Study characteristics

Methods

Participants randomly assigned to intervention and control groups, matched for age, sex and previous use of a garlic supplement.

Participants

146 participants, recruited voluntarily via newspaper advertisements.

Interventions

1 allicin‐containing garlic capsule (dose unspecified) capsule per day with the main meal; or placebo

Outcomes

Primary outcome: number of occurrences of the common cold in 3‐month period.
Secondary outcomes: cold duration (number of days), the number of days 'challenged' and the number of days to recovery.

Notes

We contacted the study author to query how well garlic could be blinded to participants because of its strong smell. He replied that the tablets used had no odour with daily use. Asked how the groups were matched if they were randomly allocated, he replied that patients were "matched with a standard protocol rejecting same sex volunteers after the maximum number was reached. Age was included in the admission protocol so volunteers over 70 were rejected as were those under 20 years old." We took this to mean that the total sample recruited was balanced according to age, sex and previous garlic use and then randomly allocated.

Although the study author reported that he had no conflict of interest, we note that he was employed by a garlic supplement company at the time of the study.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

A random number generator assigned volunteers to the active or placebo group

Allocation concealment (selection bias)

Low risk

Randomisation codes were kept secure and were not broken until all the diaries had been returned

Blinding (performance bias and detection bias)
All outcomes

High risk

5 participants (4 active, 1 placebo) noticed a 'smell' when burping. This suggests that blinding of participants may not have been adequate

Incomplete outcome data (attrition bias)
All outcomes

Low risk

4 participants withdrew from the study: 3 from the active group, 1 from the placebo group

Selective reporting (reporting bias)

Unclear risk

There was no evidence of selective reporting of outcomes. However, the statistical analysis and primary outcomes do not appear to have been decided in advance

Other bias

Low risk

—