Hospitalization outcomes associated with provider reuse practices and bleach use

1
Hospitalization outcomes associated with provider
reuse practices and bleach use

• Allan J Collins MD FACP, James Ebben BS,Fred
  Dalleska MS, Shu-Cheng Chen MS
• Nephrology Analytical Services, Minneapolis
  Medical Research Foundation, University of
  Minnesota Twin Cities

2
Introduction

• Recent studies have shown that high-flux
  dialyzers reused with formalin and bleach were
  associated with better outcomes than all other
  reuse practices and membranes.
• Earlier comparisons of reuse-associated outcomes
  have shown inconsistent results over time,
  suggesting that factors other than reuse
  practices may contribute to the associations.

3
Introduction (cont.)

• More recent studies, which added information on
  dialysis unit profit status and patient
  hematocrit level, showed important differences
  between providers practicing reuse.
• In order to assess the impact of reuse on the
  risk of first hospitalization, we studied more
  recent cohorts of prevalent dialysis patients
  (1995, 1996), evaluating outcomes related to
  reuse with bleach cleaning and utilizing more
  complete adjustments for patient and provider
  characteristics.

4
Methods

• Study population prevalent Medicare EPO-treated
  hemodialysis patient cohorts from 1995 and 1996
  (199534,328 199637,595).
• Patients survived on HD from 7/1 to 12/31 of the
  cohort year. Follow-up periods of both one and
  two years were analyzed.
• Patients who were hospitalized at the beginning
  of the follow-up period were excluded from the
  analysis.
• Reuse practices were classified on a dialysis
  unit level, based on survey data from the Centers
  for Disease Control (CDC), into high-flux or
  conventional no reuse (Conv.-NR). Germicide
  groupings included formalin (Form),
  glutaraldehyde (Glut), and peracetic (PAA) acid,
  both with (B) and without (NB) bleach cleaning.

5
Methods (cont.)

• Explanatory variables included age, gender, race,
  prior ESRD time, unit ownership, and unit profit
  status.
• Comorbidity, severity of disease measures, total
  hospital days, hematocrit level, and EPO dose
  were characterized in the entry period and
  entered into the model.
• A Cox regression model, stratified on diabetes,
  assessed the first all-cause hospitalization. As
  a baseline we used units practicing high-flux
  formalin reuse with bleach cleaning.

6
Results

• Statistically significant (Chi-square, pdifferences were observed in the distribution of
  race, gender, and diabetic status between the
  reuse groups.
• Mean age varied among the reuse groups, with
  PAA-B highest in both cohorts, and Conv.-NR
  lowest in the1995 cohort and Glut-NB lowest in
  the 1996 cohort.
• A hematocrit level lower than the current DOQI
  guideline (33 36) was associated with an
  increased risk of hospitalization, as previously
  reported.

7
Results (cont.)

• EPO dosage was associated with decreased risk for
  the lowest quintile of dose (and increased risk for the highest quintile of
  dose (77,000 units per month) (Reference
  42,00055,000 units/month).
• Form-NB was associated with both better outcomes
  (19957 decreased risk) and worse
  outcomes(19969 increased risk).
• Glut-B was associated with decreased
  hospitalization risk in the 1995 cohort, but not
  in the 1996 cohort.
• Glut-NB was associated with decreased risk (13)
  in the 1996 cohort only.
• PAA-B was associated with better outcomes (10
  decreased risk) in the 1995 cohort only.

8
Percent female, by reuse practice
1995
1996
9
Percent of diabetics, by reuse practice
1995
1996
10
Distribution of races, by reuse practice

• 1995 1996
• Reuse White Black Other White Black
  Other
• Form-B 48.4 44.6 7.0 48.3 44.0 7.7
• Form-NB 45.8 47.2 7.0 46.0 47.0 7.0
• PAA-B 56.2 40.2 3.6 62.3 35.4 2.3
• PAA-NB 54.2 38.0 7.8 53.2 38.7 8.1
• Glut-B 59.0 38.6 2.4 62.3 34.9 2.8
• Glut-NB 54.7 40.5 4.8 43.8 52.2 4.1
  
• Conv.-NR 47.8 47.8 4.4 49.5 45.9 4.6

11
Mean age, by reuse practice
1995
1996
12
Mean hematocrit, by reuse practice
1995
1996
13
Mean number of comorbid conditions, by reuse
practice
1995
1996
14
Relative risk of all-cause first hospitalization,
by reuse practiceOne-year follow-up period
1995
1996
15
Relative risk of all-cause first hospitalization,
by reuse practiceTwo-year follow-up period
1995
1996
16
Conclusions

• Statistically significant differences in
  hematocrit and number of comorbid conditions were
  noted between reuse groups, and therefore must be
  adjusted for.
• The results from the germicide-bleach reuse
  groups were inconsistent. Reuse with bleach
  cleaning was associated with both higher and
  lower first-hospitalization risks.
• Contrary to earlier studies, we find no
  conclusive differences in outcomes between bleach
  and no bleach high-flux reuse units.
• The 7-8 increased risk of first hospitalization
  in the conventional no-reuse units may be related
  to lower dialysis therapy, and warrants further
  evaluation.