Nutritional Assessment in Inpatients With Cirrhosis Can Be Improved After Training and Is Associated With Lower Readmissions

Abstract

Malnutrition is widely prevalent in cirrhosis patients, which can worsen sarcopenia, hepatic encephalopathy (HE), and overall prognosis. We aimed to define the frequency of nutritional assessments of patients with cirrhosis in retrospective and prospective (after educational training) cohorts and to evaluate prospective changes along with their effects on 90-day readmissions. This study was conducted in 2 phases. Retrospectively, records of hospitalized patients with cirrhosis from the university and Veterans Affairs (VA) settings were reviewed to assess nutritional status, if a nutrition consultation occurred, the number of days patients were nil per os (npo) and received inadequate nutrition, and if nutritional management was guideline directed. In the prospective phase, after dedicated educational efforts directed at the stakeholders regarding nutritional guidelines for patients with cirrhosis, subsequently hospitalized cirrhosis patients had nutritional and 90-day readmission data collected for comparison between groups. In total, 279 patients were included in the retrospective phase (150 university/129 VA), and 102 VA patients were in the prospective phase. Cirrhosis severity, reason for admission, and hospital course were similar between groups regardless of cohort, ie, prospective versus retrospective or VA versus university. The prospective group had significantly more nutritional consultations and assessments (74.5% versus 40.1%; P < 0.0001) compared with the retrospective group regardless of comparisons between the VA and university cohorts. Both groups had a similar number of days npo, but the prospective group had fewer days of inadequate nutrition. The 90-day readmission rate was significantly lower in the prospective group versus the retrospective group (39% versus 28%; P = 0.04), which was associated with greater nutrition outpatient follow-up. In conclusion, nutritional consultation rates in inpatients with cirrhosis can be significantly improved after educational intervention, which is associated with lower 90-day readmission rates.

Nutritional management remains one of the most poorly understood aspects in the management of cirrhosis patients.⁽¹⁾ Although nutrition has been acknowledged as a key to the prognosis and treatment of cirrhosis, it has remained a subject with little agreement and poor standardization over the years.⁽²⁾ Malnutrition has been shown to be an independent predictor of survival in cirrhosis patients,^(3,4) whereas an increased protein intake can improve disease mortality rates.⁽⁵⁾ In addition to improved survival, nutrition also plays a key role in transplant and surgical outcomes, overall liver function, hepatic regeneration, and complication rates, such as sarcopenia and hepatic encephalopathy (HE).^(6–11) Prior studies have demonstrated that malnutrition is common in hospitalized patients with cirrhosis.⁽¹²⁾ Although no unanimous definition for nutritional screening or the assessment of nutritional risk⁽¹³⁾ exists, the International Society for Hepatic Encephalopathy and Nitrogen Metabolism (ISHEN)⁽¹⁴⁾ has recommended the Royal Free Hospital–Nutritional Prioritizing Tool (RFH-NPT) as a method to analyze those at nutritional risk.⁽¹⁵⁾

Although this tool can be used easily in routine clinical care, it is often not used to either trigger dietary consultations or initiate specific nutritional therapies. Therefore, the first step toward a potential correction of the major malnutrition epidemic in cirrhosis is its recognition, preferably by using the services of registered dieticians (RDs) in the inpatient setting. However, in clinical practice, this may not be the case, which could perhaps stem from the lack of attention and knowledge on the part of providers to the nutritional aspects of cirrhosis management in hospitalized patients. Our aim was to define the frequency and extent of the routine nutritional consultation of inpatients with cirrhosis in a retrospective manner and then to prospectively evaluate changes in this pattern after a dedicated training of the inpatient staff.

Patients and Methods

There were 2 phases in our study: the retrospective and prospective phases with an intervening period of training and education (Fig. 1).

FIG. 1.

Schematic of study progress.

RETROSPECTIVE STUDY

This retrospective analysis was performed on the patients admitted between 2015 and 2016 to Virginia Commonwealth University (VCU) and Hunter Holmes McGuire Veterans Affairs Medical Center (VAMC) in Richmond, VA. We included only patients who were admitted nonelectively with confirmed cirrhosis (by biopsy, history of decompensation, radiologic or endoscopic evidence of cirrhosis, or varices in chronic liver disease). We excluded patients whose cirrhosis status was unclear, those with human immuno-deficiency virus, or those with a prior organ transplant. The data collected included patient demographics, etiology of cirrhosis, prior complications, hospital length of stay (LOS), basal metabolic index (BMI; using estimated dry weight), Model for End-Stage Liver Disease–sodium (MELD-Na) at admission, reasons for admission (liver related or not), and information about extrahepatic organ failure. We gathered details pertaining to nutritional assessment including the type of diet on admission, days nil per os (npo), days with a dietary intake <75% of that expected per ideal body weight, and the reasons for the decreased dietary intake (complications, procedures, and so on).

We also analyzed whether a formal nutrition consultation was requested. In addition, we recorded the results of the nutritional risk screening using RFH-NPT (low, moderate, or severe nutritional deficiency); recommendations for calories, proteins, snacks, and supplements from the RDs; and whether this nutritional management was guideline directed.⁽¹⁸⁾ Days with inadequate nutrition were determined by a thorough chart review that included an assessment by an RD of food intake, a subjective assessment by the primary team, nursing notes, and patient quotes in the chart to determine if the patient was either npo or had days with a dietary intake <75% of that expected per ideal body weight. When available, RD assessments of nutrition were used as a first-line determining factor for nutritional status. Readmissions at 90 days were calculated. Comparisons were also made between patients enrolled at VCU and VAMC.

EDUCATIONAL INTERVENTION

Concerted efforts were made by the research members to perform dedicated teaching of the house staff team members, which included interns and medical residents who rotated through the medical wards (in small groups or on individual basis), the inpatient medicine consultants, and the medicine hospitalists at the VAMC in Richmond. These educational efforts were quality improvement sessions that focused on the importance of early nutrition consultations in inpatients with cirrhosis according to the ISHEN guidelines.⁽¹⁴⁾ We specifically focused the education toward RDs for the medical floors where most patients with cirrhosis were admitted. Verbal reminders were provided to all of the stakeholders at least weekly, and the educational sessions were repeated when the house staff teams or attendings switched off of service every month. Posters explaining the need for nutritional consultation were displayed in prominent places in the clinical work areas and team rooms reminding the providers about consulting the dieticians early in the hospital admission for the patients with cirrhosis.

PROSPECTIVE STUDY

After 2 months of training and educational intervention, the prospective phase of the study started as a quality improvement initiative between December 2017 and June 2018 at the VAMC only. The eligibility criteria were the same as those used in the retrospective cohort to reflect real-world circumstances. During this phase, no more specific reminders were sent to the house staff, attendings, consultants, hospitalists, and the RDs. Data collected during this phase mirrored the retrospective arm of the study. Patients were followed for readmissions over the next 90 days. Recommendations provided by RDs for diet and snacks were ordered and were fulfilled by the food service staff at our facility. RDs followed for patient compliance and tolerance to recommendations. Inadequate intake was defined as <75% of estimated needs. Calorie counts were used at the discretion of the RDs when they were concerned about inadequate intake. Intake was determined by a review of the patient recall and nursing documentation. The dry weight or estimated dry weights were used for all calculations. RDs also provided nutrition-focused physical assessments for all patients, which included assessments of muscle wasting of the temple, clavicle bone region, acromion bone region, interosseous, gastrocnemius, and quadriceps muscles as well as evaluation of fat wasting in the orbital region, eye region, triceps, biceps, and thoracic and lumbar regions. In both settings, if the RDs believed that the patient was not malnourished, they would not leave detailed protein or calorie recommendations.

RFH-NPT Score

The RFH-NPT is a quick way of calculating a nutritional risk score from low (0 points), moderate (1 point), and high (2–7 points) risk in patients with cirrhosis, and it can be administered easily, requiring no specific training (Supporting Table 1). The starting point was to ascertain if the patient had alcoholic hepatitis or if he or she was already being tube fed, each of which adds a score of 6 and would automatically categorize a patient as high risk. Following this, evidence of ascites or peripheral edema was sought (adds 1 point). Then, questions directed at nutritional status were asked both objectively (BMI) and subjectively about intercurrent illness affecting dietary intake, unplanned weight loss in the past 3–6 months, and other issues with each question having a score. Finally, the overall risk score was calculated, which directed the appropriate nutritional management. This scoring system was validated in other studies and has good reproducibility.^(14,17) The RDs used the RFH-NPT in the prospective arm to prioritize patients at a greater risk for nutritional compromise and to ensure that they were given nutritional supplement recommendations, whereas these supplements were estimated from the chart in the retrospective review and did not factor in decisions made to initiate or evaluate consultations or recommendations.

The institutional review board at both centers approved the retrospective part of the study, whereas the education and prospective analyses were considered as quality improvement projects.

Statistical Analysis

We analyzed patient characteristics by describing mean ± standard deviation (SD), median (range), and n (%), as appropriate. Data were compared between the retrospective and the prospective cohorts using parametric (t tests) and nonparametric (Mann-Whitney U, chi-square, and Fisher’s exact test) tests as appropriate, and a P value <0.05 was considered statistically significant. We also compared the retrospective and prospective patients in the VAMC cohort, and software (MINITAB Inc., State College, PA) was the statistical package used.

Sample Size

An analysis of the retrospective data demonstrated that 40% of patients sought nutrition consultations. Increasing the proportion of patients with a nutritional consultation to 60% in the prospective group (for a power of 0.80 with an alpha of 0.05) required the participation of 97 inpatients.

Results

PATIENT CHARACTERISTICS

There were 279 patients who were included in the retrospective study group (150 from VCU and 129 from VAMC), and 102 VAMC patients were studied in the prospective portion of the study. Overall the retrospective and prospective patient groups were similar with respect to the severity of cirrhosis, etiology of cirrhosis, and liver-related reasons for admission, although the retrospective cohort was significantly older (Table 1). The 129 veterans in the retrospective cohort were more likely to be men, have less HE, and were older than VCU patients (Supporting Table 2). The remaining parameters between the 2 retrospective cohorts were similar. Comparing the retrospective versus the prospective VAMC cohorts, the prospectively recruited veterans were older, had more HE and hepatocellular cancer. The remaining disease-related comparisons were statistically similar (Supporting Table 2).

TABLE 1.

Demographics, Clinical Course, Nutritional Assessment, and 90-Day Outcomes

Variables	Retrospective (n = 279)	Prospective (n = 102)	P Value
Age, years	57.66 ± 8.94	64.43 ± 10.22	<0.0001
Race			0.09
White	190	61
Black	79	32
Others	10	9
Ethnicity			0.41
Hispanic	13	7
Non-Hispanic	266	95
Prior ascites	188	66	0.76
Prior HE	108	49	0.08
Prior hepatocellular cancer	23	24	<0.001
On transplant list	28	5	0.09
Prior variceal bleeding	60	22	0.95
Admission MELD-Na	17.73 ± 8.65	17.51 ± 8.62	0.83
Liver-related reason for admission	198 (70.1)	70 (68.6)	0.65
GI bleeding	44	14
HE without infection	36	15
Altered electrolytes	16	3
Anasarca	13	3
Infections	79	29
Transplant workup	10	6
Liver-unrelated admissions	81 (29.9)	32 (31.4)
Hospital course
Extrahepatic organ failures
Mechanical ventilation	21 (8.2)	5 (4.9)	0.37
HE (grade 3/4)	44 (17.2)	17 (16.7)	0.83
Shock*	17 (6.64)	5 (4.9)	0.66
Renal replacement therapy	18 (7.0)	8 (7.8)	0.63
LOS, days	8.43 ± 8.98	5.7 ± 4.63	0.004
Nutritional details
Admission albumin, g/dL	2.75 ± 0.64	2.83 ± 0.66	0.29
Admission body mass index, kg/m2	28.48 ± 7.1	28.89 ± 6.4	0.61
Days of npo	1.15 ± 1.73	1.3 ± 2.49	0.51
Days of inadequate nutrition	3.65 ± 5.08	2.28 ± 2.13	0.009
RFH-NPT
Low/moderate risk (0–1)	109 (42.65)	46 (45.1)	0.43
High risk (2–7)	160 (57.35)	56 (54.9)	0.43
Reason for poor nutrition
HE (all grades)	60 (21.5)	25 (24.5)	0.53
GI bleeding	52 (18.6)	14 (13.7)	0.26
Procedures (emergent and nonemergent)	70 (25.1)	56 (54.9)	<0.0001
Ascites	41 (14.7)	27 (26.5)	0.008
Abdominal pain	48 (17.2)	18 (17.7)	0.92
Extrahepatic organ failures	44 (15.8)	12 (11.8)	0.33
Anorexia	38 (13.6)	31 (30.4)	0.002
Others	29 (10.4)	16 (15.7)	0.16
Formal nutrition consultations	112 (40.14)	76 (74.51)	<0.0001
Nutritional recommendations
Small frequent meals	104 (37.2)	33 (32.35)	0.38
Nighttime snack	28 (10.34)	19 (18.63)	0.02
Meal supplements (Ensure, Boost)	66 (23.66)	52 (50.98)	<0.0001
Multivitamin	164 (66)	76 (74.51)	0.005
Outcomes
In-hospital death or discharge to hospice	30 (10.75)	10 (9.80)	0.79
90-day readmission	110 (39.43)	29 (28.43)	0.04

NOTE: Data are given as mean ± SD or n (%).

^*Defined as refractory hypotension despite adequate fluid.

HOSPITAL COURSE AND OUTCOMEs

The LOS was lower in the prospective group compared with the retrospective cohort. The development of HE (grade 3/4) was similar in both cohorts. Again, both groups demonstrated statistically similar rates of extrahepatic organ failures, indicated by mechanical ventilation, shock (defined as refractory hypotension despite adequate fluids), and renal replacement therapy (Table 1). There was no significant difference in the in-hospital mortality or hospice referral between the retrospective group versus the prospective group. When comparisons were made between the retrospective VCU and Veterans Affairs (VA) cohorts, VCU patients had a higher likelihood of receiving mechanical ventilation, but the instances of death or hospice referral, number of other organ failures, and LOS were similar. When prospective versus retrospective VA cohorts were compared, there were no statistically demonstrable differences in inpatient outcomes.

NUTRITIONAL ASSESSMENT

The more common potential reasons for impaired nutrition in the prospective group were procedures, ascites, and anorexia. However, HE, gastrointestinal (GI) bleeding, abdominal pain, extrahepatic organ failures, and other causes as the reason for poor nutrition were similar between the 2 groups (Table 1). Using the RFH-NPT, both patient groups had similar numbers of low-/moderate-risk versus high-risk patients (Table 1). When comparisons were made between the retrospective VCU and VA cohorts, a higher proportion of VCU patients had RFH-NPT scores that were high risk with higher rates of GI bleeding. Nutritional consultation rates were similar between the retrospective VCU and VA cohorts, but veterans were more likely to receive nighttime snack advice and less likely to receive multivitamins (Supporting Table 2). The prospective VA cohort was more likely to receive a nutrition consultation compared with the retrospective cohort, which resulted in greater multivitamin and meal supplement prescription. The prospective cohort also experienced more procedures and anorexia as potential causes of poor nutrition compared with the retrospective VA cohort.

IMPACT OF EDUCATIONAL INTERVENTION ON NUTRITIONAL ASSESSMENTS

Through our educational intervention, the prospective group had received significantly more nutrition consultations by RDs (74.5% versus 40.1%; P < 0.0001; Table 1; Fig. 2A). Of the patients evaluated by the RDs, the 2 groups had similar rates of patients with none-to-mild malnutrition versus moderate-to-severe malnutrition (Fig. 2B,C). The existing staffing was adequate to perform these consultations.

FIG. 2.

A comparison of patients who received nutrition consultations, advice, nutritional recommendations, and evaluation in the retrospective and prospective arms of the study. (A) Proportion receiving nutrition consultations. (B) Proportion assessed as malnourished among all patients. (C) Proportion assessed as malnourished among those who received a nutrition consultation. Comparisons performed using chi-square tests and P values are displayed.

INTERVENTIONS PRESCRIBED BY THE RDs

Overall, in the subgroup of patients who had a nutritional consultation, the relative proportion receiving a formal caloric and protein recommendation was similar in the retrospective group compared with the prospective group (Table 2). The calorie recommendations of <25 kcal was higher and recommendations of 25–30 kcal/g and 20–25 kcal/g were lower in the retrospective group. The recommendation for 30–35 kcal/g was similar between groups as were the protein recommendations (Table 2). Both groups had similar number of days npo, but the prospective group had fewer days of inadequate nutrition, which was statistically significant (Table 1). Both groups were prescribed small frequent meals at similar rates, but the prospective group was significantly more likely to be prescribed nighttime snacks, meal supplements, and multivitamins as compared with the retrospective group after intervention (Table 1). Of those patients who got a nutritional assessment, VCU patients were less likely to receive protein or caloric intake advice (Supporting Table 3) compared with veterans in the retrospective arm. Of those who received a nutritional assessment, when prospective versus retrospective VA cohorts were compared, a lower proportion of prospectively recruited veterans received protein and caloric advice.

TABLE 2.

Calorie, Protein, and Supplement Recommendations as per RDs of Patients Receiving a Formal Consultation

Variables	Retrospective (n = 112)	Prospective (n = 76)	P Value
Calorie recommendation	83 (74)	55 (74)	0.52
20–25 kcal/g	9 (8)	0 (0)	0.01
25–30 kcal/g	30 (27)	12 (16)	0.07
30–35 kcal/g	40 (36)	38 (50)	0.05
>35 kcal/g	4 (4)	5 (7)	0.48
Protein recommendation	88 (79)	56 (74)	0.47
<1.2 g/kg	9 (8)	3 (4)	0.37
1.2–1.5 g/kg	74 (66)	48 (63)	0.68
>1.5 g/kg	5 (5)	5 (7)	0.53

NOTE: Data are given as n (%).

READMISSION DETAILS

The prospective group had a significantly lower risk of 90-day readmissions as compared with the retrospective group. However, the reasons for readmission were likely to be liver related in both groups and were similar in pattern (Table 3). The reasons for readmission were similar regardless of being in the VCU or VA cohorts or in the prospective versus retrospective VA cohorts (Supporting Table 4).

TABLE 3.

Details of 90-Day Readmissions

	Retrospective (n = 110)	Prospective (n = 29)	P Value
Number of patients seen by GI/hepatology departments	73 (66)	15 (51)	0.14
Number of patients seen by outpatient nutrition service	6 (5)	3 (10)	0.42
Number of patients seen by other doctors or services	73 (66)	22 (76)	0.64
Readmission reasons
Liver-related readmissions*	72 (65)	15 (51)	0.17
GI bleeding	11	4
HE without infection	16	2
Altered electrolytes	7	1
Anasarca	16	3
Infections	12	4
Transplant workup	7	1
Liver-unrelated readmissions	38 (35)	14 (49)

NOTE: Data are given as n (%).

^*Adds to more than total since some readmissions were due to multiple reasons related to the liver.

In patients who were readmitted, the number that was seen by GI/hepatology departments or by other specialties was statistically similar between the groups (Table 3). This similarity continued when comparisons were made between retrospective VCU/VA and retrospective versus prospective VA cohorts (Supporting Table 4).

In the prospective cohort, 47 patients were evaluated by the outpatient dietary service, of whom only 3 required readmission. Of the remaining 55 patients in the prospective group who were not seen by the RD, 26 (47%) required readmission (P < 0.0001).

Discussion

Assessing and treating malnutrition is an essential component of cirrhosis management, especially in hospitalized patients who are prone to infections with long periods of inadequate or no nutrition due to procedures (such as endoscopies), investigations, HE, and other complications. Poor attention toward nutritional needs has the potential to further worsen a patient’s inherent sarcopenia, HE, and overall prognosis. Hence, a key component of the guidelines for end-stage liver disease is focused on adequate nutritional support.

The goal of this project was to show that through simple educational intervention, nutrition assessments of inpatient patients with cirrhosis could be improved. We were successful in accomplishing this goal with a significant increase in the number of nutritional consultations, which resulted in RDs performing and documenting timely nutrition assessments along with making recommendations for calorie and protein intake. Not only did our study demonstrate that improving nutritional assessments for the hospitalized patients was possible but that there was a stronger adherence to guideline-directed nutritional recommendations as well. In the prospective group, patients were more likely to get recommendations for nighttime snacks, meal supplements, and multivitamins. After the intervention, more patients were also recommended the correct calorie recommendation per the assessment by an RD based on ISHEN guidelines.⁽⁸⁾ The number of days of inadequate nutrition were also improved after the intervention. Importantly, the complications of cirrhosis and the reasons for admission and the in-hospital course were largely similar between the cohorts.

The American Association for the Study of Liver Diseases recommends when evaluating adult patients for liver transplant nutrition assessment by a skilled nutrition specialist.⁽¹⁸⁾ Although most of these patients were not undergoing a liver transplant workup, using the opportunity for a trained RD to evaluate them while they were inpatients is a great intervention for a group with a known high level of malnutrition.⁽¹⁹⁾ Very few quality improvement projects have focused on nutritional assessment and improving nutrition in patients with cirrhosis. Recent guidelines by ISHEN and the European Association for the Study of the Liver along with recent reviews have highlighted the importance of optimum nutrition in patients with cirrhosis, which starts with a dedicated nutritional assessment.^(1,20) The nutritional risk stratification of these fragile patients can form the first step for subsequent interventions aimed at improving nutrition and sarcopenia.

This intervention can easily be initiated in the hospitalized setting where usually there are more resources (like RDs), and the efforts can be continued in the outpatient setting. We have been able to show in our quality improvement that by using a simple educational intervention, the rate of nutritional assessment of inpatient patients with cirrhosis could be improved. The training sessions culminated in largely accomplishing this goal, with a significant increase in the number of nutrition consultations and the performing and documenting of nutrition assessments by RDs. The timely involvement of the RDs helps in generating the right kind of prescription for dietary and nutritional improvements, which are directed by cirrhosis guidelines. Indeed, one of the barriers is the lack of communication between inpatient teams and the RDs rather than a lack of knowledge of the nutritional guidelines at the dietician level. Exploring the reasons behind the low rate of consultation is difficult given the retrospective nature of the group in the first part of the study. However, because the proportion of consultations made and acted upon increased significantly after the educational intervention, it is likely that nutrition is considered to be a relatively low priority when the inpatient team is faced with challenges such as infections, HE, and GI bleeding, which are the major reasons for hospitalization of these patients. Therefore, the decision to initiate a nutrition consultation is largely subjective and prone to being missed. However, these biases existed in the retrospective and prospective arms, which we hoped to correct with specific training.

We chose to use retrospective experiences from 2 separate transplant centers, VCU and VAMC, to inform future quality improvement. We focused the prospective arm at the VAMC because of its relatively more compact inpatient house staff, ward location, and nutrition service. The prospective and the retrospective VA cohort comparisons and the retrospective VCU and VA cohort comparisons were largely similar in terms of clinical severity, reasons for admission, and hospital course. The differences in nutritional consultation and readmissions remained despite the individual comparisons.

Indeed, the study showed that a simple, low-cost educational intervention could significantly improve the rate of these consultations. However, the success of these interventions may hinge on the creation of multidisciplinary teams for taking care of patients with advanced liver disease. This is relevant for all aspects of care for the patients in the current scenario, but unfortunately, these are lacking in complications other than the hepatocellular carcinoma.⁽²¹⁾ Prior studies on the other aspects of cirrhosis complications, such as ascites, have demonstrated that fostering conditions that allow for guidelines to be followed can save costs and improve outcomes.⁽²²⁾

Therefore, achieving a continuation of dietary interventions will require close cooperation between hospital attending physicians, consultants, gastroenterology/hepatology practitioners, house staff, and dieticians. Although the overall goal is to improve outcomes, such as the number of readmissions and falls and the amount of morbidity and mortality, through enhancing nutrition, a dedicated nutritional evaluation is likely the first step toward achieving this goal. The care providers should be actively seeking out these patients, ideally as early as at admission to the hospital. One possibility is to incorporate a nutritional risk stratification tool in the electronic chart system so that an automatic referral can be generated to the RDs. Although the inpatient mortality or hospice referral rates were similar between groups, we found a significant reduction in 90-day readmissions in the prospective cohort compared with the retrospective cohort. There are multiple reasons behind readmissions, some of which can be modulated by active interventions compared with others.^(23,24) Although the rate of follow-up visits between those who got readmitted in the prospective versus retrospective arms was similar, significantly fewer patients in the prospective arm who saw an RD were readmitted compared with those who did not see an RD. Given that the admission disease severity, liver-related reasons for admission, and readmission and organ failure rates were similar between the groups, we can speculate that nutritional assessment and interventions could have contributed toward this reduction.

It is difficult to compare the success of our study to others because of the limited amount of trials of this nature. A Ugandan study showed an improvement from 0% to 79% in a 2-month span for nutritional screenings of pregnant mothers.⁽²⁵⁾ Again, there are few studies to compare an intervention like ours to any other, but some studies have attempted to improve calorie and protein intake in inpatient groups with varying amounts of success.^(26,27)

Our study is limited by the scope of the patient population and by not having measurable caloric and protein consumption of patients. Although the RDs did a focused examination based on musculature, we did not perform a formal assessment of sarcopenia. There are also biases when calculating RFH-NPT from a retrospective source, and the reasons that some patients received a consultation could not be reliably explored retrospectively. The prospective arm was only performed at the VAMC, whereas the retrospective arms were in both the VA and university settings. However, both institutions are liver transplant centers, and apart from sex, the patient mix was largely similar in terms of admission, hospitalization, and readmission details. The improvement between prospective and retrospective arms persisted even when the retrospective group was divided into VA and VCU cohorts. Because we only tested the improvement to nutrition in the VA setting after educational intervention in older male patients, the generalizability of these findings to other populations needs to be investigated. We also focused on the recommendations given rather than the actual implementation of those recommendations. Regardless, adequate nutritional assessment by dedicated professionals is indeed the first step toward optimizing care in malnourished inpatients with cirrhosis. We focused on RD consultations and follow-up only for patients who were readmitted. Although a significant effort has been undertaken to improve guidelines and recommendations for this population, larger prospective trials are needed to evaluate different models and strategies for accomplishing those recommendations.

We conclude that the nutritional consultation rates in inpatients with cirrhosis can be significantly improved after dedicated educational intervention with the inpatient teams to provide guideline-based nutritional interventions to affected individuals. Future studies are needed to put these results in the context of longer-term outcomes based on nutritional improvement.

Abbreviations:

BMI

basal metabolic index

GI

gastrointestinal

HE

hepatic encephalopathy

ISHEN

International Society for Hepatic Encephalopathy and Nitrogen Metabolism

LOS

length of stay

MELD-Na

Model for End-Stage Liver Disease–sodium

npo

nil per os

RD

registered dietician

RFH-NPT

Royal Free Hospital–Nutritional Prioritizing Tool

SD

standard deviation

VA

Veterans Affairs

VAMC

Veterans Affairs Medical Center

VCU

Virginia Commonwealth University