Comparing the Liquid Heparin Syringe with Dry Bound Heparin Syringe for Blood Gas Analysis

Introduction

Blood gas (BG) analysis is an essential part of the diagnosis and management of alteration in respiratory and metabolic parameters of the body. It has been an important tool for emergency physicians, anesthesiologists, and intensivists to understand and manage the critically ill patients. Faster and accurate results are important to form rule-out and rule-in strategies in these sick patients. The usefulness of this diagnostic tool is dependent on the accuracy of sampling, volume of syringe used, amount of sample, type and amount of heparin used, storage, transport, and timing of analysis to correctly interpret the results. However, many preanalytical errors are possible while collecting blood samples, leading to erroneous BG values.¹ This may have an influence on the further management of extremely vulnerable patients.

Although the guidelines by the National Committee for Clinical & Laboratory Standards Institute (CLSI), Wayne, Pennsylvania, United States, suggested the use of commercially prepared dry bound heparin (DBH) syringes (which avoids dilution due to the use of dried heparin), still self-prepared liquid heparin (LH) syringes are used in many emergency and intensive care units (ICUs) for BG analysis to curtail heavy economic burden in high-volume hospitals with lack of resource.², ³ LH may have many effects on BG analysis due to its dilutional effect on plasma, thereby altering some parameters such as partial pressure of arterial carbon dioxide (pCO₂) and electrolytes. By binding to positive ions such as Ca²⁺, Na⁺, and K , LH may result in falsely low values.⁴ In addition, the heparin sodium or heparin calcium salts may have composition effects on BG analysis, leading to positive bias.⁴ There are no universally accepted guidelines or recommendations for the manually prepared heparinized syringes for sampling, and this remains a nonstandardized method among many ICUs.⁵ Hence, our study aimed to compare LH and DBH syringes in arterial BG analysis of postoperative adult cardiac surgical patients in our ICU.

Material and Methods

This prospective observational study was conducted in a cohort of adult patients older than 18 years admitted to the cardiac surgical ICU of a high-volume tertiary care teaching hospital. The study was conducted during a 3-month period from April 2019 to June 2019 after obtaining ethics approval from the institute. Paired samples were collected from 100 consecutive adults who underwent open-heart surgeries electively and admitted to the ICU for postoperative management. Informed consent was invariably obtained from all the patients for ICU management, including blood sampling for different laboratory tests including arterial BG analysis. One paired sample was taken from each patient at a point of time on the second postoperative day, when the patients have become fully clinically, hemodynamically, and metabolically (from the last BG analysis) stable, and ready to be discharged from ICU. Samples were taken from the existing invasive arterial lines (after discarding double the standard dead space volume) placed during surgery, and patients were not punctured separately for drawing arterial blood. BG analysis was performed at the same time soon after drawing and mixing the blood with heparin in the sampling syringes (within 60 seconds), without any waiting period or transport, as BG machines are installed in our ICU.

Results

A total of 100 consecutive postoperative cardiac surgical patients were included in our study, and the population comprised middle-aged adults with a mean age of 40.13 ± 21.91 (mean ± SD) and mixture of both genders (male:female = 68:32). A total of 200 samples were collected from these patients (100 samples each in two different types of heparinized syringes) and analyzed. The pairwise comparison between samples collected in type 1 (LH) and type 2 (DBH) syringes, such as mean with SD, mean difference, 95% LOA, proportion beyond TEa %, are shown in Table 1.

Results of the present study showed that samplings in type 1 (LH) and type 2 (DBH) syringes were comparable with a negligible mean difference. The LOA and sample outside the TEa was 8% for pO₂, pCO₂, and hematocrit (Hct), 7% for BE, 6% for magnesium (Mg²⁺), 5% for potassium (K⁺), calcium (Ca²⁺), and lactate, 4% for bicarbonate (HCO) and sodium (Na⁺), 3% for pH, chloride (Cl^–), hemoglobin (Hb), and glucose, and 0% for oxyhemoglobin saturation (SaO₂). Figures from 1 to 8 shows the Bland–Altman plots for different BG analytes such as pH (Fig. 1), pCO₂ (Fig. 2), calcium (Fig. 3), magnesium (Fig. 4), BE (Fig. 5), lactate (Fig. 6), Hct (Fig. 7), and glucose (Fig. 8) individually.

Close

Fig. 1

Bland–Altman plot of pH for the LH group was comparable with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Close

Fig. 2

Bland–Altman plot of partial pressure of carbon dioxide (mm Hg) for the liquid heparin group was comparable with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Close

Fig. 3

Bland–Altman plot of calcium (mmol/L) for the LH group compared with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Close

Fig. 4

Bland–Altman plot of magnesium (mmol/L) for the LH group was comparable with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Close

Fig. 5

Bland–Altman plot of BE (mmol/L) for the LH group was comparable with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Close

Fig. 6

Bland–Altman plot of lactate (mmol/L) for the LH group was comparable with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Close

Fig. 7

Bland–Altman plot of hematocrit for the LH group was comparable with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Close

Fig. 8

Bland–Altman plot of glucose for the LH group was comparable with that of the DBH group. DBH, dry bound heparin; LH, liquid heparin.

Export to PPT

Discussion

Unfractionated heparin (LH sodium) is widely used and the only suitable anticoagulant for blood sampling for BG analysis. Although an error is quite possible during BG analysis due to the variability of sample volume, dilution, binding, and combining effects of LH, such errors were generally thought to be of theoretical but may be of practical importance and still LH is anticoagulant of choice for BG sampling in many ICUs.¹ The World Health Organization mentioned that 8 to 12 IU/mL of LH should be used in BG analysis, but the exact quantification is not possible in many situations in this LH method, which may lead to errors in BG analysis,⁵ although it is cost-effective and readily mixes up with blood.⁶ Many studies reported that preanalytic errors are the most common form of errors in BG analysis and include hemolysis, effect of temperature, use of plastic syringes, presence of air bubble, delay in sample transport, dilution effects of LH, manufacturing differences among syringe preparation, microclotting in the sample, and error in BG analyzer.⁷, ⁸, ⁹

In our study, we compared the effects of heparin on BG parameters between customized self-prepared LH syringes (type 1) and commercially available DBH syringe (type 2) in adult patients in cardiac surgical ICU because they represented two commonly witnessed practices of syringes at our institute.

In this study, all the respiratory parameters such as a pO₂, pCO₂, and SaO₂ decreased in the LH group than the DBH group but did not show statistically significant difference and had narrow LOA, with TEa limits of sample of 8% for pO₂ and pCO₂ and zero for SaO₂, and results did not show clinical significance between the two groups. Very minimal mean difference was found between the two groups for pH (Fig. 1) and only 3% samples were displaced outside the TEa limits. Hutchison et al and Ordog et al studied the effects of increasing LH on BG analysis and showed that pH value was not affected till 40% dilution of blood samples, but the values of pCO₂, HCOshowed an inverse relation as LH volume.¹⁰, ¹¹ This may be due to the buffering effects of the oxyhemoglobin, plasma proteins, and changes in pCO₂ and HCO, and could not probably be due to heparin salts.¹⁰, ¹¹

In this study, we observed that all electrolytes were lower in the LH sampling group when compared with the DBH group, but the mean difference, LOA, and proportion of TEa beyond limits followed this order, but all were statistically insignificant: Mg²⁺ followed by K⁺, Ca²⁺, and, finally, Na⁺ and Cl^–. Shin et al observed that LH anticoagulant produces more negative bias on Mg²⁺ than Ca , and results were similar to our study.¹² But Chantler and Cox found a good correlation between LH and DBH in their study.¹³

Until now, no study has analyzed and compared all electrolytes between both types of sampling in cardiac surgical intensive care settings in the adult population. Our study results of electrolytes were comparable with previous studies that showed that excessive heparin lead to an underestimation of positively charged ions such as K⁺, Ca²⁺, and Na due to the dilutional and direct binding effects of LH syringe group.¹², ¹⁴, ¹⁵ Chhapola et al and Morimatsu et al in their studies compared the point-of-care testing using heparinized syringe and central laboratory nonheparinized testing and revealed that the decreased levels of K⁺ and Na in the heparin group may be due to the dilution effects.¹⁵, ¹⁶ Similar to previous studies, Cl^–came out to be the least affected ion with the least mean difference, with only 3% beyond TEa limits in our study too. However, all electrolytes are statistically comparable in the LH and DBH groups.

Sandler and Goldstein showed in their results that pCO₂, K⁺, Na⁺, Ca²⁺, and Hb were falling >20% beyond the total allowable error, and Ca²⁺ and Na results were most affected in the 5,000 IU/mL group of sample than the 1,000 IU/mL cohort group.¹⁷ In our customized manual syringe preparation for regular sampling for BG analysis, we use 2,500 units (0.5 mL) of heparin to rinse 2-mL syringes, which seemed to be most appropriate without affecting the results of respiratory, metabolic, and electrolyte analytes when compared with the standard commercially prepared DBH syringes.

As per our study, though the lactate levels increased in the LH group than the DBH group, it did not show statistically relevant difference, and 5% samples were beyond TEa limits. BE was surprisingly increased in the DBH group than the LH group but did not show statistical significance, with 7% samples beyond TEa limits. Delay in the analysis of BG sample can increase blood lactate levels due to anaerobic metabolism.¹⁸ Calatayud and Tenías studied the effect of time, temperature, and blood cell count on lactate levels in heparinized samples and showed that as delay time progresses for analysis at room temperature, parallel lactate levels rise.¹⁸ BG sample should be analyzed <5 minutes of collection to eliminate the unnecessary time-related error of lactate estimation.¹⁹ Chhapola et al demonstrated poor agreement between LH and DBH for the BG parameters pCO₂, pO₂, HCO^3–, Na⁺, K⁺, and Cl and thus are not comparable, but for pH and lactate, two study group syringes can be interchanged in pediatric ICU.²⁰ But as per our study for all the parameters, DBH and LH syringes can be interchanged in adult cardiac surgical ICUs. Chhapola et al also studied the effects of varying volume of heparin on blood parameters and concluded that as the amount of heparin increases in sample syringe, the values of pH, pCO₂, HCO, Na , and Hb were decreased.²¹

A study by Hedberg et al with four different volume of samples compared commercial DBH syringes and showed that low-volumes samples were associated with significant errors in BG, electrolyte, and lactate measurements.²² Yip et al also got similar results: negative bias in positive ions were more seen in the pediatric population than adults.²³

Summary

• Among respiratory parameters, pO₂ (mm Hg) and pCO₂ (mm Hg) showed wide LOA with acceptable sample beyond TEa of 8% for both, but SaO₂% did not show any mean difference between two study groups, with nil sample number beyond TEa% limits.

• pH was comparable between groups, with less mean bias and only 3% sample beyond TEa limits.

• Among the electrolyte, Mg²⁺ (mmol/L), K⁺ (mmol/L), and Ca²⁺ (mmol/L) showed negligible mean difference and good agreement with 6, 5, and 5% of sample beyond TEa of limits, respectively, and Cl^–(mmol/L) was the least affected one.

• Among the metabolic analytes, BE and lactate showed a difference between the two study groups, with 7 and 5% of sample beyond TEa limits, respectively.

• Hct showed minimal mean difference than Hb, with sample beyond TEa limits of 8 and 3%, respectively.

• Our study results did not show any clinical difference in LH and DBH syringes for BG analysis.

• None of the previous studies have proven the comparability between the LH and DBH syringes for all the parameters in BG readings until now.

• We believe that this result mainly depends on the quality of technics of sampling and titration while preparing LH syringes.

Conclusion

Our study results have clearly showed that the LH and DBH syringes were comparable for all the BG parameters analyzed from arterial blood. We did not find any significant statistical difference among the various respiratory, metabolic, and electrolytes, which may be attributed to the appropriate manual preparation of LH syringes for sampling and of course the good technic of sampling, handling, and ready analyzing.