Prospective Study on Late Consequences of Subclinical Non-Compliance with Immunosuppressive Therapy in Renal Transplant Patients


*Corresponding author: Bart Maes,


In this prospective study we compared the incidence of late acute rejections (LAR) and changes in serum-creatinine over time between compliers and noncompliers with immunosuppressive therapy more than 1 year post transplantation and explored the relative contribution of non-compliance and other risk factors in the occurrence of LAR.

One hundred and forty-six adult renal transplant recipients were followed during a 5-year period. Patients were interviewed at the beginning of the study and categorized as non-compliers if they admitted to have skipped immunosuppressive medication on a regular basis during the previous 12 months. The occurrence of LAR during the follow-up period was recorded.

We identified 22.6% non-compliers of which 21.2% experienced a late acute rejection compared with 8% in the group of compliers at 5 years postinclusion (p < 0.05). Kaplan-Meier survival analysis showed a decreased rejection free time in non-compliers compared with compliers (p = 0.03). Non-compliant patients had a 3.2 higher risk of LAR (Cox regression analysis, p = 0.005). Non-compliers experienced a higher increase in serum-creatinine over time (Linear Mixed Models, p < 0.001).

Non-compliance in renal transplant patients more than 1-year post transplantation is associated with an increased risk for LAR and a higher increase in serum-creatinine during the following 5 years.


Short- and medium-term graft survival following renal transplantation has improved considerably over the past years as result of the use of new immunosuppressive agents (1–3). One-year graft survival after cadaver kidney transplantation is now greater than 90%. However, long-term graft loss has improved to a lesser extent. Death of patients with a functioning kidney allograft accounts for greater than 50% of late graft losses. In surviving patients, chronic rejection or chronic allograft nephropathy is the most common cause of graft failure (50–80%) (4). The etiology of chronic allograft nephropathy is not fully understood; however, acute rejection is an important risk factor. Especially late acute rejection episodes have been identified to have a strong correlation with late graft loss (5–8).

Little is known about the factors that cause late graft loss (5 years or more post transplant). Risk factors associated with long-term renal allograft survival reported in literature are race, donor age > 50 years, gender mismatch (female donors in male recipients), older recipient age, older donor age, HLA mismatching, delayed graft function, >1 acute rejection and high serum creatinine at 1 year post transplant (9–12).

Increasing evidence shows that also non-compliance has a negative impact on graft function, late acute-rejection episodes, organ loss and in some transplant populations death (7–9,13). Successful outcome in renal transplantation depends on the continued use of immunosuppressive medication. Life-long use of immunosuppressive drugs is necessary for good organ function and long-term graft survival (11,14–17). Despite advances in immunosuppressive therapy a major weakness in the therapeutic chain remains the patient's behavior. Non-compliance, also referred to as non-adherence, is a major problem in chronic health care and is a challenge for the health team.

Depending on the method and the operational definition used, the incidence of medication non-compliance in the adult renal transplant recipients ranges from 4.7% clinical non-compliance (or non-compliance assessed in relation to graft loss) to 53% subclinical non-compliance (or non-compliance assessed in the absence of a rejection episode or graft loss) (8,12,16,17). Several methods are described to assess non-compliance in transplant patients but no gold standard exists (19,20). Interviews and patient self-reports do not reveal patterns of compliance behavior and are found to underestimate the incidence of non-compliance. However for the assessment of large study populations they are a good method, easy to use in daily practice and very cost-effective (19,20).

Several studies on all types of organ transplantation show that compliance declines with time (1,21–23). Patients are more compliant in the early post transplant period and less compliant as times goes by. However, most studies only focus on the non-compliance behavior in the first year post-transplant. A retrospective study on the incidence, determinants and consequences of subclinical non-compliance with immunosuppressive medication in renal transplant patients published by our group in 1995 (8) suggested a correlation between non-compliance and the occurrence of late acute rejection. In that study we found that non-compliance is a risk factor for late acute rejection and results in lower graft survival. The aim of the present prospective study was to compare the incidence of late acute rejections (acute rejection more than 1 year post transplantation) as well as the changes in serum creatinine over time between compliers and non-compliers with immunosuppressive medication, and to explore the relative contribution of non-compliance and other known risk factors in the occurrence of these late acute-rejection episodes after renal transplantation. We also wanted to investigate if it was possible to identify potential non-compliers in daily clinical practice using self-report at an arbitrary time point after transplantation. Hypotheses are that by using self-report it is possible to identify non-compliers and find differences in time to the occurrence of late acute rejection and changes in serum creatinine over time.

Materials and Methods


One hundred and fifty patients from the Renal Transplant Program of the University of Leuven were asked by an independent investigator to participate in this study. These patients had already participated in a previous study published by our group where the impact of non-compliance on graft and patient survival was demonstrated retrospectively (8). In the present study we prospectively followed these patients over a period of 5 years (Table 1: research design). End points were late acute rejection, graft loss or death. Acute rejection was defined clinically as an unexplained increase in serum creatinine and confirmed by renal biopsy scored according to the Banff 1997 criteria. Late acute rejection (acute rejection more than 1 year post-transplant) was the occurrence of an acute rejection during the 5-year follow-up period. Follow up was complete by the end of 2000. Demographics and relevant clinical information were collected from patients' records.

Table 1.  Timeline research design Thumbnail image of


Patient selection and instruments used to measure non-compliance have been published already in our previous paper (8). In summary patients had to be transplanted for more than 1 year, being 18 years or older, literate, Dutch speaking and taking CsA as the main immunosuppressive drug. Live related transplants were excluded from this study. Before entering the study, informed consent was obtained. Of the 150 patients who met the inclusion criteria, four refused to participate, yielding a convenience sample of 146 patients. Immunosuppressive therapy consisted of methylprednisolone (David Bull Laboratories Pty Ltd, Mulgrave, Australia) and Cyclosporin A (Novartis, Basel, Switzerland), with or without azathioprine (Glaxo Wellcome, Dartford, UK). Methylprednisolone was initiated 4 h before transplantation with a single intravenous dose of 500 mg, followed by a second intravenous dose of 40 mg on day 1. Oral methylprednisolone was tapered progressively over the initial 6-month treatment period to a maintenance dose of 4 mg/day. The dose of Cyclosporin A was adjusted to obtain plasma 12-h trough levels of 100–250 ng/mL according to time after transplantation. Azathioprine was given at a dose of 2 mg/kg/day. First-line antirejection treatment consisted of supplemental corticosteroid therapy: 200 mg of intravenous prednisolon (Soludacortine, Merck-Belgolabo, Overijse, Belgium) for days 1 and 2, and 150 mg for days 3 and 4, followed by oral methylprednisolone (cfr supra) 80 mg at day 5 with daily tapering of 8 mg until the prerejection dose of methylprednisolone is achieved. Repeated rejection episodes were treated either by a renewed corticosteroid treatment (cfr supra) or by intravenous bolus injection of 5 mg of OKT3 (Ortho Biotech Inc., Raritan, NJ) for 10 days.


This study used self-reports and interviews to measure non-compliance. Self-report is described by Polit and Hungler (24) as a valid technique to measure non-compliance if asked by an investigator not belonging to the therapeutic team in a non-threatening way. It is known to underestimate the incidence of non-compliance. All interviews took place at the outpatient clinic by an independent investigator. At the same time we also measured other variables known as predictors for non-compliance behavior using validated and published methods (self-efficacy, self-care agency and situational knowledge). Based on the results of the self-report and validated by these variables, patients were categorized as non-compliers if they admitted to having skipped immunosuppressive medication on a regular basis during the previous 12 months (i.e. having missed several doses a month or taking ‘drug holidays’). Patients who stated that they had only skipped their medication on two or three occasions during the past year and evaluated themselves as taking their medication on a regular basis were considered compliant because no regular pattern of non-compliant behavior could be recognized. All interviews were discussed in a researcher meeting to assess and validate whether non-compliance on a regular basis had or had not occurred. On the basis of a joint decision, taken into account the results of the other tests, patients were assigned to either the compliant or the non-compliant group. As part of a broader study a stepwise logistic regression was performed to model determinants of being a compliant patient vs. not being a compliant patient (8). The model's validity is underscored by its adequate 78.6% correct classification rate of compliers vs. non-compliers and its high sensitivity rate of 95.9%. In addition, its false-positive rate of 19.8% (non-compliers who are in fact compliers) and false-negative rate of 40% (compliers who are in fact non-compliers) makes the model helpful as an alert model to identify potential non-compliers. The model is conservative and may assist clinicians in identifying potential non-compliers, which is clinically more pertinent then being able to identify compliers. In order to investigate if a single screening at a certain point in the transplant process is able to identify non-compliers and the long-term effects of this behavior, no retests were performed.

All interviews and tests were conducted during a scheduled clinic appointment at the outpatient clinic of the University Hospitals Leuven in Belgium. All patients were told that the aim of the study was to gain insight into their medication-taking behavior to improve the nursing care of renal transplant patients. All patients were assured that the data were treated as confidential. In order to facilitate a non-threatening climate and open communication, data were collected by two investigators not belonging to the therapeutic team. All interviews were performed in the same way, using standardized questions. Patients were asked if they had skipped immunosuppressive medication during the past year and if they could quantify the number of days this happened. Following, patients were requested to fill out the assessment of self-care agency and self-efficacy.

Any episode of biopsy-proven acute rejection occurring at any point in time after inclusion in the study during the 5-year follow-up period was defined as a late acute-rejection episode. Chronic rejection was defined as biopsy-proven histological evidence of a chronic rejection process according to the Banff 1997 criteria. Factors described in the literature that have an impact on graft survival were collected from patient records: age at time of transplantation, donor age, delayed graft function (or acute tubulus necrosis = no decrease in serum creatinine within 48 h post-transplant), number of acute rejections during the first year post transplantation with persistent elevation of creatinine, creatinine at 1 year post transplantation, and number of HLA-mismatches. Values of serum creatinine for all individual patients were recorded from the patients' medical records at 1 year post transplantation, at inclusion in the study, and each year during the 5-year follow-up period.

Statistical analysis

Data were entered and analyzed with SPSS windows 11 (Academic Service, Schoonhoven, the Netherlands) and SAS 6.12 (SAS/STAT User's Guide, Release 6.12 SAS Institute Inc., Raleigh, NC, USA). Before statistical analysis data were checked for normality. Mean values, standard deviations, median, percentiles and frequencies were calculated as appropriate and needed. Kaplan-Meier survival analysis was used to compare compliers and non-compliers with regard to time to occurrence of a late acute rejection. A regression analysis was performed of factors that were thought to have a significant influence on the incidence of acute rejection (Table 2). Time-dependent Cox regression analysis explored the relative contribution of relevant clinical factors in the occurrence of a late acute rejection after controlling for time dependency. Factors with p < 0.20 were consequently included in a multivariate Cox regression analysis. Because of the design of the study, patients were enrolled at different time points following transplantation. As a result it is not possible to calculate graft survival. To deal with this problem we looked at changes in serum creatinine over time as a measure for graft function. The Linear Mixed Effects Model for Longitudinal Data (25) was used to compare these changes in serum creatinine over time as a measure for graft function between compliers and non-compliers. With this technique it is possible to compare patient groups with different follow-up times.

Table 2.  Co-factors used in the regression analysis
Donor factors
Recipient factors:
 Age at time of transplantation
 Number of HLA-DR mismatches
 Creatinine at 1 year post transplantation
 Acute rejections within 1 year post transplantation with
  persistent elevation of serum creatinine
 Delayed graft function


All patients included in this study were Caucasian with a median age of 47 years (IQR: 19); 56% were males. Average time after transplantation was 4 years (range 1–18 years). The majority of the patients (78%) lived in a stable relationship with their partner or could rely on social support. Of the 146 patients, 33 were identified as non-compliant (22.6%) with their immunosuppressive therapy. No differences were found between compliers and non-compliers in terms of recipient age, donor age, sex, time after transplantation, serum creatinine at 1 year post transplantation, serum creatinine at inclusion in the study, delayed graft function, number of acute rejections within the first year post transplantation, and number of renal transplantation or HLA mismatches (Table 3). There were more patients with no social support in the non-compliant group compared with the compliant group (p = 0.028). Social support was defined as actual perceived social support, meaning being able to rely on a significant other. Significantly more non-compliers (21.2%) experienced late acute-rejection episodes compared with 8% in the group of compliers at 5 years post inclusion (p < 0.05).

Table 3.  Demographic characteristics compliers vs. non-compliers (n = 146)
VariableCompliers (77.4%)Non-Compliers (22.6%)p
  1. To convert serum creatinine in mg/dL to µmol/L, multiply by 88.4.

Age (years)45.83 (SD 12.41)46.72 (SD 13.03)n.s.
Sex (M/F)%53.9/46.166.7/33.3n.s.
Social Support (Y/N) (%)81.7/18.363.3/36.40.03
Time after transplantation (months)45 (IQR:30–67)58 (IQR:33–72)n.s.
Serum creatinine 1 year post-tx (mg/dL)1.78 (SD 0.74)1.57 (SD 0.50)n.s.
Serum creatinine at inclusion (mg/dL)1.77 (SD 0.72)2.00 (SD 0.79)n.s.
Donor age (year)32.73 (SD 16.2)34.13 (SD 13.5)n.s.
Delayed Graft function (%)10.69.1n.s.
Acute rejection within first year post-tx0.37 (SD 0.54)0.21 (SD 0.48)n.s.
HLA-DR mismatches (≥1) (%)52.343.8n.s.
Number of transplantations1.16 (SD 0.37)1.3 (SD 0.53)n.s.
PRA (T cell)%7.36.3n.s.
Self-care agency score93.41 (SD 10.45)88.41 (SD 10.84)0.03
Self-efficacy score3.72 (SD 0.21)3.56 (SD 0.15)0.048
Situational operational knowledge56.7 (29.3)42.3 (32.7)0.02

Kaplan-Meier survival analysis showed a decreased rejection free time in non-compliers compared with compliers (log rank: 4.55; p = 0.03) (Figure 1). Non-compliance (B: 1.21; RR: 3.34; p = 0.02), donor age (B: 0.06; RR: 1.06; p = 0.08), age at time of transplantation (B: – 0.06; RR: 0.94; p = 0.007) and creatinine at 1 year post transplantation (B: 0.48; RR: 1.62; p = 0.08) were included in the multivariate Cox regression analysis, again controlling for time dependency and time post transplantation. The ultimate model (– log likelihood: 116.1; Chi-square: 18.8; d.f.: 6; p = 0.005) included time post transplantation (RR: 0.98; p = 0.31), non-compliance (RR: 3.20; p = 0.04), age at time of transplantation (RR: 0.95; p = 0.04) and donor age (RR: 1.06; p = 0.06).

Figure 1.

Late acute rejection free time in non-compliers compared with compliers (Kaplan-Meier survival analysis, p = 0.03).

Changes in serum creatinine over time were used in this study as a measure of graft function. Non-compliant patients had a significantly higher increase in serum creatinine over time compared with compliant patients (Mixed Models, F:8.84, p < 0.001) (Figure 2).

Figure 2.

Changes in serum creatinine as a function of time: compliers vs. non-compliers (Linear Mixed Models Statistics, p < 0.001.


The aim of this study was to investigate the risk factors for late acute rejections and look for changes in serum creatinine over time between compliant and non-compliant patients as a measure of graft function. Non-compliance was measured at a single arbitrary point after transplantation using self-reporting.

A number of patients with late graft dysfunction and/or graft failure have acute rejection episodes occurring months to many years after transplantation. In this study we have analyzed the factors leading to late acute rejection, as these have not been thoroughly investigated in the past. In late acute rejections, the time of onset of the first rejection episode post transplantation varies between studies (5). We (8) reported in a retrospective study that patients with non-compliant behavior regarding their immunosuppressive medication have a worse graft survival and a higher incidence of late acute rejection. The present analysis shows a threefold increase in the risk of late acute rejection in non-compliant patients.

The reported incidence of late acute rejection episodes varies between studies from 2.5% to 38% (5,7). High and low incidences have been reported in studies where protocol biopsies have been employed (26). In our study late acute rejections were defined as a biopsy-proven rejection, resulting in an incidence of 21.2% for non-compliant and 8% for compliant patients.

A variety of factors leading to the occurrence of a late acute rejection have been described (7,8,11,15,27). These include race, non-compliance, sub optimal Cyclosporin A levels and HLA mismatches. We did not include race in our study because all our patients were Caucasian. Non-compliant patients were more at risk to develop a late acute rejection than compliant patients. Cyclosporin A levels were not included in this study because a major cause of suboptimal Cyclosporin A levels is non-compliance (8). There was no difference in HLA mismatch between patients with or without late acute rejections. All patients were transplanted with kidneys allocated according to the matching and allocation criteria of Eurotransplant, meaning that kidneys were accepted with a minimum of two matches with a priority for HLA-DR matching. It is reported that patients with a late acute rejection had a significantly lower mean age compared with those without rejection episodes (5). This may be explained by stronger immunological activity in younger patients. We can confirm these findings.

In our first paper we showed retrospectively that non-compliance has an impact on kidney graft survival (8). In our prospective cohort it was not possible to calculate graft survival because of differences in time post transplant between patients at entry in the study. Therefore we used Linear Mixed Models to look at differences in serum creatinine over time. A higher serum creatinine is an indication for lower graft function. In this study we found significantly higher changes in serum creatinine over time in patients who report non-compliant behavior with their immunosuppressive therapy.

We conclude that perceived, self-reported, non-compliance in renal transplant patients more than 1 year post transplant is associated with an increased risk for late acute rejection during the following 5 years. Non-compliance seems to be the most important risk factor in the occurrence of late acute rejection, a known risk factor for chronic rejection and consequent graft loss in renal transplant recipients. Non-compliant patients have a higher increase in serum creatinine over time compared with compliant patients. Non-compliance behavior should be measured in the post-transplant period to identify patients at risk for late acute rejections and higher increase in serum creatinine over time.


This study was partially supported by a research grant from Fujisawa GmbH.