, Introduction
Growing concerns about the environmental impact of kidney care in the context of climate change and sustainable healthcare have given rise to the concept of green nephrology [1–3]. Because kidney replacement therapies have substantial resource requirements, attention has increasingly focused on hemodialysis as a major driver of environmental impact. Hemodialysis is the predominant modality of kidney replacement therapy, used by approximately 90% of dialysis patients worldwide, with an estimated 3.4 million individuals receiving treatment globally [4]. Standard hemodialysis consumes approximately 0.5 m3 of water per session, largely due to reverse osmosis (RO) reject water, resulting in an estimated annual water consumption of approximately 80 million m3 per million patients [5–7]. The global dialysis population grows at an annual rate of approximately 7%; therefore, water use and wastewater generation from dialysis facilities are expected to increase proportionally [8]. As RO water production requires substantial energy, greater water use directly results in higher electricity consumption and carbon emissions [4,9].
Although most hemodialysis treatments rely on centralized RO systems, portable RO units are used in intensive care units (ICUs), isolation wards, emergency settings, and for home hemodialysis (HHD) when a centralized infrastructure is unavailable. These systems are essential for maintaining dialysis services and controlling infection in decentralized environments. In routine practice, portable RO units often remain idle during patient turnovers, procedural delays, or system disinfection. Continuous or periodic water circulation during standby is required to prevent stagnation, bacterial adhesion, biofilm formation, and deterioration of water quality [10]. However, maintaining system readiness under these conditions consumes substantial water and electricity. From a green nephrology perspective, such standby losses represent a modifiable source of environmental burden, without compromising patient care.
In this study, an automatic Water and Power Reduction Mode (WPRM) for portable RO systems was developed to reduce water and energy consumption during Idle Standby Mode. The performance of this approach was evaluated in terms of resource savings, while maintaining adequate system responsiveness.
Methods
Internal configuration of the portable reverse osmosis system
In this study, we used a Synopex PRO 800 portable RO unit. Feedwater enters through the water inlet, and its conductivity is checked by a conductivity sensor. The water then passes through the pretreatment filters and is stored in the water tank, which has a water-level sensor. The RO pump pushes the stored water into the RO module, where it is divided into permeate and concentrate. The permeate passes through flow meter #1 and a conductivity sensor before it is used, and any unused permeate returns to the water tank through the permeate recycle line via flow meter #2. The two flow meters provide the flow signals used to detect when no water is being used, which triggers WPRM. On the concentrate line, the needle valve sets the membrane pressure and concentrate flow, helping keep a steady pressure and flow across the RO membrane. The three-way solenoid valve directs the concentrate either to drain or back to the water tank, which determines the WPRM submode (Fig. 1).
Development of the Water and Power Reduction Mode
The idle standby detection and mode switching were based on two flow sensors located in the product and return lines. Entry into WPRM was determined using a dual-verification logic that simultaneously compared flow readings from both sensors; the system classified a no-consumption state only when the difference between the two flows remained below 3% for a predefined period. This design ensured that even minimal product water usage prevented transition into WPRM. In this mode, the pump speed was reduced from 750 to 300 revolutions per minute (rpm), and the solenoid valve on the concentrate line was switched to a low-flow internal circulation loop to maintain minimal system flow. The target conductivity was set at 5 µS/cm, consistent with the quality standards for hemodialysis water. When product water use was detected by either flow sensor, the pump immediately accelerated to a normal operating profile. To avoid unnecessary and frequent transitions between operating states, a time hysteresis function was incorporated into the control logic.
Measurement of water and power consumption
The performance evaluation included power consumption, water consumption, and reactivation time to achieve the target conductivity of 5 µS/cm. The system is operated under Standby Mode, Auto RO Mode (normal operation), Idle Standby Mode, and WPRM. The WPRM was subdivided into three submodes according to concentrate water handling: WPRM-100D, all concentrated water was drained; WPRM-50R, half of the concentrate was returned to the tank; and WPRM-100R, all concentrate was returned to the tank. These operating strategies were designed to maintain the water quality and membrane stability while achieving maximum resource efficiency (Table 1). For each mode, water consumption, power consumption, and conductivity recovery time were measured in 10 replicate runs, and the results are reported as mean ± standard deviation together with the coefficient of variation (CV).
Water quality analysis
To verify the maintenance of water quality under WPRM operation, product water samples were comprehensively analyzed. Samples were collected under steady-state conditions and immediately after switching back to Auto RO following approximately 30 minutes of operation in each WPRM mode, reflecting clinically relevant scenarios. The physicochemical parameters of the produced water were evaluated, including pH, major ions, anions, residual disinfectants, and trace metals, and the results were compared with the reference limits for purified and hemodialysis water according to ISO 13959 standards. Trace and major metals were quantified by inductively coupled plasma mass spectrometry or ICP-optical emission spectrometry with the following detection limits: copper, 0.001 mg/L; sodium, 0.1 mg/L; lead, 0.001 mg/L; magnesium, 0.01 mg/L; barium, 0.001 mg/L; arsenic, 0.001 mg/L; selenium, 0.005 mg/L; mercury, 0.0002 mg/L; zinc, 0.001 mg/L; aluminum, 0.005 mg/L; silver, 0.001 mg/L; cadmium, 0.0005 mg/L; potassium, 0.01 mg/L; calcium, 0.01 mg/L; chromium, 0.001 mg/L; antimony, 0.001 mg/L; beryllium, 0.0001 mg/L; and thallium, 0.0001 mg/L. Anions were analyzed by ion chromatography: fluoride, 0.05 mg/L; nitrate, 0.05 mg/L; and sulfate, 0.01 mg/L. Free chlorine, chloramine, and total chlorine were determined by the N,N-diethyl-p-phenylenediamine colorimetric method, each with a detection limit of 0.02 mg/L.
Microbiological safety was assessed by monitoring endotoxin levels during prolonged repeated use without disinfection. The system was operated for 6 hours each day for a total of 168 hours (Auto RO for 4 hours/day and WPRM-50R for 2 hours/day) without any disinfection cycle, and product water samples were collected 5 minutes after the start of operation each day. Endotoxin measurements were performed using the Endosafe PTS150 with an Endotoxin-LAL cartridge (method KX-122), with a detection range of 0.05–5 endotoxin unit (EU)/mL. All analyses were conducted using standardized methods in an accredited laboratory.
Results
Water consumption
Water consumption varied according to the pump speed and operational mode (Fig. 2). In the Auto RO and Idle Standby modes, water consumption increased proportionally with the pump speed, whereas the WPRM maintained a constant consumption level because the pump speed was fixed at a lower value. Depending on concentrate handling, water consumption was 526.6 mL/min in WPRM-100D, 267.3 mL/min in WPRM-50R, and 0 mL/min in WPRM-100R. The WPRM-50R mode achieved an approximately 50% additional reduction in feedwater use compared with WPRM-100D, whereas WPRM-100R eliminated feedwater use during standby periods.
Power consumption
The electrical power consumption exhibited a clear dependence on the pump speed and operating conditions (Fig. 3). In the Auto RO and Idle Standby Mode, power consumption increased linearly with pump speed. At identical speeds, power consumption was higher in the Idle Standby Mode than in Auto RO because the absence of product water flow caused pressure accumulation in the line, thereby increasing the pump load. The average power consumption across the WPRM submodes ranged from 34.9 to 35.6 W, representing a substantial reduction compared with 136.7 W in Idle Standby Mode. This corresponds to a power-saving effect of up to 74%.
Reactivation performance and conductivity stabilization
The Idle Standby Mode allows for immediate use owing to continuous recirculation. Among WPRM modes, WPRM-100D consumed 526.6 mL/min of water and 34.9 W, WPRM-50R used 267.3 mL/min and 35.3 W, and WPRM-100R eliminated water consumption but required 95 seconds to stabilize conductivity. Despite slight differences, all WPRM configurations significantly outperformed the full standby shutdown in terms of reactivation speed. Table 2 summarizes water and power consumption and the time required to reach 5 µS/cm conductivity.
Water quality assessment
The product water generated by the portable RO system consistently met all the applicable water quality standards across the tested conditions. Key parameters including pH, major ions, trace metals, and residual disinfectants remained within acceptable limits, with most contaminants below the detection thresholds. These findings indicate that reduced-flow and low-power operation under WPRM did not compromise the physicochemical water quality. The detailed analytical results are presented in Table 3.
Endotoxin remained at low levels and did not show a continuous upward trend over time. A transient increase to 0.154 EU/mL was observed at 72 hours, but this decreased to 0.053 EU/mL at 96 h and remained stable at 0.059–0.075 EU/mL between 120 and 168 hours. All measured values remained below the maximum allowable limit of 0.25 EU/mL, and most stayed below the action level of 0.125 EU/mL. These results indicate that, under repeated daily RO use, endotoxin levels did not progressively accumulate but were instead maintained at a relatively stable and low level, supporting the microbiological safety of the system during prolonged operation (Fig. 4).
Discussion
The WPRM developed in this study effectively addresses excessive water and energy consumption in portable RO systems during Idle Standby Mode, while preserving rapid system reactivation. Compared with conventional standby operation, WPRM reduced electrical power and water consumption by up to 74% and 82%, respectively, without compromising dialysis water quality or operational readiness. These efficiency gains were achieved by reducing the pump speed and recirculating concentrated water through a controlled internal loop. This approach lowers the hydraulic load and minimizes unnecessary water drainage during periods of non-use. Importantly, different WPRM configurations demonstrated a clinically meaningful balance between maximal resource savings and reactivation performance, indicating that substantial environmental benefits can be realized without compromising system stability or responsiveness in critical care settings. In particular, all WPRM modes restored the target water quality markedly faster than the conventional standby mode, which is advantageous for rapid dialysis initiation in emergency or ICU settings.
Hemodialysis is highly water-intensive, requiring approximately 300–600 L of tap water per patient per week and an estimated 265 million m3 annually worldwide, which creates a fundamental barrier to access in regions affected by water scarcity, where dialysis access and dialysate microbiological quality often fall short of acceptable standards [5,11–15]. In addition, disasters and armed conflicts can abruptly disrupt water, power, and equipment supply chains, terminating ongoing dialysis sessions even when facilities remain physically intact, and highlighting the need for both long-term resource availability and short-term operational resilience [16–18]. To mitigate these challenges, complementary water conservation strategies have been proposed both at the prescription and infrastructure levels. For example, at the prescription level, incremental hemodialysis and reduced dialysate flow rates can lower water use while maintaining adequate dialysis [9,19]. At the infrastructure level, RO reject water has been reused for non-potable purposes such as toilet flushing, cleaning, and irrigation, including simple collection systems used by patients on HHD [7,20,21]. These represent only a few examples among many possible strategies, and further progress may be achieved by developing compact, integrated solutions that reduce water and energy use within the RO system itself. For portable RO systems in particular, how the concentrate is handled during standby is a key design decision, since it directly affects both water savings and product water quality. The WPRM developed in this study addresses this issue at the system level, and the way the concentrate is routed in each submode helps explain the water quality patterns we observed in our results.
In principle, modes with more concentrate drainage (such as WPRM-100D) would be expected to produce better water quality, because less concentrate is returned to the water tank, while modes with more recirculation (WPRM-50R and WPRM-100R) would be expected to be less favorable. In our results, WPRM-100D and WPRM-50R showed similar water quality within the margin of error, while WPRM-100R consistently showed less favorable values. The similarity between WPRM-100D and WPRM-50R can be explained by the fact that the overall composition of the water in the tank did not differ much between the two conditions. In WPRM-100D, the permeate is sent back to the tank while all of the concentrate is drained, and the drained volume is replaced by fresh feedwater after passing through the pretreatment filters. In WPRM-50R, half of the concentrate is returned to the tank and the other half is drained, and the drained volume is again replaced by fresh feedwater. As a result, the tank in WPRM-100D contains a mixture of recycled permeate and fresh feedwater, while the tank in WPRM-50R contains recycled permeate, half of the concentrate, and fresh feedwater. Within the short operating time and limited replenishment used in this study, the dissolved ion levels in the tank did not differ much between these two compositions, so the resulting product water quality was similar within the margin of error. In contrast, in WPRM-100R, both the permeate and the concentrate are fully returned to the tank with no external drainage, so no fresh feedwater is added during operation. Under the low pump speed (300 rpm) used in WPRM, the pressure and flow conditions across the RO membrane are also less favorable, which can explain why WPRM-100R showed somewhat lower product water quality. Even in this conservative, worst-case configuration, however, the measured water quality remained within the applicable reference limits, supporting the overall safety and robustness of the WPRM design across all three submodes.
Nonetheless, this study had several limitations. First, the evaluation was conducted under controlled experimental conditions using a portable RO system, and the long-term performance, durability, and operational stability in routine clinical environments were not assessed. Second, formal cost-effectiveness analyses and data on real-world implementation, including maintenance requirements and feasibility across different facility settings, are beyond the scope of this study. For example, the pump speed is closely related to factors such as the pump diameter, tubing configuration, and hydraulic resistance, so that the reported values represent relative differences within our RO platform and are not directly comparable to those of other RO systems. Third, microbiological monitoring relied solely on endotoxin measurement, and bacterial colony-forming unit counts were not obtained. Although endotoxin levels remained low and stable throughout prolonged repeated use, this cannot fully exclude biofilm formation or bacterial stagnation within the RO membrane and tubing. Finally, although our mechanistic interpretation suggests that WPRM-100R represents a worst-case operating condition, this was not directly verified, as our physicochemical analysis relied on single-time-point measurements. Serial monitoring of dissolved salts, trace contaminants, and disinfection byproducts during extended standby is therefore needed in future work.
Despite these limitations, this study presents a practical infrastructure-level approach to reduce water and energy consumption without requiring changes in routine dialysis workflows. Although the absolute magnitude of resource savings per individual system may be modest, this study demonstrates that environmental efficiency can be deliberately incorporated into the design and development of dialysis water treatment systems. Future efforts should integrate compact reject-water reuse functions into next-generation RO platforms and validate their performance in real-world clinical settings to incrementally reduce the environmental footprint of hemodialysis while maintaining safety and reliability.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
