Maternal obesity seems to be associated with an overall increased risk of postpartum hemorrhage, but risk estimates vary considerably (7%–70%).1–4 Inconsistent definitions of both maternal obesity and postpartum hemorrhage complicate comparisons between data sets.
An increased trend in the incidence of postpartum hemorrhage over time has been observed in high-resource countries, but the reason for this increase remains uncertain.5,6 Recently, the International Postpartum Hemorrhage Collaborative group stated a key recommendation to further investigate maternal obesity as a potential risk factor for postpartum hemorrhage.7
The number of obese women is increasing worldwide.8 The prevalence of obesity among U.S. women aged 20 to 39 years has increased from 28.4% in 1999–2000 to 34% in 2007–2008,9 and first trimester maternal obesity have doubled from 7.6% (1989) to 15.6% (2007) in the United Kingdom.10 In Sweden, the prevalence of prepregnancy obesity was 6% in 1993 compared with 12% in 2009.
The objective of the present study was to estimate, in a large data set from the Swedish Medical Birth Registry, whether maternal obesity is associated with an increased risk for postpartum hemorrhage more than 1,000 mL and whether there is an association between maternal obesity and postpartum hemorrhage with retained placenta or postpartum hemorrhage attributable to atonic uterus or postpartum hemorrhage caused by lacerations in the birth canal and whether maternal obesity affects the risk of postpartum hemorrhage related to mode of delivery. Second, maternal obesity as a potential risk factor for use of heparin-like drugs during pregnancy was analyzed in a subgroup of women.
MATERIALS AND METHODS
The study population consisted of 1,114,071 women with singleton pregnancies who gave birth in Sweden from January 1, 1997 through December 31, 2008. They were identified using the Swedish Medical Birth Registry.11 Medical and other data of almost all (99%) deliveries in Sweden are listed in the register, which also includes stillbirths after 22 weeks of gestation. It is based on copies of the standardized medical record forms completed at the maternity health care centers at the start of prenatal care (usually during gestational week 10–12), records from the delivery units, and the pediatric examinations of the newborns. The system is identical throughout the country. A description and validation of the register content are available.11
Midwives recorded prepregnancy maternal weight and height on a standardized form at the woman's first visit to the maternity health care center. Ninety percent of women who will give birth present to the antenatal clinic during the first trimester of their pregnancy.
Body mass index (BMI, calculated as weight (kg)/[height (m)]2) was calculated from maternal weight and height data.12 A total number of 959,469 (86,1%) woman had available data for both weight and height that enabled calculation of BMI. Women were grouped in six categories of BMI: underweight (more than 18.5), normal weight (18.5–24.9), overweight (25–29.9), obese class I (30–34.9), obese class II (35–39.9), and obese class III (40 or more) according to World Health Organization.13 When evaluating postpartum hemorrhage in relation to mode of delivery, obesity classes II and III were analyzed together.
Outcomes studied were: hemorrhage more than 1,000 mL associated with retained, trapped, or adherent placenta; retained placenta with hemorrhage less than 1,000 mL, and hemorrhage more than 1,000 mL after delivery of placenta, including atonic postpartum hemorrhage and hemorrhage attributable to cervical, vaginal, or perineal lacerations.
The outcomes were registered in the Swedish Medical Birth Registry using the International Classification of Diseases (ICD). Hemorrhage more than 1,000 mL associated with retained, trapped, or adherent placenta was defined as ICD-10 code O72.0. Hemorrhage after delivery of placenta was defined as ICD-10 code O72.1. The code O72.1 in the Swedish version of ICD-10 is divided into O72.1A, which is atonic bleeding more than 1,000 mL, and O72.1B, which is bleeding more than 1,000 mL because of cervical, vaginal, or perineal lacerations, and an unspecific group O72.1X, which is used when it is unclear whether the bleeding more than 1,000 mL is attributable to atony of the uterus or lacerations. Retained placenta with hemorrhage less than 1,000 mL was equivalent to ICD-10 codes O73.0–1. The overall risk of postpartum hemorrhage more than 1,000 mL irrespective of retained placenta was evaluated in relation to year of birth, maternal age, smoking in early pregnancy, and maternal BMI. The risk for bleeding more than 1,000 mL according to delivery method in each maternal BMI class was also assessed.
Information on blood transfusions was coded and available in the register. Blood transfusions were evaluated in relation to ICD code O72 (postpartum hemorrhage more than 1,000 mL), and the risk for blood transfusion was estimated over the BMI strata.
A total number of 910 (1.8%) women included in this study had two diagnoses of postpartum hemorrhage; the most common combination was O72.0 and O721A. These women were classified as having code O72.0. The second most common combination was O73.0 and O721A; these 134 women were classified as having code O73.0 and were not included in the overall group with bleeding more than 1,000 mL. It was not possible to scrutinize each patient record to detect the main reason for postpartum hemorrhage. Eighteen patients had three postpartum hemorrhage codes.
The Swedish register on prescribed drugs was established in 2005 and includes data for all dispensed prescriptions in Sweden. The register includes data on drugs prescribed and dispensed in ambulatory care, date of prescribing and dispensing, and the prescriber's specialty. All drugs are classified according to the Anatomical Therapeutic Chemical classification system.14 The two data sources, the Swedish Medical Birth Registry and the Swedish Register on Prescribed Drugs, were linked using the woman's personal identification number. The study period was July 1, 2005–December 31, 2008.15 Use of heparin-like drugs during the second and third trimesters was obtained from this register and evaluated in relation to maternal BMI. There are five such drugs in Sweden: heparin, dalteparin, enoxaparin, danaparoid, and tinzaparin. All drugs were analyzed as a group together. The risk for bleeding more than 1,000 mL with or without the use of heparin-like drugs over the maternal BMI strata was evaluated in a subgroup in which data were available.
Maternal age (seven 5-year classes), parity (1–4 or more), smoking (unknown, no smoking, fewer than 10 cigarettes/d, 10 cigarettes/d or more), and year of birth were thought to be potential confounding factors and were included as covariates in the adjusted analyses. The register information on these variables was obtained from the antenatal care center records.
Adjusted odds ratios (OR) were determined using Mantel-Haenszel technique.16 Estimates of 95% confidence intervals (CI) were made with a test-based method17 that was based on the Mantel-Haenszel χ2. The Regional Ethical Review Board in Linköping has approved the study.
The overall prevalence of postpartum hemorrhage in the study population was 4.5% (50,565 of 1,114,071). There was an increased prevalence of postpartum hemorrhage over the study period from 1997 to 2008 (Fig. 1). The yearly risk for postpartum hemorrhage calculated as OR was practically identical after adjustments for maternal age, parity, smoking, and maternal prepregnancy BMI (data not shown). A linear regression was fitted to the percentage of women with a bleeding diagnosis each year of delivery using a least-squares method. The angle coefficient was 0.12 (95% CI 0.09–0.16) and highly statistically significant (t=7.65, P<.001). If instead a line was fitted to the OR (adjusted for age, parity, smoking, and BMI), then the angle coefficient was 0.005 with an error of 0.004 (t=1.26, P=.24). The increased prevalence of postpartum hemorrhage thus seems to be attributable to changes in population characteristics.
Maternal characteristics as risk factors for postpartum hemorrhage are shown in Table 1. There was a slightly increased risk for postpartum hemorrhage among women in all 5-year classes who were older than age 30 years. Primiparous women had an increased risk for postpartum hemorrhage. Smoking seemed to decrease the risk for postpartum hemorrhage. There was an overall slightly increased risk for postpartum hemorrhage with increasing maternal BMI, although it did not reach statistical significance in obesity class III.
Maternal BMI and risk of postpartum hemorrhage either attributable to retained placenta or attributable to atony of the uterus or attributable to cervical, vaginal, or perineal lacerations are presented in Table 2. There seemed to be no association between postpartum hemorrhage associated with retained placenta and maternal obesity. The risk of atonic hemorrhage increased rapidly with increasing BMI. There was a twofold increased risk in obesity class III (1.8%) compared with normal-weight women (1.6%). Bleeding attributable to lacerations occurred more often among obese women, except in obesity class III, but there were low numbers in that group (n=49). The risk for hemorrhage less than 1,000 mL with retained placenta was increased in the overweight and obesity II class groups.
The effect of maternal BMI on postpartum hemorrhage according to delivery method is shown in Table 3. There was an increased risk for postpartum hemorrhage the heavier the woman was after normal delivery, and it was even more pronounced after instrumental delivery. The trend was the same for cesarean delivery, although it was not statistically significant for women in obesity class I. There seemed to be no effect of BMI on the risk for postpartum hemorrhage after induction of labor, irrespective of delivery mode. The absolute risk for postpartum hemorrhage in obesity class III was 5.2% for normal vaginal delivery, 13.6% for instrumental delivery, 1.8% for cesarean delivery, and 5.3% after induction of labor. The comparable risks for normal-weight women were 4.4%, 8.8%, 1.4%, and 5.2%.
A total of 2,453 women were registered as having received blood transfusion; 1,446 of them (59%) also had ICD code O72. Among women with ICD code O72, 3% received blood transfusion. The association between blood transfusion and BMI after adjustments for year of birth, maternal age, parity, and smoking for obesity class I was OR of 1.45 (95% CI 1.15–1.83), for obesity class II it was OR of 1.87 (95% CI 1.32–2.64), and for obesity class III it was OR of 2.03 (95% CI 1.19–3.45).
The majority of heparin-like drugs used during the second or third (or both) trimesters of pregnancy were dalteparin followed by tinzaparin. In Table 4, maternal BMI was evaluated as a risk factor for using heparin-like drugs. As BMI increased, the use of heparin-like drugs increased, reaching an almost threefold increased risk among women in obesity class III.
The risk for postpartum hemorrhage among women using heparin-like drugs in the second or third (or both) trimesters was an adjusted OR of 0.94 (95% CI 0.78–1.12) (Table 5). Adjustments were made for year of birth, maternal age, parity, smoking in early pregnancy, and BMI. Low numbers of women included in this subcohort must be kept in mind when interpreting results. The risk for postpartum hemorrhage among obese women (BMI 30 or more) using heparin-like drugs during the second or third (or both) trimesters of pregnancy was compared with risk among nonobese women (BMI less than 30) using heparin-like drugs during the second or third (or both) trimesters of pregnancy was an adjusted OR of 1.15 (95% CI 0.70–1.89), comparable to the risk for postpartum hemorrhage among obese women (BMI 30 or more) without heparin-like drugs related to nonobese women (adjusted OR 1.05, 95% CI 0.70–1.89) during the same study period. There could be a slightly increased risk for postpartum hemorrhage among obese women using heparin-like drugs, but this could also be attributable to chance.
This large population-based cohort study based on Swedish Medical Birth Registry data showed that the risk of postpartum hemorrhage attributable to atonic uterus markedly increased with increasing maternal BMI. There was no association between postpartum hemorrhage attributable to retained placenta and maternal obesity. The overall increased risk of postpartum hemorrhage among obese women was 8%–13%, depending on obesity class. Sebire et al2 studied pregnancy outcomes among 31,276 obese women (BMI 30 or more) and found an increased risk of major postpartum hemorrhage (OR 1.16, 99% CI 1.12 – 1.21). The definition of major postpartum hemorrhage was not clear, but the authors state that the same increase in risk was demonstrated for blood loss of more than 1,000 mL,2 which is in accordance with results in the present study. In a data set from 1990–1999 including 677 women with BMI more than 30, the risk of blood loss of more than 500 mL was an OR of 1.5 (95% CI 1.2–1.8) compared with women with BMI of 20–30.4 No data were available on blood loss of more than 500 mL in this study. Robinson et al3 studied the risk of postpartum hemorrhage, defined as blood loss of 500 mL after vaginal delivery or 1,000 mL after cesarean delivery, among moderately obese women (prepregnancy weight of 90–120 kg) and severely obese women (prepregnancy weight of more than 120 kg). The increased risk was 10% in the moderately obese group and 6% (nonsignificant) in the severely obese group.3 Because other definitions of blood loss and BMI were used, comparisons with the present study are difficult.
The overall risk of postpartum hemorrhage among obese women is only slightly increased compared with the risk for several adverse maternal and fetal outcomes.1 The interesting finding in this study, not described previously, was that postpartum hemorrhage attributable to atonic bleeding was increased 14% in obesity class I, 47% in obesity class II, and 114% in obesity class III. Other expressions of ineffective uterine contractions have been studied in relation to maternal obesity. A cohort study demonstrated a significantly longer median duration of labor from 4 to 10 cm in overweight and obese women compared with normal-weight women.18 In a study analyzing labor progress in women with labor induction, median dose and duration of predelivery oxytocin in the obese group and extremely obese group were significantly higher than those among lean women.19 The risk for nonelective caesarean delivery attributable to ineffective uterine contractility significantly increased, with increasing maternal BMI reaching a fourfold increased risk in obesity class III.20
There was an increasing trend in postpartum hemorrhage from 1997 to 2008 in Sweden. Similar observations were made in Canada, New South Wales, United States, and Scotland, although using heterogeneous definitions of postpartum hemorrhage.7 The observed increased prevalence of postpartum hemorrhage over time in the present study was mainly associated with changes in maternal characteristics. Whether changes in maternal characteristics over time explain increasing trends of postpartum hemorrhage described in other countries is unclear because this association was not determined.
There was an overall increased risk for postpartum hemorrhage among obese women after normal delivery and it was even more pronounced after instrumental delivery. The trend was the same after cesarean delivery, although not statistically significant for women in obesity class I. In a study from the Medical Birth Registry of Norway, the risk for severe postpartum hemorrhage, which was defined as a visually estimated blood loss of 1,500 mL within 24 hours after delivery or the need for a blood transfusion after delivery, was increased among primiparous women with spontaneous onset of labor and operative vaginal delivery (OR 1.40, 95% CI 1.16–1.67) and primiparous women with spontaneous onset of labor and emergency caesarean delivery (OR 2.69, 95% CI 2.26–3.19). Spontaneous labor and normal vaginal delivery were set as references. Corresponding risk ratios for multiparous women were slightly higher. Body mass index was not included as a confounding factor.21 In the present study, the highest risk for postpartum hemorrhage after operative vaginal delivery was found among morbidly obese women (BMI 35 or higher, OR 1.58, 95% CI 1.33–1.88); the risk in the same group of women after caesarean delivery (elective and nonelective) was an OR of 1.26 (95% CI 1.01–1.55).
The risk for postpartum hemorrhage after induction of labor was not affected by BMI, irrespective of delivery mode. In a recently published study concerning the effect of maternal obesity on mode of delivery after induction of labor for prolonged pregnancy, there was no significant trend for increasing postpartum hemorrhage with increasing maternal BMI.22 These results are in accordance with findings in the present study, although all women with induction of labor for various reasons were included.
The subgroup analysis concerning the association between maternal BMI and the use of heparin-like drugs showed that maternal obesity was a significant risk factor for receiving one of these drugs during pregnancy, reaching an almost threefold increased risk among the morbidly obese. The Swedish recommendations suggest maternal prepregnancy BMI more than 28 as a risk factor for thromboembolic disease; in the frequently used risk scoring system, this factor contributes 1 point. Women with a score of four or more are recommended to use prophylactic heparin-like drugs during pregnancy.23 This subgroup was too small to evaluate whether the use of heparin-like drugs increases the risk for postpartum hemorrhage among obese women; larger datasets would be needed for that purpose.
The advantage of register studies is the large number of individuals available for evaluation and the absence of selection bias. The large number of individuals available for evaluation makes it possible to divide the study population into subgroups with sufficient numbers in each stratum and gives high statistical power. A sufficient number of participants made it possible to evaluate the three subgroups of obesity suggested by World Health Organization (obesity class I–III). The drawback is the sometimes low validity of information. ICD codes have been used to define various types of postpartum hemorrhage. It is obvious, given the large size of the study and the numbers of health care units involved, that the criteria for diagnosis may not be uniform across the study population, but the variation is probably not related to maternal BMI.
Another potential source of error in studies evaluating postpartum hemorrhage is the method of estimating blood loss. Visual estimate of blood loss is often inaccurate.24 There is no reason to believe that estimation of blood loss postpartum, irrespective of method used, differs between obese and normal-weight women.
Registered blood transfusions were used as a proxy for postpartum hemorrhage in an effort to validate data. The risk for receiving blood transfusions increased markedly with increasing maternal BMI, although no validation of the accuracy of blood transfusion registration have been made in the register.
Oxytocin use and epidural anesthesia were not included as confounders in this analysis for two reasons. First, a true confounder affects both the exposure and the outcome. Exposure in this study was prepregnancy BMI. Second, the purpose was to address the outcome, not necessarily the path.
The new findings in this population-based cohort study that the risk of postpartum hemorrhage attributable to atonic uterus markedly increased with increasing maternal BMI and that there was no association between postpartum hemorrhage attributable to retained placenta have certain important clinical implications. The slightly overall increased risk for postpartum hemorrhage among obese women has been described. This new knowledge emphasizes the need for active management after delivery of the placenta. The risk for postpartum hemorrhage after vaginal delivery differs if the woman is lean or obese, which also motivates a more stringent management of obese women.
1. Cedergren MI. Maternal morbid obesity and the risk of adverse pregnancy outcome. Obstet Gynecol 2004;103:219–24.
2. Sebire NJ, Jolly M, Harris JP, Wadsworth J, Joffe M, Beard RW, et al.. Maternal obesity and pregnancy outcome: a study of 287,213 pregnancies in London. Int J Obes Relat Metab Disord 2001;25:1175–82.
3. Robinson HE, O'Connell CM, Joseph KS, McLeod NL. Maternal outcomes in pregnancies complicated by obesity. Obstet Gynecol 2005;106:1357–64.
4. Usha Kiran TS, Hemmadi S, Bethel J, Evans J. Outcome of pregnancy in a woman with an increased body mass index. BJOG 2005;112:768–72.
5. Joseph KS, Rouleau J, Kramer MS, Young DC, Liston RM, Baskett TF, et al.. Investigation of an increase in postpartum haemorrhage in Canada. BJOG 2007;114:751–9.
6. Ford JB, Roberts CL, Simpson JM, Vaughan J, Cameron CA. Increased postpartum hemorrhage rates in Australia. Int J Gynaecol Obstet 2007;98:237–43. Epub 2007 May 4.
7. Knight M, Callaghan WM, Berg C, Alexander S, Bouvier-Colle MH, Ford JB, et al.. Trends in postpartum hemorrhage in high resource countries: a review and recommendations from the International Postpartum Hemorrhage Collaborative Group. BMC Pregnancy Childbirth 2009;9:55.
8. Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ, et al.. National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9·1 million participants. Lancet 2011;377:557–67.
9. Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trends in obesity among US adults, 1999–2008. JAMA 2010;303:235–41.
10. Heslehurst N, Rankin J, Wilkinson JR, Summerbell CD. A nationally representative study of maternal obesity in England, UK: trends in incidence and demographic inequalities in 619 323 births, 1989–2007. Int J Obes 2010;34:420–8. Epub 2009 Dec 22.
12. Blomberg MI, Källén B. Maternal obesity and morbid obesity: The risk for birth defects in the offspring. Birth Defects Res A Clin Mol Teratol 2010;88:35–40.
13. World Health Organization. Obesity: preventing and managing the global epidemic. WHO technical report series 894. Geneva (Switzerland): World Health Organization; 2000.
14. World Health Organization Oslo. Guidelines for ATC classification and DDD assignment. Oslo (Norway): WHO Collaborating Centre for Drug Statistics Methodology; 2005.
15. Wettermark B, Hammar N, Fored CM, Leimanis A, Otterblad Olausson P, Bergman U, et al.. The new Swedish Prescribed Drug Register–opportunities for pharmacoepidemiological research and experience from the first six months. Pharmacoepidemiol Drug Saf 2007;16:726–35.
16. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Nat Cancer Inst 1959;22:719–48.
17. Miettinen OS. Simple interval estimation of risk ratio. Am J Epidemiol 1974;100:515–6.
18. Vahratian A, Zhang J, Troendle JF, Savitz DA, Siega-Riz AM. Maternal prepregnancy overweight and obesity and the pattern of labor progression in term nulliparous women. Obstet Gynecol 2004;104:943–51.
19. Pevzner L, Powers BL, Rayburn WF, Rumney P, Wing DA. Effects of maternal obesity on duration and outcomes of prostaglandin cervical ripening and labor induction. Obstet Gynecol 2009;114:1315–21.
20. Cedergren MI. Non-elective caesarean delivery due to ineffective uterine contractility or due to obstructed labour in relation to maternal body mass index. Eur J Obstet Gynecol Reprod Biol 2009;145:163–6.
21. Al-Zirqi I, Vangen S, Forsén L, Stray-Pedersen B. Effects of onset of labor and mode of delivery on severe postpartum hemorrhage. Am J Obstet Gynecol 2009;201:273.e1–9.
22. Arrowsmith S, Wray S, Quenby S. Maternal obesity and labour complications following induction of labour in prolonged pregnancy. BJOG 2011;118:578–88.
23. Lindqvist PG, Torsson J, Almqvist A, Björgell O. Postpartum thromboembolism: severe events might be preventable using a new risk score model. Vasc Health Risk Manag 2008;4:1081–7.
24. Stafford I, Dildy GA, Clark SL, Belfort MA. Visually estimated and calculated blood loss in vaginal and cesarean delivery. Am J Obstet Gynecol 2008;199:519.e1–7. Epub 2008 Jul 17.