Hemodialysis (HD) has progressed remarkably over the past five decades and has become the most widely used form of kidney replacement therapy worldwide. Ideally, treatment should remove small solutes such as urea and creatinine as well as middle and protein-bound molecules implicated in inflammation, oxidative stress, and cardiovascular disease, while maintaining intradialytic safety and comfort. The modern classification of uremic toxins highlights that many clinically relevant solutes - β₂-microglobulin (B2M), cytokines, and advanced glycation products - fall beyond the diffusion range of conventional membranes (1). International guidelines have therefore expanded dialysis adequacy from urea kinetics alone to include middle-molecule removal and patient-centered outcomes (2).

OL-HDF was introduced to augment conventional high-flux HD through convective transport, enhancing the clearance of solutes above 15–25 kDa. Large multicenter trials, including the CONVINCE trial (3), the ESHOL study (4), and the recent individual-patient review and meta-analysis of randomized controlled trials (5, 6), demonstrated consistent survival benefits for high-volume OL-HDF compared with high-flux HD, following a dose–response relationship with convective volume. A 2021 consensus by the ERA Dialysis Working Group concluded that HDF is associated with improved overall and cardiovascular survival, provided that convection volumes are sufficiently high (7). Despite its proven efficacy, routine implementation of OL-HDF is restricted in many low- and middle-income countries (LMICs) because of the need for ultrapure water, online substitution monitoring, and rigorous fluid-quality control (8).

On the other hand, SHF dialyzers have emerged as a pragmatic dialysis method to extend the removal of middle-large molecules besides OL-HDF, or serve as an alternative option where OL-HDF cannot be universally implemented. These membranes widen pore-size distribution and optimize fiber geometry to enhance middle-molecule permeability with acceptable albumin loss. Randomized trials in Thailand have shown that SHF-HD can achieve short-term removal of protein-bound and middle-molecule toxins comparable to post-dilution high-volume OL-HDF (9–11). Observational data from Japan indicate that SHF-HD yields survival outcomes equivalent to OL-HDF when albumin leakage is matched (12), and the 2024 registry analyses from the Japanese Society for Dialysis Therapy provided robust epidemiological evidence that use of type V dialyzer – or super high-flux dialyzer – correlates with better survival (13). Similar trends are observed with newer medium-cutoff (MCO) membranes, which extend permeability to larger middle molecules while maintaining albumin safety (14–16). Together, these findings suggest that optimized diffusive clearance may narrow the performance gap with convective therapy. The consensus so far is that if SHF-HD can remove middle molecules to the same degree as HDF, it should in theory confer similar benefits, but this has not been definitively proven, especially in low-resource countries. In Vietnam, where dialysis resources are stretched and OL-HDF is not yet widespread, nephrologists have innovated within HD programs. Nguyen Huu Dung and colleagues adopted a practical strategy by incorporating hemoperfusion into routine HD, aiming to mimic the toxin removal benefit of OL-HDF (17). Their pilot study reported a noticeable reduction in cardiovascular mortality with the HD plus hemoperfusion approach, suggesting that better removal of middle molecules may translate into meaningful clinical benefit even when resources are limited. Removing solutes is only part of what makes a dialysis technique clinically meaningful. Any modality that claims to be “advanced” must also be judged on safety, tolerability, and overall biocompatibility, since intradialytic complications remain a practical concern for both clinicians and patients. SHF dialyzers, for instance, have not been shown to have higher rates of hypotension compared with other membranes. Albumin handling is another crucial issue. These super-permeable membranes still need to retain albumin effectively; otherwise, they risk exacerbating malnutrition and worsening long-term outcomes. High-volume HDF inevitably loses a few grams of albumin per session, which is generally acceptable but still worth monitoring. In a crossover randomized trial from Thailand, Tiranathanagul and colleagues found that patients tolerated both HDF and SHF-HD used alongside hemoperfusion, without differences in intradialytic events, reinforcing the idea that these modalities can be used safely when properly implemented (10). Serum albumin level remained unchanged after an 8-week study period, despite the observation of greater albumin in dialysate during combination of SHF and hemoperfusion dialysis than during HDF.

Vietnam’s dialysis system reflects these global contrasts. Maintenance hemodialysis has expanded quickly across the country, yet the use of OL-HDF remains modest. Cost constraints, machine limitations, and the demanding requirements for water quality control all contribute to this slow uptake, leaving many centers reliant on conventional high-flux HD despite interest in more advanced modalities (8, 18–20). Although SHF-HD has been used abroad for more than a decade, its introduction into Vietnam is recent, and it is quickly becoming a practical bridge toward more modern dialysis care. When looking across the international literature, a fairly consistent pattern emerges: both OL-HDF and SHF-HD tend to outperform conventional high-flux HD in clearing middle-molecule toxins. Yet nearly all of these data come from non-Vietnamese cohorts, which leaves several questions unanswered for local practice. No published study has evaluated SHF-HD in Vietnamese patients or directly compared it with OL-HDF in this context. That gap matters. Differences in diet, body size, and center-level treatment protocols could influence how well SHF membranes perform, and we do not know whether safety profiles regarding blood pressure stability, electrolyte handling, or albumin loss hold true in Vietnamese settings. These gaps create the small but important space that our study steps into. In short, Vietnam needs evidence generated within its own dialysis units to determine whether adopting SHF-HD is both feasible and beneficial. Our work contributes to that effort by examining SHF-HD in real-world Vietnamese patients and comparing its performance with OL-HDF for the first time. We hypothesized that SHF-HD would provide comparable short-term solute removal and safety profile to OL-HDF, thereby establishing a basis for future multicenter studies and guiding modality selection in resource-limited dialysis settings.

A total of thirty ESKD patients undergoing HD were included, of that fifteen were treated with SHF-HD, and fifteen were treated with OL-HDF.

Inclusion criteria:

Exclusion criteria:

The primary outcome was the short-term efficacy of SHF-HD, assessed by reduction ratio (RR) of middle molecules after dialysis, compared with OL-HDF. Secondary outcomes comprised intradialytic complications and post-dialysis changes in serum albumin and CRP, representing treatment safety and biocompatibility, and were likewise compared with OL-HDF.

In Vietnam, the indication of OL-HDF and HD using SHF dialyzers follows the criteria established by the Ministry of Health (Official Decision No. 3365/QĐ-BYT) (21). Patients are eligible for OL-HDF when presenting with any of the following conditions: hyperphosphatemia-related disorders, malnutrition, anemia with poor response to erythropoietin, uremic pruritus, infection-related complications, amyloidosis, cardiovascular or neurological complications related to ESKD, refractory hypertension, or in acute emergencies requiring cytokine removal when conventional hemodialysis is insufficient. The indication of hemodialysis using SHF dialyzers is similar to those for conventional low-flux and high-flux dialyzers.

At the Dialysis Unit of People’s Hospital 115, treatment modality assignment adheres to both national and institutional policies. Patients are prescribed OL-HDF if they meet the national criteria. The use of SHF dialyzer is indicated for patients who meet the criteria for OL-HDF but are unable to undergo the procedure due to specific limitations such as: patient’s refusal, financial constraints, poor vascular access, or failure to achieve the sufficient Qb required for HDF. The decision to use SHF dialyzer therefore represents a technical and pragmatic institution, aimed at enhancing the middle molecule clearance in patients who cannot access OL-HDF, while maintaining the same clinical indication framework defined by the Vietnam’s Ministry of Health.

Dialysis was performed using Fresenius dialysis machines (Fresenius Medical Care, Germany) model 4008S for SHF-HD and 5008S for OL-HDF, both were connected to a two-stage reverse osmosis system with DIASAFEⓇ plus filters (Fresenius Medical Care, Germany) to ensure water purity. SHF sessions used Elisio-17HX or Elisio-19HX dialyzers (Nipro Corporation, Japan). OL-HDF was delivered in post-dilution mode. All treatments used dialysate flow rate of 500 mililiters per minute and a treatment duration of 240 minutes dialysis session. All participants received bicarbonate-based dialysate, and dialyzers were single-use during sampling to eliminate reuse bias.

For each patient, venous blood sample was drawn immediately before and after a single 240-minute dialysis session using a slow-flow technique following the 2015 KDOQI guideline (2) to prevent access-related fluctuations. Main biochemical molecules included B2M, interleukin-6 (IL6), parathyroid hormone (PTH), leptin, albumin, and CRP.

The corrected reduction ratio was calculated following the hematocrit-corrected formula according to Daniel Schneditz et al. (22), applied for PTH, β2MG, IL6 and Leptin:

The reduction ratio of Urea and Phosphate using the following formula:

H₀ and C₀ represent for the hematocrit and solute concentration from the blood line of pre-dialyzer at time 0 minute. H₁ and C₁ denote the hematocrit and solute concentration from the blood line of pre-dialyzer at time 240 minutes.

The Qb in the HDF group was significantly higher than that in the SHF group (p = 0.008). Other treatment session parameters were similar between the two groups. At the baseline, there were no statistically significant differences in any laboratory marker concentrations between the HDF group and the SHF group. Details of all biochemical marker concentrations and dialysis parameters are presented in Table 2.

The heatmap illustrates the correlations among pre-dialysis biochemical markers. B2M showed a positive correlation with both IL6 and leptin, with moderate correlation strength (r = 0.51, p = 0.004; and r = 0.41, p = 0.025, respectively). IL6 also demonstrated a positive, moderate correlation with CRP (r = 0.5, p = 0.005), and leptin also showed the a moderate correlation with calcium (r = 0.43, p = 0.017). Other marker pairs showed only weak and non-significant correlation. Further information on correlation matrix is shown in Fig. 1.

Existing evidence on SHF or expanded HD largely originates from high-income countries (Japan, Europe) or upper-middle-income settings such as Thailand. To our knowledge, this study is the first to investigate the use of SHF dialyzers in Vietnam, and one of the first from any LMICs, to evaluate the efficacy and safety profile compared with OL-HDF. In this light, it is noteworthy that the SHF dialyzer in our study achieved middle molecules reduction comparable to OL-HDF. The primary endpoint, B2M, cytokines, and peptide hormones reduction did not differ significantly between modalities, as expected in four-hour treatment with both groups. Albumin declined insignificantly following SHF-HD and nonsignificantly to OL-HDF, and CRP did not alter in either group. None of the acute safety events were observed. Our findings are in line with prior controlled studies in which SHF-HD achieved comparable clearance of solutes and middle molecules to high-volume OL-HDF, without acute safety concerns (9, 11, 12, 16).

B2M is a surrogate for middle molecules and is the basis for dialyzer flux classification in Japan. Over 90% of Japanese HD patients have been treated with type IV/V high-performance membranes for over a decade, which has been associated with lower B2M levels and a decline in amyloidosis complications. Notably, a 2024 Japanese Society for Dialysis Therapy registry study found that the use of type V dialyzers with B2M clearance ≥ 70 mL/min was independently associated with improved 2-year survival, with adjusted HR 0.86 (95% CI, 0.80–0.92, p < 0.0001) vs. standard HF membranes (13). For instance, recent studies in Thailand demonstrated that thrice-weekly SHF-HD eliminates middle molecules and protein-bound toxins as effectively as post-dilution OL-HDF. In one such trial, RRs for indoxyl sulfate and B2M in the SHF-HD group were statistically indistinguishable from those on high-volume OL-HDF with the convective volume achieved 24–26 liters per session (9). We also measured IL6 and leptin as markers of larger uremic toxins. Both showed a 20%-40% reduction per session with no differences between modalities. These larger cytokines and adipokines are poorly removed by diffusion alone, but HDF can remove some through convection. The efficacy is probably supported by the membrane properties of SHF dialyzers. Larger and more uniformly distributed pores increase the sieving coefficient for middle molecules while maintaining a cutoff below albumin’s molecular size. In the dialyzer, the secondary flows and Taylor dispersion have contributed to the amplification of the mass transfer beyond pure diffusion at higher rates of blood flow, which is sometimes called internal convection. All these phenomena, along with the long fibres and the optimised packing of fibres, can increase the transport of larger solutes, which can be transported without the presence of a substitution fluid supplied externally. Some studies even suggest that “expanded HD” with novel membranes might exceed OL-HDF in removing the largest solutes: for example, certain SHF or MCO dialyzers have shown higher clearance of cytokines than OL-HDF (11).

Another important dimension is biocompatibility and inflammation. Persistent low-grade inflammation in dialysis patients has long been tied to their high burden of cardiovascular disease. OL-HDF might attenuate this inflammation over time by two mechanisms: the requirement for ultrapure dialysate reduces exposure to endotoxins, and convective transport can remove some cytokines. In the short term, however, we observed no significant difference in inflammatory markers between SHF-HD and OL-HDF. This suggests that both modalities were similarly biocompatible in our setting with real-world dialysis fluid quality in both arms and modern synthetic membranes. Panichi et al. observed that switching from standard HD to HDF led to significant reductions in CRP and IL6 levels over 4 months, whereas reverting to conventional HD caused these markers to rise again (23). In our comparison, we did not detect any short-term difference in inflammatory markers between SHF-HD and OL-HDF, which is reasonable given that acute inflammatory shifts rarely emerge within a single session. Even so, membrane design may matter more than the modality alone. Recent work with vitamin E–coated SHF dialyzers is a good example. In a randomized study, the vitamin E-coated membrane achieved middle-molecule clearance similar as an MCO dialyzer, yet it also lowered oxidative and inflammatory biomarkers (24). Findings like these hint that incremental refinements in membrane biocompatibility might influence the inflammatory response, although whether these effects persist or translate into clinical benefit remains uncertain. Across the 30 sessions we followed, no intradialytic complications occurred, not even the small issues that sometimes surface when membranes are changed. This pattern is similar to the experience reported by Tiranathanagul and colleagues in Thailand, where patients tolerated SHF-HD just as well as OL-HDF and showed no rise in hypotension or other symptoms over their eight-week crossover periods (10). Taken together, these findings support the view that well-designed high-flux membranes can offer HDF-like clearance while remaining safe and comfortable for patients in the short term, although longer-term effects still need clearer evidence.

Our study has several limitations. The most obvious is the small sample. With only fifteen patients in each group, our ability to detect very small differences was inevitably limited. Even so, the overlap in almost every measured parameter was striking, and that consistency gives at least some confidence that the patterns we observed were not simply noise. Another limitation involves the blood-flow rates. Both groups operated at lower Qb than what is typically reported in larger trials (9, 11, 16), sometimes because of protocol constraints and sometimes because of patient factors that restricted achievable flow. This makes comparison with other studies a little tricky, and it raises the possibility that some performance differences could be masked by lower baseline flow. Lower blood-flow rates often limit the convective volume that HDF can generate, and this can blunt the clinical advantages it would otherwise offer. That concern was relevant in our setting. Although the HDF group in our study operated at a higher Qb than the SHF group, about 270 compared with 240 mL per minute, the SHF treatments still matched HDF in terms of toxin removal. This outcome is striking because, at least in theory, the lower Qb should have placed SHF-HD at a disadvantage. Yet we did not see any meaningful reduction in clearance. It raises the possibility that membrane performance in SHF-HD may compensate for modest flow limitations, although we cannot say how far this effect extends beyond the short-term observations we captured. Third, our study did not measure albumin concentration in the dialysate or ultrafiltrate, so we could not directly quantify albumin loss through the dialyzer. We did observe that serum albumin levels after treatment were unchanged in both groups, which is an encouraging indirect indicator that albumin loss was minimal. Furthermore, the mean albumin decreases during SHF-HD (−0.40 g/L) was virtually identical to that during OL-HDF (−0.44 g/L). In OL-HDF, a small amount of albumin, typically 1–3 g per session, is known to be lost into the discarded ultrafiltration fluid. It stands to reason that SHF dialyzers might allow a similar scale of loss. Evidence from other studies supports this: a crossover trial by Tiranathanagul et al. noted slightly higher albumin leakage into dialysate with SHF-HD compared to HDF, but importantly no significant drop in patients’ serum albumin over 8 weeks of treatment (10). Likewise, long-term observational work from Thailand found that 15 months on SHF-HD did not lead to hypoalbuminemia; mean serum albumin remained approximately 4.0 g/dL, unchanged from baseline (9). Still, we acknowledge that without direct measurement, we cannot definitively state the grams of albumin lost in SHF-HD. And finally, because our work was limited to a short observational window, we were not able to examine outcomes that unfold slowly, such as survival, cardiovascular events, or changes in quality of life. That is a real constraint, and it leaves open questions that a longer study would be better suited to answer. Even so, it does not undermine the central pattern we observed: over the short term, SHF-HD performed on par with OL-HDF in both solute clearance and safety. If anything, these limitations simply point toward the next steps, inviting larger and longer studies to determine whether the short-term equivalence we found holds up when patients are followed over months or years.

These results are best interpreted within the context of routine dialysis practice in settings where OL-HDF is not yet widely available, as is still the case in many centers in Vietnam. Conventional high-flux HD remains highly effective for the removal of small solutes, yet its limited capacity for middle-molecule clearance continues to be a concern given the established links between these toxins and complications such as dialysis-related amyloidosis, chronic inflammation, and cardiovascular disease. OL-HDF offers a more efficient solution for this problem, but its broader implementation is constrained by the need for ultrapure water systems, compatible machines, and higher operational costs. Against this background, SHF membranes represent a technically simpler approach that can be readily applied on standard HD platforms and may enhance middle-molecule removal under everyday clinical conditions. From a physiological standpoint, our findings are consistent with the growing body of work supporting the concept of “expanded HD,” in which enhanced diffusive transport and internal convection achieved by SHF or MCO membranes can approach the biochemical clearance profile of post-dilution OL-HDF. Nevertheless, the present study was designed to capture only short-term biochemical and safety outcomes, and any inference beyond acute solute removal would go beyond what our data can reliably support. While SHF-HD may therefore be considered a pragmatic option in centers where OL-HDF cannot be routinely implemented, its influence on longer-term outcomes such as survival, cardiovascular events, or hospitalization cannot be addressed by this study. Importantly, we observed no acute safety concerns, including hemodynamic instability or meaningful albumin loss, which supports the short-term tolerability of SHF-HD in a real-world clinical setting. Whether this favorable short-term biochemical and safety profile translates into sustained clinical benefit will require confirmation in larger studies with longitudinal follow-up.

In this cross-sectional study conducted under routine clinical conditions in Vietnam, SHF-HD achieved short-term removal of small and middle-molecule solutes comparable to OL-HDF, with minimal changes in serum albumin, stable inflammatory markers, and no intradialytic adverse events. These findings indicate that, in the short term, SHF-HD can provide a biochemical performance similar to OL-HDF in everyday practice. However, given the limited sample size and single-session design, no conclusions can be drawn regarding long-term clinical outcomes. Larger, prospective studies with extended follow-up are needed to clarify the sustained efficacy and safety of SHF-HD in comparison with OL-HDF.