Outcome and survival analysis of intestinal ischaemia following cardiac surgery

Abstract

OBJECTIVES

Intestinal ischaemia is an uncommon (<1%) but serious complication of cardiac surgery with a mortality rate exceeding 50%. Diagnosis of this potentially lethal condition can be difficult and requires a high index of suspicion. The purpose of this study was to analyse the outcomes and prognostic factors in patients who develop intestinal ischaemia following cardiac surgery.

METHODS

In a retrospective review from August 1999 to December 2010, we identified 31 out of 9925 (0.31%) consecutive patients who developed acute intestinal ischaemia following cardiac surgery at our tertiary centre.

RESULTS

The overall mortality was 71.0%. The operative mortality was 65.4% in patients who underwent a laparotomy. Survivors of this complication had surgical intervention earlier (7.4 ± 4.9 h) compared with the non-survivors (13.9 ± 11.1 h). A total of 35 perioperative variables were analysed. A univariate analysis identified 12 variables associated with an increased risk of mortality. Logistic multivariate analysis identified the preoperative logistic EuroSCORE and the base excess at the point of diagnosis of intestinal ischaemia as significant predictors of mortality. These factors may aid prognostication in this group of patients.

CONCLUSIONS

Despite the high mortality rates associated with intestinal ischaemia following cardiac surgery, early diagnosis and surgical intervention remain the only effective means to reduce mortality.

INTRODUCTION

Intestinal ischaemia post-cardiac surgery has a reported incidence of <1% [1–10]. Early diagnosis can be difficult but is important as prompt surgical intervention can be lifesaving [5, 8].

The difficulty in making the diagnosis contributes heavily to the dismal outcome. The signs and symptoms of abdominal pathology are often masked by prolonged mechanical ventilation and sedation especially for critically ill patients after cardiac surgery. A high index of suspicion is required for early diagnosis.

Despite advances in critical care and prompt surgical intervention, the mortality rate of acute intestinal ischaemia remains high, and ranges from 46.1 to 100% [1–5, 7, 8, 11–14].

This study aims to review patients who had undergone cardiac surgery at a tertiary referral centre over an 11-year period, to determine the incidence of intestinal ischaemia following cardiac surgery and investigate their outcomes and possible prognostic factors.

MATERIALS AND METHODS

Following approval from the Singhealth Centralised Institutional Review Board (reference: 2011/835/C), a retrospective case-note and database review was performed on 9925 consecutive patients who had undergone elective and emergency cardiac surgical procedures between August 1999 and December 2010 at our tertiary referral centre. This study focused exclusively on patients with acute mesenteric ischaemia involving the small and/or large bowel. Patients who developed other gastrointestinal complications (e.g. ischaemic hepatitis, gastrointestinal bleeding, pancreatitis) were excluded. Patients who presented with acute aortic dissection with evidence of intestinal ischaemia prior to cardiac surgery were excluded.

Preoperative renal insufficiency was defined by serum creatinine levels higher than 120 mmol/l. Inotropic support was defined as infusions of dopamine over 5 mcg/kg/min or any use of adrenaline, noradrenaline or vasopressin. Preoperative anticoagulation was administered via intravenous heparin infusion. Patients with a postoperatively low cardiac output were identified as those manifesting signs of cardiogenic shock and decreased end-organ perfusion (oliguria, altered mental state, hypoxia, metabolic acidosis) requiring the use of one or more vasopressors (dopamine, adrenaline, noradrenaline or vasopressin) beyond the first 24 h of cardiac surgery. Some of these patients required further support with intra-aortic balloon pump (IABP) or extracorporeal membrane oxygenation (ECMO).

Statistical analysis was performed using the SPSS version 17 statistical software. Continuous variables were expressed as a mean ± standard deviation and were compared using the two-tailed t-test. Categorical variables, expressed as percentages, were analysed with the χ² or Fisher's exact test. To identify the risk factors predictive of mortality, a univariate analysis of 35 perioperative variables was performed. Significant univariate risk factors were examined using a stepwise logistic regression analysis. A two-tailed P-value <0.05 was used to indicate statistical significance.

RESULTS

Of the total 9925 patients who had undergone cardiac surgery between August 1999 and December 2010, acute intestinal ischaemia was identified in 31 patients (0.31%), diagnosed on either laparotomy or postmortem. Of these, 29 patients (93.5%) had undergone conventional surgery with the use of cardiopulmonary bypass (CPB) and the remaining 6.5% had undergone off-pump surgery. Twenty-two patients died, resulting in a mortality rate of 71.0%. All patients had a transmural infarction of either the small (64.5%) or large bowel (19.4%), or both (16.1%).

The incidence of acute intestinal ischaemia was 0.2% for elective cases compared with 1.2% for emergency cases (P < 0.001). Ten patients (32.3%) underwent emergency cardiac surgery, all of which were coronary artery bypass grafting (CABG). There was no significant difference in the incidence of intestinal ischaemia following on-pump compared with off-pump surgery (0.32 vs. 0.21%, P = 0.562). Of the 31 patients with intestinal ischaemia, 21 (67.7%) patients had undergone isolated CABG, 7 (22.6%) had valve surgery and CABG, 2 (6.5%) had isolated valve surgery and 1 (3.2%) had undergone excision of a left atrial myxoma.

Patient demographics and comorbidities are shown in Table 1. There were no significant differences in the demographic characteristics between the two groups. Females had a higher mortality rate compared with males (83.3 vs. 63.2%, P = 0.418). Pre-existing medical conditions were similar and did not significantly affect survival.

Table 1:

Patient demographics and comorbidities

Variable	Survivors, n = 9 (%)	Non-survivors, n = 22 (%)	P-value
Demographics
Age	66.2 ± 8.1	68.3 ± 6.3	0.455
Gender (male/female)	7/2 (77.8/22.2)	12/10 (54.5/45.5)	0.418
LVEF	41.9 ± 16.2	42.3 ± 20.2	0.960
LVEF <30%	3 (33.3)	6 (27.3)	0.528
Comorbidities
Diabetes mellitus	3 (33.3)	11 (50.0)	0.329
Obesity (BMI >25)	5 (55.6)	9 (40.9)	0.363
Renal failure	2 (22.2)	7 (31.8)	0.472
Dialysis	1 (11.1)	1 (4.5)	0.499
Hypertension	8 (88.9)	21 (95.5)	0.503
Stroke	1 (11.1)	2 (9.1)	0.657
Peripheral vascular disease	2 (22.2)	7 (31.8)	0.472
COPD	1 (11.1)	1 (4.5)	0.499
Triple vessel coronary artery disease	7 (77.8)	18 (81.8)	0.577
Previous laparotomy	3 (33.3)	3 (13.6)	0.391

BMI: body mass index; COPD: chronic obstructive pulmonary disease; LVEF: left ventricular ejection fraction.

Factors significantly affecting survival following univariate analysis are shown in Table 2. The following factors did not significantly affect survival: preoperative (emergency surgery, recent myocardial infarction within 90 days, cardiogenic shock, inotropic support, IABP), intraoperative [CPB time, aortic cross-clamp (AXC) time] postoperative (re-exploration for bleeding, myocardial infarction, sepsis, IABP, ECMO, use of vasopressors beyond 24 h of surgery, limb ischaemia, disseminated intravascular coagulopathy, stroke, anticoagulation, abdominal distension, radiographic evidence of pneumatosis or extraluminal air).

Table 2:

Univariate risk analysis for mortality in patients with intestinal ischaemia

Variable	Survivors, n = 9 (%)	Non-survivors, n = 22 (%)	P-value
Preoperative
Emergency surgery	1 (11.1)	9 (40.9)	0.107
Anticoagulation	1 (11.1)	12 (54.5)	0.026
Unstable angina	2 (22.2)	10 (45.5)	0.228
IABP	1 (11.1)	9 (40.9)	0.107
Logistic EuroSCORE	5.7 ± 4.8	26.1 ± 20.6	0.007
Intraoperative
RBC transfusion (units)	0.56 ± 0.53	1.68 ± 1.36	0.023
Postoperative
Renal failure	4 (44.4)	19 (86.4)	0.015
Postoperative CVVH	4 (44.4)	18 (81.8)	0.037
Atrial fibrillation	9 (100.0)	14 (63.6)	0.036
Abdominal pain or tenderness	9 (100)	10 (54.5)	0.014
Vasopressors beyond 24 h of surgery	3 (33.3)	15 (68.2)	0.074
IABP	3 (33.3)	14 (63.6)	0.124
ECMO	0 (0.0)	3 (13.6)	0.244
Low cardiac output	5 (55.6)	20 (90.9)	0.024
Dilated bowel loops	9 (100)	12 (54.5)	0.014
Bowel wall thickening	2 (22.2)	0 (0.0)	0.022
Presence of metabolic acidosis	6 (66.7)	21 (95.5)	0.030
Base excess	−2.9 ± 4.1	−8.8 ± 4.6	0.003

ECMO: extracorporeal membrane oxygenation; EuroSCORE: European system for cardiac operative risk evaluation; CVVH: continuous veno-venous haemofiltration; RBC: red blood cell.

Univariate analysis identified the need for preoperative anticoagulation with intravenous heparin as a risk factor for mortality. These patients requiring anticoagulation also had a higher Mean logistic EuroSCORE compared with those who did not require anticoagulation (33.4 vs. 10.7%, P = 0.001) [15]. The non-survivors required significantly more intraoperative blood transfusion.

A total of 26 (83.9%) of the 31 patients underwent a laparotomy following diagnosis. The remaining five patients were either unfit for a laparotomy or declined surgery. The operative mortality in the laparotomy group was 65.4%. The mortality was 100% for the five patients who did not undergo surgery. They all died within 24 h of the clinical diagnosis, with intestinal ischaemia being confirmed post mortem. The mean time to laparotomy from the time of diagnosis was shorter in the group of survivors (7.4 ± 4.9 h) compared with the non-survivors (13.9 ± 11.1 h). However, this result was not statistically significant (P = 0.127).

Further logistic multivariate analysis identified the preoperative EuroSCORE and the base excess at the point of diagnosis of intestinal ischaemia as the two most important predictors of mortality in this series, as shown in Table 3.

Table 3:

Multivariate risk analysis for mortality in patients with intestinal ischaemia

	Odds ratio	95% confidence interval	P-value
Logistic EuroSCORE >10.5%	20.2	1.31–312.1	0.031
Base excess <−6.0	12.3	1.04–145.1	0.047
Postoperative low cardiac output	3.49	0.23–53.0	0.368

DISCUSSION

The overall incidence of intestinal ischaemia following cardiac surgery in our series was 0.31%, comparable with the 0.36% (31/8709 patients) described by Mangi but differs from the 0.15% (17/11 058 patients), 0.25% (16/6452 patients), 0.62% (30/4819 patients) and 0.73% (39/5349 patients) reported by other authors [1–3, 8, 10].

The overall mortality was 71.0% in this series, similar to the 64.1 and 67.0% reported by Ghosh et al. [8] and Allen et al. [6], respectively. Similarly to our findings, Ghosh et al. [8] reported a 100% mortality rate in patients with intestinal ischaemia who did not undergo a laparotomy. The operative mortality in the laparotomy group was 65.4%, which is comparable with the 63.6% reported by Mangi et al. [3], but lies between the 42.8 and 80.0% reported by Abboud et al. [5] and Chaudhuri et al. [7], respectively. These findings reinforce the fact that acute intestinal ischaemia after cardiac surgery is an infrequent but catastrophic event. High-risk patients with unstable angina or those supported with an IABP preoperatively were identified and anticoagulated with intravenous heparin. Compared with those who did not require preoperative anticoagulation, they had a significantly higher logistic EuroSCORE and also a higher risk of mortality upon developing acute intestinal ischaemia.

Patients who developed new-onset atrial fibrillation postoperatively had a lower risk of mortality from acute bowel ischaemia due to early postoperative anticoagulation and close monitoring. The incidence of other postoperative complications was generally higher in the group of non-survivors compared with the survivors. Notably, acute renal failure occurring concurrently with intestinal ischaemia was shown to be a lethal combination, being observed in most of the non-survivors.

The time interval to laparotomy in our series was mainly limited by the time taken to establish the diagnosis of intestinal ischaemia. The variety of clinical presentations, which was further masked by sedation in mechanically ventilated patients, made the diagnosis of intestinal ischaemia particularly difficult. Abboud et al. [5] reported the time interval to laparotomy in their series to be 13.7 ± 19.0 h. Chaudhuri et al. [7] reported a relatively longer median interval to laparotomy of 8 days for their series of 15 patients who underwent a laparotomy. Compared with non-survivors, survivors of intestinal ischaemia underwent laparotomy earlier (6.4 ± 3.8 vs. 16.9 ± 10 h, P = 0.519) [8]. Similarly to our findings, this difference was not statistically significant.

Apart from non-specific dilated bowel loops, there are often no other signs on plain abdominal X-rays, until a severe intestinal infarction occurs, during which pneumotosis, frank perforation or portal venous gas may develop. Although abdominal pain and radiological features of bowel dilatation or wall thickening were more common in the group of survivors, these findings did not lead to a decreased time interval to laparotomy. Rather, the improved outcomes may have been due to these findings facilitating closer monitoring and prompt diagnosis, allowing surgical intervention at an earlier stage of mesenteric ischaemia.

All laparotomies were performed based on sufficient clinical evidence of acute intestinal ischaemia. To avoid delaying surgery, none of these patients underwent mesenteric angiography or computed tomography. Furthermore, most of them were in a state of haemodynamic instability and thus unfit to be transferred for imaging.

The logistic multivariate analysis identified the preoperative EuroSCORE and the base excess at the point of diagnosis of intestinal ischaemia as the two most important predictors of mortality in this series. The preoperative logistic EuroSCORE was used to risk-stratify patients prior to the initial cardiac surgery. This risk stratification model was found to be applicable in our study population who developed intestinal ischaemia after cardiac surgery. Patients with a higher logistic EuroSCORE had an increased risk of mortality from intestinal ischaemia. The severity of metabolic acidosis at the time of the diagnosis of intestinal ischaemia, as measured with base excess values, was found to correlate with the extent of the bowel infarction. Base excess values may also serve as a surrogate marker to reflect the level of critical illness in this group of patients. Severe metabolic acidosis was observed in cases where an extensive bowel infarction had occurred, in which the prognosis was invariably dismal. Persistent metabolic acidosis refractory to bicarbonate infusions occurred in 37.0% of patients without associated abdominal signs or abnormal radiological findings. Thus, intestinal ischaemia must be excluded in patients with unexplained and persistent metabolic acidosis after cardiac surgery, as it may be the earliest and only diagnostic clue of mesenteric ischaemia.

Risk factors for the development of intestinal ischaemia following cardiac surgery have been well described. Some of these factors include advanced age, prolonged CPB or AXC time, IABP usage, significant inotropic support, peripheral vascular disease, emergency cardiac surgery and postoperative renal failure [2, 7, 8].

In contrast, there are few reports of predictors of mortality from gastrointestinal complications following cardiac surgery. Zacharias et al. [4] analysed a group of patients with gastrointestinal complications following cardiac surgery who underwent abdominal surgery. In this group of 37 patients, 22 patients suffered from intestinal ischaemia, 12 patients had acute cholecystitis and three patients developed perforated duodenal ulcers. The operative mortality was 68.2% for those with intestinal ischaemia. The authors reported that preoperative renal failure and prolonged CPB or AXC time were factors associated with the increased mortality in this surgically treated group. However, these factors were not statistically significant in our study.

This is a retrospective descriptive study with inherent biases in data collection. Owing to the relatively rare occurrence of intestinal ischaemia, the small sample size underpowered the statistical analysis and could have limited the number of statistically significant variables. A prospective study incorporating a larger sample size would be useful to assess the prognostic value of the risk factors identified.

CONCLUSION

The incidence of acute intestinal ischaemia following cardiac surgery is 0.31% and is associated with a mortality rate of 71.0%. It is a diagnosis that must be excluded in all patients who develop abdominal signs or persistent metabolic acidosis following cardiac surgery. Although the clinical and radiological signs are non-specific, their presence may facilitate an earlier diagnosis and surgical intervention. Severe metabolic acidosis and a high preoperative logistic EuroSCORE are factors associated with increased mortality. Although mortality rates remain high despite early surgical intervention, a low threshold for a prompt laparotomy remains the most reliable means for a potential reduction in the mortality rate.