BACKGROUND: Pulmonary embolism (PE) is an important cause of death and its seasonality has long been observed. Very few epidemiological studies have explored the potential role of ambient temperature in PE symptom onset, especially at the hourly level. METHODS: We conducted a time-stratified case-crossover study among 17,903 PE patients with hourly onset of symptom from 1590 hospitals across China between January 2015 and September 2020. Conditional logistic regression model combined with distributed lag non-linear models were used to explore the associations between hourly ambient temperature and PE symptom onset. The attributable fractions due to non-optimum temperature were calculated. RESULTS: The exposure-response relationship curve was inverse and almost linear. Lower temperature was significantly associated with higher risk of PE symptom onset when temperature was below 18 °C. This risk occurred immediately at the same hour, attenuated thereafter, and became nonsignificant at approximately 72 h after exposure. Compared with the referent temperature (P(99), 34.1 °C), the odds ratio of PE symptom onset associated with extremely low temperature (P(1), -16.1 °C) over lag 0-72 h was 1.63 (95%CI: 1.23, 2.16). Low temperature may account for 16.19 % of the symptom onset nationally with higher proportion in the south of China. The effects were stronger in older adults, males, and cold seasons. CONCLUSIONS: We provided the first-hand robust evidence that transient exposure (at the hourly level) to low temperature might trigger the symptom onset of PE and constitute a considerable burden for PE patients. Targeted protections and health education are needed for susceptible populations.