This cohort study assesses whether there are differences in adverse outcomes between pregnant patients with acute cholecystitis who do or do not receive cholecystectomy.
Key Points
Question
For pregnant patients presenting with acute cholecystitis (AC), is there a difference in adverse outcomes between those who do or do not receive cholecystectomy?
Findings
In this cohort study of 3426 pregnant patients with AC and 1-year follow-up, 34.5% underwent cholecystectomy during pregnancy. Compared with nonoperative management, patients who received cholecystectomy had lower odds of adverse outcomes (preterm delivery or pregnancy loss) across trimesters.
Meaning
These findings support cholecystectomy during pregnancy across trimesters and suggest that increased use of surgery for patients with AC may be an opportunity to improve pregnancy outcomes.
Abstract
Importance
Acute cholecystitis (AC) management during pregnancy requires balancing the risk of pregnancy loss or preterm delivery (adverse pregnancy outcomes [APOs]) with or without surgery. Guidelines recommend cholecystectomy across trimesters; however, trimester-specific evidence on the risks of AC and its management is lacking.
Objective
To assess cholecystectomy frequency in pregnant people with AC, compare the rates of APOs in pregnant people with or without AC, and compare the rates of APOs in people with AC who did or did not undergo cholecystectomy.
Design, Setting, and Participants
This retrospective, population-based cohort study used data for pregnant people with AC from the IBM MarketScan Commercial Claims and Encounters Database from January 1, 2007, to December 31, 2019, and a propensity score–matched cohort of pregnant people without AC. Trimester status (first [T1], second [T2], and third [T3]), APOs, and cholecystectomy were defined by administrative claims. Data were analyzed from October 2021 to July 2022.
Exposures
Pregnant patients with or without AC. Pregnant patients with AC who did or did not receive cholecystectomy.
Main Outcomes and Measures
The main outcomes were cholecystectomy during pregnancy and APOs (ie, preterm delivery and pregnancy loss). Pregnant patients with and without AC were compared to assess the association of AC with risk of APOs. Propensity score inverse-probability weighting was used to calculate treatment-associated APO risk among patients with 1-year follow-up.
Results
The study included 5759 pregnant patients with AC (mean [SD] age, 30.1 [6.6] years) and 23 036 controls (mean [SD] age, 29.9 [6.7] years) after propensity score matching. Among 3426 pregnant patients with AC and 1-year follow-up, 1182 (34.5%) underwent cholecystectomy during the pregnancy (684 [41.7%] presenting with AC in T1, 404 [40.4%] in T2, and 94 [12.0%] in T3). Acute cholecystitis during pregnancy, irrespective of treatment, was associated with higher odds of APO compared with no AC during pregnancy across all trimesters (odds ratio [OR], 1.69 [95% CI, 1.54-1.85]). Compared with nonoperative management, receipt of surgery was associated with lower odds of APOs across all trimesters (OR, 0.75 [95% CI, 0.63-0.87]), in T1 (OR, 0.81 [95% CI, 0.66-1.00]), in T2 (OR, 0.71 [95% CI, 0.50-1.00]), and in T3 (OR, 0.45 [95% CI, 0.28-0.70]).
Conclusions and Relevance
In this study, cholecystectomy was associated with lower risk of APO in patients with AC across all trimesters, with the greatest benefit in T3. However, only 34.5% overall and 12.0% of patients in T3 had a cholecystectomy. These findings support guidelines recommending cholecystectomy during pregnancy and should inform decision-making discussions. Greater guideline adherence and surgery use, especially in T3, may represent an opportunity to improve outcomes for pregnant people with AC.
Introduction
Diagnosis and management of nonobstetrical abdominal emergencies in pregnant patients are challenging for both clinicians and patients. The most frequent nonobstetric indication for surgery during pregnancy is infection including acute cholecystitis (AC).1,2,3,4 To our knowledge, there have been no randomized clinical trials evaluating risk or adverse pregnancy outcomes (APOs) for AC that include pregnant patients. The Society of American Gastrointestinal and Endoscopic Surgeons guidelines recommend cholecystectomy as the treatment of choice for pregnant patients presenting with symptomatic gallbladder disease regardless of trimester.5 This recommendation is based on limited evidence about maternal and pregnancy risk when cholecystectomy is not performed and focuses only on the risk of cholecystectomy.6,7 Pregnancy risk encompasses a range of APOs, including miscarriage (before 20 weeks’ gestation), intrauterine fetal demise or stillbirth (during and after 20 weeks’ gestation), and preterm delivery (birth occurring before 37 weeks’ gestation).8 While several studies have examined the risk of APOs in pregnant patients with AC, to our knowledge, no population-based study has assessed the risk of APO among pregnant patients across trimesters who did and did not have cholecystectomy.
Despite the recommendation for cholecystectomy during pregnancy, many surgeons opt to delay surgery until after delivery.9 The proportion of pregnant patients undergoing cholecystectomy for AC during pregnancy is unknown. In this study, we aimed to (1) describe how often pregnant patients with AC receive cholecystectomy, (2) assess the risk of APO in patients with AC compared with a matched cohort of pregnant patients without AC, and (3) compare the rates of APO in patients presenting with AC who did or did not receive cholecystectomy.
Methods
This cohort study adhered to reporting standards described by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.10 This analysis of deidentified data was deemed exempt from review by the University of Washington institutional review board, with a waiver of informed consent.
Data Source, Study Design, and Population
We performed a retrospective cohort study of pregnant patients with an index diagnosis of AC during pregnancy using the IBM MarketScan Commercial Claims and Encounters Database from January 1, 2007, through December 31, 2019. The MarketScan database encompasses inpatient, outpatient, and pharmacy claims, providing person-specific clinical utilization and expenditure data for employed, privately insured adults and their dependents in the US. Sourced from health plans, employers, government, and public organizations, the database includes over 100 different payers with more than 500 million fully paid and adjudicated claim records. Claims are longitudinally linked during an individual’s plan enrollment, facilitating cross–health plan longitudinal follow-up. Data standardization ensures accuracy and reliability of financial, clinical, and demographic fields. Data are collected for the MarketScan annual database releases when nearly 100% of claims are paid, enhancing data reliability. Diagnoses and procedures are represented using International Classification of Diseases, Ninth Revision (ICD-9); International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10); Current Procedural Terminology (CPT); and Healthcare Common Procedure Coding System codes (eAppendix 1 in Supplement 1).
Determining Pregnancy Intervals
All pregnancies in the MarketScan database were identified using a slightly modified version of a previously validated algorithm by Matcho et al11 (eAppendixes 2 and 3 in Supplement 1). This algorithm uses ICD-9 codes and CPT procedure codes mapped to pregnancy length and outcomes, to which we added ICD-10 codes identified by Sarayani et al.12 We identified approximately 9.7 million unique pregnancy episodes with the estimated start date (referred to as last menstrual period), end date, and pregnancy outcome for each episode (ie, live birth, miscarriage, induced abortion, stillbirth, or ectopic pregnancy).
Inclusion and Exclusion Criteria
After identifying pregnancy episodes among enrollees aged 15 to 55 years, episodes in which AC was diagnosed during pregnancy (ie, between pregnancy start date and end date) were identified using our AC case definition (eAppendix 1 in Supplement 1). The index date of AC was defined as the first date on which an AC code occurred, and this code was considered the index diagnosis code. For any enrollee with more than 1 pregnancy episode with AC, we included only the first pregnancy episode.
We excluded multiple gestations, ectopic and molar pregnancies, and those ending in elective termination of pregnancy, none of which are suspected to be associated with AC diagnosis or management. We excluded patients with cholecystectomy, gallbladder cancer, cholangiocarcinoma, or acute pancreatitis prior to the index AC diagnosis date. For aim 2, these patients were excluded if these claims occurred prior to the start of pregnancy, and these criteria applied to both AC and control participants. For aim 3, these patients were excluded if these claims occurred prior to the index diagnosis date (eFigure in Supplement 1).
To review all care related to AC for comparing patients who did or did not receive surgery, eligible patients with AC during pregnancy were required to have continuous enrollment in MarketScan for 90 days before and 1 year after the last menstrual period, with no more than 30 days not enrolled during this time to accommodate patients who were changing insurance.
Treatment Type
Treatment of AC was considered to be surgery if the patient had a CPT procedure code for laparoscopic or open cholecystectomy during pregnancy. Pregnant patients who did not have cholecystectomy during pregnancy or on the pregnancy end date were considered to be in the nonoperative management group.
Outcomes
Outcomes of interest were preterm delivery (defined as pregnancy outcome indicated by diagnosis codes for preterm birth or live birth with estimated gestational age of <37 and ≥24 weeks) and pregnancy loss, a composite outcome including both spontaneous abortion and stillbirth as assigned in the algorithm of Matcho et al.11 Adverse pregnancy outcome was a composite outcome of preterm delivery and pregnancy loss.
Statistical Analysis
Data were analyzed from October 2021 to July 2022. All statistical analysis was performed using R, version 4.0.3 (R Project for Statistical Computing). To assess the association between AC and APOs, we compared outcomes in the cohort with AC with those in a propensity score–matched cohort without AC (control group). We did not apply a continuous enrollment criterion for patients with AC or matched controls for the part of the analysis comparing patients with AC with matched controls. In the propensity score–matched analysis, we used nearest-neighbor matching without replacement.
We first selected 20 000 pregnancies at random from each year of MarketScan data from which to draw the matched cohort. We eliminated duplicate records, subsequent pregnancies in the same individual, and all individuals failing to meet inclusion criteria or experiencing AC during pregnancy.
We trained a logistic regression model using these data combined with the cohort with AC, with AC as the outcome. The variables used were maternal age, gestational age, hypertension, gestational diabetes, diabetes, alcohol use, drug use, obesity, Charlson Comorbidity Index score, prior cesarean delivery, history of preterm delivery, history of multiple births, vaginitis, urinary tract infection, cervical incompetence, and Obstetric Comorbidity Index score. The model’s fitted values in logit scale were used as propensity scores to match each patient with AC with 4 similar controls. The cohort with AC and the control cohort were randomly split into 4 groups to reduce computing requirements, and steps were repeated for each group. The difference in propensity scores between each patient with AC and each control participant was calculated to form an adjacency matrix.
Control participants were eligible to match with a given participant with AC if the pregnancy was ongoing at the gestational age at which AC had been diagnosed for that patient with AC and if the absolute difference in propensity score was less than 20% of the SD of the propensity scores in logit scale.13 For each randomly chosen participant with AC, the 4 eligible control participants with the smallest absolute difference in propensity score who had not already been matched with another participant with AC were chosen. If fewer than 4 eligible controls were available, no control participants were matched to the patient with AC. If more than 4 control participants had the same difference, 4 were chosen randomly. Patients with AC were matched 1 at a time in a random order so that unmatched patients would not be concentrated at the beginning or end of the study period.
Proportions of pregnancy loss and preterm delivery were calculated for participants with AC stratified by trimester of AC occurrence and matching controls with accompanying 95% CIs. For proportions of pregnancy loss and preterm delivery, 95% CIs were calculated by bootstrap resampling of participants with AC, as controls matched to the same participant with AC did not constitute independent observations. Odds ratios (ORs) for pregnancy loss and preterm delivery comparing the cohort with AC and matched controls for each trimester were calculated, with 95% CIs generated via bootstrap.
We calculated the incidence of preterm delivery and pregnancy loss after AC diagnosis and reported these unadjusted proportions with 95% CIs obtained using a χ2 approximation from the prop.test function in the stats package in R for those who did or did not receive surgery during pregnancy. Pregnancies ending in pregnancy loss were excluded when calculating the incidence of preterm delivery. Given the clinical importance of trimester-specific outcomes to understand the association between treatment (surgery or no surgery) and APO, we stratified outcomes by trimester of AC occurrence (first trimester [T1], second trimester [T2], and third trimester [T3]). Trimester cutoffs were determined using the American College of Obstetricians and Gynecologists guidelines (eAppendix 1 in Supplement 1).
To assess the associations between treatment and APO, we used propensity score inverse-probability weighting. Propensity scores were obtained from logistic regression models assessing potential confounding factors that might be associated with treatment choices. These included demographic characteristics (maternal age, Charlson Comorbidity Index score, hypertension, diabetes, obesity, overt obesity, smoking, alcohol abuse, and drug abuse), pregnancy characteristics (prior cesarean delivery, history of preterm delivery, history of multiple births, vaginitis, urinary tract infection, cervical incompetence, Obstetric Comorbidity Index score, and gestational age in days), and disease severity (prior gallstones, sepsis, peritonitis, cholangitis, appendicitis, and different diagnosis codes of AC). Variables were selected a priori based on clinical experience and literature review.14,15 Propensity score modeling was stratified by trimester of AC occurrence, with a separate logistic regression model used for each trimester.
Fitted values from these models were used as propensity scores to generate inverse-probability–weighted estimates of pregnancy loss and preterm delivery ORs among patients who did or did not receive cholecystectomy during pregnancy stratified by trimester of AC occurrence. We present weighted and unweighted ORs comparing preterm delivery and pregnancy loss between pregnant patients with and without AC, with 95% CIs generated using the glm function from the stats package in R.
Results
Characteristics of Study Cohorts
Between 2007 and 2019, a total of 5932 pregnant patients met the AC case definition. Among these, 5759 (mean [SD] age, 30.1 [6.6] years) could be matched using propensity scores (1:4) with 23 036 controls without AC (mean [SD] age, 29.9 [6.7] years). The baseline characteristics were balanced in the propensity score–matched samples (Table 1).
Table 1. Characteristics of Patients With and Without AC Before and After Propensity Score Matching.
| Characteristic | Before propensity score matching | After propensity score matching | |||||
|---|---|---|---|---|---|---|---|
| Patients with AC (n = 5932)a | Controls (n = 151 901)a | OR (95% CI) | Patients with AC (n = 5759)a | Controls (n = 23 036)a | OR (95% CI) | ||
| Age, mean (SD) | 30.1 (6.7) | 30.5 (5.9) | 0.07 (0.04-0.10)b | 30.1 (6.6) | 29.9 (6.7) | 0.02 (0.00-0.05)b | |
| Age ≥40 y | 370 (6.2) | 6846 (4.5) | 1.41 (1.26-1.57) | 343 (6.0) | 1321 (5.7) | 1.04 (0.92-1.18) | |
| Diabetes | 229 (3.9) | 2874 (1.9) | 2.08 (1.81-2.38) | 209 (3.6) | 770 (3.3) | 1.09 (0.93-1.27) | |
| Hypertension | 421 (7.1) | 5837 (3.8) | 1.91 (1.72-2.12) | 376 (6.5) | 1424 (6.2) | 1.06 (0.94-1.19) | |
| Obesity | 474 (9.0) | 6636 (4.4) | 1.90 (1.72-2.09) | 433 (7.5) | 1624 (7.0) | 1.07 (0.96-1.20) | |
| Prior cholelithiasis | 837 (14.1) | 5549 (3.7) | 4.33 (4.01-4.68) | 692 (12.0) | 2941 (12.8) | 0.93 (0.85-1.02) | |
| History of preterm delivery | 6 (0.1) | 94 (0.1) | 1.68 (0.65-3.51) | 6 (0.1) | 17 (0.1) | 1.44 (0.51-3.49) | |
| Hyperemesis gravidarum | 107 (1.8) | 1575 (1.0) | 1.76 (1.43-2.13) | 92 (1.6) | 341 (1.5) | 1.08 (0.85-1.36) | |
| Insurance plan type | |||||||
| Low burden | 605 (10.2) | 14 844 (9.8) | 1.05 (0.96-1.14) | 580 (10.1) | 2520 (10.9) | 0.91 (0.83-1.00) | |
| Some burden | 3335 (56.2) | 70 334 (46.3) | 1.49 (1.41-1.57) | 3202 (55.6) | 11 275 (48.9) | 1.31 (1.23-1.38) | |
| High burden | 315 (5.3) | 10 305 (6.8) | 0.77 (0.69-0.86) | 306 (5.3) | 1091(4.7) | 1.13 (0.99-1.28) | |
| Unknown | 1677 (28.3) | 56 418 (37.1) | 0.67 (0.63-0.71) | 1671 (29.0) | 8150 (35.4) | 0.75 (0.70-0.80) | |
| Geographic region | |||||||
| Northeast | 712 (12.0) | 16 335 (10.8) | 1.13 (1.04-1.23) | 682 (11.8) | 2453 (10.6) | 1.13 (1.03-1.23) | |
| North Central | 919 (15.5) | 21 938 (14.4) | 1.09 (1.01-1.17) | 876 (15.2) | 3524 (15.3) | 0.99 (0.92-1.08) | |
| South | 1885 (31.8) | 40 745 (26.8) | 1.27 (1.2-1.34) | 1815 (31.5) | 6309 (27.4) | 1.22 (1.15-1.30) | |
| West | 765 (12.9) | 17 310 (11.4) | 1.15 (1.06-1.24) | 738 (12.8) | 2655 (11.5) | 1.13 (1.03-1.23) | |
| Unknown | 1651 (27.8) | 55 573 (36.6) | 0.67 (0.63-0.71) | 1648 (28.6) | 8095 (35.1) | 0.74 (0.69-0.79) | |
Abbreviations: AC, acute cholecystitis; OR, odds ratio.
Data are presented as number (percentage) of individuals unless otherwise indicated.
Standardized mean difference (95% CI).
Of the 5932 patients with AC during pregnancy, 3426 met the insurance enrollment criterion (eAppendix 1 in Supplement 1). Among these, 1639 (47.8%) had a diagnosis of AC in T1, 1001 (29.2%) in T2, and 786 (22.9%) in T3. Overall, 1182 patients (34.5%) received cholecystectomy, including 684 patients (41.7%) presenting in T1, 404 (40.4%) in T2, and 94 (12.0%) in T3. The baseline characteristics of study participants according to AC treatment are shown in Table 2. Overall, mean (SD) age at AC diagnosis was 30.5 (6.8) years, with similar rates of prior cholelithiasis claims (210 [17.8%] vs 386 [17.2%]) among those who did or did not receive surgery. Less than half of patients who did or did not have surgery had a Charlson Comorbidity Index score of 1 or greater (526 [44.5%] vs 867 [38.6%]). Patients who received surgery had a higher rate of inpatient diagnosis codes at AC diagnosis compared with patients who did not receive surgery (436 [36.9%] vs 497 [22.1%]).
Table 2. Characteristics of Patients With Acute Cholecystitis According to Treatment Received During Pregnancy.
| Characteristic | Patientsa | ||||
|---|---|---|---|---|---|
| Total (N = 3426) | Surgery (n = 1182) | No surgery (n = 2244) | |||
| Age, mean (SD), y | 30.5 (6.8) | 30.9 (7.2) | 30.3 (6.5) | ||
| Age ≥40 y | 233 (6.8) | 106 (9.0) | 127 (5.7) | ||
| Trimester of acute cholecystitis presentation | |||||
| 1 | 1639 (47.8) | 684 (57.9) | 955 (42.6) | ||
| 2 | 1001 (29.2) | 404 (34.2) | 597 (26.6) | ||
| 3 | 786 (22.9) | 94 (8.0) | 692 (30.8) | ||
| Diabetes | 219 (6.4) | 97 (8.2) | 122 (5.4) | ||
| Hypertension | 398 (11.6) | 162 (13.7) | 236 (10.5) | ||
| Obesity | 483 (14.1) | 220 (18.6) | 263 (11.7) | ||
| History of preterm delivery | 10 (0.3) | 3 (0.3) | 7 (0.3) | ||
| Hyperemesis gravidarum | 430 (12.6) | 164 (13.9) | 266 (11.9) | ||
| Prior cholelithiasis | 596 (17.4) | 210 (17.8) | 386 (17.2) | ||
| Charlson Comorbidity Index score | |||||
| 0 | 2033 (59.3) | 656 (55.5) | 1377 (61.4) | ||
| ≥1 | 1393 (40.7) | 526 (44.5) | 867 (38.6) | ||
| Obstetric Comorbidity Index score, mean (SD) | 0.47 (1.0) | 0.54 (1.1) | 0.43 (0.9) | ||
| Inpatient setting | 933 (27.2) | 436 (36.9) | 497 (22.1) | ||
| Insurance plan type | |||||
| Low burden | 410 (12.0) | 128 (10.8) | 282 (12.6) | ||
| Some burden | 2583 (75.4) | 899 (76.1) | 1684 (75.0) | ||
| High burden | 341 (10.0) | 121 (10.2) | 220 (9.8) | ||
| Unknown | 92 (2.7) | 34 (2.9) | 58 (2.6) | ||
| Geographic region | |||||
| Northeast | 568 (16.6) | 171 (14.5) | 397 (17.7) | ||
| North Central | 753 (22.0) | 272 (23.0) | 481 (21.4) | ||
| South | 1453 (42.4) | 529 (44.8) | 924 (41.2) | ||
| West | 595 (17.4) | 193 (16.3) | 402 (17.9) | ||
| Unknown | 57 (1.7) | 17 (1.4) | 40 (1.8) | ||
Data are presented as number (percentage) of patients unless otherwise indicated.
Clinical Outcomes for the Matched Cohorts
In the propensity score–matched analysis, pregnant patients presenting with AC across all trimesters had higher rates of APOs than patients who did not have AC (1451 of 5759 patients [25.2%] vs 3832 of 23 036 patients [16.6%]; OR, 1.69 [95% CI, 1.54-1.85]). Pregnant patients with AC had higher rates of APO in each trimester than patients without AC: T1 (923 of 2518 patients [36.7%] vs 2734 of 10 072 patients [27.1%]; OR, 1.55 [95% CI, 1.36-1.77]), T2 (301 of 1736 patients [17.3%] vs 771 of 6944 patients [11.1%]; OR, 1.68 [95% CI, 1.41-2.00]), and T3 (227 of 1505 patients [15.1%] vs 327 of 6020 patients [5.4%]; OR, 3.09 [95% CI, 2.51-3.81]). Rates of the 2 individual outcomes of the composite APO (pregnancy loss or preterm delivery) were also higher in patients presenting with AC in each trimester than in patients without AC (Table 3).16
Table 3. Outcomes by Trimester in the Cohort With AC and the Propensity Score–Matched Control Cohort.
| Presentation, outcome | Events, No./total No. (%) [95% CI] | Odds ratio (95% CI)a | |
|---|---|---|---|
| Cohort with AC (n = 5759) | Propensity score–matched cohort (n = 23 036) | ||
| Trimester 1 | |||
| Pregnancy loss | 629/2518 (24.98) [23.31-26.73] | 2066/10 072 (20.51) [19.61-21.45] | 1.29 (1.16-1.43) |
| Preterm delivery | 294/1889 (15.56) [13.97-17.30] | 668/7556 (8.84) [8.19-9.50] | 1.90 (1.64-2.20) |
| Trimester 2 | |||
| Pregnancy loss | 34/1736 (1.96) [1.38-2.76] | 65/6944 (0.94) [0.71-1.18] | 2.11 (1.34-3.24) |
| Preterm delivery | 267/1702 (15.69) [14.01-17.52] | 706/6808 (10.37) [9.65-11.12] | 1.61 (1.38-1.86) |
| Trimester 3 | |||
| Pregnancy lossb | 1/1505 (0.07) [0.00-0.43] | 4/6020 (0.07) [0.02-0.13] | 1.00 (0.00-8.01) |
| Preterm delivery | 226/1504 (15.03) [13.28-16.96] | 323/6016 (5.37) [4.79-5.97] | 3.12 (2.64-3.70) |
Abbreviation: AC, acute cholecystitis.
Results from the propensity score–matched analysis.
Pregnancy loss rates in the third trimester were too low to make a comparison. This is consistent with preexisting obstetric data.16
Clinical Outcomes Associated With Treatments
In the propensity score inverse-probability–weighted analysis, surgical treatment of AC across all trimesters was associated with a lower rate of APO compared with no surgery (280 of 1182 patients [23.7%] vs 543 of 2244 patients [24.2%]; weighted OR, 0.75 [95% CI, 0.63-0.87]). When examined by trimester, surgical treatment was associated with significantly lower rates of APOs in the first and third trimesters compared with no surgery: T1 (213 of 684 patients [31.1%] vs 327 of 955 patients [34.2%]; weighted OR, 0.81 [95% CI, 0.66-1.00]), T2 (59 of 404 patients [14.6%] vs 103 of 597 patients [17.3%]; weighted OR, 0.71 [95% CI, 0.50-1.00]), and T3 (8 of 94 patients [8.5%] vs 113 of 692 patients [16.3%]; weighted OR, 0.45 [95% CI, 0.28-0.70]).
There were significant differences in the 2 outcomes of APO (pregnancy loss and preterm delivery) in different trimesters (Figure and Table 4). The lower rate of APO in the surgery group for patients diagnosed with AC in T1 compared with in the no surgery group was mostly due to the lower rate of pregnancy loss. The lower rate of APO observed in the surgery group for patients diagnosed with AC in T2 and T3 compared with the no surgery group was attributable to the lower rate of preterm delivery.
Figure. Association Between Treatment and Outcomes by Trimester of Acute Cholecystitis (AC) Presentation.
Results are from propensity score inverse-probability–weighted analysis.
aPregnancy loss rates in the third trimester were too low to make a comparison. This is consistent with preexisting obstetric data.16
Table 4. Outcomes by Trimester of Acute Cholecystitis Presentation and Treatment.
| Presentation, outcome | Events, No./total No. (%) [95% CI] | |
|---|---|---|
| Surgery (n = 1182) | No surgery (n = 2244) | |
| Trimester 1 | ||
| Pregnancy loss | 129/684 (18.86) [16.04-22.04] | 225/955 (23.56) [20.93-26.41] |
| Preterm delivery | 84/555 (15.14) [12.31-18.45] | 102/730 (13.97) [11.58-16.75] |
| Trimester 2 | ||
| Pregnancy loss | 5/404 (1.24) [0.46-3.03] | 10/597 (1.68) [0.85-3.16] |
| Preterm delivery | 54/399 (13.53) [10.41-17.38] | 93/587 (15.84) [13.03-19.11] |
| Trimester 3a | ||
| Preterm delivery | 8/94 (8.51) [4.01-16.56] | 113/692 (16.33) [13.70-19.34] |
Pregnancy loss rates in the third trimester were too low to make a comparison. This is consistent with preexisting obstetric data.16
Discussion
To our knowledge, this was the first population-based study examining the risk of APOs associated with both AC and AC treatment options and is novel for being, to date, the first to use a large cohort size to account for trimester-specific risks in both comparisons. Acute cholecystitis was associated with a higher risk of APOs compared with no diagnosis of AC, and among patients with AC, the risk of APOs was lower in those who received surgery than among those who did not receive surgery across all trimesters. These results support professional guidelines recommending cholecystectomy across all trimesters. Only 34.5% of pregnant patients with AC underwent surgery during pregnancy, indicating limited adherence to guidelines. Given the greater risk of APOs associated with AC during pregnancy and the improved outcomes associated with surgery, increased adherence to existing professional guidelines is recommended.
The low proportion of pregnant patients receiving surgery could be attributed to the lack of clinical trials data or to recent studies suggesting that delaying surgery until after delivery may have clinical benefits.17,18 Kuy and colleagues18 reported higher cost and longer hospital length of stay with cholecystectomy in pregnant patients compared with cholecystectomy in nonpregnant patients. Fong and colleagues17 showed that cholecystectomy in the third trimester was associated with worse maternal and procedure-related outcomes compared with cholecystectomy after delivery. However, neither study compared the risk of surgical treatment with the risk of nonoperative management during pregnancy. Moreover, these studies did not account for different trimesters of pregnancy. Accounting for trimesters of pregnancy is particularly important because the risk of adverse outcomes varies across trimesters regardless of disease and treatment. Patients in the first trimester are at higher risk of spontaneous abortion than patients in the second trimester, and by definition, preterm delivery only occurs in the second and third trimesters of pregnancy.11,15 In this study, we accounted for trimester, defining trimester of AC presentation using a validated, code-based algorithm keying into specific dates of diagnostic and procedure codes11,12 to describe the trimester-specific risk of APO in patients presenting with AC who did or did not receive cholecystectomy.
Cheng and colleagues19 used the National Inpatient Sample to study cholecystectomy outcomes during different trimesters of pregnancy. Their results showed that cholecystectomy in the third trimester was associated with higher maternal and fetal complications and greater economic burden compared with cholecystectomy in the first and second trimesters. However, this comparison only included patients who underwent cholecystectomy, excluding those managed nonoperatively. It also included cholecystectomy for a wide range of biliary diseases with variable clinical severity (biliary colic, choledocholithiasis, AC, and gallstone pancreatitis) without any longitudinal follow-up data, potentially leading to missed outcomes. In contrast, our longitudinal study addressed these limitations by including a 1-year follow-up period and focusing specifically on AC to reduce potential confounding due to clinical severity. We evaluated outcomes for pregnant patients presenting with AC in different trimesters for both cholecystectomy and nonoperative management. Our results showed that 22.9% of pregnant patients with an initial AC diagnosis presented in the third trimester, lacking the option for surgery in earlier trimesters. Also, receipt of surgery for patients presenting with AC in T3 was associated with a lower rate of preterm delivery compared with nonoperative management.
Limitations
A major limitation of our study is the absence of random treatment assignment. We attempted to address this by using a propensity score–matched analysis, including factors that could potentially influence treatment selection. Nevertheless, we recognize that confounding related to disease severity at presentation, surgeon preference, patient preference, and institutional differences related to resources (eg, presence of a neonatal intensive care unit) may exist and were not available for evaluation in this claims-based data set. While the measures of disease severity (Charlson Comorbidity Index) and obstetrical risk (Obstetric Comorbidity Index) were greater in surgical patients, potential confounding by indication remained. Another limitation is the exclusive focus on APOs; patient-reported outcomes may be equally if not more relevant to patients making decisions about their care.
Another major limitation of this study is its reliance on the MarketScan database, which exclusively contains claims-based data for individuals with employer-sponsored commercial health insurance. Given that 42% of pregnancies in the US are covered by Medicaid,20 our findings might not accurately reflect care patterns and outcomes for patients lacking commercial insurance. Moreover, this data set does not account for factors such as socioeconomic status and systemic racism, which substantially impact APOs. Since effective management of AC depends on access to appropriate evaluation, treatment, and monitoring for complications, it is crucial to discern which individuals have access to these components of care and which do not.
Conclusions
In this study, cholecystectomy for patients presenting with AC across trimesters was associated with a lower risk of APOs compared with nonoperative management. When examined by trimester, this association was found only in the first and third trimesters. While cholecystectomy is recommended by professional society guidelines, cholecystectomy is infrequently performed during pregnancy, especially in the third trimester when it may offer the greatest risk reduction. This may represent an opportunity for a quality improvement initiative to improve safety and patient-centered care.
eAppendix 1. Cohorts Selection and Case Definition with Codes
eAppendix 2. Vocabulary Categorized Pregnancy Concept Sets
eAppendix 3. Pregnancy Episode Algorithm Pseudocode
eFigure 1. Identification of Study Cohorts
Data Sharing Statement
References
- 1.Cheng V, Matsushima K, Sandhu K, et al. Surgical trends in the management of acute cholecystitis during pregnancy. Surg Endosc. 2021;35(10):5752-5759. doi: 10.1007/s00464-020-08054-w [DOI] [PubMed] [Google Scholar]
- 2.Gilo NB, Amini D, Landy HJ. Appendicitis and cholecystitis in pregnancy. Clin Obstet Gynecol. 2009;52(4):586-596. doi: 10.1097/GRF.0b013e3181c11d10 [DOI] [PubMed] [Google Scholar]
- 3.Lanzafame RJ. Laparoscopic cholecystectomy during pregnancy. Surgery. 1995;118(4):627-631. doi: 10.1016/S0039-6060(05)80028-5 [DOI] [PubMed] [Google Scholar]
- 4.Nasioudis D, Tsilimigras D, Economopoulos KP. Laparoscopic cholecystectomy during pregnancy: a systematic review of 590 patients. Int J Surg. 2016;27:165-175. doi: 10.1016/j.ijsu.2016.01.070 [DOI] [PubMed] [Google Scholar]
- 5.Pearl JP, Price RR, Tonkin AE, Richardson WS, Stefanidis D. SAGES guidelines for the use of laparoscopy during pregnancy. Surg Endosc. 2017;31(10):3767-3782. doi: 10.1007/s00464-017-5637-3 [DOI] [PubMed] [Google Scholar]
- 6.Date RS, Kaushal M, Ramesh A. A review of the management of gallstone disease and its complications in pregnancy. Am J Surg. 2008;196(4):599-608. doi: 10.1016/j.amjsurg.2008.01.015 [DOI] [PubMed] [Google Scholar]
- 7.Steinbrook RA, Brooks DC, Datta S. Laparoscopic cholecystectomy during pregnancy: review of anesthetic management, surgical considerations. Surg Endosc. 1996;10(5):511-515. doi: 10.1007/BF00188397 [DOI] [PubMed] [Google Scholar]
- 8.Spong CY. Defining “term” pregnancy: recommendations from the Defining “Term” Pregnancy Workgroup. JAMA. 2013;309(23):2445-2446. doi: 10.1001/jama.2013.6235 [DOI] [PubMed] [Google Scholar]
- 9.Abuabara SF, Gross GW, Sirinek KR. Laparoscopic cholecystectomy during pregnancy is safe for both mother and fetus. J Gastrointest Surg. 1997;1(1):48-52. doi: 10.1007/s11605-006-0009-7 [DOI] [PubMed] [Google Scholar]
- 10.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453-1457. doi: 10.1016/S0140-6736(07)61602-X [DOI] [PubMed] [Google Scholar]
- 11.Matcho A, Ryan P, Fife D, Gifkins D, Knoll C, Friedman A. Inferring pregnancy episodes and outcomes within a network of observational databases. PLoS One. 2018;13(2):e0192033. doi: 10.1371/journal.pone.0192033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sarayani A, Wang X, Thai TN, Albogami Y, Jeon N, Winterstein AG. Impact of the transition from ICD-9-CM to ICD-10-CM on the identification of pregnancy episodes in US health insurance claims data. Clin Epidemiol. 2020;12:1129-1138. doi: 10.2147/CLEP.S269400 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10(2):150-161. doi: 10.1002/pst.433 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sachs A, Guglielminotti J, Miller R, Landau R, Smiley R, Li G. Risk factors and risk stratification for adverse obstetrical outcomes after appendectomy or cholecystectomy during pregnancy. JAMA Surg. 2017;152(5):436-441. doi: 10.1001/jamasurg.2016.5045 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wilcox AJ, Weinberg CR, O’Connor JF, et al. Incidence of early loss of pregnancy. N Engl J Med. 1988;319(4):189-194. doi: 10.1056/NEJM198807283190401 [DOI] [PubMed] [Google Scholar]
- 16.Smith LK, Hindori-Mohangoo AD, Delnord M, et al. ; Euro-Peristat Scientific Committee . Quantifying the burden of stillbirths before 28 weeks of completed gestational age in high-income countries: a population-based study of 19 European countries. Lancet. 2018;392(10158):1639-1646. doi: 10.1016/S0140-6736(18)31651-9 [DOI] [PubMed] [Google Scholar]
- 17.Fong ZV, Pitt HA, Strasberg SM, et al. ; California Cholecystectomy Group . Cholecystectomy during the third trimester of pregnancy: proceed or delay? J Am Coll Surg. 2019;228(4):494-502.e1. doi: 10.1016/j.jamcollsurg.2018.12.024 [DOI] [PubMed] [Google Scholar]
- 18.Kuy S, Roman SA, Desai R, Sosa JA. Outcomes following cholecystectomy in pregnant and nonpregnant women. Surgery. 2009;146(2):358-366. doi: 10.1016/j.surg.2009.03.033 [DOI] [PubMed] [Google Scholar]
- 19.Cheng V, Matsushima K, Ashbrook M, et al. Association between trimester and outcomes after cholecystectomy during pregnancy. J Am Coll Surg. 2021;233(1):29-37.e1. doi: 10.1016/j.jamcollsurg.2021.03.034 [DOI] [PubMed] [Google Scholar]
- 20.Martin JA, Hamilton BE, Osterman MJK, Driscoll AK. Births: Final Data for 2019. National Vital Statistics Reports; vol 70 no 2. National Center for Health Statistics; 2021. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix 1. Cohorts Selection and Case Definition with Codes
eAppendix 2. Vocabulary Categorized Pregnancy Concept Sets
eAppendix 3. Pregnancy Episode Algorithm Pseudocode
eFigure 1. Identification of Study Cohorts
Data Sharing Statement