Optimization of Time Interval for Plasma Counting in Plasma Sampling Method for GFR Estimation

INTRODUCTION

The glomerular filtration rate (GFR) is a critical measure of kidney function, reflecting the combined filtration capacity of all functioning nephrons. The GFR is expressed as the volume of plasma filtered by the kidneys over time. Accurate estimation of GFR is crucial for evaluating kidney health, diagnosing urological disorders, monitoring renal failure, and assessing renal transplant function and therapeutic outcomes.^[1-3] Common clinical applications include the evaluation of conditions such as obstructive uropathy, renal injury, chronic kidney disease, and acute renal failure.

MATERIAL AND METHODS

Twenty-five patients (15 males and 10 females; average age 47 ± 12 years) who underwent GFR evaluation by the double-plasma sampling method were included. The methodology is described in [Fig 1], and the material and equipment used are shown in [Fig 2].

Close

Fig 1:

The methodology adopted in this study

Export to PPT

Close

Fig 2:

Material and equipment used for glomerular filtration rate estimation by plasma sampling method

Export to PPT

Counting and calculation

Counting

For optimization of time point for counting of samples for GFR estimation, three time points, i.e., 4 h, 12 h, and 24 h postsecond sampling, were selected. Before performing the counting of samples, gamma-ray spectrophotometer calibration was performed with ¹³⁷Cs – a standard source and photo peak was adjusted using ^99mTc sample. Following calibration, both background counts and sample counts were measured at the specified time intervals (4 h, 12 h, and 24 h), and the data were entered in the datasheet.

Calculation

GFR was computed using an in-house software application developed in Visual Basic 6.0 with a Microsoft Access database backend. After entering patient-specific and radiotracer data – including demographic variables, decay-corrected syringe counts, and plasma sample counts – the software automatically calculated GFR (mL/min) using the Russell double-plasma sampling slope-intercept method [Fig 3].

Close

Fig 3:

Front end of the software

Export to PPT

$GFR={\left[\frac{D\ln \left(\frac{P_1}{P_2}\right)}{T_2-T_1}\exp \left(\frac{T_{1}ln{P}_2-T_2\ln P_1}{T_2-T_1}\right)\right]}^{0.979}$

where GFR is in mL/min,

D = Injected activity (CPM),

T1 and T2 are the time of sample collection,

P1 and P2 are the plasma samples measured at time T1 and T2,

P₁ and P₂ are in counts/min/mL.

RESULTS

In this study, the average GFR estimates were 47.99 ± 31.32, 58.42 ± 32.96, and 59.88 ± 32.99 mL/min at 4 h, 12 h, and 24 h, respectively [Table 1]. The mean GFR estimated by the Gates’ method was 57.86 ± 32.15 mL/min. [Fig 4] gives a graphical comparison of average GFR (mL/min) versus time (h).

Close

Fig 4:

Glomerular filtration rate estimation at 4 h, 12 h, and 24 h using plasma sampling method. GFR: Glomerular filtration rate

Export to PPT

A near-perfect correlation (r = 0.99) was observed between GFR values at 12 and 24 h, suggesting high consistency and comparability. In contrast, the correlation between 4-h and 24-h values was weak and negative (r = −0.32), indicating considerable deviation in GFR estimates when the 4-h sample is used. The percentage difference in GFR estimates by plasma sampling method between these intervals further supports this observation, with a 3.55% difference between the 12- and 24-h values compared to a substantially higher 29.49% between 4- and 24-h values. The percentage difference between these two pairs of time points is shown in Table 2.

The percentage difference in GFR estimates between the Gates’ method and the plasma sampling method was 21% at 4 h, while the differences at 12 h and 24 h ranged from 1% to 3%.

The comparative Bland–Altman analysis was conducted to evaluate the agreement between the two measurement methods for two time intervals: 4–12 h (Dataset 1) and 12–24 h (Dataset 2).

Dataset 1 (4–12 h): The mean difference (bias) was −10.43, with 95% limits of agreement ranging from −21.09 to 0.23. This shows a larger negative bias and a wider spread of differences, indicating lower agreement between the two methods during this period.

Dataset 2 (12–24 h): The mean difference (bias) was −1.46, with 95% limits of agreement ranging from −2.98 to 0.07. This shows a minimal bias and tighter limits of agreement, indicating substantially improved agreement between the two methods in the later time interval.

The Bland–Altman plot [Fig 5] visually confirms that the 12– 24 h measurements show superior agreement with negligible systematic bias compared to the 4–12 h measurements.

Close

Fig 5:

The comparative Bland–Altman plot for two datasets

Export to PPT

DISCUSSION

This study was conducted to optimize the timing of plasma sample counting in double-plasma sampling GFR estimation using ^99mTc-DTPA, with the aim of minimizing dead-time losses in the gamma spectrophotometer and reducing the overall TAT without compromising accuracy. Our results show that 12- and 24-h GFR measurements show excellent agreement, whereas 4-h measurements show considerable deviation, indicating the need for an optimal delay in plasma counting to achieve reliable GFR estimation.

The average GFR values at 4, 12, and 24 h were 47.99 ± 31.32, 58.42 ± 32.96, and 59.88 ± 32.99 mL/min, respectively. The 4-h GFR estimate was significantly lower than the 24-h reference, with a 29.49% difference, while the 12-h GFR showed only 3.55% deviation, suggesting that delayed plasma counting improves the accuracy of GFR estimation.

The correlation analysis reinforces this observation.

A near-perfect correlation (r = 0.99) between 12- and 24-h GFR values indicates excellent reproducibility and reliability of delayed counting.

In contrast, the 4-h and 24-h comparison yielded a negative correlation (r = −0.32), highlighting poor agreement and high variability in early measurements.

The Bland–Altman analysis provided a clear visual and statistical assessment of method agreement:

Dataset 1 (4–12 h) demonstrated a large negative bias (−10.43) and wide 95% limits of agreement (−21.09 to 0.23), confirming that early plasma sampling underestimates GFR due to potential dead-time losses and incomplete plasma clearance distribution.

Dataset 2 (12–24 h) showed a minimal bias (−1.46) and tight limits of agreement (−2.98 to 0.07), indicating clinically acceptable agreement and supporting the 12-h counting interval as a reliable time point for plasma sample measurement.

These findings are in line with previous reports that emphasize the importance of delayed plasma sampling to ensure equilibrium between plasma and interstitial compartments and minimize technical errors in well-counter counting.

Our results suggest that early (4-h) plasma counting may yield misleadingly low GFR values, which could falsely categorize patients into more severe renal impairment stages. By optimizing counting to 12 h, the method achieves high accuracy and reproducibility while maintaining a reasonable TAT for clinical workflow.

LIMITATIONS

The only limitation of our study was the small number of patients; however, the technique involved is also time-consuming; therefore, we decided to conduct the study with a limited number of patients to evaluate the results.

CONCLUSION

Our observational study shows that the early plasma counting, 4 h, is prone to significant underestimation of GFR due to dead-time effects, whereas 12-h counting provides results comparable to 24-h sampling with minimal bias and excellent agreement. Implementing 12-h plasma counting can improve the accuracy of GFR estimation while optimizing clinical workflow in nuclear medicine departments. The most reliable method of double-plasma sampling method for GFR estimation postgates scan at an optimized 12-h counting time instead of the conventional 24 h can substantially reduce reporting turnaround and facilitate timely clinical decision-making for renal patient management.