Treat 2010.

Methods	Randomized multi‐center phase III trial Inclusion criteria Cytologically or histologically confirmed stage IIIB (pericardial or pleural effusion) or stage IV or recurrent non‐small cell lung cancer (NSCLC) ≥ 18 years of age Measurable or evaluable disease (according to ECOG solid tumor criteria) PS: 0 to 1 Adequate bone marrow, hepatic, and renal function Brain metastasis allowed as long as lesion is considered controlled by investigator after surgery or radiotherapy Exclusion criteria Prior chemotherapy for NSCLC Pregnant or breast‐feeding Hypersensitivity reaction to polyoxyethyl castor oil
Participants	PG arm: 377 participants (ITT population)/112 elderly participants CG arm: 379 participants (ITT population)/119 elderly participants CP arm: 379 participants (ITT population)/107 elderly participants
Interventions	PG arm: paclitaxel 200 mg/m2 over 3‐hour i.v. infusion on day 1 plus gemcitabine 1000 mg/m2 over 30‐minute i.v. infusion on days 1 and 8, every 3 weeks CG arm: carboplatin AUC5.5 over 15 to 30‐minute i.v. infusion on day 1 plus gemcitabine 1000 mg/m2 over 30‐minute i.v. infusion on days 1 and 8, every 3 weeks ‐ participants with ≥ 20% of bone marrow previously irradiated received a reduced dose of carboplatin AUC5 CP arm: carboplatin AUC6 over 15 to 30‐minute i.v. infusion on day 1 plus paclitaxel 225 mg/m2 over 3‐hour i.v. infusion on day 1, every 3 weeks In the PG and CP arms, participants received prophylactic treatment with dexamethasone 20 mg orally 12 and 6 hours before paclitaxel infusion, diphenhydramine 50 mg i.v. ≤ 1 hour before paclitaxel infusion, and cimetidine 300 mg i.v. (or equivalent) ≤ 1 hour before paclitaxel Granulocyte colony‐stimulating factor (G‐CSF) allowed only for secondary prophylaxis in case of persistent neutropenia despite dose modifications
Outcomes	Primary outcome Overall survival Secondary outcomes Response rate (according to ECOG criteria) Time‐to‐progression Adverse events (according to NCI‐CTAE v2.0) Quality of life outcome (using Functional Assessment of Cancer Therapy‐Lung tool)
Notes	Trial designed for the following pair‐wise comparison GC arm vs GP arm CG arm vs CP arm CP arm vs GP arm
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients with stage IIIB (with pleural or pericardial effusion), stage IV, or recurrent NSCLC who met all eligibility criteria were randomly allocated to receive one of the three treatment regimens as summarized in Figure 1. Patient stratification by baseline weight loss (< 5% versus ≥ 5% in previous 6 months), stage of disease (IIIB with effusion versus IV), and brain metastasis (presence versus absence) took place at the time of randomisation to ensure balance across treatment arms with respect to these characteristics"
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment
Blinding of outcome assessment (detection bias) OS and 1y OS rate outcome	Unclear risk	Open‐label study considered to have unclear impact on mortality outcomes
Blinding of outcome assessment (detection bias) Other outcomes	High risk	Open‐label study
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Study authors presented ITT analysis for survival and TTP outcomes. 23, 22, and 13 participants did not start treatment in CG, GP, and CP arms, respectively. They were not included in the ITT analysis. No information on elderly outcomes
Selective reporting (reporting bias)	Low risk	No evidence of selective reporting bias
Other bias	Unclear risk	Outcomes for elderly population not fully reported. Study designed for the general population; elderly subgroup analysis not planned