|Year : 2019 | Volume
| Issue : 1 | Page : 1-14
|Prediction models to measure transplant readiness of patients with renal failure: A systematic review
Majid Jangi1, Zahra Ebnehoseini1, Mahin Ghorban Sabbagh2, Ebrahim Khaleghi3, Mahmoud Tara1
1 Department of Medical Informatics, School of Medicine, Mashhad University of Medical Sciences, Mashhad, Iran
2 Department of Internal Medicine, School of Medicine, Ghaem Hospital, Mashhad, Iran
3 Organ Procurement Center, Montaserie Hospital, Mashhad University of Medical Sciences, Mashhad, Iran
Click here for correspondence address and email
|Date of Submission||03-May-2018|
|Date of Decision||14-Jul-2018|
|Date of Acceptance||16-Jul-2018|
|Date of Web Publication||26-Feb-2019|
| Abstract|| |
Predicting the future of illness, a patient is facing helps the physicians to choose the best strategy to manage the disease. Models for predicting the readiness of candidates for kidney transplant can be very promising. This study sought to systematically review the predictive models and algorithms that assess the readiness of renal transplant candidates in different countries. This systematic review study was according to PRISMA-P protocol in PubMed and Science Direct databases and general search engines up to March 2017. Eligible studies were those that introduced a model to assess the readiness for renal transplantation of patients with chronic renal failure from cadavers and this assessment led to scoring prioritization or superiority among patients. We found 28 studies from 11 countries that met the search criteria and >50% of them were published from 2015 onward. Of the studies, nine models and algorithms were extracted that included 12 factors. Some models, including the European and Scandinavian models, were used jointly between different countries. All the models had at least four factors, and nearly 90% of the models considered four or five factors to measure kidney transplantation readiness. More than 50% of the models had age, dialysis duration, HLA type, and emergency status factors and, dialysis duration. Predictive models are important for renal transplant because of the significant reduction in number of cadavers and longer wait of candidates for a kidney transplant. Further studies can examine the effect of these models on the survival of the kidney transplant.
|How to cite this article:|
Jangi M, Ebnehoseini Z, Sabbagh MG, Khaleghi E, Tara M. Prediction models to measure transplant readiness of patients with renal failure: A systematic review. Saudi J Kidney Dis Transpl 2019;30:1-14
|How to cite this URL:|
Jangi M, Ebnehoseini Z, Sabbagh MG, Khaleghi E, Tara M. Prediction models to measure transplant readiness of patients with renal failure: A systematic review. Saudi J Kidney Dis Transpl [serial online] 2019 [cited 2020 May 27];30:1-14. Available from: http://www.sjkdt.org/text.asp?2019/30/1/1/252899
| Introduction|| |
A modern approach to help patients with chronic disorders are to use models to predict the outcome of treatment options and choose the best strategy to manage the disease. This type of prediction is fundamentally complex and requires the physician to review and examine a variety of factors to predict and estimate the subsequent steps of the disease or status of the patient. In fact, the purpose of prediction models is to forecast future events or results based on patterns that are according to a risk score and, consist of some predictors. Of course, designing of these models in the field of medical research is increasing daily.
For example, these models in the field of gastrointestinal diseases can predict the relapse of the disease in patients with inflammatory bowel disorder and predict the risk of developing liver cancer in patients with cirrhosis by using the risk model., In addition, some similar models are used to categorize and stratify patients based on the risk of re-admission. Other common cases are the calculation of the Framingham Risk Score for cardiovascular disease, as well as the development of early death prevention models in chronic diseases.
In this regard, this study shares the design steps of a comprehensive model for measuring the readiness of patients with chronic renal failure (CRF) for transplantation. CRF is one of the most common chronic diseases that affect approximately 10% of the population worldwide, including 13.1% of the adult population of the USA and 12.6% of the population that was studied in Iran. In many cases, kidney transplantation is considered the most effective treatment and currently, this is the preferred treatment for the end-stage renal disease (ESRD).,
The superiority of renal transplantation as renal treatment makes the identification and preparation of candidates for transplantation important. To identify candidates for renal transplant, different algorithms and many factors along with different readiness prediction models exist. The important point is that studies have shown that the number of individuals waiting for kidney transplant from a deceased person and the waiting time compared to a decade before, has doubled. This ongoing process to this day has increased the accumulation of patients waiting for transplant who remain on dialysis. Therefore, a comprehensive and suitable model for predicting the readiness of candidates for kidney transplant can be very promising.
Researchers in this study sought to systematically review the literature and scientific evidence to identify all predictive models and algorithms for assessing the readiness of renal transplant candidates in different countries.
| Methods|| |
We used the PRISM-P protocol as the basis of our systematic review protocol. The search strategy was focused to find models and algorithms related to the prediction and assessment of the readiness of patients for kidney transplant and to extract the underlying factors introduced by the models.
Applying advices from our consultant transplant experts, and the review of similar articles and studies, a set of concepts were selected. These concepts were: kidney transplantation, CRF, renal dialysis, donor, transplant readiness, priority, survival, criteria, predictive factor, and deceased/cadaver. A concept-to-keyword map was created after a thorough review of the PubMed MeSH terms, connecting the above concepts to the most relevant set of MeSH-based keywords. The final produced set of keywords was shared with the experts for final approval [Table 1]. PubMed and Science Direct were searched using the keyword set without any time limitation [Table 2]. Sources of gray literature were also searched using the Google Scholar and Google search engine. We also reviewed, based on the advice of local experts, some articles, guidelines, library references, and websites such as http://who.int, https://www.uptodate.com, and http://irodat.org. The retrieved articles from our search attempts were transferred to EndNote X6 (BLD6348 Thomson Reuters) for further analysis.
Selection of articles and literature
Two researchers independently reviewed the articles. Based on the conflict resolution strategy, in the cases where there was a discrepancy between articles for the two independent researchers, a third researcher reviewed the difference and gave an opinion. The screening of articles was initially conducted based on the inclusion criteria. The retrieved abstracts of the articles were examined for reference to any prediction model or algorithm in regard to the readiness assessment of kidney transplantation. The articles that did not meet our criteria were omitted. The full-text of the finally-selected articles plus other suggested sources of prediction models such as Websites, guidelines, and library references was examined against the exclusion criteria. The final set of selected articles was marked on EndNote for our comprehensive analysis.
We considered a study eligible if it introduced a transparent model or algorithm to assess the readiness of renal transplantation from cadavers in which the assessment led to scoring prioritization or superiority determination among patients; and if the full-text version in English was available. The articles that introduced different versions of the same model were only included if it contained the latest version.
The articles were excluded if:
- They only provided one or two specific transplant-readiness factors instead of providing a comprehensive model
- They only introduced or discussed factors without presenting any scoring, prioritizing, or superiority assessment strategy
- They were focused on children under the age of 18 years as a recipient
- They only focused on transplantation from live donors.
To extract data, a general and a proprietary data table was created. We extracted the following data: the name of the author/organization or the website address, the type of evidence [including article/guideline/website, date/year of publication, country the study was conducted in, and the name of the model (if there was an official name)].The data items to be extracted in proprietary data table are shown in [Table 3].
|Table 3: Proprietary information of the articles and references used in the study.|
Click here to view
The extracted models were first categorized based on the underlying factors they contained. Then, the power of each factor in assessing transplant readiness was calculated. In addition, the models were compared based on their factor list, country of origin and their prediction scope.
| Results|| |
Our search strategy yielded 897 articles and 43 gray literatures; of them only 19 articles and nine gray literatures were selected after removing duplicates and applying inclusion and exclusion criteria [Figure 1]. These 28 articles,,,,,,,,,,,,,,,,,,,,,,,,,,, were from 11 countries and regions [Table 4].
|Table 4: General information of the articles and references used in the study.|
Click here to view
More than 50% of these studies were either conducted in the USA or were about models that assess the transplant-readiness in the USA. Other studies were performed in Portugal, France, Poland, Chile, Brazil, Turkey, the Czech Republic, European Union (EU) and Scandinavia, plus the Catalonia region. After the USA, most articles were from Portugal (22%), whereas the other remaining countries had three or less.
The first study was published in 1997 and the last studies in 2017, and >50% of the articles and guidebooks were published in 2015. In total, nine unique models were extracted [Table 3], which belonged to the USA, EU, the Czech Republic, France, Portugal, and Scandinavian countries, the Catalonia region of Spain, Chile, and Turkey. European countries that used the Euro Transplant Kidney Allocation System (ETKAS) model were Croatia, Belgium, Austria, Germany, Hungary, Luxembourg, the Netherlands and Slovenia and Scandinavian countries that used the Scandia Transplant Acceptable Mismatch Program (STAMP) model included were Denmark, Finland, Iceland, Norway, and Sweden. Some extracted models had a special name and included Europe-based models such as ETKAS and Scandinavian Models such as STAMP, as well as the USA-based model named UNOS (United Network of Organ Sharing).
Of these nine models, 12 factors were extracted. All the models had at least four factors and nearly 90% of the models considered four or five factors to measure kidney transplantation readiness. The ETKAS model in Europe, with eight factors, had the largest set of factors in calculating readiness. The ETKAS factors were: age, duration on dialysis, HLA type, emergency status, previous history of transplantation, local and national balance, candidate distance to the site of the link, and the history of kidney donation.
More than 50% of the models had age, dialysis duration, HLA type, and emergency status factors, whereas 100% of them had dialysis duration as a factor. Although the HLA type factor had been listed in all models, just one model did not include this factor in the ranking and scoring process. The blood group type was important in all the models but was only used for donor-candidate matching. Factors such as the history of kidney donation, BMI, concurrent transplantation, and diabetes were only used in one model.
The European and Turkish models accepted donors from different regions and countries in candidate ranking and considered local and national equilibrium factors., The European and Catalonian models would also consider the distance between the candidate's location and the gradient link center.,
| Discussion|| |
Prediction models, instead of making a simple risk estimation of the outcome, propose decisions. Obviously, prediction models cannot replace the physician's judgment; however, it attempts to provide objective and statistical estimates about the next stage of an illness and provide important, additional information for medical decisions and processes. Of course, prediction models have several applications in the disease management process of a patient. Some prediction models are used to categorize, rank, and stratify patients based on the potential risk such as re-admission. The models extracted in this study were also based on the risk of nonsurvival after the candidate's transplant. Such calculation and prediction may help the clinicians significantly to reduce the waiting time for kidney transplant and noneffective transplantation. Furthermore, before the transplant, the candidate's readiness should be measured and the candidate who is more likely to be prepared and likely to survive after a better link is expected to be nominated.
Selection of a candidate based on a transplantation readiness assessment can increase the survival of the transplant patients. One-year survival rates are similar in Europe and the USA; however, the survival rate at five and 10 years is better in Europe, which can be due to the use of readiness assessment models with several extra factors (8 as compared to 4 in the United States).,
The effective factors in assessing the extent of the candidate's dynamics are changing over time. Originally, the nature of the chronic illness is such that it progresses over time and, along with its development, brings new complications or exacerbates previous complications. Therefore, patients with CRF need to be constantly and promptly assessed for their transplantation readiness. One-Time use of the model cannot produce a constant, reliable score for readiness. The UNOS model performs this gradation on a daily basis.
According to the results of this study, about 90% of the studies have been conducted in the last decade, with 50% only in 2015. This points to growing attention toward criteria of transplantation readiness; and in more general sense, to an increasing application of prediction models in improving care.
Forecast models become more complicated when effective factors are increased. Based on the results of this study, about 90% of the models use only four or five factors to measure the readiness of a candidate for transplantation, and only the European model contains eight factors. The most common factor we discovered was the duration on dialysis, which was listed in all the models. Studies show that the duration of exposure to ESRD and exposure to dialysis is the main risk factor for allograft survival. It has been demonstrated that patients on dialysis have alterations in the concentration of some substances that predispose these patients to both cardiovascular and renal allograft vascular damages. In addition, the increase in the duration of pretransplant dialysis is associated with a reduction in posttransplantation patient survival. For example, cardiomegaly and infection that are causes of death, correlate with patient survival. Furthermore, they are associated with increasing time on dialysis. Therefore, in all models, a higher score, reflecting a higher emphasis is considered for the duration on dialysis.
The age of the candidate is another important factor that accounts for about 90% of the models in the study as aging is significantly associated with reduced survival. Further more, mortality in elderly people undergoing dialysis and in the transplant waiting list is higher. In the Portuguese and Catalonian models, the age difference between the recipient and the donor is measured and rated; lower the age difference between the candidate and the donor, more points are awarded to the candidate., Older kidneys, when transplanted into younger recipients, may not be sufficient to cope with increases in physiological or metabolic demands leading to a greater reduction in renal function, a significant predictor of cardiovascular and all-cause mortality., In addition, in most models, children and adult models varied in their contained factors.
HLA typing is also an important factor in survival and nonrejection rates, and because of this, many candidates are not selected for transplantation because of the mismatch of HLA, making it possible for more suitable candidates to be selected. About 90% of the models extracted in this study included this factor. Furthermore, in the UNOS model, despite not ranking this factor, it is considered when matching is performed between the candidate and the donor.
BMI, was found in just one model (in Scandinavian countries). Furthermore, in some studies, BMI has a strong association with the outcome of the transplantation and the survival rate of the transplant., But, in many studies, the relation between BMI and outcome of transplantation is not confirmed significantly.,, Additionally, high BMI and abdominal obesity are important factors for the surgeon while performing the transplant. Of course, in some models including UNOS, BMI plays a role when the donor is known. Therefore, BMI is not considered a risk factor in many models.
Diabetes is a factor that is effective in the survival of the patient after transplantation, although its effect on transplantation survival is not proven. Many studies have shown that diabetes,,, is considered as a factor in patient mortality, loss of transplantation, and other problems, but only in the UNOS model, pretransplant diabetes has been considered as a factor in terms of candidate readiness. Of course, the highest regional, diabetes prevalence was in North America.
In cases with simultaneous transplantation including kidney-liver and kidney-pancreas, if the candidate has all good criteria of concurrent-transplantation with a nonrelated donor, good quality of life has been reported. However, the surgery may be very difficult and the allograft survival may be reduced. Furthermore, the survival rate of the transplant can be affected, with a decrease in transplant survival rate., Therefore, this factor can also be considered as an effective factor in ranking the readiness of trans-plantation, which was found in the Czech model.
In the European model, the specific location of all donors is considered in candidate ranking due to the varying potential availability of donations at the local, regional, and national levels, due to maintaining a balance kidney exchange among and within the member states.
| Conclusion|| |
Predictive models are important for renal transplant because of the significant reductions in cadavers and longer waiting time of candidates for a kidney transplant. There-fore, before the transplant, the candidate's readiness should be measured. Further studies are required to examine the effect of these models on the survival of the kidney transplant.
| Acknowledgments|| |
We are grateful to Research deputy of Mashhad University of Medical Sciences for supporting this research as part of a PhD thesis of Majid Jangi bearing number 950393.
Conflicts of interest:
| References|| |
Waljee AK, Higgins PD, Singal AG. A primer on predictive models. Clin Transl Gastroenterol 2014;5:e44.
Toll DB, Janssen KJ, Vergouwe Y, Moons KG. Validation, updating and impact of clinical prediction rules: A review. J Clin Epidemiol 2008;61:1085-94.
Waljee AK, Joyce JC, Wang S, et al. Algorithms outperform metabolite tests in predicting response of patients with inflammatory bowel disease to thiopurines. Clin Gastroenterol Hepatol 2010;8:143-50.
Singal AG, Mukherjee A, Elmunzer BJ, et al. Machine learning algorithms outperform conventional regression models in predicting development of hepatocellular carcinoma. Am J Gastroenterol 2013;108:1723-30.
Kannel WB, McGee D, Gordon T. A general cardiovascular risk profile: The Framingham study. Am J Cardiol 1976;38:46-51.
Clase CM, Kiberd BA, Garg AX. Relationship between glomerular filtration rate and the prevalence of metabolic abnormalities: Results from the third National Health and Nutrition Examination Survey (NHANES III). Nephron Clin Pract 2007;105:c178-84.
Safarinejad MR. The epidemiology of adult chronic kidney disease in a population-based study in Iran: Prevalence and associated risk factors. J Nephrol 2009;22:99-108.
Tabatabaei SA, Hashemi M, Jafari M. Frequency of complications of ureteral anastomosis stenosis, urine leakage and urine obstruction after renal transplantation and its related actors in Alzahra Hospital, Isfahan, Iran; an eight-year study. J Isfahan Med Sch 2015;33:183-93.
Scandling JD. Kidney transplant candidate evaluation. Semin Dial 2005;18:487-94.
Danovitch GM, Cohen DJ, Weir MR, et al. Current status of kidney and pancreas transplantation in the United States, 1994-2003. Am J Transplant 2005;5:904-15.
Abramowicz D, Cochat P, Claas FH, et al. European renal best practice guideline on kidney donor and recipient evaluation and perioperative care. Nephrol Dial Transplant 2015;30:1790-7.
Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015;4:1.
Domínguez J, Harrison R, Atal R, Muñoz F. A new kidney allocation policy in Chile: Computer-based simulations. Transplant Proc 2013;45:1313-5.
Chapal M, Le Borgne F, Legendre C, et al. A useful scoring system for the prediction and management of delayed graft function following kidney transplantation from cadaveric donors. Kidney Int 2014;86:1130-9.
Rao PS, Schaubel DE, Guidinger MK, et al. A comprehensive risk quantification score for deceased donor kidneys: The kidney donor risk index. Transplantation 2009;88: 231-6.
Peters-Sengers H, Berger SP, Heemskerk MB, et al. Stretching the limits of renal transplantation in elderly recipients of grafts from elderly deceased donors. J Am Soc Nephrol 2017;28:621-31.
Jacquelinet C, Audry B, Golbreich C, et al. Changing kidney allocation policy in France: The value of simulation. AMIA Annu Symp Proc 2006:374-8.
Chopra B, Sureshkumar KK. Changing organ allocation policy for kidney transplantation in the United States. World J Transplant 2015; 5:38-43.
Scaife ER, Mone MC, Shelby J, Nelson EW. Characteristics of kidneys offered as paybacks in the zero-antigen mismatch sharing policy. Transplant Proc 1997;29:3441-3.
Pokorná E, Vítko S, Lasziková E. Compliance with kidney allocation criteria in the Czech Republic in 1997. Transplant Proc 2000;32:91-3.
Reese PP, Shults J, Bloom RD, et al. Functional status, time to transplantation, and survival benefit of kidney transplantation among wait-listed candidates. Am J Kidney Dis 2015;66:837-45.
Magriço R, Malheiro J, Tafulo S, et al. Implications for patients waiting for a kidney transplant of using the calculated panel reactive antibody (cPRA). Port J Nephrol Hypert 2016;30:185-93.
Lima BA, Mendes M, Alves H. Kidney transplant allocation in Portugal. Port J Nephrol Hypert 2013;27:313-6.
Tafulo S, Malheiro J, Dias L, et al. Low transplantability of 0 blood group and highly sensitized candidates in the Portuguese kidney allocation algorithm: Quantifying an old problem in search of new solutions. HLA 2016;88:232-8.
Lima BA, Dias L, Henriques AC, Alves H. The Portuguese match algorithm in the kidney paired donation program. Cells Tissues Organs 2010;13:25-32.
Stegall MD. The right kidney for the right recipient: The status of deceased donor kidney allocation reform. Semin Dial 2010;23:248-52.
Celebi ZK, Akturk S, Erdogmus S, et al. Urgency priority in kidney transplantation: Experience in Turkey. Transplant Proc 2015; 47:1269-72.
Costa MG, Garcia VD, Leirias MM, Santos SR, Oliveira DM. Urgency priority in kidney transplantation in Rio Grande Do Sul. Transplant Proc 2007;39:381-2.
Gondos A, Döhler B, Brenner H, Opelz G. Kidney graft survival in Europe and the United States: Strikingly different long-term outcomes. Transplantation 2013;95:267-74.
Shahbazian H, Ehsanpour A, Ghorbani A, et al. Epidemiology of chronic kidney disease in Iran and worldwide. Sci Med J 2010;9:517-27.
Meier-Kriesche HU, Kaplan B. Waiting time on dialysis as the strongest modifiable risk factor for renal transplant outcomes: A paired donor kidney analysis. Transplantation 2002; 74:1377-81.
Meier-Kriesche HU, Port FK, Ojo AO, et al. Effect of waiting time on renal transplant outcome. Kidney Int 2000;58:1311-7.
Cosio FG, Alamir A, Yim Set al. Patient survival after renal transplantation: I. The impact of dialysis pre-transplant. Kidney Int 1998;53:767-72.
Zhou YC, Cecka JM. Effect of age on kidney transplants. Clin Transpl 1989:369-78.
Rao PS, Merion RM, Ashby VB, Port FK, Wolfe RA, Kayler LK. Renal transplantation in elderly patients older than 70 years of age: Results from the scientific registry of transplant recipients. Transplantation 2007;83:1069-74.
Curschellas E, Landmann J, Dürig M, et al. Morphologic findings in “zero-hour” biopsies of renal transplants. Clin Nephrol 1991 ;36: 215-22.
Epstein M. Aging and the kidney. J Am Soc Nephrol 1996;7:1106-22.
Clark B, Unsworth DJ. HLA and kidney transplantation. J Clin Pathol 2010;63:21-5.
Montgomery RA, Warren DS, Segev DL, Zachary AA. HLA incompatible renal transplantation. Curr Opin Organ Transplant 2012; 17:386-92.
Meier-Kriesche HU, Arndorfer JA, Kaplan B. The impact of body mass index on renal transplant outcomes: A significant independent risk factor for graft failure and patient death. Transplantation 2002;73:70-4.
Friedman AN, Miskulin DC, Rosenberg IH, Levey AS. Demographics and trends in overweight and obesity in patients at time of kidney transplantation. Am J Kidney Dis 2003;41:480-7.
Chang SH, Coates PT, McDonald SP. Effects of body mass index at transplant on outcomes of kidney transplantation. Transplantation 2007;84:981-7.
Marks WH, Florence LS, Chapman PH, Precht AF, Perkinson DT. Morbid obesity is not a contraindication to kidney transplantation. Am J Surg 2004;187:635-8.
Kovesdy CP, Czira ME, Rudas A, et al. Body mass index, waist circumference and mortality in kidney transplant recipients. Am J Transplant 2010;10:2644-51.
Holley JL, Monaghan J, Byer B, Bronsther O. An examination of the renal transplant evaluation process focusing on cost and the reasons for patient exclusion. Am J Kidney Dis 1998; 32:567-74.
Revanur VK, Jardine AG, Kingsmore DB, Jaques BC, Hamilton DH, Jindal RM. Influence of diabetes mellitus on patient and graft survival in recipients of kidney transplantation. Clin Transplant 2001;15:89-94.
Sulanc E, Lane JT, Puumala SE, Groggel GC, Wrenshall LE, Stevens RB. New-onset diabetes after kidney transplantation: An application of 2003 International Guidelines. Transplantation 2005;80:945-52.
Cosio FG, Pesavento TE, Osei K, Henry ML, Ferguson RM. Post-transplant diabetes mellitus: Increasing incidence in renal allograft recipients transplanted in recent years. Kidney Int 2001;59:732-7.
Kasiske BL, Snyder JJ, Gilbertson D, Matas AJ. Diabetes mellitus after kidney transplantation in the United States. Am J Transplant 2003;3:178-85.
Guariguata L, Whiting DR, Hambleton I, Beagley J, Linnenkamp U, Shaw JE. Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res Clin Pract 2014;103:137-49.
Danovitch GM. Handbook of Kidney Transplantation. 5th
ed. Philadelphia: Lippincott Williams & Wilkins; 2009.
Martin MS, Hagan ME, Granger DK. Current organ allocation disadvantages kidney alone recipients over combined organ recipients. Am J Surg 2016;211:525-8.
Nadim MK, Sung RS, Davis CL, et al. Simultaneous liver-kidney transplantation summit: Current state and future directions. Am J Transplant 2012;12:2901-8.
Department of Medical Informatics, School of Medicine, Mashhad University of Medical Sciences, Mashhad
[Table 1], [Table 2], [Table 3], [Table 4]
| Article Access Statistics|
| Viewed||1252 |
| Printed||60 |
| Emailed||0 |
| PDF Downloaded||338 |
| Comments ||[Add] |