One-stage dose–response meta-analysis for aggregated data

2018 ◽  
Vol 28 (5) ◽  
pp. 1579-1596 ◽  
Author(s):  
Alessio Crippa ◽  
Andrea Discacciati ◽  
Matteo Bottai ◽  
Donna Spiegelman ◽  
Nicola Orsini
Keyword(s):  
Dose Response ◽  
Goodness Of Fit ◽  
Linear Mixed Model ◽  
Meta Analysis ◽  
Model Specification ◽  
Two Stage ◽  
Aggregated Data ◽  
One Stage ◽  
Non Linear ◽  
The One

The standard two-stage approach for estimating non-linear dose–response curves based on aggregated data typically excludes those studies with less than three exposure groups. We develop the one-stage method as a linear mixed model and present the main aspects of the methodology, including model specification, estimation, testing, prediction, goodness-of-fit, model comparison, and quantification of between-studies heterogeneity. Using both fictitious and real data from a published meta-analysis, we illustrated the main features of the proposed methodology and compared it to a traditional two-stage analysis. In a one-stage approach, the pooled curve and estimates of the between-studies heterogeneity are based on the whole set of studies without any exclusion. Thus, even complex curves (splines, spike at zero exposure) defined by several parameters can be estimated. We showed how the one-stage method may facilitate several applications, in particular quantification of heterogeneity over the exposure range, prediction of marginal and conditional curves, and comparison of alternative models. The one-stage method for meta-analysis of non-linear curves is implemented in the dosresmeta R package. It is particularly suited for dose–response meta-analyses of aggregated where the complexity of the research question is better addressed by including all the studies.

scholarly journals Flexible piecewise linear model for investigating dose-response relationship in meta-analysis: methodology, examples, and comparison

2018 ◽  
Author(s):  
Chang Xu ◽  
Lehana Thabane ◽  
Tong-Zu Liu ◽  
Ling Li ◽  
Sayem Borhan ◽  
...  
Keyword(s):  
Linear Model ◽  
Dose Response ◽  
Linear Trend ◽  
Piecewise Linear ◽  
Meta Analysis ◽  
Continuous Exposure ◽  
Response Relationship ◽  
Dose Response Relationship ◽  
Two Stage ◽  
One Stage

Objectives: Dose-response meta-analysis (DRMA) is widely employed to establishing the potential dose-response relationship between continuous exposures and disease outcomes. However, no method is readily available for exploring the relation between a discrete exposure and a binary or continuous outcome. We proposed a piecewise linear (PL) DRMA model as a solution to this issue. Methods: We illustrated the methodology of PL model in both one-stage DRMA approach and two-stage DRMA approach. The method by testing the equality of slopes of each piecewise was employed to judge if there is “piecewise effect” against simple linear trend. We then used sleep (continuous exposure) and parity (discrete exposure) data as examples to illustrate how to apply PL model in DRMA using the Stata code attached. We also empirically compared the slopes of PL model with simple linear as well as restricted cubic spline (RCS) model. Results: Both one-stage and two-stage PL DRMA model fitted well in our examples, and the results were similar. Obvious “piecewise effects” were detected in both the two examples by the method we used. In our example, the PL model showed better fitting effect and practical reliable results compared to simple linear model, while similar results for to RCS model. Conclusion: Piecewise linear function is a simple and valid method for DRMA and can be used for discrete exposures. It also represents a superior model to linear model in DRMA and may be an alternative model to non-linear model.

scholarly journals Pattern-mixture model in network meta-analysis of binary missing outcome data: one-stage or two-stage approach?

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Loukia M. Spineli ◽  
Katerina Papadimitropoulou ◽  
Chrysostomos Kalyvas
Keyword(s):  
Mixture Model ◽  
Simulation Study ◽  
Meta Analysis ◽  
Credible Interval ◽  
Outcome Data ◽  
Systematic Bias ◽  
Two Stage ◽  
Pattern Mixture Model ◽  
One Stage ◽  
The One

Abstract Background Trials with binary outcomes can be synthesised using within-trial exact likelihood or approximate normal likelihood in one-stage or two-stage approaches, respectively. The performance of the one-stage and the two-stage approaches has been documented extensively in the literature. However, little is known about how these approaches behave in the presence of missing outcome data (MOD), which are ubiquitous in clinical trials. In this work, we compare the one-stage versus two-stage approach via a pattern-mixture model in the network meta-analysis using Bayesian methods to handle MOD appropriately. Methods We used 29 published networks to empirically compare the two approaches concerning the relative treatment effects of several competing interventions and the between-trial variance (τ2), while considering the extent and level of balance of MOD in the included trials. We additionally conducted a simulation study to compare the competing approaches regarding the bias and width of the 95% credible interval of the (summary) log odds ratios (OR) and τ2 in the presence of moderate and large MOD. Results The empirical study did not reveal any systematic bias between the compared approaches regarding the log OR, but showed systematically larger uncertainty around the log OR under the one-stage approach for networks with at least one small trial or low event risk and moderate MOD. For these networks, the simulation study revealed that the bias in log OR for comparisons with the reference intervention in the network was relatively higher in the two-stage approach. Contrariwise, the bias in log OR for the remaining comparisons was relatively higher in the one-stage approach. Overall, bias increased for large MOD. For these networks, the empirical results revealed slightly higher τ2 estimates under the one-stage approach irrespective of the extent of MOD. The one-stage approach also led to less precise log OR and τ2 when compared with the two-stage approach for large MOD. Conclusions Due to considerable bias in the log ORs overall, especially for large MOD, none of the competing approaches was superior. Until a more competent model is developed, the researchers may prefer the one-stage approach to handle MOD, while acknowledging its limitations.

scholarly journals Pattern-mixture model in network meta-analysis of binary missing outcome data: one-stage or two-stage approach?

2020 ◽  
Author(s):  
Loukia Maria Spineli ◽  
Katerina Papadimitropoulou ◽  
Chrysostomos Kalyvas
Keyword(s):  
Empirical Study ◽  
Mixture Model ◽  
Simulation Study ◽  
Meta Analysis ◽  
Outcome Data ◽  
Systematic Bias ◽  
Two Stage ◽  
Pattern Mixture Model ◽  
One Stage ◽  
The One

Abstract Background Trials with binary outcomes can be synthesised using within-trial exact likelihood or approximate normal likelihood in one-stage or two-stage approaches, respectively. The advantages of the one-stage over the two-stage approach have been documented extensively in the literature. Little is known how these approaches behave in the presence of missing outcome data (MOD) which are ubiquitous in trials. In this work, we compare the one-stage versus two-stage approach via a pattern-mixture model in the network meta-analysis Bayesian framework to handle MOD appropriately. Methods We used 29 published networks to empirically compare the two approaches with respect to the relative treatment effects of several competing interventions and the between-trial variance ( {\tau }^{2} ). We categorised the networks according to the extent and balance of MOD in the included trials. To complement the empirical study, we conducted a simulation study to compare the competing approaches regarding bias and width of the 95% credible interval of the (summary) log odds ratios (OR) and {\tau }^{2} in the presence of moderate and large MOD. Results The empirical study did not reveal any systematic bias between the compared approaches regarding the log OR, but showed systematically larger uncertainty around the log OR under the one-stage approach for networks with at least one small trial or low event risk and moderate MOD. For these networks, the simulation study revealed that the bias in log OR for comparisons with the reference intervention in the network was relatively higher in the two-stage approach. Contrariwise, the bias in log OR for the remaining comparisons was relatively higher in the one-stage approach. Overall, bias increased for large MOD. Furthermore, in these networks, the empirical results revealed slightly higher {\tau }^{2} estimates under the one-stage approach irrespective of the extent of MOD. The one-stage approach also led to less precise log OR and {\tau }^{2} when compared with the two-stage approach for large MOD. Conclusions Due to considerable bias in the log ORs overall, especially for large MOD, none of the competing approaches was superior. Until a more competent model is developed, the researchers may prefer the one-stage approach to handle MOD, while acknowledging its limitations.

Non-linear Relationship between Tacrolimus Blood Concentration and Acute Rejection After Kidney Transplantation: A Systematic Review and Dose-Response Meta-Analysis of Cohort Studies

2019 ◽  
Vol 25 (21) ◽  
pp. 2394-2403
Author(s):  
Saifu Yin ◽  
Turun Song ◽  
Xingxing Li ◽  
Hanyue Xu ◽  
Xueling Zhang ◽  
...  
Keyword(s):  
Kidney Transplantation ◽  
Acute Rejection ◽  
Linear Relationship ◽  
Dose Response ◽  
Blood Concentration ◽  
Meta Analysis ◽  
West China Hospital ◽  
West China ◽  
One Stage ◽  
Non Linear

Background: Maintaining the exposure of tacrolimus (Tac) after kidney transplantation (KT) must be necessary to prevent acute rejection (AR) and improve graft survival，but there is still no clear consensus on the optimal Tac target blood concentration and concentration-effect relationship is poorly defined. Methods: We conducted a dose-response meta-analysis to quantitatively assess the association between Tac blood concentration and (AR) or adverse effects after KT. A comprehensive search of PubMed, Embase and Cochrane library databases was conducted to find eligible studies up to 10th September 2018. Unpublished data from patients receiving KT in West China Hospital (Sichuan University, China) were also collected. Both twostage dose-response and one-stage dose-response meta-analysis models were used to improve the statistical power. Results: A total of 4967 individuals from 10 original studies and 1453 individuals from West China Hospital were eligible for the ultimate analysis. In the two-stage dose-response meta-analysis model, we observed a significant non-linear relationship between Tac blood concentration and AR (P < 0.001) with moderate heterogeneity (I2 = 46.0%, P = 0.08). Tac blood concentration at 8ng/ml was associated with the lowest risk of AR (RR: 0.26, 95%CI: 0.13 - 0.54) by reference to 2ng/ml. Tac concentration at 7.0 - 11.0 ng/ml reduced the risk of AR by at least 70%, 5-14 ng/ml by at least 60%, and 4.5 – 14 ng/ml at least 50%. In the one-stage dose-response model, we also found a strong non-linear relationship between Tac and AR (P < 0.001) with moderate heterogeneity (I2 = 41.2%, P = 0.10). Tac concentration of 7.5 ng/ml was associated with the lowest risk of AR (RR: 0.35, 95%CI: 0.16 - 0.77). The blood concentration at 5.5 - 9.5 ng/ml was associated with the reduced AR by at least 60% and 4.5 - 10.5 ng/ml by at least 50% by reference to 2 ng/ml. Conclusion: Maintaining Tac blood concentration at 5 - 9.5 ng/ml within the first year may prevent AR most effectively.

scholarly journals An Individual Patient Data Meta-Analysis with Colombian Studies on the Effect of Dark Chocolate Consumption on Cardiovascular Risk Parameters

2020 ◽  
Vol 2020 ◽  
pp. 1-8
Author(s):  
Leidy Alvarez ◽  
Javier Contreras ◽  
Mónica Giraldo
Keyword(s):  
Body Mass Index ◽  
Clinical Trials ◽  
Lipid Profile ◽  
Body Mass ◽  
Meta Analysis ◽  
Abdominal Circumference ◽  
Two Stage ◽  
One Stage ◽  
Study Results ◽  
The One

Background. It is postulated that cocoa solids possess cardioprotective capacity by various mechanisms. In the different cocoa studies evaluating cardiovascular disease, there are no conclusive data on the role it plays in controlling the lipid profile and anthropometric variables, perhaps because the concentration of cocoa, the geographical origin of the population, and the different concentrations supplied lead to a high heterogeneity of results. This study aims to estimate the effect of consuming cocoa-rich chocolate compared to placebo on the lipid profile and anthropometric variables based on data from three clinical trials conducted in Colombia. Methods. Meta-analysis of individual data from three randomized clinical trials conducted in Colombia. The entire population of the primary studies was included, which was reassigned into intervention groups if they consumed 50 grams of 70% concentrated cocoa or placebo, which was considered to be cocoa-free or with a concentration less than 50 grams. The variables at the beginning of the study were analyzed with medians, interquartile ranges, means, and deviations according to whether they met the normality assumption. Multiple imputations were used to manage missing data and were analyzed using the two approaches proposed for this type of study, that of one and two stages. In the two-stage approach, the data were weighted on a conventional Forrest plot, while in the one-stage approach, linear regressions with mixed models were applied. This study is governed by the regulations described in the 2013 Declaration of Helsinki and by article 11 of Resolution 8430 of 1993, which classifies it as a risk-free study. Results. A total of 275 participants were included, who consumed cocoa or placebo for 81 days on average; 52.7% were female and few smoked at the time of the intervention (31/275). Physical activity performed in number of hours per week was comparable between the intervention groups. When evaluating total cholesterol, low-density cholesterol (LDL), high-density cholesterol (HDL), triglycerides, abdominal circumference, and final body mass index with both the one-stage and two-stage approaches, there were no significant differences between the two groups. Conclusions. According to the results obtained in the meta-analysis, the consumption of cocoa in the Colombian population does not seem to significantly modify variables such as lipid profile, abdominal circumference, and body mass index. This conclusion according to the quality of the evidence has a weak recommendation and a low-to-moderate certainty. However, the analysis through the two proposed approaches yielded similar results.

scholarly journals Pattern-mixture model in network meta-analysis of binary missing outcome data: one-stage or two-stage approach?

2020 ◽  
Author(s):  
Loukia Maria Spineli ◽  
Katerina Papadimitropoulou ◽  
Chrysostomos Kalyvas
Keyword(s):  
Mixture Model ◽  
Simulation Study ◽  
Meta Analysis ◽  
Credible Interval ◽  
Outcome Data ◽  
Systematic Bias ◽  
Two Stage ◽  
Pattern Mixture Model ◽  
One Stage ◽  
The One

Abstract Background: Trials with binary outcomes can be synthesised using within-trial exact likelihood or approximate normal likelihood in one-stage or two-stage approaches, respectively. The performance of the one-stage and the two-stage approaches has been documented extensively in the literature. However, little is known about how these approaches behave in the presence of missing outcome data (MOD), which are ubiquitous in trials. In this work, we compare the one-stage versus two-stage approach via a pattern-mixture model in the network meta-analysis using Bayesian methods to handle MOD appropriately.Methods: We used 29 published networks to empirically compare the two approaches concerning the relative treatment effects of several competing interventions and the between-trial variance ( ), while considering the extent and level of balance of MOD in the included trials. We additionally conducted a simulation study to compare the competing approaches regarding the bias and width of the 95% credible interval of the (summary) log odds ratios (OR) and in the presence of moderate and large MOD.Results: The empirical study did not reveal any systematic bias between the compared approaches regarding the log OR, but showed systematically larger uncertainty around the log OR under the one-stage approach for networks with at least one small trial or low event risk and moderate MOD. For these networks, the simulation study revealed that the bias in log OR for comparisons with the reference intervention in the network was relatively higher in the two-stage approach. Contrariwise, the bias in log OR for the remaining comparisons was relatively higher in the one-stage approach. Overall, bias increased for large MOD. For these networks, the empirical results revealed slightly higher estimates under the one-stage approach irrespective of the extent of MOD. The one-stage approach also led to less precise log OR and when compared with the two-stage approach for large MOD.Conclusions: Due to considerable bias in the log ORs overall, especially for large MOD, none of the competing approaches was superior. Until a more competent model is developed, the researchers may prefer the one-stage approach to handle MOD, while acknowledging its limitations.

scholarly journals Flexible piecewise linear model for investigating dose-response relationship in meta-analysis: methodology, examples, and comparison

2018 ◽  
Author(s):  
Chang Xu ◽  
Lehana Thabane ◽  
Tong-Zu Liu ◽  
Ling Li ◽  
Sayem Borhan ◽  
...  
Keyword(s):  
Linear Model ◽  
Dose Response ◽  
Linear Trend ◽  
Piecewise Linear ◽  
Meta Analysis ◽  
Continuous Exposure ◽  
Response Relationship ◽  
Dose Response Relationship ◽  
Two Stage ◽  
One Stage

Objectives: Dose-response meta-analysis (DRMA) is widely employed to establishing the potential dose-response relationship between continuous exposures and disease outcomes. However, no method is readily available for exploring the relation between a discrete exposure and a binary or continuous outcome. We proposed a piecewise linear (PL) DRMA model as a solution to this issue. Methods: We illustrated the methodology of PL model in both one-stage DRMA approach and two-stage DRMA approach. The method by testing the equality of slopes of each piecewise was employed to judge if there is “piecewise effect” against simple linear trend. We then used sleep (continuous exposure) and parity (discrete exposure) data as examples to illustrate how to apply PL model in DRMA using the Stata code attached. We also empirically compared the slopes of PL model with simple linear as well as restricted cubic spline (RCS) model. Results: Both one-stage and two-stage PL DRMA model fitted well in our examples, and the results were similar. Obvious “piecewise effects” were detected in both the two examples by the method we used. In our example, the PL model showed better fitting effect and practical reliable results compared to simple linear model, while similar results for to RCS model. Conclusion: Piecewise linear function is a simple and valid method for DRMA and can be used for discrete exposures. It also represents a superior model to linear model in DRMA and may be an alternative model to non-linear model.

scholarly journals Pattern-mixture model in network meta-analysis of binary missing outcome data: one-stage or two-stage approach?

2020 ◽  
Author(s):  
Loukia Maria Spineli ◽  
Katerina Papadimitropoulou ◽  
Chrysostomos Kalyvas
Keyword(s):  
Mixture Model ◽  
Simulation Study ◽  
Meta Analysis ◽  
Credible Interval ◽  
Outcome Data ◽  
Systematic Bias ◽  
Two Stage ◽  
Pattern Mixture Model ◽  
One Stage ◽  
The One

Abstract Background: Trials with binary outcomes can be synthesised using within-trial exact likelihood or approximate normal likelihood in one-stage or two-stage approaches, respectively. The performance of the one-stage and the two-stage approaches has been documented extensively in the literature. However, little is known about how these approaches behave in the presence of missing outcome data (MOD), which are ubiquitous in clinical trials. In this work, we compare the one-stage versus two-stage approach via a pattern-mixture model in the network meta-analysis using Bayesian methods to handle MOD appropriately.Methods: We used 29 published networks to empirically compare the two approaches concerning the relative treatment effects of several competing interventions and the between-trial variance ( ), while considering the extent and level of balance of MOD in the included trials. We additionally conducted a simulation study to compare the competing approaches regarding the bias and width of the 95% credible interval of the (summary) log odds ratios (OR) and in the presence of moderate and large MOD.Results: The empirical study did not reveal any systematic bias between the compared approaches regarding the log OR, but showed systematically larger uncertainty around the log OR under the one-stage approach for networks with at least one small trial or low event risk and moderate MOD. For these networks, the simulation study revealed that the bias in log OR for comparisons with the reference intervention in the network was relatively higher in the two-stage approach. Contrariwise, the bias in log OR for the remaining comparisons was relatively higher in the one-stage approach. Overall, bias increased for large MOD. For these networks, the empirical results revealed slightly higher estimates under the one-stage approach irrespective of the extent of MOD. The one-stage approach also led to less precise log OR and when compared with the two-stage approach for large MOD.Conclusions: Due to considerable bias in the log ORs overall, especially for large MOD, none of the competing approaches was superior. Until a more competent model is developed, the researchers may prefer the one-stage approach to handle MOD, while acknowledging its limitations.

Two-Stage Procedure for the Quantitative Determination of Autoprothrombin III Concentration and Some Applications

1967 ◽  
Vol 18 (01/02) ◽  
pp. 198-210 ◽  
Author(s):  
Ronald S Reno ◽  
Walter H Seegers
Keyword(s):  
Prothrombin Time ◽  
Factor Vii ◽  
Two Stage ◽  
Pronounced Increase ◽  
One Stage ◽  
Assay Procedure ◽  
Bovine Plasma ◽  
The One ◽  
Autoprothrombin C

SummaryA two-stage assay procedure was developed for the determination of the autoprothrombin C titre which can be developed from prothrombin or autoprothrombin III containing solutions. The proenzyme is activated by Russell’s viper venom and the autoprothrombin C activity that appears is measured by its ability to shorten the partial thromboplastin time of bovine plasma.Using the assay, the autoprothrombin C titre was determined in the plasma of several species, as well as the percentage of it remaining in the serum from blood clotted in glass test tubes. Much autoprothrombin III remains in human serum. With sufficient thromboplastin it was completely utilized. Plasma from selected patients with coagulation disorders was assayed and only Stuart plasma was abnormal. In so-called factor VII, IX, and P.T.A. deficiency the autoprothrombin C titre and thrombin titre that could be developed was normal. In one case (prethrombin irregularity) practically no thrombin titre developed but the amount of autoprothrombin C which generated was in the normal range.Dogs were treated with Dicumarol and the autoprothrombin C titre that could be developed from their plasmas decreased until only traces could be detected. This coincided with a lowering of the thrombin titre that could be developed and a prolongation of the one-stage prothrombin time. While the Dicumarol was acting, the dogs were given an infusion of purified bovine prothrombin and the levels of autoprothrombin C, thrombin and one-stage prothrombin time were followed for several hours. The tests became normal immediately after the infusion and then went back to preinfusion levels over a period of 24 hrs.In other dogs the effect of Dicumarol was reversed by giving vitamin K1 intravenously. The effect of the vitamin was noticed as early as 20 min after administration.In response to vitamin K the most pronounced increase was with that portion of the prothrombin molecule which yields thrombin. The proportion of that protein with respect to the precursor of autoprothrombin C increased during the first hour and then started to go down and after 3 hrs was equal to the proportion normally found in plasma.

Standardization of Factor VIII – IV. Establishment of the 3rd International Standard for Factor VIII: C Concentrate

1983 ◽  
Vol 50 (03) ◽  
pp. 697-702 ◽  
Author(s):  
T W Barrowcliffe ◽  
A D Curtis ◽  
D P Thomas
Keyword(s):  
Factor Viii ◽  
High Purity ◽  
Collaborative Study ◽  
World Health ◽  
Current Standard ◽  
Two Stage ◽  
International Standard ◽  
One Stage ◽  
The One ◽  
Health Organization

SummaryAn international collaborative study was carried out to establish a replacement for the current (2nd) international standard for Factor VIII: C, concentrate. Twenty-six laboratories took part, of which 17 performed one-stage assays, three performed two-stage assays and six used both methods. The proposed new standard, an intermediate purity concentrate, was assayed against the current standard, against a high-purity concentrate and against an International Reference Plasma, coded 80/511, previously calibrated against fresh normal plasma.Assays of the proposed new standard against the current standard gave a mean potency of 3.89 iu/ampoule, with good agreement between laboratories and between one-stage and two- stage assays. There was also no difference between assay methods in the comparison of high-purity and intermediate purity concentrates. In the comparison of the proposed standard with the plasma reference preparation, the overall mean potency was 4.03 iu/ampoule, but there were substantial differences between laboratories, and the two-stage method gave significantly higher results than the one stage method. Of the technical variables in the one-stage method, only the activation time with one reagent appeared to have any influence on the results of this comparison of concentrate against plasma.Accelerated degradation studies showed that the proposed standard is very stable. With the agreement of the participants, the material, in ampoules coded 80/556, has been established by the World Health Organization as the 3rd International Standard for Factor VIII :C, Concentrate, with an assigned potency of 3.9 iu/ampoule.

Sign in / Sign up

Export Citation Format

Share Document