Well, it depends who you ask.
According to this 2003 policy report from the American Board of Pediatrics, 200 IU per day was plenty to prevent vitamin D deficiency. If the goal is simply to prevent rickets, the skeletal condition related to overt vitamin D deficiency, this amount is probably sufficient.
A 2008 update to this policy statement updated the recommendation to 400 IU per day. This was based on a redefinition of what “normal” levels of vitamin D are; the standard was increased from greater than 27.5 nmol/L to 50 nmol/L. Currently, this is the standard of care in the pediatrics community.
A recent study from Lebanon, however, suggests that these standard recommendations are far too low. In this report, a daily dose of 2000 IU (actually, a weekly dose of 14000 IU) was far more likely than a 200 IU dose (OK, 1400 IU per week, that is) to normalize blood vitamin D levels. Better yet, there was no evidence for vitamin D toxicity in the study group.
Note that this 2000 IU per day dose is exactly the tolerable upper limit for vitamin D established by the National Academy of Sciences (upper limit is 1000 IU for infants under 1 yr). I think that safety concerns will keep the standard recommendations much lower than the 2000 IU level, even though the evidence for toxicity at this dose is lacking.
There continues to be a large disconnect between the conservative standard recommendations for vitamin D intake and the evidence emerging from clinical trials at a meteoric rate (150 clinical trials in the past year alone). In my clinical practice, I do see many people who have chosen to go up to the higher dose levels suggested by emerging research.
I continue to worry that there is a significant and as yet unidentified downside to keeping vitamin D levels up at a level upwards of 80 nmol/L. Our experiences with hormonal treatments like cortisol and estrogen should serve as reminders that the lack of acute toxic effect is not sufficient evidence of long term safety.
On the balance, I do think that the 2000 IU dose looks like a solid and safe recommendation for children and adults alike. I think that going beyond this may introduce unanticipated risks without having supportive evidence for preventive benefit.