Is venous blood a more reliable description of acid-base state following simulated hypo- and hyperventilation?

Comparison of mathematically arterialised venous blood gas sampling with arterial, capillary, and venous sampling in adult patients with hypercapnic respiratory failure: a single-centre longitudinal cohort study

BACKGROUND

Accurate arterial blood gas (ABG) analysis is essential in the management of patients with hypercapnic respiratory failure, but repeated sampling requires technical expertise and is painful. Missed sampling is common and has a negative impact on patient care. A newer venous to arterial conversion method (v-TAC, Roche) uses mathematical models of acid-base chemistry, a venous blood gas sample and peripheral blood oxygen saturation to calculate arterial acid-base status. It has the potential to replace routine ABG sampling for selected patient cohorts. The aim of this study was to compare v-TAC with ABG, capillary and venous sampling in a patient cohort referred to start non-invasive ventilation (NIV).

METHODS

Recruited patients underwent near simultaneous ABG, capillary blood gas (CBG) and venous blood gas (VBG) sampling at day 0, and up to two further occasions (day 1 NIV and discharge). The primary outcome was the reliability of v-TAC sampling compared with ABG, via Bland-Altman analysis, to identify respiratory failure (via PaCO₂) and to detect changes in PaCO₂ in response to NIV. Secondary outcomes included agreements with pH, sampling success rates and pain.

RESULTS

The agreement between ABG and v-TAC/venous PaCO₂ was assessed for 119 matched sampling episodes and 105 between ABG and CBG. Close agreement was shown for v-TAC (mean difference (SD) 0.01 (0.5) kPa), but not for CBG (-0.75 (0.69) kPa) or VBG (+1.00 (0.90) kPa). Longitudinal data for 32 patients started on NIV showed the closest agreement for ABG and v-TAC (R²=0.61). v-TAC sampling had the highest first-time success rate (88%) and was less painful than arterial (p<0.0001).

CONCLUSION

Mathematical arterialisation of venous samples was easier to obtain and less painful than ABG sampling. Results showed close agreement for PaCO2 and pH and tracked well longitudinally such that the v-TAC method could replace routine ABG testing to recognise and monitor patients with hypercapnic respiratory failure.

TRIAL REGISTRATION NUMBER

NCT04072848; www.

CLINICALTRIALS

gov.

Mathematically arterialised venous blood is a stable representation of patient acid-base status at steady state following acute transient changes in ventilation

Hyper- or hypoventilation are commonly occurring stress responses to arterial puncture around the time of blood sampling and have been shown to rapidly alter arterial blood acid–base parameters. This study aimed to evaluate a physiology-based mathematical method to transform peripheral venous blood acid–base values into mathematically arterialised equivalents following acute, transient changes in ventilation. Data from thirty patients scheduled for elective surgery were analysed using the physiology-based method. These data described ventilator changes simulating ‘hyper-’ or ‘hypoventilation’ at arterial puncture and included acid–base status from simultaneously drawn blood samples from arterial and peripheral venous catheters at baseline and following ventilatory change. Venous blood was used to calculate mathematically arterialised equivalents using the physiology-based method; baseline values were analysed using Bland–Altman plots. When compared to baseline, measured arterial and calculated arterialised values at each time point within limits of pH: ± 0.03 and PCO₂: ± 0.5 kPa, were considered ‘not different from baseline’. Percentage of values considered not different from baseline were calculated at each sampling timepoint following hyper- and hypoventilation. For the physiological method, bias and limits of agreement for pH and PCO₂ were -0.001 (-0.022 to 0.020) and -0.02 (-0.37 to 0.33) kPa at baseline, respectively. 60 s following a change in ventilation, 100% of the mathematically arterialised values of pH and PCO₂ were not different from baseline, compared to less than 40% of the measured arterial values at the same timepoint. In clinical situations where transient breath-holding or hyperventilation may compromise the accuracy of arterial blood samples, arterialised venous blood is a stable representative of steady state arterial blood.

Trending peripheral venous PCO2 in patients with respiratory failure using mathematically arterialised venous blood gas samples

Background

Trending venous blood gases (VBGs) has been suggested as an alternative to arterial blood gases (ABGs) in patients with respiratory failure, but there are limits to its utility. The aim of this study was to compare the trending of venous carbon dioxide partial pressure (pCO₂) (pCO₂v) with mathematically arterialised pCO₂ (pCO₂ca) and to further evaluate whether pCO₂ca follows change in arterial pCO₂ (pCO₂a) more accurately.

Methods

We analysed two data sets. The first was a retrospective study of patients with respiratory failure admitted to the intensive care unit. Venous samples were mathematically arterialised using the vTAC method. The change in pCO₂ between two consecutive samples (ΔpCO₂) for pCO₂v was compared with the change in calculated pCO₂ca values. In the second data set taken from previously published work, we analysed 82 trend points (difference between consecutive samples) for change in pCO₂. There were pCO₂v, pCO₂a and pCO₂ca values for each trend point. The primary outcome measures were the 95% limits of agreement (LOAs) between different sampling methods for ΔpCO₂.

Results

In the first data set, 46 patients had 203 VBG results giving 157 trend points for ΔpCO₂ analysis. The 95% LOAs for ΔpCO₂ca and ΔpCO₂v were −9.28 to 11.12 mm Hg.

In the second data set, 95% LOAs for ΔpCO₂ were −9.46 to 9.48 mm Hg for ΔpCO₂a and ΔpCO₂v, −8.94 to 8.58 mm Hg for ΔpCO₂ca and ΔpCO₂v, and −4.54 to 4.91 mm Hg for ΔpCO₂a and ΔpCO₂ca.

Conclusion

This study suggests that trending pCO₂v is not an accurate way to trend pCO₂a in patients with respiratory failure. ΔpCO₂ca via vTAC trended differently to ΔpCO₂v. Our data suggest pCO₂ca more accurately trends pCO₂a.

Comparison of two methods for converting central venous values of acid-base status to arterial values in critically ill patients

BACKGROUND

Assessment of the critically ill patient requires arterial acid-base status. Venous blood could provide a surrogate, with methods transforming venous values to arterial, improving their utility. This manuscript compares two of these methods, a statistical and a physiological method. Where these methods are inadequate to describe critically ill patients, physiological mechanisms are explored to explain discrepancies.

METHODS

1109 paired arterial and central-venous blood samples, from patients diagnosed with acute circulatory failure, were available for retrospective analysis. Of these, 386 samples were used previously to validate the statistical model. The statistical method of Boulain et al. 2016 and the physiological method of Rees et al. 2006 were applied to the 386 sample pairs, and compared using Bland-Altman analysis. A subset of the 1109 samples, where the physiological method could not accurately calculate arterial values, were analysed further to assess the necessary addition of CO₂ or strong acid at the tissues to account for arterio-venous differences.

RESULTS

Bias (LoA) for comparison of calculated and measured arterial values (n = 386) were similar for the statistical method (pH: -0.003 (-0.051 to 0.045), PCO₂: -0.02 (-1.33 to 1.29 kPa)) and physiological method (pH: 0.009 (-0.033 to 0.052), PCO₂: -0.08 (-1.20 to 1.03 kPa)). In the 381 cases (of the 1109 sample pairs) defined as not accurately described, addition of a median CO₂ concentration of 0.72 mmol/l in excess of aerobic metabolism, explained this for 333 cases, with the remainder requiring simultaneous strong acid transport.

CONCLUSION

Both methods appear equal in their ability to transform central-venous values to arterial, albeit warranting caution when using either in a critically ill population. The physiological approach was able to describe arterio-venous differences not explained by aerobic metabolism alone.