Just a moment, the page is loading...
Browse All Studies
View List of Selected Studies and choose next steps >
Clear All Studies
Close
Login | Create Account
About Us
Data Sponsors
For Researchers ▼
Researchers Perspective
How It Works
Data Sharing Agreement
FAQs
Independent Review Panel
Metrics ▼
Metrics Overview
Agreed Proposals
Published Proposals
Help/Contact Us
Evaluation of a platelet count threshold for bleeding risk in thrombocytopenic patients with chronic liver disease who undergo diagnostic and therapeutic invasive procedures.
Proposal
1180
Title of Proposed Research
Evaluation of a platelet count threshold for bleeding risk in thrombocytopenic patients with chronic liver disease who undergo diagnostic and therapeutic invasive procedures.
Lead Researcher
Edoardo G. Giannini
Affiliation
Department of Internal Medicine, University of Genoa, Genoa, Italy.
Funding Source
None
Potential Conflicts of Interest
Consultancies: GSK
Data Sharing Agreement Date
15 December 2016
Lay Summary
This research needs to be done because, despite advancement in current knowledge, the threshold for platelet transfusion in patients with liver disease is not known.
This issue is relevant, as unnecessary transfusions (i.e., transfusions perfomed in patients who are not at risk of bleeding) may expose patients to transfusion-related risks though they do not require this practice. Furthermore, performing selective rather than acritical transfusions in patients who are at higher risk might make platelet transfusions more readily available to patients who are in the need of being transfused.
The aim of this study will be to identify if, among patients who have a similar degree of severity of disase, a subgroup of patients may avoid to be transfused without running the risk of bleeding.
The research will be conducted on data that are already available. The desing of the study will not take into account various arms according to the fact that in some patients a drug was used to increase platelet counts, as we will examine the actual platelet count of the day of the invasive procedure, regardless of how that platelet count was reached.
Application of statistics will identify a platelet count cut-off above which performing invasive manouvres can be regarded as a safe procedure even in patients with decreased platelet count.
The result of this study will not only help plan better studies in the future, but will also affect clinical practice providing an evidence-based, unbiased result that can be used to manage patients.
Study Data Provided
[{ "PostingID": 25, "Title": "GSK-TPL104054", "Description": "Medicine: eltrombopag, Condition: Liver Diseases, Phase: 3, GSK Clinical Study ID: TPL104054, Sponsor: GSK." }]
Statistical Analysis Plan
For this study we will identify the absolute risk of bleeding following an invasive procedure which is associated with platelet count.
The number and proportion of patients who bled is known as this was reported in the original study. What will be evaluated is the platelet count on the day of the procedure in the whole cohort. Patients will be divided into 2 groups (bleeders and non-bleeders) and the platelet counts in the two groups will be compared (Mann-Whitney U-test). Therafter, a Receiver Operating Characteristic curve will be built (platelet count is the variable of interest) so as to identify the platelet count with the highest accuracy for the identification of bleeding patients. Accuracy data (sensitivity, specificity, positive and begative predictive values, positive and negative likelihood ratios) will be obtained using the platelet count threshold identified by ROC curve analysis. Different clinical scenarios with various bleeding rates will be applied so as to evaluate variations in PPV and NPV.
Publication Citation
The publication citation will be added after the research is published.
To help us verify your identity, a validation code will be sent to you.
Phone Number (By providing your phone number, you agree to receive text messages. Message and data rates may apply.)
Email Address (Code will be sent via email. If you don't see the email in your inbox, please check your spam folder.)
Enter code here