Familial Hypercholesterolemia (FH) is often misdiagnosed as plain old high cholesterol because of overlapping symptoms such as elevated lipid levels.
According to a recent article in MedicalView, FH is three times more likely to cause coronary artery disease than the common form of high cholesterol. There is an urgent need to detect the disease in its early stages.
A New Approach
The algorithm (formula) created by researchers in the School of Medicine at Stanford University can now detect a person’s risk of getting the rare genetic FH disease with eighty-eight percent accuracy.
The algorithm was created from the data of 197 FH patients and 6,590 patients who did not have the disease. The data collected included patients medication, lipid levels, family history, lab tests, and other details. By inputting this information the algorithm could determine what combination of factors indicate disease.
For those people flagged by this new method, the next step would be further genetic testing to confirm the AI results.
About Early Diagnosis
Early diagnosis is crucial because heart attacks are likely to occur in approximately fifty percent of men before the age of fifty who have FH and thirty percent of women before the age of sixty with the disease.
Early screening is also important because the disease is genetic. When someone in the family has FH the chance of others in the same family having the disease is relatively high.
The Practical Aspect of the Screening
The new method of screening solves the problem that a hospital would have if it attempted to screen every patient for FH who walks through the door. This would not only be costly and a burden but it would not be feasible since the ratio of cases is about one in one hundred. Therefore although there is an urgent need to locate FH carriers, this approach is impractical.
The researchers believe that they have solved this problem with their new method of screening. Their focus is on people who already show signs of high cholesterol or heart disease. The algorithm, just like a sieve, filters out all but those who show a possible predisposition to FH.
When someone is flagged by the screening, there is an eighty percent chance that they have the disease. These people would receive further testing and if FH is confirmed they would be treated immediately to lower LDL.
About FH, Cholesterol and Coronary Artery Disease
Normal cholesterol has several positive functions such as building cell membranes and producing specific hormones and substances that help digestion. Cholesterol is found in animal products such as egg yolks, poultry, meat, dairy, and fish.
The FH mutation prevents the clearance of LDL (low-density lipoprotein) and causes coronary artery disease.
The series of events in coronary artery disease begins when there is a high level of cholesterol in the bloodstream. The excess cholesterol finds its way into the walls of the blood vessels, including the coronary arteries leading to the heart.
The cholesterol forms clumps (plaque) that cause the walls of the arteries to become narrow and harden. This, in turn, can reduce the amount of blood that flows to the heart. The buildup can cause chest pain (angina) and increases the risk of a heart attack.
About Familial Hypercholesterolemia (FH)
FH is an autosomal dominant inherited disease meaning that one copy of a mutated (altered) gene in a cell is sufficient to cause the disease. The altered gene is passed down from one affected parent and a normal gene is inherited from the non-affected parent.
Inherited FH may involve excess cholesterol in tendons. This creates growths that are called tendon xanthomas. They often appear on the Achilles’ tendons or they may affect fingers and hands.
Two areas of the eyes may be affected by FH. Cholesterol may accumulate under the skin that forms the eyelids. This is known as xanthelasmata. The second area affected by an accumulation of cholesterol is at the edge of the cornea. This is called arcus cornealis.
AI in the Clinic
The newly “trained” algorithm was put through its first test by running through 70,000 records it had not previously encountered. One hundred charts that had been flagged by the algorithm were reviewed. The team found that it was eighty-eight percent accurate.
A second test included 466 patients with FH and 5,000 patients who did not have the disease. In this case, 85 percent accuracy was achieved. The researchers knew that many of these patients had an FH diagnosis that was confirmed with genetic sequencing.
These results were enough to convince them that the next step is to obtain approval for the algorithm so that it can be used in clinics.
