Enteral compared with parenteral nutrition: a meta-analysis1,2

Carol L Braunschweig, Paul Levy, Patricia M Sheean, and Xin Wang

ABSTRACT these generally accepted benefits (2). The review, although com-
Background: The difference in outcomes in patients is unclear prehensive, did not systematically compile results or evaluate the
when 2 types of enteral nutrition, ie, tube feeding and conven- quality of the studies’ methods, and it also included many stud-
tional oral diets with intravenous dextrose (standard care), are ies conducted in populations that would not be considered can-
compared with parenteral nutrition. didates for parenteral nutrition by today’s standards. Heyland et
Objective: We reviewed systematically and aggregated statisti- al (3) conducted a meta-analysis of PRCTs that evaluated the
cally the results of prospective randomized clinical trials outcomes of parenteral nutrition compared with those of stan-
(PRCTs) to examine the relations among the nutrition interven- dard care (conventional oral diets with intravenous dextrose) in
tions, complications, and mortality rates. surgical or critically ill patients. They found that parenteral
Design: We conducted a MEDLINE search for PRCTs comparing nutrition did not influence mortality rates; however, a trend
the effects of enteral and parenteral nutrition in adults. Two dif- toward fewer complications, particularly in populations that had
ferent people abstracted data for the method and outcomes sepa- protein-energy malnutrition (PEM), was reported. Their analysis
rately. We used fixed-effects meta-analysis technique to combine included studies that provided parenteral nutrition in amounts
the relative risks (RRs) of the outcomes of infection, nutrition less than the estimated energy and protein needs of the patients
support complications, other complications, and mortality. and trials conducted in populations with functional gastrointesti-
Results: Twenty-seven studies in 1828 patients met the study cri- nal tracts. Both of these factors could have reduced the influence
teria. Aggregated results showed a significantly lower RR of infec- of parenteral nutrition on outcomes. Also, they did not include
tion with tube feeding (0.64; 95% CI: 0.54, 0.76) and standard investigations that compared parenteral nutrition with tube feed-
care (0.77; 95% CI: 0.65, 0.91). A priori hypotheses showed a ing. The purpose of this paper was to review systematically and
lower RR of infection with tube feeding than with parenteral nutri- aggregate statistically the PRCTs that were conducted in popula-
tion, regardless of nutritional status, presence of cancer, year of tions appropriate for random assignment to parenteral nutrition
study publication, or quality of the study method. In studies in and to compare the effects of parenteral nutrition with those of
which participants had high rates of protein-energy malnutrition, aggressive enteral nutrition (tube feeding) and limited nutritional
there was a significantly higher risk of mortality (3.0; 95% CI: intervention (standard care) on outcomes in which parenteral
10.9, 8.56) and a trend toward a higher risk of infection with stan- nutrition was provided at or above estimated energy needs.
dard care than with parenteral nutrition (1.17; 95% CI: 0.88, 1.56).
Conclusions: Tube feeding and standard care are associated with
a lower risk of infection than is parenteral nutrition; however, METHODS
mortality is higher and the risk of infection tends to be higher
with standard care than with parenteral nutrition in malnourished Sources and criteria of PRCT selection
populations. Am J Clin Nutr 2001;74:534–42. MEDLINE (National Library of Medicine, Bethesda, MD)
was searched for PRCTs that evaluated the effects of parenteral
KEY WORDS Meta-analysis, enteral nutrition, parenteral nutrition compared with tube feeding or standard care and that
nutrition, prospective randomized clinical trials, tube feeding, were conducted from 1966 to November 1999. Only PRCTs that
oral diets, intravenous dextrose, MEDLINE evaluated the effect of parenteral nutrition administered at or
above estimated energy needs compared with those of tube feed-
ing or standard care on outcomes with clinical significance (mor-
INTRODUCTION bidity and mortality) were reviewed; studies that evaluated only
Parenteral nutrition is an invasive therapy that provides nutri-
tion support for persons who do not have adequate gastrointesti- 1
From the Department of Human Nutrition and Dietetics and the Division
nal functions; however, it does have inherent risks (1). Enteral
of Epidemiology, University of Illinois at Chicago.
nutrition, specifically tube feeding, is the preferred method of 2
Address reprint requests to CL Braunschweig, University of Illinois at
feeding because it is cheaper, has fewer complications, and has Chicago, Department of Human Nutrition, 1919 W Taylor Street (M/C 517),
better outcomes than does parenteral nutrition. A review of the Chicago, IL 60612. E-mail: braunsch@uic.edu.
prospective randomized clinical trials (PRCTs) that compared Received September 22, 2000.
tube feeding with parenteral nutrition cast doubt on some of Accepted for publication January 5, 2001.

534 Am J Clin Nutr 2001;74:534–42. Printed in USA. © 2001 American Society for Clinical Nutrition
 ENTERAL COMPARED WITH PARENTERAL NUTRITION 535

TABLE 1
Criteria met by study populations included in meta-analysis1
Criteria Examples
Malabsorptive syndromes with severe food, electrolyte, and fluid losses Severe short-bowel syndrome
not adequately managed by oral or enteral nutrition Those induced by infection, inflammatory, and immunologic
disorders, drugs, or radiation
High-output gastrointestinal fistulas that enteral intubation cannot bypass
Severe renal tubular defects with large fluid and ion losses
Motility disorders Persistent ileus (postoperative or disease related)
Severe intestinal pseudoobstruction
Severe persistent vomiting caused by medication, brain tumor, or other
disorder (eg, hyperemesis gravidarum)
Mechanical intestinal obstruction not immediately remedied by surgery —
Perioperative state with severe undernutrition —
Critically ill patients, especially those with hypermetabolism, who are not —
appropriate for enteral nutrition because it was contraindicated or failed
1
Adapted from reference 1.

the effect on nutritional outcomes (eg, nitrogen balance and sented as total complications instead of complications per patient
serum protein) were not included because these were considered were not included. Patients recruited into the study were
surrogate endpoints. Studies were limited to the ones that involved recorded as the total number included in each treatment arm.
English-speaking adult patients. Reference lists from studies found Thus, all of our relative risks (RRs) were calculated on the basis
by the search were reviewed for additional reports. Also, Ronald of intent-to-treat numbers. Two of the meta-analysis investiga-
Koretz provided a list of all the studies published about nutrition tors, whose ratings for study-quality scores and outcomes were
support that he had compiled for > 20 y; some of these were found blinded to one another, reviewed and extracted data with the use
in his manual search of the literature that was completed for of preset criteria. Disagreements among ratings in study evalua-
the Cochrane Collaboration, which publishes electronically the tions were resolved by consensus.
COCHRANE DATABASE OF SYSTEMATIC REVIEWS (The
Cochrane Library, Oxford, United Kingdom). Study-quality score and subgroup analysis
Because differences in study populations and design might
Nutritional interventions cause variations in results, 4 sources for heterogeneity were
Parenteral nutrition is designed to provide nutrition to defined a priori: 1) study-quality score, 2) year of study publica-
patients who cannot be nourished adequately by enteral nutri- tion, 3) nutritional status of patients, and 4) percentage of patients
tion for a critical period of time. Because of the inherent risk with cancer.
and higher cost of parenteral nutrition, it should not be used as The method used in each study was evaluated for the quality
a substitute for enteral nutrition if either standard care or tube of these characteristics: concealed randomization, comparability
feeding is feasible. Also, because of physiologic changes that of groups at baseline, endpoints (blinded to staff or not), well-
are caused by gastrointestinal tract dysfunction, outcomes described treatment protocols, well-defined outcomes, and analy-
observed in populations with functional gastrointestinal tracts sis by intent to treat. One point was given for each of these traits
might be different from those observed in populations without and a study-quality score that ranged from 0 to 6 for each inves-
adequate gastrointestinal function. The criteria met by popula- tigation was calculated. Studies were categorized into those with
tions included in the meta-analysis are listed in Table 1. Tube low study-quality score (< 4 points) and those with a high study-
feeding was defined as either surgical or nonsurgical placement quality score (≥ 4 points), and separate subgroup analyses were
of a small flexible tube into the gastrointestinal tract to provide performed for each category.
required nutrients. Standard care was defined as the gradual rein- PEM in hospitalized patients has been associated with high
troduction of an oral diet as tolerated after its interruption was rates of morbidity and mortality (4–6). To address this possible
caused by a disease or a surgical procedure that resulted in sev- source of heterogeneity, the nutritional status of patients at the
eral days of inadequate nutrient intake and the use of intra- time of enrollment was examined. Each investigator’s definition of
venous dextrose or fluids for hydration. PEM was used when possible. When this was not specifically
stated, an unplanned body weight loss of ≥ 15% of normal body
Outcome descriptions and data extraction weight was used as the cutoff for PEM classification. Studies were
The primary outcomes assessed were infection, nutrition sup- categorized into 2 groups on the basis of the percentage of partic-
port complications, other complications, and mortality. Only ipants with PEM (% PEM) those with < 50% (low) and those with
deaths that occurred during a patient’s hospitalization were ≥ 50% (high), and the subgroups were analyzed separately.
included. The specific disorders included within each outcome The skill with which specialized nutrition support is provided
category are shown in Table 2. “Other complications” included has improved since it was first used in patient care. To assess
any reported major or minor complication that developed during whether time-induced changes led to discrepant results, studies
hospitalization but were not included as an infection or nutrition were divided into equal groups of those that were relatively
support complication. The definitions of the variables given by recent (1992 or later) and those that were not recent (earlier than
the authors were used unless they indicated otherwise. All data 1992) publications, and separate subgroup analyses were per-
were generated on a per-patient basis. Unclear data and data pre- formed for each category.
536 BRAUNSCHWEIG ET AL

TABLE 2 centage of participants with cancer: < 50% (low) and ≥ 50%
Primary outcome categories of meta-analysis (high), and the subgroups were analyzed separately.
Primary outcome Analysis
category Disorder
The PRCTs were divided into 2 categories: 1) tube feeding
Infection Catheter sepsis
Pneumonia
compared with parenteral nutrition and 2) standard care com-
Abscess pared with parenteral nutrition. This categorization allowed for
Empyema the examination of aggregated differences observed in outcomes
Infection when the route of aggressive nutrition support varied (tube feed-
Blood ing compared with parenteral nutrition) and when aggressive
Urine intravenous nutrition support was compared with little or no
Wound nutritional intervention (standard care compared with parenteral
Intraabdomen nutrition). To avoid the loss of each group’s independence, meta-
Nutrition support Parenteral or enteral nutrition technical problems1 analysis techniques were used to prevent any group from being
complications Pneumothorax
compared with more than one other group. For example, a par-
Hemothorax
Subclavian artery puncture
enteral nutrition group could not be compared with a tube-fed
Cardiac perforation and tamponade group and then with a standard care group because it would then
Brachial plexus injury not be an independent comparison. Hence, 5 investigations used
Innominate or subclavian vein laceration a method to randomly assign patients that resulted in some of the
Carotid artery injury groups being eliminated from our analysis. Holter and Fischer (8)
Thromboembolism and Thompson et al (9) randomly assigned patients to normally
Catheter embolism nourished standard care control, malnourished parenteral nutri-
Catheter malposition tion, and malnourished standard care groups; only the malnour-
Thoracic duct laceration ished parenteral nutrition and standard care groups were included
Subclavian hematoma
in the analysis. Greenberg et al (10) randomly assigned patients
Subclavian air embolism
Mechanical problems
to parenteral nutrition, tube-fed, partial parenteral nutrition, and
Dislodged or occluded tube or catheter partial standard care groups; only those in the parenteral nutrition
Aspiration2 and tube-fed groups were included in the analysis. Dunham et al
Vomiting, diarrhea, or constipation3 (11) randomly assigned patients to parenteral nutrition, tube-fed,
Fistula at catheter or tube site partial parenteral nutrition, and partial enteral nutrition groups;
Hyperglycemia4 only those randomly assigned to the parenteral nutrition and tube-
Other complications Organ failure fed groups were included in the analysis. Von Meyenfeldt et al
Hepatic (12) randomly assigned patients to 4 groups: malnourished par-
Renal enteral nutrition, malnourished tube- fed, malnourished standard
Respiratory
care, and normally nourished standard care groups; only the mal-
Cardiac
Reoperation
nourished parenteral nutrition and tube-fed groups were included.
Pancreatitis RRs and 95% CIs were calculated for each investigation and
Anastomotic leak for each outcome variable. A fixed-effects meta-analysis tech-
Pulmonary emboli nique modeled after the Mantel-Haenszel method was used to
Myocardial infarct or arrhythmia estimate summary RRs and their 95% CIs (13). CIs for the com-
Coronary vascular accident mon RR that did not include unity were considered significant. To
Gastric outlet obstruction determine whether the investigations within each group for each
Aortic aneurysm outcome had widely discrepant RRs, a test for heterogeneity was
Deep vein thrombosis done (14). We considered P < 0.05 to be statistically significant.
Laryngeal nerve palsy
Fistulas
Intestinal obstruction
RESULTS
Hemorrhage
Aspiration Seventy-six citations were identified in the MEDLINE search
Gastrointestinal bleeding and 22 additional studies were obtained from Koretz’ list. Of
Mortality — these studies, 27 met our inclusion criteria and included a total of
1
Caused by catheter or tube insertion. 1829 patients (n = 895 enteral nutrition, n = 934 parenteral nutri-
2
Caused by feeding. tion) with compromised gastrointestinal function caused by pan-
3
Required medical treatment or a feeding that lasted for ≥ 24 h. creatitis (15–17), ulcerative colitis (18), Crohn disease (10), surgery
4
Hyperglycemia: glucose > 11 mmol/L (> 200 mg/dL). (8–9, 12, 19–31), trauma (11, 32–35), or multisystem organ failure
(36). The 27 PRCTs are referenced in Table 3, including informa-
tion on study-quality scores, PEM, and outcomes.
Finally, because previous investigations reported little or no
effect of specialized nutrition support on outcomes in cancer Effect of tube feeding compared with parenteral nutrition
patients (7), we speculated that differences in outcomes might be on risk of outcomes
caused by the presence of cancer in the patients. Therefore, the Twenty studies in 1033 patients had information that
studies were categorized into 2 groups on the basis of the per- allowed the calculation of the RRs of the primary outcome
 ENTERAL COMPARED WITH PARENTERAL NUTRITION 537

TABLE 3
Prospective randomized clinical trials of enteral nutrition (EN; tube feeding or standard care) compared with parenteral nutrition (PN)
Outcomes
Number of Study- Nutrition support Other
subjects quality Infection complications complications Mortality
Reference EN PN score PEM1 Cancer2 EN PN EN PN EN PN EN PN
% n (%)
Tube feeding
Adams et al, 1986 (32) 23 23 3 0 No 15 (65) 17 (74) 14 (61) 9 (39) 4 (17) 4 (17) 1 (4) 3 (13)
Baigrie et al, 1996 (19) 50 47 2 0 No 2 (4) 10 (21) 10 (20) 14 (30) 15 (30) 23 (49) 4 (8) 6 (13)
Bower et al, 1986 (20) 10 10 3 35 No 0 0 3 (30) 2 (20) 0 0 0 0
Cerra et al, 1988 (36) 33 37 5 0 No 0 0 25 (76) 9 (24) 7 (21) 7 (19) 7 (21) 8 (22)
Dunham et al, 1994 (11) 12 15 5 0 No 0 0 2 (17) 2 (13) 0 0 1 (8) 1 (7)
Gonzalez-Huiz et al, 1993 (18) 23 21 3 39 No 1 (4) 8 (38) 1 (4) 3 (14) 11 (52) 11 (52) 0 0
Greenberg et al, 1998 (10) 19 32 4 0 No 0 0 0 0 0 0 0 0
Hamaoui et al, 1990 (21) 11 8 4 0 Yes 1 (9) 0 1 (9) 1 (13) 0 0 1 (9) 0
Iovinelli et al, 1993 (22) 24 24 3 0 Yes 5 (241) 4 (17) 0 7 (29) 1 (4) 2 (8) 0 0
Kalfarentzos et al, 1997 (15) 18 20 3 0 No 6 (33) 15 (75) 4 (22) 9 (45) 2 (11) 7 (35) 1 (6) 2 (10)
Kudsk et al, 1992 (33) 52 46 5 0 No 9 (17) 18 (39) 1 (2) 0 3 (6) 4 (9) 1 (2) 1 (2)
Kudsk et al, 1994 (34) 34 34 5 0 No 5 (15) 14 (41) 0 0 0 0 1 (3) 0
Lim et al, 1981 (23) 12 12 2 100 Yes 5 (42) 5 (42) 0 1 (8) 7 (58) 3 (25) 2 (17) 1 (8)
McClave et al, 1997 (16) 16 16 4 0 No 2 (12) 2 (12) 5 (31) 5 (31) 0 0 0 0
Moore et al, 1989 (35) 29 30 5 0 No 5 (17) 11 (37) 0 0 6 (21) 7 (23) 0 0
Reynolds et al, 1997 (24) 33 34 4 81 Yes 10 (30) 19 (56) 3 (9) 0 11 (33) 6 (18) 2 (6) 1 (3)
Sako et al, 1981 (25) 33 36 3 38 Yes 3 (9) 2 (6) 0 2 (6) 6 (18) 6 (17) 0 3 (8)
Sand et al, 1997 (26) 13 16 3 0 Yes 3 (23) 5 (31) 0 0 3 (23) 3 (19) 0 1 (6)
Shirabe et al, 1997 (27) 13 13 3 0 Yes 1 (8) 8 (62) 0 0 0 0 0 0
Von Meyenfeldt et al, 1992 (12) 50 51 4 100 Yes 30 (60) 43 (84) 0 2 (4) 10 (20) 10 (20) 4 (8) 2 (4)
Standard care
Brennan et al, 1994 (28) 57 60 3 0 Yes 20 (35) 36 (60) 0 2 (3) 18 (32) 35 (58) 1 (2) 4 (7)
Holter and Fischer, 1977 (8) 26 30 4 100 Yes 1 (4) 2 (7) 0 0 5 (19) 4 (13) 2 (8) 2 (7)
Muller et al, 1982 (29) 59 66 2 60 Yes 38 (64) 36 (55) 0 2 (3) 0 0 11 (19) 3 (5)
Sandstrom et al, 1993 (30) 150 150 5 22 Yes 57 (38) 91 (61) 9 (6) 12 (8) 89 (59) 95 (63) 10 (7) 12 (8)
Sax et al, 1987 (17) 26 29 3 0 No 1 (4) 3 (10) 0 2 (7) 0 1 (3) 1 (4) 1 (3)
Thompson et al, 1981 (9) 9 12 3 100 Yes 2 (22) 3 (25) 0 1 (8) 0 0 0 0
Woolfson and Smith, 1989 (31) 60 62 5 0 No 4 (7) 7 (11) 22 (37) 16 (26) 4 (7) 6 (10) 8 (13) 8 (13)
1
PEM, protein-energy malnutrition ≥ 50%.
2
Studies in which ≥ 50% of the study population had cancer.

variables and compared tube feeding with parenteral nutrition and their subgroup analyses (Figure 6 and Figure 7) were aggre-
(n = 508 tube feeding, n = 525 parenteral nutrition). When the gated, they were not significantly different.
results of these trials were aggregated, tube feeding was asso-
ciated with a significantly lower risk of infection (Figure 1) Effect of standard care compared with parenteral nutrition
(RR: 0.66; 95% CI: 0.56, 0.79). The test for heterogeneity was on risk of outcomes
not significant. Risk of nutrition support complications was Seven studies in 798 patients (n = 387 standard care, n = 409
higher for tube feeding than for parenteral nutrition (Figure 2) parenteral nutrition) compared standard care with parenteral
(RR: 1.36; 95% CI: 0.96, 1.83); however, the strength of this nutrition and had information that allowed the calculation of the
association is questionable because the test for heterogeneity RRs of the primary outcome variables. When the results of these
was significant (P = 0.03). No treatment effect for tube feeding trials were aggregated, a significantly lower risk of infection
was observed for other complications (Figure 3) (RR: 0.92; (RR: 0.77; 95% CI: 0.65, 0.91; Figure 1) and a trend toward fewer
95% CI: 0.71, 1.22) or for mortality (Figure 4) (RR: 0.96; 95% other complications (RR: 0.87; 95% CI: 0.74, 1.03; Figure 3) were
CI: 0.55, 1.65). found for those randomly assigned to standard care than for
Tube feeding was associated with a lower risk of infection in those assigned to parenteral nutrition. The test for heterogeneity
all subgroup categories (Figure 5). A significantly higher risk of for the aggregated RR of infection was significant (P = 0.03).
nutrition support complications was seen for tube feeding in Only 2 studies reported data that allowed for the calculations of
studies published in 1992 or earlier (RR: 2.12; 95% CI: 1.43, the RRs of nutrition support complications. Thus, this outcome
3.14) and in studies with higher study-quality scores (RR: 2.22; could not be aggregated meaningfully. There were no significant
95% CI: 1.4, 3.53) than for parenteral nutrition. When the results differences in aggregated risk of mortality (RR: 1.14; 95% CI:
of the risk of other complications and mortality (Figures 3 and 4) 0.69, 1.88; Figure 4).
538 BRAUNSCHWEIG ET AL

FIGURE 3. Risk factors and associated 95% CIs for the effect of
enteral nutrition (tube feeding or standard care) compared with that of
parenteral nutrition on other complications.
FIGURE 1. Risk factors and associated 95% CIs for the effect of
enteral nutrition (tube feeding or standard care) compared with that of
Additional analysis for other complications could not be completed
parenteral nutrition on infection.
because of the inadequate numbers in the different categories.
To determine whether the lower risk of infection observed
In studies in populations with high percentages of PEM, stan- with enteral nutrition was due to the categorization of catheter
dard care was associated with a significantly higher risk of sepsis as an infection instead of a nutritional support complica-
mortality (RR: 3.0; 95% CI: 1.09, 8.56; Figure 7) and a trend tion, these events were removed. The RRs and 95% CIs were
toward a higher risk of infection (RR: 1.17, 95% CI: 0.88, 1.56; recalculated for studies that reported them, and new aggregated
Figure 5) than was parenteral nutrition. The risk of infection was RRs were calculated. There were a total of 36 reported events of
found to be lower for standard care than for parenteral nutrition in catheter sepsis (n = 2 tube feeding compared with n = 23 par-
the relatively recent studies (RR: 0.62; 95% CI: 0.50, 0.76), stud- enteral nutrition; n = 1 standard care compared with n = 10 par-
ies with low rates of PEM (RR: 0.61; 95% CI: 0.50, 0.76), studies enteral nutrition). The risk of infection remained significantly
with a high percentage of patients with higher study-quality scores lower in a comparison of parenteral nutrition for both tube feed-
(RR: 0.63; 95% CI: 0.49, 0.80), and studies with a high percent- ing and standard care (RR: 0.70; 95% CI: 0.58, 0.83) and stan-
age of patients with cancer (RR: 0.78; 95% CI: 0.65, 0.92; Fig- dard care (RR: 0.79; 95% CI: 0.67, 0.94). To determine whether
ure 5). Subgroup analyses could not be completed for nutrition the higher risk of nutrition support complications in tube-fed
support complications because of the insufficient number of stud- populations was due to the exclusion of catheter sepsis from
ies that reported this information. The risk of other complications
was lower for standard care than for parenteral nutrition in studies
with lower study-quality scores (RR: 0.61; 95% CI: 0.4, 0.91).

FIGURE 2. Risk factors and associated 95% CIs for the effect of FIGURE 4. Risk factors and associated 95% CIs for the effect of
enteral nutrition (tube feeding) compared with that of parenteral nutri- enteral nutrition (tube feeding or standard care) compared with that of
tion on nutrition support complications. parenteral nutrition on mortality.
 ENTERAL COMPARED WITH PARENTERAL NUTRITION 539

FIGURE 5. Results of subgroup analyses of the effect of enteral FIGURE 7. Results of subgroup analyses of the effect of enteral
nutrition (tube feeding or standard care) compared with that of par- nutrition (tube feeding or standard care) compared with that of par-
enteral nutrition on infection. PEM, protein-energy malnutrition. enteral nutrition on mortality. PEM, protein-energy malnutrition.

this category, this event was categorized as a nutrition support teria, which prevents the occurrence of bacterial translocation
complication. The RRs and 95% CIs were recalculated for stud- (37–39); however, the role that enteral nutrition plays in pre-
ies that reported them, and new aggregated RRs were calculated. venting bacterial translocation in human investigations is still
The risk of nutrition support complications was reduced from debated (40). We cannot suggest that our findings support the
1.35 to 1.05, and the CI included unity (95% CI: 0.79, 1.4). hypothesis of a protective role for aggressive enteral nutrition
(tube feeding) in preventing bacterial translocation because there
are several aspects of our meta-analysis design that prohibited
DISCUSSION this. Specifically, none of the studies included in our analysis
Our primary objective was to determine whether aggregated investigated bacterial translocation as a primary hypothesis, and
results of PRCTs supported the use of tube feeding instead of all types of infection were grouped together within our infection
parenteral nutrition. An additional objective was to determine category, which further limited interpretation. Also, we cannot
whether the risk of infection , nutrition support complications, determine from our results whether aggressive enteral nutrition
other complications, and mortality were greater with standard reduced the risk of infection or whether parenteral nutrition led
care or with parenteral nutrition. We found that both tube feed- to a higher risk of infection. Our finding of fewer infections
ing and standard care were associated with a lower risk of infec- associated with both tube feeding and standard care than with
tion than was parenteral nutrition, and this lower risk was not an parenteral nutrition in normally nourished populations suggests
artifact of whether catheter sepsis was included in our analysis. that it is not that enteral nutrition does not result in a lower risk
For tube feeding, this lower risk remained regardless of the pres- of infection but rather that parenteral nutrition results in a higher
ence of cancer, nutritional status, year of study publication, or risk of infection.
study-quality score. The higher risk of infection associated with parenteral nutri-
It was reported in animals that enteral nutrition more so than tion may be partially explained by the higher number of patients
does parenteral nutrition lowers the risk of infection by preserv- with hyperglycemia in this population. Elevated glucose con-
ing the gastrointestinal tract’s integrity and enhancing its ability centrations reduce neutrophil chemotaxis and phagocytosis and
to provide an immunocompetent barrier to endogenous gut bac- were found to be an independent risk factor for short-term infec-
tion in patients undergoing coronary artery surgery (41). Of the
20 trials that compared tube feeding with parenteral nutrition,
16 included data on infection (12, 15, 16, 18, 19, 21–27, 32–35)
and 7 reported data on hyperglycemia (15, 16, 18, 20, 21, 22,
23), 6 of which reported data on both infection and hypergly-
cemia (15, 16, 18, 21, 22, 23). In all of these investigations,
hyperglycemia occurred more frequently in patients who
received parenteral nutrition than in those who were tube fed;
however, because of the way the data were presented, it was not
possible to discern whether persons with infection were also
those with hyperglycemia. To be included in this meta-analysis,
feeding protocols for both tube feeding and parenteral nutrition
FIGURE 6. Results of subgroup analyses of the effect of enteral had to provide energy at or above estimated needs; however,
nutrition (tube feeding) compared with that of parenteral nutrition on tube-fed patients frequently receive less than the amount pre-
other complications. PEM, protein-energy malnutrition. scribed because of feeding intolerance and interruptions. To
540 BRAUNSCHWEIG ET AL

determine whether the hyperglycemia reported was caused by associated with a lower risk of infection. These findings illus-
the greater energy infusion received by parenteral nutrition than trate 2 major points: 1) failure to provide adequate nutrition to
by tube feeding, the 5 investigations that reported both infection populations with PEM is associated with untoward conse-
and hyperglycemia were reviewed. Patients randomly assigned quences, and 2) parenteral nutrition should not be initiated in
to receive parenteral nutrition received less energy than did the normally nourished populations, unless there is a good reason
tube-fed patients in one investigation (18), approximately equal to do so.
amounts in 4 investigations (15, 16, 22, 23), and greater amounts The issue of the number of days to wait before initiating
in one investigation (21). Thus, the hyperglycemia was not enteral or parenteral nutrition in normally nourished populations
caused by a difference in the amounts of energy that were dis- is complex and frequently discussed. The goal is to prevent a
pensed. Standard parenteral nutrition solutions contain 60–75% deterioration in nutritional status, which precipitate poorer out-
of energy as dextrose, whereas standard tube feeding solutions comes; however, it must be determined whether the benefit of the
contain 40–55% of energy as dextrose. Metabolic alterations intervention exceeds its risk. Ascertaining the number of days to
that accompany the stress response result in more endogenous wait before the initiation of aggressive nutrition support was not
glucose production and reduce the capacity to oxidize plasma one of our selected objectives; however, several of the PRCTs
glucose directly (42). The higher incidence of hyperglycemia reviewed in this meta-analysis addressed this issue. Sandstrom
caused by excessive glucose loads that are superimposed on et al (30) reported that 60% of their patients began eating on
the stress response in parenteral nutrition may lead to an their own 8–9 d after surgery, and those that had not begun eat-
impaired immune response that contributes to the observed ing by 14 d had significantly higher mortality rates than did
higher risk of infection. those randomly assigned to the parenteral nutrition group that
Overall, a significantly higher risk of nutrition support com- did not have complications or to the standard care group that had
plications was found to be associated with tube feeding than begun eating before 14 d. Sax et al (17) and Brennan et al (28)
with parenteral nutrition, although there was significant hetero- found that an average of 12 d was required for parenteral nutri-
geneity (P = 0.04). When separated into the a priori categories, tion patients to consume adequate energy orally. Several of the
these results remained, and heterogeneity was not significant in trials that compared tube feeding with parenteral nutrition
the relatively recent studies or in those with higher study-qual- reported the number of days required for adequate oral intake.
ity score. Many of the complications (eg, diarrhea and abdomi- McClave et al (16) found that patients who received parenteral
nal distention) associated with tube feeding occur frequently but nutrition required 7.1 ± 1.1 d and tube-fed patients required
are considered to be less severe clinically than are those associ- 5.6 ± 0.8 d to begin oral intake. Adams et al (32) reported that
ated with parenteral nutrition. To address this, the less severe 87% of parenteral nutrition patients began oral intake 10 ± 6 d
complications of tube feeding, such as diarrhea, vomiting, and after injury, and Sako et al (25) found that a mean of 20 d was
ileus, were limited to the instances in which they required med- required for parenteral nutrition patients to resume an adequate
ication or feedings to be stopped for ≥ 24 h. Thus, although 8 of oral intake. Collectively, these results suggest that most normally
the 20 studies that compared tube feeding with parenteral nutri- nourished patients begin to eat 6–10 d after surgery or a hospital
tion reported diarrhea, our definitions restricted its inclusion as admission for a disease that necessitates bowel rest can achieve
a nutrition support complication to just 3 studies, and there were an adequate oral intake within 6-20 d. Our findings suggest that
no instances of vomiting or ileus with use of the modified crite- 7–10 d is a reasonable amount of time to wait before initiating
ria. The technical risks involved with inserting and maintaining parenteral nutrition in normally nourished patients who have not
feeding tubes can be significant and are often underrated. Six of begun eating spontaneously.
the 20 trials that compared tube feeding with parenteral nutri- The limitations of our study were similar to those for any
tion reported problems with feeding tube placement and main- meta-analysis of PRCTs. For example, each author used differ-
tenance and accounted for 14% (9 of 66) of the reported nutri- ent definitions for the outcome variables evaluated. Thus, there
tion support complications. As previously discussed, catheter was a concern that we were comparing “apples with oranges.” A
sepsis, a severe complication associated with parenteral nutri- meta-analysis of small trials may overestimate treatment effects
tion, was categorized as an infection instead of as a nutrition (43), and only 5 of the 27 studies reviewed had > 50 patients
support complication. When catheter sepsis was categorized as per treatment group. Other concerns were the paucity of PRCTs
a nutrition support complication, the differences in the risk of in populations with PEM and the scarcity of PRCTs that
nutrition support complications between tube feeding and par- compared standard care with parenteral nutrition and that
enteral nutrition were eliminated (RR: 1.05; 95% CI: 0.79, 1.4). reported nutritional support complications and other complica-
Our findings of a substantial risk of these various untoward tions. These concerns limited our ability to make generalized
events caused by both tube feeding and parenteral nutrition practice recommendations.
were based on definitions that minimized the inclusion of med- A goal of this meta-analysis was to provide an interim guide
ically insignificant events. They illustrate the need to weigh for clinical decision-making until the results of large trials con-
these risks against the potential benefits before the initiation of ducted in populations with marginal gastrointestinal function
either type of nutrition support. are available. A comprehensive search of the literature was done
The nutritional status of patients influenced the risk of infec- with the use of clinically relevant criteria for both study selec-
tion and mortality in trials that compared standard care with tion and assessed outcomes. In a broad spectrum of patients
parenteral nutrition but not in trials that compared tube feeding with compromised gastrointestinal function with tube feeding,
with parenteral nutrition. When aggregated, standard care was we found fewer infections in those who were tube-fed than in
associated with a higher risk of infection and mortality in the 3 those who received parenteral nutrition. These findings were
trials of populations that had high percentages of PEM; how- similar, although not as strong, in a comparison of standard care
ever, in the 4 trials of normally nourished populations, it was with parenteral nutrition in normally nourished patients. We
 ENTERAL COMPARED WITH PARENTERAL NUTRITION 541

also found a higher risk of nutrition support complications with 18. Gonzalez-Huix F, Fernandez-Banares F, Esteve-Comas M, et al.
tube feeding than with parenteral nutrition. Collectively, these Enteral versus parenteral nutrition as adjunct therapy in acute ulcer-
results suggest that waiting 7–10 d to initiate any form of ative colitis. Am J Gastroenterol 1993;88:227–32.
aggressive nutrition intervention may be prudent for normally 19. Baigrie RJ, Devitt PG, Watkin DS. Enteral versus parenteral nutri-
nourished populations with compromised gastrointestinal func- tion after oesophagogastric surgery: a prospective randomized com-
parison. Aust N Z J Surg 1996;66:668–70.
tion. Studies that compare outcomes of tube feeding with those
20. Bower RH, Talamini MA, Sax HC, Hamilton F, Fischer JE.
of standard care in both normally nourished and malnourished Postoperative enteral vs parenteral nutrition. Arch Surg 1986;121:
populations are needed. Finally, studies comparing the costs 1040–5.
associated with of providing either tube feeding or parenteral 21. Hamaoui E, Lefkowitz R, Olender L, et al. Enteral nutrition in the
nutrition with those of standard care are needed to guide clini- early postoperative period: a new semi-elemental formula versus
cians, hospital administrators, and third-party payers in their total parenteral nutrition. JPEN J Parenter Enteral Nutr 1990;14:
decision-making. 501–7.
22. Iovinelli G, Marsil I, Varrassi G. Nutrition support after total laryn-
gectomy. JPEN J Parenter Enteral Nutr 1993;17:445–8.
REFERENCES 23. Lim STK, Choa RG, Lam KH, Wong J, Ong GB. Total parenteral
nutrition versus gastrostomy in the preoperative preparation of
1. Shils ME, Brown RO. Parenteral nutrition. In: Shils ME, Olson JA,
patients with carcinoma of the oseophagus. Br J Surg 1981;68:
Shike M, Ross AC, eds. Modern nutrition in health and disease. 9th
69–72.
ed. Baltimore: Williams & Wilkins, 1999:1658.
24. Reynolds JV, Kanwar S, Welsh FKS, et al. Does the route of feed-
2. Lipman TO. Grains or veins: is enteral nutrition really better than
ing modify gut barrier function and clinical outcome in patients
parenteral nutrition? A look at the evidence. JPEN J Parenter
after major upper gastrointestinal surgery? JPEN J Parenter Enteral
Enteral Nutr 1998;22:167–82.
Nutr 1997;21:196–201.
3. Heyland DK, MacDonald S, Keefe L, Drover JW. Total parenteral
25. Sako K, Lore JM, Kaufman S, Razack M, Bakamjian V, Reese P.
nutrition in the critically ill patient. JAMA 1998;280:2013–9. Parenteral hyperalimentation in surgical patients with head and neck
4. Naber T, Schermer T, de Bree A, et al. Prevalence of malnutrition in cancer: a randomized study. J Surg Oncol 1981;16:391–402.
nonsurgical hospitalized patients and its association with disease 26. Sand J, Luostarinen M, Matikainen M. Enteral or parenteral feeding
complications. Am J Clin Nutr 1997;66:1232–9. after total gastrectomy: prospective randomised pilot study. Eur J
5. Sullivan DH, Sun S, Walls RC. Protein-energy undernutrition Surg 1997;163:761–6.
among elderly hospitalized patients: a prospective study. JAMA 27. Shirabe K, Matsumata T, Shimada M, et al. A comparison of par-
1999;281:2013–9. enteral hyperalimentation and early enteral feeding regarding systemic
6. Braunschweig C, Gomez S, Sheehan PM. Impact of declines in immunity after major hepatic resection—the results of a randomized
nutritional status on outcomes in long-term adult hospitalized prospective study. Hepatogastroenterology 1997;44:205–9.
patients. J Am Diet Assoc 2000;100:1316–22. 28. Brennan MF, Pisters PW, Posner M, Quesada O, Hike M. A
7. Klein S, Koretz RL. Nutrition support in patients with cancer: what prospective randomized trial of total parenteral nutrition after
do the data really show? Nutr Clin Pract 1994;9:91–100. major pancreatic resection for malignancy. Ann Surg 1994;220:
8. Holter AR, Fischer JE. The effects of perioperative hyperalimenta- 436–44.
tion on complications in patients with carcinoma and weight loss. 29. Muller JM, Brenner U, Dienst C, Pichlmaier H. Preoperative par-
J Surg Res 1977;23:31–4. enteral feeding in patients with gastrointestinal carcinoma. Lancet
9. Thompson BR, Julian TB, Stremple JF. Perioperative total par- 1982;1:68–79.
enteral nutrition in patients with gastrointestinal cancer. J Surg Res 30. Sandstrom R, Drott C, Hyltander A, Arfvidsson B, et al. The effect
1981;30:497–500. of postoperative intravenous feeding (TPN) on outcome following
10. Greenberg GR, Fleming CR, Jeejeebhoy KN, Rosenberg IH, major surgery evaluated in a randomized study. Ann Surg 1993;217:
Sales D, Tremain WJ. Controlled trial of bowel rest and nutri- 185–95.
tional support in the management of Crohns disease. Gut 1988; 31. Woolfson AMJ, Smith JAR. Elective nutritional support after major
29:1309–15. surgery: a prospective randomised trial. Clin Nutr 1989;8:15–21.
11. Dunham CM, Frankenfield D, Beizberg H, Wiles C, Cushing B, Grant Z. 32. Adams S, Dellinger EP, Wertz MJ, Oreskovich MR, Simonowitz D,
Gut failure—predictor of or contributor to mortality in mechanically Johansen K. Enteral versus parenteral nutritional support following
ventilated blunt trauma patients? J Trauma 1994;37:30–4. laparotomy for trauma: a randomized prospective trial. J Trauma
12. Von Meyenfeldt MF, Meijerink WJHJ, Rouflart MMJ, Builmaassen 1986;26:882–91.
MTHJ, Soethers PB. Perioperative nutritional support: a randomised 33. Kudsk KA, Croce M, Fabian TC, et al. Enteral versus parenteral
feeding. Ann Surg 1992;215:503–11.
clinical trial. Clin Nutr 1992;11:180–6.
34. Kudsk KA, Minard G, Wojtysiak SL, Croce M, Fabian T, Brown RO.
13. Mantel N. Chi-square tests with one degree of freedom: extensions
Visceral protein response to enteral versus parenteral nutrition and
of the Mantel-Haenszel procedure. JASA 1963;58:690–700.
sepsis in patients with trauma. Surgery 1994;116:516–23.
14. Mantel N, Brown C, Byar DP. Tests for homogeneity of effect in an
35. Moore FA, Moore EE, Jones TN, McCroskey BL, Peterson VM.
epidemiologic investigation. Am J Epidemiol 1977;106:125–9.
TEN versus TPN following major abdominal trauma—reduced sep-
15. Kalfarentzos R, Kehagias J, Mead N, Kokkinis K, Gogos CA. tic morbidity. J Trauma 1986;29:916–22.
Enteral nutrition is superior to parenteral nutrition in severe acute 36. Cerra FB, McPherson JP, Konstantinides FN, Konstantinides NN,
pancreatitis: results of a randomized prospective trial. Br J Surg Teasley KM. Enteral nutrition does not prevent multiple organ fail-
1997;84:1665–9. ure syndrome after sepsis. Surgery 1988;104:727–33.
16. McClave SA, Greene LM, Snider HL, et al. Comparison of the 37. Levine GN, Derin JJ, Steiger E, Zinno R. Role of oral intake and
safety of early enteral vs. parenteral nutrition in mild acute pancre- maintenance of gut mass and disaccharide activity. Gastroenterol-
atitis. JPEN J Parenter Enteral Nutr 1997;21:14–20. ogy 1974;67:975–82.
17. Sax HC, Warner BW, Talamini MA, et al. Early total parenteral 38. Purandare S, Offenbartl K, Westerom B, Bengmark S. Increased gut
nutrition in acute pancreatitis: lack of beneficial effects. Am J Surg permeability to floresein isothiocyanate-dextran after total par-
1987;153:117–24. enteral nutrition in the rat. Scand J Gastroenterol 1989;24:678–82.
542 BRAUNSCHWEIG ET AL

39. Deitch EA, Winterton J, Ma L, Berg R. The gut is a portal of entry for 42. Shaw JH, Wolfe RR. An integrated analysis of glucose, fat and pro-
bacteremia: role of protein malnutrition. Ann Surg 1987;13:565–71. tein metabolism in severely traumatized patients. Studies in the
40. Lipman TO. Bacterial translocation and enteral nutrition in humans: basal state and the response to total parenteral nutrition. Ann Surg
An outsider looks in. JPEN J Parenter Enteral Nutr 1995;19:156–65. 1989;209:63–72.
41. Golden SH, Peart-Vigilance C, Kao WH, Brancati FL. Perioperative 43. Petitti DB. Limitations. In: Petitti DB, ed. Meta-analysis, decision
glycemic control and the risk of infectious complications in a cohort analysis and cost-effectiveness analysis. 2nd ed. Oxford, United
of adults with diabetes. Diabetes Care 1999;22:1404–8. Kingdom: Oxford University Press, 2000:263–71.