Meta-analysis of time-to-event outcomes using a hazard-based approach: Comparison with other models, robustness and meta-regression

Heterogeneity in meta-analysis: a comprehensive overview

In recent years, meta-analysis has evolved to a critically important field of Statistics, and has significant applications in Medicine and Health Sciences. In this work we briefly present existing methodologies to conduct meta-analysis along with any discussion and recent developments accompanying them. Undoubtedly, studies brought together in a systematic review will differ in one way or another. This yields a considerable amount of variability, any kind of which may be termed heterogeneity. To this end, reports of meta-analyses commonly present a statistical test of heterogeneity when attempting to establish whether the included studies are indeed similar in terms of the reported output or not. We intend to provide an overview of the topic, discuss the potential sources of heterogeneity commonly met in the literature and provide useful guidelines on how to address this issue and to detect heterogeneity. Moreover, we review the recent developments in the Bayesian approach along with the various graphical tools and statistical software that are currently available to the analyst. In addition, we discuss sensitivity analysis issues and other approaches of understanding the causes of heterogeneity. Finally, we explore heterogeneity in meta-analysis for time to event data in a nutshell, pointing out its unique characteristics.

Systematic review and meta-analysis of the prognostic impact of lymph node micrometastasis and isolated tumour cells in patients with stage I-IIIA non-small cell lung cancer

Lymph node micrometastases could be one of the reasons for the high recurrence rate after complete surgical resection in stage I-IIIA non-small cell lung cancer (NSCLC). The standard evaluation of a single haematoxylin and eosin (H&E) slide of a paraffin-embedded section of a lymph node is insufficient for the detection of micrometastases, and there is a need for additional histopathological evaluation. The association of lymph node micrometastases with survival remains as yet unresolved. The aim of this systematic review and meta-analysis is to investigate if lymph node micrometastases and isolated tumour cells in patients with stage I-IIIA NSCLC, detected with multiple sectioning and/or immunohistochemistry (IHC) and/or reverse transcriptase polymerase chain reaction (RT-PCR), are associated with overall survival (OS) and disease-free survival (DFS) after surgical resection. We performed a meta-analysis of time-to-event outcomes based on 15 articles using ancillary techniques to detect micrometastases. We extracted the OS and DFS every 3-6 months after surgery, for patients with and without occult lymph node micrometastasis, from the survival curves published in each article. These data were used to reconstruct OS and DFS for 'micrometastasis' and 'no micrometastasis' groups. Based on all included studies that used IHC, serial sectioning, or RT-PCR, we found a 5-year OS of 55% (micrometastasis) vs. 75% (no micrometastasis), and a 5-year DFS of 53% (micrometastasis) vs. 75% (no micrometastasis). Patients with stage I-IIIA NSCLC with lymph node micrometastases detected by ancillary histopathological and molecular techniques have a significantly poorer OS and DFS compared to patients without lymph node micrometastases.

Hypersensitivity to Pegylated E.colia sparaginase as first-line treatment in contemporary paediatric acute lymphoblastic leukaemia protocols: a meta-analysis of the Ponte di Legno Toxicity working group

BACKGROUND

Hypersensitivity reactions to asparaginase challenge its use and occur frequently (30-75%) after native Escherichia Coli (E.coli) asparaginase. Comparison of incidence of allergic reactions to pegylated E.coli asparaginase (PEGasparaginase) across contemporary paediatric acute lymphoblastic leukaemia (ALL) protocols is lacking.

METHOD AND PATIENTS

Questionnaires were sent to all members of the international ALL Ponte di Legno Toxicity Working Group. Meta-analyses were conducted to estimate the incidence of three types of hypersensitivity (allergy, allergic-like reaction and silent inactivation). Information on protocol level regarding PEGasparaginase dosing regimen, administration route and use of therapeutic drug monitoring was collected for risk analysis.

RESULTS

Newly diagnosed patients with ALL (n = 5880), aged 1-24 years old, were enrolled in seven different upfront ALL protocols using PEGasparaginase as first-line treatment. The incidence of allergic reactions (sum of allergies and allergic-like reactions) [95% confidence interval] was 2% [1%; 3%] during induction and 8% [5%; 11%] during postinduction. Route of administration, number of doses, dosage and number of PEGasparaginase-free weeks did not significantly influence risk of hypersensitivity. Multivariate meta-regression analysis suggests that initiation of PEGasparaginase in postinduction and higher number of PEGasparaginase-free intervals increased the risk for allergic reactions. 9-16% and 23-29% of all hypersensitivities were allergic-like reactions and silent inactivation, respectively.

CONCLUSION

The incidence of allergic reactions is lower in protocols using PEGasparaginase as first-line treatment compared with that reported for E.coli asparaginase or PEGasparaginase after E.coli asparaginase. Postinduction phase, a higher number of PEGasparaginase-free intervals, and initiation of PEGasparaginase in postinduction phase are risk factors for allergic reactions. These results are important for planning of PEGasparaginase administrations in future frontline therapy.

Assessment of ethnic differences in sunitinib outcome between Caucasian and Asian patients with metastatic renal cell carcinoma: a meta-analysis

BACKGROUND

An increasing number of studies have reported ethnic differences in sunitinib outcome in metastatic renal cell carcinoma (mRCC) patients. However, a comprehensive analysis is still lacking. Therefore, we systematically collected available published data and performed a meta-analysis to compare sunitinib efficacy and toxicity in Asian and Caucasian mRCC patients.

METHODS

Data were extracted from published results from clinical trials, expanded access program and real-world clinical practice. Progression-free survival (or time to tumor progression), overall survival, objective response rate and adverse events were used as endpoints to evaluate the differences of sunitinib outcome between the two ethnicities. For adverse events, we focused the following clinically relevant side effects: diarrhea, fatigue, mucositis/stomatitis, hand-foot syndrome, hypertension, leukopenia, neutropenia and thrombocytopenia.

RESULTS

A total of 33 publications including 9977 patients were available for meta-analysis. The efficacy of sunitinib in Asian patients was similar to that in Caucasian patients. However, Asian patients showed a higher incidence of all grades toxicity of hand-foot syndrome, > grade 2 fatigue, > grade 2 hand-foot syndrome and > grade 2 thrombocytopenia.

CONCLUSION

Ethnic differences in adverse events of sunitinib in mRCC patients existed and dose adjustment in Asian patients may be considered.

Early subsidence of shape-closed hip arthroplasty stems is associated with late revision. A systematic review and meta-analysis of 24 RSA studies and 56 survival studies

BACKGROUND AND PURPOSE

Few studies have addressed the association between early migration of femoral stems and late aseptic revision in total hip arthroplasty. We performed a meta-regression analysis on 2 parallel systematic reviews and meta-analyses to determine the association between early migration and late aseptic revision of femoral stems.

PATIENTS AND METHODS

Of the 2 reviews, one covered early migration data obtained from radiostereometric analysis (RSA) studies and the other covered long-term aseptic revision rates obtained from survival studies with endpoint revision for aseptic loosening. Stems were stratified according to the design concept: cemented shape-closed, cemented force-closed, and uncemented. A weighted regression model was used to assess the association between early migration and late aseptic revision, and to correct for confounders. Thresholds for acceptable and unacceptable migration were determined in accordance with the national joint registries (≤ 5% revision at 10 years) and the NICE criteria (≤ 10% revision at 10 years).

RESULTS

24 studies (731 stems) were included in the RSA review and 56 studies (20,599 stems) were included in the survival analysis review. Combining both reviews for the 3 design concepts showed that for every 0.1-mm increase in 2-year subsidence, as measured with RSA, there was a 4% increase in revision rate for the shape-closed stem designs. This association remained after correction for age, sex, diagnosis, hospital type, continent, and study quality. The threshold for acceptable migration of shape-closed designs was defined at 0.15 mm; stems subsiding less than 0.15 mm in 2 years had revision rates of less than 5% at 10 years, while stems exceeding 0.15 mm subsidence had revision rates of more than 5%.

INTERPRETATION

There was a clinically relevant association between early subsidence of shape-closed femoral stems and late revision for aseptic loosening. This association can be used to assess the safety of shape-closed stem designs. The published research is not sufficient to allow us to make any conclusions regarding such an association for the force-closed and uncemented stems.

Two-stage meta-analysis of survival data from individual participants using percentile ratios

Methods for individual participant data meta-analysis of survival outcomes commonly focus on the hazard ratio as a measure of treatment effect. Recently, Siannis et al. (2010, Statistics in Medicine 29:3030-3045) proposed the use of percentile ratios as an alternative to hazard ratios. We describe a novel two-stage method for the meta-analysis of percentile ratios that avoids distributional assumptions at the study level.