Adipose markers hint at neurological recovery after in-hospital cardiac arrest

Two independent cohorts of patients who experienced in-hospital cardiac arrest suggest that adipose tissue characteristics may carry information about neurological recovery at hospital discharge. In a retrospective imaging study, thicker posterior subcutaneous adipose tissue was associated with favorable outcomes in some analyses, but the overall association weakened after correction for multiple comparisons. In a separate biochemical cohort, patients with favorable neurological outcomes had higher circulating concentrations of CTRP6, CTRP9, Omentin-1 and Vaspin, together with lower Chemerin. The findings are exploratory: the two measures were not collected from the same patients, and the investigators cannot conclude that a particular adipose depot caused better recovery.

The research was conducted at West China Hospital in Chengdu, China, and reported in Frontiers in Neurology on October 20, 2026. It combined a CT Imaging Cohort of 70 adults who achieved sustained return of spontaneous circulation and a separate Biochemical Cohort of 66 patients whose arterial blood was sampled at the onset of cardiac arrest. The imaging and biochemical groups did not overlap in time or participants. Consequently, the study provides parallel anatomical and molecular observations rather than a multimodal model in which CT measurements and blood biomarkers can be linked directly.

Neurological status was evaluated at hospital discharge using the Cerebral Performance Category scale. Patients classified as CPC 1 or 2 were considered to have a favorable neurological outcome, whereas CPC 3 through 5 indicated an unfavorable outcome. In the CT cohort, 32 of 70 patients had favorable outcomes and 28 had unfavorable outcomes after 12 patients with pre-existing neurological impairment were excluded from the neurological analysis. The biochemical cohort included 24 patients with favorable outcomes and 42 with unfavorable outcomes. Survival to discharge was 80.0% in the imaging cohort and 62.1% in the biochemical cohort, while the proportions with favorable neurological outcomes were 45.7% and 36.4%, respectively.

For the imaging analysis, multidetector CT scans taken before and after resuscitation were reconstructed in three-dimensional software. Researchers measured anterior and posterior subcutaneous adipose tissue at the T3 through T12 vertebral levels, together with transverse diameter, internal anteroposterior diameter, cross-sectional area and the Haller index, a ratio describing chest shape. Posterior adipose thickness was defined as the distance from the tip of each spinous process to the skin surface, while anterior thickness extended from the sternum to the skin. Two trained assessors made the measurements, and blinded repeat assessments in 10% of the imaging cohort indicated high inter-rater reliability for adipose thickness and most chest dimensions.

In the overall imaging analysis, neither anterior nor posterior subcutaneous adipose thickness was significantly associated with survival. Unadjusted comparisons of post-arrest posterior tissue did suggest greater thickness at several vertebral levels among patients with favorable neurological outcomes, but none remained significant after Benjamini–Hochberg false discovery rate correction across the ten levels. The corresponding adjusted q-values ranged from 0.125 to 0.386, compared with a prespecified significance threshold of 0.05. A sensitivity analysis restricted to patients with the same ventilation status during both CT examinations produced a different pattern: greater posterior thickness was associated with favorable outcomes at several levels, including T8 through T12 after correction. Because this subgroup was small and the analysis was exploratory, the result requires replication rather than a causal interpretation.

The biochemical cohort tested 12 preselected adipokines using enzyme-linked immunosorbent assays performed by technicians blinded to clinical outcomes. Arterial blood was collected as cardiopulmonary resuscitation began, before return of spontaneous circulation or termination of resuscitation, and plasma was stored at −80°C. After correction across all 12 comparisons, CTRP6, CTRP9, Omentin-1 and Vaspin concentrations were higher in patients with favorable outcomes, while Chemerin was lower. Median CTRP6 concentration was 4.01 ng/mL in the favorable group compared with 3.13 ng/mL in the unfavorable group, with a false discovery rate-adjusted q-value below 0.001. Median CTRP9 values were 192.80 versus 127.78 ng/mL, and mean Vaspin values were 8.98 versus 7.43 ng/mL. Chemerin averaged 493.22 versus 623.87 ng/mL.

CTRP6, a member of the C1q/TNF-related protein family, was then selected for exploratory prediction modeling because it had the highest apparent individual discrimination among the five biomarkers that met the false discovery rate threshold. This selection was made after examining the data and was not prespecified. All 66 patients had complete data for the predictors. In a clinical logistic regression model, each additional minute of cardiopulmonary resuscitation was associated with lower odds of a favorable outcome, with an odds ratio of 0.877 and a 95% confidence interval of 0.813 to 0.946. Higher lactate was also associated with lower odds, with an odds ratio of 0.816 per mmol/L. The apparent area under the receiver operating characteristic curve for this two-variable model was 0.876, with a 95% confidence interval of 0.795 to 0.957.

Adding CTRP6 produced an apparent AUC of 0.938, with a 95% confidence interval of 0.882 to 0.993, and an optimism-corrected AUC of 0.930 after 1,000 bootstrap resamples. Within the extended model, each additional nanogram per milliliter of CTRP6 was associated with higher odds of a favorable neurological outcome, with an odds ratio of 8.880 and a wide 95% confidence interval of 2.179 to 36.187. Longer resuscitation remained inversely associated with a favorable outcome. A Firth regression sensitivity analysis preserved these directions, although the estimated CTRP6 odds ratio fell to 6.433. Importantly, the numerical AUC gain of 0.062 over the clinical model was not statistically significant: its 95% confidence interval ranged from −0.003 to 0.126, and the paired DeLong test gave P = 0.062.

Several design features limit the conclusions. The study was retrospective and single-center, and requiring paired CT scans selected a relatively small subgroup of arrest survivors. The biochemical cohort excluded patients with cachexia, end-stage chronic disease, hematological malignancy, traumatic arrest, neurological causes of arrest and out-of-hospital arrest, further narrowing its relevance. Detailed targeted-temperature-management protocols, extracorporeal CPR information, ventilator settings and withdrawal-of-life-sustaining-treatment data were unavailable. The predictive model had only eight events per variable, increasing concern about overfitting, and the data-informed choice of CTRP6 was not repeated during bootstrapping. In addition, blood concentrations were measured at one time point, adipose thickness may be affected by intubation, fluid status and CT timing, and visceral adipose tissue was not assessed.

The results nevertheless provide a testable hypothesis for larger prospective studies: regional adipose anatomy and the endocrine profile of circulating adipokines may each contain information relevant to recovery after in-hospital cardiac arrest. Future work would need concurrent CT imaging and serial blood sampling, systematic documentation of post-resuscitation treatment, and external validation in broader populations. The current data do not establish an “adipose–brain axis,” show that higher adipokines are neuroprotective, or justify using CTRP6 in clinical prognosis. They also cannot determine whether thicker posterior adipose tissue is a marker of metabolic reserve, a consequence of resuscitation and hospitalization, or merely a chance finding. For now, the clearest conclusion is that adipose-related measures deserve further study, while CTRP6 remains a candidate biomarker rather than a clinically deployable predictor.

Subject of Research: Exploratory associations of adipose tissue thickness and adipokines with neurological outcomes after in-hospital cardiac arrest

Article Title: CT-derived subcutaneous adipose tissue thickness and circulating adipokines in relation to neurological outcome after in-hospital cardiac arrest: two parallel cohort analyses

Article References: Zhou, T., Ma, W., He, Y., Tang, S., Yao, P., Gan, L., & Cao, Y. (2026). CT-derived subcutaneous adipose tissue thickness and circulating adipokines in relation to neurological outcome after in-hospital cardiac arrest: two parallel cohort analyses. Frontiers in Neurology, 17, Article 1898760. https://doi.org/10.3389/fneur.2026.1898760

Image Credits: AI Generated

DOI: 10.3389/fneur.2026.1898760

Keywords: in-hospital cardiac arrest, neurological outcome, adipokines, CT imaging, adipose tissue, CTRP6, prognostic biomarkers, hypoxic–ischemic brain injury, CT-derived, subcutaneous, adipose, tissue