Adult reference intervals for blood gases

The clinical value of any patient test result depends on the quality of the reference interval used for its comparison (interpretation), so that good laboratory medicine practice demands continuous review of reference intervals. There is a paucity of published study aimed at validating the health-associated reference intervals currently used to interpret patient blood gas results. 

This paucity, which can perhaps be best explained by the necessity of sampling arterial blood - a procedure not without risk - from healthy individuals, is addressed by results of a recent study from Aalborg hospital in Denmark. Arterial blood was collected from 182 healthy adults (96 female) aged 20-76 years. These included 118 medical students (aged 20-32) and a further 64 volunteers, principally health professionals (aged 20-76). 

Samples were collected over a 6-year period by standard protocol, and analyzed within 10 minutes of collection. Results of these analyses allowed non-parametric estimation of reference intervals (95 % limits) for: pH (7.38-7.46); pCO₂(a) (4.3-6.0 kPa); Base Excess (–3.0 - +2.8 mmol/L); Bicarbonate (21-27 mmol/L); pO₂(a) (11.1-15.1 kPa) and sO₂(a) (97-99 %). 

Analysis of pO₂(a) results revealed a statistically and clinically significant age-dependent effect, with pO₂(a) reducing at the rate of 0.29 kPa per decade. The above pO₂(a) reference interval relates only to those 131 individuals in the study group aged 20-39 years. The pO₂(a) of the remaining 37 individuals aged 40-76 years ranged from 9.3 to 14.7 kPa, but the small number in this group precluded estimation of a separate reference interval. 

The median pO₂(a) of those aged less than 40 years was 13.6 kPa; this compared with a median pO₂(a) of just 11.8 kPa for those aged 40 and above. Females had slightly higher pH (median 7.418) than males (median 7.406) and slightly lower pCO₂(a) (median 4.9 kPa) than males (median 5.3 kPa), but these differences were judged to be too small to justify gender-specific reference intervals.