ICURT PROCEEDINGS

Timing of Dialysis Initiation: When to Start? Which

Treatment?
Paul Leurs, MD, PhD,*, Anna Machowska, MSc, and Bengt Lindholm, MD, PhD
During the late 1990s early initiation of dialysis was introduced on a large scale and between 1996 and 2008, the percentage of patients
with an estimated glomerular filtration rate (eGFR) above 10 mL/minute starting dialysis rose from 25% to 54% in the United States. However, several subsequent studies showed no survival benefit for patients commencing dialysis earlier. One possible explanation for the
negative results could be that eGFR may be a flawed index; s-creatinine is lower in patients with muscle wasting or fluid overload and
these vulnerable patients with high comorbidity burden often start early, i.e., at higher eGFR. Another explanation could be that dialysis
is in fact harmful; dialysis initiation with conventional thrice weekly in-center hemodialysis clearly associates with increased initial mortality
risk especially when using temporary dialysis catheters. Interestingly, patients starting on peritoneal dialysis (PD) appear to have better
initial outcomes. More attention should be given to finding new objective mortality-predictive markers of uremia, reducing the use of temporary hemodialysis catheters, and increasing the use of PD as initial dialysis modality. PD may not only provide better initial dialysis outcomes but may also preserve renal function and vessels for vascular access for the benefit of better long-term outcomes.
2015 by the National Kidney Foundation, Inc. All rights reserved.

N PATIENTS WITH chronic kidney disease (CKD), the

risk of all-cause and cardiovascular mortality increases
with decline in renal function, especially when the glomerular
filtration rate (GFR) decreases below 60 mL/minute.1 As patients with CKD are more likely to die than progress to renal
replacement therapy,2 one may ask whether the high mortality risk in patients with CKD stages 4 to 5 can be reduced by
earlier dialysis initiation. On the other hand, if the dialysis procedure is associated with factors leading to increased risk of
morbidity and mortality,3 what is the added risk by starting
dialysis earlier rather than later and would in fact earlier dialysis initiation instead increase the mortality risk? In this review, we briefly discuss these questions and some of the
factors that could influence the decision on when to start
dialysis and the choice of dialysis treatment modality.

Early Dialysis Initiation: Not as Good as

Previously Thought
In the past, and still during the 1980s, resources for dialysis therapy were lacking, and dialysis was in general
*

Department of Nephrology, Admiraal de Ruyter Hospital, Goes, The

Netherlands.

Renal Medicine and Baxter Novum, Department of Clinical Science,

Intervention and Technology, Karolinska Institutet, Stockholm, Sweden.
Financial Disclosure: Baxter Novum is the result of a grant from Baxter
Healthcare Corporation to the Karolinska Institutet. B.L. and A.M. are affiliated
with Baxter Healthcare. P.L. does not declare any conflicts of interest.
Address correspondence to Bengt Lindholm, MD, PhD, Baxter Novum and
Renal Medicine, Department of Clinical Science, Intervention and Technology,
Karolinska Institutet, M99 Karolinska University Hospital Huddinge, 141 86
Stockholm, Sweden. E-mail: bengt.lindholm@ki.se
2015 by the National Kidney Foundation, Inc. All rights reserved.
1051-2276/$36.00
http://dx.doi.org/10.1053/j.jrn.2014.10.015

238

initiated only when patients demonstrated clear signs of

life-threatening uremic complications. In 1985, Bonomini
et al4 from Bologna, Italy, reported that earlier initiation of
dialysis could convey survival benefits; patients, who started
dialysis early at a mean creatinine clearance rate of 11 mL/
minute, had a higher survival rate than those who started
late at a clearance rate of less than 5 mL/minute. Although
this report did not appear to influence dialysis initiation
practices much, the publication of the CanadaUSA study5
in 1996, which reported that survival was greater in patients
starting peritoneal dialysis (PD) with a higher residual renal
function (RRF) resulted in major changes. Based on the
observation in the CanadaUSA study that a total (peritoneal plus renal) removal of urea corresponding to Kt/V
.2.0 per week associated with improved survival, the US
National Kidney Foundation-Dialysis Outcomes Quality
Initiative working group on the initiation of dialysis therapy
in 1997 proposed that the adequacy target for dialysis patients not yet on dialysis should not fall below this
threshold.6 Predialysis patients should have RRF equal to
a GFR of 10.5 mL/minute or more (corresponding to
weekly Kt/V .2.0); if this cut-off point was not achieved,
the patients should be started on dialysis6 This policy led to
a major change in the timing of dialysis initiation practices
in the United States andalthough to a lesser extentalso
in other countries. Thus, in the United States, between
1996 and 2008, the percentage of patients starting dialysis
with an estimated GFR (eGFR) above 10 mL/minute
rose from 25% to 54%, and the percentage of patients initiating dialysis at an eGFR above 15 mL/minute increased
from 4% to 17%.7
After these dramatic changes in the timing of dialysis initiation practices, several large observational studies were performed comparing outcomes in patients starting dialysis at

Journal of Renal Nutrition, Vol 25, No 2 (March), 2015: pp 238-241

239

INITIATION OF DIALYSIS

various levels of eGFR. These studies included large

numbers of patients; in some cases, more than 800,000, in
registry-type data sets, including the United States Renal
Data System (USRDS), Bureau of National Health Insurance in Taiwan, European Registry, the French Renal
Epidemiology and Information Network Registry, and the
Canadian Organ Replacement Registry as well as one randomized study, the Initiating Dialysis Early and Late
(IDEAL) study.7,8 However, perhaps unexpectedly, these
studies showed that an early start of dialysis had no
beneficial effect7-10 or even resulted in a worse outcome1113
; starting dialysis at lower levels of eGFR (5 late start)
thus appeared to be associated with lower mortality.
Altogether, these studies indicated that early start of dialysis
especially when using in center-hemodialysis (HD) as initial
therapy could be harmful and questioned the trend to early
dialysis initiation based primarily on eGFR; the clinical status
of the patients, besides eGFR levels, should also be taken into
account in the decision making process. Furthermore, a
meta-analysis concluded that the outcome after dialysis initiation was not only affected by GFR and patient characteristics but also appeared to associate with the type of dialysis
modality used; interestingly, although early dialysis initiation
using in-center HD associated with worse outcomes, this
was not a consistent finding when PD was used as initial
therapy.13

Possible Factors That May Explain Lack of

Favorable Effect of Early Dialysis Initiation
When analyzing factors explaining why starting dialysis
early appear to be harmful, first of all, it should be noted
that observational studies do notand cannotprove that
starting dialysis with higher eGFR is indeed causing the
observed worse outcomes. One important confounder is
that patients are only included in studies if they actually
started dialysis; only the fittest patients survive long enough
to be included in the late start groups; this is the survivor
bias. Furthermore, among patients who survived long
enough to take part in these studies, those with symptoms
or comorbidities are more likely to be started on dialysis
early. In addition, eGFR based on s-creatinine may overestimate true GFR in the latter patients who may have a low
s-creatinine concentration because of low muscle mass or
fluid overload, or both; these patients who have an
increased high mortality risk due to frailty and comorbidities are likely to be started earlier at a higher eGFR level.
In a study based on the Nederlands Co-operative Study Adequacy of Dialysis treatment cohort, it was reported that
although patients with lower eGFR had worse survival
than those with higher eGFR at start of dialysis, such a survival difference was not seen when comparing the same
patients according to their levels of renal function based
on measured GFR; muscle mass was found to associate
with eGFR but not with measured GFR.14

Another reason is that eGFRor any index based on

creatinine (or urea)may be a poor predictor of the
concentrations for a broad range of uremic toxins. Thus,
because eGFR is poorly associated with concentrations of
uremic toxins in patients with different degrees of CKD
and correlates differently with each individual solute,
eGFR cannot be considered representative for evaluating
the accumulation of solutes in the course of CKD.15
Furthermore, although eGFR (or measured GFR) is not
a good predictor of the concentrations of uremic solutes
or their biological action, other factors may be more
important such as tubular secretion of toxins, generation
of toxins by the intestinal flora, and the metabolism of
toxins.15 Because of the association between nutritional status and eGFR and the other reasons mentioned previously,
there is now consensus that renal function estimated by screatinine (eGFR) may be useless or even misleading as a
sole guide on when to start dialysis.

Harmful Effects of Dialysis

Another reason why early dialysis initiation could associate with worse outcomes could be that the dialysis procedure is harmful. Dialysis, and especially conventional
intermittent thrice-weekly HD, leads to an accelerated
loss of RRF, a powerful predictor of mortality in CKD patients. The rate of RRF loss should therefore be an important consideration for the timing of the dialysis initiation
decision and the choice of initial dialysis modality; several
studies have shown that RRF is better preserved in terms
of slower rate of decline in GFR and a longer time to loss
of RRF in patients treated with PD compared with those
treated with conventional thrice weekly HD. Furthermore,
dialysis introduces the risk of access-related infections, a
powerful predictor of worse outcomes, induces an inflammatory response andespecially when using intermittent
dialysisleads to unphysiological fluctuations in solutes
and fluid that increase the risk for sudden cardiac death,
the most common cause of death in dialysis patients according to USRDS data. In fact, mortality increases markedly
during the first month after initiation of HD suggesting
possible harmful effect of the dialysis procedure as such or
complications caused by dialysis.12 Potential biological factors contributing to poor outcomes on early dialysis initiation in HD and PD may be more often associated with HD
than with PD.9 Furthermore, mortality during the initial
years after start of dialysis was found to be similar or better
with PD than with in-center HD.16 A propensity-matched
mortality comparison of 6,337 incident HDPD patient
pairs showed that survival from the first day on dialysis
was 8% higher for PD patients than for HD patients (hazard
ratio 0.92; 95% confidence interval 0.86-1.00; P 5 .04).17
Furthermore, the latest USRDS report shows that the first
year mortality rate (in year 2010) was much lower in patients initiating dialysis with PD than with HD even
when adjusting for age, gender, race, and primary diagnosis;

240

LEURS ET AL

Table 1. Timing of Dialysis Initiation: When Should Dialysis Should Be Started? Which Treatment?
Planning for dialysis initiation
 Early referral to nephrologist is essential for planning initial dialysis modality;
 Planning for timely placement of dialysis access; use of dialysis catheters for HD is associated with high mortality risk.
When should dialysis be started?
 If uremic symptoms and signs are getting worse in patient with GFR ,15 mL/min/1.73 m2;
 Dialysis is usually indicated when GFR is 5 to 9 mL/min/1.73 m2, if the patient has uremic symptoms and signs;
 But, there is no single criterion.
Which treatment?
If no contraindications and if it is feasible and patient is able/willing:
1. renal transplantation,
2. home-based dialysis therapy, and, as last choice,
3. in-center hemodialysis; thrice weekly conventional HD is no longer the norm.

furthermore, there has been a dramatic improvement of the

first year (and the second year) survival rates in patients
starting on PD.18

Current Recommendations on Dialysis

Initiation
What have been the lessons learned and the impact of the
IDEAL study and the many observational studies showing
no benefit of early start of dialysis? First of all, it should
be noted that the dissemination of the results from these
studies have made an impact as according to USRDS, the
accelerating trend toward earlier and earlier initiation of
dialysis appears to have stopped after 2010.18 Secondly,
there is now consensus that no specific level of kidney function defined by an estimation of GFR based on serum creatinine should direct timing of the initiation of dialysis. The
European Renal Best Practice advisory board have
concluded that eGFR is not useful in determining need
for dialysis, that a GFR of 6 mL/minute/1.73 m2 is no
longer seen as an absolute lower limit to starting dialysis,
and that there may still be time preparing patients for
home dialysis therapy although GFR is low.10 Investigators
from the United States suggest that dialysis initiation is usually indicated at GFR levels of 5 to 9 mL/minute/1.73 m2 if
accompanied by uremia symptoms or fluid management issues (Table 1), but there is no single criterion.19 Third, it
should be noted that early versus late dialysis initiation
does not equal early versus late referral to nephrology
referral in CKD; early nephrology referral is reported to
associate with reduced mortality and hospitalization, better
uptake of PD, and earlier timely placement of arteriovenous
fistula for HD.20

Summary and Future Directions

Multiple factors need to be taken into account when
determining the optimal moment at which dialysis should
be started to optimize the prognosis; timing of dialysis initiation should not be determined solely based on eGFR.
eGFR is influenced by muscle mass and overhydration and
does not provide a good measure of uremic toxicity. Therefore, there is a need for novel mortality-predictive markers of

uremia including perhaps also protein-bound solutes such as

indoxyl sulfate and p-cresyl sulfate21; such novel markers
may potentially improve guidance on when to start dialysis
in the future. Still, however, clinical assessment of the patients
appears to remain the most important decision tool according to a survey among European nephrologists on factors
influencing the decision to start renal replacement therapy.22
But, there is a scarcity of studies documenting the exact reason(s) why a particular patient was initiated on dialysis at a
particular time; such studies are much needed. We await
further such studies on factors involved in the decision making to start dialysis as in the European Quality Study on when
dialysis should be initiated.23 Also, the effect of different dialysis modalities appear to be important; using conventional
thrice weekly HD therapy as initial dialysis therapy associates
with a substantial increase of mortality risk compared with
PD. Finally, it should be noted that although early dialysis
initiation may not be associated with improved survival, early
referral to nephrology care may not only improve outcomes
in terms of morbidity and mortality but also leaves enough
time for implementation of an appropriate dialysis access
and allows patients and their caregivers to choose the most
appropriate initial dialysis modality. Conventional thrice
weekly in-center HD is clearly not an ideal initial dialysis
modality (Table 1). Alternative approaches including more
frequent HD and home-based therapies (both PD and
home HD) are promising24 and according to the most recent
USRDS report these modalities are being increasingly used
in the United States.18

References
1. Matsushita K, van der Velde M, Astor BC, et al. Association of estimated
glomerular filtration rate and albuminuria with all-cause and cardiovascular
mortality in general population cohorts: a collaborative meta-analysis. Lancet.
2010;375:2073-2081.
2. Keith DS, Nichols GA, Gullion CM, et al. Longitudinal follow-up and
outcomes among a population with chronic kidney disease in a large managed
care organization. Arch Intern Med. 2004;164:659-663.
3. Collins AJ, Foley RN, Gilbertson DT, et al. The state of chronic kidney
disease, ESRD, and morbidity and mortality in the first year of dialysis. Clin J
Am Soc Nephrol. 2009;4(suppl 1):S5-S11.

INITIATION OF DIALYSIS
4. Bonomini V, Feletti C, Scolari MP, et al. Benefits of early initiation of
dialysis. Kidney Int Suppl. 1985;17:S57-S59.
5. Churchill DN, Taylor DW, Keshiviah PR. Canada-USA (CANUSA)
Peritoneal Dialysis Study Group: adequacy of dialysis and nutrition in continuous peritoneal dialysis: association with clinical outcomes. J Am Soc Nephrol.
1996;7:198-207.
6. NKF-KDOQI clinical practice guidelines. Am J Kidney Dis. 1997;30(3
suppl 2):S67-S136.
7. Rosansky SJ, Clark WF, Eggers P, et al. Initiation of dialysis at higher
GFRs: is the apparent rising tide of early dialysis harmful or helpful? Kidney
Int. 2009;76:257-261.
8. Cooper BA, Branley P, Bulfone L, et al. A randomized, controlled trial
of early versus late initiation of dialysis. N Engl J Med. 2010;363:609-619.
9. Molnar MZ, Ojo AO, Bunnapradist S, et al. Timing of dialysis initiation
in transplant-naive and failed transplant patients. Nat Rev Nephrol.
2012;8:284-292.
10. Tattersall J, Dekker F, Heimb
urger O, et al. When to start dialysis: updated guidance following publication of the Initiating Dialysis Early and Late
(IDEAL) study. Nephrol Dial Transplant. 2011;26:2082-2086.
11. Wright S, Klausner D, Baird B, et al. Timing of dialysis initiation and
survival in ESRD. Clin J Am Soc Nephrol. 2010;5:1828-1835.
12. Rosansky SJ, Eggers P, Jackson K, et al. Early start of hemodialysis may
be harmful. Arch Intern Med. 2011;171:396-403.
13. Susantitaphong P, Altamimi S, Ashkar M, et al. GFR at initiation of dialysis and mortality in CKD: a meta-analysis. Am J Kidney Dis. 2012;59:829-840.
14. Grootendorst DC, Michels WM, Richardson JD, et al. The MDRD
formula does not reflect GFR in ESRD patients. Nephrol Dial Transplant.
2011;26:1932-1937.

241

15. Eloot S, Schepers E, Barreto DV, et al. Estimated glomerular filtration

rate is a poor predictor of concentration for a broad range of uremic toxins.
Clin J Am Soc Nephrol. 2011;6:1266-1273.
16. Mehrotra R, Chiu YW, Kalantar-Zadeh K, Bargman J, Vonesh E.
Similar outcomes with hemodialysis and peritoneal dialysis in patients with
end-stage renal disease. Arch Intern Med. 2011;171:110-118.
17. Weinhandl ED, Foley RN, Gilbertson DT, Arneson TJ, Snyder JJ,
Collins AJ. Propensity-matched mortality comparison of incident hemodialysis and peritoneal dialysis patients. J Am Soc Nephrol. 2010;21:499-506.
18. United States Renal Data System, 2014 Annual Data Report: Epidemiology of Kidney Disease in the United States. National Institutes of Health,
National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda,
MD, 2014. www.usrds.org. Accessed December 4, 2014.
19. Rosansky S, Glassock RJ, Clark WF. Early start of dialysis: a critical review. Clin J Am Soc Nephrol. 2011;6:1222-1228.
20. Smart NA, Titus TT. Early versus late nephrology referral in CKD: a
systematic review. Am J Med. 2011;124:1073-1080.
21. Niwa T. Targeting protein-bound uremic toxins in chronic kidney disease. Expert Opin Ther Targets. 2013;17:1287-1301.
22. van de Luijtgaarden MW, Noordzij M, Tomson C, et al. Factors influencing the decision to start renal replacement therapy: results of a survey
among European nephrologists. Am J Kidney Dis. 2012;60:940-948.
23. Jager KJ, Ocak G, Drechsler C, et al. The EQUAL study: a European
study in chronic kidney disease stage 4 patients. Nephrol Dial Transplant.
2012;27(suppl 3):iii27-iii31.
24. Lindholm B, Davies S. End-stage renal disease in 2010: timing of dialysis initiation and choice of dialysis modality. Nat Rev Nephrol. 2011;7:
66-68.