key: cord-0687800-qo3iaxi0
authors: Wiegand, M.; Halsall, D. J.; Cowan, S. L.; Taylor, K.; Goudie, R. J.; Preller, J.; Gurnell, M.
title: Unquantifiably low aldosterone concentrations are prevalent in hospitalised COVID-19 patients but may not be revealed by chemiluminescent immunoassay
date: 2022-03-01
journal: nan
DOI: 10.1101/2022.02.28.22271645
sha: 5ee507cb30569e7786ddbf90fde9c6570dab2db4
doc_id: 687800
cord_uid: qo3iaxi0

Objective: Previous studies have reported conflicting findings regarding aldosterone levels in patients hospitalised with COVID-19. We therefore used the gold-standard technique of liquid chromatography tandem mass-spectrometry (LCMSMS) to address this uncertainty. Design: All patients admitted to Cambridge University Hospitals with COVID-19 between March 10, 2020 and May 13, 2021, and in whom a stored blood sample was available for analysis, were eligible for inclusion. Methods: Aldosterone was measured by LCMSMS and by immunoassay; cortisol and renin were determined by immunoassay. Results: Using LCMSMS, aldosterone was below the limit of detection (<70 pmol/L) in 74 (58.7%) patients. Importantly, this finding was discordant with results obtained using a commonly employed clinical immunoassay (Liaison Diasorin), which over-estimated aldosterone compared to the LCMSMS assay (intercept 14.1 [95% CI -34.4 to 54.1] + slope 3.16 [95% CI 2.09 to 4.15] pmol/L). The magnitude of this discrepancy did not clearly correlate with markers of kidney or liver function. Solvent extraction prior to immunoassay improved the agreement between methods (intercept -14.9 [95% CI -31.9 to -4.3] and slope 1.0 [95% CI 0.89 to 1.02] pmol/L) suggesting the presence of a water-soluble metabolite causing interference in the direct immunoassay. We also replicated a previous finding that blood cortisol concentrations were often increased, with increased mortality in the group with serum cortisol levels >744 nmol/L (p=0.005). Conclusion: When measured by LCMSMS, aldosterone was found to be profoundly low in a significant proportion of patients with COVID-19 at the time of hospital admission. This has likely not been detected previously due to high levels of interference with immunoassays in patients with COVID-19, and this merits further prospective investigation.

As of February 2022 the United Kingdom has recorded in excess of 18.8 million cases of 48 COVID-19, with at least 161,000 cases resulting in death [1] . The speed of onset and severity of 49 the pandemic has spurred a coordinated response from the global biomedical community on a 50 scale not previously seen. This includes attempts to better understand the pathogenesis of the 51 disease, to identify factors that can be used to predict risk and disease trajectory in individual 52 patients, and to deliver preventative and curative interventions for the world's population. 53

Given the key role of angiotensin-converting enzyme 2 (ACE2) in facilitating entry of SARS-54

CoV-2 virus particles into the lung (alveolar epithelial type II cells), gastrointestinal tract (luminal 55 intestinal epithelial cells) and other tissues [2, 3], exploration of the potential effects on the 56 renin-angiotensin-aldosterone system (RAAS) is of interest in understanding the pathogenesis 57 of COVID-19. ACE2 inhibits RAAS activation by converting angiotensin II (AngII), to angiotensin 58 1-7 (Ang 1-7). Ang 1-7 exerts anti-inflammatory, anti-oxidative and vasodilatory effects via 59 binding to the Mas receptor [4] . AngII binds AngII receptor type 1 which then exerts pro-60 inflammatory, pro-oxidative and vasoconstrictive effects [5] . It may also contribute to pro-fibrotic 61 effects, hypercoagulability and immunothrombosis by inducing tissue factor and plasminogen 62 activator inhibitor-1 expression by endothelial cells. AngII further binds to the angiotensin Ⅰ 63 receptor on the adrenal glands, stimulating the release of the mineralocorticoid aldosterone. 64 SARS-CoV-2 has the potential to activate RAAS and the secretion of aldosterone, by preventing 65 this ACE2-Ang1-7 mediated RAAS inhibition. The uninhibited Ang II may then play a role in 66 the pathogenesis of the observed hypertension [6], inflammation, immunothrombosis and 67 possible fibrosis in . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) To the best of our knowledge, all published studies measuring aldosterone in COVID-19 76 patients have used non-extraction immunoassays. These methods lack specificity and are 77 prone to interference, for example the polar aldosterone metabolite Aldosterone-18-glucuronide 78 has been shown to cross react in a non-extraction assay commonly used in clinical laboratories 79 [15, 16] . This is more apparent in patients with renal failure as hydrophilic metabolites 80 accumulate. Mass spectrometric methods for aldosterone are now increasingly available in 81 clinical laboratories and do not suffer this interference. Whilst the clinical effectiveness of non-82 extraction immunoassays in the diagnosis of primary hyperaldosteronism is still contested [17], 83 mass spectrometric methods are metrologically superior and are more likely to represent the 84 biologically active aldosterone fraction. 85

During the pandemic we observed low aldosterone levels in a number of patients, which we had 86 not anticipated. Motivated by this observation, in this study we used a tandem mass 87 spectrometric method to estimate serum aldosterone concentration in patients admitted to 88 hospital with SARS-CoV-2 infection. We correlate these serum aldosterone results with clinical 89 outcomes, and compare results from the tandem mass spectrometric method with re-90 measurements using immunoassay methods. We also evaluate the association between high 91 cortisol concentrations and 28-day survival, as previously described by Tan et al [7] . 92 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) We made use of blood samples drawn from a biobank at Cambridge University Hospitals, which 99 was established early in the pandemic to store, where available, blood samples from patients 100 with COVID-19. We retrospectively measured aldosterone, cortisol and renin levels on the 101 available stored samples from the biobank. All patients with at least one stored sample from 102 around the time of their first positive SARS-CoV-2 result (within 72 hours) were included. Within 103 this time window, the earliest available sample was used. Patients were excluded if they had 104 received glucocorticoid or mineralocorticoid therapy (e.g. dexamethasone, fludrocortisone, 105 hydrocortisone, methylprednisolone, prednisolone) prior to collection of the blood sample. The association between aldosterone, cortisol and renin in paired samples was assessed by 115 linear regression, imputing the value 35 pmol/L for aldosterone values below the LLOQ of 70 116 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint pmol/L, which is statistically equivalent to assuming that these values are uniformly distributed 117 between 0 and the LLOQ [20] . 118

To assess the association between high cortisol concentrations (> 744 nmol/L) and outcomes 119

[7], a Kaplan-Meier estimate of mortality within 28-days of a positive SARS-CoV-2 test, stratified 120 by the first cortisol concentration following the positive test (within 72 hours), was used. Tan et 121 al. selected the threshold of 744 nmol/L as the value that maximises the log-rank statistic, i.e. 122 the difference between survival curves. The selection of the cortisol threshold that optimises the 123 difference between survival curves in our cohort was also replicated and the difference in 124 survival curves was assessed by a log-rank test. Similarly the difference in survival curves of 125 cohorts with high or low aldosterone concentrations was investigated. 126

To explore potential differences between aldosterone as determined by LCMSMS vs 127 immunoassay, aldosterone was re-measured in a subset of patients, where sample volume 128 permitted, using the Diasorin LIAISON® direct competitive chemiluminescent immunoassay 129 (CLIA) (LLOQ 1.91 ng/dL). Scatter plots, with Passing-Bablok regression lines, and Bland-130

Altmann plots were used to assess the agreement between the LCMSMS and CLIA methods. 131

To explore potential association between bias in the CLIA results with renal function, we 132 examined associations with estimated glomerular filtration rate (eGFR) and creatinine clearance 133 (Cockroft-Gault) using linear regression that is robust to outliers [21] . 134

Finally, the CLIA was repeated following solvent extraction of serum to remove potential 135 aqueous interference from patient blood samples that would not be detected by the LCMSMS 136 method. In brief, 350ul of patient sample was mixed with 2250ul of Methyl Tertiary Butyl Ether 137 by vortex for 15 minutes. The lower aqueous layer was frozen and the upper MTBE layer 138 decanted and evaporated to dryness at 60℃ under nitrogen flow. The sample was reconstituted 139 in 350ul of steroid free serum (DRG Instruments GmbH Marburg Germany). Extraction 140 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint efficiency was corrected for using the mean extraction recovery from a cohort of 15 anonymised 141 COVID-19 negative control subjects (data not shown). 142

Results 143 231 patients who tested positive for SARS-CoV-2 had at least one measurement of aldosterone, 144 cortisol or renin within 72 hours of their first positive SARS-CoV-2 test. 97 patients had received 145 glucocorticoid and/or mineralocorticoid therapy prior to the collection of the blood sample, 146 leaving 134 patients in the study cohort (126 with an aldosterone measurement available; 88 147 with measured cortisol; and 50 with measured renin). 148 80 patients (59%) were male and the median age was 64 years (IQR 46, 88). 15 patients 149 (11.2%) were admitted to the intensive care unit (ICU) and the in-hospital mortality was 13.4%. 150

Other baseline parameters and markers of severity are described in Table 1 . The patients in our 151 study had similar age, gender, ethnicity, BMI and severity markers to the overall population of 152 SARS-CoV-2 patients presenting to the hospital (eTable 1), although these patients had lower 153 interleukin-6 (median 8.1 vs 11.9); were less likely to be admitted to ICU (11.2% vs 18.9%); and 154 were more likely to have diabetes (32.8% vs 20.7%) or other endocrine disease (14.2% vs 155 8.7%). 156

The distribution of aldosterone, cortisol and renin is visualised by histograms in Figure 1 . The 157 aldosterone concentrations after the positive test in Figure 1A are remarkably low, with the 158 concentration of aldosterone below the LLOQ (70 pmol/L) in 58.7% of the patients (74 patients). 159

Analysis of cortisol levels found elevated levels, with 20.5% of patients having results greater 160 than 744 nmol/L, the level proposed by Tan el al to maximise the difference between patient 161 survival curves ( Figure 1B) . Figure 1C depicts the renin concentrations, which were 162 predominantly low despite the low aldosterone concentration. 163 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint There was evidence that higher aldosterone concentrations were associated with both higher 164 cortisol (beta=0.316, 95% CI 0.06 to 0.57, p=0.02) and higher renin concentrations (beta=0.062, 165 95% CI 0.025 to 0.099, p<0.00001), as shown in Figure 2 . Stratification of the cohort according 166 to whether aldosterone was above or below the LLOQ did not reveal any significant differences 167 in clinical characteristics (eTable 2), or in the survival curves (eFigure 2). There was no obvious 168 correlation between aldosterone level and renal function (eFigure 3). 169

The survival probability of patients with a high concentration of cortisol (above 744 nmol/L) was 170 lower than for those with a low concentration of cortisol (below 744 nmol/L). As shown in Figure  171 3, there was a marked difference in 28-day mortality, with 44% of patients dying in the high 172 cortisol group, compared to 11% of patients in the low cortisol group. While the 95% confidence 173 intervals (CI) around the survival curves overlap over the entire time frame, the log-rank test 174 comparison of the complete survival curves indicates a significant difference between the two 175 groups (p = 0.005). The average length of hospital stay in the high cortisol group (19.1 days) 176 was longer than in the low cortisol group (11.2 days). The cortisol threshold maximising the log-177 rank test for the difference of survival curves in our cohort was 801 nmol/L. 178

The considerable disparity between our LCMSMS aldosterone results and previously-published 179 studies [12, 11], which used immunoassay, is explained by Figure 4 . This shows that re- between eGFR and the discrepancy between the two assays was not demonstrable in this 231 cohort. 232

To our knowledge, this study is the first to report aldosterone levels in COVID-19 patients as 233 measured by LCMSMS, along with re-measurements using CLIA. A weakness of our study is 234 that, whilst it appears that our sample is representative of the broader cohort of hospitalised 235 COVID-19 patients in terms of clinical characteristics at admission, our sample is a convenience 236 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint sample of patients with suitable available samples in the hospital's biobank, rather than a 237 prospectively-collected cohort. This also meant we were unable to measure renin in all patients, 238 due to the more specific sample requirements. Nonetheless, we believe our findings to be 239 important and hope they will spur further research on the role of RAAS in In conclusion, these data demonstrate that aldosterone cannot be accurately estimated in 241 serum from patients with SARS-CoV-2 infection using direct competitive immunoassay due to 242 the presence of a water soluble interference. When measured using gold-standard LCMSMS, 243 serum aldosterone is found to be remarkably low in most patients with COVID-19. The 244 mechanism of this reduction remains obscure with no obvious correlation with glucocorticoid 245 status or kidney function. 246 247 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint

Uk dashboard to track COVID-19 cases in the United Kingdom

Blood tests, median

CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint 365 366 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 1, 2022 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 1, 2022 CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)The copyright holder for this preprint this version posted March 1, 2022. ; https://doi.org/10.1101/2022.02.28.22271645 doi: medRxiv preprint