Malnutrition and muscle mass loss (sarcopenia) are frequent complications of liver cirrhosis and alcoholic hepatitis (AH) and have been reported in more than 50% of patients with decompensated liver cirrhosis and patients with AH (Palmer et al, 2019; Sehrawat et al, 2020). Furthermore, malnutrition and sarcopenia are associated with higher rates of complications, such as hepatic encephalopathy, spontaneous bacterial peritonitis and increased mortality (Gunsar et al, 2006; Huisman et al, 2011; Montano-Loza et al, 2012).
Patients with severe liver disease (Child-Pugh class C) (Table 1) are regarded to be at high risk of malnutrition (Merli et al, 2019). Recommended daily energy intake for a patient with liver cirrhosis is at least 35 kcal/kg, with daily protein intake no less than 1.2-1.5 g/kg (Plauth et al, 2006; Plauth et al, 2009; Merli et al, 2019). Inpatients with severe liver disease who are unable to achieve the recommended intake with oral nutrition and oral supplements could be candidates for enteral or parenteral nutrition. Single studies have shown the benefit of enteral nutrition for survival, but no systematic reviews or meta analyses have shown that enteral nutrition has an effect on mortality (Antar et al, 2012; Koretz et al, 2012). Patients with cirrhosis have a higher risk of gastroparesis (Verne et al, 2004). Furthermore, patients with liver disease are also characterised by having anorexia, early satiety and nausea (Cheung et al, 2012; Wang et al, 2018).
| Score | 1 | 2 | 3 |
|---|---|---|---|
| Encephalopathy grade | 0 | 1 or 2 | 3 or 4 |
| Ascites | None | Slight | Moderate |
| Bilirubin (mg/dL) | < 2 | 2–3 | > 3 |
| Albumin (g/dL) | > 3.5 | 2.8–3.5 | < 2.8 |
| INR* | < 1.7 | 1.7–2.2 | > 2.2 |
| Child-Pugh class A |
5–6 points |
There are no guidelines recommending a specific type of tube for delivering enteral nutrition to ensure that the patient received a sufficient amount of nutrients. A nasogastric (NG) tube delivers enteral nutrition to the stomach and is relatively easy to insert, while a nasojejunal (NJ) tube delivers nutrition to the duodenum—that is, the tube passes through the pylorus. An NJ tube is typically thinner than an NG tube. However, it can be more challenging to insert an NJ tube, and a length of tube should be left outside the nose, allowing the tube to spontaneously pass further down to the duodenum. In patients with cirrhosis, and even a low grade of hepatic encephalopathy, it can be challenging to keep the external part of the NJ tube in place, because patients might remove it. However, an NJ tube that bypasses the stomach may be preferable, considering the increased risk of gastroparesis and nausea if an NG tube is used.
The overall nutritional aim when caring for inpatients with liver cirrhosis or AH must be to ensure or deliver as much of the individual's nutritional needs as possible. Whether this can be achieved via an NG or NJ tube is unclear. The primary aim of this study was to determine whether early NJ feeding, compared with NG feeding, can increase total daily energy and protein delivery. The secondary aim was to determine the percentages of energy delivered by NG and NJ tubes compared with patients' calculated nutritional needs and to monitor the number of tube replacements.
Methods
Design
A randomised trial was conducted between January 2015 and October 2018 at Aarhus University Hospital, Denmark. Patients were randomised 1:1 into two arms: NJ tube (Flocare Bengmark, Ch10, Nutricia) or NG tube (Flocare Pur Tube Enlock, Ch12, Nutricia). Randomisation was performed by an unbiased study nurse using computerised random number generation (www.randomization.com). Due to the nature of the intervention, no blinding of participants or investigators was attempted.
All patients were followed daily until discharge or death. In addition, follow-up visits at 3 and 6 months after inclusion were undertaken, if possible. When tube replacement was necessary, the same type of tube was inserted as initially assigned. After three tube replacements, intravenous nutrition was considered as an alternative route. Patients' individual nutrition needs were calculated by a dietitian using Harris-Benedict equations (Frankenfield et al, 1998), which estimate individual basal energy needs. All participating patients had free intake of oral food, but were supplemented with concentrated enteral nutrition (Peptamen AF, Nestle Health Science), regardless of which tube was inserted. The scheduled maximum fixed doses of enteral nutrition were
The doses from day 5 were individually tailored, depending on each patient's needs and oral intake. All patients received metoclopramide 10 mg per day to prevent nausea and vomiting.
Patients
Some 40 inpatients with liver cirrhosis or/and AH were included in the trial. Eligible patients were men and women aged over 18 years, who on three consecutive days had a daily energy intake of less than 75% of their estimated need. Patients with substantial mental impairment due to hepatic encephalopathy (grade 3 or 4) were excluded because of the difficulty in obtaining their informed consent (American Association for the Study of Liver Diseases, 2014). Patients for whom the nutrition tube was contraindicated (eg due to anatomic conditions) were excluded.
Nutritional status
As per normal procedure in the department each patient's weight, body mass index (BMI), weight change and handgrip strength were used as markers of nutritional status (Álvares-da-Silva et al, 2005; Norman et al, 2011). The authors collected data regarding self-reported height and actual or self-reported body weight of patients, along with data on habitual weight, weight change (stable, weight loss or weight gain), and BMI. Low BMI was defined as <20.5 kg/m2, according to the nutritional risk score NRS-2002 (Kondrup et al, 2003).
Handgrip strength
Handgrip strength was used as a method of identifying patients with malnutrition—as a marker of nutritional status, muscle function and sarcopenia (Cruz-Jentoft et al, 2010; Norman et al, 2011). It was measured using a dynamometer and the data were expressed in kg. The best of three measurements on the non-dominant arm was used for further analysis. Low handgrip strength was determined as a handgrip strength below 95% confidence intervals for grip strength norms, according to Bohannon et al (2006).
Nutritional risk score
Nutritional risk score was assessed using the NRS-2002 (Kondrup et al, 2003), which has two components: an A score that assesses BMI, unintentional weight loss, changes in food intake; and a B score, which is a disease severity score defined as low, moderate or severe. Furthermore, in line with the NRS, the score was adjusted for patients aged >70 years. The total score is the sum of the A and B scores, plus a potential age score. Patients were defined as being at nutritional risk if their final score was more than 2 on a scale of 0 to 7.
Reported dietary intake
On admission, the Department of Hepatology and Gastroenterology dietitian made a detailed assessment of each patient's dietary intake at home, establishing what and how much the person was normally capable of eating. The assessment included types and quantity of food and the number of meals the patient normally ate, as well as calories and protein intake. Patients were asked whether their dietary intake had changed prior to admission, for how long and how had it changed. Nutritional counselling was provided for each patient. This was followed by the calculation and evaluation of each patient's estimated required daily dietary intake for 2 days.
Measurements
Throughout the study period, data on each patient's daily nutrition were recorded. We also measured the quantity of energy and protein, as well as the route of delivery (oral or tube). When possible, body weight and data on nausea and vomiting were recorded daily. Handgrip strength was recorded weekly. The patient's Child-Pugh class was calculated using the results of blood samples collected at baseline (Table 1). Furthermore, information on each patient's sex, age, diagnoses, comorbidities, and concomitant medications was collected at baseline. Details on possible tube replacements were collected throughout the study.
Efficacy and safety assessments
The efficacy analysis included all patients who remained in the study for a minimum of 7 days. The protocol-defined primary endpoint was the difference in energy delivered. The secondary endpoint was the proportion of patients who needed tube replacements. Data on possible side-effects to the tube feeding were collected continuously throughout the study.
Statistical analysis
The sample size needed to show significant results was calculated prior to study initiation. With two groups, 14 patients were required in each group, based on the following assumptions:
With an estimated potential drop-out rate of 25%, each group required 20 patients, or a total of 40 participants.
Statistical analysis was performed primarily according to the intention-to-treat principle. Summary statistics were used to describe the two groups. Continuous variables are expressed as means, medians and ranges, or confidence intervals. Outcome measures were analysed using Student's t-test, the chi-square test or non-parametric tests. The software programs EpiData and Stata 15 were used to analyse the data.
Ethical considerations
All patients provided informed written consent to take part. Overall, the study was conducted in accordance with the Declaration of Helsinki. The experimental protocol was approved by the Central Denmark Regional Committee on Health Research Ethics (j.no. 1-10-72-322-14) prior to the investigation. The study protocol was registered at www.clinicaltrials.gov (identifier: NCT02319252).
Results
A total of 146 patients were screened, of whom 49 met the inclusion criteria and had none of the exclusion criteria. Nine of the 49 did not want to participate in the study, leaving 40 patients (Figure 1). The most common reasons for exclusion were high oral intake or hepatic encephalopathy. Other reasons were planned discharge, transfer to intensive care unit, not adherent to tube application, dementia, psychosis or delirium.
The 40 patients who were enrolled in the study were all at nutritional risk according to the NRS-2002 and were randomised 1:1 into each group. Six patients were discharged, and 1 patient died prior to day 7, which left 33 patients for analysis at day 7. By day 14, most patients had either been discharged from hospital or had died. Only 13 patients remained in the study at day 14; therefore, the authors limited the data analysed to that collected prior to day 7.
Mortality in the study population was expected and was evaluated not to be related to the type of nutrition tube or participation in the study. Patients' baseline characteristics are shown in Table 2, at which point there were no significant differences between the two groups. Most patients had liver cirrhosis, and all participants were classified as Child-Pugh class B or C (Table 1).
| Total (n=40) | Gastric tube (n=20) | Jejunal tube (n=20) | P value | |
|---|---|---|---|---|
| Demographics | ||||
| Age (years), median (IQR) | 58 (54–64) | 58 (52–63) | 59 (56–64) | 0.69 |
| Gender (male), n (%) | 16 (40.0) | 9 (45.0) | 7 (35.0) | 0.53 |
| Liver-related diagnosis | ||||
| Cirrhosis | 28 | 13 | 15 | 0.57 |
| Alcoholic hepatitis | 8 | 4 | 4 | |
| Both | 4 | 3 | 1 | |
| Child-Pugh class, median (IQR) | 10 (8–12) | 10 (8–12) | 10 (8–12) | 0.79 |
| A (5–6) | 0 | 0 | 0 | |
| B (7–9) | 12 | 7 | 5 | |
| C (10–15) | 28 | 13 | 15 | |
| Body weight | ||||
| Median (kg), (IQR) [range] | 67 (58–83) [46–115] | 69 (57–83) [46–115] | 67 (55–82) [46–97] | 0.78 |
IQR=interquartile range; kg=kilogram
Nutrition intake
As an average of the first 7 days, none of the groups reached the estimated daily nutritional intake. As shown in Table 3, there were no statistically significant differences between the NG group and the NJ group. A stratification between Child-Pugh class B and C also showed no major differences. However, the percentage of actual nutritional intake when compared to the estimated requirements tended to be higher for patients allocated to the NG tube group.
| Gastric tube (n=17) | Jejunal tube (n=16) | P value | |
|---|---|---|---|
| Mean estimated require Daily nutrition intake, kJ (SD) | 8090 (1016) | 8390 (911) | 0.38 |
| Mean actual energy intake for 7 days kJ (SD) [% intake of estimated requirements] | |||
| All patients (n=33) | 6509 (2205)[80.5] | 6605 (2369)[78.7] | 0.90 |
| Child-Pugh B (n=9) | 6100 (1856)[75.4] | 7109 (2934)[84.7] | 0.54 |
| Child-Pugh C (n=24) | 6732 (2429)[83.2] | 6489 (3281)[77.3] | 0.81 |
| Mean actual protein intake, g (SD) | 75 (24) | 72 (22) | 0.73 |
| Median changes in body weight, kg (range) | 1.6 (-4.1 to 7.6) | 2.2 (-6.4 to 6.4) | 0.89 |
| Tube removals by patients | |||
| Once | 9 | 7 | |
| Twice | 4 | 5 | |
| Three times | 2 | 4 |
g=grams; kg=kilograms; kJ=kilojoule; SD=standard deviation
Tube removal
The number of tube removals was similar in both groups, regardless of whether the patient of a health professional removed the tube. In most instances, the tube was removed by patients; in other cases, staff removed tubes prior to endoscopy and when upper gastrointestinal bleeding occurred (Table 3).
Other measurements
Changes in the patients' body weights were similar in the two groups. Data on nausea and vomiting were insufficient, with fewer than 10 responses per day. Also, due to the patients' condition, it was possible to measure handgrip strength only in half the patients and these data were not included in the analysis. Reasons for missing data included: the development of hepatic encephalopathy grade 3-4, the patient being out of the department, staff busyness or oversight. The sparse data were not analysed.
Safety
During the study, no adverse events were deemed to be related to the allocated type of nutrition tube. The number of patients with diarrhoea was comparable in the two groups (20–30%). The prevalence of diarrhoea was stable within the 7 days studied.
Discussion
In this study, the authors hypothesised that NJ feeding, compared with NG feeding, would be more efficient in delivering the calculated amount of food and result in fewer accidental tube removals. However, this was not supported by the data. NJ feeding was not superior to NG feeding, and there was no significant difference in tube removals in the two groups. There were also no significant differences in food intake either. However, the results revealed a trend towards a higher intake in the case of NG tubes, regardless of a patient's Child-Pugh classification.
This randomised controlled trial looked at patients with severe liver disease. Of the patients included in the study (n=40), 70% were classified as Child-Pugh class C. The authors screened 146 patients, resulting in the inclusion of 40 patients. Only 9 patients declined to participate, and the remainder failed screening for a variety of reasons. This illustrates the complexity of the studied population. Patients with hepatic encephalopathy grade 3 or 4 were excluded, despite the fact they still required dietary supplements to ensure sufficient nutrition. Some had a complex disease history and were unable to participate; others were about to be transferred to another hospital or intensive care unit. Despite the strict inclusion and exclusion criteria, the study sample was reduced to 33 patients at day 7.
In our power calculation prior to the start of the study, we estimated a potential 25% drop-out rate. Seven patients dropped out, which was within the expected limit. We also estimated a 25% difference in nutritional intake between the groups. The difference in our study was less marked. The marked variances between patients in our study population supported a randomised study design to distribute possible bias equally. No statistically significant differences between the two groups were revealed at baseline, and the drop-out rate was comparable in the two groups. However, the results indicate that the study may have been underpowered.
To our knowledge, this is the first study in which NG feeding and NJ feeding have been compared in patients with liver cirrhosis and/or AH. However, comparisons between the two types of tube placement have been studied in other patient populations. An Australian multicentre, randomised controlled trial compared NJ with NG feeding in 181 critically ill, medically ventilated patients (Davies et al, 2012). The total amount of energy delivered was found to be equal in the two groups. In a Scottish study (Eatock et al, 2005), 50 patients with severe acute pancreatitis were included in a randomised controlled trial. Patients were allocated to either NG tube feeding or NJ tube feeding. The Eatock et al (2005) study focused on safety, efficiency and complications. Regardless of tube type used, the results revealed a similar amount of energy delivered to patients. The course of the disease was similar in the groups, and no differences in complications were observed.
The findings in these two studies (Davies et al, 2012; Eatock et al, 2005) are comparable with those of the findings reported in this article. This could indicate that individual considerations should determine which type of nutrition tube should be used. Tube feeding versus total parenteral nutrition (TPN) has been investigated in a small study of patients who had a liver transplant (Wicks et al, 1994). Patients in this study were randomised to receive either NJ tube or TPN feeding. There were no differences in efficacy or safety between the two methods, and the authors recommended tube feeding over TPN due to possible complications and costs.
Study limitations
This clinical study has some limitations, the main one being the group of patients involved. All participants had severe liver disease and some had end-stage liver disease. This fact challenged rigorous data collection, especially the collection of patient-reported outcomes. This group of patients may also be the most challenging when it comes to tube feeding. Therefore, the research question, ‘What feeding tube to choose?’, remains relevant to nurses caring for severely ill patients and the decision need to be evaluated on a case by case basis.
This study focused primarily on comparing the efficacy of delivering nutrition via two types of feeding tube. We did not include the use of a nasal bridle, which can help to reduce the number of tube removals (Lynch et al, 2018). As shown in Table 3, the prevalence of tube removals was similar in the two groups. However, the use of a bridle could have reduced the number of tube removals, regardless of tube type. It is doubtful whether this could have made a difference to nutritional delivery in the two groups. Applying a bridle at the outset of this study might also have raised ethical considerations.
Insufficient data were collected on the incidence of nausea and vomiting, and handgrip strength. This was unfortunate, because the data would be relevant for the topic investigated in this study. It might have been possible to collect data on patients' experience and possible side effects to the different tubes. However, it was acknowledged that this would have been difficult, because some patients had serious disease and others developed hepatic encephalopathy grade 3–4.
Conclusion
We investigated the efficacy of two types of feeding tube in delivering nutrition to patients with severe disease. We found no significant differences in total nutritional intake between the early NG feeding and early NJ feeding groups in the 7 days after tube insertion. The number of tube replacements was similar in the two groups. Choice of tubes for patients with severe liver disease should be assessed according to individual patient characteristics and needs and local facilities.