Differences in peritoneal response after exposure to low-GDP bicarbonate/lactate-buffered dialysis solution compared to conventional dialysis solution in a uremic mouse model.

BACKGROUND: Long-term exposure of conventional peritoneal dialysis (PD) fluid is associated with structural membrane alterations and technique failure. Previously, it has been shown that infiltrating IL-17-secreting CD4+T cells and pro-fibrotic M2 macrophages play a critical role in the PD-induced pathogenesis. Although more biocompatible PD solutions are recognized to better preserve the peritoneal membrane integrity, the impact of these fluids on the composition of the peritoneal cell infiltrate is unknown. MATERIALS AND METHODS: In a uremic PD mouse model, we compared the effects of daily instillation of standard lactate (LS) or bicarbonate/lactate-buffered solutions (BLS) and respective controls on peritoneal fibrosis, vascularisation, and inflammation. RESULTS: Daily exposure of LS fluid during a period of 8 weeks resulted in a peritoneal increase of αSMA and collagen accompanied with new vessel formation compared to the BLS group. Effluent from LS-treated mouse showed a higher percentage of CD4+ IL-17+ cell population while BLS exposure resulted in an increased macrophage population. Significantly enhanced inflammatory cytokines such as TGFß1, TNFα, INFγ, and MIP-1ß were detected in the effluent of BLS-exposed mice when compared to other groups. Further, immunohistochemistry of macrophage subset infiltrates in the BLS group confirmed a higher ratio of pro-inflammatory M1 macrophages over the pro-fibrotic M2 subset compared to LS. CONCLUSION: Development of the peritoneal fibrosis and angiogenesis was prevented in the BLS-exposed mice, which may underlie its improved biocompatibility. Peritoneal recruitment of M1 macrophages and lower number of CD4+ IL-17+ cells might explain the peritoneal integrity preservation observed in BLS-exposed mouse.

Differential effects of donor-specific HLA antibodies in living versus deceased donor transplant.

The presence of donor-specific anti-HLA antibodies (DSAs) is associated with increased risk of graft failure after kidney transplant. We hypothesized that DSAs against HLA class I, class II, or both classes indicate a different risk for graft loss between deceased and living donor transplant. In this study, we investigated the impact of pretransplant DSAs, by using single antigen bead assays, on long-term graft survival in 3237 deceased and 1487 living donor kidney transplants with a negative complement-dependent crossmatch. In living donor transplants, we found a limited effect on graft survival of DSAs against class I or II antigens after transplant. Class I and II DSAs combined resulted in decreased 10-year graft survival (84% to 75%). In contrast, after deceased donor transplant, patients with class I or class II DSAs had a 10-year graft survival of 59% and 60%, respectively, both significantly lower than the survival for patients without DSAs (76%). The combination of class I and II DSAs resulted in a 10-year survival of 54% in deceased donor transplants. In conclusion, class I and II DSAs are a clear risk factor for graft loss in deceased donor transplants, while in living donor transplants, class I and II DSAs seem to be associated with an increased risk for graft failure, but this could not be assessed due to their low prevalence.

How can we reduce costs of solid-phase multiplex-bead assays used to determine anti-HLA antibodies?

Solid-phase multiplex-bead assays are widely used in transplantation to detect anti-human leukocyte antigen (HLA) antibodies. These assays enable high resolution detection of low levels of HLA antibodies. However, multiplex-bead assays are costly and yield variable measurements that limit the comparison of results between laboratories. In the context of a Dutch national Consortium study we aimed to determine the inter-assay and inter-machine variability of multiplex-bead assays, and we assessed how to reduce the assay reagents costs. Fifteen sera containing a variety of HLA antibodies were used yielding in total 7092 median fluorescence intensities (MFI) values. The inter-assay and inter-machine mean absolute relative differences (MARD) of the screening assay were 12% and 13%, respectively. The single antigen bead (SAB) inter-assay MARD was comparable, but showed a higher lot-to-lot variability. Reduction of screening assay reagents to 50% or 40% of manufacturers' recommendations resulted in MFI values comparable to 100% of the reagents, with an MARD of 12% or 14%, respectively. The MARD of the 50% and 40% SAB assay reagent reductions were 11% and 22%, respectively. From this study, we conclude that the reagents can be reliably reduced at least to 50% of manufacturers' recommendations with virtually no differences in HLA antibody assignments.

Plasma adiponectin concentration has an inverse and a non linear association with estimated glomerular filtration rate in patients with K/DOQI 3 - 5 chronic kidney disease.

BACKGROUND: Chronic kidney disease (CKD) is associated with an increased incidence of cardiovascular disease (CVD). A few studies have demonstrated elevated plasma adiponectin and leptin levels in CKD. The aims of this study were to assess whether 1) estimated glomerular filtration rate (eGFR) is associated with plasma leptin and adiponectin; and 2) adiponectin and leptin (partly) explain associations of CKD with endothelial dysfunction, insulin resistance, and low-grade inflammation in patients with K/DOQI Stage 3 - 5 CKD. METHODS: Baseline data from 91 patients with Stage 3 - 4 CKD in the anti-oxidant therapy in chronic renal insufficiency study, a randomized, double-blind, placebo-controlled trial, in which the effects of oxidative stress-lowering treatment on vascular function and structure were studied, and from 50 dialysis naïve patients, who took part in an open-label, randomized study that compared two peritoneal dialysis regimens, used in the analysis. All subjects for both the studies were recruited in the same centres. RESULTS: The association between eGFR and adiponectin was non-linear. In multivariate analysis, log-eGFR (unstandardized beta = 8.303 microg/ml, p < 0.0001) was the strongest determinant of adiponectin, and body mass index the strongest determinant of leptin (beta = 2.477 ng/ml, p < 0.0001). Plasma adiponectin and leptin did not modify the associations between eGFR and plasma von Willebrand factor or soluble vascular adhesion molecule-1. Plasma leptin had the strongest association with the homeostatic model assessment (HOMA-IR) index. Plasma C-reactive protein had no association with adiponectin or leptin. CONCLUSIONS: In patients with K/DOQI Stage 3 - 5 CKD, renal function had a significant non-linear inverse association with and was the strongest predictor of adiponectin. BMI was the strongest predictor of plasma leptin. Plasma adiponectin and leptin did not explain, and thus presumably are not involved in, the association between eGFR and some markers of endothelial dysfunction.

Aggressive antihypertensive strategies based on hydrochlorothiazide, candesartan or lisinopril decrease left ventricular mass and improve arterial compliance in patients with type II diabetes mellitus and hypertension.

We investigated the effects of aggressive antihypertensive therapy based on hydrochlorothiazide, candesartan or lisinopril on left ventricular mass (LVM) index and arterial stiffness in hypertensive type II diabetic individuals. Seventy hypertensive type II diabetic individuals were treated with three antihypertensive strategies in a randomized, double-blind, double-dummy design. Blood pressure was titrated to levels below 130/85 mm Hg or a decrease in systolic pressure of 10% with a diastolic pressure below 85 mm Hg. After titration, patients were treated for 12 months. Mean blood pressures were 157/93, 151/94 and 149/93 mm Hg at baseline in the hydrochlorothiazide (n = 24), candesartan (n = 24) and lisinopril (n = 22) groups, and 135/80, 135/82 and 131/80 mm Hg after titration. About 70% reached target blood pressures, with the median use of three antihypertensive drugs. Left ventricular mass index and all estimates of arterial stiffness showed significant improvement after 12 months: that is, LVM index (-11 g/m(2); -8%); carotid distensibility coefficient (DC; +2.8 x 10(-3) kPa(-1); +27%), compliance coefficient (CC; +0.13 mm2/kPa; +21%) and elastic modulus (-0.19 kPa; -16%); femoral DC (+1.6 x 10(-3) kPa(-1); +50%) and CC (+0.08 mm2/kPa; +26%); brachial DC (+2.1 x 10(-3) kPa(-1); +39%) and CC (+0.03 mm2/kPa; +27%) and total systemic arterial compliance (+0.29 ml/mm Hg; +16%). No differences in outcome variables between treatment groups were observed. Aggressive antihypertensive treatment, although difficult to achieve, resulted in substantial reductions of LVM index and arterial stiffness in relatively uncomplicated hypertensive type II diabetic individuals. Strategies based on renin-angiotensin system inhibitors were not clearly superior to conventional (i.e. diuretic-based) strategies.

Aggressive antihypertensive therapy based on hydrochlorothiazide, candesartan or lisinopril as initial choice in hypertensive type II diabetic individuals: effects on albumin excretion, endothelial function and inflammation in a double-blind, randomized clinical trial.

We investigated the effects of aggressive antihypertensive therapy based on hydrochlorothiazide, candesartan or lisinopril on urinary albumin excretion, endothelial function and inflammatory activity in hypertensive type II diabetic individuals. A total of 70 hypertensive type II diabetic individuals were treated with three antihypertensive strategies in a randomized, double-blind, double-dummy design. Blood pressure was titrated to levels below 130/85 mmHg or a decrease in systolic pressure of 10% with a diastolic pressure below 85 mmHg. After titration, patients were treated for 12 months. Mean blood pressures changed from 157/93, 151/94 and 149/93 at baseline to 135/80, 135/82 and 131/80 mmHg after titration in the hydrochlorothiazide (n=24), candesartan (n=24) and lisinopril (n=22) groups. About 70% reached target blood pressures. However, only 45% had blood pressures <130/85 mmHg. Urinary albumin excretion and levels of soluble vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 decreased (GEE regression coefficients, -2.40 mg/24 h (P<0.001), -85 ng/ml (P=0.01) and -50 ng/ml (P=0.02)), but brachial artery endothelium-dependent and -independent vasodilation and levels of von Willebrand factor and C-reactive protein did not change (GEE regression coefficients, 0.21 mm (P=0.07), 0.04 mm (P=0.43), 0.04 IU/ml (P=0.33) and -1.15 mg/l (P=0.64)). No differences in outcome variables between treatment groups were observed. These data show that achievement of target blood pressures below 130/85 mmHg in hypertensive type II diabetes is difficult. Aggressive antihypertensive therapy can improve urinary albumin excretion, endothelial function and inflammatory activity in hypertensive type II diabetic individuals, regardless of the type of antihypertensive therapy used.

[From gene to disease; dopamine-beta-hydroxylase deficiency and orthostatic hypotension]. / Van gen naar ziekte; dopamine-beta-hydroxylasedeficiëntie en orthostatische hypotensie.

The DBH gene encodes dopamine-beta-hydroxylase (DbetaH), the enzyme that catalyses the formation of norepinephrine from dopamine. Inactivation of this enzyme due to a mutation of the DBH gene causes a selective (nor)-adrenergic failure of the sympathetic nervous system. This manifests as a severe orthostatic syndrome in which sweating and a normal parasympathetic function are preserved. Several mutations of the DBH gene that cause this very rare syndrome have now been identified. Diagnosis is made on the basis of clinical features and the finding of increased plasma dopamine in the near-absence of norepinephrine. A sole finding of absent plasma DbetaH is insufficient, since about 4% of the population have absent DbetaH. This trait cosegregates with a polymorphism in the promoter region of the DBH gene and is not associated with sympathetic failure. The orthostatic syndrome of DbetaH deficiency can be treated with the non-natural amino acid L-dihydroxyphenylserine, which is decarboxylated to norepinephrine by the ubiquitous aromatic-L-amino acid decarboxylase.

Autonomic nervous function, arterial stiffness and blood pressure in patients with Type I diabetes mellitus and normal urinary albumin excretion.

Type I diabetic patients (DM-1) with an elevated urinary albumin excretion (UAE>30 mg/24 h) have a high cardiovascular risk. However, DM-1 patients with normal UAE have incipient abnormalities of the cardiovascular and nervous systems, such as elevations of blood pressures, increases in arterial stiffness and deterioration of autonomic nervous function. We studied the interrelationships of these abnormalities in normoalbuminuric DM-1 patients. In 76 patients, we performed two cardiovascular reflex tests (deep in- and expiration test (IE test) and lying-to-standing test (LS test)), and determined aortic pulse wave velocity (PWV), local arterial compliances of the common carotid, femoral and brachial arteries, and 24-h blood pressures. The DeltaRRmax value of the LS test was associated with aortic PWV (negatively) and the compliance coefficients of the carotid, femoral and brachial arteries. Per 100-ms increase in DeltaRRmax, pulse wave velocity decreased by 0.39 m/s, compliance coefficients of the carotid, femoral and brachial arteries increased by 0.06, 0.08 and 0.05 mm2/kPa, respectively. These associations were independent of age, 24-h mean arterial pressure and 24-h heart rate. Increases in arterial stiffness were associated with increases in 24-h systolic and pulse pressure (per 1 m/s increase in PWV, systolic and pulse pressure increased by 2.1 and 1.7 mmHg, respectively). In normoalbuminuric DM-1 patients, deterioration of autonomic nervous function is associated with an increase in arterial stiffness, which, in turn, was associated with, and may cause, increased systolic and pulse pressure. These findings suggest that preventive strategies targeting autonomic dysfunction may reduce cardiovascular morbidity in diabetes.

Genetic polymorphisms of the renin-angiotensin system and complications of insulin-dependent diabetes mellitus.

OBJECTIVE: Patients with insulin-dependent diabetes mellitus (IDDM) have a high risk of developing diabetic nephropathy, retinopathy and cardiovascular diseases. The contribution of gene polymorphisms of the renin angiotensin system to these complications is controversial and may differ among populations. METHODS: In 257 Dutch IDDM patients (188 with urinary albumin excretion (UAE) <30 mg/24 h), logistic regression analysis was used to study the relationships among, on the one hand, the insertion/deletion gene polymorphism of the angiotensin-converting enzyme gene (ACE-ID), the M235T gene polymorphism of the angiotensinogen gene (AGT-M235T), and the A1166C gene polymorphism of the angiotensin type 1 receptor gene (AT1-A1166C), and, on the other hand, UAE, retinopathy, hypertension, and coronary heart disease. RESULTS: The T-allele of the AGT-M235T polymorphism was associated with an increased risk of an elevated UAE (odds ratio (OR) 3.03; 95% confidence interval (CI) 1.06-8.61), but only when interaction with the D-allele of the ACE-ID polymorphism was considered. A previously described positive interaction between the T-allele of the AGT-M235T polymorphism and the D-allele of the ACE-ID polymorphism could not be confirmed. The T-allele was also associated with an increased risk of retinopathy (OR 3.89, 95% CI 1.79-8.47). The CC-genotype of the AT1-A1166C polymorphism was associated with hypertension (OR 3.58; 95% CI 1. 23-10.37). CONCLUSIONS: In a Dutch IDDM population, including 69 patients with (incipient) diabetic nephropathy, the T-allele of the AGT-M235T polymorphism is associated with an elevated UAE and diabetic retinopathy and the CC-genotype of the AT1-A1166C polymorphism is associated with hypertension. A previously described interaction between the AGT-M235T and the ACE-ID polymorphisms could not be confirmed. Since the number of nephropathic patients in this study is small, these conclusions must be interpreted with caution.

Ambulatory--not office--blood pressures decline during hormone replacement therapy in healthy postmenopausal women.

Hormone replacement therapy (HRT, estrogen plus progestagen) in postmenopausal women has beneficial effects on the cardiovascular system. However, effects on blood pressure, determined with office measurements, remain controversial. We studied the effects of HRT in 29 healthy normotensive postmenopausal women (mean age 52.3 [3.8] years, median duration of amenorrhea 34.5 months), using ambulatory blood pressure monitoring at baseline and at 3 and 12 months of follow-up. Women were randomized to two groups: an HRT group (N = 14), treated with 1 mg 17beta-estradiol once daily and 5 or 10 mg dydrogesterone once daily during the third and fourth week of every 4 weeks; and a control group (C-group, N = 15), which did not receive therapy. Blood pressures did not differ between the groups at baseline (HRT group 117.1 (9.2)/74.4 (6.6) mm Hg, C-group 113.8 (11.2)/71.3 (7.4) mm Hg). During the follow-up period, changes from baseline of office blood pressures did not differ significantly between the groups. However, changes (95% CI) of mean 24-h blood pressures differed significantly between the two groups after 1 year of follow-up: a decrease of blood pressures was observed in the HRT group (delta systolic/delta diastolic = -5.54 [-8.86 to -2.21]/-4.23 [-6.66 to -1.80] mm Hg), whereas an increase was found in the C-group (+3.33 [-0.69 to +7.35]/+1.67 [-1.75 to +5.09] mm Hg; P [HRT v control group] = .001/.005). We conclude that HRT may have blood pressure lowering properties in healthy, normotensive postmenopausal women.

Determinants of the limits of agreement between the sphygmomanometer and the SpaceLabs 90207 device for blood pressure measurement in health volunteers and insulin-dependent diabetic patients.

OBJECTIVE: To compare the limits of agreement between the SpaceLabs 90207 oscillometric blood pressure measurement device and the sphygmomanometer for patients with insulin-dependent diabetes mellitus (IDDM) versus healthy control subjects. DESIGN: A retrospective analysis of six simultaneous blood pressure measurements in 106 IDDM patients aged 32.6+/-12.0 years (mean +/- SD) and 92 healthy control subjects aged 27.1+/-8.9 years with the SpaceLabs device and a sphygmomanometer. RESULTS: For both groups, overestimation of the systolic blood pressure occurred with the SpaceLabs device, being more severe for the IDDM group (1.7 mmHg more overestimation, P= 0.009). Mean arterial blood pressure was underestimated for the control group and overestimated for the IDDM group (2.7 mmHg underestimation and 0.8 mmHg overestimation, respectively). Diastolic blood pressures for both groups were underestimated by use of the SpaceLabs device, but less severely so for the IDDM group (4.2 mmHg, P< 0.0001 less underestimation). Multiple regression analysis revealed that these differences were mainly due to the presence of diabetes. Other determinants of these limits of agreement were age (for control group subjects) and duration of diabetes (for IDDM group patients). CONCLUSION: Limits of agreement between the SpaceLabs 90207 device and the sphygmomanometer differ between IDDM patients and healthy controls. These limits are influenced by the presence of diabetes, age and duration of diabetes. Available protocols for the evaluation of automated blood pressure measurement devices of the British Hypertension Society and the Association for the Advancement of Medical Instrumentation do not consider limits of agreement in subpopulations such as diabetic patients. Therefore, we recommend that tests for the limits of agreement of blood pressure measurement devices for subgroups should be added to these protocols.

Ambulatory blood pressures and autonomic nervous function in normoalbuminuric type I diabetic patients.

BACKGROUND: In insulin-dependent diabetes mellitus (IDDM) patients with normal urinary albumin excretion (UAE) controversy exists about the presence of blood pressure (BP) elevation and an attenuation of BP decline during sleep. SUBJECTS AND METHODS: These issues were studied in 60 IDDM patients and 55 healthy control subjects with 24 h ambulatory blood pressure monitoring. In addition, in the IDDM patients two cardiovascular reflex tests were performed to study autonomic nervous function. RESULTS: 55 IDDM patients had 4.4/3.1 mm Hg higher 24 h systolic/diastolic pressures when compared with 55 healthy matched controls (P = 0.005/0.009). The diastolic BP decline during sleep was significantly attenuated in IDDM patients compared to healthy volunteers (18.9 vs 22.2%, P = 0.01). The maximum/minimum (max/min) ratio of the RR' interval of the lying to standing test (lower values indicating (incipient) parasympathetic dysfunction) was positively related to the decline of the diastolic BP during sleep in the diabetic patients. This relationship did not persist after adjusting for decline of heart rate during sleep. CONCLUSIONS: IDDM patients with normal UAE, compared with healthy control subjects, have higher BPs during both the waking and sleeping periods and a decreased diastolic BP decline during sleep. In these patients both the diastolic BP decline and the heart rate decline during sleep were related to the max/min ratio. These findings are consistent with the hypothesis that attenuation of diastolic BP decline during sleep is at least partly due to (incipient) damage to the parasympathetic nervous system, which, through a blunted heart rate decline, leads to a decreased decline of cardiac output during sleep.

Sodium, blood pressure, and arterial distensibility in insulin-dependent diabetes mellitus.

We investigated 24-hour ambulatory blood pressure and arterial distensibility, a marker of biophysical vessel wall properties, in 32 normoalbuminuric type I diabetic patients and 32 healthy control subjects on diets containing 50 mmol and 200 mmol sodium per day. The increase in daytime diastolic blood pressure from 50 to 200 mmol sodium was significantly higher in the diabetic patients than in the control subjects (2.3+/-4.9 versus 0.2+/-3.7 mm Hg, P<.05). On a high sodium regimen, femoral artery distensibility was decreased in the diabetic patients compared with the control subjects (19.2+/-7.6 versus 24.1+/-9.3 10[-3]/kPa, P<.05). Angiotensin-converting enzyme inhibition in the diabetic patients on a high sodium diet decreased daytime diastolic blood pressure and increased femoral artery distensibility. The blood pressure decrease in response to angiotensin-converting enzyme inhibition correlated significantly with the blood pressure increase to sodium (for 24-hour systolic and diastolic blood pressure, r=.72, P<.001 and r=.76, P<.001). In addition, we found that in the diabetic patients on a high sodium diet, the renal blood flow response to exogenous angiotensin II was not bimodally distributed, as is the case in essential hypertension, in which a subgroup of the patients are characterized by sodium sensitivity of the blood pressure and an abnormal renal blood flow response to exogenous angiotensin II ("nonmodulator phenotype"). These results show that blood pressure in insulindependent diabetes mellitus is sodium sensitive, but that this is not related to the nonmodulator phenotype, and suggest that in IDDM a relatively high sodium intake may be a factor that predisposes to the development of diabetic vascular disease.

Analysis of twenty-four-hour ambulatory blood pressure monitoring: what time period to assess blood pressures during waking and sleeping?

OBJECTIVE: To study whether two widely used methods of analysis of ambulatory blood pressure monitoring (ABPM), which calculate means for fixed periods, provide accurate estimates of blood pressure during the true waking and sleeping periods. DESIGN AND METHODS: Ninety-five ABPM recordings were retrospectively analysed. Mean systolic and diastolic blood pressures during waking and sleeping were calculated in three ways: the waking and sleeping period according to the diary of the patient (diary method); the waking period from 0700 to 2200 h and the sleeping period from 2200 to 0700 h (9 h); and the waking period form 1000 to 2300 h and the sleeping period from 0100 to 0700 h (6 h). RESULTS: Systolic and diastolic blood pressures during the waking period were not different among the three methods of analysis. Compared with the diary method, the 9 h sleeping period method overestimated systolic blood pressure during sleep by 3.74 mmHg [99% confidence interval (CI) 2.70-4.77] and diastolic blood pressure by 2.44 mmHg (99% CI 1.75-3.14). In contrast, there was no significant difference between the diary method and the 6 h sleeping period method. CONCLUSIONS: The 6 h sleeping period method gave accurate estimates of blood pressure during sleep. The use of the long (9 h) sleeping period method should be avoided in studies in which sleeping-waking differences will be part of the analysis.

A mathematical model of continuous arterio-venous hemodiafiltration (CAVHD).

Continuous arterio-venous hemodiafiltration (CAVHD) differs from conventional hemofiltration and dialysis by the interaction of convection and diffusion, the use of very low dialysate flow rates and by the deterioration of membrane conditions during the treatment. In order to study the impact of these phenomena on diffusive transport, we developed a mathematical model of the kinetics of CAVHD solute transport from plasma water to dialysate. The model yields an expression of the diffusive mass transfer coefficient, Kd, as a function of blood, filtrate and dialysate flow rates and solute concentrations, which can be measured in the clinical setting. This paper gives a description of the model derivation. Kd is demonstrated to vary depending on dialysate flow and duration of treatment.

Solute transport in continuous arteriovenous hemodiafiltration: a new mathematical model applied to clinical data.

A mathematical model of continuous arteriovenous hemodiafiltration is presented, by which the diffusive mass transfer coefficient (Kd) for a solute may be calculated from blood, filtrate and dialysate flow rates and solute concentrations. The model was applied to clinical data obtained with 0.6-m2 AN69 capillary dialyzers that had been used for up to 5 days. The diffusive mass transfer coefficient proved to depend on dialysate flow rate. Furthermore, it was related to the membrane index of ultrafiltration, as measured in the clinic, and to the filter resistance to blood flow. Measurement of these filter characteristics allowed a reasonable prediction of solute clearance.

Determinants of blood flow and ultrafiltration in continuous arteriovenous haemodiafiltration: theoretical predictions and laboratory and clinical observations.

In continuous arteriovenous haemofiltration (CAVH) or haemodiafiltration (CAVHD), it is important to obtain an adequate blood flow through the haemofilter to minimise the risk of excessive haemoconcentration and clotting. In this study we determined the resistance to blood flow of the extracorporeal device as well as the hydraulic permeability of the filter membrane is intensive care patients treated with CAVHD. Data were obtained for CAVH catheters and Scribner shunts and for a polyacrylonitrile (AN-69) plate filter, an AN-69 capillary filter and a polysulphone (PS) capillary filter. In accordance with recent literature we also predicted the resistance to flow by using Poiseuille's law and a formula for the estimation of blood viscosity. Although with all three filters an adequate blood flow was usually obtained, the resistance to blood flow was 2-3 times greater than the predicted value. With continued use of the filter, resistance to blood flow remained largely unchanged. When, in the laboratory, the AN-69 capillary filter was perfused with saline and with a viscous sucrose solution, the resistance to flow was only 1.4 time the predicted value, a difference that might result from small deviations of the capillary diameter. When perfused with blood, the resistance was 2.6 times greater than the predicted value. This was largely explained by gross underestimation of blood viscosity in these patients. By combining laboratory data on filter resistance during saline perfusion and a more accurate estimation of blood viscosity, a reasonably accurate prediction of blood flow rate would be feasible. In the clinic the hydraulic permeability of the filters decreased with time.(ABSTRACT TRUNCATED AT 250 WORDS)