EXTERNAL QUALITY ASSESSMENT


concentrations, with CV of 14% and 30% at 8 kPa; and 10% and >50% at 4 kPa for the haemolysate and aqueous materials respectively, consistent with improved saturation characteristics of the haemolysate material.


A wider bias, exceeding the allowable APS, was observed for the IRMA and NPT7 analysers in 2007, with wider inter-laboratory variation observed for the OPTi, ABL6 and Rapid Point 400 analysers in 2007 and the iSTAT, OPTi, and Prime+ in 2021. The latest pO2


accuracy study


(OX1124) was distributed to over 3,300 instruments in November 2024 and results compared with a concentration matched aqueous sample from BG1024. Figure 3 illustrates the intra-method variation (mean ± 2 SD) shown as an error bar for each analyser. The analytical performance specification is expressed as the Weqas TAE (Total Allowable Error) of


Analyser type


Conventional electrodes Multi-use cartridge Optical sensor


Single-use cartridge % POCT analysers Total


2001 505 14 13 11 7


543


2006 853 239 34 84


29.5 1210


2010 947 541 39


137 43.1 1664


Year


2016 541


1033 28


269 71.1 1871


2020 244


1440 16


338 88


2038 Table 1. Change in blood gas analyser type and growth of POCT analysers from 2001 to 2023.


±26.5% at 2.56 kPa and ±34.1% at 6.40 kPa for the pO2


accuracy and aqueous


samples respectively. In keeping with previous studies, there was a general improvement in method bias to the all method average for most analysers for the bovine haemolysate sample, compared with the aqueous samples. A significant positive bias compared to the all-method average was


observed for the Omni/Omni S/b221, Opti, i-Stat, and i-Stat Alinity method groups for the aqueous sample, whereas the data agreed well with the all-method average for the bovine haemolysate sample. Significant negative bias compared to the all-method average was observed for the ABL90 flex and Prime+ method groups for the aqueous sample, but whereas the ABL90 flex data agreed well with the all-method average for the bovine haemolysate sample, the negative bias remained for the Prime+ method. The overall CV for pO2


was lower for


the bovine haemolysate sample (overall CV 12.9%) compared to the aqueous sample (overall CV 48.9%). The overall CV includes components of both imprecision and method bias, with the large deviation in bias observed in the aqueous sample contributing to the wider overall CV. Although a higher CV was observed for some methods for the bovine haemolysate sample, the concentration of pO2


2023 127


1876 13


427 94.8 2443


in the two samples were not exactly concentration matched, with overall mean 2.56 kPa compared with 6.4 kPa for the aqueous sample. However, apart from the Prime+ method, most methods performed well at this very low pO2


concentration, confirming that pO2


methods are accurate across the clinical range required.


Conclusions Fig 3. a) pO2 results for blood gas distribution BG1024 (aqueous sample); b) pO2 distribution OX1124 (tonometered bovine haemolysate). WWW.PATHOLOGYINPRACTICE.COM SEPTEMBER 2025 results for pO2 accuracy


In conclusion, arterial blood gas analysis is a commonly used technique in a range of emergency and critical clinical situations, as well as in assessment of patients with chronic disease. Improvements in technology have expanded the use of POCT blood gas devices into multiple clinical areas, including intensive care units, emergency departments and ambulances. Assessment of patient oxygen status is an immediate concern for clinical teams, and EQA is a key component of ensuring that devices provide accurate results that are suitable for clinical use. Previous work has shown that aqueous samples used for blood gas analysis in EQA programmes allow participants to be compared with their peer group, but


5


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6