Study design:
We conducted this study to investigate whether changes in DRT over the first two years of HD associate with dialysis adequacy and IDH. This research was conducted under a protocol reviewed by New England Independent Review Board, who confirmed the study was secondary research of existing patient data that were deidentified and did not require informed consent per title 45 of the United States Code of Federal Regulations part 46.104-d4ii (Needham Heights, MA; NEIRB#WO1-6614). The study was performed in adherence with the Declaration of Helsinki.
Setting and Participants:
We used data previously collected from ESKD patients during standard-of-care HD at a large dialysis organization (LDO) in North America (Fresenius Kidney Care, Waltham, MA, United States) during January 2014 to December 2017. Analysis included data from adults (age ≥18 years at first date of dialysis (FDD)) treated via HD who completed ≥1 DRT survey within the first 180 days of RRT. We excluded patients age ≤18 years and pregnant females.
DRT Survey:
DRT survey was a unique questionnaire used by the LDO. The DRT survey was administered conjunction with the Kidney Disease Quality of Life (KDQOL) survey. These questionnaires (KDQOL and additional DRT survey) were administered by the LDO as a standard of care in the incident dialysis period and annually thereafter.
The DRT survey includes one question and is completed by patients selecting one of five possible responses. The DRT survey question and answers are shown below.
“How long does it take you to be able to return to your normal activities after your dialysis treatment? (Circle the timeframe that best describes your answer)
- Less than 30 minutes
- Within 1 hour
- Within 1-2 hours
- Within 2-4 hours
- More than 4 hours”
DRT survey was chosen by the LDO to be administered with the KDQOL to assess additional HRQOL parameters in June of 2013. By 2014 and through 2017, DRT survey was responded to by ≥99% of HD patients who completed the KDQOL survey.
Primary Outcome:
The primary outcome of our study was the change in DRT from the incident period (≤180 days from FDD) to: 1) the first prevalent year (>365-to-≤545 days from FDD), and 2) second prevalent year (>730-to-≤910 days from FDD).
The exposure variables were the number of HD treatments with an IDH episode per month and dialysis adequacy (i.e. spKt/V). We considered the mean values for exposure variables ±30 days of the DRT survey date in the incident period and the change from the incident to the first/second year prevalent periods. A treatment with IDH episode was defined as ≥1 intradialytic systolic blood pressure (SBP) <100 mmHg.
Covariates were selected in an a priori manner. These included the incident DRT, comorbidity burden (age, congestive heart failure (CHF), diabetes, ischemic heart disease (IHD), number of comorbidities and treatment schedule (majority of HD sessions started before/after 1200 hours in the incident period and the change from incident to prevalent periods). For continuous variables, we used average values ±30 days of the DRT survey date. Categorical variables were determined from the most recent record to DRT survey date.
Exploratory Outcomes:
We performed two descriptive exploratory analyses. First, we assessed if various demographic, environmental, comorbid, clinical, and laboratory parameters (detailed in supplemental files) were associated with DRT in the first 180 days of HD. Second, we investigated whether the DRT category in the first 180 days of HD was associated with crude 6-, 12-, and 24-month hospital admissions per patient per year (ppy).
Statistical Methods:
Analyses were performed using SAS version 9.4 (SAS, Cary, NC, USA). Sankey diagrams and Forest plots constructed using R version 3.5.2 (R Foundation, Vienna, Austria).
Analysis of Descriptive Statistics:
The characteristics of HD patients in the incident period was computed for demographic, comorbid, clinical, and laboratory parameters. Categorical and continuous variables were calculated as counts/proportions and mean ± standard deviation (SD). The changes from the incident DRT to the first- or second-prevalent year DRT were calculated as proportions and tabulated, as well as visualized via Sankey diagrams.
Analysis of Primary Outcome:
Logistic regression models were constructed to calculate the odds ratio and confidence intervals for a change to a longer DRT from the incident to the first- and second- prevalent years.
We defined changes in DRT as a binary outcome with: 1) a change to a shorter DRT being any decrease from above to below a DRT category <2 hours, or a decrease from a DRT >4 hours to a lower DRT category (this was inclusive of patients who maintained a shorter DRT that was <2 hours), and 2) a change to a longer DRT being any increase from below to above a DRT category >2 hours (this was inclusive of patients who maintained a longer DRT with no reduction from the incident to follow up period). The selection of this binary outcome for changes in DRT was constructed considering a target time for DRT being <2 hours, but also considering any decrease in DRT in longer categories to be potentially favorable for the patients’ quality of life.
The exposure variables for the logistic regression models were dialysis adequacy, as measured by the spKt/V, and the number of HD treatments with IDH episodes per month; exposure variables included mean values in the incident period and the change in the mean values from the incident to the first/second year prevalent periods. We adjusted the models for incident DRT, age, CHF, diabetes, IHD and number of comorbidities.
Analysis of Exploratory Outcomes:
For the first exploratory aim, we compared demographic, environmental, comorbid, clinical, and laboratory parameters by the DRT category in the first 180 days of HD. Comparisons were made in reference to a DRT <0.5 hours using two sample t-tests for means and Chi-Square methods for proportions. For the second exploratory aim, we constructed unadjusted Poisson models to compare hospital admission rates by DRT category in the first 180 days of HD; comparisons were made in reference to a DRT <0.5 hours.