Background
Condition or domain being studied
Chimeric antigen receptors (CARs) show a high affinity to bind effector cells of the immune system, such as T cells 1 . This cancer immunotherapy enables enhancement of the immunological response against malignant cells. Although the development of CAR T-cell (CAR-T) cell therapy started more than 20 years ago 2 , the first steps in their transfer to clinical practice are now taking place, following the recent authorization of different CAR-T therapies by the U.S Food and Drug Administration and the European Medicines Agency.
Why it is important to do this review
The earliest and most extensive research with CAR-T therapy has been carried out in hematologic malignancies, where it has pointed to high response rates in patients who generally have very poor prognosis such as refractory acute lymphoblastic lymphoma, diffuse large B-cell lymphoma, or multiple myeloma 3 – 5 . Few therapeutic options are available in this setting, and CAR-T therapy is showing encouraging results.
B-cell maturation antigen (BCMA) and CD19-targeting CAR-T cell therapies are the ones that are at a more advanced stage of research and have shown the best results so far 6 . At this time, CAR-T cells directed to other targets for treating solid tumors and infectious or autoimmune diseases are emerging 7 , 8 .
Clinical studies have suggested that CAR-T cell therapies can be even curative in patients who are responsive. However response rates can vary in different pathologies or depending on the characteristics of the patient or the disease. There is also a possibility of resistance and relapses. Moreover, these therapies can associate important adverse events such as cytokine release syndrome, neurologic adverse events and B-cell aplasia, which may lead to serious consequences 1 . To date, clinical trials have been mainly performed in terms of a single-arm design. Consequently, systematic reviews have focused on synthesizing overall response rates of CAR-T cell therapy, without providing results compared with a control group 9 – 12 .
A recently published systematic review focused on randomized controlled trials (RCTs) analyzing the efficacy and safety of CAR-T therapies versus high-dose salvage chemotherapy followed by autologous stem cell transplantation as second-line treatment of relapsed or refractory large B-cell lymphoma 13 . However, this review shows several limitations, such as the lack of diverse comparators and hematologic diseases, and methodological concerns with the pooled analysis strategy. A second systematic review showing remarkable differences with our living project was also identified, providing mixed results in adults and children, an updated search 6 months shorter than our study, less detail on pre-specified exploratory outcomes and no data available on complementary observational evidence 14 .
Although more than a thousand CAR-T trials are being conducted 15 , those with the highest quality, randomized clinical trials are still scarce. Some data from RCTs in different hematological pathologies have already been published (BELINDA 16 , ZUMA-7 17 , TRANSFORM 18 , KarMMa-3 19 or CARTITUDE-4 20 ), and other studies are currently underway (KarMMa-4, CARTITUDE-5, CARTITUDE-6 or DALY 2-EU).
The encouraging results obtained in non-comparative studies have allowed the approval of these therapies and their progressive introduction into clinical practice. However, all the above limitations do not make it possible to clarify in which patients there is a clearly favorable risk-benefit ratio or what place in therapy should be reserved to the existing therapies. Numerous phase III randomized clinical trials are already underway that will shed light on these questions in the near future. The development of CAR-T cell therapies will continue to expand and new CAR-T cells will emerge. Therefore, a living synthesis of the evidence with regular updates can be positioned as the best method to incorporate emerging evidence in a timely manner.
Considering the characteristics of the available evidence, the width variability of the SRs currently published, and taking into account that the evidence on this therapy is on the rise ( i.e. new studies have been published and other studies are going to be published in the near future), we propose to carry out a living systematic review that will make it possible to develop a rigorous and updated synthesis of the evidence on the efficacy and safety of CAR-T therapy in patients suffering from hematological malignancies. This SR will be developed as part of the Living Evidence to Inform Health Decisions project, which supports health system organizations in the implementation of a living process for the development of the synthesis of the evidence to inform health decisions 21 .
Objective
To determine the efficacy and safety of CAR-T therapy in the treatment of hematologic malignancies, in comparison with other current therapies
Methods
A living systematic review. This protocol is reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P). It shares a common methodological approach that has been defined on the basis of previous developments, as part of the Living Evidence to inform health decisions project 21 for developing multiple living systematic reviews and living overviews of reviews.
Types of studies
Randomized controlled trials (RCTs) will be prioritized. In the absence of evidence from RCTs, non-randomized controlled studies (quasi-experimental studies, cohort studies, case-control studies) will be considered for inclusion. Studies with other designs, such as those with historical controls and single-arm studies will not be considered. Studies must analyze at least one primary or secondary outcome.
Types of participants
Studies including participants diagnosed with hematologic diseases, such as multiple myeloma, leukemia and lymphoma of any type will be included. Both adults (≥ 18 years) and pediatric patients will be considered. Untreated patients, as well as patients previously treated, will be included, irrespective of the type of treatment or treatment line.
Intervention
Any CAR-T therapy type will be considered regardless of the T-cell origin (allogenic or autologous), target antigen, costimulatory domain, or infusion dose.
Comparator
The comparator will consist of chemotherapy or any other active pharmacologic treatment, hematopoietic stem cell transplantation (HSCT), best supportive care or any other intervention.
Outcome variables
Efficacy assessment
The primary outcome will be overall survival (OS).
Secondary efficacy variables will be mainly based according to the Center for International Blood & Marrow Transplant Research (CIBMTR) criteria ( https://www.manula.com/manuals/cibmtr/fim/1/en/pdf ) and consist of:
Overall response rate (ORR) (proportion of patients who have a PR or CR)
Complete Response/Remission (CR) rate (including Stringent Complete Response (sCR) and negative minimal residual disease (MRD) when provided)
Partial Response/Remission (PR) rate (including Very Good Partial Response (VGPR) and positive MRD when provided)
Relapse from CR
Progression-free survival (PFS)
Time period between CAR-T administration and transplantation
Safety assessment
Secondary safety outcomes will be the following:
Serious adverse events (SAE)
Incidence of cytokine-release syndrome (grade ≥3)
Graft-versus-host disease
Neurotoxicity (grade ≥3)
Total adverse events (grade ≥3)
Quality of life (QoL)
All efficacy and safety outcomes will be measured at the longest reported follow-up. For analyzing efficacy outcomes, variables will be considered as reported by the individual studies as long as internationally accepted criteria are used. The homogeneity of the criteria used for determining the outcomes in the different studies will be analyzed. The definition established by the authors of the primary publications will be acceptable if decided by consensus among the research team and it is consistent with the internationally accepted criteria. Results will not be combined when substantial heterogeneity (I 2 >60%) is found.
The International Conference on Harmonization Guidelines defined SAE as any event that led to death, was life-threatening, caused hospitalization or prolongation of existing hospitalization, led to persistent or significant disability, or any congenital anomaly or birth defect 22 . The National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) will be followed for the grading of the adverse events ( https://ctep.cancer.gov/protocoldevelopment/electronic_applications/ctc.htm ).
Methods for identification of studies
The literature search will be designed and launched through the Epistemonikos-L·OVE platform ( https://app.iloveevidence.com ). This platform has been validated as a repository for the COVID-19 showing to be a highly comprehensive source of evidence 23 , 24 . The following approach will be used:
Identification of terms relevant to the population and intervention components of the search strategy, using Word2vec technology to the corpus of documents available in Epistemonikos Database.
Discussion of terms with content and methods experts to identify relevant, irrelevant and missing terms.
Creation of a sensitive boolean strategy encompassing all the relevant terms.
The main search source will be the Epistemonikos database ( https://www.epistemonikos.org ), a comprehensive database of systematic reviews and other types of evidence, maintained by screening multiple information sources to identify systematic reviews and their included primary studies, including Cochrane Database of Systematic Reviews, MEDLINE, EMBASE, CINAHL, PsycINFO, LILACS, DARE, HTA Database, Campbell database, JBI Database of Systematic Reviews and Implementation Reports, EPPI-Centre Evidence Library 24 .
An additional search will be performed on MEDLINE in order to identify randomized trials and non-randomized studies. The searches will cover the inception date of each database. No publication status or language restriction will be applied to the searches in Epistemonikos or to the additional searches. We will apply validated filters to identify clinical trials in the MEDLINE database.
Boolean search strategy
Epistemonikos
(leukem* OR leukaem* OR leucem* OR leucaem* OR lymphoma* OR lymphobl* OR ((hemato* OR haemato* OR lymph* OR myelo*) AND (malignan* OR malignan*))) AND ((("car-engineered" OR "car-modified" OR "receptor-modified" OR chimeric* OR adoptive* OR redirected* OR engineered*) AND ("t cell" OR "t cells" OR "t-cell" OR "t-cells")) OR "car-t" OR "car t" OR "car-ts" OR "car ts" OR "car t-cell" OR "car t-cells")
Medline. PUBMED
(leukem* OR leukaem* OR leucem* OR leucaem* OR lymphoma* OR lymphobl* OR ((hemato* OR haemato* OR lymph* OR myelo*) AND (malignan* OR malignan*))) AND ((("car-engineered" OR "car-modified" OR "receptor-modified" OR chimeric* OR adoptive* OR redirected* OR engineered*) AND ("t cell" OR "t cells" OR "t-cell" OR "t-cells")) OR "car-t" OR "car t" OR "car-ts" OR "car ts" OR "car t-cell" OR "car t-cells")
Results from these searches will be automatically included in the L.OVE platform of the Epistemonikos Foundation ( https://iloveevidence.com/ ).
Other sources
We will carry out a manual search for reviewing the reference list of included studies, guidelines, narrative reviews and any other document of interest in an attempt to identify additional references that might be potentially eligible and to keep monitoring for the new evidence.
Study selection
The L·OVE platform repository will automatically incorporate all the references that meet the criteria of the research question. The titles and abstracts of the identified references will be independently screened by at least two authors. Full texts of all potentially eligible references will be retrieved to make a final decision on their inclusion. Reasons for the exclusion of the identified references will be registered. The screening process will be conducted through Rayyan-Intelligent Systematic Review software whenever necessary ( https://www.rayyan.ai/ ). Results of the screening process will be shown in a PRISMA 25 flow diagram.
Data extraction
A specific data extraction form will be designed to incorporate data obtained from the included studies. Information will be independently extracted by two authors, and cross-checked by another author. Discrepancies will be addressed by discussion, and if necessary a third author will be also involved.
The following information will be extracted: 1) general information of the studies, study design, number of participants, follow-up time, type of centre (Unicenter, multicenter), country, study sponsor and sources of funding; 2) Trial methods and conduction, such as randomization, blinding and follow-up time; 3) characteristics of the participants including age, sex, type of pathology, stage, number of previous lines, previous HSCT, the median time from diagnosis, percentage with high-risk cytogenetics, percentage with extramedullary disease; 4) Intervention characteristics such as T-cell origin (allogenic or autologous), transfection/transduction method (lentivirus, retrovirus, electroporation, or others), target antigen, CAR costimulatory domain, infusion dose; 5) Characteristics of the comparator; 6) Results of the measured outcomes.
The need for interleukin receptor antagonists to treat cytokine-release syndrome in the participants in the CAR-T arm will be analyzed.
We will contact the authors of the primary studies in case of missing information in the retrieved studies. Whenever possible, individual patient data and clinical study reports of included studies will be obtained by contacting corresponding authors, sponsors and promoters.
Data analysis
Risk of bias assessment
In the case of RCTs, the risk of bias will be evaluated using the Cochrane Risk of Bias Tool. This tool considers the following domains: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other bias 26 . We will rate the overall risk of bias of each study as low, moderate, high or unclear risk of bias.
The risk of bias of non-randomized studies of intervention will be assessed using the ROBINS-I tool 27 . The tool evaluates the following domains: bias due to confounding, selection of participants, classification of interventions, deviation from intended interventions, missing data, measurement of outcomes, and bias in selection of the reported result. According to these domains, we will judge the overall risk of bias as low, moderate, serious, critical or no information.
Measures of treatment effect
In the case of binary outcomes results will be presented as the proportion of patients who suffered the corresponding analyzed event. We will use hazard ratio (HR), risk ratio (RR) and Odds ratio (OR) for summarizing the results as appropriate. We will also estimate the risk difference in absolute terms and also the number needed to treat for an additional beneficial or harmful outcome when applicable. We will determine the 95% confidence intervals (CI).
For quantitative data, mean difference (MD) or standardized mean difference (SMD) will be estimated as appropriate. We will also determine the 95%CI.
We will accept any general or specific validated scale used to estimate QoL. Results that refer to specific domains of a QoL scale will be analyzed separately from results that refer to the overall results of the scale. In the event the identified studies reported QoL using the same scale, these data will be meta-analyzed and reported as MD with 95%CI. On the contrary, if the studies included in the review were to use different validated scales for reporting QoL, data will be standardized before an aggregated SMD with 95%CI is calculated.
Data synthesis
Results will be analyzed separately according to each specific study design and type of hematologic disease. Data will be analyzed following an intention-to-treat approach. We will combine outcomes across trials using a fixed-effects model when applicable.
The heterogeneity of treatment effect among studies will be analyzed by estimating the I² statistic. We will consider I²>60% as indicative of substantial heterogeneity, whose possible reasons will be explored by carrying out subgroup analyses. We will carry out meta-analyses when data from more than one study are available and unless substantial heterogeneity (I²>60%) is found, whose results will be represented using forest plots. We will carry out meta-analyses using Review Manager Software (RevMan version 5.4).
Results will be presented in a narrative form in case there is one only study that provides data for an outcome variable or when data obtained from different studies cannot be combined.
Subgroup analysis
When possible, overall survival rate, overall response rate and total grade 3 or higher adverse events will be analyzed according to the following aspects:
-Type of CAR-T therapy (antigen type, costimulatory domain)
-Type of comparator
-Age group (≥18 years and <18 years)
-Treatment line
-Tumor burden
-Cancer stage
Sensitivity analysis
Sensitivity analyses will be performed in terms of the following criteria:
-Analysis excluding studies with a high risk of bias.
-Analysis excluding industry-sponsored studies.
Certainty of evidence
The quality of the evidence for each outcome variable will be analyzed using GRADE methodology, which includes the assessment of five main aspects: risk of bias, directness of the evidence, consistency among trials’ results, the precision of effect estimates, and risk of publication bias 28 .
Two reviewers will independently assess the quality of the evidence. All possible discrepancies will be discussed until consensus and by the involvement of a third author, if necessary. Results of this assessment will be presented in a Summary of Findings table generated in the tool GRADEpro ® ( https://www.gradepro.org/ )
Evidence monitoring and surveillance plan
In order to maintain the living evidence process for this review, the Epistemonikos-L.OVE platform ( https://iloveevidence.com/ ) will be used as a technological enabler to support evidence identification, screening, and selection. We will keep a living search in the L·OVE platform to detect systematic reviews and randomized controlled trials. Additionally, every three months, we will manually search for ongoing studies in the WHO International Clinical Trials Registry Platform and in clinicaltrials.gov.
One reviewer will be in charge of assessing the evidence that has entered the specific question in the L.OVE platform every month and applying the selection criteria presented above. If a potentially eligible study is found, a second reviewer will confirm its eligibility by reading the full text. Results of evidence surveillance will be collected and kept as part of the study records. Information on the PRISMA figure will be updated accordingly. The PICO question and criteria for selecting studies will be revised and changed accordingly during the Living Evidence process every time new eligible studies are incorporated into the evidence synthesis on main outcomes or every four months, whichever is reached first. .
We will carry out data extraction for all new eligible studies. The data synthesis will be updated immediately after that, and the quality of the evidence will be evaluated following the GRADE approach accordingly, looking for changes on the certainty assessment results.
The living process for this question will end when the certainty of the evidence on the updated estimates for the desirable and undesirable effects becomes high or after 36 months of surveillance, whatever is reached first.
Statistical considerations for the living evidence synthesis
The inclusion of new studies identified as part of evidence surveillance and reporting on the outcomes of interest will follow this approach: We will perform an updated meta-analysis for each of the outcomes of interest including data obtained from the new studies. Data will be analyzed using a fixed-effect model. Heterogeneity among included studies will be analyzed by using the I 2 statistics. If new heterogeneity is detected (compared to the previous meta-analysis, new heterogeneity appears or increases), we will explore its potential sources by reviewing the new studies against previously included studies in order to identify reasons that may explain inconsistent results among studies. In the presence of unexplained heterogeneity (I 2 > 60%), data will not be meta-analyzed, and the results of the evidence synthesis will be explained narratively.
Dissemination plan
We plan to communicate our review results (from baseline review and its updates) by publication in a scientific journal. We will elaborate technical reports to the healthcare providers, decision-makers, and clinical committees. We will share the results through our social media channels and in Open Science Framework ( https://osf.io/ ). The base line report as well as the periodical updates will be available on the Living Evidence to Inform Health Decisions project website ( https://livingevidenceframework.com/en/ ).
If during the living process, new relevant results that imply changes in the current clinical practice are identified, we will update the report of this review and disseminate the new evidence and conclusions among potential users through all the above described communication channels.
Study status
The searching and screening phases have been completed at the time of this submission.
Acknowledgement
We would like to acknowledge the contribution of Angie Santafe and Laura Trujillo, master students from the Autonomous University of Barcelona (UAB) for supporting the search screening and submission of this protocol.
Funding Statement
This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 894990).
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
[version 2; peer review: 1 approved, 2 approved with reservations]
Data availability
Underlying data
No data are associated with this article.
Extended data
Open Science Framework: LIVING EVIDENCE TO INFORM HEALTH DECISIONS. 10.17605/OSF.IO/V6HDX
Reporting guidelines
Open Science Framework: PRISMA-P checklist for “Efficacy and safety of Chimeric Antigen Receptor T-Cell (CAR-T) therapy in hematologic malignancies: a living systematic review (Protocol)”. 10.17605/OSF.IO/V6HDX
Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).
References
1. June CH, Sadelain M: Chimeric Antigen Receptor therapy. N Engl J Med. 2018;379(1):64–73. 10.1056/NEJMra1706169
2. Eshhar Z, Waks T, Gross G, et al. : Specific activation and targeting of cytotoxic lymphocytes through chimeric single chains consisting of antibody-binding domains and the gamma or zeta subunits of the immunoglobulin and T-cell receptors. Proc Natl Acad Sci U S A. 1993;90(2):720–4. 10.1073/pnas.90.2.720
3. Maude SL, Laetsch TW, Buechner J, et al. : Tisagenlecleucel in children and young adults with B-Cell lymphoblastic leukemia. N Engl J Med. 2018;378(5):439–48. 10.1056/NEJMoa1709866
4. Schuster SJ, Bishop MR, Tam CS, et al. : Tisagenlecleucel in adult relapsed or refractory diffuse large B-Cell lymphoma. N Engl J Med. 2019;380(1):45–56. 10.1056/NEJMoa1804980
5. Raje N, Berdeja J, Lin Y, et al. : Anti-BCMA CAR T-cell therapy bb2121 in relapsed or refractory multiple myeloma. N Engl J Med. 2019;380(18):1726–37. 10.1056/NEJMoa1817226
6. Grigor EJM, Fergusson D, Kekre N, et al. : Risks and benefits of Chimeric Antigen Receptor T-cell (CAR-T) therapy in cancer: a systematic review and meta-analysis. Transfus Med Rev. 2019;33(2):98–110. 10.1016/j.tmrv.2019.01.005
7. Tian Y, Li Y, Shao Y, et al. : Gene modification strategies for next-generation CAR T cells against solid cancers. J Hematol Oncol. 2020;13(1): 54. 10.1186/s13045-020-00890-6
8. Maldini CR, Ellis GI, Riley JL: CAR T cells for infection, autoimmunity and allotransplantation. Nat Rev Immunol. 2018;18(10):605–16. 10.1038/s41577-018-0042-2
9. Yang Q, Li X, Zhang F, et al. : Efficacy and safety of CAR-T therapy for relapse or refractory Multiple Myeloma: a systematic review and meta-analysis. Int J Med Sci. 2021;18(8):1786–97. 10.7150/ijms.46811
10. Hu L, Charwudzi A, Li Q, et al. : Anti-CD19 CAR-T cell therapy bridge to HSCT decreases the relapse rate and improves the long-term survival of R/R B-ALL patients: a systematic review and meta-analysis. Ann Hematol. 2021;100(4):1003–12. 10.1007/s00277-021-04451-w
11. Wang N, Meng Y, Wu Y, et al. : Efficacy and safety of Chimeric Antigen receptor T cell immunotherapy in B-cell non-hodgkin lymphoma: a systematic review and meta-analysis. Immunotherapy. 2021;13(4):345–57. 10.2217/imt-2020-0221
12. Anagnostou T, Riaz IB, Hashmi SK, et al. : Anti-CD19 Chimeric Antigen Receptor T-cell therapy in acute lymphocytic leukaemia: a systematic review and meta-analysis. Lancet Haematol. 2020;7(11):e816–26. 10.1016/S2352-3026(20)30277-5
13. Shargian L, Raanani P, Yeshurun M, et al. : Chimeric antigen receptor T-cell therapy is superior to standard of care as second-line therapy for large B-cell lymphoma: a systematic review and meta-analysis. Br J Haematol. 2022;198(5):838–46. 10.1111/bjh.18335
14. Pasqui DM, Latorraca COC, Pacheco RL, et al. : CAR-T cell therapy for patients with hematological malignancies. a systematic review. Eur J Haematol. 2022;109(6):601–618. 10.1111/ejh.13851
15. Wang V, Gauthier M, Decot V, et al. : Systematic review on CAR-T cell clinical trials up to 2022: academic center input. Cancers (Basel). 2023;15(4): 1003. 10.3390/cancers15041003
16. Bishop MR, Dickinson M, Purtill D, et al. : Second-line tisagenlecleucel or standard care in aggressive B-cell lymphoma. N Engl J Med. 2022;386(7):629–639. 10.1056/NEJMoa2116596
17. Locke FL, Miklos DB, Jacobson CA, et al. : Axicabtagene ciloleucel as second-line therapy for large B-cell lymphoma. N Engl J Med. 2022;386(7):640–654. 10.1056/NEJMoa2116133
18. Kamdar M, Solomon SR, Arnason J, et al. : Lisocabtagene maraleucel versus standard of care with salvage chemotherapy followed by autologous stem cell transplantation as second-line treatment in patients with relapsed or refractory large B-cell lymphoma (TRANSFORM): results from an interim analysis of an open-label, randomised, phase 3 trial. Lancet. 2022;399(10343):2294–2308. 10.1016/S0140-6736(22)00662-6
19. Rodríguez-Otero P, Ailawadhi S, Arnulf B, et al. : Ide-cel or standard regimens in relapsed and refractory multiple myeloma. N Engl J Med. 2023;388(11):1002–1014. 10.1056/NEJMoa2213614
20. San-Miguel J, Dhakal B, Yong K, et al. : Cilta-cel or standard care in lenalidomide-refractory multiple myeloma. N Engl J Med. 2023;389(4):335–347. 10.1056/NEJMoa2303379
21. Rojas-Reyes MX, Urrutia Chuchí G, Rada G, et al. : Implementing living evidence to inform health decisions: a strategy for building capacity in health sector (Protocol) [version 1; peer review: 2 approved with reservations]. Open Res Eur. 2021;1:114. 10.12688/openreseurope.14041.1
22. Agency EM: ICH E2A Clinical safety data management: definitions and standards for expedited reporting.(Reference: CPMP/ICH/377/95). Reference Source
23. Verdugo-Paiva F, Vergara C, Ávila C, et al. : COVID-19 L·OVE Repository is highly comprehensive and can be used as a single source for COVID-19 studies. medRxiv. 2021. 10.1101/2021.09.21.21263849
24. Rada G, Pérez D, Araya-Quintanilla F, et al. : Epistemonikos: a comprehensive database of systematic reviews for health decision-making. BMC Med Res Methodol. 2020;20(1): 286. 10.1186/s12874-020-01157-x
25. Page MJ, McKenzie JE, Bossuyt PM, et al. : The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. 10.1136/bmj.n71
26. Higgins JPT, Green S: Cochrane handbook for systematic reviews of interventions.Version 5.1.0. The Cochrane Collaboration,2011. Reference Source
27. Sterne JA, Hernán MA, Reeves BC, et al. : ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355: i4919. 10.1136/bmj.i4919
28. Schünemann H, Brożek J, Guyatt G, et al. : Handbook for grading the quality of evidence and the strength of recommendations using the GRADE approach.Updated2013. Reference Source