From dialysis to transplantation: a 5-year longitudinal study on self-reported quality of life

Patients with end stage renal disease (ESRD) dependent on dialysis have reduced quality of life not only compared to the general population [1‐4], but also to patients with other chronic diseases [5, 6]. Kidney transplantation (RTX) is considered to be the optimal treatment for patients with ESRD with survival advantage compared to dialysis [7, 8]. Cross-sectional studies have indicated improvement in physical, mental and social aspects of health related quality of life (HRQOL) in patients after RTX compared to dialysis treatment [9, 10]. However, there is a scarcity of longitudinal studies addressing HRQOL during treatment shift from dialysis to RTX. The few studies suggest improved HRQOL, but most have a short follow-up time and a low number of patients [11‐13]. The majority of the studies have used only generic HRQOL measures, and the limited numbers of studies using kidney-specific questionnaires included few patients with short follow-up [4, 14, 15]. Furthermore, perceptions of mental and physical aspects of HRQOL may change with time after transplantation [16‐18]. We postulate that HRQOL will improve in the transition from dialysis to RTX, and that this applies not only for generic, but also for disease-specific issues.

Whether kidney transplanted patients will obtain the same HRQOL as that of the general population is unclear, as previous observations are inconsistent, with some reporting inferior [19‐21], and others similar HRQOL scores in RTX patients and the general population [17, 22]. As renal transplanted patients may face new encounters such as side effects of a heavy drug regimen, infections and deteriorating graft function [23], we expect self-perceived HRQOL to be lower in RTX patients than in the general population.

Thus, the aims of the current longitudinal cohort study were to assess changes in HRQOL in the transition from dialysis to renal transplantation, and to compare HRQOL in transplanted patients with that of the general population.

Between August 2005 and February 2007, a total of 301 patients from 10 different hospitals in Norway were included in a cross-sectional study. All patients >18 years receiving either hemodialysis (HD) or peritoneal dialysis (PD) for > two months were invited to participate [2]. Cognitive dysfunction, major psychiatric disorder and inadequate Norwegian language skills were exclusion criteria. Hospitalization during the investigation period led to exclusion, but patients could be enrolled four weeks or more after hospital discharge if they were clin0069cally stable. Close to one third of the prevalent dialysis population from both urban and rural areas were enrolled in the study. Study nurses and physicians had been trained in applying the study instruments to the patients. The questionnaires were answered during a regular HD session or during a scheduled visit for the PD patients. Further details regarding this study are given elsewhere [2].

All included patients from the first study were invited to participate in the follow-up study, both patients who had been transplanted as well as patients who remained in dialysis (Figure 1). Inclusion and exclusion criteria were unchanged from the original study. The transplanted patients answered the questionnaires during a regular visit at the renal outpatient clinic. Clinical and demographic data were collected from hospital charts and/or direct questioning of the patients. Records of RTX were obtained from the Norwegian Renal Registry [24]. Comorbidity was assessed using the modified Charlson comorbidity index (CCI) [25]. The modified CCI is validated for dialysis patients [25] and kidney transplanted patients [26], and is a composite score of age and 17 weighted comorbid conditions including coronary artery disease, congestive heart failure, cerebrovascular disease, diabetes, malignancy and chronic pulmonary disease. In the current study all types of skin cancers, including basal cell carcinomas, were considered as tumors and given two points. Hepatitis without cirrhosis was regarded as mild liver disease. Diabetes as a comorbid condition scored one point, diabetes as cause of ESRD scored two points. CCI was also calculated without including age to evaluate the effect of age as a separate factor in statistical analyses. Data on HRQOL in the general population were collected in the Survey of Level of Living 2002 [27].

The Regional Committees for Medical and Health Research Ethics in Norway approved the protocols (2005/2010), and concession was obtained from the National Data Inspectorate. The study was accomplished according to the Helsinki Declaration [28]. Written and oral information were provided, and signed consent was obtained.

From the original dialysis population of 301 patients, 110 patients had been transplanted, while 142 died during the 55 months of follow-up (Figure 1). The clinical and demographic characteristics of the transplanted patients are given in Table 1. Time since transplantation was 40.3 ± 14.4 months. Standard immunosuppressive regimen with cyclosporine or tacrolimus and mycophenolate mofetil (MMF) combined with (86/110) or without (3/110) prednisolone was used in the majority of the patients. Everolimus combined with MMF and prednisolone was used in 11/110 patients, and other different combinations in 10/110 patients.

An important novel observation in the present longitudinal study was the overall improvement in the kidney-specific HRQOL in transplanted patients. Kidney-specific data have so far been lacking or insufficient. Improvement would be expected shortly after transplantation, as the kidney function is retrieved. Our observation showed that this improvement was present after more than three years post-RTX, and specifically the perception of complaints related to kidney failure. The majority of the current RTX patients had mild to moderate chronic kidney disease, i.e. stage 2–3, and with fewer symptoms than would be expected with more advanced renal failure. The two domains, i.e. effect of kidney disease (EKD) and burden of kidney disease (BKD) had numerically the most pronounced improvement after RTX. BKD, which addresses how kidney disease may interfere with life, cause frustrations, and burdening the family, and EKD, being more related to daily-life restrictions and worries, were in the present study lower than what have been observed in a group of never-transplanted CKD (chronic kidney disease) patients with equivalent GFR [37]. In the latter study BKD scores were 85 compared to 67 in our study, while EKD scores were 92 and 82 in the never-transplanted CKD and the RTX patients, respectively. On the other hand, our observation of the HRQOL scores in the longitudinal cohort study was in accordance with the finding in a cross-sectional study of kidney transplanted patients with GFR in the same range, i.e. GFR 30–60 ml/min/1.73 m², with unadjusted EKD scores 88 and BKD 78 [38]. These observations suggest that despite similar GFR, renal transplanted may actually fare poorer than non-transplanted patients. This assumption can only be corroborated by findings from cross-sectional studies as it is unlikely that we will get prospective longitudinal HRQOL studies including patients from CKD stage 2–3 and follow them for years until after RTX.

The KDQOL-SF has been assessed as a valid and reliable tool in RTX patients as well as in other CKD patients [32]. The internal consistency has been found to be particularly high for the domains EKD, BKD and symptoms in RTX patients [32], while some domains may be less relevant as they contain questions specifically for dialysis patients [29]. Previous experience with the use of KDQOL-SF is, however, limited in RTX patients. As far as we know, only two longitudinal studies using KDQOL-SF have been done, and with a small number of patients [14, 15]. One of the studies was published as a short abstract, and a small amount of data was revealed. Only 30 transplanted patients were included, all with kidney from living donors [15]. The other study included a highly selected small group of young transplanted patients attending a training program [14]. There are two cross-sectional studies that have used KDQOL and compared HRQOL in dialysis patients with that of RTX patients [39, 40]. While one of the studies was small and addressed ethnicity differences rather than treatment shift [40], the other large study had less than 35% response rate in the RTX group [39], which may lead to selection bias. The current study therefore may contribute noticeably to increased knowledge on kidney-specific HRQOL in the transition from dialysis to transplantation.

Patients over 65 years of age actually perceived poorer HRQOL in the domain “quality of social interaction” in the conversion from dialysis to RTX, while younger patients perceived improvement. One possible explanation to this finding could be that older patients may become more socially isolated after RTX than younger. Older patients may have less personal contacts, a smaller network of family and friends and thus participate more infrequent in social activities than younger [41, 42]. Social interaction experienced during dialysis sessions should perhaps not be underestimated.

Dialysis vintage affected the change in the domain burden of kidney disease during follow-up, with a larger improvement in patients with the longest time in dialysis. This may not be surprising, as the patients with longest dialysis vintage may have felt more burdened, and therefore perceived that RTX relieved the frustrations and interference more than in those with shortest dialysis vintage. Divergent results regarding associations between dialysis vintage and HRQOL after RTX are reported, and only for generic domains, i.e. physical domains [18, 43, 44].

One item which has gained little attention in previous studies is the question “Overall, how would you rate your health” in the kidney-specific part of the questionnaire. As far as we know, only one study has published data on this question [39], and the authors surprisingly found the score to be lower in transplanted than in dialysis patients, with no apparent explanation. In contrast to that finding, the current study showed substantially improvement after RTX. This is in accordance with our clinical experience with RTX patients, and actually confirms that beyond the survival benefit, patient-related outcome is enhanced after RTX.

The changes in the generic HRQOL domains were less impressive than the changes observed in the disease-specific domains. The present study actually contradicted previous reports that generic HRQOL is perceived substantial higher in RTX than in dialysis patients [22, 45]. However, the majority of previous studies are cross-sectional, and therefore selection bias cannot be excluded. Furthermore, adjustments for other variables were rarely undertaken in these studies. When comparing dialysis patients on the waiting list with renal transplanted patients, the difference in HRQOL was less pronounced and quite comparable to our study [20]. A possible explanation for the difference in HRQOL in the different studies could be that the majority of patients in the cross-sectional studies would never be considered for RTX due to comorbidity or high age [22, 45]. Moreover, the only previous longitudinal study (n = 20) using KDQOL-SF [14] actually reported change in generic HRQOL from dialysis to transplantation in the same range as the present study.

We cannot rule out the possibility of a time effect contributing to the discrepancies of findings in our study and in some previous studies, as those studies comparing HRQOL changes from dialysis to transplantation have been obtained in the early post-transplantation period [4, 11, 17]. Less is known about HRQOL after the initial “honeymoon” period in the transplanted patients. In the present study, PCS was perceived lower in patients who had been transplanted for more than three years compared to those with shorter time, but no such difference was found for the kidney-specific items. It has been suggested that HRQOL in kidney transplanted patients might follow a U-shaped pattern with the highest score the first year and then after five years [46], but this time effect needs to be further explored.

Statistical significant difference in HRQOL may not necessarily be perceived as clinical important change for the patient. Clinically meaningful changes in HRQOL may vary with the disease, level of severity, gender, cultural background, socioeconomic status and other factors [47]. Furthermore, one cannot rule out that the greater change in the most impaired patients could actually reflect a statistical phenomenon, the regression to the mean, rather than a clinical alteration. Determining clinically meaningful change has been a matter of substantial debate, and several methods, either anchor-based or distribution-based, have been proposed to define and calculate this dimension, as described by Crosby et al. in a comprehensive review [47]. Anchor-based methods compares changes with a reference, an anchor, which can be objective measures such as response to treatment, or subjective measures such as patient-reported global ratings of change in health status. Distribution-based methods are based on statistical properties of the scale and include effect sizes. One of the most common used distribution-based tools is Cohen’s d, where the effect size (ES) is calculated from the change in mean scores divided by the standard deviation [34]. Cohen’s guidelines suggests that ES of 0.2 - 0.5 reflect a small effect, 0.5 – 0.8 a moderate effect and > 0.8 a large effect. There are no established thresholds for clinical important changes for the KDQOL-SF questionnaire, neither for dialysis nor kidney transplanted patients. In the current study, we chose pragmatically the distribution-based method of half a standard deviation [33] to define the cut-off clinical relevant change. This is equivalent to Cohen’s d = 0.50. To evaluate whether the method used in the current study represents the “true” threshold for these patients, more studies are needed, also combining anchor-based and distribution based methods.

Most patients in dialysis may expect their lives to change dramatically after transplantation, and thus overestimate the benefits of the transplantation [48]. Actually, the generic HRQOL scores in the current study were overall much lower in the RTX patients then in the general population for all domains in SF-36 except for bodily pain and mental health. Our finding is in keeping with a previous Norwegian study of young transplant recipients [49], and with observations in other patients with chronic diseases [50, 51]. The reduced HRQOL in RTX patients compared to the general population may reflect that these patients although transplanted still have a chronic kidney disease, often associated comorbidity, and also side effects of the medication, anxiety of rejection, malignancy, and infections.

The longitudinal design, the fairly large sample size and the fact that none of the patients were lost to follow-up strengthen the present study. Data quality was good with less than 2% missing data in SF-36. The transplantation recruitment procedure was similar in all dialysis units, and transplantations took place in one center. Some particular aspects may limit the generalizability of our results as very few patients were non-Caucasian, and the time in dialysis before RTX was fairly short. The inclusion criteria may have led to selection bias towards healthier patients as patients with cognitive dysfunction, drug abuse or hospitalized patients were excluded. although they were considered representative of the Norwegian dialysis population [2], Evaluation of response shift was not addressed when designing the study. This is a phenomenon suggesting that the self-perceived perceptions of HRQOL may be incited by changes in health, e.g. transplantation, which may then lead to changes in internal standards of measurement, priorities or in the conceptualization of perceived HRQOL [52].

While the kidney-specific HRQOL improved statistically and clinically in the transition from dialysis to transplantation, the changes in generic HRQOL items were not large enough to be regarded as clinically important change. Furthermore, RTX patients did not obtain HRQOL similar to that of the general population. We suggest that patient-related outcomes such as kidney-specific and generic HRQOL should be used routinely in RTX patients in the clinical setting, and not only for research purpose. This will provide important information about the well-being of these patients. Health professionals should be aware of the reduced HRQOL in patients with ESRD, which may persist even after RTX, and focus on how HRQOL could be improved. We recommend that not only should HRQOL, including kidney-specific items, be measured in patients with CKD 4–5, but also annually after the initiation of dialysis. Repeated measurements should also be undertaken after the conversion to RTX. Future clinical trials with longitudinal design and repeated measurements of HRQOL are required to better describe the trajectory after renal transplantation.