The UK Biobank is a large prospective information resource that recruited more than half a million people aged 40–69 years (2006 to 2010) and collected their genetic information, blood samples, lifestyle, and environmental exposure data. The relevant information has been described in detail elsewhere [14, 15]. The current analysis was approved by UK Biobank with an ID of 66,536. All patients signed informed consent at the time of the first enrollment. This analysis received approval from the National Information Governance Board for Health and Social Care and the National Health Service North West Multicentre Research Ethics Committee (reference 13/NW/0382). At baseline, we included individuals with complete data on the aforementioned four working conditions (n = 73,123). Subsequently, we excluded participants with CKD (n = 813) and those with missing values in the clinical data (n = 7,241), resulting in a final analysis cohort of 65,069 participants (Fig. 1).

Flow chart of the study population. GWAS: genome-wide association

The primary outcome in the present study was incident CKD. CKD was defined according to the International Classification of Diseases-10th Edition, including diabetes mellitus with renal complications (E10.2, E11.2, E12.2, E13.2, and E14.2), hypertensive renal disease (I12 and I13), glomerular disease (N03, N04, N05, and N07), renal tubulointerstitial disease (N11, N12, N13, N14, and N15), and renal failure (N18 and N19). In the UK Biobank study, the recorded outcomes and endpoints for each participant were primarily determined by healthcare providers, who utilized a combination of hospital admissions, self-reported information, and death registration data to diagnose CKD, following the definitions provided by ICD-10 codes.

All participants collected information on age, sex, weight, smoking status, alcohol consumption, activity_MET, medication use, and physician diagnosis of chronic diseases via a touchscreen device at baseline. Biospecimens were collected and assayed in the central laboratory. UK Biobank collected the genetic information, we selected a total of 241 independent single nucleotide variations which were identified from the most recent genome-wide association study (GWAS) and were significantly associated with CKD, details information on the selected SNP is provided in Table S1 [16]. The genetic risk score (GRS) for CKD was calculated by a method that has been described elsewhere: GRS = (β1 X SNP1 + β2 X SNP2 +… + β241 X SNP241), each SNP was coded as 0, 1, and 2 according to the number of risk alleles. The β coefficient was obtained from the reported GWAS meta-analysis [17].

Participants were asked the following questions about their jobs: occupational heat exposure was assessed by the question, “Thinking about the place where you worked: Was it very hot?” with multiple choices provided: (1) never/rarely; (2) sometimes; and (3) often. Occupational secondhand cigarette smoke exposure was obtained by the question, “Thinking about the place where you worked: Was there a lot of cigarette smoke from other people smoking?” with multiple choices provided: (1) never/rarely; (2) sometimes; and (3) often. The workload was asked by the question, “Does your work involve heavy manual or physical work?” (Physical work includes work that involves handling heavy objects and use of heavy tools.) with the following choices:1) never/rarely; 2) sometimes; 3) usually; and (4) always. Shift work was assessed by the question “Does your work involve shift work?” (shift work is a work schedule that falls outside of the normal daytime working hours of 9 am-5 pm. This may involve working afternoons, evenings, or nights or rotate through these kinds of shifts) with the following choices:1) never/rarely; 2) sometimes; 3) usually; and 4) always. The working hour was reported as the hours of work time every week (Do not include hours of commute).

Based on the participants’ responses, we obtained their information on the above working conditions. Participants who answered “sometimes” or “often” exposure to occupational heat; reported “sometimes” or “often” exposure to occupational secondhand cigarette smoke; involved in shift work (“usually” or “always”); “usually” or “always” had heavy workloads, were grouped as high-risk working conditions. Each work condition was coded 1 if grouped to high-risk and 0 if not. The working conditions risk score was obtained by summing up the above four work conditions, and higher scores indicate higher levels of risk in working conditions. We also categorized it as healthy working condition (working conditions risk score 0), intermediate working condition (working conditions risk score 1–2), and poor working condition (working conditions risk score 3–4).

Demographic and lifestyle behaviors were collected using a touchscreen device. The townsend deprivation index, which is based on the participant’s postcode and reflects the degree of deprivation, was obtained. Physical activity was evaluated by calculating the total metabolic equivalent task (MET) minutes per week. This calculation encompassed minutes spent on all leisure-time activities, including mild activities such as walking, as well as moderate and vigorous activities. Subsequently, the total MET was categorized into tertiles: low, medium, and high. Household income was classified into five levels based on the family’s pre-tax annual income. Education was classified as having a college education or not. Diet was divided into healthy and unhealthy according to a healthy diet score which has been used in several studies successfully, and the healthy diet score was based on the American Heart Association (AHA) guidelines [18,19,20]. Blood samples were collected and analyzed at the central laboratory.

All of the analyses were performed by R v4.1.2 (http://www.R-project.org, The R 121 Foundation). P < 0.05 represents statistical significance. Follow-up time was calculated from the baseline date to diagnosis of CKD, death, or last follow-up date, whichever occurred first. The Cox proportional hazards model was used to calculate the HR and 95% CI between risky working conditions,working conditions risk score, and CKD. Multivariate models were established for controlling potential confounders, including age, sex, ethnic, working time, activity_MET, the townsend deprivation index, smoking status, alcohol consumption, triglyceride, body mass index, hypertension, diabetes mellitus, estimated glomerular filtration rate (eGFR).

To examine the reliability of the results, we additionally adjusted for education, healthy diet, and genetic risk of CKD (GRS_CKD). We further excluded people who had a CKD event within 2 years of follow-up to avoid reverse causality. Calculated the population attributable risk proportion (PAR%) to estimate the proportion of participants who would not theoretically develop CKD if all participants were not exposed to risky working conditions. We also constructed a weighted working conditions risk score based on the 4 work factors by using the equation: weighted work score= (β1×factor1 + β2 ×factor 2 +…+β4×factor 4) × (4/sum of the β coefficients). This weighted score also ranges from 0 to 4 points but considers magnitudes of the adjusted relative risk for each factor in each work pattern as a combination of 4 factors.