Statin use and risk of community acquired pneumonia in older people: population based case-control study

We appreciate Drs. Morimoto and Suzuki’s interest in our research.
They note that basic science experiments have demonstrated anti-
inflammatory effects of statins, which in theory provide more support for
a beneficial impact of statins after infection develops than for a role
for statins in preventing infection initially. However, we believe our
research is important because it addresses a question on which others have
already published indicating that statins do reduce the risk of
pneumonia.(1-3) These studies were likely to have been affected by
uncontrolled confounding, particularly “healthy user” bias,(4,5) and so
further examination of this question using more detailed data may be
helpful.

Beyond the specific focus on statins and pneumonia occurrence, we
believe that our study illuminates the potential bias that can occur when
studies rely on large databases that lack important information about some
potential confounders – in this case, measures relevant to “healthy user”
bias.(5,6) We do not agree that “healthy user” bias is unlikely to play an
important role in observational studies examining outcomes after the
development of infection (e.g. mortality after pneumonia or sepsis). On
the contrary, “healthy user” bias may be at least as problematic, if not
more so, in such studies as it is in studies which examine the risk of
developing infection. Groups at particularly high risk for mortality—such
as nursing home residents, persons with serious cancer or other terminal
illness, or immunosuppressed persons—may be less likely to receive statin
therapy prior to developing infection, which could result in “healthy
user” bias in studies of mortality or other outcomes following acute
infection.

In addition, Drs. Morimoto and Suzuki suggest that the presence of
COPD may modify the association of statins with pneumonia risk and that
statin use may prevent development of COPD. Thus, they propose that
analyses of the statin-pneumonia association should be stratified by COPD
status. However, we believe that this approach would not be appropriate.
The reasoning relates to how one should handle confounders that are also
intermediate variables. If we felt that statin use protected against the
development of COPD to a substantial degree, then it would be
inappropriate to adjust for or in any way stratify on the presence of COPD
in the analysis of statin use in relation to pneumonia risk (or even to
restrict the analysis to persons without COPD). This is because according
to that conceptual model, COPD is an intermediate in the causal pathway
between statin use and pneumonia, and adjusting for an intermediate in the
causal pathway is not appropriate. However, because we are unable to sort
out the temporal relationship between statin use and COPD, and as a result
do not know which of these came first, it would be hazardous NOT to adjust
for the presence of COPD, given that it has the potential to serve as a
strong confounding variable.

We hope this response answers your questions about our study.

Sincerely,

Sascha Dublin, MD, PhD

Michael L. Jackson, PhD

Jennifer C. Nelson, PhD

Noel S. Weiss, MD, DrPH

Eric B. Larson, MD, MPH

Lisa A. Jackson, MD, MPH

References

1. Myles PR, Hubbard RB, McKeever TM, Pogson Z, Smith CJ, Gibson JE.
Risk of community-acquired pneumonia and the use of statins, ace
inhibitors and gastric acid suppressants: a population-based case-control
study. Pharmacoepidemiol Drug Saf. Apr 2009;18(4):269-275.

2. Schlienger RG, Fedson DS, Jick SS, Jick H, Meier CR. Statins and
the risk of pneumonia: a population-based, nested case-control study.
Pharmacotherapy. Mar 2007;27(3):325-332.

3. van de Garde EM, Hak E, Souverein PC, Hoes AW, van den Bosch JM,
Leufkens HG. Statin therapy and reduced risk of pneumonia in patients with
diabetes. Thorax. Nov 2006;61(11):957-961.

4. Glynn RJ, Schneeweiss S, Wang PS, Levin R, Avorn J. Selective
prescribing led to overestimation of the benefits of lipid-lowering drugs.
J Clin Epidemiol. Aug 2006;59(8):819-828.

5. Jackson LA, Jackson ML, Nelson JC, Neuzil KM, Weiss NS. Evidence
of bias in estimates of influenza vaccine effectiveness in seniors. Int J
Epidemiol. Apr 2006;35(2):337-344.

6. Jackson LA, Nelson JC, Benson P, et al. Functional status is a
confounder of the association of influenza vaccine and risk of all cause
mortality in seniors. Int J Epidemiol. Apr 2006;35(2):345-352.

Competing interests:
Jennifer C Nelson did consulting work for GlaxoSmithKline