Tube
Feeding-Related Diarrhea in Acutely III Patients*
PEGGI A. GUENTER, R.N., M.S.N., C.N.S.N., R. GREGG SETTLE, PH.D., SHELLEY PERLMUTTER, R.D., C.N.S.D., PAUL L. MARINO, M.D., PH.D., GARY A. DESIMONE, B.A., ROLANDO H. ROLANDELLI, M.D.
AND
From The Graduate Hospital, University of Pennsylvania Schools of Nursing and Medicine, and Philadelphia Veterans Affairs Medical Center,
Philadelphia, Pennsylvania
ABSTRACT. Acutely ill patients received tube feeding for an average of 15.8 days and, on average, 35% of those days were spent in the intensive care unit (ICU). Patients were prospectively assigned either a fiber-free formula (FFF-OSMOLITE
was
serum
albumin
was
in
=
defined as three or more loose or watery stools per day and occurred in 30% of all patients. Diarrhea developed in 29 (41%) of the 71 patients who received antibiotics during, or within 2 weeks prior to, the feeding period, whereas only 1 (3%) of the 29 patients not receiving antibiotics developed diarrhea (p < 0.005); and this patient developed diarrhea on the day of death. Among the 30 patients with diarrhea, stool Clostridium difficile (CD) toxin was positive in 15 (50%), negative in 11 (37%), and
the factors that may contribute to the development of diarrhea and to study the effect of a tube feeding formula supplemented with a fiber source, soy polysaccharide, on the incidence of diarrhea in acutely ill patients.
Diarrhea is a commonly cited complication associated with tube feeding. The incidence of diarrhea related to enteral nutrition varies from 2.3%1 to 60%.2 The significance of this problem ranges from patient discomfort to life-threatening situations. Diarrhea complicates nursing care and adds significantly to the cost of hospitalization. Severe diarrhea leads to fluid, electrolyte, and nutrient depletions; some of these patients also require parenteral nutrition, which adds further costs and risks to their
MATERIALS AND METHODS
Patients, who
were seen
in consultation
by the
Nutri-
Support Service at The Graduate Hospital between January 1987 and June 1989, were screened for tional
treatment.3
development of diarrhea in tube-fed, acutely ill patients has been associated with multiple factors including hypoalbuminemia, hyperosmolar or high-fat forThe
mulas, and antibiotics. 4,5 Because the exact cause of diarrhea is not well understood, various therapeutic interventions have been implemented in the absence of efficacy data. One such proposed intervention is fiber supplementation for the treatment of tube feeding-
related diarrhea.66 Most commercially available tube feeding formulas do not and traditionally have not included dietary fiber. However, a recent study demonstrated that fiber-free diets induced liquid stools in normal subjects and that this effect was reversed by the addition of pectin, a fermentable fiber.’ The purpose of this study was to investigate some of
* Part of these data were presented at the 14th Clinical Congress of the American Society for Parenteral and Enteral Nutrition in San
1990.
mean
patients with diarrhea (2.43) than in those without diarrhea (2.75) (p 0.043). There were no significant differences in age, sex, diagnoses, number of feeding days, and percent ICU days between patients with and without diarrhea. While not statistically significant, patients who received FSF were observed to have a lower incidence of diarrhea, a lower percentage of diarrhea days per total feeding days, and a lower frequency of positive CD toxin assays than patients who received FFF. In this patient population, antibiotic usage was the factor most strongly associated with diarrhea during tube feedings. ( Journal of Parenteral and Enteral Nutrition :277-280, 1991) 15
HN, Ross; n=50) or a fiber-supplemented (soy polysaccharide 14.4 g/L) formula (FSF=JEVITY, Ross; n=50). Diarrhea was
Antonio, TX, January
not measured in four. The
significantly lower
admission into the study according to the following inclusion criteria: (1) the patient was a candidate for tube feeding according to the standards of the American Society for Parenteral and Enteral Nutrition’ (2) a polymeric complete isotonic formula was deemed to be the most appropriate diet, (3) tube feeding was required for at least 5 days,’ and (4) the patient was not ingesting any food during the time of tube feedings. Patients with one or more of the following conditions were excluded from the study: (1) renal or hepatic failure, (2) hemodynamic instability (shock states), (3) malabsorption from intrinsic diseases or resections of the gastrointestinal tract. The underlying illnesses of the patients entered in the study were classified into the major categories of neurological/neuromuscular diseases (NEURO), head/ neck and esophageal cancers (HNECA), chest and heart surgery (ChestS), abdominal surgery (AbdS), and primary respiratory failure (RespF). Patients entered into the study received one of two tube feeding formulas: a fiber-supplemented formula (FSF) or a fiber-free formula (FFF). The FFF used in the study was OSMOLITE HN (Ross Laboratories, Co-
277 Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
278
lumbus, OH), which is a complete, polymeric, isotonic (300 mOsm/kg) tube feeding formula containing 44.4 g of protein, 36.8 g of fat, and 141.0 g of carbohydrates per liter. The FSF used was JEVITY (Ross Laboratories, Columbus, OH), which is similarly isotonic (310 mOsm/ kg) and polymeric (44.4, 36.9, and 152.0 g of protein, fat, and carbohydrate, respectively) and contains soy polysaccharide, 14.4 g/L. The first 50 patients assigned to each diet group (FSF=50, FFF=50) were entered into the study, for a total of 100 patients. All patients received the tube feedings continuously via an enteral pump for at least 5 days. Diarrhea was defined as three or more loose or watery stools per day, a definition used in several other studies (Bliss DZ, Guenter PA, Settle RG: Defining and reporting diarrhea in tube-fed patients: Cleaning up the mess! (in preparation)).4 Subjects were monitored for (1) stool frequency and consistency, (2) usage of antibiotics (within 2 weeks prior to or during the study period),&dquo; (3) level of serum albumin, (4) days spent in intensive care, and (5) days of feeding received. Data were collected
daily from the medical record and bedside flow sheet and from interviews of the primary nurses. The bowel movement frequency was recorded on the bedside flow sheet and consistency was described in the nurses notes. Demographic data such as age, sex, and diagnoses were also recorded by the investigators. In those patients who developed diarrhea, presence of C. difficile (CD) toxin in stool was assayed in the hospital clinical laboratory by the latex slide agglutination testll using the Culturette Brand CDT (Marion Scientific, Kansas City, MO). Data were analyzed by analysis of variance (ANOVA), except for categorical data, which were analyzed by X2. Logistic stepwise multiple regression was used to determine the degree of association of various factors with diarrhea. All statistical analyses were performed using the Biomedical Data
Processingl2
software
on a
micro-
computer. RESULTS
=
a
stepwise logistic regression analysis
was
per-
and without
diarrhea
*
Mean ± standard error.
t NM = not measured.
formed. Antibiotic usage was the first variable entered into the regression equation because it accounted for the greatest amount of variability in the incidence of diarrhea. Once antibiotic usage was entered into the equation, serum albumin no longer accounted for a significant proportion of the variability associated with the incidence of diarrhea (p 0.17), suggesting that serum albumin levels covary with antibiotics. An ANOVA of antibiotic usage on serum albumin levels revealed that patients receiving antibiotics had significantly lower serum albumin levels than patients not receiving antibiotics (2.55 us 2.90, p = 0.03) (Table II). To further explore the interrelationship among antibiotic usage, serum albumin levels, and the incidence of diarrhea, an ANOVA of three patient groups (antibiot=
ics/diarrhea ; antibiotics/no diarrhea;
The average age of all the patients was 65 years, two thirds were females, two thirds had a neurological disorder, and the average serum albumin was 2.6 g/dL. The average duration of tube feeding was 15.8 days. More than two thirds of the patients received antibiotics just prior to or during the feeding period. The characteristics of the patients who developed diarrhea and those who did not are shown in Table I. Diarrhea occurred in 30% of all patients. It developed in 29 (41 % ) of the 71 patients who received antibiotics prior to or during the feeding period, whereas only 1 of the 29 patients (3%) not receiving antibiotics developed diarrhea (p < 0.005); and this patient developed diarrhea on the day of death. Among patients with diarrhea, stool CD toxin was positive in 15 (50%), negative in 11 (37%), and not measured in 4 (13%). The mean serum albumin was significantly lower in patients with diarrhea (2.43) than in those without diarrhea (2.75, p 0.043) (Table I). In order to examine the relative contribution of antibiotic usage and serum albumin level to the incidence of
diarrhea,
TABLE I
Demographic and clinical variables of patients with
diarrhea)
on serum
albumin levels
antibiotics/no performed. (Be-
no
was
the no antibiotic/diarrhea group contained only patient who died on the day she developed diarrhea, this group was excluded from the analysis.) The Newman-Keuls range test was used to make post hoc pairwise comparisons among the groups. The only groups which differed significantly in albumin levels were those pacause
one
tients without diarrhea who did not receive antibiotics (2.94) and those patients with diarrhea who did receive antibiotics (2.44; p < 0.05). For the two groups receiving antibiotics, the serum albumin of those patients with diarrhea (2.44) did not differ significantly from that of patients without diarrhea (2.63, p > 0.25), indicating that when antibiotic usage was held constant serum albumin level was not significantly associated with the incidence of diarrhea in these patients. There were no significant differences in patient’s age, sex, use of antibiotics, or number of feeding days between the dietary groups (FFF us FSF) (Table III). Diarrhea occurred in 34% of the FFF group and 26% of the FSF group. Patients who received FSF were also observed to
Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
279 Demographic
*
TABLE II and clinical variables of patients with and without antibiotics
Mean ± standard
error.
t NM = not measured. TABLE III
Demographic and clinical variables of patients receiving FFF and those receiving FSF
*
Mean ± standard error. not measured.
t NM
=
have
a lower percentage of diarrhea days/total feeding days (9.2 and 14.2, respectively) and fewer positive CD toxin assays (n 5) than patients who received FFF (n 10); however, these differences did not reach statistical significance. =
=
DISCUSSION
The present study monitored the incidence of diarrhea and various factors reported to be associated with diar-
rhea in
a diverse group of acutely ill hospitalized patients receiving tube feedings. The overall incidence of diarrhea was 30%. Diarrhea was significantly associated with the
administration of antibiotics and lower serum albumin levels. More than 50% of the patients with diarrhea who were assayed for CD toxin in their stool (15 of 26) had a positive result. In the present study, antibiotic usage was shown to have strongest association with diarrhea. This finding is consistent with those of Keohane and co-authorsl3 and other investigators&dquo; who reported a significant relationship between antibiotics and diarrhea in patients receiving enteral nutrition. These studies have demonstrated this association using fiber-free diets. Antibiotics reduce the colonic bacteria which ferment fiber and even if a large amount of fermentable fiber were available in the colon, the bacteria required for the process are limited. Perhaps FSF would be more effective after the large intestine has been recolonized with normal flora or prior to antibiotic administration. Patients who developed diarrhea had significantly lower albumin levels than those who did not. This is consistent with Brinson and Kolts15 who have cited hypoalbuminemia as a factor contributing to the development of diarrhea in acutely ill patients. Using a stepwise logistic regression analysis in the present study, the antibiotic factor was entered first and rendered the serum albumin factor insignificant, suggesting that serum albumin was associated with antibiotics. Among patients receiving antibiotics, the serum albumin of those patients with diarrhea did not differ significantly from that of patients without diarrhea, indicating that when antibiotic usage was held constant, serum albumin level was not significantly associated with the incidence of diarrhea in these patients. In the present study, patients who received FSF were observed to have a lower incidence of diarrhea, a lower percentage of diarrhea/total feeding days, and a lower incidence of positive CD toxin assays than patients who received FFF, although none of these comparisons reached statistical significance. The mechanism by which a fiber-supplemented formula may potentially reduce tube feeding-relate diarrhea is via its fermentation, which yields short-chain fatty acids (SCFA) that provide trophic stimuli and fuels for the colonic epithelium.&dquo; The presence of SCFA in the colonic lumen stimulate the absorption of sodium and water across the colonic mucosa.17 In this study the fiber supplementation had only a small, nonsignificant effect on diarrhea. The amount of soluble fiber (fermentable) in soy polysaccharide ranges from 5% to 27% depending on which chemical analytical method is used.18,19 Thus, overall, the soy polysaccharide is only moderately fermentable and the amount of SCFA produced is limited. 21 Another effect of dietary fiber is to maintain the microbiological ecology of the colon.21 Antibiotic-induced diarrhea has been strongly associated with CD bacterial toxin.22.23 Patients receiving FFF were cultured positive 10 for CD more often than patients receiving FSF (n and n 5). CD growth and toxin production are influenced by colonic pH inasmuch as both are favored by an =
=
Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
280
alkaline environment. 21 Colonic pH is determined by total bacterial count and fermentation products (SCFA). The production of SCFA is dependent on the availability of substrates (eg, fiber polysaccharides). Fiber-free diets reduce the availability of substrates for fermentation while broad spectrum antibiotics reduce the total of colonic bacteria. This may account for the weak effect of FSF on diarrhea and CD in the present study. CONCLUSION
The results of this
study demonstrate that antibiotic usage strongly associated with the incidence of diarrhea in our tube-fed patients. In the present study the small, insignificant effect of FSF on diarrhea may was
most
have resulted from the strong association of diarrhea with antibiotics, which reduces the bacterial population available for fiber fermentation. Additional studies are needed to clarify the influence of serum albumin level and FSF on the incidence of diarrhea. This could optimally be studied in a group of acutely ill patients not
prevention. JPEN 12:338-345, 1988 HA, Green LC: Successful use of a bulk laxative to control diarrhea of tube feeding. Scand J Plast Reconstr Surg 13:193-194,
6. Frank
1979 7. Zimmaro DM, Rolandelli RH, Koruda MJ, et al: Isotonic tube feeding formula induces liquid stool in normal subjects: Reversal by pectin. JPEN 13:117-123, 1989 8. ASPEN Board of Directors: Guidelines for the use of enteral nutrition in the adult patient. JPEN 11:435-439, 1987 9. Deleted in press 10. Bartlett JG: Clostridium difficile: Pseudomembraneous colitis and antibiotic-associated diarrhea. IN Infectious Diarrhea, Gorbach SL (ed). Blackwell Publications, Boston, 1986 11. Kamiya S, Nakamura S, Yamakawa K, et al: Evaluation of a commercially available latex immunoagglutination test kit for detection of Clostridium difficile D-1 toxin. Microbiol Immunol
30:177-181, 1986 13.
14. 15.
receiving antibiotics. 16.
ACKNOWLEDGMENTS 17.
This research was supported in part by the 1988 Stanley Dudrick Research Scholar Award granted to R. Rolandelli, MD, by the American Society for Parenteral and Enteral Nutrition. The authors thank Marcelle Karustis, RD, for her assistance in data collection.
(ed): BMDP Statistical Software Manual. University of California Press, Berkeley, 1988 Keohane PP, Attrill H, Jones BJM, et al: The roles of lactose and Clostridium difficile in the pathogenesis of enteral feeding associated with diarrhea. Clin Nutr 1:259-264, 1983 Bastow D, Rawlings J, Allison SP: Overnight nasogastric tube feeding. Clin Nutr 4:7-11, 1985 Brinson RR, Kolts BE: Hypoalbuminemia as an indicator of diarrheal incidence in critically ill patients. Crit Care Med 15:506-509, 1987 Roediger WEW: Role of anaerobic bacteria in the metabolic welfare of the colonic mucosa in man. Gut 21:793-798, 1980 Roediger WEW, Moore A: Effect of short-chain fatty acid on sodium absorption in isolated human colon perfused through the vascular bed. Dig Dis Sci 26:100-106, 1981 Association of Official Analytical Chemists: Total dietary fiber in foods. J Assoc Off Anal Chem 68:339, 1985 Englyst HN, Hudson GJ: Colorimetric method for routine measurement of dietary fiber as nonstarch polysaccharides: A comparison with gas-liquid chromatography. Food Chem 24:63-76, 1987 Spiller GA, Chernoff MC, Hill RA, et al: Effect of purified cellulose, pectin, and a low-residue diet on fecal volatile fatty acids, transit time, and fecal weight in humans. Am J Clin Nutr 33:754-759,
12. Dixon WJ
18. 19.
20.
REFERENCES
1980 1. Cataldi-Betcher
2.
EL, Seltzer MH, Slocum BA,
et al:
Complications
occurring during enteral nutrition support. JPEN 7:546-552, 1983 Flynn KT, Norton CC, Fisher RL: Enteral tube feeding: Indications, practices and outcomes. Image: J Nurs Scholarship 19:16-
21.
Immun 23:559-563, 1979
3. Hart 4.
5.
JG, Chang TW, Gurwith M, et al: Antibiotic-associated pseudomembranous colitis due to toxin-producing clostridia. N Engl J Med 298:531-534, 1982 Foulke GE, Silva J: Clostridium difficile in the intensive care unit: Management problems and prevention issues. Crit Care Med
22. Bartlett
19, 1987
GK, Dobb GJ: Effect of a fecal bulking agent on diarrhea during enteral feeding in the critically ill. JPEN 12:465-468, 1988 Kelly TWJ, Patrick MR, Hillman KM: Study of diarrhea in critically ill patients. Crit Care Med 11:7-9, 1983 Gottschlich MM, Warden GD, Michel MA, et al: Diarrhea in tubefed burn patients: Incidence, etiology, nutritional impact, and
Byrne BM, Dankert J: Volatile fatty acids and aerobic flora in the gastrointestinal tract of mice under various conditions. Infect
23.
17:822-826, 1989 24. Rolfe RD: Role of volatile fatty acids in colonization resistance to Clostridium difficile. Infect Immun 45:185-191, 1984
Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
Feeding-Related Diarrhea in Acutely III Patients*
PEGGI A. GUENTER, R.N., M.S.N., C.N.S.N., R. GREGG SETTLE, PH.D., SHELLEY PERLMUTTER, R.D., C.N.S.D., PAUL L. MARINO, M.D., PH.D., GARY A. DESIMONE, B.A., ROLANDO H. ROLANDELLI, M.D.
AND
From The Graduate Hospital, University of Pennsylvania Schools of Nursing and Medicine, and Philadelphia Veterans Affairs Medical Center,
Philadelphia, Pennsylvania
ABSTRACT. Acutely ill patients received tube feeding for an average of 15.8 days and, on average, 35% of those days were spent in the intensive care unit (ICU). Patients were prospectively assigned either a fiber-free formula (FFF-OSMOLITE
was
serum
albumin
was
in
=
defined as three or more loose or watery stools per day and occurred in 30% of all patients. Diarrhea developed in 29 (41%) of the 71 patients who received antibiotics during, or within 2 weeks prior to, the feeding period, whereas only 1 (3%) of the 29 patients not receiving antibiotics developed diarrhea (p < 0.005); and this patient developed diarrhea on the day of death. Among the 30 patients with diarrhea, stool Clostridium difficile (CD) toxin was positive in 15 (50%), negative in 11 (37%), and
the factors that may contribute to the development of diarrhea and to study the effect of a tube feeding formula supplemented with a fiber source, soy polysaccharide, on the incidence of diarrhea in acutely ill patients.
Diarrhea is a commonly cited complication associated with tube feeding. The incidence of diarrhea related to enteral nutrition varies from 2.3%1 to 60%.2 The significance of this problem ranges from patient discomfort to life-threatening situations. Diarrhea complicates nursing care and adds significantly to the cost of hospitalization. Severe diarrhea leads to fluid, electrolyte, and nutrient depletions; some of these patients also require parenteral nutrition, which adds further costs and risks to their
MATERIALS AND METHODS
Patients, who
were seen
in consultation
by the
Nutri-
Support Service at The Graduate Hospital between January 1987 and June 1989, were screened for tional
treatment.3
development of diarrhea in tube-fed, acutely ill patients has been associated with multiple factors including hypoalbuminemia, hyperosmolar or high-fat forThe
mulas, and antibiotics. 4,5 Because the exact cause of diarrhea is not well understood, various therapeutic interventions have been implemented in the absence of efficacy data. One such proposed intervention is fiber supplementation for the treatment of tube feeding-
related diarrhea.66 Most commercially available tube feeding formulas do not and traditionally have not included dietary fiber. However, a recent study demonstrated that fiber-free diets induced liquid stools in normal subjects and that this effect was reversed by the addition of pectin, a fermentable fiber.’ The purpose of this study was to investigate some of
* Part of these data were presented at the 14th Clinical Congress of the American Society for Parenteral and Enteral Nutrition in San
1990.
mean
patients with diarrhea (2.43) than in those without diarrhea (2.75) (p 0.043). There were no significant differences in age, sex, diagnoses, number of feeding days, and percent ICU days between patients with and without diarrhea. While not statistically significant, patients who received FSF were observed to have a lower incidence of diarrhea, a lower percentage of diarrhea days per total feeding days, and a lower frequency of positive CD toxin assays than patients who received FFF. In this patient population, antibiotic usage was the factor most strongly associated with diarrhea during tube feedings. ( Journal of Parenteral and Enteral Nutrition :277-280, 1991) 15
HN, Ross; n=50) or a fiber-supplemented (soy polysaccharide 14.4 g/L) formula (FSF=JEVITY, Ross; n=50). Diarrhea was
Antonio, TX, January
not measured in four. The
significantly lower
admission into the study according to the following inclusion criteria: (1) the patient was a candidate for tube feeding according to the standards of the American Society for Parenteral and Enteral Nutrition’ (2) a polymeric complete isotonic formula was deemed to be the most appropriate diet, (3) tube feeding was required for at least 5 days,’ and (4) the patient was not ingesting any food during the time of tube feedings. Patients with one or more of the following conditions were excluded from the study: (1) renal or hepatic failure, (2) hemodynamic instability (shock states), (3) malabsorption from intrinsic diseases or resections of the gastrointestinal tract. The underlying illnesses of the patients entered in the study were classified into the major categories of neurological/neuromuscular diseases (NEURO), head/ neck and esophageal cancers (HNECA), chest and heart surgery (ChestS), abdominal surgery (AbdS), and primary respiratory failure (RespF). Patients entered into the study received one of two tube feeding formulas: a fiber-supplemented formula (FSF) or a fiber-free formula (FFF). The FFF used in the study was OSMOLITE HN (Ross Laboratories, Co-
277 Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
278
lumbus, OH), which is a complete, polymeric, isotonic (300 mOsm/kg) tube feeding formula containing 44.4 g of protein, 36.8 g of fat, and 141.0 g of carbohydrates per liter. The FSF used was JEVITY (Ross Laboratories, Columbus, OH), which is similarly isotonic (310 mOsm/ kg) and polymeric (44.4, 36.9, and 152.0 g of protein, fat, and carbohydrate, respectively) and contains soy polysaccharide, 14.4 g/L. The first 50 patients assigned to each diet group (FSF=50, FFF=50) were entered into the study, for a total of 100 patients. All patients received the tube feedings continuously via an enteral pump for at least 5 days. Diarrhea was defined as three or more loose or watery stools per day, a definition used in several other studies (Bliss DZ, Guenter PA, Settle RG: Defining and reporting diarrhea in tube-fed patients: Cleaning up the mess! (in preparation)).4 Subjects were monitored for (1) stool frequency and consistency, (2) usage of antibiotics (within 2 weeks prior to or during the study period),&dquo; (3) level of serum albumin, (4) days spent in intensive care, and (5) days of feeding received. Data were collected
daily from the medical record and bedside flow sheet and from interviews of the primary nurses. The bowel movement frequency was recorded on the bedside flow sheet and consistency was described in the nurses notes. Demographic data such as age, sex, and diagnoses were also recorded by the investigators. In those patients who developed diarrhea, presence of C. difficile (CD) toxin in stool was assayed in the hospital clinical laboratory by the latex slide agglutination testll using the Culturette Brand CDT (Marion Scientific, Kansas City, MO). Data were analyzed by analysis of variance (ANOVA), except for categorical data, which were analyzed by X2. Logistic stepwise multiple regression was used to determine the degree of association of various factors with diarrhea. All statistical analyses were performed using the Biomedical Data
Processingl2
software
on a
micro-
computer. RESULTS
=
a
stepwise logistic regression analysis
was
per-
and without
diarrhea
*
Mean ± standard error.
t NM = not measured.
formed. Antibiotic usage was the first variable entered into the regression equation because it accounted for the greatest amount of variability in the incidence of diarrhea. Once antibiotic usage was entered into the equation, serum albumin no longer accounted for a significant proportion of the variability associated with the incidence of diarrhea (p 0.17), suggesting that serum albumin levels covary with antibiotics. An ANOVA of antibiotic usage on serum albumin levels revealed that patients receiving antibiotics had significantly lower serum albumin levels than patients not receiving antibiotics (2.55 us 2.90, p = 0.03) (Table II). To further explore the interrelationship among antibiotic usage, serum albumin levels, and the incidence of diarrhea, an ANOVA of three patient groups (antibiot=
ics/diarrhea ; antibiotics/no diarrhea;
The average age of all the patients was 65 years, two thirds were females, two thirds had a neurological disorder, and the average serum albumin was 2.6 g/dL. The average duration of tube feeding was 15.8 days. More than two thirds of the patients received antibiotics just prior to or during the feeding period. The characteristics of the patients who developed diarrhea and those who did not are shown in Table I. Diarrhea occurred in 30% of all patients. It developed in 29 (41 % ) of the 71 patients who received antibiotics prior to or during the feeding period, whereas only 1 of the 29 patients (3%) not receiving antibiotics developed diarrhea (p < 0.005); and this patient developed diarrhea on the day of death. Among patients with diarrhea, stool CD toxin was positive in 15 (50%), negative in 11 (37%), and not measured in 4 (13%). The mean serum albumin was significantly lower in patients with diarrhea (2.43) than in those without diarrhea (2.75, p 0.043) (Table I). In order to examine the relative contribution of antibiotic usage and serum albumin level to the incidence of
diarrhea,
TABLE I
Demographic and clinical variables of patients with
diarrhea)
on serum
albumin levels
antibiotics/no performed. (Be-
no
was
the no antibiotic/diarrhea group contained only patient who died on the day she developed diarrhea, this group was excluded from the analysis.) The Newman-Keuls range test was used to make post hoc pairwise comparisons among the groups. The only groups which differed significantly in albumin levels were those pacause
one
tients without diarrhea who did not receive antibiotics (2.94) and those patients with diarrhea who did receive antibiotics (2.44; p < 0.05). For the two groups receiving antibiotics, the serum albumin of those patients with diarrhea (2.44) did not differ significantly from that of patients without diarrhea (2.63, p > 0.25), indicating that when antibiotic usage was held constant serum albumin level was not significantly associated with the incidence of diarrhea in these patients. There were no significant differences in patient’s age, sex, use of antibiotics, or number of feeding days between the dietary groups (FFF us FSF) (Table III). Diarrhea occurred in 34% of the FFF group and 26% of the FSF group. Patients who received FSF were also observed to
Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
279 Demographic
*
TABLE II and clinical variables of patients with and without antibiotics
Mean ± standard
error.
t NM = not measured. TABLE III
Demographic and clinical variables of patients receiving FFF and those receiving FSF
*
Mean ± standard error. not measured.
t NM
=
have
a lower percentage of diarrhea days/total feeding days (9.2 and 14.2, respectively) and fewer positive CD toxin assays (n 5) than patients who received FFF (n 10); however, these differences did not reach statistical significance. =
=
DISCUSSION
The present study monitored the incidence of diarrhea and various factors reported to be associated with diar-
rhea in
a diverse group of acutely ill hospitalized patients receiving tube feedings. The overall incidence of diarrhea was 30%. Diarrhea was significantly associated with the
administration of antibiotics and lower serum albumin levels. More than 50% of the patients with diarrhea who were assayed for CD toxin in their stool (15 of 26) had a positive result. In the present study, antibiotic usage was shown to have strongest association with diarrhea. This finding is consistent with those of Keohane and co-authorsl3 and other investigators&dquo; who reported a significant relationship between antibiotics and diarrhea in patients receiving enteral nutrition. These studies have demonstrated this association using fiber-free diets. Antibiotics reduce the colonic bacteria which ferment fiber and even if a large amount of fermentable fiber were available in the colon, the bacteria required for the process are limited. Perhaps FSF would be more effective after the large intestine has been recolonized with normal flora or prior to antibiotic administration. Patients who developed diarrhea had significantly lower albumin levels than those who did not. This is consistent with Brinson and Kolts15 who have cited hypoalbuminemia as a factor contributing to the development of diarrhea in acutely ill patients. Using a stepwise logistic regression analysis in the present study, the antibiotic factor was entered first and rendered the serum albumin factor insignificant, suggesting that serum albumin was associated with antibiotics. Among patients receiving antibiotics, the serum albumin of those patients with diarrhea did not differ significantly from that of patients without diarrhea, indicating that when antibiotic usage was held constant, serum albumin level was not significantly associated with the incidence of diarrhea in these patients. In the present study, patients who received FSF were observed to have a lower incidence of diarrhea, a lower percentage of diarrhea/total feeding days, and a lower incidence of positive CD toxin assays than patients who received FFF, although none of these comparisons reached statistical significance. The mechanism by which a fiber-supplemented formula may potentially reduce tube feeding-relate diarrhea is via its fermentation, which yields short-chain fatty acids (SCFA) that provide trophic stimuli and fuels for the colonic epithelium.&dquo; The presence of SCFA in the colonic lumen stimulate the absorption of sodium and water across the colonic mucosa.17 In this study the fiber supplementation had only a small, nonsignificant effect on diarrhea. The amount of soluble fiber (fermentable) in soy polysaccharide ranges from 5% to 27% depending on which chemical analytical method is used.18,19 Thus, overall, the soy polysaccharide is only moderately fermentable and the amount of SCFA produced is limited. 21 Another effect of dietary fiber is to maintain the microbiological ecology of the colon.21 Antibiotic-induced diarrhea has been strongly associated with CD bacterial toxin.22.23 Patients receiving FFF were cultured positive 10 for CD more often than patients receiving FSF (n and n 5). CD growth and toxin production are influenced by colonic pH inasmuch as both are favored by an =
=
Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
280
alkaline environment. 21 Colonic pH is determined by total bacterial count and fermentation products (SCFA). The production of SCFA is dependent on the availability of substrates (eg, fiber polysaccharides). Fiber-free diets reduce the availability of substrates for fermentation while broad spectrum antibiotics reduce the total of colonic bacteria. This may account for the weak effect of FSF on diarrhea and CD in the present study. CONCLUSION
The results of this
study demonstrate that antibiotic usage strongly associated with the incidence of diarrhea in our tube-fed patients. In the present study the small, insignificant effect of FSF on diarrhea may was
most
have resulted from the strong association of diarrhea with antibiotics, which reduces the bacterial population available for fiber fermentation. Additional studies are needed to clarify the influence of serum albumin level and FSF on the incidence of diarrhea. This could optimally be studied in a group of acutely ill patients not
prevention. JPEN 12:338-345, 1988 HA, Green LC: Successful use of a bulk laxative to control diarrhea of tube feeding. Scand J Plast Reconstr Surg 13:193-194,
6. Frank
1979 7. Zimmaro DM, Rolandelli RH, Koruda MJ, et al: Isotonic tube feeding formula induces liquid stool in normal subjects: Reversal by pectin. JPEN 13:117-123, 1989 8. ASPEN Board of Directors: Guidelines for the use of enteral nutrition in the adult patient. JPEN 11:435-439, 1987 9. Deleted in press 10. Bartlett JG: Clostridium difficile: Pseudomembraneous colitis and antibiotic-associated diarrhea. IN Infectious Diarrhea, Gorbach SL (ed). Blackwell Publications, Boston, 1986 11. Kamiya S, Nakamura S, Yamakawa K, et al: Evaluation of a commercially available latex immunoagglutination test kit for detection of Clostridium difficile D-1 toxin. Microbiol Immunol
30:177-181, 1986 13.
14. 15.
receiving antibiotics. 16.
ACKNOWLEDGMENTS 17.
This research was supported in part by the 1988 Stanley Dudrick Research Scholar Award granted to R. Rolandelli, MD, by the American Society for Parenteral and Enteral Nutrition. The authors thank Marcelle Karustis, RD, for her assistance in data collection.
(ed): BMDP Statistical Software Manual. University of California Press, Berkeley, 1988 Keohane PP, Attrill H, Jones BJM, et al: The roles of lactose and Clostridium difficile in the pathogenesis of enteral feeding associated with diarrhea. Clin Nutr 1:259-264, 1983 Bastow D, Rawlings J, Allison SP: Overnight nasogastric tube feeding. Clin Nutr 4:7-11, 1985 Brinson RR, Kolts BE: Hypoalbuminemia as an indicator of diarrheal incidence in critically ill patients. Crit Care Med 15:506-509, 1987 Roediger WEW: Role of anaerobic bacteria in the metabolic welfare of the colonic mucosa in man. Gut 21:793-798, 1980 Roediger WEW, Moore A: Effect of short-chain fatty acid on sodium absorption in isolated human colon perfused through the vascular bed. Dig Dis Sci 26:100-106, 1981 Association of Official Analytical Chemists: Total dietary fiber in foods. J Assoc Off Anal Chem 68:339, 1985 Englyst HN, Hudson GJ: Colorimetric method for routine measurement of dietary fiber as nonstarch polysaccharides: A comparison with gas-liquid chromatography. Food Chem 24:63-76, 1987 Spiller GA, Chernoff MC, Hill RA, et al: Effect of purified cellulose, pectin, and a low-residue diet on fecal volatile fatty acids, transit time, and fecal weight in humans. Am J Clin Nutr 33:754-759,
12. Dixon WJ
18. 19.
20.
REFERENCES
1980 1. Cataldi-Betcher
2.
EL, Seltzer MH, Slocum BA,
et al:
Complications
occurring during enteral nutrition support. JPEN 7:546-552, 1983 Flynn KT, Norton CC, Fisher RL: Enteral tube feeding: Indications, practices and outcomes. Image: J Nurs Scholarship 19:16-
21.
Immun 23:559-563, 1979
3. Hart 4.
5.
JG, Chang TW, Gurwith M, et al: Antibiotic-associated pseudomembranous colitis due to toxin-producing clostridia. N Engl J Med 298:531-534, 1982 Foulke GE, Silva J: Clostridium difficile in the intensive care unit: Management problems and prevention issues. Crit Care Med
22. Bartlett
19, 1987
GK, Dobb GJ: Effect of a fecal bulking agent on diarrhea during enteral feeding in the critically ill. JPEN 12:465-468, 1988 Kelly TWJ, Patrick MR, Hillman KM: Study of diarrhea in critically ill patients. Crit Care Med 11:7-9, 1983 Gottschlich MM, Warden GD, Michel MA, et al: Diarrhea in tubefed burn patients: Incidence, etiology, nutritional impact, and
Byrne BM, Dankert J: Volatile fatty acids and aerobic flora in the gastrointestinal tract of mice under various conditions. Infect
23.
17:822-826, 1989 24. Rolfe RD: Role of volatile fatty acids in colonization resistance to Clostridium difficile. Infect Immun 45:185-191, 1984
Downloaded from pen.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 9, 2015
