Page 42 - Glucose Monitoring Devices

38 CHAPTER 2 Analytical performance of SMBG systems

of study design, this study evaluated meter analytical performance against the then-
latest widely accepted BGM accuracy standard, which was the ISO 15197:2013
guideline.
To further reduce the risk of bias, the study was triple-blinded such that in-clinic
BGM operators did not know plasma reference measurements, the reference lab did
not know BGM readings, and the data analysis team worked only with BGMs desig-
nated by a code number that blinded the identity of the BGMs whose accuracy was
being assessed.
Every subject had a capillary BG level reading measured by six different BGM
devices (in random order), three before and three after a simultaneous reference
capillary sampling was performed. Each of the three sites tested all 18 FDA-
approved BGM systems by testing a new set of six BGMs across three separate
studies in a round-robin fashion. 1032 of 1035 subjects completed the study, and
each BGM was tested on 115 subjects, on average. The average number of strip
lots per study was 2.1.
Of the 18 meters that were evaluated, only six (33%) met the predetermined
accuracy standard in all three studies; five met it in only two studies; two met it
in only one study. This left four FDA-cleared BGMs that did not meet the ISO
15197:2013 standard in any of the three studies.
In November 2018, the FDA released a statement promising to establish an alter-
native 510(k) approval “Safety and Performance Based Pathway” evaluated on
objective safety and performance criteria rather than on a predicate device [15].
This new approval process could potentially give the FDA more flexibility to
respond swiftly to safety concerns that arise and perform postmarket surveillance
through programs such as the recently announced National Evaluation System for
Health Technology [16].

Advances in analytical performance of SMBG devices
The previously mentioned 2018 review article [14] determined that newer BGMs
were statistically significantly more likely to meet both the ISO 15197 2003
(OR 1.06, CI 1.06e1.07, P <.001, odds ratio reported for a 1-year period) and
ISO 15197 2013 guidelines (OR ¼ 1.06, CI 1.03e1.06, P < .01, odds ratio reported
for a 1-year period). Which advances in technology best correlate temporally with
this improved analytical performance in more recently developed BGMs?
Pfutzner proposed that several technological advances in BGMs contribute to the
improved analytical performance in more recently developed BGMs [17]. First,
device manufacturers have been able to eliminate hematocrit interference by using
mathematical correction algorithms that have been refined over the course of time.
In addition, changes in test strip chemistry, changes in electrode design, and other
algorithmic improvements have contributed to improved accuracy in newer meters.
When analyzing the test strip enzyme in meters over time and their analytical
performance, King and colleagues demonstrated that glucose dehydrogenase