Petra Macaskill

Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials.

Funnel plots, and tests for funnel plot asymmetry, have been widely used to examine bias in the results of meta-analyses. Funnel plot asymmetry should not be equated with publication bias, because it has a number of other possible causes. This article describes how to interpret funnel plot asymmetry, recommends appropriate tests, and explains the implications for choice of meta-analysis model

Accuracy and Surgical Impact of Magnetic Resonance Imaging in Breast Cancer Staging: Systematic Review and Meta-Analysis in Detection of Multifocal and Multicentric Cancer

PURPOSE We review the evidence on magnetic resonance imaging (MRI) in staging the affected breast to determine its accuracy and impact on treatment. METHODS Systematic review and meta-analysis of the accuracy of MRI in detection of multifocal (MF) and/or multicentric (MC) cancer not identified on conventional imaging. We estimated summary receiver operating characteristic curves, positive predictive value (PPV), true-positive (TP) to false positive (FP) ratio, and examined their variability according to quality criteria. Pooled estimates of the proportion of women whose surgery was altered were calculated. Results Data from 19 studies showed MRI detects additional disease in 16% of women with breast cancer (N = 2,610). MRI incremental accuracy differed according to the reference standard (RS; P = .016) decreasing from 99% to 86% as the quality of the RS increased. Summary PPV was 66% (95% CI, 52% to 77%) and TP:FP ratio was 1.91 (95% CI, 1.09 to 3.34). Conversion from wide local excision (WLE) to mastectomy was 8.1% (95% CI, 5.9 to 11.3), from WLE to more extensive surgery was 11.3% in MF/MC disease (95% CI, 6.8 to 18.3). Due to MRI-detected lesions (in women who did not have additional malignancy on histology) conversion from WLE to mastectomy was 1.1% (95% CI, 0.3 to 3.6) and from WLE to more extensive surgery was 5.5% (95% CI, 3.1 to 9.5). CONCLUSION MRI staging causes more extensive breast surgery in an important proportion of women by identifying additional cancer, however there is a need to reduce FP MRI detection. Randomized trials are needed to determine the clinical value of detecting additional disease which changes surgical treatment in women with apparently localized breast cancer.

Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration.

In medicine, numerous decisions are made by care providers, often in shared decision making, on the basis of an estimated probability that a specific disease or condition is present (diagnostic setting) or a specific event will occur in the future (prognostic setting) in an individual. In the diagnostic setting, the probability that a particular disease is present can be used, for example, to inform the referral of patients for further testing, to initiate treatment directly, or to reassure patients that a serious cause for their symptoms is unlikely. In the prognostic context, predictions can be used for planning lifestyle or therapeutic decisions on the basis of the risk for developing a particular outcome or state of health within a specific period (13). Such estimates of risk can also be used to risk-stratify participants in therapeutic intervention trials (47). In both the diagnostic and prognostic setting, probability estimates are commonly based on combining information from multiple predictors observed or measured from an individual (1, 2, 810). Information from a single predictor is often insufficient to provide reliable estimates of diagnostic or prognostic probabilities or risks (8, 11). In virtually all medical domains, diagnostic and prognostic multivariable (risk) prediction models are being developed, validated, updated, and implemented with the aim to assist doctors and individuals in estimating probabilities and potentially influence their decision making. A multivariable prediction model is a mathematical equation that relates multiple predictors for a particular individual to the probability of or risk for the presence (diagnosis) or future occurrence (prognosis) of a particular outcome (10, 12). Other names for a prediction model include risk prediction model, predictive model, prognostic (or prediction) index or rule, and risk score (9). Predictors are also referred to as covariates, risk indicators, prognostic factors, determinants, test results, ormore statisticallyindependent variables. They may range from demographic characteristics (for example, age and sex), medical historytaking, and physical examination results to results from imaging, electrophysiology, blood and urine measurements, pathologic examinations, and disease stages or characteristics, or results from genomics, proteomics, transcriptomics, pharmacogenomics, metabolomics, and other new biological measurement platforms that continuously emerge. Diagnostic and Prognostic Prediction Models Multivariable prediction models fall into 2 broad categories: diagnostic and prognostic prediction models (Box A). In a diagnostic model, multiplethat is, 2 or morepredictors (often referred to as diagnostic test results) are combined to estimate the probability that a certain condition or disease is present (or absent) at the moment of prediction (Box B). They are developed from and to be used for individuals suspected of having that condition. Box A. Schematic representation of diagnostic and prognostic prediction modeling studies. The nature of the prediction in diagnosis is estimating the probability that a specific outcome or disease is present (or absent) within an individual, at this point in timethat is, the moment of prediction (T= 0). In prognosis, the prediction is about whether an individual will experience a specific event or outcome within a certain time period. In other words, in diagnostic prediction the interest is in principle a cross-sectional relationship, whereas prognostic prediction involves a longitudinal relationship. Nevertheless, in diagnostic modeling studies, for logistical reasons, a time window between predictor (index test) measurement and the reference standard is often necessary. Ideally, this interval should be as short as possible without starting any treatment within this period. Box B. Similarities and differences between diagnostic and prognostic prediction models. In a prognostic model, multiple predictors are combined to estimate the probability of a particular outcome or event (for example, mortality, disease recurrence, complication, or therapy response) occurring in a certain period in the future. This period may range from hours (for example, predicting postoperative complications [13]) to weeks or months (for example, predicting 30-day mortality after cardiac surgery [14]) or years (for example, predicting the 5-year risk for developing type 2 diabetes [15]). Prognostic models are developed and are to be used in individuals at risk for developing that outcome. They may be models for either ill or healthy individuals. For example, prognostic models include models to predict recurrence, complications, or death in a certain period after being diagnosed with a particular disease. But they may also include models for predicting the occurrence of an outcome in a certain period in individuals without a specific disease: for example, models to predict the risk for developing type 2 diabetes (16) or cardiovascular events in middle-aged nondiseased individuals (17), or the risk for preeclampsia in pregnant women (18). We thus use prognostic in the broad sense, referring to the prediction of an outcome in the future in individuals at risk for that outcome, rather than the narrower definition of predicting the course of patients who have a particular disease with or without treatment (1). The main difference between a diagnostic and prognostic prediction model is the concept of time. Diagnostic modeling studies are usually cross-sectional, whereas prognostic modeling studies are usually longitudinal. In this document, we refer to both diagnostic and prognostic prediction models as prediction models, highlighting issues that are specific to either type of model. Development, Validation, and Updating of Prediction Models Prediction model studies may address the development of a new prediction model (10), a model evaluation (often referred to as model validation) with or without updating of the model [1921]), or a combination of these (Box C and Figure 1). Box C. Types of prediction model studies. Figure 1. Types of prediction model studies covered by the TRIPOD statement. D = development data; V = validation data. Model development studies aim to derive a prediction model by selecting predictors and combining them into a multivariable model. Logistic regression is commonly used for cross-sectional (diagnostic) and short-term (for example 30-day mortality) prognostic outcomes and Cox regression for long-term (for example, 10-year risk) prognostic outcomes. Studies may also focus on quantifying the incremental or added predictive value of a specific predictor (for example, newly discovered) (22) to a prediction model. Quantifying the predictive ability of a model on the same data from which the model was developed (often referred to as apparent performance [Figure 1]) will tend to give an optimistic estimate of performance, owing to overfitting (too few outcome events relative to the number of candidate predictors) and the use of predictor selection strategies (2325). Studies developing new prediction models should therefore always include some form of internal validation to quantify any optimism in the predictive performance (for example, calibration and discrimination) of the developed model and adjust the model for overfitting. Internal validation techniques use only the original study sample and include such methods as bootstrapping or cross-validation. Internal validation is a necessary part of model development (2). After developing a prediction model, it is strongly recommended to evaluate the performance of the model in other participant data than was used for the model development. External validation (Box C and Figure 1) (20, 26) requires that for each individual in the new participant data set, outcome predictions are made using the original model (that is, the published model or regression formula) and compared with the observed outcomes. External validation may use participant data collected by the same investigators, typically using the same predictor and outcome definitions and measurements, but sampled from a later period (temporal or narrow validation); by other investigators in another hospital or country (though disappointingly rare [27]), sometimes using different definitions and measurements (geographic or broad validation); in similar participants, but from an intentionally different setting (for example, a model developed in secondary care and assessed in similar participants, but selected from primary care); or even in other types of participants (for example, model developed in adults and assessed in children, or developed for predicting fatal events and assessed for predicting nonfatal events) (19, 20, 26, 2830). In case of poor performance (for example, systematic miscalibration), when evaluated in an external validation data set, the model can be updated or adjusted (for example, recalibrating or adding a new predictor) on the basis of the validation data set (Box C) (2, 20, 21, 31). Randomly splitting a single data set into model development and model validation data sets is frequently done to develop and validate a prediction model; this is often, yet erroneously, believed to be a form of external validation. However, this approach is a weak and inefficient form of internal validation, because not all available data are used to develop the model (23, 32). If the available development data set is sufficiently large, splitting by time and developing a model using data from one period and evaluating its performance using the data from the other period (temporal validation) is a stronger approach. With a single data set, temporal splitting and model validation can be considered intermediate between internal and external validation. Incomplete and Inaccurate Reporting Prediction models are becoming increasingly abundant in the medical literature (9, 33, 34), and policymakers are incre

Meta-analytic methods for diagnostic test accuracy.

Meta-analyses of diagnostic test accuracy are uncommon and often based on separate pooling of sensitivity and specificity, which can lead to biased estimates. Recently, several appropriate methods have been developed for meta-analysing diagnostic test data from primary studies. Primary studies usually only provide binary test data, for which Moses et al. have developed a method to estimate Summary Receiver Operating Characteristic Curves, thereby taking account of possible test threshold differences between studies. Several methods are also available for analysing multicategory and continuous test data. The usefulness of applying these methods is constrained by publication bias and the generally poor quality of primary studies of diagnostic test accuracy. Meta-analysts need to highlight important defects in quality and how they affect summary estimates to ensure that better primary studies are available for meta-analysis in the future.

Serum Levels of Phosphorus, Parathyroid Hormone, and Calcium and Risks of Death and Cardiovascular Disease in Individuals With Chronic Kidney Disease: A Systematic Review and Meta-analysis

CONTEXT Clinical practice guidelines on the management of mineral and bone disorders due to chronic kidney disease recommend specific treatment target levels for serum phosphorus, parathyroid hormone, and calcium. OBJECTIVE To assess the quality of evidence for the association between levels of serum phosphorus, parathyroid hormone, and calcium and risks of death, cardiovascular mortality, and nonfatal cardiovascular events in individuals with chronic kidney disease. DATA SOURCES The databases of MEDLINE (1948 to December 2010) and EMBASE (1947 to December 2010) were searched without language restriction. Hand searches also were conducted of the reference lists of primary studies, review articles, and clinical guidelines along with full-text review of any citation that appeared relevant. STUDY SELECTION Of 8380 citations identified in the original search, 47 cohort studies (N = 327,644 patients) met the inclusion criteria. DATA EXTRACTION The characteristics of study design, participants, exposures, and covariates together with the outcomes of all-cause mortality, cardiovascular mortality, and nonfatal cardiovascular events at different levels of serum phosphorus, parathyroid hormone, and calcium were analyzed within studies. Data were summarized across studies (when possible) using random-effects meta-regression. DATA SYNTHESIS The risk of death increased 18% for every 1-mg/dL increase in serum phosphorus (relative risk [RR], 1.18 [95% confidence interval {CI}, 1.12-1.25]). There was no significant association between all-cause mortality and serum level of parathyroid hormone (RR per 100-pg/mL increase, 1.01 [95% CI, 1.00-1.02]) or serum level of calcium (RR per 1-mg/dL increase, 1.08 [95% CI, 1.00-1.16]). Data for the association between serum level of phosphorus, parathyroid hormone, and calcium and cardiovascular death were each available in only 1 adequately adjusted cohort study. Lack of adjustment for confounding variables was not a major limitation of the available studies. CONCLUSIONS The evidentiary basis for a strong, consistent, and independent association between serum levels of calcium and parathyroid hormone and the risk of death and cardiovascular events in chronic kidney disease is poor. There appears to be an association between higher serum levels of phosphorus and mortality in this population.

Integration of 3D digital mammography with tomosynthesis for population breast-cancer screening (STORM): a prospective comparison study

BACKGROUND Digital breast tomosynthesis with 3D images might overcome some of the limitations of conventional 2D mammography for detection of breast cancer. We investigated the effect of integrated 2D and 3D mammography in population breast-cancer screening. METHODS Screening with Tomosynthesis OR standard Mammography (STORM) was a prospective comparative study. We recruited asymptomatic women aged 48 years or older who attended population-based breast-cancer screening through the Trento and Verona screening services (Italy) from August, 2011, to June, 2012. We did screen-reading in two sequential phases-2D only and integrated 2D and 3D mammography-yielding paired data for each screen. Standard double-reading by breast radiologists determined whether to recall the participant based on positive mammography at either screen read. Outcomes were measured from final assessment or excision histology. Primary outcome measures were the number of detected cancers, the number of detected cancers per 1000 screens, the number and proportion of false positive recalls, and incremental cancer detection attributable to integrated 2D and 3D mammography. We compared paired binary data with McNemars test. FINDINGS 7292 women were screened (median age 58 years [IQR 54-63]). We detected 59 breast cancers (including 52 invasive cancers) in 57 women. Both 2D and integrated 2D and 3D screening detected 39 cancers. We detected 20 cancers with integrated 2D and 3D only versus none with 2D screening only (p<0.0001). Cancer detection rates were 5.3 cancers per 1000 screens (95% CI 3.8-7.3) for 2D only, and 8.1 cancers per 1000 screens (6.2-10.4) for integrated 2D and 3D screening. The incremental cancer detection rate attributable to integrated 2D and 3D mammography was 2.7 cancers per 1000 screens (1.7-4.2). 395 screens (5.5%; 95% CI 5.0-6.0) resulted in false positive recalls: 181 at both screen reads, and 141 with 2D only versus 73 with integrated 2D and 3D screening (p<0.0001). We estimated that conditional recall (positive integrated 2D and 3D mammography as a condition to recall) could have reduced false positive recalls by 17.2% (95% CI 13.6-21.3) without missing any of the cancers detected in the study population. INTERPRETATION Integrated 2D and 3D mammography improves breast-cancer detection and has the potential to reduce false positive recalls. Randomised controlled trials are needed to compare integrated 2D and 3D mammography with 2D mammography for breast cancer screening. FUNDING National Breast Cancer Foundation, Australia; National Health and Medical Research Council, Australia; Hologic, USA; Technologic, Italy.

Dermoscopy compared with naked eye examination for the diagnosis of primary melanoma: a meta‐analysis of studies performed in a clinical setting

Background  Dermoscopy is a noninvasive technique that enables the clinician to perform direct microscopic examination of diagnostic features, not seen by the naked eye, in pigmented skin lesions. Diagnostic accuracy of dermoscopy has previously been assessed in meta‐analyses including studies performed in experimental and clinical settings.

Meta-analysis of the impact of surgical margins on local recurrence in women with early-stage invasive breast cancer treated with breast-conserving therapy

PURPOSE There is no consensus on what constitutes adequate negative margins in breast-conserving therapy (BCT). We review the evidence on surgical margins in BCT for early-stage invasive breast cancer. METHODS Meta-analysis of studies reporting local recurrence (LR) relative to quantified final microscopic margin status and the threshold distance for negative margins. The proportion of LR was modelled using random effects logistic meta-regression. RESULTS Based on 21 studies (LR in 1,026 of 14,571 subjects) the odds of LR were associated with margin status [model 1: odds ratio (OR) = 2.02 for positive/close versus negative; model 2: OR = 1.80 for close versus negative, 2.42 for positive versus negative (P<0.001 both models)] but not with margin distance [1mm versus 2mm versus 5mm (P > 0.10 both models)], adjusting for median follow-up time. However, there was weak evidence in both models that the odds of LR decreased as the threshold distance for declaring negative margins increased. This bordered significance in model 2 [OR for 1mm, 2mm, 5mm: 1.0, 0.75, 0.51 (P = 0.097 for trend)], and was not significant in model 1 [OR for 1mm, 2mm, 5mm: 1.0, 0.85, 0.58 (P = 0.11 for trend)] but was evident when one study (of women ≤ 40 years) was excluded from this model [OR for 1mm, 2mm, 5mm: 1.0, 0.72, 0.52 (P = 0.058 for trend)]: this trend was rendered insignificant by adjustment for the proportion of subjects receiving a radiation boost or the proportion of subjects receiving endocrine therapy. CONCLUSIONS Margin status has a prognostic effect in all women treated for invasive breast cancer; increasing the threshold distance for declaring negative margins is weakly associated with reduced odds of LR, however adjustment for covariates (adjuvant therapy) removes the significance of this effect. Adoption of wider margins, relative to narrower widths, for declaring negative margins is unlikely to a have substantial additional benefit for long-term local control in BCT.

Effect of study design and quality on unsatisfactory rates, cytology classifications, and accuracy in liquid-based versus conventional cervical cytology: a systematic review

BACKGROUND Liquid-based cytology is reported to increase the sensitivity of cervical cytology and the proportion of slides that are satisfactory for assessment, in comparison with conventional cytology. Although some countries have changed to liquid-based cytology for cervical screening, controversy remains. We reviewed the published work to assess the performance of liquid-based cytology relative to conventional cytology in primary studies assessed to be of low, medium, or high methodological quality. METHODS 56 primary studies were reviewed and assessed with strict methodological criteria. Liquid-based cytology and conventional cytology were compared in terms of the percentage of slides classified as unsatisfactory, the percentage of slides classified in each cytology category, and the accuracy of detection of high-grade disease. Data were examined for studies overall and in strata to examine the effect of study quality on results. FINDINGS The median difference in the percentage of unsatisfactory slides between liquid-based cytology and conventional cytology was 0.17%. Only one small study was a randomised controlled trial. The classification of high-grade squamous epithelial lesion varied according to study quality (p=0.04), with conventional cytology classifying more slides in this category than did liquid-based cytology in high-quality studies (n=3) only. In medium-quality (n=30) and high-quality studies, liquid-based cytology classified more slides as atypical squamous cells of unknown significance than did conventional cytology when compared with low-quality studies (n=17; p=0.05). Only four studies provided sufficient verified data to allow estimation of sensitivity and specificity and comparison of test accuracy. INTERPRETATION We saw no evidence that liquid-based cytology reduced the proportion of unsatisfactory slides, or detected more high-grade lesions in high-quality studies, than conventional cytology. This review does not lend support to claims of better performance by liquid-based cytology. Large randomised controlled trials are needed.

Meta-analysis: vitamin D compounds in chronic kidney disease.

Context Clinicians often treat patients with kidney disease with vitamin D compounds to prevent secondary hyperparathyroidism. Contribution This meta-analysis of 76 randomized trials found no good evidence that vitamin D compounds reduced risk for death, bone pain, vascular calcification, or need for parathyroidectomy in patients with chronic kidney disease. Compared with placebo, established vitamin D sterols increased risk for hypercalcemia and hyperphosphatemia, whereas newer vitamin D analogues increased hypercalcemia but not hyperphosphatemia. Direct comparisons found no clear benefits of newer analogues over established agents. Implication Though commonly used, vitamin D compounds for chronic kidney disease have unclear benefits and potential harms. The Editors All stages of chronic kidney disease (CKD) are associated with significantly increased rates of all-cause and cardiovascular mortality (1). Several risk factors for death have been identified and targeted by interventions, but registry data have not shown substantial improvements in survival of people with end-stage kidney disease over the past 2 decades (2). Abnormalities of bone metabolism and mineralization, which are risk factors for death in CKD, occur early and become universal as kidney function declines (3). A frequent pattern of biochemical abnormalities includes increased serum phosphorus and parathyroid hormone (PTH) levels, whereas levels of serum calcium may be low, normal, or elevated. These changes are associated with alterations in bone mineral homeostasis, increased bone fragility (4, 5), vascular and soft tissue calcification (6, 7), muscle dysfunction (8), adverse cardiovascular outcomes, and increased mortality (9). Compared with PTH levels of 16.5 to 33.0 pmol/L (150 to 300 pg/mL), levels greater than 66 pmol/L (600 pg/mL) are reported to be associated with a 10% increased risk for death (10). Similar mortality data have been observed for increased serum phosphorus and calcium levels (10). Interventions that are widely used to improve biochemical markers of bone and mineral metabolism include active vitamin D compounds, calcium supplements and noncalcium-containing phosphate binders, and calcimimetics. Vitamin D therapy has historically been based on alfacalcidol (1 -hydroxyvitamin D3) or calcitriol, both of which attenuate secondary hyperparathyroidism (1114). Although these compounds may reduce PTH levels, they increase calcium and phosphorus levels (9, 10, 15, 16). Support for use of the newer vitamin D analogues (22-oxacalcitriol, doxercalciferol, paricalcitol, and falecalcitriol) is based on reports of similar or superior dose-equivalent suppression of PTH, less calcemic and phosphatemic activity, and the possibility of improved survival when compared with established vitamin D sterols (calcitriol or alfacalcidol) (17, 18). Guidelines have suggested that doxercalciferol and paricalcitol may be preferable to calcitriol and alfacalcidol (19). We evaluated available randomized trials to determine the effects of established vitamin D sterols and newer analogues on biochemical, bone, and cardiovascular end points in CKD together with their optimal dose and route of administration. Methods Data Sources and Searches Literature searches for randomized, controlled trials (RCTs) of vitamin D sterols in CKD were performed in MEDLINE (January 1966 to July 2007) and EMBASE (January 1980 to July 2007) using optimally sensitive search strategies (20). The Cochrane Renal Group Renal Health Library and the Cochrane Central Register of Controlled Trials (CENTRAL) were also searched. The complete search strategy is outlined in Appendix Table 1. Authors followed a standardized published protocol for identification of eligible trials (21). Appendix Table 1. Summary of Search Strategy Study Selection We included randomized and quasi-randomized, controlled trials conducted in patients with CKD and that compared vitamin D compounds with placebo, different vitamin D compounds directly, and different vitamin D dose and administration regimens. Studies enrolling patients with any stage of CKD and measuring the effect of these agents on surrogate biochemical end points at the end of treatment (for example, levels of PTH, calcium, phosphorus, and calciumphosphorus product) and hard patient-level end points (for example, all-cause and cardiovascular mortality, fracture, toxicity) were included. We excluded trials enrolling only participants who had parathyroidectomy or kidney transplantation. We also excluded RCTs of vitamin D compounds in osteoporosis because results of these studies were presented without reference to kidney function or CKD was an exclusion criterion. Data Extraction and Quality Assessment Two independent authors assessed each trial. They extracted data on the characteristics of participants, interventions, comparisons, and clinical outcomes, when reported. Hypercalcemia was defined as a serum calcium level of 2.63 mmol/L or greater (10.5 mg/dL), and hyperphosphatemia was defined as a serum phosphorus level greater than 1.62 mmol/L (>5.0 mg/dL). Because trial investigators generally did not report change in values from beginning to end of treatment for continuous variables, we only considered the end-of-treatment values. Where published outcome data were not provided in sufficient detail, an author contacted the trial investigators by electronic or standard mail requesting additional information. Review authors resolved discrepancies in data extraction and quality assessment through discussion. Data Synthesis and Analysis We summarized treatment effects as relative risks (RRs) for categorical variables and weighted mean differences for continuous variables, with 95% CIs. We pooled estimates from individual trials by using the DerSimonian and Laird random-effects model (22). We repeated all analyses by adding 1/n to treatment groups with zero events and using the odds ratio as the measure of effect. Neither method resulted in substantive differences in any clinical outcome. We formally assessed heterogeneity of treatment effects between studies with the Cochran Q and the I 2 statistics (23). We performed subgroup analysis and random-effects metaregression to explore the effect of potential sources of variability on observed treatment effects. We investigated the impacts of the following plausible effect modifiers on treatment outcomes: newer vitamin D analogues versus established vitamin D sterols, baseline PTH concentration (<33 pmol/L, 33 to 66 pmol/L, 66 to 110 pmol/L, and >110 pmol/L), method of PTH assay (amino-terminal, carboxy-terminal, intact, full-length PTH [1-84], or not specified), method of calcium assay (total, corrected, or ionized), baseline serum calcium concentration (<2.63 mmol/L and 2.63 mmol/L [10.5 mg/dL]), baseline serum phosphorus concentration (<1.29 mmol/L and 1.29 mmol/L [4.0 mg/dL]), dialysis modality (peritoneal or hemodialysis) and duration, stage of CKD (CKD stage 3 [estimated glomerular filtration rate, 30 to 59 mL/min per 1.73 m2], 4 [estimated glomerular filtration rate, 15 to 29 mL/min per 1.73 m2], or 5 [estimated glomerular filtration rate <15 mL/min per 1.73 m2]), pediatric versus adult cohorts, use of calcium-based phosphate binders as a co-intervention, prior use of vitamin D sterols, duration of intervention, trial quality, and dose of vitamin D compound (high or low). All analyses were undertaken using RevMan 4.2 (The Cochrane Collaboration, Copenhagen, Denmark) and STATA version 8.0 (STATA, College Station, Texas). Role of the Funding Source The funding source had no role in the study design; collection, analysis, or interpretation of data; or writing of the report. The corresponding author had full access to all study data and had final responsibility for the decision to submit the manuscript for publication. Results The combined search of MEDLINE, EMBASE, the Cochrane Controlled Trial Register, and the Renal Health Library of the Cochrane Renal Group identified a total of 1608 articles (Appendix Figure 1). After full-text analysis, we included 76 eligible RCTs that enrolled a total of 3667 participants with CKD. Two publications each represented combined data from 3 RCTs (24, 25). We identified 3 ongoing studies from trial registries (2628). The number of individuals in each trial ranged from 6 to 266 patients, and 60 out of 76 (79%) studies enrolled fewer than 60 patients. Authors of 18 trials provided additional information, which was included in our analyses (14, 24, 2944). Appendix Figure 1. Study flow diagram. Reasons for exclusions and the number of trials reporting each outcome are provided. RCT = randomized, controlled trial. Trial Characteristics Appendix Tables 2 and 3 list the characteristics of the samples and interventions in the trials of vitamin D included in this meta-analysis. We divided the 76 eligible trials into 3 major groups of studies based on the randomized interventions: vitamin D compounds versus placebo, other vitamin D compounds, or calcium; different routes of administration of vitamin D; and different doses of vitamin D. Appendix Table 2. Characteristics of Populations and Interventions in Trials Comparing Vitamin D Compounds with Placebo, No Treatment, or Other Vitamin D Compounds in People with Chronic Kidney Disease* Appendix Table 3. Characteristics of Populations and Interventions in Trials of Vitamin D and Its Analogues Comparing Different Routes and Schedules of Administration in People with Chronic Kidney Disease* The first group compared established vitamin D compounds with another vitamin D compound, placebo, or calcium alone. This group included 19 trials (981 patients) comparing established vitamin D with placebo, of which 12 trials (669 patients) administered calcitriol (11, 13, 4554), 5 trials (275 patients) administered alfacalcidol (12, 14, 42, 55, 56), and 2 trials (37 patients) administered 24,25-(OH)2 vitamin D3 (31, 44). Fifteen studies (1