Time to Initiation of Treatment with Polymyxin B Cartridge Hemoperfusion in Septic Shock Patients

The lipopolysaccharide of endotoxin has been well characterized as one of the most potent triggers of the inflammatory cascade during acute sepsis [1,2]. Hemoperfusion with a cartridge containing polymyxin B immobilized on polystyrene fibers (PMX) has been employed to selectively absorb endotoxin in patients with Gram-negative sepsis [3,4]. It has been demonstrated to effectively neutralize the pathogenic activity of endotoxin and has been shown to reduce organ dysfunction [3].

In patients with severe sepsis or septic shock, early detection of global tissue hypoxia and hemodynamic optimization within 6 h has been shown to reduce overall mortality [5]. Accordingly, the present study was carried out to verify whether early initiation of PMX improved the clinical outcome in patients with severe sepsis or septic shock.

This study was conducted at the intensive care units of 35 Japanese hospitals from April 2006 to March 2008. The ethics committee at each hospital approved the study protocol, and informed consent was obtained from each patient or a relative or other legal representative. Subjects were required to meet both of the following criteria: (1) presence of a systemic inflammatory response to infection according to the consensus definition of the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference Committee [6], and (2) either a systolic blood pressure <90 mm Hg after fluid infusion (20 ml/kg) or a sequential organ failure assessment (SOFA) cardiovascular score ≥2 (namely, at least dopamine was necessary to maintain blood pressure) [7]. The exclusion criteria included terminal cancer, terminal hematological illness, severe hepatic dysfunction (total bilirubin ≥10 mg/dl and hepaplastin test ≤40%), and refusal to give written consent.

The cartridge for PMX was produced by immobilizing polymyxin B on polystyrene fibers (0.005 g of polymyxin B to 1 g of fibers) by covalent bonding (Toraymyxin; Toray Industries, Tokyo, Japan). Patients received direct hemoperfusion with PMX in addition to standard medical therapy for sepsis. Vascular access was obtained via a double-lumen venous catheter. Hemoperfusion with PMX was usually carried out for 2 h at each session. The second PMX session was performed 24 h after the end of the first session. Nafamostat mesilate, low molecular weight heparin, or unfractionated heparin was used as the anticoagulant.

Patients were followed up for 28 days after enrollment, and clinical data were obtained from an established registry database and included the site of infection, the Acute Physiology and Chronic Health Evaluation (APACHE) II score [8], the SOFA score, results of bacterial culture and laboratory tests, hemodynamic variables, and therapeutic interventions. The plasma endotoxin level was measured by the Endospecy test [9] at baseline and 6 h after the first PMX session.

The primary end-point was the 28-day mortality rate. The secondary end-points included changes from baseline to 7 days of the mean arterial pressure (MAP), use of catecholamines, serum lactate level, PaO₂/F_iO₂ ratio, and SOFA score. The ventilator-free rate was assessed at 14 days.

Statistical analysis was performed by using commercially available SPSS software (SPSS Inc.). Results are presented as the median and 95% confidence interval (CI). Comparison of numerical values between two groups was done with either Student’s t test or the Mann-Whitney U test. Mortality was estimated by the Kaplan-Meier method and compared between two groups by the log-rank test. The influence of the time until PMX and the number of PMX sessions was estimated by Welch’s test. Ventilator-free days were compared with the Cox proportional hazards regression model after adjustment for the PaO₂/F_iO₂ ratio.

A total of 120 patients were enrolled. 28 patients were excluded because follow-up could not be continued for 28 days or because the time from the onset of shock until starting PMX was unclear. As a consequence, 92 patients were analyzed. Of these 92 patients, 41 underwent with PMX within 6 h after the diagnosis of septic shock (early group) and 51 received PMX after more than 6 h (late group). The two groups received the same basal therapy (including early goal-directed therapy, renal replacement therapy, PMX sessions, and ventilator support), had a similar distribution of age, gender, and body weight, and had a comparable severity of sepsis (APACHE II and SOFA scores), as well as similar bacterial species and etiologic factors (table 1). The time from the onset of shock to initiation of PMX was significantly shorter in the early group than in the late group.

The two groups did not differ with respect to the baseline MAP or PaO₂/F_iO₂ ratio, but the plasma level of lactate was higher in the early group than in the late group (table 2).

The 28-day mortality rate was 39% (16/41 patients) in the early group and 35% (18/51 patient) in the late group (p = 0.76; fig. 1).

In both the early and late groups, PMX improved MAP, the lactate level, the SOFA score, and the PaO₂/F_iO₂ ratio from baseline to day 7 (table 2). The plasma endotoxin level did not change significantly after the first PMX session, being 18.4 pg/ml (95% CI 3.5–33.3 pg/ml) at baseline and 27.7 pg/ml (95% CI 1.6–53.9 pg/ml) afterwards.

On day 7, the lactate level, MAP, SOFA score, and PaO₂/F_iO₂ were not significantly different between the two groups (table 2). However, weaning from catecholamines by day 7 after PMX was more frequent in the early group (22/41 patients) than in the late group (17/51 patients) (54 vs. 33%; p = 0.050). Also, the duration of ventilator support was significantly shorter in the early group than the late group (adjusted hazards ratio 2.06; 95% CI 1.04–4.09; p = 0.039) (fig. 2).

To evaluate whether the time until initiation of PMX influenced septic shock-induced organ dysfunction and the clinical outcome, we compared patients who received PMX within 6 h after the diagnosis of septic shock and patients who started treatment from 6 h onward. We found that starting PMX within 6 h after the diagnosis of septic shock significantly shortened the duration of ventilator support and also reduced the catecholamine requirement. PMX was effective for improvement of hypotension, hypoperfusion, and pulmonary oxygenation regardless of the timing of its initiation.

Our results were similar to the findings of previous studies [10,11,12], which showed that survivors commenced PMX earlier after the onset of shock than patients who died. In patients with severe sepsis or septic shock, early administration of appropriate antibiotics, corticosteroids, and recombinant human activated protein C are all associated with better survival [13]. Thus, initiation of such therapies as well as PMX as soon as possible after diagnosis of severe sepsis or septic shock may be very important for achieving a good outcome.

The clinical benefit of PMX for hypotension and hypoperfusion identified in the present study is in good agreement with previous reports [14,15,16], but the effect of PMX on pulmonary oxygenation is less clear. Only two randomized controlled trials have been published that assessed the effect of PMX on pulmonary oxygenation. Vincent et al. [17] reported that there was no significant improvement of the PaO₂/F_iO₂ ratio in the PMX group, while a more recent study by Cruz et al. [18] showed that the PaO₂/F_iO₂ ratio increased slightly in the PMX group compared with the conventional therapy group.

Both in vivo and in vitro studies have indicated that absorption of endotoxin and its removal from the circulation is a major mechanism by which PMX improves septic shock [3,19]. In this study, however, the endotoxin level did not decrease after PMX. Although some studies have demonstrated that PMX reduces the circulating endotoxin level in septic patients [2,10,14], another study did not confirm this [17]. Even though patients with little or no circulating endotoxin, including those with Gram-positive infections, were enrolled in our study, a beneficial effect of PMX on hypoperfusion and pulmonary oxygenation was observed regardless of the pathogens causing sepsis. Thus, the response to PMX may be related to another mechanism in addition to removal of endotoxin. There have been recent reports about adhesion of activated monocytes and neutrophils [20], as well as anandamide [21] and high mobility group box-1 [22], which act as paracrine mediators of hypotension [23], to the fibers of the polymyxin cartridge. Such findings may partially explain the efficacy of PMX, although the molecular mechanisms remain unclear.

Unfortunately, the small number of patients limited the power to detect significant differences of survival and catecholamine requirements. A much larger study would be required to more thoroughly investigate the potential improvement of mortality with early initiation of PMX.

The authors would like to thank Ms. Sanae Masuko, Minako Fukushima, Mr. Toru Arakane, and Hisataka Shoji for their valuable comments and advice. There was no funding source.

The members of Japan Sepsis Study Group are as follows (in alphabetical order): Dokkyo Medical University, Saitama: K. Kubota, H. Sakio; Fujita Health University, Aichi: N. Takeyama; Fukuoka University, Fukuoka: H. Ishikura; Gifu Prefectural Tajimi Hospital, Gifu: N. Mabuchi; International University of Health and Welfare Hospital, Tochigi: Y. Shinozawa; Jikei University School of Medicine, Tokyo: Y. Kase; Jikei University Kashiwa Hospital, Chiba: T. Koyama; JR Tokyo Hospital, Tokyo: H. Tanaka; Juntendo University Nerima Hospital, Tokyo: M. Suzuki; Kansai Medical University, Osaka: K. Nishi; Kariya Toyota General Hospital, Aichi: M. Miura; Kawakita General Hospital, Tokyo: Y. Murata; Kawasaki Municipal Kawasaki Hospital, Kanagawa: S. Ito; Kenwakai Hospital, Nagano: E. Kumagai; Kitakyusyu Kenwakai Ohtemachi Hospital, Fukuoka: T. Nishinaka; Koshigaya Hospital Dokkyo University School of Medicine, Saitama: K. Yamada; Masuko Memorial Hospital, Aichi: K. Shigematsu; Matsunami General Hospital, Gifu: S. Akamatsu; Mito Saiseikai General Hospital, Ibaraki: I. Ebihara; Nagano Municipal Hospital, Nagano: S. Tagami; Narita Memorial Hospital, Aichi: T. Ohbayashi; Nihon University School of Medicine Itabashi Hospital, Tokyo: K. Tanjo: Ohhama Daiichi Hospital, Okinawa: M. Uechi; Okazaki City Hospital, Aichi: M. Asada; Oita Sanai Medical Center, Oita: F. Tamai; Shinmatsudo Central General Hospital, Chiba: T. Nakamura; Sapporo-Hokuyu Hospital, Hokkaido: M. Yonekawa; Sassa General Hospital, Nagano: S. Takeuchi; Seirei Hamamatsu General Hospital, Shizuoka: T. Isozaki; Shiga University of Medical Science Hospital, Shiga: T. Obata; St. Luke’s International Hospital, Tokyo: S. Ishimatsu; Showa University, Kanagawa: Y. Takahashi; Tajima Emergency and Critical Care Medical Center, Hyogo: M. Kobayashi; Teikyo University School of Medicine, Tokyo: T. Sakamoto; Toho University Omori Medical Center, Tokyo: K. Yoshiwara; Wakayama Medical University, Wakayama: T. Naka; Yamagata Prefectural Medical Center for Acute Medicine, Yamagata: K. Morino.

The authors have no conflicts of interest to disclose.