The Impact of Rate of Feeding Advancement after Early Initiation of Enteral Nutrition in Critically Ill , Underweight Patients : A Single-Center Retrospective Chart Review

Background: The optimal rate of feeding advancement after initiation of early enteral nutrition (EEN) for underweight, critically ill patients is unknown. Methods: We conducted a retrospective chart review in intensive care unit (ICU) patients with a body mass index (BMI) < 20.0 kg/m. Patients were categorized into Group R, which reached the energy target within 3 days of EEN initiation, and Group S, which reached the energy target 4 or more days after EEN initiation. Results: A total of 65 patients with a median age of 73 years were included in the study. No significant differences were observed between the two groups for all-cause mortality, ICU-free days, or length of hospital stay. Ventilator-free days (VFDs) were significantly fewer in Group R than in Group S (18.0 [0.0 22.0] vs. 21.0 [16.3 24.8] days; P = 0.046). A significantly higher number of patients requiring mechanical ventilation (MV) at hospital discharge were observed in Group R than in Group S (29% vs. 8%; P = 0.030). Multivariable analyses with adjustment for confounders found that days required to reach target energy intake after EEN initiation were significantly and independently associated with the requirement for MV at hospital discharge, but not with VFDs. Conclusion: A slow rate of feeding advancement after initiation of EEN in critically ill patients having a BMI of <20.0 kg/m might be associated with a reduced requirement for MV at hospital discharge. These results require confirmation in a large multicenter trial of underweight, critically ill patients. How to cite this paper: Ichimaru, S., Sono, M., Fujiwara, H., Seo, R. and Ariyoshi, K. (2016) The Impact of Rate of Feeding Advancement after Early Initiation of Enteral Nutrition in Critically Ill, Underweight Patients: A Single-Center Retrospective Chart Review. Food and Nutrition Sciences, 7, 939-954. http://dx.doi.org/10.4236/fns.2016.711093 Received: April 20, 2016 Accepted: September 9, 2016 Published: September 12, 2016 Copyright © 2016 by authors and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY 4.0). http://creativecommons.org/licenses/by/4.0/ Open Access


Introduction
Nutritional support is essential for the management of intensive care unit (ICU) patients.Enteral nutrition (EN) is the preferred route for nutritional support in critically ill patients [1].Current clinical guidelines recommend that EN be initiated within 24 to 48 h of admission [1]- [3].This concept of early EN (EEN) has become widely accepted due to reduction in the rate of infectious complications and mortality [1]- [3].
Following the initiation of EEN, clinical guidelines recommend that feeding be advanced toward a goal rate over the next 48 -72 h [1].However, several studies recommend the slow advancement of energy intake, noting improved outcomes in critically ill patients with initial enteral underfeeding [4]- [8].Importantly, however, many of these studies were conducted in Western countries in patients with normal weight or excessive body weight (BW).For underweight patients with limited nutritional reserves, advancing feeding too slowly might result in undernutrition or poor clinical outcomes [9]- [14].We previously reported that reduced energy intake during the first week of ICU admission was associated with a reduced duration of mechanical ventilation (MV) in critically ill, underweight patients [15].However, that study included patients who did not receive EEN or in whom oral intake or parenteral nutrition (PN) was initiated within the first week of admission.Therefore, the association observed in our previous report might be influenced by patients who could progress to oral intake early, or who were intentionally underfed thorough PN [16].To our knowledge, the optimal rate of enteral feeding advancement after initiation of EEN in underweight and critically ill remains unclear.
Here, to examine the impact of feeding advancement rate after the initiation of EEN on the mortality and morbidity of underweight and critically ill patients, we performed a retrospective chart review of patients with a body mass index (BMI) < 20.0 kg/m 2 .screened for eligibility.If the patients in our previous study met the inclusion criteria below, they were also included in the present study, as the two study periods overlapped (Aug 2011-Dec 2012) [15].Critically ill adult patients with a BMI of <20.0 kg/m 2 who were mechanically ventilated and initiated EN within the first 48 h of admission and who remained in the EICU for 72 h or more were include in the study.The decision of when to initiate and advance EN was made following discussion among the attending and EICU physicians based on their preference and/or the patient's hemodynamic stability, clinical condition, and underlying diseases.Exclusion criteria were as follows: aged <18 years, pregnant or lactating, readmission to the EICU within the same hospitalization period, under mask ventilation, withdrawal from treatment in the first week, BMI ≥ 20.0 kg/m 2 , missing BMI data, or initiation of oral or parenteral nutrition within the first week of admission to the EICU.

Data Collection
Data were obtained from the medical records of patients.The following baseline demographic characteristics were recorded: age, sex, body weight and height on admission, and BMI.BMI was calculated based on weight (kg)/height (m 2 ) on admission.The following baseline clinical characteristics were recorded: Acute Physiology and Chronic Health Evaluation (APACHE) II Score (which ranges from 0 to 71, with higher scores indicating more severe illness) [17] and Sequential Organ Failure Assessment (SOFA) Score (which ranges from 0 to 24, with higher scores indicating more severe organ compromise) [18] in the first 24 h of admission; surgical or medical admission; admission from emergency department, emergency ward, operating room, or other hospital; primary diagnosis in EICU for cardiovascular, vascular, respiratory, gastrointestinal, neurological disorders, trauma, sepsis, burn, metabolic disorder, or renal disorder; Charlson Comorbidity Index [19]; and comorbidities including diabetes mellitus, dialysis-dependent renal failure, cardiac disease, anemia (hemoglobin level < 8.0 g/dL) [20], and chronic obstructive pulmonary disease.The following nutritional characteristics were recorded: time to EEN initiation from EICU admission, days to reach target energy intake from EEN initiation, and mean daily energy and protein intake during the first week of EICU admission.Total energy intake included intake via parenteral and enteral routes; glucose infusion used for fluid support, drug dilution and correction of hypoglycemia; and lipids delivered together with sedatives (e.g., propofol).
An energy target of 25 kcal/kg/day was set as the required energy for the majority of critically ill patients [1] [3].In the present study, patients were categorized into a rapid advancement group (Group R), which reached the energy target within 3 days of EEN initiation, or a slow advancement group (Group S), which reached the energy target 4 or more days after EEN initiation.A cut-off value of 3 days was used as clinical guidelines recommend that feeding be advanced toward a goal rate over the next 48 -72 h following the initiation of EEN [1].

Clinical Outcomes
Clinical outcomes in Groups R and S were compared.The following variables were analyzed: 1) All-cause mortality in the hospital and EICU.
2) ICU-free days and length of hospital stay.ICU-free days were defined as the number of days between successful transfer to a general ward and Day 28 after EICU admission.ICU-free days were 0 if the patient died before Day 28 or stayed in the EICU or other ICUs for ≥28 days.For the length of hospital stay, patients who died during hospitalization were excluded from the analysis.
3) Ventilator-free days (VFDs).VFDs were defined as the number of days between successful weaning from MV and Day 28 after EICU admission.VFDs were 0 if the patient died before Day 28 or required MV for ≥28 days.6) Antibiotic-free days as the number of days during the 28 days after living patients had been admitted to the EICU and had not received any antibiotics as treatment for infections or as prophylaxis.
7) Inflammatory index, as maximum plasma C-reactive protein (CRP) level during the first 28 days after EICU admission.

8) Number of patients clinically diagnosed with pneumonia by EICU physicians
during the first 28 days after EICU admission, and number of patients who developed ventilator-associated pneumonia (VAP) [21].9) Number of patients exhibiting liver dysfunction, defined as a total bilirubin level of >1.2 mg/dL during the first 28 days after EICU admission.10) Number of patients exhibiting renal dysfunction, as determined by an increase in serum creatinine level of >1.2 mg/dL or need for renal replacement therapy during the first 28 days after EICU admission.11) Number of patients exhibiting hypoglycemia, defined as a blood glucose level of <70 mg/dL [22], during the first week in the EICU and for 28 days after EICU admission.
12) Number of patients receiving insulin therapy and average daily insulin dose in the first week and 28 days after EICU admission.Target glucose level was set as 100 to 180 mg/dL.13) Number of patients exhibiting hypophosphatemia, defined as a serum phosphorus (P) concentration < 2.7 mg/d [23], during the first week and 28 days after EICU admission.However, serum P concentration was not examined in all patients, because the frequency of laboratory testing was not standardized between physicians.Further, lowest serum P concentrations and the dose of P via parenteral or enteral administration to correct for depletion were compared between groups.P content of enteral formulas given as maintenance dosing was excluded from this analysis.P dose was calcu-lated with consideration to not only serum P level, but also to the degree of total body depletion and renal function.

Statistical Analysis
Data for continuous variables are presented as median and inter quartile range and for categorical variables as numbers and percentages.Groups were compared using the Mann-Whitney U test (continuous data) and Fisher's exact test (categorical data).We used the nonparametric test for continuous variables as the relatively small sample size underestimated the distribution assumptions for parametric tests.Multiple regression and multiple logistic regression analyses were conducted with VFDs and the requirement for MV at hospital discharge as the dependent variables; and days to reach target energy intake from EEN initiation, initial SOFA score, admission source, category of admission, and time to EEN initiation from EICU admission as independent variables.
Multivariable analyses were conducted to identify the impact of days to reach target energy intake from EEN initiation on outcomes following control for differences in patients' baseline characteristic.Statistical analyses were conducted using SPSS statistical software v. 21 (IBM, Armonk, NY, USA).P values < 0.05 were considered significant.

Results
Among the 1993 patients who were assessed for eligibility, 65 met the criteria and were included in the study (Figure 1).Of these patients, the initial 29 (45%) were also included in our previous study [15].Reasons for exclusion were as follows: aged <18 years (n =

Patient Characteristics
Table 1 shows the baseline demographic, clinical, and nutritional characteristics of Group R and Group S. Demographics did not significantly differ between the two groups.Median patient age was 72.8 years and median BMI was 17.8 kg/m 2 .APACHE II score did not significantly differ between groups.Initial SOFA score was higher in Group R than in Group S (10 [9 -13] vs. 9 [8 -10]; P = 0.029).The number of surgical patients and patients who admitted to EICU from operating room were higher in Group S than in Group R. Of the patients, half (n = 33) were diagnosed with a neurological disorder, of whom 23 had intracranial hemorrhage, 4 had seizure, and 2 had spinal cord injury.Only Group R had 6 sepsis patients, and only Group S had 3 trauma patients.No patients were diagnosed with a gastrointestinal disorder.Charlson Comorbidity Index did not significantly differ between groups.

Outcomes
Table 2 shows patient outcomes.There were no significant differences in all-cause mortality in the EICU or hospital between the two groups.ICU-free days and length of hospital stay were not also significantly different.VFDs were significantly lower in Group R (18.0 [0.0 -22.0] vs. 21.0 [16.3 -24.8] days; P = 0.046).Significantly more patients required MV at hospital discharge in Group R than in Group S (29% vs. 8%; P = 0.030).In contrast, the number of patients undergoing tracheostomy during their stay in the EICU or hospital did not significantly differ between the two groups.Further, no significant differences were noted between the two groups regarding in antibiotic-free days, serum CRP level, development of pneumonia, incidence of liver or renal dysfunction, or glycemic control, or in the development of hypophosphatemia, lowest serum P  concentrations, or dose of P required for the treatment of hypophosphatemia.

Multivariable Analyses
Because VFDs might have been influenced by factors other than the rate of feeding advancement that were analyzed in our study, we also conducted a multiple regression analysis with VFDs as the dependent variable, using days to reach target energy intake from EEN initiation, initial SOFA score, category of admission, admission source and time to initiation of EEN from EICU admission as independent variables.Results showed no independent correlation between the number of days to reach target energy intake from EEN initiation and VFDs (regression coefficient = 0.500; 95% confidence interval, −0.420 -1.419).
In a similar manner, a multiple logistic regression analysis was also conducted to assess the independent effect of days to reach target energy intake from EEN initiation on the requirement for MV at hospital discharge after controlling for other factors, namely days to reach target energy intake from EEN initiation, initial SOFA score, category of admission, source of admission and time to initiation of EEN from EICU admission.
Days to reach target energy intake from EEN initiation was significantly and independently associated with the requirement for MV at hospital discharge (Table 3).

Discussion
In this study, we found no significant differences in the all-cause mortality or number of ICU-and length of hospital stay between patients with BMI < 20.0 kg/m 2 who received a rapid or slow rate of feeding advancement after initiation of EEN in the EICU.
However, VFDs were significantly lower and the requirement for MV at hospital discharge  Energy target was set at 25 kcal/kg/day for all patients.Energy intake included intake via enteral and parenteral routes; glucose infusions used for fluid support, drug dilution and correction of hypoglycemia; and lipids delivered with sedatives (e.g., propofol).Either oral or parenteral nutrition was not initiated within the first week of EICU admission.CI = confidence interval, EEN = early enteral nutrition, EICU = emergency intensive care unit, SOFA = Sequential Organ Failure Assessment.
was significantly higher in those with rapid feeding advancement.Multivariable analyses with adjustment for confounders revealed that the number of days to reach target energy intake from EEN initiation was significantly and independently associated with the requirement for MV at hospital discharge, but not with the VFDs.
In our previous study, we only focused on mean daily energy intake during the first week of EICU admission [15].In the present study, however, we focused on the rate of enteral feeding advancement after the initiation of EEN.We therefore conducted the present study in patients who initiated EN within the first 48 h of EICU admission and did not initiate oral or parenteral nutrition within the first week of EICU admission.
Patients were categorized into two groups according to the days to reach the energy target, with Group R reaching this target within 3 days of EEN initiation and Group S reaching this target 4 or more days after EEN initiation.Although both groups initiated EN within the first 48 h of admission, Group R started EN significantly earlier than Group S. This earlier initiation and rapid rate of feeding advancement might account for the larger energy intake in Group R patients.Protein intake was significantly higher in Group R than in Group S. As all patients were given enteral formulas with a fixed energy-protein ratio, protein intake might have been dependent on energy intake.
However, the present study focused on the rate of feeding advancement after initiation of EEN.
To our knowledge, this study is the first to evaluate the influence of the rate of feeding advancement after initiation of EEN on mortality and morbidity in underweight and critically ill and patients during emergency admission.Guidelines for critical care from the American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.) recommend the advancement of feeding toward the target over the 48 -72 h after the start of EEN [1], and the Canadian Clinical Practice Guidelines (CCPG) recommend starting at target feeding rate when initiating EN in critically ill patients [2].Several studies support each of these recommendations [9]- [14].Taylor et al. reported that enhanced EN (target feeding rate from Day 1) reduced infective complications compared to standard EN (gradual increase from 15 mL/h) [9].Petros et al. compared the impact of normocaloric (100% of energy expenditure) vs. hypocaloric (50% of energy expenditure) feeding in patients with a BMI of approximately 28 kg/m 2 in the first 7 days in an ICU [14], and found that the normocaloric group had a lower incidence of nosocomial infection than the hypocaloric group.
In contrast, several other studies do not support a rapid rate of feeding advancement for critically ill patients.Arabi et al. studied enterally-fed patients with a mean BMI of approximately 29 kg/m 2 and reported that underfeeding during the ICU stay (energy target: 60% -70% of calculated requirement) was associated with lower mortality than target feeding [6].The INTACT trial evaluated the outcomes of intensive medical nutrition therapy (IMNT: provision of >75% of estimated energy and protein requirements) compared to standard nutritional support in patients with a BMI of approximately 30 kg/m 2 [8].Notably, this study was terminated early due to a significant increase in mortality in the intensive IMNT group (40% vs. 16%, P = 0.017).Other studies reported no significant effects of initial feeding rate on major outcomes [7] [24]- [28].Most of the above studies were performed in patients aged from 50 to 70 years with a BMI of >25 kg/m 2 .In contrast, the patients in the present study had a BMI of <20.0 kg/m 2 and were older with median age of 73 years.These patients were admitted to the EICU to receive emergency medical or surgical treatment, while patients from general wards or were post-elective surgery were excluded.Low BMI is associated with higher complications and mortality in ICU as well as non-ICU settings [29]- [32].Elevated risk of all-cause mortality for BMI < 21 kg/m 2 level has been reported in middle-aged and elderly people in Japan [33].A negative balance of energy due to a slow rate of feeding advancement might diminish the adipose tissue mass of patients in order to meet energy expenditure.For most patients in previous studies, the impact of negative energy balance for short periods may have only been minor, as their fat reserves were normal or excessive.In contrast, fat and protein reserves in underweight patients with a BMI < 20 kg/m 2 , such as the patients in our present study, are relatively limited.The impact of negative energy balance in these patients is therefore considered to be of greater importance than that in patients with normal or excessive BMI.On the other hand, in malnourished patients, particularly those with a BMI of <18.5 kg/m 2 , a rapid delivery of energy might induce refeeding syndrome [34].Nevertheless, we observed no difference in the incidence of serum P concentration abnormalities between groups as symptomatic of refeeding syndrome (Table 2).
In our evaluation of the influence of rapid or slow feeding advancement after initiation of EEN on mortality and morbidity in critically ill medical and surgical patients with a BMI of <20 kg/m 2 , we observed that a slow advancement rate was associated with a lower requirement for MV at hospital discharge.However, the reliability of this result is statistically limited.Normally, multiple logistic regression analysis requires at least ten outcomes for one independent variable.Unfortunately, we were unable to adopt this method due to our small sample size.Several potential mechanisms might explain the association of rapid feeding advancement with an increased requirement for MV at hospital discharge.Increased feeding might have resulted in EN-related complications, for example aspiration pneumonia [35].Although several patients in the present study did develop pneumonia (Table 2), the retrospective design of the study meant that we could not determine if this was caused by aspiration.Other variables we considered to be associated with prolonged MV were the primary reason for ICU admission, acute physiology score on Day 1, age, prior patient location and hospital length of stay, activity limits due to respiratory disease, serum albumin, respiratory rate, and the partial pressure arterial oxygen/fraction inspired oxygen ratio [36].Unfortunately, the sample size of the present study was too small to allow multivariate analysis using all these variables as confounders.We therefore cannot deny the possibility that the requirement for MV was influenced by the underlying condition of the patient rather than the rate of feeding advancement.
Several limitations of the present study warrant mention.First and most importantly, sample size was small, and the study was conducted at a single institution.Further, study subjects were old Japanese with median age of 73 years.Our results are therefore limited and require confirmation in a larger sample size in patients with different generational and ethnic backgrounds.Second, the study was conducted under a retrospective design and groups were not randomized, which might have resulted in a degree of bias.Third, nutritional status was defined by BMI alone, which might be unreliable for nutritional assessment as actual body weight early in the clinical course might be influenced by non-nutrition factors such as fluid balance and inflammatory status.Further, we were unable to determine whether patients developed illness-related weight loss or were healthy despite their low body weight prior to EICU admission.Finally, an energy target was set at 25 kcal/kg/day for all patients, which was not tailored to the clinical status of each patient.In addition, the reason why physicians advanced EN rapider in Group R could not be identified.Thus, a conclusive answer to whether a slow rate of feeding advancement after the initiation of EEN improves outcomes in critically ill underweight patients requires large multicenter prospective studies with random patient assignment.

Conclusion
The results of this single-center, retrospective chart analysis demonstrate that a slow rate of feeding advancement after initiation of EEN in critically ill patients having a BMI of <20.0 kg/m 2 might be associated with a reduced requirement for MV at hospital discharge.These results require confirmation in a large multicenter trial which enrolls underweight and critically ill patients.

4 )
Number of patients requiring continuous or intermittent MV via tracheostomy at hospital discharge.The decision of when to wean from MV was made following discussion among the attending and/or EICU physicians.5) Number of patients undergoing tracheostomy in the EICU or hospital.The decision of when to perform a tracheostomy was made following discussion among the attending and/or EICU physicians.

Table 1 .
Demographics, clinical characteristics, and nutritional characteristics of group R and group S.Data were provided as median and interquartile range for continuous variables and as number and percentage for categorical variables.Mann-Whitney U test was used for continuous data and Fisher's exact test for categorical data.Group R reached 25 kcal/kg/day within 3 days of EEN initiation; Group S reached 25 kcal/kg/day 4 or more days after EEN initiation.Energy intake included intake via enteral and parenteral routes; glucose infusions used for fluid support, drug dilution and correction of hypoglycemia; and lipids delivered with sedatives (e.g., propofol).Either oral or parenteral nutrition was not initiated within the first week of EICU admission.APACHE II = The Acute Physiology and Chronic Health Evaluation II, EEN = early enteral nutrition, EICU = emergency intensive care unit, EN = enteral nutrition, IQR = interquartile range, SOFA = Sequential Organ Failure Assessment.

Table 2 .
Comparisons of clinical outcomes between group R vs group S.Data were provided as median and interquartile range for continuous variables and as number and percentage for categorical variables.Mann-Whitney U test was used for continuous data and Fisher's exact test for categorical data.Group R reached 25 kcal/kg/day within 3 days of EEN initiation; Group S reached 25 kcal/kg/day 4 or more days after EEN initiation.Energy intake included intake via enteral and parenteral routes; glucose infusions used for fluid support, drug dilution and correction of hypoglycemia; and lipids delivered with sedatives (e.g., propofol).Either oral or parenteral nutrition was not initiated within the first week of EICU admission.CRP = C-reactive protein, EEN = early enteral nutrition, EICU = emergency intensive care unit, EN = enteral nutrition, ICU = intensive care unit, IQR = interquartile range, P = phosphorus, RRT = renal replacement therapy, VAP = ventilator associated pneumonia.

Table 3 .
Multiple logistic regression analysis of the requirement for mechanical ventilation at hospital discharge in 65 underweight EICU patients who received EEN.