Impact of different extracorporeal blood purification strategies during continuous renal replacement therapy in septic shock patients

Abstract

Background

Septic shock with acute kidney injury is associated with high mortality. Hemoadsorption methods such as CytoSorb in conjunction with standard continuous renal replacement therapy (CRRT) and oXiris-based CRRT are increasingly used; however, comparative data are scarce. This study assessed the effects of both filters on vasopressor-free days and key clinical outcomes in septic shock.

Methods

This retrospective single-center cohort included adults with septic shock treated with CytoSorb in conjunction with CRRT or oXiris-based CRRT between 1st January 2023 and 31st December 2024. The primary endpoint was vasopressor-free days to day 28, with secondary outcomes including Sequential Organ Failure Assessment (SOFA) score, lactate levels, norepinephrine equivalent dose (NEED), mean arterial pressure, ventilator-free days, intensive care unit (ICU) length of stay, and mortality.

Results

A total of 97 patients were included in the analysis. (CytoSorb: n=75; oXiris: n=22). Both extracorporeal blood purification modalities were associated with comparable reductions in lactate levels (median [IQR]: −39.7% [−47.2% to −19.3%] vs. 44.5% [−54.0% to −33.3%], respectively, P=0.14), and NEED (median [IQR]: −36.1% [−66.7% to −9.1%] vs. −56.4% [−83.3% to −12.5%], respectively, P=0.36), along with similar increases in mean arterial pressure (9.6% [7.7%–18.2%] vs. 9.8% [7.7%–16.7%], respectively, P=0.94). No significant differences were found between the two modalities. Vasopressor-free days were similar (median [IQR]: 0 [0–24.0] days with CytoSorb vs. 20.5 [0–25.0] days with oXiris, P=0.55). In the multivariable competing-risks analysis, extracorporeal blood purification modality was not independently associated with vasopressor-free days after adjustment for potential confounders (subhazard ratio =0.97, 95% confidence interval: 0.48 to 1.94, P=0.93). Ventilator-free days, ICU stay, and ICU and hospital mortality were likewise comparable.

Conclusions

CytoSorb in addition with CRRT and oXiris-based CRRT demonstrated similar hemodynamic and clinical outcomes. Larger prospective studies are needed to define the optimal role of extracorporeal CRRT-based blood purification in septic shock.

Introduction

Sepsis represents a major global health challenge, accounting for approximately 31.5% of all deaths worldwide and remaining the leading cause of acute kidney injury (AKI) among critically ill patients, with disproportionately high mortality in low-resource settings (LRS).^[1,2] Despite advances in antimicrobial therapy, hemodynamic stabilization, and organ support strategies, the mortality rate in septic shock continues to be above 40%, especially among patients needing continuous renal replacement therapy (CRRT).^[3,4] The pathophysiology of sepsis involves a dysregulated immune response to infection, characterized by the excessive release of pro- and anti-inflammatory mediators that lead to microcirculatory failure, tissue damage, and multi-organ dysfunction. This uncontrolled inflammatory process, often called a “cytokine storm,” plays a key role in the worsening of septic shock and poor patient outcomes.^[5] High plasma levels of cytokines, such as interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), and interferon-gamma, are associated with disease severity, vasoplegia, and mortality.^[6]

Therefore, extracorporeal blood purification therapies have been studied as additional and potentially useful treatments to modulate the systemic inflammatory response and restore immune balance by removing circulating cytokines and endotoxins. Among these options, hemoadsorption has become a promising treatment. It works by directly adsorbing inflammatory mediators through porous sorbent materials integrated into CRRT circuits.^[7,8]

Two widely used hemoadsorptive strategies are the oXiris membrane and the CytoSorb cartridge, which differ in design and integration within CRRT circuits. The oXiris membrane serves as the primary hemofilter, combining renal replacement with cytokine and endotoxin adsorption, whereas CytoSorb is a standalone adsorptive cartridge placed in series with a conventional CRRT filter and does not provide intrinsic renal replacement therapy. The oXiris membrane, made of an acrylonitrile and methalylsulfonate (AN69) copolymer coated with polyethyleneimine and heparin, allows simultaneous removal of cytokines and endotoxins while providing renal support. In contrast, the CytoSorb cartridge contains highly porous polystyrene-divinylbenzene beads designed for hydrophobic adsorption of a wide range of middle-molecular-weight cytokines (up to 60 kDa).^[9,10] These structural and configurational differences reflect distinct extracorporeal blood purification strategies rather than equivalent hemoadsorption techniques.

In vitro comparisons have shown that oXiris achieves dual removal of endotoxins and cytokines, while CytoSorb demonstrates more efficient adsorption of selected inflammatory mediators such as IL-6.^[10] However, heterogeneity in patient selection, timing, and study design has contributed to inconsistent findings across trials, and comparative studies evaluating the effects of different extracorporeal blood purification strategies on clinical outcomes remain scarce. Despite decades of investigation, clear evidence of improved outcomes with any single extracorporeal blood purification approach remains lacking.^[1,11,12] Therefore, the present study aims to compare the impact of oXiris-based CRRT and CytoSorb used in conjunction with standard CRRT, on clinical outcomes in critically ill septic shock patients.

Methods

Study design, participants, and ethics

This retrospective, single-center, observational study was conducted in the Medical Intensive Care Unit (MICU) of the University Clinical Centre of the Republic of Srpska, a tertiary-level, university-affiliated hospital. Results are reported following the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.^[13] The study period extended from January 1 2023 to December 31 2024. The study protocol was reviewed and approved by the Ethics Committee of the University Clinical Centre of the Republic of Srpska (approval No: 01-19-373-2/23) and was performed in accordance with the Declaration of Helsinki. Informed consent was waived by the ethics committee for this retrospective study. All adult patients (≥18 years) admitted with a diagnosis of septic shock, according to the Sepsis-3 criteria,^[14] and treated with CRRT using either an oXiris membrane or CytoSorb in conjunction with conventional CRRT were eligible for inclusion. The extracorporeal blood purification strategy and the type of device were initiated at the treating intensive care physician’s discretion. Patients were excluded from the analysis if their primary diagnosis was not septic shock, if they had active malignancy, a history of heparin-induced thrombocytopenia, or heparin allergy. The management of septic shock, including antimicrobial therapy, fluid resuscitation, vasopressor support, and CRRT, was conducted in accordance with the institutional protocol and current international guidelines.

Primary endpoint

The primary endpoint was vasopressor-free days to day 28, defined as the number of days between initiation of different extracorporeal blood purification strategy to day 28 during which the patient was both alive and free from any intravenous vasopressor support.^[15] Any patient who died within 28 days was assigned 0 days. For patients who required vasopressor therapy for >28 days, the number of vasopressor-free days was 0. If vasopressors were successfully discontinued but subsequently restarted for >60 min within any 24-h window before day 28, those intervening days were not considered vasopressor-free. Vasopressor therapy included any intravenous administration of norepinephrine, phenylephrine, dopamine, epinephrine, or vasopressin.

Secondary endpoints

Secondary endpoints included intensive care unit (ICU) and hospital mortality, Sequential Organ Failure Assessment (SOFA) score, lactate clearance, norepinephrine equivalent dose (NEED), duration of vasopressor therapy, ICU length of stay, and ventilator-free days to day 28 (defined analogously to vasopressor-free days).

CRRT and hemoadsorption

CRRT was performed using the PRISMAFLEX system (Baxter International Inc., WI, USA), which offers multiple treatment modalities; however, continuous venovenous hemodiafiltration was the most commonly used mode in this study. The oXiris hemofilter (Baxter International Inc., WI,USA) utilizes an AN69-based membrane composed of acrylonitrile and sodium methallyl sulfonate copolymer fibers, surface-treated with polyethyleneimine and pre-grafted with heparin (approximately [4500 ± 1500] IU/m²). This configuration enables simultaneous CRRT, adsorption of cytokines and endotoxins, as well as antithrombogenic surface protection. The membrane has an effective surface area of 1.5 m², an internal fiber diameter of 240 µm, and a wall thickness of 50 µm. The CytoSorb cartridge (CytoSorbents Corporation, USA) is a sterile, non-pyrogenic, single-use hemoadsorption device containing highly porous divinylbenzene copolymer beads designed to remove a broad range of hydrophobic inflammatory mediators and cytokines up to approximately 55 kDa via size-exclusion and hydrophobic interactions. The cartridge was integrated in series into the CRRT circuit, positioned after the hemofilter in the post-dialyzer (venous) line, according to the manufacturer’s recommendations.^[16] CRRT modality, prescribed dose, and anticoagulation strategy were standardized according to institutional protocols and were comparable between patients treated with the CytoSorb cartridge and those treated with the oXiris membrane.

Data source, collection, and laboratory analysis

Local electronic patient charts were reviewed, and data were collected using Microsoft Excel (Microsoft, Redmond, WA, USA). Laboratory and hematological parameters in this study were routinely collected at two time points: 15–30 min before initiation of extracorporeal blood purification therapy with either CytoSorb or oXiris, and immediately after the procedure was completed. Each CRRT-based blood purification session lasted 24 h and was performed under standardized conditions for all patients. Comprehensive demographic and clinical data were recorded at the time of ICU admission and at the initiation and completion of different extracorporeal blood purification strategies. The collected variables included, but were not limited to, age, sex, comorbidities, site of infection, vasopressor use and doses, ICU length of stay, duration of vasopressor therapy and mechanical ventilation, and hospital and ICU survival. NEED was calculated as previously defined.^[17] The Simplified Acute Physiology Score Ⅱ (SAPS Ⅱ) was assessed upon ICU admission, while the SOFA score was recorded on the day of laboratory and hematological sampling. All hematological tests were performed in the hospital’s Department of Laboratory Diagnostics using the Sysmex XN-3100 hematology analyzer, which uses fluorescence flow cytometry (WBC DIFF/RET/PLT-F/WPC) with an integrated SP-50 module for automatic smear preparation and staining. Biochemical analyses were carried out using the Abbott Alinity c analyzer (Abbott Diagnostics, USA), which uses photometric and potentiometric detection methods.

Statistical analysis

The normality of data distribution was assessed using the Shapiro–Wilk test and by visually checking the distribution (histogram) of each variable. Data were expressed as mean  ±  standard deviation (SD) when they were normally distributed and as median and interquartile range (IQR) when they were non-normally distributed. Proportions were used as descriptive statistics for categorical variables. Comparisons of values between the CytoSorb and oXiris groups were performed using a two-tailed Student’s t-test or the Mann–Whitney U test, as appropriate. Analyses of discrete data were performed using the Chi-square test or Fisher’s exact test when the numbers were small. Pairwise comparisons between the different study time periods were assessed using a paired Student’s t-test, a Wilcoxon test, or McNemar’s test as appropriate.

The primary endpoint (vasopressor-free days) was analyzed using competing-risks regression based on Fine and Gray’s proportional subdistribution hazards model. Death before day 28 was treated as the competing event, and observations were right-censored at day 28. Multivariable competing-risks models were constructed to adjust for potential confounders, incorporating clinically relevant variables and those with a P value <0.1 in univariable analyses. Potential collinearity among covariates was assessed using Spearman or Pearson correlation coefficients before model fitting. Results are reported as subhazard ratios (sHRs) with corresponding 95% confidence intervals (CIs).

A value of P <0.05 was considered statistically significant, and all reported P-values are two-sided. Statistical analyses were performed using Stata 17.0 software for Windows (Stata Corp LLC, TX, USA).

Results

Study population

Over 2 years, a total of 97 patients were included in the study. The main characteristics of the cohort are summarized in Table 1. Male patients accounted 67.0% of the study population. The median age at the onset of septic shock was 60 years (IQR: 45–69). The mean SAPS Ⅱ score was 53 ± 19. Chest infections were the most common source of sepsis, and gram-negative organisms were the most frequently identified pathogens in blood cultures.

Table 1.

Baseline characteristics of the study population and between the CytoSorb and oXiris groups.

Variables	All patients (n=97)	CytoSorb group (n=75)	oXiris group (n=22)	P value
Age (years)	60 (45–69)	61 (48–69)	56 (40–69)	0.45
Male	65 (67.0)	51 (68.0)	14 (63.6)	0.80
Route of ICU admission				0.33
Emergency department	44 (45.4)	36 (48.0)	8 (36.4)
General ward	53 (54.6)	39 (52.0)	14 (63.6)
SAPS II score	53 ± 19	54 ± 19	53 ± 20	0.84
Charlson score	2 (1–4)	2 (1–4)	2 (1–4)	0.91
Comorbidity
Essential hypertension	47 (48.4)	36 (48.0)	11 (50.0)	0.87
Diabetes	26 (26.8)	20 (26.7)	6 (27.3)	0.95
Cardiomyopathy	23 (23.7)	21 (28.0)	2 (9.1)	0.09
Chronic anemia	63 (65.0)	46 (61.3)	17 (77.3)	0.21
COPD	6 (6.2)	5 (6.7)	1 (4.5)	1.00
Chronic kidney disease	11 (11.3)	8 (10.7)	3 (13.6)	0.73
Chronic liver diseases	29 (29.9)	24 (32.0)	5 (22.7)	0.60
Source of infection				0.10
Chest	50 (51.5)	33 (44.0)	17 (77.3)
Abdomen	12 (12.4)	11 (14.7)	1 (4.5)
Skin	8 (8.2)	8 (10.7)	0 (0.0)
Blood	18 (18.6)	15 (20.0)	3 (13.6)
Others	9 (9.3)	8 (10.7)	1 (4.5)
Isolated pathogen in the blood				0.79
Gram-negative bacteria	36/74 (48.6)	25/54 (46.3)	11/20 (55.0)
Gram-positive bacteria	15/74 (20.3)	11/54 (20.4)	4/20 (20.0)
Others (including unknown)	23/74 (31.1)	18/54 (33.3)	5/20 (25.0)
Hydrocortisone	97 (100)	75 (100)	22 (100)	NA
Invasive mechanical ventilation	86 (88.7)	66 (88.0)	20 (90.9)	1.00
Norepinephrine on ICU admission	97 (100)	75 (100)	22 (100)	NA
Vasopressin on ICU admission	17 (17.5)	14 (18.7)	3 (13.6)	0.75
White blood cell count (x109/L)	10.3 (6.1–17.3)	10.3 (6.4–17.1)	10.2 (5.3–18.4)	0.99
Hemoglobin (g/L)	118 ± 30	120 ± 30	111 ± 30	0.18
Platelets (x109/L)	148 (75–247)	150 (73–240)	127 (98–261)	0.97
C-reactive protein (mg/L)	203 (88–318)	230 (91–313)	209 (62–346)	0.88
Procalcitonine (ng/mL)	7.6 (2.0–27.8)	8.7 (1.9–27.8)	5.1 (2.0–43.5)	0.94
D-dimer (mg/L)	6.8 (2.4–15.1)	6.8 (2.7–19.3)	5.6 (2.2–10.5)	0.24
Creatinine (µmol/L)	175 (100–401)	202 (112–402)	140 (60–339)	0.13
Interleukin-6 (pg/mL)	517 (123–5557)	790 (211–8855)	118 (19–444)	0.01

Data are expressed as mean ± standard deviation, median (interquartile range), or n (%).

COPD: Chronic obstructive pulmonary disease; ICU: Intensive care unit; NA: Not applicable; SAPS II: Simplified acute physiologic score.

Comparisons between the Cytosorb and oXiris groups at the baseline

On ICU admission, CytoSorb combined with CRRT was performed in 75 patients (77.3%), whereas oXiris combined with RRT was used in 22 patients (22.7%). The two study cohorts were largely similar with respect to baseline characteristics. The only significant differences were a higher SOFA score and elevated IL-6 concentrations in the CytoSorb group relative to the oXiris group (Tables 1 and 2).

Table 2.

Hemodynamic and organ function parameters before and after hemoadsorption use in the CytoSorb and oXiris groups.

Variables	CytoSorb (n=75)		oXiris (n=22)
	Before	After	Before	After
SOFA score	10.2 ± 3.7	9.7 ± 3.7*	8.2 ± 4.5†	8.0 ± 4.7
Lactate (mmol/L)	3.2 (1.6–7.0)	1.9 (1.0–3.5)*	3.9 (2.2–7.1)	2.6 (1.0–3.2)*
NEED (µg/kg∙min)	0.5 (0.3–0.7)	0.3 (0.1–0.5)*	0.3 (0.2–0.6)	0.15 (0.05–0.35)*
Mean arterial pressure (mmHg)	60 ± 8	67 ± 7*	62 ± 8	69 ± 7*
Heart rate (beats/min)	111 ± 12	100 ± 11*	111 ± 13	102 ± 12*
Vasopressor use	75 (100)	13 (17.3)*	22 (100)	5 (22.7)*

Data are expressed as mean ± standard deviation, median (interquartile range), or n (%).

NEED: Norepinephrine equivalent dose; SOFA: Sequential organ failure assessment.

^⁎P <0.05 comparisons between before and after hemoadsorption.

^†P <0.05 comparisons between CytoSorb vs. oXiris.

Comparisons between the values before and after hemoadsorption use

Lactate levels, NEED, heart rate, and vasopressor use decreased significantly from pre- to post-hemoadsorption in both the CytoSorb and oXiris groups (Table 2). Mean arterial pressure increased in both cohorts. The SOFA score showed a statistically significant reduction only in the CytoSorb group, although the magnitude of this change was not clinically meaningful. Notably, the degree of change in these variables did not differ significantly between the two groups (Table 3).

Table 3.

Primary and secondary outcomes between the CytoSorb and oXiris groups.

Variables	CytoSorb group (n=75)	oXiris group (n=22)	P value
Absolute changes in NEED (µg/kg·min)	−0.15 (−0.25 to −0.05)	−0.10 (−0.20 to −0.05)	0.44
Relative changes in NEED (%)	−36.1 (−66.7 to −9.1)	−56.4 (−83.3 to −12.5)	0.36
Absolute changes in SOFA score	0 (0–0)	0 (0–0)	0.13
Relative change in SOFA score (%)	0 (0–0)	0 (0–0)	0.17
Absolute changes in lactate levels (mmol/L)	−1.2 (−2.8 to −0.5)	−1.7 (−2.8 to −0.5)	0.63
Lactate clearance ( %)	−39.7 (−47.2 to −19.3)	−44.5 (−54.0 to −33.3)	0.14
Absolute changes in MAP (mmHg)	5 (5–10)	5 (5–11)	0.69
Relative changes in MAP (%)	9.6 (7.7–18.2)	9.8 (7.7–16.7)	0.94
Absolute changes in HR (beats/min)	−10 ± 7	−10 ± 9	0.90
Relative changes in HR (%)	−9.2 ± 6.2	−8.8 ± 8.2	0.79
ICU length of stay (days)	11 (3–24)	10 (4–122)	0.81
Vasopressor duration (days)	4.0 (2.0–8.0)	4.0 (2.0–6.0)	0.72
Vasopressor-free days to day 28 (days)	0.0 (0.0–24.0)	20.5 (0.0–25.0)	0.55
Ventilator-free days to day 28 (days)	0.0 (0.0–18.0)	6.5 (0.0–21.0)	0.34
ICU mortality	40 (53.3)	10 (45.4)	0.52
Hospital mortality	46 (61.3)	11 (50.0)	0.34

Data are expressed as mean ± SD, median(interquartile range), or n (%).

HR: Heart rate; ICU: Intensive care unit; MAP: Mean arterial pressure; NEE: Norepinephrine equivalent dose; SOFA: Sequential organ failure assessment.

Absolute changes were calculated as the post-hemoadsorption value minus the pre-hemoadsorption value. Relative changes were calculated as the absolute change divided by the pre-hemoadsorption value, multiplied by 100.

Vasopressor-free days

The median (IQR) number of vasopressor-free days to day 28 did not differ significantly between the CytoSorb and oXiris groups (0 [0–24.0] vs. 20.5 [0–25.0] days, respectively) (Table 3).

In the multivariable competing-risks regression analysis, adjusting for clinically relevant confounders (age, baseline lactate, NEED, and SOFA score) as well as variables with P <0.1 in the univariable analyses (Table 1), the extracorporeal blood purification modality (CytoSorb vs. oXiris) was not independently associated with the likelihood of survival and vasopressor independence by day 28 (adjusted sHR=0.97; 95% CI: 0.48 to 1.94; P=0.93) (Table 4 and Figure 1).

Table 4.

Multivariable competing-risks regression analysis.

Variables	sHR	95% CI	P value
CytoSorb (refer: oXiris)	0.97	0.48–1.94	0.93
Age (year)	0.99	0.97–1.02	0.63
SOFA score on ICU admission	0.95	0.88–1.02	0.17
Lactate (mmol/L)	0.84	0.69–1.02	0.08
History of cardiomyopathy	0.81	0.31–2.15	0.68
Site of infection	1.03	0.83–1.29	0.78
NEED on ICU admission (µg/kg•min)	0.53	0.13–2.13	0.37
Interleukin-6 (ng/mL)	1.00	0.99–1.00	0.92
Mean arterial pressure (mmHg)	0.99	0.94–1.04	0.79

CI: Confidence interval; ICU: Intensive care unit; NEED: Norepinephrine equivalent dose; sHR: Subhazard ratio; SOFA, Sequential organ failure assessment.

Figure 1.

Cumulative incidence functions for vasopressor-free days in patients who received CytoSorb and those who received oXiris. sHR: Subhazard ratio.

Secondary end-points

Ventilator-free days, duration of vasopressor therapy, ICU length of stay, and ICU and hospital mortality did not differ significantly between the CytoSorb and oXiris groups (Table 3).

Discussion

The key results of this study indicate that both extracorporeal blood purification techniques led to significant decreases in lactate levels and vasopressor use. Nevertheless, vasopressor-free days, lactate clearance, and the change in SOFA score were comparable between the two modalities.

At the early onset of septic shock, a massive release of endotoxins, pro-inflammatory cytokines, pathogen-associated molecular patterns (PAMPs), and damage-associated molecular patterns (DAMPs) triggers an overwhelming systemic inflammatory response, ultimately leading to organ dysfunction.^[6,18,19] Thus, the removal of pro-inflammatory mediators – such as IL-6 and TNF-α – along with PAMPs and DAMPs may attenuate immune hyperactivation, mitigate the systemic inflammatory response, and improve clinical outcomes. This hypothesis underpins the rationale for hemoadsorption therapies. Among the blood purification devices currently used during CRRT in patients with septic shock, CytoSorb and oXiris are the most widely applied.

Studies assessing the efficacy of the CytoSorb cartridge during CRRT in patients with septic shock have yielded conflicting results. Small randomized and observational studies suggest that early CytoSorb use may reduce NEED requirements within 24 h, with associated improvements in lactate clearance and hemodynamic parameters, but without consistent improvement in SOFA score.^[[20], [21], [22]^] In agreement with these findings, we observed significant improvements in NEED, lactate clearance, and hemodynamic parameters after 24 h of CytoSorb use, without any clinically meaningful change in SOFA score (Table 2). However, other studies, including the only multicenter randomized controlled trial (RCT), reported no significant reduction in vasopressor requirement and IL-6 levels.^[23,24] This apparent dissociation may be explained by ongoing cytokine production and redistribution from the interstitial to the intravascular compartment.^[25]

A similar pattern is observed with oXiris. Small trials and meta-analyses describe hemodynamic benefits,^[12,26,27] a signal aligned with the significant reductions in NEED and lactate in our cohort (Table 2). However, the ENDoX trial found no improvement in endotoxin activity or organ dysfunction compared with standard care.^[28] Nonetheless, the 2024 Asia-Pacific oXiris Expert Meeting consensus recommends considering oXiris in selected septic patients requiring CRRT, albeit based on low-quality evidence.^[29]

We observed no significant differences between CytoSorb used in conjunction with CRRT and oXiris-based CRRT in terms of hemodynamic improvement or clinical outcomes (Table 4). Although not statistically significant, reductions in NEED requirements and lactate levels, as well as vasopressor-free and ventilator-free days, were numerically greater in the oXiris group than in the CytoSorb group (Table 3). This difference may be explained by greater baseline severity in the CytoSorb group, as reflected by higher SOFA scores and IL-6 levels (Table 1). After adjustment for baseline severity and other potential confounders, the extracorporeal-CRRT based blood purification technique was not associated with vasopressor-free days at 28 days (Table 4 and Figure 1). However, given the retrospective design, confounding by indication remains a major concern. Accordingly, the absence of a significant difference in vasopressor-free days between CytoSorb and oXiris should be interpreted with substantial caution. To the best of our knowledge, this is the first study to directly compare CytoSorb used in conjunction with CRRT and oXiris-based CRRT in patients with septic shock (or other populations); therefore, no prior data are available for direct comparison with our findings.

In our study, CytoSorb and oXiris were applied using the same CRRT modality, blood flow rate, effluent dose, anticoagulation strategy, and circuit lifespan. Accordingly, differences in CRRT configuration are unlikely to have influenced hemodynamic responses in this comparison. Nevertheless, despite identical CRRT prescriptions, differences in solute clearance related to intrinsic membrane properties cannot be entirely excluded. These considerations suggest that the observed findings are unlikely to be explained by hemodynamic effects of CRRT configuration.

Hospital mortality in both groups, irrespective of the extracorporeal blood purification strategy used, ranged between 50% and 60%. Reported septic shock mortality varies widely with healthcare system resources, averaging 35%–39% in high-resource settings and 50%–70% in LRS.^[[30], [31], [32]^] Given that our study was conducted in an LRS, the observed mortality rates are consistent with those reported in comparable contexts. The impact of CytoSorb on mortality in septic shock remains controversial. A recent meta-analysis (744 patients) reported an association between CytoSorb use during CRRT and lower in-hospital mortality.^[33] However, in the only multicenter RCT, unadjusted mortality was higher with CytoSorb, with no difference after adjustment.^[23] Similarly, oXiris-based CRRT was associated with lower mortality in a meta-analysis of 14 studies (695 patients), although the certainty of evidence was low to very low due to predominantly observational designs and limited RCT data.^[12] Overall, these findings underscore the need for adequately powered, multicenter RCTs to determine whether CytoSorb during CRRT or oXiris-based CRRT confer a true mortality benefit.

Several limitations should be acknowledged. This single-center retrospective design limits generalizability and is prone to selection bias and unmeasured confounding. The overall sample size – particularly in the oXiris group – was small, reducing the statistical power to detect moderate but clinically meaningful differences in the primary and secondary outcomes. Vasopressor-free days on day 28 showed a markedly skewed distribution in this small, heterogeneous cohort, limiting the interpretability of simple summary statistics. Importantly, vasopressor duration was similar between the two groups (Table 3), suggesting comparable exposure to vasopressor support over time. Because this endpoint is influenced by both persistent shock and death, we used competing-risk regression to account for death as a competing event when assessing vasopressor discontinuation. Nevertheless, residual imprecision related to outcome distribution cannot be excluded. Endotoxin levels were not available, precluding direct assessment of endotoxemia, and IL-6 was not reassessed after hemoadsorption, limiting evaluation of cytokine removal. The exact interval between shock onset and hemoadsorption initiation was not recorded, although therapy was started within 24 h of shock onset. Finally, the absence of a control group limits causal inference regarding the effects of extracorporeal blood purification during CRRT.

Conclusions

In this retrospective cohort of patients with septic shock requiring CRRT, the use CytoSorb and oXiris demonstrated comparable hemodynamic improvements, with no significant differences in vasopressor-free days, ventilator-free days, ICU length of stay, or mortality. However, these findings should not be interpreted as evidence of technical equivalence between the devices. Instead, they provide limited, exploratory head-to-head data in a small cohort and underscore the need for adequately powered, prospective multicenter studies to better define the role, comparative effectiveness, and optimal selection of extracorporeal CRRT-based blood purification strategies in septic shock.

CRediT authorship contribution statement

Pedja Kovacevic: Writing – review & editing, Writing – original draft, Conceptualization. Nikolina Spiric: Writing – review & editing, Data curation. Katarina Vucicevic: Writing – review & editing, Data curation. Sasa Dragic: Writing – review & editing, Data curation. Danica Momcicevic: Writing – review & editing, Data curation. Biljana Zlojutro: Writing – review & editing, Data curation. Milka Jandric: Writing – review & editing, Data curation. Tijana Kovacevic: Writing – review & editing, Data curation. Jihad Mallat: Writing – original draft, Methodology, Formal analysis, Conceptualization.

Managing Editor: Jingling Bao