Address reprint requests to Samer Gawrieh, M.D., Division of Gastroenterology and Hepatology, Medical College of Wisconsin, Milwaukee, WI 53226. Telephone: 414-955-6850; FAX: 414-955-6214; E-mail: firstname.lastname@example.org
Malnutrition is commonly encountered in patients with end-stage liver disease (ESLD) and has been reported in up to 80% of all patients with cirrhosis and in up to 25% of patients with Child-Pugh class A cirrhosis.[1, 2] It has also been associated with higher morbidity and mortality.[1, 2] In fact, the original Child-Turcotte classification included nutritional status as a prognostic parameter. Cirrhosis is associated with a hypermetabolic state, increased protein catabolism, decreased glycogen storage and glucose oxidation, and increased lipid oxidation, all of which contribute to a poor nutritional status, but the precise etiology of malnutrition in patients with cirrhosis is not established.
In patients with cirrhosis and chronic cholestasis, there is inadequate delivery of bile salts into the intestinal lumen. This can lead to insufficient absorption and a fat-soluble vitamin deficiency. For example, vitamin A deficiency has been observed in patients with primary biliary cirrhosis (PBC) and may be associated with night blindness.[6, 7] Several studies have also shown vitamin D deficiency in both cholestatic and noncholestatic liver disorders.[8-10] It is, however, not clear whether such vitamin deficiencies are due to poor nutritional intake, hepatic dysfunction, malabsorption, or a combination of these factors. There are few data on the prevalence of fat-soluble vitamin deficiencies among patients with ESLD awaiting liver transplantation. The aims of this study were to assess the prevalence of fat-soluble vitamin deficiencies in patients being evaluated for liver transplantation and to elucidate the predictive factors for any such deficiencies.
PATIENTS AND METHODS
We retrospectively reviewed the charts of patients who presented for an outpatient evaluation for liver transplantation at the Medical College of Wisconsin Hepatology Clinic from January 2008 to September 2011. The study protocol was reviewed and approved by the institutional review board of the Medical College of Wisconsin. The etiology of cirrhosis, Model for End-Stage Liver Disease (MELD) score, Child-Pugh score, body mass index (BMI), and vitamin A, vitamin E, and vitamin 25-OH-D levels were recorded. Patient demographics, including age, sex, smoking history, alcohol use, and drug use, were reviewed.
Serum vitamin A and E levels were determined via high-pressure liquid chromatography with ultraviolet detection (Agilent Technologies, Santa Clara, CA). Serum vitamin D 25-OH levels were determined with a radioimmunoassay (DiaSorin, Stillwater, MN). Vitamin A, D, and E deficiencies were defined as levels < 19 μg/dL, < 32 ng/mL, and < 3 mg/L, respectively, and were based on normal ranges used at our institution. A severe vitamin D deficiency was defined as a level < 10 ng/mL.
Patients were excluded if they had short gut syndrome, celiac disease, pancreatic insufficiency, a previous liver transplant, or an incomplete laboratory analysis.
Univariate comparisons of patient characteristics between vitamin-deficient and nondeficient subjects were performed with the Student t test for continuous variables (age, BMI, and MELD score), with the chi-square test for binary variables [sex; race; alanine transaminase, aspartate transaminase, bilirubin, albumin, and alkaline phosphate levels; and contributions of alcohol, nonalcoholic steatohepatitis (NASH), hepatitis C virus (HCV), and primary sclerosing cholangitis (PSC) to the disease etiology], and with the Wilcoxon rank-sum test for ordinal variables (Child-Pugh score). The exact chi-square test was used when the expected number of subjects in a category was 5 or fewer. Similar methods were used for the comparison of excluded and included subjects.
Logistic regression was used for the multivariate analysis of predictors of vitamin deficiencies. Due to the large number of potential predictors, model building was performed by forward stepwise selection with a 10% significance level for entry and a 5% significance level for staying in the model. The predictive power of the final model for vitamin A deficiency was evaluated with a receiver operating characteristic (ROC) analysis.
All analyses were performed with SAS 9.2 (SAS Institute, Cary, NC).
Sixty-three of the 211 patients referred for liver transplantation met the inclusion criteria. Incomplete laboratory data (139 patients) were the most common reason for exclusion, and this was due to the varying adherence of different providers in our practice to screening and was not based on a clinical suspicion of vitamin deficiency. Subjects' characteristics are summarized in Table 1. There were no statistically significant differences in age, race, sex, BMI, etiology of liver disease, Child-Pugh class, MELD score, alcohol use, or smoking history between the included and excluded subjects. The most common etiologies of chronic liver disease were alcohol abuse (n = 23) and HCV (n = 19; Table 2). The remaining patients had HCV and alcohol (n = 5), NASH (n = 5), PSC (n = 4), PBC (n = 2), NASH and PBC (n = 1), autoimmune hepatitis (AIH; n = 2), hepatic carcinoid tumor (n = 1), or AIH and PSC (n = 1). None of the included patients had acute liver disease or failure. The majority of the patients were deficient in vitamin A and vitamin D (69.8% and 81.0%, respectively); however, only 2 of the 63 patients (3.2%) were vitamin E–deficient, and both had PSC. More than half of the patients were deficient in both vitamin A and vitamin D (36/63 or 57.1%). A severe vitamin D deficiency (<10 ng/mL) was found in only 4 patients (6.3%). When 20 ng/mL was used as the lower limit of normal for defining vitamin D deficiency, 47 of 63 patients (75%) were deficient. The one patient who was underweight but had normal fat-soluble vitamin levels had HCV cirrhosis.
Table 1. Vitamin Deficiencies by Patient Characteristics
Total (n = 63)
Vitamin A (n = 63)
Vitamin D (n = 63)
Vitamin E (n = 63)
Deficient: <19 μg/dL (n = 44 or 69.8%)
Normal: 19-83 μg/dL (n = 19 or 30.2%)
Deficient: <32 ng/mL (n = 51 or 81.0%)
Normal: 32-100 ng/mL (n = 12 or 19.0%)
Deficient: <3 mg/L (n = 2 or 3.2%)
Normal: 3-15 mg/L (n = 61 or 96.8%)
NOTE: All percentages are based on the total group value (n = 63).
The data are presented as means and standard deviations.
There were no documented cases of night blindness in our cohort. Vertebral fractures were documented in 4 patients, all of whom were vitamin D–deficient. Bone density data, as measured by dual-energy X-ray absorptiometry scanning, were available for 55 patients. Osteopenia was noted in 25 patients, and osteoporosis was seen in 10 patients, whereas normal bone density was noted in 20 patients. In this group with bone density data, vitamin D deficiency was noted in 18 of the 25 patients (72%) with osteopenia, in 10 of the 10 patients (100%) with osteoporosis, and in 16 of the 20 patients (80%) with normal bone density. Vertebral fractures were documented in 4 patients, all of whom had vitamin D deficiency and osteopenia on a dual-energy X-ray absorptiometry scan. One of our patients who had vitamin E deficiency before transplantation had a reperfusion injury postoperatively.
Most patients were not on any vitamin supplementation (40/63 or 63.5%) before their assessment at the hepatology clinic. Seventeen patients (27.0%) were taking a multivitamin supplement. However, 13 of these 17 patients (76.5%) were vitamin D–deficient, 12 (70.6%) were vitamin A–deficient, and 3 (17.6%) had no vitamin deficiencies. The amounts of fat-soluble vitamins in the multivitamin pills and patients' compliance with taking them could not be determined from the chart review. Only 6 patients were taking vitamin D (5 of whom were on >1000 IU/day dosing), and 2 patients were taking vitamin E supplementation. Compliance with taking these supplements could not be verified. In the limited number of patients in this cohort who had follow-up data, oral supplementation with vitamins A and D resulted in the correction of vitamin A and D deficiencies in 3 of 5 patients (60%) and 12 of 16 of patients (75%), respectively.
All patients with Child-Pugh class C cirrhosis (n = 11) were deficient in vitamin A. In a multivariate analysis, there were no statistically significant predictors for vitamin D deficiency. The Child-Pugh class [odds ratio (OR) = 6.84, 95% confidence interval (CI) = 1.52-30.86, P = 0.01], an elevated total bilirubin level (OR = 44.23, 95% CI = 5.02-389.41, P < 0.001), and an elevated BMI (OR = 1.17, 95% CI = 1.00-1.36, P = 0.045) were found to be predictors of vitamin A deficiency (Table 3). With the full model (including the Child-Pugh class, bilirubin level, and BMI), the area under the ROC curve for this model was 0.9139 (95% CI = 0.84-0.98), and it was higher than the areas of models including the Child-Pugh class and bilirubin level (0.8714, 95% CI = 0.77-0.96) or the bilirubin level alone (0.7817, 95% CI = 0.66-0.89; Fig. 1).
Table 3. Multivariate Analysis of Predictors of Vitamin A Deficiency: Stepwise Logistic Regression Results
OR (95% CI)
Total bilirubin level
The majority of ESLD patients evaluated for liver transplantation at our center were deficient in vitamin A and vitamin D. The Child-Pugh class, the serum bilirubin level, and an elevated BMI were predictors of vitamin A deficiency. There were no predictors for vitamin D deficiency. Interestingly, the etiology of liver disease was not predictive of a fat-soluble vitamin deficiency.
Malnutrition has been established as a negative prognostic indicator in patients with ESLD.[12-18] Previous studies have evaluated fat-soluble vitamin deficiencies in patients with chronic cholestatic liver disorders, in which impaired bile flow is the proposed underlying mechanism leading to malabsorption of fat-soluble vitamins. Few studies have evaluated this in patients with noncholestatic liver diseases, and most have focused on vitamin D deficiency.
Fisher and Fisher evaluated 100 consecutive patients with noncholestatic liver disease for vitamin D deficiency. The most common etiologies of liver disease were alcohol (n = 40) and HCV (n = 38). Vitamin D deficiency was found in 68 patients and was more prevalent in patients with Child-Pugh class C cirrhosis. Malham et al. sought to compare vitamin D deficiency in patients with noncholestatic liver disease and patients with cholestatic liver disease. In that study, 89 patients with alcoholic cirrhosis and 34 patients with PBC were retrospectively evaluated for vitamin D deficiency. Vitamin D deficiency was more prevalent in patients with alcoholic cirrhosis versus patients with PBC (85% versus 60%). Arteh et al. evaluated vitamin D deficiency in 118 consecutive patients with HCV. They found that 109 of the 118 patients (92%) had vitamin D deficiency, with severe vitamin deficiency (<7 ng/mL) more common in patients with cirrhosis. A high incidence of vitamin A and D deficiencies was also reported in a recent study by Abbott-Johnson et al. In that study, 107 patients with ESLD who were awaiting liver transplantation were prospectively followed; 75% of the patients were deficient in vitamin A, 66% were deficient in vitamin D, and only 3% were deficient in vitamin E. Similarly to our findings, the severity of the underlying liver disease, as reflected by the Child-Pugh class, was found to be a predictor of fat-soluble vitamin deficiencies.
In agreement with prior studies, the majority of our patients with ESLD who were evaluated for liver transplantation were found to be deficient in both vitamin A (69.8%) and vitamin D (81%), but the proportion of patients with vitamin E deficiency was low (3.2%). The MELD score and the presence or absence of cholestatic liver disease were not found to be predictive of vitamin deficiencies in our patients.
A fat-soluble vitamin deficiency may have significant clinical consequences. Osteoporosis is a potential complication after liver transplantation. By causing secondary hyperparathyroidism, vitamin D deficiency before liver transplantation may increase the risk for osteoporosis and fractures. There was a high rate of osteoporosis and osteopenia in our patients, and 4 subjects (6.3%) had vertebral fractures. In a recent study, Kaemmerer et al. evaluated the impact of ibandronate, vitamin D3, and calcium supplements before and after liver transplantation on 31 patients with osteopenia or osteoporosis before transplantation. This regimen increased bone mineral density in the lumbar spine up to 12 months after transplantation, and after 24 months, there were only 2 reported cases of fractures within this group. Although there were no documented cases of night blindness in our patients with vitamin A deficiency, none of them underwent a formal ophthalmological evaluation. A study by Abbott-Johnson et al. evaluated vitamin A deficiency and night blindness in 8 patients with cirrhosis awaiting liver transplantation. Interestingly, 6 of the 8 patients were unaware of impairments in their vision. All the patients showed improvements with dark adaptation after intramuscular supplementation with vitamin A. One of our 2 patients with vitamin E deficiency had a reperfusion injury after liver transplantation. Interestingly, Goode et al. demonstrated an increased risk of reperfusion injury after liver transplantation in patients with vitamin E deficiency before transplantation. We found no predictors for vitamin D deficiency, and this was contrary to previous studies that had identified cirrhosis, female sex, and African American race as predictors for vitamin D deficiency.[10, 22] Dietary insufficiency, malabsorption, impaired hepatic hydroxylation of vitamin D, and a lack of sunlight exposure may contribute to vitamin D deficiency. Although no predictors for vitamin A deficiency were identified in prior reports, an advanced Child-Pugh class, the total bilirubin level, and an elevated BMI were associated with it in our cohort. It is unclear why Child-Pugh class C patients had a lower frequency of vitamin D deficiency and a higher frequency of vitamin A deficiency in comparison with Child-Pugh class A and B patients.
Our study has several limitations. We were unable to find data in the reviewed medical records concerning the adequacy of oral intake or other parameters for assessing nutritional status such as pre-albumin, triceps skinfold thickness, and midarm muscle circumference. Our patient population is located in Wisconsin (at a latitude of 43° N), and this may affect the prevalence of vitamin D deficiency because the reduced exposure to sunlight may contribute to lower vitamin D levels. Vitamin E deficiency was not evident in our patient population. There are conflicting studies on whether the serum measurement of α-tocopherol or the α-tocopherol/total lipid ratio is a more accurate measure of vitamin E in patients with chronic liver disease.[23, 24] Serum lipid panels from our patient population were not analyzed, and we did not calculate this ratio to elucidate whether more patients had vitamin E deficiency. Our study was designed not to evaluate the mechanism for these deficiencies but instead to identify whether these patients had significant deficiencies in fat-soluble vitamins. Although an elevated BMI was identified as a predictor of vitamin A deficiency, our BMI calculation was not adjusted for fluid retention. Finally, the exact composition of each patient's multivitamin was unknown; therefore, it is unclear how much impact, if any, these supplements may have had on serum values of fat-soluble vitamins.
If our findings are validated in other studies, universal screening for vitamins A and D in this population of patients being evaluated for liver transplantation may be reasonable because of the high prevalence of these deficiencies. Whether vitamin E deficiency is limited only to patients with PSC also needs to be examined in order to make recommendations on appropriate screening. Whether using a daily multivitamin supplement will reduce these deficiencies has not been prospectively studied; however, it did not seem to affect vitamin A and D deficiencies in a subgroup of our patients.
In summary, our study confirms that deficiencies of vitamins A and D in patients with ESLD awaiting liver transplantation are very common. The etiology of liver disease did not predict vitamin deficiencies; however, the severity of liver disease affected vitamin A deficiency. The identification and correction of these common deficiencies may be important elements in efforts to optimize this highly morbid patient population before liver transplantation.