Second-hand science

By Robert Wagner

wile: deception by playing on the victim's weak points.


A study by Pell et al. published in NEJM (2008 359:482) reached the conclusion, "the number of admissions for acute coronary syndrome decreased from 3235 to 2684 - a 17% reduction" following a smoking ban in Scotland in March 2006. This study shows why those numbers are not a measure of acute coronary syndrome, they are a scientific trick intended to deceive.

Heart Attack:

When a patient arrives at an emergency room with chest pain, often accompanied by other cardiac symptoms, he or she is tentatively registered under acute coronary syndrome (ACS), a designation that includes acute myocardial infarction (AMI, heart attack) and uncontrolled angina. The number of such admissions is widely used in studies of heart attack incidence, including the subject study. In fact, only 20% of such patients are having a heart attack, the other 80% are having angina, other heart problems, or something unrelated to the heart. Because time is of the essence in treating AMI, emergency staff quickly administer three tests in order to distinguish AMI from other causes. When two out of three are positive, the diagnosis is AMI (I-21 on the ICD-10 system) and treatment for heart attack begins. The tests are:

EKG, elevated ST and peaked T waves indicate in-progress MI
Blood test for troponin (I and T), indicates heart muscle damage
History of individual or familial heart disease


The Scottish health database at ISD shows admissions for ACS (April through March)

1999: 24,000
2000: 22,500 -6%
2001: 21,500 -4%
2002: 21,000 -2%
2003: 20,000 -5%
2004: 18,000 -10%
2005: 17,000 -6%
2006: 16,199 -5% pre ban
2007: 15,000 -7 post ban
2008: 16,212 +8%

The Pell study, using data from 63% of the patient load for 10 months (June through March), found the numbers in Summary above. Extrapolated to 100% for 12 months gives:

2006: 6,162 pre ban
2007: 5,112 (-17%) post ban

The trick:

Why are Pell's numbers one third of the official admission counts? Because she counted only patients having a non-zero troponin reading.

"Acute coronary syndrome was defined as a detectable level of cardiac ?troponin after an emergency admission for chest pain."

Wrong. ACS is defined as AMI plus angina. Elevated troponin is what distinguishes AMI from angina. She's trying to measure AMI (heart attack), but she's doing it wrong. She's actually measuring the 20% with AMI plus an additional 15% false positives. The trick is in the false positives, because they are going down every year as the troponin test is improved.

The troponin test: The troponin test came into common usage in 2000, replacing earlier tests such as creatinine (CK) which were less specific to the heart muscle. A major complaint about troponin is that it delivers too many false positives at low levels. Many conditions cause small concentrations of troponin in the blood; only heart muscle damage causes large concentrations (with a few rare exceptions). The threshold level for a positive reading must be set high enough to eliminate the false positives. Setting the threshold at zero, as Pell did, is scientific nonsense, because it opens the door to MANY false positives. "

The Joint European Society of Cardiology/ACC Committee for the Redefinition of Myocardial Infarction has recommended that an increased concentration of cTn be defined as a measurement exceeding the 99th percentile of cTn concentrations observed in a healthy reference group."

That means testing 100 healthy people with no heart damage, sorting their readings in ascending order, numbering them from 1 to 100, using the reading of person number 99 as the threshold. Each of approx. 20 tests on the market has a different threshold. The ratio between the highest and lowest is about 50; there is a lot of imprecision in this test. The 99 percentile cutoff supplied by the manufacturer is used by all ER staff and cardiologists, although some argue the cutoff can safely be set three times higher. A sample of complaints about the imprecision of the test:

"Current second generation cTnI and fourth generation cTnT assays generally have an imprecision of around 20% coefficient of variation (CV) at the 99th percentile of the reference population, which is greater than the recommended imprecision of 10% CV. As the next generation of more analytically sensitive cTn assays are developed it can be anticipated that cTn upper reference limits will decrease by approximately 10-fold."

"Because this is a very low threshold, the total imprecision (CV) at this level should be less than 10% to minimize false-positive results. This means that the measured cTn values from one serum sample, when evaluated on at least three occasions, cannot vary by more than 10%. If they do, assay imprecision should be suspected. This information is critical to understanding the current debate regarding cTn testing, because most of the 20 different assays on the market are not precise at the threshold value suggested for determining a positive result (99th percentile in a healthy reference group). Many false-positive results occur at the lower ranges of analytic sensitivity, where it is most difficult to separate clinically relevant myocardial cell damage from spurious cTnI elevation."

Test manufacturers are listening, and improving the test every year to REDUCE FALSE POSITIVES (Pell's trick). We saw above that the two tests are in their 2nd and 4th generations since 2000. A sample from marketing literature in the period 2006 to 2007, the interval between Pell's before and after tests:

"Ortho-Clinical Diagnostics, a Johnson & Johnson company, announces the availability in the United States, Europe and other international markets of an improved sensitivity troponin I assay."

"Highly sensitive and specific assays of cardiac troponins I and T are the preferred biomarkers in diagnosing myocardial infarction (MI). Assays of cardiac troponin I (cTnI) have been improved with the addition of antibodies against the cTnI molecule and may have increased sensitivity."


The 17% reduction reported by Pell consisted of 7% from an ongoing secular trend and a reduction in false positives from 15% to 10%, which produced the appearance of an additional 10% reduction in "ACS".

Robert Wagner (

Reprinted with the kind permission of the author.