Background To evaluate the accuracy and precision of a method for serum aldosterone using supported liquid extraction (SLE) for sample preparation instead of the more conventional liquid-liquid extraction (LLE) approach.
Methods Two independently developed SLE-based LC-MS/MS methods for serum aldosterone (sample volumes 250 μl and 300 μl respectively) were compared to a modification of a previously reported LLE approach (sample volume 500 μl) in two method comparisons (n=75 and n=97). SLE analyses were performed at two separate centres. Precision was evaluated at a single site using human pools in head-to-head comparison between SLE and LLE. All analyses were performed on the ABSCIEX API-5000 LC-MS/MS system.
Results At four increasing pool concentrations spanning 67-1060 pmol/l, total precision for SLE ranged from 6.8-4.1% compared with 11.1-4.3% for LLE. Differences did not reach statistical significance except at the lowest concentration where SLE was superior. Pasing Bablok regression comparisons were SLE=0.96×LLE-5.8 pmol/l (R2=0.985) and SLE=0.96×LLE-0.44 pmol/l (R2=0.969).
Conclusions For analysis of serum aldosterone on the ABSCIEX API-5000, SLE affords a smaller sample volume while maintaining the accuracy and precision performance of LLE. By avoiding specimen vortexing, SLE also allows for greater automation in the sample preparation.
Statistics from Altmetric.com
Although liquid–liquid extraction (LLE) based liquid chromatography tandem mass spectrometry (LC–MS/MS) methods for serum aldosterone1 ,2 offer a degree of simplicity in sample preparation over methods using solid phase extraction,3 ,4 the requirements for sample volume tend to be higher (500 μl1 ,2 vs 200 μl).3 ,4 This represents an inconvenience, particularly when dealing with paediatric specimens or those for which repeat analysis happens to be required. In addition, the specimen vortexing/freezing and supernatant sampling required by LLE can become taxing in high-throughput laboratories.
Supported liquid extraction (SLE) is a newer method of sample preparation available in a 96-well format.5 SLE cartridges and plates are filled with diatomaceous earth having particle size and packing density optimised for biological sample preparation. SLE sample preparation is faster and more convenient than LLE as there is no vortexing or delay for phase separation required.
We have evaluated SLE-based sample preparation for aldosterone by LC–MS/MS at two sites against an LLE-based method.
LLE analyses were performed at Lab-1 (St Paul's Hospital) using a modification of a method recently published in this journal2 in which specimens were reconstituted with 125 μl of 20 : 80 MeOH : H2O instead of 50 : 50 MeOH : H2O.
For SLE sample preparation at Lab-1, 50 μl of 55 400 pmol/l d7-aldosterone (IsoSciences, King of Prussia, Pennsylvania, USA) internal standard solution was added to 250 μl of gel-free serum in a polyethylene 96-well plate followed by brief vortexing. The mixture was then applied to a Biotage (Uppsala, Sweden) ISOLUTE SLE+ 400 μl plate with a concomitant pulse of vacuum and given 5 min to allow specimen absorbance. Two aliquots of 900 μl methyl tert-butyl ether were then applied to each well with a pulse of vacuum. The solvent was subsequently allowed to flow passively for 5 min followed by a 60 s application of vacuum to complete elution. Drydown, reconstitution and LC–MS/MS analysis followed as per the LLE method.
SLE sample preparation at Lab-2 (LabCorp) differed in that 300 μl of serum was mixed with 100 μl of 55 400 pmol/l of internal standard, of which 300 μl was loaded to the same SLE product from Biotage. A single aliquot of 900 μl methyl tert-butyl ether was used for elution. Drydown, reconstitution and LC–MS/MS analysis (proprietary) followed.
Both laboratories used the ABSCIEX (Foster City, California, USA) API-5000 LC–MS/MS system. All pipetting was performed by the Hamilton (Reno, Nevada, USA) STARlet liquid handler at Lab-1. (online supplementary figure 1) and the Tecan (Männedorf, Switzerland) Evo 150 liquid handler at Lab-2.
Precision was investigated using human serum pools at nominal concentrations of 67, 178, 409 and 1060 pmol/l (as measured by the LLE method) by a variation of the Clinical Laboratory Science Institute EP-5A2 document:6 five analyses of each level per day, 4 days. The precision of SLE and LLE was compared head to head at Lab-1.
Gel-free serum specimens for LLE versus SLE method comparisons were obtained from the primary aldosteronism screening programme of Lab-1. LLE and SLE were compared within Lab-1 with 75 gel-free serum specimens ranging in concentration from 21.8 to 3180 pmol/l (by LLE–LC–MS/MS). Comparison between Lab-1 and Lab-2 was performed using 97 gel-free serum specimens ranging in concentration from 20.6 to 1200 pmol/l (by LLE–LC–MS/MS). Statistical analysis was performed using R V.2.13.1.
Results of the precision studies performed at Lab-1 for both SLE and LLE are presented in table 1.
Two samples with particularly high values of 3180 and 2420 pmol/l were excluded from the comparisons within Lab-1 as they unduly improved the coefficient of determination (R2=0.992 vs R2=0.985 when excluded).7 The relationship between SLE (Lab-1) and LLE (Lab-1) is shown in figure 1: SLE=0.96×LLE − 5.8 pmol/l (Passing Bablok, CI slope 0.93 to 0.98, CI intercept −10.5 to −1.81, R2=0.985). The relationship between LLE (Lab-1) and SLE (Lab-2) is shown in figure 2: SLE=0.96×LLE − 0.44 pmol/l (Passing Bablok, CI slope 0.94 to 0.99, CI intercept −6.23 to 5.74, R2=0.969).
From the perspective of total precision, we have found the SLE method to be comparable to LLE in head-to-head comparison and superior at the lowest level (F-test, p=0.01, Bonferroni adjusted α=0.0125) despite requiring half the sample volume. This allows the use of sample volumes closer to those of solid phase extraction methods.3 ,4 This approach also avoids the use of atmospheric pressure photoionisation,3 which requires an additional ion source and a toluene dopant infusion.
The SLE methods of Lab-1 and Lab-2 both demonstrated small proportional and constant biases compared with LLE. Both forms of bias reached statistical significance in the comparison within Lab-1, as did the proportional bias in the comparison between Lab-1 and Lab-2. However, considering LLE as the ‘reference’ method, both (im)precision and bias of the SLE method are within Westgard's desirable specifications of 14.7% and 12.4%, based on biological variation.8
At Lab-1, signal-to-noise (S/N) ratios, as calculated using Analyst software, were used to evaluate approximate limits of quantitation (S/N=10 : 1) and limits of detection (S/N=3 : 1) on specimens having a low endogenous level. Six specimens of approximately 50 pmol/l and five of approximately 25 pmol/l were evaluated for S/N based on the limits of quantitation (50 pmol/l) and limits of detection (<22 pmol/) of our reported method.2 At median concentrations of 53 pmol/l and 27 pmol/l, the median S/N was 20 : 1 and 14 : 1, respectively. This improvement cannot necessarily be attributed to SLE because the present study used a reconstitution solvent mixture matched to the initial mixture of the mobile phases (20 : 80 MeOH/H2O vs 50 : 50 MeOH/H2O), which also improved S/N.
We propose that SLE affords a very simple alternative to LLE, which performs comparably to LLE but allows a substantial reduction in sample volume, with comparable precision across the analytical range and superior precision in the concentration range of 50–75 pmol/l.
SLE offers an alternative approach for sample preparation of serum aldosterone by LC-MS/MS.
In this context, SLE affords the use of a smaller sample volume than LLE while allowing increased automation without loss of analytical performance.
The authors would like to acknowledge Ms Norine Freedman, Ms Karen Green and Ms Lynn Coleman for performing LLE-based aldosterone analysis at Lab-1.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online figure 1
Contributors JGVDG developed the aldosterone method for St Paul's Hospital and helped write the manuscript. DTH gave medical oversight to the development of the aldosterone method at St Paul's Hospital, helped write the manuscript and prepared the figures. MC and RPG developed the aldosterone method at LabCorp and helped write the manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.