The aim for cardiovascular high risk population is < 70 mg/dL, the IMPROVE-it clinical trial evaluated that also lower LDL-C value seem to be beneficial - to my knowledge there is no LDL-C cut-off for statin-therapy. Have a look at:

http://www.ncbi.nlm.nih.gov/pubmed/25066560

http://www.eas-society.org/fileArchive/IMPROVE%20IT%20-%20PRESENTATION%202014-11-17.pdf

If you can find another statement please tell me. Regards!

The actual discussion goes into the direction "the lower the better" for secondary prevention and some people recommend the same for primary prevention. The IMPROVE-IT trial mentioned by Marcus is indeed the first end point study going down to around 50 mg/dl and showing an improvement in comparison to 70 mg/dl.

I think the ongoing endpoint trails with PCSK9 inhibitors will finally answer the question especially because in this trails patients achieve LDL-C as low as 20-30 mg.

AHHHHH !!! Trials results must NOT applied in clinical practice (see and read well the Sackett books !). So, if you lower LDL in all treated patients on the basis of the slogan "the lower the better"  (as example: mean LDLc goal 50 mg/dL after therapy) you create a cohort of hypocholesterolemic patients with near-normal distribution of LDLc around 50 mg/dL, with lower values at 30 o less of LDLc. This is not a good idea, particularly if obtained by HMCoA-R inhibitors. Obviously, if you have a patient with coronary severe and ulcerated multiple stenosis, that can not make a CABG, with high LDLsd and/or other chol-rich atherogenic particles, the goal of 50 mg/dL (or less)  of cholesterol is suitable, HOWEVER, always verifying that therapy is really lowering the "Bad Lipo". (LDLsd=LDL3,4,5 in the example). We need several and several (and several......) long term trials to verify the safety of lowering LDL as low as 20 mg/dL. So, the answer is that LDL-therapy, as all other "clinical acts" must be decided in the vision of predictive (individual !!) personalized, precision medicine and without abdicating the clinical reasoning in favour of industry-driven guideline. OVERALL: avoid to refers only to total  or LDL cholesterol (this is a BIG mistake also in some recent clinical trials) because LDLc do not exists per sè. Really, in human blood, different LDL (and IDL, Lp(a), VLDL, HDL, remnants , free lipids and so on)  are circulating, with different composition -and different effect- in each patient. So, the atherogenetic and thrombogenetic particles are a lot !! Please, avoid reductionistic approach to lipoprotein metabolism. 

Con Simpatia 

Prof AV Gaddi 

Although there is concern on the long term implications of very low LDL cholesterol, another piece of the puzzle was published a few days ago in the New England Journal of Medicine, with the ODYSSEY-Long Term trial with alirocumab. They had a cohort of patients with LDL less than 25 mg/dl and in this subset of patients there was no increase in side effects. 

Some epidemiological studies have shown that low total cholesterol is associated with hemorrhagic stroke but this has not been shown in clinical trials. In the SPARCL trial of atorvastatin in secondary prevention of stroke, there was a slight increase in hemorrhagic stroke rate in patients on statin but overall, stroke rate was lower than placebo.  

Chih Hao Chen-Ku is right !! However, Odyssey is the "longest" study on PCSK9 inhibitors, is a phase 3 Study (!) on FH heterozygous patients (plus other no well classified HLP), with a follow up of max 3 yr if I remember well ! In my opinion, low LDLc may cause problems in several cell functions in the long period. Human newborn have very low "natural" LDLc level and live well (!), but the mean values is about 50+ mg/dL of LDLc !!!!  (30-70). The "low plasma LDL cholesterol" in adults is a complex (and "thorny") problem. 

Regards

A