Professional Documents
Culture Documents
TAGEDPABSTRACT Study Objective: To compare fetal, maternal, and operative outcomes of laparoscopic surgery versus laparotomy for major
benign diseases including appendicitis, cholecystitis, adnexal masses, and uterine myoma during pregnancy.
Design: Retrospective cohort study (Canadian Task Force classification II-2).
Setting: The Diagnosis Procedure Combination database, a national inpatient database for acute care inpatients in Japan.
Patients: Eligible patients (n = 6018) underwent abdominal surgery (4047 laparotomy and 1971 laparoscopy patients) from
July 2010 through March 2016.
Interventions: The 2 groups were compared using propensity score matching analysis.
Measurements and Main Results: The primary outcome was fetal adverse events, including abortion or stillbirth within
7 days after surgery and premature delivery during hospitalization. Secondary outcomes were operative time, blood transfu-
sion, and length of hospital stay after surgery. Propensity score matching created 740 pairs. Significant difference was
observed in the primary outcome between propensity score matched patients in the laparotomy versus laparoscopy group
(1.8% vs .41%, respectively; risk difference, ¡1.4%; 95% confidence interval, ¡2.4 to ¡.30; p = .01). Compared with the
laparotomy group, the laparoscopy group had a significantly lower incidence of blood transfusion (2.3% vs .41%, p = .002),
shorter operative time (115 vs 95 minutes, p <.001), and shorter hospital stay (9.2 vs 5.9 days, p <.001).
Conclusion: Our current study using propensity score matching suggests the advantages of laparoscopic surgery for
benign diseases compared with laparotomy because laparoscopic surgery had advantages in short-term fetal adverse
events, incidence of blood transfusion, operative time, and hospital stay. Journal of Minimally Invasive Gynecology
(2018) 00, 1 6 © 2018 AAGL. All rights reserved.
Keywords: Adnexal mass resection; Appendectomy; Cholecystectomy; Laparoscopy; Pregnancy; Uterine myomectomy
Laparoscopic surgery during pregnancy has become reviews and meta-analyses have confirmed the advantages
widely used in various fields; however, open surgery is still of laparoscopic surgery in pregnant women [1 6]. The
performed because of the potential risks of laparoscopic Society of American Gastrointestinal and Endoscopic Sur-
surgery for pregnant women and fetuses. Systematic geons guideline has been widely used in surgery during
pregnancy [7].
However, most previous studies analyzed only a small
Dr. Yasunaga received grants from the Ministry of Health, Labour and
Welfare, Japan (H29-Policy-Designated-009 and H29-ICT-Genral-004);
number of pregnant patients, and each study reported on a
Ministry of Education, Culture, Sports, Science and Technology, Japan specific disease, such as appendicitis [5,6], adnexal masses
(17H04141); and the Japan Agency for Medical Research and Develop- [1,3,4], or gallbladder disease [2]. Recent reviews and a
ment (15lk1110001h0001). Dr. Fushimi received a grant from the Ministry meta-analysis on adnexal masses included no more than
of Health, Labour and Welfare, Japan (H29-Policy-Designated-009). The 500 to 600 patients [1,4]; the Cochrane Library in 2013
funding organizations had no role in planning research, analysis, interpre-
tation of results, or report writing.
reported no randomized controlled trials on the subject
Corresponding author: Daisuke Shigemi, MD, MPH, Department of Clini- since 2006 [3]. A systematic review and meta analysis on
cal Epidemiology and Health Economics, School of Public Health, The appendectomy in pregnancy analyzed 3415 patients; how-
University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo 1130033, Japan. ever, only 599 of the patients underwent laparoscopic sur-
E-mail: dshigemi@m.u-tokyo.ac.jp gery [5]. The aim of the current study was to compare
Submitted April 25, 2018, Accepted for publication June 8, 2018. maternal, fetal, and operative outcomes after laparoscopy
Available at www.sciencedirect.com and www.jmig.org versus laparotomy for appendicitis, cholecystitis, uterine
1553-4650/$ — see front matter © 2018 AAGL. All rights reserved.
https://doi.org/10.1016/j.jmig.2018.06.008
2 Journal of Minimally Invasive Gynecology. Vol 00, No 00, 2018
myoma, and adnexal masses during pregnancy, using a The primary outcome was fetal adverse events, includ-
nationwide Japanese inpatient database. ing abortion or stillbirth within 7 days after surgery and pre-
mature delivery (<37 weeks’ gestation) during
hospitalization. Secondary outcomes were operative and
maternal outcomes, including operative time, blood transfu-
Methods
sion, and length of hospital stay after surgery. In addition,
In this retrospective cohort study, we used the Diagnosis major operative complications, including bowel injury
Procedure Combination database, a national inpatient data- (International Classification of Diseases, Tenth Revision:
base for acute care inpatients in Japan, the details of which S367, S369), bladder injury (S372, S371), and pulmonary
have been described elsewhere [8]. The database was estab- thromboembolism (I269), were evaluated.
lished in 2003 with 82 academic hospitals; the number of Categorical variables (age, gestational age at surgery,
participating hospitals has gradually increased since that smoking, operative procedure, general anesthesia, emergent
time. As of 2016 more than 1000 hospitals, including all 82 maternal transport, and emergent surgery within 2 days of
academic hospitals and more than 90% of all tertiary emer- admission) were compared with the x2 or Fisher’s exact
gency centers, had adopted the database. Because the Diag- test. The average proportion of surgeries performed lapa-
nosis Procedure Combination participating hospitals are roscopically at each hospital and BMI were compared
generally large facilities, the number of included patients between the laparotomy and laparoscopy groups with a t
was approximately 7 million in 2016, representing approxi- test.
mately 50% of all patients admitted to hospitals in Japan. We conducted propensity score matching between the
Attending physicians are required to record disease diagno- laparotomy and laparoscopy groups [10]. We estimated
ses accurately because the diagnostic records are linked to the propensity scores with a logistic regression model for
a payment system that is based on medical insurance. A the likelihood of laparoscopy as a function of the follow-
previous study showed that the validity of diagnostic ing parameters: age, BMI, gestational age at surgery,
records in the Diagnosis Procedure Combination database smoking, operative procedure, emergent maternal trans-
was generally high. The sensitivity and specificity of the port, emergent surgery within 2 days of admission, and
primary diagnoses were 78.9% and 93.2%, respectively [9]. the average ratio of laparoscopies to laparotomies per-
The database includes the following data: dates of admis- formed at each hospital.
sion and discharge, patient age and sex, body weight and The C-statistic was calculated with the area under the
height, primary and secondary diagnoses, pre-existing receiver operating characteristic curve to evaluate the
comorbidities at admission and complications after admis- model predicting the receipt of laparoscopic surgery. We
sion, procedures performed, medications and devices used, performed 1-to-1 propensity score matched analysis using
in-hospital mortality, pregnancy status (pregnant or not), nearest neighbor matching without replacement within a
gestational age at admission, and delivery during hospitali- caliper of .25 standard deviations of pooled propensity
zation. Diagnoses, comorbidities, and complications are scores [11].
recorded using the codes of the International Classification After matching, we assessed the balance of baseline var-
of Diseases, Tenth Revision; text data are in Japanese. The iables in the matched patient groups with the standardized
database does not contain any laboratory data or obstetric difference. An absolute standardized difference > 10%
examination findings (including Bishop score, uterine cer- indicated a meaningful imbalance [12].
vical length, or vaginal bacteriologic culture). We compared outcomes between the laparotomy and
We identified pregnant women aged 13 to 53 years with laparoscopy groups with the x2 test. We estimated risk dif-
complete data who underwent abdominal surgery before 37 ferences and 95% confidence intervals for the primary out-
weeks of gestation and who were admitted to the hospital come and 3 secondary outcomes.
from July 2010 to March 2016. Patients who underwent All statistical analyses were performed with Stata soft-
abdominal surgery at the time of cesarean delivery were ware version 15.0 (StataCorp LP, College Station, TX). All
not included. In this study abdominal surgeries (by laparos- tests were 2-tailed, and the threshold for significance was
copy or laparotomy) included 4 operative procedures for p <.05.
benign diseases: appendectomy, cholecystectomy, uterine Written informed consent was not required because of
myomectomy, and adnexal mass resection. the anonymous nature of the data. The need for written con-
Patient age was categorized as 19, 20 to 29, 30 to 39, sent was formally waived by the ethics committee. The
or 40 years. Body mass index (BMI) was not classified Institutional Review Board at The University of Tokyo
according to the World Health Organization classification approved the study.
because all patients in the current study were pregnant. Ges-
tational age at surgery was categorized as 14, 15 to 27, or
Results
28 to 36 weeks.
Exclusion criteria included multiple pregnancy, ectopic We identified 9065 patients who underwent abdominal
pregnancy, and impaired consciousness at admission. surgery during pregnancy in the study period. A total of
Shigemi et al. Safety of Laparoscopic Surgery in Pregnancy 3
Fig. 1
Study flow diagram showing the stratification and selection of pregnant patients in the Diagnosis Procedure Combination database who underwent
abdominal surgery from 2010 to 2016.
6018 eligible patients were selected, including 4047 lapa- Table 1 shows the baseline patient characteristics in the
rotomy patients and 1971 laparoscopy patients. Propensity unmatched and propensity score matched groups. In the
score matching created 740 pairs (Fig. 1). The C-statistic of unmatched groups the laparotomy group was significantly
the model for calculating propensity scores was .90. more likely to have emergent maternal transport and
Table 1
Baseline patient characteristics in unmatched and propensity score matched groups
Table 2
Incidence of primary outcome in unmatched and propensity score matched laparotomy and laparoscopy groups
Primary outcome is fetal adverse events (abortion or stillbirth within 7 days after surgery and premature delivery).
surgery within 2 days of admission and was less likely to who underwent surgery between 28 and 36 weeks’ gesta-
have general anesthesia than the laparoscopy group. Most tion than among those who underwent surgery before 28
patients in both the laparoscopy and laparotomy groups weeks’ gestation. Among unmatched groups the primary
underwent surgery in the first or second trimester of preg- outcome incidence ranged from 1% to 9% for different pro-
nancy. A higher proportion of laparotomies than laparosco- cedures in the laparotomy group and from 0% to 1% for dif-
pies was performed in the first and third trimesters. ferent procedures in laparoscopy group.
Adnexal mass resection and cholecystectomy were signifi- Table 5 shows operative times stratified according to
cantly more likely to be performed by laparoscopy than operative procedure. Laparoscopy required shorter opera-
by laparotomy; appendectomy and uterine myomectomy tive time than laparotomy for all procedures. Compared
were significantly more likely to be performed by laparot- with appendectomy and adnexal mass resection, cholecys-
omy. The distributions of the variables in the propensity tectomy and myomectomy required longer time in the lapa-
score matched groups were well balanced. rotomy group.
There was a significant difference in the primary out- Among the unmatched groups, there was 1 case of bowel
come (fetal adverse events, including abortion or stillbirth injury and 1 case of bladder injury in the laparoscopy
within 7 days after surgery and premature delivery) group. In the laparotomy group there was 1 case of pulmo-
between the laparotomy and laparoscopy groups before nary thromboembolism. No cases of these operative com-
propensity score matching (1.34% vs .36%, respectively; plications were included in propensity score matched
risk difference, ¡.95%; 95% confidence interval, ¡1.39 to groups.
¡.52). Significant difference was also observed in the pri-
mary outcome between the propensity score matched
Discussion
groups (1.8% vs .41%, respectively; risk difference,
¡1.4%; 95% confidence interval, ¡2.4 to ¡.30) (Table 2). This study revealed that among 6018 pregnant patients
Table 3 shows the results of the t test and x2 test for the who underwent abdominal surgery, including appendec-
secondary outcomes between propensity score matched tomy, cholecystectomy, uterine myomectomy, and adnexal
groups. Compared with the laparotomy group, the laparos- mass resection, there was a significant difference in the
copy group had a significantly lower incidence of blood occurrence of fetal adverse events after laparoscopic sur-
transfusion (2.3% vs .41%, p = .002), shorter operative time gery versus laparotomy, according to propensity score-
(115 vs 95 minutes, p <.001), and shorter hospital stay matched analysis. Among propensity score matched
(9.2 vs 5.9 days, p <.001). patients, the laparoscopy group had a significantly lower
Table 4 shows the primary outcome results stratified incidence of blood transfusion, shorter operative time, and
according to gestational age and operative procedure. The shorter hospital stay compared with the laparotomy group.
primary outcome was more likely to occur among patients Previous meta-analyses and systematic reviews of lapa-
roscopic surgery during pregnancy have included only
small numbers of eligible patients or have contained signifi-
Table 3 cant biases. A meta analysis of abdominal surgery for sus-
pected adnexal masses reported in 2016 included 4
Results of t test and x2 test for secondary outcomes in propensity comparative effectiveness studies, but the study included
score matched groups
only 240 patients [1]. A systematic review and meta anal-
ysis on appendectomy in pregnancy reported in 2012
Propensity score matched groups
included 3415 women (599 in the laparoscopic group) [5];
Laparotomy Laparoscopy however, 1 of the included studies, which comprised most
(n = 740) (n = 740) p
patients in the review (3133), did not provide information
Blood transfusion, n (%) 17 (2.3) 3 (.41) .002 on anesthesia or gestational age at surgery [13].
Operative time, min 114.6 95.4 <.001 The benefits and drawbacks of laparoscopic surgery for
Length of hospital stay 9.2 5.9 <.001
after surgery, days
pregnant women have been discussed for many years. For
pregnant patients who require abdominal surgery,
Shigemi et al. Safety of Laparoscopic Surgery in Pregnancy 5
Table 4
Results of primary outcome stratified according to gestational age and operative procedure in unmatched and propensity score matched laparotomy
and laparoscopy groups
laparoscopy appears to be associated with less risk than lap- 2100 pregnant patients who underwent laparoscopic surgery
arotomy [14,15]. The benefits of laparoscopy during preg- reported no significant differences between the laparoscopy
nancy are similar to those in nonpregnant patients. These and laparotomy groups in any measured outcome, including
include less postoperative pain, less postoperative ileus, birth weight, gestational age, intrauterine growth restriction,
reduced adhesion formation, shorter hospital stay, and congenital malformation, stillbirth, or neonatal death [17].
faster return to usual activities [16]. However, these conclu- Because fetal adverse events, including postoperative
sions have been based on studies with small numbers of abortion, stillbirth, and premature delivery, are major postop-
patients. erative concerns, we chose these as the primary outcome.
The benefits and drawbacks of laparoscopic surgery for The current study showed that the significant advantage of
the fetus have also been frequently discussed. Previous stud- laparoscopy compared with laparotomy for the primary out-
ies have described laparoscopic procedures in all trimesters come in the analysis of unmatched groups was supported by
of pregnancy with minimal fetal morbidity; however, no pro- the analysis of propensity score matched groups. Our
spective study has evaluated the safety of laparoscopic sur- results suggest that laparoscopic surgery can be performed
gery in human pregnancy. The best available outcome data safely for major benign diseases during pregnancy.
are derived from retrospective studies. A retrospective study Operative and maternal outcomes were significantly bet-
from Swedish health registries that included approximately ter in the laparoscopic group compared with the laparotomy
Table 5
Results of operative time stratified according to operative procedure in unmatched and propensity score matched laparotomy and laparoscopy groups