Adjuvant Hemadsorption Therapy in the Management of Refractory Septic Shock

Introduction

Sepsis is a potentially fatal organ dysfunction resulting from a dysregulated host response to infection. It affects millions globally, posing significant healthcare challenges, with mortality in low- and middle-income countries reaching up to 60%.¹ In India, the estimated sepsis burden was 11.3 million in 2017, with mortality estimates of 30-50%.² The primary cause of sepsis-related deaths is progression to refractory septic shock (RSS) and subsequent multi-organ dysfunction (MOD).^3,4

RSS is characterized by persistent hypotension and end-organ dysfunction despite fluid resuscitation and requires oxygenation, ventilation, and high-dose vasopressors^1,5,6 RSS requires escalated treatment strategies, such as corticosteroid therapy, immunomodulatory treatments, kidney replacement therapy (KRT), and extracorporeal blood purification therapies (EBPTs).

Pathophysiologically, septic shock is characterized by cytokine storm and endotoxemia (in Gram-negative bacterial infections). Cytokine storm results from exponential amplification of pro-inflammatory [tumor necrosis factor-alpha (TNF-α) and interleukins (IL-1, IL-6, IL-8)] and anti-inflammatory (IL-10) modulators released primarily through toll-like receptors activated by pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs).⁷ The systemic inflammation leads to vasodilation, endothelial dysfunction, and tissue hypoperfusion,⁸ resulting in RSS and subsequent MOD. Endotoxins contribute to cytokine overload by triggering an immune cascade via TLR4-mediated activation of the NF-kB pathway, leading to pro-inflammatory cytokine release.⁹

EBPTs can eliminate endotoxins, PAMPs, DAMPs, activated leukocytes, cytokines, as well as pathogens from the bloodstream. Different EBPTs filter different target molecules based on their molecular weight.¹⁰ Hemadsorption, which targets 10-60 kDa molecules, can mitigate the overload of immune modulators (pro-and anti-inflammatory cytokines) as well as endotoxins. Hemadsorption involves blood circulation through an extracorporeal circuit containing a cartridge composed of a biocompatible adsorbent.¹¹

Despite the direct impact of cytokine storms on patient survival, hemadsorption is not included in standard care, and there are currently no universal guidelines for its use in sepsis.¹ Mortality benefit was observed with hemadsorption therapy within 24-48 hours of sepsis onset.^6,12-14 Despite the growing body of international research, studies in India are limited. Therefore, we investigated the time threshold for initiating hemadsorption as an adjuvant therapy for RSS by evaluating the impact of early initiation (within 24 hours of onset) on survival and duration of ICU/hospital stay. The 24-hour mark was selected because of the cytokine surge that occurs within that timeframe after onset.

Materials and Methods

This retrospective observational study included 112 patients with septic shock progressing to RSS. It was approved by the Institutional Ethics Committee (IEC) of our hospital (HNH/IEC/2022/OCS/CCM/93). Patient (or patient family) consent was obtained for the use of hemadsorption adjuvant therapy. Considering the retrospective nature of the study need for informed consent was waived by the IEC. Inclusion criteria were adult patients (aged >18 years) with septic shock progressing to RSS. RSS onset was defined as continuous escalating need for norepinephrine infusion to maintain adequate blood pressure (defined as >0.2 µg/kg/min norepinephrine for >1 hour);¹ circulating IL-6 levels >1000 pg/mL; or intra-abdominal sepsis with/without endotoxin activity assay (EAA) level >0.6 units. Exclusion criteria were transfer from another hospital on vasopressors, received continuous KRT within 24 hours before enrollment, pregnancy, limited life expectancy due to any underlying disease, or if the patient or family had chosen palliative care and/or declined aggressive treatment.

Demographic data, including age, sex, and comorbidity profiles, were recorded. The source of sepsis was identified for all patients and controlled where possible. Acute Physiology and Chronic Health Evaluation-II (APACHE-II) and Sequential Organ Failure Assessment (SOFA) scores were calculated to determine disease severity. Creatinine levels were monitored to assess kidney function. Outcomes included length of ICU and hospital stays and mortality. IL-6, procalcitonin, and C-reactive protein (CRP) levels were measured pre- and post-hemadsorption. Due to its retrospective nature, some values for the laboratory parameters were missing in the study, a major reason being that these tests were subject to the patients’ financial status and were refused by the patient/family in some cases.

All patients received standard-of-care treatment in accordance with international guidelines.¹ Despite this, patients progressed to renal failure with escalating vasopressor requirement, and eventually required KRT in accordance with the KIDGO guidelines.¹⁵ The indication for hemadsorption in patients with RSS was hemodynamic worsening despite KRT, IL-6 levels >1000 pg/mL, or intra-abdominal sepsis with/without EAA level >0.6 units. These patients were treated with either a cytokine removal hemadsorption cartridge [CytoSorb® (CytoSorbents Inc., NY, USA) and Hemofeel (Toray Medical Co., Ltd., Japan)], an endotoxin removal cartridge (Toraymyxin™, Toray Medical Co., Ltd.), or a dual cytokine and endotoxin removal cartridge (Oxiris, Baxter, USA) where appropriate, along with hemodialysis.

All filters used in this study have regulatory approval in India. The selection of filters or their combination was based on clinical/pathological characteristics and the financial situation of the patient. The use of Oxiris was subject to the availability of a compatible hemodialysis instrument. Clinical condition and IL-6 levels were re-evaluated after each filter use to determine continuation of hemadsorption. Hemadsorption was discontinued if IL-6 levels did not decrease post-hemadsorption or patient’s condition deteriorated.

Results

Of 112 patients, 27 received hemadsorption within 24 hours of RSS onset (early group) and 85 at/after 24 hours (late group). Demographic parameters did not significantly vary between the two groups, >60% patients being male. The distribution of comorbidities between the two groups was comparable. Almost half the patients had diabetes and hypertension [Table 1].

Baseline left ventricular ejection fraction was comparable in both groups (52% vs. 54%). The abdomen, lungs, and urinary tract were the most common sepsis sources in both groups. Baseline creatinine levels did not significantly vary between the two groups [Table 2]. APACHE-II and SOFA scores, measured for all patients at admission and at discharge/death, were also comparable.

IL-6 and procalcitonin levels significantly decreased in all patients post-hemadsorption [Figures 1 and 2]. IL-6 levels significantly decreased in both early and late initiation groups. A similar trend was observed for procalcitonin, with statistical significance in the late initiation group. Although CRP levels decreased post-hemadsorption in both groups, the difference did not reach statistical significance [Figure 3].

Close

Figure 1:

Interleukin-6 levels. **p<0.001

Export to PPT

Close

Figure 2:

Procalcitonin levels. *p<0.05.

Export to PPT

Close

Figure 3:

C-Reactive protein levels.

Export to PPT

Clinical outcome

Patients who received early hemadsorption had significantly shorter ICU (p=0.017) and hospital (p=0.036) stays. The median length of ICU and hospital stay for the early initiation group was six (IQR, 2-19) and 10 (IQR, 2.5-25.5) days, respectively; for the late initiation group, the length of stay was 12 (IQR, 6-23) and 15 (IQR, 9-28) days, respectively. The survivor group had significantly reduced vasopressor requirement post-hemadsorption [Table 3].

Table 3: Vasopressor requirements pre- and post-hemadsorption in survivors

Vasopressor	Pre-hemadsorption	Post-hemadsorption	p-value
Noradrenaline	10 (7.5,13.5)	3 (2.5,6.5)	0.001
Vasopressin	2.4 (2.1,2.4)	0 (0,0.6)	0.003
Adrenaline	7.5 (3,12)	0 (0,1)	0.016

Statistical difference between the pre- and post-hemadsorption vasopressor requirements was evaluated using Mann-Whitney U test.

Mortality rates significantly differed between the two groups (p=0.032). The early initiation group had a survival rate of 25.93%, compared to 10.59% in the late initiation group. Kaplan-Meier survival curve, generated and compared using the log-rank test, showed significant (p<0.05) survival benefit for early hemadsorption initiation [Figure 4].

Close

Figure 4:

Kaplan-Meier survival curve for early (<24 h) and late (≥24 h) hemadsorption initiation groups.

Export to PPT

We conducted subgroup analysis based on hemofilter type-cytokine removal (CytoSorb and Hemofeel) and endotoxin removal (Toraymyxin and Oxiris). The survival rates in these groups, depending on hemadsorption initiation (before/after 24 hours of shock onset), have been given in Table 4. Higher survival rates were observed in both groups when hemadsorption was initiated within 24 hours of RSS onset.

Table 4: Survival rates in patients when grouped based on cytokine or endotoxin removal

Cohort		Survival rate	p-value
Cytokine removal group (n=78)	Hemadsorption within 24 h of RS onset (n=17)	25.53%	0.282
	Hemadsorption at/after 24 h of RS onset (n=61)	13.11%
Endotoxin removal group (n=34)	Hemadsorption within 24 h of RS onset (n=10)	30.0%	0.067
	Hemadsorption at/after 24 h of RS onset (n=24)	4.17%

RS: Refractory shock

Discussion

Cytokine overproduction is pivotal in sepsis pathogenesis. Elevated pro- and anti-inflammatory cytokine levels, rather than their imbalance, are associated with high mortality.¹⁶ Cytokine removal can attenuate the dysregulated host response, particularly in the early phase of sepsis. Circulatory molecules are classified as small (<500 Da-10 kDa), middle (10-100 kDa), and large (>100 kDa).¹⁰ Among the EBPTs, hemodialysis eliminates only small molecules (creatinine, urea, electrolytes), while hemofiltration, plasmapheresis, and other modalities eliminate large molecules [proteins (fibrinogen, β-globulin, complement proteins), lipoproteins, and immunoglobulins (IgA, IgM, IgA)]. Hemadsorption alone eliminates the middle-sized immune modulators [ILs (21-40 kDa), TNF-α (39-51 kDa)].¹⁰ Recently, the Italian Society of Anesthesiology and Critical Care and the Italian Society of Nephrology jointly proposed recommendations for hemadsorption in clinical practice and research.¹⁷ However, the Surviving Sepsis Guidelines offer no recommendations on hemadsorption.¹

The present study found that early initiation of hemadsorption therapy increased the survival rate in patients with sepsis. A similar trend was observed when patients were sub-grouped based on the use of cytokine or endotoxin removal filters. We observed a significant reduction in procalcitonin, a biomarker of bacterial infection,¹⁸ and IL-6, an independent predictor of sepsis.¹⁹ Although CRP levels decreased post-hemadsorption, a statistical significance was not observed. It is important to note that while CRP is a sensitive marker of sepsis, it is not specific.²⁰ The survivor group showed reduced vasopressor requirement post-hemadsorption, indicating clinical improvement. Early hemadsorption was associated with reduced length of ICU and hospital stay.

Like our findings, several studies report improved outcomes and mortality benefits of early hemadsorption using CytoSorb and Hemofeel;^12,14,21-23 however, initiation timing varies. Studies have reported 75% survival when therapy was initiated within 24 hours of ICU admission, 68% for initiation between 24-48 hours,¹³ and 28% for initiation within 48 hours following septic shock onset.²¹ Further, the maximum ICU stay for 70% of survivors was <15 days.²¹ Notably, hemodynamic stabilization post-CytoSorb therapy was correlated with early initiation, with delayed treatment linked to poor outcomes.¹² Hemadsorption was associated with improved hemodynamic stabilization^12,24,25 and reduced IL-6,^13,24,25 procalcitonin,²⁵ and CRP²⁵ levels, which were consistent with our findings. Oxiris filter (suitable for septic shock with endotoxemia) was associated with significantly reduced IL-6, procalcitonin, and lactate levels as well as decreased vasopressor requirement.^26,27 However, data on its impact on length of ICU stay and hospital mortality are lacking. A meta-analysis found adjuvant Oxiris treatment to be associated with lower 28-, 14-, and 7-day mortalities, but not 90-day mortality.²⁸ The EUPHAS²⁹ and EUPHRATES³⁰ trials supported adjuvant PMX-B hemadsorption (Toraymyxin) in abdominal septic shock to improve hemodynamics, organ dysfunction, and 28-day mortality. Toraymyxin use significantly reduced SOFA scores and mortality in abdominal septic shock survivors.³¹ Thus, substantial evidence supports hemadsorption as adjuvant therapy in sepsis management. However, consensus on timing and indication for use remains lacking.

Our study highlights the importance of early hemadsorption initiation in achieving optimal outcomes. We observed limited mortality benefits when using hemadsorption therapy late in RSS treatment. Notably, most studies on hemadsorption filters report their use as a last resort to survival. This may be a contributing factor to the lower level of evidence supporting hemadsorption efficacy. In contrast, more robust, positive results have been observed in studies that report early hemadsorption use. Furthermore, most evidence comes from case reports/series; robust randomized control trial (RCT) data are lacking. Several barriers hinder such RCTs, including patient heterogeneity in terms of clinical phenotypes,¹¹ regional variation in KRT guidelines, high cartridge cost, limited availability of KRT devices in middle- and low-income countries, and lack of universal guidelines for hemadsorption use. As an alternative, Bottari et al. suggest designing studies with a specific patient population (same clinical phenotype), which may help establish standardized protocols.³² Given these limitations, our findings support a 24-hour threshold for initiating hemadsorption in RSS to improve outcomes.

A major concern against hemadsorption is theunselective removal of hydrophobic plasma components within the filter’s size-spectrum, potentially eliminating hormones, clotting factors, and some antibiotics. This undermines its benefit and may result in sub-therapeutic antibiotic levels, affecting treatment success.³³ However, the “peak concentration theory”³⁴ suggests that this non-selective immunomodulator removal is in fact beneficial, improving outcomes by reducing pro- and anti-inflammatory levels to those observed in a normal immune response. Moreover, these filters function in a concentration-dependent manner, preventing total removal of clinically relevant molecules (e.g., antibiotics). Overall, hemadsorption is safe and does not negatively impact patient outcomes, unlike other innovative treatments, such as cytokine inhibitors and monoclonal antibodies, which pose the risks of immunosuppression, toxicity, and uncertain long-term effects.

This study had several limitations, including its single-center design. We acknowledge unequal group sizes in this retrospective study (initially, we used hemofiltration as a last resort, consistent with global practice at the time, reflecting limited awareness and acceptance; however, observing better outcomes with early initiation, our practice shifted accordingly). Nevertheless, our study findings remain robust as APACHE-II and SOFA scores were comparable between the two groups, indicating similar baseline severity of illness. Despite predominantly using IL-6/EAA levels to decide hemofilter use (cytokine/endotoxin removal), financial constraints and high turnaround time for results limited the use of these tests as the sole guide for therapy in all patients. The second filter use was based on worsening clinical parameters or rising/positive IL-6 balance post-hemadsorption. Cost remains a significant barrier to hemadsorption use in Indian ICUs. In our setting, the filter selection was partly affordability-driven, limiting uniformity and leading to non-random allocation; however, there was no filter preference. Due to limited global availability of Hemofeel/Toraymyxin, fewer studies on these filters exist compared to CytoSorb, but available data show no clear advantage of one over the other. Prospective multicenter studies with balanced cohorts, addressing feasibility concerns, especially in resource-limited ICUs of low- and middle-income countries, are needed. Nevertheless, our study underscores the value of early adjunct hemadsorption in RSS.

Hemadsorption is a safe, feasible, and effective modality to mitigate cytokine/endotoxin overload, a key feature of sepsis. Early initiation of adjuvant hemadsorption therapy (within 24 hours of RSS onset) demonstrated survival benefit and reduced ICU/hospital stay. Our findings indicate that the timing of hemadsorption initiation is crucial in optimizing clinical outcomes, supporting the role of early extracorporeal cytokine removal in patients with RSS; this may be the cornerstone to improve sepsis outcomes.