Obesity has become a global epidemic, and obesity during pregnancy is associated with increased maternal and neonatal complications.1–5 To deal with complications of obesity, bariatric surgeries are being performed with increasing frequency. In Sweden 95% of bariatric surgeries are performed as a gastric bypass.6 Bariatric surgery has led to a reduction in cardiovascular morbidity, diabetes,7,8 cancer,9 and overall mortality.10 The most common long-term complication of gastric bypass is the formation of an internal abdominal hernia, defined as an intermittent or persistent herniation through a transmesenteric defect, leading to intestinal obstruction, which occurs in 2% of patients after bariatric surgery in Sweden.11
Bariatric surgery is associated with several adverse pregnancy outcomes including an increased risk of premature labor, cesarean delivery, and being born small for gestational age, but also lower risk of gestational diabetes and of being born large for gestational age.12–20 The aim with our study was to compare the rate of abdominal surgery during pregnancy in women with previous bariatric surgery as compared with women with first-trimester body mass index (BMI, calculated as weight (kg)/[height (m)]2) greater than 35 who had not had bariatric surgery. To shed light on the question whether morbidly obese women should finish childbearing before or after having bariatric surgery, we also compared the rate of abdominal surgery during pregnancy in patients with bariatric surgery before having any children as compared with women in a control group with BMI greater than 35 in their first pregnancy.
MATERIALS AND METHODS
We performed a national cohort study. Our women in the case group were identified in the Swedish National Patient Registry as having undergone bariatric surgery from 1987 to 2011 with one of the following surgical procedure codes: gastroplasty (JDF00–01), gastric bypass (JDF10–11), or gastric banding (JDF20–21). These data were merged between the Swedish National Patient Registry11 and the Swedish Medical Birth Registry20,21 to identify pregnancy and labor outcomes. Women in the control group consisted of women identified in the Swedish Medical Birth Registry as having first-trimester BMI greater than 35 regardless of whether this BMI occurred in the first or a subsequent pregnancy. Women in the control group were not excluded from the study in the event that they lost weight in subsequent pregnancies and reached BMIs less than 35. No exclusions resulting from obstetric risk factors were made. The women in our study gave birth between 1987 and 2013.
The Swedish National Patient Registry contains diagnosis and surgical procedural codes for all patients admitted to hospitals in Sweden since 1987. Codes are classified according to the International Classification of Diseases, 9th and 10th Revision (ICD-9 and -10) codes. The ICD-9 was used for the years 1987–1997 in all of Sweden and in southern Sweden until 1998; from 1998 onward, ICD-10 codes were used.
Data on maternal characteristics and neonatal outcome are from the Swedish Medical Birth Registry. The Swedish Medical Birth Registry contains prospectively collected medical information on 99% of all deliveries in Sweden. All women in Sweden are offered free prenatal care. Standardized record forms are used at all prenatal clinics, all delivery units, and pediatric clinics during the first month of life. Copies of these forms are sent to the Swedish National Board of Health, and the information is included in the Swedish Medical Birth Registry. Information on maternal smoking is collected at the prenatal care clinic, usually at gestational weeks 8–12, and has been available in the Swedish Medical Birth Registry along with BMI data since 1983. Birth weight was classified in accordance with the national ultrasound-based fetal weight growth standard22 and is expressed as standard deviation scores relative to gender and gestational age. Neonates with a birth weight more than 2 standard deviations below the mean weight for gestational age are classified as small for gestational age, and neonates weighing more than 2 standard deviations above mean weight for gestational age are classified as large for gestational age.
Primary outcome variables were maternal ICD diagnosis and surgical procedure codes identified in the Swedish National Patient Registry during pregnancy and delivery and grouped as the following five outcome categories. The diagnosis codes and surgical procedure codes are listed in detail in Appendix 1, available online at http://links.lww.com/AOG/A939. In each outcome category, the presence of ICD diagnosis codes and surgical procedure codes was analyzed separately. Surgery could be performed as either laparoscopic or laparotomy:
- 1. Intestinal obstruction or abdominal hernia
- a. Surgical procedure codes
- b. Diagnosis codes
- 2. Gallbladder disease
- a. Surgical procedure codes (cholecystectomy)
- b. Cholelithiasis and cholecystitis
- 3. Appendicitis
- a. Surgical procedure codes (appendectomy)
- b. Diagnosis codes
- 4. Hernia (inguinal, femoral, umbilical, ventral)
- a. Surgical procedure codes
- b. Diagnosis codes
- 5. Diagnostic laparoscopy or laparotomy, without the presence of any ICD diagnosis codes or any surgical procedure codes for outcomes in outcome categories 1–4
Previous surgery was defined as a surgical procedure code in outcome categories 1–5. The study was approved by the Ethical Committee, Lund University, Sweden number 2013/11.
To study the effects of bariatric surgery on abdominal surgery during pregnancy, we first compared surgical outcomes for women who had bariatric surgery and then became pregnant with outcomes for women in the control group who were candidates for bariatric surgery in that they had first-trimester BMIs greater than 35 but who had not had bariatric surgery and then became pregnant. In further analysis, we computed odds ratios (ORs) controlling for smoking, year of delivery, maternal age, and any previous abdominal surgery. A complication is that the sequencing of bariatric surgery and births varies across individuals. To hold this sequencing constant in computing ORs, we performed analysis on a subcohort of women who had bariatric surgery before having any children. Note that in the total cohort, we considered surgical outcomes for those who had bariatric surgery and then became pregnant, even if the pregnancy was not the woman's first pregnancy. In the subcohort, on the other hand, we restricted the sample to women in whom the pregnancy was the woman's first pregnancy. For consistency, we only considered women in the control group with first-trimester BMIs greater than 35 at the time of first pregnancy. The different study groups for the descriptive analyses and OR analysis are shown in Figure 1.
Odds ratios for the five surgical outcome categories were separately calculated for the first and second pregnancy (bariatric surgery or reaching BMI greater than 35, respectively) using logistic regression analyses. Odds ratios were calculated for all subsequent pregnancies after inclusion and were calculated using general estimating equations with robust variance estimation. If numbers were small, exact ORs were calculated,23 but if numbers permitted, adjusted ORs were obtained. Initially the eligible possible confounders were: year of delivery (linear, continuous), primiparity (yes or no), maternal age (linear, continuous), maternal smoking (semicontinuous: 1=no smoking, 2=smoking one to nine cigarettes/d, 3=smoking 10 or more cigarettes/d), and any previous abdominal surgery (yes or no). Backward elimination procedures were then used so that besides bariatric surgery (yes or no), the final models included factors with P<.2 only. Thus, the variables included in final models differ across the five outcome categories evaluated and are specified in the “Results” section. Missing smoking data were replaced by the overall mean. Detailed information on the final models is available in Appendix 2, available online at http://links.lww.com/AOG/A939. Differences between women in the case group and those in the control group displayed in descriptive tables were assessed using χ2 tests or Fisher exact tests; the latter was used when numbers in any cell were below 5. P values <.05 (two-sided) were regarded as statistically significant. Statistical analyses were made using SPSS 23.
Our total study cohort consisted of 2,543 women in the case group with a total of 3,383 pregnancies after bariatric surgery and 21,909 women in the control group with a total of 42,458 pregnancies (Fig. 1). Our subcohort consisted of 1,240 women who had bariatric surgery before having any children and 14,669 women in the control group who had BMIs greater than 35 during their first pregnancy. The women who were nulliparous before bariatric surgery (n=1,240) had a total of 1,757 pregnancies after bariatric surgery. Women in the control group with BMIs greater than 35 during their first pregnancy (n=14,669) had a total of 30,854 subsequent pregnancies (Table 1).
Compared with women in the control group, the women in the case group were significantly older, more often multiparous, more often from Nordic countries, and more often had undergone previous abdominal surgery. The median time from bariatric surgery to first delivery was 1,142 days with an interquartile range of 1,042 days.
Maternal risk factors increased with every pregnancy after bariatric surgery. In the third pregnancy after bariatric surgery, 54 (43.9%) had undergone previous abdominal surgery, 53 (51.3%) had BMIs greater than 35, 26 (21.1%) smoked during pregnancy, 16 (13.0%) were older than 40 years of age, seven (5.7%) gave birth to small-for-gestational-age neonates, and 23 (18.7%) delivered at less than 37 weeks of gestation.
In the third pregnancy after reaching BMIs greater than 35, 389 (8.1%) had undergone abdominal surgery, 3,785 (78.9%) had BMIs greater than 35, 917 (19.1%) smoked during pregnancy, 252 (5.3%) were older than 40 years of age, 77 (1.6%) gave birth to small-for-gestational-age neonates, and 1,150 (9.6%) delivered at less than 37 weeks of gestation.
Table 2 shows the risk of maternal diagnosis and surgery in the five outcome categories for the first, second, and all pregnancies in the total study cohort. During the first pregnancy after bariatric surgery, regardless of parity, 65 of 2,543 women in the case group (2.5%) underwent surgery in outcome categories 1–4. In addition, 37 women in the case group (1.5%) underwent laparoscopy or laparotomy as a result of other indications than the diagnosis specified in outcome categories 1–4. Among women in the case group, 39 (1.5%) had surgery for intestinal obstruction during the first pregnancy after bariatric surgery. In addition to the 39 women who underwent surgery for intestinal obstruction in the first pregnancy, 25 additional women in the case group (1.5%) received an ICD diagnosis code for intestinal obstruction, but no surgical procedure code was present.
During the first pregnancy after reaching BMIs greater than 35, 53 (0.2%) women had surgery in outcome categories 1–4. By contrast, four (0.08%) women in the control group had surgery resulting from intestinal obstruction during the first pregnancy after reaching BMIs greater than 35. In addition to the four women in the case group who underwent surgery for intestinal obstruction in the first pregnancy, 10 additional women in the case group (0.05%) received an ICD diagnosis code for intestinal obstruction, but no surgical procedure code was present.
Classification of surgical procedures during pregnancy and time (years) from bariatric surgery are shown in Table 3. The most common indication for surgery among women in the case group was intestinal obstruction and the most common among women in the control group was cholecystectomy.
After bariatric surgery, 121 of 2,543 (4.6%) became pregnant within 1 year, of whom four (3.3%) had surgery as a result of intestinal obstruction; none underwent diagnostic laparotomy or laparoscopy. The second year after bariatric surgery, 909 of 2,543 women became pregnant, of whom 17 (1.9%) were operated as a result of intestinal obstruction, 18 (2.0%) underwent diagnostic surgery, and eight (0.9%) had a cholecystectomy.
Among women in the case group, 75 women had either a diagnosis or surgical procedure code for intestinal obstruction during pregnancy (three women had surgery during two pregnancies), of whom 63 (80.4%) women had no subsequent pregnancies. Three patients had surgery as a result of intestinal obstruction in both the first and second pregnancies after bariatric surgery.
Appendix 3, available online at http://links.lww.com/AOG/A939, shows the risk of maternal abdominal surgery and diagnosis during pregnancy in our subcohort of women who underwent bariatric surgery before childbirth compared with women in the control group who had first-trimester BMIs greater than 35 at first pregnancy.
Among women who had bariatric surgery before any pregnancy, a total of 22 of 1,241 (1.8%) had surgery as a result of intestinal obstruction compared with 2 of 14,669 (0.01%) among women in the control group with BMIs greater than 35 during their first pregnancy (crude OR 132.5, 95% confidence interval [CI] 32.5–1,162) and 21 of 1,241 (1.7%) underwent diagnostic laparotomy or laparoscopy (outcome category 5) in the first pregnancy after bariatric surgery compared with 16 of 14,669 (0.11%) of women in the control group with BMIs greater than 35 during their first pregnancy (adjusted OR 8.9, 95% CI 4.4–18.0).
In the final model, the data were adjusted for surgery performed before bariatric surgery. In a subset, adjustment was made for all previous surgery, regardless whether it was performed before or after bariatric surgery, which still showed an increased risk of intestinal obstruction and diagnostic surgery during pregnancy (adjusted OR 8.9, 95% CI 5.2–15.3 and adjusted OR 3.9, 95% CI 2.4–6.4, respectively).
This national cohort study shows that pregnant women with previous bariatric surgery have an increased rate of abdominal surgery as compared with pregnant women with BMIs greater than 35. In the first pregnancy after bariatric surgery, a 34-fold increased risk of undergoing surgery during pregnancy resulting from intestinal obstruction (from 0.08% to 1.5%), and in addition, a 11-fold increased risk of undergoing diagnostic laparoscopy or laparotomy during pregnancy (from 0.1% to 1.5%), was found when compared with women with BMIs greater than 35. Furthermore, an almost 20-fold increased risk (from 0.05% to 1.0%) of receiving an ICD diagnosis code for intestinal obstruction, without a coinciding surgical procedure code, was found in the first pregnancy after bariatric surgery when compared with the first pregnancy of women in the control group. Women with previous bariatric surgery had a sixfold increased risk of a surgical procedure code for any abdominal surgery during pregnancy compared with women in the control group.
Bariatric surgery and previous abdominal surgery are both risk factors for subsequent intestinal obstruction and abdominal surgery. Hence, abdominal surgery performed after bariatric surgery may be an effect of bariatric surgery and not a true confounder. We therefore chose to differentiate between previous abdominal surgery performed before as opposed to after bariatric surgery when presenting our results. We also chose not to stratify for maternal BMI, because maternal BMI may lie in the causal pathway between bariatric surgery and surgery during pregnancy and hence is not a confounder.
Abdominal surgery was most common during the first pregnancy after bariatric surgery, and only 20% of women had additional pregnancies after having surgery or a diagnosis of intestinal obstruction during the first pregnancy. One might assume that only the healthiest mothers chose to have more than one birth after bariatric surgery. Surprisingly, maternal risk factors such as smoking and previous abdominal surgery increased with each subsequent pregnancy after bariatric surgery. Indeed, women with previous bariatric surgery represent a high-risk pregnancy.
Intestinal obstruction during pregnancy is potentially life-threatening and a fetal loss rate of 17% and maternal mortality rate of 2% have been reported.24–26 Pregnancy may increase the risk of the formation14–16 of an internal hernia after bariatric surgery as a result of increased intraabdominal pressure and displacement of organs by an enlarged uterus. Furthermore, it is difficult to diagnose internal herniation, because symptoms can be vague and abdominal pain and vomiting are common in pregnancy.25 Concerns about fetal exposure to radiation may also delay computed tomography scanning.26 Therefore, a low threshold for diagnostic laparoscopy or laparotomy remains the cornerstone of recommended treatment.27
An important clinical question is whether obese women should undergo bariatric surgery before or after finishing childbearing. With this question in mind, we analyzed ORs in a subcohort of women who had their first child after bariatric surgery compared with women who were potential candidates for bariatric surgery because they had BMIs greater than 35 at first pregnancy. The increased risk of abdominal surgery for the subcohort was in the same magnitude as for the total cohort. Because we only focused on maternal abdominal surgical outcomes and did not take fetal outcome into consideration, we cannot answer that question fully, but believe our results can be useful when counseling obese women about the putative benefits and maternal risks of pregnancy after bariatric surgery.
The strengths of our study include the large sample size and long follow-up. The large study size made it possible to estimate the ORs with reasonable precision.
A drawback to our study is that some patients seemed to have received a diagnosis code for intestinal obstruction but did not receive a surgical procedure code for surgery as a result of intestinal obstruction. Because conservative treatment for intestinal obstruction is not recommended, the true rate of surgery for intestinal obstruction could be higher than the rates we have stated. On the other hand, a possible explanation of the discrepancy between the presence of a diagnosis code but no surgical procedure code for intestinal obstruction could be that the clinical symptoms of intestinal obstruction regressed spontaneously and the patient was erroneously given the diagnosis code for intestinal obstruction.
In addition, a drawback could be incorrect coding in patient registries; women with bariatric surgery might be more accurately diagnosed with a bowel obstruction as opposed to those without bariatric surgery. Furthermore, we have not investigated the prevalence of surgery between pregnancies.
Our results are from women who underwent bariatric surgery during 1987–2011. In Sweden, the surgeons started to close the mesenteric defects after gastric bypass surgery in 2012, which is believed to reduce the incidence of internal herniation. During our study period, 95% of bariatric surgeries were performed as gastric bypass. During the last few years sleeve gastrectomy has become more frequent in Sweden and now constitutes 16% of the bariatric surgeries performed in 2014.28 As a result of this shift in operation technique, complications resulting from intestinal obstruction may decline in the future.
In conclusion, our study showed substantially increased risks of surgery for intestinal obstruction and diagnostic laparotomy or laparoscopy during pregnancy in women with previous bariatric surgery when compared with women with BMIs greater than 35 without bariatric surgery. Although the absolute risk of abdominal surgery is small, it must nevertheless be weighed against the reduction of obesity-related obstetric risk factors after bariatric surgery when advising obese women about family planning.
1. Gaillard R, Durmuş B, Hofman A, Mackenbach JP, Steegers EA, Jaddoe VW. Risk factors and outcomes of maternal obesity and excessive weight gain during pregnancy. Obesity (Silver Spring) 2013;21:1046–55.
2. Sheiner E, Menes TS, Silverberg D, Abramowicz JS, Levy I, Katz M, et al. Pregnancy outcome of patients with gestational diabetes mellitus following bariatric surgery Am J Obstet Gynecol 2006;194:431–5.
3. Sheiner E, Levy A, Silverberg D, Menes TS, Levy I, Katz M, et al. Pregnancy after bariatric surgery is not associated with adverse perinatal outcome. Am J Obstet Gynecol 2004;190:1335–40.
4. Cedergren MI. Maternal morbid obesity and the risk of adverse pregnancy outcome. Obstet Gynecol 2004;103:219–24.
5. Robinson HE, O'Connell CM, Joseph KS, McLeod NL. Maternal outcomes in pregnancies complicated by obesity. Obstet Gynecol 2005;106:1357–64.
6. Leffler E, Gustavsson S, Karlsson BM. Time trends in obesity surgery 1987 through 1996 in Sweden—a population-based study. Obes Surg 2000;10:543–8.
7. Sjöström L, Peltonen M, Jacobson P, Sjöström CD, Karason K, Wedel H, et al. Bariatric surgery and long-term cardiovascular events. JAMA 2012;307:56–65.
8. Carlsson LM, Peltonen M, Ahlin S, Anveden Å, Bouchard C, Carlsson B, et al. Bariatric surgery and prevention of type 2 diabetes in Swedish obese subjects. N Engl J Med 2012;367:695–704.
9. Sjöström L, Gummesson A, Sjöström CD, Narbro K, Peltonen M, Wedel H, et al. Effects of bariatric surgery on cancer incidence in obese patients in Sweden (Swedish Obese Subjects Study): a prospective, controlled intervention trial. Lancet Oncol 2009;10:653–62.
10. Sjöström L, Narbro K, Sjöström CD, Karason K, Larsson B, Wedel H, et al. Effects of bariatric surgery on mortality in Swedish obese subjects. N Engl J Med 2007;357:741–52.
11. Swedish National Board of Health, Patient Registry. VGR rapport 2011–11–22 https://www2.sahlgrenska.se/upload/SU/HTA-centrum/HTA-rapporter/HTA%20Obesitaskirurgi%202011-11-22.pdf
. Retrieved March 10, 2017.
12. Marsk R, Freedman J, Tynelius P, Rasmussen F, Näslund E. Antiobesity surgery in Sweden from 1990 to 2005: a population-based study with a focus on mortality. Ann Surg 2008;248:777–81.
13. Kjær MM, Lauenborg J, Breum BM, Nilas L. The risk of adverse pregnancy outcome after bariatric surgery: a nationwide register-based matched cohort study. Am J Obstet Gynecol 2013;208:465.e1–5.
14. Loer PV III, Sanchez-Ramos L, Kaunitz AM, Kerwin AJ, Diaz J. Maternal death caused by midgut volvulus after bariatric surgery. Am J Obstet Gynecol 2005;193:1748–9.
15. Moore KA, Ouyang DW, Whang EE. Maternal and fetal deaths after gastric bypass surgery for morbid obesity. N Engl J Med 2004;351:721–2.
16. Renault K, Gyrtrup HJ, Damgaard K, Hedegaard M, Sorensen JL. Pregnant woman with fatal complication after laparoscopic Roux-en-Y gastric bypass. Acta Obstet Gynecol Scand 2012;91:873–5.
17. Shai D, Shoham-Vardi I, Amsalem D, Silverberg D, Levi I, Sheiner E. Pregnancy outcome of patients following bariatric surgery as compared with obese women: a population-base study. J Matern Fetal Neonatal Med 2014;27:275–8.
18. Kjær MM, Nilas L. Timing of pregnancy after gastric bypass—a national register-based cohort study. Obes Surg 2013;23:1281–5.
19. Sheiner E, Levy A, Menes TS, Silverberg D, Katz M, Mazor M. Maternal obesity as an independent risk factor for caesarean delivery. Paediatr Perinat Epidemiol 2004;18:196–201.
20. Johansson K, Cnattingius S, Näslund I, Roos N, Trolle Lagerros Y, Granath F, et al. Outcomes of pregnancy after bariatric surgery. N Engl J Med 2015;372:814–24.
21. Källén B, Källén K, Otterblad Olausson P. The Swedish Medical Birth Register: a summary of content and quality. Res Rep EpC 2003; 2003-112-3.
22. Marsál K, Persson PH, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterine growth curves based on ultrasonically estimated foetal weights. Acta Paediatr 1996;85:843–8.
23. Rothman KJ. Computation of exact confidence intervals for the odds ratio. Int J Biomed Comput 1975;6:33–9.
24. Webster PJ, Bailey MA, Wilson J, Burke DA. Small bowel obstruction in pregnancy is a complex surgical problem with a high risk of fetal loss. Ann R Coll Surg Engl 2015;97:339–44.
25. Perdue PW, Johnson HW Jr, Stafford PW. Intestinal obstruction complicating pregnancy. Am J Surg 1992;164:384–8.
26. Leal-González R, De la Garza-Ramos R, Guajardo-Pérez H, Ayala-Aguilera F, Rumbaut R. Internal hernias in pregnant women with history of gastric bypass surgery: case series and review of literature Int J Surg Case Rep 2013;4:44–7.
27. Goudsmedt F, Deylgat B, Coenegrachts K, Van De Moortele K, Dillemans B. Internal hernia after laparoscopic Roux-en-Y gastric bypass: a correlation between radiological and operative findings. Obes Surg 2015;25:622–7.
28. Swedish Obesity Surgery Registry. Annual report 2015. Available at http://www.ucr.uu.se/soreg/in-english
. Retrieved March 10, 2017.