Advertisement
American Journal of Kidney Diseases

The Frequency of Routine Blood Sampling and Patient Outcomes Among Maintenance Hemodialysis Recipients

Published:November 12, 2019DOI:https://doi.org/10.1053/j.ajkd.2019.08.016

      Rationale & Objective

      Surveillance blood work is routinely performed in maintenance hemodialysis (HD) recipients. Although more frequent blood testing may confer better outcomes, there is little evidence to support any particular monitoring interval.

      Study Design

      Retrospective population-based cohort study.

      Setting & Participants

      All prevalent HD recipients in Ontario, Canada, as of April 1, 2011, and a cohort of incident patients commencing maintenance HD in Ontario, Canada, between April 1, 2011, and March 31, 2016.

      Exposure

      Frequency of surveillance blood work, monthly versus every 6 weeks.

      Outcomes

      The primary outcome was all-cause mortality. Secondary outcomes were major adverse cardiovascular events, all-cause hospitalization, and episodes of hyperkalemia.

      Analytical Approach

      Cox proportional hazards with adjustment for demographic and clinical characteristics was used to evaluate the association between blood testing frequency and all-cause mortality. Secondary outcomes were evaluated using the Andersen-Gill extension of the Cox model to allow for potential recurrent events.

      Results

      7,454 prevalent patients received care at 17 HD programs with monthly blood sampling protocols (n = 5,335 patients) and at 8 programs with blood sampling every 6 weeks (n = 2,119 patients). More frequent monitoring was not associated with a lower risk for all-cause mortality compared to blood sampling every 6 weeks (adjusted HR, 1.16; 95% CI, 0.99-1.38). Monthly monitoring was not associated with a lower risk for any of the secondary outcomes. Results were consistent among incident HD recipients.

      Limitations

      Unmeasured confounding; limited data for center practices unrelated to blood sampling frequency; no information on frequency of unscheduled blood work performed outside the prescribed sampling interval.

      Conclusions

      Monthly routine blood testing in HD recipients was not associated with a lower risk for death, cardiovascular events, or hospitalizations as compared with testing every 6 weeks. Given the health resource implications, the frequency of routine blood sampling in HD recipients deserves careful reassessment.

      Index Words

      Editorial, p. 465
      The routine measurement of biochemical and hematologic parameters is an established part of the surveillance provided to patients receiving maintenance hemodialysis (HD) and has several conceivable benefits. Information gleaned through routine laboratory testing is used to measure dialysis adequacy and guide the management of dialysis-associated complications, including anemia,
      • Locatelli F.
      • Nissenson A.R.
      • Barrett B.J.
      • et al.
      Clinical practice guidelines for anemia in chronic kidney disease: problems and solutions. A position statement from Kidney Disease: Improving Global Outcomes (KDIGO).
      dialysis adequacy,
      • Jindal K.
      • Chan C.T.
      • Deziel C.
      • et al.
      Hemodialysis clinical practice guidelines for the Canadian Society of Nephrology.
      electrolyte disturbances, and abnormalities associated with chronic kidney disease–mineral bone disorder (CKD-MBD).
      KDIGO clinical practice guideline for the diagnosis, evaluation, prevention, and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
      ,
      KDIGO 2017 clinical practice guideline update for the diagnosis, evaluation, prevention and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
      Routine laboratory testing can also yield actionable information that may improve patient outcomes, such as the identification of occult conditions unrelated to the patient’s kidney disease (eg, anemia from a gastrointestinal neoplasm). It is therefore plausible that more frequent routine laboratory checks will increase the likelihood of detecting remediable problems. Despite its perceived importance and the deep entrenchment of this practice in HD care, there is no evidence to support a specific interval for routine laboratory sampling. Many centers perform their measurements every month, but this interval is based on convention rather than empirical evidence of benefit to patients.
      Frequent blood testing could also have undesirable consequences by inflating the number of false-positive readings and identifying spurious abnormalities that escalate patient anxiety while generating further testing.
      • Naugler C.
      • Ma I.
      More than half of abnormal results from laboratory tests ordered by family physicians could be false-positive.
      Disproportionate attention to laboratory results may lead clinicians to overlook issues that are of higher priority to patients, such as fatigue, pain, and anxiety.
      • Urquhart-Secord R.
      • Craig J.C.
      • Hemmelgarn B.
      • et al.
      Patient and caregiver priorities for outcomes in hemodialysis: an international nominal group technique study.
      Ultimately, unless a given protocol of laboratory testing yields concrete benefits, a higher frequency of sampling may merely increase health care use and costs. Given efforts to eliminate health investigations and procedures that are of low value to patients and the introduction of initiatives such as Choosing Wisely into nephrology, we believe that current practices pertaining to routine blood sampling in HD recipients require careful examination.
      • Chan E
      • Hemmelgarn B
      • Klarenbach S
      • et al.
      Choosing Wisely: the Canadian Society of Nephrology’s list of 5 items physicians and patients should question.
      ,
      • Williams A.W.
      • Dwyer A.C.
      • Eddy A.A.
      • et al.
      Critical and honest conversations: the evidence behind the “Choosing Wisely” campaign recommendations by the American Society of Nephrology.
      We undertook this study to compare clinical outcomes across 2 testing interval strategies for routine blood testing among patients receiving maintenance HD. Specifically, we tested whether monthly blood testing was associated with lower risk for death, cardiovascular disease, and hospitalizations compared with testing every 6 weeks.

      Methods

      We conducted an Ontario-wide cohort study using linked administrative data sets. In Ontario (population of 14.6 million, area of 1.08 million km2), HD care is organized through 27 kidney programs that oversee policy at all HD units in their jurisdiction. We assembled 2 cohorts of Ontario residents 18 years or older who received maintenance HD. The first was a prevalent cohort of all patients actively receiving outpatient maintenance HD for at least 30 days as of April 1, 2011. The second comprised incident patients who commenced maintenance HD from April 1, 2011, to March 31, 2016. While the incident cohort afforded the follow-up of individuals who were exposed to a given blood sampling frequency from a common time point, the prevalent cohort enabled the assessment of individuals who would be expected to have a longer time on maintenance dialysis and were exposed to variations in practice that may have evolved over time. We excluded individuals who were registered with SickKids Hospital (assuming that these would be exclusively nonadults) and those who received care at a facility in which blood work sampling frequency changed between 2011 and 2016.
      The study was approved by the Research Ethics Board of St. Michael’s Hospital (St. Michael’s Hospital REB #17-339), and the need for individual-level patient consent was waived. The use of administrative data in this project was authorized under section 45 of Ontario’s Personal Health Information Protection Act. The reporting of this study follows guidelines for observational studies using routinely-collected health care data.
      • Benchimol E.I.
      • Smeeth L.
      • Guttmann A.
      • et al.
      The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement.

      Data Sources

      The study was completed using linked provincial health care administrative databases housed at the Institute for Clinical Evaluative Sciences (ICES). Prevalent and incident HD recipients were identified through the Ontario Renal Reporting System (ORRS). ORRS also enabled the identification of patients who transferred between dialysis programs or switched to peritoneal dialysis. The Canadian Organ Replacement Register identified patients who received a kidney transplant. We used ORRS, the Canadian Institute of Health Information Discharge Abstract Database (CIHI-DAD), and Ontario Health Insurance Plan database to characterize baseline comorbid conditions and health care use before cohort entry. The CIHI-DAD contains the date of hospital admission and discharge and 25 diagnoses in the International Classification of Diseases, Tenth Revision. The Ontario Health Insurance Plan database contains records of all physician billings for outpatient and inpatient services, including a service date and a principal diagnosis. Income quintile was based on the Statistics Canada census, which assigns income quintiles to each neighborhood community defined by residential postal code, with 1 representing the lowest and 5 representing the highest income quintile. The Johns Hopkins’ Aggregated Diagnosis Groups score was used as a global marker of each patient’s pre-existing comorbid conditions (Johns Hopkins’ Adjusted Clinical Group System, version 10.0).
      The ICES Physician Database was used to characterize physician visits. The health care databases were linked using unique encoded identifiers to safeguard patient confidentiality. All-cause mortality was ascertained from the Registered Persons Database, which contains health care identifiers for all eligible individuals, as well as age, sex, and death date. The remaining outcomes of interest were ascertained from the Discharge Abstract Database. Emergency department visits with hyperkalemia were captured in the National Ambulatory Care Reporting System. The administrative codes used are provided in Supplementary File 1.

      Definition of Exposure

      Using a web-based program (SurveyMonkey), we designed and administered an 18-question survey composed of multiple choice, short answer, and matrix questions. The survey link was sent out to the leadership of each of Ontario’s 27 kidney programs to ascertain the nature and frequency of routine blood sampling in the center’s HD program. The surveys were completed by the program’s medical director, or when not feasible, this was delegated to a senior member of the program’s nursing or administrative team. The frequency of testing core chemistry (electrolytes, markers of CKD-MBD, assessment of HD adequacy) and hematologic indexes defined the study exposure. Respondents were also asked to comment on whether any change in sampling frequency had taken place since 2010. All surveyed programs used one of the following 2 intervals for routine blood sampling: monthly (designated as high frequency) and every 6 weeks (designated as low frequency). A reproduction of the survey is found in Supplementary File 2.

      Outcomes

      The primary outcome was time to all-cause mortality through March 31, 2017, when follow-up ended. Follow-up was also censored at the time of a modality switch to peritoneal dialysis, receipt of a kidney transplant, transfer to a different dialysis program, or emigration from Ontario. Secondary outcomes were major adverse cardiovascular events (defined as a composite of hospitalization with myocardial infarction, hospitalization with nonfatal stroke, or limb amputation), all-cause hospitalization, all-cause emergency department visits, and visits to the emergency department or hospitalization with hyperkalemia.

      Covariates

      To account for potential confounding between laboratory sampling frequency and our outcomes of interest, we adjusted for age, sex, income quintile, time since dialysis initiation (relative to April 1, 2011, in the prevalent cohort), cause of end-stage kidney disease, markers of health care use (hospitalizations and emergency department visits during the year before cohort entry), and an array of comorbid conditions including the Aggregated Diagnosis Groups score.

      Statistical Analyses

      All analyses occurred at the level of the patient. Descriptive statistics were used to describe baseline variables, categorized by receipt of care in programs with monthly versus every-6-weeks testing policies. Intergroup differences were evaluated using standardized differences, with a value > 10% being considered clinically significant.
      • Austin P.C.
      Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research.
      Cox proportional hazards models were used to evaluate the relationship between blood work frequency and all-cause mortality, with receipt of care at centers with testing every 6 weeks as the referent group. Robust sandwich variance estimators allowed us to account for intraprogram correlation.
      • Lin D.Y.
      • Wei L.J.
      The robust inference for the Cox proportional hazards model.
      Follow-up time of patients was censored if there was evidence that they emigrated from Ontario (<0.5%), switched to peritoneal dialysis, received a kidney transplant, or transferred to a different dialysis program. The proportional hazards assumption was tested using the Kolmogorov-type supremum test.
      • Lin D.Y.
      • Wei L.J.
      • Ying Z.
      Model-checking techniques based on cumulative residuals.
      The remaining outcomes, all of which had the potential to occur recurrently, were evaluated using Andersen-Gill models, with adjustment for the mentioned baseline characteristics and number of previous events.
      • Amorim L.D.
      • Cai J.
      Modelling recurrent events: a tutorial for analysis in epidemiology.
      In these models, we accounted for intraprogram correlation by modeling center as a random effect and intraindividual correlation was accounted for using a robust sandwich variance estimator. We also evaluated the association between laboratory testing frequency and all-cause mortality in the following subgroups: diabetics and nondiabetics, individuals younger than 65 or 65 years and older, receipt of care in a program affiliated with a teaching hospital, and individuals in the prevalent cohort receiving HD for 3 years or less or more than 3 years as of April 1, 2011.
      We conducted a sensitivity analysis that included patients who were treated at programs in which blood work frequency practice changed during the study period. All patients at those centers were considered to be exposed to the monthly testing strategy that existed at those sites at the outset of the study period. All analyses were performed at ICES Western using SAS, version 9.4 (SAS Institute).

      Results

      Participant Characteristics

      We identified 7,454 Ontario residents who were receiving maintenance HD on April 1, 2011, who met our predefined eligibility criteria (the prevalent cohort; Fig 1A); an additional 10,666 eligible patients who commenced HD between April 1, 2011, and March 30, 2016 (Fig 1B), constituted the incident cohort. Among the 27 HD programs in the province, 17 performed routine sampling on a monthly basis (high-frequency cohort; 5,335 prevalent and 7,752 incident patients), while 8 did so every 6 weeks (low-frequency cohort; 2,119 prevalent and 2,914 incident patients) during the entire study period. Median follow-up was 1,089 (interquartile range, 426-2,176) days in the prevalent cohort and 641 (interquartile range, 230-1,132) days in the incident cohort. Two programs comprising 1,134 prevalent and 1,590 incident patients, respectively, changed their blood work policy during the study period (both transitioned from monthly to every-6-week intervals) and those patients were excluded from our primary analyses.
      Figure thumbnail gr1
      Figure 1Assembly of (A) prevalent cohort of hemodialysis recipients as of April 1, 2011, and (B) incident cohort of hemodialysis recipients between April 1, 2011, and March 31, 2016.

      Baseline Characteristics

      Demographic and clinical characteristics were generally well balanced between patients receiving HD at programs with monthly and every-6-week testing intervals, respectively, in both the prevalent and incident cohorts (Table 1). Importantly, the frequency of pre-existing conditions was similar in patients exposed to both blood sampling strategies. However, there were some notable differences. Among prevalent patients, mean time on dialysis was longer in patients receiving monthly testing versus testing every 6 weeks (940 ± 1,298 [standard deviation] vs 412 ± 695 days). Furthermore, in both cohorts, patients exposed to monthly blood sampling were more likely to be white, have lower incomes, and receive care at programs affiliated with teaching hospitals.
      Table 1Baseline Characteristics
      Prevalent CohortIncident Cohort
      Testing Every 6 wk (n = 2,119)Monthly Testing (n = 5,335)StDTesting Every 6 wk (n = 2,914)Monthly Testing (n = 7,752)StD
      Mean age, y64.6 ± 15.465.3 ± 15.4466.1 ± 14.966.0 ± 14.80
      Female sex845 (39.9%)2,241 (42.0%)41,077 (37.0%)2,922 (37.7%)2
      Income quintile
      Income quintile data missing on 46 patients in the prevalent cohort and 40 patients in the incident cohort.
       1 (lowest)455 (21.5%)1,600 (30.0%)20589 (20.2%)2,114 (27.3%)17
       2435 (20.5%)1,190 (22.3%)4621 (21.3%)1,827 (23.6%)5
       3496 (23.4%)910 (17.1%)16681 (23.4%)1,384 (17.9%)14
       4419 (19.8%)892 (16.7%)8598 (20.5%)1,293 (16.7%)10
       5 (highest)309 (14.6%)702 (13.2%)4421 (14.4%)1,098 (14.2%)1
      Race
       Black/African228 (10.8%)358 (6.7%)14166 (5.7%)388 (5.0%)3
       White1,215 (57.3%)3,655 (68.5%)231,877 (64.4%)5,746 (74.1%)21
       Missing17 (0.8%)60 (1.1%)377 (2.6%)234 (3.0%)2
       Other659 (31.1%)1,262 (23.7%)17794 (27.2%)1,384 (17.9%)23
      BMI, kg/m2
      BMI data missing on 2,741 patients in the prevalent cohort and 1,004 patients in the incident cohort.
      26 [23-30]26 [23-31]227 [24-32]28 [24-33]4
      Dialysis vintage, d412 ± 695940 ± 1,29851
      Cause of ESKD
       Cystic disease96 (4.5%)271 (5.1%)370 (2.4%)215 (2.8%)2
       Diabetes781 (36.9%)1,809 (33.9%)6995 (34.1%)2,469 (31.8%)5
       Glomerular disease325 (15.3%)872 (16.3%)3265 (9.1%)717 (9.2%)1
       Renovascular397 (18.7%)963 (18.1%)3404 (13.9%)1,017 (13.1%)2
       Other207 (9.8%)744 (13.9%)13554 (19.0%)1,672 (21.6%)6
       Unknown313 (14.8%)676 (12.7%)6626 (21.5%)1,662 (21.4%)0
      Teaching hospital affiliation274 (12.9%)2,131 (39.9%)64246 (8.4%)3,253 (42.0%)84
      Visits in past y
       Primary care5 [2-13]5 [1-12]310 [5-19]9 [4-17]14
       Hospitalizations1.05 ± 1.451.17 ± 1.6181.12 ± 1.501.18 ± 1.634
       Emergency1.84 ± 2.852.21 ± 3.22122.53 ± 2.712.78 ± 3.368
      Comorbid conditions
       Abdominal aortic aneurysm repair16 (0.8%)48 (0.9%)215 (0.5%)91 (1.2%)7
       Alcoholism37 (1.7%)100 (1.9%)178 (2.7%)242 (3.1%)3
       Depression<612 (0.2%)NR8 (0.3%)24 (0.3%)1
       Arrhythmia434 (20.5%)1,192 (22.3%)5503 (17.3%)1,538 (19.8%)7
       Pacemaker129 (6.1%)423 (7.9%)7182 (6.2%)601 (7.8%)6
       Defibrillator30 (1.4%)77 (1.4%)069 (2.4%)160 (2.1%)2
       Atrial fibrillation286 (13.5%)812 (15.2%)5333 (11.4%)1,076 (13.9%)7
       Cancer771 (36.4%)2,124 (39.8%)71,113 (38.2%)3,039 (39.2%)2
       Liver disease278 (13.1%)646 (12.1%)3366 (12.6%)921 (11.9%)2
       Lung disease748 (35.3%)1,921 (36.0%)11,016 (34.9%)2,672 (34.5%)1
       CAD1,299 (61.3%)3,049 (57.2%)81,498 (51.4%)3,597 (46.4%)10
       CABG103 (4.9%)234 (4.4%)2123 (4.2%)288 (3.7%)3
       MI300 (14.2%)751 (14.1%)0325 (11.2%)878 (11.3%)1
       PCI135 (6.4%)395 (7.4%)4186 (6.4%)449 (5.8%)2
       Dementia252 (11.9%)713 (13.4%)4325 (11.2%)809 (10.4%)2
       HF941 (44.4%)2,325 (43.6%)21,259 (43.2%)3,325 (42.9%)1
       Hypothyroidism88 (4.2%)243 (4.6%)2108 (3.7%)289 (3.7%)0
       HTN1,885 (89.0%)4,634 (86.9%)62,439 (83.7%)6,443 (83.1%)2
       Diabetes1,291 (60.9%)3,148 (59.0%)41,846 (63.3%)4,770 (61.5%)4
       PVD371 (17.5%)936 (17.5%)0200 (6.9%)501 (6.5%)2
       Stroke137 (6.5%)354 (6.6%)1143 (4.9%)378 (4.9%)0
       Venous thrombus93 (4.4%)305 (5.7%)6154 (5.3%)394 (5.1%)1
       Dyslipidemia567 (26.8%)1,210 (22.7%)9772 (26.5%)1,943 (25.1%)3
       Drug use73 (3.4%)288 (5.4%)10180 (6.2%)650 (8.4%)9
       Obesity62 (2.9%)165 (3.1%)1135 (4.6%)404 (5.2%)3
      ADG score
       0-4133 (6.3%)364 (6.8%)2122 (4.2%)379 (4.9%)3
       5-9633 (29.9%)1,519 (28.5%)3734 (25.2%)2,226 (28.7%)8
       10-14938 (44.3%)2,380 (44.6%)11,435 (49.2%)3,630 (46.8%)5
       15-19393 (18.5%)1,025 (19.2%)2595 (20.4%)1,442 (18.6%)5
       20+22 (1.0%)47 (0.9%)228 (1.0%)75 (1.0%)0
      Note: Data expressed as number (percent), mean ± standard deviation, or median [interquartile range], as appropriate.
      Abbreviations: ADG, Aggregated Diagnostic Groups; BMI, body mass index; CABG, coronary artery bypass graft; CAD, coronary artery disease; ESKD, end-stage kidney disease; HF, heart failure; HTN, hypertension; MI, myocardial infarction; PCI, percutaneous coronary intervention; PVD, peripheral vascular disease; NR, not reportable (ICES regulations prohibit the reporting of data pertaining to cell sizes < 6 individuals); StD, standardized difference expressed as percent.
      a Income quintile data missing on 46 patients in the prevalent cohort and 40 patients in the incident cohort.
      b BMI data missing on 2,741 patients in the prevalent cohort and 1,004 patients in the incident cohort.

      Blood Work Frequency and All-Cause Mortality

      Among prevalent HD recipients, there was no evidence that more frequent blood sampling was associated with a lower risk for all-cause mortality. Mortality tended to be higher in those exposed to monthly blood sampling (185 vs 150 per 1,000 patient-years at centers with testing every 6 weeks; unadjusted hazard ratio [HR], 1.24 [95% CI, 0.99-1.54]). Results were similar after adjustment (adjusted HR of 1.16 [95% CI, 0.99-1.38] in comparison to patients treated at centers with testing every 6 weeks; Table 2) . The proportional hazards assumption was not violated in the model for the prevalent cohort.
      Table 2Primary and Secondary Outcomes: Prevalent Patients
      Testing Every 6 wk (n = 2,119)Monthly Testing (n = 5,335)Unadjusted HR (95% CI)Adjusted HR
      For cardiovascular events, hospitalizations, ED visits, and episodes of hyperkalemia, adjustment was made for number of prior events.
      ,
      Models further adjusted for age, sex, income quintile (missing values imputed as middle quintile), time since dialysis initiation, cause of end-stage kidney disease, number of hospitalizations in preceding year, number of ED visits in preceding year, pre-existing comorbid conditions (abdominal aortic aneurysm repair, alcoholism, depression, arrhythmia, pacemaker implantation, defibrillator implantation, atrial fibrillation, cancer, liver disease, lung disease, coronary artery disease, prior coronary artery bypass graft surgery, myocardial infarction, prior percutaneous coronary intervention, dementia, heart failure, hypothyroidism, hypertension, diabetes mellitus, peripheral vascular disease, stroke, venous thrombus, dyslipidemia, illicit drug use, obesity), and the Aggregated Diagnostic Groups score.
      (95% CI)
      Pts With Events
      Reflects proportion of patients with at least 1 event.
      Incidence
      Per 1,000 patient-years.
      Pts With Events
      Reflects proportion of patients with at least 1 event.
      Incidence
      Per 1,000 patient-years.
      All-cause mortality1,094 (52%)150.03,106 (58%)185.31.24 (0.99-1.54)1.16 (0.99-1.38)
      Cardiovascular events561 (26%)95.61,412 (26%)87.01.08 (0.95-1.23)1.11 (0.95-1.29)
      All-cause hospitalization1,802 (85%)614.14,537 (85%)651.11.03 (0.93-1.13)1.03 (0.94-1.13)
      All-cause ED visits1,830 (86%)784.04,599 (86%)858.71.00 (0.98-1.02)1.03 (0.87-1.27)
      Hyperkalemia123 (6%)17.6362 (7%)22.61.26 (0.74-2.14)1.33 (0.87-2.04)
      Abbreviations: CI, confidence interval; ED, emergency department; HR, hazard ratio; Pt, patient.
      a For cardiovascular events, hospitalizations, ED visits, and episodes of hyperkalemia, adjustment was made for number of prior events.
      b Models further adjusted for age, sex, income quintile (missing values imputed as middle quintile), time since dialysis initiation, cause of end-stage kidney disease, number of hospitalizations in preceding year, number of ED visits in preceding year, pre-existing comorbid conditions (abdominal aortic aneurysm repair, alcoholism, depression, arrhythmia, pacemaker implantation, defibrillator implantation, atrial fibrillation, cancer, liver disease, lung disease, coronary artery disease, prior coronary artery bypass graft surgery, myocardial infarction, prior percutaneous coronary intervention, dementia, heart failure, hypothyroidism, hypertension, diabetes mellitus, peripheral vascular disease, stroke, venous thrombus, dyslipidemia, illicit drug use, obesity), and the Aggregated Diagnostic Groups score.
      c Reflects proportion of patients with at least 1 event.
      d Per 1,000 patient-years.
      In the incident cohort, crude time to mortality was not significantly different in patients treated at centers with monthly testing compared with centers with testing every 6 weeks (205 vs 175 per 1,000 patient-years; unadjusted HR, 1.15 [95% CI, 0.91-1.46]), with a similar relationship seen after adjustment (adjusted HR, 1.15; 95% CI, 0.96-1.37; Table 3). Because the proportional hazards assumption was violated in this cohort, we derived time-stratified HRs: in the year after HD initiation, the HR for all-cause mortality was 1.03 (95% CI, 0.82-1.29); however, beyond 1 year, exposure to monthly surveillance blood sampling was associated with a higher risk for death (HR, 1.31; 95% CI, 1.14-1.50).
      Table 3Primary and Secondary Outcomes: Incident Patients
      Testing Every 6 wk (n = 2,914)Monthly Testing (n = 7,752)Unadjusted HR (95% CI)Adjusted HR
      For cardiovascular events, hospitalizations, ED visits, and episodes of hyperkalemia, adjustment was made for number of prior events.
      ,
      Models further adjusted for age, sex, income quintile (missing values imputed as middle quintile), time since dialysis initiation, cause of end-stage kidney disease, number of hospitalizations in preceding year, number of ED visits in preceding year, pre-existing comorbidities (abdominal aortic aneurysm repair, alcoholism, depression, arrhythmia, pacemaker implantation, defibrillator implantation, atrial fibrillation, cancer, liver disease, lung disease, coronary artery disease, prior coronary artery bypass graft surgery, myocardial infarction, prior percutaneous coronary intervention, dementia, heart failure, hypothyroidism, hypertension, diabetes mellitus, peripheral vascular disease, stroke, venous thrombus, dyslipidemia, illicit drug use, obesity), and the Aggregated Diagnostic Groups score.
      (95% CI)
      Pts With Events
      Reflects proportion of patients with at least 1 event.
      Incidence
      Per 1,000 patient-years.
      Pts With Events
      Reflects proportion of patients with at least 1 event.
      Incidence
      Per 1,000 patient-years.
      All-cause mortality1,066 (37%)175.43,155 (41%)204.91.15 (0.91-1.46)1.15 (0.96-1.37)
      Cardiovascular events469 (16%)86.41,385 (18%)100.01.17 (0.96-1.43)1.18 (1.00-1.39)
      All-cause hospitalization2,269 (78%)990.56,220 (80%)1,051.91.10 (1.01-1.20)1.11 (1.03-1.20)
      All-cause ED visits2,152 (74%)1,038.65,926 (76%)1,133.61.29 (0.93-1.37)1.12 (0.99-1.27)
      Hyperkalemia85 (3%)14.3316 (4)21.01.18 (0.94-1.49)1.20 (1.00-1.44)
      Abbreviations: CI, confidence interval; ED, emergency department; HR, hazard ratio; Pt, patient.
      a For cardiovascular events, hospitalizations, ED visits, and episodes of hyperkalemia, adjustment was made for number of prior events.
      b Models further adjusted for age, sex, income quintile (missing values imputed as middle quintile), time since dialysis initiation, cause of end-stage kidney disease, number of hospitalizations in preceding year, number of ED visits in preceding year, pre-existing comorbidities (abdominal aortic aneurysm repair, alcoholism, depression, arrhythmia, pacemaker implantation, defibrillator implantation, atrial fibrillation, cancer, liver disease, lung disease, coronary artery disease, prior coronary artery bypass graft surgery, myocardial infarction, prior percutaneous coronary intervention, dementia, heart failure, hypothyroidism, hypertension, diabetes mellitus, peripheral vascular disease, stroke, venous thrombus, dyslipidemia, illicit drug use, obesity), and the Aggregated Diagnostic Groups score.
      c Reflects proportion of patients with at least 1 event.
      d Per 1,000 patient-years.

      Blood Work Frequency and Secondary Outcomes

      In the prevalent cohort, there was no evidence that blood work frequency was associated with any of the prespecified secondary outcomes (Table 2). However, among incident patients, exposure to monthly blood sampling was associated with a higher risk for cardiovascular events (adjusted HR, 1.18; 95% CI, 1.00-1.39), all-cause hospitalization (adjusted HR, 1.11; 95% CI, 1.03-1.20), and episodes of hyperkalemia (adjusted HR, 1.20; 95% CI, 1.00-1.44; Table 3).

      Sensitivity Analysis

      We reanalyzed our patient populations with the inclusion of patients who received care at the 2 dialysis programs that transitioned from monthly testing to every 6 weeks. These patients were considered to have been exposed to a monthly blood work policy to reflect practice in those units as of April 1, 2011. Inclusion of these patients did not substantially alter the relationship between blood work frequency and time to death in the prevalent (adjusted HR, 1.15; 95% CI, 0.97-1.36) or incident (adjusted HR, 1.14; 95% CI, 0.96-1.36) cohorts, respectively.

      Subgroup Analyses

      Mortality was accentuated in patients exposed to monthly testing at teaching hospitals (interaction P < 0.0001 and P< 0.02 in the prevalent and incident cohorts, respectively) (Fig 2). Similarly, individuals younger than 65 years in the incident cohort who were exposed to monthly blood testing were at higher risk for death (interaction P = 0.02).
      Figure thumbnail gr2
      Figure 2Unadjusted relationship between high-frequency (Freq) surveillance blood sampling and mortality in select subgroups of (A) prevalent and (B) incident hemodialysis recipients. (A, B) P value denotes interaction term between blood work frequency and subgroup variable.

      Discussion

      Maintenance HD recipients who were treated at centers in which routine blood sampling occurred at monthly intervals did not experience improved survival as compared with patients receiving care at centers in which routine blood work was performed every 6 weeks. Incident patients receiving care at centers with a policy of more frequent blood testing were more likely to be hospitalized and experience cardiovascular events. These results call into question current practices for routine surveillance blood work in the HD population.
      Though commonly referred to as “monthly blood work,” there is no meaningful evidence to guide an optimal sampling frequency for surveillance blood work in HD recipients. Ontario’s 27 HD programs operate under independent medical leadership, each with the prerogative to establish site-specific practices. Whereas 70% of programs performed monthly blood work as their usual practice, the rest have adopted an interval of 6 weeks. This created a unique opportunity to assess the association between blood testing frequency and outcomes.
      The existing literature on optimal blood sampling frequency in the dialysis population has focused exclusively on the association between different monitoring strategies and the achievement of prespecified targets for markers of interest. Gaweda et al
      • Gaweda A.E.
      • Nathanson B.H.
      • Jacobs A.A.
      • Aronoff G.R.
      • Germain M.J.
      • Brier M.E.
      Determining optimum hemoglobin sampling for anemia management from every-treatment data.
      examined the optimal frequency of hemoglobin sampling and concluded that weekly measurements associated with the lowest degree of hemoglobin variability. Greenberg et al
      • Greenberg S
      • Gadde S
      • Pagala M
      • Greenberg M
      • Shneyderman I
      • Janga K
      Optimal frequency of parathyroid hormone monitoring in chronic hemodialysis patients.
      compared quarterly versus monthly monitoring policies for parathyroid hormone (PTH) and demonstrated that monthly measurement was associated with a higher likelihood of readings being within the recommended range. Yokoyama et al
      • Yokoyama K.
      • Kurita N.
      • Fukuma S.
      • et al.
      Frequent monitoring of mineral metabolism in hemodialysis patients with secondary hyperparathyroidism: associations with achievement of treatment goals and with adjustments in therapy.
      concluded that in patients for whom serum CKD-MBD markers exceeded target range, weekly monitoring of calcium and monthly monitoring of PTH levels were associated with achievement of guideline-recommended targets. They found no evidence that more frequent measurements were of value when calcium, phosphorus, and PTH were already in target range.
      • Yokoyama K.
      • Kurita N.
      • Fukuma S.
      • et al.
      Frequent monitoring of mineral metabolism in hemodialysis patients with secondary hyperparathyroidism: associations with achievement of treatment goals and with adjustments in therapy.
      Recent work by our group highlighted the stability of target achievement for anemia and CKD-MBD parameters after a de-escalation of blood sampling frequency from monthly to every 6 weeks at a single dialysis program in Ontario.
      • Silver S.A.
      • Alaryni A.
      • Alghamdi A.
      • Digby G.
      • Wald R.
      • Iliescu E.
      Routine laboratory testing every 4 versus every 6 weeks for patients on maintenance hemodialysis: a quality improvement project.
      No change in mortality was observed after this change in practice.
      Clinical practice guidelines are likely to influence blood sampling protocols adopted by individual HD programs and providers. The 2012 KDIGO guideline on anemia in CKD included an ungraded recommendation for a minimum of monthly sampling of hemoglobin in anemic patients or those receiving therapy with an erythropoiesis-stimulating agent.
      KDIGO clinical practice guideline for anemia in chronic kidney disease.
      These recommendations have been largely echoed or endorsed by the Kidney Health Australia–Caring for Australasians With Renal Impairment,
      • McMahon L.P.
      • MacGinley R.
      • Kha C.
      KHA-CARI guideline: biochemical and haematological targets: haemoglobin concentrations in patients using erythropoietin-stimulating agents.
      the Canadian Society of Nephrology,
      • Moist L.M.
      • Troyanov S.
      • White C.T.
      • et al.
      Canadian Society of Nephrology commentary on the 2012 KDIGO clinical practice guideline for anemia in CKD.
      and the European Dialysis and Transplant Association.
      • Locatelli F.
      • Barany P.
      • Covic A.
      • et al.
      Kidney Disease: Improving Global Outcomes guidelines on anaemia management in chronic kidney disease: a European Renal Best Practice position statement.
      KDIGO issued ungraded guideline statements suggesting monthly to quarterly testing of calcium and phosphate and quarterly to biannual testing of PTH in dialysis recipients.
      KDIGO clinical practice guideline for the diagnosis, evaluation, prevention, and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
      ,
      KDIGO 2017 clinical practice guideline update for the diagnosis, evaluation, prevention and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
      These were also endorsed internationally.
      • Goldsmith D.J.
      • Covic A.
      • Fouque D.
      • et al.
      Endorsement of the Kidney Disease Improving Global Outcomes (KDIGO) chronic kidney disease-mineral and bone disorder (CKD-MBD) guidelines: a European Renal Best Practice (ERBP) commentary statement.
      • Manns B.J.
      • Hodsman A.
      • Zimmerman D.L.
      • et al.
      Canadian Society of Nephrology commentary on the 2009 KDIGO clinical practice guideline for the diagnosis, evaluation, and treatment of CKD-mineral and bone disorder (CKD-MBD).
      • Isakova T.
      • Nickolas T.L.
      • Denburg M.
      • et al.
      KDOQI US Commentary on the 2017 KDIGO clinical practice guideline update for the diagnosis, evaluation, prevention, and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
      • Uhlig K.
      • Berns J.S.
      • Kestenbaum B.
      • et al.
      KDOQI US commentary on the 2009 KDIGO clinical practice guideline for the diagnosis, evaluation, and treatment of CKD-mineral and bone disorder (CKD-MBD).
      Our findings challenge the fundamental notion that more frequent blood sampling improves clinical outcomes in HD recipients. Whereas the majority of patients in Ontario and those receiving care at large dialysis providers in the United States are exposed to sampling of routine parameters every month, we found that this approach did not yield better outcomes as compared with a strategy of blood sampling every 6 weeks. Though our results do not suggest an optimal sampling frequency for routine blood work, it should provide patients, clinicians, and administrators with a measure of confidence that widening blood testing intervals to 6 weeks may not compromise outcomes.
      If confirmed in randomized trials, our findings suggest an opportunity for a 30% reduction in the routine blood testing that is performed in maintenance HD recipients. This would represent a resource savings that could be redirected to higher value endeavors such as addressing patient-reported symptoms, transplant access, and volume optimization. Moreover, a reduction in blood tests, and the invariable need to review and act on the results, may refocus clinicians’ attention to issues that are of pressing concern to patients. Surveys have shown that maintenance dialysis recipients and their caregivers prioritize symptom control and quality of life and lifestyle over the control of biochemical and hematologic parameters.
      • Urquhart-Secord R.
      • Craig J.C.
      • Hemmelgarn B.
      • et al.
      Patient and caregiver priorities for outcomes in hemodialysis: an international nominal group technique study.
      ,
      • Tong A.
      • Manns B.
      • Hemmelgarn B.
      • et al.
      Establishing core outcome domains in hemodialysis: report of the Standardized Outcomes in Nephrology-Hemodialysis (SONG-HD) Consensus Workshop.
      Our study has several strengths. We conducted a rigorous survey of all HD programs in our province to ensure that laboratory sampling frequency was accurately classified. Our study population encompassed all Ontario residents, all of whom had free access to HD care and laboratory testing in our publicly funded health care system, thereby ensuring the inclusion of a broad and diverse patient population. Our single-payer universal health care system, and robust network of linked health care databases, enabled a granular characterization of the study cohorts with complete and reliable ascertainment of clinical outcomes. Whereas previous studies examining blood testing frequency focused on surrogate outcomes (eg, proportion of patients meeting guideline-based biochemical and hematologic targets), we examined clinical outcomes that are of unequivocal importance to patients and their health care providers.
      There are also several weaknesses to acknowledge. We were not able to account for potential variability in clinician or center-wide practices in response to blood results. As such, any conceivable benefit of more frequent blood sampling may have been obscured by unmeasurable practice differences between centers. Despite accounting for center-specific effects and adjusting for a broad array of relevant confounders in our analyses, it is possible that patients receiving care at programs with monthly (high-frequency) laboratory sampling were sicker and more complex. Though the majority of baseline covariates demonstrated good balance between the 2 groups, it is notable that patients treated at low-frequency centers had a higher proportion of African Canadian patients. This may have biased our findings in favor of those receiving blood testing every 6 weeks, noting the well-described survival advantage of African American dialysis recipients.
      • Agodoa L.
      • Eggers P.
      Racial and ethnic disparities in end-stage kidney failure-survival paradoxes in African-Americans.
      ,
      • Crews D.C.
      • Sozio S.M.
      • Liu Y.
      • Coresh J.
      • Powe N.R.
      Inflammation and the paradox of racial differences in dialysis survival.
      In addition, programs performing more frequent blood sampling were more likely to have a teaching hospital affiliation. Teaching hospitals may attract more complex patients who are susceptible to inferior clinical outcomes. There were likely other unknown and unmeasured center-specific practices, including administration of medications, that may have led to confounding.
      Finally, we did not measure the frequency of blood work done outside of routine monitoring. There were no restrictions on clinicians ordering additional blood work outside of the dialysis program’s pre-set interval, and it is possible that clinicians practicing at centers in which routine blood sampling was done every 6 weeks “compensated” for the lower blood work frequency by ordering more blood tests between routine blood draws. This would potentially offset any resource savings associated with a lower frequency blood sampling policy.
      Our findings do not allow us to recommend an optimal frequency for routine blood testing that balances patient safety and resource consumption. We are also unable to make a definitive statement about whether testing every 6 weeks provides noninferior outcomes as compared to monthly testing. Finally, despite the size and diversity of Ontario, Canada, our results may not be applicable to other jurisdictions.
      In conclusion, monthly sampling compared to sampling of blood parameters at 6-week intervals did not show evidence of associating with lower mortality, a reduction in cardiovascular events, or fewer health care encounters in maintenance HD recipients. HD programs that perform routine blood work sampling on a monthly basis may consider a de-escalation of their sampling interval. Randomized controlled trials are needed to define a frequency of blood testing that will be both optimally beneficial to the health of HD recipients and maximally impactful in the use of health care resources.

      Article Information

      Authors’ Full Names and Academic Degrees

      Alison Thomas, MN, Samuel A. Silver, MD, Jeffrey Perl, MD, SM, Megan Freeman, BA, Justin J. Slater, MSc, Danielle M. Nash, PhD, Marlee Vinegar, MPH, Eric McArthur, MSc, Amit X. Garg, MD, PhD, Ziv Harel, MD, MSc, Rahul Chanchlani, MD, Michael Zappitelli, MDCM, Eduard Iliescu, MD, Abhijat Kitchlu, MD, MSc, Daniel Blum, MDCM, William Beaubien-Souligny, MD, and Ron Wald, MDCM, MPH.

      Authors’ Contributions

      Concept: AT, SAS, JP, RW; study design: AT, SAS, JP, MF, DMN, AXG, WB-S, RW; analysis: JJS, EM, MV; interpretation: ZH, RC, MZ, EI, AK, DB. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved

      Support

      This study was funded by an unrestricted grant provided by the Ontario Renal Network and conducted under the auspices of the ICES Kidney Dialysis Transplantation Program at ICES Western. This study was supported by the ICES Western site. ICES is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). Core funding for ICES Western is provided by the Academic Medical Organization of Southwestern Ontario (AMOSO), the Schulich School of Medicine and Dentistry (SSMD), Western University, and the Lawson Health Research Institute (LHRI). This research was conducted by members of the ICES Kidney, Dialysis and Transplantation team, at the ICES Western facility, who are supported by a grant from the Canadian Institutes of Health Research (CIHR). Parts of this material are based on data and/or information compiled and provided by CIHI. Dr Garg was supported by the Dr. Adam Linton Chair in Kidney Health Analytics and a Clinician Investigator Award from the Canadian Institutes of Health Research . Dr Silver is supported by the Kidney Research Scientist Core Education and National Training (KRESCENT) Program New Investigator Award (co-funded by the Kidney Foundation of Canada , Canadian Society of Nephrology, and CIHR). The funders had no role in in study design; data collection, analysis, or reporting; or the decision to submit for publication.

      Financial Disclosure

      The authors declare that they have no relevant financial interests.

      Disclaimer

      The opinions, results, and conclusions are those of the authors and are independent from the funding and data sources. No endorsement by ICES, AMOSO, SSMD, LHRI, CIHR, CIHI, or the MOHLTC is intended or should be inferred.

      Peer Review

      Received June 6, 2019, as a submission to the expedited consideration track with 4 external peer reviews. Direct editorial input from a Statistics/Methods Editor, a Deputy Editor, and the Editor-in-Chief. Accepted in revised form August 20, 2019. Further information on expedited consideration (AJKD Express) is available in the Information for Authors & Journal Policies.

      Supplementary Material

      References

        • Locatelli F.
        • Nissenson A.R.
        • Barrett B.J.
        • et al.
        Clinical practice guidelines for anemia in chronic kidney disease: problems and solutions. A position statement from Kidney Disease: Improving Global Outcomes (KDIGO).
        Kidney Int. 2008; 74: 1237-1240
        • Jindal K.
        • Chan C.T.
        • Deziel C.
        • et al.
        Hemodialysis clinical practice guidelines for the Canadian Society of Nephrology.
        J Am Soc Nephrol. 2006; 17: S1-S27
      1. KDIGO clinical practice guideline for the diagnosis, evaluation, prevention, and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
        Kidney Int Suppl. 2009; 76: S1-S130
      2. KDIGO 2017 clinical practice guideline update for the diagnosis, evaluation, prevention and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
        Kidney Int Suppl. 2017; 7: 1-59
        • Naugler C.
        • Ma I.
        More than half of abnormal results from laboratory tests ordered by family physicians could be false-positive.
        Can Fam Physician. 2018; 64: 202-203
        • Urquhart-Secord R.
        • Craig J.C.
        • Hemmelgarn B.
        • et al.
        Patient and caregiver priorities for outcomes in hemodialysis: an international nominal group technique study.
        Am J Kidney Dis. 2016; 68: 444-454
        • Chan E
        • Hemmelgarn B
        • Klarenbach S
        • et al.
        Choosing Wisely: the Canadian Society of Nephrology’s list of 5 items physicians and patients should question.
        Can J Kidney Health Dis. 2017; 4https://doi.org/10.1177/2054358117695570
        • Williams A.W.
        • Dwyer A.C.
        • Eddy A.A.
        • et al.
        Critical and honest conversations: the evidence behind the “Choosing Wisely” campaign recommendations by the American Society of Nephrology.
        Clin J Am Soc Nephrol. 2012; 7: 1664-1672
        • Benchimol E.I.
        • Smeeth L.
        • Guttmann A.
        • et al.
        The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement.
        PLoS Med. 2015; 12e1001885
      3. The Johns Hopkins ACG Case-Mix System Version 10.0 Release Notes. Johns Hopkins University, Baltimore, MD2011
        • Austin P.C.
        Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research.
        Commun Stat Simul Comput. 2009; 38: 1228-1234
        • Lin D.Y.
        • Wei L.J.
        The robust inference for the Cox proportional hazards model.
        J Am Stat Assoc. 1989; 84: 1074-1078
        • Lin D.Y.
        • Wei L.J.
        • Ying Z.
        Model-checking techniques based on cumulative residuals.
        Biometrics. 2002; 58: 1-12
        • Amorim L.D.
        • Cai J.
        Modelling recurrent events: a tutorial for analysis in epidemiology.
        Int J Epidemiol. 2015; 44: 324-333
        • Gaweda A.E.
        • Nathanson B.H.
        • Jacobs A.A.
        • Aronoff G.R.
        • Germain M.J.
        • Brier M.E.
        Determining optimum hemoglobin sampling for anemia management from every-treatment data.
        Clin J Am Soc Nephrol. 2010; 5: 1939-1945
        • Greenberg S
        • Gadde S
        • Pagala M
        • Greenberg M
        • Shneyderman I
        • Janga K
        Optimal frequency of parathyroid hormone monitoring in chronic hemodialysis patients.
        Clin Nephrol. 2011; : 348-353
        • Yokoyama K.
        • Kurita N.
        • Fukuma S.
        • et al.
        Frequent monitoring of mineral metabolism in hemodialysis patients with secondary hyperparathyroidism: associations with achievement of treatment goals and with adjustments in therapy.
        Nephrol Dial Transplant. 2017; 32: 534-541
        • Silver S.A.
        • Alaryni A.
        • Alghamdi A.
        • Digby G.
        • Wald R.
        • Iliescu E.
        Routine laboratory testing every 4 versus every 6 weeks for patients on maintenance hemodialysis: a quality improvement project.
        Am J Kidney Dis. 2019; 73: 496-503
      4. KDIGO clinical practice guideline for anemia in chronic kidney disease.
        Kidney Int Suppl. 2012; 2: 279-335
        • McMahon L.P.
        • MacGinley R.
        • Kha C.
        KHA-CARI guideline: biochemical and haematological targets: haemoglobin concentrations in patients using erythropoietin-stimulating agents.
        Nephrology (Carlton). 2012; 17: 17-19
        • Moist L.M.
        • Troyanov S.
        • White C.T.
        • et al.
        Canadian Society of Nephrology commentary on the 2012 KDIGO clinical practice guideline for anemia in CKD.
        Am J Kidney Dis. 2013; 62: 860-873
        • Locatelli F.
        • Barany P.
        • Covic A.
        • et al.
        Kidney Disease: Improving Global Outcomes guidelines on anaemia management in chronic kidney disease: a European Renal Best Practice position statement.
        Nephrol Dial Transplant. 2013; 28: 1346-1359
        • Goldsmith D.J.
        • Covic A.
        • Fouque D.
        • et al.
        Endorsement of the Kidney Disease Improving Global Outcomes (KDIGO) chronic kidney disease-mineral and bone disorder (CKD-MBD) guidelines: a European Renal Best Practice (ERBP) commentary statement.
        Nephrol Dial Transplant. 2010; 25: 3823-3831
        • Manns B.J.
        • Hodsman A.
        • Zimmerman D.L.
        • et al.
        Canadian Society of Nephrology commentary on the 2009 KDIGO clinical practice guideline for the diagnosis, evaluation, and treatment of CKD-mineral and bone disorder (CKD-MBD).
        Am J Kidney Dis. 2010; 55: 800-812
        • Isakova T.
        • Nickolas T.L.
        • Denburg M.
        • et al.
        KDOQI US Commentary on the 2017 KDIGO clinical practice guideline update for the diagnosis, evaluation, prevention, and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD).
        Am J Kidney Dis. 2017; 70: 737-751
        • Uhlig K.
        • Berns J.S.
        • Kestenbaum B.
        • et al.
        KDOQI US commentary on the 2009 KDIGO clinical practice guideline for the diagnosis, evaluation, and treatment of CKD-mineral and bone disorder (CKD-MBD).
        Am J Kidney Dis. 2010; 55: 773-799
        • Tong A.
        • Manns B.
        • Hemmelgarn B.
        • et al.
        Establishing core outcome domains in hemodialysis: report of the Standardized Outcomes in Nephrology-Hemodialysis (SONG-HD) Consensus Workshop.
        Am J Kidney Dis. 2017; 69: 97-107
        • Agodoa L.
        • Eggers P.
        Racial and ethnic disparities in end-stage kidney failure-survival paradoxes in African-Americans.
        Semin Dial. 2007; 20: 577-585
        • Crews D.C.
        • Sozio S.M.
        • Liu Y.
        • Coresh J.
        • Powe N.R.
        Inflammation and the paradox of racial differences in dialysis survival.
        J Am Soc Nephrol. 2011; 22: 2279-2286

      Linked Article

      • Routine Monthly Blood Draws in Hemodialysis: Where Is the Evidence?
        American Journal of Kidney DiseasesVol. 75Issue 4
        • Preview
          Beginning in the late 19th century, technologic innovation gave results of laboratory investigations an increasingly prominent role in diagnosis and treatment. Our current concept of chronic kidney disease, for example, depends largely on laboratory findings.1 However, the optimal frequency of testing has been established rigorously in few if any clinical situations. In particular, the practice of testing dialysis patients monthly, rather than at some other frequency, does not reflect any evidence that this practice optimizes either patient-centered outcomes or incremental cost-effectiveness.
        • Full-Text
        • PDF