A comparison of complications associated with nutrition between the patients receiving enteral or parenteral in the intensive care unit

Nutrition support is one of the most important parameters in the Intensive Care Unit (ICU). Nutrition support is known to affect mortality and morbidity; help wound healing and have positive effects on the immune system. While it serves only as a supportive treatment for elective patients, it can increase the survival rate of critically ill patients, when applied correctly and at right time. Oral feeding is the normal route of nutrition; however, it is often not possible in ICU patients. Enteral and parenteral nutrition were developed for such cases [1,2].


Introduction
Nutrition support is one of the most important parameters in the Intensive Care Unit (ICU). Nutrition support is known to affect mortality and morbidity; help wound healing and have positive effects on the immune system. While it serves only as a supportive treatment for elective patients, it can increase the survival rate of critically ill patients, when applied correctly and at right time. Oral feeding is the normal route of nutrition; however, it is often not possible in ICU patients. Enteral and parenteral nutrition were developed for such cases [1,2].
In recent years, researchers have been emphasizing the importance of early enteral nutrition for intensive care unit patients. Enteral nutrition (EN) is usually recommended and preferred as it protects the structure and functions of the gastrointestinal tract, is the physiological route, has fewer complications and lower costs [3][4][5][6].
The role of nutrition support in patient care has become an important factor in medicine as the studies conducted over the last 50 years revealed its bene its, such as prevention of many complications. There are a number of studies comparing enteral (EN) and parental nutrition (PN) and they showed that parenteral nutrition caused more complications in intensive care patients compared to enteral nutrition [7].
In our study, we aimed to reveal the complications in patients who were only fed enterally or parenterally, and to compare the length of stay in hospital and ICU, length of stay on mechanical ventilation, the incidence rate of infection and mortality and morbidity in patients who only received enteral or parenteral nutrition.
The comparison of the reason for admission between the EN and PN groups showed that surgical reasons were signifi cantly higher in the PN group. Nosocomial infections, the presence of infection and the development of sepsis were signifi cantly higher in the EN group. The 28-day mortality rate was higher in the PN group compared to the EN group. The length of stay in the ICU and on mechanical ventilation was longer in the EN group. There was no signifi cant diff erence in the 28-day mortality, readmission to the ICU and repeated endotracheal intubation between the two groups.
Because there is no statistical diff erence between EN and PN groups in point of infection and mortality, we conclude that the length of stay in the ICU and reason for admission play a more crucial role in the development of infection and on mortality rather than enteral or parenteral nutrition route.

Materials and methods
After the approval of the local ethics committee of a university was obtained, the patients who were under follow-up and treatment in the Anesthesiology Intensive Care Unit for two years were retrospectively analyzed. Authors' institutional review board (IRB) has approved the study.
In this study, a total of 100 patients who were under follow-up or treatment in Anesthesiology Intensive Care Unit (ICU) of a university hospital for two years were examined. In our study, demographic characteristics, the reason for admission, comorbidity, initial ICU laboratory values, morbidity during the follow-up period of the patients who received enteral nutrition (EN) or parenteral nutrition (PN) were evaluated as well as examining the complications which are related to the groups.
The exclusion criteria were as follows: being under 18 years, receiving both enteral and parenteral nutrition simultaneously, oral feeding, pregnancy, and suffering from burn, malignancy, chronic renal failure (CRF), severe hepatic insuf iciency, intoxication, diabetic ketoacidosis, the patient with cardiopulmonary resuscitation, Glasgow coma score (GCS) < 4 and the patients who stayed less than three days in hospital.
EN group included patients who only had enteral nutrition during the whole period in the intensive care unit, while PN Group included patients who only had parenteral nutrition because of surgical causes. A total of 100 patients who met inclusion criteria were included in the study. EN group included 55 patients and PN group included 45 patients in CONSORT Diagram (Diagram 1).
The complications associated with enteral nutrition (EN) were determined as mechanical, metabolic and gastrointestinal complications.
The complications associated with parenteral nutrition (PN) were determined as catheter-related, metabolic and gastrointestinal complications. The bilirubin levels in the sequential organ failure assessment (SOFA) score were used to evaluate the hepatic disfunction for the gastrointestinal complications. Metabolic complications included hyperglycemia (blood glucose > 180 mg/dL), hypoglycemia (blood glucose < 70 mg/dL), hypernatremia (serum sodium > 145 mEq/L) and hypophosphatemia (serum phosphate < 2.5 mg/dL).

Statistical analysis
Statistics package program was used for statistical analyses of the data obtained in the study. Pearson's Chisquare test and Fisher's exact test were used to compare the categorical data. The t-test was used for the comparison of the groups.
The results were evaluated in a 95% con idence interval and p < 0.05 was accepted as statistically signi icant.

Results
There was no statistically signi icant difference in gender (p = 0.270 > 0.05), BMI (p = 0.488 > 0.05), age and mean GCS (p > 0.05) between the groups. There was no statistically signi icant difference in initial APACHE II score, the levels of creatinine, platelets, bilirubin, PaO 2 /FiO 2 and blood glucose levels at admission (p > 0,05) ( Table 1). There was a statistical difference in reason for admission between the EN and PN groups (p = 0.001). Of the patients in EN group, 35 (63.6%) were admitted to hospital for surgical and 20 (36.4%) patients were admitted for medical causes while of the patients in the PN group, 41 (91.1%) were admitted to hospital for surgical and 4 (8.9%) patients were admitted for medical causes. Surgical admission rate was higher in the PN group while medical admission rate was higher in the EN group.
The rate of admission due to cardiac diseases (p = 0.013) and neurological (p = 0.010) diseases were higher in the EN group.
Postoperative admission was statistically higher in the PN group (p < 0.001).
There was no signi icant difference in comorbid diseases between the groups. However, if the comorbidities were to be categorized into subgroups; no signi icant difference was found in HT, CAD, CHF, DM, renal, COPD, neurological diseases between the EN and PN groups similarly (p > 0.05). However, GIS (p = 0.014) and hematologic (p = 0.038) diseases were signi icantly higher in the PN group than the EN group.
There was no signi icant difference between the groups in terms of anemia, ES requirement, the occurrence of DIC, renal replacement therapy, hyperbilirubinemia, vasopressor requirement, steroid therapy, septic shock, septic shock recovery, 28-day mortality, readmission to the ICU and repeated ET (p > 0.05).
The length of stay on MV was signi icantly longer in the EN group (17.91 days) than the PN group (5.36 days) (t = 3.703; p < 0.001), (Figure 4).
The comparison of duration of septic shock revealed no signi icant differences between the groups (p > 0.05). Presence of infec p=0.013   There was no signi icant difference in hypoglycemia, hyperglycemia, electrolyte imbalance (hypernatremia, hypophosphatemia) between two groups (p > 0.05).
The difference between the mean APACHE II scores (18.23 ± 5.46) of the patients with 28-day mortality (n = 26) and the mean APACHE II scores (14.83 ± 5.79) of the total of 100 patients was signi icantly higher in patients with 28-day mortality (p = 0.008).
The comparison of 28-day mortality revealed no statistically signi icant difference between the groups; however, they were slightly higher in the PN group versus EN group (28.9% vs. 23.6%, p = 0.356) ( Table 2).

Discussion
Enteral nutrition (EN) should be preferred for patients with functional GI disorders [8]. Parenteral nutrition (PN) is an important method for nourishing patients who suffer gastrointestinal tract disorders of diseases or who are unable to be fed by enteral route. It is proved that the proper use of PN has positive effects on the clinical course; however, improper use may lead to increase in infectious complications, development of metabolic anomalies and increased medical costs [9]. In Liu, et al. study [10] on 60 patients with pancreatic cancer, the postoperative group only received enteral nutrition, while the other group received parenteral nutrition. Age, gender and BMI from demographic characteristics revealed no signi icant difference. Our study showed no signi icant difference in demographic characteristics.
In their randomized study conducted on 2388 patients, Harvey, et al. [11] compared the correlation of early nutrition with the EN and PN groups and reported that there was no signi icant difference in the APACHE II scores of the patients at admission. In our study, there was also no signi icant difference between the APACHE II scores at admission of 100 patients.
Grau, et al. [12] conducted a retrospective cohort study on 725 patients with hepatic dysfunction; they compared two groups who receive EN and PN. It was detected that the reason for admission in the EN group was mostly medical treatment, while surgical admissions were higher in the PN group. Similarly, the comparison of the reason for ICU admission revealed that surgical admissions were signi icantly higher in the PN group than the EN group in our study.
A retrospective study carried out by Tao, et al. [13] revealed that there was no signi icant difference between early EN and PN in terms of comorbid diseases (HT, CAD, DM, steatosis hepatitis, COPD, etc.). In our study, only comorbid diseases with higher incidence rates were GI and haematological diseases in the PN group and similarly there are no difference between the groups for HT, CAD, DM, COPD.
In the study was conducted by Harvey, et al. [11], they reported no statistical difference in infectious complications between the EN and PN groups. A randomized, multicenter NUTRIREA-2 study was carried out by Reignier, et al. [14] showed that there was no signi icant difference in the comparison of ICU acquired infections, ventilator-associated pneumonia, bacteremia and urinary infections between the EN and the PN groups. The study of Blaser, et al. [15] revealed that the risk of infection was signi icantly lower in the EN group between the ICU patients who received early EN or PN. In our study, nosocomial infections (65.5%), the presence of infection (74.5%) and sepsis (61.8%) were signi icantly higher in the EN group. In the study on 185 patients, 89 of whom received enteral nutrition and 96 received parenteral nutrition, Tao, et al. [13] revealed signi icantly higher mortality rates among the PN group. Yan, et al. [16] noted no signi icant difference in mortality rate between the EN and PN groups. In the NUTRIREA-2 study carried out by Reignier, et al. [14], there was no signi icant difference in 28-day mortality between two groups. Our study revealed that 28-day mortality was higher in the PN group (28.9%) compared to the EN group (23.6%); however, no statistically signi icant difference was noted between the two groups.
Tao, et al. [13] and Yan, et al. [16] found that the length of hospital stay was statistically shorter in patients who received enteral nutrition. Grau, et al. [12] stated that the length of stay in ICU and on MV was statistically higher in the group who received enteral nutrition. In our study, the length of stay in ICU and the length of stay on MV were higher in the EN group, and this supports the study conducted by Grau, et al. [12]. We consider that this situation resulted from the fact that the patients receiving enteral nutrition were followed-up mostly due to medical reasons rather than surgical reasons, and this lead in a more extended stay in the ICU and on the MV. We think that the fact that the length of stay in the ICU and on the MV is shorter among parenterally fed patients is mostly due to the frequent surgical follow-ups, which is associated with lower incidence rates of nosocomial infections, presence of infection and sepsis development. The studies demonstrated that enteral nutrition may lead to metabolic, gastrointestinal and mechanical complications. The possible mechanical complications include malposition, tube occlusion and pulmonary aspiration leading to death [18]. In our study, 21.8% of 55 patients in the EN group developed mechanical complications, 9.1% were due to aspiration pneumonia and 18.2% were due to tube occlusion.
The study carried out by Harvey, et al. [11] demonstrated that the rate of hypoglycemia was signi icantly lower in the group that received parenteral nutrition compared to the group that received enteral nutrition. In the study carried out by Tao, et al. [13], hyperglycemia was observed in 32 from 96 patients who received parenteral nutrition; while it was observed in 18 from 80 patients who received enteral nutrition. However, there was found no statistically signi icant difference. In a randomized study carried out by Moore Fa, et al. [19], it was determined that blood glucose levels were higher in the patient group who received parenteral nutrition compared to patients who received enteral nutrition. In our study, hyperglycemia was observed in 74.5% of the patients who receive EN and hypoglycemia was noted in 32.7% of the patients. Hyperglycemia was observed in 66.7% of the patients receiving PN, and hypoglycemia was detected in 31.1%. There was no signi icant difference in terms of hyperglycemia and hypoglycemia between the two groups. There was no signi icant difference in hypernatremia and hypophosphatemia between two groups in our study. Similar to our study, Harvey, et al. [11] reported no signi icant difference in electrolyte disturbances.
Hepatic dysfunction is a gastrointestinal complication, such as PN-associated steatosis, cholestasis and gallstones and it has a higher prevalence in patients who receive long-term PN. At the same time, it is known that parenteral nutrition may lead to changes in serum transaminase, ALP and bilirubin [18]. The study of Grau, et al [12] noted that 30% of the patients who receive PN developed hepatic dysfunction. In parallel to this study, 28.9% of the patients who receive PN developed hepatic dysfunction in our study.
The study carried out by Park, et al. [20] including early enteral and parenteral feeding in patients with pancreaticoduodenectomy noted that 5.5% of the patients who receive enteral nutrition developed vomiting, 5.5% showed diarrhoea and 5.5% had nausea. In the study was carried out by Reignier, et al. [14], 34% of the patients who receive enteral nutrition showed nausea and 36% showed diarrhea. In our study, 49.1% of the patients who receive enteral nutrition had nausea, 32.7% developed vomiting, 36.4% developed diarrhoea and 67.3% developed constipation. We presume that the high rates are caused by the longer length of hospital stay in the EN group.
In their study that they evaluated the complications and contributing factors in the patients who receive enteral pump feeding in ICU, Metin, et al. [21] noted that 72.5% of the patients were fed by nasogastric tube and 27.7% were fed by PEG tube. In our study, 87.3% of the patients were fed by nasogastric tubes and 12.7% were fed by percutaneous tubes (PEG/PEJ).
The study of Harvey, et al. [11] revealed that there was no statistically signi icant difference between the APACHE II scores and 30-day mortality in the EN and PN groups. Blaser, et al. [15] noted no signi icant difference in mortality in the comparison between the patients receiving early EN and early PN in the ESICM clinical practice guidelines regarding early enteral nutrition in ICU patients. In the study carried out by Tao, et al. [13], there was no signi icant difference in the APACHE II scores between the EN and PN groups. Although the mean APACHE II score and 28-day mortality were higher in the PN group, no statistically signi icant difference was found between the two groups in our study.
The retrospective structure of our study was limited in some aspects; however, we tried to compare the complications associated with enteral and parenteral nutrition in intensive care patients with similar demographic characteristics. In our study, there was no statistically signi icant difference between the two groups although the PN group had a higher mortality rate. Nosocomial infection rates, the incidence rate of infection and sepsis development were higher in the patients who received enteral nutrition. We consider that the high rates are associated with the fact that the patients receiving enteral nutrition have medical conditions, which result in an extended stay in the ICU and on the MV. On the other hand, the length of stay in ICU and the length of stay on MV are relatively shorter in patients who receive parenteral nutrition as the reason for admission is mostly surgery. Therefore, nosocomial infections, the presence of infection and sepsis development are lower in the PN group.

Limitations
There are some limitations of this study. First, this is a retrospective study not a prospective. Second, the sample sizes of our groups are a bit small. We need more prospective randomized studies on multicenter design in large groups.

Conclusion
In conclusion, we presume that the length of stay in ICU and the reason for admission play a more crucial role in the development of infection and mortality rather than enteral or parenteral nutrition. Further studies with a prospective randomized clinical trial design are required for clarifying on this topic.