Академический Документы
Профессиональный Документы
Культура Документы
BACKGROUND AND OBJECTIVES dialysis vintage, be male, have lower SBP, lower changes in weight
A postdialytic increase in blood pressure (BP) is a recognized but often during HD, and receive more antihypertensive medications. These
an overlooked complication. The epidemiology and predisposing predictive factors were shown to be associated with an interaction
factors are still not well defined. We studied a large sample of Italian between weight loss and dialysis, suggesting a volume-related
dialysis patients to assess the prevalence of postdialysis hypertension mechanism in its pathogenesis. PDHYPER was also associated with
(PDHYPER), defined as any increase of systolic BP (SBP) >10 mm, Hg CV mortality.
above the predialysis value, the associated factors and its role in cardio-
vascular (CV) mortality. CONCLUSIONS
In our study on a large Italian cohort of dialysis patients, the preva-
PATIENTS AND METHODS lence of PDHYPER was higher than what was previously reported
In this observational study, we assessed dialysis associated changes in and is a significant risk factor for CV mortality in dialysis patients.
BP in 4,292 hemodialysis (HD) patients over 1 month (51,504 sessions). The pathogenesis is multifactorial but hypertensive state, antihyper-
We compared the clinical characteristics of the patients with stable tensive medications, and extracellular volume expansion appear to
BP values during the HD session with those with PDHYPER. We also play a major role.
assessed the impact of PDHYPER on CV mortality.
Keywords: blood pressure; hemodialysis; hypertension; postdialysis
RESULTS hypertension; volume expansion.
A total of 994 (23.1%) patients had PDHYPER. Patients with
PDHYPER were more likely to be hypertesive, older, have a shorter doi:10.1093/ajh/hpv162
An increase in blood pressure (BP) during hemodialysis (HD), This represents a potential bias since, in general, a single
although recognized for many years, is often overlooked.1 center has a homogeneous policy on the dialytic and medi-
There is also a lack of a unifying criteria for the diagnosis of cal management of patients. A step forward toward its better
postdialytic hypertension (PDHYPER), although an increase understanding can help reduce the high rate of cardiovascu-
in systolic BP (SBP) >10 mm Hg from pre- to postdialysis has lar (CV) morbidity occurring in HD patients.6
been proposed and is widely used.2 There is also a scarcity The aim of this large, multicenter, observational study is to
of systematic studies on its prevalence and consequences.3 investigate HD- and patient-related factors associated with
Research on the mechanisms causing PDHYPER has been PDHYPER and its impact on mortality.
addressed to both the HD treatment and patient character-
istics. The results obtained are often conflicting. Among the
possible pathogenetic mechanisms, ultrafiltration-induced PATIENTS AND METHODS
hypovolemia may trigger the renin–angiotensin system and Study design
induce vasoconstriction, which is mediated by an increase
in blood viscosity/hemoconcentration favored by higher The study is a retrospective analysis of data collected for a
hemoglobin (Hb) levels and/or treatment with erythropoiesis previous prospective study.7,8 In brief, patients were assessed
stimulating agents.4 The removal of antihypertensive drugs by at enrollment and were followed thereafter. CV mortality
dialysis is another possible cause of intradialytic hypertension. was drawn from death certificate and reported by participat-
Finally, electrolyte disequilibrium, sympathetic over activity, ing centers. The information requested from participating
and endothelial cell dysfunction may play a role.5 units on the cause of death were: death due to cerebrovascu-
Unfortunately, the majority of the available studies con- lar disease, ischemic or hemorrhagic, death due to any CV
sidered 1 single hypothetical factor at a time or were per- condition, comprising ischemic heart disease, heart failure,
formed on a small number of patients from a single center. and sudden death.
Continuous variables are presented as means and Compared to patients with no changes in BP (group
SDs unless otherwise specified. Categorical variables are 1), those with PDHYPER (group 2) were more likely to
Control, group 1: Δ SBP between −20 and + 10 mm Hg, group 2: Δ SBP > −20 mm Hg, group 3: Δ SBP > +10 mm Hg.
Abbreviations: SBP, systolic blood pressure; HD, hemodialysis; PP, pulse pressure; PTH, parathyroid hormone; ESA, erythropoiesis stimulat-
ing agent; PDHYPO, postdialysis hypotension; PDHYPER, postdialysis hypertension.
*P < 0.05 in the comparison with group 1.
Table 2. Multivariable regression model results for Δ SBP (mm Hg) in dialysis patients
Abbreviations: SBP, systolic blood pressure; PP, pulse pressure; HR, heart rate; Hb, hemoglobin; BMI, body mass index.
be older, have a shorter dialysis vintage, lower body mass analysis, using PDHYPER as a dichotomous depend-
index, lower Δ weight during HD, lower Hb, lower para- ent variable, produced 2 significant models (1 adjusted
thyroid hormone, lower predialysis SBP, and heart rate. and 1 unadjusted) (Table 3). For the unadjusted model,
Furthermore, the proportion of patients on antihyperten- the area under the receiver-operating characteristic curve
sive medications was higher, as well as the percentage of was = 0.729, correctly classified as 85.34%. The Hosmer
patients with an ejection fraction <50%. Logistic regression and Lemeshow χ2 was 4.91 with probabilities > χ2 =0.767.
Unadjusted Adjusted
Abbreviations: CI, confidence interval; ESA, erythropoiesis stimulating agent; SBP, systolic blood pressure; PP, pulse pressure; HR, heart
rate; HD, hemodialysis; PTH, parathyroid hormone.
For the adjusted model, the area under receiver-operating antihypertensive medications (RERI −1.793, CI −2.572 to
characteristic curve was = 0.732, correctly classified 85.53%. 1.015). These values indicate interaction on an additive scale,
The Hosmer and Lemeshow χ2 was 5.50, with probabilities meaning that the combined effect is less than the sum of the
>χ2 = 0.702. The tests indicated a good discrimination for factors examined singly.
both models. The use of antihypertensive medications, male
sex, increased predialysis pulse pressures, increased age, and Patients with heart failure
longer length of the HD session were significant predictors
of PDHYPER. Conversely, Δ weight during HD, erythro- The logistic regression performed in patients with heart
poiesis stimulating agent use, predialysis SBP, and predialy- failure showed the following predictors of PDHYPER: pre-
sis HR were predictors, inversely associated, of postdialysis dialysis SBP (OR 1.25, CI 1.17–1.34; P < 0.001), female sex
hypertension. Diabetes and heart failure were found not to (0.17, CI = 0.05–0.63; P = 0.008).
be predictors. The analysis, repeated after stratification by
tertiles of weight loss by dialysis, showed that the follow- Survival and changes in BP with dialysis
ing variables were significantly predictive: use of antihyper-
tensive medications, predialysis SBP, and predialysis pulse Patients were followed up for a mean of 26.8 ± 11.1 months;
pressures. Logistic regression was also carried out with 2 348 patients who were lost to follow-up were either trans-
different thresholds for PDHYPER (5 and 15 mm Hg). With planted or transferred to different units. On the basis of the
the 5 mm Hg threshold, found in 28.7% of the patients, the results of the above logistic regression, we excluded from
results were concordant with those obtained with the 10 mm the analysis subjects with heart failure. We had data for CV
Hg threshold. Raising the threshold to 15 mm Hg, 10.5% of mortality in the remaining 3,196 patients, among whom
the patients presented PDHYPER. In the analysis, only age, 420 deaths were recorded. Of those, 318 were ascribed to
Δ weight, and heart rate were significant predictors. a heart disease and 102 to a cerebrovascular event. The
Cox analysis produced a significant model with the fol-
Interaction between factors lowing predictors of CV mortality: age (HR 1.05, CI 1.03–
1.07; P < 0.001), dialysis vintage (HR 1.01, CI 1.00–1.02;
We found a significant interaction between predialysis P = 0.038), PDHYPER (HR 1.69, CI 1.12–2.57; P = 0.012)
SBP and Δ weight during HD (RERI −2.830, confidence (Figure 1). Predialysis SBP (> or <140 mm Hg) was not a
interval (CI) −3.088 to 2.573), between Δ weight during HD predictor of mortality. Apart from hypertension, no other
and the use of antihypertensive medications (RERI −1.881, association was found between the etiology of kidney dis-
CI −2.631 to 1.131) and between predialysis SBP and use of ease and PHYPER or mortality.
Figure 1. Cox analysis in patients without heart failure. CV mortality plot. Controls are patients with a postdialysis SBP Δ comprised between −20 and
+ 10 mm Hg, PDHYPER: Δ SBP > +10 mm Hg, PDHYPO: Δ SBP > −20 mm Hg. Only PDHYPER shows a significant difference from controls. Abbreviations:
PDHYPO, postdialysis hypotension; PDHYPER, postdialysis hypertension.
The prevalence of heart failure in dialysis patients is high 3. Chou KJ, Lee PT, Chen CL, Chiou CW, Hsu CY, Chung HM, Liu CP,
and the prognosis is ominous, especially in the presence of Fang HC. Physiological changes during hemodialysis in patients with
intradialysis hypertension. Kidney Int 2006; 69:1833–1838.
hypertension.20In patients with heart failure in our study, the 4. Inrig JK. Intradialytic hypertension: a less-recognized cardiovascular
factors associated with PDHYPER were not the same as the complication of hemodialysis. Am J Kidney Dis 2010; 55:580–589.
ones we found in the others. Interestingly, Δ weight was not a 5. Inrig JK, Van Buren P, Kim C, Vongpatanasin W, Povsic TJ, Toto RD.
predictor, indicating a more complex mechanism operating Intradialytic hypertension and its association with endothelial cell dys-
function. Clin J Am Soc Nephrol 2011; 6:2016–2024.
in this subset of patients. Predialysis SBP was a more sig- 6. Chazot C, Jean G. Intradialytic hypertension: it is time to act. Nephron
nificant predictor. These findings show that factors associ- Clin Pract 2010; 115:c182–c188.
ated with changes in BP during dialysis vary according to the 7. Del Vecchio L, Lusenti T, Del Rosso G, Malandra R, Balducci A,
individual characteristics of patients. Losito A; Group For Arterial Hypertension of the Italian Society Of
Our study has a number of limitations. We did not per- Nephrology. Prevalence of hypertension in a large cohort of Italian
hemodialysis patients: results of a cross-sectional study. J Nephrol 2013;
form a direct examination of extracellular volume.21 Thus, 26:745–754.
we can give only indirect evidence about the causative effect 8. Losito A, Del Vecchio L, Lusenti T, Del Rosso G, Malandra R, Sturani
of volume expansion on PDHYPER and we cannot rule out A; Gruppo di Studio Dell’Ipertensione Arteriosa Società Italiana
other contributing factors, such as dialysate sodium concen- Nefrologia. Systolic blood pressure and mortality in chronic hemodi-
alysis patients: results of a nationwide Italian study. J Clin Hypertens
tration, renin–angiotensin–aldosterone system, and sympa- (Greenwich) 2013; 15:328–332.
thetic activation. We did not collect information on the class 9. K/DOQI Workgroup. K/DOQI clinical practice guidelines for cardio-
and number of antihypertensive medications. Thus, we can vascular disease in dialysis patients. Am J Kidney Dis 2005; 45(Suppl
just speculate around medication removal by dialysis. On 3):S1–S153.
the other hand, the large number of enrolled patients and 10. K/DOQI Clinical Practice Guidelines for Cardiovascular Disease in
Dialysis Patients. Intradialytic hypotension. Am J Kidney Dis 2005;
collected variables, together with the homogeneity of the 45(Suppl 3):S76–S80.
population, strengthen the validity of the results obtained. 11. Rothman, KJ: Measuring interactions. Epidemiology: An Introduction.
In conclusion, our multicenter study shows that the preva- Oxford University Press: New York, 2002, pp. 168–180.
lence of PDHYPER is higher than that previously reported 12. Andersson T, Alfredsson L, Källberg H, Zdravkovic S, Ahlbom A.
Calculating measures of biological interaction. Eur J Epidemiol 2005;
and that it is associated with high CV mortality. 20:575–579.
13. Davenport A. Intradialytic complications during hemodialysis.
Hemodial Int 2006; 10:162–167.
14. Fellner SK. Intradialytic hypertension: II. Semin Dial 1993; 6: 371–373.
15. Agarwal R, Andersen MJ, Pratt JH. On the importance of pedal edema
ACKNOWLEDGMENTS in hemodialysis patients. Clin J Am Soc Nephrol 2008; 3:153–158.
16. Cirit M, Akçiçek F, Terzioğlu E, Soydaş C, Ok E, Ozbaşli CF, Başçi A,
The study was brought about on behalf of the of hyperten- Mees EJ. ‘Paradoxical’ rise in blood pressure during ultrafiltration in
sion study group of The Italian Society of Nephrology. dialysis patients. Nephrol Dial Transplant 1995; 10:1417–1420.
17. Agarwal R, Light RP. Intradialytic hypertension is a marker of volume
excess. Nephrol Dial Transplant 2010; 25:3355–3361.
DISCLOSURE 18. Agarwal R, Alborzi P, Satyan S, Light RP. Dry-weight reduction in
hypertensive hemodialysis patients (DRIP): a randomized, controlled
The authors declared no conflict of interest. trial. Hypertension 2009; 53:500–507.
19. Nongnuch A, Campbell N, Stern E, El-Kateb S, Fuentes L, Davenport A.
Increased postdialysis systolic blood pressure is associated with extra-
cellular overhydration in hemodialysis outpatients. Kidney Int 2015;
87:452–457.
REFERENCES 20. Losito A, Del Vecchio L, Del Rosso G, Malandra R; Gruppo di Studio
Dell’Ipertensione Arteriosa, Società Italiana di Nefrologia. Blood pres-
1. Levin N. Intradialytic hypertension I. Semin Dial 1993; 6:370–371. sure and cardiovascular mortality in dialysis patients with left ventricu-
2. Inrig JK, Oddone EZ, Hasselblad V, Gillespie B, Patel UD, Reddan D, lar systolic dysfunction. Am J Hypertens 2014; 27:401–408.
Toto R, Himmelfarb J, Winchester JF, Stivelman J, Lindsay RM, Szczech 21. Locatelli F, Cavalli A, Tucci B, Viganò S, Di Filippo S. Can chronic
LA. Association of intradialytic blood pressure changes with hospitali- volume overload be recognized and prevented in hemodialysis
zation and mortality rates in prevalent ESRD patients. Kidney Int 2007; patients? Newer methods to assess volume status. Semin Dial 2009;
71:454–461. 22:492–494.