A high fraction of patients with a failing allograft return to renal replacement therapy (RRT) with a tunneled hemodialysis catheter (TC). A late-stage chronic kidney disease tracking tool for such patients may improve rates of permanent dialysis access use at the time of RRT initiation. We identified patients at risk of initiating dialysis without a permanent access and prompted providers to initiate care pathway interventions for dialysis preparation, including modality education, vein mapping, access surgeon visits, actual access placement and maturation evaluation (where applicable).

From Jan 2015-Jan 2017 we identified 69 new incident patients with a failing allograft starting RRT. In this group, 37 started with an arteriovenous fistula (23 pre-existing), one with an arteriovenous graft (AVG), 10 with a Tenckhoff catheter, and 23 with a TC. Of those starting with a TC, 6/23 had a maturing AVF and 17/23 started with a TC only, amounting to a cumulative “catheter last” rate of 17/69 or 25%.  We noted an improvement in primary permanent access rates from 2015 to 2016. A late chronic kidney disease tracking tool can successfully increase rates of primary permanent dialysis access utilization in patients with a failing renal allograft.


Patients with a failed allograft returning to dialysis suffer more morbidity and greater early mortality compared to new incident patients with ESRD without a transplant. 1-4  Several potential factors have been identified to explain this phenomenon. Among these potential factors is the high frequency with which patients with a failing or failed allograft initiate renal replacement therapy (RRT) with a tunneled catheter (TC).  A recent survey of U.S. Renal System data from 2006-2011 on patients with a failed allograft returning to any dialysis modality reported a primary arteriovenous fistula (AVF) rate of 27.7%, arteriovenous graft (AVG) rate of  6.9% and a tunneled catheter (TC) rate of 65.4%. 5

As part of routine quality improvement activities in our renal transplant clinic, we hypothesized that the implementation of a late-stage chronic kidney disease (CKD) tracking tool in patients with a failing allograft could effectively increase the rates of returning to RRT with a functioning permanent access.

Study design, setting, and participants

Our nephrology practice is unique in having a fully contained post-transplant clinic serving > 1,800 allograft recipients located throughout North Carolina and large parts of South Carolina, as well as a large CKD clinic serving the greater Charlotte metropolitan area.  Using a web-based patient registry which draws patient demographic information and laboratory data from an Allscripts–based electronic health record platform, we identified all allograft recipients with a MDRD eGFR of < 30 cc/min. Focused chart reviews were undertaken on all identified patients, and an electronic health record-based tracking tool was completed on each patient (a template of the tracking tool is included as a Supplementary Appendix to this article). Patients who were not initially identified by the registry, but who were identified as being at risk for allograft failure due to unexpected hospitalization or clinic visit which evidenced advanced allograft dysfunction could also be added to the tracking tool on an ad hoc basis.  Completion of the tracking tool for each patient created a template progress note in the patient’s electronic chart.  The template inputs were also able to be extracted and reviewed in an Excel-based spreadsheet for review at monthly Quality Improvement (QI) meetings.

Quality improvement plan

From January 2015 to January 2017, we met monthly to review the tracking tool spreadsheet. In preparation for the monthly meeting, the advanced practitioner reviewed the charts of all identified patients as well as newly identified patients in the registry or on an ad hoc basis (e.g. unexpected hospitalization or clinic visit evidencing advanced, unexpected, irreversible allograft dysfunction).  The AP ensured previously identified patients were still being tracked month-to-month, flagged patients with evidence of a precipitous deteriorating trajectory in allograft function for closer scrutiny, and annotated the tracking tool data set for individual patients to highlight any unique clinical circumstances (e.g. multiple prior failed AV accesses, need for systemic anticoagulation, a propensity for missed clinic visits, etc. See Supplemental Appendix for complete list).

After each monthly review, a template-based “Transplant CKD” note was placed in the progress note section of each patient’s chart, which included information about whether the patient had completed RRT education, whether the patient made a modality selection (including no RRT), the status of vein mapping (if applicable), date and surgeon for access referral, date of access placement, date of maturation check (if applicable), and confirmed clearance for use of the permanent access.  Future clinic visits for each patient were identified, and the relevant provider (physician or AP) received a specific reminder communication to address one or more facets of CKD/access management at the next office visit.

All patients who ended up initiating RRT with a TC were also reviewed at the monthly meeting to identify a root cause (or causes) of the failure and to secure a permanent access and identify opportunities for further process improvement.

As this tool was explicitly designed as a quality improvement project, and since the standard of care for permanent access was being pursued for all patients, there was no “control” group subjected to intervention other than the established standard of care.  The authors sought and received formal review and approval for exemption from IRB review by Chesapeake IRB Protocol #00020598.


Primary focus was placed on RRT modality education (“TOPS” or “treatment options” education).  Clinician discussion with the patient about timing and preferences regarding RRT (including conservative management only); confirming the existence of a permanent access where applicable and arranging for surgical referral, maturation check, and clearance for access use, again as applicable.   A secondary purpose of this QI activity was to review whether identified patients with a failing allograft were candidates for re-transplantation, whether this option had been discussed with the patient, and if so whether a transplant referral had been completed.

Measurements and results

Between Jan 2015-Jan 2017, a range of 100-120 patients were identified for review through the screening process outlined in the Methods section.  Patients who transiently had an eGFR <30 cc/min who subsequently recovered to > 30 cc/min were dropped from the tracking tool, hence the total number of patients reviewed month-to-month varied slightly.  A total of 69 new incident transitions to RRT transpired in the two-year time frame.   The outcomes are illustrated by calendar year 2015 and 2016, and then cumulatively in Figure 1.  One patient each in the 2015 and 2016 cohort initiated peritoneal dialysis but also had a functioning AVF.  These patients were double-counted in the AVF and Tenckhoff categories. Thus, for calendar year 2015 and 2016 there were 22 permanent accesses for 21 patients and 26 permanent accesses for 25 patients, respectively.  Patients with an AVF not suitably mature for use who started hemodialysis with a TC were counted under the general rubric of “TC,” and then broken out into TC + AVF or TC only.

In 2015, there were 33 new incident patients starting RRT. In this group,

  • 16 patients started with an AVF
  • 1 with an AVG
  • 5 with a Tenckhoff catheter

Of the 16 patients who started with an AVF, 9 patients had pre-existing AVFs.  In 2015, 21/33 (64%) of patients started RRT without a TC.  Of the 12 patients who started RRT with a TC in 2015, 2/12 had a maturing AVF at the time of initiation, and 10/12 patients did not. Thus, our “catheter last” rate in patients initiating RRT in 2015 was 10/33, or 30%.

In 2016, there were 36 new incident patients starting RRT; 21 patients started with an AVF (14 patients had pre-existing AVF), and 5 patients started with a Tenckhoff.  Of the 11 patients in 2016 who started RRT with a TC, 4/11 had a maturing AVF; 2 of these 4 patients had their AVF cleared by a formal maturation check, but the AVF was ultimately not able to be used at the time of RRT initiation.

Of the 7 patients who initiated RRT with a TC only (for a 2016 “catheter last” rate of 7/36 or 19%), 3 had chosen PD as an initial modality and had been evaluated by a surgeon for Tenckhoff placement, but had to start RRT before the Tenckhoff could be placed and used.  Of these 3 patients, 2 transitioned to PD < 90 days after RRT initiation.  In total from Jan 2015-Jan 2017, there were 69 new incident patients starting RRT; 37 started with an AVF (23 pre-existing), 1 with an AVG, 10 with a Tenckhoff, and 23 with a TC.  Of those starting with a TC, 6/23 had a maturing AVF and 17/23 started with a TC only, amounting to a cumulative “catheter last” rate of 17/69 or 25%.

We noted an improvement in rates of primary TC initiation between 2015 and 2016 with a TC only rate of 19% in 2016 compared to 31% in 2015.  We attribute this improvement to several factors.

  • earlier discussions of RRT transition with at-risk patients, including caregiver reminders to address the topic at the next office visit
  • clearer documentation of patient RRT preferences in the chart
  • a move toward providing more continuity with a single designated physician as the patient became closer to needing RRT
  • closer attention to patients with rapidly deteriorating allograft function
  • standardized care pathway for access creation, maturation checks, and formal written clearance for access use

From our 2016 experience, we noted an opportunity for improvement in our maturation check process, as 2 patients in the 2016 cohort who had an AVF and had cleared a formal ultrasound-based maturation check ultimately had an AVF which was not suitable for primary use.  Even among the patients who started with a TC only in 2016, proactive modality education and selection allowed two of these patients to transition to peritoneal dialysis in < 90 days.  Thus, in at least a few cases our tracking tool prepared the foundation for at least a few patients who initiated RRT with a TC to transition to use of a permanent access in a short time frame.

The 17 patients who initiated with a TC only were also retrospectively viewed in detail.  In a majority of cases (59%), non-adherence to outpatient clinical follow-up was found to be the salient cause of return to RRT with a TC.  In 12% of cases where scheduled outpatient clinic appointments were well attended, access placement was discussed on multiple occasions, but pre-emptive access placement was ultimately declined.

Interestingly, patients located outside the practice’s traditional geographic catchment and poor AV access venous targets were initially hypothesized to be common contributing factors for initiating RRT with a TC only, but this was the case in only a few instances.

We observed a less than expected (14%) rate of patients who returned to PD as a primary RRT modality.  For comparison, a published registry survey of Canadian patients with failed allografts returning to RRT reported an 18 % PD rate.6  In reviewing individual patient narrative accounts of reasons for specific modality selection, we found that the patient having a pre-existing and functioning AV access seemed to play a significant role in the choice of hemodialysis over peritoneal dialysis.  Medical relative contraindications to peritoneal dialysis (multiple abdominal surgeries, adhesions, and treatment non-adherence) were salient reasons in a minority of instances.  Parenthetically, we noted a higher rate of PD as a primary modality selection among patients who chose to complete a visit to a home training unit as part of treatment options education, highlighting a further opportunity for process improvement.


In this study, we show that implementing a conceptually simple tracking tool can substantially improve rates of primary permanent access in patients with a failing allograft. This represents a sizeable cohort of patients who are newly incident to RRT, a group that has been shown to be at considerably higher risk of suffering the vicissitudes of early dialysis initiation, including cardiovascular complications and sepsis. 2,7-8   One potentially modifiable risk factor for sepsis after initiation of RRT is the avoidance of initiating dialysis with a tunneled catheter.  While it is true that these patients are ostensibly followed by a transplant nephrologist, and although a consensus statement on parameters for post-transplant follow-up have been published (circa 2000), 9 there is little evidence of any uniformity in practice regarding frequency and focus of long-term post-transplant management generally, much less late-stage CKD management in this patient population.

Lest the reader mistakenly infer this is a phenomenon unique to the United States, a large UK-based renal registry study showed that patients with an allograft and advanced CKD exhibited significantly suboptimal KDOQI clinical performance targets 10 compared to non-transplanted CKD and ESRD patients, suggesting that lower rates of securing permanent access placement may be emblematic of suboptimal CKD management more generally in the population of patients with a failing renal allograft.

Long-term patient outcomes from a deliberate, effective strategy of avoiding tunneled catheters in this population is beyond the scope of our study, but the efficacy of our intervention invites the inference that improved permanent access rates will have some positive long-term impact on patient morbidity and mortality after initiation of RRT.

We observed that the nuance of integrating the data “horizontally” (memorialized as a progress note in individual patient charts) and “vertically” (in the form of an “at-a-glance” spreadsheet format) was heuristically useful in exploiting the data in different contexts.  Having a progress note with specific tasks ready to hand for the individual clinical encounter proved effective and efficient for productive face-to-face clinical interactions.  Vertical integration permitted immediate recognition of where we had made progress, where more urgent interventions were needed, and allowed us to highlight our surveillance “failures,” (TC only) to identify opportunities for future process improvement.  In our cohort, the most important opportunity for improvement we discovered was to find alternative ways to reach patients with failing allografts who exhibited reduced adherence to scheduled outpatient clinic follow-up.

Our comparatively low rate of incident PD initiation has led us to consider alternative ways of promoting peritoneal dialysis as a primary modality for this patient population. We have recently initiated an urgent PD start program modeled on an approach described by Ghaffari 11 and we plan to integrate this into our transplant chronic kidney disease quality improvement efforts.

Limitations of the study

There are several limitations to our intervention study. Our nephrology practice provides post-transplant care in the same location as we provide CKD care, and our transplant nephrologists and transplant APs are also regularly and directly involved in late-stage CKD and ESRD clinical management.  Accordingly, common barriers to transitions of care to ESRD are less prevalent in our practice model.  The majority of patients live within the geographic catchment area of our general nephrology practice.  Transplant centers with a very large geographic catchment may face greater communication and coordination challenges with general nephrology practices which are not thoroughly integrated with the transplant center.  Our electronic health record registry capture model did not identify all allograft recipients with advanced CKD, and our ad hoc approach adding patients with unexpected, severe, and irreversible allograft dysfunction to the tracking tool may be less effective in achieving full patient capture in other locales.

Finally, in our observed patient population, 33% (23/69) of patients with a failing allograft returning to RRT already had a functioning AVF, so a 54% rate of primary AVF starts built on a high baseline rate of pre-existing and functioning AVFs.


In sum, our study shows that the implementation of a late-stage chronic kidney disease tracking tool for identifying and targeting patients with a failing allograft was effective in increasing rates of return to RRT with a permanently functioning access.  Even in instances where patients returned to RRT with a TC, we were able to achieve a slight increase in patients who either had a maturing AVF or who were able to transition to peritoneal dialysis in a relatively short time frame.  The vertical integration of the data also clarifies opportunities for process improvement.  Interventions such as this one, which straddles the artificial divide between general and transplant nephrology, will be useful both in effecting improved patient care and accommodating reimbursement trends focused on value-based purchasing and population health interventions.


  1. Gill JS, Rose C, Pereira BJG, Tonelli M.  The importance of transitions between dialysis and transplantation in the care of end-stage renal disease patients.  Kidney International 71: 442–447, 2007.
  2. Kaplan B, Meier-Kriesche HU. Death after graft loss: An important late study endpoint in kidney transplantation. Am J Transplant 2: 970–974, 2002.
  3. Knoll G, Muirhead N, Trpeski L, Zhu N, Badovinac K: Patient survival following renal transplant failure in Canada. Am J Transplant 5: 1719–1724, 2005.
  4. Rao PS, Schaubel DE, Jia X, Li S, Port FK, Saran R.  Survival on dialysis post-kidney transplant failure: results from the Scientific Registry of Transplant Recipients. Am J Kidney Dis 49:294-300, 2007.
  5. Chan MR, Oza-Gajera B, Chaplan K, Djamali AX, Muth BL, Turk J, Wakeen M, Yevzlin AS, Astor BC.  Initial vascular access type in patients with a failed renal transplant. Clin J Am Soc Nephrol 9: 1225–1231, 2014.
  6. Perl J, Hasan O, Bargman JM, Jiang D, Na Y,  Gill JS, Jassal SV.  Impact of dialysis modality on survival after kidney transplant failure.  Clin J Am Soc Nephrol 6: 582–590, 2011.
  7. Johnston O , Zalunardo N, Rose C, Gill JS.  Prevention of sepsis during the transition to dialysis may improve the survival of transplant failure patients.  J Am Soc Nephrol 18: 1331–1337, 2007.
  8. Gill JS, Abichandani R, Kausz AT, Pereira BJG.  Mortality after kidney transplant failure: The impact of non-immunologic factors.  Kidney International, 62: 1875–1883, 2002.
  9. Kasiske BL, Vazquez MA, Harmon WE, Brown RS, Danovitch GM, Gaston RS, Roth D, Scandling, JD, Singer GG.  Recommendations for the outpatient surveillance of renal transplant recipients.   J Am Soc Nephrol 11: S1–S86, 2000.
  10. Ansell D, Udayaraj UP, Steenkamp R, Dudley CRK. Chronic renal failure in kidney transplant recipients. Do they receive optimum care? Data from the UK renal registry. Am J Transplant 7: 1167-1176, 2007.
  11. Ghaffari A.  Urgent-Start Peritoneal Dialysis: A Quality Improvement Report.  Am J Kidney Dis. 59: 400-408, 2012.