Revisiting the association between pretreatment thrombocytosis and cancer survival outcomes: an umbrella review of meta-analyses.

BACKGROUND

Although associations have been reported linking pretreatment thrombocytosis to cancer survival outcomes, the validity and strength of existing observational evidence have been contested. This study aimed to conduct an umbrella review to comprehensively appraise the strength, validity and credibility of these reported associations.

METHODS

We searched Medline, Embase and Cochrane Database of Systematic Reviews from inception to 8 April 2023 to retrieve meta-analyses of observational studies. Meta-analyses were re-performed using a random-effect model and the strength of evidence was graded as convincing, highly suggestive, suggestive and weak according to seven pre-defined quantitative criteria reflecting statistical significance, amount of data, heterogeneity, and evidence of bias. The quality of review was appraised using the AMSTAR2 checklist. The umbrella review was reported adhering to the PRISMA guideline and was registered on PROSPERO (CRD42023455391).

RESULTS

A total of 21 unique meta-analyses investigating ten cancer subtypes were included. All meta-analyses reported inferior survival outcome in cancer patients with pretreatment thrombocytosis, and 18 of them (85.7%) yielded statistically significant results (P < 0.05). Consistent effects were observed across meta-analyses that adopted different cut-off values (i.e. platelet count > 300 or 400 × 10 /L) to define thrombocytosis. Although evidence appraisal did not identify convincing evidence (Class I), the associations of thrombocytosis with inferior overall survival of lung, gastric, colorectal cancer and malignant mesothelioma were classified as highly suggestive evidence (Class II). According to AMSTAR2 ratings, no meta-analysis was identified with high or moderate quality.

CONCLUSIONS

Our findings consolidated the association between pretreatment thrombocytosis and poor survival outcomes in various cancers. Nonetheless, the absence of convincing associations indicates a need for further large-scale, high-quality evidence to confirm whether platelets can serve as a prognostic predictor or a therapeutic target.