Determination of Serum Aldosterone
Determination of Serum Aldosterone
Total CVs ('interday CVs' or 'between-run CVs' in the cited references) for LC-MS/MS aldosterone assays are reported as 11.8%–9.6% at levels of 28–1415 pmol/l, 7.5%–4.3% at levels of 97–993 pmol/l and 9.4%–2.7% at levels of 69–1109 pmol/l. In this regard, the present assay is comparable with or better than previous reports.
Based on our observed total coefficient of variations of ≤5% and absolute bias of ≤3.1%, the expected total error (TE), based on the formula TE=|bias|+1.65·σ, where σ represents the (im)precision, is ≤11.4%. This is well within the desirable TE specification of 36.7% based on biological variation.
As a matter of practicality, we have defined our lower reportable limit at LOQ (S:N=10:1) of 50 pmol/l (CV≈10%), a concentration well below any diagnostic threshold used in routine clinical practice. Our LOD (S:N=3:1) is <22 pmol/l, which is slightly better than the 28 pmol/l value observed using APPI.
The proportional and constant biases (LC-MS/MS=1.17×RIA−63.9 pmol/l) observed between the present method and RIA are much smaller in magnitude than those observed between individual RIA methods and between an LLE-based LC-MS/MS method and Diasorin's RIA. However, because method-dependent bias is a significant issue for this analyte, labs implementing this or any new aldosterone method should undertake a reference interval study and appropriately adjust the threshold of aldosterone to plasma renin activity/mass used to define a positive screen for PA. Likewise, the effect of bias on the absolute aldosterone concentration used to define normal suppression in provocative studies requires consideration. With respect to specimen type, biases between EDTA and serum do not reach statistical significance when the CIs of the slope and intercept are considered.
To our knowledge, this is the only LC-MS/MS aldosterone method for which interferences from related exogenous and endogenous steroids have been systematically evaluated. Of the compounds investigated, only 18-hydroxycorticosterone has the potential to hamper integration (ie, area under the chromatographic curve) at very high concentrations. This is due to a peak showing the same MRM transitions as aldosterone but eluting a short time later at ~6.8 min. In practice, we have not observed such a problem; if present, it would be readily apparent.
The strength of this method lies in its simplicity and affordability. Online or offline SPE adds significant 'disposables' cost and requires additional equipment. Although generally more efficient in ionisation and less affected by ion suppression, an APPI source creates an added expenditure and necessitates an additional LC pump for the toluene dopant. The present method can be run on a basic ABSCIEX API-5000 system with low operational costs and no extra hardware.
This method has some weaknesses. The sample volume employed is larger than the 200 μl reported elsewhere; we opted for 500 μl volume in order to ensure the low-end sensitivity required for the definitive diagnosis of PA using provocative studies.
Additionally, 10 min between injections is longer than the 7 min reported for protein precipitation and SPE. Attempts to decrease this time by steepening the solvent ramp, increasing flow, using a shorter 50 mm column or a C8 column were hampered by both failure to resolve aldosterone from a potential interference (eluting at 6.8 min and present in all patient samples) and inadequate column reconditioning between injections.
The advent of convenient, reliable LC-MS/MS based for aldosterone assays represents a major improvement in PA screening and diagnosis. Although many biological assays such as growth hormone, HbA1c and testosterone have seen major efforts in standardisation and harmonisation, no such undertaking has been made for aldosterone nor is there any standardised reference material available. Notwithstanding, the popularisation of LC-MS/MS assays for steroid analysis should facilitate the standardisation of both assays and clinical diagnostic protocols.
Discussion
Total CVs ('interday CVs' or 'between-run CVs' in the cited references) for LC-MS/MS aldosterone assays are reported as 11.8%–9.6% at levels of 28–1415 pmol/l, 7.5%–4.3% at levels of 97–993 pmol/l and 9.4%–2.7% at levels of 69–1109 pmol/l. In this regard, the present assay is comparable with or better than previous reports.
Based on our observed total coefficient of variations of ≤5% and absolute bias of ≤3.1%, the expected total error (TE), based on the formula TE=|bias|+1.65·σ, where σ represents the (im)precision, is ≤11.4%. This is well within the desirable TE specification of 36.7% based on biological variation.
As a matter of practicality, we have defined our lower reportable limit at LOQ (S:N=10:1) of 50 pmol/l (CV≈10%), a concentration well below any diagnostic threshold used in routine clinical practice. Our LOD (S:N=3:1) is <22 pmol/l, which is slightly better than the 28 pmol/l value observed using APPI.
The proportional and constant biases (LC-MS/MS=1.17×RIA−63.9 pmol/l) observed between the present method and RIA are much smaller in magnitude than those observed between individual RIA methods and between an LLE-based LC-MS/MS method and Diasorin's RIA. However, because method-dependent bias is a significant issue for this analyte, labs implementing this or any new aldosterone method should undertake a reference interval study and appropriately adjust the threshold of aldosterone to plasma renin activity/mass used to define a positive screen for PA. Likewise, the effect of bias on the absolute aldosterone concentration used to define normal suppression in provocative studies requires consideration. With respect to specimen type, biases between EDTA and serum do not reach statistical significance when the CIs of the slope and intercept are considered.
To our knowledge, this is the only LC-MS/MS aldosterone method for which interferences from related exogenous and endogenous steroids have been systematically evaluated. Of the compounds investigated, only 18-hydroxycorticosterone has the potential to hamper integration (ie, area under the chromatographic curve) at very high concentrations. This is due to a peak showing the same MRM transitions as aldosterone but eluting a short time later at ~6.8 min. In practice, we have not observed such a problem; if present, it would be readily apparent.
The strength of this method lies in its simplicity and affordability. Online or offline SPE adds significant 'disposables' cost and requires additional equipment. Although generally more efficient in ionisation and less affected by ion suppression, an APPI source creates an added expenditure and necessitates an additional LC pump for the toluene dopant. The present method can be run on a basic ABSCIEX API-5000 system with low operational costs and no extra hardware.
This method has some weaknesses. The sample volume employed is larger than the 200 μl reported elsewhere; we opted for 500 μl volume in order to ensure the low-end sensitivity required for the definitive diagnosis of PA using provocative studies.
Additionally, 10 min between injections is longer than the 7 min reported for protein precipitation and SPE. Attempts to decrease this time by steepening the solvent ramp, increasing flow, using a shorter 50 mm column or a C8 column were hampered by both failure to resolve aldosterone from a potential interference (eluting at 6.8 min and present in all patient samples) and inadequate column reconditioning between injections.
The advent of convenient, reliable LC-MS/MS based for aldosterone assays represents a major improvement in PA screening and diagnosis. Although many biological assays such as growth hormone, HbA1c and testosterone have seen major efforts in standardisation and harmonisation, no such undertaking has been made for aldosterone nor is there any standardised reference material available. Notwithstanding, the popularisation of LC-MS/MS assays for steroid analysis should facilitate the standardisation of both assays and clinical diagnostic protocols.
