Search

ObjectiveTo determine the diagnostic accuracy of serological tests for coronavirus disease-2019 (covid-19).DesignSystematic review and meta-analysis.Data sourcesMedline, bioRxiv, and medRxiv from 1 January to 30 April 2020, using subject headings or subheadings combined with text words for the concepts of covid-19 and serological tests for covid-19.Eligibility criteria and data analysisEligible studies measured sensitivity or specificity, or both of a covid-19 serological test compared with a reference standard of viral culture or reverse transcriptase polymerase chain reaction. Studies were excluded with fewer than five participants or samples. Risk of bias was assessed using quality assessment of diagnostic accuracy studies 2 (QUADAS-2). Pooled sensitivity and specificity were estimated using random effects bivariate meta-analyses.Main outcome measuresThe primary outcome was overall sensitivity and specificity, stratified by method of serological testing (enzyme linked immunosorbent assays (ELISAs), lateral flow immunoassays (LFIAs), or chemiluminescent immunoassays (CLIAs)) and immunoglobulin class (IgG, IgM, or both). Secondary outcomes were stratum specific sensitivity and specificity within subgroups defined by study or participant characteristics, including time since symptom onset.Results5016 references were identified and 40 studies included. 49 risk of bias assessments were carried out (one for each population and method evaluated). High risk of patient selection bias was found in 98% (48/49) of assessments and high or unclear risk of bias from performance or interpretation of the serological test in 73% (36/49). Only 10% (4/40) of studies included outpatients. Only two studies evaluated tests at the point of care. For each method of testing, pooled sensitivity and specificity were not associated with the immunoglobulin class measured. The pooled sensitivity of ELISAs measuring IgG or IgM was 84.3% (95% confidence interval 75.6% to 90.9%), of LFIAs was 66.0% (49.3% to 79.3%), and of CLIAs was 97.8% (46.2% to 100%). In all analyses, pooled sensitivity was lower for LFIAs, the potential point-of-care method. Pooled specificities ranged from 96.6% to 99.7%. Of the samples used for estimating specificity, 83% (10 465/12 547) were from populations tested before the epidemic or not suspected of having covid-19. Among LFIAs, pooled sensitivity of commercial kits (65.0%, 49.0% to 78.2%) was lower than that of non-commercial tests (88.2%, 83.6% to 91.3%). Heterogeneity was seen in all analyses. Sensitivity was higher at least three weeks after symptom onset (ranging from 69.9% to 98.9%) compared with within the first week (from 13.4% to 50.3%).ConclusionHigher quality clinical studies assessing the diagnostic accuracy of serological tests for covid-19 are urgently needed. Currently, available evidence does not support the continued use of existing point-of-care serological tests.Study registrationPROSPERO CRD42020179452.

Most sudden deaths in sport are caused by cardiovascular conditions. The cardiovascular benefits of exercise are well established, and epidemiology studies suggest that long term exercise programmes may reduce the risk of sudden death. Increasing leisure time and facilities promote sports participation at all ages. A few people are at risk of serious arrhythmia or sudden death with exercise. The cause of death varies with the age of participants; congenital structural abnormalities occur in younger age groups and coronary artery disease in older age groups. Identifying such abnormalities makes prevention possible. Sudden death in sport is uncommon, with an incidence of 2 cases per 100 000 subject years. Five in 100 000 athletes have a condition that might predispose them to serious cardiac problems, and of those at risk 10% (1 in 200 000) may die suddenly or unexpectedly. Considerable controversy exists about the cost effectiveness of screening all young people by examination before participation. Alternative views suggest targeting people with a family history of sudden death or premature coronary artery disease and educating all participants to see their doctors about even minor warning symptoms. Some of the cardiac causes of sudden death will be suggested by symptoms and signs on clinical examination. The importance of others may be established only by detailed cardiological investigations including electrocardiography, chest radiography, echocardiography, doppler cardiography, and, rarely, cardiac catheterisation. Exercise testing is of major value and 24-48 hour tape electrocardiographic recordings may be required. Potentially serious symptoms such as syncope, pre-syncope, palpitation, chest pain, and undue dyspnoea should stimulate taking a detailed clinical history, making a detailed examination, and referring for cardiological investigation if appropriate. A family history of cardiac abnormality would suggest a hereditary basis. Cardiac causes of sudden death Coronary artery disease Hypertrophic cardiomyopathy Idiopathic concentric left ventricular hypertrophy Congenital anomalies …

OBJECTIVE--To determine the ability of a sensitive one step urine test to detect human chorionic gonadotrophin in women with suspected complications of early pregnancy. DESIGN--Test on women presenting to accident and emergency department with gynaecological problems over six months. Results were validated using a quantitative assay for human chorionic gonadotrophin in serum and urine. SETTING--Accident and emergency department and gynaecology wards of a university teaching hospital. SUBJECTS--130 unselected women. MAIN OUTCOME MEASURES--Detection of human chorionic gonadotrophin by one step test, presence of ectopic pregnancy, and results of quantitative analysis of chorionic gonadotrophin in serum and urine. RESULTS--79 women had a positive urine test result and 51 a negative result. All 12 women with ectopic pregnancy had a positive test result, although urinary concentration varied from 191 IU/l to 47,800 IU/l. Only one woman, who had a faintly positive result, was found not to be pregnant on subsequent examination. The sensitivity and negative predictive values of the urine test were 100% respectively. 33 women were sent home from the accident and emergency department with normal clinical findings after a negative urine test result. All these women had undetectable concentrations of chorionic gonadotrophin in matched samples of urine and serum. CONCLUSIONS--A simple, rapid one step test for chorionic gonadotrophin should be available for the initial evaluation of emergency gynaecological problems. The additional cost of the test is offset by not admitting those patients whose clinical findings are normal and who have a negative urine test result and by reducing the number of women requiring quantitative assays of chorionic gonadotrophin.

OBJECTIVE--To assess the need for formal psychotherapeutic intervention in children attending a children's haemophilia clinic after some of them had been diagnosed as positive for HIV. DESIGN--Comparison of haemophiliac children with matched control groups of diabetic and healthy children. SETTING--The West of Scotland Children's Haemophilia Centre, Glasgow. PATIENTS--43 Children aged 3 to 16 years with mild, moderate, and severe clotting disorders were matched with control groups of 46 diabetic children and 42 physically healthy children. INTERVENTIONS--Parents of children aged 3-5 years were interviewed with the behaviour screening questionnaire. Children aged 6 to 16 were assessed by parental and teacher report using standardised questionnaires and self report using a computerised depression inventory. All were scored numerically according to the presence of symptoms of emotional and behavioural problems. MAIN OUTCOME MEASURES--The groups were compared for mean scores on each rating device and for number of children achieving scores within the pathological range. RESULTS--In the 6-16 age group five haemophiliac children, five diabetic children, and three healthy children scored in the pathological range on the parent questionnaire, as did two, three, and five respectively on the teacher questionnaire and four, four, and eight on the depression inventory. There was no significant difference across the three groups. Analysis of mean scores similarly showed no significant difference across groups. In contrast, the single measure used for younger children showed an increase in behavioural difficulties among the diabetic children. CONCLUSION--Haemophiliac children attending the West of Scotland Centre were no more disturbed than their diabetic or healthy peers despite the identification of HIV infection within the clinic and the widespread adverse publicity associated with AIDS and HIV infection.

Pioneer in sports science, and doctor to the Scottish football team Stewart Hillis had been actively involved in the preparation for the 2014 Commonwealth Games in Glasgow. He was diagnosed as having mesothelioma in May and continued to work till mid-June. He died two days before the opening ceremony. Stewart was born in Clydebank, the son of a foreman at John Brown’s shipyard, two years after the town was devastated in the Blitz during the second world war. He was educated in Clydebank and at Glasgow University. During his early postgraduate training with Gavin Shaw, he became interested in cardiology and in 1971 was appointed registrar in the University of Glasgow’s cardiology department with Professor T D V Lawrie. His career progressed rapidly. He showed considerable talent in invasive investigations, and that became a distinguishing feature of his career. After spending a year at Vanderbilt University in Nashville, Tennessee, he was appointed consultant cardiologist to Glasgow’s Stobhill Hospital in 1977. It was there that he began pivotal studies in relation to intracoronary and intravenous thrombolysis. He converted the back area of the critical care unit into a coronary angiographic facility and set up a 24/7 service that allowed assessment of artery …

The question “Should the NHS work at weekends as it does during the week?”1 implies that how we work during the week is the gold standard. It is not. The NHS should be working to deliver the care patients need, when needed, as close to home as possible, and …

Martin Matthew Whittet was born in Glasgow. His secondary education was at Glasgow High School, where his father taught art. His medical education was at Glasgow University Medical School, from which he graduated in 1942. After full registration, he was appointed in 1943 to Glasgow Royal Mental …

We were astounded to read that the Department of Health plans to ban doctors from wearing white coats because of the risk of cross infection.1 2 The package describes a “bare below the elbows” dress code (short sleeves, no wrist watch, no jewellery, no ties when carrying out clinical activity). What is the evidence for this? It is impossible to …

Laughter really is the best medicine Next month France begins year long celebrations of the 500th anniversary of the birth of Francois Rabelais. To readers of the BMJ he is probably best known as a rumbustious writer of ribald tales during which whole armies are drowned in urine issuing from one of his giant heroes, ingenious but coarse solutions are found to intimate problems (for example, the use of a gosling's neck as an arsewipe in Pantagruel (chapter 28),1 and “Gargantuan” feasts take place, marked by gluttony and bawdy jokes. (Gargantua (chapter 13)).1 This vulgar Rabelais, in all senses of the word, can indeed be found in his works; but there is another, wise, medical, and Christian Rabelais who is found there too. Born in the Loire valley in 1483 or 1494, Rabelais completed his medical training at the celebrated school of Montpellier, taking his MB in 1530 and his doctorate seven years later. But before settling on medicine as a career he had been successively a Franciscan friar, possibly a law student, a Benedictine monk, and a secular priest.2 3 He was thus unusually erudite, even for his day, with a strong enthusiasm for the new learning of the …

Former professor of public health Edinburgh (b Glasgow 1922; q St Andrews 1951; DPH, FRCP, FFCM), died from a heart attack on 30 April 2002. Stuart Morrison faced a …

Last Saturday the envelope containing my decree absolute landed on the mat. Two and a half years ago, after 19 years of marriage and the production of three children, my partner, a doctor, thought that he could do better elsewhere and left. But the pill which has at times been harder to swallow than the divorce wrangling is the elegant disdain with which I have been cast aside by some of his colleagues. I had been around for a long time—not in any stunningly important capacity, but always available to take telephone messages, entertain colleagues, attend functions, soothe irate private patients. Yet within an amazingly brief period the new partner was introduced on to the scene and to certain people I automatically became yesterday's news, an Orwellian unperson. > What turns intelligent members of a profession into blank eyed strangers? These are people with whom we socialised as a couple for years. People who came regularly—over eight, 10, 12 years—to our home to receive our hospitality, which generally entailed …

Editor—The lesson of the week by Bridger et al on deaths from low dose paracetamol poisoning was most welcome but probably titled incorrectly.1 None of the patients in the cases described took low doses of paracetamol. What did happen was that the current national guidelines for treating patients after poisoning with paracetamol failed to protect these patients. We suspect that many doctors will have read these cases and dismissed them as poor management, assuming, for example, that the doctors concerned failed to establish the correct timing of the drug ingestion. We fear that complacency will persist with the belief that deaths will not happen if the current guidelines are adhered to properly. One of the cases was managed in our hospital. The timing and circumstances of the poisoning seemed to be and still do seem to be clear cut. We know the time at which the paracetamol was purchased from a local shop and the time immediately after at which help was sought. These times concur with the history that was given. The concentrations of paracetamol at four hours were 22% below the national standard treatment line. We believe that the nomogram should allow for a safety margin of 22%. We agree that there are many possible explanations for our patient’s death. Our patient may have been particularly susceptible to paracetamol but equally may have taken a sequential overdose. The key point is that a treatment strategy must allow a margin of safety that allows for some degree of inaccuracy in the history or an individual patient’s susceptibility to paracetamol. We are from the hospital in the south west referred to in the article that has changed its treatment protocol to a lower treatment line. Since making this change in 1994 we have treated around an extra 60 patients each year with acetylcysteine who would not have been treated using the national guidelines. Of these extra patients treated, one third were considered to be at normal risk and two thirds at high risk of hepatotoxicity. We cannot know if any of these patients would otherwise have developed liver failure. In 1994 we were indebted to Williams for reviewing the modifications that we had made to our local guidelines. These local guidelines are contrary to those in the British National Formulary and of the National Poisons Information Service. We therefore may be vulnerable to medicolegal problems if any patient has a reaction to acetylcysteine. We welcome the article by Bridger et al and strongly advocate that the national guidelines be changed to those now used in Truro. Executive action is clearly needed to review current practice and make appropriate changes to the national recommendations.

Editor—Moore et al concluded in their study that topical non-steroidal anti-inflammatory drugs are better than placebo in acute and chronic musculoskeletal pain.1 Does this help to define their role in practice? While we accept that these agents are better than placebo (this is considered during drug licensing), this does not help to decide how clinically effective they are. To define a therapeutic role for these preparations2 we need comparisons with other interventions, such as oral non-steroidal anti-inflammatory drugs, topical rubefacients, or paracetamol. Of the 87 trials reviewed by Moore et al only five compared topical with oral non-steroidal anti-inflammatory drugs, and none had adequate design or power to enable comparison. We are also concerned about the methodology used in this review. Moore et al conclude that the trials reviewed were of a good quality. We recently reviewed this area and found that many trials had poor methodology, low numbers of patients, and short durations of treatment.3 The quality analysis used by Moore et al is simplistic and may not fully evaluate the quality of the studies. We do, however, recognise the need to use some objective quantification of quality. Secondly, combining results from different studies of patients with varied musculoskeletal conditions may introduce errors, and a reliable meta-analysis of these agents could be difficult.4 One of the reasons for this is that the heterogeneity of conditions being studied makes it difficult to compare one study with another. We are not sure whether using the random effects model compensates sufficiently. Thirdly, the authors were unable to eliminate positive publication bias. As drug companies were asked to volunteer their trial data, this bias may extend to unpublished studies. The small number of trials volunteered by drug companies reinforces this concern. While the authors concede that further comparisons of topical with oral non-steroidal anti-inflammatory drugs are required, their conclusion that topical agents are effective and safe is not supported by their paper. Furthermore, it gives no guidance on when these drugs should be used. Not only may this review give the wrong messages to prescribers but it may be used inappropriately to promote non-evidence based prescribing. Until data are published that compare these agents with other alternatives—such as paracetamol, other much cheaper rubefacients, and interventions such as rest, ice, compression, and elevation—their therapeutic role remains unclear. Without this evidence it is difficult to justify a large proportion of the £33 million spent on these drugs over the past year (Prescription Pricing Authority, data on file, Oct 1997).

EDITOR,--In their paper on the London helicopter emergency medical service J P Nicholls and colleagues, of the University of Sheffield's medical care research unit, mention several concerns1 that we expressed at the time that the unit published its first report on the service to the Department of Health in 1994.2 The helicopter is part of an integrated specialist trauma service, and it is unfortunate that the unit looked at parts of the service separately. In doing so the unit isolated the effect of the helicopter from the effect of the hospital. We …

An average of 10-5 E.C.G.s were recorded weekly in a health centre used by 32 general practitioners serving a population of almost 65,000. The main indication for an E.C.G. was chest pain (73%). 47% of the E.C.G.s were abnormal. A change in clinical diagnosis occurred in 28% of cases and in patient management in 16%. A significant number of these changes were unwarranted, however. It is recommended that the E.C.G.s should be recorded by suitably trained nurses and reported by a specially trained general practitioner. Further education of general practitioners in the clinical use of the E.C.G. is required.

The photomotogram (P.M.G.) of the Achilles tendon reflex was studied in 26 patients with hypothermia (rectal temperature 33.3 degrees C or less), 10 of whom also had myxoedema (serum protein bound iodine 2.8 mug/100 ml or less). No reflex could be elicited in eight (31%) of these patients, including three of those with myxoedema. Hypothermia increases both the contraction and the relaxation times of the reflex, the relaxation phase being particularly prolonged in those with myxoedema. In those patients from whom the reflex was elicited the ratio of the contraction time to the "half-relaxation time" in the P.M.G. was less than unity in six of the seven with myxoedema, and considerably greater than unity in eight of the 11 (73%) who were euthyroid. Thus, analysis of the Achilles tendon reflex P.M.G. correctly predicted the thyroid status in 14 of the 18 hypothermic patients in whom the Achilles tendon reflex was present (78%). The wider use of this rapid test of thyroid function would allow a more rational use of thyroid hormones in hypothermic patients and so lead to a better assessment of their value.