T. Crowe, School of Exercise and Nutrition Sciences, Deakin University, Burwood Highway, Burwood, Victoria 3125, Australia. Tel.: +61 3 9251 7266 Fax: +61 3 9244 6017 E-mail: firstname.lastname@example.org
Background: Malnutrition and its associated complications are a considerable issue for surgical patients with upper gastrointestinal and colorectal cancer. The present study aimed to determine whether specific perioperative nutritional practices and protocols are associated with improved patient outcomes in this group.
Methods: Patients admitted for elective upper gastrointestinal or colorectal cancer surgery (n = 95) over a 19-month period underwent a medical history audit assessing weight changes, nutritional intake, biochemistry, post-operative complications and length of stay. A subset of patients (n = 25) underwent nutritional assessment by subjective global assessment prior to surgery in addition to assessment of post-operative medical outcomes, nutritional intake and timing of dietetic intervention.
Results: Mean (SD) length of stay for patients was 14.0 (12.2) days, with complication rates at 35%. Length of stay was significantly longer in patients who experienced significant preoperative weight loss compared to those who did not [17.0 (15.8) days versus 10.0 (6.8) days, respectively; P < 0.05]. Low albumin and post-operative weight loss were also predictive of increased length of stay. Of patients who underwent nutritional assessment, 32% were classified as mild–moderately malnourished and 16% severely malnourished. Malnourished patients were hospitalised twice as long as well-nourished patients [15.8 (12.8) days versus 7.6 (3.5) days; P < 0.05]. Time taken [6.9 (3.6) days] to achieve adequate nutrition post surgery was a factor in post-operative outcomes, with a positive correlation with length of stay (r = 0.493; P < 0.01), a negative correlation with post-operative weight change (r = −0.417; P < 0.05) and a greater risk of complications (52% versus 13%; P < 0.01).
Conclusions: Malnutrition is prevalent among surgical patients with gastrointestinal cancer. Poor nutritional status coupled with delayed and inadequate post-operative nutrition practices are associated with worse clinical outcomes.
Several factors predispose patients undergoing surgery for upper GI and colorectal cancer to malnutrition. These factors include the catabolic effect of cancer as well as the GI side effects of nausea, vomiting, anorexia, diarrhoea and, in some cases, dysphagia and malabsorption (Nitenberg & Raynard, 2000; Fettes et al., 2002; Chate, 2006). Traditional surgical nutrition practices (for example nil-by-mouth periods) and side effects of adjuvant treatments can also reduce nutrient intake at a time when nutritional status is already compromised (Nitenberg & Raynard, 2000; Weimann et al., 2006).
Although it is evident that preoperative malnutrition is a considerable preoperative risk for patients with GI cancer undergoing surgery, the relationships between nutritional practices during the perioperative period and clinical outcomes among this patient group are less well established (Kelsen et al., 1998; Malone et al., 2002; Alves et al., 2005). Nutrition plays an important role in minimising the effects of malnutrition; however, it has not been definitively determined which of oral supplements, enteral nutrition, parenteral nutrition or diet alone, and in which combination, provides the most effective form of nutrition support (Bozzetti et al., 2001; Elia et al., 2006). Recently, the provision of specialised supplements with immune-modulating properties (immunonutrition), has added further complexity to the issue (Mccowen & Bistrian, 2003; Farreras et al., 2005). Furthermore, barriers to meeting nutritional requirements after surgery have not been well studied in this patient population, nor have the efficacy of traditional nil-by-mouth periods (Reissman et al., 1995; Stewart et al., 1998; Lewis et al., 2001; Andersen et al., 2006).
Because malnutrition and its associated complications are a considerable issue for surgical patients with upper GI and colorectal cancer, further research is needed to determine whether specific perioperative nutritional practices and protocols can improve patient outcomes. Subsequently, the present study aimed to examine the prevalence of malnutrition and surgical complications in this patient group, in the context of current nutritional practices, to establish a rationale and direction for future research in this area.
Materials and methods
St Vincent’s Hospital Melbourne, Australia, provided the setting for this two-part study, which was conducted over a 19-month period (March 2006 to October 2007). The hospital is a major teaching, research, and tertiary referral hospital. Patients aged over 18 years admitted to the upper GI and colorectal units for Category One surgery (requiring admission within 30 days) for cancer were the focus of the study. The first part of the study involved a retrospective audit of the medical histories of all patients who met the selection criteria and were admitted during the study period. The second part of the study involved the additional assessment of a subset of patients and included preoperative nutritional status and post-operative nutritional intake.
Retrospective medical audit
The hospital waitlist co-ordinator provided details of the medical histories of patients who met the selection criteria between March 2006 and October 2007. Data collected included patient’s age, gender, medical diagnosis and surgical procedure. Prior to surgery, the most recent anthropometric data (height, weight and weight history), obtained by a combination of measured and reported methods, was collected from the medical histories. Preoperative weight loss was deemed significant if there was a loss of usual body weight >2% in the week prior to admission, 5% in the last month, 7.5% in the last 3 months or 10% in the last 6 months (Blackburn et al., 1977). The most recent preoperative serum albumin was recorded from the hospital pathology database if it was collected within 1 month prior to surgery.
Subsequent to the patient’s surgery, information regarding weight change over admission, length of stay and complications were collected from the medical history. Complications that were of relevance included mortality, anastomotic leak, sepsis, wound infection, transfer to the intensive care unit, requirement for parenteral nutrition, prolonged post-operative ileus defined as an ileus lasting more than 7 post-operative days (Artinyan et al., 2008), wound dehiscence, return to theatre, pneumonia, urinary tract infection and the number of days the patient required nasogastric tube drainage. Nutritional practices were recorded, including the number of days that the patient was kept nil-by-mouth post-operatively, the time taken for the patient to achieve adequate nutrition, and the timing and nature of dietetic intervention (if any). For the purposes of the study, achieving adequate nutrition was defined as reaching >75% of estimated energy and protein requirements via either enteral or parenteral nutrition or the second day that a soft ward diet was tolerated by the patient.
Nutritional status and nutritional intake assessment
A convenience sample of a consecutive subset of patients (those admitted between May 2007 and October 2007) included in the medical history audit were invited to participate in the second part of the study. Eligible patients were recruited from the ward or the preadmission clinic. Eligible patients were defined as those who met the inclusion criteria for the study and were able to provide informed consent, had adequate English language and cognitive abilities, and were able to participate in a nutritional assessment.
Because there is no internationally accepted definition of malnutrition, the present study used the widely adopted and validated subjective global assessment (SGA) tool to assess the preoperative nutritional status of patients (Detsky et al., 1987b). The SGA involves a short interview assessing recent weight change, dietary intake, GI symptoms and functional impairment, and a physical examination (Detsky et al., 1987b). The SGA is a validated tool with a high degree of specificity and sensitivity as well as a high degree of inter-observer agreement and correlation between objective and subjective measures (Detsky et al., 1984, 1987a,b). In GI surgical patients, SGA has also been found to be highly predictive of post-operative infections (Detsky et al., 1984). On completion of the nutrition assessment interview, the patient’s SGA ranking: A (well nourished), B (mild–moderately malnourished) or C (severely malnourished), was recorded. The SGA was performed by the researchers involved with the study.
In addition to data regarding timing of dietetic intervention, details of post-operative nutritional intake were also obtained, including the use of enteral and parenteral nutrition support and the number of days until a soft ward diet was commenced. On the second day that a soft ward diet was tolerated by the patient, detailed food and fluid intake charts were recorded. The content of these food charts was analysed using nutritional analysis software (FoodWorks; Xyris, Highgate Hill, Queensland, Australia) using the Australian Food and Nutrient Database (AusNut) compiled in 1999 by Food Standards Australia and New Zealand (Australian and New Zealand Food Authority, 1999). Patients’ estimated energy requirements were determined by the Schofield equation adjusted for relevant stress and activity factors (Schofield, 1985; Ferrie & Ward, 2007). Protein requirements were estimated on a gram of protein per kilogram body weight basis with appropriate stress factors (Moghissi & Teasdale, 1980; Larsson et al., 1990). Obesity adjusted weight was used to calculate energy and protein requirements where necessary (Barak et al., 2002).
Ethical approval for the study was obtained from the St Vincent’s Hospital Human Research Ethics Committee and the Deakin University Human Research Ethics Committee. Informed consent was not required for the retrospective medical audit aspect of the study; however, signed informed consent was obtained from all participants recruited for part two of the study involving the assessment of nutritional status and nutritional intake.
Between-group comparisons were performed using unpaired Student’s t-test or one-way analysis of variance as appropriate. Pearson’s product-moment correlation was used to determine all associations. For binomial comparisons of complication rates and significant preoperative weight loss rates, Fisher’s exact test was used. An alpha error of P < 0.05 was used to determine statistical significance in all analyses. Data was analysed using spss, version 15 (SPSS Inc., Chicago, IL, USA) and the results are presented as the mean (SD).
During the study period, 95 surgical patients were identified as being eligible for inclusion. Of these patients, 37 were admitted to the upper GI unit and 58 to the colorectal unit. Table 1 presents the preoperative demographics and medical characteristics of these patients. Characteristics between patients admitted to the upper GI and colorectal unit were similar with the exception of preoperative albumin, with upper GI patients found to have significantly lower albumin. The most common medical diagnosis of admitted patients was colon cancer (61%; includes colon, sigmoid, rectal, rectosigmoid and caecal cancer), pancreatic (17%) and gastric cancer (15%). Other diagnoses include oesophageal and gall bladder cancer. Open colonic resections were the most common surgical procedure (44%) followed by laparoscopic colonic (18%), gastric (16%) and pancreatic resections (13%). Oesophageal resections, gastric bypass surgery and gall bladder resections comprised the remainder of the surgical procedures performed.
Table 1. Preoperative patient demographics and medical characteristics
Data presented as the mean (SD).
n, number of observations; BMI, body mass index; GI, gastrointestinal; RR, reference range.
*P < 0.05 between upper GI and colorectal.
†Preoperative weight loss was deemed significant if there was a loss of usual body weight >2% in the week prior to admission, 5% in the last month, 7.5% in the last 3 months or 10% in the last 6 months.
Gender (male : female)
62 : 33
22 : 15
40 : 18
BMI (kg m−2)
Albumin (g L−1; RR = 35–50 g L−1)
Significant preoperative weight loss† (n)
Table 2 shows the post-operative outcomes of patients included in the study. Patients admitted to the upper GI unit were kept nil-by-mouth significantly longer post-operatively than colorectal patients and took longer to achieve adequate nutrition. The most common complication reported was wound dehiscence (11%), followed by prolonged post-operative ileus (10%), wound infection (8%), sepsis (6%), urinary tract infection (6%), requirement for parenteral nutrition (6%), return to theatre (5%), pneumonia (5%) and transfer to intensive care (3%). No patients died during hospitalisation, whereas 2% experienced an anastomotic leak.
Table 2. Post-operative outcomes of patients
Data presented as the mean (SD).
n, number of observations; GI, gastrointestinal.
*P < 0.01 between upper GI and colorectal; Achieving adequate nutrition defined as reaching >75% of estimated nutritional requirements by either enteral or parenteral nutrition or the second day that a soft ward diet was tolerated.
Length of stay (days)
Post-operative weight change (%)
Any complication (n)
Nil-by-mouth post-operatively (days)
Time to achieve adequate nutrition (days)
Although the majority of patients (79%) had a preoperative albumin level >35 g L−1 (reference range 35–50 g L−1), significantly more upper GI compared to colorectal patients presented with albumin below 35 g L−1 (31% versus 14%, respectively; P < 0.05). There was a significant negative correlation between preoperative albumin and length of stay (r = −0.325; P < 0.05) (Fig. 1). Compared to patients who had a preoperative albumin >35 g L−1, patients who presented with a preoperative albumin <35 g L−1 were kept nil-by-mouth significantly longer [2.0 (2.1) days versus 3.3 (2.1) days; P < 0.05] and took significantly longer to achieve adequate nutrition [6.7 (3.2) days versus 8.7 (4.4) days; P < 0.05].
Length of stay was significantly longer in patients who experienced significant preoperative weight loss compared to those who did not [17.0 (15.8) days versus 10.0 (6.8) days, respectively; P < 0.05]. There were no significant differences found between weight status on admission (underweight, healthy weight range, overweight according to BMI for age) and post-operative outcomes (length of stay, complications or weight change over admission; data not shown).
Post-operative nutritional intake
Time taken to achieve adequate nutrition (>75% of estimated nutritional requirements by either enteral or parenteral nutrition or the second day that a soft ward diet was tolerated) was found to have an association with post-operative outcomes. Time taken to achieve adequate nutrition was positively correlated with length of stay (r = 0.493; P < 0.01) and negatively correlated with post-operative weight change (r = −0.417; P < 0.05) (Fig. 2). Furthermore, patients who took ≥7 days to achieve adequate nutrition were significantly more likely to experience at least one complication compared to those who achieved adequate nutrition in <7 days (52% versus 13%, respectively; P < 0.01; data not shown). Of the seven patients with post-operative ileus, only three received parenteral nutrition and it took an average of 10.3 days for these three patients to achieve adequate nutrition.
Weight change over admission was negatively correlated with length of stay (r = −0.585; P < 0.01) (Fig. 3). Furthermore, the number of days that a patient was kept nil-by-mouth post-operatively positively correlated with length of stay (r = 0.224; P < 0.05; data not shown); however, it was not associated with weight change over admission (r = 0.051; P = 0.745; data not shown).
When the effect of days to achieve adequate nutrition, weight change over admission, and number of days nil-by-mouth post-operatively was examined in relation to the length of stay by multiple linear regression, only weight change over admission remained a significant predictive factor of length of stay (P < 0.001).
Nutritional status and post-operative outcomes
Table 3 shows the preoperative characteristics for the 25 patients who underwent preoperative nutritional status assessment. According to SGA, 32% of patients were classified as mild–moderately malnourished (SGA = B) and 16% were severely malnourished (SGA = C) on admission. Patients who were malnourished (SGA = B + C) were significantly older, had a significantly lower preoperative weight, BMI and albumin, and were more likely to have significant preoperative weight loss compared to well-nourished patients. Patients admitted to the upper GI unit (n = 9) were more likely to be malnourished compared to those admitted to the colorectal unit (n = 16), although this difference was not significant (66.7% versus 37.5%, respectively; P = 0.23).
Table 3. Preoperative characteristics of patients according to subjective global assessment (SGA) rating†
SGA = A
SGA = B
SGA = C
SGA = B + C
Data presented as the mean (SD).
n, number of observations; BMI, body mass index; GI, gastrointestinal; RR, reference range.
*P < 0.05 between SGA B + C and SGA A.
†SGA rating: A (well nourished), B (mild-moderately malnourished), C (severely malnourished).
‡Preoperative weight loss was deemed significant if there was a loss of usual body weight >2% in the week prior to admission, 5% in the last month, 7.5% in the last 3 months or 10% in the last 6 months.
Gender (male : female)
9 : 4
5 : 3
4 : 0
9 : 3
BMI (kg m−2)
Albumin (g L−1, RR = 35–50 g L−1)
Significant preoperative weight loss‡ (n)
Patients who were malnourished preoperatively spent significantly longer in hospital compared to well malnourished patients [15.8 (12.8) days versus 7.6 (3.5) days; P < 0.05]. Malnourished patients were also more than twice as likely to develop one or more complications compared to patients who were well nourished preoperatively; however, this did not reach statistical significance (41.7% versus 15.4%; P = 0.20).
The audit of patient food charts showed that, as oral intake was being re-established, mean (SD) energy intake on the second day of a soft ward diet was 38.7% (20.8%) of estimated energy requirements for well-nourished patients and 47.9% (28.3%) for malnourished patients. Only 32.0% of patients met at least half of their estimated energy requirements by the second day of a soft ward diet. Protein intake on the second day of a soft ward diet was also higher for malnourished patients compared to well-nourished patients [58.4% (36.1%) and 47.6% (25.1%) of estimated requirements, respectively].
Malnourished patients were more likely to be seen by a dietitian both preoperatively (33.3% versus 0.0%; P < 0.05) and post-operatively (75.0% versus 25.0%; P < 0.05) compared to well-nourished patients. Patients who were seen by a dietitian post-operatively had a significantly higher energy and protein intake on the second day of a soft ward diet compared to patients who were not seen by the dietitian post-operatively (Fig. 4).
Despite evidence indicating that GI surgical patients are at substantially increased risk of malnutrition preoperatively, there has been limited research measuring the consequences of malnutrition in GI surgical patients with cancer. The present study aimed to explore the preoperative nutritional status of patients with GI cancer, existing post-operative nutritional practices, and their relationships with post-surgical outcomes in an Australian hospital setting. Not surprisingly, it was revealed that malnutrition is prevalent in surgical patients with GI cancer. Approximately half (48%) of patients who were assessed for malnutrition risk were identified as malnourished, and 16% were classified as severely malnourished on admission. Although the present study was only able to conduct nutritional status assessment on a small number of patients, similar malnutrition prevalence estimates have been found in larger international studies in a similar patient group (Rey-Ferro et al., 1997; Braga et al., 2002).
The present study showed that there is an association between preoperative malnutrition and adverse post-operative outcomes for patients with upper GI and colorectal cancer. Preoperative malnutrition in this cohort was associated with significantly longer hospital stays and a trend to a greater risk of complications. This finding is consistent with the findings of previous research showing that malnutrition in hospitalised patients is associated with an increased length of stay (Middleton et al., 2001; Isabel et al., 2003; Pichard et al., 2004). Although associations cannot prove causality, this finding is clinically important because previous research shows that an extended length of stay is associated with other co-morbidities as well as increased hospital costs (Chima et al., 1997).
Other indicators of poor preoperative nutritional status, such as low serum albumin and preoperative weight loss, were also found to be adversely associated with longer length of stay. It is acknowledged that these findings should be interpreted with caution because there are limitations to the use of these parameters in assessing nutritional status. Low albumin levels may reflect an inflammatory response related to disease severity; however, albumin can still provide an important marker of patients at risk of a problematic recovery who may benefit from nutrition intervention (Corish & Kennedy, 2000). The lack of association between BMI and post-operative outcomes reinforces the clinical significance of recent weight loss rather than weight status in predicting poorer outcomes. Despite acknowledged limitations, preoperative albumin and recent weight loss appear to be useful indicators of a patient’s preoperative nutritional status, as well as markers for a higher risk of post-operative morbidity and need for extended stay, and thus should be noted and acted upon as early as possible preoperatively.
Of particular note are the observations made regarding the time taken for patients to achieve adequate nutrition post-operatively. Upper GI patients were kept nil-by-mouth for a longer period of time post-operatively and took longer to achieve adequate nutrition than colorectal patients. This is not unexpected because there is stronger evidence for early feeding in colorectal patients and it is increasingly common practice to commence oral intake within 24 h and progress to soft diet within 48 h (Andersen et al., 2006). However, in the present study, the average time taken to achieve adequate nutrition via enteral or parenteral nutrition or soft diet was 7 days, whereas the average time patients were kept nil-by-mouth was only 2 days. This indicates that, despite patients not being kept nil-by-mouth for significant periods, there was a delay in progressing to soft diet or commencing alternative nutrition support during the post-operative period. When parenteral nutrition was provided, there was a notable delay before it was commenced, with observations indicating that it took 10.3 days to reach adequate nutrition for these patients. Usually, parenteral nutrition would meet nutrition requirements within 1–2 days of initiation (Krzywda et al., 1993).
The present study observed that inadequate post-operative nutrition was associated with poor clinical outcomes in patients with upper GI or colorectal cancer. Patients who took longer to achieve adequate nutrition post-operatively were more likely to lose weight or gain less weight over their inpatient admission. Further exacerbating the potential for weight loss over the admission was the finding that, although re-establishing oral intake after surgery, 68% of patients were consuming less than half of their estimated energy requirements on the second day of a soft ward diet. Because it is not uncommon for patients to be discharged soon after tolerating a soft ward diet, this creates the potential for further weight loss and decline in nutritional status post discharge outside of the monitoring of the hospital environment.
Although a prolonged post-operative nil-by-mouth period was not correlated with weight change over admission, it was adversely associated with a longer length of stay. The present study also found that an extended period before a patient achieved adequate nutrition was also associated with prolonged length of stay. An important limitation of the present study was that we were unable to control for the impact disease severity may have had on length of stay; hence, it is possible that poor nutritional intake may be a marker for an unwell patient who unsurprisingly would be expected to require a longer period of hospitalisation. In addition, although the correlations reached significance, the r-values were modest, which suggests that other factors may be influencing length of stay and weight change that were beyond the scope of the present study with respect to measurement. Our findings, however, are consistent with previous studies showing that a post-operative delay in the recommencement of nutrient intake and achievement of adequate nutrition can contribute to longer hospital stays (Shaw-Stiffel et al., 1993). Similarly, several studies have indicated that early post-operative feeding after GI surgery is associated with shorter hospital stays (Stewart et al., 1998; Lewis et al., 2001; Andersen et al., 2006). Recent evidence suggests that nationwide collaborative programs can assist with implementing change towards early feeding in this patient group; hence, such hospital-wide programs are viable to consider implementing (Maessen et al., 2009).
Of the malnourished patients identified in the study, 75% were seen by a dietitian post-operatively. Although it was encouraging that these dietitian referral rates were high, the finding that one quarter of malnourished patients were not referred to the dietitian indicates that processes for referral of malnourished patients to a dietitian could be further improved. Importantly, however, patients who were seen by a dietitian post-operatively consumed more energy and protein on the second day of a soft ward diet than patients who were not seen, indicating that dietetic intervention can have a positive influence on nutrient intake in an already ‘at risk’ group. There is limited research regarding malnutrition and its effect post discharge; however, the available research suggests that patients who have undergone GI surgery continue to experience nutritional problems post discharge, thus adversely affecting nutritional status and clinical, functional and quality of life outcomes (Beattie et al., 2000). Unfortunately a more detailed investigation into the effect of post-operative dietetic intervention on the clinical outcomes of the patient group under investigation was beyond the scope of the present study and would be a valid area of future research.
Our findings showing an association between nutritional risk and adverse clinical outcomes broadly agree with international findings reported in larger studies. Our small sample size of patients who underwent SGA assessment may have reduced the strength and ability to detect the influence of malnutrition on all clinical outcomes measured. Similarly, the data that was collected retrospectively may have been subject to unintended bias or inaccuracies that may have limited the strength of the relationships investigated. This included the relationship found between preoperative weight loss and adverse clinical outcomes. It is interesting to note that it has been previously proposed that the rate and timing of preoperative weight loss is a better predictive factor of postoperative complications than the underlying diagnosis (Detsky et al., 1987b).
It was beyond the scope of the present study to control for the patients’ medical condition or to stratify the clinical outcomes based on time from diagnosis to surgical intervention. However, because all of the patients included in the study were undergoing Category One surgery, they were required to be admitted for surgery within 30 days of presenting to the surgical specialist, minimising the variability in timing between diagnosis and procedure. Finally, the time frame for the study only covered the period from initial presentation to discharge and long-term outcomes and readmission rates were not assessed. In future, it may be useful to follow up patients discharged from hospital to further investigate their clinical, functional and quality of life outcomes.
The findings obtained in the present study show that high rates of malnutrition in GI and colorectal cancer patients are evident in this study population. Poor nutritional status coupled with delayed and inadequate post-operative nutrition practices were associated with worse clinical outcomes. Providing timely and appropriate nutritional intervention may have a positive effect on the nutritional status and clinical outcomes of patients with GI cancer undergoing surgery. More local studies that control for disease condition and surgical intervention are justified to further explore this situation and enable the development of effective nutrition protocols to improve current post-operative nutrition practices.
The authors wish to thank Alison Bowie and Sandy Mason for their assistance.
Conflict of interest, funding and authorship
None of the authors have any conflict of interest in presenting this work.
No funding declared.
AG carried out the data collection, data analysis and drafted the manuscript. CN conceived the study, participated in the design and coordination of the study, was involved with the data collection and analysis and helped draft the manuscript. NS participated in the design of the study and helped draft the manuscript. TC participated in the design of the study, assisted with data analysis, and helped draft the manuscript. All authors critically reviewed the manuscript and approved the final version submitted for publication.