Vitamin K: applied after the laser it reduced bruising; applied beforehand it made no difference

A 2002 trial at the University of Miami. Randomised, double-blind, placebo-controlled.

22 people took part, comparing half a face against the other half — vitamin K cream on one side, vehicle alone on the other. Twice daily, for two weeks.

Participants were split into two groups.

• 11 in the pretreatment group — applied it for two weeks before the laser
• 11 in the posttreatment group — applied it for two weeks after the laser

On day 0 everyone had 585 nm pulsed dye laser treatment for facial telangiectases. Bruising was rated by the patient and by the physician on a visual analogue scale on days 0, 3, 7, 10, 14 and 17.

The results.

• applied before the laser — no significant difference in bruising compared with placebo
• applied after the laser — significantly lower bruise-severity scores than placebo
• the paper notes the difference was particularly in the initial days of application

Now how to read it.

First, the design is good. It compared the two halves of the same face (own control), it was double-blind, and a physician rating is in there. This is not a trial with only self-assessment.

Second, each group is 11 people. The total is 22, and two questions were each answered by 11 people. That is small. Whether failing to find a difference with pretreatment means there is no effect, or that 11 people could not catch it, cannot be separated by this trial. This site writes those two apart — and here it is the cannot-be-separated case.

Third, bruising was rated on a visual analogue scale. Colour was not measured by instrument. But a physician rating is included, so the endpoint is marked 'graded'.

Fourth, it is from 2002, and there is one trial.

So it sits at 'early evidence'. The design is proper, the size is small, and what followed is thin.

And there is exactly one usable sentence in this entry — in this trial, no benefit was seen from applying it in advance.

A 2010 trial at Northwestern University. It is the same trial as on the arnica page — do not count it as two badges.

The design is unusual. On healthy volunteers, four standard bruises 7 mm across were created with a 595 nm pulsed dye laser on the inner upper arms (two per arm, 5 cm apart), and each of a person's four bruises was randomly assigned one of four agents.

• 5% vitamin K
• 1% vitamin K with 0.3% retinol
• 20% arnica
• white petrolatum (control)

Twice daily under occlusion for two weeks. A dermatologist not involved in assignment scored standardised photographs 0–10. Both patient and rater were blinded.

Here is only the vitamin K side of the result.

• 20% arnica was better than white petrolatum (P = 0.003)
• 20% arnica was better than 1% vitamin K with 0.3% retinol (P = 0.01)
• 20% arnica was not better than 5% vitamin K cream

Something here has to be read precisely.

'Arnica was not better than 5% vitamin K' is not the same statement as '5% vitamin K works'.

The abstract does not report a direct comparison of 5% vitamin K against white petrolatum. Without that comparison, whether 5% beat placebo is unknown. Both readings fit: 5% may have worked as well as arnica, or two of the four may have done nothing together.

So this entry sits at 'not established'. Not being beaten is not enough to raise a grade.

What this trial does say clearly is one thing — concentration.

1% vitamin K mixed with retinol came out worse than arnica. Which means vitamin K appearing on the back of a package is not enough on its own. The percentage decided the result. And products often do not state a percentage. This site does not estimate concentrations it has not read, so the concentration field on the badge above is empty.

Finally, the design's limit. These bruises were standard 7 mm bruises made deliberately with a laser. They cannot be assumed to be the same as a bruise that actually follows a procedure, or one from a knock. And the abstract does not give the number of participants.