Lowering the Thresholds of Disease—Are Any of Us Still Healthy?

One of the most pervasive trends in contemporary medicine is the lowering the threshold of markers that indicate “disease” to detect diseases at an earlier stage so they can be treated before they do too much harm. It is happening everywhere: the acceptable levels of blood pressure (BP), blood glucose, blood cholesterol, and even the prostate-specific antigen for detecting prostate cancer are all being set lower. The latest national guideline on hypertension, in the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7),1 has identified a group of people with Bps of 120–139 mm Hg systolic and 80–89 mm Hg diastolic as having “prehypertension” on the grounds that some of them will go on to develop true hypertension, and an even smaller number will suffer a stroke or heart attack, and that, on average, the level of risk for this group is higher than the risk of people whose BP is <120 mm Hg.2 The stated rationale for identifying this group was that such individuals should take hygienic steps to prevent their BP from rising any further. The main justification for the JNC 7 committee members' concern about prehypertension was two publications from the Framingham Heart Study,2, 3 the mother of all population studies. The first publication showed that people who have BP in the high-normal or prehypertensive range do have a higher risk of heart disease and stroke than people with Bps below this level. They are also more likely to progress to real hypertension (>140/90 mm Hg).2 The second paper stated that the lifetime risk of hypertension approaches 90%.3 This is somewhat misleading, however, because the criterion for “hypertension” was a BP >140/90 mm Hg or being put on antihypertensive treatment. Although a pressure >140 mm Hg in an older person predicts increased risk, there are no studies showing that treating systolic pressure from 140–160 mm Hg in people >65 years is beneficial. For the more relevant criterion of a systolic pressure >160 mm Hg or being on treatment, the lifetime risk was closer to 70% for both men and women. The numbers of prehypertensives are huge. A recent analysis of the National Health and Nutrition Examination Surveys (NHANES)4 data found that 39% of adults>20 years are normotensive, 31% prehypertensive, and 29% hypertensive. The strongest argument for treating prehypertension comes from two studies that showed that treating patients with “high-normal” BP lowers their risk. One was the Heart Outcomes Prevention Evaluation (HOPE) trial,5 which enrolled people who were at high risk, not because of their BP (the average was 139/79 mm Hg) but because of other risk factors, such as diabetes and high cholesterol, and showed that treating them with an angiotensin-converting enzyme inhibitor as part of a treatment regimen reduced the rates of heart attacks, death, and strokes by ≥20% compared to placebo. The other was the Perindopril pROtection aGainst Recurrent Stroke Study (PROGRESS),6 which examined the effects of BP reduction in people who had already had one stroke and also found that treating the high normals (average pressure 136/79 mm Hg) had benefits in terms of a lowered chance of a recurrent stroke. These were not “normal” subjects, however, since they had all experienced a previous stroke. The important thing about these two studies is that the participants were at very high risk; these studies should not be interpreted to mean that everyone who has a BP in the high-normal range would benefit from having it lowered. The message is that we should be making treatment decisions according to risk, not BP. Prehypertension has also surfaced in the recently published recommendations of the National High Blood Pressure Education Working Group on High Blood Pressure in Children and Adolescents.7 By their criteria, any child whose BP is between the 90th and 95th percentiles, or any adolescent with a pressure >120/80 mm Hg, is considered prehypertensive. For these children, as for adults, the recommendation is for therapeutic lifestyle changes, principally diet modification and exercise. A second example of lowering the threshold of disease is blood glucose. The American Diabetes Association (ADA) has recommended two changes in the past few years. In 1997, it proposed changing the criterion for diagnosing diabetes from a plasma glucose >200 mg/dL measured 2 hours after a glucose tolerance test to a fasting glucose >126 mg/dL,8 and it introduced an intermediate group of subjects whose blood glucose levels were considered too high to be considered normal. This category was officially termed impaired fasting glucose (IFG) and was defined as fasting plasma glucose that is from 110–126 mg/dL. This is now regarded as the gray zone of prediabetes, which is equivalent to prehypertension. In 2003, the ADA lowered the threshold level of normal fasting plasma glucose from 110 mg/dL to 100 mg/dL.9 As with BP, any definition of the upper limit of normal fasting glucose is arbitrary because there is no clear threshold above which cardiovascular (CV) risk starts to increase. The rationale for choosing 100 mg/dL was two-fold: first, a value of 100 gives a more optimal sensitivity and specificity (on a receiver-operating curve) than 110 for predicting future diabetes; second, the higher value of 110 identifies a smaller number of diabetics than the 2-hour glucose, which is the traditional measurement for defining diabetes, and which has been shown to be superior to fasting plasma glucose in predicting mortality.10 The new threshold of 100 mg/dL for fasting plasma glucose, however, has been vigorously challenged.11 Based on the NHANES data, it has been pointed out that the number of Americans >20 years with IFG will increase by nearly three-fold, from 12 to 44 million. For individuals >65 years, ≥40% will have it.11 Another more recent analysis of databases from Denmark, China, India, and the United States also questioned the utility of the AD As new threshold.12 The criticism was two-fold: first, IFG is a relatively poor predictor of CV risk when compared with impaired glucose tolerance; second, the new definition will increase the number of subjects with prediabetes. The authors concluded: “…the likely impact the prevalence of IFG is enormous, creating a pandemic of prediabetes, unsupported by evidence that this will positively impact preventive actions and interventions. It is also questionable and to a large extent unknown what the prognostic impact of this labeling would be, but it cannot be ruled out that this labeling would affect the quality of life as well as finances (health insurance, life insurance, etc.) of those affected." Then where does this lowering of disease thresholds bring us? The first accomplishment, obviously, is that we suddenly have many more “unhealthy” people in this country. In the case of prehypertension, it is actually only a minority of us who are truly normotensive. If we accept the Framingham data quoted above, then that number will be <10% if we happen to live long enough. With blood glucose, 25% of the population is now labeled as being unhealthy, rising to 50% in the elderly. Since obesity is common to both conditions, there will obviously be many people who have both prehypertension and prediabetes. The epidemiologists will assure us that we have identified people who may look and feel quite healthy but who are at an increased risk of developing real disease, but how big is this risk? JNC 7 emphasizes that BPrelated risk starts to increase above a systolic pressure of 115 mm Hg. We are trained to think of risk in relative terms, such that a doubling of risk is something of which we take notice, but we really need to know the reference level of risk. A trebling of a very low risk is still a low risk, and, in the case of prehypertension, it is indeed what we are discussing. An analysis of data from the Chicago Heart Association Detection Project in Industry1 may be helpful here. It reported data on 10,874 men who were originally aged 18–39 years and who were followed for 25 years. The mortality from CV disease was the same in subjects with a systolic pressure <120 mm Hg or 120–129 mm Hg (6.1 events per 10,000 person years). For individuals with systolic pressures 130-139 mm Hg (the upper half of the JNC 7 prehypertensive group), the relative risk was increased by one third (1.33), but the difference was not significant, and the absolute risk was still low (8.7 events per 10,000 person-years). Put another way, in 100 patients with a systolic pressure <120 mm Hg, there would be 1.5 CV deaths over 25 years; in individuals with pressures in the "high prehypertensive" range (130–139 mm Hg) the number would be 2.2 deaths. The next question is what we do about it. If we follow the recommendations of the ADA and JNC 7 and identify our patients who have prehypertension and prediabetes, what do we tell them? Clearly, the purposes of both JNC 7 and the ADA recommendations are not to start these individuals on medications but to encourage lifestyle changes that may help to prevent progression of the respective risk factors to levels at which drug treatment might be needed. Fortunately, the needed changes are largely the same for both blood glucose and BP, and we all know what they are: eat less and exercise more. One of the major economical and political concerns at present is the relentless rise of health care costs, a large portion of which are ultimately related to lifestyles and their consequent chronic diseases. If lowering the thresholds of disease could help stem this tide by reducing the numbers of patients with hypertension and diabetes, we would all be happy, but two features stand out of the population with “predisease.” First, very large numbers of people are involved. For both prehypertension and prediabetes, we are talking about tens of millions of people. Second, while the risk of real disease (strokes, heart attacks, etc.) is undoubtedly higher in these individuals when compared with the truly healthy individuals, on an absolute basis, their level of risk is still quite small. If, as physicians, we are successful in lowering a patient's BP from 124 mm Hg to 119 mm Hg, what have we achieved? The answer is in the numbers. If we could do it for the whole population of prehypertensives, the benefits would be huge, but if we can only do it for a small number of our patients, they would be negligible. Physician time is limited and expensive, and a good principle for the best use of it is to achieve the “biggest possible bang for your bucks.” We are doing a conspicuously bad job of controlling BP in our hypertensive patients who are at genuinely high risk, and we all know that persuading patients to take pills is much easier than convincing them to change their lifestyles. Therefore, if we need to start setting BP goals for our prehypertensive patients, we are going to have less time to attend to more urgent matters. Does this mean that we should abandon attempts to counsel our patients about lifestyle changes? Not at all. All patients should be asked and counseled about their lifestyle because the adverse effects of not exercising, being overweight, and eating junk food are so pervasive that they cut across all the processes of chronic disease. We do not need to label our overweight and sedentary patients as being prehypertensive and prediabetic before advising them to exercise more and lose weight. Perhaps the main problem with the definitions of prehypertension and prediabetes is that they give the lower limits of these categories a significance that they do not deserve. Admittedly, 120 and 100 are nice round numbers, and 120/80 has a familiar ring to it, but the inevitable implication of introducing such limits is that if a patient scores just above or below the limit our advice should be different. A patient who has a BP of 122/82 mm Hg and a blood glucose reading of 102 mg/dL should start to worry, whereas one who has values of 118/78 mm Hg and 98 mg/dL can relax. Does this really make sense?