Can tissue dielectric constant measurements assess circulating blood volume changes in patients undergoing haemodialysis?

The tissue dielectric constant (TDC) method uses an open‐ended coaxial probe to achieve non‐invasive measurement of water content in skin. The aim of our study was to test the hypothesis that the changes in circulating blood volume would be associated with the changes in TDC values in patients undergoing haemodialysis.


Introduction
The tissue dielectric constant (TDC) method is capable of taking non-invasive measurement of water content in the skin and/or subcutaneous fat using an open-ended coaxial probe with a high-frequency electromagnetic wave of 300 MHz (Nuutinen et al., 2004). This method has been applied for evaluating the grade of local tissue oedema (primarily in the limbs) in a variety of clinical settings (Mayrovitz et al., 2008(Mayrovitz et al., , 2009aMayrovitz et al., 2013Mayrovitz et al., , 2014Birkballe et al., 2014;Lahtinen et al., 2015).
Oedema of the lower limbs in patients with diabetes mellitus has been associated with an increased TDC value (Mayrovitz et al., 2013). In another study, manual lymphatic drainage therapy reduced the lymphoedema of the upper limbs in post-mastectomy patients, and this change was reflected in a decreased TDC for the upper limbs (Mayrovitz et al., 2008).
Local oedema is in part associated with an increased blood volume due to heart and/or renal failure. Decreased circulating blood volume from the use of diuretics can reduce local oedema in such patients. From this viewpoint, we hypothesized that the changes in circulating blood volume of patients would be associated with the changes in the TDC value of local tissue. In this study, we considered that maintenance haemodialysis patients were appropriate subjects for testing our hypothesis because haemodialysis causes a clear decrease in a patient's circulating blood volume. To test our hypothesis, we thus determined the TDC measurements in patients who underwent scheduled haemodialysis, and evaluated the relationship between changes in TDC values and those in the circulating blood volume.

Study design and patients
Before the study was begun, we obtained the approval of our university's Medical Ethics Committee (Hirosaki University Graduate School of Medicine, Hirosaki, Japan; approval no. 2013-118) as well as written informed consent from the 83 patients. The trial was registered with the University Hospital Medical Information Network (registration no. UMIN000017418). This prospective descriptive study was carried out at Hirosaki University Hospital, Hirosaki, Japan. Consecutive patients who underwent haemodialysis between May 2015 and February 2016 were assessed for eligibility. The candidates for inclusion in the study were maintenance dialysis patients aged ≥18 years.

TDC measurement
Each patient's TDCs were measured using a desktop Mois-tureMeter-D (Delfin Technologies, Kuopio, Finland) by a single anaesthesiologist (K.T.). The desktop MoistureMeter-D consists of a measuring unit and four probes for different measurement depths from 0.5-5 mm. The M25 probe with 2.5 mm measurement depth was used in this study. This depth is considered effective for evaluating the grade of local oedema (Mayrovitz et al., 2009a,b).
Tissue dielectric constants of the face, hand and shin of each patient were taken before and after the patient underwent haemodialysis. Baseline TDCs were measured when the patient first assumed a supine position, before the start of haemodialysis. During the dialysis, the patients were allowed to change positions on the bed at will. The hand measurements were made on the dorsal skin between the thumb and forefinger of one hand without an arteriovenous shunt. The shin measurement was made on the same side of the middle tibia as the hand measurements. Face measurements were made on the same side as the hand measurement, just below the eye. All three standardized measurement sites were marked with a dot-shaped seal. The measurements were conducted by placing the probe above the dot-shaped seal in contact with the skin using gentle pressure. Three replications were made at each measurement site. The mean of two or three consecutive readings with a coefficient of variation below 15% at one site of interest was used.

Haemodialysis procedure
Body weight measurement was performed in all patients upon their arrival at the haemodialysis room. The scheduled amount of water removal for each patient was thus determined based on the increase in body weight over the dry weight. The duration of the procedure was determined by nephrologists or urologists in consideration of the patient's dialysis tolerance. Haemodialysis was performed using a standardized haemodialysis machine in our hospital (TR7700M; Toray Medical, Tokyo, Japan). We were able to determine the amount of water removed from patients at any time point by checking the water-removal 'metre' on the screen of the dialysis machine. After patients underwent haemodialysis, their body weights were measured again.

The primary outcome measure
The primary outcome measure was the correlation between the amount of water removal and DTDC at each body site measured. DTDC was defined as the difference in the TDC values between before and after haemodialysis (DTDC = TDC after haemodialysis-TDC before haemodialysis).

Secondary outcome measure and other collected data
The secondary outcome measure was the mean difference in the TDC of each measured site between before and after haemodialysis. The following demographical and haemodialysis data were collected: age, body mass index, reasons for haemodialysis, amount of water removal, duration of haemodialysis and patients' weight loss due to water removal. We also evaluated the correlation between the amount of water removal and the patients' weight loss.

Statistical analysis
For tests of association using Pearson correlations, a moderate correlation between variables was considered meaningful. To detect a moderate correlation (r = 0.3), a sample of 82 analysable subjects will provide 80% power to discover that the correlation is statistically different from no correlation at the 0.05 significance level. For continuous variables with a normal distribution, the mean [AE standard deviation (SD) or standard error (SE)] is reported. For variables not normally distributed, the median and interquartile ranges are reported. P-values <0.05 were considered significant. Student's t-test was used for continuous variables with normal distributions. The Mann-Whitney rank-sum test was used for continuous variables without a normal distribution. Correlations between the amount of water removal and DTDC at each body site (the primary outcome) or the patients' weight loss were analysed using Pearson's product-moment correlation coefficient. Sample size calculations were performed using G*Power 3 software (Heinrich-Heine-University Institute of Experimental Psychology, D€ usseldorf, Germany; Faul et al., 2007Faul et al., , 2009). All statistical analyses were conducted with IBM SPSSâ statistics ver. 22.0 software (IBM, Tokyo, Japan).

Results
A total of 83 patients treated in the period from September 2015 to March 2016 were included in the analyses. The demographics of the patients and haemodialysis data are summarized in Table 1. Patients' weight loss after haemodialysis was compatible with the amount of water removal determined using a dialysis machine. A strong inverse correlation was also observed between the amount of water removed and the patients' weight loss (r = À0.91, P = 0.000).

Primary outcome measure
As shown in Fig. 1, there was an inverse correlation between the amount of water removed and the DTDC at the face or shin in patients with haemodialysis (at the face, r = À0.25, P = 0.028; at the shin, r = À0.26, P = 0.018). In contrast, DTDC at the hand showed no correlation with the amount of water removed.

Secondary outcome measure
The secondary outcome measure was the mean difference in the TDC of each measured site between before and after haemodialysis. The TDC value measured before haemodialysis was defined as the baseline. The TDC values measured at all body sites were significantly decreased after patients underwent haemodialysis compared with the baseline (Table 2).

Discussion
Our starting hypothesis was that the changes in the circulating blood volume of patients would be associated with the changes in the TDC value of local tissue. In the present study, we found that the TDC values measured at local tissues were significantly decreased after patients underwent haemodialysis. However, the percentage difference in the TDC values between the baseline and after haemodialysis was small (À3.2% to À6.0%). In a similar fashion, a correlation between DTDC and the amount of water removal was also weak (r = À0.25 or À0.26). The finding of a correlation of 0.25 suggests that only 6% of the variation was attributable to  water changes. Our results thus indicate that TDC measurement is unlikely to be a reliable index of changes in the circulating blood volume of patients. A possible explanation for this inconsistency with our hypothesis would be as follows. In theory, the decreased circulating blood in the vessels due to the haemodialysis draws excessive water from the interstitial tissues into the capillary vessels to be equilibrated. Such water shift normalizes the circulating blood volume and reduces the local oedema. This would require more time than the length of the present study period. From this point of view, if we had conducted the TDC measurement later, a higher correlation and a bigger difference in the TDC value would likely have been obtained in our study. Based on this theory, the results of our study can be interpreted as that the TDC measurement detects the decreased local water content in an early phase of 'fluid equilibrium' after haemodialysis. Our results thus indicate that TDC measurement can only be used as an index of changes in local water volume and detect even slight changes in the local oedema, but would be less likely to accurately represent the circulating blood volume in real time.
One prospective descriptive study showed that there was a moderate positive correlation between the increase in the TDC value and weight gained during the postoperative period in patients undergoing cardiac surgery (r = 0.60, P<0.01, n = 29) (Petaja et al., 2003). In accordance with the theory described above, Petaja's result can be taken to mean that positive fluid balance after surgery increased local tissue water (i.e. oedema) as well as the body weight of patients. Another clinical study showed that there was a highly significant correlation between the decreasing TDC value and removed water in patients who underwent haemodialysis (r = À0.99, P<0.01, n = 7) (Nuutinen et al., 2004). Probably they would have enough time of tissue water to be equilibrated.
A possible limitation of our study is that it was a descriptive study conducted in a limited number of patients. In the future, the efficacy of the TDC method will need to be tested in a variety of patients.
We concluded that TDC measurement is unlikely to be a reliable index of real-time changes in the circulating blood volume, but can be sensitive to assess changes in local water content. Our results confirmed the current understanding that TDC measurement can be used to assess changes in the water content of local tissues. However, further studies will be needed to confirm that TDC measurement is unreliable for the real-time measurement of circulating blood volume.