Among Swedish patients, most of whom were treated with statins only, prevalence of dyslipidemia decreased significantly following treatment, from 95% to 68% in patients with elevated TC and from 93% to 61% for patients with elevated LDL-C. Isolated elevated LDL-C was less prevalent in high risk population. Elevated TGs persisted in about 50% and low HDL-C persisted in about 30% at follow up for total sample, while MD defined as prevalence of low HDL-C and/or elevated TGs were prevalent in 60%.
Improvement in TGs was moderate and low HDL-C persisted, showing only modest improvement following therapy, and this was most notable in patients with T2DM. Compared with nondiabetic patients, those with diabetes in our study had a greater risk of experiencing dyslipidemia involving TGs and HDL-C. This finding is supported by observations of increased TGs and low HDL-C in other studies involving patients with diabetes [20, 21].
About 30% and 40% of all patients attained goals of TC and of LDL-C respectively following treatment, while only additional 5% and 10% patients attained HDL-C and TG normal levels respectively following treatment. Attaining lipid goal/normal levels was slightly associated to age while not significantly associated to gender except for attainment of normal levels in HDL-C. History of CHD was associated with lower odds for attainment of TC and LDL-C goals. Diagnosis of T2DM was negatively associated to attainment of goal/normal levels for any lipid parameter. Baseline lipid levels of TC, LDL-C and TGs were strongly and negatively associated to attainment of goal/normal levels in TC, LDL-C and TGs, while baseline HDL-C was positively associated to attainment of normal levels in HDL-C. Duration of statin therapy was negatively associated to attainment of goal/normal lipid levels.
Findings compared to other studies
A large cohort study in northern Sweden  showed that mean lipid levels (TC) have decreased significantly during 1986-2004 from 6.4 to 5.8 mmol/l in men and 6.3 to 5.5 mmol/l in women in age group 25-64 years and from 6.4 to 5.5 mmol/l in men and from 7.1 to 6.2 mmol/l in women in higher age group . In this study mean TC was higher for men (6.6 mmol/l) and women (7.1 mmol/l), indicating a difference of about 1.0 mmol/l compared to the findings in the above study. This could be explained by the inclusion criteria in this study where patients were selected if they were treated with LMT and had indication of elevated lipid levels. In a study using data up to 2003 on patients treated with a LMT in the same county (Uppsala Sweden) as in this study, mean TC was 7.47 mmol/l in the total population, which could be compared to 6.9 mmol/l in this study, indicating a decrease in TC lipid levels from 2003-2007. Another cohort study  in a Swedish population showed that as from 2004 there is a tendency to an increase in mean TC lipid levels among both men and women . This tendency could however be an indication for regional differences in treatment patterns, demography and socio economic factors. High cholesterol concentrations are associated to low socio economic status .
A study from France , where fibrates are more frequently used found that regardless of LDL-C levels 38.7% had MD defined as elevated TGs and/or low HDL-C, which corresponds to 57% in this study. Fibrates have been shown effective in raising HDL-C and lowering TGs  and this could explain some of the difference in findings. Another important difference could be explained by baseline lipid profiles. Baseline mean lipid values were lower (higher in HDL-C) 3.3, 1.46 and 1.59 mmol/l in LDL-C, HDL-C and TGs respectively in the French study, which could be compared to 4.5, 1.4, 2.4 mmol/l in this study. Lipid levels at baseline were found to be strongly associated to attainment of goal/normal lipid levels following treatment. This finding is in line with another study where high lipid levels at baseline were found to be strong predictors of failing to reach goal/normal levels in TC, LDL-C and TGs and the inverse for the HDL-C .
A study on Swedish patients treated with statins found that 70% of high-risk patients had at least one lipid abnormality  compared to 82% in this study, a difference that could be due to different treatment patterns at GP and specialist settings, since patients in that study were recruited by specialists (40%) whereas in this study patients were recruited in GP settings only.
In an earlier study, where goal attainment was defined as having a TC below 5.0 mmol/l and LDL-C level below 3.0 mmol/l (as recommended by the Swedish Medical Products Agency at that time) 31% reached goal levels, compared to 43% in this study, which could indicate improved treatment of TC or LDL-C disorders between study periods. In the earlier study the time period was 1993-2001 and in our study the period is 1994-2007, a time period where acceptance and use of statins was widely increased in Sweden. In 2003 the patent of simvastatin expired and the price for the generic simvastatin decreased to about 10% of the price of the patent drug, which could partly explain some of the significant increase in usage of statins.
A study with a similar design and method carried out in the US and included 5,158 patients found that therapy (primarily statins) reduced the proportion of patients not at LDL-C goals from 77% to 22% and the proportion with high TG levels from 34% to 20%. HDL cholesterol levels were unchanged (49% and 50% were less than normal levels before and after therapy respectively) in the aggregate and in high-risk subgroups (patients with coronary artery disease, diabetes, and 10-year heart disease risk >20%). After therapy 29% of high-risk patients were found having multiple lipid abnormalities . These findings are consistent to findings in our study, however proportions of patients not reaching LDL-C goal were higher in our study (93% to 61% before and after treatment). This difference could partly be due to different thresholds for dyslipidemia employed in the studies.
We also carried out analyses using the thresholds of target/normal levels for LDL-C, HDL-C and TG specified as per National Cholesterol Education Program, Adult Treatment Panel III (NCEP ATP III) guidelines, LDL-C < 100 (2.59 mmol/l), 130 (3.37 mmol/l) or 160 mg/dL (4.14 mmol/l) (depending on risk factors), HDL-C > 40 mg/dL (1.04 mmol/l) in men, > 50 mg/dL (1.30 mmol/l) in women and Triglycerides < 150 mg/dL (1.69 mmol/l), but not considered high until > 200 mg/dL (2.26 mmol/l) [32, 33]. The NCEP guidelines are slightly different from the Swedish guidelines, threshold for abnormal LDL-C is somewhat higher in the NCEP compared to Swedish guidelines (2.59 mmol/l vs to 2.5 mmol/l). Using NCEP thresholds elevated LDL-C was prevalent in less than 40% of total study population, whilst this figure was over 50% in the high risk population. These differences were the major differences compared to the results using Swedish guidelines. Using NCEP thresholds yields more consistent findings as compared to the study carried out in the US and was discussed above .
Treatment patterns and importance of addressing lipid abnormalities
Statin therapy is established and accepted as primary and secondary prevention  and the effect of statins on lowering LDL-C as well as their effect on reducing the risk of cardiovascular events has been widely documented [5, 34–36]. Since in this study, treatment was primarily targeted towards lowering LDL-C, isolated LDL-C was less prevalent in high risk population. Low HDL-C and elevated TGs are independent risk factors for CHD [8, 12, 32, 37] and abnormalities in these lipids were not much improved in this study: Elevated TGs and low HDL-C and combinations persisted to a large extent, mostly in high-risk population despite LMT.
In the logistic regression analysis we found duration of statin therapy to be negatively associated to attainment of goal/normal levels. For each year increase on statin treatment the odds of attaining LDL-C goal decreased by about 10% (OR, 0,86; 95%CI, 0,84-0,87). This could indicate lower efficacy with increasing time on treatment, however better medication possession ratio was found to be associated with a better goal attainment in TC and in LDL-C . Half of all patients on statin treatment discontinue the medication by the end of the first year . Our findings might be influenced by factors related to patient compliance to treatment and discontinuation. Patients in this study were assumed to be compliant to treatment, as they fulfilled the criteria of refilling their prescriptions for at least one year, however this might still not accurately reflect real compliance to treatment. Another possible explanation could be related to dosing. Almost 94% of all patients in this study were treated with statins, of which 61% were treated with simvastatin. The mean dose for those patients treated with simvastatin was 16.74 mg, 43% were treated with 10 mg/day, 52% with 20 mg/day, 4.9% with 40 mg/day and only 0,1% were treated with 80 mg/day. This dosing is in the lower range of what is recommended for patients at high risk . The doses of other statins were also in the lower ranges.
Therapies that modify LDL-C, HDL and/or TG exist but they are not widely utilized. Use of new medical technology varies widely between countries but also between different disease areas within a country . The causes for variations in use of new drugs could be divided into three broad groups, macro-or system level determinants, service organization determinants and clinical practice determinants . Examples of determinants could be spending on pharmaceuticals, role and impact of health technology assessment, guidelines and clinical culture and attitudes etc. Use of statins as well as use of drugs in other therapy areas was low in Sweden when compared to other countries , which could be an indicator and explanation for low uptake of new drugs in general.
Some of the therapies that modify LDL-C, HDL and/or TG are associated with limitations and might explain the low utilization of these therapies. Fibric acid and nicotinic derivatives are two classes with documented effects on increasing HDL-C and association to reduced CV events, but have limited documentation on CV outcomes and are associated with high rates of adverse events or tolerability issues . The most common side effects during fibrate therapy were skin reactions and gastrointestinal symptoms, while flushing was the most common issue during therapy on Niacin  and could be major reason behind the limited use of these treatment options.
Limitations and strengths
Total patients that met all predefined inclusion and exclusion criteria and were included in the analysis were 5424, about 58% of total study population (n = 5424/9384). The inclusion criteria of complete lipid profiles caused an exclusion of about 21% (n = 1933/9384), which may result in selection bias, since patients with high CVD risk should have better documentation and a greater probability of selection in the cohort and this could be a limitation of generalizability. The problem of missing complete lipid profiles was reported by other researchers and the analytical solution used in this study was adapted from their prior work . In short, they compared baseline lipid values for those patients with complete lipid profiles compared to the included patients. If baseline lipid profiles were similar they concluded that those patients with complete profiles were representative for the entire cohort. Using the same methodology in this study no major differences in baseline lipid values were found between excluded and included cases. Nonetheless, one should use caution when extrapolating these data to the general population of patients using LMT.
Another limitation is that our analyses were on patients who got their LMT prescription refilled for at least one year and had complete lipid profiles, which might thus represent a best-case scenario of goal/normal lipid level attainment. Another possible limitation is that it was not possible to differentiate between fasting and non-fasting TG measurements in this study. To address this issue LDL-C measurements were considered invalid if the triglyceride values were >4.5 mmol/L. Furthermore both fasting as well as non-fasting TG act as strong predictors of cardiovascular events .
We calculated 10-year CHD risk >20% using FRS, since this predicts fatal and non fatal CHD events, while the European risk classifying system, SCORE, only predicts fatal events. The FRS system is widely used and validated . However, for certain populations the FRS has been shown to overestimate risk [43–45]. In this study only 258 of all patients (not included in patients with T2DM or CHD) in total sample were found to be at 10-year CHD risk >20%.
The major strength of this study is that patients were identified from real clinical practice, which provides insight on prevalence, treatment patterns and goal attainment in patients with dyslipidema in Sweden and could be of use for decision making within the health care system. For instance, all lipid modifying agents on the Swedish market were reviewed in 2009 by the Swedish decision making body on reimbursement of pharmaceuticals (The Dental and Pharmaceutical Benefits Agency, TLV). The outcomes of the review acknowledged the need for drugs other than statins that could be useful for patients as add on to a statin therapy where statin monotherapy is not sufficient or when statins are not tolerated , which is an important stand point within health policy and may help reducing the burden of CVD.
Complete lipid profiles provide possibilities to assess different types of disorders and can be used to identify patients where additional treatment is needed to target specific dyslipidemia profiles. Our findings indicate this need in particular for patients with T2DM. In the "The Action to Control Cardiovascular Risk in Diabetes" (ACCORD) study however, combination treatment of simvastatin and fenofibrate in patients with T2DM was not found to reduce the rate of fatal cardiovascular events, nonfatal myocardial infarction, or nonfatal stroke, as compared with simvastatin alone . However, the patients in the ACCORD study were at very high risk for CVD, therefore findings can not be generalized to all patients with T2DM. Furthermore there are strong evidence for low HDL-C as well as elevated TGs [8, 12, 32, 37] besides elevated TC and LDL-C as risk factors for CVD and patients at risk should be considered for treatments in addition to statins that target multiple lipid disorders.