Intracerebral hemorrhage (ICH) is a stroke subtype with high morbidity and mortality (1). For patients with spontaneous superficial cerebral hemorrhage without intraventricular hemorrhage, early surgical intervention may confer a clinically relevant survival advantage (2). Owing to its poor prognosis, many researchers are dedicated to optimizing the treatment strategy for ICH (3–5). Early neurological deterioration (END), with an incidence of approximately 25.4%, is considered to be associated with poor prognosis among patients with ICH (6, 7). Therefore, identifying high-risk stroke patients with END is crucial for neurologists. The current methods for identifying high-risk individuals depend largely on the assessment of the clinical and neuroimaging features present at the time of onset. In recent years, researchers have focused more on END in ischemic stroke (8–10). It was reported that the maximum body temperature and fever burden within 24 h after endovascular thrombectomy were independent predictors of END in ischemic stroke patients with large vessel occlusion (11). Fever occurs in approximately 40% of patients with ICH and is independently associated with a poor outcome and increased mortality (8). Therefore, the management of fever in patients with cerebral hemorrhage is particularly important. Recently, the INTERACT-3 trail findings revealed that the implementation of the care bundle procedure, including intensive blood pressure lowering, strict glucose control, antipyrexia treatment and rapid reversal of warfarin-related anticoagulation within a few hours after the onset of symptoms can improve the functional outcome of patients with acute cerebral hemorrhage (12). Previous studies have confirmed that the body temperature is associated with END in patients with ischemic stroke after endovascular thrombectomy. Moreover, Leira et al. (6) found that a body temperature of >37.5°C was an independent predictor of END in patients with ICH. However, it remains unclear whether fever is associated with END in patients with ICH who are undergoing surgery. Therefore, this study explored the possible relationship between body temperature and END in patients with ICH after hematoma evacuation (HE).

Between January 2017 and April 2022, 2,726 patients with spontaneous cerebral hemorrhage were admitted to the Second Affiliated Hospital of Fujian Medical University. We identified 372 patients with ICH who had undergone HE. Among them, 32 patients were excluded due to secondary ICH, 9 patients were excluded due to hemorrhagic transformation of ischemic infarction, and 23 patients were excluded due to a GCS score of less than 5 at admission. Finally, 308 patients were included in the analysis (As shown in Figure 1). Twelve patients were lost to follow-up, since it did not affect the evaluation of the in-hospital outcomes; therefore, we included these 12 patients in the study. Fever occurred in 219 patients (71.1%) within 24 h post-operation. The baseline data and clinical outcomes of the patients are shown in Table 1.

The number of women (56 vs. 35, p = 0.016) and white blood cell count [8.80 (6.83–10.82) vs. 8.13 (6.19–10.29); p = 0.007] in the fever group were higher than those in the non-fever group, but age [51 (43–60) vs. 55 (50–64); p = 0.005] was lower. We plotted the body temperature distributions of patients with and without END at admission and at 6, 12, and 24 h after HE (As shown in Figure 2). In the multivariate logistic regression model, we adjusted for age (p = 0.553), sex (p = 0.041), systolic blood pressure at admission (p = 0.265), presence of hypertension (p = 0.944), hematoma volume (p = 0.872), and presence of intraventricular hemorrhage (p = 0.959) since these confounding factors were found to be associated with END in previous studies. As shown in Table 2, the burden of fever (OR = 1.055 per °C × h, 95% CI 1.008–1.103, p = 0.020) remained independently associated with END after adjusting for possible confounding factor. Fever burden was 12.17 ± 16.19 (median:5.2) in END group, and 6.29 ± 8.85 (median:2.45) in non-END group. ROC analysis was performed to determine the value of fever burden in predicting END. The cutoff value in the ROC curve (17.15°C × h) discriminated END (+)/END (−) patients with a sensitivity of 31.7%, specificity of 88.5%, PPV of 50%, and NPV of 78.1%. The area under ROC curve was 0.590 (95% CI, 0.514–0.666; Figure 3). Univariate logistic analysis showed that no factor was associated with poor functional outcomes 3 months after discharge (Supplementary Table S1).

END after ICH is associated with poor prognosis of death and major disability (7, 17). Therefore, it is important to identify the early clinical features of END. Previous studies have confirmed that blood pressure (18), uric acid (19), plasma leptin level (20), serum annexin A7 (21), and hematoma volume (17) can predict the occurrence of END after ICH. In this single-center retrospective study, END occurred in 26.6% patients. We found that fever burden was independently associated with the occurrence of END after HE in patients with ICH. Our results might help neurologists identify and deal with END promptly.

While hyperthermia is recognized as being associated with poor prognosis in patients with ICH (22, 23), discrepancies exist in the strategies for early postoperative temperature management following ICH (24, 25). Typically, a standard treatment regimen is employed in response to fever, encompassing both pharmacological and physical cooling measures. However, the effectiveness of these methods is hindered by the absence of a defined target temperature as the optimal temperature threshold remains unclear (26, 27). Our study provides a useful clinical insight: by monitoring the “fever burden” (the aggregate of the patient’s body temperature over time) within 24 h after surgery, clinicians may obtain an indicator for the potential occurrence of END.

The concept of fever burden was proposed in the 1990s and is defined as the product of fever and its duration. It has been used as a parameter to evaluate the effectiveness of cooling therapy in patients with neurological disorders (28). In previous studies, researchers found that body temperature > 37.5°C is an independent predictor of END (6). However, when the body temperature fluctuates greatly, or the peak temperature lasts for a prolonged period, the fever burden, which combines time and peak effects may better represent fever. Our results suggest that, when dealing with such patients, neurologists should pay attention not only to the highest fever temperature and/or duration of fever but also notice the combined indicators of the two aforementioned factors.

Infections have been associated with worse outcomes in patients with ICH (29) and might indirectly contribute to the onset of END by raising the patient’s body temperature. A retrospective study identified a link between infections in patients with ICH and neurological deterioration between 3 and 15 days post-ICH (30). In our study, most patients were admitted to the emergency department immediately following ICH onset and received emergent HE. Infections leading to fever within the first 24 h postoperatively were rare in this patient group. As a result, we did not include infections in our initial analysis. Future research is needed to investigate whether infections are associated with END within 24 h post-HE.

Previous research has confirmed that the fever burden is associated with END after ischemic stroke (11). In patients with ICH, our study showed that fever burden is still associated with END after adjusting for potential confounding factors. A few mechanisms may explain our results. First, central fever regulated by the hypothalamus is the main cause of early fever after cerebral hemorrhage. Hyperthermia increases brain oxygen consumption and deprives neurons of energy reserves. Neurons are significantly vulnerable to hypoxia. Increased neuronal oxygen consumption results in compensatory glycolysis, which leads to lactate accumulation and cellular acidosis (22, 23). Energy failure and subsequent ATP-dependent ion channel dysfunction cause water to enter the cells, eventually leading to cellular swelling (24). Second, in addition to the direct effect of hyperthermia on brain cells, brain hyperthermia appears to affect the permeability of the blood-brain barrier (BBB) (25), which increases edema around the hematoma (26) and even causes hematoma expansion. Existing studies confirm hematoma expansion as one of the causes of END (1, 27, 29). Finally, a damaged BBB exacerbates the inflammatory response, leading to the activation of central and peripheral immune cells, which then release more molecules, including ROS, proteases and pro-inflammatory cytokines, leading to pathological immune reaction (30, 31). We plotted the possible mechanisms by which increased body temperature leads to END in Figure 4. However, the mechanism of END in ICH is heterogeneous, and further laboratory studies are warranted to determine the exact mechanism by which fever causes END.

Our study had several limitations. First, owing to the difficulty in implementing real-time temperature monitoring for each patient, body temperature was measured with a 6-h interval scheme, which means that the real maximum or minimum body temperature may not be collected. Therefore, future studies should focus on this issue. Second, our method of temperature measurement presents another limitation. Axillary temperature measurements, while convenient, may not accurately reflect core body temperature. This discrepancy can result in a less reliable measure of the patient’s true physiological state. To overcome this limitation, future studies should consider employing methodologies that allow for more accurate core body temperature monitoring. Third, the GCS is not widely used in the diagnosis of END, and future studies should perhaps use the NIHSS score to define END. Fourth, the area under the ROC curve of the fever burden in this study was relatively small (0.590) and the sensitivity is relatively low (31.7%), which may compromise the predictive value of the fever burden for END. Fifth, this study was conducted at a single-center with a limited sample size. The relatively small sample size may have compromised the power of our primary results. Finally, our study was conducted retrospectively; thus, any conclusions drawn are subject to the limitations of the respective study designs, including recall and observer bias. Future multicenter prospective studies are warranted to address these issues.

Fever burden is associated with END in patients with ICH undergoing HE. Our study provides more data on the relationship between the fever burden and the occurrence of early neurological deterioration in patients with intracerebral hemorrhage. Future studies with larger sample sizes are required to confirm these findings.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by the Ethics Committee of the Second Affiliated Hospital of Fujian Medical University. Written informed consent from the patients/participants or patients/participants’ legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.

H-fH, W-pH, and FZ conceived and designed the research. FW, S-lH, X-hW, and X-lC collected the data and performed the research. FW and YX processed and analyzed the data. All authors contributed to the article and approved the submitted version.