Development, evaluation and validation of machine learning algorithms to detect atypical and asymptomatic presentations of Covid-19 in hospital practice.

BACKGROUND: Diagnostic methods for Covid-19 have improved, both in speed and availability. Because of atypical and asymptomatic carriage of the virus and nosocomial spread within institutions, timely diagnosis remains a challenge. Machine learning models trained on blood test results have shown promise in identifying cases of Covid-19. AIMS: To train and validate a machine learning model capable of differentiating Covid-19 positive from negative patients using routine blood tests and assess the model's accuracy against atypical and asymptomatic presentations. DESIGN AND METHODS: We conducted a retrospective analysis of medical admissions to our institution during March and April 2020. Participants were categorized into Covid-19 positive or negative groups based on clinical, radiological features or nasopharyngeal swab. A machine learning model was trained on laboratory parameters and validated for accuracy, sensitivity and specificity and externally validated at an unconnected establishment. RESULTS: An Ensemble Bagged Tree model was trained on data collected from 405 patients (212 Covid-19 positive) producing an accuracy of 81.79% (95% confidence interval (CI) 77.53-85.55%), the sensitivity of 85.85% (CI 80.42-90.24%) and specificity of 76.65% (CI 69.49-82.84%). Accuracy was preserved for atypical and asymptomatic subgroups. Using an external data set for 226 patients (141 Covid-19 positive) accuracy of 76.82% (CI 70.87-82.08%), sensitivity of 78.38% (CI 70.87-84.72%) and specificity of 74.12% (CI 63.48-83.01%) was achieved. CONCLUSION: A machine learning model using routine laboratory parameters can detect atypical and asymptomatic presentations of Covid-19 and might be an adjunct to existing screening measures.

Diuretic lounge and the impact on hospital admissions for treatment of decompensated heart failure.

BACKGROUND: Heart failure is a prevalent condition associated with frequent and costly hospital admissions. Hospitalizations are primarily related to worsening fluid retention and often require admission for decongestion with intravenous diuretics. OBJECTIVE: To assess the safety of an outpatient intravenous diuresis service for heart failure patients, and its impact on emergency admissions and the cost of treatment. METHODS: We conducted a prospective observational cohort registry study on patients referred to the diuretic lounge at our acute hospital between May 2017 and April 2018. RESULTS: We analysed 245 patients treated in the diuretic lounge, of which 190 (77.6%) avoided hospitalization or any adverse events during the 60 days of follow up (77.6% vs. 22.4%; P < 0.001). The diuretic lounge service resulted in a significant decrease in emergency heart failure admissions compared to the previous 12 months (823 vs. 715 per annum; 68.6 ± 10.1 vs. 59.6 ± 14 per month; P = 0.04), and a numerical reduction in readmission rates (17.3% vs. 16.2%). The 13.1% decrease in admissions lead to financial savings of £315 497 per annum and £2921 per admission avoided. During the same time period, at the other acute hospital site in our trust, where no diuretic lounge service is available, the number of admissions did not significantly change (457 vs. 450 per annum; 37.5 ± 7.0 vs. 38.1 ± 7.6 per month; P = 0.81). CONCLUSION: Ambulatory administration of intravenous diuretics reduces emergency admissions and is a safe and cost-effective alternative to treat acute decomposition in heart failure patients.

Registry report of the prevalence of ECG abnormalities and their relation to patient characteristics in an asymptomatic population.

BACKGROUND: Cardiovascular disease is the leading cause of morbidity and mortality in developed countries. Many patients do not experience symptoms before a first major cardiovascular event. Resting electrocardiogram (ECG) may help identify asymptomatic individuals with a high risk of cardiovascular disease. OBJECTIVE: We sought to determine the prevalence of ECG abnormalities in an apparently symptom-free adult population within a prospective registry. METHODS: The registry consisted of 4739 consecutive apparently healthy individuals [age 62.8 ± 6.2 (SD) years; 54% female], referred by their General Practitioners between 2009 and 2013. Patient demographics, and clinical data were obtained, alongside blood tests and a resting ECG. RESULTS: ECG abnormalities were present in 1509 (31.8%) subjects and were commoner with increasing age (F = 0.161, p = 0.01), systolic (F = 0.134, p = 0.01) and diastolic (F = 0.44, p = 0.01) blood pressure and waist circumference (F = 0.53, p = 0.01). Left ventricular hypertrophy (LVH) was the most common abnormality (n = 281) and was positively associated with systolic (F = 0.12, p = 0.01) and diastolic blood pressure (F = 0.99, p = 0.01) and male gender (X2 = 60.5, p < 0.01). All ECG abnormalities (except for LVH) were associated with an increasing age, while right bundle branch block (F = 0.041, p = 0.01) and atrial fibrillation (n = 29; F = 0.036, p = 0.05) were associated with the presence of diabetes mellitus (n = 211). Only left bundle branch block (n = 50) was associated with angina (F = 0.05, p = 0.01). CONCLUSIONS: Unrecognized cardiac abnormalities are common in middle-aged men and women with no overt symptoms. ECG offers the potential to identify these abnormalities and provide earlier intervention and treatment, and possibly improve cardiovascular outcome.

Renal artery stenosis: a common and important problem in patients with peripheral vascular disease.

OBJECTIVE: To study the prevalence, severity, vascular risk factors, and clinical implications of renal artery stenosis in patients with peripheral vascular disease. DESIGN: Cross-sectional study of consecutive patients who were electively referred from the department of vascular surgery for lower limb digital subtraction angiography. SETTING: St. George's Hospital, London, United Kingdom. SUBJECTS: One hundred twenty-seven patients presenting with intermittent claudication or lower limb ischemic ulceration. MAIN OUTCOME MEASURES: Prevalence and clinical importance of renal artery stenosis in patients with peripheral vascular disease adjusted for the confounding effects of age and hypertension. RESULTS: Of the 127 patients, 57 (44.9%) had renal artery disease, of whom 22 (17.3%) had mild disease, 20 (15.7%) had severe disease, and 15 (11.8%) had bilateral renal artery stenosis. There was a significant positive relationship between the presence of renal artery stenosis and the severity of peripheral vascular disease (p = 0.00015). The risk of having renal artery stenosis was nearly four times greater in those with three to four vessels affected and nearly seven times greater in those with five or more vessels affected as compared with those with a milder degree of peripheral vascular disease (one or two vessels affected). This association persisted when the confounding effect of age and hypertension was accounted for. Six patients (31.6%) with renal artery stenosis who underwent revascularization for peripheral vascular disease died during the early postoperative period of cardiac or renal complications. None of the patients with normal renal arteries who had similar surgery developed postoperative complications (p = 0.005). CONCLUSIONS: Renal artery stenosis is a common independent feature in patients with peripheral vascular disease, and its prevalence increases with the increasing severity of the peripheral vascular disease. The postoperative risk following revascularization for peripheral vascular disease appears to be greater in those patients with renal artery stenosis. All patients studied with digital subtraction angiography for peripheral vascular disease should have an aortic flush performed to image the renal arteries. This information may be used to identify those patients likely to develop postoperative complications during peripheral revascularization.

Echocardiography overestimates left ventricular mass: a comparative study with magnetic resonance imaging in patients with hypertension.

OBJECTIVE: To compare measurement of left ventricular mass (LVM) by M-mode echocardiography and magnetic resonance imaging (MRI) in hypertensive subjects. DESIGN: A prospective study. SUBJECTS: Twenty-four untreated hypertensive patients [19 men and five women, aged 51 +/- 2 (mean +/- SEM) years, supine blood pressure 159/101 +/- 3/1 mmHg]. SETTING: The Blood Pressure Unit, St Georges Hospital Medical School and Magnetic Resonance Unit, Royal Brompton National Heart and Lung Hospital, London. MAIN OUTCOME MEASURES: LVM estimated both by M-mode echocardiography and by MRI. RESULTS: Using three standard M-mode formulae, widely different values of LVM were obtained with echocardiography [American Society of Echocardiography (ASE) 319 +/- 21 g, Penn 273 +/- 19 g. Teichholz 191 +/- 11 g]. By MRI, the LVM was 232 +/- 11 g. The differences between MRI and echocardiography could not be explained in terms of the timing of measurements in the cardiac cycle. When single-slice MRI measurements at the appropriate level were applied to the ASE and Penn formulae, the LVM was again overestimated. CONCLUSION: Our study has shown major differences in LVM estimated using methods based on one-dimensional (echocardiography) compared with three-dimensional (MRI) data. These differences seem to be largely the result of the geometrical assumptions on which M-mode measurements are based. Our findings have important clinical implications for the assessment of the severity and response to treatment of left ventricular hypertrophy in hypertensive patients.

Non-invasive screening for renal artery stenosis with ultrasound contrast enhancement.

OBJECTIVE: Our aim was to evaluate duplex ultrasound imaging in the identification of renal artery stenosis using a new technique to enhance the recorded Doppler signal. DESIGN: Colour Doppler studies of interlobar renal arteries were performed before and after enhancement using an intravenous contrast of galactose microparticle suspension containing microbubbles (Levovist, Schering) in patients with angiographically confirmed renal artery stenosis. SETTING: Blood Pressure Unit, St. George's Hospital Medical School, and Department of Radiology. The Middlesex Hospital, London, UK. PARTICIPANTS: Twenty-one consecutive hypertensive patients in whom the diagnosis of renal artery stenosis was made on digital subtraction angiography. MAIN OUTCOME MEASURES: The diagnosis of haemodynamically significant renal artery stenosis (> or = 60% on angiography). RESULTS: With Levovist, there was a 20 db increase in the Doppler intensity and, as a result, intrarenal signals were much more clearly delineated and distinct spectral waveforms were obtained from all but one kidney, which was occluded. Significant associations were found between the degree of stenosis (as assessed by angiography) and the following Doppler parameters: diastolic velocity (F = 7.6; P < 0.01), acceleration time (F = 33.5, < 0.0001), peak systolic velocity (F = 37.7, P < 0.0001) and acceleration (F = 60.0; P < 0.0001). Without enhancement, there were five false-positive and two false-negative examinations (sensitivity 85%; specificity 79%) using the acceleration cut-off value of 3.5 m/s2 to identify haemodynamically significant renal artery stenosis (> or = 60% on angiography). After contrast enhancement, there were only three false-positive and one false-negative examinations (sensitivity 94% and specificity of 88%) using the acceleration cut-off value of 3.75 m/s2 and the examination time was reduced by approximately half (sensitivity and specificity of 90% using the acceleration cut-off value of 3.5 m/s2). CONCLUSIONS: Our results suggest that renal duplex scanning using contrast enhancement is a promising new non-invasive technique in screening patients with suspected renal artery stenosis. Contrast enhancement produces more reproducible spectral waveforms, improves accuracy and halves the examination time.

"Apparent" heart failure: a syndrome caused by renal artery stenoses.

OBJECTIVE: To report on renal artery stenosis presenting as congestive heart failure. DESIGN: Case series. SETTING: Tertiary referral centre. PATIENTS: Nine hypertensive subjects (five male, four female) seen in the blood pressure unit, St George's Hospital, between 1991 and 1997 with clinical signs and symptoms of congestive cardiac failure but without overt coronary or valvar heart disease. Mean (SEM) age was 67 (3) years. Eight patients had renal artery revascularisation with percutaneous angioplasty and one had surgery. RESULTS: Renal revascularisation was followed by a large fall in blood pressure from 191/94 (7/3) to 150/75 (8/5) mm Hg two days after intervention (p < 0.01). There was also a large natriuresis and weight reduction. One week after revascularisation there was a mean loss in weight of 3.8 (0.6) kg. The largest fall in weight was seen in those patients with stenosis in a single functioning kidney. Furthermore, plasma atrial natriuretic factor fell from 120 (28) to 48 (9) pg/ml (p < 0.05; n = 6; normal value = 8.6 (0.8) pg/ml), and serum creatinine fell from 200 (37) to 140 (11) micromol/l (p < 0. 025). The clinical signs and symptoms of heart failure resolved and the diuretics were then withdrawn in all patients. On long term follow up, patients remained free from symptoms and signs of heart failure and the blood pressure was better controlled. CONCLUSIONS: In hypertensive patients with symptoms and signs of congestive heart failure who do not have obvious ischaemic or valvar heart disease, renal artery stenosis should be considered as a possible underlying cause. Relief of the stenosis can result in resolution of the apparent heart failure.

Double-blind crossover study of once daily nifedipine coat core in essential hypertension.

Nineteen patients with essential hypertension on regular treatment with nifedipine tablets 20 mg twice daily and whose DBP was < 95 mmHg on at least two occasions two weeks apart were entered in a double-blind randomised crossover study of three weeks treatment with nifedipine coat core (new formulation) either as 30 mg one daily or as 60 mg once daily dose. BP and plasma nifedipine levels were measured at 24, two, four and six hours after dosing. The pattern of BP response to both doses was similar over the 24h period. However, a greater BP lowering effect was achieved with 60 mg compared with 30 mg. The BP lowering effect of both doses was less at 24h after the last dose compared with peak effect. Plasma nifedipine levels were significantly associated with the BP lowering effect in the group as a whole (i.e. the higher the nifedipine levels, the lower the BP) and were significantly less at 24 hours compared with peak. Nifedipine in the coat core formulation is effective in lowering BP in patients with essential hypertension. The 60 mg dose is more effective than the 30 mg dose and induces higher nifedipine levels which are associated with greater BP lowering effect. However, the maximum BP lowering effect is not maintained up to 24 hours.

Platelet sodium-hydrogen exchanger activity and left ventricular mass.

INTRODUCTION: The sodium-hydrogen exchanger (NHE) is integral to the processes that facilitate cell growth and may contribute to the development of left ventricular hypertrophy. The aim of this study was to examine the relationship between platelet sodium-hydrogen exchanger activity and left ventricular mass index (LVMI). METHODS: Twenty male untreated Caucasians (mean age +/- s.d.: 48 +/- 13; body mass index: 29 +/- 4 kg/m(2)) with a wide range of blood pressures were studied (mean BP: 152 +/- 22/93 +/- 15 mm Hg; range: 115-190/61-117 mm Hg). Sodium-hydrogen exchanger activity was determined as the rate of sodium-dependent recovery of intracellular pH in isolated platelets loaded with BCECF and acidified to pH 6.25 using nigericin. LVMI was calculated from measurements made by M-mode echocardiography. Associations between continuous variables were examined using parametric tests. RESULTS: The mean rate of pHi recovery was 0.15 +/- 0.03 dpHi/s (range: 0.09-0.21). Mean LVMI was 120 +/- 32 g/m(2) (range: 56-178) and was not significantly correlated with either systolic (r = 0.39, P = 0.09) or diastolic blood pressure (r = 0.27, P = 0.3). Platelet NHE activity was not significantly correlated with LVMI (r = 0.06; P = 0.8). Platelet NHE activity was not significantly different between subjects with (n = 7 with LVMI >131 g/m(2)) and without left ventricular hypertrophy (n = 13). CONCLUSIONS: The results of this study show that platelet NHE activity is not significantly correlated with LVMI. These data contrast with previously described correlations of LVMI with exchanger activity measured in leucocytes and erythrocytes; and indicate that the relationship between LVMI and exchanger activity may be specific to the cell type in which exchanger activity is measured.

Platelet 3H-imipramine binding in pregnancy and the puerperium.

Platelet 3H-imipramine binding was examined in a cross-sectional study of 70 Caucasian women in pregnancy and the early post-partum period, and in 23 nonpregnant women of childbearing age. Mood was also assessed in the pregnancy and post-partum sample. No significant differences in number of binding sites (Bmax) were found, but an increase in the equilibrium dissociation constant (Kd) was demonstrated at 5-7 days post-partum.