AACC Annual Meeting – Houston, TX – July 31, 2013

NIDDK's National Kidney Disease Education Program (NKDEP) Laboratory Working Group (LWG) joint meeting with the International Federation of Clinical Chemistry (IFCC) Working Group for Standardization of Albumin in Urine (WG-SAU)


Greg Miller, Lori Bachmann, David Bunk, Joanna Camara, Joris Delanghe, John Eckfeldt, James Fleming, Lars-Olaf Hansson, Glen Hortin, Hans-Joachim Kytzia, Anthony Killeen, Andrew Narva, Mauro Panteghini, Karen Phinney, Claus Prümper, Heinz Schimmel, David Seccombe

Meeting Minutes

Summary of Action Items

  • A task force will be created to work on creatinine performance recommendations for pediatric ranges. G. Miller will identify a clinical chemist to lead this group.
  • J. Delanghe will develop a proposal for developing guidelines for using standardized creatinine appropriately in various equations used for specialized clinical conditions.
  • An update will be made to the recommendations for Cystatin C recalibration and estimating equations on the NKDEP website.
  • A. Narva and L. Bachmann will decide on a strategy for determining decision values for urine albumin.

Creatinine Performance Recommendations Outside the 1.0-1.5 mg/dL Interval

(especially lower values seen in pediatric patients) Greg Miller, PhD, Lab Working Group Chair

  • Initial work on creatinine standardization and eGFR equations focused on performance goals in the narrow creatinine range of 1.0-1.5 mg/dL. Some manufacturers are asking for guidance in the lower creatinine values seen in the pediatric range. George Schwartz is interested in participating in a task force that will be created by the LWG to work on developing concrete recommendations.

Discussion about how to proceed with task force

  • A desirable outcome for the task force is a publication explaining issues of creatinine measurement in a pediatric population.
  • D. Seccombe stated that in their accuracy-based program there is a 30 percent failure rate in the pediatric ranges by labs with standardized creatinine methods. Many are using single point calibration and forcing the line through zero. There is benefit to be gained by improving accuracy in the lower levels.
  • J. Eckfeldt suggested that based on data generated by this group, comparing Jaffe to enzymatic, there is not enough accuracy by any method at the level of 0.3-0.5 mg/dL and recommends that the task force address whether or not creatinine is the right measurand in this range to estimate GFR.
  • J. Delanghe agreed with J. Eckfeldt and stated that apart from the analytical issues, there are physiological differences between adult and pediatric patients. In children, tubular secretion is more important than in adult patients. Even if analytical goals are reached, there is still a problem with interpretation. At lower concentrations, creatinine is no longer a good marker for GFR and tubular secretion. Therefore, using alternative markers is a better approach than trying to improve creatinine measurement at lower concentrations.
  • L-O. Hansson suggested that pre-analytical issues, such as what the patient has been eating, affect creatinine more at lower values.
  • G. Miller recommended setting performance goals and allowing the manufacturers to determine the best way to meet them.
  • J. Eckfeldt expressed that nephrologists are needed on the task force.
  • A. Narva stated that G. Schwartz is measuring GFR in pediatric patients with normal to near normal kidney function.
  • G. Schwartz, J. Fleming, J. Eckfeldt, J. Delanghe, and G. Hortin volunteered to serve on the task force. G. Miller will identify someone to lead it.
  • G. Hortin stated that the task force needs data from a referee MS method and J. Eckfeldt suggested that if G. Schwartz is determining GFR, perhaps the creatinine could be measured by an MS method to help assess if improving accuracy will improve interpretation.
  • A. Narva suggested that the task force could apply to the CKiD Study to obtain samples for this purpose.
  • G. Miller stated that Neal Dalton may be interested to work with the task force to make IDMS creatinine measurements and the task force may have to work with EQA providers to obtain data for low creatinine concentrations.

Derived Formulas Based on Creatinine in Various Medical Specialties and Impact of Standardization

Joris Delanghe MD, PhD, Ghent University Hospital, Belgium

  • Despite huge progress in creatinine standardization, some problems remain.
  • It is impossible to re-express all current drug-dosing recommendations for use with standardized creatinine values. For the majority of patients and for most drugs, there is little difference in the calculated dose according to the equation used.
  • If using eGFR in very large or very small patients, reported eGFR should be multiplied by the body surface area to obtain eGFR.
  • Assessing kidney function using alternative methods (e.g. measured CrCl or measured GFR using exogenous markers) should be considered for drugs with narrow therapeutic indices, or for individuals in whom eGFR and eCrCl provide different estimates, or for patients in whom estimates based on creatinine are likely to be inaccurate.
  • Vigilance is required to tackle the side effects of creatinine restandardization.
  • MDRD has limitations (CrCl value, age, standardization) which should be respected.
  • eGFR data are characterized by an important error. Therefore, derived calculations (e.g. ΔeGFR/Δt) are not to be recommended.
  • In children, Cystatin C determination might offer an interesting alternative for GFR estimation. The recent availability of the IFCC Cystatin C standard will contribute to the popularization of Cystatin C as a renal function marker.
  • Confusion in hospital wards remain.
  • Inventory of remaining problems is needed.
  • A task force might be a good option to tackle the problem.


  • G. Hortin suggested that most of the problems are related to eGFR calculations more than creatinine standardization; J. Delanghe replied that the MELD score used for placing patients on the liver transplant list is an example where creatinine standardization causes a problem because the equation was derived using creatinine results prior to standardization. A. Narva commented that many clinicians don’t understand that eGFR has limitations.

Update on NKDEP Pharmacy Working Group and HIT Working Group

Andrew Narva, MD, FACP, NKDEP Director

  • NKDEP has created two new working groups using LWG as the model:
    • Pharmacy Working Group (PhWG):
      • Chair is Amy Barton Pai, Pharm.D, BCPS, FASN, FCCP, from Albany College of Pharmacy and Health Sciences
      • Objectives of this working group include:
        • Increase the number of community pharmacists who have the requisite CKD knowledge to provide CKD education.
        • Provide pharmacists with the tools and resources to educate patients about CKD.
        • Establish community pharmacy sites as a place for patients to receive CKD education.
        • Empower patients to initiate discussions with their pharmacists about their health, particularly CKD.
      • Short term goals are to: 1) develop CE opportunities on CKD for pharmacists and 2) identify opportunities to incorporate messaging that positions pharmacists as a source of health information and support into existing and new NKDEP materials and activities.
      • Future direction is to incorporate eGFR in electronic prescriptions to enable pharmacists to review prescriptions and counsel patients regarding OTC drugs that should be avoided.
    • Health Information Technology (HIT) Working Group
      • Chair is Uptal Patel, MD from Duke University
      • Primary goal is to enable and support the widespread interoperability of data related to kidney health among healthcare software applications to optimize CKD detection and management. This effort will allow expansion of NKDEP’s Lab Working Group efforts by mobilizing data related to the detection, prevention, and management of kidney disease to flow freely through electronic health records, personal health records, decision support systems, disease registries, surveillance systems, etc.
      • Secondary goal is to help satisfy meaningful use requirements, promote CKD surveillance & registries, provide data that is more suitable for high-quality research on kidney disease, improve management of CKD through integration with decision support systems, PHRs, mHealth applications, ePrescribing, etc.
      • New regulations will require the ability to search electronic health records by LOINC codes. This group will try to make sure that UACR, creatinine, and eGFR are in the top group of LOINC codes that can be used for a search. When this happens, patients can be identified to improve outcomes. This has implications for surveillance of CKD, patient care and research.
      • Objectives of this working group include:
        • Agree upon data components critical to kidney disease detection and management that should be included in EHRs and readily available to patients and providers.
        • Work with major EHR companies (e.g., NextGen, Epic, Athena Health) to incorporate agreed upon data components as searchable measures in EHRs. Work with major electronic prescription companies (e.g., Express Scripts, Surescripts) to include eGFR within electronic prescriptions. (Recommended by NKDEP’s Pharmacy Working Group.)
        • Identify data components and relationships critical to management of advanced CKD that should be included in EHRs and readily available to patients and providers.
        • Engage industry (lab services, pharmacies, etc.), government, and organization partners to address selected barriers and move Working Group recommendations forward.

Discussion about estimating equations used in different settings

  • J. Delanghe would like a task force to further explore and develop educational recommendations for how to improve estimating equations that incorporate creatinine.
  • A. Narva stated that dosing recommendations in package inserts for existing drugs are based on the Cockcroft-Gault equation and are not likely to change, but the number of cases where this is critical in adults is probably overplayed.
  • D. Seccombe is working in Canada with a group that is interested in providing creatinine and other tests germane to chronic disease management to pharmacists. This is a point of care instrument with 21 analytes and includes a renal panel, cardiovascular panel, etc. with testing performed in the pharmacy and data transferred to the cloud. In Canada, there is a push to empower the patient. A. Narva commented that it is a challenge to ensure that information is reported and accessible to all different providers participating in a patient’s health care. D. Seccombe agreed and also felt that because of the complexity of getting all of the systems to communicate, it is better to move the information to the patient.
  • G. Miller suggested that the LWG focus on developing guidelines about how to use the equations appropriately and improve awareness of their limitations. G. Hortin asked if there should be investigation into the effect of the new standardized creatinine values on the Cockcroft-Gault equation. G. Miller stated that Lesley Inker et. al. published a paper indicating that for most drugs, it doesn’t matter which equation is used because although there are errors, they did not misclassify the drug dosing category, except for those with narrow therapeutic windows. New drugs are now being evaluated with measured GFR.
  • A. Narva suggested that this is a core task for NKDEP and that the Program could start by looking at what is on the website and determine if changes are needed.
  • J. Delanghe agreed to develop a proposal for how best to proceed with this issue.

Standardization of Cystatin C Measurements and eGFR Equations Based on Cystatin C

  • Report on Lund Conference, February 2013,
    John Eckfeldt, MD, PhD, University of Minnesota
    • The meeting was hosted by Anders Grubb. The goals were to discuss clinical use of Cystatin C measurements, standardization of measurement of Cystatin C, and equations to estimate eGFR using Cystatin C. Attendees were from the U.S., Asia, and Europe and were from research labs, clinical medicine, nephrology, and IVD industry.
    • Data from the pooled samples collected from patients with kidney disease that were assayed by several labs in preparation for this meeting showed that it is possible to standardize Cystatin C and manufacturers need to be sure they are traceable to ERM/IFCC DA471. At present, some assays appear not to be traceable.
    • For now, there are likely to be multiple eGFR equations recommended by different research groups and organizations.
    • T. Killeen asked if the CAP survey material used was commutable. The caveat with interpretation of the CAP data is that it is not from real patients with kidney disease, so there is likely to be matrix effect; the product has not been evaluated for commutability. J. Eckfeldt suggested that it would be preferable to have a 3 level set of frozen, pooled serum to use for an accuracy- based survey as well as for standardization by manufacturers.
  • IFCC Working Group for Standardization of Cystatin C Update, Lars-Olof Hansson, MD, PhD, Karolinska Hospital, Sweden
    • Ten patient sample pools were run in duplicate using six different manufacturer’s assays in four laboratories over five days. If data from one manufacturer’s assay is excluded, the range of results for the other five assays is very tight. Precision of Cystatin C results is especially important in the GFR range of 60–120 mL/min/1.73m2.
    • These pools have now been sent to Siemens in the U.S. and Marburg, Roche in Pensberg, Germany, and Amanda Hyre Anderson at University of Pennsylvania.
    • Gunnar Nordin at EQUALIS in Sweden is interested in managing distribution of plasma pools to diagnostic companies.
    • An interesting observation was made when creatinine results were compared to Cystatin C using pools from patient samples. There was a different correlation depending on kidney function and it appears that this may be related to the size of creatinine and Cystatin C molecules.

Report on AACC/AdvaMedDx/FDA Forum; The road to harmonization: regulatory considerations for recalibrating IVD devices, June 201

Greg Miller

  • There are challenges for U.S. manufacturers when an international program recommends that an assay be recalibrated. In the U.S., this requires a 510K submission to the FDA which is an expensive process.
  • At the 2012 LWG meeting, it was decided to work with AACC to initiate a meeting of interested parties with the FDA.
  • AACC and AdvaMedDx organized a conference for manufacturers and the FDA to discuss problems with recalibration applications. A summary of this conference will be published on the www.hamonization.net website.
  • It was agreed that it is preferable that calibrations be traceable to an accepted reference system whenever one exits.
  • It is acceptable to use calculations based on a relationship between an old calibration and the new one.
  • The FDA has a triage program that simplifies the 510K process and the FDA strongly recommends that manufacturers contact them when a recalibration is necessary to determine exactly what data is needed.
  • Manufacturers can collaborate and come to the FDA as a group to discuss recalibration issues. AdvaMedDx has offered to assist manufacturers with this collaboration.
  • The FDA prefers that manufacturers contact them when international reference systems are first being established rather than waiting until the end.
  • The next step is to develop in cooperation with the FDA a guidance document for recalibration to conform to international recommendations.

Cystatin C Education Recommendations

Greg Miller

  • It has been demonstrated that the ERM/IFCC DA471 reference material is commutable for a large number of methods (data to be published by IFCC/IRMM WG) and is suitable for standardization of methods. Everyone agreed that the LWG needs to modify the information presented on the NKDEP website to include the information presented at the meeting by J. Eckfeldt and L-O. Hansson and recommend that manufacturers recalibrate their assays.


  • G. Miller confirmed that the focus group was a generic discussion about working with the FDA when recalibration of an assay is needed and not specifically about Cystatin C.
  • J. Eckfeldt asked why we did not hear about the 510K process when creatinine recalibration was needed and G. Miller responded that it was because the volume of creatinine testing is so much greater than Cystatin C, that manufacturers proceeded to submit 510k for recalibrated methods. Cystatin C is not commonly performed at this time so the return on investment is smaller.

Urine Albumin Harmonization Investigation

Lori Bachmann, PhD, Virginia Commonwealth University

  • The objectives of the harmonization study were to assess the current state of agreement among routine urine albumin procedures using native patient samples, evaluate analytical performance characteristics of the methods, evaluate commutability characteristics of JSCC and diluted IRMM ERM-DA470k/IFCC reference materials, and assess the utility of candidate reference materials for use in standardization of routine methods.
  • The bias was greater than ± 10 percent for 13 of 16 quantitative methods indicating that standardization is needed.
  • Several methods had non-constant bias with concentration suggesting that calibration technique needs investigation.
  • Dilution caused changes in bias for some methods.
  • CVt was greater than 10% for five methods showing that improvement in precision is needed for those assays.
  • In general, sample-specific effects were not clinically significant.
  • Recommendations for manufacturers include: 1) method imprecision should be reduced for those methods with higher CVs, especially at low urine albumin concentrations that may influence CKD risk assessment, 2) methods that showed a non-constant calibration bias over the measuring interval should be investigated and this problem corrected, and 3) methods with poor dilution protocols should improve their user information to include specified dilution buffers and validate their protocols for matrix bias.
  • The harmonization assessment manuscript has been submitted to Clinical Chemistry.
  • Next steps include assessing the commutability characteristics of IRMM and JSCC candidate reference materials and evaluating the improvement in agreement that may be possible based on mathematical recalibration using these reference materials.

NIST Reference Materials Update

Karen Phinney, PhD and David Bunk, PhD, National Institute of Standards and Technology

  • IST SRM 2925 Human Serum Albumin is a primary reference material intended for calibration of the candidate reference method procedure and is not for routine assays because a commutability study has not been done. It is a frozen recombinant human albumin and should be available in 4-6 months.
  • NIST SRM 3666 is a matrix certified reference material that is in the preparation stage. It will be albumin in frozen human urine that has 4 levels in 1 mL aliquots and should be available in 10-12 months.


  • It was agreed that 40-50 aliquots of SRM 3666 should be reserved for commutability studies and possibly another 20 to validate new assays that come on the market.
  • There was a discussion about urine stability and freeze-thaw effects. A. Grubb has developed a method using HCl and EDTA to stabilize urine. With regard to freeze-thaw effects, L. Bachmann stated no effect was seen with a single thaw in the harmonization study. J. Eckfeldt commented that he has seen albumin degradation in very old samples that had been stored for decades at -70 C. In these samples, the nephelometric assay gave very low results and the mass spectroscopy assay gave results close to the original value, suggesting that there is a denaturation of albumin over this time. D. Bunk stated that they will have to evaluate stability with an assay that is not mass spectroscopy.
  • D. Bunk commented that creatinine can be measured in the frozen urine and perhaps it would be good to assess commutability of both creatinine and albumin.

Urine Albumin Reference Measurement Procedures Validation: NIST and Mayo Clinic

Karen Phinney, PhD and David Bunk, PhD, National Institute of Standards and Technology and John Leiske, MD, Mayo Clinic

  • D. Bunk was asked what was next for the candidate reference procedure, now that SRM 2925 is available. He suggested three different approaches: 1) a validation study alone, 2) a validation study and SRM 3666 certification, or 3) a validation study, SRM 3666 certification and the commutability study. He suggested that adding creatinine will complicate the protocol but most manufacturers can run urine albumin and creatinine on the same platform.
  • G. Miller expressed that option three seemed to be the most efficient.
  • J. Eckfeldt commented that option one needs to be done first to be sure that we get the same results and then perform certification and commutability.

Update on Other Urine Albumin Standardization Activities

  • Adsorption to Containers
    Greg Miller
    • Results of this project have been submitted for publication.
  • Physiologic Variability Investigation
    Greg Miller
    • This project was discussed with Jim Ritchie and Dave Koch at Emory and they both indicated that they do not have the time or the resources. Therefore, the project will be conducted at Virginia Commonwealth University. It is planned for 2014.
  • Establishing Reporting Cut-offs for Urine Albumin
    Andrew Narva
    • This project is just being started. It will be difficult to establish cut-offs when all of the data is from non-standardized assays. There were two major issues determined from a literature review: 1) a lower cut-off should be considered because there is evidence that the risk of cardiovascular disease occurs at levels lower than 30mg/g Cr, and 2) stratifying results by gender should be considered since women excrete less creatinine which could result in over reporting of women with kidney disease. The educational challenge here is made more difficult because KDIGO recommends a cut-off of 30 mg/g Cr and although this is an opinion-based recommendation, it is rapidly being cast in stone. Perhaps a working group is needed that meets online rather than in person. A. Narva and L. Bachmann will meet to determine a strategy. G. Miller commented that realistically, it will be 5-6 years before standardization of urine albumin occurs and A. Narva added that when L. Bachmann’s urine albumin harmonization paper is published it will provide compelling evidence that standardization is needed.
  • Lieske et al. A reference system for urinary albumin: current status. Clin Chem Lab Med. 2013 May;51(5):981-9. doi:10.1515/cclm-2012-0768.
    • This paper describing the status of a reference system for urinary albumin has been published and is available online.

Other items, next steps

Greg Miller

  • K. Phinney announced that NIST has issued SRM 3667 Creatinine in frozen human urine reference material which provides an initial step for a urine creatinine reference system.
  • J. Eckfeldt commented that data has just been returned from a new CAP accuracy-based survey using pooled urine samples. This will provide information about the status of urine albumin measurement in the U.S. and Canada. Three more samples will be mailed in early fall. A question was raised by L. Bachmann regarding the stability of the albumin depending on how the sample was mailed.
  • D. Seccombe stated that they did a study that has been completed in which urine pools augmented with purified human albumin were used. Two pools of urine were made, one with low creatinine and one with slightly higher creatinine. Each creatinine pool was used to prepare six levels of albumin. Target values were assigned gravimetrically and by J. Lieske’s albumin IDMS method. The intercept of the comparison between gravimetric and the IDMS method values gave the endogenous albumin in the pools. Samples were sent to 15 U.S. and 10 Canadian laboratories and assayed in triplicate on each of 3 days. Ten individual patient samples were also sent as check samples. Results showed that generally there is a uniform picture across platforms with more bias at lower levels. This study will be published.
  • H. Schimmel stated that we have to be careful when comparing immunologic methods to mass spectroscopy methods because they are measuring different analytes. The scatter of different routine assay results may be less if they are compared to each other rather than to the mass spectroscopy method. L. Bachmann has the data in the harmonization study to investigate this issue.