Serum Ferritin: Different Assays Mean Different Results, Study Reports

Hogan, Michelle

doi: 10.1097/01.NEP.0000346561.31172.fa

Serum ferritin, a surrogate marker of iron stores known to have limited predictive power, varies substantially depending on the assay used, as well as fluctuates within individual patients, rendering single values insufficient for making treatment decisions in patients with kidney disease. This was the conclusion of a study published in Kidney International (2009;75: 104–110).

“There is tremendous variation between assays,” said study author Daniel W. Coyne, MD, Professor of Medicine at Washington University in St. Louis, in a phone interview. Bradley A. Ford, MD, PhD, and Mitchell G. Scott, PhD, both of the Department of Pathology and Immunology at Washington University, were the lead and senior authors of the study, respectively.

“Our study found that when our assay, which was the Siemens Centaur assay, noted a ferritin of 500, other assays showed ferritin values as low as 439 and as high as 632,” Dr. Coyne added.

“This difference becomes even more dramatic when you look at ferritins of 800, when one assay showed a value of 700 and another showed a value of 1,040, so that if you're trying to target patients to a ferritin of 200 to 500 but you switch assays, you may have tremendously different results.”

Back to Top | Article Outline

Substantial Variation

While the limitations of ferritin as an indicator of iron stores are well recognized, these between-assay differences had not before been demonstrated by a formal study, Dr. Coyne said.

“A couple small studies in dialysis patients have examined the question of the variability of ferritin within patients, but not between different assays, and that was one reason we wanted to do this study.

“Secondly, lab experts in ferritin assays have long known that there's tremendous variation between the results that one assay provides versus another, and so we wanted to document whether that difference also existed when we looked at specimens from dialysis patients and what would be the clinical impact of that variation.”

That difference was documented in a very elegant way, said Steven Fishbane, MD, Chief of Nephrology at Winthrop-University Hospital, when asked to comment on the study in a phone interview.

“It was really a very, very carefully planned out study, and the conduct and the procedures that were used were really very highly rigorous.… It makes it, I think, a study that we can rely on, and it's got great credibility because of that.”

In the study, 13 pools of serum from patients on hemodialysis as well as those not receiving hemodialysis were evaluated by the six most common ferritin assays—Abbott Architect, Beckman Access, Ortho Vitros Eci, Roche Elecsys 2010, Siemens Centaur, and Siemens Dimension RxL, which are used by about 70% of clinical laboratories in the United States.

When compared with the most common method, the Siemens Centaur, all the assays except the Siemens Dimension came up with significantly different ferritin values.

Back to Top | Article Outline

Day-to-Day Variation

Within-patient variability in ferritin levels also was analyzed in 60 stable patients on hemodialysis and was shown to range from 2.1% to 62.0% over an initial two-week period, with a median intraindividual coefficient of variation (CV1) of 9.1%, and between 2.8% and 51.6% over six weeks, with a median intraindividual coefficient of variation of 10.6%.

There was no significant correlation between weekly erythropoietin dose and six-week CV1, or between intravenous (IV) iron use (versus nonuse) and two- or six-week mean CV1s. In the 51 patients receiving IV iron, weekly iron dose had no relationship with six-week serum ferritin CV1.

“If you look at an individual patient, because there's so much day-to-day variability in ferritin, values need to change by at least 32% for the physician to be 95% confident that it's a real change in ferritin and not just random variation,” Dr. Coyne said. “That's a huge change.”

For a ferritin of 500, that means the level must go down to 340 or less, or up to 660 or more. “That has implications for writing guidelines and also for managing patients,” Dr. Coyne said.

When it comes to making treatment decisions, a single ferritin level often is not enough, he added.

“We need to look at general ferritin trends and ideally have several values that we're looking at. If the overall trend is that the ferritin is increasing, then we're probably administering too much iron over a long period of time, and they need less maintenance iron. If it tends to be repeatedly falling, then they likely need greater iron supplementation, but a single ferritin is not very helpful unless it's low.

“At least in dialysis patients a low ferritin is going to be low on all the assays, but higher ferritins can vary a great deal and don't have very good diagnostic ability for predicting who needs iron. A low value should likely be treated with iron; a higher value doesn't really help you very much.”

Back to Top | Article Outline

Limitations of a Ferritin Cutoff

In terms of current clinical practice in iron management, there's considerable variability there, too, Dr. Coyne said.

The National Kidney Foundation Kidney Disease Outcomes Quality Initiative (NKF KDOQI) Clinical Practice Guidelines and Clinical Practice Recommendations for Anemia in Chronic Kidney Disease recommend that serum ferritin levels be kept at 200 ng/mL or higher in patients receiving erythropoiesis-stimulating agents (ESAs).

For those who have a ferritin higher than 500 ng/mL, there is not enough evidence to recommend routine iron supplementation, and the decision to use or forgo iron should be based on ESA responsiveness, hemoglobin and transferrin saturation level, and the patient's clinical status, the guidelines add.

“Some physicians are trying not to give iron when ferritin is above 500; they're trying to target ferritin to 200 to 500. Our data indicates that's virtually impossible—the variability in ferritin is almost as wide as that target range.

“Secondly it's misinterpreting the guideline, which simply said that when ferritin is 500 or higher you should consider whether continuing iron is indicated. In many cases it is. It's not saying that iron is not needed when ferritin is above 500.

“I think future guidelines should probably recommend a general ferritin below which most patients are iron deficient but not have an upper limit because ferritin is so poorly related to the need for iron. From a safety standpoint not administering iron to patients with transferrin saturations above 50% is a much safer way to avoid iron overload than any ferritin cutoff.”

Erythropoietin (EPO) dose is another important piece, Dr. Coyne agreed.

“I think the first thing to look at is their EPO dose. If the EPO dose is relatively low, less than 15,000 units per week, then the patient likely has adequate iron, and they should either receive small doses or no iron going forward. As the EPO dose is higher then it's more likely that they're iron deficient.”

Back to Top | Article Outline

Surprising Degree of Variation

While the presence of variation in serum ferritin levels was expected, the degree was not, Dr. Fishbane said.

“I was surprised. I thought that the amount of variation was more than I might have expected from just my knowledge of the way ferritin is tested.”

This sizable variability underscores the need for clinical decision making in iron management.

“We tend to assume that laboratory results that we get are very set in stone and that they indicate a level for blood tests that is something that we can really rely on, but what Coyne's group has shown is that for serum ferritin like for some other tests in medicine that the results really have a lot of variability and doctors don't appreciate that fact, and that's why this study is particularly important,” Dr. Fishbane said.

In the short-term, there needs to be more awareness among nephrologists of this variability, he added.

“In the longer term I would hope that it wouldn't be the doctors it would be the laboratories that would come up with better ways to reduce the amount of variability that's present in testing so that doctors can use the test as being more reliable.”

Research into alternative assays for iron stores are also needed, Dr. Coyne said.

“There is some thought that assays for hepcidin, which tend to correlate with ferritin, may be superior in predicting iron needs, and further research in that area is certainly needed.”


If you are a dialysis or transplant professional and are not currently receiving Nephrology Times, or if you need service on your free subscription, please e-mail or call 800-430-5450.

© 2009 Lippincott Williams & Wilkins, Inc.