A large meta-analysis has concluded that having diabetes increases the chance that a person with mild cognitive impairment will progress to dementia by 65%.
There was no consistent evidence that hypertension or cholesterol levels increased the risk of someone with MCI progressing to dementia. Smoking was similarly not associated with increased risk, although the reason for this probably lies in mortality: smokers tend to die before developing dementia.
There was some evidence that having symptoms of psychiatric conditions, including depression, increased the risk of progressing to dementia.
There was some evidence that following a Mediterranean diet decreased the risk of an individual with amnestic MCI progressing to Alzheimer's, and that higher folate levels decrease the risk of progressing from MCI to dementia. The evidence regarding homocysteine levels was inconsistent.
The evidence indicates that level of education does not affect the risk of someone with MCI progressing to dementia.
Do note that all this is solely about progression from MCI to dementia, not about overall risk of developing dementia. Risk factors are complex. For example, cholesterol levels in mid-life are associated with the later development of dementia, but cholesterol levels later in life are not. This is consistent with cholesterol levels not predicting progression from MCI to dementia. Level of education is a known risk factor for dementia, but it acts by masking the damage in the brain, not preventing it. It is not surprising, therefore, that it doesn't affect progression from MCI to dementia, because higher education helps delay the start, it doesn't slow the rate of decline.
Do note also that a meta-analysis is only as good as the studies it's reviewing! Some factors couldn't be investigated because they haven't been sufficiently studied in this particular population (those with MCI).
The long-running Cache County study has previously found that 46% of those with MCI progressed to dementia within three years; this compared with 3% of those (age-matched) with no cognitive impairment at the beginning of the study.
More recently, data from the long-running, population-based Rotterdam study revealed that those diagnosed with MCI were four times more likely to develop dementia, over seven years. compared with those without MCI. Of those with MCI (10% of the 4,198 study participants), 40% had amnestic MCI — the form of MCI that is more closely associated with Alzheimer's disease.
The 2014 study also found that older age, positive APOE-ɛ4 status, low total cholesterol levels, and stroke, were all risk factors for MCI. Having the APOE-ɛ4 genotype and smoking were related only to amnestic MCI. Waist circumference, hypertension, and diabetes were not significantly associated with MCI. This may be related to medical treatment — research has suggested that hypertension and diabetes may be significant risk factors only when untreated or managed poorly.
 . Modifiable Predictors of Dementia in Mild Cognitive Impairment: A Systematic Review and Meta-Analysis. American Journal of Psychiatry [Internet]. 2015 ;172(4):323 - 334. Available from: http://ajp.psychiatryonline.org/doi/10.1176/appi.ajp.2014.14070878
 Conversion to dementia from mild cognitive disorder The Cache County Study. Neurology [Internet]. 2006 ;67(2):229 - 234. Available from: http://www.neurology.org/content/67/2/229
de Bruijn, R.F.A.G. et al. Determinants, MRI Correlates, and Prognosis of Mild Cognitive Impairment: The Rotterdam Study. Journal of Alzheimer’s Disease, Volume 42/Supplement 3 (August 2014): 2013 International Congress on Vascular Dementia (Guest Editor: Amos D. Korczyn), DOI: 10.3233/JAD-132558.