Frequency of submicroscopic chromosomal aberrations in pregnancies without increased risk for structural chromosomal aberrations: systematic review and meta-analysis.

OBJECTIVE: To establish, based on a systematic literature review, the frequency of pathogenic submicroscopic chromosomal aberrations in fetuses that are not at increased risk for unbalanced structural chromosomal aberrations, with the aim of determining whether high-resolution testing for submicroscopic aberrations is beneficial in a general pregnant population. METHODS: EMBASE, PubMed, Web of Science and CENTRAL databases were searched systematically on 3 June 2016 for all relevant articles on the prevalence of pathogenic submicroscopic copy number variants (CNVs) in fetuses referred for prenatal invasive testing because of advanced maternal age (AMA) or parental anxiety (ANX). Relevant full-text articles were analyzed and the prevalence of submicroscopic CNVs was calculated based on the extracted data. Meta-analysis was conducted in a pooled cohort of 10 614 fetuses based on the 10 largest studies (n > 300) of a total of 19 that were relevant. RESULTS: Pooled estimate analysis indicated that 0.84% (95% CI, 0.55-1.30%) of fetuses that had invasive testing because of AMA/ANX carried a pathogenic clinically significant submicroscopic aberration. The onset/penetrance of submicroscopic findings was studied in 10 314 fetuses reported in eight papers that presented aberrant cases with all necessary details to allow assessment of the findings. The pooled estimates resulting from meta-analysis of the data indicated that an early-onset syndromic disorder was detected in 0.37% (95% CI, 0.27-0.52%) of cases, a susceptibility CNV was found in 0.30% (95% CI, 0.14-0.67%) and late-onset diseases were reported in 0.11% (95% CI, 0.05%-0.21%). The prevalence of early-onset syndromic disorders caused by a submicroscopic aberration was calculated to be 1:270. When the risk for submicroscopic aberrations is added to the individual risk for microscopic chromosomal aberrations, all pregnant women have a risk of higher than 1 in 180 for a relevant chromosomal aberration, and pregnant women under 36 years of age have a higher risk for submicroscopic pathogenic aberrations than for Down syndrome. CONCLUSION: This systematic review shows that a significant proportion of fetuses in a general pregnant population carry a submicroscopic pathogenic CNV. Based on these figures, all women should be informed on their individual risk for all pathogenic chromosomal aberrations and not only for common trisomies. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

Effects of epidural volume extension by saline injection on the efficacy and safety of intrathecal local anaesthetics: systematic review with meta-analysis, meta-regression and trial sequential analysis.

Epidural volume extension, a modification of combined spinal-epidural anaesthesia, involves the epidural injection of saline in order to increase the spread of drugs given intrathecally. Results from individual studies have so far been contradictory and we aimed to gather the available evidence for this technique. We performed a systematic literature search for randomised, controlled trials comparing epidural volume extension after spinal injection with a control group without epidural injection in patients undergoing surgery. Conventional meta-analyses, trial sequential analyses and meta-regression were performed, with the Grading of Recommendations on Assessment, Development and Evaluation (GRADE) approach used to express reliability of outcome estimates. We included 15 studies with 1177 participants. Meta-analyses for the primary outcomes, such as maximum sensory height (6 studies, 274 participants, mean difference (MD) (95%CI) -0.59 (-1.24 to 0.07) dermatomes, low-quality evidence) and hypotension (10 studies, 683 participants, risk ratio (95%CI) 0.84 (0.66-1.07), low-quality evidence), did not differ significantly between the two treatment arms, but trial sequential analysis suggested insufficient evidence to be certain of these findings. Meta-regression suggested a volume-dependent effect, with higher volumes causing a higher spread of intrathecal drugs and a higher incidence of hypotension. A sub-group analysis indicated a pronounced effect on motor block recovery time when a lower anaesthetic dose plus epidural volume extension was compared with a higher anaesthetic dose without epidural volume extension, the MD (95%CI) being -66.75 (-76.0 to -57.5) min, with trial sequential analysis suggesting the evidence was sufficient to draw this conclusion. In trials using the same anaesthetic mixture in the epidural volume extension and the control groups, motor block recovery time did not differ between groups, with a MD (95%CI) of -1.06 (-5.48 to 3.36) min, although trial sequential analysis suggested insufficient evidence. In summary, there is not enough evidence to draw definite conclusions on the effect of epidural volume extension. The quality of the current evidence is low for both efficacy (maximum sensory height) and safety (hypotension). However, there may be a significantly shorter motor block recovery time when different anaesthetic mixtures are used in epidural volume extension and control groups; this warrants further investigation.

Risk of subsequent cutaneous malignancy in patients with prior melanoma: a systematic review and meta-analysis.

Melanoma patients are known to be at risk of developing multiple cutaneous (pre-) malignancies, however, the exact dimensions of these risks are unknown. In this meta-analysis, risks of developing a melanoma, basal cell carcinoma (BCC) or squamous cell carcinoma (SCC) after a melanoma were investigated. An extensive systematic literature search was conducted (last performed on 18 January 2012). Studies reporting risks, i.e. proportions, standardized incidence ratios (SIR) and cumulative risks (CRs) were included. Fifty, of 233 fully read articles, met selection criteria. Two independent reviewers extracted data on study characteristics and risks measurements. Random-effects meta-analyses were used to pool the risk estimates for the three tumour combinations. In melanoma patients, pooled proportions for a subsequent melanoma, BCC or SCC were respectively 3.8% (n = 47), 2.8% (n = 5) and 1.0% (n = 6). The pooled SIRs for a subsequent melanoma, BCC or SCC in melanoma patients were respectively 10.4 (n = 12), 4.6 (n = 2) and 2.8 (n = 2). Mean 20-year CRs of a subsequent melanoma, BCC or SCC in melanoma patients were respectively 5.4% (n = 3), 14.0% (n = 1) and 4.0% (n = 1). Subgroup analyses showed substantial differences in reported risks between continents and study design. In conclusion, a history of a prior melanoma is a strong predictor for development of a subsequent melanoma (approximately 10-fold increased risk) and to a lesser extent BCC or SCC. This information could serve as information for health care systems. Further, secondary prevention seems pivotal in this patient group.

Hours in non-parental child care are related to language development in a longitudinal cohort study.

BACKGROUND: The effects of child care services on several domains of child development have been extensively investigated, but evidence regarding the effects of child care on language development remains inconclusive. METHODS: Within a large-scale population-based study, we examined the longitudinal associations between non-parental child care and language development from 1 to 6 years (n = 5375). RESULTS: Results showed that more hours in non-parental child care were associated with better language abilities. However, more hours in care in the first year of life were associated with less language proficiency at ages 1 to 1.5. At later ages, this effect disappeared and language proficiency increased. Furthermore, children who spent more hours in centre-based care had better language scores than children in home-based care. Ethnicity, socio-economic status, gender or parity did not change these results. CONCLUSIONS: This large, multi-ethnic study demonstrates beneficial effects of non-parental child care, particularly centre-based care, on language proficiency later in childhood.

Systematic review and meta-analysis of the association between diabetes mellitus and incidence and mortality in breast and colorectal cancer.

BACKGROUND: Increasing evidence suggests that diabetes mellitus (DM) is associated with increased cancer incidence and mortality. Several mechanisms involved in diabetes, such as promotion of cell proliferation and decreased apoptosis, may foster carcinogenesis. This study investigated the association between DM and cancer incidence and cancer-specific mortality in patients with breast and colorectal carcinoma. METHODS: A meta-analysis of controlled trials, prospective cohort studies and pooled cohort studies published after 2007 was conducted. Embase, PubMed and the Cochrane Library were searched. Summary hazard ratios (HRs) were calculated using a random-effects model. Sensitivity and subgroup analyses were performed to adjust for confounders, mode of DM assessment and follow-up time. RESULTS: Twenty studies were included to investigate the association between DM and breast and colorectal cancer incidence and cancer-specific mortality. The studies predominantly comprised patients with type II DM. The overall HR for breast cancer incidence was 1·23 (95 per cent confidence interval 1·12 to 1·34) and that for colorectal cancer was 1·26 (1·14 to 1·40) in patients with DM compared with those without diabetes. The overall HR was 1·38 (1·20 to 1·58) for breast cancer- and 1·30 (1·15 to 1·47) for colorectal cancer-specific mortality in patients with DM compared with those without diabetes. CONCLUSION: This meta-analysis indicated that DM is a risk factor for breast and colorectal cancer, and for cancer-specific mortality.

Markers of systemic inflammation in psoriasis: a systematic review and meta-analysis.

Studies investigating systemic inflammation in psoriasis use different serum markers and report discrepant results. We set out to determine whether systemic inflammation is elevated in patients with psoriasis compared with healthy controls, and to measure the extent of this elevation, by summarizing available data on serum inflammatory markers. PubMed, Embase and Web of Science were searched from inception to March 2011. We included studies comparing the serum inflammatory markers interleukin (IL)-1ß, IL-6, IL-10, C-reactive protein (CRP), intracellular adhesion molecule (ICAM)-1, E-selectin and tumour necrosis factor (TNF)-α in patients with psoriasis and healthy controls. Differences in serum marker levels between patients and controls were pooled as standardized mean differences (SMDs; Cohen's d) using a random-effects model. Seventy-eight studies were eligible. Of the 7852 individuals included, 3085 had (severe plaque) psoriasis. The pooled SMDs were higher in patients with psoriasis than in healthy controls for IL-6 [d = 1·32, 95% confidence interval (CI) 0·83-1·81], CRP (d = 1·83, 95% CI 0·76-2·90), TNF-α (d = 1·32, 95% CI 0·86-1·79), E-selectin (d = 1·78, 95% CI 1·32-2·25) and ICAM-1 (d = 1·77, 95% CI 1·15-2·39). The SMD between cases and controls for IL-1ß and IL-10 was not significant. Age had a significant effect on the SMD for IL-6 and TNF-α. For IL-6 the effect size was higher for plaque psoriasis studies (d = 1·98). The effect size was not influenced by the Psoriasis Area and Severity Index, measurement method or quality assessment. The pooled analyses suggest modest but significantly elevated levels of the proinflammatory cytokines in the serum of patients with psoriasis with predominantly severe disease. To what extent this modest increment is clinically relevant could be investigated in a synthesis of all studies measuring inflammation before and after antipsoriatic therapy.

Exhaled nitric oxide and the risk of wheezing in infancy: the Generation R Study.

We assessed whether exhaled nitric oxide fraction (F(eNO)), a marker of eosinophilic airway inflammation, at 6 months was associated with the risk of wheezing during the first 2 yrs of life. In the Generation R birth cohort, pre- and post-natal risk factors for respiratory morbidity and respiratory symptoms were assessed by questionnaires at 6 and 24 months. In 428 infants, off-line mixed oral/nasal F(eNO) was successfully measured during tidal breathing at 6 months. Complete data on F(eNO) and respiratory symptoms within the first 6 months of life were available for 294 infants. F(eNO) was higher in males, was positively associated with age and was negatively associated with upper and lower respiratory symptoms within the first 6 months. Logistic regression analysis showed that for every ppb increase of F(eNO) measured at 6 months, infants had a 1.06 (95% confidence interval 1.01-1.11)-fold increased risk of wheezing in the second year of life. High F(eNO) (>17.5 ppb) showed a limited added value in predicting wheezing in the second year. We conclude that F(eNO) at 6 months is positively associated with the risk of wheezing, but has limited added value in predicting wheezing in the second year of life in individual children.

Infant nutritional factors and functional constipation in childhood: the Generation R study.

OBJECTIVES: Food allergy and celiac disease may lead to childhood constipation. Early introduction of food allergens and gluten in the first year of life has been suggested to have a function in these food intolerances, but it is unclear whether this also holds true for development of childhood constipation. The aim of this study was to assess the association between the timing of introduction of food allergens and gluten early in life and functional constipation in childhood. METHODS: This study was embedded in the Generation R study, a population-based prospective cohort study from fetal life until young adulthood. Functional constipation at 24 months of age was defined in 4,651 children according to the Rome II criteria of defecation frequency <3 times a week or the presence of mainly hard feces for at least 2 weeks. RESULTS: At the age of 24 months, 12% of the children had functional constipation. Children with functional constipation got introduced to gluten more often before or at the age of 6 months than children without functional constipation (37% and 27%, respectively). After adjustment for birth weight, gestational age, gender, ethnicity, maternal education, and family history of atopy and chronic intestinal disorders, functional constipation was significantly associated with early gluten introduction (odds ratio (OR): 1.35; 95% confidence interval (CI): 1.10-1.65). No association was found between timing of introduction of cow's milk, hen's egg, soy, peanuts, and tree nuts with functional constipation. A history of cow's milk allergy in the first year of life was significantly associated with functional constipation in childhood (OR: 1.57; 95% CI: 1.04-2.36). CONCLUSIONS: These results suggest that early gluten introduction in the first year of life provide a trigger for functional constipation in a subset of children. In case of functional constipation, there also might be a role for cow's milk allergy initiated in the first year of life.

Systematic review and meta-analysis of sex differences in outcome after intervention for abdominal aortic aneurysm.

BACKGROUND: The aim of this study was to assess possible differences in mortality between men and women with an abdominal aortic aneurysm (AAA) treated either by elective repair or following aneurysm rupture. METHODS: A systematic literature search was performed using the MEDLINE, Cochrane and Embase databases. Data were analysed by means of bivariate random-effects meta-analysis. Data were pooled and odds ratios (ORs) calculated for women compared with men. RESULTS: Sixty-one studies (516 118 patients) met the predetermined inclusion criteria. Twenty-six reported on elective open AAA repair, 21 on elective endovascular repair, 25 on open repair for ruptured AAA and one study on endovascular repair for ruptured AAA. Mortality rates for women compared with men were 7.6 versus 5.1 per cent (OR 1.28, 95 per cent confidence interval (c.i.) 1.09 to 1.49) for elective open repair, 2.9 versus 1.5 per cent (OR 2.41, 95 per cent c.i. 1.14 to 5.15) for elective endovascular repair, and 61.8 versus 42.2 per cent (OR 1.16, 95 per cent c.i. 0.97 to 1.37) in the group that had open repair for rupture. The group that had endovascular repair for ruptured AAA was too small for meaningful analysis. CONCLUSION: Women with an AAA had a higher mortality rate following elective open and endovascular repair.

Describing fluctuating indoor aerosol dust measurements with application to house dust mite allergens.

Measuring house dust mite aeroallergen concentrations is essential in understanding mite allergen exposure. Physically, the aerolized house dust mite faeces are part of indoor particulate matter. We studied the statistical ways of summarizing measurements of fluctuating mite aeroallergen exposure inside homes through indoor particulate matter. To study emissions from beddings, we measured the time-related airborne dust concentration after shaking a duvet. Analysis was performed both by a method based on the estimated mean and by a non-linear model. Twenty-eight studies reported a sum of concentrations; only one also reported the peak. In our four experiments on shaking a duvet (245 to 275 measurements each), the peak value was two to four times higher than the mean. The mean-based and non-linear models both predicted the sum of concentrations exactly. A 1% upper prediction bound and the non-linear model predicted the peak emission rate moderately well (64 to 92%, and 63 to 93%, respectively). Mean levels of indoor particulate matter measurements differ substantially from peak concentrations. The use of the mean is only sufficient to predict the sum of concentrations. We suggest that, mite aeroallergen measurements should include information on the peak as well as the mean.

Bivariate random effects meta-analysis of ROC curves.

Meta-analysis of receiver operating characteristic (ROC)-curve data is often done with fixed-effects models, which suffer many shortcomings. Some random-effects models have been proposed to execute a meta-analysis of ROC-curve data, but these models are not often used in practice. Straightforward modeling techniques for multivariate random-effects meta-analysis of ROC-curve data are needed. The 1st aim of this article is to present a practical method that addresses the drawbacks of the fixed-effects summary ROC (SROC) method of Littenberg and Moses. Sensitivities and specificities are analyzed simultaneously using a bivariate random-effects model. The 2nd aim is to show that other SROC curves can also be derived from the bivariate model through different characterizations of the estimated bivariate normal distribution. Thereby the authors show that the bivariate random-effects approach not only extends the SROC approach but also provides a unifying framework for other approaches. The authors bring the statistical meta-analysis of ROC-curve data back into a framework of relatively standard multivariate meta-analysis with random effects. The analyses were carried out using the software package SAS (Proc NLMIXED).

Blood pressure response to calcium supplementation: a meta-analysis of randomized controlled trials.

Calcium plays a role in blood pressure (BP) regulation, but the importance of supplemental calcium intake for the prevention of hypertension is still debated. We conducted a meta-analysis of randomized controlled trials to determine the effect of calcium supplementation on BP. A systematic search for randomized trials of calcium supplementation and BP in non-pregnant subjects was performed in Medline from 1966 to June 2003. Seventy-one trials were identified, 40 of which met the criteria for meta-analysis (total of 2492 subjects). Two persons independently extracted data from original publications on changes in calcium intake and BP. In addition, data were collected on subjects' characteristics, that is, age, gender, initial BP and initial calcium intake. A random effects model was used to obtain the effect of calcium supplementation on BP, overall and in predefined population subgroups. Calcium supplementation (mean daily dose: 1200 mg) reduced systolic BP by -1.86 mm Hg (95% confidence interval: -2.91 to -0.81) and diastolic BP by -0.99 mm Hg (-1.61 to -0.37). In people with a relatively low calcium intake (< or =800 mg per day) somewhat larger BP estimates were obtained, that is, -2.63 (-4.03 to -1.24) for systolic BP and -1.30 (-2.13 to -0.47) for diastolic BP. Our study suggests that an adequate intake of calcium should be recommended for the prevention of hypertension. More research on BP in people with calcium-deficient diets is warranted.

[Roaming through methodology. XVI. What to do about missing data]. / Dwalingen in de methodologie. XVI. Wat te doen met ontbrekende waarnemingen?

Medical scientific research involving multiple measurements in patients is usually complicated by missing values. In case of missing values the choice is to limit the analysis to the complete cases or to analyse all available data. Both methods may suffer from substantial bias and may only be applied in a valid way if the rather strong assumption of 'missing completely at random' holds for the missing values, i.e. the missing value is not related to the other measured data nor to unmeasured data. Two other statistical methods may be applied to deal with missing values: the likelihood approach and the multiple imputation method. These methods make efficient use of all available data and take into account information implied by the available data. These methods are valid under the less stringent assumption of 'missing at random', i.e. the missing value is related to the other measured data, but not to unmeasured data. The best approach is to ensure that no data are missing.

Parental anthropometrics, early growth and the risk of overweight in pre-school children: the Generation R Study.

BACKGROUND: There are limited data regarding the associations of both maternal and paternal anthropometrics with longitudinally measured post-natal growth measures in early childhood. OBJECTIVE: To assess the associations of maternal and paternal anthropometrics with growth characteristics and the risk of overweight in pre-school children. STUDY DESIGN: Population-based prospective cohort study from early foetal life onwards in the Netherlands. METHODS: Maternal pre-pregnancy anthropometrics and gestational weight gain, and paternal anthropometrics were related to foetal and post-natal growth measures and the risk of overweight until the age of 4 years. Analyses were based on 5674 mothers, fathers and their children. RESULTS: Both pre-pregnancy maternal and paternal height, weight and body mass index were associated with corresponding foetal and post-natal anthropometric measures. Maternal body mass index had a significantly stronger effect on childhood body mass index than paternal body mass index. As compared to children from parents with normal body mass index, children from two obese parents had an increased risk of overweight at the age of 4 years (odds ratio 6.52 (95% confidence interval 3.44, 12.38). Maternal gestational weight gain was only among mothers with normal body mass index associated with body mass index and the risk of overweight in the children. CONCLUSION: Maternal and paternal anthropometrics affect early growth in pre-school children differently. Gestational weight gain in mothers without overweight and obesity is related to the risk of overweight in early childhood.

Happiness Scale Interval Study. Methodological Considerations.

The Happiness Scale Interval Study deals with survey questions on happiness, using verbal response options, such as 'very happy' and 'pretty happy'. The aim is to estimate what degrees of happiness are denoted by such terms in different questions and languages. These degrees are expressed in numerical values on a continuous [0,10] scale, which are then used to compute 'transformed' means and standard deviations. Transforming scores on different questions to the same scale allows to broadening the World Database of Happiness considerably. The central purpose of the Happiness Scale Interval Study is to identify the happiness values at which respondents change their judgment from e.g. 'very happy' to 'pretty happy' or the reverse. This paper deals with the methodological/statistical aspects of this approach. The central question is always how to convert the frequencies at which the different possible responses to the same question given by a sample into information on the happiness distribution in the relevant population. The primary (cl)aim of this approach is to achieve this in a (more) valid way. To this end, a model is introduced that allows for dealing with happiness as a latent continuous random variable, in spite of the fact that it is measured as a discrete one. The [0,10] scale is partitioned in as many contiguous parts as the number of possible ratings in the primary scale sums up to. Any subject with a (self-perceived) happiness in the same subinterval is assumed to select the same response. For the probability density function of this happiness random variable, two options are discussed. The first one postulates a uniform distribution within each of the different subintervals of the [0,10] scale. On the basis of these results, the mean value and variance of the complete distribution can be estimated. The method is described, including the precision of the estimates obtained in this way. The second option assumes the happiness distribution to be described as a beta distribution on the interval [0,10] with two shape parameters (α and ß). From their estimates on the basis of the primary information, the mean value and the variance of the happiness distribution in the population can be estimated. An illustration is given in which the method is applied to existing measurement results of 20 surveys in The Netherlands in the period 1990-2008. The results clarify our recommendation to apply the model with a uniform distribution within each of the category intervals, in spite of a better validity of the alternative on the basis of a beta distribution. The reason is that the recommended model allows to construct a confidence interval for the true but unknown population happiness distribution. The paper ends with a listing of actual and potential merits of this approach, which has been described here for verbal happiness questions, but which is also applicable to phenomena which are measured along similar lines.

Measures of Inequality: Application to Happiness in Nations.

What is a good measure for happiness inequality? In the context of this question, we have developed an approach in which individual happiness values in a sample are considered as elements of a set and inequality as a binary relation on that set. The total number of inequality relations, each weighed by the distance on the scale of measurement between the pair partners, has been adopted as an indicator for the inequality of the distribution as a whole. For models in which the happiness occurs as a continuous latent variable, an analogous approach has been developed on the basis of differentials. In principle, this fundamental approach results in a (zero) minimum value, and, more importantly, also in a maximum value. In the case where happiness is measured using a k-points scale, the maximum inequality is obtained if all ½N sample members select the lowest possible rating (Eq. 1) and the other ½N the highest possible one (k). This finding even applies to the truly ordinal case, i.e., if the distances between the successive ratings on the scale are unknown. It is, however, impossible to quantify the inequality of some measured sample distribution, unless all distances of the k categories of the scale of measurement are known or at least estimated, either on an empirical basis or on the basis of assumptions. In general, the numerical application of the method to continuous distributions is very complicated. An exploration on the basis of a relatively simple model with a linear probability density function suggests that the inequality of a beta probability distribution with shape parameters a and b increases as the value of these parameters decreases. A contour plot, obtained by numerical integration, demonstrates this relationship in a quantitative way. This approach is applicable to judge the aptness of common statistics of dispersion, among which the standard deviation and the Gini coefficient. The former is shown to be more appropriate than the latter for measuring inequality of happiness within nations.

Incidence of temporomandibular involvement in juvenile idiopathic arthritis.

OBJECTIVE: Temporomandibular joint (TMJ) involvement is a frequent feature in cross-sectional prevalence studies among juvenile idiopathic arthritis (JIA) patients. The cross-sectional design makes it almost impossible to study the incidence. Follow-up data on TMJ involvement are sparse. In this study patients were reviewed with an interval of a minimum of 1 year and a maximum of 2 years to study the yearly incidence of TMJ involvement and to obtain follow-up data on TMJ involvement and orthopantomogram (OPT) alterations. METHODS: Children with JIA from a previous study on TMJ involvement were included. OPTs were scored according to Rohlin's grading system (grade 0-5). A paediatric rheumatologist measured the level of disease activity during the interval. RESULTS: Eighty-nine of the 97 patients were included in this study with a mean follow-up of 14 months. The yearly incidence of TMJ involvement was 7.1% in patients with JIA. Improvement on the OPT was seen in 27 patients (66%), and 19 of these patients no longer showed any signs of TMJ involvement. Worsening on the OPT was seen in four patients (10%). Disease activity was significantly lower in the improved patients than in the patients with worsening. CONCLUSION: Condylar lesions due to arthritis can improve over time, indicating a regenerative capacity of the mandibular condyle. As condylar improvement seems to be associated with low disease activity, it is important to consider the TMJ when deciding on a therapeutic regimen.

Baseline risk as predictor of treatment benefit: three clinical meta-re-analyses.

A relationship between baseline risk and treatment effect is increasingly investigated as a possible explanation of between-study heterogeneity in clinical trial meta-analysis. An approach that is still often applied in the medical literature is to plot the estimated treatment effects against the estimated measures of risk in the control groups (as a measure of baseline risk), and to compute the ordinary weighted least squares regression line. However, it has been pointed out by several authors that this approach can be seriously flawed. The main problem is that the observed treatment effect and baseline risk measures should be viewed as estimates rather than the true values. In recent years several methods have been proposed in the statistical literature to potentially deal with the measurement errors in the estimates. In this article we propose a vague priors Bayesian solution to the problem which can be carried out using the 'Bayesian inference using Gibbs sampling' (BUGS) implementation of Markov chain Monte Carlo numerical integration techniques. Different from other proposed methods, it uses the exact rather than an approximate likelihood, while it can handle many different treatment effect measures and baseline risk measures. The method differs from a recently proposed Bayesian method in that it explicitly models the distribution of the underlying baseline risks. We apply the method to three meta-analyses published in the medical literature and compare the results with the outcomes of the other recently proposed methods. In particular we compare our approach to McIntosh's method, for which we show how it can be carried out using standard statistical software. We conclude that our proposed method offers a very general and flexible solution to the problem, which can be carried out relatively easily with existing Bayesian analysis software. A confidence band for the underlying relationship between true effect measure and baseline risk and a confidence interval for the value of the baseline risk measure for which there is no treatment effect are easily obtained by-products of our approach.

Effects of inhaled fluticasone and oral prednisolone on clinical and inflammatory parameters in patients with asthma.

BACKGROUND: Guidelines state that oral and inhaled corticosteroids are the cornerstone of asthma treatment. The effect of both types of treatment can be assessed by measuring lung and systemic parameters. Treatment for two weeks with either oral prednisolone (30 mg/day), high dose fluticasone propionate (2000 microg/day, FP2000), or lower dose FP (500 microg/day, FP500), both given by a dry powder inhaler, were compared. METHODS: One hundred and twenty patients with asthma were treated for two weeks in a double blind parallel group design. Lung function, asthma symptoms, airway hyperresponsiveness (PC(20) methacholine and adenosine-5'-monophosphate), sputum eosinophil and eosinophilic cationic protein (ECP) levels were measured as lung parameters. In addition, morning serum blood cortisol, blood eosinophil, and serum ECP levels were measured as systemic parameters. RESULTS: PC(20) methacholine and adenosine-5'-monophosphate showed significantly greater improvement with FP2000 (1.99 and 4.04 doubling concentrations (DC), respectively) than prednisolone (0.90 DC, p = 0.02; 2.15 DC, p = 0. 05) and marginally more than with FP500 (1.69 and 3.54 DC). Changes in sputum eosinophil and ECP concentrations showed similar trends; the decrease in ECP was significantly greater with FP2000 than with FP500. In contrast, the systemic parameters of steroid activity (cortisol, peripheral blood eosinophils, and serum ECP) decreased to a similar extent with FP2000 and prednisolone but significantly less with FP500. CONCLUSIONS: Oral prednisolone (30 mg/day) was inferior to FP2000 in improving airway hyperresponsiveness to both methacholine and AMP, with similar trends in forced expiratory volume in one second (FEV(1)), sputum eosinophil and ECP concentrations. Systemic effects were similar with prednisolone and FP2000 and less with FP500.