Perioperative nutritional plan. Patients were given the study diet throughout the perioperative period (5 days before surgery and 5 days after surgery). Preoperatively, patients received the study diet (1000 mL/d) orally. After the operation, patients received the enteral diets via an intraoperatively placed needle-catheter jejunostomy, starting on the first postoperative day. The postoperative oral intake started on postoperative day 5, usually with clear liquids. Intravenous fluids were given as clinically indicated. OP indicates the day of the operation.
Preoperative and postoperative serum levels of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) were significantly higher in the experimental group (ie, the immune-enhancing feed [IEF] group) compared with the control (CON) group. The asterisk indicates P<.05; bars, SEM. The levels of EPA and DHA were measured by gas chromatography.
Patient population and onset of complications in the group receiving immune-enhancing feed (IEF) and in the control (CON) group. The number of patients with complications after postoperative day 3 was significantly higher in the CON group. The asterisk indicates P<.04 (χ2 test).
Senkal M, Zumtobel V, Bauer K, Marpe B, Wolfram G, Frei A, Eickhoff U, Kemen M. Outcome and Cost-effectiveness of Perioperative Enteral Immunonutrition in Patients Undergoing Elective Upper Gastrointestinal Tract SurgeryA Prospective Randomized Study. Arch Surg. 1999;134(12):1309-1316. doi:10.1001/archsurg.134.12.1309
Perioperatively administered enteral immunonutrition will improve early postoperative morbidity and cost-effectiveness after gastrointestinal tract surgery.
A prospective, randomized, double-blind, multicenter clinical trial.
Surgical departments in German university and teaching hospitals.
One hundred fifty-four patients with upper gastrointestinal tract malignant neoplasms who were eligible for analysis.
Preoperatively, patients received 5 days of oral immunonutrition (an arginine-, RNA-, and ω3 fatty acid–supplemented diet) or an isoenergetic control diet (1 L/d). Early postoperative enteral feeding with immunonutrition or an isoenergetic, isonitrogenous control diet using a catheter jejunostomy was performed for 10 days.
Main Outcome Measures
Postoperative infectious complications, their treatment costs, and cost-effectiveness of immunonutrition were analyzed. Plasma levels of the fatty acids eicosapentaenoic acid and docosahexaenoic acid were measured.
In the immunonutrition group, significantly fewer infectious complication events occurred (14 vs 27; P = .05). The number of patients with complications was significantly lower in the supplemented diet group after postoperative day 3 (7 vs 16; P = .04). The treatment costs of complications in the supplemented diet group were suggestively lower than in the control diet group (DM 75 172 vs DM 204 273). Cost-effectiveness was DM 1503 in the experimental group vs DM 3587 in the control group, where DM denotes deutsche mark (German currency).
The perioperative administration of an enteral immunonutrition significantly (P = .05) decreased the early occurrence of postoperative infections and reduced substantially the treatment costs of the complications after major upper gastrointestinal tract surgery.
TRAUMA AND surgery may induce severe alterations of the immune system that result in a higher rate of postoperative inflammatory and septic complications.1 The sequence of proinflammatory cascades leading to an immune paralysis in the first postoperative days starts at the time of surgical injury. Many investigations2 in enteral nutrition have focused on the ability to modulate the immune response to injury via specially formulated enteral diets. Immunonutrition using nutrients such as arginine, RNA, and ω3 fatty acids (FAs) (fish oil) has been used to alter eicosanoid synthesis, cytokine production, and immune function in an attempt to limit the undesired immune reactions following surgical injury.3,4
However, even aggressively performed early postoperative immunonutrition is not able to prevent immunosuppression within the first postoperative week. In a previous postoperative study of patients with cancer who were fed an immune-enhancing diet, Daly et al3 observed no effect of immunonutrition on the first postoperative day and a significant improvement as of day 7. Similarly, in a postoperative study5 from our group in patients who underwent surgery for upper gastrointestinal tract cancer, early postoperative immunonutrition resulted in significant advantage of humoral and cellular immunity compared with the control feed after 1 week of treatment. These beneficial immunological changes were associated with an overall reduction of infectious complications in the extended clinical outcome study, but the improved outcome was not significant before 5 days after surgery.6 The application of immunonutrition can only ameliorate but not reverse the catabolic and immunological response to trauma. Therefore, starting the immunonutrition in the preoperative period is probably most effective by improving the nutritional state and reducing the postoperative complications.
Braga et al7 found improved gut function and positively modulated postsurgical immunosuppressive and inflammatory responses after perioperative administration of enteral immunonutrition. The same group has also shown that postoperative complications may further be reduced by pretreatment of patients who are supposed to undergo elective gastrointestinal tract surgery with an immune-enhancing formula.8
These clinical results suggest that perioperative immunonutrition not only has a positive clinical impact on the patients but also may transform into substantial reductions in resource use for treating postoperative complications. While the provision of a supplemented diet increases the costs of clinical nutrition of these patients, reductions of the incidence of postoperative complications lead to savings in costs of treating these complications. These savings may compensate or even overcompensate the additional costs of immunonutrition.
The hypothesis of this clinical trial was that the administration of immunonutrition should be given preoperatively and in the early postoperative period to improve the postoperative clinical outcome, suggesting that the deterioration of immune variables observed during the postoperative phase will be counteracted early enough to bring additional advantage in this patient population. To our knowledge, we showed for the first time that the choice of feed can influence postoperative treatment costs in major surgery.
The study was a randomized, double-blind, prospective trial that was carried out between April 1, 1994, and August 30, 1997, in the Departments of Surgery of the Ruhr-University Bochum, St Josef Hospital, and Marienhospital, all located in Bochum, Germany. Other participating departments were the Department of Anaesthesiology, Evangelisches Krankenhaus Hattingen, Hattingen, Germany; and the Department of Surgery, Knappschaftskrankenhaus Bochum, Bochum. Written informed consent was obtained from each patient following approval by the Human Investigation Committee of the Ruhr-University Bochum.
One hundred fifty-four consecutive patients aged between 18 and 80 years who were supposed to undergo elective upper gastrointestinal tract surgery for histologically documented malignant tumors of the upper gastrointestinal tract were enrolled into the study. Furthermore, a minimum uptake of 3000 mL of study diet preoperatively was required to enroll into the study. Exclusion criteria were the presence of endocrine and metabolic disorders; any known allergic diseases; hemorrhagic diathesis; sepsis; a preexisting severe chronic disease, including nephric or liver failure (serum creatinine level <221 µmol/L [<2.5 mg/dL] and total serum bilirubin levels <51 µmol/L [<3.0 mg/dL]); congestive heart failure; and a regimen of immunosuppressive or chemotherapeutic medication and radiotherapy within the last 3 months before the trial. In addition, drug abuse, emergency surgery, inadequate preoperative preparation, and use of medications known to affect eicosanoid metabolism during the last 2 weeks before the trial were among the exclusion criteria.
The patients were randomized and divided into the experimental supplement group, immune-enhancing feed [IEF], which was arginine, RNA, and ω3 FA enriched (IMPACT; Novartis, Bern, Switzerland), and the control (CON) group (isoenergetic). They were fed the preoperative study diet at least 5 days before surgery. Patients received freshly mixed diet, 1000 mL/d, in 250-mL portions in addition to the usual hospital diet. They were able to choose between 3 different flavors. The composition of the preoperatively and postoperatively administered diets is described in Table 1. In the postoperative period, the study diet was given for at least 5 days before the administration of the standard enteral formula.
On the day of study entrance (day 5), a complete history was obtained from and a physical examination was performed on all patients. Also, laboratory studies were performed 5 days preoperatively, on the day of surgery, and on postoperative days 1, 5, and 10. Venous blood was obtained and analyzed for liver function tests and for the determination of levels of creatinine, serum urea nitrogen, triglyceride, total protein, albumin, and electrolytes. A complete blood cell count and prothrombin and partial thromboplastin times were also obtained. Blood glucose levels were determined either in arterial or in capillary blood.
All patients underwent surgical procedures as indicated by their primary site and stage of disease (Table 2). Single-dose antibiotics (cephalosporin N) were given intravenously to each patient before the surgery. Intravenous blood and fluid replacement was continued postoperatively as clinically required. Patients did not receive immunoglobulins.
The postoperative enteral feeding was started 12 hours after surgery using continuous infusion via an intraoperatively placed needle-catheter jejunostomy. Enteral feeding started at 20 mL/h on the first postoperative day and progressed to the optimal goal (80 mL/h) by the fifth postoperative day (Figure 1). The diet plan was intended to provide approximately 1.05 × 105 J/kg per day.
Oral intake was allowed as clinically indicated between the fifth and seventh postoperative days and started with clear liquids. All patients received intravenous fluids (5% dextrose and isotonic sodium chloride solution) and other electrolytes as clinically indicated. None of the patients received parenteral nutrition before or after surgery.
Clinical assessment was performed and recorded daily, and the body weight was assessed twice weekly. Adverse gastrointestinal tract symptoms were recorded daily, ie, any nausea or vomiting that required antiemetics, diarrhea of 3 loose bowel movements per 24 hours, or abdominal cramping. If any of these symptoms occurred in a moderate to severe fashion, infusion was discontinued for 6 to 12 hours and started again at the next lower infusion rate.
Primary outcome has been defined a priori as infectious complications after postoperative day 3 or after postoperative day 5. All patients were monitored on a daily basis for the occurrence of postoperative complications. Postoperative complications were recorded prospectively during the patient's hospital course, and they were defined as septic, wound, pulmonary, or urinary complications. The definitions of complications were given according to the criteria established by the Centers for Disease Control and Prevention and to the definitions proposed by the American College of Chest Physicians and the Society of Critical Care Medicine at the Consensus Conference in August 1991.9 All complications following surgery were recorded until the patient was discharged from the hospital. When complications occurred, they were treated adequately with either antibiotics or surgical revision. Deaths were recorded when they occurred before discharge. The secondary outcomes included hospital length of stay (LOS) and treatment costs of complications. Length of stay was defined as the number of days from admission to the surgical unit to the actual date of hospital discharge.
The analysis of plasma lipid composition was done by liquid gas chromatography. Briefly, 10 mL of blood was drawn, and the separated serum was frozen at −20°C. Lipid extraction from serum was performed. Phospholipids of serum and tissues were separated by thin-layer chromatography using a combination of petrolether, ethylmethylketone, and acetic acid (84:15:1, vol/vol/vol) as a solvent. Fatty acid methyl esters were obtained by transesterification with trimethylsulphonium hydroxide and separated by use of a capillary column (model CP-Sil-88; Chrompack, Frankfurt, Germany), installed in a gas chromatograph with a flame ionization detector (model HP 5890; Hewlett Packard, Taufkirchen, Germany). For the separation and identification of the FA methyl ester peaks, standard reagents of analytical grade were used.
The data are expressed as the mean ± SD. Statistical analysis was performed by analysis of variance with repeated measures. All given P values are 2 sided. Significance testing for the univariate, between-group differences was performed using, where appropriate, the χ2 test, the Fisher exact test, the Wilcoxon 2-sample test, and the t test. The data analysis was performed on SAS statistical software, version 6.0 (SAS Institute Inc, Cary, NC).
Treatment group balance was assessed by using exact contingency tables for discrete outcomes such as sex, diagnosis, and surgery type, and t tests were used for continuous measures such as baseline nutritional levels. Data on LOS were compared by using t tests in the absence of censoring. A value of P = .05 was required for statistical significance.
The costs of clinical nutrition, the costs of treating postoperative complications, and the effectiveness of clinical nutrition were assessed in the supplemented and the CON groups. Based on these data, cost-comparison and cost-effectiveness analyses were carried out.
The economic perspective taken was that of the hospital in which the clinical study had been carried out. This is a 1000-bed university hospital.
Effectiveness was defined as the percentage of complication-free patients.
The costs of clinical nutrition were separately calculated for the preoperative and postoperative phases. The average intake per patient was multiplied by the price per liter. These costs were added, summed up over all patients, and divided by group size.
Included were direct medical costs of treating postoperative complications during the hospital stay. Indirect costs (eg, loss of productivity by the patient) were not taken into account.
To assess the cost per complication, 2 kinds of data are needed: resource use, ie, the type and frequency of goods and services rendered to the patients on the one side, and the monetary values (prices and unit costs) for each type of these variables on the other side. The following goods and services rendered to the patients specifically for diagnosis and treatment of postoperative complications were derived from the medical records of each clinical study patient who had developed a postoperative complication: (1) operative interventions and anesthetic procedures; (2) other complication-related physicians' consultations and services; (3) artificial respiration, nutritional support, and dialysis; (4) radiography and ultrasonography; (5) physical therapy; and (6) antibiotics and other complication-related medications. Prolonged hospital LOS and days in the intensive care unit were not included because these resources could not be unambiguously attributed to the postoperative complications. Diagnostic and therapeutic services were valued according to the tariff of the German hospital association. Pharmaceuticals, devices, and material costs were valued at hospital buying prices.
In a cost comparison analysis, the per-patient costs of clinical nutrition and treating complications were added for the study and the CON groups and finally compared with one another. Cost-effectiveness analysis was carried out by dividing the per-patient costs of clinical nutrition and treating postoperative complications with the percentage of complication-free patients.
All costs are in deutsche mark (DM) (German currency) values for the year 1998. At the end of December 1998, the conversion rate from deutsche marks to US dollars was DM 1.00 to US $0.5967.
Patients (n = 178) scheduled for upper gastrointestinal tract surgery were randomized into the trial (Table 2). Twenty-four patients were excluded for lack of compliance with taking the preoperative supplement, intolerance of postoperative feed, withdrawal of consent, or protocol violation. Also, inadvertent removal of the jejunostomy occurred in 3 patients. One hundred fifty-four patients completed the study. The number of patients in each group, the age ranges, the preoperative diagnoses, and the operative procedures were similar in both groups. At the point of entrance into the study, nutritional status, type of surgical pathological characteristics, and mean percentage of usual body weight, mean operative duration or number of patients who received perioperative transfusions were similar in both groups (Table 2).
Gastrointestinal tract tolerance of both formula diets was excellent in both groups, and no dropouts occurred because of intolerance.
Preoperative and postoperative serum levels of the ω3 FAs, eicosapentaenoic acid and docosahexaenoic acid, are illustrated in Figure 2. At the end of the preoperative supplementation period, ω3 FA levels, expressed in percentage of total FAs ± SEM, were significantly higher in the group receiving IEF compared with the CON group (eicosapentaenoic acid, 0.66% ± 0.17% vs 0.25% ± 0.08% [P = .04]; docosahexaenoic acid, 1.88% ± 0.41% vs 1.07% ± 0.36% [P = .04]). Up to postoperative day 10, significantly higher serum levels were maintained in the IEF-supplemented group (eicosapentaenoic acid, 0.66% ± 0.16% vs 0.18% ± 0.03% [P = .02]; docosahexaenoic acid, 2.30% ± 0.46% vs 0.61% ± 0.07% [P = .005]).
The number and timing of complications are displayed in Table 3 and Figure 3. Of the 154 patients who completed the study, 28 (18.2%) had postoperative complications. In the IEF group, fewer patients experienced postoperative complications compared with the CON group. After postoperative day 3, the number of patients who developed complications was significantly lower in the IEF group than in the CON group. The number of patients who had complications after the fifth postoperative day was suggestively lower in the IEF group compared with the CON group.
The prevalence of complication events was also significantly lower in the IEF group compared with the CON diet group.
Per-patient nutrition costs amounted to DM 347 in the supplemented group and DM 49 in the CON group.
Per-patient costs of treating postoperative complications were DM 964 in the IEF group and DM 2688 in the CON group. These cost components sum to DM 1311 per supplemented patient and to DM 2737 per CON patient, which transforms into net savings of DM 1426 per patient in favor of immunonutrition.
The type and costs of each complication that occurred in both diet groups are listed in detail in Table 4. The costs in the IEF group were less than half of the costs as in the CON group.
The beneficial effects of early postoperative enteral immunonutrition on the immune system of patients after surgery have been demonstrated in several studies.10 In most studies, however, the positive changes in the inflammation cascade did not become apparent until about 1 week after surgery. It seems, therefore, that even early and aggressive application of enteral immunonutrition after surgery is not sufficient to prevent the immunosuppression that takes place in the early postoperative period. In a previous postoperative study5 of patients with cancer who were fed an immune-enhancing diet, a significant improvement of humoral and cellular immune variables was not observed until 1 week of treatment. By contrast, perioperative administration of an ω3 FA–, arginine-, and RNA-nucleotides–enriched enteral formula for 7 days preoperatively and in the postoperative period significantly improved postsurgical immunosuppressive and inflammatory responses and gut function.7
Clinical outcome following postoperative feeding of immunonutrition has been investigated in several prospective randomized clinical trials involving patients undergoing major surgery for gastrointestinal tract cancer.3,6,11- 14 With the exception of the trials by Daly et al,3,12 no significant reduction in postoperative infectious complications was reported by early postoperative enteral immunonutrition, although a trend toward a decreased rate of postoperative infections under immunonutrition was found in most of the studies.6,13 In our previous multicenter study,6 we found a nonsignificant reduction in the overall postoperative complication rate in the patient group receiving immunonutrition vs the group receiving isoenergetic, isonitrogenous CON formula (22% vs 31%). Looking at the occurrence of late postoperative complications after the fifth postoperative day, we observed significantly fewer patients with infectious complications under immunonutrition than in the CON group. These results are consistent with the hypothesis that some days of feeding with immunonutrition are required to obtain a beneficial effect on host defense and to subsequently result in a reduction of postoperative infections.
From these findings, the question arose whether prefeeding an immunonutrition formula before the elective surgery would bring a significant amelioration of the postoperative immune depression and a concomittant reduction in infectious complications. We addressed this question in the present study by preoperatively administrating an oral immunonutrition supplement for 5 days compared with an isoenergetic CON supplement to patients undergoing major surgery for upper gastrointestinal tract cancer.
In a prospective, randomized, clinical study8 with perioperative enteral immunonutrition, a group from Italy reported clinical outcome data. In this study, a 50% reduction of postoperative infectious complications could be achieved using the same immune-enhancing product as used in our present trial. Our perioperative nutritional intervention on patients with upper gastrointestinal tract cancer who were undergoing major surgery corroborates the findings of this group,8 resulting as well in almost 50% fewer patients with postoperative complications.
Assuming that the proinflammatory cascade triggered by the surgical intervention leads to an immune paralysis immediately after surgery, and that therefore postoperative nutritional intervention comes too late to protect against ischemia-reperfusion injury and its immediate consequences, preoperative immune modulation with ω3 FA and arginine should offset some of the immune depression by controlling the inflammatory response. We have earlier described that this is possible by shifting the lipid mediator profile by means of an ω3 FA incorporation toward the less active leukotrienes LTB5, using the same product in a similar patient population.15 Serum levels of eicosapentaenoic acid and docosahexaenoic acid in Figure 2 demonstrate that higher-circulating ω3 FAs were obtained after preoperative supplementation with the fish oil–containing IEF, already evident on the day of surgery.
Finally, in our own group, we have demonstrated that preoperative immunonutrition is superior to a standard supplement and is able to offset the early postoperative immune depression in patients undergoing major abdominal surgery, as shown by significantly increased immune variables such as T lymphocytes and IgM levels.16
Clinically indeed, in the present study, postoperative complications seem to be influenced beneficially after postoperative day 3, compared with our early postoperative intervention study,6 where it was only after postoperative day 5.
Although immunonutrition increases the nutrition costs of these patients, this is overcompensated by savings in subsequent treatment cost in those patients who experienced postoperative complications. Immunonutrition is not only cost-effective in costs per complication-free patient but generates net savings and thus is the dominant strategy. These findings are consistent with evidence that early postoperative immunonutrition generates savings of treatment costs6,17 compared with early postoperative enteral nutrition only. Although the complication costs in the IEF group were less than half of the costs as in the CON group, there was no statistically significant difference. This is due to the variation of the costs as they are listed in Table 4. Costs that are based on resource use vary in patients even with similar complications. In such situations, bigger study groups would be necessary to receive statistically significant differences between the groups. Our results represent conservative estimations of the reductions in complication-related costs, as important cost components such as prolonged hospital LOS and extra days in the intensive care unit were not included. These resources could not be unambiguously attributed to the postoperative complications.
In this study, we compared the perioperative application of an arginine-, ω3 FA–, and RNA-nucleotides–supplemented diet with a CON diet lacking these substrates. A significant (P = .04) reduction of early postoperative infectious complications could be achieved with the immunonutrition formula compared with the CON diet. The hospital LOS and the costs for complication treatment were substantially reduced. Finally, despite higher product costs, the cost-effectiveness of immunonutrition was better than in the CON group.
Presented at the 19th Annual Meeting of the Surgical Infection Society, Seattle, Wash, April 30, 1999.
We thank Heinz Schneider, PhD, and Adrian Heini, MD, of Novartis Nutrition, Bern, Switzerland, Roger Greiner, MPH, of HealthEcon AG, Basel, Switzerland, Eva Brune, MD, of Marienhospital, Herne, Germany, and Andreas Kampa of Evangelisches Krankenhaus, Hattingen, Germany, for their help.
Reprints: Metin Senkal, MD, Department of Surgery, Ruhr-University Bochum, St Josef Hospital, Gudrunstr 56, 44791 Bochum, Germany.
J. Wesley Alexander, MD, Cincinnati, Ohio: Clinical studies of this type are very difficult to do, as everyone realizes, and I think that this was done about as well as could be done. The patients were randomized appropriately, and appropriate measures were given to make them comparable. The results are fairly clear.
This study correlates very well with one that was given last year by Dr Braga and his colleagues from Italy and published just this month in the Archives of Surgery. The results, I think, are pretty near the same. The studies show about a 50% reduction in the infection rate, and this is also consistent with the other 15 or 20 published studies that are well controlled, prospective, and largely double-blinded that have looked at different types of immunonutrient diets related to complication rates and infections. These have all been associated with a reduction in the infection rate, except for 2.
One of those was done by Dr Jeff Saffle, and he actually compared 2 immunonutrient diets rather than comparing IMPACT with an appropriate control without immunonutrients. The other one was from a cancer study that I think related to a better-nourished group of patients rather than failure of the diet itself.
The best results with immunonutrient diets are clearly with early enteral feeding after trauma and with a prefeeding for elective surgery. Better results have been reported in malnourished vs well-nourished patients. In the studies where it has been looked at, there has also been a cost benefit.
This study was done with the new IMPACT formula, which is a powdered formula, and contains a whey protein rather than casein, so it may be a better formula than the previous one that was available. In addition, the liquid formula tastes terrible, and the new product is a relatively palatable formula. Thus, one question I would like to ask you is how was compliance with this formula? Were there any complaints of the patients taking it over a long period of time?
The second question I would like to ask is were there any subgroups in which there seemed to be any particular benefit, eg, such as the older patient compared to the younger patient or those who were malnourished compared to those who were not malnourished?
The third question I would have for you is what do you think are the mechanisms? I think you may have alluded to this already, and certainly Dr Williams' study that was presented just before this suggested that it may not entirely be related to improvement of the immune mechanisms for killing bacteria. But perhaps as important and maybe even more important are down-regulations of the exuberant inflammatory or discordant inflammatory responses that can occur after injury and surgery.
There is another paradox that we have in the transplant population in that these same immunonutrients, that is, ω3 fatty acids and arginine, will actually improve allograft survival. Certainly, this has nothing to do with an up-regulation of immune responses, but probably there is a nonspecific down-regulation of the inflammatory responses, particularly associated with a shift from the Th+1 to the Th+2 response.
Finally, would you comment on the potential relative importance of DHA [docosahexaenoic acid] compared to EPA [eicosapentaenoic acid]. DHA is an extremely important lipid that is incorporated into brain tissues, particularly of young individuals; perhaps as much as 10% of the lipids in the brain are from DHA, and even more in the retina. In your studies that you have just presented, there was seemingly more of an increase, or more of a difference in the increase, in DHA than EPA.
Dr Senkal: The first question referred to the compliance of the patients, whether they had problems to drink that diet. We had 4 different flavors for that study, and we did not notice actually problems with compliance, and most patients were actually drinking that diet pretty well, maybe because we told them that it would be good for them to drink that diet, I don't know, but they did pretty well on that.
Let me answer the third question. Actually, it was about the alteration of the immune system. I agree with you that probably it is much more important to down-regulate the inflammatory response after operation than to kill bacteria in the early postoperative period. This is probably the main mechanism by which those diets are working.
The second question was actually if we looked at different types of patients, whether or not there were differences in older or younger patients or well nourished or malnourished. No, we did not look at this specific issue. Most of our patients were actually well nourished, and we did not compare different subgroups with each other.
Nicholas Namias, MD, Miami, Fla: I enjoyed your paper and well-designed study and a large effort, but I have some concerns about the conclusions you drew from the statistics. First of all, the total number of patients with complications showed no statistically significant difference, but you found a subset with a difference, which would mean that it is just sort of shifting around. If the total set has no difference and 1 subset has a difference, then the other subset must also have a difference. So did you have more early infectious complications in the patients who did get the immunonutrition? How would you explain that 1 subset has a difference, while the overall set doesn't?
Secondly, you mentioned in the "Conclusions" in the abstract that the length of stay was significantly shorter in the immunonutrition group; however, on your slides you showed that the addition was statistically not significant. You also said that the cost is significantly less with the immunonutrition, but in the abstract you say that statistically the cost difference was not significant. So what test was applied to measure the cost difference?
Again, I enjoyed your paper because it supports my own personal bias, but when I go to argue with people who buy things for the formulary and point to the paper and show them differences that are not statistically significant, I will have a hard time.
Dr Senkal: Well, the length of hospital stay was 3 days shorter in the group receiving the immunonutrition, and this was substantial but did not reach statistical significance. And these are the final results I am presenting here. What I have written in the abstract were primary results, actually.
The second question about the costs, again, those were DM 75 000 vs DM 200 000, over DM 200 000, and the differences varied that much that a statistical significance could not be reached, although when you compare both figures, there is a substantial reduction in the treatment costs.
Looking at the total complication costs, again, here we had 10 vs 18 patients with complications. The complication events by themselves, 14 vs 27, were significantly different, but the occurrence, the complication rate, the incidence rate of postoperative complications was substantial but not statistically significant, but when we applied this a priori to say whether or not complications occurred after the third postoperative day, which we defined as early complications, we have seen a statistical significance using the χ2 tests.