Search

Reporting guidelines, such as the consolidated standards of reporting trials (CONSORT)statement, improve the reporting of research in the medical literature (Turner et al., 2012). Many such guidelines exist and the CONSORT extension to non-pharmacological trials (Boutron et al., 2008)provides suitable guidance for reporting betweengroups intervention studies in the behavioral sciences. The CONSORT extension for N-of-1 Trials (CENT 2015)was developed for multiple crossover trials with single individuals in the medical sciences (Shamseer et al., 2015; Vohra et al., 2015), but there is no reporting guideline in the CONSORT tradition for single-case research used in the behavioral sciences. We developed the single-case reporting guideline in behavioural interventions (SCRIBE)2016 to meet this need. This statement article describes the methodology of the development of the SCRIBE 2016, along with the outcome of 2 Delphi surveys and a consensus meeting of experts. We present the resulting 26-item SCRIBE 2016 checklist. The article complements the more detailed SCRIBE 2016 explanation and elaboration article (Tate et al., 2016)that provides a rationale for each of the items and examples of adequate reporting from the literature. Both these resources will assist authors to prepare reports of single-case research with clarity, completeness, accuracy, and transparency. They will also provide journal reviewers and editors with a practical checklist against which such reports may be critically evaluated.We developed a reporting guideline to provide authors with guidance about what should be reported when writing a paper for publication in a scientific journal using a particular type of research design: the single-case experimental design. This report describes the methods used to develop the single-case reporting guideline in behavioural interventions (SCRIBE)2016. As a result of 2 online surveys and a 2-day meeting of experts, the SCRIBE 2016 checklist was developed, which is a set of 26 items that authors need to address when writing about single-case research. This article complements the more detailed SCRIBE 2016 explanation and elaboration article (Tate et al., 2016)that provides a rationale for each of the items and examples of adequate reporting from the literature. Both these resources will assist authors to prepare reports of single-case research with clarity, completeness, accuracy, and transparency. They will also provide journal reviewers and editors with a practical checklist against which such reports may be critically evaluated. We recommend that the SCRIBE 2016 is used by authors preparing manuscripts describing single-case research for publication, as well as journal reviewers and editors who are evaluating such manuscripts.Nous avons elabore des normes de presentation (sorte de lignes directrices)servant de guide pratique pour la redaction d'un article a publier dans une revue scientifique, utilisantun design de recherche particulier: le protocole experimental a cas unique. Le present article decrit la procedure utilisee afin d’elaborer ces normes: « single-case reporting guideline in behavioural interventions » (SCRIBE-2016). A l'issue de deux enquetes en ligne et d'une conference de consensus reunissant des experts durant deux jours, une liste de conformite (check-list)SCRIBE-2016 comprenant 26 items que les auteurs devraient considerer lorsqu'ils redigent des travaux utilisant les protocoles a cas unique a ete etablie. Cet article complete ainsi l'article intitule « SCRIBE-2016 explication and elaboration » (Tate et al., 2016)qui detaille et explique chaque item de la liste tout en fournissant, en guise de modele, des illustrations pratiques tires de travaux exemplaires de la litterature. Ces deux ressources aideront les auteurs a rediger avec clarte, exhaustivite, exactitude et transparence des articles de recherche portant sur des cas uniques. Elles fourniront egalement aux experts des revues ainsi qu’a leurs redacteurs en chef une liste de verification pratique qui leur servira de grille de lecture critique de ces articles. Nous recommandons de fait que le SCRIBE-2016 soit utilise aussi bien par les auteurs qui envisagent de publier des manuscrits rendant compte d'une recherche a cas unique, que par les experts et editeurs de revues qui les evaluent.Les normes en matiere de redaction d'articles, a l'instar de celles dediees a l’ecriture de travaux d'essais randomises controles (CONSORT), ameliorent la qualite des comptes rendus de la recherche dans la litterature medicale (Turner et al., 2012). Il existe de nombreuses normes de ce genre parmi lesquelles l'extension du CONSORT pour les essais non-pharmacologiques (Boutron et al., 2008)qui fournit deja un cadre approprie pour rendre compte d’etudes experimentales intergroupes en science du comportement. Une extension du CONSORT pour les essais « N = 1 » (CENT 2015)a egalement bien ete developpee en sciences medicales pour les essais transversaux multiples realises aupres de cas uniques (Shamseer et al., 2015; Vohra et al., 2015). Toutefois il n'existait pas, jusqu’a lors, de normes de presentation comparables au CONSORTet applicables aux protocoles a cas uniques en sciences du comportement. Pour repondre a ce besoin, nous avons mis au point ces « normes de presentation d'une recherche interventionnelle utilisant un protocole a cas unique » (SCRIBE-2016). Cet article decrit la procedure d’elaboration du SCRIBE-2016, ainsi que les resultats des deux enquetes Delphi et de la conference de consensus d'experts realises dans ce contexte. Nous y presentonsles normes SCRIBE-2016 en 26 items qui en resultent. La presente declaration est un preambule a l'article plus detaille intitule « SCRIBE-2016 explanation and elaboration » (Tate et al., 2016), qui detaille et apporte un eclaircissement pour chacun des items des normes de presentation et les illustre par des exemples de travaux adequats tires de la litterature. Ces deux ressources aideront les auteurs a rediger avec clarte, exhaustivite, exactitude et transparence des travaux de recherche portant sur des cas uniques. Elles fourniront egalement aux experts des revues ainsi qu'aux redacteurs en chef une liste de verification pratique qui leur servira de grille de lecture critique de ces travaux.

Objective To evaluate diagnostic tests, analysts use meta-analyses to provide inputs to parameters in decision models. Choosing parameter estimands from meta-analyses requires understanding the meta-analytic and decision-making contexts. Study Design and Setting We expand on an analysis comparing positron emission tomography (PET), PET with computed tomography (PET/CT), and conventional workup (CW) in women with suspected recurrent breast cancer. We discuss Bayesian meta-analytic summaries (posterior mean over a set of existing studies, posterior estimate in an existing study, posterior predictive mean in a new study) used to estimate diagnostic test parameters (prevalence, sensitivity, specificity) needed to calculate quality-adjusted life years in a decision model contextualizing PET, PET/CT, and CW. Results The mean and predictive mean give similar estimates, but the latter displays greater uncertainty. Namely, PET/CT outperforms CW on average but may not do better than CW when implemented in future settings. Conclusion Selecting estimands for decision model parameters from meta-analyses requires understanding the relationship between decision settings and meta-analysis studies' settings, specifically whether the former resemble one or all study settings or represents new settings. We provide an algorithm recommending appropriate estimands as input parameters in decision models for diagnostic tests to obtain output parameters consistent with the decision context.

Objectives To compare statistical methods for meta-analysis of sensitivity and specificity of medical tests (e.g., diagnostic or screening tests). Study Design and Setting We constructed a database of PubMed-indexed meta-analyses of test performance from which 2 × 2 tables for each included study could be extracted. We reanalyzed the data using univariate and bivariate random effects models fit with inverse variance and maximum likelihood methods. Analyses were performed using both normal and binomial likelihoods to describe within-study variability. The bivariate model using the binomial likelihood was also fit using a fully Bayesian approach. Results We use two worked examples—thoracic computerized tomography to detect aortic injury and rapid prescreening of Papanicolaou smears to detect cytological abnormalities—to highlight that different meta-analysis approaches can produce different results. We also present results from reanalysis of 308 meta-analyses of sensitivity and specificity. Models using the normal approximation produced sensitivity and specificity estimates closer to 50% and smaller standard errors compared to models using the binomial likelihood; absolute differences of 5% or greater were observed in 12% and 5% of meta-analyses for sensitivity and specificity, respectively. Results from univariate and bivariate random effects models were similar, regardless of estimation method. Maximum likelihood and Bayesian methods produced almost identical summary estimates under the bivariate model; however, Bayesian analyses indicated greater uncertainty around those estimates. Bivariate models produced imprecise estimates of the between-study correlation of sensitivity and specificity. Differences between methods were larger with increasing proportion of studies that were small or required a continuity correction. Conclusion The binomial likelihood should be used to model within-study variability. Univariate and bivariate models give similar estimates of the marginal distributions for sensitivity and specificity. Bayesian methods fully quantify uncertainty and their ability to incorporate external evidence may be useful for imprecisely estimated parameters.

N-of-1 trials provide a mechanism for making evidence-based treatment decisions for an individual patient. They use key methodological elements of group clinical trials to evaluate treatment effectiveness in a single patient, for situations that cannot always accommodate large-scale trials: rare diseases, comorbid conditions, or in patients using concurrent therapies. Improvement in the reporting and clarity of methods and findings in N-of-1 trials is essential for reader to gauge the validity of trials and to replicate successful findings. A Consolidated Standards of Reporting Trials (CONSORT) extension for N-of-1 trials (CENT 2015) provides guidance on the reporting of individual and series of N-of-1 trials. CENT provides additional guidance for 14 of the 25 items of the CONSORT 2010 checklist, recommends a diagram for depicting an individual N-of-1 trial, and modifies the CONSORT flow diagram to address the flow of a series of N-of-1 trials. The rationale, development process, and CENT 2015 checklist and diagrams are reported in this document.

Background Change in glomerular filtration rate (GFR) is important for clinical decision making. GFR estimates from serum creatinine level provide an unbiased but imprecise estimate of GFR at single time points. However, the accuracy of estimated GFR over time is not well known. Study Design Longitudinal study of diagnostic test accuracy. Settings & Participants 4 clinical trials with longitudinal measurements of GFR and serum creatinine on the same day, including individuals with and without kidney disease with a wide range of kidney function, diverse racial backgrounds, and varied clinical characteristics. Index Test GFR estimated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. Reference Test GFR measured using urinary clearance of 125 I-iothalamate. Results Data included 19,735 GFR measurements in 3,531 participants during a mean follow-up of 2.6 years. Mean values at baseline for measured and estimated GFR and error (measured GFR – estimated GFR) were 73.1 (95% CI, 71.6 to 74.5), 72.7 (95% CI, 71.5 to 74.0), and 0.14 (95% CI, −0.35 to 0.63) mL/min/1.73 m 2 , respectively. Mean rates of change in measured and estimated GFR and error were −2.3 (95% CI, −2.4 to −2.1), −2.2 (95% CI, −2.4 to −2.1), and −0.09 (95% CI, −0.24 to 0.05) mL/min/1.73 m 2 per year ( P P P = 0.2, respectively). Variability (ie, standard deviation) among participants in rate of change in measured GFR, estimated GFR, and error was 4.3, 3.4, and 3.3 mL/min/1.73 m 2 per year, respectively. Only 15% of participants had a rate of change in error >3 mL/min/1.73 m 2 per year, and only 2% had a rate of change in error >5% per year. Limitations Participants' characteristics were not available over time. Conclusion The accuracy of GFR estimates did not change over time. Clinicians should interpret changes in estimated GFR over time as reflecting changes in measured GFR rather than changes in errors in the GFR estimates in most individuals.

Purpose: The aim of this study was to demonstrate that modern data mining tools can be used as one step in reducing the labor necessary to produce and maintain systematic reviews. Methods: We used four continuously updated, manually curated resources that summarize MEDLINE-indexed articles in entire fields using systematic review methods (PDGene, AlzGene, and SzGene for genetic determinants of Parkinson disease, Alzheimer disease, and schizophrenia, respectively; and the Tufts Cost-Effectiveness Analysis (CEA) Registry for cost-effectiveness analyses). In each data set, we trained a classification model on citations screened up until 2009. We then evaluated the ability of the model to classify citations published in 2010 as “relevant” or “irrelevant” using human screening as the gold standard. Results: Classification models did not miss any of the 104, 65, and 179 eligible citations in PDGene, AlzGene, and SzGene, respectively, and missed only 1 of 79 in the CEA Registry (100% sensitivity for the first three and 99% for the fourth). The respective specificities were 90, 93, 90, and 73%. Had the semiautomated system been used in 2010, a human would have needed to read only 605/5,616 citations to update the PDGene registry (11%) and 555/7,298 (8%), 717/5,381 (13%), and 334/1,015 (33%) for the other three databases. Conclusion: Data mining methodologies can reduce the burden of updating systematic reviews, without missing more papers than humans.

Background Nephrology has a limited number of randomized controlled trials (RCTs). The quality of randomized trials is compromised further when not all participants randomly assigned are accounted for transparently. Objectives Systematically evaluate RCTs in individuals with chronic kidney disease regarding reporting and accounting of data missing in outcome analysis. Study Design De novo empirical evaluation. Setting & Population English-language parallel-group design RCTs in adults with chronic kidney disease on dialysis therapy or with a kidney transplant published in MEDLINE in 2007 and 2008. Outcomes & Measurements (1) How often was there loss to analysis, defined as not all randomly assigned participants included in primary outcome analysis? (2) How often was intention-to-treat analysis complete; in other words, included all randomly assigned participants in their originally allocated group? (3) How often were methods of data imputation reported? Results Of 196 eligible RCTs, 27% did not clearly describe a primary outcome, 5% did not provide numbers of patients randomly assigned and analyzed, and 12% used time-to-event analysis. Of the remaining 110 studies, 58% had some loss to analysis, with a median loss to analysis of 10%. Fifty-four percent of trials claimed to have performed an intention-to-treat analysis, but only 44% of those included all participants randomly assigned. Only 5 of 110 (5%) studies mentioned imputation of missing data. Limitations Evaluation is restricted to analysis of primary study outcome. Only English-language publications were included. Exclusion of time-to-event analyses. Conclusions In variance to the reporting standards of CONSORT (Consolidated Standards of Reporting Trials), we found primary outcome designation missing in one-fourth of trials and poor quality in reporting and accounting of primary outcome data lost to analysis. Greater attention to transparency in handling and reporting loss to analysis will enhance the quality of trials in individuals with chronic kidney disease.

Summary Objective To evaluate the therapeutic trajectory of intra-articular hyaluronic acid (IAHA) vs placebo for knee osteoarthritis (OA). Design Our data sources include Medline, EMBASE, CINAHL, BIOSIS, Web of Science, Google Scholar, Cochrane database; hand searched reviews, manuscripts, and, supplements; author contacts for unpublished data. Randomized trials that reported effects of IAHA vs placebo on knee OA were selected based on inclusion criteria. We computed effect sizes for change from baseline at 4, 8, 12, 16, 20 and 24 weeks, using Bayesian random effects model. We performed multivariate analyses adjusting for correlation between time points. Meta-regressions were performed adjusting for potential confounders. Results The 54 eligible trials included 7545 participants. The conduct and quality of these trials varied in number of aspects. The effect size (ES) favored IAHA by week 4 (0.31; 95% CI 0.17, 0.45), reaching peak at week 8 (0.46; 0.28, 0.65), and then trending downwards, with a residual detectable effect at week 24 (0.21; 0.10, 0.31). This therapeutic trajectory was consistent among the subset of high quality trials and on multivariate analysis adjusting for correlation between time points. Conclusions Our meta-analysis highlights a therapeutic trajectory of IAHA for knee OA pain over 6 months post-intervention. With this additional perspective, we are able to infer that IAHA is efficacious by 4 weeks, reaches its peak effectiveness at 8 weeks and exerts a residual detectable at 24 weeks. On the other hand, the peak effect size (0.46; 0.28, 0.65), is greater than published effects from other OA analgesics [acetaminophen (ES=0.13; 0.04, 0.22); NSAIDs (ES=0.29; 0.22, 0.35); COX-2 inhibitors (ES=0.44; 0.33, 0.55)]. An effect size above 0.20 is considered to be clinically relevant on an individual patient basis in chronic pain conditions such as knee OA. Thus, its properties could have utility for certain clinical situations, or in combination with other therapies.

An equation from the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) provides more accurate estimates of the glomerular filtration rate (eGFR) than that from the modification of diet in renal disease (MDRD) Study, although both include a two-level variable for race (Black and White and other). Since creatinine generation differs among ethnic groups, it is possible that a multilevel ethnic variable would allow more accurate estimates across all groups. To evaluate this, we developed an equation to calculate eGFR that includes a four-level race variable (Black, Asian, Native American and Hispanic, and White and other) using a database of 8254 patients pooled from 10 studies. This equation was then validated in 4014 patients using 17 additional studies from the United States and Europe (validation database), and in 1022 patients from China (675), Japan (248), and South Africa (99). Coefficients for the Black, Asian, and Native American and Hispanic groups resulted in 15, 5, and 1% higher levels of eGFR, respectively, compared with the White and other group. In the validation database, the two-level race equation had minimal bias in Black, Native American and Hispanic, and White and other cohorts. The four-level ethnicity equation significantly improved bias in Asians of the validation data set and in Chinese. Both equations had a large bias in Japanese and South African patients. Thus, heterogeneity in performance among the ethnic and geographic groups precludes use of the four-level race equation. The CKD-EPI two-level race equation can be used in the United States and Europe across a wide range of ethnicity.

Kidney disease alters the pharmacokinetic disposition of many medications, requiring dosage adjustment to maintain therapeutic serum concentrations. The Cockcroft-Gault (CG) equation is used for pharmacokinetic studies and drug dosage adjustments, but the Modification of Diet in Renal Disease (MDRD) Study equation is more accurate and more often reported by clinical laboratories than the CG equation.Diagnostic test study.Pooled data set for 5,504 participants from 6 research studies and 4 clinical populations with measured glomerular filtration rate (GFR).Estimated kidney function using the MDRD Study and CG equations incorporating actual (CG) or ideal body weight (CG(IBW)) and standardized serum creatinine concentrations.Measured GFR assessed by using iodine-125-iothalamate urinary clearance.Concordance of assigned kidney function categories designated by the Food and Drug Administration (FDA) Guidance for Industry for pharmacokinetic studies and recommended dosages of 15 medications cleared by the kidneys.Concordance of kidney function estimates with measured GFR for FDA-assigned kidney function categories was 78% for the MDRD Study equation compared with 73% for the CG equation (P0.001) and 66% for the CG(IBW) equation (P0.001). Concordance between the MDRD Study equation and CG and CG(IBW) equations was 78% and 75%, respectively (P0.001). Concordance of kidney function estimates with measured GFR for recommended drug dosages was 88% for MDRD Study equation compared with 85% for the CG equation (P0.001) and 82% for the CG(IBW) equation (P0.001), with lower concordance when dosing recommendations for drugs included narrow GFR ranges. Concordance rates between the CG and CG(IBW) equations and MDRD Study equation were 89% and 88%, respectively (P0.05).Results based on simulation rather than pharmacokinetic studies. Outcome was drug dosage recommendations, rather than observed drug efficacy and safety.The MDRD Study equation can also be used for pharmacokinetic studies and drug dosage adjustments. As more accurate GFR-estimating equations are developed, they should be used for these purposes.

Background Proteinuria may be an early marker of chronic kidney disease in human immunodeficiency virus (HIV)-infected patients with coexisting chronic hepatitis and/or drug use. Minorities are at greater risk of chronic kidney disease. Data are limited about risk factors for proteinuria in Hispanic drug users with and without HIV infection. Study Design A cross-sectional study. Setting & Participants A community-recruited Hispanic cohort to study the role of drug use in HIV-associated malnutrition composed of 4 groups (106 HIV-infected drug users, 96 HIV-uninfected drug users, 38 HIV-infected non–drug users, and 47 healthy controls). Patients on renal replacement therapy were excluded. Predictors HIV infection, chronic hepatitis, history of hypertension or diabetes, and intravenous drug use (never, prior, or current). Outcomes & Measurements The presence of proteinuria was defined as urine dipstick result of 1+ or greater. Multivariable logistic regression was used to identify independent risk factors for proteinuria. Results Of 287 patients with available data, 24 (8.4%) had proteinuria. In univariate analyses, those with HIV infection; prior, but not current, intravenous drug use; and a history of hypertension or diabetes were more likely to have proteinuria. In multivariate analyses, significant risk factors for proteinuria were HIV infection (odds ratio, 9.2; 95% confidence interval, 1.9 to 45.8; P = 0.007); prior, but not current, intravenous drug use (odds ratio, 4.7; 95% confidence interval, 1.4 to 15.3; P = 0.01); and history of hypertension or diabetes (odds ratio, 8.2; 95% confidence interval, 3.1 to 21.7; P Limitations The cross-sectional study design makes it difficult to establish the temporal relationship. The number of outcomes in relation to the number of predictors is small. Conclusions HIV and prior intravenous drug use, but not chronic hepatitis or current intravenous drug use, were independently associated with proteinuria in this Hispanic population. Longitudinal studies to assess the development of proteinuria and chronic kidney disease in this high-risk population are warranted.

We conducted a metaanalysis of published evidence on the relationship between maternal obesity and the risk of neural tube defects (NTDs). Eligible studies were identified from 3 sources: (1) PubMed search of articles that were published from January 1980 through January 2007, (2) reference lists of publications that were selected from the PubMed search, and (3) reference lists of review articles on obesity and maternal outcomes that were published from January 2000 through January 2007. Twelve studies met inclusion criteria. A Bayesian random effects model was used for the metaanalysis and metaregression. Unadjusted odds ratios for an NTD-affected pregnancy were 1.22 (95% CI, 0.99-1.49), 1.70 (95% CI, 1.34-2.15), and 3.11 (95% CI, 1.75-5.46) among overweight, obese, and severely obese women, respectively, compared with normal-weight women. None of the study characteristics included in the metaregression analysis affected the results significantly. Maternal obesity is associated with an increased risk of an NTD-affected pregnancy.

Background Serum cystatin C was proposed as a potential replacement for serum creatinine in glomerular filtration rate (GFR) estimation. We report the development and evaluation of GFR-estimating equations using serum cystatin C alone and serum cystatin C, serum creatinine, or both with demographic variables. Study Design Test of diagnostic accuracy. Setting & Participants Participants screened for 3 chronic kidney disease (CKD) studies in the United States (n = 2,980) and a clinical population in Paris, France (n = 438). Reference Test Measured GFR (mGFR). Index Test Estimated GFR using the 4 new equations based on serum cystatin C alone, serum cystatin C, serum creatinine, or both with age, sex, and race. New equations were developed by using linear regression with log GFR as the outcome in two thirds of data from US studies. Internal validation was performed in the remaining one third of data from US CKD studies; external validation was performed in the Paris study. Measurements GFR was measured by using urinary clearance of iodine-125–iothalamate in the US studies and chromium-51–EDTA in the Paris study. Serum cystatin C was measured by using Dade-Behring assay, standardized serum creatinine values were used. Results Mean mGFR, serum creatinine, and serum cystatin C values were 48 mL/min/1.73 m 2 (5th to 95th percentile, 15 to 95), 2.1 mg/dL, and 1.8 mg/L, respectively. For the new equations, coefficients for age, sex, and race were significant in the equation with serum cystatin C, but 2- to 4-fold smaller than in the equation with serum creatinine. Measures of performance in new equations were consistent across the development and internal and external validation data sets. Percentages of estimated GFR within 30% of mGFR for equations based on serum cystatin C alone, serum cystatin C, serum creatinine, or both levels with age, sex, and race were 81%, 83%, 85%, and 89%, respectively. The equation using serum cystatin C level alone yields estimates with small biases in age, sex, and race subgroups, which are improved in equations including these variables. Limitations Study population composed mainly of patients with CKD. Conclusions Serum cystatin C level alone provides GFR estimates that are nearly as accurate as serum creatinine level adjusted for age, sex, and race, thus providing an alternative GFR estimate that is not linked to muscle mass. An equation including serum cystatin C level in combination with serum creatinine level, age, sex, and race provides the most accurate estimates.

We conducted this metaanalysis to summarize the available epidemiologic evidence on the relationship between maternal overweight and obesity and the risk of stillbirth. We identified studies from 3 sources: (1) a PubMed search of relevant articles that were published between January 1980 and September 2005, (2) reference lists of publications that were selected from the PubMed search, and (3) reference lists of review articles on obesity and maternal outcomes that were published between 2000 and 2005. We used a Bayesian random effects model to perform the metaanalysis and metaregression. Nine studies were included in the metaanalysis. The unadjusted odds ratios of a stillbirth were 1.47 (95% CI, 1.08-1.94) and 2.07 (95% CI, 1.59-2.74) among overweight and obese pregnant women, respectively, compared with normal-weight pregnant women. The metaregression analysis found no evidence that these estimates were affected by selected study characteristics. Maternal obesity is associated with an increased risk of stillbirth, although the mechanisms to explain this association are not clear.

Plasma lipid profile is affected by endogenous estrogen levels and hormone replacement therapy (HRT). As plasma lipid concentrations have a significant heritable basis and the effects of both endogenous estrogen and use of HRT are mediated by estrogen receptors, we sought to investigate the relationships between polymorphisms in estrogen receptor-α ( ESR1 ) and plasma lipid and lipoprotein concentrations. We analyzed data from 854 women (mean age 52±10 years) from the Framingham Heart Study. A TA repeat in the promoter region, c.30T > C in exon 1, c.454-397T > C , and c.454-351A > G in intron 1, all in linkage disequilibrium (LD), were significantly associated with low-density lipoprotein (LDL) particle size and concentration of small LDL particles. Women with the c.454-397C allele had larger LDL particle size (21.09±0.02nm versus 21.01±0.03nm, p =0.021) concurrent with lower small LDL particle concentration (0.47±0.02mmol/L versus 0.58±0.03mmol/L, p =0.008). Moreover, the TA[L] – c.30C – c.454-397C – c.454-351G haplotype (frequency, 32%) was associated with lower small LDL particle concentrations (−0.06±0.03mmol/L change associated with each copy of this haplotype, p =0.011) when compared to the TA[S] – c.30T – c.454-397T – c.454-351A haplotype (frequency, 46%), where L and S are long and short TA repeats. Our results suggest that common ESR1 polymorphisms have a significant effect on lipoprotein metabolism in women.

Objective Describe the level of registered dietitian (RD) involvement in neonatal intensive care units (NICUs) and associations with NICU nutrition practices. Design Questionnaires were mailed to 820 NICUs in the United States with two follow-up mailings to nonresponders. Abbreviated phone surveys were conducted with a random sample of 10% of nonresponders. A nutrition care score was devised based on a sum of 10 survey questions (range 0 to 10) to summarize the intensity of reported practices. Subjects/Setting Directors of NICUs in the United States and RDs associated with them. Statistical Analyses χ 2 , analysis of variance, Bonferroni and Duncan multiple range tests, regression. Results Respondents from 417 (54%) of the 772 NICUs eligible for the study provided completed questionnaires. Among NICUs responding, 76% involved RDs in care (41% employed full- or part-time RDs, 35% employed consult RDs), and 24% had no RD. NICUs with full- or part-time RDs provided fewer kilocalories and more protein parenterally, and more kilocalories and protein enterally. NICUs with less RD involvement were more likely to provide full-term infant feedings (eg, unfortified breast milk, full-term formula) to very-low-birth-weight infants. Mean nutrition care score varied with RD involvement from 4.6±1.7 (mean±standard deviation) for NICUs with a consult RD and 4.7±1.4 for NICUs employing no RD to 5.6±1.7 for NICUs with a full- or part-time RD (overall P Conclusions More involvement of RDs in NICUs increased the intensity of important aspects of nutrition care that may improve outcomes of very-low-birth-weight infants in NICUs. These findings highlight the importance of RDs as NICU team members.

Objective To assess the effectiveness of weight-loss and weight-control interventions for adults with pre-diabetes (impaired fasting glucose and impaired glucose tolerance), an important risk factor for the development of type 2 diabetes. Methods Computerized searches were conducted of multiple electronic bibliographic databases up to August 2003. Randomized controlled trials in any language were selected that examined weight-loss or weight-control strategies using at least one dietary, physical activity, or behavioral intervention, and with a follow-up interval of ≥12 months. Effects were combined using a random effects model. Results Studies were identified, with a total of 5168 participants. Follow-up ranged from 1 to 10 years. Quantitative synthesis was limited by the heterogeneity of populations, settings, and interventions, and by the small number of studies that examined outcomes other than weight. Overall, compared to usual care, four studies with a follow-up of 1 year reduced weight by 2.8 kg (95% confidence interval [CI]=1.0–4.7) (3.3% of baseline body weight) and decreased body mass index by 1.4 kg/m 2 (CI=0.5–2.3). Weight loss at 2 years was 2.7 kg (CI=1.9–3.4) (two studies). Modest improvements were noted in the few studies that examined glycemic control, blood pressure, and lipid concentrations ( p >0.05). The incidence of diabetes was significantly lower in the intervention groups versus the controls in three of five studies examining this outcome at 3 to 6 years follow-up. Conclusions Overall, weight-loss strategies using dietary, physical activity, or behavioral interventions produced significant improvements in weight among persons with pre-diabetes, and a significant decrease in diabetes incidence. Further work is needed on the long-term effects of these interventions on morbidity and mortality and on how to implement these interventions in the community setting.

Background Most persons with type 2 diabetes are overweight, and obesity worsens the metabolic and physiologic abnormalities associated with diabetes. Our objective was to assess the effectiveness of lifestyle and behavioral weight loss and weight control interventions in adults with type 2 diabetes. Methods Studies were obtained from searches of multiple electronic bibliographic databases, supplemented with hand searches of selected journals and consultation with experts in obesity research. Studies were included if they were published or unpublished randomized controlled trials in any language that examined weight loss or weight control strategies using one or more dietary, physical activity, or behavioral interventions, with a follow-up interval of at least 12 months. Effects were combined using a random-effects model. Results The 22 studies of weight loss interventions identified yielded a total of 4659 participants with a follow-up of 1 to 5 years. The pooled weight loss for any intervention in comparison with usual care among 585 subjects was 1.7 kg (95% confidence interval [CI]: 0.3 to 3.2 kg), or 3.1% of baseline body weight among 511 subjects. Among 126 persons who underwent a physical activity and behavioral intervention, those who also received a very low-calorie diet lost 3.0 kg (95% CI: −0.5 to 6.4 kg), or 1.6% of baseline body weight, more than persons who received a low-calorie diet. Among 53 persons who received identical dietary and behavioral interventions, those who received a more intense physical activity intervention lost 3.9 kg (95% CI: −1.9 to 9.7 kg), or 3.6% of baseline body weight, more than those who received a less intense or no physical activity intervention. Comparison groups often achieved substantial weight loss (up to 10.0 kg), minimizing between-group differences. Changes in glycated hemoglobin level generally corresponded to changes in weight and were not substantial when between-group differences were examined. Conclusion Weight loss strategies involving dietary, physical activity, or behavioral interventions were associated with small between-group improvements in weight. These results were minimized by weight loss in the comparison group, however, and examination of individual study arms revealed that multicomponent interventions, including very low-calorie diets or low-calorie diets, may hold promise for achieving weight loss in adults with type 2 diabetes.

Objective Two investigations evaluate Bayesian meta-regression for detecting treatment interactions. Study design and setting The first compares analyses of aggregate and individual patient data on 1,860 subjects from 11 trials testing angiotensin converting enzyme (ACE) inhibitors for nondiabetic kidney disease. The second explores meta-regression for detecting treatment interaction on 671 covariates, including the baseline risk, from 232 meta-analyses of binary outcomes compiled from the Cochrane Collaboration and the medical literature. Results In the ACE inhibitor study, treatment effects were homogeneous so meta-regression identified no interactions. Analysis of individual patient data using a multilevel model, however, discovered that treatment reduced glomerular filtration rate (GFR) more among patients with higher baseline proteinuria. The second investigation found meta-regression most effective for detecting treatment interactions with study-level factors in meta-analyses with >10 studies, heterogeneous treatment effects, or significant overall treatment effects. Under all three conditions, 46% of meta-regressions produced strong interactions (posterior probability >0.995) compared with 6% otherwise. Baseline risk was associated with the odds ratio in 6% of meta-analyses, half the rate found using maximum likelihood. Conclusion Meta-regression can detect interactions of treatment with study-level factors when treatment effects are heterogeneous. Individual patient data are needed for patient-level factors and homogeneous effects.

Transfer of infants between hospitals or their discharge home may bias comparisons of the performance across neonatal intensive care units (NICUs). This study attempts to show the potential size of transfer bias in the context of a large cohort study and describe strategies for minimizing this type of bias.To limit transfer bias in a neonatal growth study of extremely premature infants in six tertiary NICUs, we restricted eligibility to infants30 weeks gestation at birth and substituted matched replacements for early transfers (infants transferred or discharged prior to day of life 16).The restriction strategy was successful, reducing the overall early transfer rate from 16.4 to 3.6% and the range of transfer rates among individual NICUs from 0.6-32.7% to 0-11.0%. Replacement by matched substitutes had a much smaller effect because of the small number of early transfers and our inability to match on all factors distinguishing early transfers.Sampling strategies to minimize infants lost to follow-up were more successful than replacement strategies in limiting transfer bias in a NICU growth study. Although complete elimination of bias is likely impossible, valid studies require efforts to minimize, quantify, and test the effect of transfer bias.

Individual patient data are often required to evaluate how patient-specific factors modify treatment effects. We describe our experience combining individual patient data from 1946 subjects in 11 randomized controlled trials evaluating the effect of angiotensin-enzyme converting (ACE) inhibitors for treating nondiabetic renal disease. We sought to confirm the results of our meta-analysis of group data on the efficacy of ACE inhibitors in slowing the progression of renal disease, as well as to determine whether any study or patient characteristics modified the beneficial effects of treatment. In particular, we wanted to find out if the mechanism of action of ACE inhibitors could be explained by adjusting for follow-up blood pressure and urine protein. Each trial site sent a database of multiple files and multiple records per patient containing longitudinal data of demographic, clinical, and medication variables to the data coordinating center. The databases were constructed in several different languages using different software packages with unique file formats and variable names. Over 4 years, we converted the data into a standardized database of more than 60,000 records. We overcame a variety of problems including inconsistent protocols for measurement of key variables; varying definitions of the baseline time; varying follow-up times and intervals; differing medication-reporting protocols; missing variables; incomplete, missing, and implausible data values; and concealment of key data in text fields. We discovered that it was easier and more informative to request computerized data files and merge them ourselves than to ask the investigators to abstract partial data from their files. Although combining longitudinal data from different trials based on different protocols in different languages is complex, costly, and time-intensive, analyses based on individual patient data are extremely informative. Funding agencies must be encouraged to provide support to collaborative groups combining databases.

To understand predictors of cardiac arrest early in acute myocardial infarction (AMI), for the Thrombolytic Predictive Instrument, we developed a multivariable regression model predicting primary cardiac arrest using time-dependent variables based on a case-control study of emergency department (ED) patients with AMI: 65 cases with sudden cardiac arrest and 258 without cardiac arrest. Within the first hour of AMI symptom onset, adjusting for age, systolic blood pressure, serum potassium, and infarct size, increased risk of cardiac arrest was associated with electrocardiographic prolonged QTc interval and a greater sum of ST-segment elevation. After 1 hour, the effect of ST-segment elevation was much reduced and the effect of the QTc interval was reversed, so prolonged QTc appeared protective. Accordingly, for patients presenting 30 minutes after chest pain onset, compared with a QTc of 0.44, the risk for cardiac arrest for patients with QTc of 0.50 was more than doubled (odds ratio [OR] 2.20, 95% confidence intervals [CI] 1.17 to 4.13), whereas for those presenting after an hour, it was much lower (e.g., at 1.5 hours, OR 0.21, 95% CI 0.06 to 0.73). Patients presenting 30 minutes after chest pain onset with a sum of ST elevation of 20 mm had a threefold higher risk than patients with a sum of ST elevation of 5 mm (OR 3.37, 95% CI 1.83 to 6.20). However, if presenting 1.5 hours after chest pain onset, the risk was barely elevated (OR 1.18; 95% CI 1.09 to 1.29). Thrombolytic therapy was protective, halving the odds of cardiac arrest (OR 0.51, 95% CI 0.27 to 0.93). Thus, the relation of prolonged QTc interval and substantial ST segment elevation to cardiac arrest in AMI may be obscured because patients with these risks are more likely to die soon after AMI onset, before ED presentation, and are thereby unavailable for study. Those with prolonged QTc or substantial ST elevation who survive the initial 1.5-hour period are those less susceptible to these risks.

In 1992, a review article about meta-analysis identified only 15 meta-analyses of randomized, controlled trials of cancer therapy. Since then, the total number of meta-analyses in this field has increased almost sixfold. More importantly, the number of randomized, controlled trials in this discipline has also grown tremendously. The expansion in the literature will provide a fertile ground for future meta-analyses. The quality of the recent publications has also improved. An ongoing world-wide effort, the Cochrane Collaboration, is systematically assembling and synthesizing several hundred thousand randomized, controlled trials to improve the delivery of health care. Meta-analysis has many important advantages. It allows the viewing of the complete picture of the evidence. The advent of meta-analysis has sensitized researchers to issues of quality and has improved methodology in clinical research. Detection and explanation of bias and heterogeneity are prime objectives of meta-analysis in clinical research. An array of methods has been developed that allows a better understanding of bias and heterogeneity, beyond simple averaging of results from diverse studies. Meta-analyses of individual patient data, in particular, may promote the development of international collaborations. Several examples of their application are already available in oncology. Meta-analysis may point out deficiencies in the study design of past and current studies, suggest the need for new studies, and inform researchers about the size and design of these studies. In the end, meta-analysis helps to integrate evidence and make recommendations for medical care and medical practice.

The development of policies to prevent nosocomial transmission of hepatitis C virus (HCV) infection in hemodialysis units is critically dependent on the understanding of the relationship between tests for anti-HCV, HCV RNA, and HCV genotype and the patients' clinical characteristics. We tested sera from all patients on the renal transplant waiting list at the New England Organ Bank between November 1986 and June 1990 for anti-HCV by a third-generation enzyme-linked immunosorbent assay (ELISA3) and a third-generation recombinant immunoblot assay (RIBA3). All ELISA3-positive sera were tested for HCV RNA by reverse transcriptase "nested" polymerase chain reaction, and the genotype was characterized by restriction fragment length polymorphism. Sera were available in 1,544 of 3,243 (48%) patients on the waiting list, of whom 287 (19%) tested positive for anti-HCV by ELISA3. Two hundred eighty-six randomly selected, anti-HCV-negative patients served as controls. Compared with anti-HCV-negative controls, anti-HCV-positive patients had a longer duration since initiation of renal replacement therapy, higher number of previous kidney transplants and blood transfusions, higher proportion of patients with anti-HBc, history of liver disease, history of non-A, non-B hepatitis, and elevated serum alanine aminotransferase, and lower serum albumin concentrations. Of the 287 anti-HCV-positive sera, 261 (91%) were reactive by RIBA3, 21 (7%) were indeterminate, and five (2%) were nonreactive. HCV RNA was detected in 224 of 275 (81%) ELISA3-positive patients, in whom additional sera were available. There were no significant differences in clinical or laboratory characteristics between ELISA3-positive patients with and without HCV RNA. Genotypes 1a, 1b, 2a, 2b, 3a, and 4 were present in 53%, 23%, 8%, 10%, 4%, and 2% of patients, respectively. Infection with one, two, or three different HCV genotypes was present in 92%, 7%, and 1%, respectively. There was no significant association between the type or number of HCV genotypes and RIBA3 reactivity. There were no major differences in clinical or laboratory characteristics between genotypes or between single and mixed infection. In summary, this study provides detailed information regarding the relationship between tests for anti-HCV, HCV RNA, and HCV genotypes and the clinical and laboratory characteristics of a large, well-characterized cohort of patients referred for renal transplant. (Am J Kidney Dis 1998 Jun;31(6):920-7)

Effect of hepatitis C infection and renal transplantation on survival in end-stage renal disease. Hepatitis C virus (HCV) infection is common among patients with end-stage renal disease (ESRD). However, the effect of HCV infection on survival among ESRD patients, and the impact of renal transplantation on the course of HCV infection has not been adequately defined. Sera from patients on the renal transplant waiting list at the New England Organ Bank between November 1986 and June 1990 were tested for anti-HCV using a third generation ELISA. All anti-HCV positive patients and a 1:1 ratio of randomly selected anti-HCV negative patients comprised the study sample. Duration of follow-up was calculated from the date of the first available serum specimen until death, loss to follow-up or December 31, 1995, whichever occurred earlier. Multivariate analysis of risk factors for mortality was performed using a Cox proportional hazards model which included anti-HCV as a time-independent (baseline) variable, transplantation as a time-dependent (follow-up) variable, and independently significant baseline covariates. Anti-HCV was detected in 287 (19%) of 1544 patients in whom sera were available, and 286 anti-HCV negative patients served as controls. Complete information was available in 496 (87%) of these 573 patients. Median follow-up was 73 months (range 1 to 110 months), during which time 302 (61%) patients underwent renal transplantation and 154 (31%) patients died. For anti-HCV positive patients compared to anti-HCV negative patients, the relative risk of death (and 95% confidence intervals) from all causes was 1.41 (1.01 to 1.97) and due to liver disease or infection was 2.39 (1.28 to 4.48). For patients who underwent transplantation compared to those who remained on dialysis, the relative risk of death from all causes between 0 to 3 months, 3 to 6 months, seven months to four years, and after four years was 4.75 (2.76 to 8.17), 1.76 (0.75 to 4.13), 0.31 (0.18 to 0.54) and 0.84 (0.51 to 1.37), respectively. There was no interaction between the effect of anti-HCV status at baseline and subsequent transplantation (P = 0.93), meaning that the association between treatment modality and survival was similar among anti-HCV positive and negative patients, at all intervals after transplantation. We conclude that HCV infection at the time of referral for transplantation is associated with an increased risk of death, irrespective of whether patients remain on dialysis or undergo transplantation. Transplantation has a beneficial rather than adverse effect on long-term survival in anti-HCV positive patients. Hence, anti-HCV positive status alone is not a contraindication for renal transplantation.

Interleukin-1 receptor antagonist synthesis by peripheral blood mononuclear cells in hemodialysis patients. Background Pro-inflammatory cytokines like interleukin (IL)-1β and tumor necrosis factor-α (TNF-α) are believed to play a significant role in dialysis-related morbidity. It has been previously demonstrated that the endogenous synthesis of interleukin-1 receptor antagonist (IL-1Ra) is a reliable marker of the level of IL-1β synthesis in hemodialysis (HD) patients. In this study, we assessed the impact of clinical and laboratory variables on IL-1Ra synthesis by peripheral blood mononuclear cells (PBMC) in patients on HD with unsubstituted cellulose dialyzers. Methods IL-1Ra by PBMC was measured by a specific non-cross-reactive radioimmunoassay. Day to day variation in cytokine synthesis, the correlation between cytokine synthesis under different in vitro stimulatory conditions, and the influence of clinical and laboratory variables on cytokine synthesis were studied. Results Although there was a trend towards greater IL-1Ra synthesis by unstimulated, endotoxin-stimulated and IgG-stimulated PBMC drawn before the second and third dialysis sessions of the week when compared to the first dialysis treatment, this was not statistically significant. There was a strong correlation between IL-1Ra synthesis by PBMC cultured under different stimulatory conditions that was best observed between IL-1Ra cell content and from endotoxin-stimulated PBMC ( r = 0.51, P = 0.0001), and endotoxin- and IgG-stimulated PBMC ( r = 0.44, P = 0.0001). In addition, there was a close correlation between total synthesis (cell associated and secreted) and secreted levels of IL-1Ra in unstimulated ( r = 0.59, P = 0.0001) and endotoxin-stimulated PBMC ( r = 0.69, P = 0.0001). Interestingly, there was an inverse correlation between IL-1Ra synthesis and duration of dialysis that was strongest for secreted IL-1Ra from unstimulated ( r = -0.50, P = 0.002) and endotoxin-stimulated PBMC ( r = -0.34, P = 0.04). There was no significant correlation between IL-1Ra synthesis by PBMC and other clinical and laboratory indices. Conclusions The observations from this study indicate that: ( 1 ) in HD patients, there were no significant differences in cytokine synthesis by PBMC drawn before the three different dialysis treatments during the week; ( 2 ) there is a close relationship between IL-1Ra synthesis from PBMC cultured under different stimulatory conditions; ( 3 ) the secreted levels of IL-1Ra correlate directly with total synthesis (cell-associated and secreted); ( 4 ) with the exception of duration of dialysis, none of the other clinical or laboratory parameters correlated with cytokine synthesis; and ( 5 ) the diminished endotoxin- or IgG-stimulated IL-1Ra synthesis with increasing time on dialysis is possibly another sign of the impaired host-defense system in patients on long-term hemodialysis.

Are meta-analyses the brave new world, or are the critics of such combined analyses right to say that the biases inherent in clinical trials make them uncombinable? Negative trials are often unreported, and hence can be missed by meta-analysts. And how much heterogeneity between trials is acceptable? A recent major criticism is that large randomised trials do not always agree with a prior meta-analysis. Neither individual trials nor meta-analyses, reporting as they do on population effects, tell how to treat the individual patient. Here we take a more rounded approach to meta-analyses, arguing that their strengths outweigh their weaknesses, although the latter must not be brushed aside.

When outcomes occur in clinical trials before treatment can be given, neither intent-to-treat nor according-to-protocol analyses give optimal estimates of the treatment effect. A better approach employs a time-dependent variable for treatment. Intent-to-treat analyses are conservative, biasing against treatment; according-to-protocol analyses bias in favor of treatment. We show how to measure the effect of a time-dependent variable in a logistic regression using person-time intervals as units of measurement and describe appropriate methods for reporting model performance. The method is applied to develop a model to predict the probability that a patient with a myocardial infarction will have a sudden cardiac arrest within 48 hours of presentation to emergency medical services both when treated with thrombolysis and when not treated. We use a time-dependent treatment variable because many patients went into cardiac arrest while awaiting treatment. This technique has been programmed into an electrocardiograph for real-time use in an emergency department.

Hepatitis C virus (HCV) can be transmitted by transplantation of cadaver organs from donors with antibody to HCV (anti-HCV); therefore, transplantation of organs from anti-HCV positive donors to anti-HCV-negative recipients has been discouraged. We have looked at outcomes in recipients of organs from anti-HCV positive and negative donors to determine whether this advice is well-founded. Stored sera from 716 consecutive cadaver organ donors procured by the New England Organ Bank between 1986 and 1990 were tested for anti-HCV by a first-generation ELISA (ELISA1); 13 (1·8%) were positive. 29 recipients who received organs from these donors were the study group. 37 donors were randomly selected from 703 ELISA1-negative cadaver organ donors. 74 recipients of organs from these 37 donors were the control group. Clinical records were reviewed and recipient sera were tested for anti-HCV with a second-generation ELISA (ELISA2), and HCV RNA was tested for by polymerase chain reaction. Median post-transplant follow-up was 42 and 49 months for study and control groups. Post-transplantation prevalence of anti-HCV and HCV RNA was 67% and 96% among recipients from anti-HCV-positive donors, and 20% and 18% among recipients from anti-HCV-negative donors (p

Rosner et al. ( Stat. Med. 4 (1985), 457–467) discussed the use of a first-order autoregressive model for longitudinal data with multiple time-dependent and time-independent covariates. The model assumes that both covariate and outcome variables are measured without error, a condition frequently not satisfied in practice. Here we extend their model to allow for measurement error on both covariate and outcome variables. We obtain maximum likelihood estimates of the model parameters given knowledge of the measurement error variance for both outcome and exposure variables. Our method differs from standard linear model measurement error methods which do not apply to autoregressive models. A simulation study comparing the maximum likelihood estimates (MLEs) with the ordinary least square (OLS) estimates that ignore measurement error shows that the OLS estimates are grossly biased in the presence of substantial measurement error, while the MLEs are essentially unbiased. Furthermore, whereas the coverage probability of the confidence intervals based on the OLS estimates tends to be significantly smaller than the nominal level, the MLEs provide adequate coverage. The relative advantage of MLE versus OLS increases as either the amount of measurement error or the number of visits increases. Finally, we demonstrate that the method corrects for a severe underestimate of the effect of airway responsiveness on pulmonary function decline when both are measured with error.

Objective. —To evaluate the results of large clinical trials vs the pooled results of smaller trials. Data Identification. —Mata-analyses with at least 1 "large" study were identified from the Cochrane Pregnancy and Childbirth Database and from MEDLINE (1966-1995). Study Selection. —We used a sample size approach to select 79 meta-analyses with at least 1 large study of 1000 or more patients. We used a statistical power approach to select 61 meta-analyses with at least 1 large study based on statistical power considerations. Data Extraction. —The outcome of interest for each meta-analysis was the primary one stated in the original publication or, when not clearly specified, was decided on clinically. Data Synthesis. —By random effects calculations, we found agreement between large and smaller trials in 90% of the meta-analyses selected by the sample size approach and in 82% of the meta-analyses selected by the statistical power approach. Twice as many disagreements appeared when the variability among large studies and among smaller studies was not considered (ie, fixed effects calculations). Of the 15 disagreements between results of large and smaller trials using the random effects model, plausible explanations were identified in 10 meta-analyses: 5 with differences in the control rate of events between large and smaller trials, 4 with specific protocol or study differences, and 1 with potential publication bias. Two other disagreements were not clinically important, and tentative reasons could be identified for 2 of the remaining 3 disagreements. Conclusion. —Results of smaller studies are usually compatible with the results of large studies, but discrepancies do occur even when the diversity among both large studies and smaller studies is considered. Clinically important differences without a potential explanation are extremely uncommon. Future research should further examine sources of heterogeneity between the results of large and smaller trials.