Effect of Bone Mineral Target Achievement on Mortality in Incident Dialysis Patients: An Analysis of the United Kingdom Registry - Article Review - Dr. Ross Morton
This article appraisal is part of the EMiNEM Bone and Mineral Metabolism Series. Click here to reach the EMiNEM homepage on UKidney
The achievement of guideline targets for new dialysis patients in the United Kingdom may be associated with lower mortality
Study design and study population:
This is a retrospective cohort study using incident patients from the UK Renal Registry.
10,902 incident adult dialysis patients beginning dialysis during the period January 1st, 2002 – December 31st, 2004 were enrolled and followed until December 31st 2007. Both hemodialysis (70%) and peritoneal dialysis (30%) patients were included.
Baseline demographics and pre-existing cardiac comorbidities and risk factors were derived from the UKRR database. Calcium (corrected for albumin), phosphate and iPTH were extracted on a quarterly basis. Missing data were handled by imputation.
Outcome of Interest:
The outcome of interest was all cause mortality. A reference group of patients whose calcium, phosphate and iPTH fell into the KDOQI target range on all 4 quarters for the first year of therapy was compared with groups who favoured less well in terms of achievement of target.
Target achievement was lowest for iPTH (23 – 26%) and highest for phosphate (54 – 62%), with 43 – 47% achieving the calcium target. Although not stated, it appears that approximately 2% of the population was in target for all 3 parameters for all 4 quarters. [I confirmed this with the principle author who indicated that the investigators had combined patients in the PTH group only who were in target for 3 out of 4 and 4 out of 4 quarters to give a total of 8% of patients to create the reference group].
Being out of range for all parameters for some or all of the time had no statistically significant impact on all cause mortality in this study. When the parameters were studied in isolation as linear variables, only phosphate > 2.17 mmol/L (6.5 mg/dL) was significantly associated with increased mortality, although there was a trend for an association with increasing calcium concentration.
The study has a number of strengths. The use of incident patients from a database with a high inclusion rate (> 90% for the UKRR) enhances the generalizability of the study, as does the inclusion of peritoneal dialysis patients (30%), who are frequently excluded or severely under-represented. The sensitivity analyses suggest that the results are fairly robust. Weaknesses include the lack of validated of comorbidities for just under half of the patients, low attainment of PTH results, resulting in the requirement for combining the PTH groups as described above, and lack of access to medication profiles. Obvioulsy the study is not a randomized trial and can thus only show association rather than causation.
Impact on practice:
The study runs the risk of being over-interpreted to allow more lenient phosphate and calcium control, when what is actually required is an adequately powered study showing that targeted phosphate control improves major outcomes. The fact that no association with PTH levels can be shown is consistent with the relatively weak association found in other epidemiologic studies, but is further compounded by the lack of 66% of the PTH values.
Reviewed by Reviewed by Dr. Ross Morton