INTRODUCTION
Glutamine (GLN) is the most abundant non-essential amino acid in the plasma, and it is involved in a wide variety of processes in the body. Despite its possibility to be synthesized, GLN has been described as conditionally essential in the catabolic state (1). This amino acid is the preferential substrate of fast-dividing cells such as enterocytes and lymphocytes. In situations of metabolic stress, especially after surgery, GLN tends to be rapidly depleted in muscle cells and transported to the liver in order to optimize gluconeogenesis, as well as to favor immune system cells in the healing phases (2,3).
Clinical studies have demonstrated that GLN plays a fundamental role in the induction of cellular protective pathways, also acting in the modulation of inflammatory response and preventing organic lesions (4). In contrast, other studies raised doubts about the true efficacy and safety of supplementation with this amino acid in critically ill patients (5,6). Recently, some meta-analyses reaffirmed that GLN supplementation seems to reduce mortality and length of hospitalization in critically ill patients (7-10).
Faced with such a diverse scenario and conflicting results, the objective of this systematic literature review was to analyze whether parenteral GLN supplementation, as a part of nutritional therapy, has any relevant effect on clinical outcomes for critically ill surgical patients.
METHODS
This is a systematic literature review (SLR) performed through a retrospective analysis of primary studies that focused on the use of parenteral GLN in critical surgical patients. Articles published from 1985 until May 2019 were surveyed in the MEDLINE database. Search terms included: Glutamine, Randomized, Blind, Clinical Trial, Nutrition, Nutritional Support or Dietary Supplementation or Parenteral Nutrition or Parenteral Nutrition Solutions and Surgical ICU Patients or Surgical Critical Illness. The results obtained were reviewed for identification of those studies that used intravenous or parenteral GLN supplementation. Only articles written in English, Spanish or Portuguese were eligible. Bibliographical references of important literature reviews were also reviewed. As this was a systematic review of the literature, no ethics board approval or patient consent was required.
Only original articles that met the following criteria were included in this SLR: a) Study design: randomized clinical trials; b) Population: adult patients undergoing surgical procedures admitted to the intensive care unit (ICU); c) Intervention: parenteral use of GLN versus a control group using an isonitrogenated amino acid solution or placebo; d) Outcomes: having included one of the following outcomes: mortality, length of hospitalization, infection-related complications or other significant clinical condition. Assays that used only GLN via enteral or combined enteral/parenteral routes were excluded.
The analysis of the studies was performed by an independently reviewer. The first outcome of this SLR was general mortality. As secondary outcomes, we included the occurrence of infection and length of hospital stay (LOHS).
The Revman 5.3 software was used to perform the meta-analysis. The data extracted from the studies were combined to estimate the relative risk, and a 95 % confidence interval (CI) was stipulated for the categorical variables (mortality and occurrence of infection). For continuous variables (LOHS), the weighted mean difference and 95 % CI were estimated as results of the effect. Heterogeneity was calculated using the χ2 test. A random effects model was used to estimate the overall effect. To justify a possible heterogeneity statistic, it was taken into consideration that the form or dose of the dipeptide provided might influence treatment outcome.
RESULTS
One hundred and twenty-three articles were found using the above-mentioned search terms. After initial screening through titles and abstracts 24 papers were selected. Among these, and after analyzing important reviews, seven studies were included (11-17). The articles that were excluded in this SLR did not have the necessary outcomes delimited or did not include accurate information on patient stay in the ICU after the surgical procedure.
The samples of the studies ranged from 30 to 150 patients. The GLN used in these investigations included alanyl-glutamine (ALA-GLN) (six studies) (11-15,17). Two of them also used glycyl-glutamine (GLY-GLN) for a second experimental group (12,13). Only one study did not report the type of dipeptide used (16). For the studies that included three groups (two experimental arms and a control arm) we chose to pool together the events concerning the categorical variables of both GLN groups (ALA-GLN and GLY-GLN) in one single intervention group to avoid event duplication in the control group.
The details for each of the included studies are described in table I.
ALA-GLN: alanyl-glutamine; Con: control group; GLN: glutamine group; GLY-GLN: glycyl-glutamine; LOHS: length of hospital stay; Pre-op: preoperative; Post-op: postoperative; SD: standard deviation.
*These values are presented as median (variation).
Five studies (11,14-17) described the mortality rate of participants in each group. The combination of the results of these studies shows that glutamine supplementation in parenteral nutrition seems to have no effect on mortality reduction (RR = 0.76; 95 % CI: 0.50-1.15; p = 0.19) (Fig. 1). This is evidenced by the partial deviation of the diamond in the chart. The heterogeneity test was not significant (χ2 = 2.78; p = 0.60).
Regarding the occurrence of infection five articles described the rates of this complication (Fig. 2) (12,14-17). There was no heterogeneity among the five articles (χ2 = 5.69; p = 0.22). The combined analysis of the data showed that there seems to be a reduction in the rates of infectious complications in the patients who received parenteral glutamine supplementation (RR = 0.60; 95 % CI: 0.45-0.80; p = 0.0005), as shown in the left deviation of the diamond in the graph (in favor of glutamine use).
All seven studies reported LOHS (11-17). One of them (17) recorded LOHS as a median value, which could not be combined with the results of the other studies, and was therefore excluded from the analysis (Fig. 3).
In the studies where this information was stratified between the two experimental groups, each intervention group was independently associated with the single control group. The heterogeneity test was significant (χ2 = 53.99; p < 0.00001). The use of parenteral glutamine seems to be associated with a shorter length of hospital stay (RR = -4.09; 95 % CI: -6.71 to -1.46; p = 0.002), as shown in the left deviation of the graph diamond (in favor of glutamine use).
DISCUSSION
In the last 20 years several studies were conducted and some of them reported that there seemed to be a reduction in length of hospital stay for patients who used GLN supplementation, in addition to reduced mortality and fewer infectious complications. Two large randomized, multicenter clinical trials (REDOX and METAPLUS) showed that patients in the early phase of sepsis, using vasopressor drugs, on mechanical ventilation, or with renal insufficiency should not receive GLN because of a potential increase in mortality (5,6).
This study demonstrated that the use of parenteral glutamine does not seem to have the ability to significantly decrease mortality rate in patients having undergone surgery who required intensive care. These data are consistent with other findings in the literature about general surgical patients and critically ill patients regardless of disease type. Wang and collaborators demonstrated in their meta-analysis that general surgical patients did not have a reduction in mortality with the parenteral usage of this dipeptide (18). The same was concluded by Avenell and collaborators, who observed only a tendency towards mortality reduction with the use of parenteral glutamine in critically ill patients (19). Bollhalder and their group of researchers also confirmed in their meta-analysis this trend towards reduced mortality only in patients admitted to the ICU, with no such tendency being observed among non-ICU patients (7).
Coëffier and collaborators reported in their review that GLN has a protective effect on cells, stimulating the production of heat shock proteins (HSPs), which would protect cells against toxic and pathological agents, thus minimizing their deleterious effects (1). These proteins would serve as inflammatory messengers in the plasma. Recently, Wischmeyer and collaborators demonstrated that the level of these proteins in the serum could be used as a predictor of mortality in severely ill patients. They also reported that GLN supplemented in parenteral nutrition was not able to reduce serum HSP levels, reinforcing the low impact of this nutrient on reducing mortality for critically ill patients (20).
Our analyses also demonstrated that the use of parenteral glutamine seems to have a protective effect against infections during hospitalization among critical surgical patients, due to a lower incidence (40 % less) of events with this supplementation. This information seems to be in accordance with what was reported in the meta-analysis by Wang et al. and in the article by Zheng et al. about the use of parenteral GLN in general surgical patients (10,18). Wischmeyer et al. and Bollhalder et al. also demonstrated this tendency in their meta-analyses about the use of this dipeptide in critically ill patients (7,21).
In the surgical patient the metabolic response to uncontrolled trauma seems to have a deleterious effect on wound healing and organic defenses. This response involves cardiac hyperdynamics, pulmonary repercussion, insulin resistance, hyperglycemia, muscle protein catabolism, and use of lipid reserves for accelerated gluconeogenesis, ultimately increasing oxidative stress and inflammatory mediators. This cascade of events tends to culminate in organ failure and immunosuppression, which falicitates infection (22). The use of GLN has been shown to be effective in reducing hyperglycemia and use of insulin in critically ill patients, but the mechanism of action through which this dipeptide exerts this regulation remains unknown (1). In addition, GLN proved to be effective in reducing the expression of interleukin 8 (IL-8) and C-reactive protein (CRP), thus influencing inflammatory response. Additionally, the parenteral use of GLN has been shown to induce a reduction in IL-6, improving immunodepression (23), and a strengthening of the gastrointestinal mucosal barrier, in this way preventing bacterial translocation (24).
In this meta-analysis we have also shown that the use of parenteral GLN seems to have decreased hospital stay by approximately four days among critical surgical patients. This is consistent with the data reported in the meta-analyses that evaluated the use of parenteral GLN supplementation in general surgical or general critical patients (7,10,18,21). This information is relevant since a reduction in hospital stay has a direct impact on decreasing costs. Despite this fact, the way GLN may reduce LOHS remains unknown. Some authors believe that nitrogen balance cannot in and by itself decrease hospital stay. It seems that the metabolic response to GLN induces an improvement in organ functioning, which would then minimize infection, and optimize patient mood, as well as other parameters influencing the decision to discharge (25).
CONCLUSION
Based on the analyses of this study, we conclude that the parenteral use of GLN in critical surgical patients seems to reduce infections and length of hospital stay. Supplementation with this dipeptide, however, did not result in a significant reduction in mortality among these patients. Although these results seem to be promising, it is still difficult to generalize the use of this nutrient for this population due to the heterogeneity found among clinical trials in the literature (different doses used and lack of surgery type stratification).