Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Cost-effectiveness of polymyxin B hemoperfusion for septic shock: an observational study using a Japanese nationwide administrative database

Cost-effectiveness of polymyxin B hemoperfusion for septic shock: an observational study using a... Background Polymyxin B hemoperfusion (PMX) removes endotoxin from septic shock patients. Although the treat- ment has been clinically used for more than 20 years, its cost-effectiveness has not been evaluated in detail. Methods This study used the Japanese diagnosis procedure combination (DPC) administrative database from April 2018 to March 2021. We selected adult patients with a primary diagnosis of sepsis and the SOFA score at the sepsis diagnosis was between 7 and 12. The patients were divided into the PMX group that received PMX treatment and the control group that did not receive PMX. After adjusting the patient background by propensity score matching, we calculated the incremental cost-effectiveness ratio (ICER) by determining the difference in quality-adjusted life-year (QALY ) and medical cost between the PMX and the control groups. Results Nineteen thousand two hundred eighty-three patients were included in the study. Among them, 1492 patients received PMX treatment, and 17,791 did not. As a result of 1:3 propensity score matching, 965 patients in the PMX group and 2895 patients in the control group were selected and analyzed. Twenty-eight-day mortality and hospital mortality were significantly lower in the PMX group. The average medical cost per patient of the PMX group was 31,418 ± 21,144 Euro and that of the control group was 24,483 ± 21,762 Euro, with a difference of 6935 Euro. Life expectancy, life year-gained (LYG), and the QALY were 1.70, 0.86, and 0.60 years longer in the PMX group, respectively. The ICER was calculated to be 11,592 Euro/year, which was lower than the reported willingness-to-pay threshold of 38,462 Euro/year. Conclusion Polymyxin B hemoperfusion was shown to be an acceptable treatment in terms of the medical economy. Keywords Polymyxin B hemoperfusion, PMX, Sepsis, DPC database, Cost-effectiveness, Propensity score matching Background Sepsis is a life-threatening organ dysfunction caused by a *Correspondence: dysregulated host response to infection [1]. It is the lead- Kenji Fujimori fujimori@med.tohoku.ac.jp ing cause of death in the ICU, and the mortality rate is Department of Health Administration and Policy, Tohoku University very high when it progresses to septic shock [2]. Graduate School of Medicine, Sendai, Japan 2 The treatment of septic shock includes early adminis - Department of Health Policy and Informatics, Tokyo Medical and Dental University Graduate School of Medical and Dental Sciences, Bunkyo-Ku, tration of antimicrobial agents, infusion of fluids, and Tokyo, Japan administration of vasopressors. When these standard © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits 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:// creat iveco mmons. org/ licen ses/ by/4. 0/. Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 2 of 9 treatments are not successful, one of the option treat- transferred to other hospitals within 28  days without ments is blood purification therapy to remove toxins and recovery, or received their first PMX treatment other inflammatory mediators from the patient’s circulating than on the first or second day of sepsis diagnosis. blood. Direct hemoperfusion using polymyxin B-immo- We defined the first SOFA score record as the day of bilized fibers, polymyxin B hemoperfusion (PMX), is a sepsis diagnosis. We categorized patients who received treatment targeting endotoxin. This bacterial component PMX on the first or second day of sepsis diagnosis into triggers whole-body inflammation and causes organ dys - the PMX group and patients who did not receive PMX functions [3, 4]. into the control group. Numerous studies on PMX have shown its clinical effectiveness in improving hemodynamics and pulmo - Propensity score matching nary functions of septic shock patients [5, 6]. On the other We performed a propensity score matching analysis hand, results of several randomized controlled studies between patients with PMX-treated (PMX group) and (RCT) on PMX evaluating its survival benefit are contro - non-treated (control group). We estimated the pro- versial [7, 8]. In observational studies using an extensive pensity score using a logistic regression model for the Japanese database, we have reported that PMX is effective use of PMX as a function of the following confounders: in reducing mortality and the number of days on organ age, gender, emergency versus elective hospital admis- support [9, 10]. In particular, we have found that the effec - sion, university hospital versus non-university hospitals, tiveness of PMX is higher in patients with a moderate admission to the emergency room (ER) or intensive care degree of organ dysfunction with sequential organ failure unit (ICU), Charlson Comorbidity Index (CCI), continu- assessment (SOFA) scores in the range of 7–12 [11]. ous renal replacement therapy (CRRT), hemodialysis PMX is a relatively expensive treatment compared to the (HD), mechanical ventilation, surgery, administration of standard therapies for septic shock, such as antimicrobials, γ-globulin, AT III, rTM, steroid, red blood cell transfu- fluid infusions, and vasopressors. In the Surviving Sepsis sion, platelet transfusion, and the maximum daily dose Campaign (SSC) guideline 2021 [12], PMX treatment is of noradrenaline. A one-to-three matched analysis using rated as “suggest against” (weakly recommended not to be the nearest-neighbor matching was performed based on used). One of the reasons for this recommendation is the the estimated propensity score of each patient. We used a high cost of the treatment. However, there has been little caliper width of 0.2 of the standard deviation of the pro- information on the cost-effectiveness of PMX treatment. pensity score. We evaluated the balance among covari- Therefore, it is an important issue to evaluate the validity ates using absolute standardized difference (ASD), which of the cost of this treatment to the therapeutic effect. considers a difference below 10% to be well balanced. In this study, we examined the cost-effectiveness of PMX treatment using a sizable Japanese inpatient data- Clinical effects base, the Diagnosis Procedure Combination (DPC). For the cohort after propensity score matching, we com- pared the following patient outcomes between the PMX Methods group and control group; 28-day-mortality, mortality at Data source and patients hospital discharge, length of hospital stay, ventilator-, This retrospective observational study used inpatient vasopressor-, and CRRT-free days at day 28. Patients who data included in the Japanese DPC database. The DPC were discharged alive and who died in the hospital were database contains discharge and administrative claims included in assessing the length of hospital stay. Free days data for all inpatients discharged from more than 1000 were counted as zero when a patient died before day 28. participating hospitals, covering 92% of all tertiary-care emergency hospitals in Japan [13, 14]. We extracted the patient data from April 2018 to Cost March 2021. Selected patients were aged 20  years or We obtained the medical costs for each patient from the older and whose primary diagnosis was sepsis based on day of sepsis diagnosis to the day of discharge from the the International Classification of Diseases 10th Revi - DPC data. We received the costs separately in the follow- sion (ICD-10) codes. Since we found in our previous ing categories; hospital visit fees and management fees, study that PMX is effective on the patients whose base - prescription, injection, treatment, surgery/anesthesia, line SOFA scores are 7–12, we targeted these patients in laboratory test, diagnostic imaging, rehabilitation, gen- this analysis. Therefore, we excluded the patients whose eral hospitalization fees, ER/ICU hospitalization fees, SOFA score is below seven or above 12. We also excluded and meals. The difference in total medical costs between the patients who died within 3  days after the diagno- the PMX group and the control groups was used to eval- sis of sepsis, were hospitalized for more than 125  days, uate cost-effectiveness. F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 3 of 9 All costs were obtained in Japanese yen (JPY) and con- with other sites of infection, and calculated ICERs for verted to Euro. The conversion to Euro was calculated at each combination. Patients with abdominal infections 130 JPY per Euro as the conversion rate as of March 2022. were identified primarily by using the ICD-10 code descriptions. Evaluation of cost‑effectiveness ratio The expected life expectancy after discharge was cal - Statistical analysis culated using each patient’s age and the Japanese life We reported continuous variables as mean and standard expectancy table. Patients who survived sepsis have a deviation (SD) and categorical variables as number and higher risk of death. A reduction rate of 0.51 for the life percentage. We performed statistical analysis using JMP expectancy has been reported for the adjustment of this Pro 15.2.0 (SAS Institute Inc.) We used the χ test (Pear- risk [15, 16]. We used this value to calculate the life-year son method for p value) to compare two groups for mor- gained (LYG) by multiplying the life expectancy by 0.51 tality and the Wilcoxon rank-sum test to compare the as the following formula. length of hospital stay and the free days. LYG = life expectancy after discharge × 0.51 Results Patients In addition, we multiplied LYG by a utility weight of During the study period, 108,323 patients were admit- 0.69 to estimate quality-adjusted life-year (QALY), as the ted with a primary diagnosis of sepsis. Among them, formula below, considering the lower quality of life (QOL) 19,283 patients were included in the study after exclud- after hospital discharge in sepsis survivors [15, 16]. ing 89,040 patients who were younger than 20 years old, QALY = LYG × 0.69 patients with SOFA scores less than seven or higher than 12, patients with missing SOFA score data, patients who The incremental cost-effectiveness ratio (ICER), died within three days after sepsis diagnosis, patients which is the difference in cost divided by the difference hospitalized for more than 125  days, patients trans- in expected life expectancy, is the most commonly used ferred to other facilities without recovery, or patients measure for determining the cost-effectiveness. We cal - treated with PMX after day 3. One thousand four hun- culated the ICER using the following formula dred ninety-two patients received PMX treatment among ICER =(the average cost of the PMX group − the average cost of the control group) ∕(the average QALY in the PMX group − the average QALY in the control group) the eligible patients, and 17,791 did not. The number of Sensitivity analysis shock patients, defined as cardiovascular SOFA score We obtained the adjustment mentioned above factors for of 2 or higher, was 1144/1492 (76.7%) for PMX-treated calculating LYG and QALY (0.51 and 0.69) from the old patients and 9573/17,791 (53.8%) for untreated patients. literature published before 2010, and the values are likely As a result of 1:3 propensity score matching, 965 patients to be inaccurate. In light of this uncertainty, we examined in the PMX group and 2895 patients in the control group the variation of ICERs when we changed the adjustment were selected (Fig. 1). factors. In one case, ICERs were calculated using reduced Table  1 shows the patient baseline before and after rates of 0.3 for the LYG calculation and 0.6 for the QALY propensity score matching. Before the matching, the calculation. In the other case, we used reduced rates of PMX group had a younger age, a higher proportion of 0.7 for the LYG calculation and 0.8 for the QALY calcu- emergency and ER/ICU admissions, and a higher pro- lation. Finally, we compared the results with the value portion of university hospitals than the control group. obtained under the standard condition (base case). The PMX group also had higher CRRT use, ventila - tor use, surgery rate, and higher rates of γ-globulin, Subgroup analysis rTM, AT III, steroid administration, and red blood cell As a subgroup analysis, we examined the impact of and platelet transfusions. The maximum daily dose of severity of organ damage and the cause of sepsis on noradrenaline was also higher in the PMX group. After cost-effectiveness. First, we stratified the patients by propensity score matching, the ASD was less than 10% SOFA score range into two groups, 7–9 and 10–12. for all factors, confirming that this study balanced ele - We also divided patients into two groups by the cause ments between the groups well. of sepsis, those with abdominal infection and those Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 4 of 9 Fig. 1 Patient selection flow Table 1 Baseline patient characteristics before and after propensity score matching Unmatched Matched PMX (n = 1492) Control (n = 17,791) ASD% PMX (n = 965) Control (n = 2895) ASD% Age, mean (SD) 72.7 (12.8) 77.2 (12.7) 35.1 72.9 (12.5) 73.9 (13.4) 0.2 Sex (male), n (%) 847 (46.8) 9792 (55.0) 13.6 548 (56.8) 1635 (56.5) 0.5 Emergency admission, n (%) 1336 (89.5) 17,196 (96.7) 25.3 883 (91.5) 2662 (92.0) 1.3 University hospital, n (%) 331 (22.2) 2130 (12.0) 23.3 186 (19.3) 565 (19.5) 0.5 ER/ICU admission, n (%) 1094 (73.3) 8237 (46.3) 45.9 670 (69.4) 2037 (70.4) 1.7 CCI, mean (SD) 1.9 (1.7) 1.9 (1.7) 1.6 1.9 (1.7) 2.0 (1.8) 1.3 CRRT, n (%) 950 (63.7) 1914 (10.8) 115.0 491 (50.9) 1418 (50.0) 1.5 HD, n (%) 71 (4.8) 734 (4.1) 2.5 59 (6.1) 160 (5.5) 2.0 Mechanical ventilation, n (%) 962 (64.5) 3901 (21.9) 79.6 528 (54.7) 1576 (54.4) 0.5 Surgery, n (%) 817 (57.8) 2443 (13.7) 89.8 418 (43.3) 1278 (44.2) 1.4 γ-globulin, n (%) 492 (33.9) 1313 (7.4) 60.5 242 (25.1) 672 (23.2) 3.6 rTM, n (%) 744 (49.9) 1604 (9.0) 89.8 356 (36.9) 960 (33.2) 6.4 AT III, n (%) 430 (28.8) 843 (4.7) 62.7 199 (20.6) 528 (18.2) 5.0 Steroid, n (%) 700 (46.9) 4614 (25.9) 37.3 428 (44.4) 1245 (43.0) 2.2 RBC transfusion, n (%) 572 (38.8) 2155 (12.1) 55.3 306 (31.7) 854 (29.5) 3.9 Platelet transfusion, n (%) 240 (16.1) 832 (4.7) 34.1 124 (12.9) 350 (12.1) 1.9 Maximum noradrenaline, mean (SD) 14.3 (12.9) 7.4 (9.9) 60.2 12.9 (11.6) 12.8 (13.1) 0.6 ASD Absolute standard difference, SD Standard deviation, ER Emergency room, ICU Intensive care unit, CCI Charlson Comorbidity Index, CRRT Continuous renal replacement therapy, HD Hemodialysis, rTM recombinant thrombomodulin, AT Antithrombin, RBC Red blood cell Clinical effects discharge mortality rate was 24.4% in the PMX group Table  2 shows the comparison of the clinical outcomes and 34.1% in the control group (p < 0.0001), significantly between the PMX group and the control group. The lower in the PMX group. Ventilator-free, vasopressor- 28-day mortality rate was 16.8% in the PMX group and free, and CRRT-free days at day 28 were significantly 23.9% in the control group (p < 0.0001). And the hospital longer in the PMX group, showing better recovery of F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 5 of 9 Table 2 Comparison of the clinical outcomes in the matched cohort PMX (n = 965) Control (n = 2895) P value 28-day mortality, n (%) 162 (16.8) 692 (23.9) < 0.0001 Hospital mortality, n (%) 235 (24.4) 987 (34.1) < 0.0001 Length of hospital stay, days, median (IQR) 32 (18–51) 25 (15–45) < 0.0001 Ventilator free days, days, median (IQR) 19 (0–24) 15 (0–23) < 0.0001 Vasopressor free days, days, median (IQR) 25 (20–26) 24 (0–26) < 0.0001 CRRT free days (day), days, median (IQR) 23 (13–25) 20 (0–24) < 0.0001 IQR Interquartile range, CRRT Continuous renal replacement therapy organ dysfunctions. For the length of hospital stay, the the average length of hospital stay was longer in the PMX PMX group was substantially longer than the control group than the control group. group (median 32 days vs. 25 days). This finding probably Table  3 also shows comparisons of life expectancy, reflects that more people in the control group died in the LYG, and QALY. Life expectancy was 12.93  years in hospital in a short time. the PMX group and 11.23  years in the control group, 1.70  years longer in the PMX group. LYG, calculated Cost‑effectiveness ratio by multiplying life expectancy by 0.51, was 0.86  years Table 3 shows the calculation results of the average medi- longer in the PMX group. QALY was calculated by mul- cal costs from the day of sepsis diagnosis to the day of tiplying LYG by 0.69 and was 0.60  years longer in the discharge. The total medical cost per patient of the PMX PMX group. The ICER calculated from the differences group was 31,418 ± 21,144 Euro and that of the control in cost and in QALY was 11,592 Euro/year (1,507,019 group was 24,483 ± 21,762 Euro. A difference of 6935 JPY/year). Euro indicates that the PMX group had higher medi- cal costs. Of the various cost categories, the largest dif- Sensitivity analysis ference was in the “treatment” cost, probably because it We performed a sensitivity analysis to examine the includes the cost of PMX treatment. “Injections,” “aur- effects of changing the adjustment factors used to esti - gery/anesthesia,” “diagnostic imaging,” “rehabilitation,” mate LYG and QALY from the life expectancy years. The “general hospitalization fee,” and “meals” were also more results are shown in Table 4. The ICER for the combina - expensive in the PMX group. This result may reflect that tion of 0.3 for the reduction rate and 0.6 for utility weight Table 3 Cost and clinical effects PMX (n = 965) Control (n = 2895) Difference P value Hospital visit/management (SD) 142 (119) 126 (118) 16 0.0003 Prescription (SD) 250 (497) 247 (614) 3 0.8693 Injection (SD) 5677 (8175) 4827 (6552) 850 < 0.0001 Treatment (SD) 7514 (4564) 3053 (4821) 4461 < 0.0001 Surgery/anesthesia (SD) 4321 (7032) 3724 (8208) 597 0.0261 Laboratory test (SD) 1045 (1268) 1004 (1482) 40 0.3644 Diagnostic imaging (SD) 749 (579) 663 (625) 86 < 0.0001 Rehabilitation (SD) 726 (873) 57 (819) 190 < 0.0001 General hospitalization fee (SD) 4493 (3374) 4013 (3560) 480 < 0.0001 ER/ICU hospitalization fee (SD) 6093 (4564) 5947 (4726) 146 0.3995 Meals (SD) 408 (357) 341 (362) 67 < 0.0001 Total cost (SD) 31,418 (21,144) 24,483 (21,762) 6935 < 0.0001 Life expectancy (years) 12.93 11.23 1.70 LYG (years) 6.59 5.73 0.86 QALY (years) 4.55 3.95 0.60 ICER (Euro/year) 11,592 Costs are expressed as mean (SD) in Euro SD Standard deviation, LYG Life year gained, QALY Quality-adjusted life-year, ICER Incremental cost- effectiveness ratio Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 6 of 9 Table 4 Sensitivity analyses Reduction Utility weight for Difference in Difference in life Difference in Difference in ICER (Euro/year) rate for LYG QALY estimation cost (Euro) expectancy (yrs) LYG (years) QALY (years) estimation Case 1 (base case) 0.51 0.69 6935 1.70 0.87 0.60 11,592 Case 2 0.3 0.6 6935 1.70 0.51 0.31 22,663 Case 3 0.7 0.8 6935 1.70 1.19 0.95 7285 LYG Life year gained, QALY Quality-adjusted life-year, ICER Incremental cost-effectiveness ratio was 22,663 Euro/year. The ICER for the combination of are derived from old studies, and the values are likely to 0.7 and 0.8 was estimated to be 7285 Euro/year. be inaccurate. Therefore, we calculated ICERs for cases with smaller values (0.3 and 0.6) and larger values (0.7 Subgroup analysis and 0.8) for both reduction factors as a sensitivity analy- Table  5 shows the results of the subgroup analysis com- sis to see the impact of these uncertainties. As a result, bining SOFA scores stratified into the 7–9 and 10–12 we calculated ICERs to be 7285 Euro and 22,663 Euro, ranges and the site of infection divided into abdomen respectively, showing that the ICER is below the accept- and others. The lowest ICER was observed in the patients able WTP threshold of ICER (38,462 Euro/QALY) even with abdominal infections with SOFA 10–12 (4,102 in the worst case. Euro/year). Conversely, the highest ICER was observed in In a subgroup analysis, we stratified patients by the patients with non-abdominal infections with SOFA 7–9 SOFA score (range 7–9 vs. 10–12) and the cause of sep- (13,263 Euro/year). sis (into abdominal infection vs. others). For SOFA score, ICER values were lower for 10–12 than 7–9. For the site Discussion of infection, ICERs were lower for abdominal infection This study evaluated the cost-effectiveness of PMX than for others. However, the ICER values for all sub- treatment by using the Japanese nationwide inpatient groups were within the acceptable WTP threshold. database, the DPC data. We compared the propensity One previous study on the cost-effectiveness of PMX score-matched cohort in the SOFA score range of 7–12. reported an incremental cost per LYG of 3864 Euros [20]. We found the QALY was longer in the PMX group by However, this study analyzed data from one randomized 0.60  years. The cost was higher in the PMX group by controlled trial conducted in Italy (the EUPHAS study). 6935 Euro. The ICER was calculated to be 11,592 Euro/ This study limited the number of analyzed patients to 64. year (1,507,019 JPY/year). In addition, the study was conducted in the years 2004– Willingness-to-pay threshold (WTP) is an estimate of 2007, about 15  years ago. We assume the overall treat- what a consumer of health care is prepared to pay for the ment of sepsis may have been different in some respects health benefit. Reportedly, the willingness-to-pay thresh - from what it is today. Our study used new real-world data old for QALY in Japan is 38,462 Euro/QALY (5 million obtained from 2018 to 2021, and the number of patients JPY/QALY) [17–19]. Compared to this value, the ICER of analyzed was approximately 4000 cases. Therefore, we PMX treatment is sufficiently low, indicating that it is an believe our results are more accurate and reflect actual acceptable treatment in terms of the medical economy. clinical practice in Japan. Close examination of the cost difference between There are several limitations to this study. First, patients with and without PMX treatment shows that although we performed propensity score matching, we the increase in the cost of the “treatment” is the largest at cannot rule out the existence of unadjusted confounders 4461 Euros, reflecting the cost of PMX treatment. Other not included in the DPC data. Second, the disease code of costs, such as “injection” and “hospitalization” were also sepsis is based on clinical judgement and not base on the slightly higher in the PMX group. This result may reflect SEPSIS-3 definition. Third, blood test values, including a longer hospital stay in the PMX group reflecting the blood cell counts, biomarkers, and endotoxin levels, were lower in-hospital mortality. not available from the DPC data. Fourth, survivors’ medi- Sepsis survivors have shorter life expectancy after cal conditions, treatment costs, and life expectancy vary being discharged than healthy people. To adjust for this from country to country. Fifth, the adjustment factors fact, we calculated LYG by multiplying the life expec- used for the calculation of LYG and QALY were derived tancy by a factor of 0.51, based on a previous study. In from the old literature and may not be precise. With this addition, 0.69 was used as the utility weight when calcu- in mind, we performed a sensitivity analysis with varying lating QALY from LYG. However, these reduction rates adjustment factors. Finally, the results of this study are F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 7 of 9 Table 5 Subgroup analyses SOFA range Cause of sepsis The number of Hospital mortality Total cost (Euro) Difference in Difference in life Difference of Difference of ICER (Euro/year) patients (n) (%) cost (Euro) expectancy (yrs) LYG (years) QALY (years) PMX Control PMX Control PMX Control 7–9 Abdominal 153 274 15.7 34.3 31,734 25,782 5952 3.40 1.73 1.20 4974 7–9 Others 333 1190 24.0 28.7 28,374 21,373 7001 1.50 0.77 0.53 13,263 10–12 Abdominal 153 305 24.8 44.6 35,723 31,883 3840 2.66 1.36 0.94 4102 10–12 Others 326 1126 28.5 37.0 32,359 25,448 6911 1.58 0.81 0.56 12,429 SOFA Sequential organ failure assessment, LYG Life year gained, QALY Quality-adjusted life-year, ICER Incremental cost-effectiveness ratio Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 8 of 9 Consent for publication an evaluation of cost-effectiveness using data from Japan. Not applicable. Therefore, readers may not extrapolate them directly to other countries. Competing interests The authors declare that they have no competing interests. Conclusions Our study evaluated the cost-effectiveness of polymyxin Received: 22 December 2022 Accepted: 13 January 2023 B hemoperfusion treatment for sepsis using a Japanese nationwide administrative database. The treatment’s incremental cost-effectiveness ratio (ICER) was cal - References culated as 11,592 Euro/QALY using propensity score- 1. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, matched cohorts. This cost was sufficiently low compared Bauer M et al (2016) The third international consensus definitions for to the reported willingness-to-pay threshold. However, sepsis and septic shock (sepsis-3). JAMA 315:801–810 2. Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman due to limitations as an observational study, the clinical CS et al (2016) Developing a new definition and assessing new clinical efficacy of polymyxin B hemoperfusion and its relation to criteria for septic shock: for the third international consensus defini- medical costs should be confirmed in future studies. tions for sepsis and septic shock (Sepsis-3). JAMA 315:775–787 3. Ronco C, Klein DJ (2014) Polymyxin B hemoperfusion: a mechanistic perspective. Crit Care 18:309 4. Shoji H (2003) Extracorporeal endotoxin removal for the treatment of Abbreviations sepsis: endotoxin adsorption cartridge ( Toraymyxin). Ther Apher Dial ASD Absolute standard difference 7:108–114 AT Antithrombin 5. Cruz DN, Antonelli M, Fumagalli R, Foltran F, Brienza N, Donati A et al CCI Char lson comorbidity index (2009) Early use of polymyxin B hemoperfusion in abdominal septic CRRT C ontinuous renal replacement therapy shock: the EUPHAS randomized controlled trial. JAMA 301:2445–2452 DPC Diagnosis procedure combination 6. Nakamura Y, Kitamura T, Kiyomi F, Hayakawa M, Hoshino K, Kawano Y ER Emergency room et al (2017) Potential survival benefit of polymyxin B hemoperfusion HD Hemodialysis in patients with septic shock: a propensity- matched cohort study. Crit ICER Incremental cost-effectiveness ratio Care 21:134 ICU Intensive care unit 7. Payen DM, Guilhot J, Launey Y, Lukaszewicz AC, Kaaki M, Veber B et al IQR Interquartile range (2015) Early use of polymyxin B hemoperfusion in patients with septic JPY Japanese yen shock due to peritonitis: a multicenter randomized control trial. Inten- LYG Life-year gained sive Care Med 41:975–984 PMX P olymyxin B hemoperfusion 8. Dellinger RP, Bagshaw SM, Antonelli M, Foster DM, Klein DJ, Marshall JC QALY Quality-adjusted life-year et al (2018) Eec ff t of targeted polymyxin B hemoperfusion on 28- day RBC Red blood cell mortality in patients with septic shock and elevated endotoxin level: RCT Randomized controlled trial the EUPHRATES randomized clinical trial. JAMA 320:1455–1463 rTM Recombinant thrombomodulin 9. Fujimori K, Tarasawa K, Fushimi K (2021) Eec ff ts of Polymyxin B hemop - SD Standard deviation erfusion on septic shock patients requiring noradrenaline: analysis of a SOFA Sequential organ failure assessment nationwide administrative database in Japan. Blood Purif 50:560–566 WTP Willingness to pay 10. Fujimori K, Tarasawa K, Fushimi K (2021) Eec ff ts of Polymyxin B hemop - erfusion in patients with sepsis requiring continuous hemodiafiltration: Acknowledgements analysis of a nationwide administrative database in Japan. Ther Apher The authors would like to thank Dr. Nobuo Ida , Ms. Ayuka Suzuki and Ms. Dial 25:384–389 Makiko Miyata of Toray Industries, Inc. for valuable discussions. 11. Fujimori K, Tarasawa K, Fushimi K (2021) Eec ff tiveness of polymyxin B hemoperfusion for sepsis depends on the baseline SOFA score: a Authors’ contributions nationwide observational study. Ann Intensive Care 11:141 K. Fujimori: responsible for the study concept, design, data extraction and 12. Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French analysis, and manuscript preparation. K.T.: accountable for the statistics and C et al (2021) Surviving sepsis campaign: international guidelines for manuscript preparation. K. Fushimi: responsible for the database organization management of sepsis and septic shock 2021. Intensive Care Med and review. The authors read and approved the final manuscript. 47:1181–1247 13. Matsuda S, Fujimori K, Fushimi K (2010) Development of casemix based Funding evaluation system in Japan. Asian Pac J Dis Manag 4:55–66 Ministry of Health, Labor and Welfare grant supported this study. ( Tokyo, 14. Murata A, Matsuda S, Kuwabara K, Ichimiya Y, Fujino Y, Kubo T et al Japan; 20AA2005). (2011) Equivalent clinical outcomes of bleeding peptic ulcers in teaching and non-teaching hospitals: evidence for standardization of Availability of data and materials medical care in Japan. Tohoku J Exp Med 223:1–7 The datasets used and/or analyzed during the current study are available from 15. Jones AE, Troyer JL, Kline JA (2011) Cost-effectiveness of an emergency the corresponding author on reasonable request. department-based early sepsis resuscitation protocol. Crit Care Med 39(6):1306–1312 Declarations 16. Suarez D, Ferrer R, Artigas A, Azkarate I, Garnacho-Montero J, Gomà G, Levy MM, Ruiz JC, Edusepsis Study Group (2011) Cost-effectiveness of Ethics approval and consent to participate the surviving sepsis campaign protocol for severe sepsis: a prospective The study was approved by the Institutional Review Board of Tohoku Univer- nation-wide study in Spain. Intensive Care Med 37(3):444–52 sity, which waived the requirement for informed patient consent because of 17. Shiroiwa T, Sung YK, Fukuda T, Lang HC, Bae SC, Tsutani K (2010) the anonymous nature of the data. International survey on willingness-to-pay ( WTP) for one additional QALY gained: what is the threshold of cost effectiveness? Health Econ 19(4):422–437 F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 9 of 9 18. Shiroiwa T, Igarashi A, Fukuda T, Ikeda S (2013) WTP for a QALY and health states: more money for severer health states? Cost Eff Resour Alloc 1(11):22 19. Fukuda T, Shiroiwa T (2019) Application of economic evaluation of pharmaceuticals and medical devices in Japan. J Natl Inst Public Health 68:27–33 20. Berto P, Ronco C, Cruz D, Melotti RM, Antonelli M (2011) Cost-effective - ness analysis of polymyxin-B immobilized fiber column and conventional medical therapy in the management of abdominal septic shock in Italy. Blood Purif 32(4):331–340 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Anesthesia Analgesia and Critical Care Springer Journals

Cost-effectiveness of polymyxin B hemoperfusion for septic shock: an observational study using a Japanese nationwide administrative database

Loading next page...
 
/lp/springer-journals/cost-effectiveness-of-polymyxin-b-hemoperfusion-for-septic-shock-an-0SoU7l9Pbj

References (23)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2023
eISSN
2731-3786
DOI
10.1186/s44158-023-00087-6
Publisher site
See Article on Publisher Site

Abstract

Background Polymyxin B hemoperfusion (PMX) removes endotoxin from septic shock patients. Although the treat- ment has been clinically used for more than 20 years, its cost-effectiveness has not been evaluated in detail. Methods This study used the Japanese diagnosis procedure combination (DPC) administrative database from April 2018 to March 2021. We selected adult patients with a primary diagnosis of sepsis and the SOFA score at the sepsis diagnosis was between 7 and 12. The patients were divided into the PMX group that received PMX treatment and the control group that did not receive PMX. After adjusting the patient background by propensity score matching, we calculated the incremental cost-effectiveness ratio (ICER) by determining the difference in quality-adjusted life-year (QALY ) and medical cost between the PMX and the control groups. Results Nineteen thousand two hundred eighty-three patients were included in the study. Among them, 1492 patients received PMX treatment, and 17,791 did not. As a result of 1:3 propensity score matching, 965 patients in the PMX group and 2895 patients in the control group were selected and analyzed. Twenty-eight-day mortality and hospital mortality were significantly lower in the PMX group. The average medical cost per patient of the PMX group was 31,418 ± 21,144 Euro and that of the control group was 24,483 ± 21,762 Euro, with a difference of 6935 Euro. Life expectancy, life year-gained (LYG), and the QALY were 1.70, 0.86, and 0.60 years longer in the PMX group, respectively. The ICER was calculated to be 11,592 Euro/year, which was lower than the reported willingness-to-pay threshold of 38,462 Euro/year. Conclusion Polymyxin B hemoperfusion was shown to be an acceptable treatment in terms of the medical economy. Keywords Polymyxin B hemoperfusion, PMX, Sepsis, DPC database, Cost-effectiveness, Propensity score matching Background Sepsis is a life-threatening organ dysfunction caused by a *Correspondence: dysregulated host response to infection [1]. It is the lead- Kenji Fujimori fujimori@med.tohoku.ac.jp ing cause of death in the ICU, and the mortality rate is Department of Health Administration and Policy, Tohoku University very high when it progresses to septic shock [2]. Graduate School of Medicine, Sendai, Japan 2 The treatment of septic shock includes early adminis - Department of Health Policy and Informatics, Tokyo Medical and Dental University Graduate School of Medical and Dental Sciences, Bunkyo-Ku, tration of antimicrobial agents, infusion of fluids, and Tokyo, Japan administration of vasopressors. When these standard © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits 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:// creat iveco mmons. org/ licen ses/ by/4. 0/. Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 2 of 9 treatments are not successful, one of the option treat- transferred to other hospitals within 28  days without ments is blood purification therapy to remove toxins and recovery, or received their first PMX treatment other inflammatory mediators from the patient’s circulating than on the first or second day of sepsis diagnosis. blood. Direct hemoperfusion using polymyxin B-immo- We defined the first SOFA score record as the day of bilized fibers, polymyxin B hemoperfusion (PMX), is a sepsis diagnosis. We categorized patients who received treatment targeting endotoxin. This bacterial component PMX on the first or second day of sepsis diagnosis into triggers whole-body inflammation and causes organ dys - the PMX group and patients who did not receive PMX functions [3, 4]. into the control group. Numerous studies on PMX have shown its clinical effectiveness in improving hemodynamics and pulmo - Propensity score matching nary functions of septic shock patients [5, 6]. On the other We performed a propensity score matching analysis hand, results of several randomized controlled studies between patients with PMX-treated (PMX group) and (RCT) on PMX evaluating its survival benefit are contro - non-treated (control group). We estimated the pro- versial [7, 8]. In observational studies using an extensive pensity score using a logistic regression model for the Japanese database, we have reported that PMX is effective use of PMX as a function of the following confounders: in reducing mortality and the number of days on organ age, gender, emergency versus elective hospital admis- support [9, 10]. In particular, we have found that the effec - sion, university hospital versus non-university hospitals, tiveness of PMX is higher in patients with a moderate admission to the emergency room (ER) or intensive care degree of organ dysfunction with sequential organ failure unit (ICU), Charlson Comorbidity Index (CCI), continu- assessment (SOFA) scores in the range of 7–12 [11]. ous renal replacement therapy (CRRT), hemodialysis PMX is a relatively expensive treatment compared to the (HD), mechanical ventilation, surgery, administration of standard therapies for septic shock, such as antimicrobials, γ-globulin, AT III, rTM, steroid, red blood cell transfu- fluid infusions, and vasopressors. In the Surviving Sepsis sion, platelet transfusion, and the maximum daily dose Campaign (SSC) guideline 2021 [12], PMX treatment is of noradrenaline. A one-to-three matched analysis using rated as “suggest against” (weakly recommended not to be the nearest-neighbor matching was performed based on used). One of the reasons for this recommendation is the the estimated propensity score of each patient. We used a high cost of the treatment. However, there has been little caliper width of 0.2 of the standard deviation of the pro- information on the cost-effectiveness of PMX treatment. pensity score. We evaluated the balance among covari- Therefore, it is an important issue to evaluate the validity ates using absolute standardized difference (ASD), which of the cost of this treatment to the therapeutic effect. considers a difference below 10% to be well balanced. In this study, we examined the cost-effectiveness of PMX treatment using a sizable Japanese inpatient data- Clinical effects base, the Diagnosis Procedure Combination (DPC). For the cohort after propensity score matching, we com- pared the following patient outcomes between the PMX Methods group and control group; 28-day-mortality, mortality at Data source and patients hospital discharge, length of hospital stay, ventilator-, This retrospective observational study used inpatient vasopressor-, and CRRT-free days at day 28. Patients who data included in the Japanese DPC database. The DPC were discharged alive and who died in the hospital were database contains discharge and administrative claims included in assessing the length of hospital stay. Free days data for all inpatients discharged from more than 1000 were counted as zero when a patient died before day 28. participating hospitals, covering 92% of all tertiary-care emergency hospitals in Japan [13, 14]. We extracted the patient data from April 2018 to Cost March 2021. Selected patients were aged 20  years or We obtained the medical costs for each patient from the older and whose primary diagnosis was sepsis based on day of sepsis diagnosis to the day of discharge from the the International Classification of Diseases 10th Revi - DPC data. We received the costs separately in the follow- sion (ICD-10) codes. Since we found in our previous ing categories; hospital visit fees and management fees, study that PMX is effective on the patients whose base - prescription, injection, treatment, surgery/anesthesia, line SOFA scores are 7–12, we targeted these patients in laboratory test, diagnostic imaging, rehabilitation, gen- this analysis. Therefore, we excluded the patients whose eral hospitalization fees, ER/ICU hospitalization fees, SOFA score is below seven or above 12. We also excluded and meals. The difference in total medical costs between the patients who died within 3  days after the diagno- the PMX group and the control groups was used to eval- sis of sepsis, were hospitalized for more than 125  days, uate cost-effectiveness. F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 3 of 9 All costs were obtained in Japanese yen (JPY) and con- with other sites of infection, and calculated ICERs for verted to Euro. The conversion to Euro was calculated at each combination. Patients with abdominal infections 130 JPY per Euro as the conversion rate as of March 2022. were identified primarily by using the ICD-10 code descriptions. Evaluation of cost‑effectiveness ratio The expected life expectancy after discharge was cal - Statistical analysis culated using each patient’s age and the Japanese life We reported continuous variables as mean and standard expectancy table. Patients who survived sepsis have a deviation (SD) and categorical variables as number and higher risk of death. A reduction rate of 0.51 for the life percentage. We performed statistical analysis using JMP expectancy has been reported for the adjustment of this Pro 15.2.0 (SAS Institute Inc.) We used the χ test (Pear- risk [15, 16]. We used this value to calculate the life-year son method for p value) to compare two groups for mor- gained (LYG) by multiplying the life expectancy by 0.51 tality and the Wilcoxon rank-sum test to compare the as the following formula. length of hospital stay and the free days. LYG = life expectancy after discharge × 0.51 Results Patients In addition, we multiplied LYG by a utility weight of During the study period, 108,323 patients were admit- 0.69 to estimate quality-adjusted life-year (QALY), as the ted with a primary diagnosis of sepsis. Among them, formula below, considering the lower quality of life (QOL) 19,283 patients were included in the study after exclud- after hospital discharge in sepsis survivors [15, 16]. ing 89,040 patients who were younger than 20 years old, QALY = LYG × 0.69 patients with SOFA scores less than seven or higher than 12, patients with missing SOFA score data, patients who The incremental cost-effectiveness ratio (ICER), died within three days after sepsis diagnosis, patients which is the difference in cost divided by the difference hospitalized for more than 125  days, patients trans- in expected life expectancy, is the most commonly used ferred to other facilities without recovery, or patients measure for determining the cost-effectiveness. We cal - treated with PMX after day 3. One thousand four hun- culated the ICER using the following formula dred ninety-two patients received PMX treatment among ICER =(the average cost of the PMX group − the average cost of the control group) ∕(the average QALY in the PMX group − the average QALY in the control group) the eligible patients, and 17,791 did not. The number of Sensitivity analysis shock patients, defined as cardiovascular SOFA score We obtained the adjustment mentioned above factors for of 2 or higher, was 1144/1492 (76.7%) for PMX-treated calculating LYG and QALY (0.51 and 0.69) from the old patients and 9573/17,791 (53.8%) for untreated patients. literature published before 2010, and the values are likely As a result of 1:3 propensity score matching, 965 patients to be inaccurate. In light of this uncertainty, we examined in the PMX group and 2895 patients in the control group the variation of ICERs when we changed the adjustment were selected (Fig. 1). factors. In one case, ICERs were calculated using reduced Table  1 shows the patient baseline before and after rates of 0.3 for the LYG calculation and 0.6 for the QALY propensity score matching. Before the matching, the calculation. In the other case, we used reduced rates of PMX group had a younger age, a higher proportion of 0.7 for the LYG calculation and 0.8 for the QALY calcu- emergency and ER/ICU admissions, and a higher pro- lation. Finally, we compared the results with the value portion of university hospitals than the control group. obtained under the standard condition (base case). The PMX group also had higher CRRT use, ventila - tor use, surgery rate, and higher rates of γ-globulin, Subgroup analysis rTM, AT III, steroid administration, and red blood cell As a subgroup analysis, we examined the impact of and platelet transfusions. The maximum daily dose of severity of organ damage and the cause of sepsis on noradrenaline was also higher in the PMX group. After cost-effectiveness. First, we stratified the patients by propensity score matching, the ASD was less than 10% SOFA score range into two groups, 7–9 and 10–12. for all factors, confirming that this study balanced ele - We also divided patients into two groups by the cause ments between the groups well. of sepsis, those with abdominal infection and those Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 4 of 9 Fig. 1 Patient selection flow Table 1 Baseline patient characteristics before and after propensity score matching Unmatched Matched PMX (n = 1492) Control (n = 17,791) ASD% PMX (n = 965) Control (n = 2895) ASD% Age, mean (SD) 72.7 (12.8) 77.2 (12.7) 35.1 72.9 (12.5) 73.9 (13.4) 0.2 Sex (male), n (%) 847 (46.8) 9792 (55.0) 13.6 548 (56.8) 1635 (56.5) 0.5 Emergency admission, n (%) 1336 (89.5) 17,196 (96.7) 25.3 883 (91.5) 2662 (92.0) 1.3 University hospital, n (%) 331 (22.2) 2130 (12.0) 23.3 186 (19.3) 565 (19.5) 0.5 ER/ICU admission, n (%) 1094 (73.3) 8237 (46.3) 45.9 670 (69.4) 2037 (70.4) 1.7 CCI, mean (SD) 1.9 (1.7) 1.9 (1.7) 1.6 1.9 (1.7) 2.0 (1.8) 1.3 CRRT, n (%) 950 (63.7) 1914 (10.8) 115.0 491 (50.9) 1418 (50.0) 1.5 HD, n (%) 71 (4.8) 734 (4.1) 2.5 59 (6.1) 160 (5.5) 2.0 Mechanical ventilation, n (%) 962 (64.5) 3901 (21.9) 79.6 528 (54.7) 1576 (54.4) 0.5 Surgery, n (%) 817 (57.8) 2443 (13.7) 89.8 418 (43.3) 1278 (44.2) 1.4 γ-globulin, n (%) 492 (33.9) 1313 (7.4) 60.5 242 (25.1) 672 (23.2) 3.6 rTM, n (%) 744 (49.9) 1604 (9.0) 89.8 356 (36.9) 960 (33.2) 6.4 AT III, n (%) 430 (28.8) 843 (4.7) 62.7 199 (20.6) 528 (18.2) 5.0 Steroid, n (%) 700 (46.9) 4614 (25.9) 37.3 428 (44.4) 1245 (43.0) 2.2 RBC transfusion, n (%) 572 (38.8) 2155 (12.1) 55.3 306 (31.7) 854 (29.5) 3.9 Platelet transfusion, n (%) 240 (16.1) 832 (4.7) 34.1 124 (12.9) 350 (12.1) 1.9 Maximum noradrenaline, mean (SD) 14.3 (12.9) 7.4 (9.9) 60.2 12.9 (11.6) 12.8 (13.1) 0.6 ASD Absolute standard difference, SD Standard deviation, ER Emergency room, ICU Intensive care unit, CCI Charlson Comorbidity Index, CRRT Continuous renal replacement therapy, HD Hemodialysis, rTM recombinant thrombomodulin, AT Antithrombin, RBC Red blood cell Clinical effects discharge mortality rate was 24.4% in the PMX group Table  2 shows the comparison of the clinical outcomes and 34.1% in the control group (p < 0.0001), significantly between the PMX group and the control group. The lower in the PMX group. Ventilator-free, vasopressor- 28-day mortality rate was 16.8% in the PMX group and free, and CRRT-free days at day 28 were significantly 23.9% in the control group (p < 0.0001). And the hospital longer in the PMX group, showing better recovery of F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 5 of 9 Table 2 Comparison of the clinical outcomes in the matched cohort PMX (n = 965) Control (n = 2895) P value 28-day mortality, n (%) 162 (16.8) 692 (23.9) < 0.0001 Hospital mortality, n (%) 235 (24.4) 987 (34.1) < 0.0001 Length of hospital stay, days, median (IQR) 32 (18–51) 25 (15–45) < 0.0001 Ventilator free days, days, median (IQR) 19 (0–24) 15 (0–23) < 0.0001 Vasopressor free days, days, median (IQR) 25 (20–26) 24 (0–26) < 0.0001 CRRT free days (day), days, median (IQR) 23 (13–25) 20 (0–24) < 0.0001 IQR Interquartile range, CRRT Continuous renal replacement therapy organ dysfunctions. For the length of hospital stay, the the average length of hospital stay was longer in the PMX PMX group was substantially longer than the control group than the control group. group (median 32 days vs. 25 days). This finding probably Table  3 also shows comparisons of life expectancy, reflects that more people in the control group died in the LYG, and QALY. Life expectancy was 12.93  years in hospital in a short time. the PMX group and 11.23  years in the control group, 1.70  years longer in the PMX group. LYG, calculated Cost‑effectiveness ratio by multiplying life expectancy by 0.51, was 0.86  years Table 3 shows the calculation results of the average medi- longer in the PMX group. QALY was calculated by mul- cal costs from the day of sepsis diagnosis to the day of tiplying LYG by 0.69 and was 0.60  years longer in the discharge. The total medical cost per patient of the PMX PMX group. The ICER calculated from the differences group was 31,418 ± 21,144 Euro and that of the control in cost and in QALY was 11,592 Euro/year (1,507,019 group was 24,483 ± 21,762 Euro. A difference of 6935 JPY/year). Euro indicates that the PMX group had higher medi- cal costs. Of the various cost categories, the largest dif- Sensitivity analysis ference was in the “treatment” cost, probably because it We performed a sensitivity analysis to examine the includes the cost of PMX treatment. “Injections,” “aur- effects of changing the adjustment factors used to esti - gery/anesthesia,” “diagnostic imaging,” “rehabilitation,” mate LYG and QALY from the life expectancy years. The “general hospitalization fee,” and “meals” were also more results are shown in Table 4. The ICER for the combina - expensive in the PMX group. This result may reflect that tion of 0.3 for the reduction rate and 0.6 for utility weight Table 3 Cost and clinical effects PMX (n = 965) Control (n = 2895) Difference P value Hospital visit/management (SD) 142 (119) 126 (118) 16 0.0003 Prescription (SD) 250 (497) 247 (614) 3 0.8693 Injection (SD) 5677 (8175) 4827 (6552) 850 < 0.0001 Treatment (SD) 7514 (4564) 3053 (4821) 4461 < 0.0001 Surgery/anesthesia (SD) 4321 (7032) 3724 (8208) 597 0.0261 Laboratory test (SD) 1045 (1268) 1004 (1482) 40 0.3644 Diagnostic imaging (SD) 749 (579) 663 (625) 86 < 0.0001 Rehabilitation (SD) 726 (873) 57 (819) 190 < 0.0001 General hospitalization fee (SD) 4493 (3374) 4013 (3560) 480 < 0.0001 ER/ICU hospitalization fee (SD) 6093 (4564) 5947 (4726) 146 0.3995 Meals (SD) 408 (357) 341 (362) 67 < 0.0001 Total cost (SD) 31,418 (21,144) 24,483 (21,762) 6935 < 0.0001 Life expectancy (years) 12.93 11.23 1.70 LYG (years) 6.59 5.73 0.86 QALY (years) 4.55 3.95 0.60 ICER (Euro/year) 11,592 Costs are expressed as mean (SD) in Euro SD Standard deviation, LYG Life year gained, QALY Quality-adjusted life-year, ICER Incremental cost- effectiveness ratio Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 6 of 9 Table 4 Sensitivity analyses Reduction Utility weight for Difference in Difference in life Difference in Difference in ICER (Euro/year) rate for LYG QALY estimation cost (Euro) expectancy (yrs) LYG (years) QALY (years) estimation Case 1 (base case) 0.51 0.69 6935 1.70 0.87 0.60 11,592 Case 2 0.3 0.6 6935 1.70 0.51 0.31 22,663 Case 3 0.7 0.8 6935 1.70 1.19 0.95 7285 LYG Life year gained, QALY Quality-adjusted life-year, ICER Incremental cost-effectiveness ratio was 22,663 Euro/year. The ICER for the combination of are derived from old studies, and the values are likely to 0.7 and 0.8 was estimated to be 7285 Euro/year. be inaccurate. Therefore, we calculated ICERs for cases with smaller values (0.3 and 0.6) and larger values (0.7 Subgroup analysis and 0.8) for both reduction factors as a sensitivity analy- Table  5 shows the results of the subgroup analysis com- sis to see the impact of these uncertainties. As a result, bining SOFA scores stratified into the 7–9 and 10–12 we calculated ICERs to be 7285 Euro and 22,663 Euro, ranges and the site of infection divided into abdomen respectively, showing that the ICER is below the accept- and others. The lowest ICER was observed in the patients able WTP threshold of ICER (38,462 Euro/QALY) even with abdominal infections with SOFA 10–12 (4,102 in the worst case. Euro/year). Conversely, the highest ICER was observed in In a subgroup analysis, we stratified patients by the patients with non-abdominal infections with SOFA 7–9 SOFA score (range 7–9 vs. 10–12) and the cause of sep- (13,263 Euro/year). sis (into abdominal infection vs. others). For SOFA score, ICER values were lower for 10–12 than 7–9. For the site Discussion of infection, ICERs were lower for abdominal infection This study evaluated the cost-effectiveness of PMX than for others. However, the ICER values for all sub- treatment by using the Japanese nationwide inpatient groups were within the acceptable WTP threshold. database, the DPC data. We compared the propensity One previous study on the cost-effectiveness of PMX score-matched cohort in the SOFA score range of 7–12. reported an incremental cost per LYG of 3864 Euros [20]. We found the QALY was longer in the PMX group by However, this study analyzed data from one randomized 0.60  years. The cost was higher in the PMX group by controlled trial conducted in Italy (the EUPHAS study). 6935 Euro. The ICER was calculated to be 11,592 Euro/ This study limited the number of analyzed patients to 64. year (1,507,019 JPY/year). In addition, the study was conducted in the years 2004– Willingness-to-pay threshold (WTP) is an estimate of 2007, about 15  years ago. We assume the overall treat- what a consumer of health care is prepared to pay for the ment of sepsis may have been different in some respects health benefit. Reportedly, the willingness-to-pay thresh - from what it is today. Our study used new real-world data old for QALY in Japan is 38,462 Euro/QALY (5 million obtained from 2018 to 2021, and the number of patients JPY/QALY) [17–19]. Compared to this value, the ICER of analyzed was approximately 4000 cases. Therefore, we PMX treatment is sufficiently low, indicating that it is an believe our results are more accurate and reflect actual acceptable treatment in terms of the medical economy. clinical practice in Japan. Close examination of the cost difference between There are several limitations to this study. First, patients with and without PMX treatment shows that although we performed propensity score matching, we the increase in the cost of the “treatment” is the largest at cannot rule out the existence of unadjusted confounders 4461 Euros, reflecting the cost of PMX treatment. Other not included in the DPC data. Second, the disease code of costs, such as “injection” and “hospitalization” were also sepsis is based on clinical judgement and not base on the slightly higher in the PMX group. This result may reflect SEPSIS-3 definition. Third, blood test values, including a longer hospital stay in the PMX group reflecting the blood cell counts, biomarkers, and endotoxin levels, were lower in-hospital mortality. not available from the DPC data. Fourth, survivors’ medi- Sepsis survivors have shorter life expectancy after cal conditions, treatment costs, and life expectancy vary being discharged than healthy people. To adjust for this from country to country. Fifth, the adjustment factors fact, we calculated LYG by multiplying the life expec- used for the calculation of LYG and QALY were derived tancy by a factor of 0.51, based on a previous study. In from the old literature and may not be precise. With this addition, 0.69 was used as the utility weight when calcu- in mind, we performed a sensitivity analysis with varying lating QALY from LYG. However, these reduction rates adjustment factors. Finally, the results of this study are F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 7 of 9 Table 5 Subgroup analyses SOFA range Cause of sepsis The number of Hospital mortality Total cost (Euro) Difference in Difference in life Difference of Difference of ICER (Euro/year) patients (n) (%) cost (Euro) expectancy (yrs) LYG (years) QALY (years) PMX Control PMX Control PMX Control 7–9 Abdominal 153 274 15.7 34.3 31,734 25,782 5952 3.40 1.73 1.20 4974 7–9 Others 333 1190 24.0 28.7 28,374 21,373 7001 1.50 0.77 0.53 13,263 10–12 Abdominal 153 305 24.8 44.6 35,723 31,883 3840 2.66 1.36 0.94 4102 10–12 Others 326 1126 28.5 37.0 32,359 25,448 6911 1.58 0.81 0.56 12,429 SOFA Sequential organ failure assessment, LYG Life year gained, QALY Quality-adjusted life-year, ICER Incremental cost-effectiveness ratio Fujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 8 of 9 Consent for publication an evaluation of cost-effectiveness using data from Japan. Not applicable. Therefore, readers may not extrapolate them directly to other countries. Competing interests The authors declare that they have no competing interests. Conclusions Our study evaluated the cost-effectiveness of polymyxin Received: 22 December 2022 Accepted: 13 January 2023 B hemoperfusion treatment for sepsis using a Japanese nationwide administrative database. The treatment’s incremental cost-effectiveness ratio (ICER) was cal - References culated as 11,592 Euro/QALY using propensity score- 1. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, matched cohorts. This cost was sufficiently low compared Bauer M et al (2016) The third international consensus definitions for to the reported willingness-to-pay threshold. However, sepsis and septic shock (sepsis-3). JAMA 315:801–810 2. Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman due to limitations as an observational study, the clinical CS et al (2016) Developing a new definition and assessing new clinical efficacy of polymyxin B hemoperfusion and its relation to criteria for septic shock: for the third international consensus defini- medical costs should be confirmed in future studies. tions for sepsis and septic shock (Sepsis-3). JAMA 315:775–787 3. Ronco C, Klein DJ (2014) Polymyxin B hemoperfusion: a mechanistic perspective. Crit Care 18:309 4. Shoji H (2003) Extracorporeal endotoxin removal for the treatment of Abbreviations sepsis: endotoxin adsorption cartridge ( Toraymyxin). Ther Apher Dial ASD Absolute standard difference 7:108–114 AT Antithrombin 5. Cruz DN, Antonelli M, Fumagalli R, Foltran F, Brienza N, Donati A et al CCI Char lson comorbidity index (2009) Early use of polymyxin B hemoperfusion in abdominal septic CRRT C ontinuous renal replacement therapy shock: the EUPHAS randomized controlled trial. JAMA 301:2445–2452 DPC Diagnosis procedure combination 6. Nakamura Y, Kitamura T, Kiyomi F, Hayakawa M, Hoshino K, Kawano Y ER Emergency room et al (2017) Potential survival benefit of polymyxin B hemoperfusion HD Hemodialysis in patients with septic shock: a propensity- matched cohort study. Crit ICER Incremental cost-effectiveness ratio Care 21:134 ICU Intensive care unit 7. Payen DM, Guilhot J, Launey Y, Lukaszewicz AC, Kaaki M, Veber B et al IQR Interquartile range (2015) Early use of polymyxin B hemoperfusion in patients with septic JPY Japanese yen shock due to peritonitis: a multicenter randomized control trial. Inten- LYG Life-year gained sive Care Med 41:975–984 PMX P olymyxin B hemoperfusion 8. Dellinger RP, Bagshaw SM, Antonelli M, Foster DM, Klein DJ, Marshall JC QALY Quality-adjusted life-year et al (2018) Eec ff t of targeted polymyxin B hemoperfusion on 28- day RBC Red blood cell mortality in patients with septic shock and elevated endotoxin level: RCT Randomized controlled trial the EUPHRATES randomized clinical trial. JAMA 320:1455–1463 rTM Recombinant thrombomodulin 9. Fujimori K, Tarasawa K, Fushimi K (2021) Eec ff ts of Polymyxin B hemop - SD Standard deviation erfusion on septic shock patients requiring noradrenaline: analysis of a SOFA Sequential organ failure assessment nationwide administrative database in Japan. Blood Purif 50:560–566 WTP Willingness to pay 10. Fujimori K, Tarasawa K, Fushimi K (2021) Eec ff ts of Polymyxin B hemop - erfusion in patients with sepsis requiring continuous hemodiafiltration: Acknowledgements analysis of a nationwide administrative database in Japan. Ther Apher The authors would like to thank Dr. Nobuo Ida , Ms. Ayuka Suzuki and Ms. Dial 25:384–389 Makiko Miyata of Toray Industries, Inc. for valuable discussions. 11. Fujimori K, Tarasawa K, Fushimi K (2021) Eec ff tiveness of polymyxin B hemoperfusion for sepsis depends on the baseline SOFA score: a Authors’ contributions nationwide observational study. Ann Intensive Care 11:141 K. Fujimori: responsible for the study concept, design, data extraction and 12. Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French analysis, and manuscript preparation. K.T.: accountable for the statistics and C et al (2021) Surviving sepsis campaign: international guidelines for manuscript preparation. K. Fushimi: responsible for the database organization management of sepsis and septic shock 2021. Intensive Care Med and review. The authors read and approved the final manuscript. 47:1181–1247 13. Matsuda S, Fujimori K, Fushimi K (2010) Development of casemix based Funding evaluation system in Japan. Asian Pac J Dis Manag 4:55–66 Ministry of Health, Labor and Welfare grant supported this study. ( Tokyo, 14. Murata A, Matsuda S, Kuwabara K, Ichimiya Y, Fujino Y, Kubo T et al Japan; 20AA2005). (2011) Equivalent clinical outcomes of bleeding peptic ulcers in teaching and non-teaching hospitals: evidence for standardization of Availability of data and materials medical care in Japan. Tohoku J Exp Med 223:1–7 The datasets used and/or analyzed during the current study are available from 15. Jones AE, Troyer JL, Kline JA (2011) Cost-effectiveness of an emergency the corresponding author on reasonable request. department-based early sepsis resuscitation protocol. Crit Care Med 39(6):1306–1312 Declarations 16. Suarez D, Ferrer R, Artigas A, Azkarate I, Garnacho-Montero J, Gomà G, Levy MM, Ruiz JC, Edusepsis Study Group (2011) Cost-effectiveness of Ethics approval and consent to participate the surviving sepsis campaign protocol for severe sepsis: a prospective The study was approved by the Institutional Review Board of Tohoku Univer- nation-wide study in Spain. Intensive Care Med 37(3):444–52 sity, which waived the requirement for informed patient consent because of 17. Shiroiwa T, Sung YK, Fukuda T, Lang HC, Bae SC, Tsutani K (2010) the anonymous nature of the data. International survey on willingness-to-pay ( WTP) for one additional QALY gained: what is the threshold of cost effectiveness? Health Econ 19(4):422–437 F ujimori et al. J Anesth Analg Crit Care (2023) 3:4 Page 9 of 9 18. Shiroiwa T, Igarashi A, Fukuda T, Ikeda S (2013) WTP for a QALY and health states: more money for severer health states? Cost Eff Resour Alloc 1(11):22 19. Fukuda T, Shiroiwa T (2019) Application of economic evaluation of pharmaceuticals and medical devices in Japan. J Natl Inst Public Health 68:27–33 20. Berto P, Ronco C, Cruz D, Melotti RM, Antonelli M (2011) Cost-effective - ness analysis of polymyxin-B immobilized fiber column and conventional medical therapy in the management of abdominal septic shock in Italy. Blood Purif 32(4):331–340 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

Journal of Anesthesia Analgesia and Critical CareSpringer Journals

Published: Feb 15, 2023

Keywords: Polymyxin B hemoperfusion; PMX; Sepsis; DPC database; Cost-effectiveness; Propensity score matching

There are no references for this article.