DIABETICMedicine DOI: 10.1111/dme.12717
Research: Pathophysiology Glomerular haemodynamic profile of patients with Type 1 diabetes compared with healthy control subjects M. Skrtic1, Y. Lytvyn1, G. K. Yang1, P. Yip2, V. Lai1, M. Silverman1 and D. Z. I. Cherney1 1 Department of Medicine, Division of Nephrology, Toronto General Hospital, University of Toronto and 2University Health Network, Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada
Accepted 4 February 2015
Abstract Aims To evaluate the glomerular haemodynamic profile of patients with Type 1 diabetes with either renal hyperfiltration (GFR ≥ 135 ml/min/1.73 m2) or renal normofiltration (GFR 90–134 ml/min/1.73 m2) during euglycaemic and hyperglycaemic conditions, and to compare this profile with that of a similar group of healthy control subjects. Methods Gomez’s equations were used to derive afferent and efferent arteriolar resistances, glomerular hydrostatic pressure and filtration pressure.
At baseline, during clamped euglycaemia, patients with Type 1 diabetes and hyperfiltration had lower mean SD afferent arteriolar resistance than both those with Type 1 diabetes and normofiltration (914 494 vs. 2065 597 dyne/s/cm5; P < 0.001) and healthy control subjects (1676 707 dyne/s/cm5; p < 0.001). By contrast, efferent arteriolar resistance was similar in the three groups. Patients with Type 1 diabetes and hyperfiltration also had higher mean SD glomerular hydrostatic pressure than both healthy control subjects and patients with Type 1 diabetes and normofiltration (66 6 vs. 60 3 vs. 55 3 mmHg; P < 0.05). Similar findings for afferent arteriolar resistance, efferent arteriolar resistance, glomerular hydrostatic pressure and filtration pressure were observed during clamped hyperglycaemia. Results
Hyperfiltration in Type 1 diabetes is primarily driven by alterations in afferent arteriolar resistance rather than efferent arteriolar resistance. Renal protective therapies should focus on afferent renal arteriolar mechanisms through the use of pharmacological agents that target tubuloglomerular feedback, including sodium-glucose cotransporter 2 inhibitors and incretins.
Conclusion
Diabet. Med. 32, 972–979 (2015)
Introduction Renal hyperfiltration has been associated with increased intraglomerular pressure, albuminuria and the progression of diabetic nephropathy [1]. Hyperfiltration in humans has been attributed to two major influences: 1) neurohormonal factors, such as the renin angiotensin aldosterone system (RAAS), which causes efferent vasoconstriction, as well as afferent vasodilators, such as nitric oxide, and 2) tubular factors, which consist of altered tubuloglomerular feedback as a result of increased sodium-glucose cotransporter 2 activity, leading to decreased sodium delivery to the macula densa [2,3]. This results in a perceived reduced effective circulating volume, leading to afferent renal arteriolar vasodilatation [3].
Correspondence to: David Cherney. E-mail: [email protected]
972
Evidence of early renal haemodynamic changes resulting from changes in efferent and afferent tone from neurohormonal mediators and altered tubuloglomerular feedback have been provided by animal micropuncture studies [3]. Importantly, the vasodilatory effect of RAAS inhibitors at the efferent arteriole in preclinical models has formed the foundation for therapies that protect against the development of diabetic nephropathy in humans [4]. For tubular factors, until recently, it was not known if pharmacological modulation of proximal renal tubular sodium handling affected hyperfiltration in humans. Accordingly, to determine if increasing distal tubular sodium delivery reduces hyperfiltration in humans, we treated patients with Type 1 diabetes mellitus with the sodium-glucose cotransporter 2 inhibitor empagliflozin for 8 weeks and showed that there was a significant attenuation of renal hyperfiltration that was similar in magnitude to that achieved with an angiotensin-converting enzyme
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
What’s new? • Renal hyperfiltration in Type 1 diabetes is primarily attributable to vasodilatation of the renal afferent arteriole. • Induction of clamped hyperglycaemia was associated with an increase in renal filtration pressure. • There are potential future implications of using ‘afferent constrictors’ such as sodium-glucose cotransporter 2 inhibitors in combination with traditional renal protective strategies, such as renin angiotensin aldosterone system inhibitors, as these older agents act primarily through efferent vasodilatation. inhibitor in our previous work [5,6]. Because of an inability to assess segmental (afferent vs. efferent) resistance directly in humans, however, interpretation of pharmacological responses to existing and novel therapies has been inferred based on changes in GFR, effective renal plasma flow, renal blood flow, filtration fraction and renal vascular resistance. A more detailed description of in vivo glomerular function in humans, including pre-glomerular (afferent) and post-glomerular (efferent) arteriolar function is required to understand the pathophysiology of early diabetic nephropathy, as well as the beneficial and deleterious responses to emerging renal protective therapies. Although direct measurements of human glomerular haemodynamic variables such as afferent and efferent arteriolar resistance values and glomerular hydrostatic pressure are not feasible in vivo in humans, Gomez. [7] derived equations for these variables. Subsequent work has used these equations to evaluate patients with a variety of conditions, including hypertension and endocrine disorders [8–10]. To our knowledge, these techniques have not been used to characterize the renal haemodynamic profile in patients with diabetes mellitus, or to evaluate kidney function in patients with Type 1 diabetes and either hyperfiltration or normal filtration. Accordingly, the aim of the present exploratory analysis was to determine the glomerular haemodynamic profile of patients with Type 1 diabetes and either hyperfiltration or normofiltration and to compare them with a similar group of normal healthy control subjects using baseline data from several previous physiology studies [6,11,12]. Our aim was to determine if there were fundamental differences in glomerular haemodynamic variables (afferent and efferent arteriolar resistance values and glomerular hydrostatic pressure) between the three groups, based on filtration status, that could potentially place some individuals at risk of future renal disease, and potentially explain differences in response to renal protective therapies such as RAAS inhibitors.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
DIABETICMedicine
Patients and methods A total of 25 healthy control subjects, 35 patients with Type 1 diabetes and normofiltration and 38 patients with Type 1 diabetes and hyperfiltration were included in this exploratory analysis. Subjects were included from four previously published studies that had the same inclusion criteria: Type 1 diabetes duration > 1 year, age > 18 years, blood pressure < 140/90 mmHg and normoalbuminuria. Patients were not taking medications that alter blood pressure or cardiovascular variables. Patients were originally identified using office and hospital advertisements, through social networking and from review of clinic charts. The study was approved by the Research Ethics Board at the University Health Network. All subjects gave informed consent before the start of the study procedures. Subjects adhered to a sodium replete (> 140 mmol/day) and moderate protein (< 1.5 g/kg/day) diet during the 7 days preceding the experiment as described [6,11,12]. All studies were performed during 2 consecutive days at the Renal Physiology Laboratory at Toronto General Hospital. A modified glucose clamp technique was used to maintain euglycaemic (4–6 mmol/l) conditions on the first day, followed by hyperglycaemic (9–11 mmol/l) conditions on the second day as previously described. Studies were performed under these conditions because hyperfiltration is based on GFR during euglycaemia, and because of the effect of hyperglycaemia on GFR and blood pressure [13]. Constant inulin and para-aminohippurate infusions were used to determine GFR and effective renal plasma flow as previously described. After a 10-min equilibration period, blood samples were collected for inulin, para-aminohippurate, haematocrit and total protein, with additional blood evaluation carried out 30 min later. Two independent clearance periods were used for calculations, and the results averaged. All haemodynamic measurements were adjusted for body surface area. Intrarenal haemodynamics were calculated using formulae established by Gomez [7] to indirectly determine intraglomerular haemodynamics in humans based on data from animal studies. Gomez’s equations necessitate the following: i) that intrarenal vascular resistances be divided into afferent, postglomerular and efferent resistance; ii) that hydrostatic pressures within the renal tubules, venules, Bowman’s space and interstitium are in equilibrium and given a value of 10 mmHg; iii) that filtration disequilibrium in the glomerulus is assumed; and iv) that the gross filtration coefficient is assumed to be 0.0867 ml/s per mmHg, given a normal kidney physiology (GFR 130 ml/min, glomerular hydrostatic pressure 60 mmHg), and given Winton’s indirect estimates in a dog model that glomerular pressure is roughly two-thirds of mean arterial pressure (MAP) [7], and normal glomerular oncotic pressure is 25 mmHg. Clinical variables, including mean arterial pressure (mmHg), filtration fraction (FF), effective renal plasma flow (RPF) (ml/s), GFR (ml/s), renal
973
c et al. Glomerular haemodynamics in diabetes M. Skrti
DIABETICMedicine
blood flow (RBF) (ml/s), haematocrit (Hct) (%) and total protein (TP) (g/dl), were used to calculate efferent and afferent renal arteriolar resistance values (dyne/s/cm5), glomerular hydrostatic pressure (mmHg), glomerular filtration pressure (mmHg) and glomerular oncotic pressure (mmHg): FF ¼ GFR/RPF RBF ¼ RPF=ð1 HctÞ The filtration pressure across glomerular capillaries (DPF) is given by: DPF ¼ GFR/KFG where KFG is the gross filtration coefficient. The glomerular oncotic pressure (pG) is calculated from the plasma protein mean concentration (CM) within the capillaries: CM ¼ TP/FFlnð1=1 FFÞ pG ¼ 5ðCM 2Þ By substituting DPF, pG, and given that hydrostatic pressure in Bowman’s space (PBow) is assumed to be 10 mmHg, PGLO is calculated: PGLO ¼ DPF þ PBow þ pG PGLO ¼ðGFR/KFG Þ þ 10 mmHg þ ½5ðTP/FFlnð1=1 FFÞ 2Þ To calculate efferent (RE) and afferent renal arteriolar resistance (RA) values using principles of Ohm’s law, where 1328 is the conversion factor to dyne/s/cm5 [7]: RA ¼ ½ðMAP PGLO Þ=RBF1328 RE ¼ ½GFR/(KFG ðRBF GFRÞ1328
Statistical analysis
All data are expressed as mean SD values. Statistical analyses were performed using one-way ANOVA and post hoc Tukey’s tests were used for between-group differences as indicated. Differences were considered statistically significant at a P value of < 0.05. All statistical analyses were carried out using GRAPHPAD PRISM software (version 5.0).
hyperglycaemic conditions. As expected, the patients with Type 1 diabetes and normofiltration and those with hyperfiltration had significantly lower levels of circulating RAAS mediators compared with healthy control subjects. Also as expected, patients with Type 1 diabetes and hyperfiltration had higher GFR, effective renal plasma flow and renal blood flow and lower renal vascular resistance compared with both the Type 1 diabetes and normofiltration group and the healthy control group (P < 0.001 for all comparisons; Table 1).
Glomerular haemodynamic variables: euglycaemia
At baseline, during clamped euglycaemia, Gomez’s calculations of glomerular haemodynamic variables showed that afferent renal arteriolar resistance was significantly lower in the Type 1 diabetes and hyperfiltration group compared with both the Type 1 diabetes and normofiltration group and the healthy control group (P < 0.001; Figure 1). Glomerular hydrostatic pressure and glomerular filtration pressure were also higher in the Type 1 diabetes and hyperfiltration group than in both the Type 1 diabetes and normofiltration group and the healthy control group (P < 0.001; Figure 1). Glomerular hydrostatic pressure was lower in the Type 1 diabetes and normofiltration group than in the healthy control group (55 3 vs. 60 3 mmHg; P < 0.001), while afferent renal arteriolar resistance, efferent renal arteriolar resistance and glomerular filtration pressure values were similar between the two groups. There were no significant differences for efferent renal arteriolar resistance in any of the three groups (Figure 1).
Glomerular haemodynamic variables: hyperglycaemia
Similarly to the measurements obtained during euglycaemia, afferent renal arteriolar resistance was significantly lower during clamped hyperglycaemia in patients with Type 1 diabetes and hyperfiltration compared with patients with Type 1 diabetes and normofiltration, while efferent renal arteriolar resistance was the same in the two groups. In addition, glomerular hydrostatic pressure (67 7 vs. 57 6 mmHg; P < 0.001) and glomerular filtration pressure (34 7 vs. 25 5 mmHg; P < 0.001) were significantly higher in the Type 1 diabetes and hyperfiltration vs the Type 1 diabetes and normofiltration group during hyperglycaemia (Figure 1).
Results Renal haemodynamic responses to clamped hyperglycaemia Baseline characteristics
Clinical and demographic characteristics were similar among the healthy control group, the Type 1 diabetes and normofiltration group and the Type 1 diabetes and hyperfiltration group (Table 1). All subjects had baseline renal haemodynamic studies performed during clamped euglycaemic and
974
In response to clamped hyperglycaemia, GFR increased significantly in the Type 1 diabetes and normofiltration group (117 11 to 132 11 ml/min/1.73 m2; P < 0.005), but not in the Type 1 diabetes and hyperfiltration group (177 35 to 168 25 ml/min/1.73 m2; P > 0.05). There were no significant changes in response to hyperglycaemia for effective renal
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
DIABETICMedicine
Table 1 Clinical baseline characteristics of normal healthy control subjects and patients with Type 1 diabetes and either renal normofiltration or hyperfiltration
Healthy control group, n = 25 Baseline characteristic Males, n (%) Mean SD age, years Diabetes duration, years Weight, kg, mean SD Height, m, mean SD BMI, kg/m2, mean SD HbA1c, mmol/mol, mean SD HbA1c, %, mean SD Sodium Excretion (mmol/24 hr), mean SD Urea Excretion (mmol/24 hr), mean SD Systemic haemodynamic function, mean SD Heart rate, beats per min Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Renal haemodynamic function, mean SD Effective renal plasma flow, ml/min/1.73 m2 GFR, ml/min/1.73 m2 Filtration fraction Renal blood flow, ml/min/1.73 m2 Renal vascular resistance, mmHg/l/min Circulating neurohormonal factors, mean SD Aldosterone, pmol/L Plasma renin activity, ng/ml/h Angiotensinogen, ng/ml Intraglomerular haemodynamic variables, mean Filtration pressure, mmHg Glomerular hydrostatic pressure, mmHg Oncotic pressure, mmHg Afferent arteriolar resistance, dyne/s/cm5 Efferent arteriolar resistance, dyne/s/cm5
–
–
11 (44) 26.8 5.6
Type 1 diabetes and normofiltration group, n = 35
Type 1 diabetes and hyperfiltration group, n = 38
183 104 298 128
21 (60) 24.2 5.3 18.9 5.4 73.1 12.7 1.73 0.09 24.2 3.5 68 10 8.4 1.6 174 103 324 105
17 (45) 23.3 3.9 16.0 6.7 70.5 11.5 1.70 0.08 24.2 2.9 69 13 8.5 1.4 160 85.8 307 135
61.3 9.2 110 9.8 65.7 6.9
68.1 11.7* 113 8.5 64.9 6.6
74.0 9.9† 116 11.0‡ 65.5 6.9
69.2 11.4 1.74 0.08 22.8 3.1
628 117 0.19 1039 0.081
162 11 0.04 287 0.019
340 314 1.60 0.96 1338 791
642 116 0.18 1031 0.080
93 10.4 0.03 171 0.014
46 35* 0.57 0.56* 1298 930
970 177 0.18 1552 0.057
39 23* 0.55 0.4* 1017 558
SD
22.4 59.6 27.3 1676 2043
2.2 2.7 2.6 706 463
22.3 54.5 22.3 2065 1993
2.0 3.4* 2.6* 597* 367
268*§ 35*§ 0.06 466*§ 0.017*§
32.3 65.9 23.5 914 2047
4.9*§ 6.1‡§ 2.7* 494*§ 747
Values are mean SD, unless otherwise indicated. *P < 0.01 vs healthy control group. *P < 0.001 vs healthy control group. † P < 0.05 vs Type 1 diabetes and normofiltration group. ‡ P < 0.05 vs healthy control group. § P < 0.001 vs Type 1 diabetes and normofiltration group.
plasma flow (Type 1 diabetes and normofiltration group: 642 93 to 690 129 ml/min/1.73 m2; Type 1 diabetes and hyperfiltration group: 970 268 to 955 257 ml/min/ 1.73 m2) and renal blood flow (Type 1 diabetes and normofiltration group: 1039 171 to 1099 208 ml/min/ 1.73 m2; Type 1 diabetes and hyperfiltration: 1503 435 to 1552 466 ml/min/1.73 m2). For glomerular haemodynamic function, afferent renal arteriolar resistance tended to decrease (Type 1 diabetes and normofiltration group: 2065 597 to 1834 684 dyne/s/ cm5; Type 1 diabetes and hyperfiltration group: 914 494 to 856 652 dyne/s/cm5), while efferent renal arteriolar resistance tended to increase (Type 1 diabetes and normofiltration group: 1993 366 to 2151 475 dyne/s/cm5; Type 1 diabetes and hyperfiltration group: 2047 747 to 2255 885 dyne/s/cm5) in response to clamped hyperglycaemia, but these changes were not significant. Changes in glomerular filtration pressure (Type 1 diabetes and normofiltration group: 22 2 to 25 5 mmHg; Type 1 diabetes and hyperfiltration group: 32 5 to 34 7 mmHg) and ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
glomerular hydrostatic pressure (Type 1 diabetes and normofiltration group: 55 3 to 57 6 mmHg; Type 1 diabetes and hyperfiltration group: 66 6 to 67 7 mmHg) were also not significant. When patients with Type 1 diabetes and either normofiltration or hyperfiltration were examined as a single group, hyperglycaemia was found to have induced a significant increase in GFR (144 33 to 156 31 ml/min/1.73 m2; P < 0.05) and thereby also in filtration pressure (28 6 to 30 7 mmHg; P < 0.05). There were no other significant changes during hyperglycaemia in the combined cohort of patients with Type 1 diabetes (Table 2).
Discussion Direct measurements of segmental renal arteriolar resistance are not technically feasible in humans. As a consequence, the physiological basis for early diabetic nephropathy remains incompletely understood. In particular, the relative contribution of afferent vasodilatation vs. efferent vasoconstriction
975
DIABETICMedicine
c et al. Glomerular haemodynamics in diabetes M. Skrti
FIGURE 1 Glomerular haemodynamics estimated using Gomez’s equations (gross filtration coefficient = 0.0867) in normal healthy controls and patients with Type 1 diabetes and either renal normofiltration or hyperfiltration at baseline during clamped euglycaemia (open circles) and hyperglycaemia (filled circles) conditions. *P < 0.001 vs. healthy control subjects, †P < 0.001 vs Type 1 diabetes and normofiltration using post hoc Tukey’s test after ANOVA. Horizontal lines represent means (healthy controls, n = 25; patients with normofiltration, n = 35; patients with hyperfiltration, n = 38) of haemodynamic characteristics.
to the pathogenesis of renal hyperfiltration remains unknown. To further clarify the role of pre-glomerular vs. post-glomerular factors in the pathogenesis of haemodynamic abnormalities typical of early diabetes, we applied the formulae derived by Gomez to calculate glomerular haemodynamic variables in both healthy control subjects and in patients with Type 1 diabetes. In the group with Type 1 diabetes, it was critical to analyse patients on the basis of hyperfiltration status. Although the patients with Type 1 diabetes and hyperfiltration and those with Type 1 diabetes and normofiltration did not differ in age, diabetes duration, gender distribution or HbA1c concentration, those with Type 1 diabetes and hyperfiltration had physiologically distinct responses to sodium-glucose cotransporter 2 inhibitors, RAAS inhibitors and other pharmacological agents, highlighting their distinct physiological and renal haemodynamic profile [5,6,11,12,14]. Our first major observation was that hyperfiltration in Type 1 diabetes was primarily associated with haemodynamic changes at the afferent rather than the efferent arteriole. Previous micropuncture studies in animal models of diabetes have suggested a prominent role for afferent vasodilation, leading to hyperfiltration, through increased nitric oxide bioactivity, vasodilatory prostanoid production and changes in tubuloglomerular feedback related to increased proximal tubular sodium-glucose cotransport [3,15]. In patients with Type 1 diabetes and hyperfiltration, we have shown similar relationships between nitric oxide, prostanoids and sodiumglucose cotransport (as a tool for tubuloglomerular feedback) and hyperfiltration, suggesting a prominent role for afferent arteriolar vasodilatation in the pathogenesis of this condition
976
in humans [6,11,14]. To our knowledge, the present study is the first to show a significantly lower afferent renal arteriolar resistance in patients with Type 1 diabetes and hyperfiltration compared with both patients with Type 1 diabetes and normofiltration and healthy control subjects. In conjunction with higher glomerular filtration pressure and glomerular hydrostatic pressure in the group with Type 1 diabetes and hyperfiltration compared with the other two groups, and no difference in efferent renal arteriolar resistance, our results suggest that changes in glomerular haemodynamic function in early Type 1 diabetes are in large part related to changes in afferent rather than efferent tone. Our second finding was that, even though afferent renal arteriolar resistance, efferent renal arteriolar resistance and glomerular filtration pressure tended to be similar in the group with Type 1 diabetes and normofiltration compared with the healthy control group, afferent renal arteriolar resistance was 18.8% higher (P = non-significant) than in healthy control subjects. Furthermore, glomerular hydrostatic pressure was lowest in the Type 1 diabetes and normofiltration group, and this difference was significant compared with the other two groups. In our previous work reporting GFR, effective renal plasma flow, renal blood flow and renal vascular resistance, healthy control subjects and patients with Type 1 diabetes and normofiltration appeared to be identical, and had similar responses to afferent arteriolar vasoactive agents, suggesting that Type 1 diabetes and normofiltration exhibit normal renal physiology [11]; however, using the Gomez equations to perform a more indepth analysis of renal haemodynamic function, we were able to detect more subtle, early, ‘pre-clinical’ changes in
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
DIABETICMedicine
Table 2 Renal haemodynamic responses to clamped hyperglycaemia in 74 patients with Type 1 diabetes Type 1 diabetes: euglycaemia Systemic haemodynamic function Heart rate, beats 72 12 per min Systolic blood pressure, 114 0.4 mmHg Diastolic blood pressure, 65.5 5.3 mmHg Renal haemodynamic function Effective renal plasma 810 269 flow (ml/min/1.73 m2) 144 32.9 GFR, ml/min/1.73 m2 Filtration fraction 0.18 0.04 Renal blood flow, 1299 439 ml/min/1.73 m2 0.068 0.02 Renal vascular resistance, mmHg/l/min Circulating neurohormonal factors Aldosterone, pmol/L 43 30 Plasma renin 0.56 0.48 activity, ng/ml/h Angiotensinogen, 1158 773 ng/ml Intraglomerular haemodynamic variables Filtration pressure, 28 6 mmHg Glomerular hydrostatic 61 8 pressure, mmHg Oncotic pressure, 23 3 mmHg 1473 798 Afferent arteriolar resistance, dyne/s/cm5 2021 590 Efferent arteriolar resistance, dyne/s/cm5
Type 1 diabetes: hyperglycaemia
70 10.6 116 10.6 65.2 7.2 831 236 156 30.7† 0.19 0.05 1309 399 0.067 0.017
32 13* 0.87 1.4 1059 550 30 7† 62 8 22 3 1325 825 2205 716
Values are mean SD. *P < 0.01, †P < 0.05, for comparison of hyperglycaemia vs. euglycaemia.
glomerular hydrostatic pressure in those with Type 1 diabetes and normofiltration that may help to distinguish this group from healthy control subjects. Furthermore, higher afferent renal arteriolar resistance values in the Type 1 diabetes and normofiltration group, although not statistically significant compared with healthy control subjects, were suggestive of increased afferent arteriolar tone in Type 1 diabetes and normofiltration, which may account for the significantly lower glomerular hydrostatic pressure. Future mechanistic studies should further explore the haemodynamic differences between healthy control subjects and patients with Type 1 diabetes and normofiltration using dynamic pharmacological probes, including drugs that influence tubular function (sodium glucose cotransport inhibitors and incretin-based agents) and neurohormonal pathways (RAAS inhibitors). In addition, longitudinal follow-up is required to determine if the high afferent renal arteriolar resistance phenotype in patients with Type 1 diabetes and normofiltration changes to one of low afferent renal arteriolar resistance as these patients develop hyperfiltration, or, if young patients
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
with Type 1 diabetes and hyperfiltration exhibit a rise in afferent renal arteriolar resistance over time as renal function is lost. Finally, while our analysis implicates low afferent renal arteriolar resistance as the mechanism leading to hyperfiltration in patients with relatively long duration of disease, we can only speculate about possible afferent vasodilators causing the decline in afferent renal arteriolar resistance, listed above. Since glycaemic control and urinary measures of sodium and protein intake were similar at the time of haemodynamic testing, these factors were unlikely to have confounded our results, although we cannot be certain about the influence of these variables prior to the study period since these were not measured. The lack of observed differences in efferent renal arteriolar resistance deserves special comment. Previous studies using agents that act at the efferent arteriole—such as RAAS inhibitors and endothelin-1 type A receptor antagonists—have been shown to have significant effects on renal protection and anti-proteinuric effects [5,16]. Furthermore, haemodynamic studies in patients with Type 1 diabetes have shown a reduction in hyperfiltration with RAAS inhibition that is attributable to efferent arteriolar vasodilatation [5,17], suggesting that underlying, baseline efferent vasoconstriction was ‘corrected’ with angiotensin-converting enzyme inhibition. It was therefore surprising that efferent renal arteriolar resistance was similar in the three groups. The apparent similarity in baseline efferent renal arteriolar resistance may be attributable to: 1) a true lack of differences in efferent renal arteriolar resistance in the three groups, which may suggest that differences in KFG, which can also be influenced by RAAS blockade, are responsible for haemodynamic and GFR effects linked with this drug class; 2) insensitivity of the Gomez equations to detect differences in efferent renal arteriolar resistance; and 3) a distinct form of hyperfiltration in this cohort. As the cohort in this analysis included patients with a relatively long duration of diabetes, the low afferent renal arteriolar resistance phenotype may be different from that of patients with hyperfiltration and a shorter duration of disease, or other clinical features such as poor glycaemic control, who may have other haemodynamic abnormalities, such as elevated efferent renal arteriolar resistance. Future work using dynamic probes of efferent renal arteriolar resistance, such as RAAS inhibitors, might help to further clarify the role of high efferent renal arteriolar resistance as a contributor to hyperfiltration [17]. Furthermore, it would be interesting to explore in future studies whether a combination of agents that sequentially alter afferent renal arteriolar resistance (sodium-glucose cotransporter 2 inhibition) and efferent renal arteriolar resistance (RAAS blockade) in people with Type 1 diabetes and hyperfiltration lead to additive or synergistic renal effects as a result of blockade of both neurohormonal and tubular factors that promote hyperfiltration, thereby delaying the progression of diabetic nephropathy.
977
DIABETICMedicine
As a final comment, in previous work we reported that induction of clamped hyperglycaemia acutely increases GFR in patients with Type 1 diabetes and normofiltration, whereas there is very little change in those with Type 1 diabetes and hyperfiltration [13]. In keeping with this previous work, GFR rose significantly in the Type 1 diabetes and normofiltration group but not in the Type 1 diabetes and hyperfiltration group. We were unable, however, to determine if hyperglycaemia-induced changes in GFR were on the basis of efferent arteriolar constriction caused by RAAS activation, or rather effects on afferent vasodilatation from increased proximal tubular sodium-glucose cotransport, leading to decreased macula densa sodium delivery and suppressed tubuloglomerular feedback. Alternatively, changes in GFR may have been on the basis of changes in KFG, which we were not able to measure using the current methodology. Overall, in the combined Type 1 diabetes group comprising 74 patients, hyperglycaemia was associated with a significant increase in GFR, possibly because of a rise in glomerular filtration pressure. The present study has some limitations. First, the small sample size probably limited our ability to detect some between-group differences, such as afferent renal arteriolar resistance during clamped euglycaemia in healthy control subjects vs. patients with Type 1 diabetes and normofiltration. Second, the methodology we used in the calculation of glomerular haemodynamics was indirect. Although these formulae have been carefully tested and applied in other clinical contexts over the past 60 years, they are based on physiological assumptions, as described above. Third, while we have proposed several possible explanations, we were not able to determine the mechanistic basis for differential afferent renal arteriolar resistance levels in the three groups in the present study. Fourth, our observation that renal hyperfiltration in patients with Type 1 diabetes and hyperfiltration is associated with afferent vasodilatation may not be generalizable to the hyperfiltration that occurs in the context of other renal pathologies such as pregnancy, obesity, Type 2 diabetes, focal sclerosis and sickle cell disease. We also recognize that our present observations were made using data from several previous studies, and as such should be considered hypothesisgenerating only. Importantly, however, all studies were completed under the supervision of the same investigator, study nurse and central laboratory, making it unlikely that measurement bias played a significant role. Furthermore, as patients included in the analyses had normal blood pressure and normoalbuminuria, routine clinical care would not have changed appreciably over the last 10 years. Changes in patient management over time were probably minor. In conclusion, we have provided the first evidence in humans suggesting that glomerular haemodynamic function differences in patients with Type 1 diabetes vs. healthy control subjects are primarily based on alterations in afferent arteriolar tone. Future studies are required to determine the mechanistic basis for differences in afferent renal arteriolar
978
c et al. Glomerular haemodynamics in diabetes M. Skrti
resistance, and to assess how glomerular haemodynamic function changes over the course of the natural history of diabetes. Finally, clarification of the role of afferent vs. efferent arteriolar abnormalities in Type 1 diabetes is important in order to understand the potential renal protective effects of existing efferent-related treatments such as RAAS inhibitors, and how they interact with emerging strategies that influence afferent tone such as sodium-glucose cotransporter 2 inhibitors.
Funding sources
D.Z.I.C. was supported by a Kidney Foundation of Canada Scholarship and a Canadian Diabetes Association-KRESCENT Program Joint New Investigator Award, and receives operating support from the Canadian Institutes of Health Research, the Kidney Foundation of Canada.
Competing interests
D.Z.I.C. has received consulting fees from Boehringer Ingelheim and Eli Lilly, Janssen and Merck and operational grant support from Boehringer Ingelheim and Eli Lilly.
Acknowledgements
The authors would also like to thank Jenny Cheung for her assistance with the biochemical assays included in this work. Finally, the authors are grateful to the study participants whose time and effort are critical to the success of our research program.
References 1 Magee GM, Bilous RW, Cardwell CR, Hunter SJ, Kee F, Fogarty DG. Is hyperfiltration associated with the future risk of developing diabetic nephropathy? A meta-analysis. Diabetologia 2009; 52: 691–697. 2 Cherney DZ, Scholey JW, Miller JA. Insights into the regulation of renal hemodynamic function in diabetic mellitus. Curr Diabetes Rev 2008; 4: 280–290. 3 Vallon V, Thomson SC. Renal function in diabetic disease models: the tubular system in the pathophysiology of the diabetic kidney. Annu Rev Physiol 2012; 74: 351–375. 4 Zatz R, Dunn BR, Meyer TW, Anderson S, Rennke HG, Brenner BM. Prevention of diabetic glomerulopathy by pharmacological amelioration of glomerular capillary hypertension. J Clin Invest 1986; 77: 1925–1930. 5 Sochett EB, Cherney DZ, Curtis JR, Dekker MG, Scholey JW, Miller JA. Impact of renin angiotensin system modulation on the hyperfiltration state in type 1 diabetes. J Am Soc Nephrol 2006; 17: 1703–1709. 6 Cherney DZI, Perkins BA, Soleymanlou N, Maione M, Lai V, Lee A et al. The Renal Hemodynamic Effect of SGLT2 Inhibition in Patients with Type 1 Diabetes. Circulation 2014; 129: 587–597. 7 Gomez DM. Evaluation of renal resistances, with special reference to changes in essential hypertension. J Clin Invest 1951; 30: 1143– 1155.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
8 Tsuda A, Ishimura E, Ohno Y, Ichii M, Nakatani S, Mori K et al. Significant association of poor glycemic control with increased resistance in efferent arterioles–study of inulin and para-aminohippuric acid clearance in humans. Diabetes Res Clin Pract 2014; 104: 234–240. 9 Ott C, Schneider MP, Raff U, Ritt M, Striepe K, Alberici M et al. Effects of manidipine vs. amlodipine on intrarenal haemodynamics in patients with arterial hypertension. Br J Clin Pharmacol 2013; 75: 129–135. 10 Tsuda A, Inaba M, Ichii M, Ochi A, Ohno Y, Nakatani S et al. Relationship between serum TSH levels and intrarenal hemodynamic parameters in euthyroid subjects. Eur J Endocrinol 2013; 169: 45–50. 11 Cherney DZ, Reich HN, Jiang S, Har R, Nasrallah R, Hebert RL et al. Hyperfiltration and the effect of nitric oxide inhibition on renal and endothelial function in humans with uncomplicated type 1 diabetes mellitus. Am J Physiol Regul Integr Comp Physiol 2012; 303: R710–R718. 12 Cherney DZ, Scholey JW, Jiang S, Har R, Lai V, Sochett EB et al. The effect of direct renin inhibition alone and in combination
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
DIABETICMedicine
13
14
15
16
17
with ACE inhibition on endothelial function, arterial stiffness, and renal function in type 1 diabetes. Diabetes Care 2012; 35: 2324– 2330. Yang GK, Maahs DM, Perkins B, Cherney DZI. Renal Hyperfiltration and Systemic Blood Pressure in Patients with Uncomplicated Type 1 Diabetes Mellitus. PloS One 2013; 8: e68908. Cherney DZ, Miller JA, Scholey JW, Bradley TJ, Slorach C, Curtis JR et al. The effect of cyclooxygenase-2 inhibition on renal hemodynamic function in humans with type 1 diabetes. Diabetes 2008; 57: 688–695. Komers R, Anderson S, Epstein M. Renal and cardiovascular effects of selective cyclooxygenase-2 inhibitors. Am J Kid Dis 2001; 38: 1145–1157. Mann JF, Green D, Jamerson K, Ruilope LM, Kuranoff SJ, Littke T et al. Avosentan for overt diabetic nephropathy. J Am Soc Nephrol 2010; 21: 527–535. Raij L, Shultz PJ, Tolins JP. Possible mechanism for the renoprotective effect of angiotensin converting enzyme inhibitors. J Hypertens Suppl 1989; 7: S33–S36.
979
Research: Pathophysiology Glomerular haemodynamic profile of patients with Type 1 diabetes compared with healthy control subjects M. Skrtic1, Y. Lytvyn1, G. K. Yang1, P. Yip2, V. Lai1, M. Silverman1 and D. Z. I. Cherney1 1 Department of Medicine, Division of Nephrology, Toronto General Hospital, University of Toronto and 2University Health Network, Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada
Accepted 4 February 2015
Abstract Aims To evaluate the glomerular haemodynamic profile of patients with Type 1 diabetes with either renal hyperfiltration (GFR ≥ 135 ml/min/1.73 m2) or renal normofiltration (GFR 90–134 ml/min/1.73 m2) during euglycaemic and hyperglycaemic conditions, and to compare this profile with that of a similar group of healthy control subjects. Methods Gomez’s equations were used to derive afferent and efferent arteriolar resistances, glomerular hydrostatic pressure and filtration pressure.
At baseline, during clamped euglycaemia, patients with Type 1 diabetes and hyperfiltration had lower mean SD afferent arteriolar resistance than both those with Type 1 diabetes and normofiltration (914 494 vs. 2065 597 dyne/s/cm5; P < 0.001) and healthy control subjects (1676 707 dyne/s/cm5; p < 0.001). By contrast, efferent arteriolar resistance was similar in the three groups. Patients with Type 1 diabetes and hyperfiltration also had higher mean SD glomerular hydrostatic pressure than both healthy control subjects and patients with Type 1 diabetes and normofiltration (66 6 vs. 60 3 vs. 55 3 mmHg; P < 0.05). Similar findings for afferent arteriolar resistance, efferent arteriolar resistance, glomerular hydrostatic pressure and filtration pressure were observed during clamped hyperglycaemia. Results
Hyperfiltration in Type 1 diabetes is primarily driven by alterations in afferent arteriolar resistance rather than efferent arteriolar resistance. Renal protective therapies should focus on afferent renal arteriolar mechanisms through the use of pharmacological agents that target tubuloglomerular feedback, including sodium-glucose cotransporter 2 inhibitors and incretins.
Conclusion
Diabet. Med. 32, 972–979 (2015)
Introduction Renal hyperfiltration has been associated with increased intraglomerular pressure, albuminuria and the progression of diabetic nephropathy [1]. Hyperfiltration in humans has been attributed to two major influences: 1) neurohormonal factors, such as the renin angiotensin aldosterone system (RAAS), which causes efferent vasoconstriction, as well as afferent vasodilators, such as nitric oxide, and 2) tubular factors, which consist of altered tubuloglomerular feedback as a result of increased sodium-glucose cotransporter 2 activity, leading to decreased sodium delivery to the macula densa [2,3]. This results in a perceived reduced effective circulating volume, leading to afferent renal arteriolar vasodilatation [3].
Correspondence to: David Cherney. E-mail: [email protected]
972
Evidence of early renal haemodynamic changes resulting from changes in efferent and afferent tone from neurohormonal mediators and altered tubuloglomerular feedback have been provided by animal micropuncture studies [3]. Importantly, the vasodilatory effect of RAAS inhibitors at the efferent arteriole in preclinical models has formed the foundation for therapies that protect against the development of diabetic nephropathy in humans [4]. For tubular factors, until recently, it was not known if pharmacological modulation of proximal renal tubular sodium handling affected hyperfiltration in humans. Accordingly, to determine if increasing distal tubular sodium delivery reduces hyperfiltration in humans, we treated patients with Type 1 diabetes mellitus with the sodium-glucose cotransporter 2 inhibitor empagliflozin for 8 weeks and showed that there was a significant attenuation of renal hyperfiltration that was similar in magnitude to that achieved with an angiotensin-converting enzyme
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
What’s new? • Renal hyperfiltration in Type 1 diabetes is primarily attributable to vasodilatation of the renal afferent arteriole. • Induction of clamped hyperglycaemia was associated with an increase in renal filtration pressure. • There are potential future implications of using ‘afferent constrictors’ such as sodium-glucose cotransporter 2 inhibitors in combination with traditional renal protective strategies, such as renin angiotensin aldosterone system inhibitors, as these older agents act primarily through efferent vasodilatation. inhibitor in our previous work [5,6]. Because of an inability to assess segmental (afferent vs. efferent) resistance directly in humans, however, interpretation of pharmacological responses to existing and novel therapies has been inferred based on changes in GFR, effective renal plasma flow, renal blood flow, filtration fraction and renal vascular resistance. A more detailed description of in vivo glomerular function in humans, including pre-glomerular (afferent) and post-glomerular (efferent) arteriolar function is required to understand the pathophysiology of early diabetic nephropathy, as well as the beneficial and deleterious responses to emerging renal protective therapies. Although direct measurements of human glomerular haemodynamic variables such as afferent and efferent arteriolar resistance values and glomerular hydrostatic pressure are not feasible in vivo in humans, Gomez. [7] derived equations for these variables. Subsequent work has used these equations to evaluate patients with a variety of conditions, including hypertension and endocrine disorders [8–10]. To our knowledge, these techniques have not been used to characterize the renal haemodynamic profile in patients with diabetes mellitus, or to evaluate kidney function in patients with Type 1 diabetes and either hyperfiltration or normal filtration. Accordingly, the aim of the present exploratory analysis was to determine the glomerular haemodynamic profile of patients with Type 1 diabetes and either hyperfiltration or normofiltration and to compare them with a similar group of normal healthy control subjects using baseline data from several previous physiology studies [6,11,12]. Our aim was to determine if there were fundamental differences in glomerular haemodynamic variables (afferent and efferent arteriolar resistance values and glomerular hydrostatic pressure) between the three groups, based on filtration status, that could potentially place some individuals at risk of future renal disease, and potentially explain differences in response to renal protective therapies such as RAAS inhibitors.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
DIABETICMedicine
Patients and methods A total of 25 healthy control subjects, 35 patients with Type 1 diabetes and normofiltration and 38 patients with Type 1 diabetes and hyperfiltration were included in this exploratory analysis. Subjects were included from four previously published studies that had the same inclusion criteria: Type 1 diabetes duration > 1 year, age > 18 years, blood pressure < 140/90 mmHg and normoalbuminuria. Patients were not taking medications that alter blood pressure or cardiovascular variables. Patients were originally identified using office and hospital advertisements, through social networking and from review of clinic charts. The study was approved by the Research Ethics Board at the University Health Network. All subjects gave informed consent before the start of the study procedures. Subjects adhered to a sodium replete (> 140 mmol/day) and moderate protein (< 1.5 g/kg/day) diet during the 7 days preceding the experiment as described [6,11,12]. All studies were performed during 2 consecutive days at the Renal Physiology Laboratory at Toronto General Hospital. A modified glucose clamp technique was used to maintain euglycaemic (4–6 mmol/l) conditions on the first day, followed by hyperglycaemic (9–11 mmol/l) conditions on the second day as previously described. Studies were performed under these conditions because hyperfiltration is based on GFR during euglycaemia, and because of the effect of hyperglycaemia on GFR and blood pressure [13]. Constant inulin and para-aminohippurate infusions were used to determine GFR and effective renal plasma flow as previously described. After a 10-min equilibration period, blood samples were collected for inulin, para-aminohippurate, haematocrit and total protein, with additional blood evaluation carried out 30 min later. Two independent clearance periods were used for calculations, and the results averaged. All haemodynamic measurements were adjusted for body surface area. Intrarenal haemodynamics were calculated using formulae established by Gomez [7] to indirectly determine intraglomerular haemodynamics in humans based on data from animal studies. Gomez’s equations necessitate the following: i) that intrarenal vascular resistances be divided into afferent, postglomerular and efferent resistance; ii) that hydrostatic pressures within the renal tubules, venules, Bowman’s space and interstitium are in equilibrium and given a value of 10 mmHg; iii) that filtration disequilibrium in the glomerulus is assumed; and iv) that the gross filtration coefficient is assumed to be 0.0867 ml/s per mmHg, given a normal kidney physiology (GFR 130 ml/min, glomerular hydrostatic pressure 60 mmHg), and given Winton’s indirect estimates in a dog model that glomerular pressure is roughly two-thirds of mean arterial pressure (MAP) [7], and normal glomerular oncotic pressure is 25 mmHg. Clinical variables, including mean arterial pressure (mmHg), filtration fraction (FF), effective renal plasma flow (RPF) (ml/s), GFR (ml/s), renal
973
c et al. Glomerular haemodynamics in diabetes M. Skrti
DIABETICMedicine
blood flow (RBF) (ml/s), haematocrit (Hct) (%) and total protein (TP) (g/dl), were used to calculate efferent and afferent renal arteriolar resistance values (dyne/s/cm5), glomerular hydrostatic pressure (mmHg), glomerular filtration pressure (mmHg) and glomerular oncotic pressure (mmHg): FF ¼ GFR/RPF RBF ¼ RPF=ð1 HctÞ The filtration pressure across glomerular capillaries (DPF) is given by: DPF ¼ GFR/KFG where KFG is the gross filtration coefficient. The glomerular oncotic pressure (pG) is calculated from the plasma protein mean concentration (CM) within the capillaries: CM ¼ TP/FFlnð1=1 FFÞ pG ¼ 5ðCM 2Þ By substituting DPF, pG, and given that hydrostatic pressure in Bowman’s space (PBow) is assumed to be 10 mmHg, PGLO is calculated: PGLO ¼ DPF þ PBow þ pG PGLO ¼ðGFR/KFG Þ þ 10 mmHg þ ½5ðTP/FFlnð1=1 FFÞ 2Þ To calculate efferent (RE) and afferent renal arteriolar resistance (RA) values using principles of Ohm’s law, where 1328 is the conversion factor to dyne/s/cm5 [7]: RA ¼ ½ðMAP PGLO Þ=RBF1328 RE ¼ ½GFR/(KFG ðRBF GFRÞ1328
Statistical analysis
All data are expressed as mean SD values. Statistical analyses were performed using one-way ANOVA and post hoc Tukey’s tests were used for between-group differences as indicated. Differences were considered statistically significant at a P value of < 0.05. All statistical analyses were carried out using GRAPHPAD PRISM software (version 5.0).
hyperglycaemic conditions. As expected, the patients with Type 1 diabetes and normofiltration and those with hyperfiltration had significantly lower levels of circulating RAAS mediators compared with healthy control subjects. Also as expected, patients with Type 1 diabetes and hyperfiltration had higher GFR, effective renal plasma flow and renal blood flow and lower renal vascular resistance compared with both the Type 1 diabetes and normofiltration group and the healthy control group (P < 0.001 for all comparisons; Table 1).
Glomerular haemodynamic variables: euglycaemia
At baseline, during clamped euglycaemia, Gomez’s calculations of glomerular haemodynamic variables showed that afferent renal arteriolar resistance was significantly lower in the Type 1 diabetes and hyperfiltration group compared with both the Type 1 diabetes and normofiltration group and the healthy control group (P < 0.001; Figure 1). Glomerular hydrostatic pressure and glomerular filtration pressure were also higher in the Type 1 diabetes and hyperfiltration group than in both the Type 1 diabetes and normofiltration group and the healthy control group (P < 0.001; Figure 1). Glomerular hydrostatic pressure was lower in the Type 1 diabetes and normofiltration group than in the healthy control group (55 3 vs. 60 3 mmHg; P < 0.001), while afferent renal arteriolar resistance, efferent renal arteriolar resistance and glomerular filtration pressure values were similar between the two groups. There were no significant differences for efferent renal arteriolar resistance in any of the three groups (Figure 1).
Glomerular haemodynamic variables: hyperglycaemia
Similarly to the measurements obtained during euglycaemia, afferent renal arteriolar resistance was significantly lower during clamped hyperglycaemia in patients with Type 1 diabetes and hyperfiltration compared with patients with Type 1 diabetes and normofiltration, while efferent renal arteriolar resistance was the same in the two groups. In addition, glomerular hydrostatic pressure (67 7 vs. 57 6 mmHg; P < 0.001) and glomerular filtration pressure (34 7 vs. 25 5 mmHg; P < 0.001) were significantly higher in the Type 1 diabetes and hyperfiltration vs the Type 1 diabetes and normofiltration group during hyperglycaemia (Figure 1).
Results Renal haemodynamic responses to clamped hyperglycaemia Baseline characteristics
Clinical and demographic characteristics were similar among the healthy control group, the Type 1 diabetes and normofiltration group and the Type 1 diabetes and hyperfiltration group (Table 1). All subjects had baseline renal haemodynamic studies performed during clamped euglycaemic and
974
In response to clamped hyperglycaemia, GFR increased significantly in the Type 1 diabetes and normofiltration group (117 11 to 132 11 ml/min/1.73 m2; P < 0.005), but not in the Type 1 diabetes and hyperfiltration group (177 35 to 168 25 ml/min/1.73 m2; P > 0.05). There were no significant changes in response to hyperglycaemia for effective renal
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
DIABETICMedicine
Table 1 Clinical baseline characteristics of normal healthy control subjects and patients with Type 1 diabetes and either renal normofiltration or hyperfiltration
Healthy control group, n = 25 Baseline characteristic Males, n (%) Mean SD age, years Diabetes duration, years Weight, kg, mean SD Height, m, mean SD BMI, kg/m2, mean SD HbA1c, mmol/mol, mean SD HbA1c, %, mean SD Sodium Excretion (mmol/24 hr), mean SD Urea Excretion (mmol/24 hr), mean SD Systemic haemodynamic function, mean SD Heart rate, beats per min Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Renal haemodynamic function, mean SD Effective renal plasma flow, ml/min/1.73 m2 GFR, ml/min/1.73 m2 Filtration fraction Renal blood flow, ml/min/1.73 m2 Renal vascular resistance, mmHg/l/min Circulating neurohormonal factors, mean SD Aldosterone, pmol/L Plasma renin activity, ng/ml/h Angiotensinogen, ng/ml Intraglomerular haemodynamic variables, mean Filtration pressure, mmHg Glomerular hydrostatic pressure, mmHg Oncotic pressure, mmHg Afferent arteriolar resistance, dyne/s/cm5 Efferent arteriolar resistance, dyne/s/cm5
–
–
11 (44) 26.8 5.6
Type 1 diabetes and normofiltration group, n = 35
Type 1 diabetes and hyperfiltration group, n = 38
183 104 298 128
21 (60) 24.2 5.3 18.9 5.4 73.1 12.7 1.73 0.09 24.2 3.5 68 10 8.4 1.6 174 103 324 105
17 (45) 23.3 3.9 16.0 6.7 70.5 11.5 1.70 0.08 24.2 2.9 69 13 8.5 1.4 160 85.8 307 135
61.3 9.2 110 9.8 65.7 6.9
68.1 11.7* 113 8.5 64.9 6.6
74.0 9.9† 116 11.0‡ 65.5 6.9
69.2 11.4 1.74 0.08 22.8 3.1
628 117 0.19 1039 0.081
162 11 0.04 287 0.019
340 314 1.60 0.96 1338 791
642 116 0.18 1031 0.080
93 10.4 0.03 171 0.014
46 35* 0.57 0.56* 1298 930
970 177 0.18 1552 0.057
39 23* 0.55 0.4* 1017 558
SD
22.4 59.6 27.3 1676 2043
2.2 2.7 2.6 706 463
22.3 54.5 22.3 2065 1993
2.0 3.4* 2.6* 597* 367
268*§ 35*§ 0.06 466*§ 0.017*§
32.3 65.9 23.5 914 2047
4.9*§ 6.1‡§ 2.7* 494*§ 747
Values are mean SD, unless otherwise indicated. *P < 0.01 vs healthy control group. *P < 0.001 vs healthy control group. † P < 0.05 vs Type 1 diabetes and normofiltration group. ‡ P < 0.05 vs healthy control group. § P < 0.001 vs Type 1 diabetes and normofiltration group.
plasma flow (Type 1 diabetes and normofiltration group: 642 93 to 690 129 ml/min/1.73 m2; Type 1 diabetes and hyperfiltration group: 970 268 to 955 257 ml/min/ 1.73 m2) and renal blood flow (Type 1 diabetes and normofiltration group: 1039 171 to 1099 208 ml/min/ 1.73 m2; Type 1 diabetes and hyperfiltration: 1503 435 to 1552 466 ml/min/1.73 m2). For glomerular haemodynamic function, afferent renal arteriolar resistance tended to decrease (Type 1 diabetes and normofiltration group: 2065 597 to 1834 684 dyne/s/ cm5; Type 1 diabetes and hyperfiltration group: 914 494 to 856 652 dyne/s/cm5), while efferent renal arteriolar resistance tended to increase (Type 1 diabetes and normofiltration group: 1993 366 to 2151 475 dyne/s/cm5; Type 1 diabetes and hyperfiltration group: 2047 747 to 2255 885 dyne/s/cm5) in response to clamped hyperglycaemia, but these changes were not significant. Changes in glomerular filtration pressure (Type 1 diabetes and normofiltration group: 22 2 to 25 5 mmHg; Type 1 diabetes and hyperfiltration group: 32 5 to 34 7 mmHg) and ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
glomerular hydrostatic pressure (Type 1 diabetes and normofiltration group: 55 3 to 57 6 mmHg; Type 1 diabetes and hyperfiltration group: 66 6 to 67 7 mmHg) were also not significant. When patients with Type 1 diabetes and either normofiltration or hyperfiltration were examined as a single group, hyperglycaemia was found to have induced a significant increase in GFR (144 33 to 156 31 ml/min/1.73 m2; P < 0.05) and thereby also in filtration pressure (28 6 to 30 7 mmHg; P < 0.05). There were no other significant changes during hyperglycaemia in the combined cohort of patients with Type 1 diabetes (Table 2).
Discussion Direct measurements of segmental renal arteriolar resistance are not technically feasible in humans. As a consequence, the physiological basis for early diabetic nephropathy remains incompletely understood. In particular, the relative contribution of afferent vasodilatation vs. efferent vasoconstriction
975
DIABETICMedicine
c et al. Glomerular haemodynamics in diabetes M. Skrti
FIGURE 1 Glomerular haemodynamics estimated using Gomez’s equations (gross filtration coefficient = 0.0867) in normal healthy controls and patients with Type 1 diabetes and either renal normofiltration or hyperfiltration at baseline during clamped euglycaemia (open circles) and hyperglycaemia (filled circles) conditions. *P < 0.001 vs. healthy control subjects, †P < 0.001 vs Type 1 diabetes and normofiltration using post hoc Tukey’s test after ANOVA. Horizontal lines represent means (healthy controls, n = 25; patients with normofiltration, n = 35; patients with hyperfiltration, n = 38) of haemodynamic characteristics.
to the pathogenesis of renal hyperfiltration remains unknown. To further clarify the role of pre-glomerular vs. post-glomerular factors in the pathogenesis of haemodynamic abnormalities typical of early diabetes, we applied the formulae derived by Gomez to calculate glomerular haemodynamic variables in both healthy control subjects and in patients with Type 1 diabetes. In the group with Type 1 diabetes, it was critical to analyse patients on the basis of hyperfiltration status. Although the patients with Type 1 diabetes and hyperfiltration and those with Type 1 diabetes and normofiltration did not differ in age, diabetes duration, gender distribution or HbA1c concentration, those with Type 1 diabetes and hyperfiltration had physiologically distinct responses to sodium-glucose cotransporter 2 inhibitors, RAAS inhibitors and other pharmacological agents, highlighting their distinct physiological and renal haemodynamic profile [5,6,11,12,14]. Our first major observation was that hyperfiltration in Type 1 diabetes was primarily associated with haemodynamic changes at the afferent rather than the efferent arteriole. Previous micropuncture studies in animal models of diabetes have suggested a prominent role for afferent vasodilation, leading to hyperfiltration, through increased nitric oxide bioactivity, vasodilatory prostanoid production and changes in tubuloglomerular feedback related to increased proximal tubular sodium-glucose cotransport [3,15]. In patients with Type 1 diabetes and hyperfiltration, we have shown similar relationships between nitric oxide, prostanoids and sodiumglucose cotransport (as a tool for tubuloglomerular feedback) and hyperfiltration, suggesting a prominent role for afferent arteriolar vasodilatation in the pathogenesis of this condition
976
in humans [6,11,14]. To our knowledge, the present study is the first to show a significantly lower afferent renal arteriolar resistance in patients with Type 1 diabetes and hyperfiltration compared with both patients with Type 1 diabetes and normofiltration and healthy control subjects. In conjunction with higher glomerular filtration pressure and glomerular hydrostatic pressure in the group with Type 1 diabetes and hyperfiltration compared with the other two groups, and no difference in efferent renal arteriolar resistance, our results suggest that changes in glomerular haemodynamic function in early Type 1 diabetes are in large part related to changes in afferent rather than efferent tone. Our second finding was that, even though afferent renal arteriolar resistance, efferent renal arteriolar resistance and glomerular filtration pressure tended to be similar in the group with Type 1 diabetes and normofiltration compared with the healthy control group, afferent renal arteriolar resistance was 18.8% higher (P = non-significant) than in healthy control subjects. Furthermore, glomerular hydrostatic pressure was lowest in the Type 1 diabetes and normofiltration group, and this difference was significant compared with the other two groups. In our previous work reporting GFR, effective renal plasma flow, renal blood flow and renal vascular resistance, healthy control subjects and patients with Type 1 diabetes and normofiltration appeared to be identical, and had similar responses to afferent arteriolar vasoactive agents, suggesting that Type 1 diabetes and normofiltration exhibit normal renal physiology [11]; however, using the Gomez equations to perform a more indepth analysis of renal haemodynamic function, we were able to detect more subtle, early, ‘pre-clinical’ changes in
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
DIABETICMedicine
Table 2 Renal haemodynamic responses to clamped hyperglycaemia in 74 patients with Type 1 diabetes Type 1 diabetes: euglycaemia Systemic haemodynamic function Heart rate, beats 72 12 per min Systolic blood pressure, 114 0.4 mmHg Diastolic blood pressure, 65.5 5.3 mmHg Renal haemodynamic function Effective renal plasma 810 269 flow (ml/min/1.73 m2) 144 32.9 GFR, ml/min/1.73 m2 Filtration fraction 0.18 0.04 Renal blood flow, 1299 439 ml/min/1.73 m2 0.068 0.02 Renal vascular resistance, mmHg/l/min Circulating neurohormonal factors Aldosterone, pmol/L 43 30 Plasma renin 0.56 0.48 activity, ng/ml/h Angiotensinogen, 1158 773 ng/ml Intraglomerular haemodynamic variables Filtration pressure, 28 6 mmHg Glomerular hydrostatic 61 8 pressure, mmHg Oncotic pressure, 23 3 mmHg 1473 798 Afferent arteriolar resistance, dyne/s/cm5 2021 590 Efferent arteriolar resistance, dyne/s/cm5
Type 1 diabetes: hyperglycaemia
70 10.6 116 10.6 65.2 7.2 831 236 156 30.7† 0.19 0.05 1309 399 0.067 0.017
32 13* 0.87 1.4 1059 550 30 7† 62 8 22 3 1325 825 2205 716
Values are mean SD. *P < 0.01, †P < 0.05, for comparison of hyperglycaemia vs. euglycaemia.
glomerular hydrostatic pressure in those with Type 1 diabetes and normofiltration that may help to distinguish this group from healthy control subjects. Furthermore, higher afferent renal arteriolar resistance values in the Type 1 diabetes and normofiltration group, although not statistically significant compared with healthy control subjects, were suggestive of increased afferent arteriolar tone in Type 1 diabetes and normofiltration, which may account for the significantly lower glomerular hydrostatic pressure. Future mechanistic studies should further explore the haemodynamic differences between healthy control subjects and patients with Type 1 diabetes and normofiltration using dynamic pharmacological probes, including drugs that influence tubular function (sodium glucose cotransport inhibitors and incretin-based agents) and neurohormonal pathways (RAAS inhibitors). In addition, longitudinal follow-up is required to determine if the high afferent renal arteriolar resistance phenotype in patients with Type 1 diabetes and normofiltration changes to one of low afferent renal arteriolar resistance as these patients develop hyperfiltration, or, if young patients
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
with Type 1 diabetes and hyperfiltration exhibit a rise in afferent renal arteriolar resistance over time as renal function is lost. Finally, while our analysis implicates low afferent renal arteriolar resistance as the mechanism leading to hyperfiltration in patients with relatively long duration of disease, we can only speculate about possible afferent vasodilators causing the decline in afferent renal arteriolar resistance, listed above. Since glycaemic control and urinary measures of sodium and protein intake were similar at the time of haemodynamic testing, these factors were unlikely to have confounded our results, although we cannot be certain about the influence of these variables prior to the study period since these were not measured. The lack of observed differences in efferent renal arteriolar resistance deserves special comment. Previous studies using agents that act at the efferent arteriole—such as RAAS inhibitors and endothelin-1 type A receptor antagonists—have been shown to have significant effects on renal protection and anti-proteinuric effects [5,16]. Furthermore, haemodynamic studies in patients with Type 1 diabetes have shown a reduction in hyperfiltration with RAAS inhibition that is attributable to efferent arteriolar vasodilatation [5,17], suggesting that underlying, baseline efferent vasoconstriction was ‘corrected’ with angiotensin-converting enzyme inhibition. It was therefore surprising that efferent renal arteriolar resistance was similar in the three groups. The apparent similarity in baseline efferent renal arteriolar resistance may be attributable to: 1) a true lack of differences in efferent renal arteriolar resistance in the three groups, which may suggest that differences in KFG, which can also be influenced by RAAS blockade, are responsible for haemodynamic and GFR effects linked with this drug class; 2) insensitivity of the Gomez equations to detect differences in efferent renal arteriolar resistance; and 3) a distinct form of hyperfiltration in this cohort. As the cohort in this analysis included patients with a relatively long duration of diabetes, the low afferent renal arteriolar resistance phenotype may be different from that of patients with hyperfiltration and a shorter duration of disease, or other clinical features such as poor glycaemic control, who may have other haemodynamic abnormalities, such as elevated efferent renal arteriolar resistance. Future work using dynamic probes of efferent renal arteriolar resistance, such as RAAS inhibitors, might help to further clarify the role of high efferent renal arteriolar resistance as a contributor to hyperfiltration [17]. Furthermore, it would be interesting to explore in future studies whether a combination of agents that sequentially alter afferent renal arteriolar resistance (sodium-glucose cotransporter 2 inhibition) and efferent renal arteriolar resistance (RAAS blockade) in people with Type 1 diabetes and hyperfiltration lead to additive or synergistic renal effects as a result of blockade of both neurohormonal and tubular factors that promote hyperfiltration, thereby delaying the progression of diabetic nephropathy.
977
DIABETICMedicine
As a final comment, in previous work we reported that induction of clamped hyperglycaemia acutely increases GFR in patients with Type 1 diabetes and normofiltration, whereas there is very little change in those with Type 1 diabetes and hyperfiltration [13]. In keeping with this previous work, GFR rose significantly in the Type 1 diabetes and normofiltration group but not in the Type 1 diabetes and hyperfiltration group. We were unable, however, to determine if hyperglycaemia-induced changes in GFR were on the basis of efferent arteriolar constriction caused by RAAS activation, or rather effects on afferent vasodilatation from increased proximal tubular sodium-glucose cotransport, leading to decreased macula densa sodium delivery and suppressed tubuloglomerular feedback. Alternatively, changes in GFR may have been on the basis of changes in KFG, which we were not able to measure using the current methodology. Overall, in the combined Type 1 diabetes group comprising 74 patients, hyperglycaemia was associated with a significant increase in GFR, possibly because of a rise in glomerular filtration pressure. The present study has some limitations. First, the small sample size probably limited our ability to detect some between-group differences, such as afferent renal arteriolar resistance during clamped euglycaemia in healthy control subjects vs. patients with Type 1 diabetes and normofiltration. Second, the methodology we used in the calculation of glomerular haemodynamics was indirect. Although these formulae have been carefully tested and applied in other clinical contexts over the past 60 years, they are based on physiological assumptions, as described above. Third, while we have proposed several possible explanations, we were not able to determine the mechanistic basis for differential afferent renal arteriolar resistance levels in the three groups in the present study. Fourth, our observation that renal hyperfiltration in patients with Type 1 diabetes and hyperfiltration is associated with afferent vasodilatation may not be generalizable to the hyperfiltration that occurs in the context of other renal pathologies such as pregnancy, obesity, Type 2 diabetes, focal sclerosis and sickle cell disease. We also recognize that our present observations were made using data from several previous studies, and as such should be considered hypothesisgenerating only. Importantly, however, all studies were completed under the supervision of the same investigator, study nurse and central laboratory, making it unlikely that measurement bias played a significant role. Furthermore, as patients included in the analyses had normal blood pressure and normoalbuminuria, routine clinical care would not have changed appreciably over the last 10 years. Changes in patient management over time were probably minor. In conclusion, we have provided the first evidence in humans suggesting that glomerular haemodynamic function differences in patients with Type 1 diabetes vs. healthy control subjects are primarily based on alterations in afferent arteriolar tone. Future studies are required to determine the mechanistic basis for differences in afferent renal arteriolar
978
c et al. Glomerular haemodynamics in diabetes M. Skrti
resistance, and to assess how glomerular haemodynamic function changes over the course of the natural history of diabetes. Finally, clarification of the role of afferent vs. efferent arteriolar abnormalities in Type 1 diabetes is important in order to understand the potential renal protective effects of existing efferent-related treatments such as RAAS inhibitors, and how they interact with emerging strategies that influence afferent tone such as sodium-glucose cotransporter 2 inhibitors.
Funding sources
D.Z.I.C. was supported by a Kidney Foundation of Canada Scholarship and a Canadian Diabetes Association-KRESCENT Program Joint New Investigator Award, and receives operating support from the Canadian Institutes of Health Research, the Kidney Foundation of Canada.
Competing interests
D.Z.I.C. has received consulting fees from Boehringer Ingelheim and Eli Lilly, Janssen and Merck and operational grant support from Boehringer Ingelheim and Eli Lilly.
Acknowledgements
The authors would also like to thank Jenny Cheung for her assistance with the biochemical assays included in this work. Finally, the authors are grateful to the study participants whose time and effort are critical to the success of our research program.
References 1 Magee GM, Bilous RW, Cardwell CR, Hunter SJ, Kee F, Fogarty DG. Is hyperfiltration associated with the future risk of developing diabetic nephropathy? A meta-analysis. Diabetologia 2009; 52: 691–697. 2 Cherney DZ, Scholey JW, Miller JA. Insights into the regulation of renal hemodynamic function in diabetic mellitus. Curr Diabetes Rev 2008; 4: 280–290. 3 Vallon V, Thomson SC. Renal function in diabetic disease models: the tubular system in the pathophysiology of the diabetic kidney. Annu Rev Physiol 2012; 74: 351–375. 4 Zatz R, Dunn BR, Meyer TW, Anderson S, Rennke HG, Brenner BM. Prevention of diabetic glomerulopathy by pharmacological amelioration of glomerular capillary hypertension. J Clin Invest 1986; 77: 1925–1930. 5 Sochett EB, Cherney DZ, Curtis JR, Dekker MG, Scholey JW, Miller JA. Impact of renin angiotensin system modulation on the hyperfiltration state in type 1 diabetes. J Am Soc Nephrol 2006; 17: 1703–1709. 6 Cherney DZI, Perkins BA, Soleymanlou N, Maione M, Lai V, Lee A et al. The Renal Hemodynamic Effect of SGLT2 Inhibition in Patients with Type 1 Diabetes. Circulation 2014; 129: 587–597. 7 Gomez DM. Evaluation of renal resistances, with special reference to changes in essential hypertension. J Clin Invest 1951; 30: 1143– 1155.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
8 Tsuda A, Ishimura E, Ohno Y, Ichii M, Nakatani S, Mori K et al. Significant association of poor glycemic control with increased resistance in efferent arterioles–study of inulin and para-aminohippuric acid clearance in humans. Diabetes Res Clin Pract 2014; 104: 234–240. 9 Ott C, Schneider MP, Raff U, Ritt M, Striepe K, Alberici M et al. Effects of manidipine vs. amlodipine on intrarenal haemodynamics in patients with arterial hypertension. Br J Clin Pharmacol 2013; 75: 129–135. 10 Tsuda A, Inaba M, Ichii M, Ochi A, Ohno Y, Nakatani S et al. Relationship between serum TSH levels and intrarenal hemodynamic parameters in euthyroid subjects. Eur J Endocrinol 2013; 169: 45–50. 11 Cherney DZ, Reich HN, Jiang S, Har R, Nasrallah R, Hebert RL et al. Hyperfiltration and the effect of nitric oxide inhibition on renal and endothelial function in humans with uncomplicated type 1 diabetes mellitus. Am J Physiol Regul Integr Comp Physiol 2012; 303: R710–R718. 12 Cherney DZ, Scholey JW, Jiang S, Har R, Lai V, Sochett EB et al. The effect of direct renin inhibition alone and in combination
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
DIABETICMedicine
13
14
15
16
17
with ACE inhibition on endothelial function, arterial stiffness, and renal function in type 1 diabetes. Diabetes Care 2012; 35: 2324– 2330. Yang GK, Maahs DM, Perkins B, Cherney DZI. Renal Hyperfiltration and Systemic Blood Pressure in Patients with Uncomplicated Type 1 Diabetes Mellitus. PloS One 2013; 8: e68908. Cherney DZ, Miller JA, Scholey JW, Bradley TJ, Slorach C, Curtis JR et al. The effect of cyclooxygenase-2 inhibition on renal hemodynamic function in humans with type 1 diabetes. Diabetes 2008; 57: 688–695. Komers R, Anderson S, Epstein M. Renal and cardiovascular effects of selective cyclooxygenase-2 inhibitors. Am J Kid Dis 2001; 38: 1145–1157. Mann JF, Green D, Jamerson K, Ruilope LM, Kuranoff SJ, Littke T et al. Avosentan for overt diabetic nephropathy. J Am Soc Nephrol 2010; 21: 527–535. Raij L, Shultz PJ, Tolins JP. Possible mechanism for the renoprotective effect of angiotensin converting enzyme inhibitors. J Hypertens Suppl 1989; 7: S33–S36.
979
