Quantifying Risk of Adverse Clinical Events With One Set of Vital Signs Among Primary Care Patients with Hypertension

I applaud the authors' hard work in demonstrating the natural history of the single initial blood pressure reading. Their results are sobering indeed.

However, in my experience as a faculty member in a heavily inpatient- oriented family medicine residency program, I have found residents only too eager to tinker pharmacologically with single elevated blood pressure readings recorded on an outpatient visit. The deeper problem as I see it is that all too many residents do not hesitate to provide a few weeks' supply of pills from the medication sample closet (albeit with charitable intentions). Some of these are combination drugs, and many are the newest and most costly products. Stepped therapy starting with a diuretic is underutilized. Nor do residents routinely recheck the readings obtained by ancillary personnel.

Patient education, too, often goes by the wayside. I believe that only a relative handful of patients truly understands the concept of blood pressure, and teaching this is a skill. Residents, faculty, professional societies, pharmaceutical companies, and the mass media alike could all be doing a far better job in educating patients about lifestyle changes, as well as in motivating patients to be well. Discussions of the relative advantages and disadvantages of the countless pharmacological regimens for treating hypertension consume a disporportionate amount of the time we spend considering ways to reduce the morbidity from heart disease and stroke in individual patients and in communities as a whole.

Competing interests:   None declared

Brian Crownover asked how we defined "elevated blood pressure and pulse." We did not, because recent evidence suggests that cardiovascular risk is higher with higher blood pressure across its entire physiologic range. The recent recommendations to keep systolic blood pressure less than 130 among patients with diabetes also reflects this notion. So we analyzed systolic blood pressure (and diastolic blood pressure and heart rate) as continuous parameters.

How then did we come up with the excess risk associated with 10mmHg higher blood pressure? We simply used the hazard ratios generated by the proportional hazard models, which gives the risk for each 1mmHg. For example, in Table 3 of our article, the hazard ratio for systolic blood pressure in our model for first myocardial infarction was 1.010. That means every 1mmHg higher systolic blood pressure, regardless of the pressure and holding all other variables unchanged, would increase the risk of having a myocardial infarction by 1%. We simply multiplied that by 10 to get the 10% excess risk associated with a 10mmHg rise in systolic blood pressure.

Competing interests: The study described in this work was funded by a grant from Bristol-Myers Squibb

To the authors: thank you!

As a faculty member who reviews a never ending supply of resident charts, I see progress notes frequently reflect inaction on single elevated blood pressures. Often the physician is unwilling to "bite the bullet" and make a medication adjustment without multiple readings. 5-day blood pressure checks are ordered instead, despite having prior visits in the chart that document other "single" elevated blood pressures.

I am grateful that primary care based research now describes the prognostic value of an isolated set of abnormal vital signs. Justification to appropriately modify cardiovascular risk with single visit data is welcome. Kudos.

Can you clarify one point? How did you define elevated blood pressure and pulse? Were the 10 mmHg and 10 bpm elevations applied to an upper limit of normal of 130 mmHg and 100 bpm?

Sincerely,

Brian Crownover, MAJ, USAF, MC, FAAFP

Family Medicine Residency, Offutt AFB NE

Competing interests:   None declared

I wholeheartedly agree with Dr. Green and admit that before completing this study I, too, tended to overlook individual blood pressure readings. My patients often say "I ran in from the parking lot" or "I'm in pain today" or "I didn't take my medicine this morning." And I believed them and believed that this might transiently increase their otherwise controlled blood pressure.

This study stemmed from a request from a drug company, Bristol-Myers Squibb, to model the effects of blood pressure on vascular outcomes to help them decide how many patients they needed for a prospective study for a new drug. They suggested I use a single blood pressure measurement to assess risk, and I said, "No way that it will be predictive of long-term vascular outcomes. Blood pressure readings vary too much from moment to moment in a given patient." This is what we are taught and I believed.

So I was quite surprised when the first blood pressure record in 1993 (done by regular clinic personnel with regular equipment) was so consistently predictive of so many bad outcomes an average of more than 5 years down the road. And I must admit that it has changed the way that I care for my own primary care patients. I no longer ignore single blood pressure readings but usually adjust patients' medications up, or add another medication, if the blood pressure is elevated, especially among patients with diabetes (which is almost half of my practice). I must admit that if the blood pressure reading is borderline, I will peek at other prior readings or maybe bring the patient back for another reading. But I am now much more aggressive with my treatment in response to single elevated readings. I think other primary care physicians should be, too.

Competing interests:   I received grant funding from Bristol-Myers Squibb to perform the research reported in this study.