乌鸦传媒

This article is a collaboration between 乌鸦传媒 and:

For patients with acute kidney injury (AKI), dialysis is associated with better survival among those with higher creatinine levels, but higher mortality for those with lower levels, researchers found.

In a single-center study, risk of death with dialysis initiation fell by about 20% with each mg/dL increase in creatinine (P=0.001), , of the University of Pennsylvania in Philadelphia, and colleagues reported in the .

Increased creatinine may be a marker of higher muscle mass and possibly a healthier subset of AKI patients, they said.

"Many clinicians feel that, although acute dialysis may not help a critically ill patient, it is unlikely to cause any harm," Wilson said in a statement. "Through this study, we have been able to show for the first time among an equally matched group of patients that dialysis for AKI may cause more harm than good in the subgroup of people who are frail and have lower muscle mass, and more benefit than harm in more robust patients."

There's been a dearth of evidence as to which AKI patients are likely to benefit most from starting dialysis. Nor are there any randomized trials evaluating early versus late initiation of dialysis, although early initiation has now become standard of care.

To assess which AKI patients make the best candidates for dialysis, Wilson and colleagues looked at data on adults admitted to three acute care hospitals within the University of Pennsylvania Health System over 6 years who developed severe AKI between Jan. 1, 2004 and Aug. 31, 2010.

Overall, 602 were started on dialysis, and 545 of these patients were matched with an equal number of patients who didn't receive dialysis.

In the full cohort, Wilson and colleagues found that initiation of dialysis was strongly associated with mortality (HR 3.42, 95% CI 3.09-3.78, P<0.001) -- but after propensity matching, there was no risk associated with dialysis (HR 1.01, 95% CI 0.85-1.21, P=0.89).

This indicates that patient factors account for the observed harm associated with dialysis in the full cohort, so the researchers looked further into various patient characteristics.

They found that as serum creatinine increased, the risk of death with dialysis initiation fell by 20% for each mg/dL increase in creatinine levels (P=0.001).

Among patients with creatinine levels above 4.2 mg/dL, 57% of those who weren't on dialysis died, compared with 47% of those who did get dialysis (P=0.05).

And among patients with creatinine levels below 2.8 mg/dL, 64% of those who didn't get dialysis died during hospitalization compared with 78% who did go on dialysis (P=0.005).

"Patients with a higher creatinine at initiation of dialysis fare better not only than those who initiate dialysis at a lower creatinine, but also those who do not initiate dialysis at a high creatinine," the researchers wrote. "Conversely, patients who initiate dialysis at a low creatinine do worse than matched patients who do not initiate dialysis at a low creatinine."

The difference is likely due to an increased creatinine generation rate, Wilson and colleagues wrote. It's a proxy for muscle mass, and higher levels may represent a healthier subset of patients with AKI.

Lower creatinine, on the other hand, could be a marker of frailty, with reduced muscle mass and greater weakness.

"This finding is analogous to chemotherapy for cancer," Wilson said. "If you are strong enough to cope with the adverse effects, it's good for you, but if you're weak, it may kill you."

In an accompanying editorial, , of the University of Alberta in Canada, said including nondialyzed patients addresses "much of the criticism leveled at other observational studies of dialysis initiation."

But there are limitations to the degree of matching, Pannu wrote. In particular, the study didn't include urine output and other measures of fluid overload, which could have lead to residual confounding.

Still, Pannu concluded that the study is "an important contribution to the ongoing debate about the initiation of dialysis in AKI. The complexity of this analysis demonstrates the difficulty of identifying who will ultimately benefit from treatment."

Editor's Note: Lead author Perry Wilson is a medical reviewer for 乌鸦传媒.

CORRECTION: This article, which was originally published March 20, 2014 at 5 p.m., has been corrected (March 21, 2014, at 12:14 p.m. ).