The long-term impact of pediatric liver transplantation (LT) and its complications on general health, health-related quality of life (HRQoL) and sexual health were assessed. We conducted a national cross-sectional study of all pediatric recipients who underwent LT between 1987 and 2007. Of 66 survivors, 57 participants (86%) were compared to randomly chosen healthy controls (n = 141) at 10.7 ± 6.6 years posttransplant. PedsQL4.0, SF-36, DISF-SR and AUDIT questionnaires for appropriate age groups were used. Patients and controls <7 years had similar HRQoL and 54% of patients aged over 7 scored within the controls’ normal range on all HRQoL domains. In adult survivors, physical functioning and general health were decreased (p < 0.05). Biliary complications, reoperations and obesity were independently associated with reduced HRQoL (p < 0.05 for all). Still 64% of adult survivors considered their health excellent. Sexual health was similar to controls but LT recipients may experience problems with their orgasm strength (p = 0.050) and condom-based contraception was more common after LT than among controls (58% and 12%, p < 0.001). In conclusion, normal HRQoL and sexual health are achievable post-LT.
Derogatis interview for sexual functioning self-report
glomerular filtration rate
health-related quality of life
posttransplant lymphoproliferative disease
Liver transplantation (LT) is the only remedial therapy for end-stage liver disease and some malignancies and metabolic diseases in children (1–3). As the survival rates have increased along with improved surgical techniques and immunosuppressive medications, the focus of interest has shifted to long-term complications (4,5) and health-related quality of life (HRQoL; Refs. 6–8). The goal of LT is no longer only survival but a state of health enabling the recipients to study, work and have family-life comparable to their peers (9,10).
HRQoL can be measured directly from the patient's perspective by validated questionnaires, or in younger children from the perspective of a close proxy, such as a parent. Previous studies have suggested that HRQoL in pediatric LT recipients is lower than that of healthy children (6,11). Exact reasons for this finding remain speculative. HRQoL includes assessments of physical, psychosocial and functional dimensions, but many studies report only the incidence of surgical complications (12,13) or the immediate effects of LT on HRQoL (14) without combining these data nor assessing the actual effects of surgical complications on long-term HRQoL. If a healthy control population has been included, it has been a historical cohort not specifically matched to the LT population (6,15,16). Several studies have reported on adolescent HRQoL after LT (15,17) but follow-up periods extending into adulthood are still very rare (10). Also, we found no reports on sexual health after pediatric LT and only one in adults (18).
The purpose of this study was to evaluate the long-term general health, HRQoL and sexual health after pediatric LT using validated instruments to compare the patient results to those of healthy controls, and to combine these results with clinical data on surgical and medical complications to evaluate their effects on health and HRQoL for a population-based comprehensive view of the long-term results of pediatric LT.
Patients and Methods
Patients and controls
Between 1987 and 2007, a total of 99 pediatric patients (age <18) underwent deceased donor LT in Finland. All Finnish patients are transplanted at the Helsinki University Central Hospital to which they also return for regular follow-up visits. We conducted a cross-sectional study of all 66 survivors, of whom 57 (86%) participated a mean of 10.7 ± 6.6 years after LT. Of the nonparticipants, one patient's parents notified the researchers of their unwillingness to participate but eight patients merely did not return their questionnaires. Insurance type plays no role in the patients’ treatment or follow-up as health care in Finland is based on national social security ensuring equal access to medical care for all citizens.
Of 33 deceased patients, six were under 1 year, sixteen 1–10 years, four 11–16 years and seven 18–23 years at the time of death. Causes of death included infections (13), arterial thrombosis (3), primary nonfunction (3), cerebral edema (3), tumor recurrence (3), GI hemorrhage (2), respiratory failure (2) and noncompliance (4). The noncompliant patients were 16–23 years old.
A control group of 396 persons matched for age, sex and place of residence was randomly picked by the Finnish Population Register Centre. After one round of reminders mailed to nonresponders, 141 (36%) controls answered. Mild allergies were not considered a criterion for exclusion and no other illnesses were reported.
Cyclosporine (CsA), azathioprine (AZA) and methylprednisolone were used for maintenance of immunosuppression. CsA was switched to tacrolimus and AZA to mycophenolate mofetil, if clinically indicated. Methylprednisolone was tapered to 0.25 mg/kg day at 2 weeks, switched to every other day at 6 months (0.1 mg/kg/day) and usually discontinued after the patient reached adulthood. Individual changes were made when clinically indicated but the principles of immunosuppressive medication remained constant during the 20-year period.
Annual hospital visits included height and weight measurements, blood pressure, lumbar spine bone mineral density (BMD) measured with dual-energy X-ray absorptiometry, glomerular filtration rate (GFR) measured using 51Cr-EDTA clearance or creatinine clearance, ultrasound imaging of the graft and blood levels of liver function tests, such as total and conjugated bilirubin, aspartate transaminase (AST), alanine transaminase (ALT), gamma-glutamyltransferase (GGT), prealbumin, international normalized ratio and creatinine, glucose and cholesterol. In addition to these results, demographic and clinical data were collected from patient records and the Finnish LT registry, including diagnosis of liver disease and other possible illnesses, LT date, age at LT, vascular and biliary complications and their treatment, acute and chronic rejections, retransplantations, infections, growth delay requiring use of growth hormones, posttransplant lymphoproliferative disease (PTLD) and neurological and psychiatric problems. Body mass index (BMI) was calculated and both patients and controls were classified according to the BMI-for-age percentiles into underweight (<5th percentile), normal weight, overweight (>85th percentile) and obese (>95th percentile; Ref. 19).
PedsQL4.0 and SF-36 questionnaires were used to measure HRQoL. PedsQL4.0 is a generic, standardized and internationally widely used questionnaire with proxy forms for the youngest age groups (children <7) and both proxy and patient forms for school-age children (20). It has been validated in both Finnish (21) and Swedish (22) that are the two official languages in Finland. PedsQL4.0 measures dimensions of physical, emotional, social and school functioning by a scale of 0–100, with higher scores reflecting better health. In children younger than 7 years, the school functioning dimension is somewhat irrelevant in Finland as compulsory education starts at the age of seven.
SF-36 is a generic, standardized questionnaire for adults, which has also been validated in both Finnish and Swedish. The SF-36 consists of eight scaled scores, which are transformed into a 0–100 scale, with higher scores indicating better HRQoL. The eight sections are vitality, physical functioning, bodily pain, general health perceptions, mental health and physical, emotional and social role functioning (23).
Derogatis interview for sexual functioning self report (DISF-SR) was used to quantify the quality of the adult patients’ quality of sexual functioning (24). The 25 questions of the DISF-SR are arranged into five domains: sexual cognition/fantasy, sexual arousal, sexual behavior, orgasm and sexual drive/relationship. Each item is rated on a five- or nine-point Likert scale and total scores are calculated for each domain in addition to an aggregate total score. Higher scores only indicate better functioning on the orgasm domain whereas on other domains they imply a higher activity level.
We also asked the adult patients and controls to fill the Alcohol Use Disorders Identification Test (AUDIT) in their native language to assess alcohol consumption, possible alcohol abuse and its effects on HRQoL. All 10 questions are rated on a three- or five-point Likert scale with a maximum of four points per question yielding a maximum sum of 40 points. A total of ≥8 points is associated with increased health risks (25).
The patients and controls filled a survey on the patients’ and parental education levels (compulsory education, i.e. grades 1–9; high school, vocational school, college or university) and employment status. Adult patients and controls were also asked about possible contraception use, pregnancy wishes, pregnancies and sexually transmitted diseases. Adult patients’ forms included a final question about the sufficiency of information they received about effects of LT and immunosuppression on reproductive health and fertility. The patients’ questionnaires were returned directly to the primary researcher (S.K.), who does not participate in the patients’ care.
Statistical analyses were performed with SPSS Statistics software, version 17.0 (IBM, Somers, NY, USA) and R (R Foundation, Vienna, Austria). Frequencies, percentages, means and standard deviations (SD) were used as descriptive statistics of clinical and sociodemographic variables and in cases of less than 10 patients in a group, median and range were recorded. For ordinal variables of HRQoL measurements, differences between groups were tested with the Wilcoxon matched pair test and the false discovery rate was used to correct for multiple comparisons. The McNemar test was used for categorical variables. Correlations were calculated using the Spearman rho two-tailed test and a final p-value of ≤0.05 was considered statistically significant.
This study was approved by the ethics committee of the Helsinki and Uusimaa Hospital district. All participating patients and controls, and in case of minors also their parents, signed an informed consent form.
Baseline characteristics of the patients
The median ages at LT and HRQoL assessment and length of follow-up for all age groups are presented in Table 1. The nine nonparticipants were mainly males (8 of 9) but showed no difference in primary diagnoses, complication rates or length of follow-up. The primary diagnoses leading to LT are presented in Table 2. Laboratory tests at the time of HRQoL assessment showed stable graft function with mean values including total bilirubin 11 ± 6 μmol/L, AST 33 ± 11 U/L, ALT 27 ± 17 U/L, GGT 40 ± 32 U/L and prealbumin 214 ± 49 mg/L.
Table 1. Demographics of patients and controls
LT = liver transplantation; HRQoL = health-related quality of life.
Age at LT, age at HRQoL and follow-up time presented in years as median (range).
Total number in group
Female, n (%)
Age group, years
Age at LT
Age at HRQoL
Table 2. Diagnoses leading to LT stratified according to age groups
Age group, years 7–17
Total, n (%)
Metabolic includes tyrosinemia (6), Wilson's disease (2), hyperoxaluria (2), ornithine transcarbamylase deficiency (1) and familial hypercholesterolemia (1).
Other includes autosomal recessive polycystic kidney disease (4), acute liver failure (2), Budd-Chiari syndrome (2), hemolytic uremic syndrome (2), iron poisoning (2), primary sclerosing cholangitis (1) and congenital fibrosis (1).
Surgical and medical complications and comorbidities: Complication rates are presented in Table 3 with follow-up times after LT and the time passed between complication occurrence and HRQoL assessment. Six patients (11%) had a clean record of no surgical or medical complications. The mean number of any complications per patient was 2.3 and five patients (9%) had five different complications (maximum in our patient population). Vascular complications requiring surgical treatment included early hepatic artery thrombosis in four patients (7%; two required retransplantation), hepatic artery stenosis in one, portal stenosis in three and hepatic vein stenosis in one. Biliary complications comprised bile leaks (two) and strictures (nine patients). Bile leaks were treated by biliary drainage in both cases. Strictures of five patients were treated by percutaneous transhepatic or endoscopic stents whereas four patients underwent reconstructive surgery. In total, four patients underwent re-LT, two because of hepatic artery thrombosis and two because of chronic rejection leading to graft fibrosis.
Table 3. Complication rates and follow-up times
LT to complication (years)
Age at complication (years)
Complication to HRQoL (years)
HAT = hepatic artery thrombosis; PTLD = posttransplant lymphoproliferative disease.
Renal insufficiency defined as glomerular filtration rate ≤60 mL/min/1.73 m2.
Follow-up times (LT to complication and complication to HRQoL assessment) and age at diagnosis of complication presented as median (range) in years.
Table includes all individual complications. Most patients had several, and 11% had none.
Osteoporosis was diagnosed from lumbar spine BMD measurements in 11 patients, 6 of whom also had compression fractures of their thoracic or lumbar spine. PTLD developed and was successfully treated in four patients, and no other malignancies have been diagnosed after LT. Neurological complications included epilepsy in two patients, hearing loss in two, polyneuropathy in one and peroneal paresis in one patient. Seven patients have received psychiatric treatment, six for depression and one for anxiety. Six patients are on antihypertensive medication. The measured GFR was above 60 mL/min/1.73 m2 in 48 patients (84%). No one had GFR < 30 mL/min/1.73 m2. None of the patients developed diabetes during the follow-up period.
HRQoL in children
LT patients and controls younger than seven had similar HRQoL (Table 4). Both LT recipients aged 7–17 and their parents reported significantly lower scores for school functioning (p = 0.004 and <0.001, respectively). A third (31%) of the 19 school-aged patients was one grade behind in school compared to 5% of controls (p = 0.009). Ten of the 19 school-aged patients (53%) scored below the –2SD level of controls at least in one HRQoL domain (range 15–55, p = 0.017). Nine patients’ (47%) scores were in the normal range of controls (mean ± 2SD), of whom three scored higher than the mean level of controls on all domains.
Table 4. Scores of QoL domains, presented as means (SD)
Significant p-values between patients and controls in bold.
1Based on parent-proxy report; school functioning irrelevant as compulsory education starts at age 7.
PedsQL <7 years1
PedsQL 7–17 years
The parents of school-aged children estimated physical, social and school dimensions to be significantly worse for children who were either overweight or obese (p < 0.05 for all). Gender, diagnosis, age at LT, the parents’ education level or work status had no statistically significant effect on HRQoL. In school-aged children, HRQoL showed a significant increasing trend with longer follow-up time in the child's physical (R = 0.503) and social (R = 0.593) PedsQL scores (p = 0.028 and 0.009, respectively).
HRQoL in adults
In adults, statistically significant differences between LT recipients and controls were only present in scores of physical functioning and general health (Table 4). Although patients and controls scored similarly on the emotional well-being domain, significant differences were found in individual questions. LT recipients believed they get sick easier than other people (50% answered true or mostly true compared with 13% of controls; p < 0.001) and expected their health to get worse (28% answered true or mostly true compared with 2% of controls; p < 0.001). Still 76% of patients considered themselves as healthy or almost as healthy as other people and 64% considered their current health status excellent.
No difference was found in smoking or drinking habits according to total AUDIT scores (Table 5). LT recipients tended to drink less frequently (Figure 1A) but their alcohol use raised more concern (Figure 1B). LT recipients lived alone more often than controls (39% and 15%, respectively, p = 0.016). The patients’ education and employment levels tended to be lower than those of controls as 18% of patients only had elementary school education and 7% had studied at university level (compared with 5% and 15% of controls, respectively), although these data failed to reach statistical significance. No correlation was found between the education level of patients and their parents. Among adults, diagnosis of liver disease, age at LT, length of follow-up, education, employment status, smoking, the amount of alcohol used, parental education or parental employment status had no significant effect on HRQoL. Smokers had higher scores on the AUDIT than nonsmokers (mean sum 9.0 vs. 5.8, respectively; p = 0.043). Overweight or obese adults gave significantly lower scores for physical functioning (p = 0.009) but other dimensions of HRQoL were unaffected.
Table 5. Demographics for adult patients and controls
Significant p-value between patients and controls in bold.
Only compulsory education(%)
Employed at least part-time(%)
AUDIT sum, mean
On analysis of the lowest scoring patients (score of any HRQoL domain below –2SD level of controls), biliary complications and psychiatric issues were overrepresented. All seven patients with psychiatric problems and six of ten patients with biliary complications had impaired HRQoL. The number of complications negatively correlated especially with the emotional domains of HRQoL (Table 6). Of the 29 adult patients, 12 (41%) scored lower than the –2SD level of controls at least in one HRQoL domain (range 15–50, p = 0.083). In five of these patients, the HRQoL impairment was severe as more than half of the eight domains were affected. Seventeen (59%) adult patients’ scores were in the normal range of controls (mean ± 2SD), of whom, seven scored higher than the mean level of controls on all domains. School-aged and adult patients combined, 54% of the patients scored within the normal range of healthy controls.
Table 6. Correlation of the total number of complications per patient and HRQoL domain scores
Significant p-values in bold.
Total number of complications includes all different types of medical and surgical complications as listed in Table 3. The amount of complications per patient ranged between zero and five.
Parent-proxy scores for PedsQL not shown because no statistically significant correlations were found.
PedsQL 7–17 years
Sexual health was similar between LT recipients and controls according to DISF-SR scores (Table 7). LT recipients seemed, however, less frequently satisfied with the strength of their orgasms (44% vs. 82%, p = 0.050). This finding was unrelated to the use of condoms. No differences were found in the frequency of fantasies or sexual experience even on analysis of individual questions. Gender and BMI were unrelated to DISF-SR scores among patients and controls (data not shown). Smokers had higher scores in the sexual drive/relationship domain than did nonsmokers (p = 0.043), and AUDIT scores showed positive correlation with both orgasm (R = 0.313) and sexual drive (R = 0.224) domains of the DISF-SR (p = 0.036 and 0.039, respectively).
Table 7. Scores of DISF-SR domains presented as means (SD)
Ranges of possible scores for individual DISF-SR domains are: fantasy 0–40, arousal 0–32, behavior 0–40, orgasm 0–24 and relationship 0–24.
Sexually transmitted diseases were rare in both patients and controls with self-reported rates of 8% and 7%, respectively. LT recipients and their partners used condom-based contraception more often than controls (58% and 12%, p < 0.001) and 61% of patients reported having received insufficient information on the effects of LT and immunosuppression on fertility. Four of 24 (17%) adult patients and 12 of 58 (21%) controls had hoped for pregnancy and three patients have become parents.
This is to our best knowledge the first population-based study combining data on medical and surgical complications and HRQoL measurements and the first study assessing sexual health after pediatric LT. Patients and controls under 7 years had comparable HRQoL. In school-aged children, the school domain of HRQoL was significantly lower than in controls. Physical functioning and general health yielded reduced scores also among the patients who had reached adulthood. Still 54% of patients aged over 7 scored within the controls’ normal range on all HRQoL domains. Sexual health was comparable to healthy peers according to DISF-SR scores.
Our patient population is relatively small and some differences between patients and controls failed to reach statistical significance. The patients also represent a fairly wide age range. Our strengths, however, include the length of follow-up (mean 10.7 years) and the population-based comprehensive nature of data collection with a high response rate (86%). Moreover, an investigator uninvolved in clinical care approached the patients to assure truthful responses. In their study of 20-year survivors of LT, Duffy et al. were able to contact only 52% of LT recipients and the final response rate was 42% (10). Several studies have excluded patients who hadn't maintained regular follow-up with their transplant center (16,26) or patients of different language groups (27). Low response rates and strict exclusion criteria may skew the results. Our results should also be reliable because the controls were randomly selected for this specific study and we eliminated the bias of patients answering to their caretakers. In addition, our study includes the self-reports of all patients ≥7 years as the validity of only proxy-reported HRQoL is questionable. The difference between self- and proxy-reported HRQoL is a common finding (28) and clearly represented also in our results (Table 4).
Several studies have reported that HRQoL of LT patients is lower than that of healthy controls (8,11,16) or in fact comparable to children receiving cancer therapy (26). Half of the patients in our study group, however, scored within the normal range defined by the control group on all HRQoL domains. Also, 68% of adult survivors considered their current health excellent. We hope to raise the question of whether LT recipients actually represent two distinct groups: those with HRQoL impairment due to complications and those with normal HRQoL. Variance is important; some patients thrive as well as or even better than their healthy peers, thus reaching the original LT goal.
Our complication rates are quite similar to those from larger centers (13,29). The total number of surgical and medical complications was negatively associated with emotional HRQoL and general health. Biliary complications independently impaired HRQoL even years post-LT. This may be due to visible jaundice, prolonged treatment with a percutaneous transhepatic transanastomotic catheter or repeated endoscopic procedures enhancing the stigma of illness. Although vascular complications are severe problems, after successful treatment they leave no long-term health impairment. Most patients with renal insufficiency also had normal HRQoL. Grade 3 chronic renal disease may allow a patient to lead a normal life whereas both peritoneal and hemodialysis reduce HRQoL (8,30). The awareness of developing renal insufficiency may, however, contribute to the effect multiple complications have on the emotional domains of HRQoL (Table 6). It may also be one factor contributing to the patients’ expectations of poorer health in the future. HRQoL reduction due to obesity in our study was visible only in the parent reports and in adults. Obesity in children and adolescents has previously been associated with self-reported reduced HRQoL (31–33). These studies have, however, been unable to reach a consensus on whether obesity reduces physical or emotional domains of HRQoL, possibly explaining why we found no association in the children's self-reports.
Risk-taking behavior has a tendency to polarize in youths and young adults (34,35), which we also noted in the correlations between smoking, alcohol consumption and sexual activity. Our adult survivors actually consumed alcohol less frequently than controls, but still raised due concern in health care providers. Smoking reduces HRQoL (36,37) but the effect may come at a later age than in our study group. Our results of possibly lower education in LT recipients are in line with a recent study from France (38). Specific cognitive, for example, visuospatial, impairment after pediatric LT may affect both achievement level and education (39). Åberg et al. showed employment status to correlate with HRQoL in adults after LT (9), but our adult population may be too young to show this effect. Interestingly, we found no difference in rates of chronic pain between LT recipients and controls, a finding which has also been published previously (10) but received fairly little attention. Another noteworthy finding was the significant improvement of HRQoL with longer follow-up. As time passes and life-threatening operations and infections become less frequent, patients may feel more content and confident about their lives.
All noncompliant deceased LT recipients (4 of 33, 12%) fell into the age group of 16–23 years, which is also the age of transition from pediatric to adult services. We have had a transition-clinic from the beginning of 2005, which all patients attend after they turn 13. A list of topics to be discussed is available to both the LT recipient and doctor and besides vaccination, travel and tattoos, it also includes several issues related to sexual health and fertility. Still 61% of LT recipients over 18 reported having received insufficient information on effects of LT and immunosuppression on fertility. Some of the patients transferred to adult services before the clinic and possibly the delicate subject was addressed inadequately or precociously, or repetition after transition was deficient. French patients, who reported that LT affected their love life, disclosed no physical problems but felt greatly embarrassed by their scars (38). Scars were an issue also for 30% of patients in Taylor's study of adolescent LT recipients, and high self-esteem was associated with better HRQoL (7). Self-esteem and appearances may also play an important role in the difference we found between the patients’ and controls’ orgasm scores. A well-established, functioning transition clinic is essential in assuring the good results of pediatric LT are carried on to the future of these patients. Our work continues in ensuring the quality of our transition program and developing the cooperation of pediatric and adult care further for the advantage of our patients.
In conclusion, long-term HRQoL after LT can approach that of healthy controls in a majority of patients. Specific targets for improvement include psychosocial adjustment and consideration of school issues, as these will have long-term effects on the overall life of the LT recipients. Sexual issues should also be addressed more often, and the value and quality of a systematic transition program should be high lighted.
We thank Janne Pitkäniemi for indispensable statistical expertise. This study was supported by grants from the Helsinki Central Hospital Fund, the Sigrid Juselius Foundation, the Päivikki and Sakari Sohlberg Foundation and the Foundation for Pediatric Research.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.