The number of depression patients has exceeded 300 million worldwide, and its high disability rate has attracted global attention. However, the relationship between temperature-humidity index (Humidex) and depression remains to be explored.

To explore the relationship between Humidex of the twenty-four solar terms and depression.

Outpatient data of depression patients (n=18 289) admitted to the Psychology Department, Mental Health Center of Jiading District from 2016-10-08 to 2019-10-07 were collected, including gender, age, time of treatment, number of outpatient visits and specialist visits. The data of daily mean temperature, relative humidity, rainfall, air pressure and mean wind speed in Jiading District from 2016-10-08 to 2019-10-07 published by Shanghai Meteorological Bureau were collected to calculate Humidex. The included patients were divided into the groups of ≤18 years (n=157) , 19-44 years (n=3 099) , 45-59 years (n=4 848) , 60-74 years (n=6 270) and≥75 years (n=3 915) according to different ages. The relationship between Humidex of twenty-four solar terms and depression was estimated by using the combination of nonhomogeneous Poisson distribution and distributed-lag nonlinear model. Correlation between Humidex and meteorological variables was calculated using Pearson correlation analysis. Relative risk rates (RR) of Humidex and depression were calculated of twenty-four solar terms using median Humidex as control.

A total of 18 289 patients with depression were included from 2016-10-08 to 2019-10-07, including 6 900 males and 11 389 females. The Lesser Cold solar term corresponded to the lowest average Humidex and highest number of depression patients, the Great Heat solar term corresponded corresponded to the highest average Humidex and lowest number of depression patients from 2016-10-08 to 2017-10-07. The Great Cold solar term corresponded to the lowest average Humidex and highest number of depression patients, the Great Heat solar term corresponded to the highest average Humidex and lowest number of depression patients from 2017-10-08 to 2018-10-07. The Great Cold solar term corresponded to the lowest average Humidex and highest number of depression patients, the Great Heat corresponded to the highest average Humidex and lowest number of depression patients from 2018-10-08 to 2019-10-07. Humidex in 24 solar terms showed a nonlinear relationship with the overall population risk of depression. The top two solar terms corresponding to Humidex of depression risk distribution for total population, male, female, ≤18 years, 19-44 years, 45-59 years, 60-74 years, and ≥75 years were the Great Cold and Beginning of Spring solar terms, Great Heat solar term corresponded to the lowest depression risk distribution. The maximum lag effect was observed on the 10th day of the Great Cold and Beginning of Spring solar terms (RR=1.020, 95%CI=1.001-1.040) , and the lag effect was observed on the 8th day of Great Heat solar term and lasted to the 9th day. The lag effect was observed on the 7th day after the Great Cold and Beginning of Spring and the maximum lag effect was observed on the 9th day (RR=1.054, 95%CI=1.007-1.104) , which lasted until the 11th day in the 60-74 years group. The lag effect was observed on the 1st day and 8th day after the Great Heat, and lasted to the 2nd day and 10th day, respectively, with the maximum lag effect on the 10th day (RR=0.952, 95%CI=0.911-0.994) in male depression patients. The lag effect was observed on the 8th day after the Great Heat solar term and lasted to the 9th day in the 60-74 years group.

Humidex in the Great Cold and Beginning of Spring solar terms are risk factors for depression. Clinically, accurate diagnosis and treatment should be provided for different depression patients according to Humidex in different solar terms, and individualized intervention programs should be formulated.