Implementation of the 2013 US Lipid Guidelines in a Canadian Family Practice: a four week experience.
Completion date: 2014/07/09
Research Question: Will use of the 2013 Lipid Guidelines change clinical decision making, and in what manner?
The research was conducted in Lincoln, ON, a Niagara region town of 23,000 on southern Lake Ontario. The demographics of this area is diverse: Niagara is the second most popular retirement area in Canada behind Vancouver; from here, a short commute is possible to Toronto, Canada’s biggest city; Lincoln has residential, commercial, agricultural, and light industrial areas.
The researcher is part clinic owner of one of the Lincoln groups, this particular group comprised of seven physicians. The research site is this family practice clinic only, and the people involved will be primarily that senior physician, and his patients that identify themselves as needing assessment during the weeks of implementation. Most patients presented during implementation period, but some decisions were made based on chart data, and collected lab work. In this circumstance, nursing staff cooperated in patient negotiation by phone.
Those patients where consideration is being given to lipid screening and possible treatment, be it mandated by any circumstance.
Worldwide, physicians are still treating lipids to target. The mantra has always been, “lower is better.” Not only do high risk patients get this treatment, but health oriented young people with family history of cardiac events, for example, are being put through Framingham risk analysis, and being put on statins.
How often should one monitor lipids? How aggressive should one be? What is an acceptable risk level? Would the new US guidelines, suggesting that there is no proof of treating to target, ease the approach to lipid treatment? How would decision making change in regards choice of agent, dose, and frequency of testing advised?
According to a 2014 review of the frequency of testing actually ordered in Canada vs the actual current recommendations (Canadian, not based on US lipid Guidelines), large segments of the population are actually under screened (THETA, 2014). This document suggests there would be a marked financial impact if all Canadian lipid testing guidelines were actually followed, requiring an extra 3.6 million tests at almost $15 a piece. Currently, on average, very low risk (<1-5%) patients are being screened every 4.4 years, low risk (6-9%) q 2 yrs, intermediate (10-19%) q 1.4 years, and high risk (≥ 20%) yearly. This study did not look at the cost of LFT testing, or CK.
Current Canadian standards for monitoring lipids indicate tests yearly for those with a Framingham risk score of over 5%, and every 3-5 years if less than that (THETA, 2014). Out of the 21,000 persons who completed the Cdn Community Health Survey in 2005, 37.7% were in the very low risk group, 41.3% in the low risk group, 16.9% in the intermediate group, and 4.1% in the high risk group.
Although not expressly stated in this study (THETA, 2014), it appears that this work only speaks to monitoring and screening, ie following patients that have hit and acquired lipid targets, or screening non treated patients. My habit as a practising GP would be to retest lipid levels, and LFTs, 6/12 after a medication change. If goals were not hit, the dose would go up, and I would again retest. I have personally stopped doing CK testing unless there is myalgia.
In late 2013, the new US Lipid Guidelines were released (Stone, et al). There were major changes. After a very large literature review, this group reported that no current clear evidence existed that supported treating to target. There were substantial changes brought into new equations to determine risk, including differentiation in regards to race, and emphasis on chronological age. LFT’s were deemed not necessary to screen, unless there were symptoms. The Framingham risk calculator was replaced. Treatment in the past was recommended at the 10% risk level (for patients with risk factors). Now the risk is set for 7.5%.
The Framingham risk calculator was based on data from the Framingham Heart Study, and gave points, or subtracted points based on the following criteria: age, gender, total cholesterol, smoking, HDL level, systolic BP, if the patient was on BP medication. Newer Framingham scores excluded DM as it was considered an equivalent to CAD.
New Pooled Cohort Equations are presented to estimate 10 year ASCVD risk (Stone, et al, 2013). Stroke now is included in ASCVD risk assessment, in addition to myocardial infarction (MI). Separate equations now are advocated for nonwhite populations (African American). Inadequate data existed to work out calculators for other specific groups, such as Hispanic, Asians, native Americans. Authors suggest to use the calculator for Caucasian populations for these groups, but estimates may under or overestimate the risk for persons from many race/ethnic groups. American Indians, some Asian Americans (ie. south Asia), and some Hispanics (ie. Puerto Rico) may actually have a higher risk than is estimated. Calculations seem to overestimate the risk for some other Asian Americans (ie. east Asia) and some other Hispanics (ie. Mexico) (2013 Prevention Guidelines Tools: CV Risk Calculator. My American Heart for Professionals).
Although Guideline writers take pains to emphasize a desired conversation between physician and patient, a consensus type approach to picking an intensity of statin treatment, really, there is little room to move. Of the 4 statin benefit groups, 2 demand high intensity treatment, and 2 moderate to high intensity.
This overall change in orientation away from targets could have major impact in decision making on a day to day basis. Otherwise young, healthy people could be taken off statins. Older patients, previously drug free, may find themselves going on statins as this risk level has been lowered. Less lab work seems needed in following patients once the decision has been made to embark on treatment, as targets are not supported.
According to 2013 ACC/AHA Guidelines (Goff, et al, 2013), other patient groups are also considered high risk, including ALL diabetics (regardless of age), cerebrovascular disease, AAA, and chronic kidney disease, regardless of lipid levels.
Summary of the 2013 US Lipid Guidelines:
There is good evidence to treat in the following 4 statin benefit groups:
A) Patients with clinical ASCVD, without class II-IV heart failure, or on hemodialysis.
> this group gets high intensity statin Rx.
B) Patients with LDL-C ≥ 4.9 (mmol/l understood).
> this group gets high intensity statin Rx.
C) Diabetic Patients 40-75 with LDL 1.8-4.89.
> as far as I can see, this is a bit controversial. If all Diabetics are “high risk” as per AHA/ACC, that infers, demands, high intensity statin Rx. US Lipid guidelines however state here moderate intensity, with an increase to high intensity with an ASCVD risk calculation of ≥ 7.5%.
D) Non DM, Non ASCVD Patients 40-75 with LDL -C 1.8-4.89, and with a 10 year ASCVD risk of 7.5% or more.
> authors recommend Rx with moderate to high intensity statin.
High intensity statin doses: crestor 20, (only one reported RCT had crestor doses > 20 mg) or lipitor 40-80 mg. High risk patient, high dose statin. NB the other high risk groups listed above, including renal disease, CVD, AAA, all diabetics.
Moderate intensity statin doses: Lipitor 10-20, Crestor 5- 10 mg, zocor 20-40, pravachol 40-80, lovastatin 40, fluvastatin 40 BID. Low dose Pravastatin 10-20, Lovastatin 20, zocor 10, fluvastatin 20-40, pitavastatin 1 mg. Low dose is used if unacceptable side effects at higher doses.
Intermediate risk patients get a lower dose of statin. Again, there is no evidence for treating to target. Use the lower dose, ie crestor 10, and do not push higher to attain lower lipid values.
For ALL Patients, Regardless of Risk Factors, Implement lifestyle recommendations:
• DASH -like diet
• Physical activity
• Weight management
• Complete smoking cessation
Does one treat outside these four well proven groups?
In intermediate risk (defined as 5% to 7.5% 10 year ASCVD risk by Pooled Cohort Equation),
consider hs CRP, and /or CT scan for coronary artery calcium (CAC) score and/ or FHx of premature heart disease in considering whether or not to treat (Stone et al, 2013). If added risk factors, consider to treat. One may consider statins in those with LDL-C ≥ 4.14, genetic hyperlipidemias, FHx of ASCVD first degree male relative <55 or female < 65, hsCRP ≥ 2, CAC score ≥ 300 Agatston units (calcium scoring in the results from aCT Test for Coronary Calcification) or ≥ 75th percentile, ABI < 0.9, or elevated lifetime risk of ASCVD.
No statin RCTs reporting on-treatment non-HDL–C levels were identified, for either primary or secondary prevention (Stone, et al, 2013).
There is no clear evidence for using non statins (Stone, et al, 2013): nystatin, ezetimibe, Omega 3’s, Co E Q 10, etc.
Who matches these criteria in the US? Who has > 7.5% risk? (Based on the US National Health and Nutrition Examination Survey, or NHANES data) (McKibben, et al):
In men over 50: 50 % of all African American men and 30% of white men.
In men over 70: Almost all men.
In women in their 60s: 70% of African American women and 60% of white women.
A decision was made to implement the US Guidelines over a four week period, and document where changes were made to current care, or advised and declined. Prior to implementation, the expectation was that overall, statin use may actually decline, and that less lab monitoring would be necessary.
Some collaboration with staff was necessary, as nurses would on occasion pass on advice about changes in statin dose, etc. When directions were passed to the RN, pertinent aspects of the Guidelines were reviewed at that time.
Flow sheets were created to track data: diagnosis, ASCVD risk, drug chosen, dose, and a large area for individual consideration/ comments. An iPhone App was used to calculate 10 year US ASCVD risk (see below, under references).
After appropriate patient consultation, those that fell into the high risk group (clinical ASCVD, or DM, etc, or patients with LDL-C ≥ 4.9) would be moved to high strength statins. In these situations, 10 year risk was not calculated. Those that fell in the other groups would have a calculation done.
During the four weeks of implementation, patients presented as they normally would. During these encounters, effort was taken to assess whether or not the patient was using a statin, and to examine past laboratory work. Visits did not have to be explicitly for this reason. For example, a patient visiting with a sore throat on a statin would have his dose examined, and risk calculated.
Discussions in regards personal risk, and treatment, were made during that appointment.
Daily, lab results were received, many ordered many weeks previously. In review of that data, patient charts were pulled up, and risk assessed. Where there was clear indication to change treatment path, this information was sent to the nurse for a patient phone call. Approximately two thirds of the data here represent a patient / physician direct interaction. The others were handled by the RN on the telephone. This is an estimation: this information was not explicitly gathered, and could have been helpful, especially if there were refusals for change in treatment. However, as can be seen in results below, there were only 2 declines, and they were in the physician group.
Conversations in regards personal patient risk, and choice of treatment path added approximately 5 minutes on to each interaction. This time was not formally recorded, but in all circumstances, was relatively brief. On all occasions, the patients appeared pleased to have this discussion initiated.
On every occasion, after discussion of the possible risks of statin therapy, the patients involved preferred to have LFTs reassessed.
Using the iPhone app was a bit of a bother, compared to working through the built in risk calculator in PS Suite. During the week before actual implementation, attempt were made to make decisions based on the built in calculator. This is a clear error, that may have led to less than ideal advice. Using the US guidelines mandates the use of the new pooled equations. It is possible that several of the decisions made prior to the actual implementation were in error, because of the inappropriate calculator. During this study, two of these patients were found. New calculations actually indicated to treat.
The iPhone app (below, by Nguyen) actually facilitated using the metric measurements.
Documentation of adjustments:
During the four weeks of implementation, very rapidly it became apparent that less lab work was being ordered by myself. Although flow sheets initially had a column expressly for documenting when lab work was ordered, this was quickly ignored, as the same pattern developed for every patient. Reassessment of LFT’s was ordered for six months. No reassessment of lipids was booked. Some patients steadfastly wanted to know their numbers.
Although the calculators are meant for those not currently being treated with statins, on many occasions 10 year ASCVD risk assessments were done using lab values from patients actually being treated. The majority of these results were fasting, but there was an occasion where a random, non HDL-C blood results appeared pertinent, where it was felt a decision could be made. This hypertensive patient using zocor had an over 22 % calculated risk. That patient chose high intensity treatment.
On one patient with uncontrolled BP, a 74 year old male on crestor 10, calculations were done using ideal BP. A calculated risk of over 20% helped this patient decide on dosing.
Analysis and Reporting:
On 44 occasions during the four week implementation period, decisions were made to change treatment. Of these patients, all but one were from my personal practice. One patient was from a partner’s practice, who prefered to return to his primary provider for his advice on the dose change.
All patients were Caucasian, except for one, a native from Barbados, diabetic.
Some decisions were made without a direct patient contact, after blood work and chart review. These patients were not so identified, but were probably about a third of the 44.
The age range involved was from 42-85. Average age (for those with age documented) was 68.9 years. On 15 occasions, there was no sex identified on the data chart. Of the 44, half did not have sex identified on the chart. Of those identified, there were 10 females, and 12 males.
Of the 44 changes, 15 were of a dose increase. 3 had dose decreases. There were 18 new statin starts, and 2 decisions to stop. Ezetrol was stopped in one patient. 4 patients switched medications, and 2 were left on the same dose after patient consultation (ie declined dose change).
13 of the 44 patients where decisions were made to change treatment did not have a 10 year ASCVD risk calculated. These patients had one or more of DM, ASHD, or vasculopathy, and were changed to high intensity statin based on diagnosis. ASCVD calculations ranged from 0.9% to 42.5%. Interestingly, the patient with a 0.9% risk was 42, female, and had a LDL clearly above 4.9. The male with the 42.5% risk was 80, and the risk was actually calculated on crestor 10.
On 8 occasions, treatment was changed based on ASCVD calculations done on blood results drawn while the patient was on a statin. The calculator is intended for those NOT on treatment.
Of the 44 patients, 4 had ASHD, 5 DM, 8 HT, 13 hyperlipidemia. 8 had HT plus DM or glucose intolerance. 2 had ASHD and DM. 3 lipids, HT. One HT/ vasculopath.
Of the 44, 38 ended up on Crestor. 2 came off Crestor. 3 Lipitor, 1 Pravachol. Ezetrol was discontinued in one patient. 2 patients moved from Zocor to Crestor.
There were fourteen 10 mg crestor starts, and five 20 mg starts. Of these 20 mg starts, one had cerebral ischemia (no risk calculated); one hypertensive had a 22.3% risk calculated on non fasting results while on zocor; one hypertensive had a ASCVD risk of 26% and a systolic BP of 170; one patient HT past 25 pack year smoker with a calculation of 22.3% also on zocor; and one hypertensive with a 13% risk chose to start with 20 mg.
The patient started on Pravachol had previous past strong adverse effects with both Crestor and Lipitor. Female, 78, recent TIA. Although a strong statin would be called for, 20 mg Pravachol was decided upon as a low dose attempt to treat.
Patients were universally accepting of the US Guidelines, despite the fact they had not yet been accepted in this country. The fact that Canadian Guidelines usually follow the US path was readily appreciated by every patient. There was absolutely no “resentment”, nationalistic or negative reaction to the “US” guidelines.
Out of the 44 people where there was a change in treatment recommended, there were a couple that preferred a higher dose after the above “conversation.” One man was previously struggling with accepting a surgical treatment for prostate cancer, and had successfully embraced treatment, with a good outcome. It seemed to signify an acceptance of medical intervention itself, and a more aggressive stance towards prevention that he had decided upon. Another had a very poor family history, and a risk of > 20 % calculated on lipids performed while on zocor.
2 patients rejected the advice to treat with a statin at the indicated dose. All patients accepted treatment with a statin if advised. No patient wanted to read the guidelines.
Although new US Guidelines suggest that screening LFTs is not necessary unless there are appropriate symptoms, patients were much more comfortable accepting medication if such tests were ordered. Patients universally preferred non fasting blood work, and were disappointed when asked for fasting to do the calculations. They were happy to not fast on recheck blood work.
No patients called in or reported adverse effects to the statins during the implementation period, and I have yet to have a patient report same since this time (*** weeks). There have been no cardiac, etc, “events” reported.
Conclusions for a practising Family MD:
Facts to consider:
> The 10 yr ASCVD risk calculator replaces the Framingham risk calculator in recommendations.
> The Framingham risk score (SI units) varies from the 10 yr ASCVD risk calculator (imperial units). The worked out scores are different. As an example, on one test comparison, when the same data was entered into both calculators, the Framingham calculation led to a risk score of 5.6%, and the ASCVD a score of 2% (of patients having a cardiac event in 10 yrs).
> The PS Suite cardiovascular risk calculator is NOT the new recommended Pooled Cohort Equations. The built in PS Suite risk calculator is convenient to use, but is not exactly the Framingham calculator, either. All these calculators do allow one to identify high risk patients, by different calculations. However, in previous (pre 2013) recommendations, the AHA threshold for treatment of with statins was a 20% risk of MI over 10 years or > 10% with multiple risk factors. (Compare this with the 2013 recommendations of treatment > 7.5% risk.) Clearly, the designation of “high risk” is now redefined. Most feel that this may increase statin use.
> None of the calculators include CRP, genetic hyperlipidemias, coronary calcification calculations, or ABI. These are other factors acknowledged to be important to consider.
> Risk calculators vary and have their own advantages. If one is to implement US lipid guidelines, the 10 year ASCVD risk calculator based on new pooled cohort equations should be used. A major problem inherent in this approach are the SI units reported by Canadian labs, and the defined actual calculators that are in imperial units. There is no easily found conversion table between calculators. Decisions made with the wrong calculator can indicate a completely different course of action. There is a convenient iPhone app, be it “unofficial,” with SI units.
> Renal patients of course could need dose adjustments. Atorvastatin (Lipitor) has < 2% renal excretion. Crestor, 10%. Pravachol, 20%, Zocor, 13% (Harper and Jacobson, 2008). Treatment of hemodialysis patients is more controversial it appears, but clearly outside of a family doctor’s bailiwick.
> CTs for coronary calcification seem to be rarely ordered by Canadian cardiologists.
>There is a new push to do non fasting blood work, looking at non HDL-C blood work. No statin RCTs reporting on-treatment non-HDL–C levels were identified, for either primary or secondary prevention. If one is to adhere to US lipid guidelines, one should consider doing FASTING blood work for risk screening. Why are we screening people with these tests? They are appropriate after the decision has been made to treat, and the dose decided upon. However, if one does not treat to target, what’s the point of ordering them at all in follow up?
> High risk patients get crestor 20 or lipitor 40-80, regardless of lipid values. There is no data that supports treating to target, only theory. If such a patient is on crestor 20, there is little value in lipid testing. If a patient is on lipitor 40, is that enough? If targets are meaningless, decisions to choose a certain lipitor dose would be based on other risk factors, none of this being clearly defined.
> Lower risk, lower statin dose. However, in intermediate risk, Guidelines certainly support using high intensity doses. There are multiple factors that are recognized, that encourage a physician to use a higher dose of statin.
Conclusions from this study’s small cohort, time frame:
Results from this small practice study did not bear out my prediction that the new US Lipid Guidelines would lead to many patients coming off statins. Of the 44 changes, there were 18 starts, and 15 increases, with only 3 decreases in dose.
3 patients were seen during this implementation period that were previously told to discontinue their statin based on the US Lipid Guidelines, after having calculations made with the built in PS Suite calculator. All of these patients had their statin restarted. If one chooses to use the US Guidelines, the US calculator must be used.
The 13 patients that had a statin started based only (without blood work) on their diagnosis of CAD, DM, vasculopathy, etc, were completely accepting of this treatment path.
The use of the Guidelines facilitated the choice of a patient, previously frightened of using statins, to again make an attempt.
It was found that decisions were possible based on bloods done while already on a statin in 8 circumstances. Calculations based on these numbers were helpful in counselling patients.
There was a large decrease in the number of lab tests I ordered, although patients continued to desire LFTs to be done at six months. Recheck cholesterols were not ordered. I have plans of doing yearly checks, as evidence and directions may change.
A) The basic drug chosen was Crestor:
I found myself always coming back to this drug. The decision making was easier. It had been my favourite before, since the Jupiter study, but dosing decisions are simply easy with this drug. Some non entered patients (as no change occurred with their treatment, a qualifier for this study) stayed on Lipitor 80, clearly a top end dose for a high risk situation.
One renal failure patient (also not on this study as I didn’t change anything), remained on lipitor 5 mg (the renal friendly strong statin) as directed by his nephrologist.
One patient with past adverse effects to both crestor and lipitor, 33% risk, current possible TIAs and previous refusal to use statins accepted a trial of pravachol 20 mg, a lower strength medication at a dose considered to be low.
B) Reactions to “Disease Proof” (Katz, 2013):
This book, as enlightening as I have found it to be, is possibly over the head of many patients, and may not be the ideal resource for them. I loaned this book out to two people, and recommended it to many. It may be that many patients just don’t want to hear the advice. Many came in toting other books more heavily illustrated, with less text.
Goff DC Jr, Lloyd-Jones DM, Bennett G, Coady S, D’Agostino RB Sr, Gibbons R, Greenland P, Lackland DT, Levy D, O’Donnell CJ, Robinson JG, Schwartz JS, Shero ST, Smith SC Jr, Sorlie P, Stone NJ, Wilson PW. (2014, Jul 1). 2013 ACC/AHA Guidelineon the Assessment of Cardiovascular Risk: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.J Am Coll Cardiol. 2014, Jul 1 ;63(25 Pt B):2935-59. doi: 10.1016/j.jacc.2013.11.005. Epub 2013 Nov 12.
Stone NJ,Robinson J,Lichtenstein AH,Bairey Merz CN, Lloyd-Jones DM, Blum CB, McBride P, Eckel RH, Schwartz JS, Goldberg AC, Shero ST, Gordon D, Smith SC Jr, Levy D, Watson K, Wilson PW. (2013, Nov 7). 2013 ACC/AHA Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults. A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2013 Nov 7. pii: S0735-1097(13)06028-2. doi: 10.1016/j.jacc.2013.11.002.
2013 ACC/AHA Guidelines on the Assessment of Atherosclerotic Cardiovascular Risk:
Overview and Commentary :The Johns Hopkins Ciccarone Center for the Prevention of Heart Disease. Becky McKibben, MPH; Seth S. Martin, MD; Roger S. Blumenthal, MD
Harper, C.R., Jacobson, T.A. (2008). Managing Dyslipidemia in Chronic Kidney Disease.
J Am Coll Cardiol.;51(25):2375-2384. doi:10.1016/j.jacc.2008.03.025
THETA. Frequency of Testing for Dyslipidemia: A Systematic Review and Budget Impact Analysis. Ont Health Technol Assess Ser [Internet]. 2014 May:14(7):1-27.
Available from: http://www.hqontario.ca/evidence/publications-and-ohtac-recommendations/ontario-health-
(Published jointly by ACC and AHA | © 2014)
> this calculator is set up for imperial units. It has an excellent infographic, and clear brief summaries of recommendations.
2013 Prevention Guidelines Tools: CV Risk Calculator. My American Heart for Professionals.
> links to download apple android and web versions, official calculators from AHA. Imperial units.
Cardiac Risk Assist calculator, free app for iPhone, optimized for iPhone 5, works for 4s.
Tin T.D. Nguyen, MD@ tinm9n; www.cvriskassist.com
> this iPhone app can be used with SI units.
Katz, D. L. (2013). Disease Proof. The remarkable truth about what makes us well. New York: Penguin.
There were many shortcomings in this study. Obvious ones include the many interactions where age and sex were not entered into the table. I’m now sitting here wondering about all the patients where there was no change in treatment. I feel there was such an opportunity, partially lost. I take solace in the fact that my practice is large, and busy. I feel it was a reasonable attempt at scrutiny. Many will point to the small cohort size, and the lack of hard statistics. However, this is “action research”, a type of semi qualitative study that many physicians have not encountered. This is the kind of research that some social scientists, such as teachers, do to improve performance, not only of their students, but of themselves.
Implementation of this study led to several personal insights as a health care provider.
Initiating a semi formal study of one’s practice improves quality, in many aspects. Although efforts are always made to be complete in every patient interaction, this study forced focus to a different perspective, and broadened the interaction in many circumstances. Formally documenting one’s efforts leads to self challenge, further reading, and more complete understanding. Initiating a study such as this for a four week period forces a clear physician behaviour change that lasts after the study is finished.
I have to ask myself, that if I had not forced myself to formally write up this endeavour, would I still be using the built in PS Suite risk calculator, and attempting to use that data to apply the 2013 US Lipid Guidelines? Writing this up is painful, and laborious, and worth every minute.
Finally, I call upon David Katz to produce a website with animation limited to less than 10 minutes, using exploding colours and sound effects.