A useful scoring system for the prediction and management of delayed graft function following kidney transplantation from cadaveric donors
2014; Elsevier BV; Volume: 86; Issue: 6 Linguagem: Inglês
10.1038/ki.2014.188
ISSN1523-1755
AutoresM. Chapal, Florent Le Borgne, Christophe Legendre, Henri Kreis, Georges Mourad, Valérie Garrigue, Emmanuel Morélon, Fanny Buron, Lionel Rostaing, Nassim Kamar, M. Kessler, Marc Ladrière, Jean‐Paul Soulillou, Katy Trébern-Launay, Pascal Daguin, Lucile Offredo, Magali Giral, Yohann Foucher,
Tópico(s)Organ Donation and Transplantation
ResumoDelayed graft function (DGF) is a common complication in kidney transplantation and is known to be correlated with short- and long-term graft outcomes. Here we explored the possibility of developing a simple tool that could predict with good confidence the occurrence of DGF and could be helpful in current clinical practice. We built a score, tentatively called DGFS, from a French multicenter and prospective cohort of 1844 adult recipients of deceased donor kidneys collected since 2007, and computerized in the Données Informatisées et VAlidées en Transplantation databank. Only five explicative variables (cold ischemia time, donor age, donor serum creatinine, recipient body mass index, and induction therapy) contributed significantly to the DGF prediction. These were associated with a good predictive capacity (area under the ROC curve at 0.73). The DGFS calculation is facilitated by an application available on smartphones, tablets, or computers at www.divat.fr/en/online-calculators/dgfs. The DGFS should allow the simple classification of patients according to their DGF risk at the time of transplantation, and thus allow tailored-specific management or therapeutic strategies. Delayed graft function (DGF) is a common complication in kidney transplantation and is known to be correlated with short- and long-term graft outcomes. Here we explored the possibility of developing a simple tool that could predict with good confidence the occurrence of DGF and could be helpful in current clinical practice. We built a score, tentatively called DGFS, from a French multicenter and prospective cohort of 1844 adult recipients of deceased donor kidneys collected since 2007, and computerized in the Données Informatisées et VAlidées en Transplantation databank. Only five explicative variables (cold ischemia time, donor age, donor serum creatinine, recipient body mass index, and induction therapy) contributed significantly to the DGF prediction. These were associated with a good predictive capacity (area under the ROC curve at 0.73). The DGFS calculation is facilitated by an application available on smartphones, tablets, or computers at www.divat.fr/en/online-calculators/dgfs. The DGFS should allow the simple classification of patients according to their DGF risk at the time of transplantation, and thus allow tailored-specific management or therapeutic strategies. Within the past decade, while the frequency of acute allograft rejection episodes has dramatically decreased under modern immunosuppressive regimens,1.Miller J. Mendez R. Pirsch J.D. et al.Safety and efficacy of tacrolimus in combination with mycophenolate mofetil (MMF) in cadaveric renal transplant recipients. FK506/MMF Dose-Ranging Kidney Transplant Study Group.Transplantation. 2000; 69: 875-880Crossref PubMed Scopus (162) Google Scholar the incidence and severity of delayed graft function (DGF) has remained stable. The DGF is generally defined by the need for dialysis within the first seven days post transplantation,2.Yarlagadda S.G. Coca S.G. Garg A.X. et al.Marked variation in the definition and diagnosis of delayed graft function: a systematic review.Nephrol Dial Transplant. 2008; 23: 2995-3003Crossref PubMed Scopus (282) Google Scholar its frequency range from 25 to 50%3.Perico N. Cattaneo D. Sayegh M.H. et al.Delayed graft function in kidney transplantation.Lancet. 2004; 364: 1814-1827Abstract Full Text Full Text PDF PubMed Scopus (759) Google Scholar possibility explained by the lack of a single definition and/or the more widespread use of expanded criteria donors (ECDs) or non-heart-beating donors.3.Perico N. Cattaneo D. Sayegh M.H. et al.Delayed graft function in kidney transplantation.Lancet. 2004; 364: 1814-1827Abstract Full Text Full Text PDF PubMed Scopus (759) Google Scholar DGF is known to be associated with a lower one-year post transplantation renal function, a decreasing long-term graft and recipient survival, and an increasing patient management cost.4.Tapiawala S.N. Tinckam K.J. Cardella C.J. et al.Delayed graft function and the risk for death with a functioning graft.J Am Soc Nephrol. 2010; 21: 153-161Crossref PubMed Scopus (143) Google Scholar,5.Yarlagadda S.G. Coca S.G. Formica Jr., R.N. et al.Association between delayed graft function and allograft and patient survival: a systematic review and meta-analysis.Nephrol Dial Transplant. 2009; 24: 1039-1047Crossref PubMed Scopus (537) Google Scholar Reducing the incidence of DGF by controlling risk factors related to both donors and recipients is among the most beneficial strategies.6.Giral M. Bertola J.P. Foucher Y. et al.Effect of brain-dead donor resuscitation on delayed graft function: results of a monocentric analysis.Transplantation. 2007; 83: 1174-1181Crossref PubMed Scopus (51) Google Scholar,7.Schnuelle P. Gottmann U. Hoeger S. et al.Effects of donor pretreatment with dopamine on graft function after kidney transplantation: a randomized controlled trial.JAMA. 2009; 302: 1067-1075Crossref PubMed Scopus (200) Google Scholar Providing drugs is also promising either in pre-clinical studies8.Arumugam T.V. Shiels I.A. Strachan A.J. et al.A small molecule C5a receptor antagonist protects kidneys from ischemia/reperfusion injury in rats.Kidney Int. 2003; 63: 134-142Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar, 9.Kelly K.J. Williams Jr., W.W. Colvin R.B. et al.Antibody to intercellular adhesion molecule 1 protects the kidney against ischemic injury.Proc Natl Acad Sci USA. 1994; 91: 812-816Crossref PubMed Scopus (460) Google Scholar, 10.Knight S.F. Kundu K. Joseph G. et al.Folate receptor-targeted antioxidant therapy ameliorates renal ischemia-reperfusion injury.J Am Soc Nephrol. 2012; 23: 793-800Crossref PubMed Scopus (20) Google Scholar, 11.Thurman J.M. Royer P.A. Ljubanovic D. et al.Treatment with an inhibitory monoclonal antibody to mouse factor B protects mice from induction of apoptosis and renal ischemia/reperfusion injury.J Am Soc Nephrol. 2006; 17: 707-715Crossref PubMed Scopus (104) Google Scholar or using already well-known immunosuppressive therapy such anti-thymocyte globulin (ATG).12.Kaden J. May G. Strobelt V. et al.Intraoperative T-cell depletion prior to completion of anastomoses by high-dose single ATG bolus as a new approach to improve long-term results after kidney transplantation.Transplant Proc. 1997; 29: 344-347Abstract Full Text PDF PubMed Scopus (28) Google Scholar, 13.Kyllonen L.E. Eklund B.H. Pesonen E.J. et al.Single bolus antithymocyte globulin versus basiliximab induction in kidney transplantation with cyclosporine triple immunosuppression: efficacy and safety.Transplantation. 2007; 84: 75-82Crossref PubMed Scopus (72) Google Scholar, 14.Turunen A.J. Lindgren L. Salmela K.T. et al.Association of graft neutrophil sequestration with delayed graft function in clinical renal transplantation.Transplantation. 2004; 77: 1821-1826Crossref PubMed Scopus (30) Google Scholar Nevertheless, these therapeutic strategies are still under debate.15.Haririan A. Morawski K. Sillix D.H. et al.Induction therapy with basiliximab versus Thymoglobulin in African-American kidney transplant recipients.Transplantation. 2005; 79: 716-721Crossref PubMed Scopus (49) Google Scholar, 16.Mourad G. Rostaing L. Legendre C. et al.Sequential protocols using basiliximab versus antithymocyte globulins in renal-transplant patients receiving mycophenolate mofetil and steroids.Transplantation. 2004; 78: 584-590Crossref PubMed Scopus (151) Google Scholar, 17.Sollinger H. Kaplan B. Pescovitz M.D. et al.Basiliximab versus antithymocyte globulin for prevention of acute renal allograft rejection.Transplantation. 2001; 72: 1915-1919Crossref PubMed Scopus (130) Google Scholar, 18.Ulrich F. Niedzwiecki S. Pascher A. et al.Long-term outcome of ATG vs. basiliximab induction.Eur J Clin Invest. 2011; 41: 971-978Crossref PubMed Scopus (23) Google Scholar To propose strategies to prevent the occurrence of a DGF for a better management of kidney transplantation recipients, it is of importance to screen as far as possible patients at-risk of DGF. Several DGF-scoring systems have been proposed within the last few years. First published in 200319.Irish W.D. McCollum D.A. Tesi R.J. et al.Nomogram for predicting the likelihood of delayed graft function in adult cadaveric renal transplant recipients.J Am Soc Nephrol. 2003; 14: 2967-2974Crossref PubMed Scopus (177) Google Scholar and then refined in 2010,20.Irish W.D. Ilsley J.N. Schnitzler M.A. et al.A risk prediction model for delayed graft function in the current era of deceased donor renal transplantation.Am J Transplant. 2010; 10: 2279-2286Crossref PubMed Scopus (275) Google Scholar from the United State Renal Data System registry, Irish et al. proposed a predictive score that could be calculated using 18 parameters at the time of transplantation (human leukocyte antigen (HLA) mismatch, donor serum creatinine, donor age, donor weight, peak panel reactive antibodies (PRAs), recipient previous transplant, cold ischemia time (CIT), warm ischemia time, recipient race, pre-transplant dialysis, single organ transplant, donor cause of death, non-heart-beating donor, donor hypertension, recipient gender, diabetic recipient, pre-transplant transfusion, and recipient body mass index (BMI)), with an area under the receiver-operating characteristic curve (AUC) at 0.70. Despite the high quality of this methodology, this predictive model was developed and validated from North American recipients, whereby the patients’ profile differs substantially from European recipients. For instance, according to the OPTN & SRTR annual report of 2011, 30% of kidney transplantations were from living donors, 25% of the deceased donors were older than 50 years, 62% of US recipients received a depleting induction therapy, and 17% received no induction therapy compared with 13%, 58%, 53%, and 8%, respectively in the DIVAT (Données Informatisées et VAlidées en Transplantation) cohort that gathers 30% of the French kidney recipients throughout France. No decision threshold is proposed to use the Irish score to classify patients according to their DGF risk. Jeldres et al.21.Jeldres C. Cardinal H. Duclos A. et al.Prediction of delayed graft function after renal transplantation.Can Urol Assoc J. 2009; 3: 377-382PubMed Google Scholar proposed a simpler but equally accurate scoring system (six variables, AUC=0.74); however, these results were based on a monocentric study of North American recipients transplanted since 1979. Finally, both scoring systems,20.Irish W.D. Ilsley J.N. Schnitzler M.A. et al.A risk prediction model for delayed graft function in the current era of deceased donor renal transplantation.Am J Transplant. 2010; 10: 2279-2286Crossref PubMed Scopus (275) Google Scholar,21.Jeldres C. Cardinal H. Duclos A. et al.Prediction of delayed graft function after renal transplantation.Can Urol Assoc J. 2009; 3: 377-382PubMed Google Scholar did not take into account the induction therapy. We thus proposed to develop a complementary DGF score (DGFS), from DIVAT with patients transplanted since 2007. The DGFS allows to predict, with a good confidence, which patients will be at high risk of DGF according to only five variables at the transplantation time and possibly use it as a decision-making tool to decide which induction therapy could be prescribed according to the individual DGF risk profile. Qualitative and quantitative variables are described according to the training and validation samples in Tables 1 and 2. As the allocation was random, the two groups were similar. In the training sample used for the scoring system definition, the mean age of recipients was 51.9 (±13.2) years and 60.4% were men. For 30.2% of the recipients, the primary indication for renal transplantation was a possible recurrent disease of the renal graft. For 20.6% and 4.5% of them, it was respectively a second transplantation and a third or more transplantation. The mean duration in dialysis before transplantation was 4.1 (±3.9) years. Historical anti-HLA class I and II PRAs were detectable in 37.8% and 36.0% of the patients, respectively. Equal to or more than five HLA incompatibilities on A-B-DR loci were observed in 10.7% of the recipients. The mean donor age was 51.9 (±15.6) years, 58.6% were men and 57.9% were dead because of a vascular cause. The mean terminal serum creatinine was 91.0 (±60.7)μmol/l. The mean duration of CIT was 19.2 (±7.0)h. A total of 25.4% of the recipients had a DGF.Table 1Recipient, donor, and kidney transplant continuous characteristics (minimum, maximum, mean, and standard deviation) at time of transplantation according to both training and validation samplesTraining (n=1238)Validation (n=606)MinimumMaximumMeans.d.MinimumMaximumMeans.d.Cold ischemia time (hours)6.0058.6219.217.026.3347.3018.996.80Recipient age (years)19.0084.0051.9213.1919.0084.0051.9612.77Donor age (years)4.0088.0051.9115.601.0090.0052.9240.72Body mass index (kg/m2)14.2046.5724.074.4414.8738.5824.234.16Donor serum creatinine (μmol/l)20.00999.0091.0260.6720.00335.0086.3944.93Duration in dialysis (years)0.0236.604.073.920.0434.023.793.64 Open table in a new tab Table 2Recipient, donor, and kidney transplant continuous characteristics (effective and percentage) at time of transplantation according to both training and validation samplesTraining (n=1238)Validation (n=606)EffectivePercentage (%)EffectivePercentage (%)Recipient gender Male74860.438463.4 Female49039.622236.6Donor gender Male72058.638162.9 Female50941.422537.1Rank of the kidney transplantation First92774.946476.6 Second25520.612019.8 Third or more564.5223.6Donor cause of death Vascular causes71557.934356.6 Other causes52042.126343.4Immunosuppressive induction therapy ATG68455.333555.3 No ATG55444.727144.7History of diabetes Yes15712.78213.5 No108187.352486.5History of hypertension Yes100080.848379.7 No23819.212320.3History of cardiovascular events Yes44836.223238.3 No79063.837461.7History of dyslipidemia Yes38431.018230.0 No85469.042470.0History of hepatitis B Yes816.5345.6 No115793.557294.4History of neoplasia Yes1239.97412.2 No111590.153287.8Anti-class I panel reactive antibodies (historic pic) Detectable (>0%)46837.822537.1 Undetectable (0%)77062.238162.9Anti-class II panel reactive antibodies (historic pic) Detectable (>0%)44636.022336.8 Undetectable (0%)79264.038363.2Initial kidney disease With risk of recurrence on graft37330.219131.6 Without risk of recurrence on graft86369.841468.4HLA mismatches at HLA-A+B+DR 0 to 4110589.354890.4 5 to 613310.7589.6Abbreviations: ATG, anti-thymocyte globulin; HLA, human leukocyte antigen. Open table in a new tab Abbreviations: ATG, anti-thymocyte globulin; HLA, human leukocyte antigen. In the univariate analyses detailed in the web (Supplementary Table S1 online), possible risk factors of DGF (P<0.20) were identified: donor serum creatinine, recipient age, duration of dialysis before transplantation, donor gender, number of previous transplants, immunosuppressive induction therapy, history of cardiovascular events (except HTA) and dyslipidemia, donor treatment with epinephrine and duration of CIT, donor age, and recipient BMI. Without taking into account the other risk factors, the CIT was the main predictor of DGF. More precisely, the odds ratio (OR) was multiplied by 1.05 for each hour (95% confidence interval (CI)=(1.04, 1.07)). The corresponding AUC was 0.60 (95% CI=(0.56, 0.64)). Download .doc (.15 MB) Help with doc files Supplementary Tables The results of the final model are presented in Table 3. Five independent explicative variables seemed significantly associated with the risk of DGF. As expected, the probability of DGF increased with the CIT (OR=1.06, P<0.0001). High recipient BMI was also associated with higher DGF probability (P=0.0004). An increase in donor age of 10 years was associated with an OR multiplied by 1.16 (P=0.0014). Higher risk of DGF was observed when induction therapy was not ATG-based (OR=1.70, P=0.0001), and if donor creatinine levels were higher than 108μmol/l (OR=1.76, P=0.0004). The induction with ATG can be analyzed according to two different protocols depending on inter-center variability: delayed (n=289) or non-delayed (n=160) introduction of calcineurin inhibitors. We did not distinguish between these treatment protocols in the score as the corresponding risks of DGF were not significantly different (P=0.7704).Table 3Results of the final logistic regression (n=1222, 16 observations removed because of missing data)log odds ratio (±s.d.)Odds ratio95% confidence intervalP-valueCold ischemia time (hours)0.058 (±0.010)1.06(1.04, 1.08) 108μmol/l (*)0.565 (±0.160)1.76(1.29, 2.41)0.0004No anti-thymocyte globulin (*)0.531 (±0.138)1.70(1.30, 2.23)0.0001The score is the sum of the logarithm of the odds ratios multiplied by the values of the variables. For the variables indexed by *, the values are 1 if the patient had no depleting induction or if his/her donor had a creatinemia level higher than 108μmol/l, and 0 otherwise. Open table in a new tab The score is the sum of the logarithm of the odds ratios multiplied by the values of the variables. For the variables indexed by *, the values are 1 if the patient had no depleting induction or if his/her donor had a creatinemia level higher than 108μmol/l, and 0 otherwise. The corresponding DGFS can be calculated as follows: (1) by multiplying the logarithm of the OR and the values of the explicative variables, (2) by summing the five previous numbers, (3) by subtracting 3.546 to the previous number, and (4) by dividing the previous number by 0.626. For example, considering a patient with a BMI at 25kg/m2 (25 × 0.054), with 12h of CIT (12 × 0.058), with induction therapy based on lymphocyte-depleting agent (0 × 0.531), with a graft from a 50-year-old donor (50 × 0.015) for which the last serum creatinine level was 120μmol/l (1 × 0.565), the sum is 3.361 and the DGFS value is −0.296 (3.361–3.546)/0.626). The risk of DGF increases with the DGFS value. Useful advantages in terms of interpretations are associated with the standard normal distribution of this score (mean at 0 and standard deviation at 1). A patient with a positive DGFS can be considered at-risk of DGF higher than the mean risk. In contrast, a patient with a negative score can be considered at-risk of DGF lower than the mean. About 70% of individuals are within one s.d., i.e., with a score value between −1 and +1. About 95% of individuals are within two standard deviations, i.e., with a score value between −2 and +2. In addition to these results, it is of primary importance to propose thresholds for clinical intervention decision making. This estimation is by definition subjective since it depends on the consequent interpretations of such decision.22.Irwin R.J. Irwin T.C. A principled approach to setting optimal diagnostic thresholds: where ROC and indifference curves meet.Eur J Intern Med. 2011; 22: 230-234Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar Figure 1 illustrates the positive and negative predictive values regarding all the possible threshold values in order to help clinicians to define the most relevant decision depending on their context. One can also propose two generic decision rules. The percentage of patients without DGF was around 88% if the DGFS was lower than −0.50. For patients with a DGFS lower than this threshold, clinicians have a 12% chance to incorrectly predict DGF. Thus, one could use this threshold with a good confidence in order to identify patients with a low risk of DGF (32% of recipients had a DGFS lower than −0.50). In contrast, the percentage of patients with DGF was around 50% if the DGFS was higher than 1.20. One could use this threshold in order to identify high-risk patients (11% of the patients had a DGFS higher than 1.20). Importantly, there is one-in-two chance to incorrectly predict DGF in patients with a DGFS higher than 1.20. As a consequence, patients with a DGFS between −0.5 and 1.20 can be considered at medium risk. The characteristics of the DGFS are described in Table 4 according to the three groups, i.e., low-, medium-, and high risk of DGF.Table 4Description of the variables included in the delayed graft function score from the training sample according to the three delayed graft function score-strata, i.e., low risk with delayed graft function score lower than 0.5, medium risk with delayed graft function score in-between −0.5 and 1.2, and high risk with delayed graft function score>1.2Low risk of delayed graft function (n=782)Medium risk of delayed graft function (n=1392)High risk of delayed graft function (n=280)MinimumMaximumMeans.d.MinimumMaximumMeans.d.MinimumMaximumMeans.d.Cold ischemia time (hours)6.0030.5015.134.386.0041.2219.686.2011.1858.6228.137.68Donor age (years)4.0079.0044.3916.0914.0084.0054.4614.1514.0084.0059.4112.79Body mass index (kg/m2)14.2034.0621.713.3915.0143.1624.754.1717.7646.5727.265.16EffectivePercentage (%)EffectivePercentage (%)EffectivePercentage (%)Donor serum creatinine level >108μmol/l465.8835425.4313648.57Induction therapy; no anti-thymocyte globulin16020.4673452.7320272.14 Open table in a new tab To evaluate the DGFS predictive capacity, we used this scoring system on the independent data set composed by 606 recipients. More precisely, 561 patients without missing data for the five parameters of the score were studied. As illustrated in Figure 2, the AUCs were 0.64 (95% CI=(0.59, 0.69)) and 0.73 (95% CI=(0.68, 0.77)) for the CIT and the DGFS, respectively. The predictive capacity of the DGFS was very close to that obtained by Irish et al.,20.Irish W.D. Ilsley J.N. Schnitzler M.A. et al.A risk prediction model for delayed graft function in the current era of deceased donor renal transplantation.Am J Transplant. 2010; 10: 2279-2286Crossref PubMed Scopus (275) Google Scholar in which 18 parameters were included with an AUC at 0.70 (no CI). Out of these 18 parameters, not all were available in our database, in particular the race (the French law does not authorize the storage of patient ethnicity) and the warm ischemia time (not always collected by surgeons). Nevertheless, we computed the score proposed by Jeldres et al.21.Jeldres C. Cardinal H. Duclos A. et al.Prediction of delayed graft function after renal transplantation.Can Urol Assoc J. 2009; 3: 377-382PubMed Google Scholar based on patients without missing data on the six following parameters: CIT, recipient age, recipient weight, HLA mismatches, PRAs, and donor age. The predictive capacity of the score by Jeldres et al.21.Jeldres C. Cardinal H. Duclos A. et al.Prediction of delayed graft function after renal transplantation.Can Urol Assoc J. 2009; 3: 377-382PubMed Google Scholar seemed equivalent to a decision based only on the CIT in this validation sample (n=226, AUC=0.64, 95% CI=(0.56, 0.72)). The same equivalence was found from the training sample (Jeldres’ score: n=454, AUC=0.58, 95% CI=(0.52, 0.73); CIT: n=1227, AUC=0.60, 95% CI=(0.56, 0.64)). In addition, Figure 3 illustrates the good calibration of the model. The predicted and observed probabilities were close regardless of the DGFS level (P=0.1054, Hosmer–Lemeshow statistic). In this validation sample, the positive predictive value of a DGFS value higher than 1.20 was 0.61 (95% CI=(0.50, 0.72)). The negative predictive value of a DGFS value lower than −0.50 was 0.88 (95% CI=(0.83, 0.93)). Another major bias of our results could be the estimation of the relationship between the induction therapy (ATG or not) and the DGF probability. As shown in the Supplementary Tables S2 and S3 online in Supplementary Web Data online, patients’ characteristics were significantly different when we compared patients with and without ATG treatment. The dialysis duration before transplantation (P=0.0050) and the CIT (P=0.0016) seemed higher for patients who received ATG induction. In addition, the following characteristics were more frequent for patients with ATG induction: female recipient, re-transplantation, past history of hypertension, detectable PRAs on class I and II, and kidney disease with risk of recurrence. Regarding this description and as expected, induction by ATG was preferentially prescribed for patients at-risk of acute rejection and/or DGF. Therefore, the effect of ATG treatment presented in Table 3 (OR=1.70) may be underestimated. In order to decipher this possible bias, we firstly decided to include into the score all these possible confounding factors. The results of this ‘full-adjusted model’ are presented in Table 5. By considering the same characteristic at baseline (CIT, donor age, BMI, donor creatinine, duration in dialysis, male recipient, graft rank, hypertension, and PRAs), the estimated benefit of induction therapy with ATG was unchanged (OR=1.78) in this full-adjusted model. Moreover, by using the validation sample (n=597, nine observations removed due to missing value), the AUC of the full-adjusted model was 0.71 (95% CI=(0.66, 0.76)), which illustrates the uselessness of this full-adjusted model in terms of predictive capacities compared with the five-variable DGFS (AUC=0.73, 95% CI=(0.68, 0.77)).Table 5Results of the full logistic regression (n=1219, 19 observations removed because of missing data)log odds ratio (±s.d.)Odds ratioP-valueΔ log odds ratioCold ischemia time (hours)0.056 (±0.010)1.06 108μmol/l0.575 (±0.161)1.770.00041.18%No anti-thymocyte globulin0.575 (±0.150)1.780.00018.42%Duration in dialysis (years)0.030 (±0.017)1.030.0844—Male recipient0.132 (±0.147)1.140.0369—Rank of the kidney transplantation >1−0.265 (±0.209)0.770.2041—History of hypertension−0.216 (±0.175)0.810.2181—Detectable anti-class I panel reactive antibodies0.191 (±0.183)1.210.2966—Detectable anti-class II panel reactive antibodies0.271 (±0.194)1.310.1630—Δ log odds ratio represents the relative variation of the regression coefficients between this full model and the five-variable-based model (Table 3). Open table in a new tab Δ log odds ratio represents the relative variation of the regression coefficients between this full model and the five-variable-based model (Table 3). Secondly, we decided to perform a pairwise analysis. Among the 675 patients who received ATG for induction therapy and the 547 recipients without ATG in the training sample, we identified 121 pairs with the same characteristics for variables listed in the Table 5 (for continuous variables as donor age, CIT, recipient BMI, and duration in dialysis; three classes were defined to obtain balanced effectiveness). The estimated benefit of induction therapy with ATG was similar (OR=1.66) in this pairwise analysis compared with the DGFS (OR=1.70). Finally, the estimated benefit of induction by ATG obtained from our observational cohort from several approaches (OR=1.70 for the DGFS, OR=1.78 for the full-adjusted model, OR=1.66 for the pairwise analysis) was similar to the results recently published by Noel et al.23.Noel C. Abramowicz D. Durand D. et al.Daclizumab versus antithymocyte globulin in high-immunological-risk renal transplant recipients.J Am Soc Nephrol. 2009; 20: 1385-1392Crossref PubMed Scopus (152) Google Scholar in a randomized trials comparing anti IL2R (n=114, 44.6% of DGF) vs. ATG (n=113, 31.5% of DGF). Indeed, based on these data, one can estimate that the risk of DGF was 1.73-fold reduced for patients with ATG (OR=1.73, 95% CI=(1.01, 2.97)). The prediction of the DGF occurrence is important in preventing its possible deleterious consequences on long-term graft outcome. For instance, given that a causal pathway from CIT to graft failure through DGF was demonstrated,24.Mikhalski D. Wissing K.M. Ghisdal L. et al.Cold ischemia is a major determinant of acute rejection and renal graft survival in the modern era of immunosuppression.Transplantation. 2008; 85: S3-S9Crossref PubMed Scopus (143) Google Scholar preventing DGF may constitute a means to prevent CIT-related graft failure. Recently, Irish et al.20.Irish W.D. Ilsley J.N. Schnitzler M.A. et al.A risk prediction model for delayed graft function in the current era of deceased donor renal transplantation.Am J Transplant. 2010; 10: 2279-2286Crossref PubMed Scopus (275) Google Scholar developed a scoring system based on variables collected at the time of transplantation (AUC=0.70). This score has also been validated in North American and European cohorts of patients.25.Moore J. Ramakrishna S. Tan K. et al.Identification of the optimal donor quality scoring system and measure of early renal function in kidney transplantation.Transplantation. 2009; 87: 578-586Crossref PubMed Scopus (31) Google Scholar, 26.Rodrigo E. Minambres E. Ruiz J.C. et al.Prediction of delayed graft function by means of a novel web-based calculator: a single-center experience.Am J Transplant. 2012; 12: 240-244Crossref PubMed Scopus (16) Google Scholar, 27.Tiong H.Y. Goldfarb D.A. Kattan M.W. et al.Nomograms for predicting graft function and survival in living donor kidney transplantation based on the UNOS registry.J Urol. 2009; 181: 1248-1255Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar Unfortunately,
Referência(s)