Recent epidemiological studies have explored effects of light at night (LAN) exposure on breast cancer, but reported inconsistent findings. We performed a systematic review and meta-analysis of available evidence regarding the association of LAN assessed by satellite data with breast cancer. We conducted a systematic PubMed, Web of Science, and EMBASE database literature search until August 2020. Random-effects meta-analysis was applied to synthesis risk estimates. Heterogeneity was measured using statistics of Cochran’s Q, I², and Tau² (τ²). We assessed publication bias through funnel plot and Egger’s test. Moreover, subgroup analyses according to study design and menopausal status were performed. Risk of bias (RoB) of each included study was assessed using a domain-based RoB assessment tool. The confidence in the body of evidence was appraised using the GRADE approach for level-of-evidence translation. A total of 1157 studies were identified referring to LAN and breast cancer, from which 6 were included for quantitative synthesis. We found a significantly higher odds of breast cancer in the highest versus the lowest category of LAN exposure (OR = 1.11, 95% CI: 1.06, 1.16; I² = 0.0%). In the subgroup analyses stratified by menopausal status and study design, significant association was found in postmenopausal women (OR = 1.07, 95% CI = 1.00, 1.13) and cohort studies (OR = 1.11, 95% CI = 1.05, 1.18), while the summary estimates of premenopausal women and case-control studies showed no significance. The level of evidence for the association of LAN exposure and breast cancer risk was graded as “moderate” with “probably low” RoB according to the NTP/OHAT framework. In conclusion, this study suggests a link of LAN exposure with risk of breast cancer. Further high-quality prospective studies, especially performed in low-to middle-income countries with improvement in the area of LAN exposure assessment are needed to advance this field.

最近的流行病学研究探索了夜间光照 (LAN) 暴露对乳腺癌的影响，但报告的结果不一致。我们对通过卫星数据评估的 LAN 与乳腺癌之间的关联的现有证据进行了系统回顾和荟萃分析。我们进行了系统的 PubMed、Web of Science 和 EMBASE 数据库文献搜索，直至 2020 年 8 月。随机效应荟萃分析用于综合风险估计。使用 Cochran's Q、I ²和 Tau ² (τ ²）。我们通过漏斗图和 Egger 检验评估发表偏倚。此外，还根据研究设计和绝经状态进行了亚组分析。使用基于领域的 RoB 评估工具评估每项纳入研究的偏倚风险 (RoB)。使用 GRADE 方法评估证据体的置信度以进行证据水平转换。共确定了 1157 项涉及 LAN 和乳腺癌的研究，其中 6 项被纳入定量合成。我们发现 LAN 暴露最高与最低类别相比，乳腺癌的发病率显着更高（OR = 1.11, 95% CI: 1.06, 1.16; I ²= 0.0%）。在按绝经状态和研究设计分层的亚组分析中，绝经后妇女 (OR = 1.07, 95% CI = 1.00, 1.13) 和队列研究 (OR = 1.11, 95% CI = 1.05, 1.18) 发现了显着相关性，而绝经前妇女的总结估计和病例对照研究显示没有意义。根据 NTP/OHAT 框架，LAN 暴露与乳腺癌风险相关的证据级别被评为“中等”和“可能低”的 RoB。总之，这项研究表明 LAN 暴露与乳腺癌风险之间存在联系。需要进一步进行高质量的前瞻性研究，特别是在 LAN 暴露评估领域有所改进的中低收入国家进行的研究，以推动这一领域的发展。