Acute Kidney Injury is a major cause of death and morbidity in the UK. Earlier this month RCPE UK Consenus Conference on management of acute kidney injury (AKI): the role of fluids, e-alerts and biomarkers issued a consensus statement that e-alerts should prove valuable in the early identification and management of AKI. A national approach was recommended to developing this approach with appropriate audit and research to identify an impact of e-alerts on patient outcomes. The statement is welcome and timely however given the fact that national approaches to addressing such issues are often slow, and the need to manage the issues around early identification of AKI, or developing AKI, pressing, it would seam sensible that some simple measures were put in place that have a positive impact on patient care that do not conflict with a national approach. The simple expedient of laboratory reports indicating that a patient's creatinine measurement has changed significantly since the last time it was measured should enable this. The change may or may not indicate established or developing AKI, but at least in indicates that something is happening that requires the medical professional to take note. At this point they can apply whatever criteria that are extant to make a judgement around the particular patient and deal with their condition appropriately.
The use of Reference Change Values provides a tool to identify significant change. This concept enables identification of a percentage change in consecutive results that can be identified as significant in a patient. The formula takes acount of the variability we see in creatinine measurements and also the average amount of fluctuation that you see in a measurement because of the natural rhythms. Simply applied, if the serum creatinine changes by 14% you can be 95% sure that change in creatinine is significant and not due to random variation resulting from the measurement and the natural biological variation combined. This is a sit up and take notice change. If it changes by 20% you can be 99% sure that this is a significant change and something is really going on. Laboratory computer systems have differing capabilities in managing the flagging of results such as this, but using the commonly fetured delta check functions using values of 14% or 20% means that they can employ the concept of RCV in a way that hides any complexity from those that receive the report. Simply the user sees the flag and think has gone up I should be thinking about why. The rest is about education and training.
The concept of RCV is internationally recognised as a concept. It is identified as a concept with HL7 messaging and as such it can be carried electronically and could be used to trigger flow of patients into clinical pathways. It is not proposed that RCV, in the context of AKI, is any way a substitute for the appropriate application of international diagnostic criteria or e-alerts based on those criteria, it is simply proposed that the RCVs for a rise in creatinine are adopted as flags to identify patients who need attention. It will enable some patients with established AKI to be identified and present a warning to clinicians that that some patients may be heading in that direction. It is a blunt tool, but it has value (see Khalid et al).
The use of RCV was advocated in the 2006 RPE Consensus statement dealing with the early chronic kidney disease (CKD). Use of RCVs thus has a role in acute and chronic kidney disease management. As CKD patients who develop AKI have a poorer prognosis than those with out, it is important to identify significant detrioration in renal function quickly. The best way to follow the renal function of a patient once a patient has been classified to a CKD stage on the basis of e-GFR is by changes measurments with time. If the creatinine changes by more than the RCV that this should trigger the attending medical profession to assess whether this is an expected rate of rise or something acute.
The reporting of significance of change in biochemical measurements by UK laboratories is uncommon, whereas induction of population based reference ranges to assess "normality" is almost universal. The issue here is that most of the work carried out by Clinical Biochemistry laboratories is about monitoring disease and therefore the focus of users is more often on change. The difficulty that we have with creatinine measurement is that the average variability within a person is quite narrow, but if you look at one creatinine result in a large population of people you will see a very wide spread of values. The reference ranges that laboratories use are usually derived from large populations. This means that an elderly lady can have a creatinine result that falls inside the range of values for the population and be judged "normal, but if you look at her previous values you might find it has risen by 30 %or even 40% and in fact she is far from normal. It is in situations like this that the RCV comes of value as the patient would have clearly transgressed the 20% rise flag that indicates that you can be 99% certain that there is an event that needs to be investigated further here.
So strongly advocating the use of percentage change flags based on RCV values for serum creatinine are adopted by UK labs as a prerunner of the availability of e-alerts based on guidelines such as KDIGO. Also advocating greater use of RCVs by British laboratories in other contexts.
WA Bartlett November 2012
Please note this is a personal view and does not necessarliy rereflect those of any organisation with which I have connections.