Multicentre Evaluation of a Newly Developed Anti- Calcium Dobesilate-Enzymatic Creatinine Assay Based on Real-World Data in China


 BackgroundPreviously, we reported that calcium dobesilate (CaD), a vasoprotective agent mainly used for diabetic retinopathy, negatively interferes with enzymatic creatinine assays. Anewly developed enzymatic creatinine assay, the “Cr-R assay”, is commercially available and claims to have no interference from CaD. This study aimed to verify the performance of the Cr-R assay on clinical samples and to investigate the degree of CaD interference in a multicentre, real-world study.MethodsThe precision and accuracy ofthe Cr-R assay was evaluated in 23 hospitals in different regions of China. Interference was then calculated both in vitro and in clinical samples. Samples with potential interference were screened based on the different creatinine results between Cr-R and the creatinine assays commonly used in each participating hospital (Cr-C). Interference was confirmed by determining serum CaD concentration by directly using ultra-performance liquid chromatography (UPLC) methodand by assessing the “true” creatinine levels using the standard isotope dilution mass spectrometry (ID-MS/MS) method. ResultsTheprecision and bias and anti-interference ability of Cr-R assay on most platforms fulfilled the specification criteria. In the clinical setting, 67,469 samples were tested in 23 centres, and CaD concentration was measured in 818 samples, of which 257 contained CaD, accounting for 0.38% of the total and 31% of screened patients. In these 257 cases, the median CaD concentration was 11.356 (range, 1.350–121.519; IQR, 7.162–21.175) μg/mL. Creatinine levels measured using Cr-C assay were up to 35.1% (95%CI, 37.9–32.4 %; P < 0.001) lower than the “real” creatinine levels using the ID-MS/MSmethod. For Cr-R reagent, the average bias from the reference method was -6.2% (95%CI, -7.9–4.6%; P < 0.001), ranging from -17.9% to 7.1%.ConclusionsMulticentre tests confirmed that the precision, bias, and anti-CaD ability of the Cr-R assay met clinical needs. The prevalence of CaD interference should be considered when performing clinical enzymatic creatinine measurements.Other manufacturers should improve the anti-interference performance of creatinine reagents.


Abstract Background
Previously, we reported that calcium dobesilate (CaD), a vasoprotective agent mainly used for diabetic retinopathy, negatively interferes with enzymatic creatinine assays. Anewly developed enzymatic creatinine assay, the "Cr-R assay", is commercially available and claims to have no interference from CaD. This study aimed to verify the performance of the Cr-R assay on clinical samples and to investigate the degree of CaD interference in a multicentre, real-world study.

Methods
The precision and accuracy ofthe Cr-R assay was evaluated in 23 hospitals in different regions of China.
Interference was then calculated both in vitro and in clinical samples. Samples with potential interference were screened based on the different creatinine results between Cr-R and the creatinine assays commonly used in each participating hospital (Cr-C). Interference was con rmed by determining serum CaD concentration by directly using ultra-performance liquid chromatography (UPLC) methodand by assessing the "true" creatinine levels using the standard isotope dilution mass spectrometry (ID-MS/MS) method.

Results
Theprecision and bias and anti-interference ability of Cr-R assay on most platforms ful lled the speci cation criteria. In the clinical setting, 67,469 samples were tested in 23 centres, and CaD concentration was measured in 818 samples, of which 257 contained CaD, accounting for 0.38% of the total and 31% of screened patients. In these 257 cases, the median CaD concentration was 11.356 (range, 1.350-121.519; IQR, 7.162-21.175) μg/mL. Creatinine levels measured using Cr-C assay were up to 35.1% (95%CI, 37.9-32.4 %; P < 0.001) lower than the "real" creatinine levels using the ID-MS/MSmethod. For Cr-R reagent, the average bias from the reference method was -6.2% (95%CI, -7.9-4.6%; P < 0.001), ranging from -17.9% to 7.1%.

Conclusions
Multicentre tests con rmed that the precision, bias, and anti-CaD ability of the Cr-R assay met clinical needs. The prevalence of CaD interference should be considered when performing clinical enzymatic creatinine measurements.Other manufacturers should improve the anti-interference performance of creatinine reagents.

Background
Creatinine is the basic indicator for monitoring the renal function of patients and the most critical index for calculating the estimated glomerular ltration rate in both prognosis and treatment monitoring. The traditional alkaline picrate method used to determine creatinine concentrations in the plasma or serum is affected by interference from numerous endogenous and exogenous substances [1,2]. Recently, the Trinder reaction-based enzymatic creatinine assay has been increasingly utilised because of its sensitivity and speci city and its compatibility with automated analysis [3]. According to the National Centre for Clinical Laboratories (NCCL) of China, more than 75% of clinical laboratories (2548/3116) in China used enzymatic assays to quantify creatinine in 2019 [4]. In recent years, manufacturers have made efforts to standardise their creatinine measurements to make the calibration traceable to reference methods, such as isotope dilution mass spectrometry/mass spectrometry (IDMS/MS), or to standard materials [5]. However, calibration traceability still cannot address issues such as interference.
As early as 1986, the vasoprotective agent calcium dobesilate (CaD, calcium 2,5dihydroxybenzenesulfonate) was reported to interfere with enzymatic creatinine detection [2]. This issue was not resolved through additional research [6,7]. CaD is a synthetic vasoprotectant that can remarkably reduce capillary permeability, blood viscosity, and high platelet activity, as well as improve abnormal hemorheology and microcirculation [8]. It is used to treat microangiopathies, particularly diabetic retinopathy (DR), worldwide [8][9][10][11][12][13]. Recent studies showed its protective effects on diabetic nephropathy and its ability to reduce albuminuria [14,15]. Treatment strategies that include CaD may prevent or delay the progression of DR and nephropathy [16]. However, when CaD is used clinically to prevent and treat DR and/or DN, negative interference with enzymatic creatinine may lead to incorrect evaluation of renal function in patients with diabetes, which can mask the patient's condition and delay treatment, leading to poor clinical outcomes. Increased clinical application of CaD drugs has exacerbated the interference with creatinine measurements. In 2015, we reported clinically signi cant negative interference with CaD in eight Trinder reaction-based enzymatic creatinine assays [17]. According to our survey of 638 laboratory staff-members in different regions of China in 2019, 11% of respondents identi ed cases of creatinine interference caused by CaD. Thus, CaD interference in creatinine measurement is a common clinical problem that urgently requires resolution. Reducing the interference of the enzymatic creatinine reagent may be su cient to overcome this issue.
In May 2019, an assay to measure enzymatic creatinine claiming no interference with CaD concentrations below 100 μg/mL became commercially available in China (referred to as the Cr-R assay). This is the rst enzymatic creatinine reagent worldwide that claims to have no interference from CaD. The present study was conducted to verify the performance of this new Cr-R assay based on a multicentre investigation and real-world data. Because other commercially available creatinine assays (referred to as Cr-C assays) have signi cant negative interference with CaD, serum samples from patients being administered CaD contain the drug, resulting in a signi cantly lower creatinine value using traditional Cr-C assays. Clinical recommendations based on the present ndings should guide decision making going forward, particularly in those patients with higher CaD levels.

Participating laboratories and experimental design
The present study was performed in 2019 and involved 23 clinical laboratories located in 23 hospitals from 15 cities in different regions of China. The biochemical analysers used in the 23 laboratories included the Abbott Architect (n = 1) (Abbott Diagnostics, Park City, IL, USA), Beckman AU (n = 15) (Beckman Coulter, Brea, CA, USA), Hitachi 7600 or 7180 (n = 3) (Tokyo, Japan), and Roche Cobas (n = 4) (Roche, Basel, Switzerland). The analytical performance of the Cr-R assays was evaluated on these 23 platforms. The Cr-R assays and calibrators were provided by Maccura Biotechnology (Chengdu, China). All reagents and calibrators used in this study were standardised across participating laboratories, with matching lot numbers. The detailed characteristics of the instruments and Cr-C assays used by the 23 participating laboratories are shown in Supplementary Table 1. Precision, method comparison and bias estimation, and anti-CaD ability of the Cr-R assay were evaluated based on the experimental work ow in EP15-A guidelines [18]. The high-and low-level serum pools were evaluated using the Cr-R assay four times daily for ve consecutive days in each of the 23 participating clinical laboratories. An internal quality control was included to ensure test quality. The standard deviation (SD) and coe cient of variation (%CV) were calculated to determine the repeatability (within-run precision) and total precision (within-lab precision) for each assay. The inter-laboratory variation in the 23 participating laboratories was also calculated.
The within subject (CV I ) (4.4%) and between-subject (CV g ) (14.3%) variation of creatinine from the

European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) Biological Variation
Database [19] was used for the assessment of the performance characteristics of the creatinine assays.

Method comparison and bias estimation
The creatinine assays were compared according to CLSI EP9-A3 guidelines [21]. Forty samples were measured in the 23 laboratories using the Cr-R assayson their automated analysers in 5 days, with 8 samples per day. Each sample was tested twice and within 3 h of thawing. The target values of serum pools were assigned by a reference measurement based on isotope dilution liquid chromatographytandem mass spectrometry (LC-IDMS/MS) in two joint committees on traceability in laboratory medicinelisted reference laboratories: the NCCL, China, and the Reference Laboratory of Maccura Biotechnology (Chengdu, China). The average of these two laboratory measurements was used as the target value for each sample.

In vitro interference study
Three drug-free serum pools (Cr = 80, 150, 442 μmol/L) were spiked with a CaD standard to prepare a dose-response series according to CLSI EP7-A2 guidelines [22]. The nal CaD concentrations in each sample were 0, 8, 15, 30, 60, and 100 μg/mL. The CaD standard (100573, 94.8%) used in in vitro experiments was purchased from the National Institutes for Food and Drug Control of China (Beijing, China). In each participating hospital, the CaD-containing serum series was analysed using the Cr-C and Cr-R assay in triplicate within one analytical run to obtain an average value. The percent bias at each concentration of CaD was calculated relative to that of the drug-free specimen. Based on the biological variation, the acceptable limit of deviation for the in vitro experiments was de ned as ±5.6%.
Interference study based on real-world clinical data Twenty-three participating laboratories used the Cr-C and Cr-R assays to simultaneously detect creatinine in the remaining serum of the clinical samples for 3-5 consecutive days. Each laboratory evaluated approximately 3000 samples, which is clinically required to detect creatinine. If the creatinine measurement according to the Cr-C assay was 10% lower than that of the Cr-R measurement, the remaining serum was collected and stored in aliquots at -80 °C. CaD levels were accurately measured in these serum samples at PUMCH as described below. Fifty samples were randomly selected from those samples containing CaD to measure the creatinine levels using the above reference method, and deviations between the results of the two assays and the reference method were calculated.

Precision evaluation
The mean creatinine concentrations of the low and high-level serum pools as determined by the 23 Cr-R assays were 82.5 ± 3.4 μmol/L and 309.4 ± 7.2 μmol/L, respectively. The inter-laboratory variation of the low-and high-level pooled samples were 4.2% and 2.3%, respectively. Repeatability (within-run precision) and total precision (within-lab precision) for each assay are shown in Fig. 2. The within-laboratory CV% of the Cr-R assays reached the minimal speci cations (3.3%) for all assays, except for in three laboratories (Beckman AU5400-2, Beckman AU5800-9, Roche Cobas 702-1) for the low-level creatinine serum pool.
The high-level creatinine serum pool had less overall variation among laboratories, and the repeatability of the 22 Cr-R assays reached the minimum speci cation of 1.1% (Fig. 2B). The within-laboratory CV% was 0.4-1.9% (Fig. 2D), ful lling the speci cation criteria (2.2%).

Method comparison and bias estimation
Passing-Bablok regression analysis was performed for each creatinine assay and the LC-IDMS/MS method. Bias at the medical decision levels of the individual creatinine assay for each of the 23 participating clinical laboratories is presented in Fig. 3. The regression equation of the Passing-Bablok regression and 95%CI of bias at the four medical decision levels are shown in Supplementary Table 2 and 3, respectively. The optimal speci cation for bias (1.9%) was reached in three Cr-R assays (Beckman AU5400-2, Roche Cobas 702-2, Beckman AU5800-2), whereas the desirable goal for bias (3.7%) was reached in three Cr-R assays at all medical decision levels. The other seven Cr-R assays ful lled the minimal speci cation criterion (5.6%) at all medical decision levels. The lowest creatinine concentration showed the largest bias for most Cr-R assays. At the medical decision level of 88.4 μmol/L, 34.8% (8/23) of Cr-R assays failed to reach the minimal bias speci cation of 5.6% (Fig. 3A). At the medical decision level of 265 and 442 μmol/L, all Cr-R assays showed the desirable bias speci cation of 5.6% except for the Beckman AU5400-4 (Fig. 3C, D).

In vitrointerference study
The results of the in vitro interference test are shown in Fig. 4. Deviations in the Cr-R assays in determining the median and high-Cr interference pools were -2.5% to 4.95% and -5.2% to -0.12%, respectively, both meeting the minimal bias speci cation of 5.6% (Fig. 4B, C). For low-Cr interference pools, deviations in the Cr-R assays ranged from 1.82% to 15.95% (Fig. 4A). Exogenous addition of CaD clearly exhibited concentration-dependent negative interference in creatinine measurement in the Cr-C assays in all of the 23 participating laboratories. When the CaD concentration was 100 μg/mL, the percentage bias relative to the drug-free specimen of the low, median, and high creatinine concentration interference serum samples according to Cr-C ranged from -77.78% to -38.55%, -71.47% to -33.33%, and -58.72% to -22.47%, respectively ( Fig. 4A-C).
Interference study based on real-world data A total of 67 469 samples were evaluated using the Cr-R and Cr-C assays simultaneously to measure creatinine concentrations in 23 participating clinical laboratories. Following this, 818 samples were screened out because the results of the Cr-C assay were 10% lower than the Cr-R measurement. CaD was detected in the serum of 257 patients out of the screened samples, accounting for 0.39% of all patients and 31% of the screened samples. In these 257 cases, the median CaD concentration was 11.356 (range, Fifty samples were randomly selected from the CaD-containing samples to accurately measure creatinine levels by LC-IDMS/MS. A higher serum creatinine level was associated with a higher CaD concentration (Fig. 5A). Thus, creatinine levels according to the Cr-C assay were up to 35.1% (95% CI, 37.9 to 32.4%; P < 0.001) lower than the actual creatinine levels. For Cr-R reagent, the average bias from the reference method was -6.2% (95% CI, -7.9 to -4.6%; P < 0.001), ranging from -17.9 to 7.1% (Fig. 5B).

Discussion
To overcome the problem of CaD interference with enzymatic creatinine measurements, and thus reduce clinical risk, an enzymatic creatinine test, the Cr-R assay, claiming to have no interference from CaD below100 μg/mL was introduced in China. In the present study, we investigated the basic analytical performance of the Cr-R assay on different platforms and in different hospitals. According to the 2014 EFLM European Biological Variation Study, the performance speci cations of a clinical creatinine test depend on the within-subject (3.4%) and between-subject (14.3%) variation of creatinine [19], which were obtained from the latest data from the EFLM database. Pooled serum samples were used to determine the within-lab %CV variation of the Cr-R assay, with 3 out of the 23 participating laboratories not meeting the minimum standards on pooled samples. The status of each biochemical analyser, including the age of the sample loading stage, the use of the light source, and the cleanliness of the cuvettes may affect the precision of the results. The within-laboratory CV% of the laboratory using a Beckman AU5400-2 was as high as 4.5%. Further investigation revealed that the instrument had been in use for more than 6 years. In the present study, only 13Cr-R assays ful lled the minimal speci cation criterion (5.6%) at all medical decision levels. The accuracy of creatinine measurement at low levels therefore still remains a major challenge. From the summary report of the Creatinine Accuracy Calibration Veri cation/Linearity LN24-A 2019, organised by the College of American Pathologists, the allowable error at low creatinine levels is ±14.5%, and at high creatinine levels it is ±7.5%. When using the minimal speci cation criterion (5.6%) proposed by this study, the results from approximately half of the laboratories assessed in the LN24-A 2019 may not meet minimum standards for low creatinine levels. Although all 23 laboratories used the same lot of Cr-R reagents and calibrators, the bias obtained by different research centres varied widely, even when using the same model of biochemical analyser. This may have occurred for the following reasons: First, creatinine uses a single-point calibration, and any deviations, such as improper reconstitution of the calibrator or differences in temperature and time exposure when the calibrator is opened, may therefore cause systemic changes in the outcomes of the assay. Second, the large bias in some centres was caused by large day-to-day variation, which may be related to instrument maintenance or reagent stability. We instructed participating laboratories to operate according to their internal QA/QC guidelines, without requiring multiple calibrations for Cr-R assays. The large levels of variation obtained using standard reagents should therefore be alarming to laboratories indicating the need to give more attention to daily QC and calibration.
We also spiked pooled serum samples with increasing concentrations of CaD to evaluate alongside these pooled samples to determine the critical concentration at which CaD levels affect the results of the Cr-R and Cr-C assays. According to these results, CaD levels > 8 μg/mL result in the under-representation of serum creatinine levels according to the Cr-C assay (Fig.4). For the Cr-R assay, increased CaD concentrations tended to show slightly higher values only in low-Cr interference pools, which was obvious at 60 μg/mL or higher CaD (Fig.4A). However, based on pharmacokinetic information (23) and actual measured results, when the patient's creatinine level is close to the low-Cr interference pools, the CaD concentration is not expected to reach 60 μg/mL (Fig. 5A). As the usual recommended dosage of CaD is 500 mg 3 times per day, the blood serum concentration is estimated to be 15 μg/mL at the steady state [23]. However, CaD is a small molecule that is primarily excreted through the kidneys. Patients with renal failure have exceedingly low glomerular ltration rates and are thus prone to drug accumulation (Fig. 4A), which may result in higher serum CaD levels and thus, more severe interference with the results of the Cr-C assay (Fig. 4B). At the highest CaD concentration (121.519 μg/mL) in the screened samples, the creatinine levels according to the Cr-R assay did not show strong deviation (-5.2%) from the levels as determined by LC-IDMS/MS. In summary, the new Cr-R assay met clinical needs according to its ability to measure serum creatinine levels without interference from CaD.
When applying the Cr-R and Cr-C assays to a large clinical cohort comprised of 67 469 samples, we identi ed 818 patients with large differences in creatinine levels according to the Cr-R and Cr-C assays, while only one-third of these patients had detectable CaD in their serum. For the remaining samples, the observed differences may be partially attributable to systemic differences between different reagents or other unknown interference factors in the serum samples. A number of substances may interfere with creatinine quanti cation using enzymatic assays [24][25][26][27][28][29]. For some patients with large deviating Cr-R and Cr-C assay results who did not have detectable serum CaD, we evaluated the patient's medical records and compared the creatinine levels with those determined by the reference method and identi ed ten cases where this discrepancy may have been caused by the presence of etamsylate. Etamsylate is known to interfere with enzymatic creatinine assays [28][29]. The Cr-R assay also solved the problem of interference by etamsylate, and generally gave more accurate serum creatinine results in these cases. The manufacturer of the Cr-R assay claims that serum etamsylate levels below 250 μg/mL will not interfere with the assay results, and that the Cr-R assay may also work in the presence of other analogues, which interfere with enzymatic creatinine assays.
There were some limitations to the present study. Frist, in the precision assessment, we used EP15A but not the latest EP15A3 due to the volume limitation of the serum pools. Second, the serum pools used in the precision evaluation were not checked for homogeneity and freeze-thaw effects. However, previously published data showed that a freeze-thaw cycle does not alter the stability of creatinine in pooled serum [30].

Conclusions
Both in vitro addition tests and real-world data revealed that the new Cr-R assay can overcome clinical CaD interference with creatinine detection. Multicentre tests con rmed that the precision and accuracy meet the needs of clinical testing. Additionally, it is important to further raise awareness of such interference so that clinical physicians can make judgements on the levels of creatinine in patients under CaD therapy. Other manufacturers should improve the anti-interference performance of creatinine reagents.  Figure 1 The experimental work ow of the study. LC-IDMS/MS, isotope dilution liquid chromatography-tandem mass spectrometry; UPLC, ultra-performance liquid chromatography.