Blood Pressure and Mortality in Hemodialysis Patients: A Systematic Review of an Ongoing Debate

Authors

  • Yusra Habib Khan,

    Corresponding author
    1. Discipline of Clinical Pharmacy, School of Pharmaceutical Sciences, University Sains Malaysia, Penang
    2. Chronic Kidney Disease Resource Centre, School of Medical Sciences, Health Campus, University Sains Malaysia, Kubang Kerain, Kelantan, Malaysia
    • Address correspondence and reprint requests to Dr Yusra Habib Khan, Discipline of Clinical Pharmacy, School of Pharmaceutical Sciences, University Sains Malaysia, Penang 11800, Malaysia. Email: yusrahabib@ymail.com

    Search for more papers by this author
  • Azmi Sarriff,

    1. Discipline of Clinical Pharmacy, School of Pharmaceutical Sciences, University Sains Malaysia, Penang
    Search for more papers by this author
  • Azreen Syazril Adnan,

    1. Chronic Kidney Disease Resource Centre, School of Medical Sciences, Health Campus, University Sains Malaysia, Kubang Kerain, Kelantan, Malaysia
    Search for more papers by this author
  • Amer Hayat Khan,

    1. Discipline of Clinical Pharmacy, School of Pharmaceutical Sciences, University Sains Malaysia, Penang
    Search for more papers by this author
  • Tauqeer Hussain Mallhi

    1. Discipline of Clinical Pharmacy, School of Pharmaceutical Sciences, University Sains Malaysia, Penang
    Search for more papers by this author

Abstract

Hypertension is prevalent in 75–80% of hemodialysis patients and remains the most controversial prognostic marker in end stage kidney disease patients. In contrast to the general population where systolic blood pressure of ≤120 mm Hg is considered normal, a debate remains regarding the ideal target blood pressure in hemodialysis patients. Using the PUBMED and EMBASE databases, the research studies that evaluated the relationship between blood pressure measurements and mortality in hemodialysis patients were searched. Thirteen studies were identified from different regions of the world. Five studies reported low predialysis systolic blood pressure as a prognostic marker of mortality. Other studies showed varying results and reported postdialysis systolic blood pressure as well as ambulatory blood pressure as better predictors of mortality and emphasized their optimized control. One study in this review concluded that there is no direct relationship between mortality and blood pressure if the patients are on anti-hypertensive medications. The observed all-cause mortality varied from 12% to 36%, whereas the cardiovascular mortality varied from 16% to 60%. On the basis of studies included in the current review, a low predialysis systolic blood pressure (<120 mm Hg) is shown to be a widely accepted prognostic marker of mortality while ambulatory blood pressure best predicts CV mortality. Therefore, we recommend that apart from routine BP (pre, post and intradialysis) monitoring in centers, assessment of ambulatory BP must be mandatory for all patients to reduce CV mortality in hemodialysis patients.

In comparison to the general population, hemodialysis patients are characterized by a remarkably high mortality rate [1]. Mortality assessment in hemodialysis patients is typically based on an estimation of prognostic markers. Prognostic markers are characteristics that classify a study population with different risk levels in relationship to a specific outcome [2]. In end stage kidney disease (ESKD) patients, the specific outcome of interest is the expected mortality or survival.

Despite recommendations by the American Society of Nephrology (ASN), prognostic research has received less attention than therapeutic and diagnostic research [3]. In recent years, several mortality predicting prognostic markers have been identified in hemodialysis patients. Because of the lack of consensus, uncertainty regarding the prognostic significance of the established markers persists. Although, hypertension is prevalent in 75– 80% of hemodialysis patients, it remains the most controversial prognostic marker in ESKD patients [4]. In contrast to the guidelines for the general population, in which systolic blood pressure of ≤120 mmHg is considered normal, a debate remains regarding the ideal target blood pressure in hemodialysis patients [1]. Moreover, literature research shows disparity regarding prognostic significance of type of blood pressure measurement with respect to mortality in hemodialysis patients.

On the basis of the above mentioned, this review was carried out to examine the research studies that have analyzed the relationship of different types of blood pressure measurements and mortality in hemodialysis patients.

Methodology

Search strategy

A 10-year systematic review of publications that evaluated the relationship between blood pressure and mortality in hemodialysis patients was conducted (Fig. 1). Two reviewers (Y.H.K, T.H.M) independently searched the PUBMED and EMBASE databases (from 2004 to the present) for relevant studies. Disagreement or discrepancies were resolved by third reviewer (A.H.K). In conjunction with blood pressure, systolic blood pressure, diastolic blood pressure, hypertension, predialysis and postdialysis, the following search terms were used: end stage kidney disease (ESKD), mortality, survival, prognosis, predictive, prognostic markers. Medical subject headings and text terms were used. The search was limited to studies on hemodialysis patients published in the English language.

Figure 1.

Review process flow sheet diagram.

Inclusion/Exclusion criteria

All the original investigations that evaluated the relationship between blood pressure and mortality in hemodialysis patients (>18 years) and were published from 2004–2014 were included in this review. For a study to be selected the minimum number of participants was 100 and the minimum follow-up was 1 year. Meta-analyses, systematic reviews, editorials, commentaries and case reports were excluded, as were studies with patients on other forms of renal replacement therapy, patients younger than 18 years of age and studies with no relevant outcome.

Data extraction and quality assessment:

The relevant data were extracted and tabulated for convenience in Table 2 (January 2004 until January 2014). The methodological quality of the studies was assessed using the STARD recommendations (Statement for Reporting Studies of Diagnostic Accuracies). The original checklist for the STARD recommendations consisted of 25 items for quality assessment [18]. The checklist was further modified and scored according to the scoring system used by Coca et al. 2007 [19]. Ten validity criteria were used for this review, which are summarized in Table 1.

Table 1. Scoring system for assessing validity of studies included in systematic review
SectionCriteriaScoringComments
  1. CI, Confidence interval.
IntroductionClearly defined research question/ study aims

Yes = 1

No = 0

All studies

------------

MethodologyDescription of study population, inclusion exclusion criteria, study settings

Yes =1

No = 0

8,10–20

9

 Participants recruitment

Incident cohort = 1

Prevalent cohort = 0

9,10,11,13,14,17,20

8,12,15,16,18,19

 Sample size calculation

Yes = 1

No = 0

None

All studies

 Data collection

Prospective = 1

Retrospective = 0

8-10,13,14,19,20

11,12,15,16,17,18

Statistical analysisDescription of methods for calculating prognostic accuracy, survival and methods used to quantify uncertainty i.e. CI

Yes = 1

No = 0

8,9,11,12,14–17,19

10,13,18,20

 Report when study was done including beginning and ending dates of recruitment

Yes = 1

No = 0

All studies

------------

 Report clinical and demographic characteristics of study population, missing participants, reason of missing

Yes = 1

No = 0

8,10–12,14–20

9,13

 Report population using antihypertensives

Yes = 1

No = 0

8,10–12, 14,16,17,18,20

9,13,15,19

DiscussionDiscuss clinical applicability of study findings

Yes = 1

No = 0

All studies

------------

Results

A total of 13 studies that met our inclusion criteria were included (Table 2). These studies were conducted in different geographical locations (five US studies two studies in Asia; three UK studies) with a maximum number of participants of 67 085 and a highest quality score of 9. One study scored 9 [11] whereas the majority of studies attained a score of 8. The only domain in which the studies scored less was that of participant recruitment because most of the studies were retrospective, with a cohort of prevalent hemodialysis patients.

Table 2. Summaries of studies demonstrating association of blood pressure with mortality in hemodialysis patients
ReferencePopulationDuration and siteObjectiveConfounder adjustmentType of BP measurementMortality rateRef groupMortality predictor (BP at which maximum mortality was observed)P-valueHR (95% CI)
All causeCVS
Shoji T et al., (2004) [5] QS: 81244 prevalent HD patients, 55% male, mean age 60 ± 13 y1999–2001 28 dialysis units JapanTo evaluate impact of HD associated hypotension on mortalityAge, gender, diabetes, serum creatinine, ultrafiltration /body weight, body weight post HDIn center predialysis (supine, standing)Intradialysis, postdialysis (supine, standing)12%51%Intradialysis SBP 130–139 DBP 70–79Intradialysis hypotensionSBP:110-119 mmHg DBP: <59 mmHg Orthostatic hypotension SBP: < 99 mmHg<0.05 <0.05 <0.053.03 (1.14–8.04) 2.62 (1.08–6.40) 2.47 (1.11–5.49)
Li Z et al., (2006) [6] QS: 656 338 incident, 69 590 prevalent HD patients, 52.4% male, mean age: 61.5y, Total:125 928Prevalent:1997–2001 Incident:2002–2004 North AmericaRelationship between SBP and mortalityAge at initiation of HD, sex, diabetic status, raceIn center seated predialysis, postdialysis15.7%59.7%SBP ≥160- ≤ 180Low predialysis SBP: <120 mmHg0.00013.14 (2.92–3.37)
Stidley T A et al., (2006) [7] QS: 816 959 incident HD patients, 52.3% male, age: 20- >75y1993–2003 MexicoTo evaluate relationship between BP and mortality over timeSerum albumin, hematocrit, spKt/VIn center seated predialysis, postdialysis36%42%SBP 140––149Low predialysis SBP:<120 mmHg (for first 2 years) >150 mmHg (3 years and above)NANA
Agarwal R (2010) [8] QS: 8326 incident HD patients, 66% male, Age: 54.9 ± 12.932 months (median follow-up) 4 dialysis units Indianopolis, IN, USARelationship between home, ambulatory, pre and post-dialysis BP with mortalityAge, ethnicity, sex, diabetes mellitus, cardiovascular disease, antihypertensive medications, serum albumin, hemoglobin, and dialysis vintage.Ambulatory BP (every 20 min during day, every 30 min during night) In center predialysis, postdialysis Home BP ( 3 x/d by patient)31%NAABP 79-119 HBP 88-133Ambulatory b.p 134–146 mmHg Home b.p 139–149 mm0.0001 0.023.43 (1.73–6.79) 2.15 (1.13–4.11)
Tripepi G et al., (2005) [9] QS: 7CVS free HD patients, 30% male, Mean age 57.2 ± 15.8, 168 non-diabeticAvg. follow up: 38 ± 22 months 3 dialysis units ItalyPrognostic power of 24 hr ambulatory BP with mortalityAge, gender, smoking, duration of regular HD treatment, antihypertensive therapy, hematocrit, albumin, cholesterol, calcium phosphate product, Kt/V, and LVHAmbulatory BP (every 15 min during day and night)28%60%Night/day SBP ratio <0.93Night/day SBP ratio: >10.022.88 (1.21–6.87)
Myers OB et al., (2010) [10] QS: 616 283 incident HD patients, 54% male, age: 20–80 years2001–2006 Median follow-up: 1.5 years New Mexico, USExamine whether age affects association between predialysis SBP and mortalityAge; gender; race; cause of ESRD; and time-varying serum albumin, hemoglobin, creatinine, dialysis dosage, and postdialysis weight.In center predialysis, postdialysis38%NASBP 140Age and predialysis SBP: Age > 50 years: low systolic bp (<140 mmHg) Age < 50 years, high systolic bp (>160 mmHg)0.001 0.001NA
Inrig JK et al., (2009) [11] QS: 91748 incident HD patients, 51% male, Age: 61.3 ± 15.4, 85% hypertensive2 years USAAssociation between intradialysis BP and mortalityAge, body mass index less than 26 kg/m2, interdialytic weight gain greater than 5%, diabetes mellitus, hypertension, peripheral vascular disease, congestive heart disease, coronary artery disease, serum albumin level, creatinine level, phosphorus level, and use of nitrates.In center predialysis, postdialysis32.6%58.9%SBP unchanged±10 mmHg during HDIntradialysis SBP: Increase intradialysis SBP: >10 mmHg from predialysis SBP if predialysis SBP <120 mmHg0.031.01-1.12
Molnar MZ et al., (2010) [12] QS: 667 085 prevalent HD patients, 56% male, Mean age: 62.5 ± 12.6, 43.2% diabetics2001–2006 USABlood pressure and survivalAge, gender, race-ethnicity, diabetes mellitus, vintage categories, primary insurance, marital status, residual renal function, Kt/VIn center seated predialysis and postdialysis (mean of BP taken from 3 consecutive HD sessions)12.3%NASBP 140-159Low Pre and post dialysis SBP: predialysis SBP <120 mmHg postdialysis SBP < 120 mmHg0.0001 0.00011.71 (1.60–1.82) 1.26 (1.22–1.30)
Yang CY et al., (2012) QS: 7 [13]115 prevalent HD patients, Mean age: 64 y, 45% male, 91% hypertensive2006–2010 Follow-up: 4 years TaiwanInfluence of postdialysis BP on mortalityAdjustment for confounders not indicatedIn center seated predialysis and postdialysis (mean of BP taken from 25 consecutive HD sessions)35%16%Change in postdialysis SBP of -5 to 5 mmHgPostdialysis SBP Increase in postdialysis SBP of more than 5 mmHg from baseline0.0083.92 (1.41–10.84)
Selvarajah V et al., (2013) [14] QS: 8203 incident HD patients, 66% male, Mean age: 66 ± 15y2005-2010 Mean follow up: 2 yearsUKIntra individual BP variability (SBP, DBP) with all-cause mortalityAge, sex, mean SBP, previous cardio-vascular events, diabetesIn center seated predialysis and postdialysis for consecutive 3 months18.2%NAVIM of SBP below meanVariability in predialysis SBP: For one unit mmHg increase in SD of SBP (SD 5.1 mmHg)NA1.09 (1.02-1.06)
Losito A et al., (2012) [15] QS: 63674 HD patients, Male: 61.7% Mean age: 67.2 ± 14.1 years Diabetic: 19.5%2006-2009 Follow up: 26.5 ± 10.5 months 73 dialysis units, ItalyRelationship between BP and mortalityHemoglobin, serum cholesterol, parathyroid hormone, Kt/V, interdialytic weight gain, and all BP componentsIn center seated predialysis and postdialysis26%NA--------------No relationship between bp and mortalityNANA
Chang TI et al., (2011) [16] QS: 71846 prevalent HD patients, Age: 57.6 years, Male: 44% Diabetics: 44.6%1995–2000 Follow up 2.5 years USATime from randomization to all-cause mortalityAge, sex, black race, vintage, diabetes, serum albumin,residual kidney function, intervention group, ICED scoreIn center seated predialysis, postdialysis47.2%47%Predialysis SBP 140–159Prediaysis SBP: 120 mm Hg0.00011.84 (1.45–2.30)
Bos WJ et al., (2010) QS: 8 [17]1111 Incident HD patients, both with and without CVD 59% male, Mean age: 63 ± 14 yearsFollow-up: 7.5 years The NetherlandsRelationship between SBP and mortalityAge, sex, comorbidity scorePre and postdialysis Bp measurements at 3 months, 6 months and then every6 months (mean of 6 BP readings)43%NASBP 120–140Predialysis SBP: <120 mm Hg for first months long term effects not clear<0.051

Of the 13 studies, six concluded that low predialysis systolic blood pressure is associated with an increase all-cause as well as cardiovascular mortality [6, 7, 10, 12, 16, 17]. Of six studies, five concluded that SBP of <120 mmHg is associated with mortality whereas Myers et al. observed that predialysis SBP of <140 mmHg is associated with mortality [10]. Among the authors who reported low predialysis systolic blood pressure as a predictor of mortality, the highest all-cause (AC) mortality was reported by Chang et al. (47.2%) whereas the highest cardiovascular (CV) mortality (59.7%) was reported by Li et al. (2006) over a follow-up of 2.5 years and 1 year, respectively [6, 16].

Two studies showed a relationship between postdialysis systolic blood pressure and mortality [12, 13]. Monar et al. concluded that, in addition to a low predialysis SBP <120 mm Hg, a low postdialysis SBP of <120 mm Hg is associated with all-cause and cardiovascular mortality [12]. Yang et al. concluded that a 5 mmHg increase in the postdialysis SBP from the baseline is associated with CV (16%) and all-cause (35%) mortality during a follow-up of 4 years [13].

Two studies in this review concluded that the intradialysis variation in SBP is a better predictor of mortality than predialysis or postdialysis SBP [5, 11]. A study by Inrig et al. concluded that for every 10 mmHg rise in the SBP, there is an increased risk of mortality. The reported all-cause and cardiovascular mortality over the 2-year follow-up were 32.6% and 58.9%, respectively [5]. In contrast to the findings of Inrig et al., Shoji et al. showed that a decrease in the intradialysis SBP (110–119 mmHg), instead of an increase in the intradialysis SBP, is associated with mortality [11]. The contrary findings of these studies could be attributed to the differences in the study populations under observation because Inrig et al. studied incident hemodialysis patients, of which a majority of patients (85%) were hypertensive whereas Shoji et al. observed prevalent hemodialysis patients among whom 58.9% of the patients were diabetics.

Studies by Agarwal et al. and Tripepi et al. concluded that ambulatory blood pressure affects mortality [8, 9] whereas predialysis and postdialysis blood pressure are not significant prognostic markers of mortality. Although these studies were conducted on a limited number of patients (overall: 494), the main strength of these studies is their prospective nature with incident hemodialysis patients. The participants in the Tripepi et al. [9] study were free from baseline cardiovascular disease that could influence blood pressure and mortality.

One study included in this review demonstrated no relationship between mortality and blood pressure [15]. The authors conducted a prospective nationwide multicenter study of chronic hemodialysis patients and observed a relationship between different measurements of blood pressure and mortality. The study findings study suggested that there is no direct relationship between blood pressure and mortality among patients taking anti-hypertensive medications; however, mortality was inversely associated with postdialysis systolic blood pressure in the patients who were not on anti-hypertensive medications.

Discussion

This review evaluates the relationship between different types of BP measurements and mortality in hemodialysis patients. Since the early 1990s, several studies have reported a relationship between blood pressure measurements and outcomes in hemodialysis patients; still there is no consensus regarding the target blood pressure for chronic hemodialysis patients [20, 21]. A lack of clearly defined guidelines presents challenges to practicing physicians in management of such patients, particularly in emergency hypertensive cases.

According to the National Clinical Practice guidelines of the United States, a predialysis BP of less than 140/90 mmHg and a postdialysis BP of less than 130/80 mmHg is recommended [22]. These guidelines acknowledge that the evidence to support this recommendation is weak. Similar BP targets were proposed by UK guidelines; however, because the evidence was limited, these recommendations were removed [23]. Davenport et al. reported a higher incidence of intradialytic hypotension in patients who achieved the targeted postdialysis BP [24]. Intradialytic hypotension is, in turn, associated with a high incidence of cardiac complications as well as death [6, 25]. Many observational studies have consistently demonstrated a linear relationship between the risk of CVD and an increase in BP in the general population. In contrast to the general population, the same linear relationship does not exist in ESKD patients. Over the past few decades, many studies have demonstrated a “U-shaped” or “reverse J-shaped” relationship between BP and mortality, suggesting high mortality with low pre- and postdialysis SBP (<120 mmHg) and a slightly higher risk of death with a higher SBP (>180 mmHg) [6, 25, 26].

Systolic blood pressure

A low predialysis as well as postdialysis systolic BP of <120 mm Hg is statistically significant and is the most documented prognostic marker of all-cause and CV mortality.

The association between SBP and mortality is complicated because this relationship changes over time. Stidley et al. reported that during the first 2 years of hemodialysis, a predialysis SBP of less than 120 mmHg is strongly associated with a threefold-higher risk of mortality; however, this relationship diminished during the 3rd and 4th year of hemodialysis [7]. Similar findings were stated by other studies [12, 16, 17]. Second, this complex relationship is further modified by confounding demographic factors such as age and concurrent diabetes mellitus. Myers et al. reported an association between low predialysis SBP (<140 mm Hg) and mortality. However, when the cohort was stratified by age, the authors found that the risk of low SBP and mortality remains for patients aged 50 years and above only, whereas the risk was insignificant for patients younger than 50 years. Moreover, the risk was stronger in patients with diabetes mellitus than in non-diabetics [10]. It is well documented that mortality caused by low SBP is attributed to cardiovascular complications.

Ambulatory blood pressure

In a study by Tripepi et al., 168 non-diabetic hemodialysis patients were followed up for 38 months [9]. The authors found an association between the highest night/day BP ratio and cardiovascular mortality. The etiology behind the relationship between night to day BP and CV mortality could be explained by the phenomenon of non-dipping. In the general population, there is an approximate 10–20% fall in BP during sleep, which is referred to as dipping. In CKD (50%) and ESKD (80%) patients, loss in diurnal variation in blood pressure, a phenomenon known as non-dipping, has been reported [26]. A loss of diurnal variation in blood pressure has been associated with a poor renal prognosis and left ventricular hypertrophy, which, in turn, causes adverse cardiovascular outcomes that lead to all-cause and cardiovascular system (CVS) mortality in ESKD patients. Similar results were reported by Agarwal. Agarwal also reported that ambulatory BP (134–146 mmHg) is a significant predictor of all-cause mortality (31%) over a median follow-up of 32 months; CV mortality was not mentioned in his study [9].

Intradialytic variations in BP

The prevalence of intradialytic hypertension varies from 10–15% in hemodialysis patients [26, 27]. Intradialytic hypertension is the paradoxical increase in SBP during or immediately after a hemodialysis session [11]. The possible mechanism behind intradialytic hypertension is the removal of large amounts of fluid (approximately 4 L) during ultrafiltration. Using the data from the Dialysis Morbidity and Mortality Wave 2 Study, Inrig et al. reported that a 10 mmHg increase in SBP during dialysis is associated with a 6% (95% CI, 1%–11%) increased risk of all-cause death at 2 years. This association was found to interact significantly with baseline predialysis SBP, with a higher risk of death in patients with a baseline predialysis SBP of less than 120 mmHg [11]. With a random sample of 6393 hemodialysis patients, Flythe et al. conducted a prospective study to find out the relationship between intradialytic SBP variations with all cause and CV mortality. Authors concluded that higher intradialytic SBP variations are strongly associated with all cause (HR 1.26; P =  0.001) and CV mortality (HR 1.32; P =  0.04) [28]. Apart from intradialytic hypertension, intradialytic hypotension is also associated with higher risk of mortality [29].

Diastolic blood pressure

The 13 studies included in this review focused on systolic blood pressure (predialyis, postdialysis, intradialytic) and diastolic blood pressure did not gain significant importance as a mortality predicting prognostic marker. Based on a comprehensive literature review, Agarwal (2005) reported that when systolic and diastolic blood pressures are considered separately, an inverse relationship is observed between the total mortality and blood pressure. However, when systolic and diastolic blood pressure levels are considered together, a direct relationship between the total mortality and systolic blood pressure is observed whereas an inverse relationship is observed between diastolic blood pressure and total mortality [27].

Blood pressure measurements

All the studies except one [9] have used blood pressure measurements at dialysis centers or at home to evaluate the relationship between blood pressure and mortality. Agarwal (2010) conducted a study in which blood pressure measurements were taken at different places, i.e., dialysis centers (predialysis, postdialysis, intradialytic), home and ambulatory BP [30]. He concluded that ambulatory BP measurements have better prognostic significance than predialysis and postdialysis BP measurements. These findings were supported by other studies that suggested a better prognostic significance of ambulatory and home BP compared to dialysis BP measurements [9, 31]. The possible reason for the superiority of BP measurement outside dialysis centers could be explained by a variation in BP because of extreme changes in the volemic state (predialytic, intradylatic, post dialytic) that cause difficulty in obtaining a representative value of blood pressure at dialysis centers. Additionally, the white coat effect contributes to fluctuations in blood pressure when measured at dialysis centers [32]. Gerin et al. further reported that the better prognostic significance of BP measurements observed outside of dialysis units could be attributed to a potential detection of masked hypertension (elevated BP at home and normal at the dialysis unit) and recording of BPs samples from a broader pool of conditions that makes these readings more representative of individuals' BP [33].

Therefore, BP readings taken at home and ambulatory blood pressure should be recorded to draw a conclusive statement on the ideal BP in hemodialysis patients and to understand the complex role of blood pressure in hemodialysis patients.

Limitations

The main limitation of this review is selection bias. Efforts were made to ensure inclusion of all the relevant studies and to clearly elaborate the results of studies; however, there are chances of error. Furthermore, due to methodological heterogeneity of studies, a meta-analysis was not done (I2 > 75%). Some of the limitations of the studies included in this review are summarized in Table 3.

Table 3. Potential limitations of existing studies included in current review
Limitations
Very low number of patients
Inadequate follow-up
Lack of cardiovascular system (CVS) mortality reporting
Consideration of systolic and diastolic blood pressure separately
Use of only one type of blood pressure measurement
Prevalent cohorts
Lack of consideration of cardiac dysfunction
No information on predialysis serum sodium content and volume overload

Unidentified confounders might have influenced the relationship between BP and mortality. Other than hypertension, adjustment of risk factors (such as diabetes, smoking) that are strong predictors of CVS mortality might dilute the effect of hypertension as a risk factor. In addition, a formal assessment of the comorbidities was not performed in all the studies, and the data on the overall use of medications were very limited. An assessment of the baseline comorbidities is crucial because the relationship between low BP and mortality is stronger in diabetics than in non-diabetics [10].

Anti-hypertensive treatment varies from 59–83% in hemodialysis patients. Of the 12 studies included in this review, four studies did not report the use of anti-hypertensive medications [7, 11, 13, 17]. Heerspink et al. conducted a meta-analysis of randomized controlled trials to assess the effect of anti-hypertensive therapy in dialysis patients and found that, regardless of BP control, anti-hypertensive medications are associated with a lower risk of all-cause and CV mortality [34]. Such confounding factors might affect the relationship between blood pressure and mortality and should not be ignored in mortality reports. To show blood pressure as an independent prognostic marker of mortality in hemodialysis patients, the known confounders require adjustment by statistical analysis either by stratification or via multivariable techniques.

Conclusion

Conclusively, the relationship between blood pressure and mortality is complex in hemodialysis patients with SBP as better prognostic marker of all-cause mortality as compared to DBP. A low in-center SBP (<120 mm Hg) measured before and after hemodialysis demonstrates maximum risk of mortality especially in younger (less than 50) and diabetic population. However, half of the deaths in hemodialysis patients are attributed to CV causes and ambulatory BP is the best predictor of CV mortality. Therefore, we recommend that apart from routine BP (pre, post and intradialysis) monitoring in centers, assessment of ambulatory BP must be mandatory for all patients to reduce CV mortality in hemodialysis patients.

Recommendation

The following recommendations are based on the study results: randomized trials with well-defined BP goals are needed. Such trials should be designed with detailed methodologies, preferably recruiting incident cohorts and using elaborative techniques for longer follow-up periods. Furthermore, standardization of BP measurements, consideration of the pharmacological treatments and important covariates (diabetes, smoking) of BP and pulse pressure should be kept in mind. These measures would facilitate a better understanding of the relationship between blood pressure and mortality, as well as the CVD risk stratification in hemodialysis patients.

Acknowledgment

We would like to thank Institute of Postgradute Studies (IPS) of Universiti Sains Malaysia for providing fellowship support [Ref.no.P-FD0013/13(R)].

Author's Contribution

Y.H.K and A.S conceived the idea and designed the methodology. Literature search was done by Y.H.K, T.H.M and A.H.K. Manuscript writing was done by Y.H.K and A.S.A. Manuscript was approved by A.S and A.H.K.

Conflicts of Interest

: None.

Sources of support

: None

Ancillary