Materials and methods

PATIENTS

Two groups of patients were recruited to the study by mail. The first group comprised 58 patients with intestinal failure (corresponding to approximately 75% of all patients in Denmark7 receiving or having initiated HPN during the period November 1995 to March 1997). The other group comprised 76 non-HPN patients with intestinal insufficiency, defined as faecal energy excretion of more than 2.0 MJ/day (measured at a previous admission), a remnant small intestine of 200 cm or less (measured peroperatively from the ligament of Treitz) or as having undergone consecutive small intestinal resections exceeding 150 cm.

Forty six of the 58 HPN patients and 45 of the 76 non-HPN patients with intestinal insufficiency consented to participate in the study. One of the non-HPN patients was excluded as she experienced gastroenteritis and vomiting during admission. Also, one HPN patient with protein losing enteropathy was excluded as she used her catheter for intravenous calcium supplements only. All HPN patients received parenteral supplements of nutrients or saline at least three times a week. Thus the study comprised 45 HPN and 44 non-HPN patients. None of the patients with inflammatory bowel disease showed evidence of active disease.

STUDY PROTOCOL

Patients were admitted for a period of 2.5 days. On arrival on the first afternoon, patients were given three containers and an electronic precision balance (measuring in grams). Two of the containers were designated for collection of faeces and urine, respectively. In the third container, patients were told to collect a duplicate of their oral intake (both liquid and solid foods). When instructed, patients were told that they could eat what they pleased from breakfast, lunch, and supper buffets, each containing a wide range of food items. The beverages available included water, tea, coffee, milk products, soft drinks, juice, glucose-saline solutions, etc. Sandwiches, biscuits, and beverages were available in the kitchen between meals.

Patients were told to abstain from food between 2200 and 0800 on the day of admission so that faecal excretion during the first 24 hours of the study would not be a reflection of intake prior to the balance period. The study and collection period began at 0800 on the second day of admission and patients were requested to urinate, defecate, or empty their stoma bags. During the next 48 hours, patients collected their faeces, urine, and duplicate diets in the three containers. Before admission, intestinal transit was measured by timing a Brilliant Blue marker from oral administration to its appearance in the faeces of non-HPN patients and HPN patients with a preserved colon. These measurements revealed intestinal transit times of less than 10 hours, except in three HPN patients with oral failure and intestinal dysmotility. Thus the limited duration of intestinal transit indicated that 10 hours of fasting between 2200 and 0800 were sufficient so that the events prior to the balance study did not influence the 48 hour balance period. On retrieval, faeces were deposited on ice, immediately frozen, and kept at −20°C until the samples were analysed.

Patients were interviewed about the composition and volume of parenteral support, and information on daily medicine use was obtained. During admission, patients received their usual parenteral supplements and medication. None of the patients was receiving sodium chloride capsules, but patients drinking a glucose-saline solution were told to place a similar portion in their food container. None of the patients required tube feeding (enteral nutrition).

ANALYTICAL METHODS

Patient height and fasting body weight were measured on admission. Body composition was measured during or before admission by dual energyx ray absorptiometry (Norland XR-26 DXA densitometer, Norland Corp., Fort Atkinson, Wisconsin, USA). In addition, patient weight six months before admission was obtained from weight curves drawn at ambulatory visits. Basal metabolic rate (BMR) was calculated according to the Harris-Benedict equations using actual body weights.8

The weight of the 48 hour oral intake was measured together with faecal and urinary losses. Intestinal wet weight absorption was calculated as equivalent to the difference between the weight of the oral intake and faecal weight. Analyses of the diet and faeces were performed on homogenised, freeze dried samples. Dietary and faecal energy contents were determined by bomb calorimetry using approximately 1 g of freeze dried sample which was combusted in an IKA adiabatic calorimeter (model C 4000 A; IKA-Analysentechnik, Heitersheim, Germany). Intestinal energy absorption was calculated as equivalent to the difference between ingested and excreted energy, and relative absorption as diet-faecal energy/diet energy×100%. The precision of the analytical methods used was demonstrated previously as high, with high reproducibility.9 Questionnaires regarding habitual dietary intake of patients were not used. Weight and energy content in parenteral supplements was calculated from information given by the manufacturers. The remnant small intestine was measured peroperatively in 56 patients whereas the length of resection was measured in 31 patients. One of the non-HPN and two of the HPN patients did not have resection of the small intestine. The length of the colon was expressed in terms of percentage of the usual length according to the method of Cummings and colleagues.10

ETHICS

The study was approved by the Ethics Committee for Medical Research, Copenhagen, Denmark. The procedures followed were in accordance with the ethical standards of the Helsinki Declaration of 1975, as revised in 1983. All patients gave informed consent before entering the study.

STATISTICAL ANALYSIS

Results are expressed as median (25–75% percentiles). Differences between groups were tested using the Mann-Whitney rank sum test and comparisons of frequencies were evaluated by the χ² test or alternatively by Fisher's exact test using SigmaStat for Windows version 2.0 (copyright 1992–1995, Jandel Corporation, Erkrath, Germany); p<0.05 was considered significant.

Results

PATIENT DEMOGRAPHICS

Patient demographics are shown in table 1. Most patients had Crohn's disease (37 of 44 in the non-HPN group and 26 of 45 in the HPN group). The two groups had a similar BMR, calculated from the Harris-Benedict equations. Sex and age distributions did not differ between groups. The remnant gut structures of patients are described in table 2.

ENERGY

Median dietary energy intake was considerably higher in the non-HPN compared with the HPN patients (11.69 MJ/day (∼2780 kcal/day) (range 6.71–24.42 MJ/day) compared with 7.56 MJ/day (∼1800 kcal/day) (range 5.76–21.13 MJ/day); p<0.001) (table 3), corresponding to 178% and 134% of BMR. In the HPN patients, parenteral support was 3.86 MJ/day, corresponding to 72% of BMR, and as a result the energy provided by the diet in the non-HPN patients closely matched the sum of enteral and parenteral nutrition in the HPN patients (11.69 MJ/day (178% of BMR) and 11.30 MJ/day (213% of BMR), respectively; p=0.98) (table 3). Faecal energy excretion did not differ significantly between the two groups and consequently the abundant intake of the non-HPN patients resulted in intestinal absorption of energy (7.92 MJ/day, 1885 kcal/day) that was twofold higher than in the HPN patients (3.96 MJ/day, 940 kcal/day; p<0.001). Energy absorption as a percentage of BMR was 71% in the HPN patients and 130% in the non-HPN patients (p<0.001). Intestinal absorption, when added to parenteral supplies, resulted in an overall balance of 7.97 MJ/day (141% of BMR) in the HPN patients, equivalent to intestinally absorbed energy of 7.92 MJ/day (130% of BMR) in the non-HPN group (table 3).

WET WEIGHT

Intestinal wet weight absorption was threefold higher in the non-HPN patients (2.48 kg/day) compared with the HPN patients (0.84 kg/day; p<0.001) (table 4). Oral intakes were 3.56 and 2.81 kg/day in the non-HPN and HPN patients, respectively. Parenteral fluid supply was 2.25 kg/day. The overall fluid balance was approximately 1 litre/day in both groups because urine volume in the HPN patients (2.23 kg/day) was twice that of the non-HPN patients (1.16 kg/day; p<0.001) (table4).

SODIUM

Dietary sodium intake was similar in the non-HPN and HPN patients (135 and 134 mmol/day, respectively; p=0.24) (table 5). Parenteral support provided the HPN patients with 177 mmol/day of sodium, resulting in a significantly higher total supply compared with the non-HPN patients (302 v 135 mmol/day; p<0.001). Faecal sodium excretion was higher in HPN patients (129v 105 mmol/day; p<0.05) (table 5) whereas intestinal sodium absorption was significantly lower in the HPN patients compared with that found in the non-HPN patients (−14v 32 mmol/day; p<0.01). Urinary sodium excretion was almost fourfold higher in the HPN patients (151 mmol/day) compared with the non-HPN patients (41 mmol/day; p<0.001), and overall sodium balance was positive (29 mmol/day) compared with the non-HPN patients (−31 mmol/day; p<0.001) (table 5).

THE BORDER BETWEEN INTESTINAL INSUFFICIENCY AND INTESTINAL FAILURE

The border between intestinal insufficiency in the 44 non-HPN patients and intestinal failure in the 45 patients who received HPN was defined as the 5% lower confidence interval of energy and wet weight absorption found in the non-HPN patients. This arbitrary confidence interval was found to be reasonable. This definition was based solely on data obtained from non-HPN patients to avoid confounding factors from HPN, which may reduce dietary intake and consequently the amount of energy and wet weight absorbed in patients receiving HPN.

The 5% lower limit of energy absorption was calculated as equivalent to 84% of BMR (fig 1). The 5% lower limit of wet weight absorption was 1.41 kg/day (fig 1). Using the calculated cut off values, fig 1could be divided into four areas. Patients in the upper right quadrant were those whose intestinal absorption was above the limit (i.e. where HPN was not necessary). Patients located to the left of a wet weight absorption of 1.41 kg/day were those requiring saline supplements, whereas those with an energy absorption of less than 84% of BMR required parenteral energy supplements. Patients in the lower left quadrant had a combined need for saline and energy. Thus 5% of the non-HPN patients had a BMR below 84% and 5% were below the 1.41 kg/day cut off; this was equivalent to four non-HPN patients, of whom three were borderline (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Absorption of wet weight and energy in relation to basal metabolic rate (BMR) calculated by the Harris-Benedict equations in 44 non-HPN patients and in 45 HPN patients. The 5% confidence limits of the non-HPN patients, defining intestinal failure, are given by the lines. Energy absorption/BMR was 84% and wet weight absorption 1.41 kg/day.

Subsequent insertion of the 45 HPN patients into this frame (fig 1) demonstrated that the borders between intestinal insufficiency and failure defined by the non-HPN patients resulted in a good distinction between non-HPN and HPN patients. Only five HPN patients were located in the upper right quadrant in the area of non-HPN patients.

ORAL INTAKE AND PERCENTAGE ABSORBED

The data presented in fig 1 do not disclose if intestinal failure is due to insufficient absorption of a sufficient amount of food or insufficient intake of food (oral failure). The purpose of figs 2 and 3was to differentiate between these two different clinical problems, both resulting in intestinal insufficiency or failure.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Relative energy absorption in relation to dietary energy intake/basal metabolic rate (BMR) in 44 non-HPN patients and 45 HPN patients. The open triangles indicate HPN patients who received HPN due to a wet weight absorption of less than the 5% limit (1.41 kg/day).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Relative wet weight absorption in relation to the dietary wet weight intake in 44 non-HPN patients and 45 HPN patients. The open triangles indicate HPN patients who received HPN because of energy absorption less than the 5% limit for basal metabolic rate (BMR) (84%).

Figure 2 illustrates the relation between percentage of the diet absorbed (diet-faecal energy/diet energy×100%) and adequacy of dietary intake (dietary energy as a percentage of BMR). The limits of 84% of BMR in fig 1, used to distinguish between intestinal insufficiency and failure, were fitted to the plots from the HPN and non-HPN patients. Several of the HPN patients had an energy intake of less than 84% of BMR and therefore even 100% absorption would not help their dependence on HPN as long as energy intake was that low. Conversely, several of the non-HPN patients had compensatory increased energy intake above 200% of BMR (hyperphagia) which meant that they did not require HPN despite absorption of less than 50%. In fact, one patient had a relative energy absorption (diet-faecal energy/diet energy×100%) of only 24% but consumed 22.18 MJ/day and thus managed to avoid HPN. Fourteen of the HPN patients absorbed more than the limit of 84% of BMR. However, nine of these patients (indicated by the open triangles in fig 2) had an isolated problem due to wet weight absorption of less than 1.41 kg/day and did not receive HPN because of insufficient energy absorption. Four of the remaining five patients were borderline and absorbed less than 105% of their BMR (see below). The last outlying patient received no parenteral energy and less than 1 litre of saline daily.

Figure 3 illustrates the relation between intake of wet weight and percentage absorbed. The limit of 1.41 kg/day in fig 1 to discriminate between intestinal insufficiency and failure was fitted. Some patients had good absorption above 50% but an intake of less than 1.5–3 kg/day, making them dependent on parenteral saline. Other patients who only absorbed 25–50% did not need parenteral saline as their intakes were 3–7 kg/day. However, as illustrated in fig 1, the cut off (defined by a final absorption of more than 1.41 kg/day) was accurate in differentiating between HPN and non-HPN patients over a considerable range of intakes. Thirteen HPN patients absorbed more than the limit of 1.41 kg/day for wet weight absorption. Eight received HPN to compensate for energy absorption below the upper limit of 84% of BMR (indicated by the open triangles). The remaining five were the patients described above.

Discussion

Adequate gut function is needed to maintain intestinal autonomy thus avoiding long term parenteral nutrition. Two principal gut functions are absorption of energy from fat, carbohydrates, and protein sufficient to meet the metabolic needs of the body, and absorption of wet weight and electrolytes, mainly sodium, sufficient to avoid dehydration and electrolyte depletion.

In this study, we invited our total cohort of 58 patient with intestinal failure7 (HPN patients, defined as patients in need of HPN at the time of the study) and all of our 76 patients with intestinal insufficiency (non-HPN patients, defined as patients with a short small bowel of less than 200 cm or considerable malabsorption of at least 2 MJ/day, or both) to participate. A central referral system connecting the small local hospitals throughout Denmark to the intestinal failure unit at the Rigshospitalet ensured referral of patients with the most severe intestinal insufficiency who nevertheless were able to manage without HPN. The non-HPN and HPN patients are described in details in tables 1 and 2. It is important to emphasise that the actual remnant bowel length was disregarded as it was irrelevant to measurements of intestinal function and to the limits defined. Clinicians in charge of patients had no access to the study results and treatment was guided by clinical evaluation of the patients, and by weight, blood tests, and occasional measurements of faecal and urine volumes. In total, 46 HPN patients and 45 non-HPN patients were included.

The main aim of the study was to define the border between intestinal insufficiency and failure using objective measurements of intestinal function (i.e. to identify the minimal level of gut function necessary to avoid long term parenteral support). Intestinal absorption is often measured as a percentage of intake which (as shown in figs 2 and 3) lacks strict clinical bearing because high absorption of an insufficient amount of food still results in intestinal failure.

The border between intestinal insufficiency and failure was a wet weight absorption of 1.41 kg/day (i.e. patients not able to absorb this quantity would most often need parenteral saline supplements). Most clinicians in charge of these patients would probably agree that 1.5 litres of water seems to be a reasonable minimal amount to compensate for fluids lost from sweat and urine. Because the wet weight absorption includes absorption of approximately 0.40 kg/day of dry weight, fluid absorption is even less. However, one has to bear in mind the production of water by oxidation of nutrients of approximately 0.03 kg/MJ, thus accounting for approximately 0.27 kg/day in the non-HPN patients.

Surprisingly, the calculated 5% lower confidence level for intestinal sodium absorption was −85 mmol/day in the non-HPN patients, and their median overall balance was −31 mmol/day (corresponding to less than −2 g/day) (table 5). A plausible explanation of the overall negative sodium balance in the non-HPN patients may be a low degree of patient compliance regarding the addition of salt to the solid food containers.

The border between intestinal insufficiency and failure was 84% of BMR. At this level, absorption would have to cover BMR and in addition at least 20–30% to perform daily activities. Even though these patients are known to compensate by low physical activity, it does not explain how they are able to survive at such a low level of energy absorption. Actual BMR was not measured but our data suggest that it was lower in the non-HPN patient closest to the borderline value compared with the estimated BMR calculated by the Harris-Benedict equations. Similar findings have been demonstrated in patients with anorexia nervosa where the resting energy expenditure was reduced to 49–91% of values predicted by the Harris-Benedict equations.11-13

Figure 1 does not demonstrate oral intake and consequently does not discriminate between patients with oral failure (e.g. pseudo-obstruction) who may or may not absorb a high percentage of the small amount of food they eat, and short bowel patients, who often absorb a low percentage of a large food intake. Thus to describe the spectrum between these two extremes, the percentage of absorption was plotted against dietary intake for energy (fig 2) and wet weight (fig3).

Figure 2 clearly demonstrates that many of the non-HPN patients were able to compensate by hyperphagia, as previously described.14 15 Thirty six percent had an energy intake of 200–419% of their BMR. Examining fig 2, it is clear that absorption expressed as a percentage does not help in the classification of a patient before the amount of intake is known, except for those who absorbed less than 25% of their intake. Patients may absorb up to 100% but if energy intake is less than 84% of BMR they will have intestinal failure. Many of these patients had dysmotility disorders combined with pseudo-obstruction. However, if patients who absorb between 100% and 50% for example, could eat a minimum of 84–168% of their calculated BMR, respectively, they may be weaned from HPN and regain intestinal autonomy. Patients with an absorption of 50–25% would correspondingly have to eat at least 168–336% of their BMR every day to avoid HPN. The same principles regarding wet weight rather than energy absorption were considered in fig 3. Hence the position of a given patient, non-HPN or HPN, in figs 2and 3 illustrates the efforts which have to be made by the patient to discontinue or remain on HPN. Patients whose energy or wet weight absorption is less than 25% are unlikely to be weaned from HPN.

Rodrigues et al investigated energy absorption as a measure of intestinal failure in the short bowel syndrome by bomb calorimetry.6 They measured energy absorption from a liquid meal over a six hour period. A median of 87% (range 82–92%) of the energy was absorbed in a group of five patients with ileostomy. Seven short bowel patients not on HPN absorbed 67% (range 59–78%), and five short bowel patients receiving HPN absorbed 27% (range 2–63%). Messing et al reported a mean coefficient of digestive absorption of energy in five HPN patients of 64% (range 41–85%).16 A median of 71% of energy was absorbed by the non-HPN patients in our study (table 3) but the range was wide (24–86%; fig 2). The HPN patients absorbed 49% of dietary energy (table 3), with values ranging from a net loss of energy of 40% to complete absorption of 100% (fig 2). Therefore, percentage absorption values obtained in different studies are characterised by a wide range, and thus medians are highly dependent on patient selection.

Nightingale et al focused on wet weight absorption in 15 patients with a short residual length of jejunum and no colon.17 In six of the patients who did not require HPN, wet weight absorption was 1.56–2.42 kg/day, closely matching the lower limit in our study. Furthermore, all nine patients who received parenteral supplements were located below the limit suggested.

In summery, patients with intestinal insufficiency who maintained intestinal autonomy and did not depend on parenteral supplements tended to have a wet weight absorption of more than 1.41 kg/day and an energy absorption exceeding 84% of their BMR, as calculated by the Harrison-Benedict equations. Patients with intestinal failure, by definition depending on HPN, tended to absorb less of either or both. Preservation of intestinal autonomy was consistent with a large variation in relative intestinal absorption of 25–100% of either wet weight or energy, as long as oral intake was proportionally increased to meet the total demand, as outlined. Hence the border between intestinal insufficiency and failure was narrow and well defined. Longitudinal studies of the Danish cohort of HPN patients have shown that adaptation occurs in a considerable proportion of patients who are weaned off HPN temporarily or permanently over a period of five years.7 Future prospective studies will assess the value of the measurements of energy and wet weight absorption for predicting if patients with intestinal failure will require lifelong HPN. Furthermore, the short term balance tests may become important educational tools for borderline and selected patients in whom weaning from HPN is desired.