Our Team

OBJECTIVESThe rising diversity of food preferences and the desire to provide better personalized care provide challenges to renal dietitians working in dialysis clinics. To address this situation, we explored the use of a large language model, specifically, ChatGPT using the GPT-4 model (openai.com), to support nutritional advice given to dialysis patients.RESULTSChatGPT generated a daily menu with five recipes. The renal dietitian rated the recipes at 3 (3, 3) [median (Q1, Q3)], the cooking instructions at 5 (5,5), and the nutritional analysis at 2 (2, 2) on the five-point Likert scale. ChatGPT's nutritional analysis underestimated calories by 36% (95% CI: 44-88%), protein by 28% (25-167%), fat 48% (29-81%), phosphorus 54% (15-102%), potassium 49% (40-68%), and sodium 53% (14-139%). The nutritional analysis of online available recipes differed only by 0 to 35%. The translations were rated as reliable by native speakers (4 on the five-point Likert scale).CONCLUSIONWhile ChatGPT-4 shows promise in providing personalized nutritional guidance for diverse dialysis patients, improvements are necessary. This study highlights the importance of thorough qualitative and quantitative evaluation of artificial intelligence-generated content, especially regarding medical use cases.METHODSWe tasked ChatGPT-4 with generating a personalized daily meal plan, including nutritional information. Virtual "patients" were generated through Monte Carlo simulation; data from a randomly selected virtual patient were presented to ChatGPT. We provided to ChatGPT patient demographics, food preferences, laboratory data, clinical characteristics, and available budget, to generate a one-day sample menu with recipes and nutritional analyses. The resulting daily recipe recommendations, cooking instructions, and nutritional analyses were reviewed and rated on a five-point Likert scale by an experienced renal dietitian. In addition, the generated content was rated by a renal dietitian and compared with a U. S. Department of Agriculture-approved nutrient analysis software. ChatGPT also analyzed nutrition information of two recipes published online. We also requested a translation of the output into Spanish, Mandarin, Hungarian, German, and Dutch.

RESULTSThe intervention group showed an improved median percentage of hemoglobin measurements within target at 47% (interquartile range, 39–58), with a 10% point median difference between the two groups (95% confidence interval, 3 to 16; P = 0.008). The odds ratio of being within the hemoglobin target in the standard-of-care group compared with the group receiving the personalized ESA recommendations was 0.68 (95% confidence interval, 0.51 to 0.92). The variability of hemoglobin levels decreased in the intervention group, with the percentage of patients experiencing fluctuating hemoglobin levels being 45% versus 82% in the standard-of-care group. ESA usage was reduced by approximately 25% in the intervention group.KEY POINTSWe conducted a randomized controlled pilot trial in patients on hemodialysis using a physiology-based individualized anemia therapy assistance software. Patients in the group receiving erythropoiesis-stimulating agent dose recommendations from the novel software showed improvement in hemoglobin stability and erythropoiesis-stimulating agent utilization.CONCLUSIONSOur results demonstrated an improved hemoglobin target attainment and variability by using personalized ESA recommendations using the physiology-based anemia therapy assistance software.CLINICAL TRIAL REGISTRATION NUMBER:NCT04360902.BACKGROUNDAnemia is common among patients on hemodialysis. Maintaining stable hemoglobin levels within predefined target levels can be challenging, particularly in patients with frequent hemoglobin fluctuations both above and below the desired targets. We conducted a multicenter, randomized controlled trial comparing our anemia therapy assistance software against a standard population-based anemia treatment protocol. We hypothesized that personalized dosing of erythropoiesis-stimulating agents (ESAs) improves hemoglobin target attainment.METHODSNinety-six patients undergoing hemodialysis and receiving methoxy polyethylene glycol-epoetin beta were randomized 1:1 to the intervention group (personalized ESA dose recommendations computed by the software) or the standard-of-care group for 26 weeks. The therapy assistance software combined a physiology-based mathematical model and a model predictive controller designed to stabilize hemoglobin levels within a tight target range (10–11 g/dl). The primary outcome measure was the percentage of hemoglobin measurements within the target. Secondary outcome measures included measures of hemoglobin variability and ESA utilization.

RESULTSWe utilized data from 693 patients who contributed 42 656 hemodialysis sessions and 355 693 intradialytic SBP measurements. IDH occurred in 16.2% of hemodialysis treatments. Our model predicted IDH 15-75 min in advance with an AUROC of 0.89. Top IDH predictors were the most recent intradialytic SBP and IDH rate, as well as mean nadir SBP of the previous 10 dialysis sessions.CONCLUSIONSReal-time prediction of IDH during an ongoing hemodialysis session is feasible and has a clinically actionable predictive performance. If and to what degree this predictive information facilitates the timely deployment of preventive interventions and translates into lower IDH rates and improved patient outcomes warrants prospective studies.BACKGROUNDIn maintenance hemodialysis patients, intradialytic hypotension (IDH) is a frequent complication that has been associated with poor clinical outcomes. Prediction of IDH may facilitate timely interventions and eventually reduce IDH rates.METHODSWe developed a machine learning model to predict IDH in in-center hemodialysis patients 15-75 min in advance. IDH was defined as systolic blood pressure (SBP) <90 mmHg. Demographic, clinical, treatment-related and laboratory data were retrieved from electronic health records and merged with intradialytic machine data that were sent in real-time to the cloud. For model development, dialysis sessions were randomly split into training (80%) and testing (20%) sets. The area under the receiver operating characteristic curve (AUROC) was used as a measure of the model's predictive performance.

DISCUSSIONAlthough segmental eight-point bioimpedance techniques provided comparable TBW measurements not affected by standing over a period of 10-15 min, the ECW/TBW ratio appeared to be significantly lower in InBody compared with Seca and Hydra. Results from our study showed lack of agreement between different bioimpedance devices; direct comparison of ECW, ICW, and ECW/TBW between different devices should be avoided and clinicians should use the same device to track the fluid status in their HD population in a longitudinal direction.INTRODUCTIONSegmental eight-point bioimpedance has been increasingly used in practice. However, whether changes in bioimpedance analysis components before and after hemodialysis (HD) using this technique in a standing position is comparable to traditional whole-body wrist-to-ankle method is still unclear. We aimed to investigate the differences between two eight-point devices (InBody 770 and Seca mBCA 514) and one wrist-to-ankle (Hydra 4200) in HD patients and healthy subjects in a standing position.FINDINGSOverall, total body water (TBW) was not different between the three devices, but InBody showed lower extracellular water (ECW) and higher intracellular water (ICW) compared to the other two devices. When intradialytic weight loss was used as a surrogate for changes in ECW (∆ECW) and changes in TBW (∆TBW), ∆ECW was underestimated by Hydra (-0.79 ± 0.89 L, p < 0.01), InBody (-1.44 ± 0.65 L, p < 0.0001), and Seca (-0.32 ± 1.34, n.s.). ∆TBW was underestimated by Hydra (-1.14 ± 2.81 L, n.s.) and InBody (-0.52 ± 0.85 L, p < 0.05) but overestimated by Seca (+0.93 ± 3.55 L, n.s.).METHODSThirteen HD patients were studied pre- and post-HD, and 12 healthy subjects once. Four measurements were performed in the following order: InBody; Seca; Hydra; and InBody again. Electrical equivalent models by each bioimpedance method and the fluid volume estimates by each device were also compared.

RESULTSThe SEV group reported a 3.3-fold higher frequency of BSS stool types 1 and 2 (more likely constipated, p < 0.05), whereas the SFO group reported a 1.5-fold higher frequency of BSS stool types 5-7 (more likely loose stool and diarrhea, not significant). Participants in the SFO group showed a trend toward better adherence to phosphate binder therapy (SFO: 87.6% vs. SEV: 66.6%, not significant). UTOX, serum phosphorus, nutritional and liver function markers, and tryptophan were not different between the two groups.CONCLUSIONThere was no difference in the gut microbiome-derived UTOX levels between phosphate binders (SFO vs. SEV), despite SFO therapy resulting in fewer constipated participants. This pilot study may inform study design of future clinical trials and highlights the importance of including factors beyond bowel habits and their association with UTOX levels.INTRODUCTIONConstipation is prevalent in patients with kidney failure partly due to the use of medication, such as phosphate binders. We hypothesized that serum levels of gut microbiome-derived uremic toxins (UTOX) may be affected by the choice of phosphate binder putatively through its impact on colonic transit time. We investigated two commonly prescribed phosphate binders, sevelamer carbonate (SEV) and sucroferric oxyhydroxide (SFO), and their association with gut microbiome-derived UTOX levels in hemodialysis (HD) patients.METHODSWeekly blood samples were collected from 16 anuric HD participants during the 5-week observational period. All participants were on active phosphate binder monotherapy with either SFO or SEV for at least 4 weeks prior to enrollment. Eight UTOX (7 gut microbiome-derived) and tryptophan were quantified using liquid chromatography-mass spectrometry. Serum phosphorus, nutritional, and liver function markers were also measured. For each substance, weekly individual levels, the median concentration per participant, and differences between SFO and SEV groups were reported. Patient-reported bowel movements, by the Bristol Stool Scale (BSS), and pill usage were assessed weekly.