Abstract
Background
Sepsis remains a major cause of morbidity and mortality, yet current strategies for identifying susceptible individuals provide limited discriminatory performance. Although serum uric acid (SUA) and gout have been implicated in inflammatory and immune dysregulation, the associations of clinical urate phenotypes with incident sepsis risk and prognosis remain incompletely understood.
Methods
A prospective cohort analysis was conducted using UK Biobank data including 466,611 participants, in which clinical urate phenotypes (normal uric acid, asymptomatic HUA, and gout) were evaluated as primary exposures, with SUA quartiles and separate HUA and gout status analyses used as alternative exposure definitions, using multivariable Cox proportional hazards models, restricted cubic spline analyses, subgroup analyses, and sensitivity analyses including Fine-Gray competing-risk models.
Results
During a median follow-up of 14.53 years, 16,210 incident sepsis cases were identified. Compared with the normal uric acid group, asymptomatic HUA and gout were associated with higher sepsis risk after full adjustment, with HRs of 1.31 (95% CI 1.26–1.36; P < 0.001) and 1.35 (95% CI 1.25–1.46; P < 0.001), respectively. Sepsis risk increased progressively across SUA quartiles, with HRs of 1.06 (95% CI 1.01–1.12), 1.12 (95% CI 1.07–1.19), and 1.27 (95% CI 1.20–1.34) for Q2, Q3, and Q4 relative to Q1, respectively. Restricted cubic spline analysis shown a nonlinear association between SUA and sepsis risk (P < 0.001; P for nonlinearity < 0.001), with risk increasing above approximately 384.8 μmol/L. For 28-day all-cause mortality among sepsis patients, asymptomatic HUA remained associated with a higher risk in the fully adjusted model (HR 1.17, 95% CI 1.07–1.29, P < 0.01), whereas the association for baseline gout was not significant after adjustment.
Conclusion
Higher SUA levels, HUA, and gout were associated with a higher incidence of sepsis in this observational cohort. Among patients who developed sepsis, asymptomatic HUA was associated with modestly higher short-term mortality, whereas this risk was not observed in the gout group.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s40560-026-00925-z.
Keywords: Sepsis, Hyperuricemia, Uric acid, Gout, Cohort studies
Introduction
Sepsis, a life-threatening acute organ dysfunction caused by a dysregulated host response to infection [1], has become a major global public health burden, with its incidence, mortality, and associated economic costs continuing to rise [2–4]. Concurrently, sepsis is shifting from a terminal event of severe infection to a common fatal complication of multiple major chronic non-communicable diseases [5]. In response to this trend, real-world, data-driven cohort studies are urgently needed to establish prevention and treatment frameworks.
However, the current capacity for sepsis risk prediction and the identification of vulnerable populations remains markedly insufficient. Although the early warning scores endorsed by contemporary international guidelines have partially supplanted earlier standards, each instrument carries intrinsic limitations, and an ideal screening tool that combines high sensitivity with high specificity is still lacking [6]. These limitations are further magnified when applied to particularly susceptible populations. Older and immunocompromised patients more frequently manifest atypical symptoms, substantially eroding the discriminatory performance of conventional vital-sign-based scores. Moreover, biomarkers, such as C-reactive protein and interleukin-6 lose considerable warning capacity in such patients due to a blunted inflammatory response [7, 8]. Therefore, there is a compelling need to introduce novel discriminative dimensions capable of capturing occult pathological progression.
Uric acid (UA) is the end product of purine nucleotide metabolism in humans [9]. Traditionally, hyperuricemia (HUA) has been regarded as being primarily associated with the onset of gouty arthritis. However, mounting evidence indicates that the pathological significance of HUA extends well beyond the spectrum of rheumatic diseases [10]. In recent years, the potential link between HUA and sepsis has drawn increasing attention, revealing a bidirectional vicious cycle. On the one hand, massive cell death, upregulation of xanthine oxidase activity, impaired renal excretion, and systemic metabolic disturbances triggered by sepsis can lead to a sharp rise in serum uric acid (SUA) levels [11, 12]. On the other hand, biological mechanisms suggest that SUA may participate in and exacerbate the pathological process of sepsis through activation of inflammasomes, induction of oxidative stress, and other pathways [13, 14]. Several clinical investigations conducted in intensive care unit populations have shown that HUA at admission is associated with elevated short- and long-term mortality, and this association has remained robust in subgroups with preserved renal function [15–17]. Nevertheless, most of these studies are constrained by small sample sizes and insufficient control for confounders, and they lack a systematic evaluation of consistency across clinically relevant subgroups; consequently, the independent prognostic value of HUA in patients with sepsis remains unresolved.
To fill the aforementioned research gaps, the present study conducted a large-scale prospective cohort study based on the UK Biobank. This study features a large sample size and a long follow-up period, thus ensuring adequate statistical power. In addition, this study also performed meticulous stratification of SUA levels and applied refined clinical diagnostic criteria, aiming to clearly delineate the graded relationship between different urate metabolic states and the risk of sepsis. The primary objective was to investigate the association between clinical urate phenotypes (normal SUA, asymptomatic HUA, and gout) and the risk of incident sepsis. Secondary analyses evaluated SUA levels and alternative urate classifications to further characterize dose–response relationships and assess the robustness of the findings. We additionally examined the association of urate status with short-term mortality following sepsis.
Methods
Data source and study population
The UK Biobank is a large-scale prospective cohort comprising over 500,000 residents of the United Kingdom. This study was conducted based on this resource (data from this database). All follow-ups ended on October 1, 2023.
Exposure measurement
Serum uric acid (SUA) concentration was measured using an enzymatic method. The UK Biobank implemented strict quality control procedures. Participants were initially divided into quartiles based on SUA concentration (Q1 < 250.4 μmol/L, 250.4 ≤ Q2 < 302.90 μmol/L, 302.90 ≤ Q3 < 360.70 μmol/L, and Q4 ≥ 360.70 μmol/L). Additionally, we defined HUA as SUA levels ≥ 420 μmol/L (7 mg/dL) in men and postmenopausal women, and ≥ 360 μmol/L (6 mg/dL) in premenopausal women.
Baseline gout was identified by integrating primary care records, hospital inpatient data, and self-reported medical history and death register records. Identified diagnoses were coded according to the International Classification of Diseases, 10th Revision (ICD-10: M10).
For the primary analyses, participants were classified into three mutually exclusive clinical urate phenotypes: (1) normal uric acid, defined as absence of gout and no HUA; (2) asymptomatic HUA, defined as HUA without gout; and (3) gout. In secondary analyses, SUA was additionally examined as quartiles (Q1–Q4), HUA status and gout status were evaluated separately as alternative exposure definitions.
Outcome definition
The primary outcome was incident sepsis, identified using the following International Classification of Diseases, 10th Revision (ICD-10) codes: A02, A39, A40, and A41, consistent with the previous literature [18–23]. These admission and diagnosis data were used to identify incident sepsis. Censoring was applied at the date of the first sepsis occurrence, death, withdrawal from the study, or the end of follow-up, whichever came first. This definition was adopted as a pragmatic epidemiologic proxy, as the UK Biobank lacks sufficiently complete, consistently time-stamped clinical measurements to operationalize Sepsis-3 criteria uniformly across the full cohort. Accordingly, the outcome examined in this study should be interpreted as ICD-coded sepsis, rather than clinically adjudicated sepsis per the Sepsis-3 consensus definition.
Sepsis-related mortality was assessed only among participants who developed sepsis during follow-up. Two prespecified mortality outcomes were evaluated: (1) all-cause mortality within 28 days of sepsis onset; (2) all-cause mortality within 60 days of sepsis onset. The date of sepsis onset was defined as the date of the first eligible hospitalization record. Mortality information was obtained from the national death registry. Participants contributed person-time from the date of sepsis onset until death, withdrawal from the study, or the end of follow-up, whichever came first.
Covariate assessment
Information on age, sex, ethnicity, socioeconomic status (Townsend deprivation index), education level, household income, employment status, smoking status, alcohol consumption, sleep duration, physical activity (MET-hours/week), body mass index (BMI), and history of hypertension (HTN), cardiovascular disease (CVD), diabetes mellitus (DM), and cancer was collected at baseline through nurse-led interviews and touchscreen questionnaires. We also collected data on the following biomarkers: low-density lipoprotein cholesterol (LDL-C), triglycerides, creatinine, and glycated hemoglobin (HbA1c). Full details of the covariates are provided in the Supplementary Methods.
Exclusion criteria
As shown in Figure S1, starting from the total baseline population of the UK Biobank (N = 502,318), we sequentially excluded individuals with a pre-existing sepsis onset at baseline (N = 2,268) and those lacking SUA measurements or lost to follow-up (N = 33,259), resulting in a final analytic cohort of 466,611 participants for this prospective cohort analysis.
Statistical analysis
Baseline characteristics were summarized according to the primary exposure categories: normal SUA, asymptomatic HUA, and gout. Data are reported as numbers (percentages) or means (standard deviations). Continuous variables were compared using the Analysis of Variance (ANOVA) or Kruskal–Wallis test, and categorical variables were compared using the chi-square test. The Bonferroni test was used for between-group comparisons.
Multivariable Cox proportional hazards models with varying degrees of covariate adjustment were constructed. The primary analysis evaluated the association between mutually exclusive clinical urate phenotypes (normal SUA, asymptomatic HUA, and gout) and incident sepsis using multivariable Cox proportional hazards models. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using normal SUA as the reference category. Secondary analyses were conducted using alternative exposure definitions, including SUA quartiles, HUA status, and gout status separately, to assess the consistency of the findings across different classifications of urate metabolism. The proportional hazards assumption was assessed using scaled Schoenfeld residuals. The primary exposure framework based on clinical urate phenotypes satisfied the proportional hazards assumption (P = 0.06). Similar results were observed for the secondary exposure definitions of HUA status (P = 0.16) and gout status (P = 0.37). A deviation from proportionality was observed for SUA quartiles (P < 0.05). Follow-up time was calculated from the date of enrollment to the date of sepsis onset, death, withdrawal from the study, or the end of follow-up, whichever occurred first. We generated Kaplan–Meier cumulative incidence curves for sepsis stratified by SUA/gout status and compared them using the log-rank test. Among the secondary analyses, restricted cubic spline models with five knots (5%, 27.5%, 50%, 72.5%, and 95%) were used to examine potential nonlinear associations between continuous SUA levels and sepsis risk in Model 3.
We assessed the consistency of the primary association across prespecified subgroups, including age group (< 65 years, ≥ 65 years), sex (male, female), alcohol consumption (current, previous, never), smoking status (current, previous, never), BMI (< 25 kg/m2, 25 kg/m2 ~ 30 kg/m2, ≥ 30 kg/m2), history of HTN, CVD and DM (yes, no). The P values for the interaction terms between gout status and each stratification variable were used to evaluate the significance of interactions.
To assess robustness, the primary incidence models were re-estimated after sequentially excluding participants who developed sepsis within 1, 2, and 3 years of baseline to mitigate potential reverse causation. Because the primary objective of this study was to evaluate the association between urate phenotypes and the hazard of a first sepsis event, Cox proportional hazards models were prespecified as the primary Analytic approach. Given the potential influence of renal function and medication use on both urate metabolism and infection susceptibility, an additional sensitivity analysis was performed using extended adjustment models. Model 4 was based on Model 3, with serum creatinine replaced by estimated glomerular filtration rate (eGFR, mL/min/1.73 m2) [24]. Model 5 was further adjusted for baseline use of corticosteroids (dexamethasone, prednisone, prednisolone, methylprednisolone, budesonide, hydrocortisone, triamcinolone, and betamethasone), urate-lowering therapy (allopurinol and probenecid), and diuretics (furosemide, bumetanide, torasemide, hydrochlorothiazide, bendroflumethiazide, indapamide, and spironolactone), colchicine, disease-modifying antirheumatic drugs (DMARDs) (methotrexate, leflunomide, mycophenolate mofetil, tacrolimus, cyclosporine, hydroxychloroquine, cyclophosphamide, and azathioprine), all coded as binary variables (yes/no), in addition to the covariates included in Model 4. Considering that all-cause mortality may act as a competing event for incident sepsis during long-term follow-up, additional competing-risk analyses were performed using the Fine-Gray subdistribution hazard model to evaluate the potential impact of competing mortality on the primary findings. Furthermore, we conducted landmark analyses at 3-year and 5-year time points. Participants who developed gout during the first 3 or 5 years of follow-up were treated as a separate category, and participants were classified into four groups: normal, asymptomatic HUA, baseline gout, and future incident gout. These landmark analyses were then performed to further evaluate the robustness of the associations over time.
To evaluate the impact of SUA categories and gout phenotypes on short-term survival among patients with sepsis, we examined all-cause mortality at 28 and 60 days following the sepsis onset. Participants were categorized based on baseline SUA levels and gout status. The associations between gout status and mortality risk were assessed using multivariable Cox proportional hazards models, adjusted for the full set of covariates specified in Model 3, including age at recruitment, sex, ethnic group, Townsend deprivation index, education level, income, smoking status, alcohol consumption, sleep duration/sleep status, physical activity, employment status, BMI, LDL-C, total cholesterol, triglycerides, creatinine, HbA1c, and history of HTN, cardiovascular disease, and cancer. Results were reported as HRs with corresponding 95% CIs.
All tests were two-sided, and P < 0.05 was considered statistically significant. All analyses were performed using R 4.5.1 and SPSS 27.
Results
Analysis of baseline characteristics of the study population (N = 466,611) consistently showed that patients with asymptomatic HUA and gout had a significantly greater disease burden (Tables 1, S1–S3). As shown in Table 1, compared with participants with normal SUA, individuals with asymptomatic HUA or gout were generally older and more frequently male, and tended to have less favorable socioeconomic profiles, including lower educational attainment and lower income levels. Differences were also observed in lifestyle characteristics, with a greater proportion of previous smokers and shorter sleep duration among participants with asymptomatic HUA and gout. Marked differences were observed in anthropometric and metabolic characteristics. Participants with asymptomatic HUA and gout had substantially higher levels of adiposity, with obesity occurring in nearly half of affected individuals compared with approximately one-fifth of those with normal urate status. Metabolic abnormalities were also more pronounced in these groups, characterized by higher triglyceride, creatinine and HbA1c levels, whereas LDL-C levels were comparatively lower in the gout group. Similarly, the burden of comorbidities increased progressively across urate phenotypes. The prevalence of hypertension rose from 23.6% in participants with normal uric acid status to 46.7% in asymptomatic HUA and 55.3% in gout, while cardiovascular disease increased from 5.0 to 9.7% and 14.7%, respectively. Incident sepsis during follow-up also showed a graded increase across groups (3.1%, 5.5%, and 7.2%, respectively). When participants were further stratified by clinical status (Tables S1 and S2) and by SUA quartiles (Table S3), similar patterns were observed, indicating that higher SUA levels were accompanied by a progressively less favorable risk profile.
Table 1.
Baseline characteristics of the study population stratified by clinical urate phenotypes in the UK Biobank
| Normal | Asymptomatic HUA | Baseline gout | P | |
|---|---|---|---|---|
| n | 400464 | 56242 | 9905 | |
| Age at recruitment [mean (SD)]a | 56.25 (8.12) | 57.93 (7.79) | 59.87 (6.96) | < 0.001 |
| Age group (%)b | < 0.001 | |||
| < 65 | 327757 (81.8) | 42845 (76.2) | 6914 (69.8) | |
| ≥ 65 | 72707 (18.2) | 13397 (23.8) | 2991 (30.2) | |
| Sex (%)b | < 0.001 | |||
| Male | 171767 (42.9) | 32835 (58.4) | 8891 (89.8) | |
| Female | 228697 (57.1) | 23407 (41.6) | 1014 (10.2) | |
| Race (%)b | < 0.001 | |||
| White | 377625 (94.3) | 52756 (93.8) | 9416 (95.1) | |
| Townsend deprivation [mean (SD)]a | −1.35 (3.07) | −1.06 (3.19) | −1.19 (3.12) | < 0.001 |
| College or university degree (%)b | 133195 (33.3) | 15458 (27.5) | 2513 (25.4) | < 0.001 |
| Income (%)b | < 0.001 | |||
| Less than 18,000 | 75847 (18.9) | 12419 (22.1) | 2257 (22.8) | |
| 18,000–30,999 | 86554 (21.6) | 12531 (22.3) | 2286 (23.1) | |
| 31,000–51,999 | 90246 (22.5) | 11643 (20.7) | 2075 (20.9) | |
| 52,000–100,000 | 70878 (17.7) | 8780 (15.6) | 1528 (15.4) | |
| Greater than 100, 000 | 18761 (4.7) | 2388 (4.2) | 466 (4.7) | |
| Employment status (%)b | 0.017 | |||
| Employed | 366651 (91.6) | 51210 (91.1) | 8997 (90.8) | |
| Unemployed | 29509 (7.4) | 4477 (8.0) | 797 (8.0) | |
| Smoke (%)b | < 0.001 | |||
| Current | 42793 (10.7) | 5278 (9.4) | 894 (9.0) | |
| Previous | 132609 (33.1) | 23682 (42.1) | 4782 (48.3) | |
| Never | 223079 (55.7) | 26948 (47.9) | 4180 (42.2) | |
| Drink (%)b | < 0.001 | |||
| Current | 367118 (91.7) | 51784 (92.1) | 9321 (94.1) | |
| Previous | 14421 (3.6) | 1926 (3.4) | 361 (3.6) | |
| Never | 17933 (4.5) | 2383 (4.2) | 197 (2.0) | |
| Sleep (%)b | < 0.001 | |||
| 7–8 h/d | 271259 (67.7) | 35781 (63.6) | 6327 (63.9) | |
| < 7 h/d | 104287 (26.0) | 16199 (28.8) | 2769 (28.0) | |
| > 8 h/d | 24918 (6.2) | 4262 (7.6) | 809 (8.2) | |
| MET (hours/week)a | 41.31 (39.53) | 38.21 (38.16) | 40.14 (40.96) | < 0.001 |
| BMI cate (%)b | < 0.001 | |||
| < 25 | 146968 (36.7) | 5885 (10.5) | 1020 (10.3) | |
| 25 ~ 30 | 172039 (43.0) | 23280 (41.4) | 4347 (43.9) | |
| ≥ 30 | 81457 (20.3) | 27077 (48.1) | 4538 (45.8) | |
| LDL-C [mean (SD)]a | 3.56 (0.86) | 3.61 (0.92) | 3.28 (0.88) | < 0.001 |
| Triglycerides [mean (SD)]a | 1.66 (0.95) | 2.27 (1.24) | 2.42 (1.39) | < 0.001 |
| Creatinine [mean (SD)]a | 70.58 (15.66) | 81.67 (23.34) | 87.30 (43.24) | < 0.001 |
| HbA1C [mean (SD)]a | 3.48 (0.59) | 3.61 (0.61) | 3.67 (0.76) | < 0.001 |
| CVD (%)b | 19894 (5.0) | 5434 (9.7) | 1455 (14.7) | < 0.001 |
| HTN (%)b | 94274 (23.6) | 26147 (46.7) | 5459 (55.3) | < 0.001 |
| Cancer (%)b | 29497 (7.4) | 4841 (8.6) | 778 (7.9) | 0.005 |
| Follow up time [mean (SD)]a | 14.10 (2.24) | 13.77 (2.71) | 13.40 (3.02) | < 0.001 |
| incident sepsis (%)b | 12402 (3.1) | 3098 (5.5) | 710 (7.2) | < 0.001 |
HUA Hyperuricaemia, MET Metabolic Equivalent of Task, BMI Body mass index, CVD Cardiovascular disease, HTN Hypertension, LDL-C Low-density lipoprotein-cholesterol, HbA1C Hemoglobin A1c
aANOVA test was used to compare the continuous variables among different groups
bChi-square test was used to compare the categorical variables among different groups
During a median follow-up of 14.53 years, a total of 16,210 incident sepsis cases were recorded. Kaplan–Meier curves in Fig. 1 show statistically significant differences in the cumulative incidence of sepsis among groups with different urate/gout statuses or clinical urate phenotypes (log-rank test P < 0.001). When participants were categorized by clinical urate phenotypes, individuals with asymptomatic HUA showed a higher sepsis risk than those with normal SUA, and those with baseline gout had an even higher risk. Correspondingly, sepsis risk escalated across SUA quartiles (Q4 > Q3 > Q2 > Q1) and was significantly higher in HUA and baseline gout than in normal SUA.
Fig. 1.
Uric acid and baseline gout linked to stepwise increase in long-term sepsis risk. A Incidence curves categorized by clinical urate phenotypes: normal, asymptomatic HUA and baseline gout. B Kaplan–Meier estimates of sepsis incidence across quartiles of SUA levels (Q1–Q4). C Comparison of sepsis risk between participants with normal SUA and those with HUA. D Incidence curves categorized by gout status: normal and baseline gout. The cumulative incidence of sepsis was significantly higher in groups with elevated SUA and gout phenotypes compared to controls (P < 0.001, log-rank test)
Figure 2 illustrates a significant nonlinear association between SUA and sepsis risk, as revealed by restricted cubic spline analysis (P < 0.001, P for nonlinearity < 0.001). A model-derived reference point was identified at an SUA concentration of 301.6 μmol/L. The RCS curve remained relatively flat across a broad range of SUA levels, with a statistically significant increase in sepsis risk observed only at higher SUA concentrations exceeding approximately 384.8 μmol/L, where the lower bound of the 95% CIs surpassed 1.0.
Fig. 2.
Elevated serum uric acid is independently associated with increased risk of incident sepsis. Restricted cubic splines illustrate a non-linear relationship between baseline SUA and sepsis risk (P < 0.001). Data are presented as adjusted hazard ratios (solid line) with 95% CIs (shaded area)
To examine the associations between serum urate quartiles, clinical urate phenotypes, and incident sepsis, multivariable Cox proportional hazards models were fitted, with results summarized in Table 2. Individuals with asymptomatic HUA showed an increased risk of sepsis compared to those with normal SUA across all models, although the magnitude of association was progressively attenuated after covariate adjustment, with HRs decreasing from 1.83 (95% CI 1.76–1.91) in the crude model to 1.31 (95% CI 1.26–1.36) in the fully adjusted model (Model 3). Similarly, participants with baseline gout exhibited a significantly elevated risk, with HRs decreasing from 2.46 (95% CI 2.28–2.65) in the crude model to 1.35 (95% CI 1.25–1.46) after full adjustment (all P < 0.001). Consistent findings were observed when SUA was analyzed by quartiles. Compared with the lowest quartile (Q1), the fully adjusted HRs (95% CIs) for Q2, Q3, and Q4 were 1.06 (1.01–1.12), 1.12 (1.07–1.19), and 1.27 (1.20–1.34), respectively. Consistent with the quartile-based findings, in the fully adjusted model, HUA was associated with a higher risk of incident sepsis compared with normal urate (HR 1.29; 95% CI 1.24–1.35), while baseline gout was associated with a higher risk compared the normal group (those without gout) (HR 1.25; 95% CI 1.16–1.35). Notably, the HRs for all aforementioned groups showed substantial attenuation after multivariable adjustment (despite all P values remaining < 0.001), suggesting that differences in baseline demographic, metabolic, and comorbidity profiles partially accounted for the crude associations observed in the unadjusted model.
Table 2.
Association between serum uric acid quartiles, clinical urate phenotypes and the risk of incident sepsis
| Crude mode | Model 1 | Model 2 | Model 3 | |
|---|---|---|---|---|
| Clinical urate phenotypes | ||||
| Normal | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Asymptomatic HUA | 1.83 (1.76, 1.91)* | 1.65 (1.58, 1.71)* | 1.55 (1.49, 1.62)* | 1.31 (1.26, 1.36)* |
| Baseline gout | 2.46 (2.28, 2.65) * | 1.85 (1.71, 1.99)* | 1.73 (1.60, 1.86)* | 1.35 (1.25, 1.46)* |
| Serum uric acid quartiles | ||||
| Q1 | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Q2 | 1.28 (1.22, 1.35)* | 1.15 (1.09, 1.21)* | 1.13 (1.07, 1.19)* | 1.06 (1.01, 1.12) |
| Q3 | 1.59 (1.51, 1.66)* | 1.29 (1.23, 1.36)* | 1.26 (1.20, 1.33)* | 1.12 (1.07, 1.19)* |
| Q4 | 2.14 (2.05, 2.24)* | 1.64 (1.55, 1.72)* | 1.56 (1.48, 1.64)* | 1.27 (1.20, 1.34)* |
| Hyperuricemia status | ||||
| Normal uric acid | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Hyperuricemia | 1.83 (1.76, 1.90)* | 1.62 (1.56, 1.69)* | 1.54 (1.48, 1.60)* | 1.29 (1.24, 1.35)* |
| Gout status | ||||
| Normal | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Baseline gout | 2.24 (2.07, 2.41) * | 1.67 (1.55, 1.81)* | 1.58 (1.46, 1.70)* | 1.25 (1.16, 1.35)* |
Quartiles of uric acid: Q1 < 250.4 μmol/L, 250.4 ≤ Q2 < 302.90 μmol/L, 302.90 ≤ Q3 < 360.70 μmol/L, Q4 ≥ 360.70 μmol/L
Model 1 was adjusted agegroup (< 65, 65 and over), sex (men, women), ethnic group (White and non-white)
Model 2: Model 1 + Townsend deprivation index (continuous) + education (less than 12th grade, high school graduate, some college or AA degree, and college graduate or above), income (less than 18,000, 18,000–30,999, 31,000–51, 999, 52,000–100,000), smoking (never, previous, current), drink (current, previous, never), sleep (7–8 h/d, < 7 h/d, > 8 h/d), physical activity (MET-hours/week), employment status (employed, unemployed and not answer)
Model 3: Model 2 + BMI(≤ 25 kg/m2, 25 ~ 30 kg/m2, > 30 kg/m2), LDL-C (mmol/L), Cholesterol (mmol/L), Triglycerides (mmol/L), Creatinine (μmol/L), HbA1C (%), history of hypertension, cardiovascular disease, cancer (yes, no, not answer)
*P < 0.001
To evaluate the robustness of the associations between clinical urate phenotypes, serum uric acid quartiles, HUA, gout status and to mitigate potential reverse causality, a series of sensitivity analyses were conducted (Tables S4-S8), excluding participants who developed sepsis within 1, 2, and 3 years of follow-up. Across all analyses, the results remained consistent with the primary findings. In Table S4, compared with participants with normal SUA, both asymptomatic HUA and baseline gout were consistently associated with a higher risk of incident sepsis across all adjustment models. The effect estimates showed progressive attenuation from minimally adjusted to fully adjusted models, but the associations remained statistically significant. In the fully adjusted model (Model 3), excluding participants who developed sepsis within the first year yielded HRs of 1.31 (95% CI 1.25–1.36) for asymptomatic HUA and 1.35 (95% CI 1.25–1.46) for gout group. Similar estimates were observed after excluding cases occurring within the first two and three years of follow-up. In Table S5, in the Model 3, increasing SUA quartiles were associated with progressively higher sepsis risk, with participants in the highest quartile showing an approximately 27–28% increased risk compared with the lowest quartile (all P for trend < 0.05). Table S6 showed consistent associations between HUA and incident sepsis in the fully adjusted model (HR: 1.29–1.30), while Table S7 showed similar findings for gout group (HR: 1.23–1.25). In Table S8, further adjustment for eGFR and medications (steroids, urate-lowering therapy, diuretics, colchicine and DMARDS) yielded similar patterns (Model 5), with asymptomatic HUA remaining significantly associated with incident sepsis (HR 1.21, 95% CI 1.16–1.27), supporting the consistency of the observed associations.
The results of the competing-risk analyses are presented in Table S9. Our findings showed that, even after accounting for the competing-risk of death, participants with HUA and gout remained significantly associated with an increased risk of incident sepsis. In Model 3, the HRs were 1.31 (95% CI 1.23–1.40) for asymptomatic HUA and 1.35 (95% CI 1.20–1.52) for participants with baseline gout. The results of the 3-year and 5-year landmark analyses are presented in Table S10. The associations between clinical urate phenotypes, and incident sepsis remained consistent in both landmark analyses. In Model 3 of the 3-year landmark analysis, the HRs (95% CIs) for incident sepsis were 1.30 (1.25–1.36) for participants with asymptomatic HUA, 1.33 (1.23–1.45) for participants with baseline gout, and 1.57 (1.32–1.88) for participants who later developed gout. Similar findings were observed in the 5-year landmark analysis, with corresponding HRs (95% CIs) of 1.31 (1.25–1.37), 1.29 (1.18–1.40), and 1.57 (1.36–1.80), respectively.
Stratified analyses by demographic characteristics, lifestyle factors, and comorbidities are presented in Tables 3 and S11. Overall, the associations between clinical urate phenotypes, gout status, HUA and incident sepsis were consistent across most subgroups. In Table 3, sex and BMI significantly modified the association between clinical urate phenotypes and sepsis risk (both P for interaction < 0.01). Compared with individuals with normal SUA levels, the HRs for sepsis were 1.24 (95% CI 1.18–1.31) for asymptomatic HUA and 1.29 (95% CI 1.18–1.40) for baseline gout in males, while the corresponding HRs were stronger in females at 1.39 (95% CI 1.30–1.49) and 1.89 (95% CI 1.52–2.34), respectively. For BMI subgroups, the association between asymptomatic HUA and sepsis was most pronounced in individuals with BMI < 25 kg/m2 (HR 1.54, 95% CI 1.36–1.74), and gradually attenuated with increasing BMI. Table S11 further confirmed these findings, showing that sex significantly modified both the HUA-sepsis and gout-sepsis associations (both P for interaction < 0.01), while BMI significantly modified only the HUA-sepsis association (P for interaction < 0.01).
Table 3.
Association between clinical urate phenotypes and the risk of incident sepsis stratified by subgroups (Model 3)
| Normal | Asymptomatic HUA | Baseline gout | P for interaction | |
|---|---|---|---|---|
| Age group | 0.42 | |||
| < 65 | 1.00 (Ref.) | 1.30 (1.23, 1.37) | 1.52 (1.37, 1.68) | |
| ≥ 65 | 1.00 (Ref.) | 1.33 (1.24, 1.43) | 1.14 (1.01, 1.30) | |
| Sex | < 0.01 | |||
| Male | 1.00 (Ref.) | 1.24 (1.18, 1.31) | 1.29 (1.18, 1.40) | |
| Female | 1.00 (Ref.) | 1.39 (1.30, 1.49) | 1.89 (1.52, 2.34) | |
| BMI | < 0.01 | |||
| < 25 | 1.00 (Ref.) | 1.54 (1.36, 1.74) | 1.25 (0.96, 1.64) | |
| 25 ~ 30 | 1.00 (Ref.) | 1.32 (1.24, 1.41) | 1.46 (1.29, 1.65) | |
| ≥ 30 | 1.00 (Ref.) | 1.25 (1.18, 1.32) | 1.29 (1.16, 1.44) | |
| CVD | 0.28 | |||
| Yes | 1.00 (Ref.) | 1.30 (1.17, 1.44) | 1.24 (1.04, 1.47) | |
| No | 1.00 (Ref.) | 1.31 (1.25, 1.37) | 1.38 (1.26, 1.51) | |
| HTN | 0.50 | |||
| Yes | 1.00 (Ref.) | 1.32 (1.25, 1.40) | 1.40 (1.27, 1.54) | |
| No | 1.00 (Ref.) | 1.31 (1.23, 1.39) | 1.31 (1.14, 1.50) | |
| DM | 0.62 | |||
| Yes | 1.00 (Ref.) | 1.39 (1.25, 1.54) | 1.28 (1.09, 1.51) | |
| No | 1.00 (Ref.) | 1.29 (1.23, 1.35) | 1.39 (1.27, 1.51) | |
| Smoke | 0.19 | |||
| Current | 1.00 (Ref.) | 1.29 (1.14, 1.45) | 1.21 (0.94, 1.54) | |
| Previous | 1.00 (Ref.) | 1.34 (1.26, 1.43) | 1.31 (1.18, 1.46) | |
| Never | 1.00 (Ref.) | 1.27 (1.19, 1.36) | 1.48 (1.30, 1.68) | |
| Drink | 0.88 | |||
| Current | 1.00 (Ref.) | 1.31 (1.25, 1.37) | 1.38 (1.27, 1.50) | |
| Previous | 1.00 (Ref.) | 1.36 (1.14, 1.62) | 1.01 (0.71, 1.45) | |
| Never | 1.00 (Ref.) | 1.21 (1.00, 1.46) | 0.84 (0.47, 1.53) |
HUA Hyperuricemia, BMI Body mass index, CVD Cardiovascular disease, HTN Hypertension, DM diabetes
Model 3 was adjusted agegroup (< 65, 65 and over), sex (men, women), ethnic group (White and non-White), Townsend deprivation index (continuous) + education (less than 12th grade, high school graduate, some college or AA degree, and college graduate or above), income (less than 18,000, 18,000–30,999, 31,000–51,999, 52,000–100,000), smoking (never, previous, current), drink (current, previous, never), sleep (7–8 h/d, < 7 h/d, > 8 h/d), physical activity (MET-hours/week), employment status (employed, unemployed and not answer), BMI (≤ 25 kg/m2, 25 ~ 30 kg/m2, > 30 kg/m2), LDL-C (mmol/L), Cholesterol (mmol/L), Triglycerides (mmol/L), Creatinine (μmol/L), HbA1C (%), history of hypertension, cardiovascular disease, cancer (yes, no, not answer)
Tables 4 and S12 present the associations of SUA levels and clinical urate phenotypes with 28-day and 60-day all-cause mortality among patients with sepsis. The asymptomatic HUA remained statistically associated with higher 28-day mortality after multivariable adjustment (Model 3: HR 1.17, 95% CI 1.07–1.29; P < 0.01). Baseline gout showed a significant crude association (HR 1.20, 95% CI 1.02–1.42; P < 0.05) but became nonsignificant after adjustment for confounders (Model 3: HR 1.13, 95% CI 0.95–1.35; P > 0.05). Similar patterns were observed for 60-day mortality. Asymptomatic HUA remained independently associated with a higher risk (Model 3: HR 1.19, 95% CI 1.10–1.30; P < 0.01), while baseline gout lost statistical significance after full adjustment (Model 3: HR 1.15, 95% CI 0.98–1.34; P > 0.05). Participants in the highest SUA quartile had significantly elevated risks of both 28-day and 60-day mortality compared with the lowest quartile. Overall HUA was also independently associated with increased mortality at both time points (28-day Model 3: HR 1.15, 95% CI 1.05–1.26; P < 0.01; 60-day Model 3: HR 1.18, 95% CI 1.08–1.27; P < 0.01). After further adjustment for eGFR and medications (Table S13), the associations remained significant for asymptomatic HUA (28-day Model 5: HR 1.13, 95% CI 1.03–1.25; P < 0.05; 60-day Model 5: HR 1.15, 95% CI 1.05–1.25; P < 0.01) and overall HUA (28-day Model 5: HR 1.11, 95% CI 1.01–1.22; P < 0.05; 60-day Model 5: HR 1.14, 95% CI 1.05–1.24; P < 0.01), whereas the association for the highest SUA quartile lost statistical significance at both time points (all P > 0.05), and baseline gout consistently showed no independent association.
Table 4.
Hazard ratios (95% CIs) for 28-day sepsis-related death according to serum uric acid levels and clinical urate phenotypes
| Crude mode | Model 1 | Model 2 | Model 3 | |
|---|---|---|---|---|
| Normal | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Asymptomatic HUA | 1.20 (1.10, 1.31)** | 1.17 (1.07, 1.28)** | 1.15 (1.06, 1.26)** | 1.17 (1.07, 1.29)** |
| Baseline gout | 1.20 (1.02, 1.42)* | 1.13 (0.95, 1.34) | 1.13 (0.95, 1.33) | 1.13 (0.95, 1.35) |
| Q4SUA | ||||
| Q1 | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Q2 | 1.03 (0.92, 1.14) | 0.97 (0.87, 1.08) | 0.97 (0.88, 1.08) | 1.00 (0.89, 1.11) |
| Q3 | 1.05 (0.95, 1.17) | 0.97 (0.87, 1.08) | 0.98 (0.88, 1.10) | 1.02 (0.91, 1.14) |
| Q4 | 1.20 (1.09, 1.33)** | 1.09 (0.98, 1.22) | 1.09 (0.98, 1.21) | 1.13 (1.00, 1.27)* |
| Normal uric acid | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Hyperuricemia | 1.18 (1.08, 1.28)** | 1.15 (1.06, 1.25)** | 1.13 (1.04, 1.24)** | 1.15 (1.05, 1.26)** |
| Normal | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) | 1.00 (Ref.) |
| Baseline gout | 1.16 (0.98, 1.37) | 1.09 (0.92, 1.29) | 1.09 (0.92, 1.29) | 1.07 (0.90, 1.27) |
HUA Hyperuricaemia
Model 1 was adjusted agegroup (< 65, 65 and over), sex (men, women), ethnic group (White and non-white)
Model 2: Model 1 + Townsend deprivation index (continuous) + education (less than 12th grade, high school graduate, some college or AA degree, and college graduate or above), income (less than 18,000, 18,000–30,999, 31,000–51, 999, 52,000–100, 000), smoking (never, previous, current), drink (current, previous, never), sleep (7–8 h/d, < 7 h/d, > 8 h/d), physical activity (MET-hours/week), employment status (employed, unemployed and not answer)
Model 3: Model 2 + BMI (≤ 25 kg/m2, 25 ~ 30 kg/m2, > 30 kg/m2), LDL-C (mmol/L), Cholesterol (mmol/L), Triglycerides (mmol/L), Creatinine (μmol/L), HbA1C (%), history of hypertension, cardiovascular disease, cancer (yes, no, not answer)
*P < 0.05; **P < 0.01
Discussion
Based on data from the large-scale prospective UK Biobank cohort, this study systematically investigated the associations of clinical urate phenotypes, serum uric acid quartiles, HUA and gout status with the risk of incident sepsis and short-term prognosis of sepsis. We observed that clinical urate phenotypes during the first 3 or 5 years of follow-up were associated with a higher risk of incident sepsis. Stratified analyses suggested that these associations were particularly pronounced in women and individuals with normal BMI. Furthermore, restricted cubic spline analysis of SUA levels revealed a nonlinear dose–response relationship between SUA and sepsis risk, suggesting that higher SUA levels may serve as markers of increased sepsis susceptibility. Finally, regarding prognosis, HUA status, particularly asymptomatic HUA, was significantly associated with an increased risk of 28-day and 60-day mortality in sepsis patients, whereas gout status itself was not independently associated with short-term mortality.
In prior research, Qin et al. [25] provided genetic evidence supporting a positive association between SUA levels and sepsis. However, that study cannot directly capture the impact of short-to-medium-term clinical interventions on sepsis risk. The present study better reflects individuals’ biological states and dynamic changes in real-world clinical settings. Furthermore, the findings of Liu et al. using the MIMIC-III database [15] corroborate the short-term prognostic findings of the present study. Alshehri et al. [17] reported conflicting results from a study of 599 ICU sepsis patients, which may be attributed to the retrospective nature of their study and the inherent limitations in data completeness and consistency. Moreover, ICU-based retrospective studies involve critically ill populations, among whom risk associations may differ from those observed in the general population with HUA. In contrast, our prospective design allowed for more complete and standardized data collection, minimizing recall and measurement bias. Our larger sample size also improves the representativeness and generalizability of the findings.
The present study showed a significant nonlinear association between SUA and sepsis risk, and the inflection point lay notably within the range conventionally regarded as normouricemic, a finding consistent with evidence from multiple large-scale cohort studies. In a study drawing on the NHANES cohort, the U-shaped relationships of UA with all-cause and cardiovascular mortality exhibited inflection points at approximately 5.7 mg/dL and 5.9 mg/dL, respectively [26, 27]. Similarly, Virdis et al. [28] showed in the URRAH study that the SUA thresholds predicting all-cause and cardiovascular mortality risk were 4.7 mg/dL and 5.6 mg/dL. Collectively, these observations point toward an optimal SUA window of roughly 5–6 mg/dL. Mechanistically, these findings align with the concentration-dependent dual effect of SUA on health outcomes. At physiological concentrations, soluble UA is recognized as a potent free radical scavenger, capable of neutralizing reactive oxygen and nitrogen species such as peroxynitrite, chelating metal ions, and conferring protection through activation of the Nrf2/HO-1 pathway [29]. Once SUA exceeds a critical threshold, however, it shifts toward a pro-oxidant state, promoting oxidative stress [30], endothelial dysfunction, vascular smooth muscle cell proliferation [31], inflammation [10, 32], and insulin resistance [33]. In this study, patients with HUA showed an increased risk of incident sepsis and significantly higher short-term mortality.
Beyond the mechanisms discussed above, this association may also be explained by the following: sustained HUA and monosodium urate crystal deposition aberrantly activate multiple immune pathways and perpetuate a chronic low-grade inflammatory state, maintaining immune cells in a sensitized, partially activated condition [34, 35]. A secondary insult, such as an infection precipitating sepsis, then triggers the rapid and intense release of inflammatory mediators from these primed cells, precipitating an uncontrolled systemic inflammatory response [36]. However, it should be emphasized that although several biological pathways may plausibly underlie these observed associations, the present study was not designed to establish mechanistic or causal relationships.
The stratified analysis revealed stronger associations between clinical urate phenotypes, HUA, gout status and incident sepsis in females and individuals with normal body weight (BMI < 25). Regarding gender difference, a possible explanation is the modulating effect of estrogen on UA metabolism [37, 38] and immune responses [39], potentially leading to more pronounced inflammatory responses in postmenopausal women or those with HUA. The stronger association in individuals with normal BMI may reflect the absence of obesity as a dominant confounding factor, thereby more clearly revealing the independent pathological contribution of HUA. This suggests that screening and managing HUA in nonobese populations may be of unique importance for the primary prevention of sepsis.
The findings of this study suggest a risk gradient of incident sepsis across clinical urate phenotypes, with the strongest association observed in individuals who subsequently developed gout. This pattern is consistent with an underlying continuous metabolic-inflammatory process, supporting that serum urate dysregulation primarily affects host susceptibility to sepsis. This phenomenon may reflect more sustained or progressive metabolic dysregulation preceding clinically overt disease. In contrast, although HUA remained statistically associated with short-term mortality after full adjustment, baseline gout showed no significant relationship. From a clinical perspective, this suggests that urate-related metabolic dysfunction is more likely to act at the stage of disease initiation rather than determining outcomes once sepsis has developed. Notably, the effect sizes for mortality outcomes were modest, indicating limited clinical impact. The attenuation of the association between the highest SUA quartile and incident sepsis after multivariable adjustment suggests that the crude association may partly reflect shared cardiometabolic comorbidities, rather than a direct causal effect of extreme SUA levels.
This study utilizes the UK Biobank, a large-scale, prospective cohort with a substantial sample size and long follow-up, providing rich clinical and biochemical data that ensure statistical power and the reliability of the findings. The primary outcome was incident sepsis, which carries public health and clinical significance for primary prevention. Throughout the analysis, extensive adjustments were made for potential confounding factors, reducing the influence of measured confounding. We also performed refined stratification of HUA and gout based on clinical standards, clearly distinguishing between "asymptomatic HUA," "baseline gout," and "future incident gout," thereby more comprehensively characterizing the differential risks of sepsis associated with different urate metabolic states. This framework provides a theoretical basis for implementing stratified prevention and intervention strategies based on different urate phenotypes to improve sepsis outcomes. In terms of statistical analysis, the study employed multiple analytic approaches, such as Cox proportional hazards models and restricted cubic splines to systematically assess the dose–response relationship and nonlinear trend between SUA and sepsis risk, supplemented by sensitivity and subgroup analyses to ensure the robustness of the observed associations.
Nevertheless, this study has several limitations worth acknowledging. First, SUA levels were based on a single measurement which may be insufficient to fully reflect intraindividual dynamic changes over time. Second, despite adjustment for a broad range of covariates, the observational design precludes causal inference, and residual or unmeasured confounding may persist. Third, given the differences in genetic background, lifestyle, and UA metabolism across different racial and ethnic groups, caution is warranted when extrapolating the findings of this study to other populations. Therefore, future validation in diverse, independent multiethnic and multi-center cohorts is warranted. Additionally, sepsis was ascertained using prespecified ICD-10 codes instead of clinical adjudication or direct application of the Sepsis-3 criteria. Variation in diagnostic recognition, clinical documentation, and coding practices may have resulted in outcome misclassification, although the direction and magnitude of any resulting bias cannot be determined from the available data. Accordingly, alternative analyses that could not be adequately implemented and validated using the available data were not presented. The findings should therefore be interpreted in the context of an administratively ascertained, hospital-recorded sepsis endpoint, and their applicability to clinically adjudicated or strictly Sepsis-3-defined sepsis remains to be established. Further studies incorporating detailed clinical data are warranted to evaluate the robustness and generalizability of these findings.
In summary, higher SUA levels, HUA, and gout were associated with a higher incidence of sepsis in this large prospective cohort. Asymptomatic HUA was also associated with poorer short-term outcomes among patients with sepsis. These findings suggest that urate metabolism may represent a clinically relevant marker for sepsis risk stratification and provide a basis for future research exploring whether urate pathway-targeted strategies may influence sepsis risk and prognosis. These results should be interpreted as observational associations rather than evidence of causality, and substantial residual or unmeasured confounding cannot be excluded. Further mechanistic and interventional studies are needed to determine whether urate metabolism plays a causal role in sepsis development or prognosis.
Supplementary Information
Acknowledgements
This research was conducted using the UK Biobank resource under application number 479734. We are grateful to all study participants for their cooperation. We are grateful for the support of National Natural Science Foundation of China, Beijing Key Clinical Specialty Program and Peking Union Medical College Hospital Talent Cultivation Program.
Abbreviations
- BMI
Body mass index
- CI
Confidence interval
- CVD
Cardiovascular disease
- DM
Diabetes mellitus
- HbA1c
Glycated hemoglobin
- HR
Hazard ratio
- HTN
Hypertension
- HUA
Hyperuricemia
- ICD-10
International Classification of Diseases, Tenth Revision
- ICU
Intensive care unit
- LDL-C
Low-density lipoprotein cholesterol
- MET
Metabolic equivalent of task
- RCS
Restricted cubic splines
- SUA
Serum uric acid
- UA
Uric acid
Author contributions
All authors helped to perform the research. Conceptualization, Yingdong Han, Yun Zhang and Xuejun Zeng; Data Curation, Yingdong Han and Yudian Zhang; Formal Analysis, Yingdong Han, Juan Wu and Menghui Yao; Investigation, Yingdong Han, Tiange Xie and Yudian Zhang; Methodology, Yingdong Han, Yudian Zhang, Juan Wu and Menghui Yao; Project Administration, Yingdong Han, Yudian Zhang, Chunyue Chen, Yun Zhang and Xuejun Zeng; Resources, Yingdong Han, Yun Zhang and Xuejun Zeng; Software, Yingdong Han, Yudian Zhang, Tiange Xie and Menghui Yao; Supervision, Yun Zhang and Xuejun Zeng; Validation, Chunyue Chen, Juan Wu; Visualization, Yingdong Han, Yudian Zhang; Writing—Original Draft Preparation, Yingdong Han, Yudian Zhang; Writing—Review & Editing, Yun Zhang and Xuejun Zeng; Funding Acquisition, Yun Zhang and Xuejun Zeng. All authors read and approved the final manuscript.
Funding
This work was supported by the National Natural Science Foundation of China (Grant No.82071841), Beijing Key Clinical Specialty Program and Peking Union Medical College Hospital Talent Cultivation Program (Category C) No. UBJ10806.
Data availability
Data are available to other researchers by applying to the UK Biobank (https://www.ukbiobank.ac.uk/). Analytic methods will be made available to other researchers upon request.
Declarations
Ethics approval and consent to participate
The UK Biobank received ethical approval from the North West Multi-Centre Research Ethics Committee (Manchester, U.K.). All participants gave informed consent at recruitment. We declared that this study followed the Declaration of Helsinki.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yingdong Han and Yudian Zhang are co-first authors.
Contributor Information
Yun Zhang, Email: zhangyun10806@pumch.cn.
Xuejun Zeng, Email: zxjpumch@126.com.
References
- 1.Singer M, Angus DC, Annane D, Bauer M, Kalil AC, Klompas M, et al. Sepsis. Lancet. 2026;407(10535):1276–88. [DOI] [PubMed] [Google Scholar]
- 2.GBD 2021 Global Sepsis Collaborators. 2025 Global, regional, and national sepsis incidence and mortality, 1990–2021: a systematic analysis. Lancet Glob Health. 2025;13(12):e2013–26. [DOI] [PubMed] [Google Scholar]
- 3.Meyer NJ, Prescott HC. Sepsis and septic shock. N Engl J Med. 2024;391(22):2133–46. [DOI] [PubMed] [Google Scholar]
- 4.Wang Y, Zhang H, Miao C. Unraveling immunosenescence in sepsis: from cellular mechanisms to therapeutics. Cell Death Dis. 2025;16(1):393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zador Z, Landry A, Cusimano MD, Geifman N. Multimorbidity states associated with higher mortality rates in organ dysfunction and sepsis: a data-driven analysis in critical care. Crit Care. 2019;23(1):247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Prescott HC, Antonelli M, Alhazzani W, Møller MH, Alshamsi F, Azevedo LCP, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2026. Crit Care Med. 2026;54(4):725–812. [DOI] [PubMed] [Google Scholar]
- 7.Deinhardt-Emmer S, Chousterman BG, Schefold JC, Flohé SB, Skirecki T, Kox M, et al. Sepsis in patients who are immunocompromised: diagnostic challenges and future therapies. Lancet Respir Med. 2025;13(7):623–37. [DOI] [PubMed] [Google Scholar]
- 8.Putot A, Prendki V. New horizons in sepsis management in older patients. Age Ageing. 2023. 10.1093/ageing/afad016. [DOI] [PubMed] [Google Scholar]
- 9.Yanai H, Adachi H, Hakoshima M, Katsuyama H. Molecular biological and clinical understanding of the pathophysiology and treatments of hyperuricemia and its association with metabolic syndrome, cardiovascular diseases and chronic kidney disease. Int J Mol Sci. 2021. 10.3390/ijms22179221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Joosten LAB, Crişan TO, Bjornstad P, Johnson RJ. Asymptomatic hyperuricaemia: a silent activator of the innate immune system. Nat Rev Rheumatol. 2020;16(2):75–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhou M, Aziz M, Wang P. Damage-associated molecular patterns as double-edged swords in sepsis. Antioxid Redox Signal. 2021;35(15):1308–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Galley HF, Davies MJ, Webster NR. Xanthine oxidase activity and free radical generation in patients with sepsis syndrome. Crit Care Med. 1996;24(10):1649–53. [DOI] [PubMed] [Google Scholar]
- 13.Poulsen R, Dalbeth N. Gout and NLRP3 inflammasome biology. Arthritis Rheumatol. 2025;77(10):1317–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ma M, Jiang W, Zhou R. DAMPs and DAMP-sensing receptors in inflammation and diseases. Immunity. 2024;57(4):752–71. [DOI] [PubMed] [Google Scholar]
- 15.Liu S, Zhong Z, Liu F. Prognostic value of hyperuricemia for patients with sepsis in the intensive care unit. Sci Rep. 2022;12(1):1070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Jiang YX, Gong CL, Tang Y, Yi Y, Liu FG, Zhou JW, et al. Association between hyperuricemia and acute kidney injury in critically ill patients with sepsis. BMC Nephrol. 2023;24(1):128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Alshehri AM, Alrashed M, Shawaqfeh M, Almutairi F, Alanazi A, Alfaifi M, et al. Impact of hyperuricemia on clinical outcomes in sepsis patients: a retrospective cohort study. J Clin Med. 2024. 10.3390/jcm13216548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hamilton FW, Thomas M, Arnold D, Palmer T, Moran E, Mentzer AJ, et al. Therapeutic potential of IL6R blockade for the treatment of sepsis and sepsis-related death: a Mendelian randomisation study. PLoS Med. 2023;20(1):e1004174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zou M, Lu D, Luo Z, Huang N, Wang W, Zhuang Z, et al. Association of healthy sleep patterns with incident sepsis: a large population-based prospective cohort study. Crit Care. 2025;29(1):51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Han Y, Chen C, Yao M, Wu J, Xie T, Zhang Y, et al. Association of vitamin D with sepsis incidence and mortality: a large population-based prospective cohort study. Crit Care. 2026. 10.1186/s13054-026-05977-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Zekavat SM, Lin SH, Bick AG, Liu A, Paruchuri K, Wang C, et al. Hematopoietic mosaic chromosomal alterations increase the risk for diverse types of infection. Nat Med. 2021;27(6):1012–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Nan W, Huang Q, He B, Liu Y, Peng Z. Adherence to the EAT-lancet diet and risk of sepsis: a prospective cohort study from the UK Biobank. NPJ Sci Food. 2026;10(1):153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.You S, Xing Z, Hu J. Association of TyG index with sepsis incidence and mortality: a prospective study with diabetes stratification. Front Endocrinol (Lausanne). 2026;17:1834832. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Qin Y, Yang X, Ning Z. Causal roles of serum uric acid levels and gout in sepsis: a mendelian randomization study. Shock. 2024;62(1):44–50. [DOI] [PubMed] [Google Scholar]
- 26.Hu L, Hu G, Xu BP, Zhu L, Zhou W, Wang T, et al. U-shaped association of serum uric acid with all-cause and cause-specific mortality in US adults: a cohort study. J Clin Endocrinol Metab. 2020. 10.1210/clinem/dgz068. [DOI] [PubMed] [Google Scholar]
- 27.Chen R, Pang M, Zhang Y, Zhang H, Zhang G, Wei Y, et al. Associations of serum uric acid, risk of atherosclerotic cardiovascular disease, and mortality: results from NHANES. Eur J Med Res. 2025;30(1):283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Virdis A, Masi S, Casiglia E, Tikhonoff V, Cicero AFG, Ungar A, et al. Identification of the uric acid thresholds predicting an increased total and cardiovascular mortality over 20 years. Hypertension. 2020;75(2):302–8. [DOI] [PubMed] [Google Scholar]
- 29.Wang M, Wu J, Jiao H, Oluwabiyi C, Li H, Zhao J, et al. Enterocyte synthesizes and secrets uric acid as antioxidant to protect against oxidative stress via the involvement of Nrf pathway. Free Radic Biol Med. 2022;179:95–108. [DOI] [PubMed] [Google Scholar]
- 30.Lee TS, Lu TM, Chen CH, Guo BC, Hsu CP. Hyperuricemia induces endothelial dysfunction and accelerates atherosclerosis by disturbing the asymmetric dimethylarginine/dimethylarginine dimethylaminotransferase 2 pathway. Redox Biol. 2021;46:102108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Dempsey B, da Silva BP, Cruz LC, Vileigas D, Silva ARM, da Silva RP, et al. Unraveling the effects of uric acid on endothelial cells: a global proteomic study. Redox Biol. 2025;82:103625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ma Q, Immler R, Pruenster M, Sellmayr M, Li C, von Brunn A, et al. Soluble uric acid inhibits β2 integrin-mediated neutrophil recruitment in innate immunity. Blood. 2022;139(23):3402–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Du L, Zong Y, Li H, Wang Q, Xie L, Yang B, et al. Hyperuricemia and its related diseases: mechanisms and advances in therapy. Signal Transduct Target Ther. 2024;9(1):212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zhang P, Wang R, Qu Y, Guo ZN, Zhen Q, Yang Y. Serum uric acid levels and outcome of acute ischemic stroke: a dose-response meta-analysis. Mol Neurobiol. 2024;61(3):1704–13. [DOI] [PubMed] [Google Scholar]
- 35.Mira JC, Gentile LF, Mathias BJ, Efron PA, Brakenridge SC, Mohr AM, et al. Sepsis pathophysiology, chronic critical illness, and persistent inflammation-immunosuppression and catabolism syndrome. Crit Care Med. 2017;45(2):253–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Gonsales Spindola D, Bahr A, Clark S, Pin de Jesus G, Martino N, Lowery A, et al. Endothelial cells retain inflammatory memory through chromatin remodeling in a two-hit model of infection-induced inflammation. Am J Physiol Cell Physiol. 2025;329(6):C1994-c2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ayesh H, Nasser SA, Ferdinand KC, Carranza Leon BG. Sex-specific factors influencing obesity in women: bridging the gap between science and clinical practice. Circ Res. 2025;136(6):594–605. [DOI] [PubMed] [Google Scholar]
- 38.Nappi RE, Chedraui P, Lambrinoudaki I, Simoncini T. Menopause: a cardiometabolic transition. Lancet Diabetes Endocrinol. 2022;10(6):442–56. [DOI] [PubMed] [Google Scholar]
- 39.Chakraborty B, Byemerwa J, Krebs T, Lim F, Chang CY, McDonnell DP. Estrogen receptor signaling in the immune system. Endocr Rev. 2023;44(1):117–41. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data are available to other researchers by applying to the UK Biobank (https://www.ukbiobank.ac.uk/). Analytic methods will be made available to other researchers upon request.