Transcript Journal Club Alcohol and Health: Current Evidence March-April 2007 www.alcoholandhealth.org Featured Article Alcohol consumption as a trigger of recurrent gout attacks. Zhang Y, et al.

Journal Club
Alcohol and Health: Current Evidence
March-April 2007
www.alcoholandhealth.org
1
Featured Article
Alcohol consumption as a
trigger of recurrent gout
attacks.
Zhang Y, et al. Am J Med. 2006;119(9):800.e13–800.e18.
www.alcoholandhealth.org
2
Study Objective
To examine whether…
– alcohol use may trigger recurrent
gout attacks
www.alcoholandhealth.org
3
Study Design
Web-based case-crossover study
– A case-crossover study is a method to assess transient
effects of exposures on the risk of acute events. It
addresses whether an event was triggered by something
that happened just before the event.
– To estimate relative risk, the exposure frequency just
before the event occurred is compared with exposure
frequencies during one or more control times (each case
serves as its own control).
– This design, therefore, controls for confounding by
permanent subject characteristics and also between the
trigger's acute and chronic effects.
Maclure M, et al. Annu Rev Pub Health. 2000;21:193–221.
www.alcoholandhealth.org
4
Study Design (cont.)
• All subjects had a gout attack in the past year and
were recruited online.
• Medical records were reviewed to confirm the
diagnosis.
• Online surveys assessed subjects’ alcohol use and
other risk factors for gout attacks during…
• a hazard period (2 days before an attack) and
• 4 attack-free control periods (study entry and then 3, 6, and 9
months later)
www.alcoholandhealth.org
5
Assessing Validity of an
Article About Harm
• Are the results valid?
• What are the results?
• How can I apply the results to
patient care?
www.alcoholandhealth.org
6
Are the Results Valid?
• Did the investigators demonstrate similarity in all
known determinants of outcomes? Did they
adjust for differences in the analysis?
• Were exposed patients equally likely to be
identified in the two groups?
• Were the outcomes measured in the same way in
the groups being compared?
• Was follow-up sufficiently complete?
www.alcoholandhealth.org
7
Did the investigators demonstrate similarity in
all known determinants of outcomes? Did they
adjust for differences in the analysis?
• The case-crossover design is similar to a case-control study.
– But, the case-crossover design does not have 2 groups of
people because cases serve as their own controls.
– Therefore, demonstrating similarity in static determinants
of outcomes is unnecessary given that cases and controls
are the same.
• Researchers did not report the similarity of characteristics
(diuretic use and total purine intake) in the hazard and
control periods (the relevant timeframes) but did adjust for
them in analyses.
www.alcoholandhealth.org
8
Were exposed patients equally likely
to be identified in the two groups?
 The case-crossover design eliminates this concern,
which applies to a classic case-control study.
 All subjects had an attack; some attacks occurred
with prior alcohol exposure, others did not.
 All subjects contributed data from both hazard and
control time periods. Therefore, exposed and
unexposed periods are equally identified.
www.alcoholandhealth.org
9
Were the outcomes measured in the
same way in the groups being compared?
•
Cases and controls are the same so the
outcomes could not be measured differently.
•
Hazard and control time periods are
compared. Risk factors were measured in the
same way during both of these periods.
www.alcoholandhealth.org
10
Was follow-up sufficiently complete?
• The report does not state how many people
completed all assessments.
• However, 197 subjects provided data on hazard
and control periods; some provided more than
others.
www.alcoholandhealth.org
11
What are the Results?
• How strong is the association between
exposure and outcomes?
• How precise is the estimate of the risk?
www.alcoholandhealth.org
12
What are the Results?
• 197 subjects enrolled and completed both hazard and
control period questionnaires.
– 91% met American College of Rheumatology criteria
for gout.
– 95% of 172 medical records reviewed confirmed the
diagnosis.
• Over 1 year of follow-up, 321 gout attacks occurred.
• 53 subjects consumed no alcohol.
www.alcoholandhealth.org
13
How strong is the association between
exposure and outcome?
P for
Time Before
Attack
Sample Odds Ratios* Comparing
Drinking Vs. Not Drinking
24 hours before
1.4 (95% CI, 0.6–2.4) for 1–2 drinks
3.1 (95% CI, 1.0–11.0) for >7 drinks
<0.02
48 hours before
1.1 (95% CI, 0.7–2.0) for 1–2 drinks
2.5 (95% CI, 1.1–5.9) for >7 drinks
<0.005
trend
*Adjusted for diuretic use and purine intake
www.alcoholandhealth.org
14
How strong is the association between
exposure and outcome? (cont.)
• Risk did not increase with drinking in the 25-48
hours before an attack…
– although the odds ratio was 7.3 [95% CI, 1.8–29.2]
for >7 drinks.
• In analyses also adjusted for total alcohol
consumption, the risk of an attack was not
associated with any specific alcoholic beverage.
www.alcoholandhealth.org
15
How precise is the estimate of the risk?
• Confidence intervals for the odds ratio risk estimates
for all drinking levels within the 48 hours of an attack
included 1…
– except for >7 drinks within the last 25–48 hours:
OR 7.3, 95% CI, 1.8–29.2.
• At this and other drinking levels, the intervals are wide.
• So, estimates for each level are not precise.
• However, the trend was statistically significant.
www.alcoholandhealth.org
16
How Can I Apply the Results to
Patient Care?
• Were the study patients similar to the patients in
my practice?
• Was the duration of follow-up adequate?
• What was the magnitude of the risk?
• Should I attempt to stop the exposure?
www.alcoholandhealth.org
17
Were the study patients similar to the
patients in my practice?
• The subjects were recruited online with paid
advertisements linking to a search term “gout.”
– How this might affect selection is not known.
• 91% fulfilled criteria of gout.
• Most were middle-aged white males with more than a
high school education.
• Drinking patterns over time (e.g., heavy occasional,
heavy constant, always moderate) were not reported.
www.alcoholandhealth.org
18
Was the duration of follow-up adequate?
• Follow-up duration was one year.
– However, follow-up is not very relevant to the
question at hand.
• There were many hazard and control time periods
for comparison.
• The time periods of clinical interest were short but
adequate to answer the question about recent
alcohol exposure as a risk factor.
www.alcoholandhealth.org
19
What was the magnitude of the risk?
• The magnitude of the absolute risk cannot be
calculated.
• But the odds ratio provides an estimate: 2–3 fold
odds of an attack after drinking >5 drinks.
– The trend towards increased risk with recent drinking
(within 24 hours of the attack), however, began with
any drinking.
• The attributable and absolute risks would depend on
the number of attacks an individual had and the
prevalence of drinking in people with gout (73% 1year prevalence of any drinking in this sample).
www.alcoholandhealth.org
20
Should I attempt to stop the exposure?
• You should probably attempt to stop the
exposure if the goal is to reduce the likelihood of
a gout attack.
• Physician advice can lead to decreased drinking.
– And decreased drinking would likely be
associated with fewer gout attacks, according
to this study and prior assumptions.
www.alcoholandhealth.org
21