04219

2013

PENXXX10.1177/0148607113504219Journal of Parenteral and Enteral NutritionFlordelís Lasierra et al

Original Communication

Early Enteral Nutrition in Patients With Hemodynamic Failure Following Cardiac Surgery José Luis Flordelís Lasierra, MD1; José Luis Pérez-Vela, MD1; Luis Daniel Umezawa Makikado, MD1; Enrique Torres Sánchez, MD1; Lara Colino Gómez, MD1; Borja Maroto Rodríguez, MD1; Primitivo Arribas López, MD1; Agustín Gómez de la Cámara, MD, PhD2; and Juan Carlos Montejo González, MD, PhD1

Journal of Parenteral and Enteral Nutrition Volume 39 Number 2 February 2015 154­–162 © 2013 American Society for Parenteral and Enteral Nutrition DOI: 10.1177/0148607113504219 jpen.sagepub.com hosted at online.sagepub.com

Abstract Background: Enteral nutrition (EN) is controversial in patients with circulatory compromise. This study assesses the feasibility and safety of EN given early after cardiac surgery in patients with hemodynamic failure. Methods: Prospective observational study conducted in a surgical intensive care unit (ICU) of a tertiary hospital over 17 months. Inclusion Criteria: Cardiac surgery patients with hemodynamic failure (dependence on 2 or more vasoactive drugs and/or mechanical circulatory support) requiring more than 24 hours of mechanical ventilation. Variables Examined: Descriptive data, daily hemodynamic data, and variables related to the efficacy and safety of EN. EN was performed according to our EN protocol. Results: Of 642 patients admitted to the ICU, 37 (5.8%) met the inclusion criteria. Of these, 11 (29.7%) required mechanical circulatory support, and 25 (68.0%) met the criteria for early multiorgan dysfunction. Mortality was 13.5%. Mean EN duration was 12.3 days (95% confidence interval [CI], 9.6–15.0). The mean EN diet volume delivered/patient/d was 1199 mL (95% CI, 1118.7–1278.8), and mean EN energy delivered/patient/d was 1228.4 kcal (95% CI, 1145.8–1311). The set energy target was achieved in 15 patients (40.4%). The most common EN-related complication was constipation. No case of mesenteric ischemia was detected. Conclusions: Our findings indicate that early EN is feasible in this type of patients and not associated with serious complications. However, it is difficult to attain an appropriate energy target by EN alone. These observations point to a need for monitoring of daily energy delivery and balance, as well as careful monitoring of warning signs of intestinal ischemia. (JPEN J Parenter Enteral Nutr. 2015;39:154-162)

Keywords enteral nutrition; critical care; cardiac disease

Clinical Relevancy Statement In patients with severe circulatory compromise, enteral nutrition (EN) is usually considered hazardous as it may aggravate gut ischemia. However, the available evidence supporting this hypothesis is scarce and contradictory. We describe our experience with early EN in 37 critically ill post–cardiac surgery patients with hemodynamic failure. Under proper medical supervision, early EN emerged as feasible and safe. An intervention clinical trial is needed to confirm these findings.

Introduction Most patients undergoing cardiac surgery do not require nutrition therapy since they are able to resume oral feeding within 2–3 days. However, in some of these patients, the postoperative course is more complicated, and prolonged mechanical ventilation, as well as vasoactive drugs and/or mechanical support, is needed as a consequence of hemodynamic failure. These patients are often hypercatabolic and are unable to ingest

food for 5–7 days, thus increasing the risk of malnutrition. In critically ill patients, this situation has been linked to several possible complications, prolonging hospital stay and elevating healthcare costs.1-4 In critically ill mechanically ventilated patients, particularly the sickest patients, early enteral nutrition (EN) has been correlated with reductions in intensive care unit (ICU) and From the 1Intensive Care Medicine Service and 2Clinical Research Support Unit, Hospital Universitario 12 de Octubre, Madrid, Spain. Financial disclosure: None declared. Received for publication February 26, 2013; accepted for publication August 15, 2013. This article originally appeared online on October 4, 2013. Corresponding Author: José Luis Flordelís Lasierra, MD, Intensive Care Medicine Service, Hospital Universitario 12 de Octubre, Paseo de Las Acacias 30, Block 4, Floor 8th-A, 28005, Madrid, Spain. Email: [email protected]

Flordelís Lasierra et al hospital mortality,5 hospital stay, and infectious complications.6,7 Nevertheless, EN is often contraindicated and is even considered dangerous in patients with circulatory failure. Hemodynamic instability affects the splanchnic circulation, reducing splanchnic flow relative to cardiac output, and this, in turn, increases the risk of splanchnic ischemia.8 Reported rates of acute mesenteric ischemia following cardiac surgery are 0.5%−1.4%, and mortality attributed to this complication is in the range of 11%−27%.9,10 Current guidelines issued by the American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.) and European Society for Clinical Nutrition and Metabolism (ESPEN)11,12 state that EN should be withheld in patients with circulatory compromise until they are fully resuscitated and/or hemodynamically stable. These recommendations are based on the results of nonrandomized trials, studies using retrospective control data, the findings of several uncontrolled studies examining case series, and expert opinions. However, several experimental animal models13-17 have indicated that EN may be possible despite hemodynamic failure and may even improve splanchnic perfusion and reduce bacterial translocation. Moreover, following cardiac surgery, the gut appears to remain functional even when there is circulatory failure.18 In addition, the results of 3 studies have indicated that EN may be safe in these patients under close clinical monitoring.19-21 More recently, a study conducted in patients with severe respiratory and hemodynamic failure requiring extracorporeal membrane oxygenation (ECMO) has highlighted the feasibility of EN in these patients.22 Following cardiac surgery, patients in the ICU of our center routinely receive EN therapy. However, despite growing evidence in favor of early EN in these patients, to date, few prospective studies have addressed this issue, and the decision to initiate early EN in patients with circulatory compromise after cardiac surgery remains controversial.23,24 The present study was designed to examine the feasibility and efficacy of EN support and its associated complications in cardiac surgery patients with hemodynamic failure.

Methods In a prospective observational study performed at our post– cardiac surgery ICU, we recruited cardiac surgery patients with hemodynamic failure subjected to at least 24 hours of mechanical ventilation. Hemodynamic failure was defined as dependence on 2 or more vasoactive drugs and/or mechanical support such as intra-aortic balloon pump (IABP), mechanical circulatory assistance, or venoarterial ECMO. Hemodynamic failure was diagnosed upon ICU admission and persisted in the patients recruited for at least 48 hours after admission. Exclusion criteria were a history of gastrointestinal (GI) ischemia or aortic dissection involving the mesenteric vessels.

155

Hemodynamic Monitoring In the first few ICU days, patients underwent advanced hemodynamic monitoring of arterial pressure (MAP), central venous pressure (CVP), and cardiac index (CI, using a Swan-Ganz catheter). Using these data, we determined the lowest daily MAP/CI and the highest daily CVP, eliminating erroneous measurements. After invasive monitoring, MAP was noninvasively measured every hour. As for the invasive recordings, the lowest daily MAP was recorded once erroneous measurements were eliminated. Vasoactive and inotropic drug doses were calculated every hour in µg/kg/min. From these data, the highest daily dose was recorded for each patient. Blood lactate was determined every 8 hours and the daily peak recorded for each patient.

Nutrition Support Nutrition support was given in the form of EN according to the protocol established at our ICU.25 Based on the results of metabolic studies,19,26 the energy target is 25 kcal/kg/d. For patients with a body mass index 30 kg/m2, this was calculated as (25 kcal/kg of ideal weight) + 30%.23 Enteral access was nasogastric. Feeds were adjusted to gradually achieve 100% of the energy target in daily 25% steps over the first 4 days. No prokinetic agents were routinely used. The enteral formulas prescribed by the attending physician were Jevity (Abbott Nutrition, North Chicago, IL), Impact (Nestlé SA, Vevey, Switzerland), Isosource Protein (Nestlé), Novasource GI control (Nestlé), or Novasource Diabet Plus (Nestlé). EN was started after checking the patient’s GI tolerance of fluids (a gastric residual volume [GRV] 500 mL was obtained in each assessment.25 To reduce the risk of pulmonary aspiration and/or ventilation-associated pneumonia, the patient’s head was elevated >30° from the horizontal during enteral feeding. Abdominal distention was defined as a change in the abdomen detected in a physical examination and was usually an increase in abdominal cavity size relative to that recorded in the pre-EN examination. Regurgitation was defined as the presence of EN feed in the oral cavity or oropharynx, as well as its spontaneous drainage by the oral and/or nasal route. EN-related diarrhea was defined as 5 or more liquid stools in 24 hours or more than two 1000-mL stool volumes, each deposited over a 24-hour period. Bronchoaspiration was defined as the presence of respiratory secretions of similar characteristics to the prescribed EN feed, confirmed by the glucose-oxidase technique in tracheal secretion. Constipation was defined as a lack of bowel movements in 7 days from the onset of EN or for 3 days in the first week of admission.

156

Journal of Parenteral and Enteral Nutrition 39(2)

Table 1.  Variables Considered in This Study. Variable General patient characteristics Surgery related Hemodynamics

Other

EN efficacy related

EN complications

Description Demographic data, weight in kilograms, height in meters, EUROSCORE,27 and SAPS II28 Intervention performed, cardiopulmonary bypass time, and reintervention Blood lactate (daily peak), cardiac index (daily lowest), dose of vasoactive drugs (highest daily), mechanical support, mean arterial pressure (lowest daily), and central venous pressure (highest daily) Preoperative serum albumin, multiple transfusions (transfusion ≥10 red blood cell units in 24 hours), maximum daily SOFA,29 perioperative acute myocardial infarction,30 acute kidney failure, continuous renal replacement therapy, infectious complications,a days of mechanical ventilation, ICU stay, and mortality EN diet, days of EN, volume delivered, kilocalories (kcal) delivered by enteral/nonenteral route,b energy balance (kcal delivered by EN—nutrition target in kcal), nutrition tolerance (kcal delivered by EN/nutrition target in kcal), and daily GRV. Contribution of the oral route was not assessed. Mesenteric ischemia, high GRV, abdominal distention, diarrhea, vomiting/regurgitation, bronchoaspiration, nasogastric tube complications (obstruction or misplacement/accidental extubation), and need to interrupt or discontinue EN (and reasons)

EN, enteral nutrition; EUROSCORE, European System for Cardiac Operative Risk Evaluation27; GRV, gastric residual volume; ICU, intensive care unit; SAPS II, Simplified Acute Physiology Score II28; SOFA, Sequential Organ Failure Assessment.29 a Infectious complications were monitored daily and defined according to definitions from the National Study of Nosocomial Infection in ICU and Hospital in Europe Link for Infection Control through Surveillance (ENVIN-HELICS).31 b kcal delivered by the nonenteral route included parenteral nutrition and nonnutrition energy delivered to the patient as intravenous infusions and fluids used to solubilize drugs (glucose, glucosaline, etc), including the lipid vehicle of the sedative propofol (Diprivan 1%; AstraZeneca, Zug, Switzerland).

All possible complications were assessed daily by the attending physician and recorded in the data collection forms.

Patient Follow-up Patients were followed from the start to day 28 of EN. The variables examined are described in Table 1. According to our institutional review board guidelines, informed consent is not required for this type of study. The database used was maintained by the principal researcher (J.L.F.L.). All data were confidential and encoded so that only the research team members were aware of patient identities.

Statistical Analysis Data are provided as means and their corresponding 95% confidence intervals (CIs) for continuous variables and as absolute and relative frequencies for categorical variables. Mean energy delivered, nutrition tolerance, and blood lactate levels during the course of EN are graphically represented for the entire study sample. Qualitative variables in contingency tables were compared using the χ2 or Fisher exact tests. The Wilcoxon-Mann-Whitney test or Student t test, as appropriate, was used to compare ordinal and continuous measurement distributions. To account for subject effects on repeat measurements that could lead to bias (eg, blood lactate, cardiac index, or Sequential Organ Failure Assessment [SOFA] score), a mixed-effects regression model was constructed using 2-level data for observations repeated over time (level 1) nested within patients (level 2). We estimated the mean and 95% confidence interval of the dependent variable (eg, SOFA or blood lactate) measured over time, using the model with or without correction for patient characteristics (ICU stay, need for IABP/

mechanic support, and complications). Because of the limited sample size, a multivariate analysis to identify variables independently associated with EN efficacy and safety was not considered appropriate. All statistical tests were performed using SAS software (SAS Institute, Inc, Cary, NC). Our hospital’s Unit of Research Support was responsible for the statistical treatment of data.

Results Patient Characteristics In total, 642 consecutive post–cardiac surgery patients recruited over a period of 17 months were examined (for characteristics, see Table 2). Of these, 37 (5.8%) met the inclusion criteria (Figure 1). Participants showed high severity scores (means of 37 and 8 in the Simplified Acute Physiology Score II [SAPS II] and SOFA, respectively), a high incidence of early multiorgan dysfunction (failure of 2 or more organ systems in the first 48 hours), and prolonged ICU stay (mean 25 days; in 81%, ICU stay was >7 days). Valve replacement surgery was the most common surgical procedure.

Hemodynamic Status As indicated in the Methods section, all patients had hemodynamic failure during the first 48 hours of the ICU stay (see Table 3), which is when EN was started in most patients (n = 36, 97.3%). Moreover, in this period, a significant proportion of the patients was severely hemodynamically unstable (3 vasoactive drugs were required in 38% of the patients, 4 drugs in 24%, and 4 vasoactive drugs plus mechanical assistance in 16%).

Flordelís Lasierra et al

157

Table 2.  Data Recorded for the Variables Considered (n = 37). Variable a

Age, y Male sexb Preoperative serum albumin, g/dLa SAPS IIa SOFAa EUROSCOREa CBP time, mina CABGb Valve replacement surgeryb Ascending aortic surgeryb Endocarditis surgeryb Heart transplantb Other surgeriesb,c Reinterventionb,d Multiple transfusionsb Perioperative AMI Need for mechanical circulatory supportb Early MODSb Acute renal failure without CRRTb Acute renal failure with CRRTb Ventilator-associated pneumoniab Bloodstream infectionsb Surgical site infectionsb Urinary tract infectionsb Days of mechanical ventilationa ICU stay, da “Short stay” patientsb “Long stay” patientsb ICU mortalityb

Value 56.8 (51.3–62.3) 18 (49.0) 3.9 (3.7–4.1) 37.4 (34.3–40.4) 7.7 (6.9–8.6) 8.8 (7.8–9.9) 126.3 (103.5–149.1) 5 (13.5) 13 (35.1) 4 (10.8) 1 (2.7) 4 (10.8) 8 (21.6) 14 (37.8) 2 (5.4) 8 (21.6) 11 (29.7) 25 (67.6) 8 (21.6) 7 (18.9) 6 (16.2) 5 (13.2) 1 (2.7) 2 (5.4) 19.6 (13.2–25.9) 25.5 (18.3–32.6) 7 (18.9) 30 (81.1) 5 (13.5)

AMI, acute myocardial infarction; CABG, coronary artery bypass graft; CPB, cardiopulmonary bypass; CRRT, continuous renal replacement therapy; EUROSCORE, European System for Cardiac Operative Risk Evaluation; ICU, intensive care unit; MODS, multiple organ dysfunction syndrome; SAPS II, Simplified Acute Physiology Score; SOFA, Sequential Organ Failure Assessment. Mechanical circulatory support included intra-aortic balloon pump, ventricular assist device, or extracorporeal membrane oxygenation (ECMO). “Short stay” refers to patients whose ICU stay was ≤7 days. “Long stay” refers to patients spending > 7 days in the ICU. a Data expressed as mean (95% confidence interval). b Data expressed as number (%) of patients. c “Other surgeries” refers to other cardiac surgical procedures: pericardial window (n = 1), pulmonary endarterectomy (n = 1), surgery for mechanical complications of acute myocardial infarction (n = 1), or ECMO/ ventricular assist device placement upon ICU admission for cardiogenic shock (n = 6). d These patients were enrolled in the study at the time of the initial surgery.

Efficacy of EN Support EN was the only form of nutrition support given in 33 (89%) patients (see Table 4 and Figure 2). In 4 patients, supplementary parenteral nutrition (PN) was needed to achieve caloric goals. EN was initiated on the second day after ICU admission in 36

Figure 1.  Patient recruitment. ICU, intensive care unit.

Table 3.  Hemodynamic Data Recorded in the First 48 Hours Spent in the ICU. Variable Lactate, mg/dL Cardiac index, L/min/m2 MAP, mm Hg CVP, mm Hg NE dose, µg/kg/min DA dose, µg/kg/min DBT dose, µg/kg/min Ep dose, µg/kg/min

Mean (95% CI) 4 (3.2–4.7) 2.3 (2–2.5) 64 (60–68) 14 (13–16) 0.32 (0.23–0.41) 6.35 (4.71–7.99) 7.19 (5.33–9.05) 0.11 (0.04–0.18)

CI, confidence interval; CVP, central venous pressure; DA, dopamine; DBT, dobutamine; Ep, epinephrine; MAP, mean arterial pressure; NE, norepinephrine.

(97.3%) patients. In the remaining patient, EN was delayed because of a high GRV. The most frequently prescribed enteral formula was Impact (210 EN days, 28 patients), followed by Novasource GI control (172 EN days, 19 patients), Jevity (43 EN days, 33 patients), Novasource Diabet Plus (29 EN days, 5 patients), and Isosource Protein (19 EN days, 3 patients).

Safety of EN Support EN-related complications were observed in 23 patients (62%), although no serious GI complications or clinical signs of mesenteric ischemia were recorded (see Table 5). The most common complication was constipation (46%, n = 17). In 1 patient with GI bleeding, a diagnosis of ischemic colitis attributed to prior vascular disease was made upon colonoscopy.

Patient Groups of Special Interest Short (≤7 days) and long (>7 days) ICU stays for patients (see Table 6 and Figure 2c). In the short-stay group, the most

158

Journal of Parenteral and Enteral Nutrition 39(2)

Table 4.  Efficacy of EN. Variable

Value a

Days of EN Diet volume delivered/patient/d, mLa kcal delivered/patient/da Patients achieving the energy targetb Daily energy balance, kcala,c Cumulated energy balance (days 1–4), kcala Cumulated energy balance (days 4–7), kcala Cumulated energy balance (days 8–14), kcala Cumulated energy balance (days 15–28), kcala Nutrition tolerance, %

12.3 (9.6 to 15.0) 1198.8 (1118.7 to 1278.8) 1228.4 (1145.8 to 1311) 15 (40.4) −196.4 (−300.3 to −92.6) 354 (35.5 to 672.6) −883.6 (−1268.1 to −499.1) −1917.9 (−2854.5 to −981.3) −3508.5 (−5273.5 to −1743.4) 92 (84–100)

Nutrition tolerance (%) = kcal delivered by enteral nutrition (EN)/nutrition target in kcal. Energy balance and nutrition tolerance represent the means and 95% confidence intervals recorded over the course of the study. a Data expressed as means (95% confidence intervals). b Data expressed as number (%) of patients. c Nonnutrition calories were not included in the calculation of energy balances. The mean additional delivery by this route was 113.9 (95% confidence interval, 88.8–139) kcal.

frequently performed surgery was valve replacement (n = 3, 42.9%), followed by coronary artery bypass graft (n = 2, 28.6%). In contrast, patients in the long-stay group were those with more serious conditions undergoing more complex surgical procedures such as heart transplant (n = 4, 13.3%) or ECMO/ventricular assist device placement on ICU admission due to cardiogenic shock (n = 6, 20%). No significant difference in illness severity scores or ICU mortality was observed between the 2 groups. However, the long-stay patients showed a significantly higher rate of complications attributable to EN. None of these complications were serious, and most did not require EN discontinuation or interruption. In addition, a greater energy deficit and a lower nutrition tolerance were observed in this group of patients. Patients requiring IABP/ECMO or a ventricular assist device.  No significant differences in the calories administered by the enteral route, energy balance, or nutrition tolerance were observed between patients requiring or not requiring mechanical assistance. Those requiring mechanical assistance showed a higher rate of EN-related complications (82%). These minor complications were managed according to our EN protocol. Patients with EN-related complications.  No significant differences in illness severity scores, mortality, or hemodynamic variables were detected in patients with or without EN-related complications.

Discussion The findings of our study indicate that early EN is feasible in postoperative cardiac surgery patients with hemodynamic failure. We observed that early EN was effective and safe, even in the more severely ill and unstable patients included in our

study (those requiring ECMO, IABP, or mechanical assistance). In our experience, the patients in our cardiac surgery ICU normally show good EN tolerance. The main problem generally encountered is an inability to achieve the established nutrition target. This target was attained in only 40% of our study patients, leading to a high cumulative energy deficit. Berger et al21 also observed that EN support usually leads to hypocaloric feeding in post–cardiac surgery patients. The same conclusion was reached in 2 recent reviews.23,24 The causes of underfeeding in patients receiving EN are multiple. Our cautious approach to attaining the preset energy target was, in our opinion, the main cause of the undernutrition observed in our patients. The practice of supplementary PN in critically ill patients is controversial,21,23,24,32,33 and we have observed its limited use in our cardiac surgery ICU. In effect, volume intake restriction in cardiac surgery patients showing hemodynamic instability is common practice. When we compared our shortstay (≤7 days) and long-stay (>7 days) ICU patients (Table 6), it was noted that although nutrition tolerance in the short-stay group was not a problem, the energy target was not generally reached. This can be explained by our cautious approach to EN diet adjustment and attempts at oral feeding. Such attempts are frequently an inefficient source of calories.34 In the long-stay ICU group, a higher rate of complications led to a significant reduction in energy delivery and broadly negative cumulated energy balances. Maybe this subgroup of patients might especially benefit from supplementary PN. According to the EN safety-related variables assessed (Table 5), no patient experienced serious GI complications. There was a single case of GI bleeding due to ischemic colitis detected and confirmed by colonoscopy. Constipation was the most frequent complication and perhaps indicates a need for preventive treatment in these patients. In the short-stay ICU patients, this complication was less common (Table 6). In

Flordelís Lasierra et al

159

Figure 2.  Efficacy and safety of enteral nutrition (EN): (A) energy delivery, energy balance, and energy target expressed as means and the corresponding 95% confidence intervals, (B) blood lactate levels expressed as means and the corresponding 95% confidence intervals, and nutrition tolerance expressed as percentage. (C) Efficacy of EN in short-stay (≤7 days) and long-stay (>7 days) patients in the intensive care unit (ICU). Energy delivery, energy balance, and energy target expressed as means and the corresponding 95% confidence intervals.

160

Journal of Parenteral and Enteral Nutrition 39(2)

Table 5.  Safety of Enteral Nutrition. Variable High GRV Abdominal distention Diarrhea Constipation NGT complications Vomiting/regurgitation Bronchoaspiration Gastrointestinal bleeding Patients with complications Patients with complications requiring   EN discontinuation

No. (%) of Patients 3 (8.1) 8 (21.6) 10 (27.0) 17 (46.0) 2 (5.4) 1 (2.7) 1 (2.7) 1 (2.7) 23 (62.0) 9 (24.3)

EN, enteral nutrition; GRV, gastric residual volume; NGT, nasogastric tube.

general, the complications detected were mild and could be resolved by interrupting EN and/or reducing the infusion rate. At this point, we should stress that clinical, analytical, and radiological follow-up of these patients during early EN is critical. Nutrition support must be performed according to a previously established protocol. Moreover, careful and repeated monitoring of warning signs of intestinal ischemia such as a high GRV, abdominal distention, intra-abdominal pressure >15 mm Hg, oliguria, leukocytosis, hemodynamic instability with lactic acidosis, or a radiographic finding of dilated, thickened bowel loops is required.23,35 According to U.S. and European guidelines,11,12 early EN in critically ill postoperative cardiac surgery patients is controversial. The pathophysiology of splanchnic circulation is complex in conditions of both hemodynamic stability and instability. The presence of nutrients in the gut lumen is important for normal gut structure and function. In hemodynamically stable individuals, this increases splanchnic blood flow known as “postprandial hyperemia.”36 Shock induces intense visceral vasoconstriction, decreasing oxygen delivery and increasing tissue oxygen consumption, eventually leading to mesenteric ischemia.37 However, effects on the splanchnic circulation are complex and vary with the etiology of shock. Furthermore, the vasoactive and inotropic support used in these patients affects the GI tract and splanchnic microcirculation.38-40 There is evidence that lowered splanchnic blood flow in response to cardiogenic shock may persist even after hemodynamic variables have normalized, suggesting a role played by other factors in regulating blood flow to the tissues. This evidence has changed our concept of cardiogenic shock from being essentially a cardiac problem to being a microcirculatory disease.24,41-43 In effect, microcirculatory alterations have been related to an increase in mortality in humans.43 In animal models of shock, reduced splanchnic blood flow has been linked to bowel ischemic injury, bacterial translocation, and multiorgan dysfunction.44,45 It may thus be hypothesized that interventions able to

improve splanchnic blood flow will play a protective role.35,46 This has been demonstrated in animal models.13-17 The findings of studies conducted in critically ill patients have reinforced the hypothesis of a beneficial effect of EN support in conditions of circulatory compromise. In a study performed in postoperative cardiac surgery patients in 2000, Berger et al18 observed that the intestinal tract maintains its absorptive capacity, even when there is hemodynamic failure. One year later, Revelly et al19 reported beneficial effects on hemodynamic parameters of EN in 9 postoperative cardiac surgery patients. This study was followed by an observational study including 73 post–cardiac surgery patients by Kesek et al20 in which initiating EN within 72 hours of ICU admission was found to be feasible and resulted in no serious complications. However, hemodynamic instability and illness severity were not quantified. Two years later in 2005, Berger et al21 published the results of a prospective, descriptive study including 70 patients with circulatory failure, 25.7% of whom needed mechanical support. Patients were similarly assessed as in our study, although hemodynamic monitoring was less thorough. These authors reported that EN was possible in most patients and not related to serious complications, although energy delivery was often insufficient, as observed in our study. More recently, Lukas et al22 addressed the tolerability and safety of EN in 48 patients with respiratory or hemodynamic failure requiring ECMO. Nutrition tolerance was 68%, and no serious complications were associated with nutrition support. In a report of our initial experience with early EN in 7 patients requiring venoarterial ECMO for severe hemodynamic failure after cardiac surgery,47 we were able to conclude that the procedure was possible and safe in these patients. Our study has several limitations. Being an observational, single-center study with no control group conducted on a small sample size, we are aware that an intervention clinical trial is needed to confirm its findings. Although we included patients undergoing different types cardiac surgery procedures, they all showed a complicated postoperative course, including a high mortality rate, prolonged ICU stay in most cases, and a high rate of multiorgan dysfunction. The patients also had in common hemodynamic failure at least within the first 48 hours of the ICU stay. In our opinion, this patient population is highly representative of that generally observed in a cardiac surgery ICU.

Conclusions Our findings indicate that early EN in critically ill post–cardiac surgery patients with hemodynamic failure is feasible, safe, and not related to any serious complications. However, it is difficult to meet the nutrition requirements of patients when this is the only feeding route. These observations highlight the need for an established EN protocol including daily monitoring of energy delivery and energy balance to achieve preset calorie targets. These high-risk patients also need to be carefully monitored for warning signs of gut ischemia.

Flordelís Lasierra et al

161

Table 6.  Long (>7 Days) vs Short (≤7 Days) ICU Stays for Patients. Variable a

SAPS II SOFAa Mortalityb kcal delivered by enteral route/da Energy balance, kcala Nutrition tolerance, % Overall complicationsb Complications requiring EN  discontinuationb

Short Stay (n = 7) 38.7 (30.6 to 46.9) 6.9 (4.9 to 8.9) 1 (14.3) 1009.4 (757.5 to 1261.3) 129.3 (–144.5 to 403) 118 (95 to 142) 0 0

Long Stay (n = 30) 37.1 (33.6 to 40.5) 7.9 (7 to 8.6) 4 (13.3) 1254.6 (1166.9 to 1342.4) −245.8 (−349.5 to −142) 89 (81 to 97) 23 (62.1) 9 (24.3)

P Value NS NS NS .06