Feature Story

Squeezing out platelet contamination

November 2003
Anne Paxton

First gradually, then suddenly. That was F. Scott Fitzgerald’s characterization of how one goes broke. But in the blood banking world, it’s an apt description of something far more positive: how the first standards for screening platelets for bacterial contamination have become a reality.

Only 12 months ago, a group of blood bankers issued an alarm that while contamination with bacteria is the largest transfusion risk to platelet recipients, there had been nothing but talk about the issue for seven years. But the CAP added a related question to its Laboratory Accreditation Program transfusion medicine checklist in December 2002, and in March 2003 the American Association of Blood Banks approved a standard.

The AABB’s requirement doesn’t become effective until March 1, 2004, and at the moment the CAP’s new checklist question (TRM.44955) is a Phase I question. "That means your accreditation won’t be revoked but you have to have a plan in place. Eventually it will be a Phase II question," explains Richard C. Friedberg, MD, PhD, chairman of the Department of Pathology at Baystate Medical Center, Springfield, Mass., member of the CAPTransfusion Medicine Resource Committee, and chair of the CAP Patient Safety and Performance Measures Committee.

The checklist question asks, "Does the laboratory have a system to detect the presence of bacteria in platelet components?" This requirement can be fulfilled, the checklist says, if the blood establishment that supplies platelet components to the transfusion service tests for the presence of bacteria in the units and notifies the transfusion service of units that are suspected of containing bacteria.

Dr. Friedberg, who says even an imperfect plan is an improvement, sees the new standard as having several pluses. "One, it’s the right thing to do. Two, it’s in the interests of the patient. And three, it may actually lead to a net increase in platelets’ shelf life."

Despite the committee’s consensus, however, the checklist question has not been added without controversy, particularly among clinicians. Some have expressed doubt that the need is all that great, others that available technology is up to the task.

James P. AuBuchon, MD, vice chair of the Transfusion Medicine Resource Committee, says he hasn’t heard of any qualms among College members but has heard from other quarters. "The concerns appear to be coming from hospital transfusion services more than from blood collection agencies. Although blood collection agencies are not used to dealing with bacterial tests, they are accustomed to having to implement new tests to deal with pathogens," says Dr. AuBuchon, professor of pathology and medicine and chair of the Department of Pathology at Dartmouth Hitchcock Medical Center, Lebanon, NH.

There is a significant difference in perception among the stakeholders, observes Gerald A. Hoeltge, MD, head of the section on transfusion medicine at the Cleveland Clinic Foundation and a member of the CAPCommission on Laboratory Accreditation. "The clinical medical community measures the risk according to the rate of transfusion reactions, which is low. The blood banking community, the people who collect, process, and store platelets, and of course the FDA, are looking at it from the standpoint of impurity."

"Nobody wants bacteria in their blood, but the truth is it doesn’t always hurt you," he says. "In fact, most of the time it’s innocuous, because in the hospital blood transfusion is highly correlated with antibiotic treatment. So you may not even see any effects whatsoever, and if you do, the effect may just be fever."

Laurence A. Sherman, MD, JD, emeritus professor of pathology at Northwestern University, Chicago, and a member of the CAPtransfusion medicine committee, finds it natural that the need for screening would fly under clinicians’ radar, so to speak. "Patients who receive platelets tend to be more complex, and it’s often difficult to sort these things out, particularly if the bag has been discarded," he says.

"There is, to a degree, an underreporting of transfusion reactions in general, and in patients who are complex where there are other things going on that would change the temperature, they may not think about transfusion," Dr. Sherman adds.

Ira A. Shulman, MD, chair of the CAPtransfusion medicine committee, says when data are collected in an "organized, prospective, systematic fashion, you’re more likely to detect and report outcomes or reactions than when you wait for people to tell you.

"That’s why institutions that have looked at all the outcomes of transfusions systematically, whether the problem is mislabeled specimens, transfusion mistakes like giving blood to the wrong patient, or viral transmission after transfusion, appear to have a higher incidence of adverse outcomes," says Dr. Shulman, director of laboratories for the Los Angeles County-USC Healthcare Network.

"It’s totally true that some patients can clear bacteria without clinical consequences," he continues. "And many patients are on antibiotics at the time of their blood transfusion, which can serve as a mitigating factor in the event a transfusion contains bacteria. But do you think every patient getting a transfusion is on antibiotics? I don’t think so. Just because some are doesn’t mean everyone will have a protective mechanism to shield them against potential infection."

But the clinicians’ questions are legitimate, Dr. Friedberg says. "It makes perfect sense from the standpoint of small group practices who say, ’Well, we transfuse 500 platelets a year and have had no problems.’"

"Part of the problem is that platelet transfusion is a relatively recent event," he notes. "We’ve had red cells for decades, since back in the ’30s and ’40s, but we only started doing random donor platelets in the ’60s and ’70s, and apheresis in the ’80s. So we don’t have the same history as with red cells. And platelets come in at a time when there’s increased concern about viral transmission."

Because platelets have to be kept warm, they create a highly "bug-ogenic" environment, he adds. "It’s a great medium to grow things in when you keep them warm, and with good aeration you have everything needed for bacteria to grow. And the average front-line clinicians don’t recognize that platelets are not stored like red cells, in the cold."

Regarding the belief that the incidence level is overblown, Dr. Friedberg cautions: "It’s a matter of seeing enough cases to know what the problem is. If you used the same logic, you’d argue there is no chance of transmission of hepatitis or HIV, because we’re at a low enough level that you’ll not see them without looking for it. And bacterial contamination is still an orders-of-magnitude greater risk than everything we’re doing for viral load."

The charge has also been leveled that the number of false-positives with current testing technology will lead to significant waste. Says Dr. Shulman: "There’s no test we do that doesn’t have a risk of false-positivity. In fact, that isn’t sufficient reason not to do a test. One has to consider what is the sensitivity, what is the specificity, and what is the incidence of the condition you are testing for among the population being tested, to determine the negative and positive predictive value of a test."

"If we were going to test 10,000 consecutive platelet products, how many would be truly contaminated? There would be five to 10. If we assume whatever test we use has 99 percent specificity, that means we’d have 100 false-positives. So there would be 10 times the likelihood of a false-positive over a true positive."

"Basically," he says, "the lower the specificity, the more likely a positive result will be a false alarm. But that’s how testing is when dealing with low-incidence situations. If the sensitivity is 99 percent and the specificity is 99 percent, then you’ll detect almost all the contaminated products and you’ll have one in 100 products tossed or not used because of false-positives. These are the kinds of calculations one needs to do when deciding which test to use."

Addressing bacterial contamination of platelets could reopen the possibility of returning platelet products to a seven-day shelf life, the standard in place before the FDA changed it to five days several years ago, Dr. Friedberg says. "If the concern back then was contamination, and not the quality of the platelets that survived, and you could prove there was no bug on day one, or even with quality control be fairly certain of it, I think that would reopen the whole issue of shelf life," he says. "And seven days would get you around a lot of problems. One of the biggest problems in any blood bank is what to do on a four-day weekend."

Every large center has issues on three-day weekends, he notes, but "between New Year’s, July 4th, and other holidays there are at least four times a year when you have a good shot at having a four-day problem, so that by Tuesday you’re on the last platelets. Every place I’ve been it’s been a problem." With platelets in short supply, reverting to a seven-day shelf life would be beneficial, he says.

Thus, whether the bacterial detection standard will increase the costs of platelet products must be considered in light of the potential improvements to supply, Dr. Friedberg says. "If contamination is identified up front and shelf life can be increased, the overall net impact could well be negligible-with significant improvements in patient safety."

Until then, some worry that the new bacterial detection standard could result in platelet shortages. Dr. Shulman says it depends on the test strategy. "For example, if the platelets were tested in the hospital laboratory at around the time of their dispensing, there should be no impact on supply necessarily."

"On the other hand," he says, "if they are tested by a blood collection agency and held in quarantine an extra day before being shipped to the hospital, that could conceivably cause a crimp in supply. Whatever routine they come up with that causes delay in getting the platelets into the field causes a shorter shelf life between receipt and outdating, which could result in extra waste."

That supply issue could be addressed or remedied, of course, if platelets were allowed to be stored for an extra few days. But he says moving to such a standard would have to be based on data.

Culturing the whole-blood-derived product in the donor center is one approach, but this might not be practical for every institution, he adds. "What many blood bankers really want is to use some sort of rapid, reliable, highly accurate, highly reproducible bacteria detection test at around the time the platelets are readied for dispensing for transfusion. If that test is okay, then the platelets go; if it says there’s a question of contamination, the platelets don’t go."

But many transfusion services are grappling with the lack of a clear, highly reliable set of methods, some blood bankers say.

The two automated systems that have been licensed-Pall BDS and BioMérieux BacT/Alert-are designed for apheresis units, while the point-of-care technology is clumsy, says Dr. Hoeltge. "Those transfusion services using pheresis units have an automated technology available to them that hospitals using random platelets don’t." There are big differences around the country on this score, he adds.

The data suggest that each unit of platelet concentrate carries a risk of bacterial contamination of about 1:1,000 to 1:2,000. A pool of five such concentrates would have a risk of 5:1,000 to 5:2,000. On the other hand, the risk that a plateletpheresis unit has bacterial contamination is about 1:1,000 to 1:2,000, which is one fifth the risk of a five-platelet-concentrate pool. "So even though the risk of using pheresis units is very low, it’s still not zero, no matter how you cut it," Dr. Shulman says.

Admittedly, there are logistical problems, and the testing will inhibit timely patient care, he says. "The real cost of random platelets is time. Under the simplest technology available, testing adds 10 to 15 minutes per platelet pool, and if the patient is hemorrhaging, 10 to 15 minutes is expensive. The FDA has not approved any mechanism for pooling these things ahead of time, so they’re stored to be dispensed when there’s a clinical need, and you have to pool them when there’s a clinical need."

Another concern is that current bacterial detection technologies are intended only for leukoreduced products, and increasing numbers of blood centers are moving away from leukoreduction. But Dr. Friedberg doesn’t consider this a serious obstacle.

"It’s really a red herring," he says. "It’s true at this time that the FDA has approved commercial culture systems only for leukoreduced products. It’s easier to process without a lot of white cells, and the systems are approved to perform that way. But it doesn’t mean that eventually the systems can’t be tweaked to accept nonleukoreduced products."

Blood bankers don’t have data yet on how many units will be discarded under the bacterial testing requirement, but Dr. Hoeltge estimates it will be up to two percent of platelet units. "Platelets are in chronically short supply. Some days we call all around the country just trying to get a few units of platelets, and this will reduce the number available."

"We also know some nonautomated techniques for bacterial detection, like using dipstick, identify many more that are not bacterially contaminated compared to those that are."

"What we’re trying to do here," he says, "is meet an AABB standard, and that standard is goal-directed. It doesn’t define the technology for contamination; the methodology is the responsibility of each transfusion service."

It appears that most blood centers using apheresis platelets will culture them and the hospitals receiving them will not be involved in the bacterial detection process. But Dr. AuBuchon says the problems arise with the consideration of what to do with platelet concentrates. "Some blood centers have decided to culture these; others have felt that bacterial detection methods would be better placed in hospitals for these units," he says.

"That places hospital transfusion services in an unusual situation," he adds, "because they’re not accustomed to testing a unit to qualify it for transfusion, but only to match it to the patient for transfusion. The tests that hospital transfusion services would do would require some validation to implement, but beyond that, it’s a mindset issue: ’This is something that blood centers do; we don’t do it in hospitals.’"

Noting that a recent AABB bulletin addresses many of the technical issues hospital transfusion services might encounter, he says a range of options is available. "For example, some hospital transfusion services are considering using urine dipsticks to measure pH or glucose concentrations in platelet concentrates to detect the presence of bacteria."

"There are several papers in the literature documenting the applicability of this approach," Dr. AuBuchon says. "But there are some validation steps that a transfusion service would have to undertake." For example, the initial glucose level is different in different anticoagulation systems, he points out.

"So the rate of fall of glucose may well change with different kinds of plastic bags. Therefore, a hospital transfusion service or blood center would have to determine what the expected range of glucose would be for platelet units in a particular system at each day of storage. This is not an onerous task, but it will require a little work to set up a detection system that’s useful."

At the Cleveland Clinic, Dr. Hoeltge says, there is no plan on the drawing board yet. "I’m hoping something will become available. Right now we’re waiting for the technology to get up to speed, or for the FDA to license pooled platelets; that would solve the problem in a great way. But I have three grandchildren, and I’ll probably be a great-grandfather before that happens."

"Nobody has confidence that this is going to be straightforward," Dr. Hoeltge adds. "I suppose there may be places where they’re using 100 percent pheresis and the blood supply will then be bacterially tested. But for transfusion services that are relying on their own resources, they’re the ones that have the problem because they don’t have technology that’s sensitive enough. They’re not exactly sure how they’re going to meet the standard."

Dr. Sherman agrees there is a multiplicity of technologies out there, none of which is 100 percent accurate. "But what the committee has felt is we should, even if the methods are imperfect, at least take a step. More standardized things will be coming along shortly, and several measures like looking at swirling or pH will detect a fair number of contaminated units."

In particular for CAP-accredited facilities, Drs. AuBuchon and Sherman recommend checking for swirling. "It is something that can be done today that costs nothing and is easy, and it meets the Phase I requirement," Dr. AuBuchon says. "It requires no reagents, takes just a second, and, frankly, is fun. It’s not the most sensitive technique, but it’s not substantially different from other techniques."

Several companies have rapid bacterial detection systems under development, and it won’t be long before they start selling them. "These techniques will probably offer better sensitivity and greater rapidity and greater simplicity than what hospital transfusion services have today," Dr. Sherman says. Some are expected to be on the market within a few months.

Anne Paxton is a writer in Seattle. Shortly before CAPTODAY press time, the FDA cleared platelets collected with Gambro apheresis technology for seven-day storage, provided the products undergo bacterial detection testing by an approved method before release.