Presentación

Comité de Honor 

Comité Científico  

Comité Técnico 

Registrarse

Programa  

Comunicaciones

Instrucciones para el IRC (chat)









A cost-effectiveness study of enteral immune modulating nutrition in intensive care patients

Elizabeth Coates BA (Hons) Research Associate Medical Economics and Research Centre, Clare L Hibbert BA (Hons) MRC Fellow in Health Services Research Medical Economics and Research Centre, Sheffield (MERCS) Intensive Care Unit Royal Hallamshire Hospital Glossop Road Sheffield S10 2JF United Kingdom

 

Introduction

Malnutrition is a major problem for patients admitted to the Intensive Care Unit (ICU). There are several reasons for this. Firstly, the development of aggressive surgical and medical interventions, and the increasingly ageing population have resulted in higher levels of malnutrition at hospitalisation. Secondly, advancements in intensive care may ensure the prolonged survival of severely ill patients, but this also results in malnutrition [1]. Malnutrition is due to the nature of patients' conditions, that is, chronic wasting from the advanced catabolic state that generally accompanies critical illness. Acute malnutrition slows down recovery rates and increases hospital and ICU length of stay, as well as ventilator dependency [1-2]. In addition, the compromise of the immune system makes patients more vulnerable to nosocomial infections [1, 3]. These infections further increase the likelihood of more serious infections, and therefore, single or multiple organ failure (MOF). This is important as MOF can also lead mortality, for example, the first studies of MOF recorded a consistently positive relationship with infection and concluded that MOF is the "fatal expression of uncontrolled infection" [4, 5]. Severe infections carry significant economic implications in that they further increase length of stay and also result in greater resource consumption [3].

One way to overcome the problems associated with malnutrition is to provide enteral nutritional support. This treatment is beneficial to patients who cannot receive food or drink by mouth [6], and involves the nasal insertion of a tube with the aim of reaching the stomach, for gastric feeding, or, the jejunum, in order to feed the small bowel [7]. Enteral feeding is generally considered to be preferable to the parenteral route as it enables the maintenance of structure and integrity of the gut. It builds up a patient's resistance to bacteria and endotoxins, and consequently, protects against infection and metabolic irregularities. However, evidence demonstrates that this treatment alone is unable to affect the incidence of nosocomial infections [2].

Nevertheless, the development of immunonutrition (IMN), a special form of enteral feed supplemented with specific nutrients (omega-3 fatty acids, arginine, nucleotides and sometimes glutamine) has demonstrated a beneficial effect on patients' immune systems [8]. The advantages of IMN have been demonstrated in a number of studies [9-20]. Two recent meta-analyses have concluded that the use of IMN results in a significant reduction in infection rates, and as a consequence, shorter durations of hospital stay [21-22].

Despite the clinical evidence however, the cost-effectiveness of IMN has rarely been explored. Given the significant resources needed to provide intensive care, for example, this specialty costs, on average, four times that of ward care, it is imperative that any potential means of reducing expenditure are investigated. The importance of this is amplified by the opportunity for infection to generate further resource use (in that patients become more severely ill and stay longer in both the ICU and the hospital). Furthermore, an European ICU infection prevalence rate of more than 20% [11] within the context of increasing pressure to reduce expenditure in the health service clearly demonstrates that both the costs and the consequences (or effectiveness) of any such treatments should be examined.

The acquisition costs of IMN are higher than that of standard enteral nutrition (SEN), and this factor is often attributed to its' limited use within the ICU. However, IMN's ability to reduce infection demonstrates the potential to make relative cost savings, and as such, may present a more cost-effective option.

 

BACKGROUND AND LITERATURE REVIEW

Definition of immunonutrition

As previously described, immunonutrition is the term given to describe special enteral feeds containing arginine, omega-3 fatty acids, nucleotides and sometimes glutamine [8]. Arginine is a non-essential amino acid that improves immune responses to bacteria, viruses and tumour cells, as well as promoting wound healing and protein turnover [23-24]. High levels of omega-3 fatty acids can be found in fish oil and rape seed, and this nutrient is key to cell functioning, e.g. dilation and contraction, inhibition and clotting, as well as cell growth and division [25]. Nucleotides form the basis of deoxyribonucleic acid (DNA) and ribonucleic acid (RNA), and are required for almost all cellular activities, for example, the transfer of energy, catalysis ad the coordination of hormonal signals [26]. Nucleotides also enhance repair cell production, and prevent the loss of important bacteria in the gut [27]. Glutamine, the most important acid in the body, is an important source of nitrogen and calories [28]. Based on animal studies that demonstrated the beneficial effects of these nutrients on immune function, similar feeds were developed for humans [29].

Meta-analyses

There have been many randomised controlled trials (RCTs) that have assessed the clinical efficacy of IMN in a critically ill population [9-20, 30-35]. As a consequence, at least two meta-analyses on the effects of IMN have been recently published [21-22]; although many other smaller scale studies and discussions exist [2, 8, 29, 36-41]. Beale et al [21] studied twelve RCTs [9-20] for differences in mortality, infection, ventilator days, ICU and hospital length of stay, diarrhoea days and calorie and nitrogen intake. Immunonutrition was found not to influence survival but to significantly reduce hospital length of stay, the number of ventilator days and infection rates. Heys et al [22] studied eleven RCTs [10, 12-17, 20, 30, 34-35] for differences in major infectious complications, nosocomial pneumonia and ICU and hospital length of stay between patients receiving IMN with those receiving SEN. The results were favourable, with patients in the immunonutrition groups experiencing fewer infectious complications and days in hospital.

Clinical evidence

Cerra and colleagues conducted one of the earliest trials examining the effects of IMN in ICU patients, and concluded that feeding with an enriched enteral formula was associated with an improvement in immune function [19]. In two similar trials, Daly and colleagues compared IMN with SEN in patients who had undergone surgery for upper gastrointestinal cancer [15-16]. Both studies demonstrated the ability of the experimental feed (IMN) to reduce the incidence of infectious and wound complications, for example, in the first study: 11% vs. 37% (p= 0.02) and in the second study: 10% vs. 42% (p< 0.01). Length of hospital stay was also shorter in the group who received IMN.

Key studies

However, three studies appear to have attracted the most attention in the debate surrounding IMN [9-11]. In 1998, Atkinson et al demonstrated that despite no significant difference in hospital mortality rate (the primary outcome measure), analysis in a predefined sub group demonstrated several interesting benefits. That is, the patients who successfully achieved early enteral nutrition (³2.5 litres in the first 72 hours after ICU admission) experienced a reduction in ICU and hospital length of stay, and also the duration of mechanical ventilation [9]. Improved outcomes were also evident in a RCT conducted by Bower [10]. This study was able to demonstrate in a subgroup of septic ICU patients that IMN could be associated with a reduction in the frequency of acquired infections (p< 0.01), and hospital length of stay (p< 0.05). However, it is important to note that the mortality rate was higher in the IMN group, i.e. 23/147 vs. 10/132 in the control group. Galban and colleagues also looked at the effects of IMN in septic ICU patients [11]. There was a reduction in the number of bacteremias (8% vs 22%, p< 0.05), and although the different groups developed the same number of late infections, fewer patients receiving IMN developed more than one infection (5/89 vs. 17/87; p= 0.01). There was a lower mortality rate in the group receiving the experimental feed: 17/89 vs. 28/87 in the control group (p< 0.05).

Negative findings

Despite the benefits demonstrated by the above studies, not all of the clinical trials of IMN have been favourable [32, 34-35]. One major critic of the treatment is Heyland, whose numerous reviews of the evidence have highlighted several important issues [37-39]. In a recently published review, the author tackles the aforementioned meta-analyses [21-22], and expresses concern at their production of different estimates of IMN's effect on mortality [39]. As already discussed, the results from Beale et al's meta-analysis did not show any significant effect of IMN on mortality; Relative Risk (RR) =1.05; 95% Confidence Intervals (CI) = 0.78-1.41[21]. However, the results of the meta-analysis from Heys et al [22] (Odds Ratio (OR) = 1.77; 95% CI = 1.00-3.12) are 'consistent with more harm than good' [39]. The application of different search strategies ensured that different papers were included, and as such, Heys et al made reference to two RCTs that do not support the use of IMN in critically ill patients [34-35]. Mendez and colleagues found that trauma patients who received IMN actually spent more time in the hospital (32.9 vs. 22 days), and longer on a ventilator than patients fed with SEN (16.4 vs. 9.7 days). The incidence of infectious complications was also greater (but not significant) with IMN: 86% vs. 57% overall, although, there was little difference in mortality (4.5% vs. 5%) [34]. A trial of burns ICU patients compared IMN with a high protein and lower cost enteral feed, and found no significant differences. For example, the groups were comparable with regard to hospital length of stay, costs, ventilator support and mortality (20% vs. 12.5%) [35]. These results are supported by a trial comparing IMN with conventional postoperative treatment in patients after upper gastrointestinal cancer surgery [32]. Despite these issues, the demonstration of the beneficial effect of IMN on infection rates (and length of hospital stay) in both meta-analyses [21-22] permits an evaluation of the cost-effectiveness of IMN in reducing infection in an ICU population.

 

AIM OF THE STUDY

The aim of this study is to estimate the cost-effectiveness of immunonutrition in reducing infection rates in critically ill adult patients with sepsis, when compared to standard enteral nutrition. Evidence of effectiveness will be derived from Beale et al's meta-analysis of twelve published RCTs [21], and costs will derived from a previous study conducted in the 6-bedded adult general ICU of the Royal Hallamshire Hospital (RHH), Sheffield [42].

The main outcome measure in this study will be a reduction in the days of infection and it will be conducted from the perspective of the health care provider (NHS) with a focus on the costs incurred in the ICU.

 

METHODS

Clinical effectiveness

Evidence of the clinical effectiveness of IMN in comparison with SEN was based on the results of the meta-analysis undertaken by Beale et al [21]. This method was chosen because meta-analyses aggregate the results and conclusions of a large number of individual studies to produce a more precise and reliable estimate of a treatment's true efficacy [43]. The findings of a meta-analysis have greater generalizability than individual studies because their conclusions are based on larger patient populations within wider treatment settings. As such, the statistical power of a meta-analysis' results is increased [44].

The methods used by Beale et al to identify relevant studies for inclusion in the meta-analysis have been previously described [21]. In summary, the authors conducted a search of MEDLINE (1967 to Jan 1998) and identified 15 RCTs that met the inclusion criteria, i.e. they were conducted on an intent-to-treat basis (defined as the inclusion of all patients randomised and eligible to take part in the study, independent of their results). Three trials were excluded from the final analysis, due to bias in one study [45], and the inclusion of the results of the other two studies [30, 46] within other larger trials [13, 15]. Of the 9 studies that reported infection rates, 8 demonstrated a beneficial effect in the treatment group (Table 1).

 

Consolidated Standards of Reporting Trials  (CONSORT) statement

 

Although the quality of the studies was assessed during the meta-analysis, I felt it necessary to independently evaluate the studies using the Consolidated Standards of Reporting Trials (CONSORT) statement [47]. The CONSORT statement was developed as a standard way for the authors of RCTs to report their work, to improve the conduct of research, and also to enable the readers of published work to critically evaluate and interpret studies. The validity of an RCT can be assessed through the use of the 25-point checklist that details all the most important aspects of trial design (Table 2). The issue of bias is central to the CONSORT statement; how to prevent this, and how to make its' occurrence transparent to the reader.

Results of CONSORT statement review

The results of the review of the RCTs included in the meta-analysis are shown in Table 3. The CONSORT statement provided an effective method for assessing quality and highlighted the various shortcomings and omissions of the RCTs.

  • Several of the RCTs did not fully detail the hypotheses, objectives and planned sub-group analysis in their introduction (Criteria 3) [9-11, 13-14, 20].
  • Almost all of the trials did not provide details of any prospectively defined stopping rules for the administration of the feed (Criteria 9) [9-18].
  • Varying amounts of information were given about the methods of assigning patients to different treatments (Criteria 11-13) and how and when this assignment was concealed and broken, and whether this concealment was successful (Criteria 16-17).
  • Several RCTs did not mention the method used to generate the allocation schedule [9, 13, 18, 20], or how this was concealed during the trial [14-15, 18].
  • Many reports did not refer to how the generation of the assignment schedule was separated from its' execution [9, 11, 13-16, 18, 20].

Despite these issues, and those relating to different methods of presenting the results of the trial, as demonstrated through problems with criteria 18-19 and 22-23 for many studies, it is possible to recommend the quality of these studies. This is because many of the omissions can be attributed to the organization of material in the papers, i.e. inserting text in the methods as opposed to the introduction and so on. Several studies did not refer to sub-group analysis or detail their clinical objectives in the introduction as required, however, this was due to its' inclusion in the methods. (Table 3)

Cost-effectiveness analysis

A cost-effectiveness analysis identifies both the costs and consequences of the two treatments and was the preferred approach for this study [48]. A fully stochastic cost-effectiveness analysis was undertaken because it was possible to identify the mean size effect with an associated variance within the evidence of the effectiveness of IMN (from the meta-analysis) and, within the cost data (from ICU patients at RHH) [42]. The incremental ICU costs of using IMN were assessed by relating the additional cost of using IMN over SEN, to any reduction in infection rates and associated resource use.

This analysis enabled the results to be expressed as cost-effectiveness ratios; the cost incurred/saved per unit of outcome changed, e.g. per patient and per days of infection avoided. Due to the level of controversy surrounding IMN, it was also necessary to test the robustness of any findings through a sensitivity analysis. Sensitivity analysis is used in economic evaluations to account for the uncertainty in the data, i.e. when average estimates are used it is often necessary to test the validity of the results over a series of plausible ranges [48]. In this study, the sensitivity analysis will concentrate on two main areas of uncertainty: the costs of treating sepsis and the ability of IMN to reduce the duration of infection. The mean and 95% CIs will be used to explore how changes in the analysis can affect the generalizability of the results.

Patient population

The costs used in this study were taken from a previous study looking at the patient related costs of 213 patients admitted to the ICU in RHH between September 2nd 1995 and June 30th 1996 [42]. The medical records of these patients were retrospectively evaluated according to whether they had severe sepsis or early septic shock (sepsis) at any point during their ICU stay.  Thirty-six (16.9%) of these patients were classified as having sepsis. During the study sepsis was classified according to the definition given in Table 4.

Sepsis is a serious infection that is associated with high levels of morbidity and mortality, for example, rates of approximately 50% [49-51] and even up to 90% in some studies [52]. As such, the occurrence of sepsis can further increase length of stay and resource consumption in the ICU [42]. Further, the demographic characteristics of these patients can be considered as similar to those in previously published studies [53-54], a factor which therefore strengthens the representativeness of the patient population. (Table 4)

Resource data

The costs for this study were determined from retrospective data produced by an activity-based costing methodology [55] and based on the resource utilisation of the aforementioned patients [42]. The activity-based costing method measures the patient-related costs of delivering care, i.e. costs attributable to individual patients. These costs include nurses, medical ward rounds, drug treatments, disposables, equipment utilisation and clinical support services such as physiotherapy, radiology and laboratory services. The care delivered to a patient is partitioned into discreet elements termed 'activities of care'. An activity of care is defined as any patient-related task requiring the use of ICU resources. At the time of study completion, there were over 300 activities of care configured on the system, for example, all the drugs, treatments and major monitoring procedures used on the ICU, as well as the 'background' nursing care and discussions with relatives. Patient care is costed using the activities of care by identifying the resources (and therefore costs) necessary to deliver each activity. For each day of ICU stay a 'daily care record' is produced which documents the care delivered to a patient and its' associated cost. The total patient-related cost of care for an individual patient is determined from the sum of the costs of the activities delivered to that patient [42]. As a result, the costs collected in this study are as follows; nurses, medical staff, drugs, fluids and consumables, medical imaging (x-rays, scans etc) and other staff costs (technical and administrative support), however, overhead costs were not included in this study.

 

RESULTS

Resource use associated with the treatment of sepsis

The estimated differences in costs of intensive care for patients with and without sepsis are shown in Table 5. These costs are summarised under seven headings: drugs, fluids, consumables, nursing, medical staff and other staff (e.g. technical and administrative support). The average costs per day and total costs per patient were, respectively £111.81 and £6151.21 higher for patients who had sepsis. On a daily basis, several of the cost components were actually lower for the sepsis patients, i.e., consumables, medical imaging, medical and other staff. However, the significantly higher costs of the other areas of expenditure (drugs, fluids and nursing) ensured the overall greater daily cost. Most of the differences per patient can be attributed to the nursing expenditure, which at £2213.47 accounts for more than one third of the total difference in cost (£6151.21). Patients who had sepsis also tended to spend longer in the ICU than patients without sepsis, for example, with mean±standard deviation (SD) lengths of stay of 14.08±12.05 and 3.31±4.46 days respectively. (Table 5)

Cost-effectiveness of IMN

Table 6 demonstrates the estimates of the effect on introducing IMN on the costs of treating sepsis. The results (mean, 95% CI) of the meta-analysis were used to simulate how a reduction in the duration of sepsis would influence the cost per patient (as shown in Table 5 and including the additional cost of the treatment at £35 per day). This was calculated in several stages. First it was necessary to determine the adjusted duration of the infection using the formula: length of stay / 100 * reduction in infection (as a percentage). So, for example, using the mean size effect: 14.08 / 100 *40 = 5.63 non-sepsis days, with a remaining 8.45 days with sepsis. To calculate the adjusted cost (in accordance with this change in infection duration), the following formula was used: (no. of non-sepsis days * cost per non-sepsis day) + (no. of sepsis days * [cost per sepsis days + cost of feed]). For example, using the mean to generate a reduction in the cost per patient by £136.70, and a reduction of £24.28 / day of infection avoided (= £136.70 / 5.63 days). Although, when the lower bound of the 95% CIs is used to estimate a reduction in the duration of infection this actually increases the cost per patient by £272.53, or £138.34 per day of infection avoided. The most favourable cost-effectiveness ratio is created using the upper bound of the 95% confidence intervals, for example, per patient, the cost is reduced by £436.35, and by £52.51per day of infection avoided.

Cost-effectiveness of IMN (per 100 patients)

By applying the costs given in Table 6, the total cost for 100 patients, with 16.9% developing sepsis was calculated at £244621.43. Table 7 estimates the effects of introducing IMN based on the mean and 95% CI reduction in the duration of infection. In accordance with the cost per patient analysis, when IMN reduced infection by only 14%, this actually increased the cost of treating 100 patients by £4605.70. If a 40% reduction in infection was achieved the cost would be reduced by £2310.19 (at £242311.24), and if a 59% reduction was achieved the total cost would be reduced by £7109.75 (at £237511.68). (Table 7)

Sensitivity analysis

Table 8 illustrates the results of the sensitivity analysis using the lower and upper limits of the 95% CIs of costs to illustrate the variation in expenditure on different patients. This was in order to provide a more accurate representation of the clinical setting. When the lower bound is used, the introduction of IMN is always associated with an increased cost per patient, and the consequent per day of infection avoided. The application of the upper bound results in a reduction of £698.39 per patient and £124.05 per day of infection avoided (based on a mean reduction in infection duration). The most beneficial cost-effectiveness ratio is shown when a 59% reduction in infection duration occurs, i.e. a decrease of £1265.42 per patient and £152.28 per day of infection avoided.

 

DISCUSSION

This study has shown that the introduction of immune-enhancing enteral feed can be cost-effective through the reduction in the duration of infection in septic intensive care patients. The external validity of the study is limited by several assumptions in the analysis.

  1. As the measure of cost-effectiveness centres around the reduction in the duration of infection, it was necessary to determine the average length of ICU stay for patients with sepsis. The mean of 14.08 days was used as a baseline upon which to test the mean and 95% CI results of the meta-analysis. However, the 95% CI limits for length of stay at 10 and 18.15 days clearly demonstrates that many patients would actually spend considerably longer and shorter times in the ICU, and therefore use more or less resources than the current analysis may suggest.
  2. The sensitivity analysis which applied the 95% CI for costs represents an attempt to overcome this limitation. This also proved a valuable way to recognise the variability in the level of cost-effectiveness that using this treatment in this patient population would permit. Indeed, under several conditions the introduction of IMN actually resulted in greater costs, that is, when the IMN only produced a 14% reduction in infection and when the lower bound for costs were used.
  3. Another significant assumption in the analysis is based on the cost of the feed at an additional £35 per day compared to SEN. In the analysis it was necessary to work on average estimates. However, the assumption that the feed would be given at a maximum and equal dosage to all patients ignores one facet of the potential cost savings offered by IMN. Different patients would require, or be able to receive different amounts of feed, and this means that further cost savings could be achieved via those patients who were given lesser amounts of IMN.
  4. On a similar note, previous RCTs of IMN have demonstrated its ability to reduce ICU length of stay [9, 12, 14, 17-18]. Therefore, the assumption that patients receive IMN for the full 14 ICU days overlooks the possibility that they may leave the ICU earlier to produce additional cost savings.

The analysis focused on a reduction in the days of sepsis, however, this disregards the potential effect on reducing other infections, e.g. pneumonia. Furthermore, the meta-analyses from Beale and Heys both found that IMN was associated with a reduced length of stay in hospital [21-22]. As such, this also presents the opportunity for further cost savings, however, it was not possible to determine daily expenditure for these patients following their discharge from the ICU.

Each of the study's limitations highlights the need for detailed decision analysis to reflect the differences in patients' course of treatment, its' effectiveness and the associated expenditure. Decision analysis uses a 'decision tree' to illustrate the choices and pathways of health care [56]. A decision tree runs from left to right and begins with an initial decision or choice (depicted by a square) about a group of patients. This decision results in a variety of outcomes with a predetermined probability of occurrence (indicated by a circle) [48]. The decision analysis approach would be beneficial here, however, I did not feel equipped to undertake such a high level of analysis at the time of study completion. The completion of a sensitivity analysis which incorporated the 95% CI of costs represents a basic, yet valid attempt to account for uncertainty and variability in the data. The findings of this cost-effectiveness analysis support previous research [12] concluding that despite its additional cost, the clinical benefits of IMN ensure that it can be associated with reduced overall costs when compared to SEN (under certain conditions). One key point about IMN is that its' greatest benefits have been demonstrated in patients who are able to receive early enteral feeding of a certain amount [9]. Atkinson and colleagues found that the so-called early enteral feeders demonstrated a significant reduction in their requirements for mechanical ventilation, and duration of both ICU and hospital length of stay. This illustrates that the timing of the feed's administration and the amount of feed received are key to whether or not IMN can be cost-effective. However, the use of the results shown in a small group to form the basis of a conclusion or recommendation for a wider clinical setting should be done so with caution [39]. The generalizability of results demonstrated in a small number of patients is obviously limited, and could mask the negative effects on the group as a whole, i.e. a higher mortality rate in the group receiving IMN [9].

Despite this, the greatest benefits of IMN were also demonstrated in a specific group in the study from Galban and colleagues [11]. There was a decrease in mortality rate in the group of patients who were the least sick, i.e. with APACHE II (Acute Physiology and Chronic Health Evaluation) scores of less than 15. However, this difference was not observed in the sicker group of patients (APACHE II scores >15), which creates uncertainty about the effects of IMN on the sickest patients in the ICU [39]. However, the benefits of IMN on infection were demonstrated in the whole study population [11].

 

CONCLUSIONS AND RECOMMENDATIONS

The issues discussed above are significant as they further highlight the need for an economic evaluation of immunonutrition to be conducted alongside a RCT evaluating its clinical efficacy. This would overcome the problems relating to the number of assumptions that are inherent in this analysis. For example, the cost of the feed regarding the amount given to patients, an association between IMN and other infections could also be determined due to closer monitoring of the patients, and the patients could be followed up to hospital discharge in order to provide a cost-effectiveness ratio based on this outcome measure also. This would also assist in the close monitoring of patients in order to facilitate the completion of a decision analysis, which could reflect the numerous decisions, and corresponding outcomes that are associated with the care of the critically ill. The completion of a multi-centre RCT would also assist in the determination of a population who can safely receive the feed, to ensure that clinical as well as financial recommendations could be made accordingly.

So, in conclusion, this small-scale economic evaluation has been able to demonstrate that the use of IMN is cost-effective under certain conditions. The analysis shows that this outcome is dependent upon the achievement of a certain reduction in the duration of infection (i.e. more than 14%). In turn, this factor is dependent on the determination of a patient population who can receive the additional benefits that are necessary to warrant the additional expenditure on IMN.

 

REFERENCES

    1. Jolliet, P., Pichard, D. and Biolo, G. et al on behalf of the Working Group on Nutrition and Metabolism of the ESICM, Enteral nutrition in intensive care patients: a practical approach. Intensive Care Medicine, 1998. 24: p. 848-859.

    2. Jolliet, P. and Pichard, C., Immunonutrition in the critically ill. Intensive Care Medicine, 1999. 25: p. 631-633.

    3. Vincent, J., Bihari, D. and Suter, P. et al for the EPIC International Advisory Committee, The prevalence of nosocomial infection in the intensive care units of Europe: Results of the European Prevalence of Infection in Intensive Care (EPIC) study. JAMA, 1995. 274: p. 639-644.

    4. Eiseman, B., Beart, R. and Norton, L.,  Multiple Organ Failure. Surgery, Gynecology and Obstetrics, 1977. 144: p. 323-326.

    5. Fry, D., Pearlstein, L. and Fulton, R. et al, Multiple system organ failure: the role of uncontrolled infection. Archives of Surgery, 1980. 115: p. 136-140.

    6. Moore, F., Feliciano, D. and Andrassy, R. et al, Early enteral feeding, compared with parenteral, reduces septic post-operative complications: The results of a meta-analysis. Annals of Surgery, 1992. 216: p. 172-183.

    7. DeLegge, M., Enteral access - the foundation of feeding. Journal of Parenteral and Enteral Nutrition, 2001. 25(2): p. 8-13.

    8. Barbul, A., Immunonutrition comes of age. Critical Care Medicine, 2000. 28(3): p. 884-885.

    9. Atkinson, S., Sieffert, E. and Bihari, D. on behalf of the Guy's Hospital Intensive Care Group, A prospective randomised double-blind clinical trial of enteral immunonutrition in the critically ill. Critical Care Medicine, 1998. 26: p. 1164-1172.

    10. Bower, R., Cerra, F. and Bershadsky, B. et al, Early enteral administration of a formula (IMPACT) supplemented with arginine, nucleotides, and fish oil in intensive care unit patients: Results of a multicenter, prospective, randomised, clinical trial. Critical Care Medicine, 1995. 23: p. 436-449.

    11. Galban, C., Montejo, J. and Mesejo, A. et al, An immune-enhancing enteral diet reduces mortality and episodes of bacteremia in septic intensive care unit patients. Critical Care Medicine, 2000. 28(3): p. 643-648.

    12. Senkal, M., Mumme, A. and Eickhoff, U. et al, Early postoperative enteral immunonutrition: Clinical outcome and cost-comparison analysis in surgical patients. Critical Care Medicine, 1997. 25: p. 1489-1496.

    13. Braga, M., Gianotti, L. and Vignali, A. et al, Artificial nutrition after major abdominal surgery: Impact of route of administration and composition of the diet. Critical Care Medicine, 1998. 26: p. 26-30.

    14. Moore, F., Moore, E. and Kudsk, K. et al, Clinical benefits of an immune-enhancing diet for early postinjury enteral feeding. Journal of Trauma, 1994. 37(4): p. 607-615.

    15. Daly, J., Lieberman, M. and Goldfine, J. et al, Enteral nutrition with supplemental arginine, RNA and omega-3 fatty acids in patients after operation: Immunologic, metabolic and clinical outcome. Surgery, 1992. 112: p. 56-67.

    16. Daly, J., Weintraub, F. and Shou, J. et al, Enteral nutrition during multimodality therapy in upper gastrointestinal cancer patients. Annals of Surgery, 1995. 221: p. 327-338.

    17. Kudsk, K., Minard, G. and Croce, M. et al, A randomised trial of isonitrogeous enteral diets after severe trauma. Annals of Surgery, 1996. 224: p. 531-540.

    18. Weimann, A., Bastian, L. and Bischoff, W. et al, Influence of arginine, omega-3 fatty acids and nucleotide-supplemented enteral support on systemic inflammatory response syndrome and multiple organ failure in patients after severe trauma. Nutrition, 1998. 14(2): p. 165-172.

    19. Cerra, F., Lehman, S. and Konstantinides, N. et al, Effect of enteral nutrient on in vitro tests of immune function in ICU patients: A preliminary report. Nutrition, 1990. 6: p. 84-87.

    20. Schilling, J., Vranjes, N. and Fierz, W. et al, Clinical outcome and immunology of postoperative arginine, omega-3 fatty acids and nucleotide-enriched enteral feeding: A randomised prospective comparison with standard enteral and low calorie/low fat solutions. Nutrition, 1996. 12: p. 423-429.

    21. Beale, R., Bryg, D. and Bihari, D., Immunonutrition in the critically ill: A systematic review on clinical outcome. Critical Care Medicine, 1999. 27: p. 2799-2805.

    22. Heys, S., Walker, L. Smith, I. and Eremin, O., Enteral nutritional supplementation with key nutrients in patients with critical illness and cancer: a meta-analysis of randomised controlled clinical trials. Annals of Surgery, 1999. 229(4): p. 467-477.

    23. Seifter, D., Rettura, G., Berbul, A. and Levenson, S., Arginine: an essential amino acid for injured rats. Surgery, 1978. 84(2): p. 224-230.

    24. Sitren, H. and Fisher, H., Nitrogen retention in rats fed on diets enriched with arginine and glycine. 1 Improved N retention after trauma. British Journal of Nutrition, 1977. 37(2): p. 195-208.

    25. Cerra, F., Nutrient modulation of inflammatory and immune function. American Journal of Surgery, 1991. 161: p. 230-234.

    26. Schloerb, P., Immune-enhancing diets: products, components, and their rationales. Journal of Parenteral and Enteral Nutrition, 2001. 25(2): p. 3-7.

    27. Ortega, M., Nunez, M., Gil, A. and Sanchez-Pozo, A., Dietary nucleotides acclerate intestinal recovery after food deprivation in old rats. J Nutr, 1995. 125: p. 1413-1418.

    28. Gentilini, O., Braga, M. and Gianotti, L., Rational base and clinical results of immunonutrition. Minerva Anestesiol, 2000. 66(5): p. 362-369.

    29. Zaloga, G., Immune-enhancing enteral diets: Where's the beef? Critical Care Medicine, 1998. 26(7): p. 1143-1146.

    30. Braga, M., Vignali, A. and Gianotti, L. et al, Benefits of early postoperative enteral feeding in cancer patients. Infusionsther Transfusionsmed, 1995. 22: p. 280-284.

    31. Cerra, F., Lehmann, S. and Konstantinides, N. et al, Improvement in immune function in ICU patients by enteral nutrition supplemented with arginine, RNA and menhaden oil is independent of nitrogen balance. Nutrition, 1991. 7(3): p. 193-199.

    32. Heslin, M., Latkany, L. and Leung, D. et al, A prospective randomised trial of early enteral feeding after resection of upper gastrointestinal malignancy. Annals of Surgery, 1997. 226(4): p. 567-580.

    33. Kemen, M., Senkal, M. and Homann, H. et al, Early postoperative enteral nutrition with arginine-w-3 fatty acids and ribonucleic acid-supplemented diet versus placebo in cancer patients: An immunologic evaluation of Impact. Critical Care Medicine, 1995. 23(4): p. 652-659.

    34. Mendez, C., Jurkovich, G. and Garcia, I. et al, Effects of an immune-enhancing diet in critically injured patients. The Journal of Trauma: Injury, Infection and Critical Care, 1997. 42(5): p. 933-940.

    35. Saffle, J., Wiebke, G. and Jennings, K. et al, Randomized trial of immune-enhancing enteral nutrition in burn patients. The Journal of Trauma: Injury, Infection and Critical Care, 1997. 42(5): p. 793-800.

    36. Beale, R., Immunonutrition in the ICU. International Journal of Intensive Care, 1999. Spring.

    37. Heyland, D., Cook, D. and Guyatt, G., Does the formulation of enteral feeding products influence infectious morbidity and mortality rates in the critically ill patient? A critical review of the evidence. Critical Care Medicine, 1994. 22(7): p. 1192-1202.

    38. Heyland, D. and Drover, J., Does immunonutrition make an impact? It depends on the analysis. Critical Care Medicine, 2000. 28(3): p. 906-7.

    39. Heyland, D. and Novak, F., Immunonutrition in the critically ill patient: more harm than good. Journal of Parenteral and Enteral Nutrition, 2001. 25(2): p. 51-56.

    40. Heys, S., Recent advances in intensive care. More still needs to be known about immunonutrition. British Medical Journal, 2000. 320: p. 1669.

    41. Moore, F., Effects of immune-enhancing diets on infectious morbidity and multiple organ failure. Journal of Parenteral and Enteral Nutrition, 2001. 25(2): p. 36-43.

    42. Edbrooke, D., Hibbert, C. and Kingsley, J. et al, The patient-related costs of care for sepsis patients in a United Kingdom adult general intensive care unit. Critical Care Medicine, 1999. 27(9): p. 1760-1767.

    43. Beck, C., Facilitating the work of a meta-analyst. Research in Nursing and Health, 1999. 22: p. 523-530.

    44. Sim, J. and Wright, C., Research in health care. Concepts, designs and methods. 2000: Stanley Thornes (Publishers) Ltd.

    45. Hasselmann, M., Kummerlen, C. and Bryg, D., Enteral feeding supplemented with arginine, fish oil and nucleotides (IMPACT) reduces infection rate and improves immunitary status in critically ill patients. Abstract. Intensive Care Medicine, 1997. 23: p. 136.

    46. Lieberman, M., Shou, J. and Torres, A. et al, Effects of nutrient substrates on immune function. Nutrition, 1990. 6: p. 88-91.

    47. The Standards of Reporting Trials Group, A proposal for structured reporting of randomized controlled trials. JAMA, 1994. 272: p. 1926-1931.

    48. Drummond, M., O'Brien, B., Stoddart, G. and Torrance, G., Methods for the economic evaluation of health care programmes. 2nd ed. 1997: Oxford University Press.

    49. Sasse, K., Nauenberg, E. and Long, A. et al, Long-term survival after intensive care unit admission with sepsis. Critical Care Medicine, 1995. 23: p. 1040-1047.

    50. Brun-Buisson, C., Doyon, F. and Carlet, J. et al, Incidence, risk factors, and outcome or severe sepsis, and septic shock in adults: A multicenter prospective study in intensive care. JAMA, 1995. 274: p. 968-974.

    51. Le-Gall, J.-R., Lemeshow, S. and Leleu, G. et al, Customised probability models for early severe sepsis in adult intensive care patients. JAMA, 1995. 273: p. 644-650.

    52. Chalfin, D. and Fein, A., The economic impact and cost-effectiveness of monoclonal antibody therapy for sepsis. Clinical Infectious Diseases, 1993. 17 (Suppl 2): p. 520-524.

    53. Abraham, E., Glauser, M. and Butler, T. et al, P55 tumor necrosis factor receptor fusion protein in the treatment of patients with severe sepsis and septic shock: A randomised controlled multi-center trial. JAMA, 1997. 277: p. 1531-1538.

    54. Sands, K., Bates, D. and Lanken, P. et al, Epidemiology of sepsis syndrome in 8 academic medical centers. JAMA, 1997. 278: p. 234-240.

    55. Edbrooke, D., Stevens, V. and Hibbert, C. et al, A new method of accurately identifying costs of individual patients in intensive care: the initial results. Intensive Care Medicine, 1997. 23: p. 645-650.

    56. Jefferson, T., Demicheli, V. and Mugford, M.,  Elementary economic evaluation in health care. 2nd ed. 2000, London: BMJ Books.