Incidence of GI complications
The incidence of GI complications following cardiac surgery in our study was 2.16%. Previous studies have reported rates of GI complications post-cardiac surgery, between 0.41–3.7% [3, 4, 5, 6, 7, 8]. In an Australian cohort, Saxena et al reported an incidence of GI complications following aortic valve replacement of 1.3% in patients aged < 80 years and 3% in patients aged 80–89 years. Furthermore, Viana et al reported an incidence of GI complications of 1.1% following cardiac surgery conducted between 2001–2011 [8]. The higher incidence of GI complications in those aged 18 years and over in our study may be due to a higher prevalence of co-morbidities and increased complexity in patients receiving cardiac surgery.
GI bleeding
In our study, the most common GI complication was GI bleeding, with lower GI bleeding occurring most frequently, and no patients experiencing both upper and lower GI bleeding. These results confirm data from previous studies with similar numbers of patients and incidence of GI complications which have shown GI bleeding to be the most common GI complication observed following cardiac surgery [9, 10, 11, 12]. However, these results differ from Vianna et al who observed 3 of 21 patients to have both upper and lower GI bleeding [8]. Gastro-duodenal ulcer and erosive gastritis are the two most common aetiologies of upper GI bleeding [13]. At our hospital, 40 milligram pantoprazole once daily is used as routine care post-operatively in all cardiac surgical patients. In the absence of the routine pantoprazole, the incidence if GI bleeding complication may be higher, explaining the conflicting data from prior studies.
Mesenteric ischemia
Mesenteric ischemia occurred in just 16% of patients with a GI complication in our cohort (equivalent to 0.03% of the total cardiac surgical population) but had a higher mortality (n = 7/16; 43.7%) and greater operative intervention (n = 15/16; 94%) as compared to other GI complications. In other studies, the incidence of acute mesenteric ischemia ranges from 0.06–0.2% of all cardiac surgical patients and is similarly associated with significant mortality 46–100% [14–16]. In a systematic review of 3692 patients Schoots et al reported the overall mortality in non-surgically treated patients to be almost 95% [15]. While the mechanism for increased mortality following mesenteric ischemia in a cardiac surgery population is poorly understood, it has been postulated that the presence of coronary artery disease is associated with vasculopathy in the mesenteric bed [16, 17], which predisposes the patient to ischemia in a peri-operative period. In addition, the signs and symptoms for mesenteric ischemia are non-specific or may be obscured by sedation or analgesia, which may increase the incidence of mesenteric ischemia following cardiac surgery when compared to a non-surgical population.
Acute cholecystitis
In our study, 4% of patients with a GI complication exhibited acute cholecystitis. This is similar to previous reports in which acute cholecystitis accounts for 3–8% of GI complications following cardiac surgery [13, 18, 19]. The precise pathophysiology of the cholecystitis is not known, although several mechanisms have been suggested which include visceral hypoperfusion of the gallbladder, increased viscosity of bile because of the stasis, endotoxemia and overproduction of inflammatory mediators [20]. Furthermore, diagnosis may be challenging as common symptoms of fever and right upper quadrant pain are highly non-specific following cardiac surgery.
Paralytic ileus and perforation
Nearly one quarter of patients with a GI complication in our study had a paralytic ileus, of which three died within 30 days. Paralytic ileus is a rare, poorly understood surgical complication with multifactorial origins [21]. Paralytic ileus is associated with the disturbance of the autonomic innervation of the colon [22]. In our series, we had five cases of perforation and only one death. Perforated duodenal or gastric ulcer post-cardiac surgery accounts for approximately 6% of all GI complications with a reported mortality of around 38% [23]. Existing data would suggest the use of proton pump inhibitors for prophylaxis of gastric or duodenal stress ulcers for high-risk patients may be appropriate [23].
Hepatitis
In our series, 22% of patients with a GI complication had hepatitis, of which three patients died within 30 days. The incidence of liver failure post cardiac surgery in the international literature is 2.2% of all GI complications and 0.026% of all cardiac operations [24]. Liver injury following cardiopulmonary bypass may be due to ischemia as a result of arterial hypotension and venous congestion [25]. Attention to hepatic perfusion pressure and avoidance of hepatotoxic drugs may mitigate the development of liver injury post cardiac surgery; further data regarding liver protective therapies such as N-acetyl cysteine are needed.
Pre-operative risk factor for GI complication
In our study, advanced age, higher illness severity, diabetes, renal dysfunction, and arrythmias (mostly atrial fibrillation) were associated with the presence of a GI complication. Many studies have identified various risk factors trying to elucidate the correlation with GI complication. A few authors have used multivariate analysis, but with little concordance except for age, renal dysfunction, low EF, prolonged ventilation, and NYHA functional class [26, 27]. Aithoussa et al has shown that smoking, diabetes, obesity, hypertension, and hyperlipidaemia occurred more frequently in patients with a GI complication and patients with a GI complication were more likely to have a history of previous gastric ulcer, peripheral vascular disease (PVD), pre-operative renal impairment, and anaemia. Our study confirms these data.
Intra-operative risk factor for GI complications
In our study, patients with a GI complication were more likely to have a cardiopulmonary bypass time of greater than 120 min than in those with no GI complications 30% of patients with a GI complication had CPB > 120 min as compared to the 15% in cohort with no GI complication. This is in accordance with previous data [11, 13, 27, 28]. McSweeny et al have shown that the presence of hypotension and duration of the CPB have been implicated as the strongest predictors of an adverse GI outcome (ref). In this study, 7% of patients with GI complications had a CPB > 100 min as compared to 4% of patients with no GI complications.
Post-operative risk factors for GI complications
Post-operative factors such as need for inotropic agents, requirement of ventilation for more than 24 hours, and stroke were more common in patients with GI complications. This corresponds well with previous reports [1, 10–12]. Filsoufi et al has shown that age, MI, haemodynamic stability, CPB time > 120 min, PVD, renal and hepatic failure as the independent predictors of the GI complications. Aithoussa et al suggested that severe infection is a risk factor for GI complication in univariate analysis but not multivariate analysis [29]. Septicaemia, pneumonia and multiorgan failure were also more prevalent in patients with GI complications in our study.
Clinical outcomes
Our study also revealed that in terms of resource utilisation perspective, patients with a GI complication had a longer stay in ICU and hospital as compared to those without a GI complication. McSweeny et al showed that the adverse GI outcome more than doubled the mean post-surgical stay to 3.5 week and also doubled the median hospital length of stay [28]. In our study, the median length of the ICU stay was three times longer in the GI complication group when compared to those with no GI complications. Similarly, the median length of hospital stay is doubled in the GI complication group as compared to the non-GI complication group.
The strengths of this study are that data from a large number of patients collected in the recent past were analysed. The patient population is likely to be representative of the cardiac surgery population in most Australian and western cardiac surgical centres given the breath of cardiac surgery performed at this site. This study also collected data on GI complications using an established, pre-defined set of criteria ensuring consistency. This study was a retrospective review of prospectively collected data which means any inferences are associations and not causations. In addition, while data for patients with a GI complication were cross-checked against patient case notes following data extraction, data for patients initially identified as not having a GI complication were not cross-checked which may have led to the underestimation of the percentage of GI complications in the entire cohort.