Prognosis of patients with nonmalignant chronic intestinal failure receiving long-term home parenteral nutrition

GASTROENTEROLOGY1995;108:1005-1010

Prognosis of Patients With Nonmalignant Chronic Intestinal Failure Receiving Long-term Home Parenteral Nutrition BERNARD MESSING,* MARC LI~MANN, * PAUL LANDAIS,* MARIE-CLAUDE GOUTTEBEL, § MICHELE GI~RARD-BONCOMPAIN, II FRANCOIS SAUDIN, II ANDRI~ VANGOSSUM, ~ PHILIPPE BEAU, # CLAIRE GUI~DON,** DIDIER BARNOUD, t* MARTINE BELIAH,* HENRI JOYEUX, § PAUL BOULETREAU, II DOMINIQUE ROBERT, Ik CLAUDE MATUCHANSKY, # XAVIER LEVERVE, *t ERIC LEREBOURS,** YVON CARPENTIER, ~ and JEAN-CLAUDE RAMBAUD* • H6pital Saint-Lazare,Paris, France;tD6partementde Biostatistiques,H6pital Necker, Paris, France;§H6pital Paul Lamarque, Montpellier, France; I/H6pitalde la Croix-Rousseet H6pital de I'H6tel Dieu, Lyon, France;~H6pitalErasme, Bruxelles,Belgium; #H6pital La Mil6trie, Poitiers, France; **H6pital Charles-Nicolle,Rouen, France; and ~'tH6pitalde la Tronche, Grenoble,France

I

See editorial on page 1302.

I

Background~Aims: Long-term survival of patients with intestinal failure requiring home parenteral nutrition (HPN) has been only partly shown. Therefore, we described the survival of these patients and explored prognosis factors. Methods: Two hundred seventeen noncancer non-acquired immunodeficiency syndrome adult patients presenting with chronic intestinal failure enrolled from January 1980 to December 1989 in approved HPN programs in Belgium and France; prognosis factors of survival were explored using multivariate analysis. Data were updated in March 1991; not one of the patients was lost to followup. Results: Seventy-three patients died during the survey, and the mortality rate related to HPN complications accounted for 11% of deaths. Probabilities of survival at 1, 3, and 5 years were 91%, 70%, and 62%, respectively. Three independent variables were associated with a decreased risk of death: age of patients younger than 40 years, start of HPN after 1987, and absence of chronic intestinal obstruction. In patients younger than 60 years of age included after 1983 with a very short bowel, who could represent suitable candidates for small bowel transplantation, the 2-year survival rate was 90%, a prognosis that compared favorably with recent reports of survival after small bowel transplantation. Conclusions: HPN prognosis compares favorably with recent reports of survival after small bowel transplantation.

survival of noncancer, non-acquired immunodeficiency syndrome patients with intestinal failure has been only partly shown in two reports. 2'7 In the H P N program at Toronto, survival rate in a cohort of 73 patients enrolled between 1970 and 1982 was 62% at 5 years. 2 The OASIS Registry, coordinated under the hospices of the American Society for Parenteral and Enteral Nutrition, described the outcome of 761 nonmalignant patients during a 4-year period between 1984 and 1987. 7 This study showed that the 3-year survival rate ranged from 65% to 80% according to the underlying disease with the best prognosis for Crohn's disease and the worst for radiation enteritis. However, the OASIS Registry sampled, at the most, 7.9% of the total H P N population, the ultimate survival postH P N course was not known, and prognosis factors other than diagnosis of primary disease were not analyzed] Recent reports of encouraging results with small bowel transplantation 8-1° reinforce the need for more precise evaluation of the outcome of patients receiving long-term HPN. The purpose of the present study conducted in two European countries was therefore to describe the survival of 217 noncancer non-acquired immunodeficiency syndrome adult patients presenting with chronic intestinal failure and enrolled in H P N programs between January 1980 and December 1989. Prognosis factors of survival were explored by multivariate analysis. Materials

he concept of parenteral nutrition as an "artificial gut" and the transfer of parenteral nutrition to the home environment began in North America and Europe in 1970 and 1975, respectively) -6 Home parenteral nutrition (HPN) for nonmalignant patients resulted in a 3.3-year increase in survival compared with the alternative of treating these patients in the hospital with intermittent parenteral nutrition when needed] Long-term

T

and Methods

Patients We analyzed the data from adult patients enrolled in HPN programs of approved centers in Belgium (n = 2) and Abbreviations used in this paper: CI, confidence interval; HPN, home parenteral nutrition. © 1995 by the AmericanGastroenterological Association 0016-5085/95/$3.00

1006

MESSING ET AL.

GASTROENTEROLOGY Vol. 108, No. 4

Table 1. Demographic Characteristics of Patients Sex Male Female Age at HPN initiation Younger than 40 years 4 0 - 6 0 years Older than 60 years Period of HPN initiation 1981-1983 1984-1986 1987-1989 Primary disease Mesenteric infarction Crohn's disease Radiation enteritis Villous atrophy resistant to gluten-free diet Small bowel volvulus Chronic pseudo-obstruction Postoperative complications Miscellaneous

103 (47%) 114 (53%) 86 (40%) 80 (37%) 51 (23%) 36 (17%) 67 (31%) 114 (52%) 59 54 47 12 11 8 8 18

(27%) (25%) (22%) (6%) (5%) (4%) (4%) (8%)

NOTE. n = 217.

France (n = 7) from January 1980 to December 1989. These centers conducted the only available H P N programs in both countries and therefore provided the intervention to virtually all patients undergoing HPN. Patients 15 years of age or older who presented with intestinal failure and submitted to H P N for more than 1 month were considered for inclusion. Patients with evidence of active malignancy at the start of the H P N program, those presenting a malignant tumor diagnosed before initiation of H P N that recurred during the follow-up, and those with human immunodeficiency virus infection were excluded from this study. The decision to institute and pursue H P N was based on the patient's inability to maintain nutritional status without intravenous support because of intestinal failure. H-14 H P N was always delivered through tunneled Silastic catheters (Vygon, Ecouen, France) positioned into the superior vena cava. Nocturnal infusion of solutions from disposable ethylvinyl acetate plastic bags were used in all patients. Composition and volume of the solutions given to the patients were adjusted to individual requirements, and the frequency of infusion ranged from 3 to 7 per week.

Evaluation Methodology All participating H P N centers filled out a four-page questionnaire for all patients matching the inclusion criteria. Data records included demographic characteristics, diagnosis of primary disease, causes of death (if so), and a description of the digestive tract at the initiation of HPN. Special attention was focused on the length of the small bowel and on the presence (or absence) of chronic intestinal obstruction, either functional or mechanical. Data were updated in March 1991. Survival time was calculated from the date H P N started to March 1991 or to the date of death. Probability of survival was calculated using the Kaplan-Meier method. The influence

on survival of concomitant variables was examined using univariate and multivariate analysis. Survival distributions were compared using the log rank test according to sex, age at the start of HPN, date of inclusion in the H P N program, center, diagnosis of primary disease, length of small bowel from Treitz's ligament, and existing chronic intestinal obstruction. A P value of <0.05 was considered as the level of significance. Then, to identify independent prognostic factors, a Cox proportional hazards model 15 was applied with a stepwise procedure to the above variables using the BMDP statistical software. 16 In the multivariate procedure, P < 0.1 was considered, as usual, as the level of significance. Quantitative variables were expressed as mean _+ SD. Both probabilities of survival and relative risks of death were presented with a 95% confidence interval (CI).

Results Two hundred seventeen patients (mean age, 46.5 + 16.9 years; range, 1 5 - 8 7 years) were included in the study. They represented 64% of the patients referred to our H P N centers; 120 patients with malignancy and 2 patients with h u m a n immunodeficiency virus infection were excluded from the study, and no benign but chronic gastrointestinal patient was submitted to H P N for < 1 month. N o patient was lost to follow-up. Main demographic characteristics, diagnosis of primary disease, and digestive status of patients are listed in Tables 1 and 2. Duration of follow-up ranged from 2 to 137 months (median, 72 months). Overall H P N duration ranged from 1 to 137 months (median, 19 months); it ranged from 1 to 104 months (median, 8 months) in the 88 patients who stopped H P N because they no longer needed it, from 2 to 70 months (median, 19 months) in the 77 patients who died during the survey, and from 4 to 137 months (median, 42 months) in the 59 patients who were still receiving H P N in March 1991. Causes of deaths are indi-

Table 2. Digestive Status of Patients Small bowel length < 5 0 cm 5 0 - 1 5 0 cm > 1 5 0 cm No resection Small bowel disease Chronic intestinal obstruction Active disease without obstruction No residual lesion Other features Jejunostomy Colostomy Chronic fistula Gastrectomy Pancreatectomy NOTE. n = 217.

71 44 14 88

(33%) (20%) (7%) (40%)

36 (17%) 75 (35%) 105 (48%) 50 49 12 8 5

(23%) (22%) (6%) (4%) (2%)

April 1 9 9 5

PROGNOSIS OF HOME PARENTERAL NUTRITION

T a b l e 3. Causes of Death

Related to primary disease Related to HPN Sepsis Venous thrombosis Liver failure Related to other diseases Sepsis Cardiovascular disease Secondary cancer Renal failure Respiratory failure Miscellaneous Unknown

22 8 4 1 3 37 12 7 5 4 3 6 6

(30%) (11%)

(51%)

(8%)

NOTE. n = 73.

cated in Table 3. Survival curve for overall patients is shown in Figure 1. Probabilities of survival at 1, 3, and 5 years were 91% (95% CI, 87%-95%), 70% (95% CI, 64%-76%), and 62% (95% CI, 54%-70%), respectively. Figure 2 shows the influence on survival of age at the start of HPN, date of inclusion in the H P N program, diagnosis of primary disease, and existence of a chronic intestinal obstruction. Diagnosis of Crohn's disease, age younger than 41 years at the start of HPN, H P N initiation in 1984 or later, and absence of chronic intestinal obstruction were associated with a better prognosis. In contrast, neither the length of small bowel nor sex seemed significantly associated with an increased risk of death. In the multivariate analysis, three of the variables tested were significantly associated with an increased risk of death: age at the start of HPN (P < 0.001), date of inclusion in the H P N program (P < 0.01), and presence of chronic intestinal obstruction (P < 0.01). The risk of death increased with age; in comparison with patients younger than 40 years of age, relative risk of death was 3.1 (95% CI, 1.5-6.7) for patients 4 0 - 6 0 years of age and 4.9 (95% CI, 2.1-11.4) for those older than 60 years of age. The risk of death was 5.6 (95% CI, 2.4-21) higher in patients who began the H P N program before 1984 than in those included after 1987. Lastly, patients with chronic intestinal obstruction had a 2.6 (95% CI, 1.1-5.8) higher risk of death compared with those without obstruction. The 41 patients younger than 60 years of age included later than 1983 and presenting with a very short bowel ( < 50 cm) who theoretically represent suitable candidates for small bowel transplantation were also studied. Their probability of survival after 1 and 2 years was 98% and 90%, respectively.

Discussion The present study describes the prognosis of 217 adult patients enrolled in all H P N centers from Belgium

1007

and France between January 1980 and December 1989 with an update in March 1991. In France, H P N received a formal recognition from the Ministry of Health in 1984 and was then officially performed through a limited number of regionally distributed "approved" academic centers. In Belgium, H P N was mainly organized from academic "reference" centers. Because of this organization, *v'ls the present study performed in seven centers in France and two centers in Belgium explored virtually all of the nonmalignant H P N population. Moreover, data were exhaustively collected, and none of the patients was lost to follow-up. A doubling of inclusion in H P N programs was observed in each 3-year period of the survey. In the late 1980s, the H P N incidence in France and Belgium reached 0.1/100,000 per year, a figure very close to the one quoted in the United Kingdom, where unregistered cases are also unlikely to account for more than 10%) 9 During our 10-year survey, the probability of survival increased; in the latest period of inclusions (1987-1989), it was 92% and 75% at 1 and 3 years, respectively. Because the date of inclusion was found to be an independent prognosis factor in the multivariate analysis, this improvement in survival is probably not explained by the inclusion of less severe patients with time but reflects the advance in knowledge and experience of H P N centers. It highlights the role of H P N protocols and team approach as prerequisites for H P N program approval to decrease the rate of H P N complications and avoid isolation of patients, v'8'1~'~8-22 In previous studies of H P N prognosis, 2'7 univariate analysis suggested that outcome was principally related to the nature of the underlying disease. In the Toronto study, 2 the 73 noncancer adult patients were divided in two "condition classes"; 64% were labeled as "chronic"

100 8o

%

60 >"

40

.o

e

213

0.

years Number of patients at risk

217

181

153

98

67

46

27

Figure 1. Probability of survival for the 217 patients included in the

HPN programs in France and Belgium between 1980 and 1989.

±008

MESSING ET AL.

GASTROENTEROLOGY Vol. 108, No. 4

A

B 100

100 -

~'""" '" 8O

~'";~""

•

< 40 yrs

,. -. '---1

i=.

80

~

1987-1989

.l

,.~

" " -" •

40- 60 yrs

° . o

.=.==

60

~'l.m.

=.

-. |-i

60

|,

" ' " ~"

I

..... ~ . . . . . . . . . . . . . . . . . . . :..-.. =:

i. I

> 60 yrs..=-.=" ;

40

=l=..

20

i

i

2

40

1981-1983

.................................

20

p = 0.0005

1

i

i

3 years

....ii . . . . . .

4

1984- 198"6- "

5

p = 0.03

l

l

6

1

" i

2

.....

l

!

|

3 years

i

4

5

6

D

C -=,

......

Crohn's disease

80

no obstruction

80

"1 60 / mesenteric infarction

•

..... ~.,

|---~,=

/~ :'- ..... -""I ..... • ....' ° m / other diagnosis

40 20

60

~

-.~....

3 years

[ .....

i

radiation enteritis

i ........

20 0

2

chronic obstruction

40

p = 0.0001

1

"h

=.

4

5

6

1 |

p = 0.002

!

i

1

2

I

3 years

!

I

/

4

5

6

Figure 2. Survival according to (A) age at the start of HPN, (B) date of inclusion in HPN program, (C) diagnosis of primary disease, and (D) presence of chronic intestinal obstruction. Comparisons of survival distributions were made using the log rank test.

(i.e., mainly Crohn's disease and pseudo-obstruction), whereas 36% had "acute" loss of bowel necessitating H P N within 1 year of their insult (i.e., mainly small bowel infarction, volvulus, and radiation enteritis). 2 The 5-year probability of survival was better in the chronic group (75%) than in the acute group (42%), and the investigators suggested that comorbid conditions (in particular, extensive atherosclerosis) observed in 37% of the acute group could account for its worse prognosis. In the OASIS Registry, the 3-year probability of survival was better in Crohn's disease and congenital bowel disease (80%) than in motility disorders, ischemic bowel disease,

and radiation enteritis (72%, 70%, and 65%, respectively), v The U.K. Registry data were not analyzed using an actuarial method but also indicated a trend for a higher death rate in patients with ischemic bowel disease. 19'21 Because the age of the patients is obviously different in the disease categories, it may be hypothesized that difference of age is the main explanation for the effect of primary diagnosis on survival. This appeared clearly in the OASIS report, in which a better survival rate was found in patients with Crohn's disease (median age, 36 years) and congenital bowel dysfunction, v Moreover, in our study, Crohn's disease diagnosis and younger

April 1995

age bore a significantly better prognosis in the univariate analysis, but age only remained a significantly independent factor in the multivariate analysis. Deaths were mainly caused by either n o n - H P N - r e lated diseases (51%) or the evolution of the primary disease (30%). Death related to H P N per se (caused by sepsis, extensive thrombosis of the superior vena cava with pulmonary emboli, or liver failure) played a minor role in overall mortality; it occurred in 8 of 217 patients (4%), representing 11% of deaths, data in the range of the previously reported rates of 3 % - 6 . 5 % and 8 % 30%, respectively)8-2°'23-25 It is of interest to note that probability of death was not dependent on the length of small bowel and that survival was similar for patients with or without short bowel syndrome, even in those with < 5 0 cm of residual gut. In contrast, the presence of chronic intestinal stasis caused by either a mechanical obstacle or intestinal pseudo-obstruction was associated with a higher rate of mortality. Causes of death in patients with chronic intestinal obstruction were mainly represented by sepsis. This result may have practical implications in patients with chronic intestinal obstruction, in whom a surgical approach should be considered even if it lead to a very short bowel syndrome. 17 Until recent years, H P N was the only available treatment for patients with irreversible intestinal failure, and our data confirm the long-term efficacy and safety of this therapy when conducted under the control of specialized centers. 2'v'ls-2° This approach might change in the near future with the improvement of surgical strategies such as reversed loop, 26'27 intestinal lengthening procedures, 28 and the development of small bowel transplantation. 8 lo Although this latter approach is still considered experimental, encouraging results using FK506 as the major immunosuppressive agent have recently been presented. 9'1° Todo et al. and Abu-Elmagd et al. reported a 2-year survival rate of 74% in 43 patients treated with small bowel transplantation during the 1990-1993 period, 65 % of them also receiving liver or multiple organ transplantation. In our study, considering the patients younger than 60 years of age, included later than 1983, with a very short bowel syndrome ( < 5 0 cm) who would represent suitable candidates for small bowel transplantation, we found that the 2-year survival rate was 90%. Thus, at the present time, it seems reasonable to consider for small bowel transplantation only the subgroup of patients receiving H P N with an increased risk of mortality caused by H P N complications such as extensive thrombosis of the superior vena cava, severe metabolic disturbances, or advanced liver disease.

PROGNOSIS OF HOME PARENTERAL NUTRITION

1009

References 1. Detsky AS, McLaughlin JR, Abrams HB, Wittaker JS, Whitwell J, L'Abbe K, Jeejeeboy KN. A cost-utility analysis of the HPN programme at Toronto General Hospital: 1970-1982. JPEN 1986; 10:49-57. 2. Dudrick SJ, Englert DM, Van Buren CT, Rowlands BJ, MacFayden BV. New concepts of ambulatory home hyperalimentation. JPEN 1979;3:72-76. 3. Ladefeged K, Jarnum S. Long-term parenteral nutrition. Br Med J 1978;2:262-266. 4. Ricour C, NihouI-F6f6t6 C. Nutrition parent~rale prolong6e ehez I'enfant. Arch Fr Pediatr 1973;30:469-480. 5. Scribner BH, Cole JJ, Christopher TG. Long-term total parenteral nutrition. The concept of an artificial gut. JAMA 1970; 212:457463. 6. Solassol C, Joyeux H. Ambulatory parenteral nutrition. In: Manni C, Magalini SI, Scrascia E, eds. Parenteral alimentation. The international symposium on intensive therapy. New York: Elsevier, 1976:138-152. 7. Howard L, Heaphey LL, Fleming CR, Lininger L, Steiger E. Four years of North American Registry home parenteral nutrition: outcome data and their implications for patient management. JPEN 1991; 15:384-394. 8. Grant D, Wall W, Mimeault R, Zhong R, Ghent C, Garcia B, Stiller C, Duff J. Successful small-bowel/liver transplantation. Lancet 1990;335:181-200. 9. Todo S, Tsakis AG, Abu-Elmagd K, Reyes J, Nakamura K, Casavilla A, Selby R, Nour B, Wright H, Fung JJ, Demetris A, Van Thiel DH, Starzl TE. Intestinal transplantation in composite viscera[ grafts or alone. Ann Surg 1992;216:223-234. 10. Abu-Elmagd K, Todo S, Tsakis A, Reyes J, Nour B, Fuzukawa H, Fung JJ, Demetris A, Starzl TE. Three years clinical experience with intestinal transplantation. J Am Coil Surg 1994;179: 385-400. 11. American Society for Parenteral and Enteral Nutrition. Standards for home nutrition support. JPEN 1993; 17:10-11SA. 12. Fleming CR, Remington M, Intestinal failure. In: Hill GI, ed. Nutrition and the surgical patient. Clinical Surgery International. Edinburgh: Churchill Livingstone, 1981:219-235. 13. Irving MH. Intestinal failure and its treatment by home parenteral nutrition. In: Johnston IDA, ed. Home parenteral nutrition. Lancaster, England: MTP Press, 1983:1-13. 14. Messing B, Beliah M, Girard-Pipau F, Leleve D, Bemier JJ. Technical hazards of using nutritive mixtures in bags for cyclical intravenous nutrition: comparison with standard intravenous nutrition in 48 gastroenterological patients. Gut 1982;23:297-303. 15. Cox DR, Snell EJ. Analysis of binary data. 2nd ed. London: Chapman et Hall, 1989. 16. BMDP Statistical software. Dixon WJ, ed. Berkeley: University of California, 1988. 17. L6mann M, Messing B, Medhi A, Thuillier F, Pignon JP, Beliah M, Rambaud JC. Long-term prognosis in patients on HPN: a 10year experience (abstr). Clin Nutr 1991;10(Suppl 2):43A. 18. Messing B, Landais P, Goldfarb B, L6mann M, Joyeux H, Gouttebel M-C, Robert D, Bouletreau P, Matuchausky C, Beau P, Colin R, Lezebouzs E. Nutrition parent~rale ~ domicile chez I'adu[te: resultats d'une enquete multicentrique en France. Presse Med 1988; 17:845-849. 19. Mughal M, Irving MH. Home parenteral nutrition in the United Kingdom and Ireland. Lancet 1986; 328:383-386. 20. Burnes JU, O'Keefe JD, Fleming CR, Devine RM, Berkner S, Herrick L. Home parenterat nutrition: a three year analysis of clinical and laboratory monitoring. JPEN 1992;16:327-332. 21. O'Hanrahan T, Irving M. The role of home parenteral nutrition in

1010

22. 23.

24.

25. 26.

MESSING ET AL.

the management of intestinal failure--report of 400 cases. Clin Nutr 1992; 11:331-336. Burgess P, Irving MH. Problems and organization of a home parenteral nutrition service. Clin Gastroenterol 1988;2:905-914. Messing B, Landais P, Goldfarb B, Irving MH. Home parenteral nutrition in adults: a multicentre survey in Europe. Clin Nutr 1989;8:3-9. Steiger E, Srp F. Morbidity and mortality related to home parenteral nutrition in patients with gut failure. Am J Surg 1983; 145:102-105. Stokes MA, Irving MH. Mortality in patients on home parenteral nutrition. JPEN 1989;13:172-175. Coffin B, Hautefeuille P, Flouri6 B, Valleur P, Rambaud JC, Messing B. Consequences of the interposition of a reversed small bowel loop in patients with short bowel syndrome requiring parenteral nutrition. Clin Nutr 1994; 13(Suppl 1):039.

GASTROENTEROLOGY Vol. 108, No. 4

27. Pigot F, Messing B, Chaussade S, Pfeiffer A, Pouliquen X, Jian R. Severe short bowel syndrome with a surgically reversed small bowell segment. Dig Dis Sci 1990;35:137-144. 28. Thompson JS, Pinch LW, Murray ND, Vanderhoof JA, Schultz LR. Experience with intestine lengthening for the short bowel syndrome. J Pediatr Surg 1991;26:721-724.

Received July 7, 1994. Accepted December 20, 1994. Address requests for reprints to: Bernard Messing, M.D., HSpital Saint-Lazare, 107 rue du Fbg Saint-Denis, 75010 Paris, France. Fax: (33) 1-48-00-55-81. Presented in part at the European Society for Parenteral and Enteral Nutrition in Budapest, Hungary, in September 1993 and published in abstract form (Clin Nutr 1993;12[Suppl 2]:9A).