- Email: [email protected]
Enteral feeding increases sepsis in infants with short bowel syndrome
Enteral
Feeding Increases Short Bowel
Sepsis in Infants Syndrome
With
By Thomas R. Weber St Louis, Missouri 0 Sepsis secondary to bacterial translocation is common in infants with short bowel syndrome (SBS). Although early feeding is advocated to enhance adaptation in SBS, the effects of feeding on sepsis in SBS patients have not been examined. Twenty-one infants and children (aged 2 months to 3 years) with SBS (~80 cm small bowel length) from a variety of causes (15 necrotizing enterocolitis, 2 atresia, 2 gastroschisis, 2 volvulus) had follow-up prospectively for septic episodes before and after feedings were initiated, while still receiving total parenteral nutrition. The incidence and number of septic episodes and microbiology (blood cultures) were tabulated and compared with those of 20 patients with similar ages, and diagnoses without SBS. Statistically significant differences among infants with SBS were noted with respect to sepsis incidence (6 of 21 [2g%] NPO v 16 of 21 [76%] feeding) number of septic episodes (1.3 + .2 NPO Y 4.2 + .4 feeding), and presence of gramnegative rods causing bacteriemia (1 of 6 [17%] NPO v 13 of 16 [81%] feeding) (all: P < .05). There were similar differences between SBS and non-SBS infants. These data show that enteral feeding increases the incidence and number of episodes of sepsis in SBS infants, but not in matched non-SBS patients. The predominance of gram-negative organisms in sepsis in SBS suggests increased gut bacterial translocation in these patients, implying that selective gut decontamination may reduce the episodes of bacteremia. Copyright o 7995 by W.B. Saunders Company INDEX tion.
WORDS:
Short
bowel
syndrome,
bacterial
transloca-
HORT BOWEL length, from a variety of causes, S is a common condition in the pediatric age group. Total parenteral nutrition (TPN), usually provided through a central venous catheter, is an important part of the short and long-term therapy of these infants, enabling relatively normal growth and development while the remaining gut undergoes a process of adaptation, and enteral feedings are gradually advanced.’ Recent studies suggest that bacteremia and catheter sepsis are common in infants with short bowel syndrome (SBS),2 and probably are caused by bacterial translocation across the intestinal From the Division of Pedratric Surgery, St Louis University School of Medicine and Cardinal Glennon Children’s Hospital, St LOUIS, MO. Presented at the 1994 Annual Meeting of the Section on Surgery of the American Academy of Pediatrics, Dallas, Texas, October 21-23, 1994. Address reprint requests to Thomas R. Weber, MD, Cardinal Glennon ChrldrenS Hospital, I465 S Grand Blvd, St Louis, MO 63104. Copyright Q 1995 by WB. Saunders Company
0022-3468/9513007-0037$03.OOlO 1086
mucosa.3 Although the use of parenteral nutrition has been shown to increase bacterial translocatioq4 the effects of enteral feeding on bacterial translocation and resultant bacteremia in patients with short bowel length have not been investigated. This study prospectively examines these relationships in a series of infants and children with SBS. MATERIALS
AND METHODS
Twenty-one infants and children with short bowel length ( < 80 cm) were evaluated prospectively for septic episodes while central venous catheters were in place and TPN was being administered. The causes of short bowel length included necrotizing enterocolitis (15), jejunal or ileal atresia (2) gastroschisis (2), and midgut volvulus (2). The ages, extent and location of bowel resection, and bowel length remaining after resection and closure of ostomies are shown in Table 1. The clinical management of the SBS patient has been outlined previously.’ Briefly, during the postoperative period of no enteral feedings (generally 7 to 14 days, depending on diagnosis), the child is supported by TPN. Enteral feedings are begun by a continuous drip technique, at low volumes and concentrations of an elemental formula (Pregestamil; Mead-Johnson, Evansville, IN). Oral feedings of small amounts of an identical formula are also encouraged. The formula is first advanced to full strength, then the volume and rate are advanced as stool output allows. Serum electrolytes and stool-reducing substances are monitored frequently. The continuous feedings are begun while the child has an ostomy, and are advanced until the ostomy is closed (generally 3 to 6 weeks). However, the feeding volumes remain small to limit fluid and electrolyte losses from the ostomy. The enteral feeding is resumed after the onset of stools per rectum. Septic episodes were tabulated after closure of all ostomies. Sepsis episodes during “no enteral feeding” periods and during the early onset of feeding with low-volume dilute formula were tabulated together, generally over a 3-to-6-week period. Septic episodes “during feeding” were counted only after full-strength formula was used, usually at 20% to 30% of total caloric needs, for another 4 to 6 weeks. Episodes of sepsis and bacteremia were suspected when an infant demonstrated nonspecific signs and symptoms (fever, lethargy, ileus, vomiting), but were always confirmed by blood cultures obtained from the central venous catheter and from a peripheral site. The organisms were identified by routine culture practices, but quantification of bacteria was not performed. Catheter sepsis was routinely treated with systemic antibiotics without removal of the catheter, except for fungal sepsis which invariably necessitated catheter removal to successfully treat the infection. For this group of patients, the incidence of sepsis, mean number of episodes of sepsis per patient, and type of bacteria isolated were tabulated over a 2-month to 3-year period. These parameters were compared with those of a concurrent group of 20 newborns and infants with similar ages and diagnoses (aged 6 months to 2.5 years; 13 NEC, 4 atresia, 1 gastroschisis, 2 volvulus). This latter group required operative therapy, had central venous catheters in place for a prolonged period both before and after the onset of enteral feeding, but did not have short bowel length. JournalofPecliatr~
Surgery,
Vol30, No 7 (July), 1995: pp
1086-1089
ENTERAL
FEEDING
Table
AND
1. Clinical
Etmlogy of Short Bowel
Necrotizing
SHORT
BOWEL
Characteristics NO
enterocolitis
15
of 21 Patients Resected
30-55 40-80 25-75
(3)
Jejunum-ileum-colon Jejunum-ileum
Gastroschisrs
2
Jejunum-ileum
Midgut
2
Jejunum-ileum
volvulus
SBS Bowel Length or Range (cm)
(6)
2
atresia
With
Bowel
Jejunum-ileum Ileum (5) Ileum-colon
Jejunal-rleal
1087
SYNDROME
(1)
20 45 60 40 80 50 75
RESULTS
The results are summarized in Table 2. In six of the 21 patients (29%) with short bowel length, catheter sepsis developed before enteral feeding was begun. These six patients had eight separate episodes of sepsis (incidence, 1.3 episodes per patient). Five of the six (83%) had infection caused by gram-positive bacteria (StaphyZococc~s), and the other had infection with gram-negative bacteria (Escherichia coli). These data were not significantly different from those of the non-SBS patients. In contrast, after feeding was initiated, 16 of the 21 patients (76%) had 67 episodes of bacteremia (incidence, 4.2 per patient). Both of these were significantly higher than the values before enteral feeding was begun, and also significantly more than those of patients without short bowel length. In addition, gram-negative organisms (E coli, Klebsiella, Pseudomonas)were significantly more common than gram-positive ones when compared with non-SBS patients and with the SBS patients who did not receive enteral feedings (P < .05). DISCUSSION
TPN administered through a central venous catheter is a vital part of the management of infants and Table
2. Characteristics of Septic Episodes in 21 Patients and 20 Matched Patients Without SBS SBS (n = 21)
Sepsrs
With
Non-SBS (n = 20)
incidence
NPO Feeding No. of sepsrs
Gram-posrtive NPO Feeding *P < .05 versus
4 of 20 (20%)
16 of 21 (76%)*
5 of 20 (25%)
eprsodes
NPO Feeding Gram-negative NPO Feedmg
6 of 21 (29%)
1.3 k .2
1.5 + .I
4.2 2 .4*
1.6 + .2
1 of 6 (17%) 13 of 16 (Sl%)*
0 of 4 (0%) 1 of 4 (25%)
5 of 6 (83%)
4of4(100%)
5of16(31%)
3 of 5 (60%)
bacteria
bacteria
NPO and non-SBS.
SBS
children with SBS. Sepsis and bacteremia are well known complications of central venous catheters used for a variety of reasons, but recent data suggest that patients with SBS have an increased incidence of sepsis, especially with enteric gram-negative organisms.2,3 Because enteric organisms are frequently involved in sepsis associated with short bowel length, bacterial translocation has been implicated.3 The present study suggests that enteral feeding increases septic episodes in the eary postoperative period in infants with SBS. Although not measured in this study, the septic episodes appeared to decrease with time, suggesting that the explanation for increased sepsis was, in fact, related to the feeding and not simply to other factors (ie, bacterial overgrowth) associated with increasing time postoperatively. A number of factors have been clinically and experimentally associated with increased incidence of bacterial translocation, many of which are present in the child with short bowel length.5,6 These include prolonged parenteral nutrition, elemental-diet enteric feeding, alteration and overgrowth of gut microflora, enteric stasis, cholestasis, and altered mucosal integrity. Early enteral feedings have been advocated for patients with short bowel length as a means of promoting adaptation and hypertrophy of the remaining gut.’ In addition, some studies suggest that enteric feeding might reduce postoperative septic complications when compared with parenteral feeding.* Although this later observation may be true in the patient with normal or near-normal bowel length, the data in the present study suggest that feeding promotes rather than limits septic episodes in the patient with short bowel length. Enteral feeding with elemental diets has been associated with increased bacterial translocation,9 perhaps because of cytokine activation, mucosal damage, or bacterial overgrowth. All the patients in the present study received continuous elemental diets during the early phases of adaptation, and many continued to receive some elemental diet even after beginning a nonelemental oral diet. It has been suggested9 that elemental diets high in glutamine might maintain mucosal integrity and decrease bacterial translocation, and we are conducting a trial of these formulas in several patients over 1 year of age, because they are not suitable for infant use. Selective modification of the gut flora has been used in other studies with limited success. Nonantibiotic measures, such as high-fiber dietary supplements, have been shown experimentally to reduce bacterial translocation,9 but their usefulness has not been demonstrated clinically. Poorly absorbed oral antibi-
1088
THOMAS
otics, used both cyclically and continuously,7 may have a role in selected patients, but the risk of the development of resistant strains of bacteria and of fungal overgrowth limits the applicability of this approach. Noninfectious inflammatory bowel disease, occasionally present in patients with short bowel length, may respond to oral antinflammatory agents.lO Because the central venous line can be lifesaving in many of these patients, and the number of access sites
R. WEBER
for these catheters is limited, we agree with investigators29 who advocate attempts at salvaging infected catheters with specific antibiotic therapy, rather than immediately removing them. Recognizing that many (if not most) of these infections in patients with short bowel length are caused by enteric organisms, initial antibiotic therapy should be broad-spectrum and include coverage for gram-negative bacteria, until the specific organisms and antibiotic sensitivities are identified.
REFERENCES 1. Weber TF, Tracy TF, Connors RH: Short bowel syndrome in children: Quality of life in an era of improved survival. Arch Surg 126:841-846,199l 2. Piedra PA, Dryja DM, LaScolea LJ: Incidence of catheterassociated gram-negative bacteremia in children with short bowel syndrome. J Clin Microbial 27:1317-1319,1989 3. Kurkchubasche AG, Smith SD, Rowe MI: Catheter sepsis in short bowel syndrome. Arch Surg 127:21-25,1992 4. Alverdy JC, Aoys E, Moss GS: Total parenteral nutrition promotes bacterial translocation from the gut. Ann Surg 202:681684,1985 5. Edmiston CE, Condon RE: Bacterial translocation. Surg Gynecol Obstet 173:73+X3,1991
6. Alexander JW, Boyce ST, Babcock GF, et al: The process of microbial translocation. Ann Surg 212:496-512,199O 7. Vanderhoof JA, Langnas AN, Pinch LW, et al: Short bowel syndrome. J Pediatr Gastroenterol Nutr 14:359-370,1992 8. Moore FA, Feliciano DV, Andrassy RJ, et al: Early enteral feeding, compared with parenteral, reduces postoperative septic complications-The results of a META analysis. Ann Surg 216: 172183,1992 9. Alverdy JC, Aoys E, Moss GS: Effect of commercially available chemically defined liquid diets on the intestinal microflora and bacterial translocation from the gut. JPEN 14:1-6,199O 10. Taylor SF, Sondheimer JM, Sokol RJ, et al: Noninfectious colitis associated with short gut syndrome in infants. J Pediatr 119:24-28,1991
Discussion MM. Ziegler (Cincinnati, OH): This, in my way of thinking, has become a tremendously commonplace problem in the management of these kinds of patients. The one thing in addition to what you have reported that we have seen in Cincinnati is an increased incidence of yeast infection. My practical question is very simple, and that is that now that you have exposed the problem and have suggested some alternate therapies for it, what is the current cocktail that you would suggest be used day to day in the management of these patients? L.A. Martinez (Denver, CO): I have two brief questions. One is, how many of your 21 patients had an ileocecal valve or had lost it? And second, in how many of your patients was the sepsis severe enough that you had to remove the central catheter? KE. Georgeson (Birmingham, AL): We have just reviewed this in our patients and have found that gut lengthening, which was one of the suggestions made, cut our incidence to less than 50% of gram-negative septicemia. The other thing I am wondering about is, have you looked at gut clearance? Very often we look at transit time-that is, how long it takes for the contrast to go
from the tongue to the tail-but then we fail to see how rapidly the material is cleared from the gut. Very often in these patients who have gone through extensive adaptation their bowel dilates, and if you get an x-ray the next day they still have barium hanging around. In those patients I think we’re seeing a much higher incidence of stasis and associated gramnegative septicemia because of translocation. D.H. Teitelbaum (Ann Arbor, MI): How did you express your incidence of sepsis? How long were the patients NPO and how long were they feeding? In other words, if they were NPO for a month and they were feeding for a year, did you express the number of septic episodes per, say, thousand catheter days, and would that change the relative incidence of your septic episodes and decrease it perhaps in that group? R.J. Touloukian (New Haven, CT): This is really an extension of what Dr Georgeson asked you regarding the possibility that you are basically manifesting a partial obstruction in the bowel in those patients who are at high risk for the development bacterial translocation when they are fed. The question is, have you studied these patients radiographically to be certain that their anastomoses are nonobstructed?
ENTERAL
FEEDING
AND
SHORT
BOWEL
SYNDROME
T.R. Weber (response): We continue to use elemental diets. We use oral, poorly absorbed antibiotics and try to switch the patients to either breast milk, if available, or a glutamine-enriched formula as soon as we can. We have noticed in several patients with the glutamine-enriched formula that there is a great decrease in the incidence of translocation. To answer some of the other questions, 50% of our patients in the SBS group have no ileocecal valve, a much greater incidence than in our non-SBS group. That may be related to our incidence of translocation. We have not had to remove a catheter for bacterial
sepsis. However, yeast sepsis certainly is an indication for removing a catheter. We have noticed some decrease in the incidence of translocation and sepsis with gut lengthening and gut tapering procedures, which also, I think, somewhat answers the question about gut clearance. If you can decrease the diameter of the bowel and improve peristalsis, there’s no question that gut clearance of barium and presumably of bacteria also improves. We did not specify our septic episodes per thousand days of catheter use. However, when we looked at that, it didn’t make any difference in the data.
Feeding Increases Short Bowel
Sepsis in Infants Syndrome
With
By Thomas R. Weber St Louis, Missouri 0 Sepsis secondary to bacterial translocation is common in infants with short bowel syndrome (SBS). Although early feeding is advocated to enhance adaptation in SBS, the effects of feeding on sepsis in SBS patients have not been examined. Twenty-one infants and children (aged 2 months to 3 years) with SBS (~80 cm small bowel length) from a variety of causes (15 necrotizing enterocolitis, 2 atresia, 2 gastroschisis, 2 volvulus) had follow-up prospectively for septic episodes before and after feedings were initiated, while still receiving total parenteral nutrition. The incidence and number of septic episodes and microbiology (blood cultures) were tabulated and compared with those of 20 patients with similar ages, and diagnoses without SBS. Statistically significant differences among infants with SBS were noted with respect to sepsis incidence (6 of 21 [2g%] NPO v 16 of 21 [76%] feeding) number of septic episodes (1.3 + .2 NPO Y 4.2 + .4 feeding), and presence of gramnegative rods causing bacteriemia (1 of 6 [17%] NPO v 13 of 16 [81%] feeding) (all: P < .05). There were similar differences between SBS and non-SBS infants. These data show that enteral feeding increases the incidence and number of episodes of sepsis in SBS infants, but not in matched non-SBS patients. The predominance of gram-negative organisms in sepsis in SBS suggests increased gut bacterial translocation in these patients, implying that selective gut decontamination may reduce the episodes of bacteremia. Copyright o 7995 by W.B. Saunders Company INDEX tion.
WORDS:
Short
bowel
syndrome,
bacterial
transloca-
HORT BOWEL length, from a variety of causes, S is a common condition in the pediatric age group. Total parenteral nutrition (TPN), usually provided through a central venous catheter, is an important part of the short and long-term therapy of these infants, enabling relatively normal growth and development while the remaining gut undergoes a process of adaptation, and enteral feedings are gradually advanced.’ Recent studies suggest that bacteremia and catheter sepsis are common in infants with short bowel syndrome (SBS),2 and probably are caused by bacterial translocation across the intestinal From the Division of Pedratric Surgery, St Louis University School of Medicine and Cardinal Glennon Children’s Hospital, St LOUIS, MO. Presented at the 1994 Annual Meeting of the Section on Surgery of the American Academy of Pediatrics, Dallas, Texas, October 21-23, 1994. Address reprint requests to Thomas R. Weber, MD, Cardinal Glennon ChrldrenS Hospital, I465 S Grand Blvd, St Louis, MO 63104. Copyright Q 1995 by WB. Saunders Company
0022-3468/9513007-0037$03.OOlO 1086
mucosa.3 Although the use of parenteral nutrition has been shown to increase bacterial translocatioq4 the effects of enteral feeding on bacterial translocation and resultant bacteremia in patients with short bowel length have not been investigated. This study prospectively examines these relationships in a series of infants and children with SBS. MATERIALS
AND METHODS
Twenty-one infants and children with short bowel length ( < 80 cm) were evaluated prospectively for septic episodes while central venous catheters were in place and TPN was being administered. The causes of short bowel length included necrotizing enterocolitis (15), jejunal or ileal atresia (2) gastroschisis (2), and midgut volvulus (2). The ages, extent and location of bowel resection, and bowel length remaining after resection and closure of ostomies are shown in Table 1. The clinical management of the SBS patient has been outlined previously.’ Briefly, during the postoperative period of no enteral feedings (generally 7 to 14 days, depending on diagnosis), the child is supported by TPN. Enteral feedings are begun by a continuous drip technique, at low volumes and concentrations of an elemental formula (Pregestamil; Mead-Johnson, Evansville, IN). Oral feedings of small amounts of an identical formula are also encouraged. The formula is first advanced to full strength, then the volume and rate are advanced as stool output allows. Serum electrolytes and stool-reducing substances are monitored frequently. The continuous feedings are begun while the child has an ostomy, and are advanced until the ostomy is closed (generally 3 to 6 weeks). However, the feeding volumes remain small to limit fluid and electrolyte losses from the ostomy. The enteral feeding is resumed after the onset of stools per rectum. Septic episodes were tabulated after closure of all ostomies. Sepsis episodes during “no enteral feeding” periods and during the early onset of feeding with low-volume dilute formula were tabulated together, generally over a 3-to-6-week period. Septic episodes “during feeding” were counted only after full-strength formula was used, usually at 20% to 30% of total caloric needs, for another 4 to 6 weeks. Episodes of sepsis and bacteremia were suspected when an infant demonstrated nonspecific signs and symptoms (fever, lethargy, ileus, vomiting), but were always confirmed by blood cultures obtained from the central venous catheter and from a peripheral site. The organisms were identified by routine culture practices, but quantification of bacteria was not performed. Catheter sepsis was routinely treated with systemic antibiotics without removal of the catheter, except for fungal sepsis which invariably necessitated catheter removal to successfully treat the infection. For this group of patients, the incidence of sepsis, mean number of episodes of sepsis per patient, and type of bacteria isolated were tabulated over a 2-month to 3-year period. These parameters were compared with those of a concurrent group of 20 newborns and infants with similar ages and diagnoses (aged 6 months to 2.5 years; 13 NEC, 4 atresia, 1 gastroschisis, 2 volvulus). This latter group required operative therapy, had central venous catheters in place for a prolonged period both before and after the onset of enteral feeding, but did not have short bowel length. JournalofPecliatr~
Surgery,
Vol30, No 7 (July), 1995: pp
1086-1089
ENTERAL
FEEDING
Table
AND
1. Clinical
Etmlogy of Short Bowel
Necrotizing
SHORT
BOWEL
Characteristics NO
enterocolitis
15
of 21 Patients Resected
30-55 40-80 25-75
(3)
Jejunum-ileum-colon Jejunum-ileum
Gastroschisrs
2
Jejunum-ileum
Midgut
2
Jejunum-ileum
volvulus
SBS Bowel Length or Range (cm)
(6)
2
atresia
With
Bowel
Jejunum-ileum Ileum (5) Ileum-colon
Jejunal-rleal
1087
SYNDROME
(1)
20 45 60 40 80 50 75
RESULTS
The results are summarized in Table 2. In six of the 21 patients (29%) with short bowel length, catheter sepsis developed before enteral feeding was begun. These six patients had eight separate episodes of sepsis (incidence, 1.3 episodes per patient). Five of the six (83%) had infection caused by gram-positive bacteria (StaphyZococc~s), and the other had infection with gram-negative bacteria (Escherichia coli). These data were not significantly different from those of the non-SBS patients. In contrast, after feeding was initiated, 16 of the 21 patients (76%) had 67 episodes of bacteremia (incidence, 4.2 per patient). Both of these were significantly higher than the values before enteral feeding was begun, and also significantly more than those of patients without short bowel length. In addition, gram-negative organisms (E coli, Klebsiella, Pseudomonas)were significantly more common than gram-positive ones when compared with non-SBS patients and with the SBS patients who did not receive enteral feedings (P < .05). DISCUSSION
TPN administered through a central venous catheter is a vital part of the management of infants and Table
2. Characteristics of Septic Episodes in 21 Patients and 20 Matched Patients Without SBS SBS (n = 21)
Sepsrs
With
Non-SBS (n = 20)
incidence
NPO Feeding No. of sepsrs
Gram-posrtive NPO Feeding *P < .05 versus
4 of 20 (20%)
16 of 21 (76%)*
5 of 20 (25%)
eprsodes
NPO Feeding Gram-negative NPO Feedmg
6 of 21 (29%)
1.3 k .2
1.5 + .I
4.2 2 .4*
1.6 + .2
1 of 6 (17%) 13 of 16 (Sl%)*
0 of 4 (0%) 1 of 4 (25%)
5 of 6 (83%)
4of4(100%)
5of16(31%)
3 of 5 (60%)
bacteria
bacteria
NPO and non-SBS.
SBS
children with SBS. Sepsis and bacteremia are well known complications of central venous catheters used for a variety of reasons, but recent data suggest that patients with SBS have an increased incidence of sepsis, especially with enteric gram-negative organisms.2,3 Because enteric organisms are frequently involved in sepsis associated with short bowel length, bacterial translocation has been implicated.3 The present study suggests that enteral feeding increases septic episodes in the eary postoperative period in infants with SBS. Although not measured in this study, the septic episodes appeared to decrease with time, suggesting that the explanation for increased sepsis was, in fact, related to the feeding and not simply to other factors (ie, bacterial overgrowth) associated with increasing time postoperatively. A number of factors have been clinically and experimentally associated with increased incidence of bacterial translocation, many of which are present in the child with short bowel length.5,6 These include prolonged parenteral nutrition, elemental-diet enteric feeding, alteration and overgrowth of gut microflora, enteric stasis, cholestasis, and altered mucosal integrity. Early enteral feedings have been advocated for patients with short bowel length as a means of promoting adaptation and hypertrophy of the remaining gut.’ In addition, some studies suggest that enteric feeding might reduce postoperative septic complications when compared with parenteral feeding.* Although this later observation may be true in the patient with normal or near-normal bowel length, the data in the present study suggest that feeding promotes rather than limits septic episodes in the patient with short bowel length. Enteral feeding with elemental diets has been associated with increased bacterial translocation,9 perhaps because of cytokine activation, mucosal damage, or bacterial overgrowth. All the patients in the present study received continuous elemental diets during the early phases of adaptation, and many continued to receive some elemental diet even after beginning a nonelemental oral diet. It has been suggested9 that elemental diets high in glutamine might maintain mucosal integrity and decrease bacterial translocation, and we are conducting a trial of these formulas in several patients over 1 year of age, because they are not suitable for infant use. Selective modification of the gut flora has been used in other studies with limited success. Nonantibiotic measures, such as high-fiber dietary supplements, have been shown experimentally to reduce bacterial translocation,9 but their usefulness has not been demonstrated clinically. Poorly absorbed oral antibi-
1088
THOMAS
otics, used both cyclically and continuously,7 may have a role in selected patients, but the risk of the development of resistant strains of bacteria and of fungal overgrowth limits the applicability of this approach. Noninfectious inflammatory bowel disease, occasionally present in patients with short bowel length, may respond to oral antinflammatory agents.lO Because the central venous line can be lifesaving in many of these patients, and the number of access sites
R. WEBER
for these catheters is limited, we agree with investigators29 who advocate attempts at salvaging infected catheters with specific antibiotic therapy, rather than immediately removing them. Recognizing that many (if not most) of these infections in patients with short bowel length are caused by enteric organisms, initial antibiotic therapy should be broad-spectrum and include coverage for gram-negative bacteria, until the specific organisms and antibiotic sensitivities are identified.
REFERENCES 1. Weber TF, Tracy TF, Connors RH: Short bowel syndrome in children: Quality of life in an era of improved survival. Arch Surg 126:841-846,199l 2. Piedra PA, Dryja DM, LaScolea LJ: Incidence of catheterassociated gram-negative bacteremia in children with short bowel syndrome. J Clin Microbial 27:1317-1319,1989 3. Kurkchubasche AG, Smith SD, Rowe MI: Catheter sepsis in short bowel syndrome. Arch Surg 127:21-25,1992 4. Alverdy JC, Aoys E, Moss GS: Total parenteral nutrition promotes bacterial translocation from the gut. Ann Surg 202:681684,1985 5. Edmiston CE, Condon RE: Bacterial translocation. Surg Gynecol Obstet 173:73+X3,1991
6. Alexander JW, Boyce ST, Babcock GF, et al: The process of microbial translocation. Ann Surg 212:496-512,199O 7. Vanderhoof JA, Langnas AN, Pinch LW, et al: Short bowel syndrome. J Pediatr Gastroenterol Nutr 14:359-370,1992 8. Moore FA, Feliciano DV, Andrassy RJ, et al: Early enteral feeding, compared with parenteral, reduces postoperative septic complications-The results of a META analysis. Ann Surg 216: 172183,1992 9. Alverdy JC, Aoys E, Moss GS: Effect of commercially available chemically defined liquid diets on the intestinal microflora and bacterial translocation from the gut. JPEN 14:1-6,199O 10. Taylor SF, Sondheimer JM, Sokol RJ, et al: Noninfectious colitis associated with short gut syndrome in infants. J Pediatr 119:24-28,1991
Discussion MM. Ziegler (Cincinnati, OH): This, in my way of thinking, has become a tremendously commonplace problem in the management of these kinds of patients. The one thing in addition to what you have reported that we have seen in Cincinnati is an increased incidence of yeast infection. My practical question is very simple, and that is that now that you have exposed the problem and have suggested some alternate therapies for it, what is the current cocktail that you would suggest be used day to day in the management of these patients? L.A. Martinez (Denver, CO): I have two brief questions. One is, how many of your 21 patients had an ileocecal valve or had lost it? And second, in how many of your patients was the sepsis severe enough that you had to remove the central catheter? KE. Georgeson (Birmingham, AL): We have just reviewed this in our patients and have found that gut lengthening, which was one of the suggestions made, cut our incidence to less than 50% of gram-negative septicemia. The other thing I am wondering about is, have you looked at gut clearance? Very often we look at transit time-that is, how long it takes for the contrast to go
from the tongue to the tail-but then we fail to see how rapidly the material is cleared from the gut. Very often in these patients who have gone through extensive adaptation their bowel dilates, and if you get an x-ray the next day they still have barium hanging around. In those patients I think we’re seeing a much higher incidence of stasis and associated gramnegative septicemia because of translocation. D.H. Teitelbaum (Ann Arbor, MI): How did you express your incidence of sepsis? How long were the patients NPO and how long were they feeding? In other words, if they were NPO for a month and they were feeding for a year, did you express the number of septic episodes per, say, thousand catheter days, and would that change the relative incidence of your septic episodes and decrease it perhaps in that group? R.J. Touloukian (New Haven, CT): This is really an extension of what Dr Georgeson asked you regarding the possibility that you are basically manifesting a partial obstruction in the bowel in those patients who are at high risk for the development bacterial translocation when they are fed. The question is, have you studied these patients radiographically to be certain that their anastomoses are nonobstructed?
ENTERAL
FEEDING
AND
SHORT
BOWEL
SYNDROME
T.R. Weber (response): We continue to use elemental diets. We use oral, poorly absorbed antibiotics and try to switch the patients to either breast milk, if available, or a glutamine-enriched formula as soon as we can. We have noticed in several patients with the glutamine-enriched formula that there is a great decrease in the incidence of translocation. To answer some of the other questions, 50% of our patients in the SBS group have no ileocecal valve, a much greater incidence than in our non-SBS group. That may be related to our incidence of translocation. We have not had to remove a catheter for bacterial
sepsis. However, yeast sepsis certainly is an indication for removing a catheter. We have noticed some decrease in the incidence of translocation and sepsis with gut lengthening and gut tapering procedures, which also, I think, somewhat answers the question about gut clearance. If you can decrease the diameter of the bowel and improve peristalsis, there’s no question that gut clearance of barium and presumably of bacteria also improves. We did not specify our septic episodes per thousand days of catheter use. However, when we looked at that, it didn’t make any difference in the data.