Original Contribution

emergency endoscopy for Upper gastrointestinal bleeding in patients with coronary artery disease^?,??

Ping-Huei Tseng MD^a, Jyh-Ming Liou MD^b, Yi-Chia Lee MD, MSc^b,c, Lian-Yu Lin MD, PhD^b, Alyssa Yan-Zhen Liu BBA, MSc^c, Dun-Cheng Chang BBA, MSc^c, Han-Mo Chiu MD^b,

Ming-Shiang Wu MD, PhD^b, Jaw-Town Lin MD, PhD^d, Hsiu-Po Wang MD^a,b,?

^aDepartment of Internal Medicine, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin County 640, Taiwan

^bDepartments of Internal Medicine, National Taiwan University Hospital and College of Medicine, National Taiwan University,

Taipei 100, Taiwan

^cDivision of Biostatistics, Graduate Institute of Epidemiology, and Center of Biostatistics Consultation,

College of Public Health, National Taiwan University, Taipei 100, Taiwan

^dDepartment of Internal Medicine, E-DA Hospital and I-Shou University, Kaohsiung County 824, Taiwan

Received 9 June 2008; revised 22 June 2008; accepted 24 June 2008

Abstract

Background: Endoscopy is useful for diagnosis and treatment of upper gastrointestinal bleeding . However, both endoscopy and UGIB may compromise the cardiovascular function. The present study is to investigate the cardiovascular responses of emergency endoscopy for patients with UGIB and stable coronary artery disease (CAD).

Methods: Consecutive 50 patients with known CAD and 50 patients without CAD history (non-CAD group) in whom emergency endoscopy was requested for UGIB were prospectively enrolled. All patients received ambulatory electrocardiographic monitoring before, during, and after endoscopies. cardiac indices including supraventricular and ventricular arrhythmia, ST ischemic change, and autonomic nervous function evaluated by Heart rate variability were compared.

Results: All patients in both groups had successful primary hemostasis, and Peptic ulcer bleeding was the main etiology (82%). Compared with the non-CAD group, patients with CAD had a significantly higher incidence (42% vs 16%, P = .004) and frequency (1.19 vs 0.12 events per minute, P = .003) of ventricular arrhythmias during endoscopy. Nine patients with CAD and 1 patient without CAD had ischemic STchanges (P = .016). Comorbidity with congestive heart failure was not only associated with a higher frequency (P = .02) but also a more severe fluctuation (P = .002) of ventricular arrhythmia. None in both groups had angina or MI before, during, or after endoscopy. Heart rate variability did not show a difference. Conclusions: Ventricular arrhythmias and myocardial ischemia, although mostly subclinical, were common in patients with stable CAD undergoing emergent endoscopy for UGIB, especially in those with concomitant congestive heart failure.

(C) 2009

^? The study was supported by Research Grants from the National Taiwan University Hospital, Taipei, Taiwan (94S162).

^?? ClinicalTrials.gov (Identifier: no., NCT00173511).

* Corresponding author. Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan.

E-mail address: [email protected] (H.-P. Wang).

0735-6757/$ - see front matter (C) 2009 doi:10.1016/j.ajem.2008.06.018

Introduction

Upper gastrointestinal bleeding (UGIB), a common disorder encountered in the emergency department, may cause hypotension and tachycardia capable of compromis- ing cardiovascular function [1-5]. Endoscopy is an impor- tant diagnostic therapeutic tool in managing patients with UGIB; this strategy can identify the bleeding source(s), achieve subsequent hemostasis, and limit the need for surgical intervention [6-8]. Nonetheless, endoscopy itself may precipitate Cardiovascular complications, including arrhythmias, myocardial ischemia, and infarction [9-22]. These complications are thought to occur with an estimated frequency of 0.005% to 0.5% [23-25]. Endoscopic insertion, pneumatic inflation, and one or more Therapeutic procedures may provoke Catecholamine release and increase oxygen demand, which may contribute to the development of cardiovascular events [26].

For patients with stable coronary artery disease (CAD), defined as patients who have previously diagnosed CAD without angina or symptom complex remaining stable for at least 60 days, the risk of UGIB increases because of the frequent use of Antiplatelet agents such as aspirin and/or clopidogrel with an overall prevalence of 1.5% per year in patients with CAD on low-dose aspirin [27-29]. However, the safety and efficacy of emergency endoscopy in patients with stable CAD and concomitant UGIB remains a clinical dilemma. Wilcox et al [30] studied 25 hospitalized patients with CAD who underwent Endoscopic procedures while sedated. In these patients, endoscopy rarely resulted in serious cardiovascular events. In contrast, Lee et al [31] reported that procedure-related myocardial ischemia occurred in 16% of patients with CAD undergoing therapeutic endoscopy. In addition to cardiac events, substantial autonomic imbalance has also been detected during endoscopy [12]. Thus, cardiac arrhythmia, ischemia, and autonomic derangement may precipitate the occurrence of life-threatening cardiac events during emergent endo- scopy and ultimately lead to resuscitation efforts. Therefore, it is imperative to detect these electrophysiologic changes in the preclinical phase to prevent irReversible changes.

Goals of this investigation

Using a comprehensive evaluation of cardiac indices, the purpose of this study was to examine the cardiovascular responses, including preclinical cardiac arrhythmias and ischemia, and autonomic derangements at different stages of emergencyendoscopyinpatientswithstable CADand UGIB.

Methods

Study population

This prospective study was performed in the endoscopy department of a Tertiary university hospital from April 2005

to November 2006. Adult patients with a history of CAD requiring emergency endoscopy for UGIB were consecu- tively enrolled. During the same study period, UGIB patients without a history of CAD or typical angina were also recruited as a comparative non-CAD group. Upper gastro- intestinal bleeding was diagnosed based on the presence of the following symptoms: melena, coffee-ground vomitus, and frank hematemesis. The diagnosis of CAD was based on the following criteria: more than 50% stenosis in the lumen of at least one coronary artery on prior coronary angiography, positive treadmill exercise test, or a history of myocardial infarction (MI). Exclusion criteria included patients with acute MI, shock unresponsive to resuscitation, underlying advanced malignancy, pregnancy, diagnosed liver cirrhosis, and active lung disease requiring ventilator support. Institu- tional review board of the hospital approved the study protocol before implementation, and a written informed consent was obtained in all patients.

Study protocol

Before endoscopic consultation, all patients were stabi- lized via fluid resuscitation. The medical histories were recorded in detail, including demographic data, history of CAD, comorbid diseases, and current use of medication. Emergency endoscopy was defined as an endoscopic procedure performed within 24 hours after admission to the emergency department and/or consultation of the endoscopist. Laboratory tests included complete blood cell counts, cardiac enzymes, and a 12-lead electrocardiography before and after endoscopy. Myocardial infarction was defined by the presence of at least 2 of the following criteria: angina, electrocardiographic changes, and elevation of cardiac enzymes. Those who met the criteria of acute MI were monitored closely with clinical symptoms and serial follow-up of cardiac enzymes and 12-lead electrocardiogra- phy. Cardiologic consultation and further intensive care would be arranged. A 3-channel ambulatory continuous ECG recorder was carried by each patient for at least 30 minutes before the endoscopic procedure until 2 hours after the completion of the procedure.

Endoscopy

Because of the increased risk of aspiration in patients with active UGIB [32], premedication was limited to butyl scopolamine and oropharyngeal Topical anesthesia. All patients received 3 L/min nasal oxygen during endoscopic examination, and oxygen saturation and blood pressure were monitored continuously during the procedure.

Endoscopy and subsequent hemostasis were performed by experienced endoscopists. Bleeding was classified according to the Forrest classification as follows: Ia and Ib defined active spurting and oozing vessels, respec- tively; IIa, IIb, and IIc described exposed vessel, adherent

blood clot, and hematin coatings over the ulcer base, respectively; and III denoted clean ulcer base with no signs of recent bleeding [33]. The method used to achieve endoscopic hemostasis was recorded (eg, epinephrine injection, heater probe thermocoagulation, hemoclipping, or a combination approach).

Electrocardiographic evaluation

The ECG recordings were analyzed with a standard analysis system (Zymed 2010 Plus, Philips, Andover, MA, USA) and interpreted by a cardiologist (LYL). Three cardiac domains were evaluated, including the cardiac arrhythmia, ischemic STchange, and heart rate variability . Cardiac arrhythmia indicated ventricular ectopy (VE) and supraven- tricular ectopy (SVE). Ventricular ectopy included premature ventricular contractions (PVCs), couplets (2 consecutive PVCs), triplets (3 consecutive PVCs), ventricular runs (>=4 consecutive PVCs), and the R on T phenomenon. Supraventricular ectopy included premature atrial contrac- tions (PACs), atrial pairs (2 consecutive PACs), and atrial runs (>=3 consecutive PACs). Ischemic ST changes indicated the presence of ST-segment elevation or depression of more than 1 mm for at least 1 minute.

Autonomic nervous function

Heart rate variability was used to assess the autonomic nervous function as previously reported [34]. Briefly, the normal-to-normal R-R intervals (N-N intervals) were deduced from adjacent normal sinus beats. The N-N interval time series was transferred to a personal computer and processed by a program written in Matlab language (version 6.0, Mathworks, Natick, Mass) [35,36]. Using Fourier transformation, these time series data were transformed into the power spectra, which included the low-frequency (LF) power (reflecting sympathetic and parasympathetic tones) and high-frequency (HF) power (indicating para- sympathetic tone). These HRV indices were calculated and compared in the before, during, and after endoscopy periods, respectively [37].

Clinical outcomes

Myocardial infarction was defined as described before. Other clinical outcomes included successful primary hemostasis, in-hospital recurrent bleeding, duration of hospital stay, and cardiovascular events during hospitali- zation. Successful primary hemostasis was defined as endoscopically verified cessation of bleeding for at least 1 minute after the first endoscopic hemostasis. Recurrent bleeding was defined as the presence of clinical hematemesis/melena persisting for more than 48 hours and/or a decrease in hemoglobin level greater than 2 g/dL after successful endoscopic hemostasis.

Statistical analysis

Categorical data were expressed as percentage (%) and compared using the ?2 test or Fisher exact test when appropriate. Continuous data were expressed as mean +- SD

Table 1 Demographic data from patients with and without CAD

Characteristic CAD Non-CAD P (n = 50) (n = 50)

Age, mean +- SD (y) 71.1 +- 11.5 67.1 +- 11.3 .085 Sex (male/female) 36/14 26/24 .040 ? Smoking (n [%]) 13 (26) 4 (8) .017 ? Previous PUD (n [%]) 20 (40) 20 (40) 1.0 Comorbidity Diabetes (n [%]) 25 (50) 18 (36) .157 Hypertension (n [%]) 41 (82) 30 (60) .015 ? Hyperlipidemia (n [%]) 20 (40) 6 (12) .001 ? Obesity (BMI >=27), 9 (18) 9 (18) 1.0 n (%) Heart failure (n [%]) 14 (28) 2 (4) .001 ? Atrial fibrillation 5 (10) 0 (0) .022 ? (n [%]) Renal failure (n [%]) 8 (16) 1 (2) .031 ? Medication Aspirin (n [%]) 29 (58) 10 (20) b.001 ? Clopidogrel (n [%]) 29 (58) 1 (2) b.001 ? NSAID (n [%]) 7 (14) 10 (20) .425 Warfarin (n [%]) 4 (8) 0 (0) .118 ?-blocker (n [%]) 19 (38) 3 (6) b.001 ? Calcium channel 27 (54) 10 (20) b.001 ? blocker (n [%]) Preendoscopy condition Hematemesis (n [%]) 5 (10) 6 (12) .752 Coffee-ground 11 (22) 13 (26) .644 vomitus (n [%]) Melena (n [%]) 40 (80) 42 (84) .602 Lowest SBP, 109.4 +- 21.7 118.6 +- 25.7 .058 mean +- SD (mm Hg) Lowest DBP, 59.8 +- 14.7 68.5 +- 16.6 .007 ? mean +- SD (mm Hg) Fastest HR, 96.1 +- 18.3 95.2 +- 20.9 .833 mean +- SD (bpm) Hemoglobin, 9.8 +- 3.1 9.7 +- 2.8 .889 mean +- SD (g/dL) Troponin I, 0.17 +- 0.5 0.03 +- 0.1 .060 mean +- SD (ng/mL) Blood transfusion 32 (64) 29 (58) .539 (n [%]) Blood transfusion 3.8 +- 2.01 3.7 +- 2.1 .913 volume, mean +- SD (U) Butyl scopolamine 48 (96) 48 (96) 1.0 use (n [%])

PUD indicates Peptic ulcer disease; BMI, body mass index, indicating weight in kilogram divided by body surface area; bpm, beats per minute; NSAID, nonsteroidal antiinflammatory drug; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate. * P b .05 with the t test or ?2 test.

and compared with the independent t test or Wilcoxon rank sum test when appropriate. Risk factors associating with cardiac events were also assessed. To achieve this, Linear regression analysis was used. We treated the SVE, VE, ischemic ST change, and autonomic functional change as outcome variables and treated factors of interest as predicting variables. All P values were 2-tailed, with the level of statistical significance specified as .05. All analyses were performed using a statistical software package (SAS version 9.1; SAS Institute Inc, Cary, NC).

Parameters

CAD

(n = 50)

Non-CAD P

(n = 50)

SVE total number >=1 (n [%]) Before 32 (64)

During 23 (46)

SVE frequency >=1/min (n [%])

32 (64)

13 (26)

After 34 (68) 41 (82)

1.0

.038 ?

.106

Before During

4 (8)

8 (16)

VE total number >=1 (n [%])

4 (8)

1 (2)

After 4 (8) 3 (6)

1.0

.014 ?

.699

VE frequency >=1/min (n [%])

Before During After

25 (50)

21 (42)

32 (64)

16 (32)

8 (16)

15 (30)

.067

.004 ?

.001 ?

Before During After

9 (18)

10 (20)

10 (20)

2 (4)

2 (4)

0 (0)

.025 ?

.014 ?

.001 ?

Ischemic ST changes (n [%])

Before During After

Maximal HR, mean (bpm)

Minimal HR, mean (bpm)

6 (12)

9 (18)

9 (18)

126.8

1 (2)

1 (2)

1 (2)

131.7

.112

.016 ?

.016 ?

.304

69.5

68.9

.847

SBP, mean +- SD (mm Hg)

Before After

133.0 +- 25.0 135.8 +- 25.7 .583

140.4 +- 28.0 146.5 +- 32.7 .318

DBP, mean +- SD (mm Hg) Before

After

HR, mean +- SD (bpm) Before

After

SpO2, mean +- SD (%) Before

After

73.3 +- 14.6

74.8 +- 16.6

76.5 +- 16.2 .302

81.8 +- 20.0 .060

91.7 +- 15.7

96.5 +- 21.9

99.1 +- 15.8 106.5 +- 19.4

.210

.041 ?

LF, mean +- SD (ln msec²/Hz)

98.5 +- 1.5 98.3 +- 1.7 .704

98.4 +- 1.6 98.5 +- 1.3 .734

HF, mean +- SD (ln msec²/Hz)

Before 3.38 +- 0.18 3.41 +- 0.18 .389

During 3.15 +- 0.35 3.16 +- 0.23 .928

After 3.29 +- 0.17 3.28 +- 0.37 .830

Before During After

LF/HF ratio (mean +- SD)

2.77 +- 0.18

2.53 +- 0.36

2.67 +- 0.17

2.80 +- 0.20 .480

2.53 +- 0.23 .956

2.65 +- 0.36 .734

Results

Study population

Of the 104 patients recruited into this study, 100 patients completed the examination, including 50 patients with known CAD and 50 patients without CAD. Reasons for withdrawing included intolerance to endoscopy (n = 1), failed ECG recording (n = 1), and refusal for further evaluation (n = 2).

Baseline demographics are summarized in Table 1. Male sex, smoking, the presence of comorbid diseases, and

Table 2 Endoscopic bleeders, hemostatic methods, outcomes after endoscopy, and comparisons between the 2 groups

Parameters	CAD (n = 50)	Non-CAD (n = 50)	P
Peptic ulcer Ulcer size, mean (cm) Forrest Ia (n [%]) Forrest Ib (n [%]) Forrest IIa (n [%]) Forrest IIb (n [%]) Forrest IIc (n [%]) Forrest III (n [%]) Angiodysplasia (n [%]) Therapeutic hemostasis (n [%]) Epinephrine injection (n [%]) Epinephrine volume, mean (mL) Heater probe thermocoagulation Endoscopy duration, mean (min) Successful initial hemostasis (n [%]) In-hospital recurrent bleeding (n [%]) Hospital stay (d) In-hospital cardiac death (n [%])	43 (86)	39 (78)	.3
	0.81	0.82	.967
	2 (4)	0 (0)	.242
	3 (6.)	7 (14)	.318
	6 (12)	12 (24)	.130
	1 (2)	2 (4)	1.000
	8 (16)	4 (8)	.204
	23 (46)	14 (28)	.051
	1 (2)	1 (2)	1.0
	13 (26)	21 (42)	.093
	12 (24)	16 (32)	.378
	9.9	8.8	.581
	6 (12)	6 (12)	1.0
	8.6	10.9	.083
	50 (100)	50 (100)	1.0
	2 (4)	3 (6)	.65
	8.0 +- 17.7	5.3 +- 5.5	.306
	1 (2)	0 (0)	-

Before	4.09 +- 0.37	4.13 +- 0.39	.607
During After	4.18 +- 0.37 4.18 +- 0.25	4.20 +- 0.30 4.25 +- 0.20	.831 .145

prescribed medication were more frequent in the CAD group. Most CAD diagnoses were based on coronary angiography (n = 47). Among them, 8 patients were

Table 3 Electrocardiographic, hemodynamic, and autonomic abnormalities before, during, and after emergency endoscopy, stratified by CAD and non-CAD groups

* P b .05 with the Wilcoxon rank sum test or Fisher’s exact test.

diagnosed with 1-vessel disease (VD), 17 with 2-VD, and 22 with 3-VD. Of the remaining 3 patients, one was diagnosed based on a positive exercise ECG testing, whereas the other 2 patients had histories of MI. Among the 35 patients who received coronary angioplasty, 30 underwent coronary stenting and 14 were treated via coronary artery bypass surgery.

Endoscopic hemostasis

Table 2 illustrates the characteristics of bleeders and the hemostatic methods used in this study. Peptic ulcer disease was the most commonly identified etiology (86%), and Forrest III lesions were the most common bleeding stigma (46%) in the CAD group. Thirteen (26%) patients in the CAD group received therapeutic hemostasis, and local injection with diluted epinephrine (1:10 000) was the most frequent method used to arrest hemorrhage (12 patients). None of the endoscopic parameters showed significant difference between the 2 groups.

Clinical outcomes

The comparison of clinical outcomes between patients with and without CAD was also shown in Table 2. None of the patients in both groups had angina or MI before, during, or after endoscopy. Successful primary hemostasis was achieved in all patients. However, 2 patients in the CAD group and 3 in the non-CAD group developed recurrent bleeding. These bleeding episodes were successfully managed in 4 of the 5 patients via conservative treatment,

Fig. 1 Mean frequencies of SVE of patients, stratified by group (CAD vs non-CAD) and endoscopy periods. The P values of the before, during, and after endoscopy period for the 2 groups were

.498, .018, and .773, respectively, with the Wilcoxon rank sum test.

Fig. 2 Mean frequencies of VE of patients, stratified by group (CAD vs non-CAD) and endoscopy periods. The P values of the before, during, and after endoscopy period for the 2 groups were .007, .003, and .0001, respectively, with the Wilcoxon rank sum test.

and the other patient with CAD required emergent angiography to arrest the hemorrhage. One patient with CAD and 3-VD developed angina 5 days later and died of ventricular fibrillation. Overall, the length of hospitalization was similar between these 2 groups.

Cardiac arrhythmia and ischemic ST change

Table 3 describes the ECG abnormalities before, during, and after endoscopy, stratified by the presence or absence of CAD. Patients with CAD had more arrhythmic and Ischemic events, especially during the actual endo- scopic procedure.

For Hemodynamic fluctuations, the slopes in the before/ during and during/after endoscopic periods between the CAD and non-CAD groups were evaluated. The fluctuation of SVE frequencies (Fig. 1) was significant in the CAD group, either in the before-and-during period (P = .048) or the during-and-after period (P = .049). However, in the non- CAD group, the fluctuation of SVE frequencies was only significant in the during-and-after period (P = .038) but not in the before-and-during period (P = .048). The fluctuation of VE frequencies (Fig. 2) in the 2 groups was not statistically significant both in the before-and-during and the during-and-after periods.

Ischemic ST changes between patients with and without CAD in the 3 endoscopy periods were also evaluated (Fig. 3). Although more episodes were detected in the CAD group during and after endoscopy (P = .016), the fluctuation was not significant different (P = .18).

Autonomic nervous function

Fig. 4 shows the power spectral analysis of HRV in the CAD and non-CAD group before, during, and after endoscopy. The LF and HF powers decreased significantly during the endoscopy period (P b .01). When the indices were compared between the 2 study groups, there was no significant difference between the Baseline functions and follow-up autonomic nervous functions.

Risk factor assessment

Baseline factors in Table 1 were assessed one-by-one for their ability to predict cardiac events. With the outcome variable of VE frequency during endoscopy, the presence of congestive heart failure was identified as a significant risk factor (regression coefficient, 1.79; 95% confidence interval, 1.03-2.55; P = .02). Using the increment of cardiac indices as outcome variables, the presence of congestive heart failure was still significantly associated with the increase of VE in the before/during endoscopic period (regression coefficient, 1.45; 95% confidence interval, 1.01-1.89; P = .002). No other risk factor showed significant association with the increase of cardiac events of SVE, ischemic ST change, or autonomic nervous derangement.

Discussion

To our knowledge, this is the first prospective study to evaluate the cardiovascular and autonomic responses in

Fig. 3 The proportion of patients with ST ischemia, stratified by the group (CAD and non-CAD) and endoscopy period. The P values of the before, during, and after endoscopy period for the 2 groups were .111, .015, and .015, respectively, with the Fisher exact test.

Fig. 4 Power spectral analysis of HRV, including the mean LF and HF powers of patients, stratified by the groups of CAD and non-CAD and the endoscopy periods.

patients with CAD receiving emergency endoscopy for acute UGIB. Upper gastrointestinal bleeding may affect myocar- dial function and a high Incidence of arrhythmias and myocardial ischemia in patients with CAD undergoing emergency endoscopy was identified in the present study.

Several mechanisms may account for the adverse effect of UGIB on cardiac function. First, acute blood loss may cause hypovolemia, hypoperfusion, and hypoxemia of Major organs [38]. The heart rate may secondarily increase, leading to an excessive workload. In a retrospective study of

113 patients with severe UGIB, 16 (12.3%) patients developed MI [2]. Lee et al [5] evaluated 103 bleeding patients (29 patients had CAD) in the same institute as our study. Similarly, Lee et al [5] found that 5 (4.6%) patients developed MI after emergency endoscopy for UGIB, including 2 patients with CAD. In contrast, none of the patients in this current study developed MI after the endoscopic procedure. This discrepancy may result from the different inclusion criteria used between the 2 studies. Of the 5 patients, 4 (80%) with postendoscopy MI in the study by Lee et al [5] were critically ill patients with refractory shock before the endoscopic procedure. This current study was only interested in the UGIB that physicians encountered in daily practice, whereas patients with shock and who were unresponsive to resuscitation were excluded.

Theoretically, endoscopic hemostasis might be more difficult to achieve in patients with CAD because of antiplatelet agents are commonly prescribed in this popula- tion. These agents also impair gastric mucosa repair and lead to a higher rebleeding rate [28,39]. In the present study, the use of standard hemostatic methods was not less effective in patients with CAD-successful hemostasis was achieved in all patients. Another area of potential concern relates to the

systemic absorption of epinephrine [40]. In this study, 12 patients with CAD received epinephrine injection with an average volume of 9.9 mL (range, 3-20 mL). The incidence of arrhythmia and myocardial ischemia was not different compared to patients treated via the alternate hemostatic techniques.

The higher incidences of subclinical myocardial ischemia and cardiac arrhythmias in our study deserve special attention. In the study by Johnston et al [13], myocardial ischemia occurred in 25% of 41 patients receiving Endoscopic retrograde cholangiopancreatography. Kounis et al [14] also found transient myocardial ischemia, and arrhythmias occurred frequently in elderly patients under- going endoscopic retrograde cholangiopancreatography. In yet another study enrolling elderly patients with a median age of 80 years, 48% developed a new arrhythmia or experienced an exacerbation of an existing arrhythmia during endoscopy [16]. Studying 71 patients with CAD who received Upper endoscopy without sedation or supplemental oxygen, Schenck et al [41] reported that 30 patients (42%) had silent ischemia, but only 1 (1%) was symptomatic. The occurrence of ischemia in the study of Schenck [41] was associated with an increase in heart rate. In the study described herein, 18% of patients with CAD had silent ischemia and as high as 80% of patients with CAD had one episode of SVE or VE. More impressively, up to 20% of patients with CAD had more than one episode of ventricular ectopy per minute (the so-called sinister arrhythmia), which carries a higher risk of cardiac complications in patients with underlying ischemic heart disease [42,43].

The presence of autonomic imbalance is associated with the development of cardiovascular complications after endoscopy [44,45]. Indeed, Hayashi et al [12] used HRV to evaluate autonomic nervous function during endoscopy and reported a significant decrease of HF and LF powers during endoscopy. This same finding was also identified in our study. Furthermore, other studies report the presence of impaired cardiac autonomic regulation and reduced HRV in patients with uncomplicated CAD [46,47].

Several factors might account for the insignificant difference in autonomic function between the CAD and non-CAD group. First, we analyzed a relatively short Holter recording (3-4 hours) in this study, whereas other studies obtained 24 hours of readings. Second, the sample size of both groups (n = 50 each) may not be large enough. Third, some of the patients with CAD in the present study had previous MI or coronary revascularization, both of which could alter HRV [48]. Last, the use of ?-blocker and calcium channel blockers in a number of patients with CAD in the present study might also have influences on HRV [49]. Further research is clearly warranted.

The strength of this study lies in that a comprehensive and detailed recording of ECG abnormalities before, during, and after endoscopic procedures was obtained. Furthermore, we enrolled a control group for comparison, which might reveal the importance of underlying CAD in the precipitation of

cardiac complications. Finally, we had a paired comparison of ECG changes between the before/during endoscopy period, which helps to establish the risk factors that might be valuable in predicting clinical outcomes before perform- ing endoscopic assessment.

There are some limitations in this study. First, this study did not assess the use of sedation. Although sedation may help reduce patient anxiety and discomfort, which subse- quently reduces sympathetic tone and the incidence of arrhythmia, use of sedation during endoscopy increases the risk of pulmonary aspiration and hypotension, which might further impair coronary perfusion in patients with UGIB [15,32]. Certainly this is an area requiring further evaluation. Second, although patients in the control group has declared no CAD history, some of them indeed had the risk factors of CAD, such as advanced age, diabetes, and hypertension. Whether these “non-CAD” patients had occult CAD and thus affect the comparative analysis with the CAD group in aspects of cardiac events is not clear and deserves further studies. Third, the sample size in the current study was relatively small, which might contribute to a statistic type 2 error and prevent us identifying significant differences. Furthermore, whether the common and more frequent subclinical arrhythmia and Ischemic changes in the patients with CAD demonstrated in the present study would have an actual clinical impact is uncertain and awaits further large- scale studies. Finally, the diagnosis of congestive heart failure in the present study relied on clinical History taking and might not reflect the true cardiac function before endoscopy. A more objective tool such as left ventricular ejection fraction measured by echocardiography may be helpful in the preendoscopy evaluation for these high-risk patients with CAD in future practice.

In conclusion, this study demonstrated that emergency endoscopy carries a higher incidence of subclinical arrhyth- mia and ischemic changes in patients with stable CAD and UGIB. For patients with CAD and congestive heart failure, it would be imperative to consider potential risks and benefits before making a decision regarding endoscopic management of UGIB in this high-risk cohort.

References

1. Anthopoulos LP, Stamatelopoulos SF, Sideris DA, et al. Incidence of coronary disease in patients with gastroDuodenal ulcer or upper gastrointestinal tract hemorrhage. Angiology 1972;23:358-64.
2. Bhatti N, Amoateng-Adjepong Y, Qamar A, et al. Myocardial infarction in critically ill patients presenting with gastrointestinal hemorrhage: retrospective analysis of risks and outcomes. Chest 1998; 114:1137-42.
3. Colleran JA, Papademetriou V, Narayan P, et al. electrocardiographic abnormalities suggestive of myocardial ischemia during upper gastrointestinal bleeding. Am J Cardiol 1995;75:312-4.
4. Emenike E, Srivastava S, Amoateng-Adjepong Y, et al. Myocardial infarction complicating gastrointestinal hemorrhage. Mayo Clin Proc 1999;74:235-41.
5. Lee CT, Huang SP, Cheng TY, et al. Factors associated with myocardial infarction after emergency endoscopy for upper gastro- intestinal bleeding in high-risk patients: a prospective observational study. Am J Emerg Med 2007;25:49-52.
6. Cook DJ, Guyatt GH, Salena BJ, et al. Endoscopic therapy for acute nonvariceal Upper gastrointestinal hemorrhage: a meta-analysis. Gastroenterology 1992;102:139-48.
7. Sacks HS, Chalmers TC, Blum AL, et al. Endoscopic hemostasis. An effective therapy for bleeding peptic ulcers. JAMA 1990;264:494-9.
8. Spiegel BM, Vakil NB, Ofman JJ. Endoscopy for acute nonvariceal upper gastrointestinal tract hemorrhage: is sooner better? A systematic review. Arch Intern Med 2001;161:1393-404.
9. Bough EW, Meyers S. Cardiovascular responses to upper gastro- intestinal endoscopy. Am J Gastroenterol 1978;69:655-61.
10. Fisher L, Fisher A, Thomson A. Cardiopulmonary complications of ERCP in older patients. Gastrointest Endosc 2006;63:948-55.
11. Gangi S, Saidi F, Patel K, et al. Cardiovascular complications after GI endoscopy: occurrence and risks in a large hospital system. Gastrointest Endosc 2004;60:679-85.
12. Hayashi T, Nomura M, Honda H, et al. Evaluation of autonomic nervous function during Upper gastrointestinal endoscopy using heart rate variability. J Gastroenterol 2000;35:815-23.
13. Johnston SD, McKenna A, Tham TC. Silent myocardial ischaemia during endoscopic retrograde cholangiopancreatography. Endoscopy 2003;35:1039-42.
14. Kounis NG, Zavras GM, Papadaki PJ, et al. Electrocardiographic changes in elderly patients during endoscopic retrograde cholangio- pancreatography. Can J Gastroenterol 2003;17:539-44.
15. Lieberman DA, Wuerker CK, Katon RM. Cardiopulmonary risk of esophagogastroduodenoscopy. Role of endoscope diameter and systemic sedation. Gastroenterology 1985;88:468-72.
16. McAlpine JK, Martin BJ, Devine BL. Cardiac arrhythmias associated with upper gastrointestinal endoscopy in elderly subjects. Scott Med J 1990;35:102-4.
17. McEwan-Alvarado G, Barnes RN, Wallace TI. Electrocardiographic response to upper gastrointestinal endoscopy. Am J Gastroenterol 1972;57:26-33.
18. Segawa K, Nakazawa S, Yamao K, et al. Cardiac response to upper gastrointestinal endoscopy. Am J Gastroenterol 1989;84:13-6.
19. Seinela L, Reinikainen P, Ahvenainen J. Effect of upper gastro- intestinal endoscopy on cardiopulmonary changes in very old patients. Arch Gerontol Geriatr 2003;37:25-32.
20. Sturges HF, Krone CL. Cardiovascular stress of peroral gastrointestinal endoscopy. Gastrointest Endosc 1973;19:119-22.
21. Yazawa K, Adachi W, Koide N, et al. Changes in cardiopulmonary parameters during upper gastrointestinal endoscopy in patients with heart disease: towards safer endoscopy. Endoscopy 2000;32:287-93.
22. Froehlich F, Schwizer W, Thorens J, et al. Conscious sedation for gastroscopy: patient tolerance and cardiorespiratory parameters. Gastroenterology 1995;108:697-704.
23. Arrowsmith JB, Gerstman BB, Fleischer DE, et al. Results from the American Society for Gastrointestinal Endoscopy/U.S. Food and Drug Administration collaborative study on complication rates and drug use during gastrointestinal endoscopy. Gastrointest Endosc 1991;37: 421-7.
24. Sieg A, Hachmoeller-Eisenbach U, Eisenbach T. Prospective evalua- tion of complications in outpatient GI endoscopy: a survey among German gastroenterologists. Gastrointest Endosc 2001;53:620-7.
25. Froehlich F, Gonvers JJ, Fried M. Conscious sedation, clinically relevant complications and monitoring of endoscopy: results of a nationwide survey in Switzerland. Endoscopy 1994;26:231-4.
26. Levy N, Abinader E. Continuous electrocardiographic Monitoring with Holter electrocardiocorder throughout all stages of gastroscopy. Am J Dig Dis 1977;22:1091-6.
27. Serrano P, Lanas A, Arroyo MT, et al. Risk of upper gastrointestinal bleeding in patients taking low-dose aspirin for the prevention of cardiovascular diseases. Aliment Pharmacol Ther 2002;16:1945-53.
28. Yusuf S, Zhao F, Mehta SR, et al. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494-502.
29. Ng W, Wong WM, Chen WH, et al. Incidence and predictors of upper gastrointestinal bleeding in patients receiving low-dose aspirin for secondary prevention of cardiovascular events in patients with coronary artery disease. World J Gastroenterol 2006;12:2923-7.
30. Wilcox CM, Faibicher M, Wenger NK, et al. Prevalence of Silent myocardial ischemia and arrhythmias in patients with coronary heart disease undergoing gastrointestinal tract endoscopic procedures. Arch Intern Med 1993;153:2325-30.
31. Lee JG, Krucoff MW, Brazer SR. Periprocedural myocardial ischemia in patients with severe symptomatic coronary artery disease under- going endoscopy: prevalence and risk factors. Am J Med 1995;99: 270-5.
32. Lipper B, Simon D, Cerrone F. Pulmonary aspiration during emergency endoscopy in patients with upper gastrointestinal hemor- rhage. Crit Care Med 1991;19:330-3.
33. Forrest JA, Finlayson ND, Shearman DJ. Endoscopy in gastrointest- inal bleeding. Lancet 1974;2:394-7.
34. Lee YC, Wang HP, Lin LY, et al. Heart rate variability in patients with different manifestations of gastroesophageal reflux disease. Auton Neurosci 2004;116:39-45.
35. Lin LY, Lin JL, Du CC, et al. Reversal of deteriorated fractal behavior of heart rate variability by beta-blocker therapy in patients with advanced congestive heart failure. J Cardiovasc Electrophysiol 2001;12:26-32.
36. Welch P. The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Trans Audio Electroacoust 1967;15:70-3.
37. Berger RD, Akselrod S, Gordon D, et al. An efficient algorithm for spectral analysis of heart rate variability. IEEE Trans Biomed Eng 1986;33:900-4.
38. Cappell MS. Gastrointestinal bleeding associated with myocardial infarction. Gastroenterol Clin North Am 2000;29:423-44 vi.
39. Silverstein FE, Feld AD, Gilbert DA. Upper gastrointestinal tract bleeding. Predisposing factors, diagnosis, and therapy. Arch Intern Med 1981;141:322-7.
40. Sung JY, Chung SC, Low JM, et al. Systemic absorption of epinephrine after endoscopic submucosal injection in patients with bleeding peptic ulcers. Gastrointest Endosc 1993;39:20-2.
41. Schenck J, Muller CH, Lubbers H, et al. Does gastroscopy induce myocardial ischemia in patients with coronary heart disease. Endo- scopy 2000;32:373-6.
42. Ruberman W, Weinblatt E, Goldberg JD, et al. Ventricular premature complexes and sudden death after myocardial infarction. Circulation 1981;64:297-305.
43. Follansbee WP, Michelson EL, Morganroth J. Nonsustained ventri- cular tachycardia in ambulatory patients: characteristics and associa- tion with sudden cardiac death. Ann Intern Med 1980;92:741-7.
44. Palmer ED. The abnormal upper gastrointestinal vagovagal reflexes that affect the heart. Am J Gastroenterol 1976;66:513-22.
45. Myers GA, Martin GJ, Magid NM, et al. Power spectral analysis of heart rate variability in sudden cardiac death: comparison to other methods. IEEE Trans Biomed Eng 1986;33:1149-56.
46. Huikuri HV, Niemela MJ, Ojala S, et al. Circadian rhythms of frequency domain measures of heart rate variability in healthy subjects and patients with coronary artery disease. Effects of arousal and Upright posture. Circulation 1994;90:121-6.
47. Wennerblom B, Lurje L, Tygesen H, et al. Patients with uncomplicated coronary artery disease have reduced heart rate variability mainly affecting Vagal tone. Heart 2000;83:290-4.
48. Wennerblom B, Lurje L, Solem J, et al. Reduced heart rate variability in ischemic heart disease is only partially caused by ischemia. An HRV study before and after PTCA. Cardiology 2000;94:146-51.
49. Cook JR, Bigger Jr JT, Kleiger RE, et al. Effect of atenolol and diltiazem on heart period variability in normal persons. J Am Coll Cardiol 1991;17:480-4.