Abstract
Objectives
The aim of this study was to quantify the consumables waste for continuous renal replacement therapy (CRRT) in intensive care, determine the carbon dioxide equivalent (CO2e) per patient per day, and identify the impact of blood (Qb) and effluent flow rates, citrate anticoagulation, autoeffluent, and circuit lifespan on emissions and cost.
Design
A retrospective observational study was conducted.
Setting
The study included two major tertiary intensive care units in Melbourne from January 2017 to December 2024.
Participants
We analysed 70,000 h of treatment data stored on CRRT machines, captured from 2464 patients.
Interventions
Quantities of all major consumables were obtained. A chemical composition analysis was performed, and life cycle analysis was done for each component.
Main outcome measures
Carbon dioxide–equivalent values and financial cost estimates were calculated using local, commercial pricing.
Results
Combined impact of autoeffluent, lowest effluent rate (<30 mL/kg/hr), and doubling of filter life linked to Qb (130–150 mL/h) and citrate anticoagulation was associated with a 56.8% [54.4%–59.1%] decrease in CO2e.kg/day and 38.1% [34.8%–40.9%] cost reduction.
Median CO2e.kg/day fell from 25.3 [21.5–29.7] in 2017 to 19.3 [16.3 – 22.8] in 2024, p <0.0001. The total cost per day of consumables also fell from $995.89 ($820.40 – $1210.67) in 2017 to $829.07 ($689.64 – $1069.19) in 2024.
Conclusion
Extended circuit lifespan, autoeffluent, and moderated effluent doses translate to the most sustainable delivery of CRRT with associated cost savings.
Keywords: Continuous renal replacement therapy (CRRT), Carbon dioxide equivalent (CO2-e), Life cycle analysis (LCA), Sustainability in intensive care, Consumables
1. Introduction
Health care accounts for 5% of Australia’s greenhouse gas emissions,1 and intensive care generates high volumes of consumables waste, contributing to carbon emissions and costs.2 Environment, social, and governance reporting requirements of health services demand strong data-led strategies to drive this figure down to reduce health care’s impact on climate change. As such, comprehensive analyses of consumables waste are important to inform decision-making at manufacturing, procurement, and clinical levels.
Renal replacement therapy is recognised as a significant contributor to consumables waste in the healthcare setting. In patients with chronic kidney disease, intermittent haemodialysis alone is estimated to produce 1.5 billion tonnes of waste annually, more than any other healthcare procedure.3 Nephrologists have reported the substantial environmental effects of long-term renal support for the last two decades.4 In response, programs such as the National Health Service’s Sustainable Healthcare Programme have achieved annual savings of 10 million euros through improvements in water sterilisation efficiency and reduced electricity usage.
Unlike intermittent haemodialysis, the environmental burden of continuous renal replacement therapy (CRRT) is not well understood, despite being delivered to 8000 intensive care unit (ICU) patients in Australia and New Zealand per year (Appendix 1, Table A1 – ANZICS CRRT and total admissions data). Accordingly, we aimed to quantify major consumables during CRRT, estimate the associated carbon emissions and costs, and measure the impact of (i) filter life; (ii) blood flow and effluent flow rates; (iii) citrate anticoagulation; and (iv) use of an automatic effluent drainage system (auto-effluent). We used benchmarks derived from current clinical evidence around timing of initiation, dialysis dose, citrate anticoagulation, and blood flow rates, with cost and carbon dioxide emissions presented alongside the evolution of practice over time. We hypothesised that lower effluent rates, use of citrate anticoagulation, longer filter life, and use of auto-effluent were likely to reduce financial costs and carbon emissions.
2. Methods
2.1. Design
We undertook a retrospective observational study evaluating the major consumables waste generated during CRRT for intensive care patients at two large tertiary ICUs in Melbourne from January 2017 to December 2024 and estimated the carbon impact and cost.
2.2. Participants, devices, and CRRT protocol
Between January 2017 and December 2024, we included adult patients undergoing CRRT if they had a recorded treatment on the Prismaflex® or PrisMax® device (Vantive, IL, USA) (Appendix 1, Figure A1 – Images of machines with circuits connected). CRRT was prescribed as continuous venovenous haemodiafiltration (generally a 1:1 dialysis-to-filtration ratio, predominantly predilution) or continuous venovenous haemodialysis, with citrate, heparin, low-molecular-weight heparin, heparin-protamine, or no circuit anticoagulation. Fluid removal (or net ultrafiltration) was at the discretion of the treating clinician, and effluent rate was 25–50 mL/kg/hr. Blood flow (Qb) was 150–250 mL/min at hospital A and 130 mL/min at hospital B, the former changing practice to reduced dialysate and blood flow rates in 2022. For citrated circuits, an 18-mmol/L trisodium citrate solution was used as predilution (pre–blood pump) to deliver either 2.7 mmol/L citrate predilution (hospital A) or 3 mmol/L (initially titrated up or down to a postfilter ionised calcium of 0.25–0.4, hospital B). Predilution volume was calculated by the machine whereby it was fixed by Qb and citrate dose. Dialysate utilised a calcium-free bicarbonate-buffered solution. Postdilution was done either with a bicarbonate-buffered solution or with a calcium-free lower bicarbonate solution, generally set to 200 mL/h, and dialysate flow was either the balance to achieve total effluent rate of 30 mL/kg/hr or set to match the total convective volume (i.e., predilution + postdilution). Noncitrate circuits used bicarbonate-buffered solutions as predilution, postdilution, and dialysate. Membranes used were AN69ST100 or AN69ST150 (Vantive, IL, USA). All pre-prepared dialysis solutions were presented in 5-L bags and connected inline to membrane sets. Effluent was discharged either into single-use 5-L collection bags or directly into the hospital sewerage drainage system via an autoeffluent circuit, the latter option only being available with PrisMax® devices (Vantive, IL, USA, Appendix 1, Figure A2 – Autoeffluent circuit diagram).
2.3. Variables and data sources
2.3.1. Primary data source and patient data
Prismaflex® or PrisMax® devices (Vantive, IL, USA) record and store comprehensive data for every circuit run, which can be used for quality and training purposes. Parameters for Qb, net ultrafiltration, effluent rate, and filter life were transferred from machine data storage and imported into “R” and “RStudio” software (Appendix 1, Table A2 – R packages used), for this study. Each unit record number per patient, entered by nursing staff into machines when commencing CRRT, is encrypted to generate a unique identifier. All CRRT circuit runs per patient are linked through this identifier. Aggregate patient characteristics during the period studied were extracted from the Australian and New Zealand Intensive Care Society (ANZICS) Adult Patient Database (APD) (Table 1 - Patient Characteristics during the period of study from ANZICS Adult Patient Database) and not linked to machine data or unique identifiers.
Table 1.
Patient characteristics in the study period from ANZICS Adult Patient Database.
| Variable | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|---|---|---|
| N | 343 | 451 | 492 | 465 | 488 | 480 | 497 | 528 |
| Age, years [95% CI]a | 60.4 [47.5 to 69.8] | 60.2 [47.3 to 70.0] | 59.9 [48.3 to 69.6] | 61.0 [48.8 to 70.6] | 61.8 [49.0 to 71.7] | 59.5 [47.3 to 70.2] | 61.1 [48.0 to 71.1] | 61.0 [48.5 to 69.2] |
| Males, n (%)b | 225 (65.6) | 304 (67.4) | 337 (68.5) | 305 (65.6) | 310 (63.5) | 318 (66.2) | 349 (70.2) | 361 (68.4) |
| Hospital A, n (%)b | 220 (64.1) | 318 (70.5) | 321 (65.2) | 271 (58.3) | 302 (61.9) | 281 (58.5) | 295 (59.4) | 315 (59.7) |
| Length of stay, days [IQR]c | 7.7 [3.2 to 15.6] | 8.3 [3.8 to 15.9] | 7.0 [2.9 to 14.8] | 7.2 [3.0 to 14.9] | 8.0 [3.2 to 17.1] | 7.6 [3.1 to 15.0] | 7.4 [2.8 to 16.7] | 7.8 [3.2 to 16.0] |
| AP3, ROD [IQR]c | 41.0 [17.8 to 75.6] | 34.6 [15.8 to 66.8] | 36.3 [15.8 to 72.6] | 35.3 [17.0 to 68.5] | 36.5 [16.0 to 71.1] | 37.1 [16.7 to 74.2] | 32.1 [13.3 to 68.7] | 34.5 [13.1 to 66.0] |
| Mechanical ventilation, n (%)b | 287 (83.7) | 348 (77.2) | 352 (71.5) | 309 (66.5) | 341 (69.9) | 315 (65.6) | 339 (68.2) | 353 (66.9) |
| Noninvasive ventilation, n (%)b | 216 (63) | 272 (60.3) | 274 (55.7) | 229 (49.2) | 256 (52.5) | 258 (53.8) | 276 (55.5) | 278 (52.7) |
| Mortality, n (%)b | 120 (35) | 135 (29.9) | 156 (31.7) | 110 (23.7) | 138 (28.3) | 134 (27.9) | 130 (26.2) | 138 (26.1) |
| Admission diagnosis—surgicalb | ||||||||
| Cardiothoracic surgery, n (%) | 36 (10.5%) | 43 (9.5%) | 30 (6.1%) | 39 (8.4%) | 43 (8.8%) | 32 (6.7%) | 48 (9.7%) | 44 (8.3%) |
| Other surgical, n (%) | 17 (5%) | 21 (4.7%) | 19 (3.9%) | 26 (5.6%) | 25 (5.1%) | 24 (5%) | 30 (6%) | 20 (3.8%) |
| Admission diagnosis—medicalb | ||||||||
| Cardiovascular, n (%) | 80 (23.3%) | 137 (30.4%) | 150 (30.5%) | 131 (28.2%) | 126 (25.8%) | 143 (29.8%) | 130 (26.2%) | 164 (31.1%) |
| Haematological, n (%) | 10 (2.9%) | 10 (2.2%) | 4 (0.8%) | 8 (1.7%) | 11 (2.3%) | 7 (1.5%) | 8 (1.6%) | 6 (1.1%) |
| Liver failure or transplant, n (%) | 4 (1.2%) | 5 (1.1%) | 3 (0.6%) | 6 (1.3%) | 6 (1.2%) | 6 (1.2%) | 6 (1.2%) | 7 (1.3%) |
| Other medical, n (%) | 22 (6.4%) | 28 (6.2%) | 38 (7.7%) | 28 (6%) | 32 (6.6%) | 27 (5.6%) | 41 (8.2%) | 42 (8%) |
| Pneumonia, n (%) | 20 (5.8%) | 10 (2.2%) | 7 (1.4%) | 13 (2.8%) | 41 (8.4%) | 35 (7.3%) | 15 (3%) | 20 (3.8%) |
| Renal, n (%) | 16 (4.7%) | 34 (7.5%) | 37 (7.5%) | 26 (5.6%) | 30 (6.1%) | 29 (6%) | 38 (7.6%) | 45 (8.5%) |
| Respiratory, n (%) | 20 (5.8%) | 24 (5.3%) | 36 (7.3%) | 17 (3.7%) | 30 (6.1%) | 20 (4.2%) | 20 (4%) | 26 (4.9%) |
| Sepsis—shock, n (%) | 62 (18.1%) | 55 (12.2%) | 74 (15%) | 71 (15.3%) | 47 (9.6%) | 48 (10%) | 51 (10.3%) | 49 (9.3%) |
| Sepsis, n (%) | 32 (9.3%) | 51 (11.3%) | 58 (11.8%) | 75 (16.1%) | 61 (12.5%) | 67 (14%) | 72 (14.5%) | 76 (14.4%) |
| Toxicology, n (%) | 5 (1.5%) | 6 (1.3%) | 7 (1.4%) | 4 (0.9%) | 5 (1%) | 11 (2.3%) | 8 (1.6%) | 5 (0.9%) |
ANZICS: Australian and New Zealand Intensive Care Society; CI: confidence interval; IQR: interquartile range.
Values are presented as mean [95% confidence interval].
Values are presented as number (percentage of total for that year).
Values are presented as median [interquartile range].
Test runs and training/demonstration circuits were identified through knowledge of the encrypted unique identifiers for “demo” and “test” entries and removed. Any “patient” where their total therapy time was less than 5 h was also excluded as these were likely also test/training runs.
2.3.2. Materials usage
Total consumables usage was calculated on a per-patient basis and presented per day (24 h) of CRRT delivered. Volumes per CRRT circuit run were divided by 5000 mL (i.e., one bag of dialysate/filtrate) to obtain the number of bags used (and rounded up to the nearest integer) for effluent (where auto-effluent not used), predilution, postdilution, and dialysate. The number of membrane and circuit tubing sets per circuit run was “1”, with set type extracted from machine data. Use of auto-effluent was established from machine events records, and for every 7 days of therapy, it was assumed a new auto-effluent set would be used, as per manufacturer recommendations. This was established as a total number of auto-effluent sets used for a patient. All materials’ usage was then divided by the number of days of treatment.
2.3.3. Compound analysis, life cycle analysis, and carbon emission estimation
Consumable components were analysed by Fourier transform infrared (FTIR) spectroscopy with attenuated total reflectance at the University of Melbourne to determine absorbance peaks and identify composite compounds, using the Bruker® FTIR attenuated total reflectance device, Germany (Appendix 1, Figure A3/1–8—Spectrogram for Prismaflex 5-Litre effluent bag; Appendix 1, Figure A4 - FTIR Basic Organic Functional Group Reference Chart).
Process-based (PB) life cycle assessment (LCA) quantifies the environmental impact of a product or service throughout its entire life cycle, from raw material extraction to end-of-life disposal. It is a standardised process divided into four key stages: goal and scope definition, life cycle inventory analysis, life cycle impact assessment, and interpretation.[5], [6], [7] Background data on the environmental impacts of materials (e.g., plastics) and manufacture were obtained from the ecoinvent 3.11 APOS database (Appendix 1, Table A3 – Life Cycle Assessment Values).
The functional unit is “Patient dialysis day carbon dioxide equivalent”. Results are in kg carbon dioxide equivalents, using IPCC 2021GWP100 characterisation factors. CO2e derives from the emissions related to materials within the bags, extraction, production, transportation, and distribution (wholesale, storage, and delivery to site).5
2.3.4. Cost estimation
Local retail purchase pricing in Australian Dollar (AUD) was used as a cost estimate, with each material usage calculated on a per-circuit basis and then per-patient basis. Waste costs were estimated based on state clinical and nonclinical waste pricing, $3.50 (AUD) per kilogram of clinical waste and $0.30 (AUD) per kilogram for nonclinical waste. Hospital waste in the state of Victoria is disposed of via landfill (clinical waste being sterilised prior to disposal) (Appendix 1 – Table A4 - Cost and CO2 emissions, unit calculations).
2.4. Statistics
Baseline and process observations were summarised as proportions, means (95% confidence interval [CI]), or medians (interquartile range), as appropriate.
Multivariate analyses were performed on a per-patient basis utilising mixed linear modelling with hospital as a random effect. Outcomes of interest were dialysis-day carbon dioxide equivalent (CO2e.kg/day) and total cost per day (AUD/day). Variables of interest were use of citrate, use of auto-effluent, approximate effluent rate tertile (<30 mL/kg/hr, 30–40 mL/kg/hr and >40 mL/kg/hr), and log-transformed filter life. Effects were considered significant at a p value <0.005. The sensitivity analysis was performed with a subset of patients who were treated with an effluent rate of 20–25 mL/kg/hr, as per the Kidney Disease Improving Global Outcomes (KDIGO) Clinical Practice Guideline for Acute Kidney Injury. Furthermore, we performed a dose-intensity modelling analysis using mixed-effectlinear regression. Model predictions were generated to obtain expected daily values at effluent doses ranging from 10 to 50 mL/kg/hr.
To estimate the combined effect of multiple predictors, we computed linear contrasts from the fixed effects of the mixed-effect model. Predicted values for the combinations of interest were calculated on the log-transformed scale, and differences were used to derive contrast estimates with corresponding standard errors and 95% CIs based on the model’s variance–covariance structure. We evaluated effect modification using interaction terms between citrate and filter life in linear mixed-effect models. The mediation analysis was performed to decompose total effects of citrate on cost and CO2 emissions into direct and indirect (via filter life) pathways, using quasi-Bayesian estimation with 1000 simulations.
2.5. Ethical approval
The study was approved as a multisite low-/negligible-risk project (QA2024156) by the local Human Research and Ethics Committee (HREC). Patient consent was not required for this noninterventional retrospective observational study.
3. Results
3.1. Patient characteristics
A total of 3744 patients underwent CRRT across both sites in the study period. We observed reasonably consistent patient characteristics across the 8-year period (Table 1 – ANZICS Adult Patient Database) with a male predominance and median age of 60 years. This was an expectedly sick cohort of mixed medical/surgical patients with mortality rates between 25 and 35%, often receiving more-than-one organ supports with a prolonged length of stay. Hospital A managed 10–20% more patients requiring CRRT.
3.2. Data loss
Of the 3744 patients recorded to have received CRRT, machine-recorded data were available for 2464 patients. Complete information was available for 2017, 2023, and 2024, approximately 80% for 2018 and 2019, and only 25–40% of episodes during 2020–2022. Restricted access to machines by industry during the pandemic, decommissioning of older Prismaflex® devices prior to data access, and software upgrades with failure to save pre-existing data accounted for some data loss. In total, 2464 patients with median circuit lifespans of 25–35 h provided approximately 70,000 h of therapy for analysis and calculations.
3.3. Consumable weights, carbon emissions, and cost calculations
The heaviest consumables were the sets themselves (approximately 1 kg) followed by the auto-effluent sets (546 g); while effluent and individual fluid bags were all relatively light (88–139 g), their contributing total weight was due to numbers used (Appendix 1, Table A4). CO2 emission estimates/life cycle analyses were derived using the spectrographic analysis shown in Appendix 1, Figures A3.1–A3.8, with Fourier Transform Infrared (FTIR) reference chart Appendix 1, Figure A4, and life cycle analysis values Appendix 1, Table A3. Carbon emissions were the highest per unit for the membrane/tubing sets (6.22–6.95 CO2e.kg/day), followed by the auto-effluent devices (2.98 CO2e.kg/day), with that for fluid bags being similar around 0.60 CO2e.kg/day. Purchase cost contributed most costs, and unit costs were the highest for the membrane/tubing sets followed by the auto-effluent apparatus. Waste cost was the highest for clinical waste elements of consumables (Appendix 1, Table A4).
3.4. Consumable usage, carbon emissions, and cost over time and associated circuit variables
Over time, there was a downwards trend of CO2 emissions and cost (Fig. 1 and Appendix 1, Table A5). This was associated with use of autoeffluent, reduced effluent rates (predilution and dialysis volumes), and increasing filter life, the latter in turn correlated with lower Qb and increased use of citrate. Median CO2e.kg/day fell from 25.3 [21.5–29.7] in 2017 to 19.3 [16.4–22.8] in 2024. The total cost per day of consumables fell in parallel from $995.89 ($820.40–$1210.67) in 2017 to $829.07 ($689.64–$1069.19) in 2024.
Fig. 1.
Changes by year, blood flow, effluent rate, circuit life, and CO2 emissions
Note different Y-axis for blood flow (left) versus effluent rate, circuit life, and CO2-e (right).
A trend in improved filter life and carbon emissions was seen with reductions in effluent rate and blood pump speed.
3.5. Consumable usage, carbon emission, cost, and associated circuit variables by anticoagulation and autoeffluent use
Circuit life values for citrate versus noncitrate circuits without auto-effluent were 32.0 [20.9–44.9] versus 23.7 [14.5–36.6] hours and that with auto-effluent were 34.4 [20.4–50.3] versus 30.7 [18.6–43.5], respectively. Citrate circuits had higher effluent rates, from predilution and dialysate volumes, leading to higher cost and CO2 emissions (Table 2). Hospital B used auto-effluent and lower blood flow rates; the former negated the need for effluent bags, reducing waste cost from $10 per day to $5 per day and CO2 emissions from approximately 24 CO2e.kg/day to 16 CO2e.kg/day. Auto-effluent purchase costs outweighed effluent bag costs for short-duration treatments (Table 2).
Table 2.
Circuit variables, CO2 emissions, and cost by citrate and use of autoeffluent.
| Variable |
Citrate + No autoeffluent |
Citrate + Autoeffluent |
Noncitrate + No autoeffluent |
Noncitrate + Autoeffluent |
|---|---|---|---|---|
| Total Patients, n | 1471 | 197 | 699 | 98 |
| Proportion hospital A, % | 90.1 | 0 | 63.1 | 0 |
| Proportion PrisMax, % | 36.8 | 100 | 29.3 | 100 |
| Blood flow rate, mL/min | 191.0 [150.0–199.5] | 129.8 [128.6–130.0] | 172.9 [130.0–198.9] | 129.6 [126.0–130.0] |
| Effluent flow rate, mL/kg/hr | 38.1 [31.1–43.4] | 33.7 [27.9–39.6] | 32.4 [27.9–40.9] | 30.9 [29.1–35.6] |
| Net ultrafiltration rate, mL/kg/hr | 1.6 [1.1–2.1] | 1.1 [0.61–1.6] | 1.5 [0.8–2.1] | 1.1 [0.3–1.6] |
| Predilution rate, mL/kg/hr | 17.6 [15.7–19.9] | 14.1 [11.5–16.4] | 13.8 [8.9–16.0] | 12.6 [11.5–14.6] |
| Postdilution rate, mL/kg/hr | 2.3 [2.0–2.7] | 2.2 [1.8–2.5] | 2.4 [2.1–2.8] | 2.4 [1.8–2.8] |
| Dialysis rate, mL/kg/hr | 17.6 [9.7–19.3] | 16.3 [13.5–18.8] | 19.0 [13.2–23.5] | 14.8 [14.5–17.4] |
| Circuit life, hours | 32.0 [20.9–44.9] | 34.4 [20.4–50.3] | 23.7 [14.5–36.6] | 30.7 [18.6–43.5] |
| Predilution bags per day, n | 7.6 [6.5–8.0] | 6.4 [6.2–6.7] | 6.0 [4.0–6.7] | 6.2 [4.9–7.5] |
| Postdilution bags per day, n | 1.3 [1.1–1.5] | 1.3 [1.1–1.5] | 1.4 [1.2–1.9] | 1.4 [1.2–1.7] |
| Dialysis bags per day, n | 6.9 [4.6–8.4] | 7.4 [7.2–7.8] | 7.8 [5.6–9.8] | 7.1 [5.9–8.2] |
| Effluent bags per day, n | 15.3 [13.1–17.4] | 0.00 [0.00–0.00] | 13.7 [11.5–16.1] | 0.00 [0.00–0.00] |
| Sets per day, n | 0.7 [0.5–1.1] | 0.7 [0.5–1.2] | 1.0 [0.7–1.6] | 0.8 [0.5–1.3] |
| Autoeffluent per day, n | 0.00 [0.00–0.00] | 0.8 [0.5–1.3] | 0.00 [0.00–0.00] | 0.9 [0.6–1.4] |
| Predilution cost, $AUD | 258.4 [218.3–272.9] | 217.0 [211.4–228.2] | 158.4 [104.8–180.0] | 166.7 [128.3–199.9] |
| Postdilution cost, $AUD | 41.9 [37.0–50.0] | 33.2 [28.9–40.0] | 38.7 [31.4–50.5] | 35.9 [30.5–43.8] |
| Dialysis cost, $AUD | 233.6 [155.5–283.9] | 251.3 [243.7–265.3] | 209.3 [150.6–260.3] | 189.7 [154.5–216.3] |
| Effluent cost, $AUD | 120.74 [102.9–136.8] | 0.00 [0.00–0.00] | 107.73 [90.3–127.1] | 0.00 [0.00–0.00] |
| Sets cost, $AUD | 285.9 [203.7–438.8] | 267.7 [183.2–452.5] | 385.2 [250.4–622.7] | 3 00.3 [212.2–494.6] |
| Autoeffluent cost, $AUD | 0 [0-0] | 235.1 [145.1–385.8] | 0 [0-0] | 251.9 [180.4–409.8] |
| Waste cost, $AUD | 9.8 [8.2–11.5] | 4.6 [3.2–6.9] | 10.2 [8.2–13.1] | 5.0 [3.7–7.7] |
| Total cost, $AUD | 956.0 [800.0–1133.8] | 997.7 [845.1–1355.5] | 901.2 [708.6–1193.9] | 964.6 [766.2–1313.8] |
| Predilution CO2 emissions, CO2 e.kg/day | 4.6 [4.0–4.9] | 3.9 [3.8–4.1] | 3.6 [2.4–4.0] | 3.7 [2.9–4.5] |
| Postdilution CO2 emissions, CO2 e.kg/day | 0.8 [0.7–0.9] | 0.8 [0.7–0.9] | 0.9 [0.7–1.1] | 0.8 [0.7–1.0] |
| Dialysis CO2 emissions, CO2e.kg/day | 4.1 [2.8–5.1] | 4.4 [4.3–4.7] | 4.7 [3.4–5.9] | 4.3 [3.5–4.9] |
| Effluent CO2 emissions, CO2e.kg/day | 9.2 [7.8–10.4] | 0.00 [0.00–0.00] | 8.2 [6.9–9.7] | 0.00 [0.00–0.00] |
| Sets CO2 emissions, CO2e.kg/day | 5.2 [3.7–7.8] | 4.8 [3.3–8.2] | 6.8 [4.5–10.9] | 5.4 [3.8–9.0] |
| Autoeffluent CO2 emissions, CO2e.kg/day | 0.00 [0.00–0.00] | 1.1 [0.7–2.2] | 0.00 [0.00–0.00] | 1.2 [0.7–2.3] |
| Total CO2 emissions, CO2e.kg/day | 24.2 [20.3–28.1] | 15.6 [13.2–19.3] | 24.1 [19.8–30.4] | 15.9 [12.2–20.4] |
All values excluding total numbers and proportions are median [interquartile range].
Waste costs are an aggregate of all consumables waste.
Costs of individual consumable items represent commercial purchase costs.
3.6. Impact of effluent rate, anticoagulation, auto-effluent, and filter life on CO2 emissions and cost
Associations between anticoagulation, auto-effluent, and effluent rate tertile versus CO2 emissions and cost are shown in violin plots (Appendix 1, Fig. 5A–F); additionally, a log linear trend was observed between filter life and both CO2 emissions and cost (Fig. 2).
Fig. 2.
Association of filter life with reduced CO2 emissions and cost per day of treatment
Cost/day and CO2eq.kg/day showed log linear relationship; note both scales are logarithmic for both plots. Each data point represents a patient total cost divided by treatment days and CO2eq.kg divided by treatment days. Locally estimated scatterplot smoothing (LOESS) curve has been applied to each.
The multivariate analysis revealed auto-effluent and longer filter life correlated with reductions in CO2-e of 22.2% [20.2%–24.2%] and 34.3% [33.5%–35.0%], respectively (Table 3), while citrate use and higher effluent rates demonstrated increases in CO2e of 5.4% [3.8%–7.0%] and 18.3% [16.2%–20.4%], respectively. Prolonged filter life provided a cost reduction of 42% [40.4%–42.6%], with increased costs linked to citrate (15.8% [12.5%–15.8%]), auto-effluent (23.8% [20.9%–26.7%]), and higher effluent rates (>40 versus <30 mL/kg/hr) 15.6% [13.7%–17.4%]). Combination of auto-effluent, lowest effluent tertile (<30 mL/kg/hr), and doubling of filter life gave a 56.8% [54.4%–59.1%] decrease in CO2e.kg/day and a reduced cost of 37.9% [ 34.8%–40.9%]; however, the greatest cost benefit was with no auto-effluent, lowest effluent tertile, and longer filter life, 49.8% [48.5%–51.1%], all p < 0.0001. Citrate anticoagulation was associated with lower CO2 emissions and mixed effects on cost, driven by opposing direct and indirect pathways. As anticipated, a strong interaction between citrate and filter life was observed within the CO2 and cost models. The mediation analysis confirmed that citrate reduced CO2 and cost indirectly via filter life extension, but these gains were partly offset by direct increases in CO2 and cost attributable to citrate (Appendix 1, Table A6).
Table 3.
Mixed linear models: Anticoagulation, autoeffluent, effluent rate tertile, and filter life associations with CO2 emissions and cost.
| Variable | CO2-equivalent effect estimate | Interpretation | p value | Cost effect estimate | Interpretation | p value |
|---|---|---|---|---|---|---|
| Anticoagulation: citratea | 1.054 [1.038–1.070] | ↑ 5.4% increase in CO2e.kg/day | <0.0001 | 1.158 [1.142–1.175] | ↑ 15.8% increase in cost/day | <0.0001 |
| Autoeffluentb | 0.778 [0.758–0.798] | ↓ 22.2% decrease in CO2e.kg/day | <0.0001 | 1.238 [1.209–1.267] | ↑ 23.8% increase in cost/day | <0.0001 |
| Effluent rate: 30–40 mL/kg/hrc | 1.074 [1.056–1.092] | ↑ 7.4% increase in CO2e.kg/day | <0.0001 | 1.061 [1.045–1.078] | ↑ 6.1% increase in cost/day | <0.0001 |
| Effluent rate: >40 mL/kg/hrc | 1.183 [1.162–1.204] | ↑ 18.3% increase in CO2e.kg/day | <0.0001 | 1.156 [1.137–1.174] | ↑ 15.6% increase in cost/day | <0.0001 |
| Filter lifed | 0.657 [0.650–0.665] | ↓ 34.3% decrease in CO2e.kg/day | <0.0001 | 0.580 [0.574–0.586] | ↓ 42% decrease in cost/day | <0.0001 |
| Combination of autoeffluenta, Effluent rate <30 mL/kg/hre, filter lifed |
0.432 (0.418–0.446) | ↓ 56.8% decrease in CO2e.kg/day | <0.0001 | 0.621 [0.604 to 0.640] | ↓ 37.9% decrease in cost/day | <0.0001 |
All results were statistically significant (p < 0.0001).
Versus noncitrate anticoagulation.
Versus no autoeffluent/effluent bags.
Versus <30 mL/kg/hr (lowest tertile).
Estimate is effect for each doubling of filter life.
Versus >40 mL/kg/hr.
An additional sensitivity analysis of patients receiving 20–25 mL/kg/hr showed a lower overall cost, $923 [769–1292], and CO2-e, 22.1 [18.6–28.6] kg/day.
Predicted cost and CO2e rose progressively with increasing effluent dose (Appendix A1, Table A8)., An estimated daily consumable cost of approximately $816–$867 per patient-day (95% CI ∼ $410–$1650) and CO2 emissions of 22–24 kg CO2e per day (95% CI ∼ 13–40 kg) was demonstrated for doses of 20-25mL/kg/hr. At the lowest modelled dose of 10 mL/kg/hr, both cost and CO2 were roughly 15–20% lower, whereas at the highest dose of 50 mL/kg/hr, they were 20–25% higher than baseline.
4. Discussion
4.1. Key findings
This comprehensive analysis of consumable waste with CRRT highlights several key findings. In 2024, CRRT delivered for 24 h translated to a total of 19.3 kg [16.4–22.8] CO2e. The membrane and tubing sets accounted for the highest CO2e, followed by auto-effluent, and purchase cost dominated total cost. Between 2017 and 2024, there was a downward trend in both CO2e and cost. This is most likely explained by focused adjustments to CRRT prescriptions over the same time (Qb, lower effluent dose, citrate anticoagulation, and increased auto-effluent), which align with the clinical evidence base and are plausibly associated with reduced waste, cost, and CO2e.
4.2. Relationship to previous studies
Previous studies have not specifically quantified the burden of consumables waste attributable to CRRT. However, data are available for comparison in related settings. The daily environmental footprint of an ICU patient is estimated at 138 kg CO2e in the United States8 and 61 kg CO2e in France.9 This figure is even higher in critically ill patients with sepsis, at 178 kg CO2e/day in the US and 88 kg CO2e/day in Australia.4 When using PB LCA, the highest contributors to carbon emissions are consumables and energy consumption.10
From our data, it appears that important modifications in CRRT practice and prescription have translated into related benefits in terms of waste and cost. Studies in support of deferred initiation of CRRT and controlled effluent rates uniformly demonstrate the lack of mortality benefit using more liberal measures.[11], [12], [13], [14], [15], [16], [17], [18] Citrate anticoagulation benefits circuit lifespan,19,20 as does a lower Qb,21 through reduced access dysfunction, resulting in fewer circuit changes and lower predilution fluids volumes. These collectively translate to reduced waste and cost. Our analysis highlights the parallel impact on reduced carbon emissions through implementing this evidence base. The BEST study found that limiting effluent doses alone to 25 mL/kg/hr has the effect of reducing fluid costs in CRRT by approximately 43.3%,22 but in contrast, we found only found a <10% cost reduction.
4.3. Implications
Application of sustainability principles of reduce, recycle, reuse, and repurpose assists with improving the overall carbon footprint of CRRT. Dialysis is the highest consumables waste generator in health care.3,23 Baxter Healthcare (now Vantive) identified that 78% of CO2 emissions for CRRT was associated with the supply chain of fluid bags, 49% being linked to transport. The products we analysed were all manufactured overseas. Changing the point of manufacture to Australia or Asia (India or China) could be considered. Problematically, their primary fuel source for electricity generation are fossil fuels, compared with Italy and the rest of Europe using predominantly renewables and nuclear.24
Recycled polypropylene is more expensive than the virgin product, owing to cost of separation and sterilisation, is less robust and stable with each recycling process, and has higher CO2 emissions (Appendix 1, Table A7 - Chemical composition of consumables). Product manufacture using polyvinyl chloride (PVC) facilitates recyclability of all nonhazardous plastic waste. Every tonne of recycled PVC replaces about one tonne of virgin PVC compound in new products, consuming 80% less energy and reducing carbon emissions,24 in line with national climate action objectives. In Victoria in 2024, 1818 patients received CRRT, which produced 1.2 kg, per patient dialysis day, of nonhazardous waste, translating to upwards of 2 tonnes of potentially recycled PVC that would avoid landfill and the production of virgin plastic (Appendix 1, Table A1 - ANZICS total admissions and CRRT numbers per year).
While not included in our analysis, cardboard boxes transporting all consumables need to be addressed more broadly. For CRRT alone, between these two hospitals in 2024, 3000–4000 boxes were discarded.
Refer to Fig. 3 for some focus points of change management in the transition to the sustainable practice of CRRT.
Fig. 3.
Change management in sustainable CRRT. CRRT: continuous renal replacement therapy.
4.4. Strengths and limitations
This large data set of 70,000 h over 8 years, enabled identification of associations between different factors surrounding the delivery of CRRT and CO2 emissions. Filter life has been previously assumed to benefit resource usage, but this is the first study to demonstrate that prolonged filter life is associated with reduced costs and carbon emissions. The median values of consumables waste will be representative for CRRT in most Australian ICUs, and therefore, CO2e and cost data will have local applicability. The patterns will be externally valid, but absolute values will vary per region.
This analysis singularly used Vantive products, which hold a 70% market share of Australia and New Zealand ICUs, creating commercial bias with these results. The alternative companies, B. Braun Avitum AG (OMNI ® and Diapact®, Melsungen, Germany) and Fresenius (multiFiltratePRO® Bad Hornburg, Germany), manufacture fluid bags from almost identical raw materials, and the volumes of waste are likely equivalent to that of Vantive-run therapies as fluid volumes used are similar.
The centre that exclusively uses auto-effluent also uses lower blood flow rates, whereby the impact auto-effluent independently makes on circuit lifespan and waste may be amplified.
Our PB LCA introduces a risk of truncation bias compared with economic input–output LCA; however, the former provides relevant and interpretable information. This retrospective observational study poses the risk of confounders and therefore limits causal interpretation. Significant data loss, with the data obtained over 2020–2022, predominantly from hospital A, may affect overall validity of findings; however, the sheer volume of data likely mitigates this. Some data were rejected for potentially representing test or training runs, deemed advantageous for integrity of the dataset on balance. We also did not look at the comparative CO2e of providing intermittent haemodialysis in the ICU, which is a feasible clinical alternative, and potentially both carbon and cost saving.
5. Conclusions
CRRT is associated with significant generation of consumable waste, although the overall burden has decreased over time. Cohorted use of auto-effluent, a lower effluent rate (<30 mL/kg/hr), and longer treatment duration (every doubling of filter life) were associated with a 56.8% reduction in CO2e per 24 h (estimate: 0.397 [0.365–0.427]) compared to the reference group, a combination that may significantly reduce environmental impact. This is potentially mediated by low blood flow rates and citrate anticoagulation. Sustainably delivered CRRT in the ICU is a realistic vision for the near future.
Conflict of interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Co-authors Associate editor—Prof Andrew A. Udy, Alfred Hospital Intensive Care Unit—Consultant and Head of research. Monash University Australia and New Zealand Intensive Care Research Centre, Boonwurrung Country Victoria, Australia Editorial board—Dr Emily See, Royal Melbourne Hospital Intensive Care Unit and Department of Nephrology Parkville, Victoria, Australia—Consultant Adjunct senior research fellow—Monash University Senior Honorary Fellow—University of Melbourne. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. There were no external sources of funding for this rese
CRediT authorship contribution statement
1. Dr Benjamin W. Sansom, Senior Intensive Care Registrar, Royal Melbourne Intensive Care Unit – data interrogation, statistical analysis, editing of manuscript, mentorship.
2. Ms Catherine J. O'Shea, Program Manager of Healthcare Carbon Lab, University of Melbourne – composition analysis of each component of consumables.
3. Dr Scott McAlister, University of Melbourne, Senior Research Officer–Life cycle analysis of each component of consumables.
4. Dr Mayuri G.W. Wijayasundara, Honorary Fellow Deakin University, CEO Anvarta Pty Ltd. - Expertise on sustainability, circular economy and change management.
5. A/Professor Forbes McGain, Associate Dean Healthcare Sustainability, University of Melbourne – subject expertise, guidance of carbon emissions calculations.
6. Professor Andrew A. Udy – Consultant Intensivist and Head of Research Alfred hospital – mentorship for initiation of project.
7. Dr Emily J. See – Consultant Intensivist and Nephrologist Royal Melbourne Hospital and Senior Honorary Fellow, University of Melbourne and Adjunct Senior Research Fellow, Monash University – project supervisor.
8. Dr Patrick R. Joyce – Consultant Intensivist and ICU Renal clinical lead Alfred Hospital – supervisor for ethics application at Alfred hospital.
9. Dr Christian Karcher – Consultant Intensivist Royal Melbourne Hospital – Supervisor of training.
10. Ms Natalie Adams – Clinical Nurse Consultant, Renal Support Therapies, Intensive Care Unit, Alfred Hospital – assistance with data retrieval.
11. Ms Susan R. James – Critical Care Registered Nurse, Sustainability lead, Intensive care unit, Alfred Hospital – assistance with communication with commercial recycling companies and environmental services at Alfred Hospital.
12. Mrs Megan Taylor – Vantive Product Specialist – Assistance with data retrieval and provisions of consumables for analysis.
APPENDIX
Table A1.
Total ICU admissions, number of patients receiving CRRT, and percentage receiving CRRT per state per year, between 2017 and 2024.
| State | Year | Total | CRRT | Percentage |
|---|---|---|---|---|
| VIC | 2017 | 41,089 | 735 | 0.02 |
| VIC | 2018 | 43,563 | 1490 | 0.03 |
| VIC | 2019 | 44,790 | 1668 | 0.04 |
| VIC | 2020 | 42,665 | 1840 | 0.04 |
| VIC | 2021 | 43,878 | 1868 | 0.04 |
| VIC | 2022 | 41,877 | 1828 | 0.04 |
| VIC | 2023 | 43,690 | 1777 | 0.04 |
| VIC | 2024 | 45,990 | 1818 | 0.04 |
| ACT | 2017 | 4495 | 127 | 0.03 |
| ACT | 2018 | 4721 | 175 | 0.04 |
| ACT | 2019 | 4943 | 174 | 0.04 |
| ACT | 2020 | 4927 | 148 | 0.03 |
| ACT | 2021 | 5024 | 185 | 0.04 |
| ACT | 2022 | 4752 | 192 | 0.04 |
| ACT | 2023 | 5040 | 218 | 0.04 |
| ACT | 2024 | 5322 | 221 | 0.04 |
| NSW | 2017 | 56,498 | 442 | 0.01 |
| NSW | 2018 | 59,315 | 1480 | 0.02 |
| NSW | 2019 | 64,217 | 1741 | 0.03 |
| NSW | 2020 | 64,989 | 1706 | 0.03 |
| NSW | 2021 | 65,409 | 2084 | 0.03 |
| NSW | 2022 | 62,843 | 2156 | 0.03 |
| NSW | 2023 | 65,896 | 2367 | 0.04 |
| NSW | 2024 | 61,870 | 2001 | 0.03 |
| NT | 2017 | 1858 | 51 | 0.03 |
| NT | 2018 | 1742 | 218 | 0.13 |
| NT | 2019 | 1771 | 200 | 0.11 |
| NT | 2020 | 1721 | 241 | 0.14 |
| NT | 2021 | 1676 | 176 | 0.11 |
| NT | 2022 | 1706 | 170 | 0.10 |
| NT | 2023 | 1679 | 166 | 0.10 |
| NT | 2024 | 1744 | 171 | 0.10 |
| QLD | 2017 | 29,346 | 465 | 0.02 |
| QLD | 2018 | 31,225 | 1034 | 0.03 |
| QLD | 2019 | 32,402 | 1277 | 0.04 |
| QLD | 2020 | 31,112 | 1215 | 0.04 |
| QLD | 2021 | 31,217 | 1455 | 0.05 |
| QLD | 2022 | 30,529 | 1476 | 0.05 |
| QLD | 2023 | 32,177 | 1441 | 0.04 |
| QLD | 2024 | 31,366 | 1494 | 0.05 |
| SA | 2017 | 12,513 | 73 | 0.01 |
| SA | 2018 | 13,403 | 302 | 0.02 |
| SA | 2019 | 14,606 | 370 | 0.03 |
| SA | 2020 | 15,475 | 357 | 0.02 |
| SA | 2021 | 17,443 | 381 | 0.02 |
| SA | 2022 | 17,726 | 451 | 0.03 |
| SA | 2023 | 17,764 | 406 | 0.02 |
| SA | 2024 | 16,396 | 421 | 0.03 |
| TAS | 2017 | 3187 | 18 | 0.01 |
| TAS | 2018 | 3584 | 104 | 0.03 |
| TAS | 2019 | 3542 | 99 | 0.03 |
| TAS | 2020 | 2259 | 84 | 0.04 |
| TAS | 2021 | 2746 | 115 | 0.04 |
| TAS | 2022 | 3210 | 129 | 0.04 |
| TAS | 2023 | 3420 | 158 | 0.05 |
| TAS | 2024 | 2183 | 112 | 0.05 |
| WA | 2017 | 12,001 | 53 | 0.00 |
| WA | 2018 | 12,981 | 408 | 0.03 |
| WA | 2019 | 14,115 | 552 | 0.04 |
| WA | 2020 | 14,277 | 551 | 0.04 |
| WA | 2021 | 14,366 | 605 | 0.04 |
| WA | 2022 | 12,979 | 614 | 0.05 |
| WA | 2023 | 13,847 | 686 | 0.05 |
| WA | 2024 | 14,399 | 685 | 0.05 |
| Total | 1,411,223 | 47,491 | 0.03 |
Data obtained from ANZICS CORE database, revealing total numbers of patients admitted to the ICU per year and those registered to have received CRRT: 0.03% of ICU patients receive CRRT in Australia.
Transition of databases occurred in 2016/17.
ANZICS: Australian and New Zealand Intensive Care Society; CRRT: continuous renal replacement therapy; ICU: intensive care unit; VIC, Victoria; ACT, Australian Capital Territory; NSW, New South Wales; NT, Northern Territory; QLD, Queensland; SA, South Australia; TAS, Tasmania; WA, Western Australia; ANZICS CORE, Australia and New Zealand Intensive Care Society Centre for Outcome and Resource Evaluation.
Fig. A1.
Prismaflex® or PrisMax® device (Vantive, IL, USA) Image reproduced from www.baxter.com
Left hand side - Prismaflex®, Right hand side - PrisMax®
Each machine connected to circuit with 3 solution bags and 1 effluent collection bag
Fig. A2.
Autoeffluent circuit diagram (courtesy Baxter/Vantive)
Autoeffluent circuit as available for PrisMax® machine
Table A2.
R Packages used
| Package | Reference |
|---|---|
| broom.mixed | Bolker B, Robinson D (2024). _broom.mixed: Tidying Methods for Mixed Models_. R package version 0.2.9.6, <https://CRAN.R-project.org/package=broom.mixed>. |
| data.table | Dowle M, Srinivasan A (2023). _data.table: Extension of “data.frame”_. R package version 1.14.8, <https://CRAN.R-project.org/package=data.table>. |
| Dbplyr | Wickham H, Girlich M, Ruiz E (2023). _dbplyr: A “dplyr” Back End for Databases_. R package version 2.3.2, <https://CRAN.R-project.org/package=dbplyr>. |
| Dplyr | Wickham H, François R, Henry L, Müller K, Vaughan D (2023). _dplyr: A Grammar of Data Manipulation_. R package version 1.1.2, <https://CRAN.R-project.org/package=dplyr>. |
| Forcats | Wickham H (2023). _forcats: Tools for Working with Categorical Variables (Factors)_. R package version 1.0.0, <https://CRAN.R-project.org/package=forcats>. |
| ggplot2 | H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York, 2016. |
| Ggpubr | Kassambara A (2023). _ggpubr: ‘ggplot2’ Based Publication Ready Plots_. R package version 0.6.0, <https://CRAN.R-project.org/package=ggpubr>. |
| gridExtra | Auguie B (2017). _gridExtra: Miscellaneous Functions for "Grid" Graphics_. R package version 2.3, <https://CRAN.R-project.org/package=gridExtra>. |
| Hmisc | Harrell Jr F (2023). _Hmisc: Harrell Miscellaneous_. R package version 5.1-0, <https://CRAN.R-project.org/package=Hmisc>. |
| lme4 | Douglas Bates, Martin Maechler, Ben Bolker, Steve Walker (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1), 1-48. https://doi.org/10.18637/jss.v067.i01. |
| lmerTest | Kuznetsova A, Brockhoff PB, Christensen RHB (2017). “lmerTest Package: Tests in Linear Mixed Effects Models”. _Journal of Statistical Software_, ∗82∗(13), 1-26. https://doi.org/10.18637/jss.v082.i13 <https://doi.org/10.18637/jss.v082.i13>. |
| Lubridate | Garrett Grolemund, Hadley Wickham (2011). Dates and Times Made Easy with lubridate. Journal of Statistical Software, 40(3), 1-25. URL https://www.jstatsoft.org/v40/i03/. |
| Mediation | Dustin Tingley, Teppei Yamamoto, Kentaro Hirose, Luke Keele, Kosuke Imai (2014). mediation: R Package for Causal Mediation Analysis. Journal of Statistical Software, 59(5), 1-38. URL https://www.jstatsoft.org/v59/i05/. |
| Parallel | R Core Team (2023). _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria. <https://www.R-project.org/>. |
| Purr | Wickham H, Henry L (2023). _purrr: Functional Programming Tools_. R package version 1.0.1, <https://CRAN.R-project.org/package=purrr>. |
| Rmisc | Hope RM (2022). _Rmisc: Ryan Miscellaneous_. R package version 1.5.1, <https://CRAN.R-project.org/package=Rmisc>. |
| RSQLite | Müller K, Wickham H, James DA, Falcon S (2023). _RSQLite: SQLite Interface for R_. R package version 2.3.3, <https://CRAN.R-project.org/package=RSQLite>. |
| Tidyr | Wickham H, Vaughan D, Girlich M (2023). _tidyr: Tidy Messy Data_. R package version 1.3.0, <https://CRAN.R-project.org/package=tidyr>. |
| Tidyverse | Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen TL, Miller E, Bache SM, Müller K, Ooms J, Robinson D, Seidel DP, Spinu V, Takahashi K, Vaughan D, Wilke C, Woo K, Yutani H (2019). “Welcome to the tidyverse”. _Journal of Open Source Software_, ∗4∗(43), 1686. https://doi.org/10.21105/joss.01686 <https://doi.org/10.21105/joss.01686>. |
Fig. A3.
Spectrogram for Prismaflex 5-Litre effluent bag
Y axes – Absorbance units; relative peak intensities
Individual composition of components detailed in Table A6 presented later, determined from reads-outs mentioned earlier
Fig. A4.
FTIR basic organic functional group reference chart (reproduced from www.thermofisher.com)
Industry reference chart for peak absorbances and related functional groups, revealing composite materials. FTIR: Fourier transform infrared
Table A3.
Life cycle analysis values
| Damage category | Unit | Haemosol BP 5000 mL Product number 955426 | Prismocitrate 18/0 Non-PVC (APAC) | Prism0cal EN/VN | SP 414: Prismaflex 5-Litre effluent bag | Auto-Effluent set (Prismax) 115370 | Haemofilter set ST100 (Prismaflex) | Haemofilter set ST150 |
|---|---|---|---|---|---|---|---|---|
| GWP100 | kg CO2-eq | 0.6008679 | 0.6109661 | 0.6008679 | 0.5985857 | 2.976586 | 6.224895 | 6.949951 |
GWP100 : Global warming potential over a 100-year period, a metric used to compare the warming effect of difference greenhouse gases over a 100-year period relative to carbon dioxide. Inputs used for the calculations are from global supply data of the raw materials and average global values of electricity usage for induction and moulding of the plastic, integrated with the above analyses.
Table A4.
Cost and CO2 emissions, unit calculations
| Item | Weight (total) | Purchase price (AUD) | CO2-equivalent per 24 h | Clinical waste weight (g) | Nonclinical waste weight (g) | Clinical waste cost (AUD) | Nonclinical waste cost (AUD) | Total waste cost (AUD) |
|---|---|---|---|---|---|---|---|---|
| AN69 ST100 set | 1060.58 | 362.76 | 6.22 | 1060.58 | 0 | 3.71 | 0 | 3.71 |
| AN69 ST150 set | 1103.58 | 384.31 | 6.95 | 1103.58 | 0 | 3.86 | 0 | 3.86 |
| Auto-effluent set | 546.35 | 288.1 | 2.98 | 461.73 | 84.62 | 1.61 | 0.025 | 1.64 |
| Effluent bags | 139.87 | 7.88 | 0.60 | 113.87 | 26 | 0.40 | 0.008 | 0.40 |
| Haemosol | 110.69 | 26.36 | 0.60 | 0 | 110.69 | 0 | 0.03 | 0.03 |
| Prism0cal B22 | 110.69 | 34.11 | 0.60 | 0 | 110.69 | 0 | 0.03 | 0.03 |
| Prismocitrate | 88.23 | 34.11 | 0.61 | 0 | 88.23 | 0 | 0.03 | 0.03 |
1. Weight per item.
2. Individual commercial prices.
3. LCA values.
4. Weight of waste.
5. Distinguishing clinical versus nonclinical waste in both weight and cost of disposal.
6. Total cost of waste per item.
Table A5.
Results by year: circuit parameters, quantities of consumables, CO2-e/day, and costs
| Variable | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|---|---|---|
| Total Patients, n | 336 | 368 | 393 | 204 | 122 | 121 | 409 | 512 |
| Proportion hospital A, % | 69.6 | 67.9 | 59.5 | 93.1 | 96.7 | 63.6 | 72.9 | 71.4 |
| Proportion PrisMax, % | 0 | 0 | 0 | 0 | 0 | 99.17 | 100 | 100 |
| Proportion citrate, % | 53.9 | 62.2 | 70.2 | 74.0 | 76.2 | 70.2 | 71.9 | 70.1 |
| Blood flow rate, mL/min | 194.4 [149.6–199.8] | 192.5 [132.1–199.6] | 182.6 [129.9–199.3] | 196.1 [162.8–199.9] | 196.1 [171.3–199.9] | 149.9 [129.9–195.1] | 150 [130–195.2] | 150.0 [130–189.4] |
| Effluent flow rate, mL/kg/hr | 39.1 [31.1–44.0] | 40.0 [32.7–45.0] | 40.4 [35.4–45.0] | 41.5 [38.0–44.6] | 41.9 [36.9–45.0] | 31.0 [28.7–35.2] | 31.0 [28.1–35.2] | 30.2 [27.5–33.6] |
| Net ultrafiltration rate, mL/kg/hr | 1.5 [0.9–1.9] | 1.6 [1.0–2.0] | 1.5 [0.8–2.1] | 1.7 [1.3–2.2] | 1.7 [1.3–2.2] | 1.5 [0.8–2.0] | 1.5 [0.9–2.0] | 1.5 [1.0–2.0] |
| Predilution rate, mL/hr | 16.4 [5.6–19.6] | 17.1 [12.4–20.0] | 17.5 [14.8–19.8] | 18.0 [15.6–20.1] | 17.6 [15.3–20.1] | 15.0 [13.7–17.1] | 15.6 [13.3–17.6] | 14.9 [13.0–17.1] |
| Postdilution rate, mL/hr | 2.4 [2.0–2.8] | 2.4 [2.0–2.8] | 2.3 [2.0–2.8] | 2.4 [2.1–2.7] | 2.4 [2.0–2.7] | 2.2 [1.9–2.6] | 2.3 [1.9–2.7] | 2.3 [1.9–2.7] |
| Dialysis rate, mL/hr | 19.3 [18.4–22.8] | 19.3 [18.0–22.4] | 19.0 [16.9–20.4] | 18.9 [18.2–20.2] | 19.1 [18.0–20.3] | 10.4 [9.5–15.8] | 10.1 [9.5–14.7] | 9.7 [9.4–14.5] |
| Circuit life, hours | 26.7 [16.8–38.7] | 27.0 [17.3–40.6] | 26.8 [17.6–38.1] | 27.4 [17.3–41.2] | 28.4 [17.9–38.4] | 32.6 [19.1–51.9] | 34.3 [19.1–47.3] | 35.2 [22.1–49.2] |
| Predilution bags per day, n | 6.5 [2.9–7.8] | 7.2 [5.9–8.0] | 7.3 [6.3–8.0] | 7.6 [6.4–8.0] | 7.5 [6.4–8.0] | 6.8 [6.2–7.6] | 6.5 [6.0–7.6] | 6.5 [6.0–7.4] |
| Postdilution bags per day, n | 1.4 [1.2–1.8] | 1.3 [1.1–1.7] | 1.3 [1.2–1.7] | 1.3 [1.1–1.6] | 1.3 [1.1–1.6] | 1.3 [1.1–1.6] | 1.3 [1.1–1.5] | 1.2 [1.1–1.5] |
| Dialysis bags per day, n | 8.3 [7.0–9.6] | 8.2 [7.0–9.5] | 7.9 [7.1–9.1] | 8.1 [6.9–9.2] | 8.4 [7.1–9.4] | 5.1 [4.4–7.4] | 4.8 [3.9–7.2] | 4.8 [3.9–7.1] |
| Effluent bags per day, n | 15.3 [12.5–17.4] | 16.1 [13.3–17.6] | 16.4 [14.5–17.8] | 16.7 [15.0–18.2] | 16.6 [15.2–18.2] | 11.3 [0.0–13.6] | 11.1 [0.0–13.7] | 11.1 [0.0–13.5] |
| Sets per day, n | 0.9 [0.6–1.4] | 0.9 [0.6–1.4] | 0.9 [0.6–1.4] | 0.9 [0.6–1.4] | 0.8 [0.6–1.3] | 0.7 [0.5–1.2] | 0.7 [0.5–1.2] | 0.7 [0.5–1.1] |
| Auto effluent per day, n | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.5] | 0.0 [0.0–0.4] | 0 [0–0.4] |
| Predilution cost, $AUD | 209.59 [75.77–266.99] | 232.96 [156.65–270.65] | 237.14 [200.57–267.91] | 253.96 [208.30–272.40] | 254.36 [203.66–273.44] | 217.92 [188.31–257.47] | 213.95 [188.80–252.24] | 211.59 [190.80–241.21] |
| Postdilution cost, $AUD | 42.4 [36.31–53.06] | 40.84 [34.6–50.63] | 40.73 [34.8–50.24] | 42.16 [37.56–51.33] | 41.76 [35.81–50.19] | 38.86 [33.00–48.60] | 38.84 [33.89–46.74] | 38.48 [33.72–45.47] |
| Dialysis cost, $AUD | 253.64 [210.15–294.67] | 258.66 [220.07–299.14] | 258.12 [219.28–293.19] | 260.21 [221.42–303.03] | 255.83 [226.16–307.84] | 165.69 [134.48–238.85] | 156.75 [124.38–235.54] | 154.5 [123.69–217.64] |
| Effluent cost, $AUD | 120.47 [98.25–137.14] | 126.64 [105.09–138.36] | 128.92 [114.52–139.92] | 131.97 [117.78–143.61] | 131.23 [120.02–143.82] | 88.84 [0.00–107.32] | 87.88 [0.00–107.90] | 87.63 [0.00–106.35] |
| Sets cost, $AUD | 344.51 [236.80–541.59] | 336.16 [220.92–518.30] | 336.8 [237.31–512.84] | 336.17 [223.90–527.12] | 325.12 [239.90–516.57] | 283.18 [177.80–481.61] | 268.51 [195.00–482.00] | 261.88 [187.62–418.15] |
| Auto effluent cost, $AUD | 0.00 [0.00–0.00] | 0.00 [0.00–0.00] | 0.00 [0.00–0.00] | 0.00 [0.00–0.00] | 0.00 [0.00–0.00] | 0.00 [0.00–145.11] | 0.00 [0.00–105.27] | 0.00 [0.00–113.23] |
| Waste cost, $AUD | 10.46 [8.95–12.51] | 10.44 [8.92–12.44] | 10.69 [9.44–12.86] | 10.72 [9.22–12.83] | 10.75 [9.53–12.52] | 7.96 [6.02–9.92] | 7.91 [6.42–10.07] | 7.72 [6.34–9.54] |
| Total cost, $AUD | 995.89 [820.40–1210.67] | 992.48 [841.78–1200.82] | 1024.25 [886.58–1230.87] | 1037.93 [888.71–1232.12] | 1027.44 [888.20–1217.06] | 902.22 [750.88–1144.75] | 832.32 [723.77–1102.70] | 829.07 [689.64–1069.19] |
| Predilution CO2 emissions, CO2 equivalents per day | 3.9 [1.7–4.8] | 4.4 [3.5–4.9] | 4.4 [3.8–4.9] | 4.6 [3.9–4.9] | 4.6 [3.9–4.9] | 4.1 [3.8–4.7] | 3.9 [3.6–4.6] | 4.0 [3.6–4.5] |
| Postdilution CO2 emissions, CO2 equivalents per day | 0.8 [0.7–1.1] | 0.8 [0.7–1.0] | 0.8 [0.7–1.0] | 0.8 [0.7–1.0] | 0.8 [0.7–1.0] | 0.8 [0.7–0.9] | 0.8 [0.7–0.9] | 0.7 [0.7–0.9] |
| Dialysis CO2 emissions, CO2 equivalents per day | 5.0 [4.2–5.8] | 4.9 [4.2–5.7] | 4.7 [4.2–5.5] | 4.9 [4.2–5.6] | 5.0 [4.2–5.6] | 3.1 [2.7–4.4] | 2.9 [2.4–4.3] | 2.9 [2.4–4.3] |
| Effluent CO2 emissions, CO2 equivalents per day | 9.1 [7.5–10.4] | 9.6 [8.0–10.5] | 9.8 [8.7–10.6] | 10.0 [8.9–10.9] | 10.0 [9.1–10.9] | 6.7 [0.00–8.1] | 6.7 [0.00–8.2] | 6.7 [0.0–8.1] |
| Sets CO2 emissions, CO2 equivalents per day | 6.1 [4.3–9.6] | 6.1 [3.9–9.1] | 6.0 [4.3–9.0] | 6.1 [4.0–9.4] | 5.9 [4.3–9.3] | 5.1 [3.2–8.7] | 4.9 [3.5–8.7] | 4.7 [3.4–7.6] |
| Auto effluent CO2 emissions, CO2 equivalents per day | 0 [0-0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–0.0] | 0.0 [0.0–1.0] | 0.0 [0.0–0.5] | 0.0 [0.0–0.5] |
| Total CO2 emissions, CO2 equivalents per day | 25.3 [21.5–29.7] | 25.3 [22.1–29.9] | 26.0 [23.3–30.1] | 26.4 [23.1–30.4] | 26.5 [23.8–30.0] | 20.4 [16.7–23.4] | 19.6 [16.2–23.8] | 19.3 [16.4–22.8] |
Fig. A5.
Violin plots of autoefflient, anticoagulation, effluent rate tertile versus CO2 emissions and cost
1. CO2 emissions for: A. Autoeffluent, B. Anticoagulation, C. Effluent tertile (ml/kg/hr)
2. Cost for D. Autoeffluent, E. Anticoagulation, F. Autoeffluent tertile (ml/kg/hr)
Table A6.
Mediation models for citrate, filter life and CO2-e, and cost
| CO2 equivalent | ||||
|---|---|---|---|---|
| Term | Estimate | 95% CI | p value | Interpretation |
| ACME (indirect) | −0.099 | [−0.120, −0.076] | <0.0001 | Citrate reduces CO2 by prolonging filter life |
| ADE (direct) | +0.063 | [+0.047, +0.079] | <0.0001 | Citrate increases CO2 through other means |
| Total effect | −0.034 | [−0.060, −0.008] | 0.01 | Net CO2 reduction, largely mediated by filter life |
| Proportion mediated | 2.799 | [1.81, 8.11] | 0.01 | 280% mediated. Mediation dominates; direct and indirect effects in opposite directions |
| Cost | ||||
| Term | Estimate | 95% CI | p value | Interpretation |
| ACME (indirect) | −0.127 | [−0.155, −0.100] | <0.0001 | Citrate reduces cost indirectly by extending filter life |
| ADE (direct) | +0.156 | [+0.142, +0.170] | <0.0001 | Citrate directly increases cost |
| Total effect | +0.029 | [−0.004, +0.059] | 0.08 | Net effect not significant; opposing effects largely cancel |
| Proportion mediated | −3.904 | [−31.2, +20.5] | 0.08 | High variability due to opposing direct and indirect effects, not significant. |
ACME: average causal mediation effect; ADE: average direct effect; CI: confidence interval.
Table A7.
Composition and weights of fluid bags, autoeffluent kit and outer packaging; indication of quantity of nonhazardous versus hazardous waste
| Component/material | Weight in grams |
||||||
|---|---|---|---|---|---|---|---|
| Bag | Connections/ports/tubing | Outer packaging | Caps/clamps | Miscellaneous | Total hazardous waste | Total nonhazardous waste | |
| Prism0cal B22 | |||||||
| Polypropylene | 55.62 | 47.64 (inner layer) | |||||
| Polyamide | 6.6 (outer layer) | ||||||
| Polypropylene | 6.6 (outer layer of injection port + 2 connections to patient) | ||||||
| Polyethylene | 0.2 (interface of inner and outer layers) | ||||||
| Polycarbonate | 0.63 (inner layer of injection port) | ||||||
| Total | 110.69 | ||||||
| Haemosol | |||||||
| Polypropylene | 55.62 | 47.64 (inner layer) | |||||
| Polyamide | 6.6 (outer layer) | ||||||
| Polypropylene | 6.6 (outer layer of injection port + 2 connections to patient) | ||||||
| Polyethylene | 0.2 (interface of inner and outer layers) | ||||||
| Polycarbonate | 0.63 (inner layer of injection port) | ||||||
| Total | 110.69 | ||||||
| Prisomocitrate 18/0 | |||||||
| Polyamide | 6.6 (outer layer of packaging) | ||||||
| Polypropylene | 42.51 (irrigation bag + connectors) | 39.12 (inner layer of packaging) | |||||
| Total | 88.23 | ||||||
| Autoeffluent | |||||||
| PVC (polyvinyl chloride) | 137.59 (tubing, line junctions) | ||||||
| Polyethylene | 7.65 (printed packaging) | ||||||
| MBS | 1.24 (connector) | ||||||
| PET (polyethylentherephthlate copolyester) | 76.97 (packaging) | 150.73 (casette) | |||||
| PAA | 0.9 (caps) | ||||||
| Polypropylene | 30.6 (clamps) | ||||||
| PCT | 135.56 | ||||||
| Paper | 1.33 (adhesive strip) | ||||||
| Polyethylene | 3.78 (support for line) | ||||||
| Total | 461.73 | 84.62 | |||||
| Effluent | |||||||
| Low-density polyethylene (fossil fuel based) | 5.4 (clear packaging) | ||||||
| High density polyethylene (fossil fuel based) | 0.39 (caps) | ||||||
| Paper | 6.54 (printer packaging and printed sticker) | ||||||
| Polypropylene | 8.85 (connectors) | 5.4 (clips) | |||||
| Polyamide | 6.6 (packaging) | ||||||
| PP-coated cardboard | 1.67 (instructions for clip) | ||||||
| PVC (polyvinyl chloride) | 99.62 (effluent bag, connectors, and upper pipe—small proportion of rubber) | ||||||
| Total | 113.87 | 20.6 | |||||
Values obtained during composition analysis at Melbourne University Healthcare Carbon lab.
Individual components separated and analysed using Fourier transform infrared spectroscopy, including attenuated total reflectance on Bruker ALPHA II compact FT-IR Spectrometer, to generate absorbance spectrograms (Figure A3 earlier).
Table A8.
Predicted cost and CO2-e with increasing effluent dose
| Effluent dose (mL/kg/hr) | Predicted daily cost ($) | Cost 95% CI | Predicted CO2-e (kg/day) | CO2-e 95% CI |
|---|---|---|---|---|
| 10 | 705.4 | [345.5] to [1406.2] | 18.5 | [10.2] to [32.7] |
| 15 | 758.4 | [372.2] to [1531.4] | 20.5 | [11.5] to [35.8] |
| 20 | 815.7 | [412.2] to [1637.9] | 22.4 | [12.6] to [38.5] |
| 25 | 866.8 | [438.3] to [1655.9] | 23.7 | [13.8] to [40.4] |
| 30 | 905.7 | [453.1] to [1814.7] | 24.7 | [13.8] to [42.4] |
| 35 | 924.6 | [455.5] to [1827.2] | 25.63 | [15.2] to [45.3] |
| 45 | 987.4 | [495.5] to [1968.1] | 28.1 | [15.8] to [51.5] |
| 50 | 1003.2 | [501.6] to [2043.1] | 28.2 | [16.0] to [51.2] |
CI: confidence interval.
References
- 1.Environmental sustainability and climate resilience healthcare model, A.C.o.S.a.Q.i. Healthcare, Editor. 2024. [Google Scholar]
- 2.McGain F., Story D., Kayak E., Kashima Y., McAlister S. Workplace sustainability: the "cradle to grave" view of what we do. Anesth Analg. 2012 May;114(5):1134–1139. doi: 10.1213/ANE.0b013e31824ddfef. [DOI] [PubMed] [Google Scholar]
- 3.M. Ben Hmida, M. Torreggiani, N. Berman-Parks, M. Salomone, G. Barbara Piccoli., Roadmaps to green nephrology: a mediterranean point of view, Curr Opin Nephrol Hypertens, 33 (1), 2024, 115–121. [DOI] [PubMed]
- 4.Barraclough K.A., Talbot B., Knight J., Blair S., McGain F., Masterson R., et al. Carbon emissions from different dialysis modalities: a life cycle assessment. Am J Kidney Dis. 2025 Jun;30 doi: 10.1053/j.ajkd.2025.04.019. S0272-6386(25)00920-5 Epub ahead of print. PMID: 40602678. [DOI] [PubMed] [Google Scholar]
- 5.Golsteijn L. 2022. Life cycle assessment explained.www.sustainability.com [Google Scholar]
- 6.McGain F., Sheridan N., Wickramarachchi K., Yates S., Chan B., McAlister S. Carbon footprint of general, regional, and combined Anaesthesia for total knee replacements. Anaesthesiology. 2021 Dec 1;135(6):976–991. doi: 10.1097/ALN.0000000000003967. [DOI] [PubMed] [Google Scholar]
- 7.Prasad P.A., Joshi D., Lighter J., Agins J., Allen R., Collins M., et al. Environmental footprint of regular and intensive inpatient care in a large US hospital. Int J Life Cycle Assess. 2022;27:38–49. [Google Scholar]
- 8.Bardoult P., Cadic E., Brichory O., Marie V., Rouxel C., Millet C., et al. Which carbon footprint for my ICU? Benchmark, hot spots and perspectives. Ann Intensive Care. 2025;15:35. doi: 10.1186/s13613-025-01445-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.M. Gaetani, E.Uleryk, C.Halgren, C.Maratta The carbon footprint of critical care: a systematic review, Intensive Care Med, 50 (5), 2024, 731–745. [DOI] [PubMed]
- 10.The STARRT-AKI Investigators, C.C.C.T.G., Australian and New Zealand Intensive Care Society Clinical Trials Group, United Kingdom Critical Care Research Group, Canadian Nephrology Trials Network, and Irish Critical Care Trials Group∗ Timing of initiation of renal-replacement therapy in acute kidney injury. N Engl J Med. 2020;383:240–251. doi: 10.1056/NEJMoa2000741. [DOI] [PubMed] [Google Scholar]
- 11.Barbar S.D., Clere-Jehl R., Bourredjem A., Hernu R., Montini F., Bruyère R., et al. Timing of renal-replacement therapy in patients with acute kidney injury and sepsis. N Engl J Med. 2018;379:1431–1442. doi: 10.1056/NEJMoa1803213. [DOI] [PubMed] [Google Scholar]
- 12.Zarbock Alexander, Schmidt Christoph, Van Aken Hugo, Wempe Carola, Pavenstädt Hermann, Boanta Andreea, et al. Effect of early vs delayed initiation of renal replacement therapy on mortality in critically ill patients with acute kidney injury: the ELAIN randomized clinical trial. JAMA. 2005;315(20):2190–2199. doi: 10.1001/jama.2016.5828. [DOI] [PubMed] [Google Scholar]
- 13.Meersch M., Küllmar M., Schmidt C., Gerss J., Weinhage T., Margraf A., et al. Long-term clinical outcomes after early initiation of RRT in critically ill patients with AKI. J Am Soc Nephrol. 2018;29(3):1011–1019. doi: 10.1681/ASN.2017060694. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gaudry S., Hajage D., Schortgen F., Martin-Lefevre L., Pons B., Boulet E., et al. Initiation strategies for renal-replacement therapy in the intensive care unit. N Engl J Med. 2016;375:122–133. doi: 10.1056/NEJMoa1603017. [DOI] [PubMed] [Google Scholar]
- 15.Investigators∗, T.R.R.T.S. Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med. 2009;361:1627–1638. doi: 10.1056/NEJMoa0902413. [DOI] [PubMed] [Google Scholar]
- 16.Network∗, T.V.N.A.R.F.T. Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med. 2008;359:7–20. doi: 10.1056/NEJMoa0802639. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Gallagher M., Cass A., Bellomo R., Finfer S., Gattas D., Lee J., Lo S., et al. POST-RENAL Study Investigators and the ANZICS Clinical Trials Group Long-term survival and dialysis dependency following acute kidney injury in intensive care: extended follow-up of a randomized controlled trial. PLoS Med. 2014;11(2) doi: 10.1371/journal.pmed.1001601. 11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Gattas D.J., Rajbhandari D., Bradford C., Buhr H., Lo S., Bellomo R. A randomized controlled trial of regional citrate versus regional heparin anticoagulation for continuous renal replacement therapy in critically ill adults. Crit Care Med. 2015;43(8):1622–1629. doi: 10.1097/CCM.0000000000001004. [DOI] [PubMed] [Google Scholar]
- 19.Zarbock A., Küllmar M., Kindgen-Milles D., Wempe C., Gerss J., Brandenburger T., et al. RICH Investigators and the Sepnet Trial Group Effect of regional citrate anticoagulation vs systemic heparin anticoagulation during continuous kidney replacement therapy on dialysis filter life span and mortality among critically ill patients with acute kidney injury: a randomized clinical trial. JAMA. 2020;324(16):1629–1639. doi: 10.1001/jama.2020.18618. Oct 27; PMID: 33095849; PMCID: PMC7585036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sansom B., Udy A., Sriram S., Presneill J., Bellomo R. Circuit haemodynamics during non-citrate and regional citrate continuous renal replacement, and impact of blood flow on filter life. Int J Artif Organs. 2022;45(12):988–996. doi: 10.1177/03913988221118585. [DOI] [PubMed] [Google Scholar]
- 21.Uchino S., Kellum J.A., Bellomo R., Doig G.S., Morimatsu H., Morgera S., Schetz M., et al. Acute renal failure in critically ill patients: a multinational, multicentre study. Beginning and ending supportive therapy for the kidney (BEST kidney) investigators. JAMA. 2005;294(7):813–818. doi: 10.1001/jama.294.7.813. [DOI] [PubMed] [Google Scholar]
- 22.G.B. Piccoli, A.Cupisti, F. Aucelia, G.Regolisti, C.Lomonte, M.Ferraresi et al, Green nephrology and eco-dialysis: a position statement by the Italian society of nephrology, J Nephrol, 33 (4), 2020, 681–698. [DOI] [PMC free article] [PubMed]
- 23.International Energy Agency Website; www.iea.org.
- 24.Vinyl Council of Australia; www.vinyl.org.au.