Psychosocial adjustment to orthotopic liver transplantation in 266 recipients

Authors


Abstract

Although the survival rate of patients undergoing orthotopic liver transplantation (OLT) is highly satisfactory, one of the most important objectives for liver transplantation teams at the present time is to achieve the best possible quality of life and psychosocial functioning for these patients after transplantation. We present the preliminary results of a study designed to determine which domains of psychosocial functioning are most affected in liver transplant recipients, and to examine the factors associated with poorer adjustment after OLT, using a utility-based standardized measure. Patients who had undergone liver transplant more than 12 months previously were eligible. They were administered the Psychosocial Adjustment to Illness Scale (PAIS), and they provided the answers themselves. Multivariate regression models showed that attitudes toward health care were poorer in women (β = 0.916, P < .001), in patients who were employed at the moment of transplantation (β = 0.530, P = .032), and in patients of lower social class (β = 0.722, P = .026) than in men, unemployed patients, and patients of higher social class. Sexual functioning was worse in women (β = 0.907, P = .001) and older patients (β = 0.999, P < .001) than in men or younger patients. Psychological distress was higher in women (β = 0.981, P = .001) than in men, and lower in currently employed patients (β = −0.937, P = .001) than in the unemployed. Only gender remained significantly associated with the total PAIS score (β = 0.969, P < .001), with women showing a poorer overall psychosocial adjustment to OLT. In conclusion, there seems to be no doubt that liver transplantation improves quality of life, but special attention should be paid to female recipients, who seem to have more difficulty than their male counterparts in adjusting to the psychosocial consequences of the procedure. (Liver Transpl 2004;10:228–234.)

The survival rate of patients undergoing orthotopic liver transplantation (OLT) is considered highly satisfactory, and the attention of liver transplantation teams is now turning to achieving the best possible quality of life (QOL) and psychosocial functioning for these patients after transplantation.1 The term QOL is very broad and includes physical and mental status as well as psychosocial adjustment. Physical and mental disturbances after OLT may be produced by medical conditions such as the neuropsychiatric side effects of certain immunosuppressants, and opportunistic diseases affecting the central nervous system and other organs.2, 3 In these cases, it is the duty of hepatologists, psychiatrists, and neurologists to treat or palliate these disturbances.

For its part, psychosocial adjustment—that is, the “capacity of an individual to perform social and domestic roles so as to meet the challenges of everyday living without emotional distress or physical disability”4—depends on personality traits such as coping strategies or personal locus of control. Psychotherapeutic intervention is needed when patients have problems adjusting to the new situation after OLT, and the participation of psychologists, social workers, and support nurses in liver transplantation teams1 has been recommended in order to facilitate improvement in psychosocial functioning. Observational studies are needed to identify which patient characteristics or transplantation-related factors interfere with correct overall functioning after OLT. Information on this topic will help in the design of adequate psychotherapeutic intervention strategies aimed to improve adjustment to the new situation.

There are two principal methods for measuring QOL: health profiles (also called psychometric instruments), and utility-based measures. The two elicit related but different information. Health profiles measure a patient's functional status in a given health state, and utility-based approaches assess how a patient values that health state. It is highly likely that liver transplant recipients who have similar levels of functioning will vary widely with respect to how they value different aspects of QOL. Utility-based measures are sensitive to these differences in preferences, whereas psychometric measures are not.5

Almost all studies of QOL of patients before and after liver transplantation have used generic health profiles such as the Medical Outcome Survey Short Form (SF-36), the Karnofsky Performance Scale, the Nottingham Health Profile, and the Sickness Impact Profile.2 Few utility-based evaluations of liver transplant recipients have been published to date.5

In general, QOL is significantly impaired in liver transplant recipients, but improves posttransplantation.5 A meta-analysis published in 1999 reported larger gains in the areas of QOL most affected by physical health and more modest improvements in areas affected by psychosocial functioning,2 such as worry about illness, employment, home, sexuality, family, social environment, and psychological distress. We present the preliminary results of a study designed to identify which domains of psychosocial functioning are most affected in liver transplant recipients, and to examine the factors associated with poorer adjustment after OLT, using a utility-based standardized measure.

Abbreviations:

OLT, orthotopic liver transplantation; QOL, quality of life; CNS, central nervous system; PAIS, Psychosocial Adjustment to Illness Scale.

Methods

Patients

Patients undergoing OLT at both the Hospital Clínic and the Vall d'Hebró, two of the leading liver transplantation centers in Barcelona, Spain, were eligible for this study. Consecutive patients who had undergone liver transplant more than 12 months previously and who were attending routine control visits at the outpatient clinics were invited to participate and to be interviewed by one of the investigators. Strict confidentiality was ensured.

Materials

Demographic and clinical data were collected from patients themselves.

Given the predominance of physical dimensions in most of the instruments used to measure quality of life in OLT, and the need to provide a greater emphasis on psychosocial domains, each patient was administered the Psychosocial Adjustment to Illness Scale (PAIS) and provided the answers themselves.6 The PAIS was designed to assess recent psychological and social adjustment to illness (during the previous 4 weeks), and it is generally preferred to other questionnaires or scales for the assessment of functional, emotional, and social dimensions of QOL in patients with medical illnesses.7 We used the validated Spanish version of the questionnaire,8 replacing the original term “illness” with the term “liver transplantation” in each question, to emphasize the fact that we aimed to measure the impact of the transplantation itself and not of the liver disease and its complications. The PAIS assesses 46 items grouped together in seven domains: (I) health care orientation, (II) vocational environment, (III) domestic environment, (IV) sexual relations, (V) extended family relationships, (VI) social environment, and (VII) psychological distress. Each answer is evaluated for each domain and for all 46 items together. This instrument has shown a good degree of reliability for both individuals and groups.6 The seven domains are relatively independent of each other in the assessment of the overall score, and each gives a measure of the PAIS validity. Higher scores indicate poorer functioning.

Data Analysis

Proportions were compared by using the chi-square test with Yates correction or Fisher exact test; continuous variables were compared by parametric (t test and analysis of variance) or nonparametric (Mann-Whitney and Kruskal-Wallis tests) statistics, as appropriate. Missing data on the PAIS were treated following the specific guidelines for the instrument.9

To examine the impact of several patients' characteristics on the overall PAIS score and on each domain, we used simple and stepwise multiple linear regression analyses. Due to the skewed distribution, we dichotomized the age of patients at interview and the age of patients at OLT according to their medians (≤56 vs. >56, and ≤53 vs. >53, respectively). Variables with a significant association (P < .005) in simple linear regression analysis were entered as candidate risk factors in multivariate linear regression models. Performance of the model was assessed by the adjusted explained variance (R2). All P values reported are 2-tailed. P values were considered significant if they were less than .05. Data were analyzed using the Statistical Program for Social Sciences Version 10.0 software (SPSS Inc., Chicago, IL).

Results

The final study group consisted of 266 patients. Patients' characteristics are described in Table 1.

Table 1. Demographic and Clinical Characteristics of Study Patients (N = 266)
  • *

    Data expressed as median (interquartile range).

Demographics
Age at OLT (yrs)*53 (45–58)
Age at interview (yrs)*56 (49–62)
Gender (male), n (%)168 (63.2)
With Partner, n (%)206 (77.4)
Social Status 
High, n (%)8 (3.0)
Medium, n (%)223 (83.8)
Low, n (%)35 (13.2)
Employed Before OLT, n (%)193 (72.6)
Employed at OLT, n (%)88 (33.1)
Employed at Interview, n (%)88 (33.1)
Place of Birth (urban), n (%)150 (56.4)
Place of Residency (urban), n (%)211 (79.3)
Time Since OLT (mos)*46 (34–61)
 
PAIS Subscales
Health Care Orientation (n = 266)*4 (2–6)
Vocational Environment (n = 166)*4 (2–7)
Domestic Environment (n = 266)*2 (1–5)
Sexual Relations (n = 224)*4 (0–7)
Extended Family Relations (n = 266)*0 (0–1)
Social Environment (n = 266)*2 (0–5)
Psychological Distress (n = 266)*3 (1–6)
Total (n = 166)*20.5 (12–34)

Overall PAIS Score

Only 126 (47%) of the interviewed patients answered all PAIS items correctly. This was mainly due to difficulties in answering the items in the “vocational environment” subscale; as most of the patients were unemployed before OLT, they were not able to assess whether their professional functioning was impaired due to the transplant itself or due to the previous liver disease. Because of similar conceptual difficulties, 42 patients did not answer certain items related to sexual functioning. Excluding the “vocational functioning” subscale, the participation rate was high (84–99%), suggesting that our study group represents a reasonably unbiased sample of Spanish OLT survivors.

The median overall PAIS score was 20.50 (range, 12–34); adjustment to transplantation was considered good in 27.8% of patients (≤25 percentile score), intermediate in 49.2% (26–75 percentile score), and poor in 23.0% (>75 percentile score). Female gender was the only factor associated with a worse overall PAIS score (P = .008). Women showed significantly poorer adjustment to OLT than men (Fig. 1). The proportion of women with poor adjustment was higher (31.5% vs. 16.7%), and the rate of women with good adjustment was lower( (16.7% vs. 36.1%).

Figure 1.

The effect of gender in psychosocial adjustment to OLT. Good adjustment, overall PAIS score < 25 percentile; intermediate adjustment, overall PAIS score from 26 to 75 percentiles; and poor adjustment, overall PAIS score >75 percentile. A difference was seen between men (gray bars) and women (black bars). Chi-square (2) = 21.6; P < 0.001.

Specific Domains

Median and interquartile ranges, and number of completing patients for each PAIS subscale or dimension are presented in Table 1. The impact of age at time of OLT, age at interview, gender, employment before OLT, employment at time of OLT, place of birth, place of residence, having a partner, social status, and follow-up time after OLT on each PAIS domain is shown in Table 2.

Table 2. Demographic and Clinical Variables Associated With Psychosocial Functioning (Measured by PAIS)
 PAIS Domains*
I Health Care (N = 266)II Vocational Environment (N = 166)III Domestic Environment (N = 266)IV Sexual Relations (N = 224)V Family Relations (N = 266)VI Social Environment (N = 266)VII Psychological Distress (N = 266)Total (N = 224)
PMeanPMeanPMeanPMeanPMeanPMeanPMeanPMean
  • Abbreviation: NS, not significant.

  • *

    Simple ordered logit regression model was performed. The outcome variables (subscales and total scale) were categorized using percentile 33 and percentile 66 as cut-off scores. Only P values of significant association (P < 0.05) were presented.

  • For the total PAIS score, we excluded the items of the subscale “vocational environment.”

  • The variables age at interview and age at OLT were dichotomized according to the median.

AgeNS NS NS <0.001 NS NS NS NS 
 ≤ 56 years 4.62 4.62 3.71 3.24 1.17 2.97 3.76 25.04
 > 56 years 4.37 4.63 3.16 5.17 0.79 3.56 3.89 22.00
Age at OLTNS NS NS <0.001 NS NS NS NS 
 ≤ 53 years 4.66 4.65 3.68 3.08 1.03 3.10 3.79 23.84
 > 53 years 4.39 4.60 3.28 4.85 0.95 3.37 3.85 23.79
Recipient Gender<0.001 NS NS 0.002 0.022 NS 0.010 <0.001 
 Female 5.34 4.70 3.93 5.36 1.27 3.56 4.78 28.30
 Male 4.00 4.56 3.14 3.62 0.81 3.10 3.27 20.44
Employed 3 yrs Before OLTNS NS 0.001 0.020 NS NS NS NS 
 Yes 4.54 4.50 3.06 3.60 0.88 2.90 3.58 22.84
 No 4.38 5.08 4.44 5.71 1.25 4.31 4.47 27.37
Employed at OLT0.021 NS NS NS NS NS 0.048 NS 
 Yes 4.91 4.18 3.06 3.45 1.01 2.59 3.43 23.74
 No 4.29 4.98 3.63 4.55 0.97 3.63 4.02 23.87
Employed at Interview0.009 <0.001 0.027 NS NS 0.001 0.036 NS 
 Yes 4.92 3.14 2.85 3.67 1.06 2.14 3.32 20.39
 No 4.28 6.19 3.73 4.43 0.94 3.84 4.07 27.34
Place of ResidencyNS NS NS NS NS 0.046 NS NS 
 Urban 4.38 4.60 3.45 4.05 1.02 3.45 4.03 24.37
 Nonurban 4.93 4.73 3.37 4.70 0.82 2.56 3.05 21.42
PartnershipNS NS NS NS NS NS NS NS 
 Yes 4.41 4.76 3.53 4.22 1.00 3.39 3.90 23.84
 No 4.80 4.28 3.09 3.96 0.90 2.84 3.58 23.59
Social Status0.033 NS 0.032 NS NS NS NS NS 
 High 4.40 4.59 3.20 4.01 0.92 3.16 3.69 22.75
 Low 5.02 4.86 4.63 5.17 1.32 3.83 4.59 30.59
Time After OLTNS NS NS NS NS NS NS NS 
 ≤ 5 years 4.42 4.65 3.44 4.10 1.00 3.40 4.02 23.14
 > 5 years 4.67 4.62 3.46 4.30 0.94 2.94 3.18 25.56

Older age at OLT (P < .001) and older age at interview (P < .001) were associated with poorer adjustment (higher scores) in sexual relations (domain IV). Female gender was associated with poorer adjustment to health care (domain I; P < .001), sexual relations (domain IV; P = .002), extended family relationships (domain V; P = .022), and psychological distress (domain VII) (P = .010). Patients who had been unemployed for three years before OLT (i.e., long-term unemployed) but not patients unemployed at time of OLT (i.e., short-term unemployed) scored higher (i.e., had poorer adjustment) in the domestic environment (domain III; P = .001) and sexual relations (domain IV; P = .020) subscales. Patients who lived in urban areas showed a poorer social adjustment (domain VI) (P = .046) than patients from nonurban areas. Partnership and time after OLT did not have a significant impact on adjustment.

Stepwise multivariate ordered logit regression models for each subscale and for the total score are shown in Table 3. Adjustment to health care was worse in women (β = 0.916, P < .001), in patients who were employed at the time of transplantation (β =0.530, P= .032), and in patients of lower social class (β = 0.722, P = .026) than in men, unemployed patients, and patients of higher socioeconomic status. Sexual functioning was worse in women (β = 0.907, P = .001) and older (age > 56 years) patients (β = 0.999, P < 0.001) than in men or younger patients. Social functioning was worse in patients living in urban areas (β = 0.590, P = .036) than in nonurban patients and better in patients who were employed at time of interview (β = −0.883, P < 0.001) than in unemployed patients. Psychological distress was higher in women (β =0.981, P = .001) than in men and lower in currently employed patients (β = −0.937, P = .001) than in unemployed patients. Multivariate analysis of the total PAIS score as the outcome variable showed that only gender remained significantly associated with the total PAIS score (β = 0.969, P < .001). Women showed poorer overall psychosocial adjustment to OLT.

Table 3. Demographic and Clinical Variables Associated With Psychosocial Functioning (Measured by PAIS)
 PAIS Domains*CategoriesAdjusted Coefficient95% CIP Value
  • Abbreviation: NS, not significant.

  • *

    Multiple ordered logit regression model was performed. The outcome variables (subscales and total scale) were categorized using percentile 33 and percentile 66 as cut-off scores. Only P values of significant association (P < 0.05) were presented.

  • For the total PAIS score, we excluded the items of the subscale “vocational environment.”

IHealth Care (n = 266)recipient gender (female)0.9160.439–1.394<.001
  employed at OLT0.5300.045–1.015.032
  social status (high)−0.722−1.357–−0.087.026
IIVocational Environment (n = 166)employed at interview−1.678−2.298–1.059<.001
IIIDomestic Environment (n = 266)employed 3 years before OLT−0.866−1.377–−0.354.001
  social status (high)−0685−1.329–−0.042.037
IVSexual Relations (n = 224)age (> 56 years)0.9990.497–1.502<.001
  recipient gender (female)0.9070.366–1.448.001
VFamily Relations (n = 266)recipient gender (female)1.8631.095–3.168.022
VISocial Environment (n = 266)employed at interview−0.883−1.376–−0.390<.001
  place of residency (urban)0.5900.037–1.143.036
VIIPsychological Distress (n = 266)recipient gender (female)0.9810.458–1.503<.001
  employed at interview−0.937−1.483–−0.390.001
 Total (n = 224)recipient gender (female)0.9690.439–1.499<.001

Discussion

The Concept of Psychosocial Adjustment

Previous studies have presented substantial evidence that OLT improves QOL. However, the term “QOL” is very broad and includes physical and mental status, as well as psychosocial adaptation.1 Although several studies of psychosocial functioning in post-OLT patients have used only measures of psychological or mental health status, it is obvious from the term “psychosocial” that more than just intrapsychic processes are involved. The concept includes interactions between the individual and other individuals and the institutions that make up his or her sociocultural environment. Such interactions are usually achieved through loosely prescribed behavioral patterns. termed “roles.” The functional efficiencies of an individual's role behavior (e.g., spouse, parent, professional) tend to be highly correlated with judgments concerning his or her level of psychosocial adjustment.9 Further, it has been consistently observed in many major diseases that the nature of the patient's psychosocial adjustment can be just as important as the status of his or her physical disease in determining the quality of an illness experience.9 There is a small body of literature on the effects of OLT on QOL,10 but few studies have specifically focused on the assessment of the psychosocial adjustment of OLT recipients to their new life conditions.

Overall PAIS Score

The PAIS has been used to evaluate outcomes of coronary artery bypass surgery,11 hemodialysis,12 diabetes retinopathy,13 bone marrow transplantation,14 and heart transplantation.15 In these studies, pretreatment global PAIS scores ranged from 31.313 to 58.5,11 and posttreatment scores ranged from 23 to 34. The posttreatment scores for the study by Pinson et al. ranged from 24 to 31.16 Two studies previously showed an improvement in certain PAIS domain scale scores and the total score after liver transplantation using bivariate correlations on cross-sectional samples ranging from pretransplantation to 5 years after.17, 18 In a study of 19 heart transplant patients, Walden et al. found a PAIS global score of approximately 54 before surgery, improving to approximately 34 at a mean of 30 months after transplant.15

Comparing our sample's overall PAIS score with those of other groups of patients studied by the author of the questionnaire,6, 9 we observed that patients submitted to OLT reported better functioning than patients undergoing hemodialysis or patients with chronic heart disease. Nevertheless, our patients showed a higher mean score (poorer functioning) than the norm-group of patients with hypertension.9 Our sample also showed a better overall functioning than the recipients of different types of solid organ transplantations presented in a previous study16 and another sample of patients after allogenic bone marrow transplantation.19 In the studies by Jenkins et al.20 and Caccamo et al.,21 no data about the overall PAIS score were given. To our knowledge, no other studies have administered the PAIS to posttransplant patients.

Factors Associated With Poorer Outcome

In our study, a substantial proportion of patients (23%) reported poor adjustment to OLT, although the only factor associated with overall poor adjustment was female gender. Gender differences in psychosocial functioning were also found in studies with bone marrow transplantation survivors.19, 22 In our previous study, women reported higher psychiatric morbidity and a higher percentage of medical problems than men, and a lower percentage of women were active.22

In the present study, women showed a greater dysfunction in health care orientation (domain I), sexual relations (domain IV), extended family relationships (domain VI), and psychological distress (domain VII). Given the lack of scientific evidence for these phenomena, the following conclusions are drawn from our clinical experience. The health care orientation dimension of the PAIS addresses the nature of the respondent's attitude to health care and whether this attitude will promote a positive or a negative adjustment to the OLT. In our study women seemed to suffer more from their illness state. This may be because, in our milieu, while men with a severe illness take support from their spouse or partner for granted, sick women are expected to care not only for themselves, but for their husband (or partner) and their children as well; they are also responsible for domestic chores. After transplantation, female patients are no longer regarded as sick and are expected to gradually return to nonpatient status.23 This imbalance is also reflected in the state's pensions and benefits system. Male transplant patients are considered permanently disabled and do not need to work because they receive a pension from the Spanish National Health Service, but female transplant patients are expected to continue to perform their family duties. The nonfulfillment of a woman's role at home may produce real domestic conflict. Furthermore, whereas male transplant patients expect to be cared for by the female members of their family, the reverse is not the case: Women may not receive the care they require unless they have daughters.

Greater problems with sexuality in women than in men has already been described in previous studies done in bone marrow transplant patients by other authors and also by our group.22, 24, 25 Decrease in sexuality in women after OLT may have a variety of causes: loss of body image due to hypertrichosis and Cushing's syndrome caused by prolonged immunotherapy, hypoestrogenism, and fear of relapse.19 In addition, many women do not like to show their surgical scars.

Another remarkable finding of our study is that the unemployment rate in our patients had not changed 1 year after OLT, similar to previous studies performed in the United States and the UK (Table 1).26, 27 There are two main explanations for this phenomenon in our setting. First, patients with permanent disability due to OLT receive adequate economic coverage from the Spanish National Health Service. Second, given that transplant patients tend to be relatively old (mean age, > 50) and will need frequent medical follow-up visits, it is very difficult for them to obtain employment. Analyzing the different domains separately, we were able to obtain more information about the psychosocial functioning of OLT patients with specific characteristics. We observed that older patients and patients who underwent OLT at later stages of their lives also reported a greater impact on sexuality.

As in previous research carried out in bone marrow transplantation by other authors28, 29 and also by our group,22 older age at OLT was found to be associated with poorer post-OLT outcome. Although in our previous study in bone marrow transplants we also found that quality of life and psychosocial distress improved with the passage of time,22 in the present study time since OLT was not significantly associated with psychosocial functioning.

There seems to be no doubt that liver transplantation improves QOL, but special attention should be paid to women recipients, who seem to have more difficulty than men in adjusting to the psychosocial consequences of the operation. Female patients may be in particular need of psychological intervention after OLT. Support, perhaps in the form of patient groups, could be offered to recently transplanted patients.

Although this study has shed light on a number of psychosocial aspects of OLT, its retrospective status means that the results can be regarded only as preliminary. Prospective studies are now required to establish how to identify patients at increased risk and to explore the best means of intervention to relieve immediate distress and to prevent long-term problems.

Ancillary