Medical Device Daily Washington Editor
GAITHERSBURG, Maryland – FDA has waived a fair number of diagnostics under the Clinical Laboratory Improvements Act (CLIA) of 1988, but very few have gone to panel. Last Friday's meeting of the hematology and pathology devices advisory committee, however, took a look at a white blood cell (WBC) counter made by HemoCue (Malmo, Sweden), and the application went down in flames, receiving a 9-0 vote against approving the diagnostic's use in waived settings.
The sponsor knew for at least a year that its application faced substantial headwinds, although FDA documents indicate that HemoCue had initially applied for the waiver in 2007. The committee met last year – almost a year to the day prior – to discuss whether blood cell counters in general ought to be waived for use in doctors' offices (Medical Device Daily, July 22, 2008). During that July 18 hearing, the panel concluded that cell counters are not ready for waived-setting use, in part because of concerns over the high rate of turnover in physician offices and the effect that turnover might have on the use of such tests.
According to FDA documents, HemoCue obtained a 510(k) for its WBC unit in October 2007 based on the predicate of the XS-1000i made by Sysmex (Kobe, Japan). At that time, FDA deemed the counter a moderate complexity diagnostic, putting it out of reach for most physician practices. FDA's briefing documents state that the HemoCue is less accurate than the Sysmex unit when white blood cell counts fall below 1,000 per cubic millimeter, but the company's application requested waived status for detection of WBC counts as low as 300.
HemoCue had arrayed five witnesses to speak at the open public hearing, but there were two others who spoke in disapproving terms, including Judy Yost, director of the division of clinical lab services at the Centers for Medicare & Medicaid Services (CMS). Yost said that compliance with quality control standards for point-of-care (POC) diagnostics was not particularly good in most waived settings, and in reference to the HemoCue application, remarked that CMS sees "significant potential areas of risk that must be addressed to reduce the likelihood of harm to the patient."
The morning open public hearing began with a broadside against the application. A man who identified himself as medical technologist Ray Osmond railed against the application, arguing at one point that "by voting no today, the panel will send a shot across FDA's bow" regarding the approval process.
"Laboratory tests should be free of error," Osmond said, asserting that waived testing "has been a bonanza for industry and a cash cow for FDA."
"Anyone can push a button, anyone can get a result," Osmond noted, adding that the interpretation of the result is the key issue. "The matter before you is not a medical issue, it's a money issue" that risks patient safety, he asserted.
Andrew Osei-Bimpong, MD, of Hammersmith Hospital (London), followed Osmond, making the opposite case. Osei-Bimpong referred to a study he conducted of the company's WBC counter that was published in the August 28, 2008, edition of the International Journal of Laboratory Hematology in which he and two others compared the HemoCue WBC counter to a reference standard and concluded that with the exceptions of instances in which normoblasts or reticulocytosis might interfere with the analysis, the HemoCue WBC counter meets the CLIA requirement.
The authors noted the presence of a "negligible differences in blood count parameters between venous and capillary blood, provided that the capillary sample is free-flowing and obtained without squeezing." The issue of venous versus capillary blood was not a major feature of the panel's discussions, but did come up on several occasions.
Osei-Bimpong said he and his colleagues saw some inexactness in the diagnostic's assessment of white blood cell counts, but that "all these results were within the acceptable error" of 15%.
"We noticed it is a relatively simple instrument to operate and maintain," Osei-Bimpong said, adding that it's use "requires very little technical knowledge" although clinical knowledge is needed to interpret results. He described the unit as "an ideal point of care test for screening."
Tan Nguyen, MD, presented FDA's views of the application, noting of HemoCue's supporting clinical trial that "there were very few – only five – from pediatric populations," and "none of the fingerstick samples in this study was from patients younger than 15 years." This was a major issue for the application because of the probability that the test will be used extensively to rule out various pathogens in pursuit of a reduction in the use of antibiotics in pediatric populations.
Nguyen charged that the unit "is unable to detect interfering nucleated RBC [nRBCs] from normal WBC," a consideration because the presence of nRBCs, "even a few, is the first hint of a long list of important conditions."
Nguyen said that FDA is concerned about the absence of a backstop test in doctors' offices. WBC reflex testing, he said, is "a well established practice in most labs" and the agency is "concerned that reflex testing is not readily available in waived settings." As for allowable total error (ATE), Nguyen said, "for WBC counts less than 5,000, only 86% of fingersticks fell within the ATE zone."
In summarizing the panel's response to FDA's questions regarding whether the HemoCue is sufficiently simple to qualify for a CLIA waiver, panel chairwoman Dorothy Adcock, MD, medical and laboratory director of Esoterix (Denver) said "the panel generally believes the system is not a simple test. While it may be easy to perform, the absence of additional information that may lead to a differential [count] is a concern in that you might misinterpret the WBC count." However, she continued, if "the parameters of the reporting are tightened ... the potential for misreporting may be significantly reduced."
The panel was unsure as to how to address whether a backup test for WBCs would change their thinking about a waiver, and the final vote on whether to recommend a waiver reflected that uncertainty.