International Journal of Gynecology and Obstetrics 130 (2015) 3–9
Contents lists available at ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
REVIEW ARTICLE
A meta-analysis of maternal and fetal outcomes of pregnancy after bariatric surgery Xiao-yan Yi a, Qi-fu Li a, Jun Zhang b, Zhi-hong Wang a,⁎ a b
Department of Endocrinology, First Affiliated Hospital of Chongqing Medical University, Chongqing, China Department of General Surgery, First Affiliated Hospital of Chongqing Medical University, Chongqing, China
a r t i c l e
i n f o
Article history: Received 9 October 2014 Received in revised form 7 January 2015 Accepted 16 March 2015 Keywords: Bariatric surgery Meta-analysis Obesity Outcome Pregnancy
a b s t r a c t Background: The effects of bariatric surgery (BS) on outcomes in subsequent pregnancies are unclear. Objectives: To compare maternal and fetal outcomes among women who become pregnant after BS and obese women who have not undergone BS before pregnancy. Search strategy: PubMed and Embase were searched for relevant reports, and the reference lists of identified articles were hand-searched. Selection criteria: Cohort studies that compared outcomes among women who had undergone any type of BS and obese women who had not undergone surgery were included when results were reported as risk ratios or odds ratios (ORs). Data collection and analysis: Summary ORs were estimated using a random effects model. Main results: Eleven studies were included. Compared with obese women who had not undergone BS, women who had undergone BS had significantly lower odds of gestational diabetes (OR 0.31; 95% CI 0.15–0.65), hypertensive disorders (OR 0.42; 95% CI 0.23–0.78), and macrosomia (OR 0.40; 95% CI 0.24–0.67). However, their odds of small-for-gestationalage newborns were increased (OR 2.16; 95% CI 1.28–3.66). No significant differences were recorded for cesarean, postpartum hemorrhage, and preterm delivery. Conclusions: BS reduces the odds of some adverse maternal and fetal outcomes among obese women. © 2015 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction The prevalence of obesity is a public health concern worldwide [1]. Obesity is associated with an increased risk of obstetric complications, such as gestational diabetes, pre-eclampsia, eclampsia, macrosomia, and fetal growth restriction [2,3]. Therefore, in view of the substantial increase in the prevalence of obesity among women of child-bearing age, a rise in the frequency of adverse maternal and fetal outcomes should be expected [4]. Weight loss can reduce the likelihood of these adverse outcomes [5]. However, dietary changes, exercise, and medical management result in only short-term benefits, which are not sustained in the long term [6]. Bariatric surgery (BS) is thought to be an effective intervention to sustain weight loss [7]. BS procedures are generally categorized into three groups. Restrictive procedures (e.g. laparoscopic adjustable gastric banding [LAGB] and sleeve gastrectomy [SG]) lead to weight loss by reducing gastric capacity which in turn restricts energy intake [8,9]. Malabsorptive procedures (e.g. biliopancreatic diversion [BPD])
⁎ Corresponding author at: Department of Endocrinology, First Affiliated Hospital of Chongqing Medical University, Chongqing 400016, China. Tel.: + 86 23 89011554; fax: +86 23 89011552. E-mail address: [email protected] (Z. Wang).
lead to weight loss by restricting absorption of nutrients [10]. Finally, malabsorptive and restrictive procedures (e.g. Roux-en-Y gastric bypass [RYGB]) reduce stomach capacity, thereby causing malabsorption and a certain degree of restriction of food intake [8]. BPD is rarely used because it is associated with substantial long-term complications—hepatic failure, calcium oxalate kidney stones, renal failure, arthritis, and malnutrition [11]. The most performed procedures today are LAGB and RYGB, although SG is becoming the principal treatment option in many countries for obese women [9]. Although these procedures are beneficial in terms of weight reduction, reports of their implications on maternal and fetal outcomes have been inconsistent. The aim of the present meta-analysis was to compare maternal and fetal outcomes among women who have undergone BS with those among obese women who have not undergone BS. 2. Materials and methods 2.1. Search strategy The PubMed and Embase databases were searched from inception to October 7, 2014, with the keywords “bariatric surgery,” “pregnancy,” “obstetric,” “maternal,” “neonatal,” “perinatal,” and “fetal.” There were no language restrictions. The references of identified articles were hand-searched for further relevant reports.
http://dx.doi.org/10.1016/j.ijgo.2015.01.011 0020-7292/© 2015 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
4
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
2.2. Study selection Cohort studies that reported on maternal and/or fetal outcomes among pregnant women after BS and among obese women before/ without BS were included if the results were reported as risk ratios (RRs) or odds ratios (ORs) with corresponding 95% confidence intervals (CIs). Studies were excluded when they were of any type other than cohort studies, they compared deliveries before and after BS in the same individuals, data were not available, or they included women who were not obese or who underwent BS during pregnancy. If two studies included overlapping populations, the study for which most information was available was included and the other was excluded. If similar studies from the same authors in which data were duplicated were identified, the largest was included and the others were excluded.
Records identified (n=1004) In electronic databases (n=991) Through hand searching (n=13) Duplicates excluded (n=223) Unique records (n=781) Excluded on basis of titles and abstracts (n=760) Full-text review (n=21) Excluded (n=10) Study design (n=1) Overlapping population (n=3) Unavailable data (n=1) Not all women who underwent bariatric surgery were obese (n=4) Some women who underwent bariatric surgery were pregnant (n=1)
2.3. Data extraction and quality assessment For all included studies, two reviewers (X-y.Y. and Q-f.L.) independently extracted the first author’s name, country, publication year, sample size, maternal age, BMI, type of procedures, time from surgery to conception (S-C time), and maternal and/or fetal outcomes. When data were missing, J.Z. would email the authors of the relevant articles. A body mass index (BMI; calculated as weight in kilograms divided by the square of height in meters) of 30 or greater was considered obese. BS included any type of weight-loss procedure. The maternal outcomes assessed were gestational diabetes mellitus (GDM), hypertensive disorders (including gestational hypertension, pre-eclampsia, and eclampsia), postpartum hemorrhage, and cesarean delivery. The fetal outcomes assessed were preterm delivery, macrosomia, and being small for gestational age (SGA). The quality of the included studies was evaluated by the Newcastle– Ottawa Scale, with some modifications to match the needs of the present review. The highest score was nine points. Disagreements were resolved by discussion. 2.4. Data analysis The meta-analysis was performed using Stata 12.0 (StataCorp, College Station, TX, USA). Summary ORs were estimated using a random effects model by the Mantel-Haenszel method. Because the OR is equivalent to the RR when events are rare, it was possible to interpret the OR as the RR. Heterogeneity was assessed with the test of inconsistency (I2): when the value was greater than 50%, it was deemed statistically significant. Meta-regression, and sensitivity and subgroup analyses were planned to identify possible sources of the between-study heterogeneity if necessary and possible. The subgroups assessed were maternal age (≤ 32 years vs N32 years), type of BS (restrictive vs malabsorptive and restrictive vs mixed), and S-C time (≤ 2 years vs N2 years). The publication bias of included studies was assessed using the funnel plot with the Begg and Egger tests. A two-sided P value of less than 0.05 was considered statistically significant. The metaanalysis was conducted according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) criteria [12]. 3. Results 3.1. Description of studies A total of 781 publications were identified, of which 21 underwent full-text review (Fig. 1). After exclusion of 10 studies (Supplementary Material S1), 11 [13–23] were included in the meta-analysis (Table 1). The outcomes analyzed for each study are shown in Supplementary Material S2, along with the definitions used for each of the outcomes. Two studies from the USA [13,15] had overlapping populations, but they were included because data for hypertensive disorders were
Included in meta-analysis (n=11)
Fig. 1. Identification of eligible articles.
extracted from one [13] and data for GDM and postpartum hemorrhage were extracted for the other [15]. Among the 11 included studies, five were from Europe, four from the USA, one from the Middle East, and one from Australia (Table 1). The mean age of participants was generally older than 30 years. Among the studies reporting relevant data, generally more than half the participants had given birth previously (Table 1). The procedures performed varied; some studies stated that they included only LAGB or RYGB [16–18,20,22]. Among women who had undergone BS, there was an apparent reduction in average BMI from 40–50 to 32–35 (Table 1). Additionally, birth weight was generally lower for the neonates delivered by women who had undergone BS than for those delivered by women who had not undergone BS (Table 1). Only four studies [16,18,20,22] mentioned post-surgery nutritional recommendations and follow-up. Results of quality assessment are shown in Supplementary Material S3. 3.2. GDM Nine articles that reported the maternal outcome of GDM in women with or without surgery [14–20,22,23] were included in the analysis of GDM, with 711 cases overall. The random effects model showed that GDM was significantly less likely among women who had undergone BS than among those who had not (pooled OR 0.31, 95% CI 0.15–0.65; I2 = 85.2%, P b 0.001) (Fig. 2A). No publication bias was found with either the Begg (0.602) or Egger (P = 0.240) tests. Because the heterogeneity was significant, a sensitivity analysis was performed for GDM. It showed that no article significantly affected the results (data not shown). Subgroup analysis indicated that some heterogeneity could be a result of differences in S-C time. In some studies, the time from surgery to delivery (S-D time) was provided instead of the S-C time. For such studies, an assumption was made that their mean gestational age was 40 weeks, and S-C time was estimated as SD time minus the length of pregnancy (Supplementary Material S4). Among studies in which women who had undergone BS conceived up to 2 years after surgery, GDM was significantly less likely among women who had undergone BS than among those who had not (OR 0.22, 95% CI 0.14–0.34; I2 = 0.0%, P = 0.844). Similar results were recorded for the subgroup analyses including studies of women who conceived more than 2 years after surgery (OR 0.19, 95% CI 0.04–0.94; I2 = 70.8%, P = 0.016) (Supplementary Material S5).
5
NR 3258/3577 NR 3397/3297 3271/3305 3178/3373 2951/3463 2951/3240 NR 2948/3442 3206/3454
3.3. Hypertensive disorders
NR NR NR 9.6/15.5 5.5/7.1 6.1/6.4 12.0/11.6 14.6/9.1 NR NR NR
Nine studies [14–20,22,23] were included in the analysis of hypertensive disorders, with 594 patients with hypertensive disorders among 4993 participants overall. Hypertensive disorders were less likely among women who had undergone BS than among those who had not (summary OR 0.42, 95% CI 0.23–0.78; I2 = 83.3%, P b 0.001) (Fig. 2B). There was no publication bias noted on Begg (1.000) or Egger (P = 0.757) tests. The sensitivity analysis presented a robust result that was not influenced by individual studies (data not shown). Subgroup analysis indicated that some heterogeneity could be a result of differences in SC time. Hypertensive disorders were significantly less likely among women who underwent BS and conceived up to 2 years after surgery than among those who had not undergone surgery (OR 0.14, 95% CI 0.08–0.25; I2 = 26.1%, P = 0.258). Similarly, women who had conceived more than 2 years after surgery had a reduced likelihood of hypertensive disorders (OR 0.45, 95% CI 0.29–0.72; I2 = 0.0%, P = 0.530) (Supplementary Material S5).
23.6 mo NR 20.9 mo 20 mo NR NR NR NR 5.2 y NR NR
3.4. Postpartum hemorrhage
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters); NR, not reported. a Values are given as women who have undergone bariatric surgery/obese women who have not undergone bariatric surgery. b Exposed group only. c ≥2 categories of bariatric surgery.
NA NR NR NR 17 mo 3.1 y 3.61 y 25.4 mo NR 26.6 mo NR NR/N35 31.6/31.2 NR/N35 NR/N35 34.8/35.0 35.0/39.4 33.7/48.4 32.5/39.0 32.5/38.5 32.7/32.7 NR/N30 N35 NR N35 45.9 44 45.2 49.31 NR N35 50.4 NR NR 42.5/42.7 NR 49.4/49.4 46.1/35.0 NR 48.6/42.9 NR NR 33.3/35.8 0/0.7 2010 2014 2010 2005 2007 2010 2012 2008 2013 2010 2014 Bennett et al. [13] Berlac et al. [14] Burke et al. [15] Dixon et al. [16] Ducarme et al. [17] Lapolla et al. [18] Lesko, Peaceman [19] Patel et al. [20] Roos et al. [21] Santulli et al. [22] Shai et al. [23]
USA Denmark USA Australia France Italy USA USA Sweden France Israel
316/269 413/823 354/346 79/79 13/414 83/120 70/140 26/66 2474/12 027 24/120 326/1612
32.5/31.3 31.2/31.1 32.5/31.1 29.9/30.9 31.5/31 31.4/33.0 33.2/30.5 34.1/30.6 NR 31.7/31.7 32.9/32.3
Mixedc Malabsorptive–restrictive Mixedc Restrictive Restrictive Restrictive Mixedc Malabsorptive–restrictive Mixedc Malabsorptive–restrictive Restrictive
Mean S-C timeb Mean BMI before surgeryb Nulliparous, %a Country Year Study
Table 1 Characteristics of the 11 included studies.
Sample sizea
Mean maternal age, ya
Type of bariatric surgery
Mean BMI before pregnancya
Mean S-D timeb
Mean weight gain during pregnancy, kga
Mean birth weight, ga
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
Six studies [14,15,17,19,22,23] were included in the analysis of postpartum hemorrhage, with 181 cases in a sample of 4655 participants. Postpartum hemorrhage was less likely among women who had undergone BS than among those who had not, but the difference was not significant (summary OR 0.62, 95% CI 0.38–1.00; I2 = 8.7%, P = 0.361) (Fig. 2C). No publication bias was observed with Begg (1.000) or Egger (P = 0.439) tests, and a sensitivity analysis did not show a result sensitive to any individual studies (data not shown). No subgroup analysis was undertaken because between-study heterogeneity was not significant. 3.5. Cesarean delivery Seven studies [14,15,17–20,22] were included in the analysis of incidence of cesarean delivery. Cesarean was not significantly more likely among women who had undergone BS than among those who had not (summary OR 0.75, 95% CI 0.50–1.13; I2 = 74.2%, P = 0.001) (Fig. 2D). However, cesarean deliveries were common in both groups, which means that the OR would differ from the RR. Therefore, this result should be interpreted with caution. No publication bias was observed with Begg (1.000) or Egger (P = 0.701) tests, and a sensitivity analysis presented a robust result (data not shown). Subgroup analysis did not give any significant results because of the small number of studies (data not shown). 3.6. Preterm birth Eight studies [16–23] investigating the risk of preterm birth in women who had and had not undergone BS were included, with a total of 1366 preterm deliveries among 17 673 participants. Preterm birth was significantly more likely among women who had undergone BS than among those who had not (summary OR 1.33, 95% CI 1.16–1.52; I2 = 0.0%, P = 0.695). No publication bias was found with Begg (0.386) or Egger (P = 0.448) tests. In a sensitivity analysis, the study by Roos et al. [21] was shown to have a weight of 81.0% and influenced the results significantly. When this article was omitted, the difference between groups was no longer significant (summary OR 1.18, 95% CI 0.86–1.61; I2 = 0.0%, P = 0.671) (Fig. 3A). No subgroup analysis was undertaken because betweenstudy heterogeneity was not significant. 3.7. Macrosomia Six studies [16–20,23] concerning the risk of macrosomia in women with or without BS were included, with 368 macrosomia cases among 3028 participants. Macrosomia was significantly less likely among
6
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
Fig. 2. Forest plots of maternal outcomes. (A) Gestational diabetes. (B) Hypertensive disorders. (C) Postpartum hemorrhage. (D) Cesarean delivery. Abbreviations: OR, odds ratio; CI, confidence interval.
women who had undergone BS than among those who had not (summary OR 0.40, 95% CI 0.24–0.67; I2 = 23.9%, P = 0.255) (Fig. 3B). No publication bias was found with either Begg (1.000) or Egger (P = 0.925) tests. No study was found to influence the results significantly in a sensitivity analysis (data not shown). No subgroup analysis was undertaken because between-study heterogeneity was not significant.
3.8. SGA Five studies [18–22] were included in the analysis of incidence of SGA, with a total of 479 cases among 15 098 participants. An SGA newborn was significantly more likely among women who had undergone BS than among those who had not (summary OR 2.16, 95% CI
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
7
Fig. 3. Forest plots of fetal outcomes. (A) Preterm delivery. (B) Macrosomia. (C) SGA. Abbreviations: SGA, small for gestational age; OR, odds ratio; CI, confidence interval.
1.28–3.66; I2 = 27.8%, P = 0.236) (Fig. 3C). No publication bias was observed with Begg (0.806) or Egger (P = 0.949) tests, and a sensitivity analysis presented a robust result (data not shown). No subgroup analysis was undertaken because between-study heterogeneity was not significant.
4. Discussion In the present meta-analysis, women who had undergone BS had a reduced risk of GDM, hypertensive disorders, and macrosomia when compared with obese women who had not undergone BS. However, they had an increased risk of SGA newborns. No significant differences were recorded for cesarean delivery, postpartum hemorrhage, and preterm delivery. GDM and hypertensive disorders are two important obstetric complications. In the present study, women who became pregnant after BS had at least 50% lower odds of these complications than did obese women without a history of a bariatric procedure. However, significant heterogeneity was also observed for these two outcomes. Because of the limited data available, subgroup analyses could only be stratified by age, type of bariatric surgery, and S-C time. Heterogeneity was
lower in the subgroup analyses by S-C time for both GDM and hypertensive disorders. According to our findings, pregnancy soon after weight-loss surgery (≤2 years) was associated with a reduced incidence of GDM and hypertensive disorders. Moreover, pregnancy soon after BS had an even greater reduction in the odds of hypertensive disorder than did pregnancy more than 2 years after BS. These findings are in contrast with those of Parikh et al. [24], who reported an increased risk of GDM and decreased weight loss among women who became pregnant within the first 2 years after BS compared with those who became pregnant after 2 years. No significant difference was seen in the incidence of hypertensive disorders or anemia. Kjær et al. [25] and Dao et al. [26] did not observe any differences regarding maternal or fetal outcomes according to timing of pregnancy after surgery. In line with the present findings, Patel et al. [20] found adverse pregnancy outcomes were significantly less likely during the first 2 years. Nonetheless, they admitted that the women who became pregnant during this period required oral protein supplementation more frequently than did those who became pregnant later. Thus, they recommended postponing pregnancy for 2 years after BS. A woman is in a rapid weight-loss phase during the first 12 months after BS; the rate of weight loss decreases in the second year and then
8
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
tends to stabilize [27,28]. Most researchers believe that pregnancy during the rapid weight-loss period would raise the risks of adverse maternal and fetal outcomes [24,29,30]. Therefore, it is generally advised to delay pregnancy until at least 1 year after BS [27]. However, no conclusive evidence exists suggesting that pregnancy during the first postoperative year is unsafe [31]. On the basis of the present analysis, an early pregnancy might not be related to a higher risk of adverse outcomes, especially when adequate nutritional was supplemented. Further studies are required to better understand the influence of pregnancy timing on obstetric outcomes and set the best timeline for conception after BS. It is known that GDM increases the risk of macrosomia [32]. Macrosomia is related to serious obstetric problems, including shoulder dystocia, an increased rate of cesarean delivery, and postpartum hemorrhage [33,34]. As for GDM, the present review has shown that odds of macrosomia decreased by more than 50% after BS. However, no significant difference was recorded for cesarean delivery or postpartum hemorrhage. BS leads to weight loss through restriction of energy intake and malabsorption [28,35,36]. Therefore, risks of fetal growth restriction are a concern. In most of the studies included in the present analysis, birth weight was lower among women who had undergone BS than among obese women who had not. Meanwhile, risk of SGA was significantly increased among women who had undergone BS. In four of the included studies, mean S-C time was longer than 24 months, which indicates that malnutrition can still be severe even when conception occurs 2 years after BS, when weight loss should have slowed down. The frequency of cesarean among women who had undergone BS varied between 15.4% [17] and 61.5% [20], compared with between 29.8% [14] and 65.8% [18] of obese women who have not undergone BS. The difference in frequency of cesarean between women who had and had not undergone BS was even bigger in Dell’Agnolo et al.’s study [37]. Importantly, the difference between groups might be affected more by the local recommendations, economic benefits, and previous cesareans than by BS. Therefore, the present analysis concerning this outcome might not accurately reflect the actual influence of BS on cesarean delivery. There are several limitations of the present meta-analysis. Firstly, the included articles were all observational studies. The possibilities of selection and confounding bias are unavoidable. Indeed, patients who chose to undergo BS could differ from patients who did not undergo the procedure in terms of socioeconomic status, access to healthcare services, attention to their own health, and good after-surgery care. Alternatively, patients who have undergone BS are more likely to be severely obese or have increased comorbidity than are obese women who have not undergone this procedure. Only three included articles mentioned that patients in the unexposed group met the criteria for BS. As to confounding bias, all the confounders could not be extracted to allow an assessment of which one influenced the results most, because of unavailable detailed data. In addition to the subgroup analyses performed (maternal age, type of bariatric surgery, and S-C time), differences in obesity, BMI level, parity, previous cesarean delivery, multiple birth, adequate nutrition supplement after surgery, and weight gain during pregnancy could all bias results. Secondly, small sample sizes could limit the strength of the conclusions, especially those for the individual outcomes. One study by Ducarme et al. [17] included only 13 women who had undergone BS compared with 414 obese women who had not undergone such a procedure. However, when that study was excluded during sensitivity analysis, the results were robust. Another five of the eleven included articles included fewer than 100 women who had undergone BS. Finally, only 11 studies were included. Additionally, most of the studies were from high-income countries. In summary, BS reduces the odds of some adverse maternal-fetal outcomes among pregnant women. Compared with obese women who had not undergone BS, women who became pregnant after BS
had a lower risk of GDM, hypertensive disorders, and macrosomia. However, risk of SGA newborns increased after BS. Therefore, women who have undergone BS and subsequently become pregnant need to receive adequate supplements and their nutritional status should be closely monitored to prevent fetal growth restriction. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijgo.2015.01.011.
Acknowledgments The present analysis was supported by research grants from the Project of National Clinical Key Specialties Construction of China (2011), the General Program of Chongqing Municipal Health Bureau (2011-2-062), the Natural Science Foundation of Chongqing (CSTC2012jjA10040), and the Tackling Project of Science and Technology of Chongqing Committee of Science and Technology (CSTC2012ggyyjs10038). Conflict of interest The authors have no conflicts of interest.
References [1] Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. JAMA 2005;293(15):1861–7. [2] Weiss JL, Malone FD, Emig D, Ball RH, Nyberg DA, Comstock CH, et al. Obesity, obstetric complications and cesarean delivery rate–a population-based screening study. Am J Obstet Gynecol 2004;190(4):1091–7. [3] Eerdekens A, Debeer A, Van Hoey G, De Borger C, Sachar V, Guelinckx I, et al. Maternal bariatric surgery: adverse outcomes in neonates. Eur J Pediatr 2010;169(2): 191–6. [4] 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(8):1175–82. [5] Cnattingius S, Bergström R, Lipworth L, Kramer MS. Prepregnancy weight and the risk of adverse pregnancy outcomes. N Engl J Med 1998;338(3):147–52. [6] Catalano PM. Management of obesity in pregnancy. Obstet Gynecol 2007;109(2 Pt 1):419–33. [7] Buchwald H, Avidor Y, Braunwald E, Jensen MD, Pories W, Fahrbach K, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;292(14):1724–37. [8] Guelinckx I, Devlieger R, Vansant G. Reproductive outcome after bariatric surgery: a critical review. Hum Reprod Update 2009;15(2):189–201. [9] Buchwald H, Oien D. Metabolic/bariatric surgery worldwide 2008. Obes Surg 2009; 19(12):1605–11. [10] Scopinaro N, Gianetta E, Civalleri D, Bonalumi U, Bachi V. Bilio-pancreatic bypass for obesity: II.Initial experience in man. Br J Surg 1979;66(9):618–20. [11] Livingston EH. Obesity and its surgical management. Am J Surg 2002;184(2): 103–13. [12] Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Metaanalysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000; 283(15):2008–12. [13] Bennett WL, Gilson MM, Jamshidi R, Burke AE, Segal JB, Steele KE, et al. Impact of bariatric surgery on hypertensive disorders in pregnancy: retrospective analysis of insurance claims data. BMJ 2010;340:c1662. [14] Berlac JF, Skovlund CW, Lidegaard O. Obstetrical and neonatal outcomes in women following gastric bypass: a Danish national cohort study. Acta Obstet Gynecol Scand 2014;93(5):447–53. [15] Burke AE, Bennett WL, Jamshidi RM, Gilson MM, Clark JM, Segal JB, et al. Reduced incidence of gestational diabetes with bariatric surgery. J Am Coll Surg 2010;211(2): 169–75. [16] Dixon JB, Dixon ME, O'Brien PE. Birth outcomes in obese women after laparoscopic adjustable gastric banding. Obstet Gynecol 2005;106(5 Pt 1):965–72. [17] Ducarme G, Revaux A, Rodrigues A, Aissaoui F, Pharisien I, Uzan M. Obstetric outcome following laparoscopic adjustable gastric banding. Int J Gynecol Obstet 2007; 98(3):244–7. [18] Lapolla A, Marangon M, Dalfrà MG, Segato G, De Luca M, Fedele D, et al. Pregnancy outcome in morbidly obese women before and after laparoscopic gastric banding. Obes Surg 2010;20(9):1251–7. [19] Lesko J, Peaceman A. Pregnancy outcomes in women after bariatric surgery compared with obese and morbidly obese controls. Obstet Gynecol 2012;119(3):547–54. [20] Patel JA, Patel NA, Thomas RL, Nelms JK, Colella JJ. Pregnancy outcomes after laparoscopic Roux-en-Y gastric bypass. Surg Obes Relat Dis 2008;4(1):39–45. [21] Roos N, Neovius M, Cnattingius S, Trolle Lagerros Y, Sääf M, Granath F, et al. Perinatal outcomes after bariatric surgery: nationwide population based matched cohort study. BMJ 2013;347:f6460.
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9 [22] Santulli P, Mandelbrot L, Facchiano E, Dussaux C, Ceccaldi PF, Ledoux S, et al. Obstetrical and neonatal outcomes of pregnancies following gastric bypass surgery: a retrospective cohort study in a French referral centre. Obes Surg 2010;20(11):1501–8. [23] 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-based study. J Matern Fetal Neonatal Med 2014;27(3):275–8. [24] Parikh R, Lavoie M, Wilson L, Maranda L, Leung K, Simas TM. Timing of pregnancy after bariatric surgery. Obstet Gynecol 2014;123(Suppl. 1):165S. [25] Kjær MM, Nilas L. Timing of pregnancy after gastric bypass-a national register-based cohort study. Obes Surg 2013;23(8):1281–5. [26] Dao T, Kuhn J, Ehmer D, Fisher T, McCarty T. Pregnancy outcomes after gastricbypass surgery. Am J Surg 2006;192(6):762–6. [27] American College of Obstetricians and Gynecologists. ACOG practice bulletin no. 105: bariatric surgery and pregnancy. Obstet Gynecol 2009;113(6):1405–13. [28] 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(8):741–52. [29] Martin LF, Finigan KM, Nolan TE. Pregnancy after adjustable gastric banding. Obstet Gynecol 2000;95(6 Pt 1):927–30. [30] Wittgrove AC, Jester L, Wittgrove P, Clark GW. Pregnancy following gastric bypass for morbid obesity. Obes Surg 1998;8(4):461–4.
9
[31] Karmon A, Sheiner E. Timing of gestation after bariatric surgery: should women delay pregnancy for at least 1 postoperative year? Am J Perinatol 2008;25(6): 331–3. [32] Karmon A, Levy A, Holcberg G, Wiznitzer A, Mazor M, Sheiner E. Decreased perinatal mortality among women with diet-controlled gestational diabetes mellitus. Int J Gynecol Obstet 2009;104(3):199–202. [33] Tsur A, Sergienko R, Wiznitzer A, Zlotnik A, Sheiner E. Critical analysis of risk factors for shoulder dystocia. Arch Gynecol Obstet 2012;285(5):1225–9. [34] Wispelwey BP, Sheiner E. Cesarean delivery in obese women: a comprehensive review. J Matern Fetal Neonatal Med 2013;26(6):547–51. [35] Dixon AF, Dixon JB, O'Brien PE. Laparoscopic adjustable gastric banding induces prolonged satiety: a randomized blind crossover study. J Clin Endocrinol Metab 2005;90(2):813–9. [36] Maggard MA, Shugarman LR, Suttorp M, Maglione M, Sugerman HJ, Livingston EH, et al. Meta-analysis: surgical treatment of obesity. Ann Intern Med 2005;142(7): 547–59. [37] Dell’Agnolo CM, Carvalho MD, Pelloso SM. Pregnancy after bariatric surgery: implications for mother and newborn. Obes Surg 2011;21(6):699–706.
Contents lists available at ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
REVIEW ARTICLE
A meta-analysis of maternal and fetal outcomes of pregnancy after bariatric surgery Xiao-yan Yi a, Qi-fu Li a, Jun Zhang b, Zhi-hong Wang a,⁎ a b
Department of Endocrinology, First Affiliated Hospital of Chongqing Medical University, Chongqing, China Department of General Surgery, First Affiliated Hospital of Chongqing Medical University, Chongqing, China
a r t i c l e
i n f o
Article history: Received 9 October 2014 Received in revised form 7 January 2015 Accepted 16 March 2015 Keywords: Bariatric surgery Meta-analysis Obesity Outcome Pregnancy
a b s t r a c t Background: The effects of bariatric surgery (BS) on outcomes in subsequent pregnancies are unclear. Objectives: To compare maternal and fetal outcomes among women who become pregnant after BS and obese women who have not undergone BS before pregnancy. Search strategy: PubMed and Embase were searched for relevant reports, and the reference lists of identified articles were hand-searched. Selection criteria: Cohort studies that compared outcomes among women who had undergone any type of BS and obese women who had not undergone surgery were included when results were reported as risk ratios or odds ratios (ORs). Data collection and analysis: Summary ORs were estimated using a random effects model. Main results: Eleven studies were included. Compared with obese women who had not undergone BS, women who had undergone BS had significantly lower odds of gestational diabetes (OR 0.31; 95% CI 0.15–0.65), hypertensive disorders (OR 0.42; 95% CI 0.23–0.78), and macrosomia (OR 0.40; 95% CI 0.24–0.67). However, their odds of small-for-gestationalage newborns were increased (OR 2.16; 95% CI 1.28–3.66). No significant differences were recorded for cesarean, postpartum hemorrhage, and preterm delivery. Conclusions: BS reduces the odds of some adverse maternal and fetal outcomes among obese women. © 2015 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction The prevalence of obesity is a public health concern worldwide [1]. Obesity is associated with an increased risk of obstetric complications, such as gestational diabetes, pre-eclampsia, eclampsia, macrosomia, and fetal growth restriction [2,3]. Therefore, in view of the substantial increase in the prevalence of obesity among women of child-bearing age, a rise in the frequency of adverse maternal and fetal outcomes should be expected [4]. Weight loss can reduce the likelihood of these adverse outcomes [5]. However, dietary changes, exercise, and medical management result in only short-term benefits, which are not sustained in the long term [6]. Bariatric surgery (BS) is thought to be an effective intervention to sustain weight loss [7]. BS procedures are generally categorized into three groups. Restrictive procedures (e.g. laparoscopic adjustable gastric banding [LAGB] and sleeve gastrectomy [SG]) lead to weight loss by reducing gastric capacity which in turn restricts energy intake [8,9]. Malabsorptive procedures (e.g. biliopancreatic diversion [BPD])
⁎ Corresponding author at: Department of Endocrinology, First Affiliated Hospital of Chongqing Medical University, Chongqing 400016, China. Tel.: + 86 23 89011554; fax: +86 23 89011552. E-mail address: [email protected] (Z. Wang).
lead to weight loss by restricting absorption of nutrients [10]. Finally, malabsorptive and restrictive procedures (e.g. Roux-en-Y gastric bypass [RYGB]) reduce stomach capacity, thereby causing malabsorption and a certain degree of restriction of food intake [8]. BPD is rarely used because it is associated with substantial long-term complications—hepatic failure, calcium oxalate kidney stones, renal failure, arthritis, and malnutrition [11]. The most performed procedures today are LAGB and RYGB, although SG is becoming the principal treatment option in many countries for obese women [9]. Although these procedures are beneficial in terms of weight reduction, reports of their implications on maternal and fetal outcomes have been inconsistent. The aim of the present meta-analysis was to compare maternal and fetal outcomes among women who have undergone BS with those among obese women who have not undergone BS. 2. Materials and methods 2.1. Search strategy The PubMed and Embase databases were searched from inception to October 7, 2014, with the keywords “bariatric surgery,” “pregnancy,” “obstetric,” “maternal,” “neonatal,” “perinatal,” and “fetal.” There were no language restrictions. The references of identified articles were hand-searched for further relevant reports.
http://dx.doi.org/10.1016/j.ijgo.2015.01.011 0020-7292/© 2015 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
4
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
2.2. Study selection Cohort studies that reported on maternal and/or fetal outcomes among pregnant women after BS and among obese women before/ without BS were included if the results were reported as risk ratios (RRs) or odds ratios (ORs) with corresponding 95% confidence intervals (CIs). Studies were excluded when they were of any type other than cohort studies, they compared deliveries before and after BS in the same individuals, data were not available, or they included women who were not obese or who underwent BS during pregnancy. If two studies included overlapping populations, the study for which most information was available was included and the other was excluded. If similar studies from the same authors in which data were duplicated were identified, the largest was included and the others were excluded.
Records identified (n=1004) In electronic databases (n=991) Through hand searching (n=13) Duplicates excluded (n=223) Unique records (n=781) Excluded on basis of titles and abstracts (n=760) Full-text review (n=21) Excluded (n=10) Study design (n=1) Overlapping population (n=3) Unavailable data (n=1) Not all women who underwent bariatric surgery were obese (n=4) Some women who underwent bariatric surgery were pregnant (n=1)
2.3. Data extraction and quality assessment For all included studies, two reviewers (X-y.Y. and Q-f.L.) independently extracted the first author’s name, country, publication year, sample size, maternal age, BMI, type of procedures, time from surgery to conception (S-C time), and maternal and/or fetal outcomes. When data were missing, J.Z. would email the authors of the relevant articles. A body mass index (BMI; calculated as weight in kilograms divided by the square of height in meters) of 30 or greater was considered obese. BS included any type of weight-loss procedure. The maternal outcomes assessed were gestational diabetes mellitus (GDM), hypertensive disorders (including gestational hypertension, pre-eclampsia, and eclampsia), postpartum hemorrhage, and cesarean delivery. The fetal outcomes assessed were preterm delivery, macrosomia, and being small for gestational age (SGA). The quality of the included studies was evaluated by the Newcastle– Ottawa Scale, with some modifications to match the needs of the present review. The highest score was nine points. Disagreements were resolved by discussion. 2.4. Data analysis The meta-analysis was performed using Stata 12.0 (StataCorp, College Station, TX, USA). Summary ORs were estimated using a random effects model by the Mantel-Haenszel method. Because the OR is equivalent to the RR when events are rare, it was possible to interpret the OR as the RR. Heterogeneity was assessed with the test of inconsistency (I2): when the value was greater than 50%, it was deemed statistically significant. Meta-regression, and sensitivity and subgroup analyses were planned to identify possible sources of the between-study heterogeneity if necessary and possible. The subgroups assessed were maternal age (≤ 32 years vs N32 years), type of BS (restrictive vs malabsorptive and restrictive vs mixed), and S-C time (≤ 2 years vs N2 years). The publication bias of included studies was assessed using the funnel plot with the Begg and Egger tests. A two-sided P value of less than 0.05 was considered statistically significant. The metaanalysis was conducted according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) criteria [12]. 3. Results 3.1. Description of studies A total of 781 publications were identified, of which 21 underwent full-text review (Fig. 1). After exclusion of 10 studies (Supplementary Material S1), 11 [13–23] were included in the meta-analysis (Table 1). The outcomes analyzed for each study are shown in Supplementary Material S2, along with the definitions used for each of the outcomes. Two studies from the USA [13,15] had overlapping populations, but they were included because data for hypertensive disorders were
Included in meta-analysis (n=11)
Fig. 1. Identification of eligible articles.
extracted from one [13] and data for GDM and postpartum hemorrhage were extracted for the other [15]. Among the 11 included studies, five were from Europe, four from the USA, one from the Middle East, and one from Australia (Table 1). The mean age of participants was generally older than 30 years. Among the studies reporting relevant data, generally more than half the participants had given birth previously (Table 1). The procedures performed varied; some studies stated that they included only LAGB or RYGB [16–18,20,22]. Among women who had undergone BS, there was an apparent reduction in average BMI from 40–50 to 32–35 (Table 1). Additionally, birth weight was generally lower for the neonates delivered by women who had undergone BS than for those delivered by women who had not undergone BS (Table 1). Only four studies [16,18,20,22] mentioned post-surgery nutritional recommendations and follow-up. Results of quality assessment are shown in Supplementary Material S3. 3.2. GDM Nine articles that reported the maternal outcome of GDM in women with or without surgery [14–20,22,23] were included in the analysis of GDM, with 711 cases overall. The random effects model showed that GDM was significantly less likely among women who had undergone BS than among those who had not (pooled OR 0.31, 95% CI 0.15–0.65; I2 = 85.2%, P b 0.001) (Fig. 2A). No publication bias was found with either the Begg (0.602) or Egger (P = 0.240) tests. Because the heterogeneity was significant, a sensitivity analysis was performed for GDM. It showed that no article significantly affected the results (data not shown). Subgroup analysis indicated that some heterogeneity could be a result of differences in S-C time. In some studies, the time from surgery to delivery (S-D time) was provided instead of the S-C time. For such studies, an assumption was made that their mean gestational age was 40 weeks, and S-C time was estimated as SD time minus the length of pregnancy (Supplementary Material S4). Among studies in which women who had undergone BS conceived up to 2 years after surgery, GDM was significantly less likely among women who had undergone BS than among those who had not (OR 0.22, 95% CI 0.14–0.34; I2 = 0.0%, P = 0.844). Similar results were recorded for the subgroup analyses including studies of women who conceived more than 2 years after surgery (OR 0.19, 95% CI 0.04–0.94; I2 = 70.8%, P = 0.016) (Supplementary Material S5).
5
NR 3258/3577 NR 3397/3297 3271/3305 3178/3373 2951/3463 2951/3240 NR 2948/3442 3206/3454
3.3. Hypertensive disorders
NR NR NR 9.6/15.5 5.5/7.1 6.1/6.4 12.0/11.6 14.6/9.1 NR NR NR
Nine studies [14–20,22,23] were included in the analysis of hypertensive disorders, with 594 patients with hypertensive disorders among 4993 participants overall. Hypertensive disorders were less likely among women who had undergone BS than among those who had not (summary OR 0.42, 95% CI 0.23–0.78; I2 = 83.3%, P b 0.001) (Fig. 2B). There was no publication bias noted on Begg (1.000) or Egger (P = 0.757) tests. The sensitivity analysis presented a robust result that was not influenced by individual studies (data not shown). Subgroup analysis indicated that some heterogeneity could be a result of differences in SC time. Hypertensive disorders were significantly less likely among women who underwent BS and conceived up to 2 years after surgery than among those who had not undergone surgery (OR 0.14, 95% CI 0.08–0.25; I2 = 26.1%, P = 0.258). Similarly, women who had conceived more than 2 years after surgery had a reduced likelihood of hypertensive disorders (OR 0.45, 95% CI 0.29–0.72; I2 = 0.0%, P = 0.530) (Supplementary Material S5).
23.6 mo NR 20.9 mo 20 mo NR NR NR NR 5.2 y NR NR
3.4. Postpartum hemorrhage
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters); NR, not reported. a Values are given as women who have undergone bariatric surgery/obese women who have not undergone bariatric surgery. b Exposed group only. c ≥2 categories of bariatric surgery.
NA NR NR NR 17 mo 3.1 y 3.61 y 25.4 mo NR 26.6 mo NR NR/N35 31.6/31.2 NR/N35 NR/N35 34.8/35.0 35.0/39.4 33.7/48.4 32.5/39.0 32.5/38.5 32.7/32.7 NR/N30 N35 NR N35 45.9 44 45.2 49.31 NR N35 50.4 NR NR 42.5/42.7 NR 49.4/49.4 46.1/35.0 NR 48.6/42.9 NR NR 33.3/35.8 0/0.7 2010 2014 2010 2005 2007 2010 2012 2008 2013 2010 2014 Bennett et al. [13] Berlac et al. [14] Burke et al. [15] Dixon et al. [16] Ducarme et al. [17] Lapolla et al. [18] Lesko, Peaceman [19] Patel et al. [20] Roos et al. [21] Santulli et al. [22] Shai et al. [23]
USA Denmark USA Australia France Italy USA USA Sweden France Israel
316/269 413/823 354/346 79/79 13/414 83/120 70/140 26/66 2474/12 027 24/120 326/1612
32.5/31.3 31.2/31.1 32.5/31.1 29.9/30.9 31.5/31 31.4/33.0 33.2/30.5 34.1/30.6 NR 31.7/31.7 32.9/32.3
Mixedc Malabsorptive–restrictive Mixedc Restrictive Restrictive Restrictive Mixedc Malabsorptive–restrictive Mixedc Malabsorptive–restrictive Restrictive
Mean S-C timeb Mean BMI before surgeryb Nulliparous, %a Country Year Study
Table 1 Characteristics of the 11 included studies.
Sample sizea
Mean maternal age, ya
Type of bariatric surgery
Mean BMI before pregnancya
Mean S-D timeb
Mean weight gain during pregnancy, kga
Mean birth weight, ga
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
Six studies [14,15,17,19,22,23] were included in the analysis of postpartum hemorrhage, with 181 cases in a sample of 4655 participants. Postpartum hemorrhage was less likely among women who had undergone BS than among those who had not, but the difference was not significant (summary OR 0.62, 95% CI 0.38–1.00; I2 = 8.7%, P = 0.361) (Fig. 2C). No publication bias was observed with Begg (1.000) or Egger (P = 0.439) tests, and a sensitivity analysis did not show a result sensitive to any individual studies (data not shown). No subgroup analysis was undertaken because between-study heterogeneity was not significant. 3.5. Cesarean delivery Seven studies [14,15,17–20,22] were included in the analysis of incidence of cesarean delivery. Cesarean was not significantly more likely among women who had undergone BS than among those who had not (summary OR 0.75, 95% CI 0.50–1.13; I2 = 74.2%, P = 0.001) (Fig. 2D). However, cesarean deliveries were common in both groups, which means that the OR would differ from the RR. Therefore, this result should be interpreted with caution. No publication bias was observed with Begg (1.000) or Egger (P = 0.701) tests, and a sensitivity analysis presented a robust result (data not shown). Subgroup analysis did not give any significant results because of the small number of studies (data not shown). 3.6. Preterm birth Eight studies [16–23] investigating the risk of preterm birth in women who had and had not undergone BS were included, with a total of 1366 preterm deliveries among 17 673 participants. Preterm birth was significantly more likely among women who had undergone BS than among those who had not (summary OR 1.33, 95% CI 1.16–1.52; I2 = 0.0%, P = 0.695). No publication bias was found with Begg (0.386) or Egger (P = 0.448) tests. In a sensitivity analysis, the study by Roos et al. [21] was shown to have a weight of 81.0% and influenced the results significantly. When this article was omitted, the difference between groups was no longer significant (summary OR 1.18, 95% CI 0.86–1.61; I2 = 0.0%, P = 0.671) (Fig. 3A). No subgroup analysis was undertaken because betweenstudy heterogeneity was not significant. 3.7. Macrosomia Six studies [16–20,23] concerning the risk of macrosomia in women with or without BS were included, with 368 macrosomia cases among 3028 participants. Macrosomia was significantly less likely among
6
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
Fig. 2. Forest plots of maternal outcomes. (A) Gestational diabetes. (B) Hypertensive disorders. (C) Postpartum hemorrhage. (D) Cesarean delivery. Abbreviations: OR, odds ratio; CI, confidence interval.
women who had undergone BS than among those who had not (summary OR 0.40, 95% CI 0.24–0.67; I2 = 23.9%, P = 0.255) (Fig. 3B). No publication bias was found with either Begg (1.000) or Egger (P = 0.925) tests. No study was found to influence the results significantly in a sensitivity analysis (data not shown). No subgroup analysis was undertaken because between-study heterogeneity was not significant.
3.8. SGA Five studies [18–22] were included in the analysis of incidence of SGA, with a total of 479 cases among 15 098 participants. An SGA newborn was significantly more likely among women who had undergone BS than among those who had not (summary OR 2.16, 95% CI
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
7
Fig. 3. Forest plots of fetal outcomes. (A) Preterm delivery. (B) Macrosomia. (C) SGA. Abbreviations: SGA, small for gestational age; OR, odds ratio; CI, confidence interval.
1.28–3.66; I2 = 27.8%, P = 0.236) (Fig. 3C). No publication bias was observed with Begg (0.806) or Egger (P = 0.949) tests, and a sensitivity analysis presented a robust result (data not shown). No subgroup analysis was undertaken because between-study heterogeneity was not significant.
4. Discussion In the present meta-analysis, women who had undergone BS had a reduced risk of GDM, hypertensive disorders, and macrosomia when compared with obese women who had not undergone BS. However, they had an increased risk of SGA newborns. No significant differences were recorded for cesarean delivery, postpartum hemorrhage, and preterm delivery. GDM and hypertensive disorders are two important obstetric complications. In the present study, women who became pregnant after BS had at least 50% lower odds of these complications than did obese women without a history of a bariatric procedure. However, significant heterogeneity was also observed for these two outcomes. Because of the limited data available, subgroup analyses could only be stratified by age, type of bariatric surgery, and S-C time. Heterogeneity was
lower in the subgroup analyses by S-C time for both GDM and hypertensive disorders. According to our findings, pregnancy soon after weight-loss surgery (≤2 years) was associated with a reduced incidence of GDM and hypertensive disorders. Moreover, pregnancy soon after BS had an even greater reduction in the odds of hypertensive disorder than did pregnancy more than 2 years after BS. These findings are in contrast with those of Parikh et al. [24], who reported an increased risk of GDM and decreased weight loss among women who became pregnant within the first 2 years after BS compared with those who became pregnant after 2 years. No significant difference was seen in the incidence of hypertensive disorders or anemia. Kjær et al. [25] and Dao et al. [26] did not observe any differences regarding maternal or fetal outcomes according to timing of pregnancy after surgery. In line with the present findings, Patel et al. [20] found adverse pregnancy outcomes were significantly less likely during the first 2 years. Nonetheless, they admitted that the women who became pregnant during this period required oral protein supplementation more frequently than did those who became pregnant later. Thus, they recommended postponing pregnancy for 2 years after BS. A woman is in a rapid weight-loss phase during the first 12 months after BS; the rate of weight loss decreases in the second year and then
8
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9
tends to stabilize [27,28]. Most researchers believe that pregnancy during the rapid weight-loss period would raise the risks of adverse maternal and fetal outcomes [24,29,30]. Therefore, it is generally advised to delay pregnancy until at least 1 year after BS [27]. However, no conclusive evidence exists suggesting that pregnancy during the first postoperative year is unsafe [31]. On the basis of the present analysis, an early pregnancy might not be related to a higher risk of adverse outcomes, especially when adequate nutritional was supplemented. Further studies are required to better understand the influence of pregnancy timing on obstetric outcomes and set the best timeline for conception after BS. It is known that GDM increases the risk of macrosomia [32]. Macrosomia is related to serious obstetric problems, including shoulder dystocia, an increased rate of cesarean delivery, and postpartum hemorrhage [33,34]. As for GDM, the present review has shown that odds of macrosomia decreased by more than 50% after BS. However, no significant difference was recorded for cesarean delivery or postpartum hemorrhage. BS leads to weight loss through restriction of energy intake and malabsorption [28,35,36]. Therefore, risks of fetal growth restriction are a concern. In most of the studies included in the present analysis, birth weight was lower among women who had undergone BS than among obese women who had not. Meanwhile, risk of SGA was significantly increased among women who had undergone BS. In four of the included studies, mean S-C time was longer than 24 months, which indicates that malnutrition can still be severe even when conception occurs 2 years after BS, when weight loss should have slowed down. The frequency of cesarean among women who had undergone BS varied between 15.4% [17] and 61.5% [20], compared with between 29.8% [14] and 65.8% [18] of obese women who have not undergone BS. The difference in frequency of cesarean between women who had and had not undergone BS was even bigger in Dell’Agnolo et al.’s study [37]. Importantly, the difference between groups might be affected more by the local recommendations, economic benefits, and previous cesareans than by BS. Therefore, the present analysis concerning this outcome might not accurately reflect the actual influence of BS on cesarean delivery. There are several limitations of the present meta-analysis. Firstly, the included articles were all observational studies. The possibilities of selection and confounding bias are unavoidable. Indeed, patients who chose to undergo BS could differ from patients who did not undergo the procedure in terms of socioeconomic status, access to healthcare services, attention to their own health, and good after-surgery care. Alternatively, patients who have undergone BS are more likely to be severely obese or have increased comorbidity than are obese women who have not undergone this procedure. Only three included articles mentioned that patients in the unexposed group met the criteria for BS. As to confounding bias, all the confounders could not be extracted to allow an assessment of which one influenced the results most, because of unavailable detailed data. In addition to the subgroup analyses performed (maternal age, type of bariatric surgery, and S-C time), differences in obesity, BMI level, parity, previous cesarean delivery, multiple birth, adequate nutrition supplement after surgery, and weight gain during pregnancy could all bias results. Secondly, small sample sizes could limit the strength of the conclusions, especially those for the individual outcomes. One study by Ducarme et al. [17] included only 13 women who had undergone BS compared with 414 obese women who had not undergone such a procedure. However, when that study was excluded during sensitivity analysis, the results were robust. Another five of the eleven included articles included fewer than 100 women who had undergone BS. Finally, only 11 studies were included. Additionally, most of the studies were from high-income countries. In summary, BS reduces the odds of some adverse maternal-fetal outcomes among pregnant women. Compared with obese women who had not undergone BS, women who became pregnant after BS
had a lower risk of GDM, hypertensive disorders, and macrosomia. However, risk of SGA newborns increased after BS. Therefore, women who have undergone BS and subsequently become pregnant need to receive adequate supplements and their nutritional status should be closely monitored to prevent fetal growth restriction. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijgo.2015.01.011.
Acknowledgments The present analysis was supported by research grants from the Project of National Clinical Key Specialties Construction of China (2011), the General Program of Chongqing Municipal Health Bureau (2011-2-062), the Natural Science Foundation of Chongqing (CSTC2012jjA10040), and the Tackling Project of Science and Technology of Chongqing Committee of Science and Technology (CSTC2012ggyyjs10038). Conflict of interest The authors have no conflicts of interest.
References [1] Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. JAMA 2005;293(15):1861–7. [2] Weiss JL, Malone FD, Emig D, Ball RH, Nyberg DA, Comstock CH, et al. Obesity, obstetric complications and cesarean delivery rate–a population-based screening study. Am J Obstet Gynecol 2004;190(4):1091–7. [3] Eerdekens A, Debeer A, Van Hoey G, De Borger C, Sachar V, Guelinckx I, et al. Maternal bariatric surgery: adverse outcomes in neonates. Eur J Pediatr 2010;169(2): 191–6. [4] 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(8):1175–82. [5] Cnattingius S, Bergström R, Lipworth L, Kramer MS. Prepregnancy weight and the risk of adverse pregnancy outcomes. N Engl J Med 1998;338(3):147–52. [6] Catalano PM. Management of obesity in pregnancy. Obstet Gynecol 2007;109(2 Pt 1):419–33. [7] Buchwald H, Avidor Y, Braunwald E, Jensen MD, Pories W, Fahrbach K, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;292(14):1724–37. [8] Guelinckx I, Devlieger R, Vansant G. Reproductive outcome after bariatric surgery: a critical review. Hum Reprod Update 2009;15(2):189–201. [9] Buchwald H, Oien D. Metabolic/bariatric surgery worldwide 2008. Obes Surg 2009; 19(12):1605–11. [10] Scopinaro N, Gianetta E, Civalleri D, Bonalumi U, Bachi V. Bilio-pancreatic bypass for obesity: II.Initial experience in man. Br J Surg 1979;66(9):618–20. [11] Livingston EH. Obesity and its surgical management. Am J Surg 2002;184(2): 103–13. [12] Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Metaanalysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000; 283(15):2008–12. [13] Bennett WL, Gilson MM, Jamshidi R, Burke AE, Segal JB, Steele KE, et al. Impact of bariatric surgery on hypertensive disorders in pregnancy: retrospective analysis of insurance claims data. BMJ 2010;340:c1662. [14] Berlac JF, Skovlund CW, Lidegaard O. Obstetrical and neonatal outcomes in women following gastric bypass: a Danish national cohort study. Acta Obstet Gynecol Scand 2014;93(5):447–53. [15] Burke AE, Bennett WL, Jamshidi RM, Gilson MM, Clark JM, Segal JB, et al. Reduced incidence of gestational diabetes with bariatric surgery. J Am Coll Surg 2010;211(2): 169–75. [16] Dixon JB, Dixon ME, O'Brien PE. Birth outcomes in obese women after laparoscopic adjustable gastric banding. Obstet Gynecol 2005;106(5 Pt 1):965–72. [17] Ducarme G, Revaux A, Rodrigues A, Aissaoui F, Pharisien I, Uzan M. Obstetric outcome following laparoscopic adjustable gastric banding. Int J Gynecol Obstet 2007; 98(3):244–7. [18] Lapolla A, Marangon M, Dalfrà MG, Segato G, De Luca M, Fedele D, et al. Pregnancy outcome in morbidly obese women before and after laparoscopic gastric banding. Obes Surg 2010;20(9):1251–7. [19] Lesko J, Peaceman A. Pregnancy outcomes in women after bariatric surgery compared with obese and morbidly obese controls. Obstet Gynecol 2012;119(3):547–54. [20] Patel JA, Patel NA, Thomas RL, Nelms JK, Colella JJ. Pregnancy outcomes after laparoscopic Roux-en-Y gastric bypass. Surg Obes Relat Dis 2008;4(1):39–45. [21] Roos N, Neovius M, Cnattingius S, Trolle Lagerros Y, Sääf M, Granath F, et al. Perinatal outcomes after bariatric surgery: nationwide population based matched cohort study. BMJ 2013;347:f6460.
X. Yi et al. / International Journal of Gynecology and Obstetrics 130 (2015) 3–9 [22] Santulli P, Mandelbrot L, Facchiano E, Dussaux C, Ceccaldi PF, Ledoux S, et al. Obstetrical and neonatal outcomes of pregnancies following gastric bypass surgery: a retrospective cohort study in a French referral centre. Obes Surg 2010;20(11):1501–8. [23] 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-based study. J Matern Fetal Neonatal Med 2014;27(3):275–8. [24] Parikh R, Lavoie M, Wilson L, Maranda L, Leung K, Simas TM. Timing of pregnancy after bariatric surgery. Obstet Gynecol 2014;123(Suppl. 1):165S. [25] Kjær MM, Nilas L. Timing of pregnancy after gastric bypass-a national register-based cohort study. Obes Surg 2013;23(8):1281–5. [26] Dao T, Kuhn J, Ehmer D, Fisher T, McCarty T. Pregnancy outcomes after gastricbypass surgery. Am J Surg 2006;192(6):762–6. [27] American College of Obstetricians and Gynecologists. ACOG practice bulletin no. 105: bariatric surgery and pregnancy. Obstet Gynecol 2009;113(6):1405–13. [28] 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(8):741–52. [29] Martin LF, Finigan KM, Nolan TE. Pregnancy after adjustable gastric banding. Obstet Gynecol 2000;95(6 Pt 1):927–30. [30] Wittgrove AC, Jester L, Wittgrove P, Clark GW. Pregnancy following gastric bypass for morbid obesity. Obes Surg 1998;8(4):461–4.
9
[31] Karmon A, Sheiner E. Timing of gestation after bariatric surgery: should women delay pregnancy for at least 1 postoperative year? Am J Perinatol 2008;25(6): 331–3. [32] Karmon A, Levy A, Holcberg G, Wiznitzer A, Mazor M, Sheiner E. Decreased perinatal mortality among women with diet-controlled gestational diabetes mellitus. Int J Gynecol Obstet 2009;104(3):199–202. [33] Tsur A, Sergienko R, Wiznitzer A, Zlotnik A, Sheiner E. Critical analysis of risk factors for shoulder dystocia. Arch Gynecol Obstet 2012;285(5):1225–9. [34] Wispelwey BP, Sheiner E. Cesarean delivery in obese women: a comprehensive review. J Matern Fetal Neonatal Med 2013;26(6):547–51. [35] Dixon AF, Dixon JB, O'Brien PE. Laparoscopic adjustable gastric banding induces prolonged satiety: a randomized blind crossover study. J Clin Endocrinol Metab 2005;90(2):813–9. [36] Maggard MA, Shugarman LR, Suttorp M, Maglione M, Sugerman HJ, Livingston EH, et al. Meta-analysis: surgical treatment of obesity. Ann Intern Med 2005;142(7): 547–59. [37] Dell’Agnolo CM, Carvalho MD, Pelloso SM. Pregnancy after bariatric surgery: implications for mother and newborn. Obes Surg 2011;21(6):699–706.
