Scholarlink

Background <p>Point-of-care testing (POCT) using blood gas analyzers provides rapid results and can reduce diagnostic delays. Discrepancies between POCT and central laboratory (CL) measurements have been observed, particularly in patients with sodium disorders. This study evaluated the agreement between POCT and CL measurements for sodium and potassium in patients with hyponatremia, normonatremia and hypernatremia.</p> Methods <p>In this retrospective single-center analysis, all paired measurements of sodium and potassium performed by POCT (Radiometer ABL 90 Flex Plus) and the CL (Roche Cobas Pro ISE) within ± 2 h between January and December 2024 were included. The correlation, Bland-Altman analysis, paired tests and intraclass correlation coefficients (ICC) were used to assess agreement. Regression analysis evaluated proportional bias and ANOVA tested differences among dysnatremic and normonatremic groups.</p> Results <p>A&#xa0;total of 6404 paired sodium and 5622 paired potassium measurements were analyzed. Sodium was 140.3 mmol/L (SD&#xa0;4.5) (POCT) and 137.8 mmol/L (SD&#xa0;6.6) (CL), showing a&#xa0;mean difference of 2.4 mmol/L (SD&#xa0;1.5). The correlation was significant (r = 0.583, <i>p</i> &lt; 0.001) but ICC indicated poor agreement (0.473). The ANOVA demonstrated statistically significant differences between sodium groups (<i>p</i> = 0.007, effect size 0.002), indicating more pronounced deviations in hypernatremic patients. For potassium, the mean difference was 0.12 mmol/L (interquartile range, IQR 0.05–0.22 mmol/L) with good agreement (ICC = 0.85) but a&#xa0;significant proportional bias (<i>p</i> &lt; 0.001). Clinically relevant deviations (&gt; 4 mmol/L sodium or &gt; 0.5 mmol/L potassium) occurred in 8.5% and 4.5% of cases, respectively.</p> Conclusion <p>While potassium showed a&#xa0;very strong correlation and good agreement, sodium demonstrated only moderate correlation and poor agreement according to ICC criteria. Systematic and clinically relevant deviations were frequent, particularly in dysnatremic and dyskalemic patients.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comparison of sodium and potassium measurements in the emergency department by point-of-care vs. central laboratory testing

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月13日

查看全文
  • Svenja Ravioli,
  • Annemarie Edenhofner,
  • Christoph Schwarz,
  • Susanne Oswald,
  • Gregor Lindner

摘要

Background

Point-of-care testing (POCT) using blood gas analyzers provides rapid results and can reduce diagnostic delays. Discrepancies between POCT and central laboratory (CL) measurements have been observed, particularly in patients with sodium disorders. This study evaluated the agreement between POCT and CL measurements for sodium and potassium in patients with hyponatremia, normonatremia and hypernatremia.

Methods

In this retrospective single-center analysis, all paired measurements of sodium and potassium performed by POCT (Radiometer ABL 90 Flex Plus) and the CL (Roche Cobas Pro ISE) within ± 2 h between January and December 2024 were included. The correlation, Bland-Altman analysis, paired tests and intraclass correlation coefficients (ICC) were used to assess agreement. Regression analysis evaluated proportional bias and ANOVA tested differences among dysnatremic and normonatremic groups.

Results

A total of 6404 paired sodium and 5622 paired potassium measurements were analyzed. Sodium was 140.3 mmol/L (SD 4.5) (POCT) and 137.8 mmol/L (SD 6.6) (CL), showing a mean difference of 2.4 mmol/L (SD 1.5). The correlation was significant (r = 0.583, p < 0.001) but ICC indicated poor agreement (0.473). The ANOVA demonstrated statistically significant differences between sodium groups (p = 0.007, effect size 0.002), indicating more pronounced deviations in hypernatremic patients. For potassium, the mean difference was 0.12 mmol/L (interquartile range, IQR 0.05–0.22 mmol/L) with good agreement (ICC = 0.85) but a significant proportional bias (p < 0.001). Clinically relevant deviations (> 4 mmol/L sodium or > 0.5 mmol/L potassium) occurred in 8.5% and 4.5% of cases, respectively.

Conclusion

While potassium showed a very strong correlation and good agreement, sodium demonstrated only moderate correlation and poor agreement according to ICC criteria. Systematic and clinically relevant deviations were frequent, particularly in dysnatremic and dyskalemic patients.