Retracted: Early supplementation of parenteral nutrition is capable of improving quality of life, chemotherapy-related toxicity and body composition in patients with advanced colorectal carcinoma undergoing palliative treatment: results from a prospective, randomized clinical trial



This article corrects:

  1. Retraction Volume 13, Issue 12, 1447, Article first published online: 8 November 2011

Till Hasenberg, Department of Surgery, University Hospital Mannheim, Ruprecht Karls University Heidelberg, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany.


Aim  Patients suffering from advanced colorectal cancer can experience unintended weight loss and/or treatment-induced gastrointestinal toxicity. Based on current evidence, the routine use of parenteral nutrition (PN) for patients with colorectal cancer is not recommended. This study evaluates the effect of PN supplementation on body composition, quality of life (QoL), chemotherapy-associated side effects and survival in patients with advanced colorectal cancer.

Method  Eighty-two patients with advanced colorectal cancer receiving a palliative chemotherapy were prospectively randomized to either oral enteral nutrition supplement (PN-) or oral enteral nutrition supplement plus supplemental PN (PN+). Every 6 weeks body weight, body mass index (BMI), chemotherapy-associated side effects and caloric intake were assessed, haemoglobin and serum albumin were measured. Body composition was assessed by body impedance analysis, and QoL was evaluated by European Organization for Research and Treatment of Cancer (EORTC) QLQC30 questionnaire.

Results  No differences were evident at baseline between the groups for age, sex, diagnosis, weight, BMI or QoL. A difference in BMI was observed by week 36, whereas differences of the mean body cell mass could be observed from week 6, albumin dropped significantly in the PN- group in week 36 and QoL showed significant differences from week 18. Chemotherapy-associated side effects were higher in PN-. The survival rate was significantly greater in the PN+ group.

Conclusion  A supplementation with PN slows weight loss, stabilizes body-composition and improves QoL in patients with advanced colorectal cancer. Furthermore, it can reduce chemotherapy-related side effects.


Treating patients suffering from advanced cancer is often complicated by a variety of major clinical problems. The dilemma is they are partly caused by the tumour itself and partly by the anticancer treatment. Two major clinical problems include weight loss and the occurrence of chemotherapy-induced gastrointestinal toxicity.

The most common prodromal finding in patients with cancer, occurring in 31–37% [1–3], is weight loss. Greatest losses are seen in patients with cancer of the upper gastrointestinal tract [1,2]. Cytokine-mediated cachexia and side effects of toxicities of antineoplastic therapy can further exacerbate the existing involuntary weight loss [1,2,4–9]. Proinflammatory cytokines such as tumour necrosis factor-α, interleukin-1 (IL-1), IL-6 and interferon-γ are putative mediators of tumour-related catabolism [9]. These are also mediators for increased energy and protein requirements, which often cannot be met simply by consuming a regular diet. Tumour-related cachexia is associated with compromised substrate intake and absorption, as well as alterations of protein, carbohydrate and fat metabolism. The metabolic changes can occur before clinical evidence of cachexia [10–18]. The difference between simple starvation and catabolic weight loss in cancer is the progressive loss of muscle mass. Besides the loss of total body weight, disproportionate losses of lean tissue weight or body cell mass (BCM) are common in patients with malignancy [19]. BCM is the metabolically active component of lean tissue weight and changes rapidly with either involuntary weight loss or with nutrition repletion. Losses of fat and lean tissue can be assessed by anthropometrics, whereas losses of fat and BCM can be assessed by bioelectrical impedance analysis (BIA) [20,21]. In addition to the lack of adequate intake to meet caloric and protein requirements induced by the metabolic abnormalities associated with proinflammatory cytokines [9], there is often significant loss of appetite that is multifactorial in aetiology: cytokine-mediated changes in appetite [9], side effects of antineoplastic therapy [4–8], and psychogenic factors [22]. Gastrointestinal symptoms especially are a major oncological problem, caused by the cytotoxic effects of cancer chemotherapy and radiotherapy, and they influence the nutritional status tremendously. Because of a major cellular renewal [23] and fast protein turnover in the mucosa [24], the gut is highly sensitive to anticancer drugs. These treatments often induce mucositis, erosive lesions of the mucosa from mouth to anus. About 40% of patients receiving standard dose chemotherapy exhibit symptoms of mucositis [25–27]. 5-Fluorouracil (5-FU) is one of the most commonly used chemotherapy agents in clinical oncology practice and is associated with significant mucositis [13]. The combination of progressive physical weakness and fatigue, the psychological stress of the cancer diagnosis, toxicities of therapy and anorexia can all adversely affect the patient’s quality of life (QoL). Intensified nutrition intervention has been shown to ease some of these symptoms [28] and thereby improve QoL. Malnutrition and involuntary weight loss adversely affects prognosis in patients with cancer [1]. However, most patients do not receive intensified nutrition intervention until oral intake is significantly compromised and the patient has experienced a significant degree of weight loss [15].

The discussion of intensive nutritional intervention in patients with cancer is controversial [29–32]. Although some studies have demonstrated no significant benefit for patients with cancer [31,32], others have shown definitive value [33–40]. According to current evidence, most organizations, including the American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.), do not recommend the routine use of artificial nutrition for patients with cancer [41]. A recent study from our group showed, however, that parenteral nutrition (PN) is able to stabilized the body composition, enhance QoL and even prolong survival [42]. Problems of the initial study included the heterogeneity of the study population, the small sample size of the cohorts and the differing oncologic therapy scheme. For these reasons, the present study was designed to reproduce the initial study but only in patients with the same malignancy and a defined oncological therapy scheme. We also aimed to show whether supplemental PN influences the incidence of gastrointestinal and chemotherapy-related symptoms.



Eighty-two patients with advanced incurable colorectal carcinoma (UICC stages IV) were prospectively randomized to one of two nutrition groups between November 2000 and November 2005. All were managed by the nutrition support team (NST) team of the University Hospital Mannheim. All patients had pending or clinically evident malnutrition. Malnutrition was diagnosed by either weight loss of at least 5% in 3 months or a body mass index (BMI) of 20 kg/m2 or below. All patients were intermittently receiving palliative chemotherapy according to the Ardalan scheme (folinic acid 500 mg/m2 and 5-FU 2400 mg/m2) or the Mayo scheme (folinic acid 20 mg/m2 and 5-FU 425 mg/m2). Patients were excluded from randomization for any of the following conditions: immune deficiency diseases, infection or sepsis, corticosteroids therapy, incompatibility with components of PN or gastrointestinal obstruction. All patients or their carers gave their informed consent to participate in the study. Every patient or his carer had the right to withdraw from participation at any time during the trial. The study protocol was reviewed and approved by the Ethical Committee of the University Hospital of Mannheim.

Study design, therapy plan

After tumour staging and the decision for palliative chemotherapy by a tumour board and also the detection of pending or clinically evident malnutrition, patients were immediately seen by the NST. They were randomized to two treatment groups. One group (PN-) received intense oral enteral nutrition and the other (PN+) received the same intense oral enteral nutrition supplemented with PN. The intense oral enteral nutrition was defined as normal oral diet ad libitum, supplemented with oral enteral formulae as follows: Fresubin Energy Drink (Fresenius Kabi, Bad Homburg, Germany; 1.5 calorie/ml, 5.6 g/100 ml protein, 18.8 g/100 ml carbohydrate, 5.8 g/100 ml fat); Fortimel (Nutricia; Pfrimmer, Erlangen, Germany; 1.0 calorie/ml, 10 g/100 ml protein, 10.3 g/100 ml carbohydrate, 2.1 g/100 ml fat). The PN supplement regimen consisted of central venous infusion of fat, carbohydrate and amino acids (Nutriflex Lipid Plus; B. Braun, Melsungen, Germany; 1.0 calorie/ml, 38.4 g/l amino acids, 120 g/l carbohydrate, 40 g/l fat). PN was administered at night through a central venous port and was calculated to supply 30% of the individual’s energy and protein requirement (Table 1). Both groups received an adapted, customized supplement of vitamins (Cernevit; Baxter, Unterschlei ßheim, Germany) and trace elements (Addel N; Baxter, Unterschlei ßheim, Germany).

Table 1.   Nutrition regimen.
 PN+ (% of total intake)PN- (% of total intake)
  1. EN, enteral formula; PNS, supplemented parenteral nutrition.

Oral diet50–6050–6060–8060–80
Total intake100100100100

Every patient kept a daily nutrition diary for assessment of calorie and protein consumption. Patients underwent physical examination and nutritional assessment every 6 weeks. Parameters assessed included calorie and protein intake, measurement of body weight, BMI, body composition by BIA including total body fat, BCM, extracellular mass (ECM), total body water (TBW), gastrointestinal symptoms, chemotherapy side-effects and QoL using the European Organization for Research and Treatment of Cancer (EORTC) QLQ-C30 instrument. Laboratory assessment included complete blood count, serum electrolytes and serum albumin. Nutrition regimen was adjusted to meet individual patient requirements according to the results of the assessments.

Bioelectrical impedance analysis

Body composition was determined by BIA. All patients were required to rest for 30 min, after which BIA was performed by applying four silver electrodes, with two detecting electrodes placed at the ulnar aspect of the right wrist and the right medial malleolus. After connecting the electrodes to the multiple-frequency BIA instrument (BIA 2000; Data Input, Darmstadt, Germany), the measurements were recorded in real time via computer. The calculations for BCM, ECM, TBW and fat were performed using nutri 2000 software (Data Input).

Resting energy expenditure

Resting energy expenditure (REE) was assessed by indirect calorimetry using a ventilated hood system (Deltatrac MBM-100; Datex Instrumentarium Corp., Helsinki, Finland). After an overnight fast and a resting period of approximately 30 min, CO2 production and O2 consumption were measured during a 30-min period between 9.00 and 11.00 a.m., whereas the patient was lying at complete rest. REE was calculated by using the abbreviated Weir formula. Measured REE was compared with predicted REE using the Harris-Benedict equations, which are age-, height-, weight- and gender-specific. The equipment was calibrated at the start of each experiment.

EORTC QLQ-C30 questionnaire

In 1986, the EORTC initiated a research programme to develop an integrated, modular approach for evaluating the QoL of patients participating in international clinical trials. The QLQ-C30 incorporates nine multi-item scales: five functional scales (physical, role, cognitive, emotional and social); three symptom scales (fatigue, pain and nausea and vomiting) and a global health and QoL scale. Several single-item symptom measures are also included. The EORTC QLQ-C30 is a reliable and valid measure of the QoL of cancer patients in multicultural clinical research settings.[43–47]


The software-supported data evaluation was conducted with the assistance of spss, release 15.0.0 (SPSS Inc., Chicago, Illinois, USA). Results for quantitative data were expressed as means ± SD. Two-sample t-test was used to compare the groups for age, weight, weight loss, BMI, haemoglobin, gastrointestinal symptoms, chemotherapy-related side effects and albumin at onset of the trial. Qualitative parameters such as gender or medical diagnosis were compared using the χ2 test. Weight, BMI, BCM, ECM, TBW, fat, albumin levels, haemoglobin and QoL over time, both for treatment and control groups, were analyzed by using repeated-measures analysis of variance (ANOVA). Comparisons between two groups at a single time point were performed using the two-sample t-test. The cumulative rate of survival for both groups was displayed via Kaplan–Meier, and the difference of the death rates was determined via log rank test. P value 0.05 was considered as statistically significant.


Of the 82 evaluable subjects, 42 were randomized to PN+ and 40 to PN-. At randomization, subjects in both groups were comparable in terms of gender, age, medical diagnoses, UICC (L’Union Internationale Contre le Cancer) stage, weight, BMI, haemoglobin, REE and albumin (Table 2). The protocol used for chemotherapy in patients with colorectal carcinoma (MAYO; folinic acid 20 mg/m2, 5-FU 425 mg/m2): PN+ 18; PN- 19, ARDALAN (folinic acid 500 mg/m2, 5-FU 2400 mg/m2): PN+ 24, PN- 21. Median follow-up time of all participants was 12.7 months. The calculated calorie and protein goal was attained to within 95% at baseline and throughout the study in both groups. Mean volumes of enteral feed consumed per day were significantly different, with 255 ± 23 ml/day for PN+ and 498 ± 68 ml/day for PN-, P = 0.0001. The PN+ group did not suffer a negative effect on QoL owing to the PN itself. All patients of PN+ group continued their PN through the whole study period. No complications of PN (e.g. catheter-associated inflammation) were observed. The calculated caloric and protein goal was reached in both groups 95–100% of all times (Table 3). In the premortal period, the intensified clinical nutrition was stopped. This period was on an average 13 + 4 days. The PN+ group did not report a negative influence on QoL because of the PNS.

Table 2.   Baseline characteristics of the patients.
 PN+ (n = 42)PN- (n = 40)P-value
  1. *Mean ± SD.

  2. †Mean weight loss in last 3 months, n = 82.

  3. BCM, body cell mass; BMI, body mass index; ECM, extracellular mass; TBW, total body water; REE, resting energy expenditure.

Age (years)*59.0 ± 12.357.4 ± 10.40.529 (ANOVA)
Female subject22 (52%)21 (53%)0.991 (χ2 test)
Male subject20 (48%)19 (47%)
Weight (kg)*66.8 ± 12.268.1 ± 11.70.621 (ANOVA)
Weight loss (kg) † 7.0 ± 2.16.9 ± 2.00.789 (ANOVA)
BMI (kg/m2)*23.1 ± 2.122.8 ± 2.10.496 (ANOVA)
BCM (%)*52.8 ± 1.652.1 ± 2.10.062 (ANOVA)
ECM (%)*50.6 ± 2.251.0 ± 1.70.310 (ANOVA)
Fat (%)*17.9 ± 1.617.0 ± 2.20.024 (ANOVA)
TBW*61.5 ± 2.261.0 ± 1.90.350 (ANOVA)
REE (kcal/day)1545 ± 561564 ± 630.2753 (ANOVA)
Haemoglobin (g/dl)13.1 ± 1.212.7 ± 0.90.076 (ANOVA)
Albumin (g/l)40.6 ± 4.539.2 ± 5.30.198 (ANOVA)
Rectal carcinoma16 (38%)14 (35%)0.771 (χ2 test)
Colon carcinoma26 (62%)26 (65%)
Liver metastasis28 (67%)27 (67%)0.736 (χ2 test)
Table 3.   Mean energy and protein intake per day.
 PN+ (n = 42)PN- (n = 40)P-value
  1. *Comparison of the mean energy and protein intake in both groups, mean ± SD, n = 82.

Oral diet
 Energy (kcal/day)1260 ± 89*1614 ± 84*0.001
 Protein (g/day)54.0 + 4*61.1 + 6*0.001
Enteral formula
 Volume (ml/day)255 ± 23*498 ± 68*0.001
 Energy (kcal/day)312 ± 23*619 ± 31*0.001
 Protein (g/day)21.3 + 5*41.6 + 6*0.001
 Energy (kcal/day)675 + 49*  
 Protein (g/day)28.0 + 5*  
 Energy (kcal/day)2247 ± 70*2233 ± 51*NS
 Protein (g/day)103.3 ± 5*102.7 ± 6*NS

The BMI of the participants in both groups showed no significant difference at the time of entry into the study. Both groups were able to maintain their BMI stable for 6 month. After 6 month of nutritional therapy, the PN- group showed a decline in BMI, which reached significance after 36 weeks. Patients in the PN+ group however were able to maintain a stable BMI over the entire period of observation (Fig. 1).

Figure 1.

 Changes of body mass index (BMI, kg/m2) over time. n = 82, *significance level P < 0.05.

Bioelectric impedance analysis

At the time of entry, the patients in both groups showed equal values for BCM, ECM and TBW. Only the values for fat showed a significant difference (P = 0.024). The further BIA measurements showed a decline of BCM in the PN- group, which persisted over the entire observation period and reached significance after 6 weeks of nutritional therapy (Fig. 2). The PN+ patients were however able to keep their BCM stable over time.

Figure 2.

 Changes of body cell mass (BCM, %) over time. n = 82, *significance level P < 0.05.

The course of the ECM showed a continuous increase in the PN- group and stable values in the PN+ group (Fig. 3). The difference between the two groups reached significance in the 18th and 24th week of treatment and from the 42nd week until the end of our study. The TBW rose in the PN- group throughout the study, whereas the TBW in the PN+ group showed a stable course (Fig. 4). The difference between the two groups reached a significant level in the 48th week of treatment.

Figure 3.

 Changes of extra cellular mass (ECM, %) over time. n = 82, *significance level P < 0.05.

Figure 4.

 Changes of total body water (TBW, %) over time. n = 82, *significance level P < 0.05.

The total body fat showed stable results in the PN+ group throughout the study, whereas the PN- group showed a continuous decline over time (Fig. 5). The differences between the two groups were significant during the whole study.

Figure 5.

 Changes of body fat (%) over time. n = 82, *significance level P < 0.05.

Laboratory values

Both treatment groups showed stable haemoglobin values for 36 weeks. From the 42nd week of treatment, the patients in the PN- group showed a significant decline in haemoglobin, whereas the patients in the PN+ group presented stable values during the treatment.

Regarding the albumin levels, an equally stable course was observed in both groups up to the 18th week. Beginning at the 24th week the albumin levels of the PN- group showed a declining trend, which reached significance in the 36th week. Although the albumin levels of the PN- group showed an increase in the 48th week, the difference between the two groups was still significant. The PN+ group showed stable albumin levels throughout the entire observation.

Quality of life

Both groups showed an improvement in QoL (Fig. 6). In patients in the PN+ group, however, this was greater than in the PN- group. This difference reached significance in the 18th week and remained significant from the 30th throughout the rest of the study.

Figure 6.

 Changes in quality of life (QL) over time. n = 82, *significance level P < 0.05.

Gastrointestinal symptoms

Thirty-two per cent of the patients in the PN+ group and 69% of the PN- group experienced early satiety. Twenty-four per cent of the PN+ group and 59% of the PN- group complained of constipation, whereas 9% in the PN+ and 24% in the PN- group reported diarrhoea. Nausea and vomiting were experienced by 22% and 12%, respectively, in the PN+ group and 49% and 38% in the PN- group. Abdominal pain was reported by 18% in the PN+ group and 37% of the PN- group (Table 4).

Table 4.   Gastrointestinal symptoms.
 PN + (n = 42)PN- (n = 40)P-value
  1. n = 82.

  2. *Significance level P < 0.05.

Early Satiety (%)32*69*<0.001
Constipation (%) 24*59*<0.001
Nausea (%)22*49*<0.001
Vomiting (%)12*38*<0.001
Abdominal pain (%)18*37*<0.001
Diarrhoea (%)9*24*<0.001

Chemotherapy related side effects

Twenty-two per cent of the PN+ and 32% of the PN- group experienced symptoms, which could be grouped under the term mucositis. Polyneuropathy developed in 6% of the PN+ group and 5% of the PN- group. Leucopoenia occurred in 6% of the PN+ group and 7% of the PN- group while on chemotherapy (Table 5).

Table 5.   Chemotherapy-related side effects.
 PN+ (n = 42)PN- (n = 40)P-value
  1. n = 82.

  2. *Significance level P < 0.05.

Mucositis (%)22*32*<0.013
Diarrhoea (%)9*24*<0.020
Polyneuropathy (%)6*5*NS
Leucopoenia (%)6*7*NS


Median survival rates for both groups were significantly different, being 16.7 (range 14–19) months for PN+ compared with 10.2 (range 8–12) months for PN-, (P < 0.001). In addition, there was a highly significant difference (P < 0.0001) in the cumulative rate of survival (Fig. 7).

Figure 7.

 Cumulative survival rate by Kaplan–Meier, n = 82, significance level P < 0.05.


The results of this study showed that PN supplementation slows the progression of weight loss in patients suffering from advanced colorectal cancer. This is in accordance with our initial study in which it was shown to stabilize body composition, enhance the QoL and even prolong survival [42]. One problem with the initial study was the heterogeneity of the study population, the small sample size of the cohorts and the differing oncologic therapy scheme.

The focus of the present study was therefore to reproduce the initial study setting in a more homogenous way with only one malignant disease and a defined oncological therapy scheme. Parenteral supplementation had a positive influence on the body composition. Although the BMI did not show any significant difference in both treatment groups for up to 36 weeks, the analysis of the body composition demonstrated significant changes that occurred within the first weeks of the trial. The BCM fell in the PN- group, which reached significance at 6 weeks of treatment. The curve for ECM and TBW showed an inverse shape, with a stable course in the PN+ group and significant higher values in the PN- group after 12 weeks (ECM) respectively 48 weeks (TBW). These findings are in accordance with earlier studies [20,21,48,49], which showed that significant changes in body composition appear much early than changes in body weight or BMI.

Other than in our previous report [42], the positive influence of parenteral supplementation has not been previously described. A key feature of the altered body composition in patients with malignant diseases is severe protein loss [50,51]. It has been shown that the early initiation of PN can stabilize body protein [15,18]. In our study, we were able to show that the serum albumin remained stable in the PN+ group, whereas the serum albumin levels in the PN- group declined over time. This effect can be considered as a stabilization of body proteins by a supplemental PN. Even without eliciting the nitrogen balance in this study, it is likely that supplementary PN can improve the severe protein loss in patients with malignant disease.

One feature of the study protocol was the close follow-up. We aimed to insure that the caloric goals were reached by each patient and that side effects were detected as early as possible. Interestingly, even with a normal calorie intake the progressive loss of weight, BCM, fat and protein still occurred in the PN- group, whereas the PN+ group was able to keep these measures stable. This is in accordance with the anabolic effects of PN described by others [34,35,37,38]. The deterioration of body-composition in the PN- group and the stabilization in the PN+ group cannot be ascribed to any type of hypo- or hyperalimentation. The study protocol assured an adequately adjusted isocaloric nutrition therapy in both study groups. This is also in accordance with the finding that a hypercaloric nutritional intervention in cancer patients does not provide any benefit [34]. It is possible that the nutrition diary, which was kept by the patients, could be biased by the appearance of chemotherapy-related gastrointestinal side effects. In our study, the PN+ group reached 70% of their caloric needs enterally. In the PN- group, 100% of the calculated energy requirements were met by mouth. In the PN+ group, we found chemotherapy-related side effects to a much lesser extent than in the enteral group. We cannot rule out the possibility that patients in both groups, who suffer from such gastrointestinal side effects, over-reported their enteral intake.

Besides the unintended weight loss and the changes in body composition, cancer patients suffer from symptoms partly related to the disease and its metabolism and partly to the effect of the treatment. These result in an unfavourable clinical outcome, psychological and socioeconomic impairment and poor QoL [52]. The improved QoL is a known effect because of the ‘special care’ of a study. However, our study showed that a PN was not only able to stabilize body composition but to reduce gastrointestinal symptoms and chemotherapy-related symptoms and to improve QoL.

Quality of life can be assessed by measuring physical, psychological and social factors. Malignant disease and treatment-induced changes in metabolism [53] can cause alterations in physiological and psychological function [54], which may reduce it. Indeed, nutritional status and food intake are often adversely affected by acute and chronic symptoms secondary to antineoplastic treatments [55–60]. Progressive wasting may occur [61] and results in cachexia [54,60–62].

In this study, we used the European Organization for Research and Treatment of Cancer Quality of Life Core Questionnaire (EORTC QLQ-C30) [63], which contains 30 items developed for cancer patients enrolled in clinical trials. It assesses several factors that contribute to QoL, including physical ability, cognitive status and emotional and social factors. Symptoms and financial impact are also considered by the questionnaire. Several items of the EORTC QLQ 30 focus gastrointestinal symptoms that can be influenced by the malignant disease or the therapy.

Beyond global QoL improvement, there were also significantly fewer gastrointestinal symptoms in the PN+ group and rather than staying stable, the score actually improved during the study. This trend was seen to a lesser degree in the PN- group. During the trial, patients in the PN+ group reported fewer symptoms such as early satiety, constipation, nausea, vomiting, abdominal pain and diarrhoea. This resulted in a significant increase in appetite in the PN+ group.

Gastrointestinal and/or chemotherapy-related symptoms not only affect the QoL of cancer patients [64,65] but also lead to a reduction in dose which may contribute to morbidity or mortality. The intestine is especially susceptible to toxicity [66,67]. Mucositis is associated with gut barrier disruption [68,69] being related to an imbalance between proliferation and apoptosis of gut epithelial cells. It affects from 40% to 76% of patients on chemotherapy [67]. The nutritional impact of mucositis is controversial. Various publications have studied on the effect of glutamine on gastrointestinal mucositis [70–73]. Numerous studies have established that hypercatabolic and hypermetabolic states are associated with profound glutamine deprivation [74,75]. Thus, glutamine supplements have been used to treat mucositis but clinical trials have shown conflicting results [70–72]. In a blinded randomized study, parenteral alanyl-glutamine supplementation was shown to reduce mucositis and patients receiving glutamine had a higher incidence of disease relapse. The reduced incidence of mucositis rate in our patients cannot be explained easily. One factor may be the stabilization of body protein achieved in the PN+ in contrast to the PN- group. Low albumin levels may be associated with the development of mucositis. Shoval et al. [76] demonstrated that patients after bone marrow transplantation show a marked increase in mouthrinse albumin before the development of oral mucositis. In the same study, serum albumin level were correlated with the mouthrinse albumin level, suggesting that oral albumin leakage is a significant source of serum albumin loss [76] as noted in critically ill patients related to an increase in capillary leakage [77].

Survival analysis shows a significant advantage for the PN+ group. Both the number of long-term survivors and the median duration of survival were greater (PN+ 16.7 months vs PN- 10.2 months). The recent randomized prospective study of nutrition intervention with home PN by Lindholm et al. [78] demonstrated improved survival in 399 patients with malignant disease and progressive cachexia. Patients were randomized to the COX inhibitor indomethacin, 50 mg twice daily, and erythropoietin (15,000–40,000 units per week), with specialized, nutrition-focused patient care (oral NS and home PN). Control patients received the same indomethacin and erythropoietin doses without added NS. All patients were treated and followed until death. An intention-to-treat analysis revealed an improvement in energy balance for nutritionally supported patients (P < 0.03) and no other significant differences between study and control patients were observed. Patients receiving nutrition experienced prolonged survival (P < 0.01) accompanied by improved energy balance (P < 0.001), increasing body fat (P < 0.05), and a greater maximum exercise capacity (P < 0.04). Although the design of this and the study of Lindholm et al. [78] are different, the results of both support the contention that nutrition and lean tissue status may be limiting factors influencing survival.

In conclusion, our study showed that PN supplement slowed weight loss and improved QoL in patients with advanced colorectal cancer. Furthermore, it reduced chemotherapy-related side effects and mucositis. With regard to the positive effect of PN on survival, body composition and QoL, additional controlled studies need to be conducted to confirm these findings.