FO presents a promising alternative to RO for primary dialysate production. Zhao et al. demonstrated the feasibility of using FO for osmotic dilution to prepare dialysate directly from tap water.4 Their research found that tailor-made hollow-fiber thin-film composite (TFC) membranes outperformed commercial cellulose triacetate (CTA) membranes, achieving higher water flux, lower reverse salt flux, which is a key metric where solutes from the draw solution diffuse back into the feed, and better overall ion rejection. Although minor draw solute loss occurred, compensation mechanisms ensured the final solution met dialysate suitability standards. However, practical studies using repurposed spiral-wound RO elements have shown that flow distribution and pressure limitations can constrain performance, indicating that modules must be specifically engineered for the low-pressure conditions of FO.5 Further innovating on module design, the same group later developed a vacuum-assisted interfacial polymerization technique to fabricate high-performance FO membranes directly onto commercial dialyzer modules, demonstrating a direct pathway for integrating FO technology into existing dialysis hardware.6 This highlights FO's potential for scalable and cost-effective dialysate production, though further membrane optimization is necessary to achieve the required real-time production rates for a full dialysis session, approximately 30–50L/h.4,6

A highly impactful application of FO is in the regeneration of spent dialysate, which could drastically reduce the total pure water demand by recycling a portion of the water from the waste stream. Talaat proposed FO as a feasible option for ambulatory dialysis systems, suggesting that using a sodium chloride draw solution could recover up to 50% of water from spent dialysate.7 A performance evaluation by Kim et al. confirmed the feasibility of using FO for spent dialysate reuse, but critically highlighted that organic fouling, from uremic toxins and proteins, and inorganic scaling pose significant challenges to maintaining stable water flux, underscoring the need for effective pre-treatment or advanced anti-fouling membranes.8 The integration of advanced membranes, such as those enhanced with Aquaporin proteins, could further improve water permeability and solute rejection. However, a significant challenge identified by Dou et al. is the limited urea rejection of commercial FO membranes.9 This reduced urea concentration gradient across the membrane could potentially diminish the efficiency of toxin removal, possibly necessitating an extension of dialysis treatment time to achieve adequate clearance, thereby presenting a critical clinical trade-off that requires further investigation.9

A primary advantage of FO over RO is its significantly lower energy requirement. While RO for similar feedwater requires about 2.5kWh/m3 of water processed, the FO process itself consumes minimal energy. Zhao et al. estimated the electrical energy consumption for their FO process at atmospheric pressure to be approximately 0.05kWh/m3, which includes energy for membrane pumping (∼0.03kWh/m3) and for pumping into the draw solution recovery column (∼0.02kWh/m3).3 The total energy cost is primarily associated with the subsequent recovery and reconcentration of the draw solution, which can often be achieved using low-grade or waste heat, further enhancing its energy efficiency profile in certain settings.

The feasibility of FO in hemodialysis is contingent on membrane performance. As illustrated in Fig. 2, membrane-related issues like concentration polarization and fouling are central problems. Current-generation FO membranes, such as CTA, typically achieve water fluxes of 10–20L/m2/h, which may be sufficient for small-scale or regenerative applications but fall short of the high fluxes needed for real-time, high-volume dialysate production.4 Thin-film composite membranes show promise with fluxes approaching 40L/m2/h, but they remain susceptible to fouling by uremic toxins, proteins, and salts present in spent dialysate.8 A pilot-scale study on hollow fiber FO modules identified that the selected draw solute and its concentration have the biggest impact on membrane performance due to internal concentration polarization, a finding that underscores the challenge of achieving high, stable flux in real-world systems.10 The design of the selective layer is therefore a critical research frontier, as its properties, including thickness, hydrophilicity, and charge, dictate the fundamental trade-off between water flux and solute rejection that is so crucial for dialysis efficacy.2

Fig. 2.

Key challenges and influencing factors in forward osmosis process performance.

A highly promising avenue to overcome this trade-off is the development of biomimetic membranes. This approach draws inspiration from nature, incorporating biological components to create advanced membrane systems.11 The most significant advancement in this area involves the integration of aquaporins (AQPs), which are transmembrane proteins that act as highly selective and ultra-efficient water channels in all living organisms.12,13 First discovered by Peter Agre,14 AQPs are known as water channels as they do not function as ion/molecule carriers but facilitate water transport at an estimated rate of 700l per second per gram of protein, offering the potential for simultaneously high water flux and near-perfect solute rejection.14,15 The primary challenge lies in stabilizing these proteins within a synthetic membrane matrix. A successful strategy has been to encapsulate AQPs within robust synthetic vesicles known as polymersomes, which mimic the natural lipid bilayer and overcome the degradability of natural liposomes.16,17 These AQP-embedded polymersomes are then incorporated as additives during the interfacial polymerization process used to fabricate the selective polyamide layer of thin-film composite (TFC) membranes, significantly enhancing their permeability and selectivity.18,19 As an alternative to biological proteins, research is also advancing in the design of synthetic polymeric water channels that aim to mimic the exceptional performance of AQPs.20,21

Future development must focus on optimizing these next-generation membranes for dialysis-specific conditions. This includes further refinement of biomimetic additives, the application of anti-fouling coatings, improved pre-filtration, or novel materials that offer simultaneously high water flux and superior solute rejection, particularly for key uremic toxins like urea.

The draw solution is central to the FO process, and its efficient recovery is crucial for the system's overall efficiency and economic viability. Thermal methods, such as heating an ammonium bicarbonate solution to around 60°C to decompose it into gaseous CO2 and NH3 for separation, add an energy cost of 0.5–1kWh/m3 that partially offsets FO's inherent energy advantage.22 Alternative methods like low-pressure RO reduce energy use but add system complexity. Furthermore, for dialysis applications, any potential leakage of draw solutes into the dialysate stream must be absolutely safe for patients. Therefore, the development of novel, biocompatible draw solutions, for example, non-toxic osmotic agents like glucose-based solutions where trace remnants would be harmless, is a critical area for future research.7,23

• •

  Barrier: Current FO membrane fluxes, ranging from 10 to 40L/m2/h, fall short of the real-time dialysate production requirement of 30–50L per hour for a full system. Furthermore, fouling by uremic solutes rapidly diminishes this performance.8

• •

  Solution: Research must prioritize engineering next-generation membranes with dialysis-specific fluxes targeting 50–100L/m2/h. This could involve novel materials like graphene or advanced thin-film composites integrated with anti-fouling coatings, for example, hydrogel layers. A targeted approach to optimizing the selective layer's architecture and chemistry is essential to break the perennial flux-rejection trade-off.2 The efficacy of these membranes must be rigorously validated in trials using spent dialysate to simulate real-world conditions.

• •

  Barrier: The energy cost of thermal draw solution recovery partially negates FO's core energy advantage. Furthermore, the library of draw solutions proven to be safe and biocompatible for dialysis applications is severely underdeveloped.

• •

  Solution: Innovation is needed in low-energy recovery processes, such as osmotic backwashing or advanced membrane distillation, aiming to reduce energy demand to 0.1–0.3kWh/m3. Concurrently, a major initiative must focus on developing and testing dialysis-safe draw solutions, for example, glucose or specific electrolytes, to ensure any trace reverse solute flux meets strict AAMI water quality standards. Industry partnerships with dialysis equipment and consumable manufacturers are crucial to drive this development.

• •

  Barrier: A significant hurdle is the absence of commercially available, integrated FO systems designed for dialysis, which delays clinical adoption and widespread use.

• •

  Solution: The most viable path to scalability is through the development and deployment of pilot-scale hybrid systems, for example, FO-RO for pre-concentration or FO-sorbent for regeneration, in partner dialysis centers. Innovative module designs, such as those fabricated directly onto dialyzers,6 represent a critical step toward this integration. A strategy of incremental deployment, supported by techno-economic analyses and cost-sharing models with manufacturers, can de-risk the investment and build the necessary evidence for full-scale implementation.

Finally, comprehensive life cycle assessments are needed to quantify the environmental gains in water and energy savings, and clinical trials are essential to validate final water quality and patient outcomes. Initial costs for FO membranes and draw solution recovery may exceed RO setups, though long-term water and energy savings could offset this. Pilot studies are needed to assess this economic feasibility.