点击显示 收起
1 From the Department of Gastroenterology, Copenhagen University Hospital, Rigshospitalet, Copenhagen (KVH, PBJ, PBM, and MS), and the Department of Endocrinology, Hvidovre Hospital, Hvidovre, Denmark (HAS).
2 Supported by The P.A. Messerschmidt and Wife Foundation and The Danish Hospital Foundation for Medical Research, Region of Copenhagen, The Faeroe Islands and Greenland. 3 Address reprint requests to KV Haderslev, Department of Gastroenterology CA 2121, Copenhagen University Hospital, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Denmark. E-mail: khaderslev{at}dadlnet.dk.
ABSTRACT
Background: Patients who have undergone resection of the small intestine have lower body weight than do healthy persons. It remains unclear whether it is the body fat mass or the lean tissue mass that is reduced.
Objective: We compared body-composition values in patients who had undergone small-intestinal resection with reference values obtained in healthy volunteers, and we studied the relation between body-composition estimates and the net intestinal absorption of energy.
Design: In a cross-sectional study, we included 20 men and 24 women who had undergone small-intestinal resection and had malabsorption of energy > 2000 kJ/d. Diagnoses were Crohn disease (n = 37) and other conditions (n = 7). Body composition was estimated by dual-energy X-ray absorptiometry, and data were compared with those from a reference group of 173 healthy volunteers. Energy absorption was measured during 48-h balance studies by using bomb calorimetry, and individual values were expressed relative to the basal metabolic rate.
Results: Body weight and body mass index in patients were significantly (P < 0.05) lower than the reference values. Fat mass was 6.4 kg (30%) lower (95% CI: -8.8, -3.9 kg), but lean tissue mass was only slightly and insignificantly lower (1.5 kg, or 3.3%; 95% CI: -3.7, 0.60 kg). Weight, body mass index, and body-composition estimates by dual-energy X-ray absorptiometry did not correlate significantly with the net energy absorption relative to the basal metabolic rate, expressed as a percentage.
Conclusions: Patients who had undergone small-intestinal resection had significantly lower body weights and body mass indexes than did healthy persons, and they had significant changes in body composition, mainly decreased body fat mass.
Key Words: Body composition • small-intestinal resection • dual-energy X-ray absorptiometry • fat mass • lean tissue mass
INTRODUCTION
Crohn disease, ischemic bowel infarction, and complications after surgery are some of the most frequent reasons for extensive resection of the small intestine. Patients who have undergone that procedure often have a lower body weight and body mass index (BMI; in kg/m2) than do healthy persons (1). This difference is thought to result mainly from malabsorption of nutrients due to a reduced area of absorptive surface, although a reduced food intake due to anorexia or abdominal pain and the restriction of dietary fat intake to diminish diarrhea may also contribute in some patients.
Many studies examined whole-body composition in terms of fat and lean tissue masses in patients who undergo small-intestinal resection (1–9), but only a few have included data from healthy volunteers for direct comparison. It remains unclear whether it is principally the body stores of fat mass (FM) or the lean tissue components such as muscle mass or total body water that are reduced in patients after small-intestinal resection. Some studies showed that lean tissue mass (LTM) is reduced more than is FM (6, 7), whereas other studies have shown a reduction predominantly in FM (8, 9).
Patients who undergo small-intestinal resection may overcome the nutritional effect of malabsorption of energy by substantially increasing oral intake (8), but the possibility for hyperphagia to fully compensate for fecal losses of energy varies among individuals. If the energy requirements of the body are not met by a sufficient net absorption of energy, loss of body weight and changes in body composition may ensue. To date, the relation between net energy absorption and body composition in patients who have undergone small-intestinal resection has not been explored.
Several techniques are currently available for the accurate assessment of body composition (10), but some are associated with a high dose of irradiation or are unpleasant for the patient, and others are laborious or require access to advanced technical equipment. Dual-energy X-ray absorptiometry (DXA) is a relatively new method of body-composition analysis that has gained wide acceptance because of its high degree of accuracy, safety, and convenience (11).
The purposes of the present cross-sectional study were to investigate body composition by means of DXA in weight-stable patients with a history of at least one intestinal resection procedure and to determine whether body-composition variables in these patients deviate significantly from those in a healthy reference group. We also assessed the extent to which the magnitude of these alterations in body composition was related to net energy absorption.
SUBJECTS AND METHODS
Patients and reference material
This work was carried out in conjunction with a study that aimed to classify intestinal failure on the basis of energy and wet weight absorption (12). Patients were eligible for entry to the study if they had a fecal energy excretion of > 2.0 MJ/d (measured at a previous admission), which corresponds to a relative energy loss of > 15–20% of total daily energy consumption. Patients with a remnant small intestine of ≤ 200 cm, measured intraoperatively from the ligament of Treitz, or in whom > 150 cm of the small intestine was resected were assumed to have a fecal energy excretion of > 2.0 MJ/d and were also included.
Exclusion criteria were a current need for parenteral nutrition, evidence of inflammatory activity in patients with inflammatory bowel disease, and current steroid treatment. Patients were also excluded if the time since their last small-intestinal resection was < 2 y. On the basis of information obtained from the case records, 76 patients were selected and invited by letter to participate in the study; 44 patients (20 men and 24 women, of whom 9 were postmenopausal) accepted. Clinical assessment included data concerning time since the last small-intestinal surgery, the length of the remaining bowel (based on specification in the case records), and body weight 6 mo before admission (obtained from weight curves drawn at ambulatory visits). The basal metabolic rate (BMR) was calculated according to the Harris-Benedict equation with the use of the actual body weights (13).
The patients were compared with a group of healthy volunteers described previously (14); briefly, the control group comprised 173 healthy subjects, 84 men and 89 women aged 20–79 y who were recruited from among staff members at the Hvidovre University Hospital. In all healthy volunteers, anthropometry and measurements of body composition had been performed with the use of procedures and a DXA instrument (and software) similar to those used in this study.
The Ethics Committee for Medical Research in Copenhagen approved the study protocol, and the study was conducted in accordance with the Declaration of Helsinki of 1975, as revised in 1983. Written and oral informed consent were obtained from all patients before inclusion.
Study protocol
Anthropometric measurement was performed in all patients after an overnight fast; all patients were weighed (after voiding and defecating or emptying their stoma bags) on a calibrated digital scale accurate within 0.1 kg while the subjects were wearing light clothes, and their heights were measured with the use of a wall-mounted stadiometer to the nearest 0.1 cm. Body composition was then measured by DXA as described below. During the next 48 h, the patients collected their feces and a duplicate of the unrestricted diet in 2 separate containers for measurement of ingested and excreted energy.
Analytic methods
Dual-energy X-ray absorptiometry
Measurements of body composition were performed with the Norland XR-36 DXA densitometer (Norland Corporation, Fort Atkinson, WI). The host software used was version 2.5.2, and the scanner software used was version 2.0.0. The theory and method for measuring body composition by DXA were described previously (11). Briefly, while lightly dressed, the patient lay on the scan table for 20 min, and transverse scans < 1 cm apart were performed from head to foot. The instrument uses X-rays of 2 distinct energy levels that are attenuated to different extents by fat, bone, and lean mass. By computerizations of data inputs from < 11 000 pixels, DXA estimates body composition on the basis of a tree-compartment model, measuring total-body bone mineral content (TBBMC), FM, and nonskeletal fat-free tissue mass (ie, LTM). The fat-free mass measured by DXA is the sum of the nonfatty components of the body, namely, TBBMC and LTM. The accuracy (SEEs) of in vivo measurements by DXA are 2.9%, 1.9 kg, and 2.7 kg for percentage FM (%FM), FM, and LTM, respectively (15). Hendel et al (16) reported precision errors (CVs) of body-composition measurements taken with the Norland XR-36 DXA densitometer: 2.2% for TBBMC, 2.7% for fat-free mass, and 2.6% for %FM. In our hands, the between-measurement CVs of TBBMC, LTM, and FM were 1.5%, 1.6%, and 3.9%, respectively (17).
Analysis of energy in diet and feces
On retrieval, the containers with diet and feces were placed on ice, immediately frozen, and stored at -20 °C. Dietary and fecal energy contents were determined by bomb calorimetry with the use of < 1 g homogenized freeze-dried sample, which was combusted in an adiabatic calorimeter (model C 4000 A; IKA-Analysentechnik, Heitersheim, Germany). The net intestinal energy absorption was calculated as being equivalent to the difference between ingested and excreted energy and was expressed relative to the individual BMR, to obtain an estimate independent of height, weight, and sex. The minimum energy requirement in healthy adults is considered to be > 130–140% of the BMR (18).
Statistical analysis
Results are expressed as means ± SDs. For each patient, reference values for BMI, weight, fat-free mass, FM, %FM, and TBBMC were interpolated from the age- and sex-stratified reference group data with a linear adjustment for differences in height (no adjustment for height differences was done for BMI and %FM values). Each value of LTM, FM, %FM, TBBMC, BMI, and weight was also converted to a z score [(patient value - age- and sex-matched reference values)/SD of the age- and sex-matched reference group]. Student’s paired t test was used for comparison between patients’ values and reference values, and Student’s unpaired t test was used to investigate differences between the men and the women. Student’s one-sample t test was used for calculating 95% CIs.
For between-group comparison, we used a one-way analysis of variance with test for linear trend, and the Bonferroni method was used for post hoc analysis if the P value of the overall test was < 0.05. Pearson’s correlation coefficients and linear regression analysis were used to establish association between variables. All statistical tests were two-tailed, and a P value of < 0.05 was considered statistically significant. SPSS statistical software (version 11.0; SPSS Inc, Chicago) was used for all analyses.
RESULTS
Patient demographics
A total of 44 patients (20 men and 24 women, of whom 9 were postmenopausal) were included in the study. The mean age of the study population was 50.6 ± 12.3 y. Diagnoses were Crohn disease (n = 37), mesenteric thrombosis (n = 3), and complications after surgery (n = 4). The mean time since the last bowel resection was 11.0 ± 7.5 y. On average, the length of the remaining small intestine was 201 ± 74 cm, and the length of the remaining colon expressed as a percentage of the original length was 35 ± 40%. The study population was considered weight-stable as documented by an essentially unchanged body weight for 6 mo.
The average energy intake was 12 200 ± 4060 kJ/d (range: 8120–24 420 kJ/d for men, 6700–16 180 kJ/d for women), and the mean fecal energy excretion was 4122 ± 3001 kJ/d, which corresponded to an average absorption of 68 ± 13% of the ingested energy. The absorption expressed relative to the BMR was on average 134 ± 34% (the minimum energy requirement in healthy adults is > 130–140% of the BMR; 19, 20). The mean fecal weight was 1416 ± 1011 g/d, which is significantly above the normal range (0–200 g/d). The patients absorbed on average 48% of the ingested fat and excreted a mean of 48 ± 40 g/d (normal range: 0–7 g/d).
Anthropometric measures and body-composition variables by dual-energy X-ray absorptiometry
The anthropometric measures and body-composition variables (LTM, FM, %FM, and TBBMC) by DXA in the men and the women, along with the calculated age- and sex-matched and height-adjusted reference values for healthy volunteers, are shown in Table 1. The weight by DXA correlated smoothly with weight by scale (r = 0.993, SEE = 1.53 kg, P < 0.001), and the intercept did not differ significantly from zero. Weight by scale and BMI were significantly lower in both the men and the women compared with the reference values of healthy volunteers, and there were no significant differences in the magnitude of these decreases between the sexes. On average, the weight by scale was 7.6 kg lower in the women and 7.4 kg lower in the men, which corresponded to a difference of 12.0% and 9.3%, respectively. The difference in weight by scale correlated excellently (r = 0.990, SEE = 1.46 kg, P < 0.001) with that by DXA, which was 7.7 kg in the women and 8.5 kg in the men. By category of BMI, 29.5% of the patients had a BMI < 20 (low), 50% had a BMI of 20–25 (normal), 20.5% had a BMI > 25 kg/m2 (overweight), and only one patient had a BMI slightly > 30 (obese).
View this table:
TABLE 1 . Comparison of anthropometric measures and body-composition values analyzed with dual-energy X-ray absorptiometry in 44 patients who have undergone small-intestinal resection with reference values from paired healthy volunteers1
The FM and %FM measurements by DXA were significantly lower in both the men and the women, with no significant differences between sexes in the magnitude of difference in these variables. The FM tissue was ≤ 30% lower in both sexes, corresponding to %FM that was 20% lower. The difference in FM of 6.4 kg (95% CI: -8.8, -3.9 kg) accounted for > 80% of the difference in weight between patients’ values and references values. Conversely, the average difference in LTM of 1.5 kg (95% CI: -3.7, 0.60 kg), which was 3.3% lower, was not statistically significant. TBBMC was lower by a mean of 6.6% (P < 0.05) in the men and by slightly less in the women (mean difference = 2.7%, P > 0.05). To confirm the appropriateness of the linear adjustment of differences in height between patients and the healthy volunteers, we also analyzed the data by using analysis of covariance with height entered as a covariate, and this analysis found similar results. When the results for the men and the women were combined in the analysis, FM, %FM, and TBBMC z scores were significantly (P < 0.05, for all comparisons) lower than those for the reference values, whereas the z score for LTM in patients was only slightly and insignificantly lower. We found no significant correlations between any of the body-composition estimated z scores and the time since the last intestinal resection or the length of the remaining small intestine.
The influence of the intestinal net energy absorption relative to the basal metabolic rate on body weight, BMI, and body-composition variables (z scores)
We found no significant correlations between the intestinal net energy absorption (NEA) relative to the BMR, expressed as a percentage [(NEA/BMR)%], and the body weight, BMI, or body-composition variables (DXA z scores). To further explore the relation, the data were reduced by grouping patients according to the (NEA/BMR)% by increments of 20%. The clinical characteristics of patients in the resulting 5 groups are given in Table 2. The bowel anatomy among the patients in the 5 groups did not differ statistically, and patients ingested the same amount of energy, except for those who absorbed energy at a rate > 160% of their BMR, who had a significantly greater energy intake. The relations between the (NEA/BMR)% and body weight, BMI, and body-composition variables (z scores) are shown in Figure 1. The within-group variation was very large, as indicated by the large 95% CIs, and no significant differences between groups could be detected. The curves connecting the mean values in each group indicated a weak positive relation between the LTM, TBBMC, weight, and BMI z scores and (NEA/BMR)%. Thus, a test for linear trend across groups showed a significant positive trend between (NEA/BMR)% and TBBMC, and a positive trend between (NEA/BMR)% and LTM and weight that was nearly significant. The data also showed that, for body weight, BMI, and most body-composition variables, the average reference value from healthy volunteers (z score = 0) was reached when absorption was > 140% of the BMR.
View this table:
TABLE 2 . Clinical characteristics of 44 patients who have undergone small-intestinal resection1
FIGURE 1. . SD scores (z scores) for weight, BMI, and body-composition variables measured by dual-energy X-ray absorptiometry in 44 patients who had undergone small-intestinal resection in relation to the net energy absorption relative to the basal metabolic rate, expressed as a percentage [(NEA/BMR)%]. Each symbol indicates the group mean value, and the bars represent the 95% CIs. The dotted lines indicate average values for age- and sex-matched healthy volunteers. There were no significant differences among the groups based on one-way ANOVA, but a significant linear trend across groups was observed between total-body bone mineral content (TBBMC) and (NEA/BMR)%.
LTM correlated highly significant with BMR (r = 0.83, SEE = 6.6 kg, P < 0.001), and the correlation was even more significant after correction for differences in sex (r = 0.86, R2 = 0.75, SEE = 5.9 kg, P < 0.001). Thus, 75% of the variation in the LTM can be explained by variations in BMR and sex. Likewise, LTM correlated significantly with the amount of energy absorbed (r = 0.80, SEE = 7.3 kg, P < 0.001), and even more so after correction for differences in sex (r = 0.83, R2 = 0.69, SEE = 6.6 kg, P = 0.002).
DISCUSSION
This study showed that the mean body weight of patients who had undergone small-intestinal resection was 7.9 kg below that of age- and sex-matched healthy volunteers (P < 0.05). Measurements of body composition by DXA showed that changes in FM accounted for > 80% of the total difference between the values for body weight. Thus, the FM in the patients was significantly lower (< 6.4 kg) than that in the healthy volunteers, and, because of the small size of the FM compartments, this corresponded to a difference of ≤ 30%. Conversely, the LTM of patients was on average only slightly and insignificantly lower (1.5 kg), and that corresponded to a difference of only 3.3%.
It must be emphasized that, on average, 10 y had elapsed since these patients had undergone small-intestinal resection, and all were considered weight-stable, a condition that indicated a balance between energy requirements and intake. The results presented here agree with several other body-composition studies showing that a steady state was reached at a lower body weight and a significantly smaller FM in patients who had undergone small-intestinal resection than in healthy volunteers (1, 8, 9). The primary reduction in the FM can probably be explained by appropriate physiologic responses related to the reduced amount of energy available. Thus, with restricted energy, very little or no fat tissue is built, and when energy requirements exceed intake stores of energy in the form of fat depots are combusted, whereas lean tissue components such as muscle mass and other proteins are spared. The latter result was clearly documented in the classic experimental starvation study by Keys et al (21), in which lean young males during a 24-wk period of semi-starvation had a 70% reduction in FM, whereas their fat-free tissue mass decreased only 15%. It is intriguing that, although the indication for small-intestinal resection in our patients was very different, longitudinal studies in patients who have undergone jejuno-ileal bypass for morbid obesity also showed that the reduction in body weight resulted almost exclusively from a loss of FM (22, 23).
A previous study from this department in patients with small-intestinal resections due to Crohn disease showed, in contrast with our results, a low body weight that was primarily the result of a significant reduction in LTM (6). The discrepancy between the studies may be partly explained by differences in the 2 study populations. Hence, the patients in the present study had more extensive small-intestinal resections and consequently a more severe degree of malabsorption, as reflected in the greater fecal loss of fat. Second and probably not less important, the reference population used in the present study had a much higher mean BMI than the reference population in the previous study (6), which agreed closely with reported data on BMI in larger Scandinavian population studies (24).
LTM as measured by DXA is the sum of all nonfatty, nonbone components of the body, but DXA does not provide any evaluation of the chemical composition of the LTM. Therefore, it must be kept in mind that, although the LTM in patients was quite similar in size to that in the healthy volunteers, a similar proportion of the main constituents within the LTM compartment, ie, protein and water, cannot necessarily be inferred. Thus, an increased amount of total body water, which may occur in severely malnourished patients, could potentially mask a concomitant loss of muscle mass and protein and vice versa.
This study confirms several reports that patients who undergo small-intestinal resection, particularly those with Crohn disease, frequently have low bone mass and are at risk of developing osteoporosis (25). Our data further indicate that a significant reduction in TBBMC might be related in part to a lower (NEA/BMR)%. An explanation for this finding may be that patients with a lower (NEA/BMR)% have a concomitant lower absorption of minerals or other nutrients that are essential for maintenance of normal bone homeostasis, although we provide no data to support such a theory. However, caution should be used in interpreting the TBBMC data, because a range of factors, such as cumulative lifetime steroid dose, duration of disease, sex and hormonal status, and differences in osteoporosis therapy, are not accounted for in this study.
The mean daily energy intake of the patients was > 12 000 kJ, which is substantially above the recommended daily energy intake for healthy persons (18), and none of the patients had a deficient oral intake. However, patients absorbed only a mean of 68% of the ingested energy and the (NEA/BMR)% averaged 134%, which is barely in the lower end of the normal range (18–20). Therefore, malabsorption of energy not fully compensated for by increased intake is likely to have played an important role in determining the lower body weight of these patients. In spite of this, we were unable to establish any significant correlation between the (NEA/BMR)% and the standardized values of weight, BMI, and body-composition variables measured by DXA. This may be explained by differences among patients in the level of physical activity and thus in their energy expenditure, which is an important factor not accounted for in the study. Yet there was a tendency for a positive relation between (NEA/BMR)% and weight, BMI, and the z scores for most DXA variables, and it appeared that most of these measures were normal in patients who had (NEA/BMR)% above the minimum energy requirement in healthy adults. It is interesting that 14% of the patients were able to maintain daily activities although their intestinal energy absorption actually was below their BMR as estimated by the Harris-Benedict equation. There is evidence to suggest, however, that, in seriously underweight patients, such as those with anorexia nervosa or those who are starved, the actual BMR may be far below that predicted by the Harris-Benedict equation (26). Nevertheless, our data showed that patients with (NEA/BMR)% < 100% had the lowest LTM and TBBMC values among all groups of patients and had mean body weight and BMI values > 2 SD below those of healthy volunteers.
In recent years, several hormones and neuropeptides have been identified that are implicated in the regulation of the mass of body fat (27). One of these, leptin, acts on appetite-regulating neurons and circuits in the hypothalamus that regulate energy balance. Thus, theoretically, differences in the individual concentrations of leptin may have influenced the food intake and, indirectly, the body composition of our patients. However, a study in patients with short-bowel syndrome found no association between leptin concentrations and hyperphagia (7).
In conclusion, this study showed that the lower body weight of patients who have undergone small-intestinal resection was predominantly the result of a significant decrease in the FM, whereas the decrease in the LTM was relatively small and insignificant. We found no clear effect of the magnitude of the (NEA/BMR)% on body weight, BMI, and body-composition variables as measured by DXA, which indicates that other factors, eg, energy expenditure, are also important to the body composition of these patients.
ACKNOWLEDGMENTS
The technical assistance of Jette Christiansen and Bodil Petersen was greatly appreciated.
KVH, PBJ, PBM, and MS were responsible for the conception and design of the study; KVH and MS were responsible for data interpretation and manuscript preparation; and PBJ, PBM, and HAS assisted in data interpretation and manuscript preparation. None of the authors had a personal or financial interest in any organization sponsoring the research.
REFERENCES