# Body weight and vitamin K for raised INR

## Clinical bottom line

The response to subcutaneous vitamin K given to reduce raised INR can be variable. Adjusting the dose according to body weight may be more reliable in lowering INR over 24 hours.

### Study

This was a retrospective study of patients on warfarin with INR greater than 6 who received 1 mg vitamin K subcutaneously, had a follow INR within 48 hours, with body weight documented, and without active bleeding or obvious drug interactions, or comorbid condition needing hospital admission. Injections were into right or left deltoid fat pad.

INR values above 6 were routinely duplicated (to check for blunders) and clinic nurse and patients alerted.

Change in INR between initial and follow up values was divided by elapsed time in hours and multiplied by 24 to give standard change over one day. The dose of vitamin K was divided by 'adjusted' body weight. Adjusted body weight was given by the ideal body weight + 0.4(actual body weight minus ideal body weight). The ideal body weight was calculated as 50kg + 2.3(height in inches over five feet), for men.

The relationship between change in INR at 24 hours and the dose of vitamin K in mg/kg was analysed by linear regression.

## Results

There were 15 patients identified. INR values initially ranged from 6 to 14. Body weights ranged from 67 to 98 kg. Actual data for all important variables are given for individual patients.

Several relationships were examined, including those between change in INR and actual body weight. The best fit was found for that of change in INR and mg/kg of adjusted body weight (Figure 1), for which the Pearson regression coefficient (R) was 0.85, showing a good fit with P=0.00005.

### Figure 1: Correlation between change in INR over 24 hours with adjusted body weight dose of subcutaneous vitamin K

The authors suggest that Figure 1 can be used to adjust the subcutaneous dose of vitamin K. The desired change in 24 hours is drawn from the y-axis to the line of best fit, and the dose on the vertical intersection of the x-axis is then chosen.

## Comment

This is interesting hypothesis generating, and may well be useful. One problem is that the rationale behind choosing an adjusted body weight is not well explained, though a reference is given. The small numbers and retrospective nature may limit its usefulness, and it should be tested prospectively.