Prelude to CHD: Risk estimation furor

Column-Dr Ramon Abarquez photoMedicine is Not an Exact Science

Ramon F. Abarquez, Jr., MD, EFACC, FAsCC, FPCP, FPCC, CSPSH

Current risk-screening tools are not perfect as previously implied and no prospective study has been performed in order to test the hypothesis of risk stratification with subsequent tailored treatment strategy.

Atherogenic dyslipidemia, smoking, or hypertension is influence by both the magnitude and deviation from ‘normal’ as well as the exposure duration. (Sniderman, Lancet 2008;371:1547–1549) Risk influences vary over time, i.e. smoking may be stopped and restarted, therapy changed and living attitudes altered. Compounding risk as increasing obesity with declining HDL and high triglycerides, and insulin resistance provide poor prognosis. (Carroll, JAMA 2005;294:1773–1781) What then are the limitations of risk profiling?

Implications of risk assessments

PROCAM: An important feature of the PROCAM study is the inclusion only of “hard” end points of definite myocardial infarction or sudden coronary death that are unlikely to be overlooked or misdiagnosed even without angiographic documentations. However, the calculated score still included 63 men with CHD diagnosed by angiography or other means. And the score for each risk factor was virtually identical. The Framingham score to determine ACS relative risk (77.8 percent) was significantly less than the PROCAM score (82.4 percent, p < 0.001) since the latest Framingham score does not include information on family history of CHD, triglycerides, or LDL cholesterol. In addition, the validation of both the Framingham and PROCAM scoring systems would require their application to a third independent data set. (Assmann, Circulation. 2002 Jan 22;105(3):310-5.)

FRS: A review of 1,655 abstracts with five cohorts had a total of 13,142 subjects. Mexican cohort event-rate ratio for CHD FRS = 1.68,(1.26-2.11); incident myocardial infarction, 1.36 (0.90- 1.83); and CHD death, 1.21 (0.43-2.00). Among a United State’s cohort that included Hispanics, FRS overestimated CVD risk for Hispanics (0.69).Thus, the FRS overestimates CVD risk in Hispanics when not properly recalibrated. (Cortes-Bergoderi, Rev Panam Salud Publica. 2012 Aug;32(2):131-9)

Is the accuracy of CVD risk estimation similar based on Framingham 1991, Framingham 2008, Reynolds risk, ASSIGN, SCORE and QRISK2), since each differ in the risk factors used to estimate CVD outcomes? The six algorithms were applied to a sample of 500,000 people in England, aged 40-74 to estimate CVD event in 10 years.

“At a 20 percent false-positive rate the detection rates of the six algorithms ranged from 72 percent to 79 percent. The estimated risk cut-offs to achieve the same false-positive rate varied fivefold, from 4 percent to 21 percent because of the different risk factors and outcomes considered.” The predictive variations may be due to the different risk factors and outcomes considered. Nonetheless, all six risk algorithms had similar screening performances.

The accuracy (calibration) of CVD risk estimation does not materially affect screening performance, in distinguishing who will and will not develop CVD. It is screening performance that matters rather than the accuracy of the risk estimation” (Simmonds, J Med Screen. 2012 Dec;19 (4):201-5) However, the classical CV risk assessments failed to account for several emerging CV risk factors. CV risk indexes and risk estimates are computed from different populations without considering specific population features.

Attending physicians’ estimates

Physicians’ perceptions often underestimate their patient’s real CV risk with only 13 percent based on use of a risk chart. (Erahrdt, Eur Heart J Suppl . 2005;7:L11–L15) . In another study, CV risk calculation was not possible due to information deficiency in 43 percent and paper recoded risk computation performed after and out of the medical consultation. (Sheerin, NZ Med J. 2007;120:U2714) . Moreover, in general practice, in 50 percent of patients low levels of information recordings exist. (Holmes, Med Care. 2005;43:I33–I41)

More importantly, less than 25 percent of patients with hypertension have adequate BP control. (Assmann, International Task Force for Prevention of Coronary Heart Disease.http://www.chdtaskforce. com/pocketpc_, Jan 2003:1–128) And, less than 10 percent have both blood pressure and cholesterol controlled. (Kotis, Am J Med. 2007;120:746–747)

Finally, 44 percent of patients have more than one additional CV risk factor (eg, mental stress, obesity, physical activity). Thus, the classical algorithms used for CV risk assessment fail to account for several emerging CV risk factors; CV risk indexes are computed from different populations without considering specific population features; CV risk estimates and electronic medical records are separate; centralized medical database with patient consent allow sharing of patient clinical status between different doctors and specialists leading to real time update/upgrades of medical guidelines, patient data, and standardized medical consultation with personalized therapy and lifestyle changes. (Franchi, Ther Clin Risk Manag. 2011;7:59-68)

Furthermore, current risk-screening tools are not necessarily perfect since the concordance of CV risk scores (PROCAM Score, Framingham Score and ESC Score) with the treating physicians’ risk assessment remains unclear. A study among 8,957 primary-care patients without known CVD (age 40-65 years) to establish AMI or CVD death, the mean 10-year coronary morbidity risk estimated by Prospective Cardiovascular Munster score (PROCAM) (4.9 percent), Framingham Score (10.1 percent), ESC Score (2.9 percent) and primary-care physicians (2.7 percent), concordance in all was only 34 percent.

Among high risk patient groups, Framingham Score (22.6 percent), a medium risk (48 percent) and a low risk (41-46 percent) were compared with only eight percent as recognized by the primarycare physicians. Thus, substantial discordance can occur among risk estimates in the four groups. (Silber, Med Klin (Munich) 2008;103:638–645)

Thus, what matters is whether the risk assessment can discriminate who and who will not have a future event more than the comparative calibrated accuracy. By the way, who examined the patient, the attending physician or the risk scoring system?

VitalSigns Issue 66 Vol. 3, August 1-31, 2014

Be the first to comment

Leave a Reply

Your email address will not be published.