Individual patient data meta-analysis in cancer

Immunotherapy Efficacy in the Initial Lines of Treatment in Advanced Upper Gastrointestinal Malignancies: A Systematic Review of the Literature

Background

The therapeutic role of immune checkpoint inhibitors (ICIs) has represented the cutting edge of clinical research in upper gastrointestinal (GI) malignancies, with these agents now included in the armamentarium of treatment options for advanced gastric and esophageal cancers.

Methods

We performed a systematic literature review and pooled analysis to map out the currently available robust clinical evidence for the use of ICIs in upper GI cancers. Immunotherapy (IO), either as monotherapy or in combination with chemotherapy, and its role in first-line, maintenance, and second-line settings, as well as in specific clinical and biological subgroups, were critically appraised. All statistical tests were 2-sided.

Results

ICIs, in combination with chemotherapy, have provided statistically significant overall survival benefit in the first-line setting in gastric and gastro-esophageal adenocarcinomas (hazard ratio [HR] = 0.83, 95% confidence interval [CI] = 0.76 to 0.90, P < .001; based on 4 studies) and esophageal squamous cell carcinoma (HR = 0.72, 95% CI = 0.64 to 0.81, P < .001; based on 3 studies), albeit with heterogeneous efficacy according to biomarker expression. Patients with esophageal squamous cell carcinoma, and in particular high programmed cell death ligand-1 expression, derive survival benefit when treated with IO in the second-line setting (HR = 0.74, 95% CI = 0.68 to 0.82, P < .001; for any level of programmed cell death ligand-1 expression). Clinical trials interrogating the combination of IO with chemotherapy in second-line treatment should be seriously considered in upper GI adenocarcinomas. The role of maintenance IO after initial disease control is still unclear and cannot be recommended. Impressive response rates and survival benefit from IO have been reported in patients with microsatellite instability-high tumors (HR = 0.33, 95% CI = 0.19 to 0.57, P < .001), and this warrants further prospective biomarker-driven studies.

Conclusions

IO is changing the treatment landscape in upper GI malignancies. The rapidly developing evidence in the field needs to be critically appraised while further validation of the existing information from ongoing trials is awaited.

Comparing two approaches for meta-analysis of binary outcomes

In the present study, meta-analysis of binary outcome was undertaken by using two approaches namely Summary Statistics (SS) and Individual Animal Data (IAD) approach for obtaining more reliable estimates of the association of risk factors [breed (crossbred & indigenous), parity (primiparous & multiparous), age (< 5 years & > 5 years) and milk yield (kg)] with clinical mastitis [binary outcome (yes or No)] in dairy cows in India. For the present study the data on mastitis were compiled from three organized cattle farms. The results of bootstrapping showed that the pooled estimates under Individual Animal Data (IAD) approach were significantly higher than Summary Statistics (SS) appraoch for all unadjusted risk factors. However the results of both approaches were similar under covariate-adjusted circumstances. In case of heterogeneity of effects across farms, Individual Animal Data (IAD) approach provides more reliable information than Summary Statistics (SS) approach. Therfore it was also concluded that the crossbreds (adjusted for age) cows (1.47 times), older cows (1.85 times) multiparous cows (2.21 times) and high yielders (1.67 times) cows were at higher risk of mastitis than their respective reference categories. Therefore Individual Animal Data (IAD) approach is an appropriate approach for animal science data as it is more reliable and perform better in heterogenous conditions.

Effects and moderators of exercise on sleep in adults with cancer: Individual patient data and aggregated meta-analyses

OBJECTIVES

To evaluate the effects of exercise interventions on sleep disturbances and sleep quality in patients with mixed cancer diagnoses, and identify demographic, clinical, and intervention-related moderators of these effects.

METHODS

Individual patient data (IPD) and aggregated meta-analyses of randomized controlled trials (RCTs). Using data from the Predicting OptimaL cAncer RehabIlitation and Supportive care project, IPD of 2173 adults (mean age = 54.8) with cancer from 17 RCTs were analyzed. A complementary systematic search was conducted (until November 2018) to study the overall effects and test the representativeness of analyzed IPD. Effect sizes of exercise effects on self-reported sleep outcomes were calculated for all included RCTs. Linear mixed-effect models were used to evaluate the effects of exercise on post-intervention outcome values, adjusting for baseline values. Moderator effects were studied by testing interactions for demographic, clinical and intervention-related characteristics.

RESULTS

For all 27 eligible RCTs from the updated search, exercise interventions significantly decreased sleep disturbances in adults with cancer (g = -0.09, 95% CI [-0.16; -0.02]). No significant effect was obtained for sleep quality. RCTs included in IPD analyses constituted a representative sample of the published literature. The intervention effects on sleep disturbances were not significantly moderated by any demographic, clinical, or intervention-related factor, nor by sleep disturbances.

CONCLUSIONS

This meta-analysis provides some evidence that, compared to control conditions, exercise interventions may improve sleep disturbances, but not sleep quality, in cancer patients, although this effect is of a small magnitude. Among the investigated variables, none was found to significantly moderate the effect of exercise interventions on sleep disturbances.

Partially systematic thoughts on the history of systematic reviews

Six years after the launch of Systematic Reviews by Biomed Central, this article is part of the celebration of the journal. It contains personal reflections on the past, present and future of systematic reviews, using examples relevant to the role of systematic reviews in cataloguing and analysing research, assessing quality and planning new studies. The focus is on the most common of the various types of systematic review in health and social care, namely those assessing the effects of interventions.

Weekly versus three weeks chemotherapy for advanced ovarian cancer: a meta-analysis

Aim

Three weeks paclitaxel and carboplatin has been considered the standard of care for primary treatment of ovarian cancer (OC). Whether weekly therapy will further improve the clinical outcomes or not is still unclear. We conducted a meta-analysis to compare the two regimens.

Method

Articles were selected with a systematic approach, using PubMed databases. Trials concerning comparison between carboplatin plus weekly paclitaxel (dose-dense regimen) and carboplatin plus paclitaxel every 3 weeks were considered. Outcomes included overall survival (OS), progression free survival (PFS) and severe acute toxicity.

Results

Dose-dense regimen was associated with significant improvement of PFS compared with standard schedule, with HR of 0.73 (95% CI 0.61-0.88, p = 0.001). There was no difference in OS between treatment regimens (HR 0.95, 95% CI 0.77-1.16, p=0.06), as well as in term of severe acute toxicity.

Conclusion

Dose-dense regimen is superior to standard schedule in terms of PFS. Further studies are necessary to firmly confirm this evidence in advanced OC treatment.

Technical success, technique efficacy and complications of minimally-invasive imaging-guided percutaneous ablation procedures of breast cancer: A systematic review and meta-analysis

OBJECTIVES

To systematically review studies concerning imaging-guided minimally-invasive breast cancer treatments.

METHODS

An online database search was performed for English-language articles evaluating percutaneous breast cancer ablation. Pooled data and 95% confidence intervals (CIs) were calculated. Technical success, technique efficacy, minor and major complications were analysed, including ablation technique subgroup analysis and effect of tumour size on outcome.

RESULTS

Forty-five studies were analysed, including 1,156 patients and 1,168 lesions. Radiofrequency (n=577; 50%), microwaves (n=78; 7%), laser (n=227; 19%), cryoablation (n=156; 13%) and high-intensity focused ultrasound (HIFU, n=129; 11%) were used. Pooled technical success was 96% (95%CI 94-97%) [laser=98% (95-99%); HIFU=96% (90-98%); radiofrequency=96% (93-97%); cryoablation=95% (90-98%); microwave=93% (81-98%)]. Pooled technique efficacy was 75% (67-81%) [radiofrequency=82% (74-88); cryoablation=75% (51-90); laser=59% (35-79); HIFU=49% (26-74)]. Major complications pooled rate was 6% (4-8). Minor complications pooled rate was 8% (5-13%). Differences between techniques were not significant for technical success (p=0.449), major complications (p=0.181) or minor complications (p=0.762), but significant for technique efficacy (p=0.009). Tumour size did not impact on variables (p>0.142).

CONCLUSIONS

Imaging-guided percutaneous ablation techniques of breast cancer have a high rate of technical success, while technique efficacy remains suboptimal. Complication rates are relatively low.

KEY POINTS

• Imaging-guided ablation techniques for breast cancer are 96% technically successful. • Overall technique efficacy rate is 75% but largely inhomogeneous among studies. • Overall major and minor complication rates are low (6-8%).

Individual patient data meta-analysis of time-to-event outcomes: one-stage versus two-stage approaches for estimating the hazard ratio under a random effects model

Meta-analyses of individual patient data (IPD) provide a strong and authoritative basis for evidence synthesis. IPD are particularly useful when the outcome of interest is the time to an event. Methodological developments now enable the meta-analysis of time-to-event IPD using a single model, allowing treatment effect and across-trial heterogeneity parameters to be estimated simultaneously. This differs from the standard approaches used with aggregate data, and also predominantly with IPD. Facilitated by a simulation study, we investigate what these new 'one-stage' random-effects models offer over standard 'two-stage' approaches. We find that two-stage approaches represent a robust, reliable and easily implementable way to estimate treatment effects and account for heterogeneity. Nevertheless, one-stage models can be used to provide a deeper insight into the data. Software for fitting one-stage Cox models with random effects using Restricted Maximum Likelihood methodology is made available, and its use demonstrated on an IPD meta-analysis assessing post-operative radio therapy for patients with non-small cell lung cancer. Copyright © 2011 John Wiley & Sons, Ltd.

Individual patient data meta-analysis is needed in Chinese medical research

Publication biases and collection limitations are the main disadvantages of a traditional meta-analysis based on aggregate patient data (APD) from published articles. Individual patient data (IPD) meta-analysis, as the gold standard of systematic review, is a possible alternative in this context. However, the publications relative to IPD meta-analyses are still rare compared with the traditional ones, especially in the research oriented to Chinese medicine (CM). In this article, the strengths and detailed functioning of IPD meta-analysis are described. Furthermore, the need for IPD meta-analysis to assess the treatments based on CM was also discussed. Compared with the traditional APD meta-analysis, the IPD meta-analysis might give a more accurate and unbiased assessment and is worth to be recommended to CM researchers.

Locoregional recurrence after breast cancer surgery: a systematic review by receptor phenotype

Molecular subtyping confirms that breast cancer comprises at least four genetically distinct entities based on the expression of specific genes including estrogen receptor (ER), progesterone receptor (PR), and HER2/neu receptor. The quantitative influence of subtype on ipsilateral locoregional recurrence (LRR) is unknown. The aim of this study was to systematically appraise the influence of breast cancer subtype on LRR following breast conserving therapy (BCT) and mastectomy. A comprehensive search for studies examining outcomes after BCT and/or mastectomy according to breast cancer subtype was performed using Medline and cross-referencing available data. Reviews of each study were conducted and data extracted to perform meta-analysis. Primary outcome was LRR related to breast cancer subtype. A total of 12,592 breast cancer patients who underwent either BCT (n = 7,174) or mastectomy (n = 5,418) were identified from 15 studies. Patients with luminal subtype tumors (ER/PR +ve) had a lower risk of LRR than both triple-negative (RR 0.38; 95% CI 0.23-0.61); and HER2/neu-overexpressing (RR 0.34; 95% CI 0.26-0.45) tumors following BCT. Luminal tumors were also less likely to develop LRR than HER2/neu-overexpressing (OR 0.69; 95% CI 0.54-0.89) or triple-negative tumors (OR 0.61; 95% CI 0.46-0.79) after mastectomy. HER2/neu-overexpressing tumors have increased risk of LRR compared to triple-negative tumors (RR 1.44; 95% CI 1.06-1.95) following BCT but there was no difference in LRR between HER2/neu-overexpressing and triple-negative tumors following mastectomy (RR 0.91; 95% CI 0.68-1.22). Luminal tumors exhibit the lowest rates of LRR. Patients with triple-negative and HER2/neu-overexpressing breast tumors are at increased risk of developing LRR following BCT or mastectomy. Breast cancer subtype should be taken into account when considering local control and identifies those at increased risk of LRR, who may benefit from more aggressive local treatment.

Meta-analysis: A brief introduction

Meta-analysis is the process of combining data from multiple sources and analyzing it together to increase power and provide a clearer picture of the effect of intervention or exposure on an outcome. The process is not complicated, but requires a great deal of attention to detail. A specific set of inclusion criteria for studies must be defined. Published or available study results may be affected by publication bias of several different types, so the researcher should be sure to conduct a thorough search of available databases in order to include unpublished findings. Following selection, the group of studies should be examined using funnel plots or statistical tests. Meta-analysis models themselves must be selected to properly reflect the combined studies. Both fixed- and random-effects modeling are discussed. Two case studies are presented, illustrating a well-conducted meta-analysis and a meta-analysis that was more controversial.

Overall survival benefit for weekly vs. three-weekly taxanes regimens in advanced breast cancer: A meta-analysis

BACKGROUND

Taxanes have been extensively tested in patients with advanced breast cancer, but it is unclear whether their weekly use might offer any benefits against standard every three weeks administration. We therefore performed a meta-analysis of randomized controlled trials that compared weekly and every three weeks taxanes regimens in advanced breast cancer.

METHODS

The endpoints that we assessed were objective response rate, progression free survival (PFS) and overall survival. Efficacy data for paclitaxel and docetaxel were separately analyzed. Trials were located through PubMed and Cochrane Library searches and abstracts of major international conferences.

RESULTS

Omicronbjective response rate was notably better when paclitaxel was used as every three weeks regimen (7 studies, 1772 patients, fixed effect model pooled RR 1.20 95%CI 1.08-1.32 p<0.001). No difference were found for PFS (6 studies, 1610 patients, random effect model HR 1.02, 95%CI 0.81-1.30 p=0.860); while OS was statistically higher among patients receiving weekly paclitaxel (5 studies, 1471 patients, fixed effect model pooled HR 0.78, 95%CI 0.67-0.89 p=0.001). No differences were observed for the weekly compared to the every three weeks use of docetaxel either for objective response, PFS and OS. Overall, the incidence of serious adverse events, neutropenia, neutropenic fever, and peripheral neuropathy were significantly lower in weekly taxanes schedules. The incidence of nail changes and epiphora were significantly lower in the every three weeks docetaxel regimens.

CONCLUSIONS

Use of paclitaxel in weekly regimen give overall survival advantages compared with the standard every three weeks regimen. The observed survival benefit does not seem to stem from an increased potency of the drug with weekly regimens. The use of weekly paclitaxel regimens is therefore recommended for the treatment of locally advanced/metastatic breast cancer.

Effect of Margin Status on Local Recurrence After Breast Conservation and Radiation Therapy for Ductal Carcinoma In Situ

Purpose

There is no consensus on what constitutes an adequate surgical margin in patients receiving breast-conserving surgery (BCS) and postoperative radiation therapy (RT) for ductal carcinoma in situ (DCIS). Inadequate margins may result in high local recurrence, and excessively large resections may lead to poor cosmetic outcome without oncologic benefit.

Methods

A comprehensive search for published trials that examined outcomes after adjuvant RT after BCS for DCIS was performed using MEDLINE and cross referencing available data. Reviews of each study were conducted, and data were extracted. Primary outcome was ipsilateral breast tumor recurrence (IBTR) related to surgical margins.

Results

A total of 4,660 patients were identified from trials examining BCS and RT for DCIS. Patients with negative margins were significantly less likely to experience recurrence than patients with positive margins after RT (odds ratio [OR] = 0.36; 95% CI, 0.27 to 0.47). A negative margin significantly reduced the risk of IBTR when compared with a close (OR = 0.59; 95% CI, 0.42 to 0.83) or unknown margin (OR = 0.56; 95% CI, 0.36 to 0.87). When specific margin thresholds were examined, a 2-mm margin was superior to a margin less than 2 mm (OR = 0.53; 95% CI, 0.26 to 0.96); however, we saw no significant difference in the rate of IBTR with margins between 2 mm and more than 5 mm (OR = 1.51; 95% CI, 0.51 to 5.0; P > .05).

Conclusion

Surgical margins negative for DCIS should be obtained after BCS for DCIS. A margin threshold of 2 mm seems to be as good as a larger margin when BCS for DCIS is combined with RT.

A comparison of methods for fixed effects meta-analysis of individual patient data with time to event outcomes

BACKGROUND

Alternative methods for individual patient data (IPD) meta-analysis of time-to-event outcomes have been established and utilized in practice. The most common approach is a stratified log-rank analysis. The IPD approach is considered to be the gold standard approach for meta-analysis and is becoming increasingly more popular but the performance of different methods has not been studied previously.

PURPOSE

To compare commonly used methods for fixed effects meta-analysis of individual patient time-to-event data.

METHODS

The stratified log-rank analysis, an inverse variance weighted average of Cox model estimates, and the stratified Cox regression model are compared. First, a theoretical comparison of approaches is undertaken. Second, the bias and coverage are assessed for the pooled hazard ratio using simulated data under commonly encountered meta-analysis conditions. Finally, a comparison is presented using empirical data from four separate systematic reviews of anti-epileptic drug trials where IPD are available for two time-to-event outcomes.

RESULTS

For hazard ratio close to 1 with minimal heterogeneity between trials, theoretical results suggest similar results should be expected from all the three methods. Results for empirical and simulated data are in keeping with the theoretical results and show all the three methods perform well under these conditions. When there is no heterogeneity and the proportional hazards assumption holds, the stratified Cox model and inverse variance weighted average produce similar estimates of the pooled treatment effect and are to be preferred to the stratified log-rank analysis when the underlying treatment effect is large. Coverage values diminish for all the three methods and are below 95% for low or moderate heterogeneity. The low coverage values highlight the need for models that appropriately account for or explore the between trial variation.

LIMITATIONS

Until larger simulations can be undertaken, conclusions based on the simulated and empirical data should only be applied to small meta-analyses of four or five trials.

CONCLUSIONS

These investigations suggest that under normal conditions all three methods provide similar results. For moderate heterogeneity coverage for all the three fixed effects models depreciates.

Practical methodology of meta-analysis of individual patient data using a survival outcome

Meta-analysis of individual patient data (MIPD) is considered as one of the statistical approaches to provide integrated information on the effect of a treatment or an intervention. Statistical analysis of such meta-analyses should account for the clustered structure of data which is induced by all factors varying across the trials. For survival analysis, several models can handle such clustering under proportional hazards. This comprises models with fixed or random trial effects, stratified models and marginal models. In this paper, we review these models and compare their performances using a numerical simulation study. Results show that frailty models, and particularly those with random treatment by trial interactions, are well suited for meta-analyses on individual patient data. This is further exemplified on a meta-analysis of three trials comparing high-dose therapy to conventional chemotherapy in multiple myeloma.

Microvessel density as a prognostic factor in non-small-cell lung carcinoma: a meta-analysis of individual patient data

BACKGROUND

Angiogenesis is a potential prognostic factor that has been investigated in patients with non-small-cell lung carcinoma. However, published studies of the role of angiogenesis as a prognostic factor are inconclusive. We aimed to collect individual patient data to assess microvessel-density counts (ie, a measure of angiogenesis) as a prognostic factor in non-small-cell lung carcinoma.

METHODS

We obtained published and unpublished datasets and extracted appropriate data, taking particular care to ensure data quality. Detailed information was obtained for the laboratory methods used by every research centre that generated the data. The outcome of interest was overall survival. We did a meta-analysis to estimate the prognostic role of microvessel density by combining separately estimated hazard ratios (HR) from every study, which were adjusted for tumour stage and age. Analyses were done separately for studies that used the Chalkley method or for those that counted all microvessels.

FINDINGS

17 centres provided data for 3200 patients, 2719 of which were included in the analysis. All but three centres (datasets 9, 10, and 13-367 cases) had already published their findings, and six had updated follow-up information (datasets 1, 2, 3, 6, 7, and 8-1273 cases). For all but three centres (datasets 4, 11, and 13) some data corrections were necessary. For microvessel density counts obtained by the Chalkley method, the HR for death per extra microvessel was 1.05 (95% CI 1.01-1.09, p=0.03) when analysed as a continuous variable. For microvessel density counts obtained by the all vessels method, the HR for death per ten extra microvessels was 1.03 (0.97-1.09, p=0.3) when analysed as a continuous variable.

INTERPRETATION

Microvessel density does not seem to be a prognostic factor in patients with non-metastatic surgically treated non-small-cell lung carcinoma. This conclusion contradicts the results of a meta-analysis of published data only. Therefore, the methodology used to assess prognostic factors should be assessed carefully.

A Meta-Analysis of Unrelated Donor Umbilical Cord Blood Transplantation versus Unrelated Donor Bone Marrow Transplantation in Adult and Pediatric Patients

Several studies have compared the results of unrelated donor bone marrow transplantation (UBMT) and unrelated donor cord blood transplantation (UCBT). To objectively analyze these data, we performed a systematic review and meta-analysis of pooled data on comparative studies of UCBT and UBMT in patients requiring hematopoietic stem cell transplantation. Combining the studies, 161 children and 316 adults undergoing UCBT (mostly 1 or 2 antigen-mismatched), along with 316 children and 996 adults undergoing UBMT (almost entirely fully matched with the recipient), were analyzed. T-cell-depleted UBMT was excluded; where data were available, only fully matched UBMT was used in the analysis. Pooled comparisons of studies of UCBT and UBMT in children found that the incidence of chronic graft-versus-host disease (GVHD) was lower with UCBT (relative risk [RR] = 0.26; 95% confidence interval [CI] = 0.12-0.57; P = .16), but the incidence of grade III-IV acute GVHD did not differ (RR = 1.46; 95% CI = 0.42-5.03; P = .55). There was no difference in 2-year OS in children when studies were pooled (RR = 0.76; 95% CI = 0.31-1.87; P = .55). For adults, transplantation-related mortality (pooled estimate, 1.04; 95% CI = 0.52-2.08; P = .91) and disease-free survival (DFS) (pooled estimate, 0.59; 95% CI = 0.18-1.96; P = .39) were not statistically different. Because of the unavailability of randomized controlled trials, pooled analysis of nonrandomized comparative studies was performed. Thus, our meta-analysis confirmed that UCBT in children and adults had consistently equivalent survival outcomes compared with UBMT despite greater donor-recipient HLA disparity with UCBT.

Survival with aromatase inhibitors and inactivators versus standard hormonal therapy in advanced breast cancer: meta-analysis

BACKGROUND

Aromatase inhibitors and inactivators have been extensively tested in patients with advanced breast cancer, but it is unclear whether they offer any survival benefits compared with standard hormonal treatment with tamoxifen or progestagens. We performed a meta-analysis of randomized controlled trials that compared several generations of aromatase inhibitors and inactivators with standard hormonal treatment in patients with advanced breast cancer.

METHODS

The endpoint that we assessed was survival. Trials were located through searches of PubMed and Cochrane Library (last update March 2006). Relative hazards (RHs) were summarized across trials through fixed- and random-effects analyses, and heterogeneity was assessed with the Q and I2 statistics. All statistical tests were two-sided.

RESULTS

Twenty-five different comparisons, with a total of 8504 patients, were included in the meta-analysis. We found statistically significant survival benefits with third-generation aromatase inhibitors and inactivators (vorozole, letrozole, examestane, and anastrazole) (RH = 0.87, 95% confidence interval [CI] = 0.82 to 0.93; P<.001) but not with first-generation (aminoglutethimide) or second-generation (formestane and fadrozole) agents. The difference in the summary effects between these two groups of trials was statistically significant (P = .04). The survival benefit with third-generation agents in first-line trials, in which these agents were compared with tamoxifen (11% RH reduction, 95% CI = 1% to 19%; P = .03), was identical to their benefit in second- and subsequent-line trials in which these agents were compared with other treatments (14% RH reduction, 95% CI = 6% to 21%; P<.001).

CONCLUSIONS

Inhibition of the aromatase system, in particular with third-generation aromatase inhibitors and inactivators, appears to be associated with statistically significant improved survival of patients with advanced breast cancer compared with standard hormonal treatments.

Median and ulnar nerve injuries: a meta-analysis of predictors of motor and sensory recovery after modern microsurgical nerve repair

BACKGROUND

The aim of this study was to quantify variables that influence outcome after median and ulnar nerve transection injuries. The authors present a meta-analysis based on individual patient data on motor and sensory recovery after microsurgical nerve repair.

METHODS

From 130 studies found after literature review, 23 articles were ultimately included, giving individual data for 623 median or ulnar nerve injuries. The variables age, sex, nerve, site of injury, type of repair, use of grafts, delay between injury and repair, follow-up period, and outcome were extracted. Satisfactory motor recovery was defined as British Medical Research Council motor scale grade 4 and 5, and satisfactory sensory recovery was defined as British Medical Research Council grade 3+ and 4. For motor and sensory recovery, complete data were available for 281 and 380 nerve injuries, respectively.

RESULTS

Motor and sensory recovery were significantly associated (Spearman r = 0.62, p < 0.001). Multivariate logistic regression analysis showed that age (< 16 years versus > 40 years: odds ratio, 4.3; 95 percent confidence interval, 1.6 to 11.2), site (proximal versus distal: odds ratio, 0.46; 95 percent confidence interval, 0.20 to 1.10), and delay (per month: odds ratio, 0.94; 95 percent confidence interval, 0.90 to 0.98) were significant predictors of successful motor recovery. In ulnar nerve injuries, the chance of motor recovery was 71 percent lower than in median nerve injuries (odds ratio, 0.29; 95 percent confidence interval, 0.15 to 0.55). For sensory recovery, age (odds ratio, 27.0; 95 percent confidence interval, 9.4 to 77.6) and delay (per month: odds ratio, 0.92; 95 percent confidence interval, 0.87 to 0.98) were found to be significant predictors.

CONCLUSIONS

In this individual patient data meta-analysis, age, site, injured nerve, and delay significantly influenced prognosis after microsurgical repair of median and ulnar nerve injuries.

Synthesizing medical evidence: systematic reviews and metaanalyses

Systematic reviews and metaanalyses have become increasingly popular ways of summarizing, and sometimes extending, existing medical knowledge. In this review the authors summarize current methods of performing meta-analyses, including the following: formulating a research question; performing a structured literature search and a search for trials not published in the formal medical literature; summarizing and, where appropriate, combining results from several trials; and reporting and presenting results. Topics such as cumulative and Bayesian metaanalysis and metaregression are also addressed. References to textbooks, articles, and Internet resources are also provided. The goal is to assist readers who wish to perform their own metaanalysis or to interpret critically a published example.

Neoadjuvant versus adjuvant systemic treatment in breast cancer: a meta-analysis

BACKGROUND

Interest in the use of preoperative systemic treatment in the management of breast cancer has increased because such neoadjuvant therapy appears to reduce the extent of local surgery required. We compared the clinical end points of patients with breast cancer treated preoperatively with systemic therapy (neoadjuvant therapy) and of those treated postoperatively with the same regimen (adjuvant therapy) in a meta-analysis of randomized trials.

METHODS

We evaluated nine randomized studies, including a total of 3946 patients with breast cancer, that compared neoadjuvant therapy with adjuvant therapy regardless of what additional surgery and/or radiation treatment was used. Fixed and random effects methods were used to combine data. Primary outcomes were death, disease progression, distant disease recurrence, and loco-regional disease recurrence. Secondary outcomes were local response and conservative local treatment. All statistical tests were two-sided.

RESULTS

We found no statistically or clinically significant difference between neoadjuvant therapy and adjuvant therapy arms associated with death (summary risk ratio [RR] = 1.00, 95% confidence interval [CI] = 0.90 to 1.12), disease progression (summary RR = 0.99, 95% CI = 0.91 to 1.07), or distant disease recurrence (summary RR = 0.94, 95% CI = 0.83 to 1.06). However, neoadjuvant therapy was statistically significantly associated with an increased risk of loco-regional disease recurrences (RR = 1.22, 95% CI = 1.04 to 1.43), compared with adjuvant therapy, especially in trials where more patients in the neoadjuvant, than the adjuvant, arm received radiation therapy without surgery (RR = 1.53, 95% CI = 1.11 to 2.10). Across trials, we observed heterogeneity in the rates of complete clinical response (range = 7%-65%; P for heterogeneity of <.001), pathologic response (range = 4%-29%; P for heterogeneity of <.001), and adoption of conservative local treatment (range = 28%-89% in neoadjuvant arms, P for heterogeneity of <.001).

CONCLUSIONS

Neoadjuvant therapy was apparently equivalent to adjuvant therapy in terms of survival and overall disease progression. Neoadjuvant therapy, compared with adjuvant therapy, was associated with a statistically significant increased risk of loco-regional recurrence when radiotherapy without surgery was adopted.

Randomised trials comparing chemotherapy regimens for advanced non-small cell lung cancer: biases and evolution over time

We systematically evaluated the evidence from randomised trials comparing various chemotherapy regimens for advanced non-small cell lung cancer. Across 254 eligible trials (42661 patients), no regimens were compared in >6 studies. Twenty-six trials (10%) found statistically significant differences in survival between the compared arms. Only five reported the randomisation mode, and four reported adequate allocation concealment; nine performed unaccounted interim analyses. Statistical significance was more common in larger (P=0.003), more recent studies (P=0.031), and trials from countries with only one published eligible study (P=0.008). Increased reported median survival was independently associated with platinum and/or taxane and combination regimens, but also with the year of publication, smaller sample size, and larger representation of non-stage IV patients and patients with a better performance status. The proportion of enrolled patients with a performance status of 2 or worse decreased significantly over time (12.9% per decade, P<0.001). Randomised evidence in this field is fragmented and subject to considerable selection biases.

Evidence-based mucogingival therapy. Part 2: Ordinary and individual patient data meta-analyses of surgical treatment of recession using complete root coverage as the outcome variable

BACKGROUND

The literature (1970-2000) on the outcome of surgical root coverage has been revised and summarized in a companion paper. The overall conclusion was that the various procedures are effective, but it was not possible to determine which procedure was best indicated in different clinical conditions. In this study, meta-analysis techniques were used to seek evidence for guiding clinical decisions when planning root coverage surgery. The aim of this study was to illustrate the differences between meta-analyses applied to summarized and individual patient data (IPD) and to present suggestions for reducing the costs of IPD meta-analysis.

METHODS

Only clinical trials and case series that included data on the number of teeth treated, baseline recession depth (BRD) and the proportions of postoperative complete root coverage (CRC) were considered. The first group of meta-analyses (the outcome of each procedure based on summarized data) covered 65 studies dealing with coronally advanced flap (CAF), epithelial free gingival graft (EFGG), connective tissue graft (CTG), and guided tissue regeneration (GTR) procedures. The second group of meta-analyses was done to determine the outcome of each procedure on the basis of 26 studies that reported IPD for at least baseline recession depth (BRD) and final CRC for each site. The third group of meta-analysis compared the outcomes of CTG and GTR in 5 randomized studies, 4 of which reported only summarized data.

RESULTS

The first analysis showed that CRC was achieved more often in non-randomized than in randomized studies. The heterogeneity tests revealed great variability of results in both the randomized and non-randomized studies, which makes it difficult to draw any definite conclusions. In the second analysis all the tested techniques revealed similar trends: greater baseline recession depths were always associated with a decreased CRC. The third analysis showed that CRC was achieved more frequently in the sites treated with CTG as opposed to GTR. The small sample size and the lack of IPD rendered the analyses inconclusive despite the randomized design.

CONCLUSIONS

Few studies reported individual patient data; they are a valuable contribution to clinical decision making, but IPD published in the literature are still insufficient to provide a reliable guide for clinical decision making. Therefore, decisive steps should be taken to facilitate the publication of IPD, in electronic format, whenever a clinical study is published in a leading journal.

Bone marrow transplantation for paediatric AML in first remission: a systematic review and meta-analysis

For children with AML in CR1, the major consolidation therapies are BMT, ABMT and intensive chemotherapy. The relative effectiveness of these strategies is still debated. We conducted a systematic review and meta-analysis of trials to determine the effectiveness of BMT and ABMT in CR1 in paediatric AML. Eligible studies enrolled patients <21 years from 1985 to 2000 with AML in CR1. Two groups of studies were identified: (1) Those comparing the outcome of patients with and without a histocompatible family donor; and (2) Randomised controlled trials (RCT) comparing ABMT with non-myeloablative chemotherapy. The relative risk statistic was calculated for outcomes of interest in each trial. If there was no excessive heterogeneity between trials the results were pooled, and an overall relative risk and risk difference for treatment effect across trials were calculated. Results of the analysis showed that allocation to BMT reduced risk of relapse and improved disease-free and overall survival. For ABMT, heterogeneity of effect between RCTs prevented pooling of results. In conclusion, BMT from a histocompatible family donor improves patient outcome. Data are insufficient to determine whether this is true for all subgroups of AML, and whether ABMT is superior to non-myeloablative chemotherapy. An individual patient data meta-analysis is required to further evaluate the available data.

To IPD or not to IPD? Advantages and disadvantages of systematic reviews using individual patient data

Systematic reviews and meta-analyses that obtain original research data on individual participants enrolled in trials have been described as the gold standard of review. However, they may take longer and be more resource intensive than other types of review. The authors describe potential advantages and disadvantages of the individual patient data (IPD) approach, including benefits from improved data quality, benefits afforded by the type of analyses that can be done, and advantages in achieving consensus around results and interpretation by an international multidisciplinary team. Disadvantages and barriers relating to resource and expertise, negotiating collaboration, and software requirements are also discussed. At the outset, reviewers should consider the methodological factors likely to influence results in their particular review setting, together with time and resource constraints, so that an active decision can be made about whether to extract data from published reports, collect additional or replacement summary data from trialists, or collect IPD.

Meta-analyses of randomised clinical trials in oncology

A meta-analysis is a quantitative synthesis of randomised clinical trials, used to evaluate moderate treatment effects in oncology. It is complementary to large-scale trials. We describes the principles, methods, and limits of meta-analyses. The gold standard for a meta-analysis is to obtain individual patient data directly from each principal investigator, but this is time-consuming and costly. The main steps of a meta-analysis using individual patient data are described. Multidisciplinary collaboration is needed for clinical insight and critical review of the data and results. Meta-analysis should include an evaluation of the trial quality, a quantification of the overall treatment effect, a study of the variations seen in this effect between trials, and pre-planned exploratory analyses to identify groups of patients who may benefit more from the treatment. Statistical methods are explained using real working examples. Since literature-based meta-analysis can lead to seriously biased assessments, meta-analyses of individual patient data should be undertaken systematically when long-term follow-up is needed, when a detailed analysis is important, or when the literature-based meta-analyses are not in agreement. The main factors which influence the quality of a meta-analysis are discussed.

Methods for exploring heterogeneity in meta-analysis

In meta-analysis, when the difference in results between studies is greater than would be expected by chance, one needs to investigate whether the observed variation in results across studies is associated with clinical and/or methodological differences between studies. This article reviews methods used in meta-analysis for exploring heterogeneity, including statistical tests for homogeneity, methods for visually displaying results of primary studies, methods for reducing heterogeneity, methods for investigating sources of heterogeneity, and identification of moderator variables or effect modifiers. The investigation of sources of heterogeneity in meta-analysis is by nature exploratory, and therefore its results should always be interpreted with caution. However, careful investigation of heterogeneity may provide an important second level of evidence that can be useful in suggesting direction of future research. Sometimes, it may provide clinically important results by indicating who might benefit more or less from a treatment or how an intervention should be applied.

Estimating number of events from the Kaplan-Meier curve for incorporation in a literature-based meta-analysis: what you don't see you can't get!

In literature-based meta-analyses of time-to-event data, the number of events in the treated and control groups together with the total number of patients randomized to the two treatment arms are often used as summary statistics. If interest is in mortality at a specified moment in time, the number of events can, in most cases, only be obtained from the Kaplan-Meier curve. The estimated number of events, however, is typically larger than the true number of events. The effect of this overestimation on the Mantel-Haenszel test and the odds ratio is studied in this paper. From these results, it can be concluded that the number of events should not be estimated from the Kaplan-Meier curves for meta-analytic purposes unless virtually no patients are lost to follow-up or censored and there are still many patients at risk in the two groups at the time at which the number of events is to be determined.

The role of meta-analyses in assessing cancer treatments

The results of individual phase III cancer clinical trials are often inconclusive due to the overly optimistic size of treatment differences that are sought. Increased power and precision can generally be obtained if the data from several different trials studying the same or similar questions are analysed together. Individual patient data meta-analyses, which combine together the quantitative results from all properly randomised studies, provide an overall estimate of the size of treatment differences. Individual patient data meta-analyses have played an especially important role in breast and gastrointestinal tract cancers where many important questions have been addressed. Although meta-analyses have been subject to considerable criticism, individual patient data meta-analyses provide the best overall evidence of treatment effect in the absence of large-scale trials and have been instrumental in providing objective data that can be used in the design of new studies.

Individual patient data meta-analysis of randomized anti-epileptic drug monotherapy trials

Meta-analysis may be based on either aggregate data or individual patient data (IPD). Three reasons why IPD are desirable for the meta-analysis of anti-epileptic drug (AED) monotherapy trials are: (1) to undertake a more complete analysis of time-to-event outcomes; (2) to investigate the interaction between AED and type of epilepsy; and (3) to undertake re-analysis of the trial to obtain results for all relevant outcomes. We demonstrate that IPD meta-analysis is possible in AED research. Problems arose from missing data at four levels: (1) unknown trials; (2) known trials but no IPD supplied; (3) known trials but missing outcome data for some individuals within trials; and (4) known trials but missing covariate data for some individuals within trials. Empirical evidence of the reliability of meta-analyses based on aggregate rather than individual patient data is still lacking. Examples of other benefits such projects may bring include improvements to the design of a new trial in the area, in terms of the sample size considerations, the definition of outcomes and data collection.

Meta-analyses in head and neck squamous cell carcinoma. What is the role of chemotherapy?

Despite more than 70 randomized clinical trials comparing locoregional treatment plus chemotherapy with the same locoregional treatment alone, the effect of chemotherapy on the survival of patients with head and neck squamous cell cancer is uncertain. Indeed, most of the randomized trials have been inconclusive in this regard, and only a few have reported significant results either in favor of or against chemotherapy. This article focuses on four meta-analyses that included a comparison between chemotherapy (neoadjuvant, adjuvant, or concomitant) plus locoregional treatment versus the same locoregional treatment alone and in which survival was the main end-point.