Clinical characteristics of HD patients just prior to starting HD session
The baseline clinical characteristics of the 49 enrolled HD patients (24 non-DM, 25 T2DM), determined just prior to a Monday or Tuesday morning HD session, three days after the previous session, are shown in Table 1. There were no significant differences for age, gender, HD duration, or inter-dialytic weight gain between the non-DM and T2DM groups, while BMI and albumin were significantly higher in the T2DM HD patients. Serum creatinine did not differ significantly between the groups. Casual plasma glucose and GA, parameters for glycemic control, were significantly higher in the T2DM group.
As shown in Figure 1, arterial b-HB was significantly higher in the T2DM HD [29.0 (range 17.0-59.5) mmol/L] as compared to the non-DM [17.0 (11.5-36.0) mmol/L] patients, whereas arterial AcAc was not significantly different between the groups [28.0 (22.0-49.3) vs. 26.0 (20.0-36.5) mmol/L], resulting in a significantly lower arterial AcAc/b-HB ratio in the T2DM HD [0.91 (0.73-1.24) mmol/L] than in the non-DM HD [1.35 (1.06-2.17) mmol/L] patients. An arterial AcAc/b-HB ratio <1.0, known to indicate significant risk for organ dysfunction (liver, heart) and increased mortality (12), was noted in 13 (56%) of the T2DM HD patients, which was significantly higher as compared to the non-DM group (n=5, 20.8%) (p<0.05, c2 test). Neither arterial blood pH nor HCO3- was significantly different between the groups.
Regression analysis of correlation of arterial serum AcAc, b-HB, and arterial AcAc/b-HB ratio with various clinical parameters
Arterial AcAc, b-HB, and AcAc/b-HB ratio were examined for their correlation with various clinical parameters in all of the present cohort, as well as separately in the non-DM and T2DM groups (Table 2). Arterial AcAc was correlated with serum Pi in all as well as in the non-DM HD patients, and arterial b-HB was correlated with the rate of interdialytic weight gain, and urea nitrogen and uric acid in serum in all patients, and with casual plasma glucose and LDL-C in the non-DM HD patients. Additionally, arterial AcAc/b-HB ratio was correlated with rate of interdialytic weight gain as well as urea nitrogen, creatinine, and uric acid in serum in all HD patients, and also with albumin, casual plasma glucose, and LDL-C in the non-DM HD patients. Notably, casual plasma glucose before the HD session was significantly correlated in a negative manner with b-HB and in a positive manner with AcAc/b-HB ratio in the non-DM, but not the T2DM group. The only significant correlations noted in the T2DM group was between arterial b-HB and rate of interdialytic weight gain, and between arterial AcAc/b-HB ratio and serum urea nitrogen and Pi.
Increased arterial ketone body levels during a single HD session
During a single session, all of the present HD patients exhibited significant increases in arterial levels of AcAc from 26.0 (range 20.0-36.5) mmol/L to 82.0 (39.8 -157.8) mmol/L (p<0.0001) (Fig. 1A) as well as b-HB from 20.0 (14.8-42.0) mmol/L to 104.0 (32.0-297.3) mmol/L (p<0.0001) (Fig. 1B). Those increases resulted in a significant reduction in arterial AcAc/b-HB ratio from 1.18 (0.76-1.51) (mmol/mmol) to 0.70 (0.44-1.13) (mmol/mmol) (p<0.0001) (Fig. 1C). When all HD patients were divided into T2DM and non-DM groups, the non-DM HD patients retained significant increases in arterial AcAc from 26.0 (20.0-36.5) mmol/L to 57.0 (37.5-162.5) mmol/L (p<0.0001) (Fig. 1D) and b-HB from 17.0 (11.5-36.0) to 80.0 (28.5-291.5) mmol/L (p<0.0001) (Fig. 1E). In T2DM HD patients, arterial AcAc and b-HB were significantly increased from 28.0 (22.0-49.3) mmol/L to 96.0 (54.0-158.0) mmol/L (p<0.0001) Fig. 1G) and from 29.0 (17.0-59.5) mmol/L to 105.0 (63.5-309.0) mmol/L (p<0.0001). Of interest, the reduction in arterial AcAc/b-HB ratio in the non-DM HD patients (Fig. 1F) during an HD session [from 1.35 (1.06-2.17) mmol/mmol to 0.70 (0.44-1.41) mmol/mmol] became statistically significant (p<0.0001), while that change in T2DM HD patients (Fig. 1I) [from 0.91 (0.73-1.24) mmol/mmol to 0.68 (0.44-1.06) mmol/mmol] did not (p=0.1078). As a result, though the arterial AcAc/b-HB ratio was significantly higher in the non-DM HD as compared to the T2DM HD patients before (p=0.0134), it did not differ significantly between those groups after the HD session. Furthermore, arterial b-HB was found to be significantly lower in the non-DM HD patients before but not after the session. Together, these results suggest that the rate of increase in AcAc/b-HB ratio during the inter-dialytic period was higher in the non-DM HD patients, possibly due to a greater rate of b-HB reduction.
Changes in arterial blood pH and HCO3- were not significantly correlated with those of arterial AcAc, b-HB, or AcAc/b-HB ratio in the full cohort, as well as after dividing into the T2DM and non-DM groups (data not shown). Notably, the baseline levels of both casual plasma glucose and glycoalbumin before the HD session were significantly correlated in a negative manner with change in AcAc/b-HB ratio during the HD session in the non-DM HD patients (Figure 2), but not in the T2DM HD patients (data not shown).
Multiple regression analysis of b-HB and arterial AcAc/b-HB ratio with serum log albumin, Pi, and UA
Multiple regression analysis was performed to examine whether arterial b-HB or AcAc/b-HB ratio had a significant association with serum levels of albumin and uric acid in all of the present HD patients. When log b-HB was included as an independent variable, in addition to age, gender, presence/absence of T2DM, and log HD duration (Model 1), it emerged as a significant and independent factor showing an association with uric acid, but not albumin. When log b-HB was replaced with log AcAc/b-HB ratio (Model 2), that ratio showed a significant positive relationship with both albumin and uric acid. When log b-HB and log AcAc/b-HB ratio were simultaneously included as independent variables in Model 3, log AcAc/b-HB ratio, but not log b-HB, retained its independent and significant association in a positive manner with albumin and uric acid, clearly indicating that arterial AcAc/b-HB ratio is superior to arterial b-HB as a clinically relevant marker for nutritional status in HD patients.