Because of a number of advances in the field of liver transplantation (LT), the length of posttransplant survival and the number of long-term survivors are steadily increasing. The focus of concern has thus now been shifted to long-term complications, such as chronic renal failure, osteoporosis, malignancies, and cardiovascular disease. The effects of such complications on quality of life may, however, be perceived differently by individual patients1–3 and at different posttransplant time points. LT patients are also faced with life-long immunosuppressive medication and follow-up visits, which may further influence their quality of life. Traditional medical outcome variables, such as survival rates and specific morbidity rates, are important, but they can be considered insufficient in the evaluation of general outcome and long-term health status after LT.
Health-related quality of life (HRQoL), reflecting the physical, psychological, and emotional dimensions of health, is considered a comprehensive and important measure of outcome.3–6
Deteriorated HRQoL has been associated with increased health resource use and costs, and, consequently, a need for future studies to identify subgroups of LT patients at particular risk of impaired HRQoL and reduced working capacity after LT has been pointed out.6–11 This is important to the challenge of developing purposeful intervention programs aimed at facilitating recovery after transplantation and maximizing favorable long-term outcomes after LT while minimizing care-related costs.1, 6–8, 11, 12 Previous studies in the field have been limited by small cohorts or short follow-up periods and have mostly not compared results with an age-standardized and gender-standardized general population. They also differ in the instruments and methods used and have often not included subgroup analyses.1, 6–8, 13–17
We conducted a cross-sectional study on posttransplant HRQoL among all LT patients alive in Finland, using the validated 15D questionnaire and comparing the results with those of an age-standardized and gender-standardized general population. Furthermore, we studied whether HRQoL and subjective working capacity differed between genders, age groups, or transplant indication groups or at different time points of posttransplant follow-up.
Δmean, difference in means between groups; ALF, acute liver failure; CI, confidence interval; CLD, chronic liver disease; HRQoL, health-related quality of life; LT, liver transplantation; SD, standard deviation.
PATIENTS AND METHODS
From our population of LT patients transplanted between 1982 and May 31, 2007 (674 patients), we identified all living patients 18 years old and older (401 patients). All patients in Finland are transplanted at the Helsinki University Central Hospital, to which they also return regularly for follow-up visits. Clinical and demographic data were collected from the Finnish LT registry and from patient records when necessary. Survival times were from first transplant to May 31, 2007. One-, 3-, 5-, and 10-year patient survival rates were 90%, 86%, 82%, and 72%, respectively. One-, 3-, 5-, and 10-year graft survival rates were 85%, 80%, 75%, and 65%, respectively.
HRQoL was assessed with the 15D questionnaire, which was mailed for self-administration to all patients alive in June 2007 along with a prepaid envelope for returning it. Questionnaires were in each patient's native language, Finnish or Swedish, which are the 2 official languages of Finland. The letter also included a questionnaire for the assessment of employment status and a consent form that the patients were asked to sign and return. A reminder was sent in July 2007 to those who did not respond to the first letter. Strict confidentiality was ensured, and in order to obtain truthful responses, it was stressed to the patients that their transplant physicians would not have access to answers from any individual patient. Letters were returned to a department unassociated with the transplantation center (Helsinki and Uusimaa Hospital Group, Group Administration), and responses were registered by persons uninvolved in analyses of the results.
HRQoL was measured by the 15D instrument.18–20 The 15D is a generic, 15-dimensional, standardized, and self-administered measure of HRQoL that can be used both as a profile and as a single-index score measure.20, 21
For each of the 15 dimensions, the respondent must choose from 1 of 5 levels that best describes his or her present health status (the best level being 1 and the worst level being 5). The valuation system of the 15D is based on an application of the multiattribute utility theory. A set of utility or preference weights, obtained from the general public through a 3-stage valuation procedure, is used to generate the utility score, that is, the 15D score (single index number), over all the dimensions and the dimension level values on a 0 to 1 scale (1 = no problems in any dimension, 0 = being dead).21 The minimal clinically important difference in the 15D score is ≥0.03.18 No corresponding difference has been established for separate dimensions. In most of the important properties, the 15D compares favorably with other similar instruments.20, 22, 23
If a respondent left up to 3 questions unanswered, the missing data were imputed by regression models with the responses on the other dimensions, as well as age and gender as explanatory variables, according to the 15D instructions.21
The 15D data for the general population came from the National Health 2000 Health Examination Surveys, which represent the Finnish population.24 For this analysis, those individuals who were in the age range of the LT patients were selected (n = 6050). This sample of the general population was weighted to reflect the age and gender distribution of the patients.
Employment Status and Working Capacity
Employment was defined as engagement in such work from which a person receives economic and taxable income. The typical age for old-age retirement in Finland is 65 years. Persons in the age range of 16 to 64 years who as the result of a persisting illness are unable to engage in gainful employment may be granted a disability pension according to Finnish legislation, which results in what is called “early retirement.”
A categorical and descriptive evaluation of patients' current employment status and subjective impact of LT on their working capacity was assessed with 5 questions:
1Are you currently employed (yes or no)?
2If the answer to the previous question was “no,” then choose one of following alternatives:
cEarly retirement for a cause associated with my liver disease.
dEarly retirement for a cause not associated with my liver disease.
eOther cause (what?).
3How soon after your transplant were you able to return to work?
aLess than 3 months.
fI have not returned.
4If the answer to the previous question was option f, then choose one of the following alternatives:
aAge—retirement before transplantation.
bAge—retirement shortly after transplantation.
cStudent before transplantation.
dUnemployed before transplantation.
eEarly retirement for a cause associated with my liver disease.
fEarly retirement for a cause not associated with my liver disease.
gOther cause (what?).
5How do you estimate your current working capacity and functional capacity in everyday life compared to that at 1 week before transplantation? (the respondent must choose 1 of 5 options on a bipolar scale, where a is “much better” and e is “much worse”).
Ethical approval for the study was obtained from the ethical committee of the Helsinki and Uusimaa Hospital district. All patients were asked to sign an informed consent form.
Statistical analyses were performed with SPSS statistical software, version 14.0 (SPSS, Inc., Chicago, IL). Differences in the mean 15D score between groups were tested with the independent sample t test or 1-way analysis of variance, as appropriate. The chi-square test was used for categorical variables. The difference in means between groups (Δmean) and the 95% confidence intervals (CIs) for this difference are reported for the main results.
Subgroup analyses were performed by age, gender, survival time, transplant number, and employment status. These subgroups were considered in separate linear regression models, with the 15D score as the dependent variable and age, gender, and the respective subgroup variable as independent variables. It was also tested separately whether there is a linear or an inverted U-shape relationship between HRQoL and survival time. The latter was tested by the inclusion of both survival time and its square as explanatory variables in the respective linear regression model.
The LT population was further combined with the age-standardized and gender-standardized sample of the general population in order to test whether the HRQoLs of different transplant indication groups [chronic liver disease (CLD), acute liver failure (ALF), or liver tumor] deviate from that of the general population. This was tested by the computation of a linear regression model adjusted for current age and gender, in which each of the 3 LT indication groups was coded with a different indicator variable and the general population was held as a reference. P values < 0.05 were considered statistically significant from one another. However, to test whether there existed significant differences between LT patients and the general population on the individual 15D dimensions according to the independent sample t test and to adjust for multiple testing, we used the Bonferroni correction. Here a P value < 0.003 (0.05/15 = 0.003) was considered statistically significant.
Three hundred fifty-six of 401 patients returned the questionnaire, and this made the response rate 88.5%. Three patients were excluded from analyses because they left more than 3 questions unanswered in the 15D questionnaire. For another 23 cases, 1 to 3 missing data were imputed [in 2 cases in 3 domains, in 2 cases in 2 domains, and in 19 cases in only 1 domain (most often sexual activity)]. The characteristics of the remaining 353 patients included in the final analyses are presented in Table 1. The nonresponders (45 patients) were significantly more often children at the time of LT (13% of patients < 20 years among nonresponders versus 5% among responders, respectively; P = 0.034, chi-square test), but otherwise the groups did not differ significantly.
Table 1. Characteristics of the Patients Included in the Analysis According to Age Group
Age at LT
Gender, n (%)
Age at May 2007, mean (range)
Age at liver transplantation, mean (range)
Follow-up in years, mean (range)
Liver transplant indication group, n (%)
Chronic liver disease
Acute liver failure
Time period of liver transplantation, n (%)
Retransplantation, n (%)
HRQoL in Comparison with the General Population
The mean 15D score of LT patients [0.889; standard deviation (SD), 0.103] was slightly but statistically significantly lower than that of the age-standardized and gender-standardized general population (0.907; SD, 0.089; Δmean between groups, 0.018; 95% CI, 0.007-0.029; P < 0.002; Fig. 1). In light of the mean dimension level values, the LT patients were statistically significantly worse off on the dimensions of moving (Δmean, 0.040; 95% CI, 0.022-0.058), sleeping (Δmean, 0.029; 95% CI, 0.010-0.048), elimination (ie, urination and defecation; Δmean, 0.048; 95% CI, 0.027-0.069), usual activities (ie, keeping up with work, studies, household activities, and leisure activities; Δmean, 0.066; 95% CI, 0.044-0.088), and sexual activity (Δmean, 0.045; 95% CI, 0.020-0.070) but better off on the dimension of discomfort and symptoms (Δmean, 0.037; 95% CI, 0.014-0.059). The groups did not differ significantly on the other dimensions.
HRQoL in Subgroups
Regression analyses showed no statistically significant difference in the 15D scores between genders when they were adjusted for current age (regression coefficient for male gender, −0.001; 95% CI, −0.023 to 0.021). There was a small but statistically significant decline in 15D scores along with increasing current age (regression coefficient, −0.001; 95% CI, −0.002 to 0.000) when they were adjusted for gender. Age at the time of LT, on the other hand, did not influence 15D scores significantly when they were adjusted for current age and gender (regression coefficient, −0.000; 95% CI, −0.002 to 0.003). Retransplantation likewise did not alter 15D scores in a statistically significant way when they were adjusted for current age and gender (regression coefficient, −0.029; 95% CI, −0.073 to 0.014).
Figure 2 shows the mean 15D score according to the follow-up period in which patients were at the time of this study. HRQoL was worse during the first posttransplant year than at later times, and then it peaked to its best at 1 to 5 years post-transplant, after which a gradual decrease in HRQoL was seen with time. With current age and gender standardized, a slightly inverted U-shape relationship between HRQoL and survival time after LT remained, but the relationship was not robust enough to be statistically significant (Table 2). The coefficient of linear survival time was not significant either (−0.001; 95% CI, −0.003 to 0.002).
Table 2. Results of an Age-Adjusted and Gender-Adjusted Linear Regression Model Estimating the Existence of an Inverted U-Shape Relationship Between the 15D Scores and Survival Time Since LT Among LT Patients
The mean 15D scores of the CLD, ALF, and liver tumor groups were 0.892 (SD, 0.101), 0.883 (SD, 0.102), and 0.877 (SD, 0.131), respectively. The 15D scores of the CLD and ALF groups were statistically significantly lower than that of the age-standardized and gender-standardized general population. The respective linear regression coefficients were −0.0108 (95% CI, −0.0211 to −0.0005; P = 0.040) for the CLD group and −0.0236 (95% CI, −0.0423 to −0.0048; P = 0.014) for the ALF group. The regression coefficient of liver tumor patients was −0.0118 (95% CI, −0.0492 to 0.0255; P = 0.535).
Employment Status and Working Capacity
Three hundred forty-seven patients responded to all or most of the questions in the employment questionnaire. The results are given in Table 3. One-third of all patients were employed at the time of the study. The percentages of patients currently employed did not differ according to the indication group (36%, 34%, and 32% in the CLD, ALF, and tumor groups, respectively). Among working-age respondents (20-65 years of age at the time of study, n = 268), 44% (119/268 patients) were currently employed, and those patients had better HRQoL than working-age respondents that were not employed (mean 15D score, 0.934 versus 0.859, respectively). This difference (Δmean, 0.074; 95% CI, 0.053-0.096) was clinically important and statistically significant according to linear regression analysis adjusted for gender and current age (P < 0.0001).
Table 3. Results for the 350 Patients That Responded to the Questionnaire on Current Employment Status and Change in Working Capacity
Answers to the Specific Question (%)
Home maker in 5 cases, sickness allowance in 3 cases, rehabilitation in 1 case, and no reason given in 4 cases.
Child at LT in 7 cases, sickness allowance in 4 cases, home maker in 2 cases, rehabilitation in 2 cases, maternity allowance in 1 case, and no reason given in 4 cases.
Reason if the answer was “no”
Early retirement associated with liver disease
Early retirement not associated with liver disease
Current working/functional capacity compared to 1 week before LT
The underlying cause of not being employed at the time of the study was more often early retirement (56%) than retirement due to age (31%; Table 3). The mean 15D score was lower among patients with early retirement due to either a cause related (82/219 patients; mean score, 0.866; SD, 0.103) or not related (41/219 patients; mean score, 0.828; SD, 0.122) to their liver disease than among patients currently not employed for any of the other reasons presented in Table 3 (96/219 patients; mean score, 0.882; SD, 0.109). The differences between groups were statistically significant according to a 1-way analysis of variance (P = 0.030).
Figure 3 shows that, among working-age patients at LT (20-65 years of age at the time of transplantation), patients transplanted at a younger age were more often able to return to work after LT. The proportion of patients returning to work in less than 6 months after LT was also higher among the younger patients.
Patients that had returned to work early after LT showed better HRQoL than patients that had returned to work at later posttransplant time points. This is demonstrated by the fact that the mean 15D scores for patients returning to work <3 months, 3 to 6 months, 6 to 12 months, 1 to 2 years, and >2 years after LT were 0.967 (SD, 0.038), 0.923 (SD, 0.097), 0.917 (SD, 0.089), 0.924 (SD, 0.061), and 0.897 (SD, 0.069), respectively. These subgroups were, however, too small for reliable statistical testing. Of the patients that had not returned to work after LT (57%, 194/342 patients), a noteworthy proportion had done so because of early retirement (60%; Table 3). A larger percentage of CLD patients returned to work in less than 6 months (18%, 46/258 patients) compared to ALF patients (11%, 8/76 patients), but the difference between these proportions (0.07; 95% CI, −0.03 to 0.15) was not statistically significant according to the chi-square test (liver tumor patients were excluded from the analysis because of small numbers).
Eighty-seven percent of respondents (290/333 patients) felt that their current working/functional capacity had improved much or slightly compared to that 1 week before LT. In CLD and ALF patients, the proportions were similar (90% and 81%, respectively), but in liver tumor patients, it was significantly less (72%, P = 0.012, chi-square test). None of the patients transplanted at 20 years of age or less reported that their current working/functional capacity had deteriorated compared to that before LT. In the 2 other age groups, the proportions of patients reporting improved working/functional capacity were similar, that is, 88% (147/168 of patients) in the age group of 20 to 50 years and 86% (131/153 of patients) in the age group > 50 years.
This cross-sectional study is unique in the sense that it comprises all adult LT patients alive from an entire country, Finland, with no patient lost to follow-up, and a comparison is made with an age-standardized and gender-standardized population with the same ethnic characteristics. With an over 88% response rate, the representativeness of the study population is excellent. Also, as all questionnaires could be answered by patients in their native language, significant selection or technical bias does not occur. Bias is naturally caused by the fact that deceased patients are not included in this cross-sectional study. Clearly, inclusion of deceased patients would change quality of life results. On the other hand, patient survival is high in our LT program.
HRQoL reflects the physical, psychological, and emotional dimensions of health. HRQoL is affected by several factors, such as the general condition of an individual, his or her health problems, phase of life, tasks and goals in life, social status, education, and specific diseases and treatments. As it is impossible to control for all the factors influencing HRQoL, it is of utmost importance to compare HRQoL in specific disease groups with a control population. In this study, HRQoL was assessed with the 15D instrument, which has been shown to be a valid, sensitive, and feasible tool for assessing HRQoL in various fields of human life.20–23 As a generic HRQoL instrument that is not disease-specific, the 15D provides a more general HRQoL score and allows for comparisons with the general population. A simple power analysis with the statistical significance level set at 0.05 and power set at 80% indicates that it generally would require a sample size of approximately 110 patients to achieve a clinically important difference in mean 15D scores when compared to the reference values of the general population. The sample size in this study was therefore adequate.
According to our data, LT patients experience surprisingly good HRQoL compared to that of the general population. Although a statistically significant difference in HRQoL scores was found among LT patients compared to the general population, the difference was not clinically important. The greatest differences were seen in some physical dimensions, such as moving, sleeping, elimination, usual activities, and sexual activity, in which LT patients were worse off compared to the general population. On the other hand, LT patients did not differ from the general population in mental and psychological dimensions, and LT patients scored better on discomfort and symptoms. It is difficult, and perhaps not even appropriate, to compare the current results with those of previous reports because of the heterogeneity of instruments used for HRQoL assessment in other studies.6, 7 It should be noted, however, that Younossi et al.,8 Hellgren et al.16 and van der Plas et al.25 reported similarly that LT patients differed from healthy controls mostly on the physical domains of HRQoL. In the study by Hellgren et al., the Swedish LT patients suffered more from pain than the healthy controls. Future studies should assess which conditions or complications specifically account for these deteriorated scores on the physical dimensions of HRQoL.
In subgroup analyses, we found that gender, age at LT, and retransplantation did not influence HRQoL. Older age, on the other hand, was an independent factor impairing HRQoL. HRQoL deteriorates with age also in the general population.24
Patients transplanted for ALF differ markedly from patients with CLD. CLD patients have experienced life with a chronic disease prior to LT, have had time to prepare for the coming LT, and have often received extensive patient education regarding life after transplantation.12 ALF patients, on the other hand, are often previously healthy individuals with a sudden onset of liver disease and a rapid deterioration of their general condition. Therefore, it could be expected that ALF patients would perceive their quality of life after LT differently from CLD patients. Sargent and Wainwright26 reported that posttransplant HRQoL was comparable between 26 ALF and 34 CLD patients. In line with that, we found that although the HRQoL of ALF patients deviated somewhat more from population norms than the HRQoL of CLD patients, this difference between ALF and CLD patients was not clinically important, as shown by the similar CIs for both indication groups. The fact that the liver tumor group did not experience significantly decreased HRQoL, in comparison with the general population, may be explained by the small number of patients in this group and thus the lack of statistical power.
The majority of studies have reported that HRQoL in general improves after LT compared to that before LT, but in most of the studies, posttransplant follow-up times have been only 1 or 2 years.6–10, 15, 17, 27 Only a few studies have assessed HRQoL more than 10 years after LT.1, 13 In this study, a trend of gradual deterioration of HRQoL was seen after the first posttransplant year, but the age-adjusted and gender-adjusted regression analyses revealed that this deterioration was mostly due to the HRQoL deteriorating effect of increasing age itself. The HRQoL deteriorating effect of increasing posttransplant survival time, when adjusted for age and gender, was not significant. This is interesting as several studies have shown that longer post-LT survival time is associated with an accumulation of several long-term posttransplant complications, such as renal dysfunction, malignancies, and cardiovascular disease.28–30 It could thus be expected that the accumulation of complications would be reflected in a deterioration of HRQoL along with longer survival time. The results from this study therefore indicate that long-term posttransplant complications may not markedly decrease HRQoL. As the results are, however, based on different patients at different follow-up periods, they can be regarded as only indicative and need to be confirmed in prospective studies with long-term follow-up.
Although the majority of patients (87%) felt that their working/functional capacity improved after LT, only 43% of all patients had returned to work after LT. In this study, patients were asked to estimate the change in their working capacity or functional capacity in everyday life also if they had not actually returned to paid employment but were, for example, students or homemakers instead. In line with our results, Loinaz et al.31 reported that 41% of patients returned to work an average of 2.6 months after LT in a study from their center in Spain. Our study further indicated a trend in which returning to work was dependent on age; the younger the patient was at the time of LT, the more likely and sooner he or she was to return to work.
In our study, 44% of working-age patients were employed at the time of the study. This is undeniably less than in the general Finnish population, in which approximately 70% of the working-age population is employed,24 but similar to that reported in other studies on LT patients.32–33 Saab et al.32 reported an employment rate of 27% in their LT population, whereas Cowling et al.33 reported employment rates of 40% for men and 25% for women at 2 years post-LT. Comparisons to other studies are, however, problematic because of different definitions used for the term employment and because of different age distributions in the patient populations. Patients currently employed had significantly better HRQoL than those unemployed in the present study.
More than half of patients unemployed at the time of the study and more than half of patients that had not returned to work after LT reported early retirement as the cause. Early retirement was also associated with the worst HRQoL scores. Particular concern should thus be paid to this group of patients when intervention strategies are developed with the aim of facilitating post-LT recovery. It would also be of interest in future studies to review which medical conditions and complications in particular most often underlie early retirement in LT patients.
Results from this study regarding employment are based on self-developed questions, which to some extent limit the reliability. We did, for example, not distinguish between part-time and full-time employment or whether patients had returned to their previous work and salary level. It can also be criticized that we did not include students in the employed group. However, only 0.5% of the patients were students at the time of study, so the exclusion of this group from employment analyses should not markedly influence the results. Also, as patient numbers were quite small in many of the subgroups, there was not enough statistical power for statistical testing of some of the trends. The results on employment should thus be seen as indicative and should be confirmed in future research with larger study populations and with validated employment questionnaires.
In conclusion, LT patients experienced only slightly worse HRQoL than the age-standardized and gender-standardized general population. Gender, age at LT, survival time, and retransplantation did not affect HRQoL significantly. HRQoL deteriorated somewhat more after LT due to ALF than after LT due to CLD. Most patients experienced improved working capacity after LT. Employment status was found to be a predictor of HRQoL. Patients unemployed because of early retirement presented with worse HRQoL than patients unemployed for other reasons.