Sensitive Ion-Chromatographic Determination of Citric Acid in Urine

Abstract

Urine citrate analysis is relevant in the screening and monitoring of patients with calcium nephrolithiasis. A sensitive, fast, easy, and low-maintenance ion chromatographic (IC) method with conductivity detection for the analysis of urine citrate is developed and validated. Its application on true samples is also reported. Sample urine is diluted with a water solution containing internal standard (IS) before the chromatographic assay. The isocratic chromatographic run time is twenty-five minutes, using sodium hydroxide aqueous solution as the mobile phase. The method is fully validated as a quantitative method to objectively demonstrate its applicability for the intended use. The analytical response is linear in the 0.08–10.4 mmol/L concentration range. Precision and accuracy studies carried out on spiked urine and internal quality control samples reveal an imprecision CV% lower than 11% and an accuracy between 85 and 103%. The stability of citrate in urine samples is also evaluated. An easy, rapid, and low-maintenance, cost-effective IC method for urinary citrate determination is developed and validated. Internal standardization improves reliability and precision. The method has been currently used in our laboratory over recent years to analyze more than 1000 samples per year.

1. Introduction

Citric acid is an important intermediate in metabolic function. Its urinary excretion in mammals varies under different physiological conditions. In humans, 10–35% of filtered citrate is excreted and variations in the acid–base balance, or changes in intracellular pH, influence citrate excretion. Metabolic alkalosis increases citrate excretion, while metabolic acidosis decreases it [1,2,3].

Citrate in urine contributes to the inhibitory potential toward crystallization of calcium salts [4,5,6], acting through both surface-controlled mechanisms [7,8,9] and the formation of stable soluble complexes with calcium [10,11]. Therefore, citrate is a relevant component of ionic equilibria in urine and its determination is essential to assess the state of saturation with respect to calcium oxalates and phosphates [12,13].

Indeed, low citrate may be crucial in lowering the inhibition potential toward calcium stone formation in the urine environment [14,15,16]. Hypocitraturia has been reported as a common metabolic abnormality in patients with calcium stones, occurring in as many as 15% to 63% of cases. While it is often associated with other metabolic defects, in isolation, it accounts for about 10% of the patients [5,17].

There is convincing clinical evidence that the oral administration of alkaline citrate salts significantly reduces the calcium stone recurrence rate, taking into account both calcium complexation and crystal growth, and aggregation inhibition [18,19,20,21].

The traditional enzymatic spectrophotometric method for the determination of citrate uses citrate lyase (EC 4.1.3.6), malate dehydrogenase (EC 1.1.1.37), lactate dehydrogenase (EC 1.1.1.27), and NADH [22,23,24,25]. We have previously reported some modifications of earlier methods that eased routine use and improved sensitivity [26].

Ion chromatographic (IC) methods have been previously described for the determination of urinary citrate [27,28,29,30]. Other methods have been proposed, including capillary electrophoresis (CE) [31,32]. Holmes described a method for measuring urinary oxalate and citrate using CE and indirect UV absorbance detection [33], while Munoz et al. validated a method for measuring urine oxalate, citrate, uric acid, and creatinine [34]. In more recent studies, liquid chromatography–tandem mass spectrometry (LC–MS/MS) was employed, in which urine was either pre-treated by solid-phase extraction [35] or simply diluted with water [36].

Herein, we describe a simple IC method for the quantitative determination of urinary citrate. As well as using internal standardization, it is fast, cost-effective, and requires low instrumental maintenance. It has been validated as a quantitative method, according to the main guidelines on bioanalytical method validation [37,38,39], and successfully applied to the analysis of thousands of urine samples.

2. Materials and Methods

2.1. Chemicals

All reagents were of analytical grade. Hydrochloric acid 37%, sulfuric acid 98%, sodium hydroxide solution 50% (w/w), 100% w/v trichloroacetic acid, 5.20 mmol/L of IC-grade citric acid (Citrate Standard for IC) certified standard solution, trisodium trimetaphosphate purity >95% (IS), trichloroacetic acid 30% w/v, bis-tris purity ≥99.0%, trichloroacetic acid purity ≥99.0%, synthetic urine (Surine, Negative Urine Control), and citrate lyase from Aerobacter aerogenes (EC 4.1.3.6) lyophilized with a specific activity ≥0.25 U/mg (at 25 °C with citrate as the substrate) were purchased from Merck (Darmstadt, Germany). Hibitane (chlorhexidine gluconate 20%) was purchased from VWR International S.r.l. (Milano, Italy). Ultrapure water (conductivity <0.05 µS/cm; ASTM parameter Type I) was obtained by using a 1720 Deionizer Device (G. Maina, Pecetto Torinese, Torino, Italy).

2.2. Standard and Working Solutions

A certified 5.20 mmol/L citric acid standard solution was stored at 4 °C and used for the two-point, 2.60 and 5.20 mmol/L, routine calibration curve. The 2.60 mmol/L solution was prepared by diluting the concentrated 5.20 mmol/L standard.

A 100 mmol/L IS stock standard solution was prepared in 10 mL of ultrapure water and 0.5 mL aliquots were stored at −20 °C for one year. The operative 0.05 mmol/L IS solution was prepared by dissolving a 0.5 mL aliquot of stock solution in 1000 mL of water in a dark glass bottle with a top dispenser.

The eluent, a 20 mmol/L NaOH aqueous solution, was freshly prepared by mixing 1.0 mL of sodium hydroxide 50% (w/w) into 1000 mL of degassed ultrapure water.

The solution for the suppression of conductivity was prepared by mixing 3.6 mL of sulfuric acid 98% into 5 L of ultrapure water.

Citrate lyase from Aerobacter aerogenes (120 U, 302 mg) was suspended in 9 mL of ultrapure water, and 1 mL aliquots were stored at −20 °C. Suspended enzyme was stable for at least 1 month at +4 °C or 6 months at −20 °C.

2.3. Internal Quality Control

Quality control samples were prepared from true urine samples with a citrate concentration <1.5 mmol/L (low control) and >5.0 mmol/L (high control). An amount of 2 L of urine containing the proper citrate concentration was collected, mixed, acidified by adding 5 mL/L of hydrochloric acid 37%, and subjected to five cycles of subsequent freezing, defrosting, and filtration steps. Once ready, the control samples were subdivided into 10 mL aliquots and stored at −20 °C in the dark for one year. Weekly, one aliquot was defrosted and stored at +4 °C.

2.4. Sample Preparation

2.4.1. Urine

Both second-morning fasting (2 h) and 24 h (24 h) urine samples were analyzed. The 2 h urine specimens were acidified with hydrochloric acid 37% (50 µL of 37% hydrochloric acid/5 mL of urine) and stored at 4 °C. The 24 h urines were collected upon addition of 10 mL of 37% hydrochloric acid and stored at 4 °C. Samples were analyzed within one week of the collection.

Each analytical series included two controls (low and high) and two points of calibration (2.60 and 5.20 mmol/L). Amounts of 25 µL aliquots of the urine sample, standard solution, or control were transferred into 2 mL glass capped autosampler vials. An amount of 1.5 mL of IS working solution was added. The mixtures were vortexed and analyzed.

2.4.2. Plasma

Plasma samples can be analyzed as well using the aforementioned instrumental method. Blood samples were collected into heparinized vials and centrifuged. One 200 µL aliquot of plasma, placed in a 1.5 mL plastic-capped vial, was deproteinized by adding 40 µL of trichloroacetic acid, 30% w/v, vortex-mixed, and centrifuged at 4.000 g for 10 min. The clear supernatant was stored at −20 °C until analysis. An amount of 50 µL of the extract was diluted with 1.5 mL of IS and the mixture was injected into the chromatograph.

2.5. Apparatus and Chromatographic Conditions

Chromatographic separation was performed using a Dionex, Series 4500i, Ion Chromatograph (IC) system (Thermo Fisher Scientific, Waltham, MA, USA), including an eluent degasser, a gradient pump, and a conductivity detector. The system was equipped with a Dionex ACRS 500 conductivity suppressor. A Merck Hitachi AS 2000A (Hitachi Ltd., Tokyo, Japan) autosampler was used. The IC was equipped with a Dionex IonPac^TM Fast Anion IIIA column (3 mm × 250 mm) and a Dionex IonPac^TM Fast Anion IIIA guard-column (3 mm × 50 mm) (Thermo Fisher Scientific, Waltham, MA, USA). The mobile phase was a 20 mmol/L aqueous sodium hydroxide solution, at a flow rate of 1 mL/min, at a 23 °C temperature. The injection volume was 20 µL. The eluent conductivity background was suppressed by using a Dionex AMMS/LS-ll cation-exchange membrane set in-line after the separator. It was regenerated with 12.5 mmol/L of sulfuric acid flowing countercurrently at a flow rate of 1.5 mL/min.

2.6. Method Validation

All the validation parameters usually required for quantitative bioanalytical procedures were determined [37,38,39], including selectivity, carry-over, lower limit of quantification (LLOQ), limit of detection (LOD), calibration curve, accuracy, precision, matrix effect, and stability. Synthetic urine was used for the validation experiments following the analytical protocol described above.

2.6.1. Selectivity

The synthetic urine was analyzed in order to verify the effective absence of both citrate and other potential interferences at the expected retention time interval of analyte and IS. The correct identification of the analyte was evaluated considering the within- and between-run relative retention time (RRT) precision in urine samples spiked at three concentration levels (0.32, 1.30, and 5.20 mmol/L) and in two sets of internal quality control samples (low and high level, respectively, of 1.06 and 5.93 mmol/L).

Selectivity was confirmed by analyzing citrate standard solutions and true urine samples that were incubated either in the presence of citrate lyase or without it.

An amount of 25 µL of each sample or standard solution was placed in double into a/s glass vials. Amounts of 50 µL aliquots of Bis-Tris buffer (200 mmol/L, pH 8.0) were added. Each series of mixtures were added with 25 µL of suspended citrate lyase (0.33 Units) or 25 µL of water.

After 60 min of incubation at 23 °C, 1.5 mL of IS solution was added to all samples and the mixtures were injected into the chromatograph.

2.6.2. Carry-Over

Carry-over was evaluated by injecting an alternate sequence of high-internal-quality control samples and ultra-pure water. To ensure the absence of any carry-over effect, no analytical signals had to be present at the attended citrate and IS retention times in ultrapure water chromatograms.

2.6.3. Calibration Curve

The linear calibration model was checked by analyzing synthetic urine samples spiked at six concentration levels (0.08, 0.32, 1.30, 2.60, 5.20, 10.4 mmol/L). Each concentration level was measured in triplicate. A calibration curve was obtained by plotting the relative peak area (citrate/IS) against concentration of citrate. In addition, a blank sample (processed matrix sample without analyte and without IS) and a zero sample (processed matrix with IS) were processed, but not considered among the calibration curve parameters.

2.6.4. LOD and LLOQ

An estimation of the LOD and LLOQ values was obtained by analyzing 10 replicates of blank samples (synthetic urine), as the concentrations corresponding to the average signal plus 3 and 10 times the standard deviation, respectively. A three-level and two-replicates calibration curve, in the range of LLOQ (0.04–0.08–0.16 µmol/L), was used to calculate the LOD and LLOQ concentrations.

LOD and LLOQ values were then experimentally verified. At a LLOQ concentration, the accuracy (expressed as percentage) should be within ±20% of the nominal value and the precision (expressed as percent variation coefficient, CV%) should be below 20% [39]. The analyte signal of samples at the LLOQ level should be at least 5 times higher than the signal of blanks [37].

2.6.5. Accuracy and Precision

Accuracy (%) was evaluated on synthetic urine samples spiked at four concentration levels (0.08, 0.32, 1.30, and 5.20 mmol/L; n = 10).

The within- and between-run precisions (CV%) were evaluated on quality control samples (low and high levels; n = 10) and analyzed in the same batch and across different days. For the between-run precision estimation, the last ten analytical sessions were considered. Within- and between-run precisions were evaluated also on spiked urine samples for accuracy determination.

2.6.6. Matrix Effect

The matrix effect was evaluated by comparing peak areas of citrate (mean value from five replicates) obtained by analyzing both synthetic urine and ultrapure water at three concentration levels (0.32, 1.30, and 5.20 mmol/L). Differences highlight matrix suppression (values below 100%) or signal enhancement (values above 100%).

2.6.7. Urinary Citrate Stability

We collected spot urine samples from 8 male and 4 female volunteers. These samples were pooled (about 2 L collected), mixed, analyzed immediately for citric acid, and subdivided into five aliquots. Each aliquot was spiked with: (i) Hibitane (chlorhexidine gluconate 20%)—2 mL/L of urine; (ii) HCl 37% (w/w; 10 mol/L)—10 mL/L of urine; (iii) acetic acid 100% (w/w; 10 mol/L)—10 mL/L of urine; (iv) trichloroacetic acid (solid)—10 g/L of urine; (v) no preservative. Each differently preserved urine was subdivided into five 5 mL aliquots, using 10 mL capped plastic vials (25 vials in total). Vials were stored at +4 °C. Analyses of each of the differently preserved urine were performed in the following 1, 2, 5, 10, 30 days.

2.7. Application to Real Samples

Citrate and creatinine were analyzed in both 2 h and 24 h urine samples. Urines were from 16 healthy subjects, 7 females, aged 28 through 58 years, on a free home diet and with normal renal function. The subjects provided written informed consent before attending the study, and an anonymous code was attributed to each subject participating in the study to adhere to privacy regulations.

3. Results

3.1. Method Development

We analyzed citrate in urine by using an enzyme-spectrophotometric method, based on citrate lyase-catalyzed and phenylhydrazine reactions [26]. A suppressed conductivity IC method operating with a weak anion exchanger and a diluted alkaline mobile phase was subsequently developed. Under these conditions, the complete dissociation of citrate, its chromatographic handling, and efficient suppression of conductivity could be achieved.

We tested two different weak anion exchangers that showed useful performances (DIONEX^TM IonPac^TM Fast Anion IIIA and IonPac^TM AS22 Fast). The flow of the mobile phase (1 mL/min) was selected in agreement with the suggested operative conditions, which are compatible with a “low-pressure” chromatographic system.

For both columns, the main interference for citrate was isocitrate, which is excreted in urine at a lower extent (5–10%).

Different concentrations of NaOH in aqueous solution (10, 20, 30, and 40 mmol/L) were used as the mobile phase. The chromatographic performances of the columns were compared by evaluating the logarithm capacity factors (log k′) and resolution (R) for the two citrate-isocitrate peaks. The data obtained for each column are reported in

Table 1. Method development preliminary proofs. Measured rt and log k′ for citrate, isocitrate, and IS peaks, and R (citr-isocitr) as a function of mobile phase concentration.

Column	IonPacTM Fast Anion IIIA							IonPacTM AS22 Fast
Analyte	Citrate		Isocitrate		IS		R(CITR-ISOCITR)	Citrate		Isocitrate		IS		R(CITR- ISOCITR)
Mobile Phase [NaOH] (mmol/L)	rt (min)	log k′	rt (min)	log k′	rt (min)	log k′		rt (min)	log k′	rt (min)	log k′	rt (min)	log k′
40	2.65	−1.222	2.73	−1.036	3.27	−0.511	0.87	2.83	−0.258	3.06	−0.170	4.93	0.230	1.28
30	3.51	−0.396	3.73	−0.310	4.98	−0.003	1.61	4.73	0.206	5.43	0.300	11.04	0.706	3.12
20	6.99	0.254	7.72	0.319	12.0	0.581	2.97	11.9	0.755	14.3	0.849	34.5	1.267	4.55
10	38.6	1.159	43.6	1.216	77.1	1.475	3.41	75.5	1.62	92.0	1.71	>110	-	2.13

.

By decreasing the eluent concentration, the resolution generally increased, except for the IonPac^TM AS22 fast column that, with NaOH 10 mmol/L as the mobile phase, yielded excessive peaks widening. Furthermore, with this column, peaks were quite asymmetrical. Thus, we chose to proceed with IonPac^TM Fast Anion IIIA.

Complete separation (R ≥ 1.5) was achieved already at 30 mmol/L. Notwithstanding, in the routine, we preferred to operate with 20 mmol/L. Under this condition, despite the longer run time, satisfactory separations were obtained even after thousands of analyzed samples.

Thanks to the higher specificity of the chromatographic method and the relatively high urinary citrate concentrations, a simple urine dilution was applied, making further sample pre-treatments unnecessary.

3.2. Method Validation

3.2.1. Selectivity

The chromatographic profiles of the synthetic urine sample did not show significant signals at retention times of citrate and IS. The analyte was clearly identified in all the spiked samples and quality control samples, with RRT within- and between-run imprecisions under 1%. Retention times gradually decreased over time as a function of the number of analyzed samples with the same column.

The high specificity of citrate lyase toward citrate is known [22]. A standard solution and three urine samples were incubated in the presence of citrate lyase at pH 8. Figure 1 shows analytical traces of a sample either incubated or not in the presence of the enzyme: all citrate was destroyed within 60 min of incubation. Complete disappearance of the citrate peaks indicates that the method was selective and not affected by interfering substances. The isocitrate peak, eluting 1 min later, did not interfere (see Figure 1). It normally represents 5–10% of the citrate peak and could be taken into account in the estimation of citrate contribution to supersaturation [13].

3.2.2. Carry-Over

The chromatographic profiles monitored during the analysis of ultra-pure water injected after a high-level quality control sample did not show the presence of significant signals at the peak retention times expected for both analyte and IS.

3.2.3. Calibration Curve

Quantitative data resulting from area counts were corrected using the respective IS areas. The results were linearly correlated according to the equation y = 0.0687x − 0.0100; R² = 0.9999 (y = relative peak area; x = concentration of citrate in mmol/L).

3.2.4. LOD and LLOQ

LOD and LLOQ estimated values were 0.03 mmol/L and 0.07 mmol/L respectively. The equation used to convert the analytical signal into concentration is y = 0.0772x − 0.0003. In addition, 10 replicates of synthetic urine spiked with citrate at 0.04 mmol/L and 0.08 mmol/L were analyzed. At 0.08 mmol/L, the CV% was 11%, the accuracy was 85%, and the citrate signal was more than 5 times the signal of the blank samples. At 0.04 mmol/L, the citrate peak was clearly detectable (S/N ≥ 3).

3.2.5. Accuracy and Precision

Within- and between-run data on accuracy are reported in

Table 2. Accuracy (n = 10) and matrix effect (n = 5) data of citrate-fortified synthetic urines.

Expected Concentration (mmol/L)	Accuracy (Percentage)			Matrix Effect (%)
	Day 1	Day 7	Inter-Day
0.08	85	-	-	-
0.32	103	99.2	101	−1.0
1.30	97.1	93.5	95.4	−2.5
5.20	94.1	94.1	94.1	+7.3

. The results were satisfactory, as the percentage (%) ranged from 85% to 103%.

Within- and between-run data on precision are reported in

Table 3. Precision data of n.10 repeated assays, either in the series or in different days, of citrate-fortified synthetic urines and internal quality controls (low and high levels).

	Concentration (mmol/L)	Calculated Concentration (mmol/L)		Relative Retention Time
		Within-Run Precision (CV%)	Between-Run Precision (CV%)	Within-Run Precision (CV%)	Between-Run Precision (CV%)
Citrate-fortified synthetic urine	0.08	11	-	0.73	-
	0.32	4.4	6.9	0.16	0.20
	1.30	5.5	6.9	0.20	0.18
	5.20	5.6	9.7	0.21	0.17
IQC level	1.06 (low level)	6.4	8.3	0.27	0.86
	5.93 (high level)	5.2	6.0	0.21	0.94

. The results show satisfactory repeatability, as the CV% is lower than 10% at all levels, both in spiked urines and internal quality controls, except for concentrations at the LLOQ level (CV = 11%).

3.2.6. Matrix Effect

Matrix effect values are given in Table 2. It was acceptable at all concentrations, allowing the correct determination of the analyte.

3.2.7. Urinary Citrate Stability

Citrate in pooled urine was stable at +4 °C using any preservative over the 30-day storing period. Without preservative, the citrate concentration was stable for 4 days, disappearing afterward within ten days (see

Table 4. Urinary citrate stability, 4 °C storage.

Sample	Preservative Type	Preservative Concentration	Day 1	Day 2	Day 4	Day 10	Day 30	CV (%)
1	Hibitane (chlorexidine gluconate 20%)	2 mL/L of urine	1.72	1.70	1.69	1.65	1.70	1.4
2	HCl 37%	10 mL/L of urine	1.59	1.63	1.64	1.63	1.57	1.7
3	CH3COOH	10 mL/L of urine	1.65	1.72	1.73	1.75	1.75	2.9
4	CCl3COOH	10 g/L of urine	1.62	1.64	1.74	1.63	1.64	3.1
5	No preservative	-	1.75	1.73	1.78	0.07	0.04	73.9

). Our results are in contrast with data reported by Saude et al., who observed only a slight decrease (−7%) in citrate concentration of raw female urine samples stored at +4 °C for 4 weeks [40].

3.3. Application to Real Samples

Citrate-to-creatinine ratios from males and females were compared for both 24 h and 2 h samples. Female values were significantly higher (unpaired t-test: p < 0.05 for both 24 h and 2 h samples). These differences disappeared when net citrate excretions were compared.

Comparing 24 h and 2 h citrate-to-creatinine ratios in individual subjects (paired t-test) failed to find significant differences. The two sets of data showed a close correlation (p = 0.146; R² = 0.733) (

Table 5. Urinary citrate concentration in real samples from healthy subjects.

Group	Age (Years)	Citrate/Creatinine 24 h (mol/mol)	Citrate Excretion 24 h (mmol/24 h)	Citrate/Creatinine 2 h (mol/mol)
All (n = 16)	41 ± 9 (28–58)	0.25 ± 0.12 (0.11–0.53)	3.32 ± 1.40 (1.59–6.93)	0.23 ± 0.12 (0.10–0.53)
Males (n = 9)	42 ± 10 (28–58)	0.19 ± 0.08 * (0.11–0.30)	2.89 ± 1.14 ‡ (1.59–5.05)	0.17 ± 0.07 † (0.10–0.32)
Females (n = 7)	41 ± 8 (31–51)	0.33 ± 0.12 * (0.21–0.53)	3.87 ± 1.59 ‡ (2.08–6.93)	0.30 ± 0.13 † (0.20–0.53)

X medium ± 1 SD; range in parentheses; *^,†,‡ p < 0.05, unpaired t-test.

).

4. Discussion

We aimed to develop a simple, sensitive, and rugged method, useful for high throughput. Among the current analytical techniques, IC appears to be the most adequate in fulfilling our requirements. Although this technique has already been widely used for citrate determination, the availability of new separators could improve its performance.

We tested different separators and Dionex IonPacTM Fast Anion IIIA performed better in terms of column efficiency. Specifically, it was observed that Dionex IonPacTM Fast Anion IIIA easily separated citrate from other polycarboxylic acids, and particularly from isocitrate, its main interferer.

The following necessary step was the choice of an appropriate exogenous substance to be used as an IS. Among the eligible ones, trimetaphosphate was chosen because of its chromatographic behavior and its stability. It is remarkable that up to now, no other published studies about citrate determination with internal standardization were found, except for a recent LC–MS/MS method [35].

This fully validated method is currently operating in our laboratory for monitoring citraturia in patients with calcium nephrolithiasis, and in line with previous works, the urine analysis from healthy subjects showed a higher ratio of citric acid rather than creatinine in females [41,42].

5. Conclusions

The method described here is suitable for the laboratory routines and is effective in detecting hypocitraturia (LLOQ = 0.08 mmol/L), which is the main concern for renal stone laboratories when undertaking citrate determination.

It is easy, cost-effective, and requires little maintenance. The chromatographic run is completed within twenty-five minutes and the sample preparation only requires a simple dilution with an aqueous IS solution. Internal standardization allows us to improve the precision and to prevent any variations in injection volumes.

The validated method appears adequate to measure the real concentration of this analyte in patients’ urine samples. High sensitivity referring to the relatively high citrate concentration in urine allows this test to be performed using only some fractions of a microliter of urine, which ensures no system perturbations, a long column life, and, as a consequence, cost-effectiveness. It has shown easy suitability in our laboratory where it is currently adopted to analyze over 1000 urine samples per year.