Spironolactone use and renal toxicity: population based longitudinal analysis

The study by Wei et al [1] led the editors of the BMJ to conclude
“spironolactone prescribing seems safe in the NHS, probably owing to
careful monitoring”. However, the authors omit a number of important
comments which the journal’s readership may wish to consider. Wei et al
draw direct and extensive comparison with a much larger Canadian study,
the results of which conflict with their own [2]. There are important
differences between the Wei and Canadian studies.

In the Canadian study, Juurlink et al restricted their analysis to
patients older than 65 years who had been hospitalised for heart failure,
whereas Wei et al did not, 2) Juurlink et al used 4 month intervals for
time series analyses compared with 6 month intervals in the Wei study,
making conclusions of the latter more liable to type II error, and 3)
Juurlink et al considered potential confounders in some detail, whereas
Wei and colleagues did not.
These factors show firstly that the two cohorts are not directly
comparable. The inclusion of younger patients by Wei et al will bias
results towards null. Secondly, the use of a 6 month time interval may be
open to selection bias, a concern which is supported by the marked
variation in ‘rates of hyperkalaemia’ from one 6 month period to the next
shown in figure 3b. The start and end points of the pre- and post RALES
comparison are likely important in determining whether the results are
significant or not.

Concerning rates of biochemical measurements, the study interval
overlaps with the April 2004 introduction of the Quality and Outcomes
Framework (QOF) which remunerated general practitioners for making the
same measurements assessed as outcome by Wei et al. Indeed, their data
support a temporal relationship between the introduction of QOF and an
increase in biochemistry sampling. The conclusion that spironolactone
prescribing per se is monitored better in Tayside than Canada is therefore
suspect.

Finally and importantly, Wei and colleagues report that 76% of
patients with heart failure diagnosed with hyperkalaemia had a prior
creatinine measurement greater than 220µmol/L. Patients with renal
impairment were excluded from the RALES trial [3], and the effects of
spironolactone on mortality in heart failure patients with renal
impairment therefore remains unknown. Hyperkalaemia is a serious
complication of spironolactone therapy, and rather than contradicting the
findings of Juurlink et al, the study by Wei et al should bring into
sharper relief the risks of spironolactone in patients with heart failure
who are elderly or have renal impairment.

1. Wei L, Struthers AD, Fahey T, Watson AD, Macdonald TM:
Spironolactone use and renal toxicity: population based longitudinal
analysis. BMJ (Clinical research ed, 340:c1768.

2. Juurlink DN, Mamdani MM, Lee DS, Kopp A, Austin PC, Laupacis A,
Redelmeier DA: Rates of hyperkalemia after publication of the Randomized
Aldactone Evaluation Study. The New England journal of medicine 2004,
351(6):543-551.

3. Pitt B, Zannad F, Remme WJ, Cody R, Castaigne A, Perez A, Palensky
J, Wittes J: The effect of spironolactone on morbidity and mortality in
patients with severe heart failure. Randomized Aldactone Evaluation Study
Investigators. The New England journal of medicine 1999, 341(10):709-717.

Competing interests:
None declared