The Epidemiology of Systolic Blood Pressure and Death Risk in Hemodialysis Patients
2006; Elsevier BV; Volume: 48; Issue: 4 Linguagem: Inglês
10.1053/j.ajkd.2006.07.005
ISSN1523-6838
AutoresZhensheng Li, Eduardo Lacson, Edmund G. Lowrie, Norma J. Ofsthun, Martin K. Kuhlmann, J. Michael Lazarus, Nathan W. Levin,
Tópico(s)Heart Failure Treatment and Management
ResumoBackground: This study compares the associations of predialysis systolic blood pressure (SBP) with mortality risk in both incident and prevalent hemodialysis (HD) cohorts by using both conventional and time-varying Cox analyses, thus addressing limitations of prior studies. Methods: A total of 56,338 incident patients starting HD therapy during 1997 to 2001 and 69,590 prevalent HD patients on January 1, 2002, were grouped into the following categories: (1) SBP less than 120 mm Hg, (2) 120 ≤ SBP < 140 mm Hg, (3) 140 ≤ SBP < 160 mm Hg, (4) 160 ≤ SBP < 180 mm Hg, (5) 180 ≤ SBP < 200 mm Hg, and (6) SBP of 200 mm Hg or greater. Conventional and time-varying models evaluated 1-year and 3-year (incident patients only) survival. Results: Nine percent and 26.0% of incident patients and 5.7% and 20.1% of prevalent patients were in categories 1 and 2, respectively. Their associated 1-year hazard ratios (HRs) were 2.63 to 3.68 and 1.57 to 1.68 compared with category 4, the reference group. HRs for categories 3, 5, and 6 were not different from category 4. Time-varying models magnified category 1 and 2 HRs to 5.54 to 7.42 and 1.92 to 2.21, such that 25% to 35% of patients in the target SBP range (<140 mm Hg) had the greatest risk. A "reversed J-shaped" risk profile emerged in the time-varying models, with very high SBP (category 6) associated with HRs of 1.52 to 1.55, but only 1% of patients were in category 6. Three-year outcomes were similar. Conclusion: Epidemiological characteristics of predialysis SBP consistently differ from those in the general population despite different analytic perspectives. The data suggest a need for greater investigative, diagnostic, and therapeutic focus on HD patients with normal and prehypertensive blood pressure ranges. Background: This study compares the associations of predialysis systolic blood pressure (SBP) with mortality risk in both incident and prevalent hemodialysis (HD) cohorts by using both conventional and time-varying Cox analyses, thus addressing limitations of prior studies. Methods: A total of 56,338 incident patients starting HD therapy during 1997 to 2001 and 69,590 prevalent HD patients on January 1, 2002, were grouped into the following categories: (1) SBP less than 120 mm Hg, (2) 120 ≤ SBP < 140 mm Hg, (3) 140 ≤ SBP < 160 mm Hg, (4) 160 ≤ SBP < 180 mm Hg, (5) 180 ≤ SBP < 200 mm Hg, and (6) SBP of 200 mm Hg or greater. Conventional and time-varying models evaluated 1-year and 3-year (incident patients only) survival. Results: Nine percent and 26.0% of incident patients and 5.7% and 20.1% of prevalent patients were in categories 1 and 2, respectively. Their associated 1-year hazard ratios (HRs) were 2.63 to 3.68 and 1.57 to 1.68 compared with category 4, the reference group. HRs for categories 3, 5, and 6 were not different from category 4. Time-varying models magnified category 1 and 2 HRs to 5.54 to 7.42 and 1.92 to 2.21, such that 25% to 35% of patients in the target SBP range (<140 mm Hg) had the greatest risk. A "reversed J-shaped" risk profile emerged in the time-varying models, with very high SBP (category 6) associated with HRs of 1.52 to 1.55, but only 1% of patients were in category 6. Three-year outcomes were similar. Conclusion: Epidemiological characteristics of predialysis SBP consistently differ from those in the general population despite different analytic perspectives. The data suggest a need for greater investigative, diagnostic, and therapeutic focus on HD patients with normal and prehypertensive blood pressure ranges. HYPERTENSION USUALLY BEGINS in the early stages of renal disease and is still prevalent in up to 80% of hemodialysis (HD) patients.1US Renal Data SystemUSRDS 2003 Annual Data Report. The National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD2003Google Scholar, 2Rocco M.V. Yan G. Heyka R.J. et al.Risk factors for hypertension in chronic hemodialysis patients: Baseline data from the HEMO Study.Am J Nephrol. 2001; 21: 280-288Crossref PubMed Scopus (103) Google Scholar, 3Hyman D.J. Pavlik V.N. Characteristics of patients with uncontrolled hypertension in the United States.N Engl J Med. 2001; 345: 479-486Crossref PubMed Scopus (691) Google Scholar, 4Buckalew Jr, V.M. Berg R.L. Wang S.R. et al.Prevalence of hypertension in 1,795 subjects with chronic renal disease: The Modification of Diet in Renal Disease Study baseline cohort.Am J Kidney Dis. 1996; 6: 811-821Abstract Full Text PDF Scopus (205) Google Scholar However, the association of greater mortal risk in patients with "normal" blood pressure (BP) than high BP was observed more than a decade ago.5Lowrie E.G. Huang W.H. Lew N.L. Liu Y. The relative contribution of measured variables to death risk among hemodialysis patients.in: Friedman E. Death on Hemodialysis: Preventable or Inevitable? Kluwer, Hingham, MA1994: 121-141Crossref Google Scholar More recent reports confirmed the association.6Salem M.M. Bower J. Hypertension in the hemodialysis population: Any relation to one-year survival?.Am J Kidney Dis. 1996; 28: 737-740Abstract Full Text PDF PubMed Scopus (84) Google Scholar, 7Zager P.G. Nikolic J. Brown R.H. et al."U" curve association of blood pressure and mortality in hemodialysis patients Medical Directors of Dialysis Clinic, Inc.Kidney Int. 1998; 54: 561-569Crossref PubMed Scopus (590) Google Scholar, 8Salem M.M. Hypertension in the hemodialysis population: Any relationship to 2-years survival?.Nephrol Dial Transplant. 1999; 14: 125-128Crossref PubMed Scopus (76) Google Scholar, 9Klassen P.S. Lowrie E.G. Reddan D.N. et al.Association between pulse pressure and mortality in patients undergoing maintenance hemodialysis.JAMA. 2002; 287: 1548-1555Crossref PubMed Scopus (359) Google Scholar, 10Kalantar-Zadeh K. Block G. Humphreys M.H. Kopple J.D. Reverse epidemiology of cardiovascular risk factors in maintenance dialysis patients.Kidney Int. 2003; 63: 793-808Crossref PubMed Scopus (980) Google Scholar The term reverse epidemiology10Kalantar-Zadeh K. Block G. Humphreys M.H. Kopple J.D. Reverse epidemiology of cardiovascular risk factors in maintenance dialysis patients.Kidney Int. 2003; 63: 793-808Crossref PubMed Scopus (980) Google Scholar has been used to describe this counterintuitive phenomenon. Those findings recently were questioned on methodological grounds,11Agarwal R. Hypertension and survival in chronic hemodialysis patients—Past lessons and future opportunities.Kidney Int. 2005; 67: 1-13Crossref PubMed Scopus (139) Google Scholar and current clinical guidelines for BP management in HD patients address only the management of BP exceeding 140/90 mm Hg.12National Kidney Foundation: K/DOQI Clinical Practice Guidelines for Cardiovascular Disease in Dialysis.Am J Kidney Dis. 2005; 45: S49-S57Google ScholarThe Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure defined a 4-level classification for systolic BP (SBP) in the general population as follows: normal (SBP < 120 mm Hg), prehypertension (120 ≤ SBP < 140 mm Hg), stage 1 hypertension (140 ≤ SBP < 160 mm Hg), and stage 2 hypertension (SBP ≥ 160 mm Hg).13Chobanian A.V. Bakris G.L. Black H.R. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The JNC 7 Report.JAMA. 2003; 289: 2560-2571Crossref PubMed Scopus (16370) Google Scholar The practice guideline for HD patients therefore requires SBP to be in the prehypertensive range or less.The guidelines,12National Kidney Foundation: K/DOQI Clinical Practice Guidelines for Cardiovascular Disease in Dialysis.Am J Kidney Dis. 2005; 45: S49-S57Google Scholar lack of understanding about the meaning of normal SBP in HD patients, and questioning of prior literature11Agarwal R. Hypertension and survival in chronic hemodialysis patients—Past lessons and future opportunities.Kidney Int. 2005; 67: 1-13Crossref PubMed Scopus (139) Google Scholar prompted us to reevaluate the epidemiological associations of SBP with death risk in HD patients by using a combination of patient sampling and analytic strategies. A clear understanding of those associations could lay foundations for future physiological studies, randomized trials comparing specific cardiovascular and BP management strategies, and, perhaps, changes in clinical practice.MethodsTwo SamplesIncident and prevalent patient samples were obtained from the Fresenius Medical Care, North America (FMCNA) files. The incident sample included 56,338 adult (age ≥ 18 years at first HD treatment) patients who started HD therapy during 1997 through 2001 and continued to be treated in an FMCNA clinic for at least 30 days, with at least 1 recorded SBP measurement. The prevalent sample included 69,590 adult patients receiving in-center HD treatment on January 1, 2002, with at least 1 predialysis SBP measurement within the prior 3 months. A total of 22,118 patients (33.7%) included in the incident patient cohort survived long enough to also contribute data to the prevalent patient cohort. However, for the majority of these patients, SBP values used in prevalent analyses did not overlap with the earlier SBP values used in incident cohort analyses.Follow-UpIncident patients were followed up until discharge or death or were censored at 3 years (the absolute last day of follow-up was December 31, 2004, for patients enrolled in December 31, 2001). Prevalent patients were followed up to either discharge from an FMCNA facility, death, or December 31, 2002. Survival time in days for conventional Cox models is defined as the date observation ended minus the date it started (after the initial 30 days for incident cohort and beginning January 1, 2002, for the prevalent cohort). Survival time was transformed to weeks for time-varying Cox models by dividing survival time in days by 7.Other VariablesAge at the first HD date (incident patients) or age on January 1, 2002 (prevalent patients), sex, diabetic status, and race (white/black/other) were used as case-mix measures. All BP measurements during both the initial and follow-up periods were collected. BP measurements usually were obtained by using the standard module in the Fresenius 2008H or 2008K dialysis machines (Fresenius USA, Walnut Creek, CA). Predialysis BP was recorded once, with the patient sitting in the dialysis chair, before puncturing the dialysis access and/or connecting to the extracorporeal dialysis circuit. Median values for predialysis SBP were calculated from the first month of dialysis treatment for incident patients and from October through December 2001 for prevalent patients. Those median values were used to classify patients into 6 categories, expanding the Joint National Committee Report13Chobanian A.V. Bakris G.L. Black H.R. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The JNC 7 Report.JAMA. 2003; 289: 2560-2571Crossref PubMed Scopus (16370) Google Scholar groups: category 1, SBP less than 120 mm Hg (normal); category 2, 120 ≤ SBP < 140 mm Hg (prehypertensive); category 3, 140 ≤ SBP < 160 mm Hg (stage 1 hypertension); category 4, 160 ≤ SBP < 180 mm Hg (stage 2); category 5, 180 ≤ SBP < 200 mm Hg (stage 2+); and category 6, SBP of 200 mm Hg or greater (stage 2++). Weekly predialysis SBP median values were used to classify patients into SBP categories for the time-varying models. Medians were chosen a priori because the researchers did not have a preliminary expectation about the distribution of SBP in the study population (although mean and median values turned out to be nearly identical, as listed in Table 1). Similarly, initial and weekly body surface areas (BSAs; in square meters) were calculated by using the DuBois equation14Kannel W.B. Cupples A. Epidemiology and risk profile of cardiac failure.Cardiovasc Drugs Ther. 1988; 2: S387-S395Crossref PubMed Scopus (31) Google Scholar with postdialysis weight (kilograms) and height (centimeters). Median values for predialysis diastolic BP, mean BP, pulse pressure, and postdialysis measurements obtained upon study initiation are reported for descriptive purposes only. For prevalent patients, vintage is the time between starting HD therapy to January 1, 2002. Incident patients were subclassified by year of admission to dialysis treatment.Table 1Distributions of MeasuresIncident Patient Sample (N = 56,338)Prevalent Patient Sample (N = 69,590)VariablesMeanSD25thMedian75thMeanSD25thMedian75thPredialysis SBP (mm Hg)147.421.4133147161153.121.2139153168 Diastolic BP (mm Hg)75.512.067758379.212.3718087 Mean BP (mm Hg)123.417.3112123135128.417.4117129140 Pulse pressure (mm Hg)71.815.261718173.814.6647384Postdialysis SBP (mm Hg)146.221.7130145160138.019.9124137151 Diastolic BP (mm Hg)74.711.767748271.511.1647179 Mean BP (mm Hg)122.317.4110121133115.916.2104115126 Pulse pressure (mm Hg)71.615.760708266.513.5576575Weight loss (kg)2.031.411.301.902.652.891.212.102.803.60Age (y)62.215.152657460.915.1506273BSA (m2)1.820.261.651.811.971.830.251.661.821.99Vintage (y)3.43.61.02.34.5Women (%)47.947.3Diabetes (%)52.750.8Race (%) White61.451.3 Black32.441.6 Other6.27.1NOTE. For categorical variables, mean indicates percentage of patients (excluding missing data); 25th, median (50th), and 75th indicate percentiles. Age is age at first HD treatment ever (incident) or on January 1, 2002 (prevalent). BSA is calculated from postdialysis weight and height. Open table in a new tab Survival StatisticsSurvival curves and Cox analyses were performed using SAS 9.1 procedure LIFETEST and PHREG (SAS Institute, Cary, NC) in addition to standard descriptive analysis. Conventional models evaluated death in terms of initial SBP category. Time-varying analyses used weekly SBP category in association with concurrent death(s) or censor event(s). Both analytic strategies were used for 1-year survival analyses in prevalent and incident cohorts and 3-year survival analyses for incident patients. Three levels of statistical adjustment were used for all analyses: unadjusted, case-mix adjusted, and case-mix plus BSA adjusted.SBP Change During Follow-UpFindings from primary survival analyses led us to perform secondary analyses comparing change in SBP between patients who survived and died. The difference between the first and last SBP measurement during the full year for survivors and up to the time of death for nonsurvivors, grouped by initial SBP category, was evaluated.ResultsResults are presented in 6 sections. The first section describes distributions of relevant measures in the 2 samples. Predialysis SBP risk profiles for the incident and prevalent samples (and subsets) using both conventional and time-varying Cox models are described in the next 4 sections. The last section describes the evaluation of predialysis SBP change in both incident and prevalent patient samples.Patient DescriptionsTable 1 lists measurement distributions in both samples. Mean ages were 61 (prevalent) and 62 years (incident), likely reflecting a survival advantage for younger age. Similarly, the greater percentage of black race in prevalent patients (41.6%) compared with incident patients (32.4%) may reflect lower mortality associated with black race. Incident patients had a lower average SBP than prevalent patients. Averages of predialysis diastolic BPs were 79.2 and 75.5 mm Hg in prevalent and incident samples, respectively. Other BP measures, demographics, and weight loss during dialysis also are listed in Table 1 for descriptive purposes.SBP correlated highly with diastolic BP, mean BP, and pulse pressure values (incident: r = 0.70, r = 0.98, and r = 0.82; prevalent: r = 0.74, r = 0.99, and r = 0.83, respectively; all P < 0.0001). Postdialysis values for BP-related measures also correlated with predialysis SBP (incident: r = 0.74, r = 0.54, r = 0.62, and r = 0.73 for postdialysis SBP, diastolic BP, mean BP, and pulse pressure; prevalent: r = 0.68, r = 0.52, r = 0.57, and r = 0.67, respectively; all P < 0.0001). Postdialysis BP measures also correlated with their respective predialysis values (incident: r = 0.73, r = 0.77, r = 0.74, and r = 0.76 for SBP, diastolic BP, mean BP, and pulse pressures; prevalent: r = 0.68, r = 0.75, r = 0.69, and r = 0.73, respectively; all P < 0.0001).The initial SBP distribution histograms and percentages of patients in each of the 6 categories are shown in Fig 1 and suggested approximately normal curves. Nine percent and 5.7% of incident and prevalent patients were classified as category 1, normal; and 26.0% and 20.1% were classified as category 2, prehypertension, respectively. Therefore, SBP in 35.0% and 25.8% of patients was less than the guideline target (SBP < 140 mm Hg).12National Kidney Foundation: K/DOQI Clinical Practice Guidelines for Cardiovascular Disease in Dialysis.Am J Kidney Dis. 2005; 45: S49-S57Google Scholar Stage 1 hypertension, category 3, had the greatest percentage of patients in both samples. The 3 expanded hypertension categories, category 4 (stage 2), category 5 (stage 2+), and category 6 (stage 2++), had 21% (28%), 6% (9%), and 1% (1%) of patients in the incident and (prevalent) samples, respectively.Survival Curves in Both SamplesFigure 2 shows 1-year survival curves for the 6 initial SBP categories. In incident patients, 35% of patients met the current SBP target (<140 mm Hg), but experienced worse survival than those with SBP in ranges generally considered to be elevated. The same observation applies to the 25% of prevalent patients with SBP less than 140 mm Hg. Survival curve patterns were similar between incident and prevalent samples. Overall, there were 9,234 (16.4%) and 10,414 total deaths (15.0%) within 1 year for the incident and prevalent cohorts, respectively. Proportions of patients for whom death was attributed directly to cardiovascular events (eg, cardiac arrest, congestive heart failure, cerebrovascular accidents, and other cardiac events) were 59.2% in the incident cohort and 60.2% in the prevalent cohort.Fig 2One-year survival curves for the initial predialysis SBP categories in incident and prevalent patient samples.View Large Image Figure ViewerDownload (PPT)SBP Risk Profiles in the Incident SampleResults of the 1-year Cox models for incident patients are listed in Table 2. Conventional 1-year models showed a clear association between SBP and death risk. The initial SBP categories of normal (SBP < 120 mm Hg), prehypertension (SBP, 120 to 140 mm Hg), and stage 1 hypertension (SBP, 140 to 160 mm Hg) were associated with significant hazard ratios (HRs) of 3.68, 1.68, and 1.08 compared with the stage 2 reference group (SBP, 160 to 180 mm Hg). Of patients with higher SBPs, only stage 2++ (SBP ≥ 200 mm Hg) was associated with a higher HR that was not significant statistically. Statistical adjustment for case-mix with or without BSA did not alter the findings.Table 2One-Year Death Hazard Analyses: Incident Patient SampleAnalysisConventional Cox ModelTime-Varying Cox ModelSBP Group (mm Hg)Wald Chi-SquarePHR95% Confidence IntervalWald Chi-SquarePHR95% Confidence IntervalUnadjusted <1201,355.6<0.00013.6823.435-3.9462,247.1<0.00017.4196.829-8.060 120-140263.3<0.00011.6821.580-1.792335.0<0.00012.2142.033-2.411 140-1606.40.0121.0851.018-1.15519.8<0.00011.2211.118-1.333 160-1800.0Reference1.0000.0Reference1.000 180-2000.2NS0.9760.877-1.0880.2NS0.9720.853-1.108 ≥2000.7NS1.1040.873-1.3974.40.03571.2481.015-1.536Case-mix adjusted <120994.4<0.00013.1432.927-3.3751,872.3<0.00016.4975.969-7.071 120-140194.0<0.00011.5731.476-1.676283.8<0.00012.0961.923-2.285 140-1603.60.0591.0630.998-1.13316.2<0.00011.1991.098-1.311 160-1800.0Reference1.0000.0Reference1.000 180-2000.3NS0.9700.870-1.0810.1NS0.9830.862-1.120 ≥2000.7NS1.1080.876-1.4014.90.02741.2641.026-1.557Case-mix + BSA adjusted <120898.6<0.00013.1912.957-3.4421,594.8<0.00016.2735.733-6.865 120-140171.9<0.00011.5761.472-1.687233.8 0.10). Open table in a new tab Time-varying models used "time-linked" SBP. HRs increased to 7.42, 2.21, and 1.22 for the 3 lower SBP categories of patients and to 1.25 for stage 2++ hypertension, now significantly associated with risk, respectively. Statistical adjustments did not noticeably influence risk patterns. Comparison of the 2 types of Cox models suggests that the recent SBP was more important (sensitive) than remote SBP when evaluating survival regardless of whether SBP is normal or very high. Figure 3 shows risk profiles graphically.Fig 3Risk profiles of predialysis SBPs in the incident sample.View Large Image Figure ViewerDownload (PPT)SBP Risk Profiles in the Prevalent SampleResults of the 2 Cox models for prevalent patients are listed in Table 3, with graphic illustrations shown in Fig 4. Risk patterns and change in patterns from conventional to time-varying Cox models were similar to those observed for incident patients, except that higher HRs were observed in the stage 2+ category in 2 of 3 analyses using the conventional model. The profiles suggest improvement in death risk with increasing SBP into the range of 160 to 180 mm Hg. Similarly, category 6, stage 2++, was associated significantly with greater relative death risk only in time-varying analyses. Addition of the vintage category to the models did not result in a meaningful change in risk profiles (analyses not shown).Table 3One-Year Death Hazard Analyses: Prevalent Patient SampleAnalysisConventional Cox ModelTime-Varying Cox ModelSBP Group (mm Hg)Wald Chi-SquarePHR95% Confidence IntervalWald Chi-SquarePHR95% Confidence IntervalUnadjusted <120793.1<0.00012.7692.579-2.9721,682.2<0.00015.9495.463-6.479 120-140295.0<0.00011.6271.539-1.720254.5<0.00012.0291.860-2.213 140-16015.5<0.00011.1111.054-1.17119.8<0.00011.2241.120-1.338 160-1800.0Reference1.0000.0Reference1.000 180-2004.80.0291.0901.009-1.1780.7NS1.0540.933-1.191 ≥2003.4NS1.1920.990-1.43520.5<0.00011.5201.268-1.823Case-mix adjusted <120640.9<0.00012.6442.452-2.8511,403.8<0.00015.5875.106-6.114 120-140223.9<0.00011.5691.479-1.664192.8<0.00011.9171.749-2.101 140-16011.20.00081.1001.040-1.16316.8<0.00011.2161.107-1.334 160-1800.0Reference1.0000.0Reference1.000 180-2004.50.0341.0921.006-1.1851.0NS1.0690.940-1.214 ≥2002.5NS1.1710.961-1.42820.8<0.00011.5541.286-1.879Case-mix + BSA adjusted <120636.0<0.00012.6342.443-2.8401,365.6<0.00015.5355.054-6.060 120-140232.3<0.00011.5821.491-1.678192.7<0.00011.9261.756-2.113 140-16012.90.00031.1071.048-1.17117.1<0.00011.2201.110-1.340 160-1800.0Reference1.0000.0Reference1.000 180-2003.8NS1.0840.999-1.1771.0NS1.0670.938-1.213 ≥2001.5NS1.1330.930-1.38118.8 0.10). Open table in a new tab Fig 4Risk profiles of predialysis SBPs in the prevalent sample.View Large Image Figure ViewerDownload (PPT)Incident Sample 3-Year SurvivalFigure 5 shows corresponding Cox model results for the 3-year follow-up of the incident cohort. Patients with initial SBP in the normal and prehypertensive range experienced noticeably worse survival than patients with hypertension, including those with stage 2++ hypertension. The Cox analyses again suggest that SBP in the normal and prehypertensive range is associated with the greatest risk; the risk profile is accentuated by time-varying analyses, very high SBP becomes associated significantly with risk only in the time-varying analyses, and statistical adjustments for case-mix and BSA did not alter results. Finally, addition of admission year to the models did not alter these findings (analyses not shown).Fig 5Three-year survival risk profiles of predialysis categorical SBP in the incident sample.View Large Image Figure ViewerDownload (PPT)SBP Change During Follow-UpSurvival analyses were supplemented with simple SBP change analyses to better understand why the apparent risks associated with normal and stage 2++ hypertension were amplified by the time-varying Cox models compared with conventional models. The top panels of Fig 6 show comparisons of the first and last SBPs measured in patients who survived 1 year on dialysis therapy with those who died. The pairwise change in SBPs is shown in the bottom panels.Fig 6(Top) First and last SBPs and (bottom) BP changes in patients who lived 1 year or died for the (left) incident and (right) prevalent patient samples.View Large Image Figure ViewerDownload (PPT)Average SBP of incident patients who survived for 1 year increased from 149.9 to 153.2 mm Hg, a difference of 3.26 mm Hg (P < 0.0001). Conversely, incident patients who died showed a decrease in average SBP from 140.6 to 134.7 mm Hg, a difference of −5.44 mm Hg (P < 0.0001). In the prevalent cohort, survivors maintained their average SBP from 154.3 to 154.1 mm Hg (difference, P > 0.05), whereas those who died had a decrease from 148.4 to 137.9 mm Hg, a difference of −10.5 mm Hg (P < 0.0001).SBP for patients with SBP less than 120 mm Hg who lived 1 year increased by averages of 19.8 (incident patients) and 10.4 mm Hg (prevalent patients). Conversely, it was relatively unchanged in dying patients, increasing by only 1.9 mm Hg (incident patients) or decreasing by 2.7 mm Hg (prevalent patients). SBP tended to decrease more dramatically in dying than living patients in all SBP categories (P < 0.0001), with the possible exception of stage 2++ hypertension. The decrease in SBPs for incident patients in the reference category (160 < SBP ≤ 180 mm Hg) who lived and died averaged 5.8 and 14.1 mm Hg, respectively: a difference of 8.3 mm Hg. Similar values for prevalent patients in that category were 5.1 and 14.2 mm Hg, a similar difference of 9.1 mm Hg. However, living-dying differences were much less in patients with stage 2++ hypertension: 4.0 mm Hg for incident patients (−31.2 and −35.2 mm Hg, respectively) and 2.7 mm Hg for prevalent patients (−24.5 and −27.2 mm Hg, respectively).DiscussionThis study evaluates the epidemiological associations of SBP with death risk by using a variety of analytic approaches to ascertain robustness of the observations. We used 2 sampling methods (incident and prevalent patients) and 2 Cox modeling techniques (conventional and time varying). Furthermore, we performed a series of subsample analyses using both patient samples and both modeling methods. Findings were remarkably consistent across all analyses and are similar to findings reported in the literature.5Lowrie E.G. Huang W.H. Lew N.L. Liu Y. The relative contribution of measured variables to death risk among hemodialysis patients.in: Friedman E. Death on Hemodialysis: Preventable or Inevitable? Kluwer, Hingham, MA1994: 121-141Crossref Google Scholar, 6Salem M.M. Bower J. Hypertension in the hemodialysis population: Any relation to one-year survival?.Am J Kidney Dis. 1996; 28: 737-740Abstract Full Text PDF PubMed Scopus (84) Google Scholar, 7Zager P.G. Nikolic J. Brown R.H. et al."U" curve association of blood pressure and mortality in hemodialysis patients Medical Directors of Dialysis Clinic, Inc.Kidney Int. 1998; 54: 561-569Crossref PubMed Scopus (590) Google Scholar, 8Salem M.M. Hypertension in the hemodialysis population: Any relationship to 2-years survival?.Nephrol Dial Transplant. 1999; 14: 125-128Crossref PubMed Scopus (76) Google Scholar, 9Klassen P.S. Lowrie E.G. Reddan D.N. et al.Association between pulse pressure and mortality in patients undergoing maintenance hemodialysis.JAMA. 2002; 287: 1548-1555Crossref PubMed Scopus (359) Google Scholar SBPs considered normal and prehypertensive13Chobanian A.V. Bakris G.L. Black H.R. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The JNC 7 Report.JAMA. 2003; 289: 2560-2571Crossref PubMed Scopus (16370) Google Scholar were most strongly associated with the greatest death risk. Conversely, severe hypertension, our stage 2++, was associated inconsistently with greater risk, attaining statistical significance only in time-varying models.Between 25% and 35% of patients were in the normal or prehypertensive range of SBP, whereas only 1% had stage 2++ hypertension. When SBP distributions are viewed in the context of much greater death risk for patients with normal SBP and prehypertensive patients, it seems reasonable for clinical investigators and practicing physicians to reevaluate the usual interpretation of BP in HD patients, devise appropriate investigative strategies, and use patient-specific management goals for dialysis patients, even those with normal or prehypertensive SBP. Therefore, an early insight evolving from these findings is the current absence of evidence-based recommendations for evaluating patients with normal and prehypertensive BP because current clinical BP guidelines for dialysis patients12National Kidney Foundation: K/DOQI Clinical Practice Guidelines for Cardiovascular Disease in Dialysis.Am J Kidney Dis. 2005; 45: S49-S57Google Scholar focus solely on the management of high BP. They likely are patterned after guidelines established for the general population.13Chobanian A.V. Bakris G.L.
Referência(s)