Accuracy of Bioimpedance Spectroscopy in the Detection of Hydration Changes in Patients on Hemodialysis

OBJECTIVES

The body composition monitor (BCM) is a bioimpedance spectroscopy device, specifically developed for patients on hemodialysis (HD) to improve ultrafiltration (UF) programming, based on an objective assessment of the degree of overhydration (OH) at the start of HD. However, its acceptance in clinical practice remains limited because of concerns about the accuracy at the individual level. The aim of this study is to examine the performance of the BCM and to identify means of improvement.

METHODS

Precision of the OH estimate was assessed by 6 consecutive measurements in 24 patients on HD. Accuracy was examined in 45 patients, by comparing the change in OH (ΔOH) during HD with UF volume. Accuracy was considered acceptable if the volume error in individual patients was ≤0.5 L.

RESULTS

The OH estimate had an analytical precision of 1.0 ± 0.4%. The correlation between UF volume and ΔOH was moderate (Slope = 0.66, R² = 0.44, P < .001) and indicated underestimation of UF volume, in particular for high UF volumes. Accuracy at individual level was highly variable. A volume error >0.5 L occurred in 44% of patients. Accuracy improved over the course of HD, with a decrease in total error range from 2.3 L in the first hour to 1.1 L in the final hour of HD.

CONCLUSIONS

The accuracy of BCM volume change estimates is highly variable and below requirements of daily practice. Improvement may be achieved by a switch to an end-of-HD measurement.

Impact of measured versus estimated glomerular filtration rate-based screening on living kidney donor characteristics: A study of multiple cohorts

Background

Most transplant centers in the Netherlands use estimated glomerular filtration rate (eGFR) for evaluation of potential living kidney donors. Whereas eGFR often underestimates GFR, especially in healthy donors, measured GFR (mGFR) allows more precise kidney function assessment, and therefore holds potential to increase the living donor pool. We hypothesized that mGFR-based donor screening leads to acceptance of donors with lower pre-donation eGFR than eGFR-based screening.

Methods

In this longitudinal cohort study, we compared eGFR (CKD-EPI) before donation in one center using mGFR-based screening (mGFR-cohort, n = 250) with two centers using eGFR-based screening (eGFR-cohort1, n = 466 and eGFR-cohort2, n = 160). We also compared differences in eGFR at five years after donation.

Results

Donor age was similar among the cohorts (mean±standard deviation (SD) mGFR-cohort 53±10 years, eGFR-cohort1 52±13 years, P = 0.16 vs. mGFR-cohort, and eGFR-cohort2 53±9 years, P = 0.61 vs. mGFR-cohort). Estimated GFR underestimated mGFR by 10±12 mL/min/1.73m² (mean±SD), with more underestimation in younger donors. In the overall cohorts, mean±SD pre-donation eGFR was lower in the mGFR-cohort (91±13 mL/min/1.73m²) than in eGFR-cohort1 (93±15 mL/min/1.73m², P<0.05) and eGFR-cohort2 (94±12 mL/min/1.73m², P<0.05). However, these differences disappeared when focusing on more recent years, which can be explained by acceptance of more older donors with lower pre-donation eGFR over time in both eGFR-cohorts. Five years post-donation, mean±SD eGFR was similar among the centers (mGFR-cohort 62±12 mL/min/1.73m², eGFR-cohort1 61±14 mL/min/1.73m², eGFR-cohort2 62±11 mL/min/1.73m², P = 0.76 and 0.95 vs. mGFR-cohort respectively). In the mGFR-cohort, 38 (22%) donors were excluded from donation due to insufficient mGFR with mean±SD mGFR of 71±9 mL/min/1.73m².

Conclusions

Despite the known underestimation of mGFR by eGFR, we did not show that the routine use of mGFR in donor screening leads to inclusion of donors with a lower pre-donation eGFR. Therefore eGFR-based screening will be sufficient for the majority of the donors. Future studies should investigate whether there is a group (e.g. young donors with insufficient eGFR) that might benefit from confirmatory mGFR testing.

Impact of diffusion, ultrafiltration, and posture on total body electrical resistance in patients on hemodialysis

Monitoring of hydration in patients on hemodialysis (HD) by currently available bioelectrical impedance analysis (BIA) methods is hampered by limited accuracy. This may be caused by changes in total body electrical resistance (TBER) that are induced by processes other than ultrafiltration (UF). To identify these sources of error, we examined the impact of UF, diffusion, and postural change (PC), separately. Extracellular TBER (TBER_e) was measured by bioimpedance spectroscopy every 30 min in 23 patients on HD, for 2 h during diffusion-only (DO), followed by 2-h UF-only (UFO). The impact of PC from upright to semi-recumbent position was assessed by a 2-h TBER_e measurement on the day after HD. TBER_e increased by 23.5 ± 12.4 Ω (P < 0.001) during DO and by 40.0 ± 16.2 Ω (P < 0.001) during UFO. PC, evaluated on a separate day, was associated with an increase in TBER_e of 27.6 ± 26.0 Ω (P < 0.001). TBER_e changes during DO were mainly attributed to PC and to a lesser extent to electrolyte exchange. Extrapolation of the data to a conventional 4-h HD session indicates that about 32% of the total increase in TBER_e is not related to UF. In conclusion, a significant part of the increase in TBER during HD is not related to UF but can be attributed to other processes such as the effects of PC and diffusion-related electrolyte exchange. These factors have to be taken into account when TBER-guided UF is considered.NEW & NOTEWORTHY Current BIA methods have limited accuracy in patients on HD. This may be related to the incorrect assumption that all changes in total body electrical resistance (TBER) are caused by changes in body water volumes. The present study indicates that two-thirds of the change in TBER during a conventional 4-h HD session can be attributed to fluid extraction, and that the remaining part is caused by other processes such as postural change and electrolyte exchange. This may cause volume prediction errors when not recognized.

Chronic pain following laparoscopic living-donor nephrectomy: Prevalence and impact on quality of life

Chronic postsurgical pain (CPSP) following laparoscopic donor nephrectomy (LDN) is a disregarded topic. In this cross-sectional study, all consecutive patients who underwent an LDN at the Radboud University Medical Center (Radboudumc; 2003-2016) were approached for participation. Five hundred twelve living kidney donors were included and asked to complete two questionnaires, including the McGill Pain Questionnaire and the RAND Short Form-36 Health Status Inventory (RAND SF-36) regarding their health-related quality of life (HRQoL). The mean prevalence of CPSP following LDN was 5.7%, with a mean follow-up time of 6 years. Possible predictors of CPSP following LDN are severe early postoperative pain, previous abdominal surgery, and preexisting backache. The RAND SF-36 revealed an impaired HRQoL in patients with CPSP when compared to patients without CPSP. In conclusion, this study revealed that the prevalence of CPSP following LDN is substantial. Given the possible association between the presence of CPSP and impaired HRQoL scores, living kidney donors should be well informed in the preoperative phase about the risk of CPSP.

Preoperative CT-Angiography Predicts Ex Vivo Vein Length for Right Kidneys After Laparoscopic Donor Nephrectomy

BACKGROUND Implantation of a kidney with a short renal vein is technically more challenging and therefore prone for technique-related complications. It remains unclear whether pre-operative computed tomography angiography (CTA), to assess vascular anatomy of the donor kidney, can be used to predict renal vein length. MATERIAL AND METHODS Right and left renal vein lengths of 100 consecutive kidney donors were measured in an oblique-coronal plane multiplanar reconstruction image of 100 consecutive kidney donors in whom ex vivo vein length was measured after recovery. In a second retrospective cohort of 100 consecutive kidney donors donating a right kidney, preoperative CTA vein length measurements were correlated to anastomosis time and early graft outcome. RESULTS Left and right renal vein lengths, measured on CTA, were 43.2 mm and 30.0 mm, respectively. No correlation was found between CTA and ex vivo measurements for the left renal vein (p=.610), whereas a significant correlation was found for the right renal vein (p=.021). In the retrospective cohort, right renal vein length was significantly correlated with the anastomosis time but not with early graft outcome. CONCLUSIONS The length of the right, but not the left, renal vein can be predicted by preoperative CTA, but this does not hold true for the left renal vein.