Red flags to screen for vertebral fracture in patients presenting with low-back pain.

EDITORIAL NOTE: See https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD014461.pub2/full for a more recent review that covers this topic and has superseded this review. BACKGROUND: Low-back pain (LBP) is a common condition seen in primary care. A principal aim during a clinical examination is to identify patients with a higher likelihood of underlying serious pathology, such as vertebral fracture, who may require additional investigation and specific treatment. All 'evidence-based' clinical practice guidelines recommend the use of red flags to screen for serious causes of back pain. However, it remains unclear if the diagnostic accuracy of red flags is sufficient to support this recommendation. OBJECTIVES: To assess the diagnostic accuracy of red flags obtained in a clinical history or physical examination to screen for vertebral fracture in patients presenting with LBP. SEARCH METHODS: Electronic databases were searched for primary studies between the earliest date and 7 March 2012. Forward and backward citation searching of eligible studies was also conducted. SELECTION CRITERIA: Studies were considered if they compared the results of any aspect of the history or test conducted in the physical examination of patients presenting for LBP or examination of the lumbar spine, with a reference standard (diagnostic imaging). The selection criteria were independently applied by two review authors. DATA COLLECTION AND ANALYSIS: Three review authors independently conducted 'Risk of bias' assessment and data extraction. Risk of bias was assessed using the 11-item QUADAS tool. Characteristics of studies, patients, index tests and reference standards were extracted. Where available, raw data were used to calculate sensitivity and specificity with 95% confidence intervals (CI). Due to the heterogeneity of studies and tests, statistical pooling was not appropriate and the analysis for the review was descriptive only. Likelihood ratios for each test were calculated and used as an indication of clinical usefulness. MAIN RESULTS: Eight studies set in primary (four), secondary (one) and tertiary care (accident and emergency = three) were included in the review. Overall, the risk of bias of studies was moderate with high risk of selection and verification bias the predominant flaws. Reporting of index and reference tests was poor. The prevalence of vertebral fracture in accident and emergency settings ranged from 6.5% to 11% and in primary care from 0.7% to 4.5%. There were 29 groups of index tests investigated however, only two featured in more than two studies. Descriptive analyses revealed that three red flags in primary care were potentially useful with meaningful positive likelihood ratios (LR+) but mostly imprecise estimates (significant trauma, older age, corticosteroid use; LR+ point estimate ranging 3.42 to 12.85, 3.69 to 9.39, 3.97 to 48.50 respectively). One red flag in tertiary care appeared informative (contusion/abrasion; LR+ 31.09, 95% CI 18.25 to 52.96). The results of combined tests appeared more informative than individual red flags with LR+ estimates generally greater in magnitude and precision. AUTHORS' CONCLUSIONS: The available evidence does not support the use of many red flags to specifically screen for vertebral fracture in patients presenting for LBP. Based on evidence from single studies, few individual red flags appear informative as most have poor diagnostic accuracy as indicated by imprecise estimates of likelihood ratios. When combinations of red flags were used the performance appeared to improve. From the limited evidence, the findings give rise to a weak recommendation that a combination of a small subset of red flags may be useful to screen for vertebral fracture. It should also be noted that many red flags have high false positive rates; and if acted upon uncritically there would be consequences for the cost of management and outcomes of patients with LBP. Further research should focus on appropriate sets of red flags and adequate reporting of both index and reference tests.

Factors Associated With Advanced Colorectal Neoplasia in Patients With CKD.

RATIONALE & OBJECTIVE: The risk of developing colorectal cancer in patients with chronic kidney disease (CKD) is twice that of the general population, but the factors associated with colorectal cancer are poorly understood. The aim of this study was to identify factors associated with advanced colorectal neoplasia in patients with CKD. STUDY DESIGN: Prospective cohort study. SETTING & PARTICIPANTS: Patients with CKD stages 3-5, including those treated with maintenance dialysis or transplantation across 11 sites in Australia, New Zealand, Canada, and Spain, were screened for colorectal neoplasia using a fecal immunochemical test (FIT) as part of the Detecting Bowel Cancer in CKD (DETECT) Study. EXPOSURE: Baseline characteristics for patients at the time of study enrollment were ascertained, including duration of CKD, comorbidities, and medications. OUTCOME: Advanced colorectal neoplasia was identified through a 2-step verification process with colonoscopy following positive FIT and 2-year clinical follow-up for all patients. ANALYTICAL APPROACH: Potential factors associated with advanced colorectal neoplasia were explored using multivariable logistic regression. Sensitivity analyses were performed using grouped LASSO (least absolute shrinkage and selection operator) logistic regression. RESULTS: Among 1,706 patients who received FIT-based screening-791 with CKD stages 3-5 not receiving kidney replacement therapy (KRT), 418 receiving dialysis, and 497 patients with a functioning kidney transplant-117 patients (6.9%) were detected to have advanced colorectal neoplasia (54 with CKD stages 3-5 without KRT, 34 receiving dialysis, and 29 transplant recipients), including 9 colorectal cancers. The factors found to be associated with advanced colorectal neoplasia included older age (OR per year older, 1.05 [95% CI, 1.03-1.07], P<0.001), male sex (OR, 2.27 [95% CI, 1.45-3.54], P<0.001), azathioprine use (OR, 2.99 [95% CI, 1.40-6.37], P=0.005), and erythropoiesis-stimulating agent use (OR, 1.92 [95% CI, 1.22-3.03], P=0.005). Grouped LASSO logistic regression revealed similar associations between these factors and advanced colorectal neoplasia. LIMITATIONS: Unmeasured confounding factors. CONCLUSIONS: Older age, male sex, erythropoiesis-stimulating agents, and azathioprine were found to be significantly associated with advanced colorectal neoplasia in patients with CKD.

QUADAS-C: A Tool for Assessing Risk of Bias in Comparative Diagnostic Accuracy Studies.

Comparative diagnostic test accuracy studies assess and compare the accuracy of 2 or more tests in the same study. Although these studies have the potential to yield reliable evidence regarding comparative accuracy, shortcomings in the design, conduct, and analysis may bias their results. The currently recommended quality assessment tool for diagnostic test accuracy studies, QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2), is not designed for the assessment of test comparisons. The QUADAS-C (Quality Assessment of Diagnostic Accuracy Studies-Comparative) tool was developed as an extension of QUADAS-2 to assess the risk of bias in comparative diagnostic test accuracy studies. Through a 4-round Delphi study involving 24 international experts in test evaluation and a face-to-face consensus meeting, an initial version of the tool was developed that was revised and finalized following a pilot study among potential users. The QUADAS-C tool retains the same 4-domain structure of QUADAS-2 (Patient Selection, Index Test, Reference Standard, and Flow and Timing) and comprises additional questions to each QUADAS-2 domain. A risk-of-bias judgment for comparative accuracy requires a risk-of-bias judgment for the accuracy of each test (resulting from QUADAS-2) and additional criteria specific to test comparisons. Examples of such additional criteria include whether participants either received all index tests or were randomly assigned to index tests, and whether index tests were interpreted with blinding to the results of other index tests. The QUADAS-C tool will be useful for systematic reviews of diagnostic test accuracy addressing comparative questions. Furthermore, researchers may use this tool to identify and avoid risk of bias when designing a comparative diagnostic test accuracy study.

Health-Related Quality of Life in People Across the Spectrum of CKD.

INTRODUCTION: People with chronic kidney disease (CKD) experience reduced quality of life (QoL) because of the high symptom and treatment burden. Limited data exist on the factors associated with overall and domain-specific QoL across all CKD stages. METHODS: Using data from a prospective, multinational study (Australia, New Zealand, Canada, and Spain) in 1696 participants with CKD, we measured overall and domain-specific QoL (pain, self-care, activity, mobility, anxiety/depression) using the EuroQoL, 5 dimension, 3 level. Multivariable linear regression and logistic modeling were used to determine factors associated with overall and domain-specific QoL. RESULTS: QoL for patients with CKD stages 3 to 5 (n = 787; mean, 0.81; SD, 0.20) was higher than in patients on dialysis (n = 415; mean, 0.76; SD, 0.24) but lower than in kidney transplant recipients (n = 494; mean, 0.84; SD, 0.21). Factors associated with reduced overall QoL (ß [95% confidence intervals]) included being on dialysis (compared with CKD stages 3-5: -0.06 [-0.08 to -0.03]), female sex (-0.03 [-0.05 to -0.006]), lower educational attainment (- 0.04 [-0.06 to -0.02), lacking a partner (-0.04 [-0.06 to -0.02]), having diabetes (-0.05 [-0.07 to -0.02]), history of stroke (-0.09 [-0.13 to -0.05]), cardiovascular disease (-0.06 [-0.08 to -0.03]), and cancer (-0.03 [-0.06 to -0.009]). Pain (43%) and anxiety/depression (30%) were the most commonly affected domains, with dialysis patients reporting decrements in all 5 domains. Predictors for domain-specific QoL included being on dialysis, presence of comorbidities, lower education, female sex, and lack of a partner. CONCLUSIONS: Being on dialysis, women with CKD, those with multiple comorbidities, lack of a partner, and lower educational attainment were associated with lower QoL across all stages of CKD.

Effectiveness and cost-effectiveness of a progressive, individualised walking and education programme for prevention of low back pain recurrence in adults: study protocol for the WalkBack randomised controlled trial.

INTRODUCTION: Low back pain (LBP) is recognised globally as a prevalent, costly and disabling condition. Recurrences are common and contribute to much of the burden of LBP. Current evidence favours exercise and education for prevention of LBP recurrence, but an optimal intervention has not yet been established. Walking is a simple, widely accessible, low-cost intervention that has yet to be evaluated. This randomised controlled trial (RCT) aims to establish the effectiveness and cost-effectiveness of a progressive and individualised walking and education programme (intervention) for the prevention of LBP recurrences in adults compared with no treatment (control). METHODS AND ANALYSIS: A pragmatic, two-armed RCT comparing walking and education (n=349) with a no treatment control group (n=349). Inclusion criteria are adults recovered from an episode of non-specific LBP within the last 6 months. Those allocated to the intervention group will receive six sessions (three face to face and three telephone delivered) with a trained physiotherapist to facilitate a progressive walking programme and education over a 6-month period. The primary outcome will be days to first recurrence of an episode of activity-limiting LBP. The secondary outcomes include days to recurrence of an episode of LBP, days to recurrence of an episode of LBP leading to care seeking, disability and quality of life measured at 3, 6, 9 and 12 months and costs associated with LBP recurrence. All participants will be followed up monthly for a minimum of 12 months. The primary intention-to-treat analysis will assess difference in survival curves (days to recurrence) using the log-rank statistic. The cost-effectiveness analysis will be conducted from the societal perspective. ETHICS AND DISSEMINATION: Approved by Macquarie University Human Research Ethics Committee (Reference: 5201949218164, May 2019). Findings will be disseminated through publication in peer-reviewed journals and conference presentations. TRIAL REGISTRATION NUMBER: ACTRN12619001134112.

Correction to: Estimation of treatment preference effects in clinical trials when some participants are indifferent to treatment choice.

In the original publication of this article [1], the number "- 0.49" in the below sentence in the Results section should be changed to "-3.23", and this typo does not affect the wider conclusions.

Pilot trial of digital breast tomosynthesis (3D mammography) for population-based screening in BreastScreen Victoria.

OBJECTIVES: To estimate detection measures for tomosynthesis and standard mammography; to assess the feasibility of using tomosynthesis in population-based screening for breast cancer. DESIGN, SETTING: Prospective pilot trial comparing tomosynthesis (with synthesised 2D images) and standard mammography screening of women attending Maroondah BreastScreen, a BreastScreen Victoria service in the eastern suburbs of Melbourne. PARTICIPANTS: Women at least 40 years of age who presented for routine breast screening between 18 August 2017 and 8 November 2018. MAIN OUTCOME MEASURES: Cancer detection rate (CDR); proportion of screens that led to recall for further assessment. RESULTS: 5018 tomosynthesis and 5166 standard mammography screens were undertaken in 10 146 women; 508 women (5.0% of screens) opted not to undergo tomosynthesis screening. With tomosynthesis, 49 cancers (40 invasive, 9 in situ) were detected (CDR, 9.8 [95% CI, 7.2-13] per 1000 screens); with standard mammography, 34 cancers (30 invasive, 4 in situ) were detected (CDR, 6.6 [95% CI, 4.6-9.2] per 1000 screens). The estimated difference in CDR was 3.2 more detections (95% CI, -0.32 to 6.8) per 1000 screens with tomosynthesis; the difference was greater for repeat screens and for women aged 60 years or more. The recall rate was greater for tomosynthesis (4.2%; 95% CI, 3.6-4.8%) than standard mammography (3.0%; 95% CI, 2.6-3.5%; estimated difference, 1.2%; 95% CI, 0.46-1.9%). The median screen reading time for tomosynthesis was 67 seconds (interquartile range [IQR] 46-105 seconds); for standard mammography, 16 seconds (IQR, 10-29 seconds). CONCLUSIONS: Breast cancer detection, recall for assessment, and screen reading time were each higher for tomosynthesis than for standard mammography. Our preliminary findings could form the basis of a large scale comparative evaluation of tomosynthesis and standard mammography for breast screening in Australia. TRIAL REGISTRATION: Australian New Zealand Clinical Trials Registry, ACTRN12617000947303.

One-Time Fecal Immunochemical Screening for Advanced Colorectal Neoplasia in Patients with CKD (DETECT Study).

BACKGROUND: In patients with CKD, the risk of developing colorectal cancer is high and outcomes are poor. Screening using fecal immunochemical testing (FIT) is effective in reducing mortality from colorectal cancer, but performance characteristics of FIT in CKD are unknown. METHODS: To determine the detection rates and performance characteristics of FIT for advanced colorectal neoplasia (ACN) in patients with CKD, we used FIT to prospectively screen patients aged 35-74 years with CKD (stages 3-5 CKD, dialysis, and renal transplant) from 11 sites in Australia, New Zealand, Canada, and Spain. All participants received clinical follow-up at 2 years. We used a two-step reference standard approach to estimate disease status. RESULTS: Overall, 369 out of 1706 patients who completed FIT (21.6%) tested positive; 323 (87.5%) underwent colonoscopies. A total of 1553 (91.0%) completed follow-up; 82 (4.8%) had died and 71 (4.2%) were lost. The detection rate of ACN using FIT was 6.0% (5.6%, 7.4%, and 5.6% for stages 3-5 CKD, dialysis, and transplant). Sensitivity, specificity, and positive and negative predictive values of FIT for ACN were 0.90, 0.83, 0.30, and 0.99, respectively. Of participants who underwent colonoscopy, five (1.5%) experienced major colonoscopy-related complications, including bowel perforation and major bleeding. CONCLUSIONS: FIT appears to be an accurate screening test for patients with CKD, such that a negative test may rule out the diagnosis of colorectal cancer within 2 years. However, the risk of major complications from work-up colonoscopy are at least ten-fold higher than in the general population.

Optimising the two-stage randomised trial design when some participants are indifferent in their treatment preferences.

Outcomes in a clinical trial can be affected by any underlying preferences that its participants have for the treatments under comparison and by whether they actually receive their preferred treatment. These effects cannot be evaluated in standard trial designs but are estimable in the alternative two-stage randomised trial design, in which some patients can choose their treatment, while the rest are randomly assigned. We have previously shown that, when all two-stage trial participants have a preferred treatment, the preference effects can be evaluated, in addition to the usual direct effect of treatment. We also determined criteria by which to optimise how many participants should be given a choice of treatment vs being randomised. More recently, we extended our methodology to allow for participants who are unable or unwilling to express a treatment preference if they are assigned to the choice group. In this paper, we show how to optimise the two-stage design when some participants are undecided about their treatment. We demonstrate that the undecided group should be regarded as distinct in the analysis, to obtain valid estimates of the preference effects. We derive the optimal proportion of participants who should be offered a choice of treatment, which in many cases will be close to 50%. More generally, the optima depend on the preference rates for treatments and the proportion of undecided participants, and the parameters of primary interest. We discuss some advantages and disadvantages of the two-stage trial design in this situation and describe a practical example.

Predicting pain recovery in patients with acute low back pain: Updating and validation of a clinical prediction model.

BACKGROUND: The prognosis of acute low back pain (LBP) is typically good; however, there is substantial variation in individual patient's outcomes. We recently developed a prediction model that was able to predict the likelihood of pain recovery in patients with acute LBP who continue to have pain approximately 1 week after initially seeking care. The aims of the current study were to (a) re-categorize the variables in the developmental dataset to be able to validate the model in the validation dataset; (b) refit the existing model in the developmental dataset; and (c) validate the model in the validation dataset. METHODS: The validation study sample comprised 737 patients with acute LBP, with a pain score of ≥2/10, 1 week after initially seeking care and with duration of current episode of ≤4 weeks. The primary outcome measure was days to pain recovery. Some of the variables from the development dataset were re-categorized prior to refitting the existing model in the developmental dataset using Cox regression. The performance (calibration and discrimination) of the prediction model was then tested in the validation dataset. RESULTS: Three variables of the development dataset were re-categorized. The performance of the prediction model with re-categorized variables in the development dataset was good (C-statistic = 0.76, 95% CI 0.70-0.82). The discrimination of the model using the validation dataset resulted in a C-statistic of 0.71 (95% CI 0.63-0.78). The calibration for the validation sample was acceptable at 1 month. However, at 1 week the predicted proportions within quintiles tended to overestimate the observed recovery proportions, and at 3 months, the predicted proportions tended to underestimate the observed recovery proportions. CONCLUSIONS: The updated prediction model demonstrated reasonably good external validity and may be useful in practice, but further validation and impact studies in relevant populations should be conducted. SIGNIFICANCE: A clinical prediction model based on five easily collected variables demonstrated reasonable external validity. The prediction model has the potential to inform patients and clinicians of the likely prognosis of individuals with acute LBP but requires impact studies to assess its clinical usefulness.

Agreement between digital breast tomosynthesis and pathologic tumour size for staging breast cancer, and comparison with standard mammography.

PURPOSE: Tomosynthesis is proposed to improve breast cancer assessment and staging. We compared tomosynthesis and mammography in estimating the size of newly-diagnosed breast cancers. METHODS: All pathologically-confirmed cancers detected in the STORM-2 trial (90 cancers, 85 women) were retrospectively measured on tomosynthesis by two independent readers. One reader also measured cancers on mammography. Relative mean differences (MDs) and 95% limits of agreement (LOA) with pathology were estimated for tomosynthesis and mammography within a single reader (Analysis 1) and between two readers (Analysis 2). RESULTS: Where cancers were detected and hence measured by both tests, tomosynthesis overestimated pathologic size relative to mammography (Analysis 1: MD 5% versus 1%, Analysis 2: 7% versus 3%; P = 0.10 both analyses). There was similar, large measurement variability for both tests (LOA range: -60% to +166%). Overestimation by tomosynthesis was attributable to the subgroup with dense breasts (MDs = 12-13% versus 4% for mammography). There was low average bias for both tests in the low-density subgroup (MDs = 0-4%). LOA were larger in dense breasts for both tomosynthesis and mammography (P ≤ 0.02 all comparisons). Cancers detected only by tomosynthesis were more frequently in dense breasts (60-68%): for those tumours size was estimated with increased measurement variability (LOA ranging from -75% to +293%). CONCLUSIONS: On average, tomosynthesis overestimates pathologic tumour size in women with dense breasts; that difference is more likely to impact management in women with larger tumours. The main advantage of tomosynthesis appears to be detecting mammographically-occult cancers; however tomosynthesis less accurately measured those cancers in dense breasts (large measurement variability).

The Validity of Drug Effects on Proteinuria, Albuminuria, Serum Creatinine, and Estimated GFR as Surrogate End Points for ESKD: A Systematic Review.

RATIONALE & OBJECTIVE: Proteinuria, albuminuria, and serum creatinine level are widely used as surrogate end point measures of end-stage kidney disease (ESKD). We evaluated the correlation between antihypertensive drug effects on surrogate renal end points and ESKD. STUDY DESIGN: Systematic review. SETTING & PARTICIPANTS: Randomized controlled trials of blood pressure-lowering therapy. SELECTION CRITERIA FOR STUDIES: Trials of pharmacological blood pressure-lowering strategies reporting drug effects on albuminuria, proteinuria, or serum creatinine level and ESKD through March 26, 2018. ANALYTICAL APPROACH: Bayesian bivariate meta-analysis to calculate correlations between drug effects on surrogate end points and drug effects on ESKD. Risks of bias were adjudicated using the Cochrane tool. RESULTS: 22 randomized controlled trials involving 69,642 participants were eligible. Risks of bias in the included trials were frequently unclear due to incomplete reporting. Relative risk for ESKD was statistically significant in 1 of 29 (3.4%) treatment comparisons. There appeared to be little or no correlation between antihypertensive drug effects on serum creatinine level, albuminuria, proteinuria, and the corresponding effects on ESKD. All correlations had wide 95% credible intervals that included the null effect. LIMITATIONS: Low power due to infrequent outcomes of ESKD and incomplete data reporting in primary trials. CONCLUSIONS: The association between antihypertensive drug effects on doubling of serum creatinine level and albuminuria or proteinuria with ESKD in treatment trials is not sufficiently certain to enable the confident use of these markers to guide clinical or regulatory decision making.

Effect of integrating digital breast tomosynthesis (3D-mammography) with acquired or synthetic 2D-mammography on radiologists' true-positive and false-positive detection in a population screening trial: A descriptive study.

BACKGROUND: We previously reported the Screening with tomosynthesis or standard mammography-2 (STORM-2) trial, showing that tomosynthesis (3D-mammography) screening detected more cancers than 2D-mammography in double-reading practice. In this study, we report reader-specific detection measures for radiologists who performed the screen-reading in this trial. METHODS: This is a sub-study of the STORM-2 trial which prospectively integrated 3D-mammography with acquired or synthetized 2D-mammograms in parallel double-reading arms. Asymptomatic women ≥49 years who attended population-based screening (Trento, 2013-2015) were recruited. Screening participants were recalled at any positive sequential screen-read in either reading arm of the trial. Radiologist-specific detection measures were calculated for each of seven radiologists who performed screen-reads: number of detected cancers, proportion of true-positive (TP) detection, and number and rate of false-positive (FP) recalls (FPR). We estimated incremental cancer detection rate (CDR) from integrating 3D-mammography in screen-reading. RESULTS: Across all radiologists, TP detection (relative sensitivity) ranged between: 46% and 100% (median 59.5%) for 2D-mammography; 75% and 100% (median 76%) for integrated 2D/3D-mammography screening; 56% and 76% (median 64%) for 2Dsynthetic; 67% and 88% (median 78%) for 2Dsynthetic/3D-mammography. Integrating 3D-mammography led to incremental CDRs between 0/1000 and 3.5/1000 screens. FPR ranged between: 1.2% and 2.7% (median 2.25%) for 2D-mammography; 1.5% and 3.4% (median 2.75%) for 2D/3D-mammography; 1.6% and 4.6% (median 2.4%) for 2Dsynthetic; and 1.8% and 6.7% (median 3.0%) for 2Dsynthetic/3D-mammography. CONCLUSIONS: There was variability in the magnitude of effect from integrating 3D-mammography (relative to screen-reading with acquired or synthesised 2D-mammography alone) on individual radiologist's TP and FP detection, although there was an overall pattern of increasing cancer detection and also increasing FP recall for most readers.

Breast Cancer Screening Using Tomosynthesis or Mammography: A Meta-analysis of Cancer Detection and Recall.

Background: Tomosynthesis approximates a 3D mammogram of the breast, reducing parenchymal overlap that masks cancers or creates false "lesions" on 2D mammography, and potentially enabling more accurate detection of breast cancer. We compared breast cancer screening detection and recall in asymptomatic women for tomosynthesis vs 2D mammography. Methods: A systematic review and random effects meta-analysis were undertaken. Electronic databases (2009-July 2017) were searched for studies comparing tomosynthesis and 2D mammography in asymptomatic women who attended population breast cancer screening and reporting cancer detection rate (CDR) and recall rate. All statistical tests were two-sided. Results: Seventeen studies (1 009 790 participants) were included from 413 citations. The pooled incremental CDR for tomosynthesis was 1.6 cancers per 1000 screens (95% confidence interval [CI] = 1.1 to 2.0, P < .001, I2 = 36.9%). Incremental CDR was statistically significantly higher for European/Scandinavian studies, all using a "paired" design where women had both tests (2.4 per 1000 screens, 95% CI = 1.9 to 2.9, P < .001, I2 = 0.0%) compared with US ("unpaired") studies (1.1 per 1000 screens, 95% CI = 0.8 to 1.5, P < .001, I2 = 0.0%; P < .001 between strata). The recall rate for tomosynthesis was statistically significantly lower than for 2D mammography (pooled absolute reduction = -2.2%, 95% CI = -3.0 to -1.4, P < .001, I2 = 98.2%). Stratified analyses showed a decrease in US studies (pooled difference in recall rate = -2.9%, 95% CI = -3.5 to -2.4, P < .001, I2 = 92.9%) but not European/Scandinavian studies (0.5% increase in recall, 95% CI = -0.1 to 1.2, P = .12, I2 = 93.5%; P < .001 between strata). Results were similar in sensitivity analyses excluding studies with overlapping cohorts. Conclusions: Tomosynthesis improves CDR and reduces recall; however, effects are dependent on screening setting, with greater improvement in CDR in European/Scandinavian studies (biennial screening) and reduction in recall in US studies with high baseline recall.