Quality Requirements and Standards

It's becoming clear that the 2014 Draft Guidelines on Glucose Meters have raised the ire of nearly everyone in the laboratory marketplace. Here's an overview of the key opinions.

May 2014
STEN WESTGARD, MS

Earlier this year, the FDA made available a new set of Draft Guidelines for Glucose devices, a set of new regulations that would apply to Manufacturers. In the months that followed, criticism has been mounting. Not only have manufacturers complained that the new requirements will be so onerous as to stifle the development of new meters into the market, laboratories have complained loudly that the regulations would ultimately have the effect of hurting patients.

[A little backstory: back in March 16-17 of 2010, the Food and Drug Administration held a public meeting to discuss the quality of Blood Glucose Meters (BMG).  Transcripts of that meeting have become available, as well as the slide presentations.  The consensus from that meeting was that requirements needed to be tightened.]

With the Draft Guidelines published, the FDA asked for public comment, and many of those responses have been posted on regulations.gov. In addition, there have been lively discussion on many POC-related listservs, as well as articles in the industry publications like Clinical Laboratory News. Below follows a sampling of the concerns voiced by different segments of the laboratory marketplace.

A sample of Professional Organizations Responses:

Perhaps the biggest shoe to drop on the response to the Draft Guidelines came from the AACC in May:

"Although we support the agency's objective, reclassifying prescription-use blood glucose monitors will add to the regulatory burden for healthcare providers by increasing their personnel documentation, proficiency testing, and method performance certification requirements under CLIA'88," said James H. Nichols, PhD, medical director of clinical chemistry at Vanderbilt University Medical Center and chair of AACC's Government and Regulatory Affairs Committee. "We do not believe that this will improve patient outcomes. In fact, we are concerned that the opposite may occur if glucose meters are suddenly removed from critical care settings, where they are effectively used to measure and manage glucose levels, thus contributing to better patient outcomes."

AACC suggests improving the clinical protocols followed by the personnel performing blood glucose monitoring rather than increasing the regulatory requirements associated with using these devices. Overall, AACC recommends that FDA set its new performance criteria for prescription blood glucose meters high enough to improve patient care, but not so high that they encumber the use and production of these devices with no benefit to patients.

The American Society of Clinical Laboratory Scientists (ASCLS) responded a little differently:

We agree with the agency that POC devices to be used for prescription point-of-care use must meet different standards than devices used for self-monitoring. Recent incidents of adverse patient outcomes emphasize the need to re-evaluate the limitations of these devices. We support that additional guidance addressing the quality issues with the performance of these devices.

We agree that new devices designated for prescription use should be classified as moderately complex. In theory there is a compelling need for the structureof a total quality management system to ensure that these devices are usedand maintained properly. Upgrading the qualifications of the supervising
personnel and requiring proficiency testing will overtime lead to better patient outcomes. We applaud the guidance that is intended to:

• • Improve the analytical performance
  • Address the issues seen in different patient populations in an acute care setting
  • Protect patient safety by improving design elements that will minimize pathogen transmission and
  • Ensure precision and accuracy

The FDA’s emphasis on minimizing bloodborne pathogen transmission is excellent. We assume that the concern extends to other pathogens that can be transmitted via contaminated surface contact and cause healthcare associated infections (HIA). If that is not the intent, we urge FDA to extend their guidance recommendations to pathogens such as MRSA, Clostridium difficile, etc.

The performance evaluation criteria described in the guidance is very comprehensive and offers excellent advice to manufacturers. We applaud the requirements to expand the sample numbers as well as including patients in intensive care units. The therapeutic substances in this patient population were never anticipated when most of these devices were brought to market and the potential interference has not been well documented or understood by users.

The College of American Pathologists (CAP) also responded:

 First, we believe that the term “critically ill” is too vague and should be avoided. We think it is more appropriate, and more scientifically valid, to focus on the  specific limitations cited by the manufacturers and the literature, as specific situations in which the meters should not be used for capillary samples, including, among others, severe hypotension, dehydration, shock, hyperglycemic-hyperosmolar state. We agree with, and strongly support, the specific exclusion of Blood Glucose Monitoring (BGM) devices, under any and all circumstances, for so-called tight glycemic control. We mention this here because that exclusion is independent of the limitations discussed above.

The second issue is the implication of the FDA draft document on the manufacturers’ current intended use package inserts, in that it could result in the immediate removal of BGM devices from acute care situations. Removing these devices from a variety of hospital settings (including, but not limited to, ICUs, ORs, recovery rooms, and ERs) will compromise some types of care because of the time delays associated with getting results from the central laboratory. We encourage the FDA to work with manufacturers and regulatory agencies to prevent negative patient outcomes.

A sample of Manufacturer Responses

The NOVA Biomedical Corporation, manufacturer of the highly regarded StatStrip, submitted a lengthy response to the FDA. We find notable the company's objections to the specific quality requirements for glucose at the point-of-care:

As a POC blood glucose manufacturer, we are in favor of tightening the performance acceptance standard used to grant 510(k) clearance to a POC blood glucose monitoring test system. That stated, we have reviewed the clinical data currently in published clinical studies and have also reviewed clinical data from ongoing or unpublished investigator initiated clinical studies. Our conclusion is that it will be difficult for current POC blood glucose monitoring test systems to meet the proposed standard.

We are concerned the proposed < 70 mg/dL standard (99% within +/- 7 mg/dL) is too tight. The proposed standard is less than half of the current standard of +/- 15 mg/dL. In addition, the proposed percent of compliance is increased from 95% to 99%. For a typical 350 subject clinical study, the outlier allowance has been reduced from 17 to 3 outliers. The proposed standard is too tight for even the best POC blood glucose monitoring test systems to achieve.

We are also concerned the proposed > 70 mg/dL standard (99% within +/- 10%) is too tight. The proposed standard is half of the current standard of +/- 20. In addition, the proposed percent of compliance has also been increased from 95% to 99%. Again, the proposed standard is too tight for even the best POC blood glucose monitoring test systems to achieve.

An unintended outcome of establishing a new accuracy standard that POC blood glucose manufacturers are unable to achieve is that manufacturers may decide not to invest in future product development because the accuracy standard is set too high. Instead, POC blood glucose manufacturers may decide to do product  updates/refreshes that do not require a new 510(k) submission based upon FDA's published guidance document "Deciding When to Submit a 510(k) for a Change to an Existing Device"....

NOVA further suggests a different quality requirement instead:

Factoring these clinical considerations into a recommended accuracy standard, we believe a quality blood glucose monitoring test system should be able to achieve the following standard within a controlled clinical study

• 95% of all values should be within +/- 12.5% of the reference method for glucose concentrations > 70 mg/dL, and within +/- 10 mg/dL at glucose concentrations < 70 mg/dL.
• 100% of all values should be within +/- 20% of the reference method for glucose concentrations > 70 mg/dL or +/- 15 mg/dL < 70 mg/dL.
• All values outside of +/- 20% of the reference method for samples >70 mg/dL or +/- 15 mg/dL < 70 mg/dL should be thoroughly investigated to determine the reason for the clinical discrepancy.

To standardize device performance within all new submissions, the clinical study should be performed against an IDMS aligned and certified IDMS equivalent (definitive) hexokinase method.

Roche Diagnostics wins the longest-submission award, submitting a response that extends more than 60 pages. They, too, suggest a loosening of the quality requirements suggested by FDA:

Comment 1: Roche recommends that FDA change the method comparison criteria in the Draft Guidance to: 95% of all results must be within +/-10 mg/dL of the reference at glucose concentrations <100 mg/dL and within +/-10% of the reference at glucose concentrations ≥100 mg/dL, based on comparison to a control.

A sample of the Laboratory Responses

From RUSH:

We are specifically concerned about reclassification of the inpatient use of a simple glucose meter as "categorized upon clearance as moderate complexity," a change that will significantly limit the use of the meters. If the new regulation will limit the overall use of meters in the hospital setting, this will lead to chaos and confusion, this will lead to a decrease in the safe use of insulin therapy and thus will lead to increased patient harm.

From North Oaks Medical Center:

• We recommend the regulation clearly place the burden of validating devices for use in all care settings (critical care, pediatrics, medicine, etc.) be placed on the manufacturers. If this is not done, hospitals would have to assume that burden. This is cost prohibitive for providers, and from a methodology standpoint, facilities of our size may not have adequate patient volumes in all care settings to produce statistically significant, valid results.
• If the manufacturers do perform the additional testing necessary for validation, it is likely those costs will be shifted to hospitals. It will be difficult to absorb the increased costs with the budget cuts we already have endured. It should be noted that CMS is evaluating reimbursement for the top 20 laboratory tests (of which this is one) to identify potential costs savings. If that analysis relies on historical cost data, and does not consider the possibility that the cost of these tests will increase because of these new requirements, it will be another financial hit for hospitals.
• Validation of these devices in the critical care arena is essential. If hospitals are not permitted to use these devices for critical care patients, the implications for patient safety issues are tremendous. The ability to quickly adjust insulin medications in our most vulnerable patients increases patient safety and improves outcomes. The other option is to run a serum glucose test, which has a minimum turn-around time in our hospital of 45 minutes, which would significantly delay needed care.

From Cleveland Clinic:

1)We understand the need for improvements in the POC blood glucose testing, and that the draft guidance is intended to apply to manufacturers. While some specific limitations on the use of POC glucose testing should be considered, the unintended effects of implementing this guidance raise serious concerns. The guidance is already having a direct impact on the availability of blood glucose testing in some states, and this is likely, in some circumstances, to impact patient care and outcomes negatively.
2) POC blood glucose testing is an integral part of patient care in almost every clinical setting of the modern healthcare environment. Clinicians are well aware of the limitations of the testing paradigm and, despite those limitations, continue to use POC glucose testing because overall, it results in improved patient care.
3) Clinical manifestations of hypoglycemia are varied and include seizure, altered mental status, and neurologic deficits, making rapid assessment of serum glucose essential in multiple settings; failure to be able to use a POC testing device in the hospital setting may impact patient care due to delays in obtaining results quickly from the laboratory.

The Last Words, for now

David Koch, valued colleage of ours and incoming AACC President, was quoted in Clinical Laboratory News:

“Hospital laboratorians and endocrinologists who are working with surgeons and others in the intensive care units know that this close monitoring is beneficial, both in decreasing length-of-stay and improving outcomes for patients,” Koch said. “We monitor the meter, we monitor the people who are using them, and we’ve been pleased with the performance....

It’s more than meters and test strips. If the meter doesn’t work, it’s often that the sample wasn’t acquired properly, or the patient is in shock and their capillary circulation isn’t adequately in agreement with plasma. For the most part, the performance of the meters themselves is good.”

Elissa Passiment, Executive Vice President, summarized the quandry of the need for better regulations with the burden of compliance with those needed rules:

While we and the FDA do not always agree on their decisions, it should be noted that the issues that precipitated this guidance surround an analytical problem exhibited by some glucometers. The methods are subject to producing inaccurate results due to the interference of IV solutions (containing non-glucose sugars for example) and/or drugs. The lack of knowledge/recognition of these analytical issues by users at the bedside have caused the deaths of patients in intensive care units from insulin overdoses. (the glucometer gives falsely elevated readings). The most recent patient death made all of the news outlets in the Northeast as the physicians and nurses used the glucose meter readings to dose insulin and ignored the laboratory's results which showed the patient was quite hypoglycemic. Thus it is not a simple capillary vs venous sample issue.

To understand the difference between how these meters come to market vs the validity testing done on the methods we use in the laboratory, would require a boring tour through FDA clearance of over the counter devices. And even after that, you are correct - our methods are not always tested against every patient population we encounter. However, the FDA assumes that we as laboratory physicians and scientists understand the limitations of our test methods, document the interfering substances, pick up issues with methods and, because we are compulsive, read labeling, run appropriate quality systems, etc. but have found out that is not consistently the case with folks outside our profession.

Yes they need to define "critically ill" better - or we will need to do that. And for obvious legal reasons, manufacturers of current devices will now have that additional labeling until such time as a newer model that has been validated differently is available. By the way this is a guidance document not a regulation; if the manufacturers follow the suggested guidance our patients will benefit. And in the meantime we are now all on alert.

There is undoubtedly much more to come...

More than 150 comments were submitted to the FDA, most of them critical. There are multiple aspects that are concerning, not just the exact technical requirements proposed for new meters, but also the economic burden of the studies, and above all, the potential danger of removing meters from areas now considered "off-label."

 How the FDA responds and readjusts these guidelines means we have to stay tuned. This is one issue where the response from labs and organizations and industry will make a major impact in the final outcome.