点击显示 收起
1 From the Beth Israel Deaconess Medical Center and Harvard Medical School, Boston.
See corresponding articles on page160 and 265.
Two articles on total parenteral nutrition (TPN) in this issue of the Journal (1, 2), one an original research communication and the other a review, can be used as a springboard for discussing several important points about TPN in general and about research in this area specifically. First, TPN, a major therapeutic advance in the modern medical era, still provides ample opportunity for continued exploration. Zauner et al (1) used indirect calorimetry to evaluate energy and substrate metabolism in septic and nonseptic patients before and after the administration of a standard TPN formula. Their conclusion was that no major differences in metabolism exist between these patient groups, and thus that disease-specific TPN formulas are not required. Recognizing that the study was well conducted and provided new and interesting information and that the statistical evaluations were appropriate, a closer look at the data provides insight into how such results could subsequently be used in support of clinical practices that may be erroneous. Such is the case, I believe, that led to the still widely held belief that TPN is inferior to enteral nutrition, a belief that is clearly debunked in the review by Jeejeebhoy (2).
Which characteristics of the study by Zauner et al might make the data difficult to generalize to all septic and nonseptic patients receiving TPN? First, these were patients in a medical intensive care unit with mostly medical conditions; yet, most patients receiving TPN have surgical illnesses. The resting energy expenditures of these elderly medical patients were 98.3 and 92.5 kJ/kg body wt (23.5 and 22.1 kcal/kg body wt) in the septic and nonseptic groups compared with 176 and 151 kJ/kg body wt (42 and 36 kcal/kg body wt) in a group of septic and nonseptic trauma patients aged 30 y (3). These differences in age and diagnosis had a major metabolic effect as well. For example, mean blood glucose concentrations in the study of medical patients were 8.77 and 7.77 mmol/L (158 and 140 mg/dL) in septic and nonseptic patients, respectively (1), compared with 10.8 and 9.10 mmol/L (194 and 164 mg/dL), respectively, in a study of surgical patients (3), which might be expected to affect the risk of nosocomial infection through the development of significant hyperglycemia (4). A second characteristic that limits the generalizability of the study by Zauner et al is the choice of TPN formula. The fixed formula used in the study by Zauner et al contained 45% glucose, 41% fat, and 14% amino acids, a much higher fat content than conventionally used in the United States and a protein intake of 1 g/kg body wt, which would be considered marginal (5).
The final 2 concerns that limit the extrapolation of these results are related to study design. Although the study was prospectively conducted, there is no reason to suppose that septic patients were different from those without sepsis by that characteristic alone. For example, 4 patients in one group had severe liver disease and 4 patients in the other group had severe respiratory disease, with no patients with either diagnosis in the comparison group. Second, lactate concentrations were significantly higher in the septic group than in the nonseptic group and serum urea and creatinine were significantly reduced in the nonseptic group only. This suggests a strong possibility of a type II error in detecting differences in metabolic responses between the 2 groups. For these reasons, the conclusions of this study should not be more broadly applied to surgical patients receiving TPN without further study.
After more than a decade of general acceptance of the dictum that TPN is more likely to lead to complications than is enteral nutrition, a careful review of the evidence by Lipman (6) and now Jeejeebhoy (2) and a recent pragmatic comparative study (7) suggest that the pendulum is beginning to swing to a middle ground. It appears that there is no substantial difference in outcome between the 2 approaches, and somewhat more difficulty in achieving full feeding rates and slightly more complications with enteral feeding. Two factors may account for this reversal of opinion. As Jeejeebhoy states, in the early years of TPN, energy intakes far in excess of energy expenditure were the norm. For instance, in the Veterans Affairs trial of TPN efficacy in a largely malnourished population, the estimated total energy intake in the TPN group was 192 kJ/kg body wt (46 kcal/kg body wt), whereas the ad libitum group consumed >84 kJ/kg body wt (>20 kcal/kg body wt) (8), which approaches the total energy needs of such patients (9). Overfeeding can be harmful to critically ill patients in part through hyperglycemia (4), and overfeeding is easy to accomplish with TPN and difficult to accomplish with enteral nutrition. Energy intakes >125–146 kJ/kg body wt (>30–35 kcal/kg body wt) lead to high rates of hyperglycemia (10), and most trials of TPN in critically ill patients in which positive results were reported used intakes below this approximate amount, in most cases 125 kJ/kg body wt (30 kcal/kg body wt) (11–17). A more general appreciation of this phenomenon has led to more modest provision of TPN energy and, with this, fewer complications.
The results of 3 trials are largely responsible for the still commonly held conviction that TPN predisposes to infectious complications (18–20). In each of these studies of trauma patients, the TPN group received significantly more energy: 141 kJ/kg body wt (33.6 kcal/kg body wt) on day 5 in one study (18) and 157 and 146 kJ/kg body wt (37.7 and 35 kcal/kg body wt) at maximal rates in the others (19, 20). In a fourth study of similar patients, equivalent but high intakes were provided to both patient groups by increasing enteral [134 kJ (32 kcal)/kg body wt] and parenteral [146 kJ (35 kcal)/kg body wt] intakes, and infection rates were similar but very high (65%) in both groups (21). Other trials of TPN and enteral nutrition in nontrauma patients who received lower intakes reviewed by Lipman (6) and Jeejeebhoy (2) did not report a difference in outcome. It is interesting to note that in the study by Zauner et al (1), <126 kJ/kg body wt (<30 kcal/kg body wt) was provided to both groups (1), presumably reflecting the now general trend in TPN administration.
The article by Zauner et al and the review by Jeejeebhoy provide both new information and an object lesson that careful review of study design and interpretation is necessary to fully appreciate advances in knowledge. Furthermore, it is always helpful to have a superb thinker like Jeejeebhoy as a guide.
REFERENCES