Post-pyloric versus gastric enteral nutrition in critically ill children – a single-center retrospective cohort study

Abstract

Objective-

We aimed to describe enteral nutrition (EN) delivery in patients receiving post-pyloric EN (PPEN) versus gastric EN (GEN).

Methods-

Single-center, retrospective study including patients <21 years of age, admitted to an ICU in a pediatric quaternary care hospital for ≧48 hours and who received PPEN or GEN as a first approach, as guided by a nutrition algorithm. PPEN patients were 1:1 propensity-score matched to GEN patients for demographics, clinical characteristics, and disease severity. Days to EN initiation from admission, percent EN adequacy (delivered EN volume/ prescribed EN volume) on days 1-3 and 7 after EN initiation, and time to achieving 60% of prescribed EN volume were compared between the two groups using Wilcoxon Mann-Whitney tests and a Cox proportional hazards model. Data are presented as median (IQR1, IQR3).

Main Results-

Forty-six PPEN and 46 GEN patients were matched. Median time to EN initiation was 3.25 (2, 6.8) days for PPEN and 4.15 (1.5, 7.1) days for GEN, p=0.6. Percent EN adequacy was greater for PPEN than GEN patients, [Day 1 PPEN 59.4% (18.8, 87.5) vs GEN 21.1% (7.8, 62.8), Day 2 PPEN 54.3% (16.7, 95.8) vs GEN 24% (5.4, 56.7), Day 3 PPEN 65.4% (14.7, 100) vs GEN 16% (0, 64.6), Day 7 PPEN 77.8% (11.1, 100) vs GEN 13.8% (0, 74.5), p<0.05]. PPEN patients had greater likelihood of achieving 60% of their prescribed EN volume than GEN patients, [HR 1.84, 95% CI 1.07-3.15, p=0.028].

Conclusions –

Post-pyloric EN was associated with greater EN delivery compared to GEN.

INTRODUCTION

In pediatric critical illness, early initiation and delivery of a minimum of 60% energy adequacy via enteral nutrition (EN) is associated with improved clinical outcomes.(1-3) Enteral nutrition can be delivered via the post-pyloric or gastric route. Gastric EN (GEN) delivery is thought to be more physiologic and is often the default route for EN in critically ill children, but may be associated with risk of aspiration.(4-6) Post-pyloric EN (PPEN) feeding bypasses the stomach, potentially reducing the risk of aspiration and improving EN delivery due to reduced intolerance.(7) The ASPEN and SCCM nutrition guideline recommends GEN as the preferred site for nutrition, but states that PPEN may be considered for patients with gastric EN intolerance or risk of aspiration.(1) The choice of post-pyloric versus (vs) gastric tube feeding varies by institutional practice, and the availability of resources and experience in placing post-pyloric tubes. Post-pyloric tube placement, for example, has been associated with delays in EN initiation due to difficulties in placing them.(8) In our institution, PPEN is recommended as the first route of EN for patients at risk for aspiration or with history of EN intolerance.(9, 10) In a cohort of 100 pediatric intensive care unit (ICU) patients, the indication for post-pyloric tube placement was risk for or history of aspiration in 48%, failed gastric EN (including intolerance) in 25% and non-invasive ventilation in 22% of patients.(10) We also reported an 85% success rate of post-pyloric tube placement at the bedside.(10)

In this study, we aimed to describe the characteristics of patients for whom PPEN was provided as the first mode of EN and compare nutritional and clinical outcomes in critically ill children who received PPEN vs GEN. We hypothesized that post-pyloric feeding would result in greater daily EN adequacy compared to gastric feeding in critically ill children receiving EN.

MATERIALS & METHODS

We performed a single-center, retrospective cohort study using propensity-score matched analysis of data extracted from an administrative database and the electronic medical record (EMR) to compare two modes of EN delivery in children admitted to a multidisciplinary ICU in a free-standing quaternary care pediatric hospital. The exposure in this study was PPEN as first approach for nutrition delivery, and control patients were those who received GEN. This study was approved by the local Institutional Review Board (IRB) and informed consent was waived.

Patient Selection

Patients less than 21 years of age, admitted to the ICU between 2013 and 2018 for a minimum of 48 hours, without a gastrostomy or jejunostomy, who received EN via nasoenteric tube were included. A cohort of 100 patients with post-pyloric tubes were identified using records of radiographs to examine tube placement practices, as previously published.(10) In 54 of these patients, a post-pyloric feeding tube was placed as the first approach for nutrition delivery and these were selected as cases for this propensity-score matched analysis. We subsequently utilized the Pediatric Health Information System (PHIS), further described below, to screen for eligible GEN patients at our institution between 2013-2018. Eligibility, including GEN as the first approach for EN delivery, was confirmed through a review of the EMR and patients were enrolled consecutively starting from 2018 through 2013. Gastric EN patients served as controls for the propensity-score matched analysis.

Gastric or post-pyloric tube placement in our ICU is guided by a stepwise EN algorithm, as previously published.(9)(Supplemental File 1 including the algorithm) Specifically, post-pyloric tubes are considered for patients with risk for aspiration as detailed in the algorithm and history of EN intolerance. (9, 10) The nutrition algorithm also guides time to initiation and advancement of EN, and recommendations are identical for patients fed via post-pyloric or gastric tubes.(9) Per the algorithm all patients are fed continuously in the acute phase of illness and fasted for procedures. Indications for EN interruptions include EN intolerance defined as abdominal distension, abdominal discomfort, emesis and/or diarrhea for both PPEN and GEN patients, and also gastric residual volume (GRV) for GEN patients. Gastric residual volume cutoffs of >3ml/kg or >150mL for patients 50kg or greater were considered EN intolerance during the time of this study, and refeeding of the GRV was recommended when possible. Further details regarding criteria for EN initiation, advancement and interruption are in Supplemental File 1. Post-pyloric feeding tubes are placed at the bedside by trained registered nurses or nurse specialists without special devices and the position of the tip of the tube is confirmed by X-Ray, as previously described.(10) Gastric feeding tubes are also placed at the bedside without special devices and confirmed by X-Ray.

Nutrition prescriptions are determined by the unit’s dietitian, who primarily employs the WHO equation for patients <2 years of age, the Schofield equation for patients 2-18 years of age and the Harris-Benedict equation for patients >18 years of age.(11-13) Standard non-hydrolyzed formulas are used for the majority of patients with a caloric density of 20 kcal per ounce for patients less than 1 year of age and 30 kcal per ounce for patients older than 1 year of age.

Data Abstraction

Patient demographic and clinical data were extracted for the specific ICU admission during which the post-pyloric or gastric tube was placed. We used the PHIS database to extract these data for both cohorts. The PHIS database is an administrative database with clinical and resource utilization data on patients treated at 49 nonprofit, tertiary care children’s hospitals in the United States. The Children’s Hospital Association (CHA) (Overland Park, KS) maintains the PHIS database and assures data quality and reliability. Although the PHIS database holds data from multiple hospitals, in this study data were extracted only from patients who received care at our institution during the study period. Data extracted from PHIS include patient demographics (age at admission, race and ethnicity, sex, insurance payor), clinical information (priority of admission, medical or surgical admission, mechanical ventilation, pharmaceutical utilization, complex chronic condition based on Feudtner classifications(14)), and resource intensity weight based on total CHA Children’s Hospital Relative Cost Weight as a proxy for acuity.

Nutritional data were extracted from the EMR and included: prescribed goal EN volume, time to EN initiation from admission, EN adequacy days 1-3 and 7 after EN initiation, hours of EN interruption per day on days 1-3 of EN administration, time to 60% and 100% of prescribed EN goal, and whether the patient achieved 100% of EN goal at any time during their admission.

Propensity-Score Matched Analysis

We used a multivariate logistic regression model to generate propensity scores in order to match patients who received PPEN with patients who received GEN. The model included patient demographics (age, sex, race and ethnicity, payor) and clinical characteristics (discharge year, priority of admission, admission diagnosis, mode of respiratory support at time of tube placement, extracorporeal membrane oxygenation (ECMO) use at time of tube placement, ICU unit, medical or surgical admission, and mechanical ventilation). We also matched by category of medications administered (paralytics, sedatives, vasoactive agents, narcotics, etc.). Full list of medications per category available in Supplemental File 2. We matched for presence of a complex chronic condition, resource utilization cost weights, and PIM2 probability of death (POD) scores as surrogates of disease severity.(15) These variables were selected to include known confounders that might influence the decision to choose PPEN vs GEN and those that influence the outcomes. We used the Greedy matching algorithm to achieve 1:1 propensity score matching between the two groups.

We used descriptive analyses to compare patient, visit, and clinical level characteristics among the two groups in both the unmatched and matched cohorts. Continuous variables were summarized using medians (interquartile range) and categorical variables were summarized using frequencies and percentages. Standardized mean differences were calculated for all variables. For both the pre and post-matched samples, we assessed differences across clinical and demographic characteristics of patients in the two groups using Chi Square and Fisher’s exact tests for categorical variables and Wilcoxon Mann-Whitney tests for continuous variables.

Primary outcome was daily EN adequacy days 1-3 and 7 of EN delivery. Daily EN adequacy was defined as administered EN volume divided by prescribed EN volume to achieve caloric goal and presented as a percentage. Secondary nutritional outcomes included time from admission to EN initiation, time from EN initiation to achieving 60% and 100% of prescribed EN goal, and hours of EN interruption on Days 1-3. Secondary clinical outcomes included in-hospital mortality, ICU length of stay (LOS), hospital LOS, and mechanical ventilation days.

Differences in EN adequacy achieved on Days 1, 2, 3 and 7 of EN delivery between the matched post-pyloric and gastric groups were assessed using Chi Square tests. Hours from admission to EN initiation, hours from EN initiation to 100% goal EN volume and hours of EN interruptions were assessed using Wilcoxon Mann-Whitney tests for continuous variables. We created a Kaplan-Meier plot to illustrate differences in days to achieving 60% or greater of nutritional goal among the matched PPEN and GEN patients, and used a Cox proportional hazard model to estimate these differences with death, discharge or transfer out of the ICU as competing risk variables. For all nutrition analyses if a patient transitioned from the initial mode of nutrition to another mode of nutrition (eg. PPEN or GEN to oral diet, or PPEN to GEN or vice versa) they were censored from subsequent analyses on the day of transition.

Generalized linear mixed effect models, with matching indicator as random effect, were used to assess effect of mode of EN (post-pyloric vs gastric) with ICU LOS, hospital LOS, and mechanical ventilation days including correction per 100 ICU days. Effect estimates were described using Odds Ratios (95% Confidence intervals) for in-hospital mortality and Rate Ratios (95% confidence intervals) for ICU and hospital LOS, mechanical ventilation days, and mechanical ventilation rate (mechanical vent days per 100 ICU days

All analyses were performed using SAS v.9.4 (SAS Institute, Cary, NC), and p-values <0.05 were considered statistically significant.

RESULTS

Fifty-four patients from the previously identified cohort of 100 patients for whom a post-pyloric tube was placed received PPEN as the first mode of EN delivery. Table 1S depicts the demographics and clinical characteristics of the post-pyloric and gastric fed cohorts before and after matching. In the unmatched cohorts, post-pyloric fed patients had a lower prevalence of respiratory illness, infection or sepsis, and gastrointestinal conditions as their diagnosis of admission. The percent of post-pyloric fed patients requiring paralytics, sedatives, gastrointestinal medications or narcotics was also lower than the percent of gastric fed patients. Other characteristics, including severity of illness by PIM2 score, need for mechanical ventilation and presence of an underlying complex chronic condition, were similar between the unmatched cohorts.

Eight PPEN patients could not be adequately matched to a comparable control, leaving 46 PPEN patients matched 1:1 to 46 GEN patients using propensity-score matching, with no statistically significant differences between the two groups. The mean (standard deviation) age in years of the post-pyloric vs gastric fed patients was 6.5 (7.1) years vs 5.2 (6.4) years, respectively, and 28% and 39% were female, respectively. The three most common diagnoses of admission for both cohorts were respiratory, gastrointestinal or neurologic disease and 67% of PPEN and 71% of GEN patients were invasively- mechanically ventilated, whether via orotracheal or tracheostomy tube, Table 1S.

Table 1 depicts the nutritional outcomes. Days to EN initiation was not statistically different between the PPEN and GEN cohorts, [PPEN 3.25 (2, 6.8) days and GEN 4.15 (1.5, 7.1) days, p=0.6]. Patients that were fed via post-pyloric tube received greater percent of goal EN volume on Days 1-3 and 7 of nutrition delivery than GEN patients. The percent of goal EN delivered increased from Day 1 (59.4%) of nutrition through Day 3 (65.4%) for patients who were fed via post-pyloric tube, whereas gastric tube fed patients did not have an overall increase in percent goal EN volume delivered (Day 1 21.1% and Day 3 16%). Patients fed via post-pyloric tube also received a significantly greater percent of EN volume goal on Day 7 of nutrition than patients fed via gastric tube, [PPEN 78% (11, 100) EN volume goal and GEN 14% (0, 75) EN volume goal, p=0.017]. More patients fed via post-pyloric tube achieved >60% of prescribed EN volume goal after EN initiation than gastric fed patients, [PPEN 34/46, 74% and GEN 22/44, 50%, p=0.019]. Post-pyloric fed patients had less hours of EN interruptions on Day 1 and 3 of nutrition [Post-pyloric- 5 hours (3, 8.5) and Gastric- 17.5 (5,24), p=0.003; Day 3 Post-pyloric- 7 hours (1.5, 21.5) and Gastric-13 (7, 24), p=0.25] but not on day 2 [Day 2 Post-pyloric- 10 hours (2, 23) and Gastric- 6 (3,22), p=0.58].

Table 1.

Enteral nutrition outcomes in the matched gastric and post-pyloric fed cohorts

Outcome	Group	Matched
		N	Median	Q1	Q3	p value
Goal EN volume prescribed by RD, mL	Gastric	46	865.5	528.0	1440.0	0.568
	Post-pyloric	46	1020.0	672.0	1200.0
EN Adequacy: Day 1 Goal %	Gastric	44	21.1	7.8	62.8	0.016
	Post-pyloric	46	59.4	18.8	87.5
EN Adequacy: Day 2 Goal %	Gastric	42	24.0	5.4	56.7	0.014
	Post-pyloric	43	54.3	16.7	95.8
EN Adequacy: Day 3 Goal %	Gastric	38	16.0	0.0	64.6	0.009
	Post-pyloric	40	65.4	14.7	100.0
EN Adequacy: Day 7 Goal %	Gastric	33	13.8	0.0	74.5	0.017
	Post-pyloric	34	77.8	11.1	100.0
Time from admission to initiation of EN	Gastric	46	4.15	1.50	7.10	0.601
	Post-pyloric	46	3.25	2.00	6.80
Time from initiation of EN to 100% EN goal (hours)	Gastric	23	28.0	0.0	181.0	0.831
	Post-pyloric	37	35.0	17.0	77.0

EN adequacy defined as EN volume delivered divided by EN volume prescribed to meet caloric goal.

EN, enteral nutrition; RD, registered dietitian

We examined the time to achieve 60% of prescribed EN goal using a time-to-event analysis, Figure 1. Patients who were fed via post-pyloric tube had greater likelihood of achieving 60% of their goal EN volume earlier than patients who were fed via gastric tube, [HR 1.84, 95% CI 1.07-3.15, p=0.028]. The median number of days to achieve 60% of EN goal volume was 2 (1,2) days for PPEN patients and 7 (2, -) days for GEN patients, p=0.01.

Figure 1.

Time-to-event analysis of the days to achieving 60% of prescribed enteral nutrition goal volume between post-pyloric and gastric tube fed critically ill children. Patients were censored for discharge, transfer out of the intensive care unit or death. Panel A. graphically depicts the time to achieving 60% of prescribed EN goal between post-pyloric (black line) and gastric (dashed line) fed patients over the first 3 days of EN nutrition and on day 7. Panel B. includes the statistical analysis details. Post-pyloric tube fed patients were 84% more likely to achieve 60% enteral nutrition goal over a 7-day period than gastric red patients (HR 1.84, (95% CI 1.07, 3.15), p=0.028). Post-pyloric tube fed patients achieved 60% of enteral nutrition goal in a median of 2 days (1,2) compared to 7 days (2, --) for gastric tube fed patients, p=0.010.

There were no differences in the prevalence of secondary clinical outcomes between PPEN and GEN patients, Table 2.

Table 2.

Clinical outcomes in the matched post-pyloric and gastric fed cohorts

Clinical Outcomes	Matched cohort
	Post Pyloric (N=46)	Gastric (N= 46)	Odds Ratio/Rate Ratio (95% CI)	p value
Death in hospital, N (%)	15 (32.6%)	13 (28.3%)	1.18 (0.53,2.64)	0.683
ICU LOS days, Mean (95%CI)	35.5 (27.4, 46.0)	28.9 (22.3,37.4)	1.23 (0.85, 1.77)	0.261
Total LOS days, Mean (95%CI)	45.3 (33.1, 62.2)	36.6 (26.7,50.2)	1.24 (0.79, 1.94)	0.341
Mechanical Ventilation days, Mean (95%CI)	31.5 (24.0,41.2)	21.7 (14.7,32.1)	1.45 (0.90, 2.33)	0.123
Mechanical Ventilation rate (per 100 ICU days), Mean (95%CI)	76.7 (67.8, 86.7)	65.7 (56.9,75.7)	1.17 (0.97, 1.41)	0.103

ICU, intensive care unit; LOS, length of stay

DISCUSSION

We have described a cohort of critically ill children in whom EN was initiated via the post pyloric route as a first approach. This group had similar demographics and clinical severity compared to the group initiated on EN via the gastric route, but had a different distribution in the diagnoses of admission and in the use of medications prior to matching. After propensity-score matching, we assessed nutrition outcomes and identified that critically ill children who received post-pyloric nutrition had greater EN delivery when compared to gastric nutrition. Patients fed via post-pyloric tube received greater than 50% of their prescribed volume to meet caloric goal in the first three days of EN compared to patients fed via gastric tube who received an average of 20% of their prescribed volume. Median number of days to achieve 60% of prescribed EN goal volume was 2 days for patients fed via post-pyloric tube compared to 7 days for patients fed via gastric tube. There were no statistically significant differences in clinical outcomes between the two cohorts such as LOS or mechanical ventilation days. The results of this study support the use of post-pyloric tubes in a select group of children, particularly those identified based on an EN algorithm. Future studies are needed to examine whether improved EN delivery due to use of the post-pyloric route as first approach is associated with improvement of other clinical outcomes and to understand the physiologic underpinnings of this association.

In this study, we have shown that patients who received PPEN achieved on average 60% of their prescribed EN goal volume in the first 3 days of EN delivery, whereas GEN patients received an average of 20% of their prescribed EN goal volume. These findings are similar to a pediatric randomized controlled trial (RCT) with a primary outcome of EN provision, and summary results from a Cochrane review including multiple adult RCTs examining PPEN vs GEN.(7, 8) In the pediatric study, patients randomized to PPEN received ~50% of their goal over a 14-day period and received 2 more days of nutrition compared to GEN patients who received an average 30% of their nutritional goal.(7) In the Cochrane review, critically ill adults who were fed via a post-pyloric tube received between 1.4 and 14.2 times higher mean percentage of nutrition compared to gastric fed patients.(8)

One reason why PPEN is thought to be associated with greater EN provision is the possible lower incidence of EN intolerance and therefore less frequent EN interruptions. In our study, PPEN patients had a median 12 hours and 6 hours less of EN interruptions on Days 1 and 3 of EN delivery, respectively, than GEN patients. Although, statistical significance regarding hours of EN interruptions was only met on Day 1, a median of 6 hours more of EN interruptions in the GEN group on Day 3 may be clinically relevant and could contribute to lower EN delivery. As a retrospective study, we could not reliably ascertain the reason for EN interruption, however per our institutional algorithm, the indications for EN interruption and signs of EN intolerance are the same for PPEN and GEN patients. It is possible that the difference in EN interruptions between our study groups was due to a difference in the frequency of procedures although the NPO guidelines for PPEN and GEN patients are the same. Furthermore, in this study, matching for severity of illness, and type of and reason for admission may have indirectly accounted for the number of procedures between groups. Therefore, hours of EN interruptions may serve as a proxy for intolerance, and our results hint at greater EN tolerance for the PPEN group on day 1 and possibly day 3 of nutrition. Previous studies have not reported significant differences in EN interruptions or intolerance between post-pyloric and gastric fed patients. A pediatric RCT reported no difference between post-pyloric and gastric feeding in the number of patients whose feeds were interrupted or in the prevalence of symptoms of intolerance (vomiting, diarrhea, distension) over a 14-day period.(7) A Cochrane review of critically ill adults reported that no individual study identified a difference in vomiting and diarrhea, common signs of EN intolerance, between post-pyloric and gastric fed patients.(8) Direct comparisons regarding interruptions or intolerance between our study and the aforementioned RCTs cannot be made due to differences in study design and timeline. The precise reasons for greater EN adequacy in post-pyloric tube feeding remain unclear, and may be due to factors beyond EN intolerance. Additional factors such as perceptions of tolerance and safety by bedside providers, regional EN practice algorithms, and actual difference in gastrointestinal function might be responsible for differences in the achievement of nutritional goal between PPEN and GEN groups.

Previous studies comparing PPEN vs GEN have reported difficulties with placing post-pyloric tubes and associated delays in EN initiation. In a pediatric RCT, post-pyloric tube placement was not successful in 30% of patients randomized to post-pyloric nutrition.(7) Two adult RCTs reported a greater than 12-hour delay in post-pyloric tube placement and EN initiation compared to gastric feeding tube placement, and between 23% and 70% of post-pyloric tube placements required advanced techniques due to failed bedside placement.(16, 17) Our study was different from these prior studies in that the time to initiate EN was similar between patients who received PPEN and GEN, which allowed for comparison of PPEN and GEN as a first approach without any delays associated with tube placement between groups.

Our study is unique from prior pediatric studies due to its focus on the post-pyloric route as a first approach for EN delivery and its use of propensity-score matched analysis to evaluate for differences between PPEN and GEN delivery in critically ill children. Propensity-score matched analysis balances the distribution of potential confounders between the two cohorts. In this study, we included a robust list of potential confounders including demographics, clinical diagnoses, markers of severity of illness and interventions known to influence our outcomes of interest such as medications. However, the possibility for unknown confounders is still present and must be acknowledged. The presence of covariates in both cohorts, as was the case in our study, is important to ensure there are no intrinsic differences in the cohorts that could account for the prevalence of the outcomes of interest. As a retrospective cohort study design, our study is limited by potential missing data, as is possible in the dataset associated with EN interruptions and the lack of data on clinician decision-making. The selection of patients with post-pyloric tube placement was dependent on natural language processing of radiograph documentation in the EMR, and therefore may have missed patients with post-pyloric tubes. However, our results reflect nutritional outcomes in the setting of routine practice in a pediatric ICU and under the guidance of an established nutrition algorithm that recommends the same EN initiation and advancement strategies and guidelines for EN interruptions and NPO status between post-pyloric and gastric fed patients.(9) Our study population was heterogenous in diagnosis and included patients with and without respiratory support, whereas prior pediatric studies have only included mechanically ventilated patients. Our sample size though larger than that of previously published pediatric studies on post-pyloric nutrition, was still limited and may not be powered to detect some differences among the study groups that may be identified in a larger study population. Studies including larger sample sizes should be completed. Studies to date, including ours, have failed to explain why PPEN results in improved EN delivery. Future studies should consider examining differences in gastrointestinal physiology between PPEN and GEN. Nevertheless, our study and prior studies suggest that empiric PPEN may benefit critically ill children as long as post-pyloric tube placement doesn’t result in significant delays in EN initiation and selection of the population that will benefit the most for this approach is considered.

CONCLUSIONS

In a single-center, propensity-score matched analysis, PPEN was initiated without delays, and was associated with greater percent goal EN volume delivered during days 1-3 and 7 of nutrition provision compared to GEN patients. Post-pyloric EN was also associated with achieving 60% of goal EN volume over a shorter period of time. Less EN interruptions and EN intolerance are possible reasons for greater EN delivery in post-pyloric fed patients though future studies should explore additional potentially contributing factors such as clinician’s perceptions on safety or differences in gastrointestinal physiology between the two modes of EN. The impact of route of EN on clinical outcomes needs to be examined in larger studies.

CLINICAL RELEVANCY.

Achieving 60% of EN goal in critically ill children is associated with improved clinical outcomes. Enteral nutrition delivery however is often limited by EN intolerance. In this study, using a propensity-score matched analysis, we show than post-pyloric tube fed patients achieved 60% of their prescribed enteral goal 5 days earlier and received an average of 40% more of their EN goal in their first 3 days of nutrition delivery compared to gastric fed patients. This study suggests that the use of post-pyloric feeding as a first approach and as guided by a nutrition algorithm can result in improved EN delivery as long as there are no delays in tube placement and EN initiation.