Biomedical subjects
Publications and source records attributed to S J Sharp.
Narcolepsy, although frequently beginning in childhood, is usually diagnosed in young adults. The diagnostic symptoms of narcolepsy are usually less typical in the young child, and sleep studies have not been standardized. We present a 12-month-old child with symptoms typical for narcolepsy who shows improvement with nonpharmacologic narcolepsy therapy.
Three approaches for estimating the relationship between treatment effect and underlying risk in a meta-analysis of clinical trials have recently been published. The aim of each is to overcome the bias inherent in conventional regressions of treatment effect on control group risk, which arises from the measurement error in the observed control group risks in different trials. Here we describe these published approaches, and compare them with respect to their underlying models and methods of implementation. The underlying model for one of them is shown to be seriously flawed, while the other two are both statistically more appropriate than the conventional approaches, and differ from each other in only two assumptions. Both may be implemented using the Gibbs sampling algorithm in BUGS, and are exemplified here using a meta-analysis of mortality and bleeding data in trials of sclerotherapy for patients with cirrhosis. One approach is developed further; for the illustrative example considered, it is shown to be robust to different choices of prior distributions for the model parameters, and to the assumption of a linear relationship on a log-odds scale. It can also be used to estimate the level of underlying risk (and its standard error) at which the treatment effect crosses from benefit to harm, and other trial-level covariates may be included in the model as confounders. The BUGS code is provided in an Appendix, to enable applied researchers to perform the various analyses described.
BACKGROUND: As prophylaxis against influenza in families, amantadine and rimantadine have had inconsistent effectiveness, partly because of the transmission of drug-resistant variants from treated index patients. We performed a double-blind, placebo-controlled study of inhaled zanamivir for the treatment and prevention of influenza in families. METHODS: We enrolled families (with two to five members and at least one child who was five years of age or older) before the 1998-1999 influenza season. If an influenza-like illness developed in one member, the family was randomly assigned to receive either inhaled zanamivir or placebo. The family member with the index illness was treated with either 10 mg of inhaled zanamivir (163 subjects) or placebo (158) twice a day for 5 days, and the other family members received either 10 mg of zanamivir (414 subjects) or placebo (423) once a day as prophylaxis for 10 days. The primary end point was the proportion of families in which at least one household contact had symptomatic, laboratory-confirmed influenza. RESULTS: The proportion of families with at least one initially healthy household contact in whom influenza developed was smaller in the zanamivir group than in the placebo group (4 percent vs. 19 percent, P<0.001); the difference represented a 79 percent reduction in the proportion of families with at least one affected contact. Zanamivir provided protection against both influenza A and influenza B. A neuraminidase-inhibition assay and sequencing of the neuraminidase and hemagglutinin genes revealed no zanamivir-resistant variants. Among the subjects with index cases of laboratory-confirmed influenza, the median duration of symptoms was 2.5 days shorter in the zanamivir group than in the placebo group (5.0 vs. 7.5 days, P=0.01). Zanamivir was well tolerated. CONCLUSIONS: When combined with the treatment of index cases, prophylactic treatment of family members with once-daily inhaled zanamivir is well tolerated and prevents the development of influenza. In this study there was no evidence of the emergence of resistant influenza variants.
BACKGROUND: Previous work has suggested that the long-term regular use of inhaled beta2-agonist bronchodilators might lead to a deterioration in asthma control. The aim of TRUST (The Regular Use of Salbutamol Trial) was to study the effects of regular use of inhaled salbutamol, the most widely prescribed bronchodilator in the UK, on the control of asthma. METHODS: A randomised, double-blind, placebo-controlled trial was undertaken in 983 patients with asthma being treated at least twice a week with short-acting beta2-agonist, alone or in combination with inhaled steroids (2 mg or less) daily. Patients were aged 18 years and over and were recruited from 115 general practices in the UK. 90% (881) of the patients used inhaled corticosteroid therapy, and all patients continued to use their usual inhaled beta2-agonist for symptomatic relief. Patients were randomised to receive 400 microg salbutamol or matched placebo via a Diskhaler four times per day for 12 months. The primary outcome measure was rate of exacerbations of asthma, with criteria based on data from diary cards completed daily by each patient, treatment with additional corticosteroids, or both. FINDINGS: There were no differences in the annual rate, timing, or duration of exacerbations between the two groups. The mean morning peak expiratory flow was similar for the two groups. The mean evening peak expiratory flow (p<0.001) and the diurnal variation (p<0.001) were greater, and the use of rescue bronchodilator was less (p<0.001), in the group receiving regular salbutamol. INTERPRETATION: There was no evidence that regular use of inhaled salbutamol 400 microg four times daily for a year increased the exacerbation rate of asthma in the population studied.
BACKGROUND: Zanamivir is a neuraminidase inhibitor, the first of a new class of drugs with potent, specific antiviral activity against influenza A and B. Administration by inhalation results in direct delivery to the respiratory tract, the principal site of viral replication. OBJECTIVE: This study was undertaken to determine the effectiveness of zanamivir on duration and resolution of influenza symptoms. METHODS: Using a method similar to that employed in amantadine treatment studies to obtain supporting evidence of efficacy for US Food and Drug Administration consideration, pooled data from 6 zanamivir phase II and III clinical trials involving primarily previously healthy adults were analyzed to categorize patients as accelerated resolvers (temperature <37.8 degrees C < or = 24 hours after dosing, plus a > or = 50% reduction in symptom score by 36 hours), early resolvers (temperature <37.8 degrees C < or = 36 hours after dosing), and nonaccelerated resolvers (febrile >36 hours after dosing). Patients in the accelerated and early categories were termed rapid resolvers; the others were slow resolvers. Patients recorded their symptom severity, temperature, ability to perform normal daily activities, and use of relief medication on a diary card. The primary end point of median time to alleviation of symptoms was also reexamined, with the additional requirement that patients were not taking relief medication when their symptoms were alleviated. This analysis was intended to control for the possible effect of relief medication on the primary end point. RESULTS: In the influenza-positive population (n = 1572), significantly more zanamivir-treated patients were rapid resolvers compared with those receiving placebo (807 [72%] vs 765 [64%], P < 0.001). Significant benefits of zanamivir treatment were observed in patients with a baseline temperature of > or = 37.8 degrees C (630 [68%] vs 595 [57%], P < 0.001) or > or = 38.3 degrees C (382 [67%] vs 365 [52%], P < 0.001) and in patients considered by their physi- cian to have severe symptoms at the start of therapy (252 [70%] vs 222 [63%], P = 0.02). Differences were even more apparent in patients with high-risk conditions (74% vs 53%, P = 0.014) and in those aged > or = 50 years (70% vs 54%, P = 0.005). Zanamivir treatment was also associated with a significant reduction in time to alleviation of symptoms with no use of relief medication. This was particularly noticeable in those aged > or = 50 years; time to alleviation was 13 days in patients receiving placebo and 6 days in patients receiving zanamivir (P < 0.001). In these trials, adverse events were reported at a similar frequency in patients receiving zanamivir and those receiving placebo. CONCLUSIONS: Zanamivir is more effective than placebo in patients with influenza at providing early symptom relief and a reduced duration of illness at a time when use of relief medication has ended. These benefits are seen across different patient groups but appear to be particularly marked in patients who are aged > or = 50 years, who have underlying illnesses, who are considered high risk, or who are more severely ill at the beginning of therapy.
Exploring the possible reasons for heterogeneity between studies is an important aspect of conducting a meta-analysis. This paper compares a number of methods which can be used to investigate whether a particular covariate, with a value defined for each study in the meta-analysis, explains any heterogeneity. The main example is from a meta-analysis of randomized trials of serum cholesterol reduction, in which the log-odds ratio for coronary events is related to the average extent of cholesterol reduction achieved in each trial. Different forms of weighted normal errors regression and random effects logistic regression are compared. These analyses quantify the extent to which heterogeneity is explained, as well as the effect of cholesterol reduction on the risk of coronary events. In a second example, the relationship between treatment effect estimates and their precision is examined, in order to assess the evidence for publication bias. We conclude that methods which allow for an additive component of residual heterogeneity should be used. In weighted regression, a restricted maximum likelihood estimator is appropriate, although a number of other estimators are also available. Methods which use the original form of the data explicitly, for example the binomial model for observed proportions rather than assuming normality of the log-odds ratios, are now computationally feasible. Although such methods are preferable in principle, they often give similar results in practice.
Many biological processes give outcome data which show a curvilinear association with time which tends to an asymptote. We show how autoregressive models can be used to describe this association within individual subjects. We also present a Bayesian approach implemented using statistical software, BUGS, to fit these models in a multi-level (hierarchical) setting that describes variation in the association between subjects. Peak expiratory flow data from a clinical trial involving subjects with asthma are used to illustrate the methods.
BACKGROUND: Percutaneous transluminal coronary angioplasty (PTCA) and coronary-artery bypass grafting (CABG) are both effective intervention strategies for patients with coronary heart disease. We report comparative long-term clinical and health-service cost findings for these interventions in the first Randomised Intervention Treatment of Angina (RITA-1) trial. METHODS: 1011 patients with coronary heart disease (45% single-vessel, 55% multivessel) were randomly assigned initial treatment strategies of PTCA or CABG. Information on clinical events, subsequent intervention, symptomatic status, exercise testing, and use of health-care resources is available for a median 6.5 years of follow-up. Analyses were by intention to treat. FINDINGS: The predefined primary endpoint of death or nonfatal myocardial infarction occurred in 87 (17%) PTCA-group patients and 80 (16%) CABG-group patients (p=0.64). Similarly, there was no significant treatment difference in deaths alone (39 PTCA, 45 CABG), of which 46% were cardiac related. In both groups, the risk of cardiac death or myocardial infarction was more than five times higher in the first year than in subsequent years of follow-up. 26% of patients assigned PTCA subsequently also had CABG, and a further 19% required additional nonrandomised PTCA. Most of these reinterventions occurred within a year of randomisation, and from 3 years onwards the reintervention rate averaged 4% per year. In the CABG group the reintervention rate averaged 2% per year. The prevalence of angina was consistently higher in the PTCA group, with an absolute average 10% excess compared with the CABG group (p<0.001). Total health-service costs over 5 years showed no significant difference between initial strategies of PTCA and CABG (mean difference pounds sterling 426 [95% Cl -pounds sterling 383 to pounds sterling 1235]; p=0.30). The clinical and cost comparisons showed similar patterns for patients with single-vessel and multivessel disease. INTERPRETATION: Initial strategies of PTCA and CABG led to similar long-term results in terms of survival and avoidance of myocardial infarction and to similar long-term health-care costs. Choice of approach, therefore, rests on weighing the more invasive nature of CABG against the greater risk of recurrent angina and reintervention over many years after PTCA.
In a meta-analysis of clinical trials, an important issue is whether the treatment benefit varies according to the underlying risk of the patients in the different trials. The usual naive analyses employed to investigate this question use either the observed risk of events in the control groups, or the average risk in the control and treatment groups, as a measure of underlying risk. These analyses are flawed and can produce seriously misleading results. We show how their biases depend on three components of variability, the within-trial and between-trial variances of the control group risks, and the between-trial variance of the treatment effects. We propose a Bayesian solution to the problem which can be carried out using the BUGS implementation of Gibbs sampling. The analysis is illustrated for a meta-analysis of bleeding and mortality data in trials of sclerotherapy for patients with cirrhosis, and the results contrasted with those from the naive approaches. Comparisons with other methods recently proposed for this problem are also made. We conclude that the Bayesian solution presented in this paper is not only more appropriate than other proposed methods, but is also sufficiently easy to implement that it can be used by applied researchers undertaking meta-analyses.
The severity of cutaneous leishmaniasis may be determined by host immunity, parasite virulence, and host or vector behavior. We performed a multivariate analysis to identify the main causes of the variability in clinical symptoms, response to treatment, and parasite isolation rate among Peruvian patients. The effect of host immunity was demonstrated first by the finding that secondary infections induced smaller lesions associated with a lower parasite isolation rate than did primary infections and, second, by the finding of fewer lesions in older patients. Phenotypic differences between parasite populations were suggested by the observation that the mean scar size and number varied between villages: patients had more scars in villages where the transmission rates were higher. Human behavior probably determined the site of lesions on the body, since most lesions in the cooler South were on the head, whereas in the North, lesions were equally frequent on the extremities. In addition, older patients, who were more likely infected through occupational exposure, had fewer head lesions. Geographic variation in the pattern of exposure to sandflies indicates that uta control strategies should be region specific.
Following evidence of an association between low serum cholesterol and cancer from several prospective studies, this paper concentrates on individual time trends in serial measurements of serum cholesterol before a cancer-related death. The Framingham Heart Study contains repeated measurements of cholesterol at approximately 2-year intervals in 5,209 subjects, of whom 539 died from cancer during 30 years of follow-up. We quantify (1) the change in serum total cholesterol level before cancer death, and (2) the association between fall in serum total cholesterol and odds of cancer death. The mean fall in serum total cholesterol in the 4- to 6-year period before cancer death is 8.06 (95% confidence interval = 4.58-11.54) mg per dl, with some evidence of lowered cholesterol before that period. This pattern is corroborated by evidence of a substantially increased odds of cancer death if a large fall in cholesterol occurs over any 4- to 6-year period. We suggest that these time trends can plausibly be attributed to the effects of prevalent cancer on lowering serum cholesterol; our findings add weight to the argument that the low cholesterol-cancer mortality relation does not arise because of any causal contribution of low serum total cholesterol to the risk of cancer.
In meta-analyses of clinical trials comparing a treated group with a control group it has been common to ask whether the treatment benefit varies according to the underlying risk of the patients in the different trials, with the hope of defining which patients would benefit most and which least from medical interventions. The usual analysis used to investigate this issue, however, which uses the observed proportions of events in the control groups of the trials as a measure of the underlying risk, is flawed and produces seriously misleading results. This arises through a bias due to regression to the mean and will be particularly acute in meta-analyses which include some small trials or in which the variability in the true underlying risks across trials is small. Approaches which previously have been thought to be more appropriate are to substitute the average proportion of events in the control and treated groups as the measure of underlying risk or to plot the proportion of events in the treated group against that in the control group (L'Abbé plot). However, these are still subject to bias in most circumstances. Because of the potentially seriously flawed conclusions that can result from such analyses, they should be replaced either by statistically appropriate (but more complex) approaches or, preferably, by analyses which investigate the dependence of the treatment effect on measured baseline characteristics of the patients in each trial.
Differential gene expression occurs in the process of development, maintenance, injury, and death of unicellular as well as complex organisms. Differentially expressed genes are usually identified by comparing steady-state mRNA concentrations. Electronic subtraction (ES), subtractive hybridization (SH), and differential display (DD) are methods commonly used for this purpose. A rigorous examination has been lacking and therefore quantitative aspects of these methods remain speculative. We compare these methods by identifying a total of 58 unique differentially expressed mRNAs within the same experimental system (HeLa cells treated with interferon-gamma). ES yields digital, reusable data that quantitated steady-state mRNA concentrations but only identified abundant mRNAs (seven were identified), which represent a small fraction of the total number of differentially expressed mRNAs. SH and DD identified abundant and rare mRNAs (33 and 23 unique mRNAs respectively) with redundancy. The redundancy is mRNA abundance-dependent for SH and primer-dependent for DD. We conclude that DD is the method of choice because it identifies mRNAs independent of prevalence, uses small amounts of RNA, identifies increases and decreases of mRNA steady-state levels simultaneously, and has rapid output.
The RNA polymerase III transcription initiation factor TFIIIB contains the TATA-box-binding protein (TBP) and polymerase III-specific TBP-associated factors (TAFs). Previous studies have shown that DNA oligonucleotides containing the consensus TATA-box sequence inhibit polymerase III transcription, implying that the DNA binding domain of TBP is exposed in TFIIIB. We have investigated the TATA-box DNA binding activity of Xenopus TFIIIB, using transcription inhibition assays and a gel mobility shift assay. Gel shift competition assays with mutant and nonspecific DNAs demonstrate the specificity of the TFIIIB-TATA box DNA complex. The apparent dissociation constant for this protein-DNA interaction is approximately 0.4 nM, similar to the affinity of yeast TBP for the same sequence. TFIIIB transcriptional activity and TATA-box binding activity cofractionate during a series of four ion-exchange chromatographic steps, and reconstituted transcription reactions demonstrate that the TATA-box DNA-protein complex contains TFIIIB TAF activity. Polypeptides with apparent molecular masses of 75 and 92 kDa are associated with TBP in this complex. These polypeptides were renatured after elution from sodium dodecyl sulfate-gels and tested individually and in combination for TFIIIB TAF activity. Recombinant TBP along with protein fractions containing the 75- and 92-kDa polypeptides were sufficient to reconstitute TFIIIB transcriptional activity and DNA binding activity, suggesting that Xenopus TFIIIB is composed of TBP along with these polypeptides.
The ability of antibodies to neutralize diverse primary isolates of human immunodeficiency virus-type 1 in vitro has been questioned, with implications for the likely efficacy of vaccines. A recombinant human antibody to envelope glycoprotein gp120 was generated and used to show that primary isolates are not refractory to antibody neutralization. The recombinant antibody neutralized more than 75 percent of the primary isolates tested at concentrations that could be achieved by passive immunization, for example, to interrupt maternal-fetal transmission of virus. The broad specificity and efficacy of the antibody implies the conservation of a structural feature on gp120, which could be important in vaccine design.
Transcription of eukaryotic tRNA genes is dependent on the A- and B-Box internal control regions (ICRs) and the upstream transcription modulatory region. The B-Box ICR spans nucleotides 52 to 62 and directs the primary binding of transcription factor C as the first step in the formation of a transcription complex. The conservation of the sequence of the B-Box in all tRNA species reflects its importance in both the expression of the gene and the processing, structure and function of the gene product. In order to identify the nucleotides essential to the promoter function of the B-Box ICR, site-directed mutagenesis was used to generate all the possible single point mutations at positions 52 to 58, 61 and 62 of a Drosophila melanogaster tRNA(Arg) gene. The effect of these mutations on gene transcription was evaluated using in vitro transcription and template exclusion competition assays. Optimal activity was displayed by the wild type tDNA(Arg) B-Box sequence but several other sequences supported in vitro transcription at wild type levels. The majority of mutants, however, showed lower efficiency in the in vitro transcription assay. Of the single point mutations, those at positions 53, 55, and 56 had a critical effect on gene function in Drosophila and HeLa transcription extracts and transcription factor interaction most likely requires base contacts at these positions. Since the effect of several of the point mutations cannot be explained in terms of possible major or minor groove contributions the possibility is raised that local DNA geometry also is an important determinant in specifying B-Box function.
We have studied the mechanism by which 5'-flanking sequences modulate the in vitro transcription of eucaryotic tRNA genes. Using deletion and linker substitution mutagenesis, we have found that the 5'-flanking sequences responsible for the different in vitro transcription levels of three Drosophila tRNA5Asn genes are contained within a discrete region centered 22 nucleotides upstream from the transcription initiation site. In conjunction with the A-box intragenic control region, this upstream transcription-modulatory region functions in the selection mechanism for the site of transcription initiation. Since the transcription-modulatory region directs the position of the start site and the actual sequence of the transcription-modulatory region determines the level of tRNAAsn gene transcription, the possibility is raised that the transcription-modulatory region directs a transcription initiation event similar to open complex formation at procaryotic promoters.
Linker-scanning (LS) mutations were constructed spanning the length of the Drosophila melanogaster 5S RNA gene. In vitro transcription analysis of the LS 5S DNAs revealed five transcription control regions. One control region essential for transcription initiation was identified in the 5'-flanking sequence. The major sequence determinants of this upstream promoter region were located between coordinates -39 and -26 (-30 region), but important sequences extended to the transcription start site at position 1. Since mutations in the upstream promoter did not alter the ability of 5S DNA to sequester transcription factors into a stable transcription complex, it appears that this control region involved the interaction of RNA polymerase III. Active 5S DNA transcription additionally required the four intragenic control regions (ICRs) located between coordinates 3 and 18 (ICR I), 37 and 44 (ICR II), 48 and 61 (ICR III), and 78 and 98 (ICR IV). LS mutations in each ICR decreased the ability of 5S DNA to sequester transcription factors. ICR III, ICR IV, and the spacer sequence between were similar in sequence and position to the determinant elements of the multipartite ICR of Xenopus 5S DNA. The importance of ICR III and ICR IV in transcription initiation and in sequestering transcription factors suggests the presence of an activity in D. melanogaster similar to transcription factor TFIIIA of Xenopus laevis and HeLa cells. Transcription initiation of Drosophila 5S DNA was not eliminated by LS mutations in the spacer region even though these mutations reduced the ability of the TFIIIA-like activity to bind. The previously unidentified control regions ICR I and ICR II appear to be important for the interaction of a transcription factor activity, or multiple-factor activities, distinct from the TFIIIA-like activity. The interaction of this activity with ICR I directed the selection of the transcription start site.