- Email: [email protected]
Endoscopic management and outcomes of pregnant women hospitalized for nonvariceal upper GI bleeding: a nationwide analysis
ORIGINAL ARTICLE: Clinical Endoscopy
Endoscopic management and outcomes of pregnant women hospitalized for nonvariceal upper GI bleeding: a nationwide analysis Geoffrey C. Nguyen, MD, PhD, Amreen M. Dinani, MD, Kevin Pivovarov Toronto, Ontario, Canada; Baltimore, Maryland, USA
Background: Upper GI endoscopy has an important diagnostic and therapeutic role in the management of nonvariceal upper GI bleeding (NVUGB). Objective: To characterize nationwide patterns of utilization of upper GI endoscopy in pregnant women with NVUGB and to assess health outcomes. Design: Retrospective cohort study. Setting: Participating hospitals from the Nationwide Inpatient Sample, 1998-2007. Patients: Pregnant and age-matched nonpregnant women admitted for NVUGB. Intervention: The study population was classified as pregnant women with NVUGB (n ⫽ 1210) and nonpregnant women with NVUGB (n ⫽ 6050). Main Outcome Measurements: Rate of upper GI endoscopy, maternal mortality, fetal death/complications, and premature delivery. Results: Pregnant women were less likely than nonpregnant women to undergo upper GI endoscopy (26% vs 69%; P ⬍ .0001) even after adjustment for comorbidities, transfusion requirement, and the presence of hypovolemic shock (adjusted odds ratio, 0.19; 95% confidence interval, 0.16-0.22). Among those who underwent endoscopy, pregnant women were less likely to undergo the procedure within 24 hours of admission (50% vs 57%; P ⫽ .02). Mortality was lower among pregnant women compared with nonpregnant women (0% vs 0.6%; P ⫽ .006). In comparing outcomes between those who did and did not undergo endoscopy, there was no difference in fetal loss (0.2% vs 0.6%), fetal distress/complications (2.7% vs 2.6%), or premature delivery (7.3% vs 6.4%). Limitations: The study was based on administrative data. Conclusion: A conservative nonendoscopic approach is common in the management of pregnant women with NVUGB and is not associated with worse maternal or fetal outcomes. Upper GI endoscopy is, however, safe when judiciously implemented in the actively bleeding patient. (Gastrointest Endosc 2010;72:954-9.)
Acute nonvariceal upper GI bleeding (NVUGB) is a common clinical emergency leading to 50 to 160 hospitalizations per 100,000 adults. Mortality may be as high as 10% to 14%, although this appears to be decreasing in
many countries.1 Upper GI endoscopy has been endorsed as essential in the diagnosis and management of NVUGB. Endoscopy allows assessment of the risk of rebleeding and enables therapeutic hemostasis that reduces bleeding, the
Abbreviations: aOR, adjusted odds ratio; ICD-9-CM, International Classification of Diseases, 9th edition, Clinical Modification; NIS, Nationwide Inpatient Sample; NVUGB, nonvariceal upper GI bleeding.
doi:10.1016/j.gie.2010.07.018
DISCLOSURE: This work was supported by an American Gastroenterological Association Research Scholar Award by the Foundation for Digestive Health and Nutrition to G. Nguyen. The sponsor had no role in the study design or data collection, analysis, or interpretation. No other financial relationships relevant to this publication were disclosed.
Current affiliations: Mount Sinai Division of Gastroenterology, University of Toronto, Toronto, Ontario, Canada (G.C.N., A.M.D., K.P.); Division of Gastroenterology and Hepatology, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA (G.C.N.).
Copyright © 2010 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00
954 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 5 : 2010
Received June 1, 2010. Accepted July 12, 2010.
Reprint requests: Geoffrey C. Nguyen, MD, PhD, 600 University Ave, Ste 437, Toronto, Ontario M5G 1X5, Canada.
www.giejournal.org
Nguyen et al
need for surgery, and mortality.2,3 Although the safety and efficacy of upper GI endoscopy are well established in the general population, the role and use of the procedure in pregnant women is less well characterized. The decision to perform upper GI endoscopy during pregnancy warrants careful deliberation of the risk-benefit profile for both mother and fetus and consideration of alternative therapeutic options. The clinical indications for upper GI endoscopy in pregnancy include (1) significant or ongoing GI bleeding, (2) refractory nausea or vomiting or abdominal pain, and (3) dysphagia or odynophagia.4 During the periprocedural period, the fetus can be particularly vulnerable to fetal hypoxia resulting from maternal oversedation, maternal positioning, or complications of the procedure.5 There are additional risks of premature delivery and teratogenicity associated with certain sedatives. Consequently, it is generally recommended that endoscopic procedures should be delayed until the second trimester when possible.4 Despite these caveats, under circumstances in which therapeutic intervention may be necessary, endoscopy is generally safer in pregnancy than alternative radiologic and surgical interventions.5 In a multicenter retrospective study of 83 pregnant women, upper GI endoscopy was not associated with adverse fetal outcomes and resulted in a 95% diagnostic yield for GI bleeding.6 However, that study may have been underpowered to detect small endoscopy-associated risks to the fetus. It also should be noted that those results were from tertiary-care academic centers, and we have little, if any, population-based data on the maternal and fetal outcomes related to NVUGB. Therefore, we sought to use U.S. population– based discharge data from the Nationwide Inpatient Sample (NIS) to (1) characterize nationwide patterns of utilization of upper GI endoscopy for NVUGB in the pregnant population and (2) assess in-hospital maternal and fetal outcomes associated with NVUGB.
Pregnancy and upper GI bleeding
Take-home Message ●
There was a much higher threshold for performing upper GI endoscopy in pregnant women who were hospitalized for nonvariceal upper GI bleeding. However, this did not necessarily reflect a lower quality of care, because pregnant women presented with less severe bleeding and experienced outcomes similar to or better than nonpregnant control women. Upper GI endoscopy in pregnant women, however, appears to be safe when clinically indicated.
TABLE 1. ICD-9-CM codes for non-variceal upper GI bleeding Diagnosis Esophagitis or esophageal ulcer Mallory-Weiss tear
Code(s) 530.21, 530.82 530.7
Gastric or duodenal ulcer
531.0x, 531.2x, 531.4x, 531.6x, 532.0x, 532.2x, 532.4x, 532.6x, 533.0x, 533.2x, 533.4x, 533.6x
Gastritis
535.01, 535.11, 535.21, 535.31, 535.41, 535.51, 535.61, 535.71
Dieulafoy lesion or arteriovenous malformation Anastomotic ulcer Hematemesis, unspecified
537.83, 537.84 534.0x, 534.2x, 534.4x, 534.6x 578.0
ICD-9-CM, International Classification of Diseases, 9th edition, Clinical Modification.
METHODS Data source All data were extracted from the NIS from 1998 to 2007. The NIS is maintained as part of the Healthcare Cost and Utilization Project sponsored by the Agency for Healthcare Research and Quality. This database contains discharge abstracts from a 20% stratified sample of nonfederal acutecare hospitals in the United States. The sampling frame includes community and general hospitals and academic medical centers comprising ⬃90% of all hospital discharges in the United States. Each data entry includes a unique identifier, demographic variables (defined as age, gender, race/ethnicity, and median income for ZIP code), source of admission, discharge disposition, primary and secondary diagnoses (up to 15), primary and secondary procedures (up to 15), primary insurance payers, total hospital charges, and length of stay. NIS data concur with www.giejournal.org
the National Hospital Discharge Survey, supporting data reliability.7
Eligibility criteria We included all women hospitalized for NVUGB identified by using diagnostic International Classification of Diseases, 9th edition, Clinical Modification (ICD-9-CM) codes (Table 1) with an indicator variable for pregnancy during admission. These codes were previously validated and implemented and were shown to have good sensitivity and positive predictive value.8-10 A control population was selected among nonpregnant women admitted with NVUGB by using stratified sampling in which 5 nonpregnant control women were randomly selected for each pregnant woman drawn from the same 5-year age group. This sampling schema is depicted in Figure 1. Admissions with missing age data were excluded. A second compariVolume 72, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY
955
Pregnancy and upper GI bleeding
Nguyen et al
as “Hispanic.” All of the other race categories were non-Hispanic. Health insurance type was derived from the primary payer indicated in the hospital discharge abstract, and “self-pay” referred to individuals who had no health insurance coverage. Case-mix adjustment was performed by using the Charlson/Deyo Index.11,12 The Charlson/Deyo Index is a widely used instrument used to characterize and adjust for disease burden, comorbidity, and case mix in administrative data.
Statistical analysis Data were analyzed by using the Stata 10.0 SE software package (Stata Corp, College Station, Texas). Analyses took into account the stratified 2-stage cluster design by using Stata’s SVY (survey data) commands and incorporating individual discharge-level weights. Weighting functions using these hospital and discharge weights were applied to the 20% NIS sample to estimate the total number of obstetric hospitalizations for NVUGB. We used logistic regression to assess the association between pregnancy and likelihood of undergoing upper GI endoscopy while adjusting for gender, age, race/ethnicity, health insurance, comorbidity, teaching hospital status, presence of hypovolemic shock, and need for blood transfusion.
Ethical considerations Figure 1. Diagram of study population showing sampling scheme for the nonpregnant control group with nonvariceal upper GI bleeding. For each pregnant woman, 5 nonpregnant women were randomly selected from the source population of all nonpregnant women from the same 5-year age strata.
The analysis of the NIS uses completely unidentified data with no risk of loss of confidentiality, and an initial expedited review by the Institutional Review Board of the Johns Hopkins Medical Institutions deemed it to be exempt from further ethical review.
RESULTS son group comprised a 10% random sample of all pregnant women admitted for all causes between 1998 and 2005.
Outcome and predictor variables The primary measured outcomes were proportion of women with NVUGB undergoing upper GI endoscopy and in-hospital mortality. We identified upper GI endoscopy procedures by using the following ICD-9-CM procedure codes: 45.13, 45.14, 45.16, 42.23, 42.33, 44.13, and 44.43. Other outcomes measured included proportion of upper GI endoscopies that were performed within 24 hours of admission, requirements for blood transfusion, the presence of hypovolemic shock, and obstetric outcomes including premature delivery, fetal complications/ distress, and fetal death. Data on race and ethnicity were derived from hospital administrative data and included the following categories: White, African American, Hispanic, Asian or Pacific Islander, Native American, or other. For Hispanic patients, ethnicity took precedence over race and was coded 956 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 5 : 2010
Demographic and clinical characteristics There were 1210 pregnant women with NVUGB identified in the NIS, yielding a national estimate of 5923 NVUGB-related obstetric admissions in the United States between 1998 and 2007. The demographic characteristics of pregnant women and sampled nonpregnant control women admitted with NVUGB are summarized in Table 2. Obstetric patients were less likely to have other underlying medical conditions, with 90% having no comorbidities, compared with only 65% in the control group. Pregnant women also were less likely to be non-Hispanic white, more likely to have Medicaid coverage, and more likely to be admitted to teaching hospitals. Table 3 presents the type and location of NVUGB in both the pregnant group and the nonpregnant control group. Mallory-Weiss tear was the most common identified cause of NVUGB among pregnant woman, but more than half were discharged with an unspecified etiology for hematemesis. In contrast, peptic ulcer disease and gastritis were the predominant etiologies of NVUGB in the age-matched nonpregnant control women. Pregnant women were less likely to rewww.giejournal.org
Nguyen et al
Pregnancy and upper GI bleeding
TABLE 2. Patient demographics of pregnant and nonpregnant controls hospitalized for non-variceal upper GI bleeding
Age, y, mean ⫾ SD
Pregnant population (n ⴝ 1210)
Nonpregnant control women (n ⴝ 6050)
25.7 ⫾ 6.4
25.8 ⫾ 6.1
Ethnicity, % White
42.2
60.5*
African American
32.2
22.0*
Hispanic
18.2
11.8*
Asian
1.8
1.9*
Other
5.8
3.7*
Charlson/Deyo Index, %
Pregnant Nonpregnant population control (%) women (%) (n ⴝ 1210) (n ⴝ 6050) Peptic ulcer disease
6.8
22.8
Gastritis
12.1
26.7
Mallory-Weiss tear
25.4
14.2*
Esophagitis or esophageal ulcer
2.5
3.0
Dieulafoy lesion or arteriovenous malformation
0.1
0.6
Anastomotic ulcer
0.2
2.6
Hematemesis, unspecified
53
30.2*
*P ⬍ .0001.
0
90.5
64.9*
1
8.0
19.8*
⬎2
1.6
15.3*
Private
38.4
45.3*
Medicare
1.7
6.3*
Medicaid
49.1
28.9*
Self-pay
8.0
14.4*
Other
2.7
5.1*
Admitted to teaching hospital, %
49.8
43.8*
Urban location, %
83.4
84.5
Weekend admission, %
21.1
23.2
Health insurance, %
*P ⬍ .001.
quire blood transfusion (4.3% vs 15.3%; P ⬍ .0001) and were less likely to present with hypovolemic shock compared with nonpregnant women (7.6% vs 13.8%; P ⬍ .0001).
Endoscopic management of NVUGB Upper GI endoscopy was performed substantially less frequently in pregnant women compared with control women (26% [n ⫽ 319] vs 69% [n ⫽ 4169]; P ⬍ .0001). The proportion of upper GI endoscopies that led to therapeutic intervention was similar between pregnant and nonpregnant women (8.9% vs 7.2%, P ⫽ .3). Among those who underwent upper GI endoscopy during hospitalization, those who were pregnant were less likely to have their procedures within 24 hours of admission (50% vs 57%; P ⫽ .02). The mean time interval from admission to www.giejournal.org
TABLE 3. Etiologies of nonvariceal upper GI bleeding
upper endoscopy was longer for pregnant women compared with nonpregnant women (1.4 vs 1.2 days; P ⫽ .04). Even after restricting our analysis to women who required blood transfusions, only 56% of pregnant women underwent upper GI endoscopy, whereas 87% of nonpregnant women had the procedure (P ⬍ .0001). After adjustment for age, race, health insurance payer, comorbidity, teaching hospital status, requirement for blood transfusion, and presence of hypovolemic shock, the adjusted odds ratio (aOR) for undergoing upper GI endoscopy among pregnant women versus nonpregnant women was 0.19 (95% confidence interval [CI], 0.16-0.22). Having ⱖ1 underlying comorbid condition was associated with higher likelihood of receiving upper GI endoscopy compared with those without any comorbidity (aOR, 1.50; 95% CI, 1.32-1.69). African-American women were less likely to receive upper GI endoscopy than white women (aOR, 0.84; 95% CI, 0.72-0.98). Additionally, those insured by Medicare were less likely than privately insured women to undergo endoscopy (aOR 0.75; 95% CI, 0.59-0.95). Requiring blood transfusion was a strong predictor for undergoing upper GI endoscopy (aOR 3.24; 95% CI, 2.634.00), whereas the presence of hypovolemic shock was a modest negative predictor (aOR, 0.83; 95% CI, 0.71-0.98).
Maternal and fetal outcomes There were no in-hospital deaths among pregnant women with NVUGB, whereas there were 36 in-hospital deaths among nonpregnant women (0.6%; P ⫽ .006). The proportions of pregnant and nonpregnant women requiring surgery for upper GI bleeding were not statistically different (0.1% vs 0.5%, respectively; P ⫽ .06). Average hospital length of stay was shorter among pregnant women compared with nonpregnant women (3.2 vs 3.6 Volume 72, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY
957
Pregnancy and upper GI bleeding
Nguyen et al
TABLE 4. Maternal-fetal outcomes among pregnant women with and without nonvariceal upper GI bleeding (NVUGB) in the Nationwide Inpatient Sample Pregnant population with NVUGB (%)
Pregnant population without NVUGB (%)
0
0.01
Fetal loss
0.4
0.6
Fetal distress/ complications
2.7
9.3*
Premature delivery
7.0
9.1†
Outcome Maternal mortality
*P ⬍ .0001. †P ⫽ .03.
days; P ⫽ .001). However, after adjustment for comorbidity, health insurance payer, calendar year, and hospital characteristics (including teaching status and size), there was no difference in length of stay (⫺3% difference in length of stay; 95% CI, ⫺8% to 3% difference). Having ⱖ2 underlying illnesses was the most significant confounder, contributing to a 54% increase in length of stay compared with those without a comorbidity. To gain a point of reference for maternal-fetal outcomes among pregnant women with NVUGB, we compared their outcomes with those of a random sample from all obstetric admissions in the NIS. These are summarized in Table 4. Although the rates of maternal mortality and fetal loss were well below 1% in both groups, fetal distress/ complications were lower in the pregnant group admitted with NVUGB (2.7% vs 9.3%; P ⬍ .0001), as was premature delivery (7.0% vs 9.1%; P ⫽ .03). Among pregnant women with NVUGB, there were no differences in outcomes between those who did and did not undergo upper GI endoscopy regarding fetal loss (0.2% vs 0.6%, respectively; P ⫽ .13), fetal distress and complications (2.7% vs 2.6%, respectively; P ⫽ .9), and premature delivery (7.3% vs 6.4%, respectively; P ⫽ .6).
DISCUSSION We have conducted the largest analysis of obstetric hospitalizations for NVUGB, and, to the best of our knowledge, this is the first reported population-based study. Our study demonstrated substantially more conservative therapy with lower use of upper GI endoscopy among pregnant women compared with nonpregnant control women. However, we showed no deaths among pregnant women hospitalized for NVUGB, and obstetric outcomes were similar to and even favorable compared with the general obstetric population. 958 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 5 : 2010
The underlying reasons for lower rates of endoscopy use in the pregnant population are likely to be multifactorial. First, there was a higher prevalence of MalloryWeiss tear among pregnant women, which is not unexpected, given the higher occurrence of nausea and vomiting during pregnancy. This etiology of NVUGB tends to be less severe, typically is self-limited, and usually heals spontaneously.13,14 Nonpregnant women, in contrast, were more likely to have peptic ulcer disease, which is more likely to require endoscopic hemostasis. Unfortunately, because the underlying etiology of bleeding is usually dependent on findings of upper GI endoscopy and is therefore influenced by whether upper GI endoscopy is performed, we cannot assess the likelihood of upper GI endoscopy while adjusting for type of bleeding. This drawback is particularly evident when we consider that only one-half of the pregnant women had a specified etiology of bleeding because of the use of endoscopy. However, in a secondary analysis, we accounted for bleeding severity by restricting our comparison to those who required blood transfusions and found lower use of upper GI endoscopy among pregnant women compared with nonpregnant control women. The magnitude of difference in endoscopy rates did decrease among those requiring blood transfusions, which is a more compelling indication for upper GI endoscopy. After simultaneous multivariate adjustment for multiple demographic and clinical confounders, we found that pregnant women were only one-fifth as likely to undergo upper GI endoscopy. These findings would suggest that the threshold for performing upper GI endoscopy was higher in pregnant women than in nonpregnant women, even when patients had similar indications. This more conservative approach may reflect both physician and maternal preferences. It should be noted that among pregnant women who ultimately required upper GI endoscopy, a smaller proportion underwent the procedure within 24 hours of admission compared with nonpregnant women who required endoscopy. This observation, again, may reflect a “wait and see” strategy for pregnant women. International consensus guidelines generally recommend early upper GI endoscopy within 24 hours of admission for NVUGB, because that protocol is associated with better clinical and economic outcomes.1 However, the pregnant women in the present study population may fall into one of the few categories of exceptions, including those patients with a high risk of complications or low likelihood of benefiting from endoscopic intervention.1 Lower rates of upper GI endoscopy among pregnant women did not translate into poorer outcomes. Crude mortality was actually higher in the age-matched nonpregnant women with NVUGB than in those who were pregnant. This observation is likely because of nonpregnant women having greater comorbidity and more severe bleeding on presentation, as manifested by a higher rate of transfusion requirement and the presence of hypovolemic www.giejournal.org
Nguyen et al
shock. Similarly, the fetal outcomes (fetal loss, fetal distress, and premature delivery) were similar to or more favorable than those of the general hospitalized obstetric population. There also was no difference in fetal outcomes comparing pregnant patients with NVUGB who underwent upper GI endoscopy and those who did not. This finding supports conclusions from prior non–populationbased studies that upper GI endoscopy was not associated with adverse fetal outcomes.6 There are several limitations to the present study that are inherent to the use of administrative data. First, we did not have access to detailed clinical data regarding the gestational period of the pregnant women, which could influence the threshold for performing upper GI endoscopy, particularly if a woman was in the first trimester of pregnancy. Second, the data were limited regarding clinical presentation of NVUGB, because we did not have access to patient vital signs or degree of anemia. Instead, we used less refined indicators of severity, such as need for blood transfusion, which implies significant anemia. Moreover, the database was limited to outcomes within the hospital admission and could not assess postdischarge maternal and fetal outcomes. Despite these drawbacks, we were able to use validated coding algorithms to characterize endoscopic management and outcomes of pregnant women with NVUGB by using nationwide population-based data, including data from community hospitals. These data complement and increase the generalizability of an earlier study from tertiary-care centers.6 Furthermore, the use of this large dataset allowed us to draw a sufficient sample size of pregnant patients with NVUGB to enable regression analyses that account for important confounders. In summary, our nationwide study depicted nationwide patterns of endoscopy utilization among pregnant women with NVUGB. It is important to note that pregnant women were more likely to have Mallory-Weiss tears, which tend to be self-limited, and thus were associated with lower mortality, lower requirement for blood transfusion, and less frequent presentation with hypovolemic shock. An important consideration that should be emphasized is that a higher threshold for endoscopy in pregnant women does not imply lower quality of care. The American Society for Gastrointestinal Endoscopy (ASGE) guidelines recommend upper GI endoscopy in pregnant women only if there is persistent or significant bleeding. It may have been quite appropriate to defer endoscopy in a significant proportion of cases, especially because doing so did not affect maternal or fetal outcomes.4 Therefore, a conservative “wait and see” approach appears to be reasonable in pregnant women who remain hemodynamically stable with self-limited NVUGB. The present data, however, support previous studies showing that upper GI endoscopy is
www.giejournal.org
Pregnancy and upper GI bleeding
safe for the mother and fetus when it is clearly warranted. Certainly, there were a small number of women who required blood transfusions and likely had a definitive indication for endoscopy but never received it. Failure to perform upper GI endoscopy under such circumstances runs counter to the ASGE guidelines and may reflect a shortcoming in optimal care. However, it should be acknowledged that maternal preferences also may have played a significant role in the choice to forego endoscopy. The decision of whether to proceed with upper GI endoscopy in pregnant women with NVUGB is inherently more complicated because of the potential risks to the fetus. Therefore, the decision requires careful deliberation of the risk-benefit profile with input from both gastroenterologists and obstetricians. REFERENCES 1. Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2010;19;152:101-13. 2. Storey DW, Bown SG, Swain CP, et al. Endoscopic prediction of recurrent bleeding in peptic ulcers. N Engl J Med 1981;305:915-6. 3. Laine L, McQuaid KR. Endoscopic therapy for bleeding ulcers: an evidence-based approach based on meta-analyses of randomized controlled trials. Clin Gastroenterol Hepatol 2009;7:33-47. 4. Qureshi WA, Rajan E, Adler DG, et al. ASGE guideline: guidelines for endoscopy in pregnant and lactating women. Gastrointest Endosc 2005;61:357-62. 5. Cappell MS. The fetal safety and clinical efficacy of gastrointestinal endoscopy during pregnancy. Gastroenterol Clin North Am 2003;32: 123-79. 6. Cappell MS, Colon VJ, Sidhom OA. A study of eight medical centers of the safety and clinical efficacy of esophagogastroduodenoscopy in 83 pregnant females with follow-up of fetal outcome with comparison control groups. Am J Gastroenterol 1996;91:348-54. 7. Whalen D, Houchens R, Elixhauser A. 2002 HCUP Nationwide Inpatient Sample (NIS) comparison report. No. 2005-03. Rockville (Md.): U.S. Agency for Healthcare Research and Quality, 2005:1-89. 8. Cattaruzzi C, Troncon MG, Agostinis L, et al. Positive predictive value of ICD-9th codes for upper gastrointestinal bleeding and perforation in the Sistema Informativo Sanitario Regionale database. J Clin Epidemiol 1999;52:499-502. 9. Cooper GS, Chak A, Lloyd LE, et al. The accuracy of diagnosis and procedural codes for patients with upper GI hemorrhage. Gastrointest Endosc 2000;51:423-6. 10. Lopushinsky SR, Covarrubia KA, Rabeneck L, et al. Accuracy of administrative health data for the diagnosis of upper gastrointestinal diseases. Surg Endosc 2007;21:1733-7. 11. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130-9. 12. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45: 613-9. 13. Knauer CM. Mallory-Weiss syndrome. Characterization of 75 MalloryWeiss lacerations in 528 patients with upper gastrointestinal hemorrhage. Gastroenterology 1976;71:5-8. 14. Ansari A. Mallory-Weiss syndrome. Experience in a community hospital. Postgrad Med 1984;76:189-95.
Volume 72, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY
959
Endoscopic management and outcomes of pregnant women hospitalized for nonvariceal upper GI bleeding: a nationwide analysis Geoffrey C. Nguyen, MD, PhD, Amreen M. Dinani, MD, Kevin Pivovarov Toronto, Ontario, Canada; Baltimore, Maryland, USA
Background: Upper GI endoscopy has an important diagnostic and therapeutic role in the management of nonvariceal upper GI bleeding (NVUGB). Objective: To characterize nationwide patterns of utilization of upper GI endoscopy in pregnant women with NVUGB and to assess health outcomes. Design: Retrospective cohort study. Setting: Participating hospitals from the Nationwide Inpatient Sample, 1998-2007. Patients: Pregnant and age-matched nonpregnant women admitted for NVUGB. Intervention: The study population was classified as pregnant women with NVUGB (n ⫽ 1210) and nonpregnant women with NVUGB (n ⫽ 6050). Main Outcome Measurements: Rate of upper GI endoscopy, maternal mortality, fetal death/complications, and premature delivery. Results: Pregnant women were less likely than nonpregnant women to undergo upper GI endoscopy (26% vs 69%; P ⬍ .0001) even after adjustment for comorbidities, transfusion requirement, and the presence of hypovolemic shock (adjusted odds ratio, 0.19; 95% confidence interval, 0.16-0.22). Among those who underwent endoscopy, pregnant women were less likely to undergo the procedure within 24 hours of admission (50% vs 57%; P ⫽ .02). Mortality was lower among pregnant women compared with nonpregnant women (0% vs 0.6%; P ⫽ .006). In comparing outcomes between those who did and did not undergo endoscopy, there was no difference in fetal loss (0.2% vs 0.6%), fetal distress/complications (2.7% vs 2.6%), or premature delivery (7.3% vs 6.4%). Limitations: The study was based on administrative data. Conclusion: A conservative nonendoscopic approach is common in the management of pregnant women with NVUGB and is not associated with worse maternal or fetal outcomes. Upper GI endoscopy is, however, safe when judiciously implemented in the actively bleeding patient. (Gastrointest Endosc 2010;72:954-9.)
Acute nonvariceal upper GI bleeding (NVUGB) is a common clinical emergency leading to 50 to 160 hospitalizations per 100,000 adults. Mortality may be as high as 10% to 14%, although this appears to be decreasing in
many countries.1 Upper GI endoscopy has been endorsed as essential in the diagnosis and management of NVUGB. Endoscopy allows assessment of the risk of rebleeding and enables therapeutic hemostasis that reduces bleeding, the
Abbreviations: aOR, adjusted odds ratio; ICD-9-CM, International Classification of Diseases, 9th edition, Clinical Modification; NIS, Nationwide Inpatient Sample; NVUGB, nonvariceal upper GI bleeding.
doi:10.1016/j.gie.2010.07.018
DISCLOSURE: This work was supported by an American Gastroenterological Association Research Scholar Award by the Foundation for Digestive Health and Nutrition to G. Nguyen. The sponsor had no role in the study design or data collection, analysis, or interpretation. No other financial relationships relevant to this publication were disclosed.
Current affiliations: Mount Sinai Division of Gastroenterology, University of Toronto, Toronto, Ontario, Canada (G.C.N., A.M.D., K.P.); Division of Gastroenterology and Hepatology, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA (G.C.N.).
Copyright © 2010 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00
954 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 5 : 2010
Received June 1, 2010. Accepted July 12, 2010.
Reprint requests: Geoffrey C. Nguyen, MD, PhD, 600 University Ave, Ste 437, Toronto, Ontario M5G 1X5, Canada.
www.giejournal.org
Nguyen et al
need for surgery, and mortality.2,3 Although the safety and efficacy of upper GI endoscopy are well established in the general population, the role and use of the procedure in pregnant women is less well characterized. The decision to perform upper GI endoscopy during pregnancy warrants careful deliberation of the risk-benefit profile for both mother and fetus and consideration of alternative therapeutic options. The clinical indications for upper GI endoscopy in pregnancy include (1) significant or ongoing GI bleeding, (2) refractory nausea or vomiting or abdominal pain, and (3) dysphagia or odynophagia.4 During the periprocedural period, the fetus can be particularly vulnerable to fetal hypoxia resulting from maternal oversedation, maternal positioning, or complications of the procedure.5 There are additional risks of premature delivery and teratogenicity associated with certain sedatives. Consequently, it is generally recommended that endoscopic procedures should be delayed until the second trimester when possible.4 Despite these caveats, under circumstances in which therapeutic intervention may be necessary, endoscopy is generally safer in pregnancy than alternative radiologic and surgical interventions.5 In a multicenter retrospective study of 83 pregnant women, upper GI endoscopy was not associated with adverse fetal outcomes and resulted in a 95% diagnostic yield for GI bleeding.6 However, that study may have been underpowered to detect small endoscopy-associated risks to the fetus. It also should be noted that those results were from tertiary-care academic centers, and we have little, if any, population-based data on the maternal and fetal outcomes related to NVUGB. Therefore, we sought to use U.S. population– based discharge data from the Nationwide Inpatient Sample (NIS) to (1) characterize nationwide patterns of utilization of upper GI endoscopy for NVUGB in the pregnant population and (2) assess in-hospital maternal and fetal outcomes associated with NVUGB.
Pregnancy and upper GI bleeding
Take-home Message ●
There was a much higher threshold for performing upper GI endoscopy in pregnant women who were hospitalized for nonvariceal upper GI bleeding. However, this did not necessarily reflect a lower quality of care, because pregnant women presented with less severe bleeding and experienced outcomes similar to or better than nonpregnant control women. Upper GI endoscopy in pregnant women, however, appears to be safe when clinically indicated.
TABLE 1. ICD-9-CM codes for non-variceal upper GI bleeding Diagnosis Esophagitis or esophageal ulcer Mallory-Weiss tear
Code(s) 530.21, 530.82 530.7
Gastric or duodenal ulcer
531.0x, 531.2x, 531.4x, 531.6x, 532.0x, 532.2x, 532.4x, 532.6x, 533.0x, 533.2x, 533.4x, 533.6x
Gastritis
535.01, 535.11, 535.21, 535.31, 535.41, 535.51, 535.61, 535.71
Dieulafoy lesion or arteriovenous malformation Anastomotic ulcer Hematemesis, unspecified
537.83, 537.84 534.0x, 534.2x, 534.4x, 534.6x 578.0
ICD-9-CM, International Classification of Diseases, 9th edition, Clinical Modification.
METHODS Data source All data were extracted from the NIS from 1998 to 2007. The NIS is maintained as part of the Healthcare Cost and Utilization Project sponsored by the Agency for Healthcare Research and Quality. This database contains discharge abstracts from a 20% stratified sample of nonfederal acutecare hospitals in the United States. The sampling frame includes community and general hospitals and academic medical centers comprising ⬃90% of all hospital discharges in the United States. Each data entry includes a unique identifier, demographic variables (defined as age, gender, race/ethnicity, and median income for ZIP code), source of admission, discharge disposition, primary and secondary diagnoses (up to 15), primary and secondary procedures (up to 15), primary insurance payers, total hospital charges, and length of stay. NIS data concur with www.giejournal.org
the National Hospital Discharge Survey, supporting data reliability.7
Eligibility criteria We included all women hospitalized for NVUGB identified by using diagnostic International Classification of Diseases, 9th edition, Clinical Modification (ICD-9-CM) codes (Table 1) with an indicator variable for pregnancy during admission. These codes were previously validated and implemented and were shown to have good sensitivity and positive predictive value.8-10 A control population was selected among nonpregnant women admitted with NVUGB by using stratified sampling in which 5 nonpregnant control women were randomly selected for each pregnant woman drawn from the same 5-year age group. This sampling schema is depicted in Figure 1. Admissions with missing age data were excluded. A second compariVolume 72, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY
955
Pregnancy and upper GI bleeding
Nguyen et al
as “Hispanic.” All of the other race categories were non-Hispanic. Health insurance type was derived from the primary payer indicated in the hospital discharge abstract, and “self-pay” referred to individuals who had no health insurance coverage. Case-mix adjustment was performed by using the Charlson/Deyo Index.11,12 The Charlson/Deyo Index is a widely used instrument used to characterize and adjust for disease burden, comorbidity, and case mix in administrative data.
Statistical analysis Data were analyzed by using the Stata 10.0 SE software package (Stata Corp, College Station, Texas). Analyses took into account the stratified 2-stage cluster design by using Stata’s SVY (survey data) commands and incorporating individual discharge-level weights. Weighting functions using these hospital and discharge weights were applied to the 20% NIS sample to estimate the total number of obstetric hospitalizations for NVUGB. We used logistic regression to assess the association between pregnancy and likelihood of undergoing upper GI endoscopy while adjusting for gender, age, race/ethnicity, health insurance, comorbidity, teaching hospital status, presence of hypovolemic shock, and need for blood transfusion.
Ethical considerations Figure 1. Diagram of study population showing sampling scheme for the nonpregnant control group with nonvariceal upper GI bleeding. For each pregnant woman, 5 nonpregnant women were randomly selected from the source population of all nonpregnant women from the same 5-year age strata.
The analysis of the NIS uses completely unidentified data with no risk of loss of confidentiality, and an initial expedited review by the Institutional Review Board of the Johns Hopkins Medical Institutions deemed it to be exempt from further ethical review.
RESULTS son group comprised a 10% random sample of all pregnant women admitted for all causes between 1998 and 2005.
Outcome and predictor variables The primary measured outcomes were proportion of women with NVUGB undergoing upper GI endoscopy and in-hospital mortality. We identified upper GI endoscopy procedures by using the following ICD-9-CM procedure codes: 45.13, 45.14, 45.16, 42.23, 42.33, 44.13, and 44.43. Other outcomes measured included proportion of upper GI endoscopies that were performed within 24 hours of admission, requirements for blood transfusion, the presence of hypovolemic shock, and obstetric outcomes including premature delivery, fetal complications/ distress, and fetal death. Data on race and ethnicity were derived from hospital administrative data and included the following categories: White, African American, Hispanic, Asian or Pacific Islander, Native American, or other. For Hispanic patients, ethnicity took precedence over race and was coded 956 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 5 : 2010
Demographic and clinical characteristics There were 1210 pregnant women with NVUGB identified in the NIS, yielding a national estimate of 5923 NVUGB-related obstetric admissions in the United States between 1998 and 2007. The demographic characteristics of pregnant women and sampled nonpregnant control women admitted with NVUGB are summarized in Table 2. Obstetric patients were less likely to have other underlying medical conditions, with 90% having no comorbidities, compared with only 65% in the control group. Pregnant women also were less likely to be non-Hispanic white, more likely to have Medicaid coverage, and more likely to be admitted to teaching hospitals. Table 3 presents the type and location of NVUGB in both the pregnant group and the nonpregnant control group. Mallory-Weiss tear was the most common identified cause of NVUGB among pregnant woman, but more than half were discharged with an unspecified etiology for hematemesis. In contrast, peptic ulcer disease and gastritis were the predominant etiologies of NVUGB in the age-matched nonpregnant control women. Pregnant women were less likely to rewww.giejournal.org
Nguyen et al
Pregnancy and upper GI bleeding
TABLE 2. Patient demographics of pregnant and nonpregnant controls hospitalized for non-variceal upper GI bleeding
Age, y, mean ⫾ SD
Pregnant population (n ⴝ 1210)
Nonpregnant control women (n ⴝ 6050)
25.7 ⫾ 6.4
25.8 ⫾ 6.1
Ethnicity, % White
42.2
60.5*
African American
32.2
22.0*
Hispanic
18.2
11.8*
Asian
1.8
1.9*
Other
5.8
3.7*
Charlson/Deyo Index, %
Pregnant Nonpregnant population control (%) women (%) (n ⴝ 1210) (n ⴝ 6050) Peptic ulcer disease
6.8
22.8
Gastritis
12.1
26.7
Mallory-Weiss tear
25.4
14.2*
Esophagitis or esophageal ulcer
2.5
3.0
Dieulafoy lesion or arteriovenous malformation
0.1
0.6
Anastomotic ulcer
0.2
2.6
Hematemesis, unspecified
53
30.2*
*P ⬍ .0001.
0
90.5
64.9*
1
8.0
19.8*
⬎2
1.6
15.3*
Private
38.4
45.3*
Medicare
1.7
6.3*
Medicaid
49.1
28.9*
Self-pay
8.0
14.4*
Other
2.7
5.1*
Admitted to teaching hospital, %
49.8
43.8*
Urban location, %
83.4
84.5
Weekend admission, %
21.1
23.2
Health insurance, %
*P ⬍ .001.
quire blood transfusion (4.3% vs 15.3%; P ⬍ .0001) and were less likely to present with hypovolemic shock compared with nonpregnant women (7.6% vs 13.8%; P ⬍ .0001).
Endoscopic management of NVUGB Upper GI endoscopy was performed substantially less frequently in pregnant women compared with control women (26% [n ⫽ 319] vs 69% [n ⫽ 4169]; P ⬍ .0001). The proportion of upper GI endoscopies that led to therapeutic intervention was similar between pregnant and nonpregnant women (8.9% vs 7.2%, P ⫽ .3). Among those who underwent upper GI endoscopy during hospitalization, those who were pregnant were less likely to have their procedures within 24 hours of admission (50% vs 57%; P ⫽ .02). The mean time interval from admission to www.giejournal.org
TABLE 3. Etiologies of nonvariceal upper GI bleeding
upper endoscopy was longer for pregnant women compared with nonpregnant women (1.4 vs 1.2 days; P ⫽ .04). Even after restricting our analysis to women who required blood transfusions, only 56% of pregnant women underwent upper GI endoscopy, whereas 87% of nonpregnant women had the procedure (P ⬍ .0001). After adjustment for age, race, health insurance payer, comorbidity, teaching hospital status, requirement for blood transfusion, and presence of hypovolemic shock, the adjusted odds ratio (aOR) for undergoing upper GI endoscopy among pregnant women versus nonpregnant women was 0.19 (95% confidence interval [CI], 0.16-0.22). Having ⱖ1 underlying comorbid condition was associated with higher likelihood of receiving upper GI endoscopy compared with those without any comorbidity (aOR, 1.50; 95% CI, 1.32-1.69). African-American women were less likely to receive upper GI endoscopy than white women (aOR, 0.84; 95% CI, 0.72-0.98). Additionally, those insured by Medicare were less likely than privately insured women to undergo endoscopy (aOR 0.75; 95% CI, 0.59-0.95). Requiring blood transfusion was a strong predictor for undergoing upper GI endoscopy (aOR 3.24; 95% CI, 2.634.00), whereas the presence of hypovolemic shock was a modest negative predictor (aOR, 0.83; 95% CI, 0.71-0.98).
Maternal and fetal outcomes There were no in-hospital deaths among pregnant women with NVUGB, whereas there were 36 in-hospital deaths among nonpregnant women (0.6%; P ⫽ .006). The proportions of pregnant and nonpregnant women requiring surgery for upper GI bleeding were not statistically different (0.1% vs 0.5%, respectively; P ⫽ .06). Average hospital length of stay was shorter among pregnant women compared with nonpregnant women (3.2 vs 3.6 Volume 72, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY
957
Pregnancy and upper GI bleeding
Nguyen et al
TABLE 4. Maternal-fetal outcomes among pregnant women with and without nonvariceal upper GI bleeding (NVUGB) in the Nationwide Inpatient Sample Pregnant population with NVUGB (%)
Pregnant population without NVUGB (%)
0
0.01
Fetal loss
0.4
0.6
Fetal distress/ complications
2.7
9.3*
Premature delivery
7.0
9.1†
Outcome Maternal mortality
*P ⬍ .0001. †P ⫽ .03.
days; P ⫽ .001). However, after adjustment for comorbidity, health insurance payer, calendar year, and hospital characteristics (including teaching status and size), there was no difference in length of stay (⫺3% difference in length of stay; 95% CI, ⫺8% to 3% difference). Having ⱖ2 underlying illnesses was the most significant confounder, contributing to a 54% increase in length of stay compared with those without a comorbidity. To gain a point of reference for maternal-fetal outcomes among pregnant women with NVUGB, we compared their outcomes with those of a random sample from all obstetric admissions in the NIS. These are summarized in Table 4. Although the rates of maternal mortality and fetal loss were well below 1% in both groups, fetal distress/ complications were lower in the pregnant group admitted with NVUGB (2.7% vs 9.3%; P ⬍ .0001), as was premature delivery (7.0% vs 9.1%; P ⫽ .03). Among pregnant women with NVUGB, there were no differences in outcomes between those who did and did not undergo upper GI endoscopy regarding fetal loss (0.2% vs 0.6%, respectively; P ⫽ .13), fetal distress and complications (2.7% vs 2.6%, respectively; P ⫽ .9), and premature delivery (7.3% vs 6.4%, respectively; P ⫽ .6).
DISCUSSION We have conducted the largest analysis of obstetric hospitalizations for NVUGB, and, to the best of our knowledge, this is the first reported population-based study. Our study demonstrated substantially more conservative therapy with lower use of upper GI endoscopy among pregnant women compared with nonpregnant control women. However, we showed no deaths among pregnant women hospitalized for NVUGB, and obstetric outcomes were similar to and even favorable compared with the general obstetric population. 958 GASTROINTESTINAL ENDOSCOPY
Volume 72, No. 5 : 2010
The underlying reasons for lower rates of endoscopy use in the pregnant population are likely to be multifactorial. First, there was a higher prevalence of MalloryWeiss tear among pregnant women, which is not unexpected, given the higher occurrence of nausea and vomiting during pregnancy. This etiology of NVUGB tends to be less severe, typically is self-limited, and usually heals spontaneously.13,14 Nonpregnant women, in contrast, were more likely to have peptic ulcer disease, which is more likely to require endoscopic hemostasis. Unfortunately, because the underlying etiology of bleeding is usually dependent on findings of upper GI endoscopy and is therefore influenced by whether upper GI endoscopy is performed, we cannot assess the likelihood of upper GI endoscopy while adjusting for type of bleeding. This drawback is particularly evident when we consider that only one-half of the pregnant women had a specified etiology of bleeding because of the use of endoscopy. However, in a secondary analysis, we accounted for bleeding severity by restricting our comparison to those who required blood transfusions and found lower use of upper GI endoscopy among pregnant women compared with nonpregnant control women. The magnitude of difference in endoscopy rates did decrease among those requiring blood transfusions, which is a more compelling indication for upper GI endoscopy. After simultaneous multivariate adjustment for multiple demographic and clinical confounders, we found that pregnant women were only one-fifth as likely to undergo upper GI endoscopy. These findings would suggest that the threshold for performing upper GI endoscopy was higher in pregnant women than in nonpregnant women, even when patients had similar indications. This more conservative approach may reflect both physician and maternal preferences. It should be noted that among pregnant women who ultimately required upper GI endoscopy, a smaller proportion underwent the procedure within 24 hours of admission compared with nonpregnant women who required endoscopy. This observation, again, may reflect a “wait and see” strategy for pregnant women. International consensus guidelines generally recommend early upper GI endoscopy within 24 hours of admission for NVUGB, because that protocol is associated with better clinical and economic outcomes.1 However, the pregnant women in the present study population may fall into one of the few categories of exceptions, including those patients with a high risk of complications or low likelihood of benefiting from endoscopic intervention.1 Lower rates of upper GI endoscopy among pregnant women did not translate into poorer outcomes. Crude mortality was actually higher in the age-matched nonpregnant women with NVUGB than in those who were pregnant. This observation is likely because of nonpregnant women having greater comorbidity and more severe bleeding on presentation, as manifested by a higher rate of transfusion requirement and the presence of hypovolemic www.giejournal.org
Nguyen et al
shock. Similarly, the fetal outcomes (fetal loss, fetal distress, and premature delivery) were similar to or more favorable than those of the general hospitalized obstetric population. There also was no difference in fetal outcomes comparing pregnant patients with NVUGB who underwent upper GI endoscopy and those who did not. This finding supports conclusions from prior non–populationbased studies that upper GI endoscopy was not associated with adverse fetal outcomes.6 There are several limitations to the present study that are inherent to the use of administrative data. First, we did not have access to detailed clinical data regarding the gestational period of the pregnant women, which could influence the threshold for performing upper GI endoscopy, particularly if a woman was in the first trimester of pregnancy. Second, the data were limited regarding clinical presentation of NVUGB, because we did not have access to patient vital signs or degree of anemia. Instead, we used less refined indicators of severity, such as need for blood transfusion, which implies significant anemia. Moreover, the database was limited to outcomes within the hospital admission and could not assess postdischarge maternal and fetal outcomes. Despite these drawbacks, we were able to use validated coding algorithms to characterize endoscopic management and outcomes of pregnant women with NVUGB by using nationwide population-based data, including data from community hospitals. These data complement and increase the generalizability of an earlier study from tertiary-care centers.6 Furthermore, the use of this large dataset allowed us to draw a sufficient sample size of pregnant patients with NVUGB to enable regression analyses that account for important confounders. In summary, our nationwide study depicted nationwide patterns of endoscopy utilization among pregnant women with NVUGB. It is important to note that pregnant women were more likely to have Mallory-Weiss tears, which tend to be self-limited, and thus were associated with lower mortality, lower requirement for blood transfusion, and less frequent presentation with hypovolemic shock. An important consideration that should be emphasized is that a higher threshold for endoscopy in pregnant women does not imply lower quality of care. The American Society for Gastrointestinal Endoscopy (ASGE) guidelines recommend upper GI endoscopy in pregnant women only if there is persistent or significant bleeding. It may have been quite appropriate to defer endoscopy in a significant proportion of cases, especially because doing so did not affect maternal or fetal outcomes.4 Therefore, a conservative “wait and see” approach appears to be reasonable in pregnant women who remain hemodynamically stable with self-limited NVUGB. The present data, however, support previous studies showing that upper GI endoscopy is
www.giejournal.org
Pregnancy and upper GI bleeding
safe for the mother and fetus when it is clearly warranted. Certainly, there were a small number of women who required blood transfusions and likely had a definitive indication for endoscopy but never received it. Failure to perform upper GI endoscopy under such circumstances runs counter to the ASGE guidelines and may reflect a shortcoming in optimal care. However, it should be acknowledged that maternal preferences also may have played a significant role in the choice to forego endoscopy. The decision of whether to proceed with upper GI endoscopy in pregnant women with NVUGB is inherently more complicated because of the potential risks to the fetus. Therefore, the decision requires careful deliberation of the risk-benefit profile with input from both gastroenterologists and obstetricians. REFERENCES 1. Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2010;19;152:101-13. 2. Storey DW, Bown SG, Swain CP, et al. Endoscopic prediction of recurrent bleeding in peptic ulcers. N Engl J Med 1981;305:915-6. 3. Laine L, McQuaid KR. Endoscopic therapy for bleeding ulcers: an evidence-based approach based on meta-analyses of randomized controlled trials. Clin Gastroenterol Hepatol 2009;7:33-47. 4. Qureshi WA, Rajan E, Adler DG, et al. ASGE guideline: guidelines for endoscopy in pregnant and lactating women. Gastrointest Endosc 2005;61:357-62. 5. Cappell MS. The fetal safety and clinical efficacy of gastrointestinal endoscopy during pregnancy. Gastroenterol Clin North Am 2003;32: 123-79. 6. Cappell MS, Colon VJ, Sidhom OA. A study of eight medical centers of the safety and clinical efficacy of esophagogastroduodenoscopy in 83 pregnant females with follow-up of fetal outcome with comparison control groups. Am J Gastroenterol 1996;91:348-54. 7. Whalen D, Houchens R, Elixhauser A. 2002 HCUP Nationwide Inpatient Sample (NIS) comparison report. No. 2005-03. Rockville (Md.): U.S. Agency for Healthcare Research and Quality, 2005:1-89. 8. Cattaruzzi C, Troncon MG, Agostinis L, et al. Positive predictive value of ICD-9th codes for upper gastrointestinal bleeding and perforation in the Sistema Informativo Sanitario Regionale database. J Clin Epidemiol 1999;52:499-502. 9. Cooper GS, Chak A, Lloyd LE, et al. The accuracy of diagnosis and procedural codes for patients with upper GI hemorrhage. Gastrointest Endosc 2000;51:423-6. 10. Lopushinsky SR, Covarrubia KA, Rabeneck L, et al. Accuracy of administrative health data for the diagnosis of upper gastrointestinal diseases. Surg Endosc 2007;21:1733-7. 11. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130-9. 12. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45: 613-9. 13. Knauer CM. Mallory-Weiss syndrome. Characterization of 75 MalloryWeiss lacerations in 528 patients with upper gastrointestinal hemorrhage. Gastroenterology 1976;71:5-8. 14. Ansari A. Mallory-Weiss syndrome. Experience in a community hospital. Postgrad Med 1984;76:189-95.
Volume 72, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY
959