Original Research Received: September 6, 2008 Accepted after revision: March 21, 2009 Published online: June 12, 2009
Cardiology 2009;114:142–149 DOI: 10.1159/000224776
Left-Ventricular Diastolic Dysfunction as © Free Author a Risk Factor for Dialytic Hypotension Copy – for personal use only
a, b Abbès Benmaadi a Guy Rostoker a, b Mireille Griuncelli a, b Christelle Loridon ANY DISTRIBUTION OF THIS c ARTICLE WITHOUT WRITTEN Eric Illouz
CONSENT FROM S. KARGER c BASEL IS A VIOLATION Division of Nephrology and Dialysis, b Laboratory of Clinical Investigation, and AG, Division of Cardiology, OF THE COPYRIGHT. Centre Hospitalier Privé Claude Galien, Quincy-sous-Sénart, France a
Written permission to distribute the PDF will be granted against payment of a permission fee, which is based on the number of accesses required. Please contact
[email protected]
Key Words Hemodialysis ⴢ Dialysis hypotension ⴢ Brain natriuretic peptide ⴢ Diastolic dysfunction ⴢ Diastolic flow propagation velocity index
Abstract Objectives: Intradialytic hypotension may adversely affect the outcome of chronic hemodialysis and thus reduce the patients’ life expectancy. The aim of this study was to assess the link between left-ventricular diastolic dysfunction and dialytic hypotension. Methods: We performed a prospective cross-sectional study of 72 hemodialysis patients with a low dialysis vintage, 36 of whom had dialysis hypotension, based on echocardiography and brain natriuretic peptide (BNP) assay. Results: There was no difference between normotensive patients and those with dialysis-associated chronic hypotension as regards BNP level, cardiac index, left-ventricular ejection fraction, or myocardial fractional shortening. Both hypotension-prone patients requiring dialysate sodium profiling and chronic refractory hypotensive patients requiring macromolecule infusion had cardiac diastolic dysfunction as shown by a similarly abnormal E/A ratio !1 in 89–91% of cases, associated with a significant decrease in color M-mode diastolic flow propagation velocity (Vp, p ! 0.05 nonpara-
© 2009 S. Karger AG, Basel 0008–6312/09/1142–0142$26.00/0 Fax +41 61 306 12 34 E-Mail
[email protected] www.karger.com
Accessible online at: www.karger.com/crd
metric ANOVA). The area under the ROC curve for Vp was 0.69. A Vp cutoff of 39.5 cm/s was optimal for predicting dialysis-associated hypotension. Conclusions: We conclude that diastolic dysfunction is associated with dialytic hypotension and that a low Vp – a preload-independent index – © Free Author Copy – for personal use is predictive of dialysis-associated hypotension.
only
ANY DISTRIBUTION OF THIS ARTICLE WITHOUT WRITTEN CONSENT FROM S. KARGER AG, BASEL IS Copyright © 2009 S. Karger AG, Basel Written permission to distribute the PDF will be granted against payment of a permission fee, which
Introduction
Occurring in up to 33% of patients, intradialytic hypotension is the most common complication of hemodialysis. There are two main clinical patterns of dialysis-associated hypotension: the first is episodic hypotension that typically occurs during the later stages of dialysis sessions and is generally favored by excessive weight gain, and the second is chronic persistent hypotension which affects about 10% of long-term dialysis patients [1–3]. Most of these latter patients experience frequent episodes of hypotension during dialysis sessions, whereas some patients have permanent hypotension with low predialysis systolic pressure (often less than 100 mm Hg) [4, 5]. Intradialytic hypotension not only causes discomfort and has a negative impact on health-related quality of life, but Dr. Guy Rostoker Service de Néphrologie et de Dialyse Centre Hospitalier Privé Claude Galien, 20 route de Boussy Saint-Antoine FR–91480 Quincy-sous-Sénart (France) Tel. +33 1 6939 9200, Fax +33 1 6939 9184, E-Mail
[email protected]
it may also adversely affect the outcome of chronic hemodialysis and thus reduce the patients’ life expectancy and favor underdialysis [6–11]. The immediate cause of intradialytic hypotension is acute central hypovolemia [3, 5]. Frank hypotension occurs when cardiovascular mechanisms do not adequately compensate for the blood volume reduction that results from the imbalance between the ultrafiltration rate and the plasma refilling rate [12]. Several factors contribute to dialytic hypotension. These include too rapid fluid removal in an attempt to reach dry weight, a rapid reduction in plasma osmolality (which causes extracellular water to move into cells), high interdialytic weight gain, anemia, poor nutritional status with hypoalbuminemia, autonomic neuropathy, anephric status, reduced pressor response to vasopressor agents, reduced cardiac reserve, increased arterial stiffness, impaired venous compliance, use of acetate rather than bicarbonate as dialysate buffer, ingestion of a meal immediately before or during the dialysis session, use of low sodium or high magnesium concentrations in the dialysate, and intake of anti-hypertensive medications before the dialysis session, which can impair cardiovascular stability (especially nitrate derivatives) [1–5, 12]. Excessive release of several endogenous vasodilatators such as nitric oxide, adrenomedullin and adenosine has been implicated in the pathogenesis of dialytic hypotension, together with an imbalance in the synthesis of the endogenous vasocontrictors endothelin and vasopressin [13–16]. Management of intradialytic hypotension involves treating the acute episode and applying measures to prevent future episodes [1, 2, 5]. Standard measures to prevent or alleviate intradialytic hypotension include accurate setting of the dry weight, avoidance of modifiable factors known to favor this phenomenon (see above), adjustment of the dialysate sodium and/or calcium concentration and temperature (cool dialysate and isothermic dialysis), use of initial ultrafiltration followed by standard dialysis, use of sodium dialysate and ultrafiltration modeling [17, 18]. Diastolic dysfunction refers to abnormal mechanical properties of the myocardium and includes abnormal left-ventricular distensibility, impaired filling and slow or delayed relaxation, regardless of whether the ejection fraction is normal or depressed [19, 20]. The term asymptomatic diastolic dysfunction is used to refer to an asymptomatic patient with a normal ejection fraction and an abnormal echo-Doppler pattern of left-ventricular filling, whereas the term diastolic heart failure is used if such a patient exhibits symptoms of effort intolerance and dyspnea, especially if there is evidence of venous con-
gestion and edema [19, 20]. The term diastolic heart failure is synonymous with heart failure with a preserved ejection fraction (150%) [19, 20]. Patients with diastolic heart failure have recently been shown by cardiac catheterization to have incomplete left-ventricular relaxation and increased left-ventricular chamber stiffness [21]. Left-ventricular diastolic dysfunction and diastolic heart failure are associated with aging and diabetes mellitus and cardiovascular conditions such as hypertensive heart disease, ischemic and valvular heart diseases, and more rarely restrictive cardiomyopathy [19, 20, 22]. The prevalence of diastolic heart failure was recently shown to have increased in the general population over the last 15 years while the associated mortality rate has remained unchanged [23]. Genetic variants in the renin-angiotensin system may determine an individual risk of developing diastolic heart failure [24]. A recent controlled trial failed to show an improvement in the outcome of patients with diastolic heart failure treated with irbesartan [25]. We postulated that changes in the epidemiology of dialysis in the last decade may have played a pivotal role in the pathogenesis of dialysis-associated hypotension through left-ventricular diastolic dysfunction and diastolic heart failure [19, 20, 26–28]. Because diastolic function can now be studied noninvasively by means of diastolic flow propagation velocity (Vp) measurement by tissue Doppler echocardiography – a new preload-independent index (in contrast to the E/A ratio and mitral annular early diastolic velocities) [29–31] – we conducted a prospective cross-sectional study of 72 hemodialysis patients with a short dialysis vintage to assess the link between left-ventricular diastolic dysfunction and dialytic hypotension.
Diastolic Dysfunction in Dialytic Hypotension
Cardiology 2009;114:142–149
Materials and Methods Working Hypothesis We postulated that changes in the epidemiology of dialysis in the last decade such as the increasing number of elderly and diabetic patients and patients with cardiovascular disease who are now starting hemodialysis may have played a pivotal role in the pathogenesis of dialysis-associated hypotension through left-ventricular diastolic dysfunction and diastolic heart failure, which are associated with aging and these pathological states [19, 20, 26–28]. Patients and Dialysis After obtaining the patients’ informed consent and approval from the ethics committee of Henri Mondor University Hospital (approval No. 02-017), 72 patients undergoing chronic intermittent bipuncture bicarbonate hemodialysis 3 times per week on a
143
Cimino-Brescia fistula (n = 60) or a Canaud central venous catheter (n = 12) were enrolled in this study. Eighteen patients had hypertensive nephritis (nephroangiosclerosis), 18 patients had diabetes mellitus nephritis, 10 had hereditary nephritis (6 polycystic renal disease, 3 Alport syndrome, 1 interstitial hereditary nephritis), 10 had inherited or acquired urologic abnormalities of the kidneys or interstitial nephritis, 9 had primary glomerulonephritis (6 IgA nephropathy, 2 segmental glomerulosclerosis, 1 membranous nephropathy), and 7 had systemic diseases (2 lupus erythematosus, 2 primary sicca syndrome, 2 multiple myeloma with AL amyloidosis, 1 Wegener syndrome). Group 1 consisted of 36 hemodialysis patients without hypotension (median age 70 years; range 34–85), 15 women/21 men, median history of dialysis 3 years (range 1–9 years). Group 2 comprised 17 hypotension-prone hemodialysis patients who successfully received cool dialysate (35–35.5 ° C), sodium dialysate and ultrafiltration profiling (median age 66 years; range 24–92), 6 women/11 men, median history of dialysis 3 years (range 1–17 years). Group 3 consisted of 19 hemodialysis patients with permanent hypotension requiring systematic administration of colloids (200 ml of 4% gelatin or 200 ml of 20% albumin) during dialysis sessions, plus cool dialysate, sodium dialysate and ultrafiltration profiling [32] (median age 74 years; range 37–93), 12 women/7 men, median history of dialysis 4 years (range 1–10 years). Dialysis sessions were performed without dialyser reuse (forbidden in France). All the patients in groups 2 and 3 had experienced hypotensive episodes more than once a week, as defined by a systolic pressure !100 mm Hg or a fall in systolic pressure 130 mm Hg associated with symptoms of hypotension (dizziness, nausea, vomiting, headache, sweating, loss of consciousness) before measures to prevent future episodes of dialytic hypotension were taken (see above). Dry weight was estimated by echography of the inferior vena cava, echocardiographic estimation of enddiastolic left-ventricular size 2 h after a dialysis session or the morning following an evening dialysis session [33, 34] and by physical evaluation. All the patients were at their optimal dry weight as shown by normovolemia on echography of the inferior vena cava and echocardiography, and by the lack of symptoms of hypovolemia or hypervolemia. Hemodialysis was performed with an Integra module (Hospal, Lyon, France) with ultrapure dialysate (Gambro, Colombes, France). The ultrafiltration rate was prescribed according to the estimated dry weight and interdialytic weight gain; ultrafiltration profiles were associated with sodium profiling. The blood flow rate was 250–300 ml/min and the dialysate flow rate was 500 ml/min. The composition of the 505A dialysate from Fresenius (Fresnes, France) was as follows: 2.0 mmol/l potassium, 3.0 mmol/l acetate, 0.5 mmol/l magnesium, 37 mmol/l bicarbonate, 1.75 mmol/l calcium and 1.0 g/l (5.6 mmol/l) glucose. Noninvasive Hemodynamic Study We conducted a prospective noninvasive cardiac hemodynamic cross-sectional study to identify predictors of dialytic hypotension. We used echocardiography and the brain natriuretic peptide (BNP) assay on EDTA plasma (AxSYM BNP-MEIA; Abbot, Rungis, France). Echocardiography was performed by the same cardiologist (E.I.), who was unaware of the patient’s dialysis modalities and of hemodynamic stability during the dialysis sessions. Standard two-dimensional M-mode echocardiography images and Doppler echocardiography of mitral flow were obtained
144
Cardiology 2009;114:142–149
2 h after the end of the dialysis session for morning and afternoon sessions and the morning after evening hemodialysis sessions, using a Hewlett Packard Sonos 5500 machine (Hewlett Packard, Issy-les-Moulineaux, France). Echocardiography was performed at least 2 h after the dialysis session to avoid spurious results related to underestimation of the degree of diastolic dysfunction seen before the hemodialysis session in relation to weight gain and volume overload as recommended by Ie et al. [35]. Wall thickness and left-ventricular diameter were derived from M-mode echocardiograms as recommended by the American Society of Echocardiography [36]. Doppler recordings of mitral inflow were performed in the apical 4-chamber view to determine the early diastolic peak velocity, late atrial diastolic peak velocity and deceleration time [37]. Vp was determined according Garcia et al. [29, 30] by measuring the slope of the first aliasing velocity from the mitral tips to a position 4 cm distal in the left ventricle automatically by tissue Doppler echocardiography and specific software on the Hewlett Packard Sonos 5500. We chose the method of Garcia et al. [29, 30] because of its higher reproducibility. Left-ventricular mass was calculated by the method of Devereux and Reichek [38]. The left-ventricular mass index (LVMI) was calculated by dividing left-ventricular mass by the body surface area. Left-ventricular hypertophy was defined as an LVMI of more than 130 g/m2 in men and more than 100 g/m2 in women [38]. Diastolic function was estimated by measuring the E/A ratio and Vp [29]. Statistical Analyses Echocardiography parameters and BNP levels in the different groups of hemodialysis patients were analyzed by parametric or nonparametric analysis of variance (ANOVA) for continuous variables, followed by post-tests depending on the normality of the distribution (in the Kolmogorov-Smirnov test) and by using the 2 test for categorial variables [39] (Instat 3 software, Graphpad, San Diego, Calif., USA). p ! 0.05 was considered significant [39]. Values are expressed as means 8 SD or medians and ranges, depending on the normality of the distribution [39]. We analyzed the interaction and contribution of several variables to the Vp by multiple regression analysis and a correlation matrix (Instat 3 software, Graphpad) [40]. We also calculated the sensitivity, specificity, positive predictive value, negative predictive value and the likelihood ratio of a Vp !55 cm/S for predicting dialysis-associated chronic hypotension by pooling data for groups 2 and 3 (Instat 3 software, Graphpad) [40]. Finally, we utilized receiver-operator characteristic (ROC) curves to examine the ability of Vp to predict dialysis-associated hypotension. ROC plots simultaneously showing the sensitivity and specificity of Vp were generated over the whole spectrum of test results, which enabled us to define the best Vp threshold separating dialysis patients without hemodynamic instability during dialysis sessions from those with dialysis-associated hypotension [40] (Prism 4 software, Graphpad). Sponsors The study and the ethics committee were sponsored by the nonprofit Physicians’ Association of Claude Galien Hospital (Association Quincy Recherche Clinique et Thérapeutique). The aim of this association is to promote clinical and cognitive research at Claude Galien Hospital.
Rostoker /Griuncelli /Loridon /Benmaadi / Illouz
Table 1. Noninvasive hemodynamic study of hypotension-prone dialysis patients
BNP, pg/ml Cardiac index, liters/min/m2 Left-ventricular ejection fraction, % Myocardial fractional shortening, % E/A ratio
