Acute ischemic stroke (AIS) significantly contributes to severe disability and mortality among the elderly. This study aims to explore the association between longitudinal fluid balance (FB) trajectories and clinical outcomes in elderly patients with AIS. Our hypothesis posits the existence of multiple latent trajectories of FB in patients with AIS during the initial 7 days following ICU admission.
Methods
Patients (age ≥ 65 years) with AIS and continuous FB records were identified from two large databases. Group-based trajectory modeling identified latent groups with similar 7-day FB trajectories. Subsequently, multivariable logistic and quasi-Poisson regression were employed to evaluate the relationship between trajectory groups and outcomes. Additionally, nonlinear associations between maximum fluid overload (FO) and outcomes were analyzed using restricted cubic spline models. To further validate our findings, subgroup and sensitivity analysis were conducted.
Results
A total of 1146 eligible patients were included in this study, revealing three trajectory patterns were identified: low FB (84.8%), decreasing FB (7.2%), and high FB (7.9%). High FB emerged as an independent risk factor for in-hospital mortality. Compared with those without FO, patients with FO had a 1.57-fold increased risk of hospital mortality (adjusted odd ratio (OR) 1.57, 95% confidence interval (CI) 1.08–2.27), 2.37-fold increased risk of adverse kidney event (adjusted OR 2.37, 95% CI 1.56–3.59), and 1.33-fold increased risk of prolonged ICU stay (adjusted incidence rate ratio (IRR) 1.33, 95% CI 1.19–1.48). The risk of hospital mortality and adverse kidney event increased linearly with rising maximum FO (P for non-linearity = 0.263 and 0.563, respectively).
Conclusion
Daily FB trajectories were associated with adverse outcomes in elderly patients with AIS. Regular assessment of daily fluid status and restriction of FO are crucial for the recovery of critically ill patients.
Various factors intensify the complexity in managing fluids for older patients with acute ischemic stroke.
This study aimed to explore the association between longitudinal fluid balance trajectories and clinical outcomes in this population.
Persistent high fluid balance trajectory was a risk factor for adverse outcomes in elderly patients with acute ischemic stroke.
Regular assessment of daily fluid status and restriction of fluid overload are crucial for the recovery of critically ill patients.
Introduction
Acute ischemic stroke (AIS) is a critical medical condition characterized by the sudden interruption of blood supply to a segment of the brain, leading to insufficient oxygen and nutrients delivery to the affected cerebral tissue and subsequent neuronal damage [1]. It stands as a significant cause of severe disability and mortality, particularly among the elderly. Annually, stroke affects 26 million individuals worldwide, ranking as the second leading cause of death [2].
Anzeige
Fluid balance (FB), the equilibrium between fluid intake and output in the body, plays a pivotal role in sustaining vital functions. In intensive care units, maintaining FB is a fundamental aspect of supportive care for critically ill patients. Patients with AIS often exhibit symptoms such as dysphagia and altered consciousness, resulting in inadequate fluid intake [3]. In older adults, the perception of thirst diminishes, and the kidney’s ability to concentrate urine declines [4]. All these factors may result in dehydration, hemoconcentration, and even thrombosis [5] exacerbating cerebral ischemia. Additionally, elderly patients frequently contend with conditions like acute and chronic heart failure, leading to lower cerebral perfusion pressure. Such factors intensify the complexity in managing fluids for older patients with AIS. A previous study showed that patients aged 85 years and older with AIS had a worse outcome (higher in-hospital mortality and more severe neurological deficit) than younger patients [6]. Consequently, timely assessment of FB and adjustment of it are crucial for improving outcomes.
In clinical practice, as a result of the variability of conditions, not all patients consistently follow prescribed fluid strategies, whether liberal or restrictive [7]. Categorizing patients based solely on single strategy is inaccurate and inappropriate. To address this limitation, our study employs group-based trajectory modeling (GBTM), a method that more accurately classifies patients by continuously monitoring their FB patterns. GBTM allows for the identification of subgroups within a population based on similar trajectories over time. This approach not only enhances the precision of patient categorization but also provides insights into the potential mechanisms linking fluid strategies to outcomes [8]. While FB assessment is crucial in patients’ management, relying solely on it for prognosis is insufficient. Considering fluid overload (FO), defined as a 10% increase in cumulative FB from baseline weight, is also imperative [9]. Early fluid resuscitation effectively enhances tissue perfusion, thereby improving outcomes [10, 11]. However, the lack of efficient indicators for fluid responsiveness often leads to the accumulation of positive FB or FO [12, 13]. Studies indicate that FO is an independent predictor for morbidity and increased hospital costs in critically ill patients [14, 15]. A recent study on patients with septic shock demonstrated that patients with FO faced a 1.4 times higher risk of hospital mortality compared to those without [16].
Current research on the dynamic changes in FB or cumulative FB and their correlation with the prognosis of elderly patients with stroke is limited. The use of group-based trajectory models holds the potential to uncover latent patterns and associations, addressing this research gap. This study aimed to explore the association between longitudinal FB trajectories and clinical outcomes in critically ill patients with AIS. We hypothesized that the trajectories of FB in patients with AIS can be classified into several patterns, some of which may be associated with better recovery, while others may indicate poor prognosis.
Methods
Data Source and Selection of Participants
The data for this study were extracted from two databases, namely the Medical Information Mart for Intensive Care IV (MIMIC-IV) database (version 2.0) and eICU Collaborative Research Database (eICU-CRD, version 2.0). MIMIC-IV database is an openly accessible, single-center repository comprising 76,943 records from 2008 to 2019 across multiple intensive care units (ICUs) at Beth Israel Deaconess Medical Center. The eICU-CRD is a multicenter database that encompasses over 200,000 ICU admissions across the USA during 2014–2015. This research adhered to the ethical principles outlined in the Declaration of Helsinki and gained endorsement from the ethics evaluation committees of both Massachusetts Institute of Technology and Beth Israel Deaconess Medical Center (researcher certification number 52759164). This study focused on elderly patients (≥ 65 years old) with AIS who spent three or more days in the ICUs. Those lacking admission weight data or continuous fluid records for the first 3 days in the ICUs were excluded. Records with only output data and no input data within each 24-h period were also considered incomplete and excluded. In cases of multiple ICU admissions, only data from the first ICU admission were considered.
Anzeige
Data Collection
Baseline characteristics, including age, sex, weight, mean arterial pressure, laboratory results (sodium, potassium, prothrombin time, and international normalized ratio), drug use (thrombolysis, antiplatelet agent, anticoagulation agent, diuretic, angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker, beta-blocker, calcium channel blocker, and sodium nitroprusside), mechanical ventilation, renal replacement therapy, decompression, and comorbidities (congestive heart failure, atrial fibrillation, chronic kidney disease, hepatic dysfunction, seizures, aspiration and sepsis) were extracted for patients with AIS. The severity of illness was assessed using the Acute Physiology Score III (APSIII). Variables with more than 20% missing data were excluded, and the remaining ones were imputed using the multiple imputation [17].
Longitudinal fluid input and output data from day 1 to 7 were recorded. In our study, insensible loss was not considered because of the challenges in its accurate assessment. FB was determined using the following formula: FB = (total fluid input − total fluid output) mL/weight at admission (kg). FB values outside the interval [Q3 + 3 IQR, Q1–3 IQR] were identified as extreme outliers and replaced by the endpoint values [18]. Cumulative FB was calculated per 24 h. FO was defined as cumulative FB (in liters) exceeding 10% of initial body weight. Maximum FO represented the highest FO over the first 7 days, reported as a percentage. The primary outcome was in-hospital mortality, with secondary outcomes including adverse kidney event and ICU stay time. Adverse kidney event was defined as new onset acute kidney injury (AKI). Data was obtained using Navicat 16 for PostgreSQL through Structured Query Language (SQL).
Group-Based Trajectory Model
Group-based trajectory model (GBTM) selection involved estimating models with latent classes, ranging from 1 to 7 classes, and comparing parameters using Bayesian information criteria (BIC), Akaike information criterion (AIC), likelihood, average post-test grouping probability (AvePP), relative entropy, and sample size (< 5% were not considered clinically meaningful) [19]. The model with the smallest BIC, AIC, and likelihood value, along with higher AvePP (> 0.7) and relative entropy was selected. We also considered the model’s complexity and clinical rationality. Differences among FB groups were compared using one-way analysis of variance (ANOVA), followed by Bonferroni post hoc test. Longitudinal trajectories of daily FB and cumulative FB during the first 7 days were presented in box plots.
Statistical Analysis
Patients were divided into several groups based on the optimal latent trajectories. The Shapiro–Wilk normality test was applied to assess all continuous variables. Normally distributed variables were presented as mean and standard deviation (SD) while non-normally distributed variables were represented by median and interquartile range (IQR). Categorical variables were expressed as numerical values and percentages (%). Group distinctions was evaluated using the t test or Wilcoxon rank-sum test for continuous variables and the chi-square test for categorical variables. Cumulative risk curves of in-hospital mortality were constructed and compared using the log-rank test. The relationship between latent trajectory groups or FO and outcomes was scrutinized through multivariate logistic and quasi-Poisson regression analysis. Furthermore, we analyzed the nonlinear relationship between maximum FO and outcomes using restricted cubic spline (RCS) regression models. In addition, analysis was conducted for various subgroups. Considering the potential impact of renal replacement therapy (RRT) on FB, an additional group-based trajectory model was constructed for patients without RRT. Three adjusted regression models were employed to determine the association between different trajectories and clinical outcomes in this specific cohort. Statistical significance was determined with a two-sided P value of < 0.05. R software (version 4.3.1) and Stata (version 17.0) were employed to perform the statistical analyses.
Results
Baseline Characteristics
This study enrolled a total of 1146 eligible patients, as illustrated in Fig. S1. The missing rate of baseline variables were below 10%, as depicted in Fig. S2 and Table S1. No statistically significant differences were observed in the comparison of baseline data before and after imputation. The characteristics of latent trajectory groups were presented in Table S2. The optimal one was the model with three latent classes (shape: quadratic, cubic, and cubic), of which the absolute BIC value was 28,972.21. In this model, patients were divided into three groups (Fig. 1). The level of the low FB trajectory (84.8%) was very close to 0 mL/kg during the first 7 days. Decreasing FB trajectory (7.2%) exhibited a pattern in which the subjects started with a high FB and then decreased rapidly from day 2 to day 4 and then maintained a low level in the following days. High FB trajectory (7.9%) showed high FB levels throughout the 7-day period. Patients in this group manifested the highest severity of illness (APSIII) (Table 1).
Fig. 1
Three fluid balance trajectories based on GBTM. GBTM group-based trajectory modeling
Table 1
Comparisons of baseline characteristics among three fluid balance trajectory subgroups
Characteristics
Overall
Low FB group
Decreasing FB group
High FB group
P value
n = 1146
n = 977
n = 81
n = 88
Age (years)a
76.95 [70.96, 83.57]
77.15 [71.00, 83.67]
75.18 [69.20, 83.67]
74.13 [69.10, 80.34]
0.029
Femaleb
617 (53.8)
522 (53.4)
52 (64.2)
43 (48.9)
0.109
Weight (kg)a
74.50 [63.52, 88.15]
75.40 [64.80, 90.00]
69.20 [58.80, 78.00]
69.25 [59.75, 80.00]
< 0.001
APSIII (score)a
51.00 [36.00, 70.00]
48.00 [35.00, 66.00]
66.00 [49.00, 85.00]
75.00 [57.00, 94.25]
< 0.001
MAP (mmHg)a
93.33 [79.67, 106.67]
94.33 [81.00, 107.00]
82.00 [70.67, 98.33]
90.33 [77.42, 104.75]
< 0.001
Sodium (mmol/L)a
139.00 [136.00, 142.00]
139.00 [137.00, 142.00]
139.00 [136.00, 142.00]
139.00 [137.00, 142.00]
0.929
Potassium (mmol/L)a
4.20 [3.80, 4.60]
4.10 [3.80, 4.50]
4.20 [3.80, 4.60]
4.25 [3.80, 4.85]
0.107
PT (s)a
13.10 [11.70, 15.30]
12.90 [11.60, 15.00]
14.30 [12.00, 16.60]
13.75 [12.07, 15.75]
0.003
INRa
1.20 [1.10, 1.40]
1.20 [1.10, 1.30]
1.30 [1.10, 1.50]
1.20 [1.10, 1.40]
0.001
Thrombolysisb
21 (1.8)
18 (1.8)
1 (1.2)
2 (2.3)
0.880
Antiplatelet agentb
473 (41.3)
416 (42.6)
34 (42.0)
23 (26.1)
0.011
Anticoagulation agentb
48 (4.2)
45 (4.6)
2 (2.5)
1 (1.1)
0.216
Diureticb
261 (22.8)
240 (24.6)
13 (16.0)
8 (9.1)
0.001
ACEI/ARBb
69 (6.0)
65 (6.7)
4 (4.9)
0 (0.0)
0.039
Beta-blockerb
513 (44.8)
447 (45.8)
29 (35.8)
37 (42.0)
0.194
CCBb
244 (21.3)
219 (22.4)
11 (13.6)
14 (15.9)
0.077
SNPb
18 (1.6)
13 (1.3)
2 (2.5)
3 (3.4)
0.258
MVb
497 (43.4)
396 (40.5)
55 (67.9)
46 (52.3)
< 0.001
RRTb
39 (3.4)
20 (2.0)
0 (0.0)
19 (21.6)
< 0.001
Decompressionb
7 (0.6)
6 (0.6)
0 (0.0)
1 (1.1)
0.638
CHFb
514 (44.9)
443 (45.3)
46 (56.8)
25 (28.4)
0.001
Atrial fibrillationb
537 (46.9)
462 (47.3)
37 (45.7)
38 (43.2)
0.743
CKDb
226 (19.7)
190 (19.4)
11 (13.6)
25 (28.4)
0.046
Hepatic dysfunctionb
55 (4.8)
40 (4.1)
3 (3.7)
12 (13.6)
< 0.001
Seizuresb
49 (4.3)
34 (3.5)
3 (3.7)
12 (13.6)
< 0.001
Aspirationb
94 (8.2)
75 (7.7)
11 (13.6)
8 (9.1)
0.168
Sepsisb
161 (14.0)
107 (11.0)
26 (32.1)
28 (31.8)
< 0.001
FOb
189 (16.5)
59 (6.0)
44 (54.3)
86 (97.7)
< 0.001
Maximum FO (%)a
3.37 [0.69, 7.42]
2.47 [0.24, 5.30]
10.38 [8.01, 13.73]
18.72 [15.04, 24.59]
< 0.001
FB fluid balance, IQR interquartile range, APSIII Acute Physiology Score III, MAP mean arterial pressure, PT prothrombin time, INR international normalized ratio, ACEI/ARB angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker, CCB calcium channel blocker, SNP sodium nitroprusside, MV mechanical ventilation, RRT renal replacement therapy, CHF congestive heart failure, CKD chronic kidney disease, FO fluid overload
aExpressed as median [IQR]
bExpressed as n (%)
×
Daily and Cumulative Fluid Balance Among the Three Trajectory Patterns
The three groups exhibited significantly varied levels of FB and cumulative FB at each time point. The daily FB level in the low FB group fluctuated slightly around 0 during this period. The decreasing FB group experienced the highest daily FB (72.4 mL/kg) on day 1, rapidly decreasing to 1.1 mL/kg by day 3 and maintaining negative FB from day 4 to 7. In the high FB group, levels exceeded 30 mL/kg on 2 days, with the rest between 20 and 30 mL/kg (Fig. S3A).
There was a rapid increase in the high FB group (from 28.8 mL/kg to 200.7 mL/kg) during the first 7 days after ICU admission. A slight increase in fluid accumulation was observed in the low FB group (from 7.2 mL/kg to 11.9 mL/kg). In the decreasing FB group, cumulative FB showed a slight increase in the first 3 days (from 72.4 mL/kg to 93.9 mL/kg), followed by a steady decrease in the subsequent days (from 93.9 mL/kg to 24.8 mL/kg) (Fig. S3B).
Association of FB Trajectories and FO with Clinical Outcomes
The high FB group exhibited the highest in-hospital mortality rate (51.1%), adverse kidney event occurrence (51.1%), and length of ICU stay (9.67 [6.45, 15.13]) among the three groups (Table 2). Figure 2 depicts the cumulative in-hospital mortality rate for different FB trajectories. Patients with the high FB trajectory had the highest cumulative mortality rate than those in other groups. In the logistic and quasi-Poisson regression analysis, two adjusted models were constructed: model I, adjusted only for age, sex, MAP, sodium, potassium, PT, INR, CHF, atrial fibrillation, CKD, hepatic dysfunction, seizures, aspiration, and sepsis; building upon model I, model II further adjusted for thrombolysis, antiplatelet agent, anticoagulation agent, diuretic, ACEI/ARB, beta-blocker, CCB, SNP, MV, and decompression (Table 3). The low FB trajectory was used as the reference group. The results demonstrated that high FB was an independent risk factor for in-hospital mortality in the full model (model II: OR 2.89, 95% CI 1.75–4.78, P < 0.001). High FB (model II: OR 1.97, 95% CI 1.10–3.52, P = 0.023) and decreasing FB (model II: OR 2.80, 95% CI 1.57–5.00, P < 0.001) were independent risk factors for adverse kidney event. Furthermore, the results showed that high FB (adjusted IRR in model II 1.31, 95% CI 1.13–1.51, P < 0.001) and decreasing FB (adjusted IRR in model II 1.23, 95% CI 1.05–1.42, P = 0.009) were significantly associated with prolonged ICU stays.
Table 2
Clinical outcomes of three fluid balance trajectory subgroups
Outcomes
All
Low FB group
Decreasing FB group
High FB group
P value
n = 1146
n = 977
n = 81
n = 88
Primary outcome
Hospital mortalityb
291 (25.4)
227 (23.2)
19 (23.5)
45 (51.1)
< 0.001
Secondary outcome
Adverse kidney eventb
350 (30.5)
264 (27.0)
41 (50.6)
45 (51.1)
< 0.001
ICU stay time (day)a
6.14 [4.14, 9.96]
5.80 [4.01, 9.30]
7.71 [5.75, 11.72]
9.67 [6.45, 15.13]
< 0.001
IQR interquartile range, FB fluid balance, ICU intensive care unit
aExpressed as median [IQR]
bExpressed as n (%)
Fig. 2
Cumulative incidence curves of patients with different trajectories of fluid balance
Table 3
Association between different trajectory groups and clinical outcomes
Clinical outcomes
Crude model
Model I
Model II
OR/IRR (95% CI)
P value
OR/IRR (95% CI)
P value
OR/IRR (95% CI)
P value
Hospital mortalitya
Low FB group
1 (Ref)
1 (Ref)
1 (Ref)
Decreasing FB group
1.01 (0.59–1.73)
0.964
0.86 (0.49–1.52)
0.608
0.69 (0.38–1.23)
0.207
High FB group
3.46 (2.22–5.39)
< 0.001
3.24 (2.02–5.21)
< 0.001
2.89 (1.75–4.78)
< 0.001
Adverse kidney eventa
Low FB group
1 (Ref)
1 (Ref)
1 (Ref)
Decreasing FB group
2.77 (1.75–4.38)
< 0.001
2.66 (1.51–4.69)
< 0.001
2.80 (1.57–5.00)
< 0.001
High FB group
2.83 (1.82–4.39)
< 0.001
1.69 (0.96–2.98)
0.071
1.97 (1.10–3.52)
0.023
ICU stay timeb
Low FB group
1 (Ref)
1 (Ref)
1 (Ref)
Decreasing FB group
1.42 (1.20–1.67)
< 0.001
1.29 (1.10–1.51)
0.002
1.23 (1.05–1.42)
0.009
High FB group
1.55 (1.32–1.80)
< 0.001
1.32 (1.13–1.53)
< 0.001
1.31 (1.13–1.51)
< 0.001
Model I = adjusted for (age + sex + MAP + sodium + potassium + PT + INR + CHF + atrial fibrillation + CKD + hepatic dysfunction + seizures + aspiration + sepsis)
Model II = model I + (thrombolysis + antiplatelet agent + anticoagulation agent + diuretic + ACEI/ARB + beta-blocker + CCB + SNP + MV + decompression)
FB fluid balance, OR odds ratio, IRR incidence rate ratio, CI confidence interval, ICU intensive care unit, MAP mean arterial pressure, PT prothrombin time, INR international normalized ratio, CHF congestive heart failure, CKD chronic kidney disease, ACEI/ARB angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker, CCB calcium channel blocker, SNP sodium nitroprusside, MV mechanical ventilation
aOR for binary outcomes (hospital mortality and adverse kidney event)
bIRR for count data outcome (ICU stay time)
×
Anzeige
A total of 189 (16.5%) patients developed FO during the first 7 days after ICU admission. Compared with those without FO, patients with FO had a 1.57-fold increased risk of hospital mortality (adjusted OR in model II 1.57, 95% CI 1.08–2.27, P = 0.018), 2.37-fold increased risk of adverse kidney event (adjusted OR in model II 2.37, 95% CI 1.56–3.59, P < 0.001), and 1.33-fold increased risk of prolonged ICU stays (adjusted IRR 1.33, 95% CI 1.19–1.48). For every 5% increase in maximum FO, the risk of in-hospital mortality is 1.3 times higher, the risk of an adverse kidney event is 1.4 times higher, and the risk of extended ICU stay is 1.1 times higher (Table 4). The restricted cubic spline regression model revealed that the risk of in-hospital mortality increased linearly with increasing maximum FO (P for non-linearity = 0.263). The risk of in-hospital mortality increased significantly when the maximum FO exceeded 3.6% (Fig. 3a). The risk of adverse kidney event increased linearly with increasing maximum FO (P for non-linearity = 0.563). When the maximum FO approached 4.2%, the risk of adverse kidney event became statistically significant (Fig. 3b).
Table 4
Association between fluid overload and clinical outcomes
Clinical outcomes
Crude model
Model I
Model II
OR/IRR (95% CI)
P value
OR/IRR (95% CI)
P value
OR/IRR (95% CI)
P value
Hospital mortalitya
FO (per 5%)
1.36 (1.23–1.50)
< 0.001
1.33 (1.20–1.48)
< 0.001
1.27 (1.13–1.41)
< 0.001
FO ≤ 10%
1 (Ref)
1 (Ref)
1 (Ref)
FO > 10%
2.02 (1.45–2.81)
< 0.001
1.83 (1.29–2.61)
< 0.001
1.57 (1.08–2.27)
0.018
Adverse kidney eventsa
FO (per 5%)
1.41 (1.28–1.55)
< 0.001
1.29 (1.14–1.47)
< 0.001
1.39 (1.22–1.60)
< 0.001
FO ≤ 10%
1 (Ref)
1 (Ref)
1 (Ref)
FO > 10%
2.71 (1.97–3.73)
< 0.001
2.05 (1.37–3.06)
< 0.001
2.37 (1.56–3.59)
< 0.001
ICU stay timeb
FO (per 5%)
1.14 (1.10–1.17)
< 0.001
1.10 (1.06–1.13)
< 0.001
1.09 (1.06–1.13)
< 0.001
FO ≤ 10%
1 (Ref)
1 (Ref)
1 (Ref)
FO > 10%
1.54 (1.37–1.72)
< 0.001
1.35 (1.20–1.51)
< 0.001
1.33 (1.19–1.48)
< 0.001
Model I = adjusted for (age + sex + MAP + sodium + potassium + PT + INR + CHF + atrial fibrillation + CKD + hepatic dysfunction + seizures + aspiration + sepsis)
Model II = model I + (thrombolysis + antiplatelet agent + anticoagulation agent + diuretic + ACEI/ARB + beta-blocker + CCB + SNP + MV + decompression)
FO fluid overload, OR odds ratio, IRR incidence rate ratio, CI confidence interval, ICU intensive care unit, MAP mean arterial pressure, PT prothrombin time, INR international normalized ratio, CHF congestive heart failure, CKD chronic kidney disease, ACEI/ARB angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker, CCB calcium channel blocker, SNP sodium nitroprusside, MV mechanical ventilation
aOR for binary outcomes (hospital mortality and adverse kidney event)
bIRR for count data outcome (ICU stay time)
Fig. 3
Restricted cubic spline curves for association between FO and clinical outcomes. Restricted cubic spline curve for association between the maximum FO and a in-hospital mortality or b adverse kidney event. FO fluid overload
×
Subgroup Analysis
Subgroup analyses, stratified by age, MV, and comorbidities, revealed a significant association between high FB trajectory pattern and in-hospital mortality across all strata. No significant interaction was observed among all groups (Fig. S4). For the outcome of adverse kidney event, patients with decreasing FB were at a significantly higher risk than other groups in all strata, except for the age > 75 years, MV ,and sepsis strata. The high FB group was significantly associated with adverse kidney event in the age ≤ 75 years, MV, non-CHF, and non-atrial fibrillation strata (Fig. S5).
Sensitivity Analysis
After excluding 39 patients receiving RRT, we remodeled a FB trajectory pattern that closely matching the original one. There were 80.1%, 7.3%, and 12.6% patients identified with low FB, decreasing FB, and high FB, respectively (Fig. S6). High FB was associated with elevated risks of hospital mortality, adverse kidney events, and extended ICU stays, consistent with pre-sensitivity analysis findings. Significant association was also observed between decreasing FB and adverse kidney events. In addition, decreasing FB trajectory was a risk factor of extended ICU stays (Table S3).
Discussion
In this retrospective multicenter study, we investigated the impact of FB trajectories on adverse outcomes in elderly patients with AIS. Fluid management, considered fundamental care for critically ill patients, aims to maintain circulatory stability, balance electrolytes and acid–base levels, and enhance tissue perfusion [9]. Various perspectives exist regarding the association between FB and prognosis. While some studies indicate that hypovolemia may lead to secondary brain injury [20], recent data also highlight the potential risks of FO [21, 22]. Patients with AIS primarily present with either normal or insufficient blood volume [23]. Although having a normal blood volume is preferable, the relationship between hydration and outcomes during AIS remains unclear.
Anzeige
Our study identified three trajectory patterns in this population, categorized as low FB, decreasing FB, and high FB groups. We found that persistent high FB was a risk factor for adverse outcomes in elderly patients with AIS, supported by multivariate logistic regression models and quasi-Poisson regression models, which was further demonstrated using subgroup analysis and sensitivity analysis. Additionally, the decreasing group was significantly associated with an increased risk of adverse kidney event and prolonged ICU stays as well. Although the FB trajectory of the decreasing group rapidly diminished over time, its cumulative FB maintained a positive equilibrium. Our research indicated that patients experiencing positive FB or FO ultimately face a poor prognosis. Raimundo et al. found that for every 1-L increase in FB, the risk of AKI progression increased sixfold [24]. In patients with acute respiratory distress syndrome, a positive cumulative FB correlated with AKI development by day 3, whereas this association was not observed in patients with an even or negative cumulative FB [25]. Therefore, for physicians, daily review of fluid status and limiting FO is essential.
In the multivariate regression models, the trajectory characterized by high FB exhibited the highest risk of in-hospital mortality and prolonged ICU stays. And the decreasing FB trajectory exhibited the highest risk of AKI. The decreasing FB trajectory was also associated with prolonged ICU stays, albeit with a lower risk compared to the high FB group. This group’s cumulative FB exceeded that of the high FB group in the first 2 days, but showed a downward trend from the third day onwards. This nuanced observation suggested that distinct degrees and durations of FO may exert varying impacts on the prognosis of patients with AIS. Our RCS models further underscored the correlation between increasing peak FO and elevated risks of in-hospital mortality and AKI. The risk of in-hospital mortality began to increase when the levels exceeded 3.6% and continued to increase. Another multicenter study demonstrates that the 28-day mortality risk for patients with AKI increases markedly when the maximum FO is over 10% [26]. This disparity indicates that in managing patients with AIS, their susceptibility to FO significantly exceeds that of patients with AKI. Therefore, intervention at a lower level of FO may be necessary to mitigate the increased risk of death compared to patients with AKI. It suggests that FO thresholds for assessing mortality risk might need adjustment based on the specific condition.
FO manifests as increased blood volume and edema [27]. Currently, osmotherapy serves as the cornerstone in the treatment of cerebral edema, with mannitol and hypertonic saline (HS) emerging as the most frequently utilized osmotic agents [28]. The contribution of FO to cerebral edema can elevate intracranial pressure [29] thereby posing significant risks to brain function. In our study, 16.5% of the patients developed FO; in comparison to those without FO, patients with FO exhibited an almost 1.6-fold higher hospital mortality risk. One of the common causes of early death in patients with AIS is cerebral edema. Malignant cerebral edema usually occurs between the second and fifth days after a stroke. As the brain is encased in a closed space, the increased intracranial pressure caused by cerebral edema can lead to brain herniation, brainstem compression, and death. A study shows that fluid intake exceeding 28 mL/kg/day in the initial days is associated with malignant cerebral edema [23]. Over the initial 4-day period, FB levels of our study within the high FB group consistently exceeded this threshold, with the decreasing FB group also recording an FB level above this threshold on the first day. A European multicenter study reveals a substantial correlation between positive daily FB in patients with traumatic brain injury and increased ICU mortality (OR 1.10 [95% CI 1.07–1.12] per 0.1-L increase) along with worse functional outcomes (OR 1.04 [95% CI 1.02–1.05] per 0.1-L increase) [29].
AKI is prevalent among critically ill patients, with fluid therapy as a conventional approach [30]. While fluid administration is commonly employed to prevent AKI [31] excessive fluid resuscitation beyond correcting the hypovolemia can also precipitate AKI [32]. Numerous studies have demonstrated a correlation between FO and AKI. Kuo et al. reported a 7.1-fold increase in AKI risk among patients with positive FB after cardiac or aortic surgery [33]. Similarly, Wang et al. found that, compared with a low FB, a high FB trajectory was associated with an increased risk of AKI in critically ill patients (OR 2.04 [95% CI 1.23–3.37]) [26]. One plausible mechanism behind this association is that FO leads to an accumulation of fluids in the extracellular space, disrupting the diffusion of oxygen and metabolites, impairing intercellular interactions, and potentially culminating in progressive organ dysfunction, including renal failure [34]. It is noteworthy that in our study, for the subgroup of patients with sepsis, a high FB was not identified as a risk factor for adverse kidney event. The inflammatory response in patients with sepsis can alter the effectiveness of fluid management. This response leads to the widening of gaps between endothelial cells in blood vessels, increasing vascular permeability and causing relative hypovolemia. In this context, appropriate fluid resuscitation is considered a crucial component of treating septic shock and hypotension [35].
Anzeige
Fluid management in neurocritical care patients poses specific challenges compared to other critically ill patients, given its focus on maintaining adequate cerebral blood flow (CBF) and oxygenation. Unfortunately, the implementation of sophisticated tools for CBF and oxygenation monitoring is not widely adopted in clinical practice [20]. In patients with ischemic stroke, the process of ischemia and subsequent reperfusion initiates a series of pathological events, with cerebral edema being the most critical. Cytotoxic edema initiates within hours of the stroke and peaks between the second and fifth days. This critical “watch period” necessitates vigilant neurological monitoring because of the inherent risk of secondary brain and brainstem compression [36].
In our study, patients’ FB levels were repeatedly documented over the initial 7 days following a stroke. This timeframe aligns with the development of cerebral edema development, ensuring the capture of significant changes in FB during the highest risk phase. Contrary to measurements taken at a single time point, longitudinal data provides a more comprehensive perspective, accurately reflecting trends in indicators over time [37]. Moreover, our study marked the pioneering use of GBTM to investigate the value of FB trajectories during the first 7 days following ICU admission in this population. This approach facilitates a deeper insight into disease progression and treatment efficacy, empowering doctors to make more precise decisions based on individual changes. However, our study still had some limitations. As a result of the retrospective nature of this study, we were unable to collect data on several important variables related to stroke prognosis, such as the National Institutes of Health Stroke Scale (NIHSS), Alberta Stroke Program Early CT Score (ASPECTS), blood pressure targets, ischemic stroke subtypes, death causes, etc. Further prospective studies are needed to incorporate these critical variables, which could provide a more detailed understanding of stroke outcomes and fluid strategies. Exploring the differences in fluid balance and prognosis among different subtypes of ischemic stroke is also an interesting direction worth further investigation. Moreover, we did not consider insensible losses while calculating FB because of the difficulty of accurate assessment. Lastly, the study did not investigate the effects of different fluid intake types on fluid accumulation and outcomes.
Conclusion
Our results identified three distinct trajectory patterns of daily FB in elderly patients with AIS. This contributes to the existing evidence linking high FB to adverse outcomes in elderly patients with AIS. Regular assessment of daily fluid status and restriction of FO are crucial for the recovery of critically ill patients.
Acknowledgements
The authors thank Dr. Huanrui Zhang for insightful discussions regarding statistical analyses. We thank the participants of the study.
Declarations
Conflicts of Interest
Jia Tang, Changdong Wu, and Zhenguang Zhong have nothing to disclose.
Ethical Approval
This research adhered to the ethical principles outlined in the Declaration of Helsinki and gained endorsement from the ethics evaluation committees of both Massachusetts Institute of Technology and Beth Israel Deaconess Medical Center (researcher certification number 52759164). Written informed consent was waived due to the non-identifiable and anonymous attributes of the database.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which permits any non-commercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc/4.0/.
Eine Therapie mit dem C40-Ligand-Blocker Frexalimab kann MS-Schübe und neue MRT-Läsionen über zwei Jahre hinweg verhindern. Dafür spricht die Auswertung einer offen fortgeführten Phase-2-Studie.
Eine auf das B-Zell-Reifungsantigen gerichtete mRNA-basierte CAR-T-Zell-Therapie wurde jetzt in einer ersten Phase IIb-Studie zur Behandlung der generalisierten Myasthenia gravis mit Placebo verglichen.
Eine Analyse aus Kanada bestätigt: Setzen ältere MS-Kranke die Behandlung mit Basistherapeutika ab, müssen sie kaum mit neuen Schüben und MRT-Auffälligkeiten rechnen.
Eine große Registerstudie bestätigt, dass Schichtarbeit mit einem erhöhten Risiko für psychische und neurologische Erkrankungen einhergeht, sowie mit einer Volumenabnahme in Gehirnarealen, die für Depression, Angst und kognitive Funktionen relevant sind.
