Background: Radiotherapy has become much better targeted since the 1980s, improving both safety and efficacy. In breast cancer, radiotherapy to regional lymph nodes aims to reduce risks of recurrence and death. Its effects have been studied in randomised trials, some before the 1980s and some after. We aimed to assess the effects of regional node radiotherapy in these two eras. Methods: In this meta-analysis of individual patient data, we sought data from all randomised trials of regional lymph node radiotherapy versus no regional lymph node radiotherapy in women with early breast cancer (including one study that irradiated lymph nodes only if the cancer was right-sided). Trials were identified through the EBCTCG's regular systematic searches of databases including MEDLINE, Embase, the Cochrane Library, and meeting abstracts. Trials were eligible if they began before Jan 1, 2009. The only systematic difference between treatment groups was in regional node radiotherapy (to the internal mammary chain, supraclavicular fossa, or axilla, or any combinations of these). Primary outcomes were recurrence at any site, breast cancer mortality, non-breast-cancer mortality, and all-cause mortality. Data were supplied by trialists and standardised into a format suitable for analysis. A summary of the formatted data was returned to trialists for verification. Log-rank analyses yielded first-event rate ratios (RRs) and confidence intervals. Findings: We found 17 eligible trials, 16 of which had available data (for 14 324 participants), and one of which (henceforth excluded), had unavailable data (for 165 participants). In the eight newer trials (12 167 patients), which started during 1989-2008, regional node radiotherapy significantly reduced recurrence (rate ratio 0·88, 95% CI 0·81-0·95; p=0·0008). The main effect was on distant recurrence as few regional node recurrences were reported. Radiotherapy significantly reduced breast cancer mortality (RR 0·87, 95% CI 0·80-0·94; p=0·0010), with no significant effect on non-breast-cancer mortality (0·97, 0·84-1·11; p=0·63), leading to significantly reduced all-cause mortality (0·90, 0·84-0·96; p=0·0022). In an illustrative calculation, estimated absolute reductions in 15-year breast cancer mortality were 1·6% for women with no positive axillary nodes, 2·7% for those with one to three positive axillary nodes, and 4·5% for those with four or more positive axillary nodes. In the eight older trials (2157 patients), which started during 1961-78, regional node radiotherapy had little effect on breast cancer mortality (RR 1·04, 95% CI 0·91-1·20; p=0·55), but significantly increased non-breast-cancer mortality (1·42, 1·18-1·71; p=0·00023), with risk mainly after year 20, and all-cause mortality (1·17, 1·04-1·31; p=0·0067). Interpretation: Regional node radiotherapy significantly reduced breast cancer mortality and all-cause mortality in trials done after the 1980s, but not in older trials. These contrasting findings could reflect radiotherapy improvements since the 1980s. Funding: Cancer Research UK, Medical Research Council.