In Which Cardiac Surgeons Practice Preventative Medicine: the PrevAKI Trial

Good morning all, Bill here. Today’s piece comes at you courtesy of budding nephrologist and my good friend John Nawn. John is a PGY-2 IM resident at Penn. Not only is his name really fun to say in a fake Boston accent (Jahhhn Nahhhhwwwn), but he also has one of the cutest babies I’ve ever seen. Oh, and is a competent physician, I guess. So, enjoy:

The Beans Are Always the First Organs to Go: Preventing Acute Kidney Injury in Post-Op Cardiac Surgery Patients

Acute kidney injury (AKI) is bad. (“Insightful!” -Bill) Defined as a sudden decrement in kidney function and manifesting with falling urine output, rising creatinine, or a need for renal replacement therapy, AKI occurs in approximately 20% of all hospitalized patients worldwide and is associated with a mortality approaching 50% if dialysis is required. Before anyone sensationalizes these statistics (“you will get AKI and die,” that one online quack said!), AKI likely serves as a bellwether for overall acuity of a patient’s illness — sicker patients have a risk of both developing AKI and dying. It’s thus difficult to determine how much of a patient’s risk of dying depends on their renal dysfunction. Still, it’s not a stretch to suggest that there are thousands of deaths yearly that might be prevented if we could reduce the incidence or progression of AKI.

When the Kidney Disease: Improving Global Outcomes group (KDIGO) published their latest guidelines in 2012, they recommended establishing preventative measures for patients deemed at higher risk for AKI. Two problems arise when attempting to apply these guidelines, however. First, preventative measures and treatments for AKI differ depending on clinical context. AKI can develop rapidly or slowly and has a heterogenous course depending on the initial insult (e.g. sepsis or surgery), so it’s difficult to identify a homogeneous population who need similar interventions prior to the development of AKI. Second, serum creatinine and urine output are essential to diagnosing and staging AKI, but no renal-specific biomarkers existed in 2012 to predict possible development of AKI. The KDIGO authors acknowledge such, recommending that “[t]he role of biomarkers other than serum creatinine in the early diagnosis, differential diagnosis, and prognosis of AKI patients should be explored.” As a result, efforts to prevent AKI have remained largely based on estimation and supposition rather than evidence. However, we finally have a trial courtesy of Melanie Meersch and colleagues that seeks to standardize an approach to AKI prevention.

If You’re Only Going to Read One Paragraph

In a single-center randomized control trial of patients undergoing cardiac surgery, the incidence of AKI was reduced significantly by use of a postoperative checklist-style “bundle” of renal protective measures. More astounding than the fact that someone cared about reducing AKI is the remarkable absolute risk reduction (ARR) of 16.6%, which rounds out to a nice number-needed-to-treat (NNT) of 6 However, there was no difference in short-term mortality (out to 90 days) or need for renal replacement therapy. All of the interventions in the bundle make intuitive sense, but the extremely selective population (cardiac surgery patients at high risk of AKI) limits the broad application of this bundle. It’s too early to tell if the interventions in this surgical ICU population can be applied to medical populations, or whether they might result in a mortality benefit.

But Get Back to the Headline: Cardiac Surgeons Did What?

Tired of guessing how to best save the kidneys, a group of cardiac surgeons decided to develop a bundle of interventions aimed at preventing AKI. In this German single-center randomized control trial — cleverly named PrevAKI — approximately 300 patients undergoing cardiac surgery at high risk for AKI based on a biomarker assay (more on that in just a moment) were randomized to either a postoperative “KDIGO bundle” (intervention, n=138) or standard postoperative care (control, n=138). Why cardiac surgery patients? Patients undergoing cardiac surgery provide a relatively homogenous population with a reasonably well-understood underlying risk factor (cardiopulmonary bypass) and incidence of AKI approaching 20-30%. One can readily identify the insult and closely monitor patients’ clinical course and laboratory parameters following that insult. These two groups were stratified with respect to heart failure, chronic kidney disease, and other major comorbid conditions, and there were no significant differences in these risk factors following randomization. Additionally, each arm had similar rates of specific cardiac surgeries, including CABG and valvular surgery. There was also equal intraoperative vasopressor use and volumes of fluid resuscitation, so patients left surgery with relatively similar exposures.

The postoperative “KDIGO bundle” included cessation of ACE inhibitors or ARBs for 48 hours postoperatively, maintenance of normotension (MAP > 65 mmHg), avoidance of volume overload (goal CVP between 8 to 10 mmHg), consideration of non-contrasted cross-sectional imaging, avoidance of nephrotoxins, and avoidance of prolonged hyperglycemia. These interventions all seem like common sense, but the results reveal how hard it is to apply even these modest recommendations. Finally, to better enrich the population so as to include those at highest risk of AKI (and a higher incidence means a smaller sample size for the same power calculation!), the investigators screened all patients four hours postoperatively with the use of Nephrocheck.

Wait, Nephrocheck? A Brief Excursus

(“Excursus (n)


  1. A detailed discussion of a particular point in a book, usually in an appendix.


I know, I didn’t believe it was a word either.” -Bill)

It’s safe to say that Nephrocheck is absolutely the new hotness about which you’ve never heard. This novel proprietary biomarker test relies on the presence of two markers of G1 phase cell cycle arrest, tissue inhibitor of metalloproteinases-2 (TIMP-2) and insulin-like growth factor binding protein 7 (IGFBP7). The presence of these biomarkers in the urine suggests tubular ischemia with associated cell cycle arrest, and thus predicts progression to acute kidney injury if the insult is sustained. In multiple validation studies, Nephrocheck has performed as well as clinical adjudication and has outperformed KDIGO scoring in diagnosing and predicting AKI in multiple populations. This extends even to patients undergoing cardiac surgery, where patients with a positive Nephrocheck test had an astounding 80% incidence of AKI. It’s worth noting that the lead authors in this particular validation study were the same as our current PrevAKI: take what you will from that (ZOMG so much conflict LOL credibility!). However, if one accepts that patients undergoing cardiac surgery represent a homogenous population with a risk of AKI at baseline, further stratifying some of these patients as high-risk with the use of early biomarkers like Nephrocheck would allow selective application of preventative measures and possible avoidance of AKI.

Fine, You Industry Shill. What Did They Observe?

The primary outcome, incidence of AKI at 72 hours postoperatively, was reported as 55.1% in the intervention group and 71.7% in the control group (p = 0.004). This corresponds to an absolute risk reduction (ARR) of 16.6%, or a number needed to treat of approximately 6. While this rate of AKI seems astronomical even compared to the incidence of 30% in all cardiac surgery patients, remember that this is a cohort of patients at higher risk as determined by Nephrocheck, which population has a reported incidence of approximately 80% in prior validation studies. Moveover, there was a statistically-significant increase of higher stage AKI (stage II or stage III) in the control arm compared to the bundled intervention.

M Meersch et al. Intensive Care Med (2017). DOI:10.1007/s00134-016-4670-3.

Reviewing the general post-operative courses in each arm, several things become evident. First, the diagnosis of AKI was made more often by falling urine output (n=143) than rising creatinine (n=24), even as patients had similar rates of diuretic use. Falling urine output can be monitored in real time and better reflects the kidneys’ “true” GFR, while serum creatinine is collected at intervals and lags behind GFR changes: thus, the diagnosis of AKI was made by the information immediately available, i.e. urine output. Second, patients in the intervention arm received more inotropic support to achieve their MAP goal. However, both arms had largely similar CVPs through the trial of approximately 9-10 mmHg. Finally, there was a non-significant reduction in contrast use, vancomycin/gentamicin use, and unspecified other nephrotoxins in the intervention arm. All of these changes are consistent with the bundle itself, which encouraged avoidance of nephrotoxins with an emphasis on renal perfusion.

Regarding secondary outcomes, there were no significant differences. Investigators noted no difference in 30-day, 60-day, or 90-day mortality, no difference in the need for renal replacement therapy, not even a difference in ICU or hospital length of stay. Before anyone points and laughs, these are all secondary outcomes. The trial proved its primary hypothesis, namely that AKI can be prevented with a straightforward bundle of interventions. Whether that prevention translates into a significant reduction in mortality merits further inquiry in a larger population, as the anticipated absolute risk reduction for mortality is probably around 1-5%. Finally, the insult in cardiac surgery is a single episode of aortic cross-clamping and cardiopulmonary bypass — in medical patients, insults tend to accrue at irregular intervals during the course of their acute illness. Developing a bundle for medical patients, while entailing similar interventions, would involve a different time course for their applications.

So What’s the Takeaway, Nerd?

Overall, this trial proves it may be possible, without the use of prayer or guessing, to prevent some cases of AKI in a more rigorous way. This is an exciting first step: if this can be extended to medical patients, and if subsequent trials can demonstrate a mortality benefit, AKI bundles may have an analogous impact on standardizing kidney care that intubation and extubation bundles have on our ventilated patients. Even constraining ourselves to the data at hand, preventing 1 in every 6 cases of “high-risk” AKI with modest interventions is remarkable.

Bottom line: don’t forget about urine output when considering a diagnosis of AKI, renal biomarkers like Nephrocheck are closer to breaking into the mainstream, and we’ll have to wait for a well-designed trial in a medical critical care or ward setting to see if bundles can continue to work for AKI, especially in proving a mortality benefit. For now, I’m encouraged — at least until my next attending requests a CT scan with contrast in our patient in septic shock on amikacin.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s