Perceptions of pediatric nephrologists regarding timing of dialysis initiation in children in Canada
This is the first report of a national survey of pediatric nephrologists to evaluate
opinions and factors driving practice concerning dialysis initiation in children.
This study has importantly identified that only 50 % of pediatric nephrologists in
Canada consider GFR important when deciding the timing of dialysis initiation in their
patients. There is also significant variability in the threshold that nephrologists
consider early (vs. late) initiation with almost 60 % choosing a threshold above 15 ml/min/1.73 m
2
and only 12 % choosing 10 ml/min/1.73 m
2
. In contrast, the threshold at which practitioners typically start asymptomatic children on dialysis is between 10 and 14 ml/min/1.73 m
2
in most cases, a threshold now considered “high or early” in the adult literature.
Important clinical factors have emerged as more relevant triggers for dialysis initiation
by pediatric care providers including fatigue, weight loss, nausea, and increasing
school absenteeism. Awaiting a pre-emptive transplant is also an important factor
that can delay the decision to start dialysis. These findings shed important light
on the reasons for variability in practice across Canada previously reported by our
group 2], 3].
The previous pediatric literature on the topic of timing of dialysis initiation in
children includes retrospective analyses of large databases in North America 15], 16] and Europe 17], 18], which have reported that 50 % of children are started on dialysis with an eGFR 10 ml/min/1.73 m
2
and about 20 % started with an eGFR 15 ml/min/1.73 m
2
. Our group also reported, utilizing the Canadian Organ Replacement Register (CORR)
data, that 30 % of children in Canada are started on dialysis with an eGFR ?10.5 ml/min/1.73 m
2
. This study showed considerable practice variation across the country, with a range
of 12–70 % of children started with an eGFR 10.5 ml/min/1.73 m
2
by treatment facility 3]. The limitation of these previous studies is a lack of information that may have
influenced clinical decision-making, such as the clinical indication for dialysis
and the presence of uremic symptoms at the time of dialysis initiation.
This survey provides insight into the likely explanations for the previously identified
variation in practice based on eGFR, as practitioners are in fact not utilizing the
same target eGFR thresholds nor do they universally agree that eGFR should be an important
tool in the decision-making process to begin with. This suggests that future studies
on timing of dialysis initiation in children based on eGFR may be challenging to perform.
Although the IDEAL trial in adults has rigorously shown a lack of clinical benefit
to starting dialysis at a higher eGFR (eGFR 10–14 vs. 5–7 ml/min/1.73 m
2
) 13], the applicability of these findings to children is at this time unknown. While most
respondents did report changing their pediatric clinic practice based on adult literature,
the lack of translation of the adult trial evidence on timing of dialysis initiation
specifically into pediatric practice is explicit. This was demonstrated by two thirds
of the respondents stating the IDEAL trial did not alter their practice.
A range of opinions were uncovered in this survey, especially in regard to factors
influencing later initiation of dialysis. Respondent answers were mixed as to whether
late dialysis initiation increases hospitalizations, has negative impacts on the quality
of life, leads to sicker patients, or is cost-effective. This reflects the lack of
evidence in this area in pediatrics and a need for outcome studies to address these
significant knowledge gaps. The only outcome-based study that has been published to
date revealed a 21 % decreased risk of hospitalizations for hypertension and pulmonary
edema in children with higher baseline eGFRs at dialysis initiation 16]. However, this study was observational, and differences between groups may reflect
the impact of unmeasured confounders. Whether early dialysis initiation is harmful
or not is not ascertainable from the available evidence, and clearly, more research
is required in order for pediatric nephrologists to make informed decisions.
The importance of patient factors such as symptoms of uremia in decision-making was
universally supported in the survey. When assessing what factors were important or
very important to dialysis providers in determining when to start dialysis, three
tiers of responses emerged. Most respondents agreed that uremic symptoms such as fatigue,
nausea, and pruritis were important in decision-making. In the same tier were measurable
functional indicators such as increased missed school days, weight loss 10 %, and
suboptimal height growth velocity. These factors are all similar in their objective
and measurable nature. On the other hand, the more subjective, less measurable factors
fell to a lower tier of importance to providers. This indicates a clear preference
of physicians for more objective factors in determining when to initiate dialysis.
This survey also evaluated important facility and practice characteristics that could
influence the feasibility of future intervention trials in children. Firstly, local
pediatric CKD and dialysis populations are quite small. Therefore, the ability to
perform single-center studies does not exist, and all research efforts will require
national and international collaborations to obtain sufficient sample sizes.
Despite small program sizes, most centers do function with a multidisciplinary CKD
clinic and have the ability to perform dialysis and transplants within their local
program; therefore, resource issues should not be significant factors influencing
dialysis starts. However, some areas for improvement identified by the survey include
only 31 % of the respondents are using a patient and family educational process, and
there is a universal lack of formal policies on timing of dialysis initiation. The
reason for this lack of formalized decision-making process is likely multifactorial,
owing to a lack of evidence from the literature, a lack of pediatric studies regarding
dialysis initiation, or perhaps due to a deficiency in knowledge translation from
the adult literature.
Our study was not without limitations. We had approximately a two-third response rate
to our survey. However, the distribution of the respondents came from all parts of
the country and is therefore likely to be representative of the spectrum of pediatric
nephrologists in Canada. Though 40 physicians responded to the survey, the most responses
received for any one question was 35. Excluding these five individuals still leaves
a response rate of 59 %. Our survey was limited by a lack of power to statistically
assess the provider or facility characteristics that may be influencing eGFR thresholds.
Dialysis initiation is clearly a decision made through the input of several important
groups including the physicians, nurses, patients, parents, and the facilities themselves.
Outside of the scope of this project but an important limitation was the lack of responses
by parents, patients, and nurses involved in the decision to initiate dialysis. More
research is necessary to understand the roles of each of these groups in the decision-making
process for dialysis initiation. Finally, our survey mostly pertains to children with
slowly progressive CKD, as opposed to those who present with acutely falling GFRs
and classical indications for dialysis. As a result, many respondents may have envisioned
older children when formulating their answers in this survey. Nonetheless, only 50 %
thought that age was an important factor in their decision-making and would start
dialysis at an earlier threshold in children less than 1 year of age. Therefore, the
findings likely apply to all children over 1 year of age.