Publications

BACKGROUND: Bisphosphonates are specific inhibitors of osteoclastic activity and are used in the treatment of patients with multiple myeloma (MM). While bisphosphonates are shown to be effective in reducing vertebral fractures and pain, their role in improving overall survival (OS) remains unclear. This is an update of a Cochrane review first published in 2002 and previously updated in 2010 and 2012.

OBJECTIVES: To assess the evidence related to benefits and harms associated with use of various types of bisphosphonates (aminobisphosphonates versus non-aminobisphosphonates) in the management of patients with MM. Our primary objective was to determine whether adding bisphosphonates to standard therapy in MM improves OS and progression-free survival (PFS), and decreases skeletal-related morbidity. Our secondary objectives were to determine the effects of bisphosphonates on pain, quality of life, incidence of hypercalcemia, incidence of bisphosphonate-related gastrointestinal toxicities, osteonecrosis of jaw (ONJ) and hypocalcemia.

SEARCH METHODS: We searched MEDLINE, Embase (September 2011 to July 2017) and the CENTRAL (2017, Issue 7) to identify all randomized controlled trial (RCT) in MM up to July 2017 using a combination of text and MeSH terms.

SELECTION CRITERIA: Any randomized controlled trial (RCT) comparing bisphosphonates versus placebo/no treatment/bisphosphonates and observational studies or case reports examining bisphosphonate-related ONJ in patients with MM were eligible for inclusion.

DATA COLLECTION AND ANALYSIS: Two review authors extracted the data. Data were pooled and reported as hazard ratio (HR) or risk ratio (RR) using a random-effects model. We used meta-regression to explore statistical heterogeneity. Network meta-analysis using Bayesian approach was conducted.

MAIN RESULTS: In this update, we included four new studies (601 participants), resulting in a total of 24 included studies.Twenty RCTs compared bisphosphonates with either placebo or no treatment and four RCTs involved another bisphosphonate as a comparator. The 24 included RCTs enrolled 7293 participants. Pooled results showed that there was moderate-quality evidence of a reduction in mortality with on OS from 41% to 31%, but the confidence interval is consistent with a larger reduction and small increase in mortality compared with placebo or no treatment (HR 0.90, 95% CI 0.76 to 1.07; 14 studies; 2706 participants). There was substantial heterogeneity among the included RCTs (I2 = 65%) for OS. To explain this heterogeneity we performed a meta-regression assessing the relationship between bisphosphonate potency and improvement in OS, which found an OS benefit with zoledronate but limited evidence of an effect on PFS. This provided a further rationale for performing a network meta-analyses of the various types of bisphosphonates that were not compared head-to-head in RCTs. Results from network meta-analyses showed evidence of a benefit for OS with zoledronate compared with etidronate (HR 0.56, 95% CI 0.29 to 0.87) and placebo (HR 0.67, 95% CI 0.46 to 0.91). However, there was no evidence for a difference between zoledronate and other bisphosphonates.The effect of bisphosphonates on disease progression (PFS) is uncertain. Based on the HR of 0.75 (95% CI 0.57 to 1.00; seven studies; 908 participants), 47% participants would experience disease progression without treatment compared with between 30% and 47% with bisphosphonates (low-quality evidence). There is probably a similar risk of non-vertebral fractures between treatment groups (RR 1.03, 95% CI 0.68 to 1.56; six studies; 1389 participants; moderate-quality evidence). Pooled analysis demonstrated evidence for a difference favoring bisphosphonates compared with placebo or no treatment on prevention of pathological vertebral fractures (RR 0.74, 95% CI 0.62 to 0.89; seven studies; 1116 participants; moderate-quality evidence) and skeletal-related events (SREs) (RR 0.74, 95% CI 0.63 to 0.88; 10 studies; 2141 participants; moderate-quality evidence). The evidence for less pain with bisphosphonates was of very low quality (RR 0.75, 95% CI 0.60 to 0.95; eight studies; 1281 participants).Bisphosphonates may increase ONJ compared with placebo but the confidence interval is very wide (RR 4.61, 95% CI 0.99 to 21.35; P = 0.05; six studies; 1284 participants; low-quality evidence). The results from the network meta-analysis did not show any evidence for a difference in the incidence of ONJ (eight RCTs, 3746 participants) between bisphosphonates. Data from nine observational studies (1400 participants) reported an incidence of 5% to 51% with combination of pamidronate and zoledronate, 3% to 11% with zoledronate alone, and 0% to 18% with pamidronate alone.The pooled results showed no evidence for a difference in increase in frequency of gastrointestinal symptoms with the use of bisphosphonates compared with placebo or no treatment (RR 1.23, 95% CI 0.95 to 1.59; seven studies; 1829 participants; low-quality evidence).The pooled results showed no evidence for a difference in increase in frequency of hypocalcemia with the use of bisphosphonates compared with placebo or no treatment (RR 2.19, 95% CI 0.49 to 9.74; three studies; 1090 participants; low-quality evidence). The results from network meta-analysis did not show any evidence for differences in the incidence of hypocalcemia, renal dysfunction and gastrointestinal toxicity between the bisphosphonates used.

AUTHORS' CONCLUSIONS: Use of bisphosphonates in participants with MM reduces pathological vertebral fractures, SREs and pain. Bisphosphonates were associated with an increased risk of developing ONJ. For every 1000 participants treated with bisphosphonates, about one patient will suffer from the ONJ. We found no evidence of superiority of any specific aminobisphosphonate (zoledronate, pamidronate or ibandronate) or non-aminobisphosphonate (etidronate or clodronate) for any outcome. However, zoledronate was found to be better than placebo and first-generation bisposphonate (etidronate) in pooled direct and indirect analyses for improving OS and other outcomes such as vertebral fractures. Direct head-to-head trials of the second-generation bisphosphonates are needed to settle the issue if zoledronate is truly the most efficacious bisphosphonate currently used in practice.

OBJECTIVE: To assess the efficacy of active treatment targeted at underlying disease (TTD)/potentially curative treatments versus palliative care (PC) in improving overall survival (OS) in terminally ill patients.

DESIGN: We performed a systematic review and meta-analysis of randomised controlled trials (RCT). Methodological quality of included RCTs was assessed using the Cochrane risk of bias tool.

DATA SOURCES: Medline and Cochrane databases were searched, with no language restriction, from inception to 19 October 2016.

ELIGIBILITY CRITERIA FOR SELECTING STUDIES: Any RCT assessing the efficacy of any active TTD versus PC in adult patients with terminal illness with a prognosis of <6-month survival were eligible for inclusion.

RESULTS: Initial search identified 8252 citations of which 10 RCTs (15 comparisons, 1549 patients) met inclusion criteria. All RCTs included patients with cancer. OS was reported in 7 RCTs (8 comparisons, 1158 patients). The pooled results showed no statistically significant difference in OS between TTD and PC (HR (95% CI) 0.85 (0.71 to 1.02)). The heterogeneity between pooled studies was high (I2=62.1%). Overall rates of adverse events were higher in the TTD arm.

CONCLUSIONS: Our systematic review of available RCTs in patients with terminal illness due to cancer shows that TTD compared with PC did not demonstrably impact OS and is associated with increased toxicity. The results provide assurance to physicians, patients and family that the patients' survival will not be compromised by referral to hospice with focus on PC.

OBJECTIVE: To assess the efficacy of active treatment targeted at underlying disease (TTD)/potentially curative treatments versus palliative care (PC) in improving overall survival (OS) in terminally ill patients.

DESIGN: We performed a systematic review and meta-analysis of randomised controlled trials (RCT). Methodological quality of included RCTs was assessed using the Cochrane risk of bias tool.

DATA SOURCES: Medline and Cochrane databases were searched, with no language restriction, from inception to 19 October 2016.

ELIGIBILITY CRITERIA FOR SELECTING STUDIES: Any RCT assessing the efficacy of any active TTD versus PC in adult patients with terminal illness with a prognosis of <6-month survival were eligible for inclusion.

RESULTS: Initial search identified 8252 citations of which 10 RCTs (15 comparisons, 1549 patients) met inclusion criteria. All RCTs included patients with cancer. OS was reported in 7 RCTs (8 comparisons, 1158 patients). The pooled results showed no statistically significant difference in OS between TTD and PC (HR (95% CI) 0.85 (0.71 to 1.02)). The heterogeneity between pooled studies was high (I2=62.1%). Overall rates of adverse events were higher in the TTD arm.

CONCLUSIONS: Our systematic review of available RCTs in patients with terminal illness due to cancer shows that TTD compared with PC did not demonstrably impact OS and is associated with increased toxicity. The results provide assurance to physicians, patients and family that the patients' survival will not be compromised by referral to hospice with focus on PC.

BACKGROUND: Next-generation sequencing has identified somatic mutations that are prognostic of cancer.

PATIENTS AND METHODS: We evaluated the incidence and prognostic significance of somatic mutations in 89 myelodysplastic syndrome (MDS) patients who received an allogeneic hematopoietic cell transplantation. Next-generation sequencing was performed on paraffin embedded bone marrow, which was obtained at a median of 31 days before initiating the preparative regimen.

RESULTS: The 3 most common subtypes of MDS were refractory anemia with excess blasts (RAEB)-1 (35%), RAEB-2 (29%), and refractory cytopenia with multilineage dysplasia (18%). Most patients (91%) received a myeloablative regimen of fludarabine with intravenous busulfan. Somatic mutations (> 0) were identified in 39 (44%) of analyzed samples. The 6 most commonly identified gene mutations were ASXL1 (8%), DNMT3A (8%), RUNX1 (7%), KRAS (6%), IDH2 (4%), and TP53 (4%). The low incidence of mutations in our study sample might be explained by tissue source and stringent variant-calling methodology. Moreover, we speculate that the low incidence of mutations might, perhaps, also be explained by previous azacitidine treatment in 82% of cases. Multivariate analysis identified TP53 (hazard ratio [HR], 3.82; 95% confidence interval [CI], 1.12-13.09; P = .03) and IDH2 mutations (HR, 4.74; 95% CI, 1.33-16.91; P = .02) as predictors of inferior 3-year overall survival.

CONCLUSION: This study furthers implementation of clinical genomics in MDS and identifies TP53 and IDH2 as targets for pre- or post-transplant therapy.

RATIONALE, AIMS AND OBJECTIVES: Theories of decision making are divided between those aiming to help decision makers in the real, 'large' world and those who study decisions in idealized 'small' world settings. For the most part, these large- and small-world decision theories remain disconnected.

METHODS: We linked the small-world decision theoretic concepts of signal detection theory (SDT) and evidence accumulation theory (EAT) to the threshold model and the large world of heuristic decision making that rely on fast-and-frugal decision trees (FFT).

RESULTS: We connected these large- and small-world theories by demonstrating that seemingly different decision-making concepts are actually equivalent. In doing so, we were able (1) to link the threshold model to EAT and FFT, thereby creating decision criteria that take into account both the classification accuracy of FFT and the consequences built in the threshold model; (2) to demonstrate how threshold criteria can be used as a strategy for optimal selection of cues when constructing FFT; and (3) to show that the compensatory strategy expressed in the threshold model can be linked to a non-compensatory FFT approach to decision making. We also showed how construction and performance of FFT depend on having reliable information - the results were highly sensitive to the estimates of benefits and harms of health interventions. We illustrate the practical usefulness of our analysis by describing an FFT we developed for prescribing statins for primary prevention of cardiovascular disease.

CONCLUSIONS: By linking SDT and EAT to the compensatory threshold model and to non-compensatory heuristic decision making (FFT), we showed how these two decision strategies are ultimately linked within a broader theoretical framework and thereby respond to calls for integrating decision theory paradigms.

RATIONALE, AIMS AND OBJECTIVES: The American College of Cardiology and American Heart Association (ACC/AHA) statin guidelines recommend that people with risk of cardio-vascular disease (CVD) ≥7.5% over 10 years should be treated with statins. This recommendation ignores individual patient CVD risks and preferences. We compared the ACC/AHA guidelines to the following management strategies a) individualized statins treatment based on Framingham Risk Score (FRS), b) treat none, c) treat all.

METHODS: We employed regret-based decision curve analysis to evaluate the optimal treatment strategy. We used data on 5013 participants from the second generation of the Framingham Heart Study. We assessed regret of each treatment strategy [treat according to FRS vs. treat none vs. treat all] as a function of emotionally felt loss of treatment benefits and incurred treatment harms. We calculated the difference between regret associated with one strategy compared with the other and expressed it as Net Expected Regret Difference (NERD). Two strategies are identical if NERD = 0.

RESULTS: Treatment according to ACC/AHA guidelines represents the optimal strategy only if the patient values avoiding heart disease 12 times more than harms related to statins. For values of benefit/harms (B/H) <12, treatment according to FRS represents the optimal strategy. For B/H <3, 'treat none' represents equally acceptable strategy. Adopting a threshold of 10% recommended by other professional organizations would decrease over-treatment by more than 60% without significantly affecting under-treatment.

CONCLUSION: Under most realistic scenarios, individualizing statins treatment, or not recommending statins at all, represents the optimal strategy for primary prevention of heart disease.

BACKGROUND: The acceptable regret model postulates that under some circumstances decision-makers may tolerate wrong decisions. The purpose of this work is to empirically evaluate the acceptable regret model of decision-making in the end-of-life care setting, where terminally ill patients consider seeking curative treatment versus accepting hospice/palliative care.

METHODS: We conducted interviews with 48 terminally ill patients to assess their preferences about end-of-life treatment choices. We first elicited the patient's regret of potentially wrong choices with regards to the recommended management and provided information on life expectancy estimated by two prognostication models. We then elicited the patients' level of acceptable regret by assessing their tolerance for potentially wrongly accepting hospice care versus continuing unnecessary treatment. Using the levels of acceptable regret, we computed: (1) the probability of death above which a patient would tolerate wrongly accepting hospice care and (2) the probability of death below which the patient would tolerate unnecessary treatment. We also assessed patients' understanding of the interview questions using a 7-point Likert scale.

RESULTS: We found that the median probability of death above which a patient would tolerate wrongly accepting hospice care was 96% (95% CI 94-98%), whereas the median probability of death below which a patient would tolerate unnecessary treatment was 2.5% (95% CI 0.3-5%). We also found that the levels of acceptable regret measured for wrong hospice referral (mean = 1.52; SD = 2.26; min = 0; max = 7.72) were similar to the levels of acceptable regret measured for unnecessary treatment (mean = 2.10; SD = 4.33; min = 0; max = 23) (KW test; p = 0.68) indicating that acceptable regret levels for either of the wrong decisions is felt similarly. The results were independent of the estimated probability of death communicated to patients before the acceptable regret interview.

CONCLUSIONS: We have elicited empirical data that corroborated the acceptable regret hypothesis. The requirement for high level of certainty before accepting recommended management may explain the difficulties related to decision-making in the end-of-life setting.

BACKGROUND: Decision curve analysis (DCA) is an increasingly used method for evaluating diagnostic tests and predictive models, but its application requires individual patient data. The Monte Carlo (MC) method can be used to simulate probabilities and outcomes of individual patients and offers an attractive option for application of DCA.

MATERIALS AND METHODS: We constructed a MC decision model to simulate individual probabilities of outcomes of interest. These probabilities were contrasted against the threshold probability at which a decision-maker is indifferent between key management strategies: treat all, treat none or use predictive model to guide treatment. We compared the results of DCA with MC simulated data against the results of DCA based on actual individual patient data for three decision models published in the literature: (i) statins for primary prevention of cardiovascular disease, (ii) hospice referral for terminally ill patients and (iii) prostate cancer surgery.

RESULTS: The results of MC DCA and patient data DCA were identical. To the extent that patient data DCA were used to inform decisions about statin use, referral to hospice or prostate surgery, the results indicate that MC DCA could have also been used. As long as the aggregate parameters on distribution of the probability of outcomes and treatment effects are accurately described in the published reports, the MC DCA will generate indistinguishable results from individual patient data DCA.

CONCLUSIONS: We provide a simple, easy-to-use model, which can facilitate wider use of DCA and better evaluation of diagnostic tests and predictive models that rely only on aggregate data reported in the literature.

INTRODUCTION: In patients with obscure gastrointestinal bleeding (OBGIB) capsule endoscopy (CE) is the initial diagnostic procedure of choice. Often patients undergo single balloon enteroscopy (SBE) with both diagnostic and therapeutic intention after CE. Although SBE offers a therapeutic benefit, long procedure times, complexity, and invasiveness are drawbacks. We aimed to evaluate the diagnostic correlation between these two modalities after an initial positive CE finding.

METHODS: We performed a retrospective review of 418 patients who underwent CE at our institution from January 2010 to May 2014. A total of 95 patients were analyzed after selecting patients that underwent SBE originally after a positive CE result for the evaluation for OGIB. Agreement beyond chance was evaluated using the κ coefficient. A p value less than 5% was considered statistically significant.

RESULTS: The mean age of our population was 65.8 ± 12.2 and it was female predominant: 57/95 (60%). The most frequent positive findings were vascular lesions found on SBE in 31.6% and on CE in 41.1%. There was a strong agreement when identifying active bleeding and clots [κ=0.97; 95% confidence interval (CI) 0.92-1.03; p ⩽ 0.0001], and a moderate agreement when diagnosing vascular lesions (0.41; 95% CI 0.21-0.61; p ⩽ 0.0001). There was fair agreement for ulcers (0.26; 95% CI 0.07-0.59; p = 0.005). There was a low correlation between masses, polyps, and others.

CONCLUSION: CE still remains the initial test of choice in evaluating stable patients with OBGIB since it has strong-to-fair concordance for the major small bowel findings. However, in cases of severe overt small bowel bleeding, balloon-assisted enteroscopy can be considered the initial procedure of choice since it is therapeutic as well as diagnostic and this approach avoids delays in treatment. Further research should focus on methods to improve interpretation of CE and enhance the ability to evaluate the entire small bowel with SBE.

OBJECTIVES: The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) is a widely used methodology for the development of clinical practice guideline. Although its reproducibility is good for evaluating the quality of evidence, it has not been tested in context of developing recommendations. The objective of this study was to assess the reproducibility of all GRADE factors that determine the direction and strength of a recommendation among the guideline panel members with limited exposure to GRADE methodology.

STUDY DESIGN AND SETTING: The study was conducted as part of the clinical practice guideline development process of American Association of Blood Banking for the use of prophylactic vs. therapeutic platelet transfusion in patients with thrombocytopenia. The results from the systematic review and meta-analysis for each question were summarized as a GRADE evidence profile. Interrater agreement for all GRADE factors and strength of recommendations was summarized using a weighted kappa statistic with 95% confidence intervals (CI).

RESULTS: Eighteen members of the panel participated in the deliberation of making recommendations and completed the online questionnaire. They were given two 1-hour lectures about GRADE. The agreement for all domains was better than chance. The interrater agreement for the domain of quality of evidence was good (kappa value: 0.68; 95% CIs: 0.54, 0.84), and fair for balance of benefit and harms (kappa value: 0.4; 95% CIs: 0.25, 0.57) and use of resources (kappa value: 0.28: 95% CIs: 0.12, 0.42). The interrater agreement was moderate for the GRADE domain of patients' values and preferences (kappa value: 0.44; 95% CI: 0.31, 0.56). The interrater agreement for making a for/against recommendation was good (kappa value: 0.74; 95% CIs: 0.33, 0.91) and fair for strong/weak recommendation (kappa value: 0.39; 95% CIs: 0.18, 0.68).

CONCLUSIONS: Although not all elements of GRADE system had good agreement, the interrater agreement for assessing the quality of evidence and issuing a recommendation of for vs. against among panel members who had limited exposure to GRADE methodology was good. This is probably because GRADE has operationalized these two areas in more detail than other domains. Further operationalization of all GRADE domains such as with the GRADE evidence to decision frameworks would likely improve its reproducibility.