How do centres begin the process to prevent contrast-induced acute kidney injury: a report from a new regional collaborative
- Jeremiah R Brown1,2,
- Peter A McCullough3,
- Mark E Splaine1,4,
- Louise Davies5,
- Cathy S Ross1,
- Harold L Dauerman6,
- John F Robb2,
- Richard Boss7,
- David J Goldberg8,
- Frank A Fedele9,
- Mirle A Kellett10,
- William J Phillips11,
- Peter N Ver Lee12,
- Eugene C Nelson1,
- Todd A MacKenzie4,
- Gerald T O'Connor1,
- Mark J Sarnak13,
- David J Malenka2,
- for the Northern New England Cardiovascular Disease Study Group
- 1The Dartmouth Institute for Health Policy and Clinical Practice, Dartmouth Medical School, Lebanon, New Hampshire, USA
- 2Section of Cardiology, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire, USA
- 3St. John Providence Health System, Providence Park Heart Institute, 47601 Grand River Avenue; Suite B-125 Novi, MI 48374, USA
- 4Department of Medicine, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire, USA
- 5VA Outcomes Group, Department of Veterans Affairs Medical Center, White River Junction, Vermont, USA
- 6Fletcher Allen Cardiology, Cardiovascular Catheterization Laboratories, Fletcher Allen Health Care and University of Vermont College of Medicine, Burlington, Vermont, USA
- 7Concord Hospital Cardiac Associates, Concord, New Hampshire, USA
- 8New England Heart Institute, Catholic Medical Center, Manchester, New Hampshire, USA
- 9Cardiovascular Medicine, Atlantic Cardiology Associates, Portsmouth Regional Hospital, Portsmouth, New Hampshire, USA
- 10Cardiac Services, Maine Medical Center, Portland, Maine, USA
- 11Central Maine Heart and Vascular Institute, Central Maine Medical Center, Lewiston, Maine, USA
- 12Northeast Cardiology Associates, Eastern Maine Medical Center, Bangor, Maine, USA
- 13Division of Nephrology, Tufts Medical Center, Boston, Massachusetts, USA
- Correspondence to Dr Jeremiah R Brown, Dartmouth-Hitchcock Medical Center, HB 7505, One Medical Center Drive, Lebanon, NH 03756, USA;
- Accepted 2 August 2011
- Published Online First 2 September 2011
Objectives This study evaluates the variation in practice patterns associated with contrast-induced acute kidney injury (CI-AKI) and identifies clinical practices that have been associated with a reduction in CI-AKI.
Background CI-AKI is recognised as a complication of invasive cardiovascular procedures and is associated with cardiovascular events, prolonged hospitalisation, end-stage renal disease, and all-cause mortality. Reducing the risk of CI-AKI is a patient safety objective set by the National Quality Forum.
Methods This study prospectively collected quantitative and qualitative data from 10 centres, which participate in the Northern New England Cardiovascular Disease Study Group PCI Registry. Quantitative data were collected from the PCI Registry. Qualitative data were obtained through clinical team meetings to map care processes related to CI-AKI and focus groups to understand attitudes towards CI-AKI prophylaxis. Fixed and random effects modelling were conducted to test the differences across centres.
Results Significant variation in rates of CI-AKI were found across 10 medical centres. Both fixed effects and mixed effects logistic regression demonstrated significant variability across centres, even after adjustment for baseline covariates (p<0.001 for both modelling approaches). Patterns were found in reported processes and clinical leadership that were attributable to centres with lower rates of CI-AKI. These included reducing nil by mouth (NPO) time to 4 h prior to case, and standardising volume administration protocols in combination with administering three to four high doses of N-acetylcysteine (1200 mg) for each patient.
Conclusions These data suggest that clinical leadership and institution-focused efforts to standardise preventive practices can help reduce the incidence of CI-AKI.
- Contrast media
- acute kidney injury
- patient safety
- quality improvement
- healthcare quality improvement
- comparative effectiveness research
- randomised controlled trial
- laboratory medicine
- hospital medicine
- quality measurement
Contrast-induced acute kidney injury (CI-AKI) results from exposure to radio-contrast agents given during invasive radiographical procedures. Radio contrast has been hypothesised to cause AKI through direct toxicity and via haemodynamic changes.1 2 CI-AKI occurs in 3–14% of patients and is associated with an increased in-hospital stay and long-term mortality.3 4 CI-AKI is the third most common cause of AKI in hospitalised patients.5 On average, patients developing CI-AKI stay 3.5 more days in hospital, costing patients and payers over $7500 in additional hospital costs.6
Reducing the prevalence of CI-AKI is a patient safety objective set out by the National Quality Forum.7 However, since 2006, preventive measures to reduce CI-AKI have been applied inconsistently.8 Reports of systematic efforts to apply this evidence in care settings are lacking in the literature. Specifically, the implementation of multidisciplinary quality improvement care teams to adopt effective evidence-based guidelines to reduce CI-AKI has been underused as a method to address this problem. Weisbord and colleagues found in the Veterans Administration system that 16% of high-risk patients who were eligible to receive prophylactic fluids failed to receive intravenous fluid expansion. When intravenous fluids were used, their dose and timing of administration varied significantly by treating specialty and procedure. Furthermore, contraindicated agents, including non-steroidal anti-inflammatory drugs and cyclooxygenase-2 inhibitors, were prescribed in 8% of patients. These data suggest that there is a considerable opportunity to improve the clinical effectiveness of CI-AKI prevention measures.9
We hypothesised variation in PCI practice patterns and the presence of institutional protocols explains the variability in CI-AKI outcomes in northern New England. To investigate this hypothesis, we conducted baseline focus groups of multidisciplinary care teams at 10 medical centres performing percutaneous coronary interventions (PCIs) in Vermont, New Hampshire, and Maine. We hypothesised that patient safety might be improved by reducing CI-AKI through modifiable practice patterns and prophylactic strategies.
Baseline procedural data were collected prospectively on 7287 consecutive patients undergoing PCI between 1 July 2008 and 30 June 2009 at 10 medical centres in northern New England. The cases are recorded in the PCI registry maintained by the Northern New England Cardiovascular Disease Study Group (NNECDSG), a voluntary regional consortium of clinicians, hospital administrators, and healthcare research personnel who seek to continually improve the quality, safety, effectiveness, and cost of medical interventions in cardiovascular disease.10 11 The NNECDSG has Institutional Review Board (IRB) approval for data collection and analysis at all participating centres. Additional IRB approval was obtained for the multidisciplinary focus groups at all participating centres.
Determination of CI-AKI incidence
All data collected in the NNECDSG PCI Registry are collected prospectively using the same definitions for all variables across centres. The last pre-PCI serum creatinine and highest post-PCI serum creatinine between the PCI procedure and discharge were used to determine the incidence of CI-AKI across medical centres. In an effort to move towards standardisation of definitions, we used the Acute Kidney Injury Network definition for CI-AKI: ≥0.3 (mg/dl) or ≥50% increase in serum creatinine from baseline within 48 h.12 Multivariable logistic regression was used to calculate adjusted rates of CI-AKI, adjusting for age, gender, priority of procedure (non-urgent, urgent, emergent), congestive heart failure, diabetes, use of pre-PCI intra-aortic balloon pump, and baseline estimated glomerular filtration rate based on National Kidney Foundation definitions using the Modification of Diet in Renal Disease equation (ml/min/1.73 m2): 186× (serum creatinine mg/dl)−1.154× (age)−0.203× (0.742 for women) × (0.180 for African American).13–15 Multilevel mixed-effects logistic regression (both fixed and random effects modelling) was conducted to test the differences across centres adjusting for the same covariates listed above.
Process mapping of CI-AKI prophylaxis
Baseline multidisciplinary clinical team meetings were conducted at all medical centres. Each centre organised their multidisciplinary team to include interventional cardiologists, cardiac catheterisation lab managers and technicians, nursing representation from the intensive care unit and/or holding areas, cardiology administration, and nephrology. Participants provided written informed consent to participate in the focus group according to each institution's IRB protocol. Each multidisciplinary team was asked to explain the process of care for an in-patient undergoing PCI (already on the hospital service) from the time PCI was recommended to 1 week after discharge. A research coordinator facilitated the meetings and the teams were then asked about any differences that typically happen for a same-day admission patient. Detailed information about key aspects of the centre's process on CI-AKI prophylaxis was collected. The teams were asked to describe their attitudes towards CI-AKI and prevention of CI-AKI. Teams were blinded to their rank-ordered adjusted rates of CI-AKI to prevent bias in reporting.
Each meeting was videotaped by a videographer and moderated by an experienced facilitator. Videotaped interviews were standardised and conducted in the same process for all centres. Ascertainment of centre protocols or standing order sets relevant to CI-AKI prophylaxis was collected (if available). Table 1 reports ‘variable’ when standard order sets or protocols were non-existent or order sets had a ‘write-in’ order for volume administration, N-acetylcysteine, or sodium bicarbonate (ie, the use of preventive therapies was at the discretion of the provider and not a team-wide mandate). Field notes were recorded. A third party transcribed each tape void of identifiable information. Videotapes were destroyed after transcription. The transcription was provided to the research team without names or identifiable information on the participants.
Qualitative analysis was performed using the grounded theory approach.16 In this approach, theory is created from the data and categories are not specified a priori. We used open coding to develop initial themes, followed by axial and selective coding to develop matrixes across medical centres. We defined aggressive CI-AKI prophylaxis as indicated by the presence of clinical leaders who devised a CI-AKI prophylactic protocol, implemented such protocol at their institution and had a heightened awareness of the need for CI-AKI prevention through volume expansion or sparing use of contrast.
Adjusted rates of CI-AKI were measured across 7286 consecutive patients undergoing PCI at 10 medical centres and varied from 1.9% to 10.1%, representing a fivefold difference in the incidence of CI-AKI (figure 1). Adjusted rates of CI-AKI are reported for patients undergoing PCI by centre; patients with a history of dialysis prior to PCI were excluded from the adjusted rates of CI-AKI. The regional average rate of CI-AKI was 6.28%. Baseline patient and disease characteristics are reported in table 2, rank-ordered by the adjusted rate of CI-AKI, denoted as A–J. Both fixed effects and mixed effects logistic regression demonstrated significant variability across centres even after adjustment for baseline covariates (p<0.001 for both modelling approaches). We report the adjusted rates of CI-AKI stratified by eGFR≥60 and <60 (ml/min/m2) in the online appendix.
Common processes of care for patients were identified from focus groups of providers at each medical centre. These processes were used to develop a matrix of procedural practice hypothesised to be associated with CI-AKI (table 1). The rates of CI-AKI are reported based on all patients undergoing PCI at the centre without pre-PCI dialysis, the process-level data in table 1 were developed in response to patients eligible for volume expansion or CI-AKI prophylaxis (serum creatinine ≥1.5 mg/dl or estimated glomerular filtration rate (eGFR) <60 ml/min/m2). The matrix was rank ordered according to the adjusted CI-AKI incidence across centres. Table 3 represents a second matrix on performing checks in the process of care.
The practice with the lowest adjusted rate of CI-AKI prevention in our region was identified as Centre A. Centre A incorporated a standardised protocol for CI-AKI prevention in 2005. The protocol included bypassing the midnight nil by mouth (NPO) orders to incorporate nil by mouth (NPO) orders limited to 4 h prior to the procedure. All patients with chronic kidney disease (eGFR<60 ml/min/m2) were started on an intravenous sodium bicarbonate protocol at 3 ml/kg/h for 1 h prior to the procedure. In addition, intravenous normal saline was given to patients until the sodium bicarbonate protocol was started approximately 1 h prior to the procedure. If a case was delayed, the sodium bicarbonate rate was reduced to 1 ml/kg/h until the procedure was performed. Intravenous sodium bicarbonate was continued after PCI at a rate of 1 ml/kg/h for 6 h. Patients were also given 1200 mg of N-acetylcysteine orally twice before and twice after PCI. Iopamidol contrast was used in all procedures.
The practice with the second lowest adjusted rate of CI-AKI was Centre B. Similar to Centre A, Centre B had an established volume administration protocol, used 1200 mg N-acetylcysteine twice before PCI and twice after PCI, and all cases were conducted with iopamidol. Centre B also had an option to bypass the nil by mouth (NPO) orders after midnight and changing the nil by mouth (NPO) status to only 2 h prior to PCI. However, Centre B differed from Centre A in that it used only intravenous normal saline for volume expansion before and after PCI.
Standardising practice reduces variation. The main contributor to success from our two centres with the lowest rates of CI-AKI was an evidence-based standardised CI-AKI prophylaxis protocol. Centre B called this their ‘Power Protocol’ and Centre A sought to ‘Keep it simple so everyone knows what to do: hydrate patients and alkalinise the kidneys’.
Centres A and B were the only centres to rank significantly below the regional average; all other centres ranged between 4% and 10.1% of the post-procedure CI-AKI rate. All centres routinely had patients nil by mouth (NPO) after midnight, with the exception of Centre F, which occasionally had patients nil by mouth (NPO) after 06:00 on the day of the procedure. All centres had variable volume administration practice, most using physician written orders or electronic written orders without a standardised protocol. Most other centres used a lower dose N-acetylcysteine at 600 mg (as opposed to 1200 mg) and sometimes this was ordered only three times (once before and twice after the procedure) as opposed to four times (twice before and twice after the procedure). There was also variability in the use of contrast agent; the two centres with the lowest CI-AKI rates used the same contrast agent on all patients (regardless of high or low risk of CI-AKI), while other centres would use a low-osmolar contrast agent for low-risk patients and an iso-osmolar agent for high-risk patients. The findings suggest that low adjusted rates of CI-AKI may be obtained with clinical leadership and aggressive prophylaxis through volume expansion.
Key patterns from centres with the lowest incidence of CI-AKI were identified as potential targets for CI-AKI prevention at other centres (box 1). Four practice patterns were identified: changing nil by mouth (NPO) liquid orders from midnight to 2–4 h prior to the procedure; standardisation of volume administration protocols; use of high-dose N-acetylcysteine (1200 mg); and iopamidol low-osmolar contrast agent.
Change strategies for evaluation and further consideration
Nil by mouth (NPO) with clear fluids allowed up to 2–4 h prior to PCI
N-acetylcysteine 1200 mg orally every 12 h×4, first dose 18:00 night before PCI: 1200 mg dose @ 18:00, 6:00, 18:00, 6:00
Standardisation of volume administration prophylaxis
Intravenous NS at 1.5 ml/kg/h at 22:00 night before PCI, continue until intravenous NaHCO3 protocol begins.
Intravenous NaHCO3 at 3 ml/kg/h for 1 h pre PCI (reduce to 1 ml/kg/h if PCI is delayed)
Intravenous NaHCO3 at 1 ml/kg/h for 6 h post PCI
Recommendation: post-discharge follow-up at 3–5 days to determine CI-AKI, if persistent, weekly labs until resolved
Iopamidol: low-osmolar contrast agent
CI-AKI, contrast-induced acute kidney injury; NaHCO3 = sodium bicarbonate (1000 ml D5W mixed with 150 mEq NaHCO3); NS, normal saline; PCI, percutaneous coronary intervention.
Clinical leadership was a key driver in Centres A and B. The above tools of CI-AKI prophylaxis may be markers for the level of aggressiveness in attitudes towards CI-AKI prophylaxis and proactive clinical leadership. One observation from the focus groups was the level of aggressiveness and attitudes towards implementing and using these prophylactic tools and proactive leadership (championing) by clinical faculty that drove institutional changes resulting in aggressive CI-AKI prophylaxis protocols. Centres A and B demonstrated aggressive attitudes towards CI-AKI prophylaxis during the baseline focus groups; these aggressive attitudes were also captured through use of standard order sets to drive the use of aggressive volume administration (Centres A and B), use of N-acetylcysteine (Centres A and B), and sodium bicarbonate (Centre A). Both centres had interventional cardiologists who proposed changes to their standard order sets and implemented those changes. Centre A called this protocol the ‘CI-AKI Prophylaxis’, Centre B called it their ‘Power Protocol’. While these tools may be useful in preventing CI-AKI, we must also look at the aggressive prophylactic culture of the practicing operators to obtain an understanding of true CI-AKI prophylaxis. It may require quality improvement tools and active leadership with attention to detail, volume administration status of the patient prior to and after the intervention, and sparing use of contrast to improve the rates of CI-AKI.
We identified collective quality improvement efforts that may explain the collection of processes and protocols aimed at CI-AKI prevention. Through systematic focus groups, we ascertained each centre's culture, processes and protocols for CI-AKI prevention. The major changes that emerged were the following: reducing nil by mouth (NPO) after midnight orders to 2–4 h prior to PCI, volume expansion protocols, use of N-acetylcysteine, contrast agent, and other procedural factors such as the limited use of left ventriculography for patients with eGFR less than 60. We considered these trends in practice and clinical leadership to exist within each centre collectively. This collective approach suggests that any of these prophylactic efforts may not be individually driving the prevention of CI-AKI. The role of the clinical champion as described above was one element that emerged from these focus groups and was attributed to the top performing centres. It is believed that the emergence of this trend in clinical leadership may well encompass a collective group of culture, process, and attitudes that were collective approaches towards the prevention of CI-AKI and deserves further investigation.
In our investigation of 10 medical centres performing routine PCI, we found fivefold variability in adjusted rates of CI-AKI. By means of facilitating focus groups and collecting qualitative data, we were able to identify key attributes of centres with the lowest rates of CI-AKI in our region. The practice identified in Centre A was achieved from establishing a simple CI-AKI prophylaxis protocol, including nil by mouth (NPO) orders for only 4 h, iopamidol contrast, a standardised sodium bicarbonate protocol (3 ml/kg/h for 1 h before and 1 ml/kg/h for 6 h after PCI) and use of 1200 mg N-acetylcysteine twice before and twice after PCI.17
The discussion over volume administration has been an ongoing debate in the literature, ranging from a volume administration protocol with 0.45% normal saline for 12 h before and after angiography18 to a sodium bicarbonate protocol with isotonic (154 mEq/litre) infusion of sodium bicarbonate before and after iopamidol administration (370 mg iodine/ml). The largest randomised trial by Maioli is consistent with our qualitative findings, in that the two centres with the lowest rates of CI-AKI used standardised protocols for CI-AKI prevention, but one used sodium bicarbonate and the other used normal saline—both with successful results.19 Alkalinisation of the urine with sodium bicarbonate has been shown to reduce labile iron-dependant free oxygen radicals to a larger degree than saline alone.17 Both protocols demonstrated reductions in CI-AKI in multiple trials and subsequent meta-analyses. Most recently, a meta-analysis by Kunadian et al demonstrated the superiority of sodium bicarbonate over normal saline with a 67% risk reduction for CI-AKI, yet no difference in other clinical endpoints.20 In our qualitative analysis the practice with the lowest rates of CI-AKI used a protocol for sodium bicarbonate, yet the second lowest incidence of CI-AKI had success using a normal saline protocol. While other centres used the sodium bicarbonate protocol from time to time, there was a lack of standardisation of this approach for all patients. Current evidence suggests that volume administration should remain a central focus for preventing CI-AKI.
The most supportive evidence for high-dose N-acetylcysteine was published by Marenzi et al and demonstrated high-dose N-acetylcysteine at 1200 mg significantly reduced CI-AKI (3%) compared with low-dose N-acetylcysteine (600 mg, 8%) or placebo (33%) after primary angioplasty.21 Trivedi et al (2009) used a meta-analysis to compare randomised trials of high-dose N-acetylcysteine (a daily dose greater than 1200 mg or a single dose greater than 600 mg within 4 h of contrast exposure). The results showed high-dose N-acetylcysteine significantly reduced the occurrence of CI-AKI by 54% when compared with low-dose N-acetylcysteine (HR 0.46; 95% CI 0.33 to 0.63).22 However, the largest trial to date, the ACT trial with 2308 patients, recently reported no effect of N-acetylcysteine over placebo (RR 1.00; 95% CI 0.81 to 1.25), suggesting this agent may have no efficacy in preventing CI-AKI.23 It should be noted that our reported research was conducted prior to the public release of the ACT trial and represented practice patterns from the current best evidence. Although N-acetylcysteine has been postulated to cause an artificial decline in serum creatinine,24 25 recent evidence has shown that in the absence of contrast, N-acetylcysteine had no effect on serum creatinine or cystatin C.26
Prevention of CI-AKI has been investigated through the combination of interventions. Recently, Brown and colleagues reported the results of a meta-analysis evaluating randomised trials using a combination of volume administration and N-acetylcysteine. They demonstrated the combination of sodium bicarbonate and N-acetylcysteine provided additional prevention of CI-AKI over normal saline with or without N-acetylcysteine.27 This finding from clinical trials was supported by the experience of Centre A, whereby a simple, yet concrete protocol of combination prophylaxis of intravenous sodium bicarbonate and high-dose N-acetylcysteine (1200 mg) resulted in the lowest rate of CI-AKI in northern New England. However, this should be balanced with the recent release of the ACT trial23 and may suggest the more aggressive approach to CI-AKI prophylaxis was driven by clinical leadership and centre culture.
Contrary to some of our a priori hypotheses, there were CI-AKI techniques that did not prevail in the best-practice approach. Surprisingly, automated contrast devices were not a key attribute of reducing CI-AKI despite initial evidence demonstrating significantly less use of contrast and a 31.1% risk reduction for CI-AKI over hand-held devices.28 Nor was the systematic use of the maximum allowable contrast dose a key factor for CI-AKI prevention.29 30
First, careful readers will note that serum creatinine may have a delayed latency prior to reaching its peak. In fact, Maioli reported that when CI-AKI is assessed up to 5 days post PCI, 40% of the patients had not reached AKI by 48 h.19 Therefore, we may be underestimating the incidence of AKI in this study. The definition of AKI by 48 h is, however, consistent with real world practice. Second, we have chosen to use the Acute Kidney Injury Network definition of CI-AKI.12 The diagnosis of CI-AKI using this definition has undergone substantial evaluation and the field of AKI is moving towards standardising the definition of AKI in all clinical contexts using the ≥0.3 (mg/dl) or ≥50% increase in serum creatinine. Some experts believe that ≥25% increase in serum creatinine is too low of an elevation in the acute setting and ≥0.5 (mg/dl) misses patients with AKI. We have included the 25%, 0.5 (mg/dl) and 25% or 0.5 (mg/dl) definitions in table 2 for comparison to previously reported randomised controlled trials and the Contrast-Induce Nephropathy Consensus31 and re-evaluated the inference from the results. The results do not change with respect to the centres with the best rates of CI-AKI; changes include Centre D to swap with Centre C and Centre G to move into the Centre E position. Upon re-evaluation of these two changes, the inference from the results remains the same, with added support for the use of high-dose N-acetylcysteine at 1200 mg. Third, we recognise the importance of contrast volume, however it is not currently reported in our registry or other national registries at this time. The NNECDSG and other registries such as the National Cardiovascular Data Registry are working towards including contrast volume in future data collection forms. However, we can assume contrast exposure based on the number of vessels intervened upon and the number of stents deployed. Fourth, we used a quantitative approach to summarise centre characteristics and to develop risk-adjusted rates of CI-AKI. We used qualitative methods to develop matrices to characterise the variation in the underlying practice. The qualitative data from tables 1 and 3 represent data collected from focus groups and therefore are captured as centre-level data. This limits the ability to conduct sophisticated regression analyses to ascertain cause and effect. While this is a limitation, there are strengths associated with having the qualitative data to ascertain process-level characteristics that may be the underlying reason for variation in adjusted rates of CI-AKI. Through the use of these methods, we suggest several strategies that may account for lower rates of CI-AKI. Further investigation with patient-level data is needed to confirm the hypotheses generated from the qualitative findings. Centres reported their practice patterns, but evidence was not collected prospectively to determine whether these practices were in fact conducted on the floor or in the catheterisation laboratory. However, this self-reporting provides a snapshot that allows for a useful comparison across practices, the goal of this study. In addition, these initial findings can serve as the hypotheses for future more detailed investigation of the processes of care for patients undergoing PCI.
Conceptual framework for prevention of CI-AKI in northern New England
The Northern New England Cardiovascular Disease Study Group has chosen to evaluate the relationship of high-intensity quality improvement efforts on patient safety and CI-AKI. This report summarises the initial steps in understanding the variation in adjusted CI-AKI rates and practice patterns in the region. We now plan to start a high-intensity quality improvement intervention based on the northern New England and Vermont Oxford frameworks.32–35 High-intensity quality improvement interventions will be focused at each centre and will include training in continuous quality improvement methods to engage clinical practice and evaluation of that practice through the following: evidence-based medicine; collaborative learning; developing a habit for systems thinking; developing a habit for change; and robust data systems to assess and manage process changes.
In conclusion, we have identified best practices associated with a reduction in CI-AKI. Future research should aim to validate these findings and determine if the reduction in CK-AKI translates into clinically meaningful outcomes.
Funding This project was supported by grant number K01 HS018443 (Dr Brown) from the Agency for Healthcare Research and Quality and K24 DK078204 (Dr Sarnak) from the National Institute of Diabetes and Digestive and Kidney Diseases. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality or the National Institute of Health.
Competing interests None.
Ethics approval Center for the protection of human subjects at Dartmouth, others.
Provenance and peer review Not commissioned; externally peer reviewed.