The Kidney Disease: Improving Global Outcomes (KDIGO) guidelines recommend delivering a continuous kidney replacement therapy (CKRT) dose of 20-25mL/kg/h. However, it is not clear if it is safe to use delivered CKRT doses below this recommendation. In an Original Investigation recently published in AJKD, Keisuke Okamoto and colleagues found that over 90% of the patients received CKRT with a delivered dose below the KDIGO recommendation and that a delivered CKRT dose below the median was associated with increased risk of death within 90 days.

AJKDBlog Interviews Editor Timothy Yau (@Maximal_Change) caught up with Dr. Okamoto to discuss the implications of a lower delivered CKRT dose, and the differences between US and Japanese dosing practices.

Dr. Keisuke Okamoto is a Clinical Assistant Professor in the Department of Nephrology at Nara Medical University. His clinical and research interests focus on acute kidney injury and kidney replacement therapy.

AJKDBlog: Thank you for agreeing to this interview! Your paper starts with a little background explaining the suggested CKRT dose based on KDIGO guidelines. Can you summarize briefly the findings from the ATN and RENAL study and how we came to the suggested delivered dose of 20-25 ml/kg/hr?

Dr. Okamoto: The ATN (Acute Renal Failure Trial Network) study and the RENAL (Randomized Evaluation of Normal versus Augmented Level) study are among the most important RCTs in establishing the current KDIGO guidelines-recommended CKRT dose. The ATN study compared 60-day mortality between two groups of patients receiving different CKRT doses: a lower dose of 20 mL/kg/hr and a higher dose of 35 mL/kg/hr. The study found no significant difference in 60-day mortality between these two groups. Similarly, the RENAL study compared the effects of a lower dose of 25 mL/kg/hr to a higher dose of 40 mL/kg/hr on 90-day mortality, and demonstrated the same conclusion of no survival difference. These findings suggested that increasing the CKRT dose beyond a certain threshold did not improve patient outcomes. Based on these results from two large multicenter trials, the KDIGO guidelines recommend delivering a CKRT dose of 20-25 mL/kg/hr.

AJKDBlog: You mention several factors that contribute to a lower actual delivered dose such as vascular access, filter changes/clotting, and patients going out of ICU for studies and surgeries. What studies have looked at how many patients end up with a CKRT delivered dose of less than the 20 ml/kg/hr goal?

Dr. Okamoto: A quality improvement study conducted at the University of Colorado reviewed 915 CKRT treatment sessions and found that 20.7% of treatment sessions were delivered CKRT doses below 20 mL/kg/hr. After interventions (modifications to the electronic medical record and standardized training for ICU nurses) the percentage of treatments delivering less than 20 mL/kg/hr decreased to 11.7%.  But it’s important to recognize that the delivered dose is often less than the prescribed dose to a variety of factors.

AJKDBlog: Can you tell us how the CKRT dose in Japan differs from other countries, and the reasons for this discrepancy? What was the question you were aiming to answer with your study design?

Dr. Okamoto: In Japan, the CKRT dose is typically lower than the KDIGO recommendations due to government health insurance restrictions designed to reduce healthcare costs for the aging population. The system covers a fixed amount of dialysate and replacement fluid per day, regardless of the patient’s body weight. In countries following the KDIGO recommendation, the target CKRT dose of 20-25 mL/kg/hr and the patient’s body weight determine the daily amount of dialysate plus replacement fluid. In contrast, in Japan, a fixed daily amount of 15-20 liters of dialysate plus replacement fluid, along with the patient’s body weight, determines the CKRT dose. Therefore, the usual CKRT dose in Japan is only 10-15 mL/kg/hr, mainly influenced by the patient’s body weight.

Another unique aspect in Japan is that various medical specialties manage CKRT: nephrology, critical care medicine, emergency medicine, anesthesiology, cardiology, and so on. Therefore, CKRT is not always managed by nephrologists in Japan, which is different from the practice pattern in the U.S., where CKRT is almost always managed by nephrologists. Due to the involvement of various specialties, there might be fewer opportunities for in-depth discussions regarding CKRT dosing in Japan.

In my case, I initially started my medical career as a cardiologist in Japan, taking care of many critically ill patients, including those with acute myocardial infarction and acute decompensated heart failure in the CCU. I used CKRT with this fixed prescription for dialysate plus replacement fluid rate, never questioning its effectiveness. After several years of cardiology experience, I transitioned my career to nephrology. After obtaining the ECFMG certificate, I secured a position as a nephrology fellow at the Medical University of South Carolina (MUSC). During my nephrology fellowship at MUSC, I had numerous opportunities to care for AKI patients requiring CKRT in the ICU. I learned about the concept of determining CKRT dose following the KDIGO recommendations. It was during this time that I developed a growing interest in the KDIGO-guided prescription of CKRT dose which led me to question the adequacy of the fixed prescribing practice in Japan.

During my nephrology fellowship in the U.S., the COVID-19 pandemic impacted medical supply shortages, including CKRT machines and solutions. This further motivated me to explore the concept of a “floor” CKRT dose, and I hope our study will contribute to future research on this topic.

AJKDBlog: One side note from your study is the use of your anticoagulant, which was nafamostat mesylate. This is quite different from US practice, where citrate or heparin are exclusively the anticoagulants used. Can you say a few words about this?

Dr. Okamoto: Nafamostat mesylate was developed by a Japanese pharmaceutical company and has been commercially available since the 1980s. It is now commonly used in Japan and South Korea. Nafamostat mesylate has a short half-life and low molecular weight, making it suitable for elimination through dialysis. It is used not only in CKRT but also in intermittent hemodialysis in Japan, with its anticoagulation ability monitored using activated clotting time. In the U.S., regional citrate anticoagulation is commonly used for CKRT in patients with a high risk of bleeding. In contrast, nafamostat mesylate is used for CKRT in patients with a high risk of bleeding in Japan. We do not use regional citrate anticoagulation for CKRT in Japan because calcium-free CKRT solutions are not commercially available here.

AJKDBlog: The median CKRT dose was noted to be 13.2 ml/kg/hr in your cohort. How did you split up your two groups in the study and what were the mean doses of CKRT in those groups? Were there any patients that received 20 ml/kg/hr or more of CKRT?

Dr. Okamoto: We divided the study cohort into two groups based on the median delivered CKRT dose of 13.2 mL/kg/hr. We thought it was more appropriate to use the median value as a cutoff since we do not have any specific target CKRT dose in our unique Japanese CKRT practice. The mean delivered CKRT dose was 10.9 mL/kg/hr (median: 11.3 mL/kg/hr) in the below-median group, whereas the mean delivered CKRT dose was 17.0 mL/kg/hr (median: 16.2 mL/kg/hr) in the above-median group. Only 38 patients (7.7%) in our cohort received a delivered CKRT dose of 20 mL/kg/hr or more.

AJKDBlog: Looking at the two groups (below-median and above-median), there are some expected differences like BMI (higher in the below-median group), but also it appeared that the below-median group was less ill overall with less sepsis, mechanical ventilation, and higher Hb levels. Is this overall an accurate representation of the differences in the groups?

Dr. Okamoto: Yes, this is accurate. In our study, the below-median group had a higher average BMI (which makes sense due to the fixed dialysate dose) and appeared to be less critically ill overall. This group had a lower incidence of sepsis, required mechanical ventilation less frequently, and had higher hemoglobin levels compared to the above-median group. One of the main factors contributing to these differences is the younger age of the below-median group. Younger patients generally have better overall health and a greater ability to recover, which likely influenced their lower severity of illness and better baseline health status.

AJKDBlog: Upon initial inspection, it seemed that there was similar 90-day mortality in the two groups, but in multivariable analysis, it appeared that the below-median CKRT group did worse overall. Is this correct, and does it reflect that this group was less ill overall at baseline, yet experienced the same mortality? 

Dr. Okamoto: That is correct. While the unadjusted Kaplan-Meier analysis showed similar 90-day mortality rates between the two groups, the multivariable Cox regression analysis revealed that the below-median CKRT dose group had significantly higher adjusted mortality. This suggests that despite being less ill at baseline, the lower CKRT dose was associated with worse outcomes, emphasizing the importance of adequate dosing.

Unadjusted and adjusted survival curves. (A) The 90-day unadjusted survival curves for each group, generated using the Kaplan-Meier method and assessed with the log-rank test. (B) The 90-day adjusted survival curves for each group, generated from the multivariable Cox regression analysis. The variables for adjustment included age, sex, mean arterial pressure, body mass index, urine output, Acute Physiology and Chronic Health Evaluation (APACHE) II score, presence of sepsis, mechanical ventilation use, hemoglobin, serum albumin, serum urea nitrogen, serum creatinine, and C-reactive protein. Figure 2 from Okamoto et al, © National Kidney Foundation.

AJKDBlog: Given the unique fixed low daily amount of dialysate/replacement in Japan, are there other studies that have examined the association between low CKRT doses and mortality? How does your study differ in terms of design and findings?

Dr. Okamoto: Yes, there have been two studies (both about 10 years old now) looking at the relationship between low CKRT doses and mortality in Japan

Fujii et al aimed to find the minimal effective dose of CKRT for patients with AKI. This retrospective study took place in two ICUs in Japan and involved 131 patients treated with CVVHDF. They divided patients into lower-dose and higher-dose groups, using the median prescribed CKRT dose of 16.7 mL/kg/hr. Their results showed no significant difference in hospital mortality between the two groups, suggesting that lower doses of CKRT did not increase mortality.

Uchino et al looked at the impact of low-intensity CKRT on patient outcomes in a multicenter retrospective study with 343 patients across 14 ICUs in Japan. They compared Japanese data with the BEST Kidney Study database, using various CKRT intensity levels. Despite a lower median prescribed CKRT dose in Japan (14.3 mL/kg/hr) compared to the BEST Kidney Study database (20.4 mL/kg/hr), ICU and hospital mortality rates were lower in the Japanese group. However, there were no statistically significant differences in mortality among the different CKRT dose groups in the combined analysis.

Our study differs from these previous studies in a few important ways. We used a different cutoff value of 13.2 mL/kg/hr, which was lower than that used by Fujii et al. Additionally, Uchino et al compared Japanese data with international data and did not compare within the Japanese low-dose CKRT cohort.

While our study results differ in some points from the findings of the previous two studies, all three studies suggest that the “floor” CKRT dose may be below the KDIGO recommendation of 20 mL/kg/hr. This indicates that the lower fixed doses used in Japan may have different implications for patient outcomes. Therefore, our study aligns with previous Japanese studies in exploring lower CKRT dose thresholds, suggesting that the minimum effective dose could be lower than the current KDIGO guidelines recommend.

AJKDBlog: One of my big takeaways from reading your study is that while we have established the “ceiling” dose of CKRT based on trials like ATN and RENAL, we have not yet established the “floor.” How might the findings from your trial inform future prospective studies to examine a threshold dose for CKRT?

Dr. Okamoto: Our study highlights the need for further research into the “floor” CKRT dose, as it suggests that the minimum effective CKRT dose might be lower than the current KDIGO recommendation of 20 mL/kg/hr. Future prospective studies could focus on identifying this “floor” CKRT dose by comparing various low-dose CKRT regimens with the KDIGO-recommended CKRT dose to determine the minimum effective CKRT dose. Identifying the “floor” CKRT dose through future prospective studies could expand the range of treatment strategies available for CKRT. This would be particularly valuable in situations where there may be a shortage of medical supplies. Additionally, understanding the minimal effective dose could provide new insights into the cost-effectiveness of CKRT, helping to optimize resource use while maintaining patient care quality.

To view Okamoto et al [Subscription Required], please visit AJKD.org.

Title: Low-Dose Continuous Kidney Replacement Therapy and Mortality in Critically Ill Patients With Acute Kidney Injury: A Retrospective Cohort Study
Authors: Keisuke Okamoto, Hidetada Fukushima, Masahiko Kawaguchi, Kazuhiko Tsuruya
DOI: 10.1053/j.ajkd.2024.01.526