., Fontaine P. L., Levy B. S., et al. 2015. “National Partnership for Maternal Safety: Consensus Bundle on Obstetric Hemorrhage.” Obstetrics and Gynecology 126(1): 155–62. 10.1097/AOG.0000000000000869. [DOI] [PubMed] [Google Scholar]
- 6. Shields, L. E. , Smalarz K., Reffigee L., Mugg S., Burdumy T. J., and Propst M.. 2011. “Comprehensive Maternal Hemorrhage Protocols Improve Patient Safety and Reduce Utilization of Blood Products.” American Journal of Obstetrics and Gynecology 205(4): 368 e1–8. 10.1016/j.ajog.2011.06.084. [DOI] [PubMed] [Google Scholar]
- 7. Boehlen, F. , Hohlfeld P., Extermann P., Perneger T. V., and de Moerloose P.. 2000. “Platelet Count at Term Pregnancy: A Reappraisal of the Threshold.” Obstetrics and Gynecology 95(1): 29–33. 10.1016/s0029-7844(99)00537-2. [DOI] [PubMed] [Google Scholar]
- 8. Reese, J. A. , Peck J. D., Deschamps D. R., McIntosh J. J., Knudtson E. J., Terrell D. R., Vesely S. K., and George J. N.. 2018. “Platelet Counts during Pregnancy.” New England Journal of Medicine 379(1): 32–43. 10.1056/NEJMoa1802897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Reese, J. A. , Peck J. D., Yu Z., Scordino T. A., Deschamps D. R., McIntosh J. J., George J. N., et al. 2019. “Platelet Sequestration and Consumption in the Placental Intervillous Space Contribute to Lower Platelet Counts during Pregnancy.” American Journal of Hematology 94(1): E8–E11. 10.1002/ajh.25321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Reese, J. A. , Peck J. D., McIntosh J. J., Vesely S. K., and George J. N.. 2017. “Platelet Counts in Women with Normal Pregnancies: A Systematic Review.” American Journal of Hematology 92(11): 1224–32. 10.1002/ajh.24829. [DOI] [PubMed] [Google Scholar]
- 11. Carlson, L. M. , Dotters‐Katz S. K., Smid M. C., and Manuck T. A.. 2017. “How Low Is Too Low? Postpartum Hemorrhage Risk among Women with Thrombocytopenia.” American Journal of Perinatology 34(11): 1135–41. 10.1055/s-0037-1604194. [DOI] [PubMed] [Google Scholar]
- 12. 2019. “ACOG Practice Bulletin No. 207: Thrombocytopenia in Pregnancy.” Obstetrics and Gynecology 133(3): e181–93. 10.1097/AOG.0000000000003100. [DOI] [PubMed] [Google Scholar]
- 13. Attali, E. , Epstein D., Lavie M., Lavie A., Reicher L., Yogev Y., Ashwal E., and Hiersch L.. 2022. “Mild Thrombocytopenia and the Risk for Postpartum Hemorrhage in Twin Pregnancies.” International Journal of Gynaecology and Obstetrics 159(3): 790–5. 10.1002/ijgo.14187. [DOI] [PubMed] [Google Scholar]
- 14. Attali, E. , Epstein D., Reicher L., Lavie M., Yogev Y., and Hiersch L.. 2021. “Mild Thrombocytopenia Prior to Elective Cesarean Section Is an Independent Risk Factor for Blood Transfusion.” Archives of Gynecology and Obstetrics 304(3): 627–32. 10.1007/s00404-021-05988-x. [DOI] [PubMed] [Google Scholar]
- 15. DiSciullo, A. , Mokhtari N., Landy H., and Kawakita T.. 2021. “Effect of Mild Preoperative Thrombocytopenia on Postpartum Hemorrhage after Cesarean Deliveries.” American Journal of Obstetrics & Gynecology MFM 3(4): 100368. 10.1016/j.ajogmf.2021.100368. [DOI] [PubMed] [Google Scholar]
- 16. Govindappagari, S. , Moyle K., and Burwick R. M.. 2020. “Mild Thrombocytopenia and Postpartum Hemorrhage in Nulliparous Women with Term, Singleton, Vertex Deliveries.” Obstetrics and Gynecology 135(6): 1338–44. 10.1097/AOG.0000000000003861. [DOI] [PubMed] [Google Scholar]
- 17. Rottenstreich, M. , Rotem R., Glick I., Shen O., Grisaru‐Granovsky S., and Sela H. Y.. 2022. “Mild Gestational Thrombocytopenia in Primiparous Women, Does It Affect Risk of Early Postpartum Hemorrhage? A Retrospective Cohort Study.” Journal of Maternal‐Fetal & Neonatal Medicine 35(25): 8426–33. 10.1080/14767058.2021.1977793. [DOI] [PubMed] [Google Scholar]
- 18. Bauer, M. E. , Arendt K., Beilin Y., Gernsheimer T., Perez Botero J., James A. H., Leffert L., et al. 2021. “The Society for Obstetric Anesthesia and Perinatology Interdisciplinary Consensus Statement on Neuraxial Procedures in Obstetric Patients With Thrombocytopenia.” Anesthesia and Analgesia 132(6): 1531–44. 10.1213/ANE.0000000000005355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Işıkalan, M. M. , Özkaya E. B., Özkaya B., Ferlibaş E., Şengül N., and Acar A.. 2021. “Does Mild Thrombocytopenia Increase Peripartum Hemorrhage in Elective Cesarean Deliveries? A Retrospective Cohort Study.” International Journal of Gynaecology and Obstetrics 153(1): 89–94. 10.1002/ijgo.13554. [DOI] [PubMed] [Google Scholar]
- 20. Merriam, A. A. , Wright J. D., Siddiq Z., D'Alton M. E., Friedman A. M., Ananth C. V., and Bateman B. T.. 2018. “Risk for Postpartum Hemorrhage, Transfusion, and Hemorrhage‐related Morbidity at Low, Moderate, and High Volume Hospitals.” Journal of Maternal‐Fetal & Neonatal Medicine 31(8): 1025–34. 10.1080/14767058.2017.1306050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Kramer, M. S. , Berg C., Abenhaim H., Dahhou M., Rouleau J., Mehrabadi A., and Joseph K. S.. 2013. “Incidence, Risk Factors, and Temporal Trends in Severe Postpartum Hemorrhage.” American Journal of Obstetrics and Gynecology 209(5): 449 e1–7. 10.1016/j.ajog.2013.07.007. [DOI] [PubMed] [Google Scholar]
- 22. Dilla, A. J. , Waters J. H., and Yazer M. H.. 2013. “Clinical Validation of Risk Stratification Criteria for Peripartum Hemorrhage.” Obstetrics and Gynecology 122(1): 120–6. 10.1097/AOG.0b013e3182941c78. [DOI] [PubMed] [Google Scholar]
- 23. Ende, H. B. . 2022. “Risk Assessment Tools to Predict Postpartum Hemorrhage.” Best Practice & Research Clinical Anaesthesiology 36(3–4): 341–8. 10.1016/j.bpa.2022.08.003. [DOI] [PubMed] [Google Scholar]
- 24. Rubenstein, A. F. , Zamudio S., Douglas C., Sledge S., and Thurer R. L.. 2021. “Automated Quantification of Blood Loss versus Visual Estimation in 274 Vaginal Deliveries.” American Journal of Perinatology 38(10): 1031–5. 10.1055/s-0040-1701507. [DOI] [PubMed] [Google Scholar]
- 25. Blosser, C. , Smith A., and Poole A. T.. 2021. “Quantification of Blood Loss Improves Detection of Postpartum Hemorrhage and Accuracy of Postpartum Hemorrhage Rates: A Retrospective Cohort Study.” Cureus 13(2): e13591. 10.7759/cureus.13591. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information