ARTICLE IN PRESS Clinical Nutrition (2005) 24, 124–132

http://intl.elsevierhealth.com/journals/clnu

ORIGINAL ARTICLE

Enteral nutrition in critically ill patients with severe hemodynamic failure after cardiopulmonary bypass Mette M. Berger, Jean-Pierre Revelly, Marie-Christine Cayeux, Rene ´ L. Chiolero Soins Intensifs Chirurgicaux et Centre des Bru ˆle´s, Centre Hospitalier Universitaire Vaudois (CHUV)-BH 08.660, CH-1011 Lausanne, Switzerland Received 13 May 2004; accepted 16 August 2004

KEYWORDS Nutritional support; Enteral nutrition; Hemodynamic failure; Energy balance; Energy deficit

Summary Background & aims: The study was designed to investigate and quantify nutritional support, and particularly enteral nutrition (EN), in critically ill patients with severe hemodynamic failure. Methods: Prospective, descriptive study in a surgical intensive care unit (ICU) in a university teaching hospital: patients aged 67713 yrs (mean7SD) admitted after cardiac surgery with extracorporeal circulation, staying X5 days in the ICU with acute cardiovascular failure. Severity of disease was assessed with SAPS II, and SOFA scores. Variables were energy delivery and balance, nutrition route, vasopressor doses, and infectious complications. Artificial feeding delivered according to ICU protocol. EN was considered from day 2–3. Energy target was set 25 kcal/kg/day to be reached stepwise over 5 days. Results: Seventy out of 1114 consecutive patients were studied, aged 67717 years, and staying 1077 days in the ICU. Median SAPS II was 43. Nine patients died (13%). All patients had circulatory failure: 18 patients required intra-aortic balloonpump support (IABP). Norepinephrine was required in 58 patients (83%). Forty patients required artificial nutrition. Energy delivery was very variable. There was no abdominal complication related to EN. As a mean, 13607620 kcal/kg/day could be delivered enterally during the first 2 weeks, corresponding to 70735% of energy target. Enteral nutrient delivery was negatively influenced by increasing dopamine and norepinephrine doses, but not by the use of IABP.

Abbreviations: CPB: Cardiopulmonary bypass; CVVH: Continuous renal replacement therapy; EN: Enteral nutrition; IABP: Intra-aortic balloon pump; ICU: Intensive care unit; SAPS: Simplified acute physiology score; SOFA: Sepsis-related organ failure assessment. Corresponding author. Tel.: +41-21314-2095; fax: +41-21314-1033. E-mail address: [email protected] (M.M. Berger). 0261-5614/$ - see front matter & 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.clnu.2004.08.005

ARTICLE IN PRESS Enteral/nutrition in hemodynamic failure

125

Conclusion: EN is possible in the majority of patients with severe hemodynamic failure, but usually results in hypocaloric feeding. EN should be considered in patients with careful abdominal and energy monitoring. & 2004 Elsevier Ltd. All rights reserved.

Introduction Cardiovascular diseases generate a large number of admissions to intensive care units (ICU). The most frequent causes of admissions are acute myocardial infarction, cardiac surgery, and acute cardiomyopathies. The vast majority of cardiac surgery patients do not require artificial nutritional support, as they generally stay briefly in the ICU and are able to resume feeding within 1–2 days after surgery. But some patients suffer complicated clinical course requiring pharmacological and/or mechanical cardiac support, and prolonged mechanical ventilation. In our unit, they represent 5–10% of the cardiac surgery patients. Such patients are frequently hyper-catabolic, and unable to feed themselves for more than 5–6 days, which makes them candidates for artificial nutritional support.1 Enteral nutrition (EN) is considered desirable in acutely ill patients for a variety of metabolic, immune and practical reasons. Nevertheless, EN is commonly considered contraindicated and potentially a threat to gut integrity in severe circulatory compromise requiring high levels of vasopressors. Such patients have a combination of hemodynamic, intestinal and practical problems which render enteral feeding difficult, and sometimes hazardous. Splanchnic hemodynamics may be altered after cardiopulmonary bypass (CPB), exposing the patient to the risk of gastrointestinal complications, particularly bowel ischemia: after cardiac surgery the prevalence of acute mesenteric ischemia varies between 0.5%2 and 1.4%.3 Mortality associated with intestinal ischemia is high ranging between 11% and 27%.2,3 Further, bowel motility is reduced due to a combination of pyloric dysfunction which is frequent in critically ill,4 and intestinal atony. Finally there are practical problems such as the numerous interruptions of enteral feeding which characterize ICU patients. All theses reasons expose the patients to insufficient energy delivery by the enteral route and to progressive malnutrition.5 The ASPEN 2002 guidelines state that EN should be deferred until the patient is hemodynamically stable.1 Defining stability may be difficult though in patients requiring prolonged inotropic therapy, as well as in patients requiring mechanical ventricular assistance such as intra-aortic balloon pump (IABP) for

many days.6 On the other hand, the gut appears functional in such patients: our group has shown that enteral absorption is maintained, even in severe hemodynamic failure,7 suggesting that partial EN is possible in such patients. In absence of conclusive data for guidance, this study aimed at observing nutritional support techniques, and particularly enteral energy delivery in consecutive ICU patients with circulatory compromise after cardiac surgery.

Patients and methods The 17 bed surgical ICU of the Lausanne tertiary care university hospital is equipped with a computerized information system enabling the collection of all physiologic, monitoring, therapeutic and laboratory variables. The study was designed as an observational study with prospective data collection in the computerized database. It was conducted with Institutional Ethics Committee approval in consecutive patients admitted after cardiovascular surgery with cardiopulmonary bypass and requiring 5 days or more ICU treatment days. A waver of individual consent was delivered for this observational study based on anonymous computer retrieved data. The severity of illness and presence of organ failure were assessed during the first 24 h of admission using (1) the SAPS II (simplified acute physiology score),8 and (2) the SOFA score (sepsis-related organ failure assessment).9 The later score defines 6 organs/systems and attributes 0 (no failure) to 4 points (maximal failure) to each according to defined criteria of severity: the cardiac failure score escalates from 1=mean arterial pressure o70 mmHg, to 4=continuous dopamine/dobutamine415 mg/kg or, epinephrine/norepinephrine40.1 mg/kg/min. Failure was stratified by 1–2 points for one organ=moderate failure, 3–4 points=severe failure. The following patient data were recorded: age, sex, weight, height, surgical intervention, daily energy delivery, route of nutritional support, use of vasopressors (total daily norepinephrine dose in mg) and mechanical myocardial support devices, length of mechanical ventilation and of ICU stay, abdominal complications, and ICU outcome. Use of IABP, was recorded, as was that of continuous renal

ARTICLE IN PRESS 126 replacement therapy (CVVH). The data retrieved from the ICU computerized information system (MetaVision, iMDsoft, Tel Aviv, Israel) included automatically recorded physiologic variables, and drug and nutrition prescription. The energy supply includes energy resulting from non-intentional energy delivery (see below) and from nutritional support, enabling exact automatic calculations of both energy delivery and balances. The use of prokinetics was recorded. The appropriate data were recorded from admission to discharge from the ICU. Patients with ‘‘short stay’’ (i.e. 5–7 days) were analysed separately from those with ‘‘long stay’’ (i.e.47 days). Nutritional support: According to the ICU feeding protocol, the enteral route was considered first for nutritional support. Energy target was set at 25 kcal/kg/day based on prior metabolic trials.12 Oral feeding was recorded on food eaten from standardized meals (eating of 0–25–50–75 or 100%). The enteral feeding protocol specifies that the target is to be reached step-by-step over 4–6 days: day of admission 0%, 1—0%, 2—25%, 3—50%, 4—75%, 5—100%). EN was started at 20 ml/h, and increased stepwise, every 12–24 h according to tolerance. Clinical criteria used to assess feeding tolerance were the volume of gastric residues (o or4300 ml), the occurrence of abdominal distension, ileus, vomiting, broncho-aspiration of gastric contents, impossibility to achieve energy target defined as energy delivery o50% of target for more than 3 days. Feeding solutions: Enteral formulas were industrial solutions: Isosources standard, fiber or energy combined with Impacts in inflammatory patients (maximum 1000 ml/ day for 8–10 days) (Novartis, Bern, Switzerland). Oral supplements were provided as Ensures (Abbott AG, Baar, Switzerland) or Fresubins (Fresenius Kabi, Bad Homburg, Germany). Parenteral nutrition solutions were prepared by the CHUV pharmacy. The ICU standard solution bag (Nutrisic&, 1960 ml with 2330 kcal) delivers 18.5% of energy as proteins, including 27 g glutamine (Glamins and Dipeptivens, Fresenius Kabi), 60% as carbohydrates (glucose 50%), and 21.5% as lipids (Lipofundin MCT/LCTs 20%, B.Braun Medical, Melsungen, Germany), 50 mg Se, 10 mg Zn (Laboratoires Aguettant, Lyon, France), and 1 A AddamelsN (Fresenius Kabi)—volume was prescribed according to energy requirements. Composition of Nutrisic was individualised in case of renal or respiratory failure: renal - reduction of protein content; respiratory - reduction of lipid content. Vitamin B1 100 mg, vitamin C 1 g, Se 100 mg and Zn 10 mg supplements were supplied to all patients.

M.M. Berger et al.

Nutritional data and calculations Energy delivery: Twenty-hour intakes included nutritional and non-nutritional energy delivered to the patient, (1) intentional, i.e. energy intakes resulting from artificial nutrition, oral nutrition (assessed by the nurses as 0–25–50–75 or 100% of standardized meals, or liquid supplements) and (2) non-intentional, i.e. energy delivered by infusions and fluids used to solubilize drugs (glucose, glucosaline, etc), as well as the 10% lipid emulsion included with the sedative drug propofol (Disoprivan 2%, Astra Zeneca, Zug, Switzerland). Estimated energy balance (24-hour) was calculated as the daily difference between total energy delivered–calculated energy target (25 kcal/kg): the energy stored was ignored. Statistical analysis: Results are presented as mean7SD, or medians as appropriate. One-way ANOVA was used to compare patient characteristics and two-way ANOVA for changes over time. Single and multiple regressions were calculated to analyse the relationship between energy intakes, energy balance and other variables such as 24-h vasoactive drug doses dose. Statistical package: JMPs Version 3.1.5., SAS Institute Inc., Cary, NC, USA.

Results Patient population: Seventy patients out of 1114 consecutive patients (6.3%) admitted after cardiac surgery with CPB were observed prospectively over a 2-year period (Fig. 1): 50 were after coronary artery bypass graft and valve surgery, 12 after thoraco-abdominal aortic surgery, and eight after heart transplant. The patients’ characteristics and therapeutic requirements are described in Tables 1 and 2. A total of 715 patient days was analysed: 39 patients stayed 5–7 days, while 31 stayed longer,

Figure 1 Patients enrolment and type of feeding.

ARTICLE IN PRESS Enteral/nutrition in hemodynamic failure

127

Table 1 Clinical data and severity of disease of the overall study population and of the subgroups (mean 7SD (median or %)). Population

N

Age

SAPS

SOFA

Length MV (days)

Length ICU (days)

Deaths

Overall Transplant IABP CVVH

70 8 18 10

67717 50713* 68710 6379

43715 40711y 45714 52720y

1073 973 1273 1272z

575 576 675 775

1077 (7) 15716 (8)y 1275 (12)y 16714(12)y

10 1 5 3

(3) (1) (4) (4)

(14%) (13%) (29%)y (33%)z

MV=mechanical ventilation. * Po0:001 versus all other patients. y P ¼ 0:04 versus all other patients. z P ¼ 0:03 versus all other patients. y P=0.051–0.06=ns, but strong trend.

Table 2

Types of cardiovascular, respiratory and renal support required by the patients (mean7SD (median)).

Variable

N

Duration of ttt (d) Mean 7SD (median)

Total number of days with support

Length of ICU stay Mechanical ventilation Vasopressor drugs (epinephrine or norepinephrine) Inotropic drugs (dobutamine or dopamine) IABP Continuous renal replacement therapy

70 70 63 67 18 10

1077 (7) 575 (3) 5.474.4 (5) 6.173.7 (5) 572 (5) 13714 (10)

715 326 376 419 88 125

including 11 for more than 2 weeks. The patients with short ICU stay were less severely ill than those with long stays as reflected by lower SAPS II (40714 versus 46716 P ¼ 0:09), and lower SOFA score (8.970.4 versus 11.570.5, Po0:0001). Organ and system failure caused prolonged ICU stay with the very frequent presence of multiorgan failure: the median number of failing organs according to SOFA score was 4. By inclusion criteria, all patients had some degree of cardiovascular failure, with severe failure in 59 patients (84%) (i.e. SOFA score 3–4, requiring vasopressors and/or IABP support—see below), and moderate failure in 11 (i.e. SOFA 1-2). The second most frequent organ failure was respiratory failure, which was present in 69 patients (severe failure in 49 and moderate in 20). Coagulation disorders were present in 60 patients (severe in 10, moderate in 50), neurological alterations in 34 patients (severe in 6, moderate in 28), liver failure in 13 patients, and renal failure in 41 patients (severe in 9, moderate in 32). Ten patients required continuous renal replacement therapy. Eight patients suffered complicated clinical course after transplantation (Table 1). They were

younger than the other patients (Po0:001), with a longer mean stay in the ICU (P ¼ 0:04). Their requirements for IABP support (2/8) and CVVH (2/8) did not differ from other patients. IABP support was required in 18 patients. Their SOFA score was higher than those of patients without IABP (P ¼ 0:009). Other variables did not differ significantly (Table 1). Mortality tended to be higher (P ¼ 0:056). Renal replacement therapy was required in 10 patients, including three patients on IABP. Compared with non-CVVH patients, they were sicker on admission by SAPS II and SOFA scores, and stayed longer in the ICU (P ¼ 0:009). Mortality was higher in CVVH patients (P ¼ 0:03).

Nutritional support—routes (Table 3) Out of the 715 days, 154 (22%) were without nutritional support of any form, corresponding to the 2–3 first days of ICU stay, during which feeding is not yet considered (Table 3). The other 561 days were with oral or artificial feeding. Altogether 40 patients received EN alone or in combination with

ARTICLE IN PRESS 128

M.M. Berger et al.

Table 3 Routes of feeding and energy delivery in all patients and according to length of stay (the numbers indicate days/patients). Routes-

No feed

Oral (pure)

Enteral (pure and with oral)

Parenteral and combined EN+IV

Total stay

154 90 64

176/30 36/18 6/2

350/36 116/11 368/25

35/4 0/0 35/4

715/70 242/39 473/31

kType of stay All Short stay Long stay

parenteral feeding in both short and long stay groups. The route and time to initiation of EN was similar in all categories of patients. Time to initiation was below 72 h in 68% patients considered for EN (27/40). The gastric route was used in 49% of artificial feeding days (276/561), while the jejunal route was used only in 74/561 days (13%).

Energy delivery (Fig. 2, Table 4) Daily individual energy delivery was variable. In absence of deliberate nutritional support during the first 48–72 h, patients received hypocaloric supply from their intravenous infusions, resulting in a mean daily energy delivery of 2757190 kcal/day (Fig. 2, Table 4). Energy delivery was highest, although below target, in those patients in severe cardiogenic shock who benefited from glucose–insulin therapy to support the failing myocardium.6 Similar hypocaloric delivery was observed in all subgroups of patients. After initial stabilization, a mean energy delivery of 13607620 kcal/day was achieved by the enteral route during the first 2 weeks (which represents 70735% of energy target), with 300 kcal/day more by jejunal than gastric route (ns). The mean daily estimated energy deficit was low with 2557370 kcal/day, although variability of daily and cumulated deficit was substantial. Short ICU stay: Mean daily energy delivery was lower and more variable in the short ICU stay patients compared with the long stay (5347517 versus 11977736, Po0:01), explained by attempting oral feeding (Fig. 2). The large variability explains the absence of significance of the difference. Long ICU stay: As the feeding progressed, the mean daily energy delivery becomes larger, reducing the daily deficit. Among the long stay patients, three differed in their nutritional management: one patient was moderately overfed during her stay (receiving a mean of 35 kcal/day: cumulated balance+9200 kcal in 19 days), and two others were

Figure 2 Mean energy delivery in patients with short (5–7 days) versus long stay (47 days), and estimated mean energy deficit: curves over the zero line show mean energy delivery in patients staying o8 days (n ¼ 39) and in those staying longer (n ¼ 41), while curves below the zero line show estimated energy balances (deficit) in the two groups of patients. The patients staying 5–7 days have smaller energy intakes, and consequently tend to have larger energy deficits.

grossly underfed (14,300 in 29 days and –18,700 kcal in 22 days): the latter patient eventually died. CVVH patients: Energy balance during the 2nd week was higher in CVVH patients (290 versus –1850 kcal from day 8 to 14: non-significant), while the cumulated energy deficit of the whole stay was similar. Transplantation patients: Enteral energy delivery and deficit were less variable in the eight cardiac transplant patients than the overall patient population: mean cumulated energy deficit at discharge was 188073600 versus 233574260 kcal in all other patients (ns). IABP patients: In those 18 patients, i.e. those with most severe cardiac failure, enteral energy delivery was initiated between 24 and 48 h in 8/18 patients, and before 72 h in 13/18, and between day 4 and 6 in 4 patients (Fig. 3). During the 1st week, the energy delivery was higher in the IABP

ARTICLE IN PRESS Enteral/nutrition in hemodynamic failure Table 4

129

Type of feeding, duration and resulting energy delivery over the 715 days.

Route

Energy delivery (kcal/d)

Estimated energy balance (kcal/d)

Route and combinations

N days (%)

No nutrition

2757190

—

—

22%

Per os

6707560

7107740

—

24%

Gastric

12507650*,y

2107680

Pure gastric Gastric+p.os

229(32%) 47 (7%)

Jejunal

15457720*,y

1707770

Pure jejunal Jejunal+p.os

59 (8%) 15 (2%)

Intravenous

19607490y

160768

Pure IV IV+EN IV+p.os

4 (1%) 33 (4.6%) 1

*

Difference gastric versus jejunal energy delivery: ns. Difference intravenous versus jejunal or gastric energy delivery: P ¼ 0:001:

y

Figure 3 Mean total energy delivery (full line) with the detail of energy provided by enteral nutrition (dotted line) in 18 patients with intra-aortic balloon pump.

patients compared to the other patients, resulting in a significantly better energy balance (270 versus 1920 kcal in 7 days, P ¼ 0:01). One patient did not receive any form of feeding during her 7day stay, including 4 days of IABP support: she was transferred to the ward and expected to feed spontaneously. The 2nd week energy deficit was similar in both IABP and other patients (1195 versus 1340 kcal from day 8 to 14). Vasoactive drugs: Most patients were on vasoactive drugs for many days (Table 2), epinephrine being stopped within 12–48 h of admission and switched to norepinephrine, before any significant enteral feeding was initiated. The most frequent combination was dobutamine plus norepinephrine. Multiple linear regression analysis between total daily dose of ‘‘maintenance vasoactive drugs’’ (dopamine, dobutamine and norepinephrine) and

daily enteral energy delivery showed that dopamine and norepinephrine were significantly negatively related with enteral feeding (F ratios of 4.69, P ¼ 0:03 and 8.96, P ¼ 0:003; respectively), while there was only a negative trend with dobutamine (F ¼ 2:82; P ¼ 0:09). Intestinal tolerance: No patient experienced any serious gastrointestinal complication. Prokinetics were used in nine patients, including five patients receiving one single isolated injection for an episode of nausea: four patients required prokinetics for 3–10 days (three on cisapride and one on metoclopramide). Complications: 39 patients (56%) required antibiotic treatment for infectious complications (bronchopneumonia). Mean cumulated energy deficit tended to be larger in those patients with infections (3’09074920 versus 126572725 kcal during the stay: P ¼ 0:068).

Discussion The present descriptive study shows that EN is possible in many patients with severe hemodynamic failure after cardiac surgery. Nevertheless such feeding usually results in hypocaloric feeding. To our knowledge, this is the first study to focus on enteral energy delivery in a group of patients with severe circulatory failure. The energy data show that enteral feeding not only is possible, but can be considered as a true nutritional support. A recent series of 73 cardiothoracic ICU patients also reports the feasibility of EN during the early postoperative period.10 Unfortunately, description of the patient

ARTICLE IN PRESS 130 population is incomplete regarding severity of failure (vasopressor and IABP support requirements) and energy delivery, rendering comparison with our results impossible. That trial mainly focused on gastrointestinal tolerance, which was good. The only information about daily energy supply is that energy target could be reached with EN only in nine patients (12%). In our collective, more than 1200 kcal per day could be delivered by the enteral route in all the patients requiring artificial nutrition. The patients included in the study were severely sick as reflected by their elevated SAPS and SOFA scores: hemodynamic failure was critical as shown by the dependence on norepinephrine and other vasoactive drugs for hemodynamic stability. Among these 70 patients, 18 were even dependent on mechanical cardiac support (IABP), i.e. in extremely severe circulatory failure. The patients had a complicated postoperative clinical course, and constitute a population at high risk of gastro-intestinal complication such as bowel ischemia.2,3 In such critically ill patients the level of energy requirement is highly variable. Hypermetabolism is frequent, but the requirements are difficult to predict, especially during the early postoperative period. In our experience, based on repeated indirect calorimetry determinations, the energy requirements can be set at 25 kcal/kg/day,11,12 resulting in a mean target of 1900 kcal/day in this group of patients. The data also show that gastric feeding was possible during severe hemodynamic failure in nearly any patient. As a mean, 1000 kcal per day could be delivered by the gastric route, and 1500 kcal by the postpyloric route with a large variability. Energy delivery data in the patients with the most severe failures, i.e. those on IABP, showed similar results, enabling the delivery of similar amounts of energy by the enteral route (15–0 kcal/kg/day). Two recent reviews suggest that EN is well tolerated and probably beneficial in most critically ill, as it contributes to restoring splanchnic perfusion and immune function.11,14 It should however be used with caution in hypotensive patients.13 Our ICU enteral feeding protocol specifies a very prudent increase of energy delivery to energy target over 5 days, with no enteral feeding on admission while in unstable shock during the first 48 h, feeding being initiated after 24–48 h. Our data collected over 350 days of EN were associated with no abdominal complications, and show that EN is reasonable and safe in such critically ill patients. This protocol results in hypocaloric feeding since parenteral nutrition was not used for top up of energy deficit: this may have deleterious consequences by itself.

M.M. Berger et al. We compared the 39 patients with short stay with the 31 patients with long stay. The short stay corresponds to the maximal starvation time in critically ill according to experts.1 Ten (14%) patients were given no nutrition while in the ICU, and 18 (25%) resumed oral feeding. In the 28 patients without artificial nutrition, a significant energy deficit developed, oral feeding being inefficient to provide significant amounts of energy: this deficit was less pronounced in the 11 who got early EN. Their hemodynamic evolution was the determinant of their ICU stay, and absence of early nutrition did not appear to influence their outcome: such ‘‘short starvation’’ can therefore be considered to have no serious deleterious consequence in case of hemodynamic recovery, but may become a problem in case of prolonged organ failure, as an ‘‘energy debt’’ is created. Nevertheless, it must be emphasized that we observed a high rate of bronchopneumonia in our patients (56%), the incidence tending to be higher in the patients with the largest energy deficits. This high rate is in part explained by the selection bias of our trial: only the sickest patients with organ failure were investigated (70/1114) and all required mechanical ventilation, which is a well-known risk factor: considering the complete population the incidence was reasonable. In addition, the patients were indeed on hypocaloric feeding, which is an additional risk of infectious complications. In acute conditions and especially in postoperative states, many authors recommend the use of parenteral nutrition, despite EN being considered the optimal route in most critically ill.1 In acute circulatory failure EN is indeed considered relatively contraindicated, as it may aggravate gut ischemia: low mesenteric blood flow is a risk factor of bowel necrosis. Such opinions are expert opinions, and the available literature is sparse and contradictory. Despite lack of evidence, the recently revised guidelines for nutritional support of the American Society for Parenteral and Enteral Nutrition recommend a very cautious approach during the introduction of enteral feeding in patients with hemodynamic failure.1 The normal hemodynamic response to feeding is complex, including an increase in cardiac output, and vasodilation of mesenteric arteries, and a decrease in peripheral resistance. This hemodynamic adaptation to feeding may induce increase in local oxygen consumption (VO2) rendering oxygen delivery inadequate. In acutely ill patients with cardiac failure, this response may be altered and worsen an already insufficient O2 delivery to the tissues and organs. During low cardiac output, splanchnic O2 delivery is reduced, while splanchnic

ARTICLE IN PRESS Enteral/nutrition in hemodynamic failure O2 consumption is maintained: therefore splanchnic O2 extraction is high.15 This is one of the factors explaining the high rate of gastrointestinal complications in this category of patients.2,3 In patients with chronic heart failure, continuous enteric feeding set at 1.4–1.5 times resting energy expenditure, compared with intermittent feeding, has been shown to be well tolerated.16 The authors concluded that EN can be provided safely, except in patients with overt cardiac failure.16 The combination of oral food and parenteral nutrition to achieve 20–30 kcal/kg per day for 2–3 weeks in patients with cardiac cachexia (severe mitral valve disease and congestive heart failure) has also been shown to be associated with stable hemodynamics, unchanged whole body VO2 and CO2 production.17 Altogether, these data suggest that careful limited hypocaloric enteral feeding can be administered in patients with acute and chronic circulatory failure, particularly if continuous feeding is used. Our team has repeatedly shown that cautious EN can be used even during severe cardiac compromise. Paracetamol absorption, which is very similar to that of peptide absorption, is maintained in postoperative cardiac surgery patients even in low output states.7 In a series of 23 cardiac surgery patients with hemodynamic failure, paracetamol jejunal absorption was maintained compared with controls without cardiac failure. Such patients can be fed by either gastric or jejunal route according to clinical tolerance of enteral feeds. In another study including cardiac surgery patients requiring inotropic support, the introduction of continuous EN caused a 10% increase in cardiac index and splanchnic blood flow, a 10% decrease in mean arterial pressure in parallel with decreased systemic vascular resistances with unchanged heart rate.12 Metabolic and endocrine responses indicated that nutrients were utilized as energy substrate: on initiation of EN, glucose turnover increased, as did plasma glucose concentrations. These data suggest that the hemodynamic response to early continuous EN is adequate after cardiac surgery. In all our studies, EN alone resulted in significant energy deficits: combined enteral and parenteral nutrition is required to achieve energy targets. The total energy delivery should be monitored due to the limited feeding volume tolerance. Gastrointestinal motility is reduced in patients with severe hemodynamic failure. Our team has shown that self-propelled feeding tubes can be used in such conditions18 although progression was lowest in those patients on highest norepinephrine doses: norepinephrine and morphine doses were indeed the most important determinants of feeding

131 tube progression. In the present trial, enteral energy delivery was inversely related to dopamine and norepinephrine dose, confirming the negative impact of these vasoactive drugs on the gastrointestinal motility: despite this negative influence enteral feeding remained possible, although hypocaloric. Interestingly, few of our patients actually needed prokinetics confirming the data by Kesek et al.:10 only four patients received such stimulation, stressing that gastro-intestinal motility although impaired in some patients, is not a major issue. Moreover, none of these 70 patients, not even the 18 patients on IABP, experienced any serious gastrointestinal complication. In conclusion the data show that: (1) EN is possible during the 1st postoperative week, and even already after 24 h, in patients with acute severe circulatory failure under careful abdominal monitoring, (2) EN generally results in insufficient energy delivery, stressing the importance of careful monitoring the total daily energy delivery, and (3) combination with parenteral nutrition may be required to achieve optimal energy delivery.

Acknowledgements The study was entirely supported by the funds of the surgical ICU-CHUV. There is no conflict of interest for any of the persons involved in the trial (specific document signed).

References 1. ASPEN Board of Directors and the Clinical Guidelines Task Force. Guidelines for the use of parenteral and enteral nutrition in adult and pediatric patients. J Parenter Enteral Nutr 2002;26(1 Suppl):1SA–138SA. 2. Venkateswaran RV, Charman SC, Goddard M, Large SR. Lethal mesenteric ischaemia after cardiopulmonary bypass: a common complication? Eur J Cardio-Thor Surg 2002;22:534–8. 3. Lazar HL, Hudson H, McCann J, Fonger JD, et al. Gastrointestinal complications following cardiac surgery. Cardiovasc Surg 1995;3:341–4. 4. Heyland DK, Tougas G, King D, Cook DJ. Impaired gastric emptying in mechanically ventilated, critically ill patients. Intens Care Med 1996;22:1339–44. 5. Berger MM, Chiole ´ro RL, Pannatier A, Cayeux C, Tappy L. A ten year survey of nutritional support in a surgical ICU: 1986–1995. Nutrition 1997;13:870–7. 6. Berger MM, Mustafa I. Metabolic and nutritional support in acute cardiac failure. Curr Opin Clin Nutr Metab Care 2003;6:195–201. 7. Berger MM, Berger-Gryllaki M, Wiesel PH, et al. Gastrointestinal absorption after cardiac surgery. Crit Care Med 2000;28:2217–23.

ARTICLE IN PRESS 132 8. LeGall JR, Lemeshow S, Saulnier F. A new simplified acute physiology score (SAPS II) based of a European/North American multicenter study. J Am Med Assoc 1993;270:2957–63. 9. Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsisrelated Organ Failure Assessment) score to describe organ dysfunction/failure. Intens Care Med 1996;22:707–10. 10. Kesek DR, Akerlind L, Karlsson T. Early enteral nutrition in the cardiothoracic intensive care unit. Clin Nutr 2002; 21:303–7. 11. Revelly JP, Berger MM, Chiole ´ro R. The hemodynamic response to enteral nutrition. In: Vincent J editor. Yearbook of intensive care and emergency medicine. Berlin: Springer; 1999. p. 105–14. 12. Revelly JP, Tappy L, Berger MM, Gersbach P, Cayeux C, Chiolero R. Metabolic, systemic and splanchnic hemodynamic responses to early enteral nutrition in postoperative patients treated for circulatory compromise. Intens Care Med 2001;27:540–7.

M.M. Berger et al. 13. McClave SA, Chang WK. Feeding the hypotensive patients: does enteral feeding precipitate or protect against ischemic bowel? Nutr Clin Practice 2003;18:279–84. 14. Zaloga GP, Roberts PR, Marik PE. Feeding hemodynamically unstable patient: a critical evaluation of the evidence. Nutr Clin Practice 2003;18:285–93. 15. Jakob SM, Ensinger H, Takala J. Metabolic changes after cardiac surgery. Curr Opin Clin Nutr Metab Care 2001; 4:149–55. 16. Heymsfield SB, Casper K. Congestive heart failure: clinical management by use of continuous nasoenteric feeding. Am J Clin Nutr 1989;50:539–44. 17. Paccagnella A, Calo ` M, Caenaro G, et al. Cardiac cachexia: preoperative and postoperative nutrition management. J Parenta Enteral Nutr 1994;18:409–16. 18. Berger MM, Bollmann MD, Revelly JP, et al. Progression rate of self-propelled feeding tubes in critically ill patients. Intens Care Med 2002;28:1768–74.