Making and Breaking Rules

A couple of weeks ago I posted on neonatal hypertension. In the course of that piece I commented that I had made a rule about when I would recommend treatment for this condition. A commenter, Dr. Isis, asked the following:

I am interested in the "I decided to make a rule..." part of it. Can you tell us more about how you made this decision?

I posted an answer about my review of the literature and other considerations, but that really does not address the root of the issue; why did I feel the need to make a rule?

So I reviewed a relatively recent book, How Doctors Think. This read from 2007 examines the ways physicians deal with uncertainty in medicine. We learn to recognize patterns, and respond to them accordingly. Neonatal hypertension bothered me because of the uncertainty, the lack of data to drive decisions. We usually cannot identify a cause, we cannot determine which tiny minority of patients will not do well, and yet we hope to do no harm using drugs not approved for use in infancy.

In my case, I wanted a rational approach to the common neonatal hypertension patient who had no evidence of kidney or vascular problems to explain the hypertension. I picked a level of blood pressure that seemed significant, and I decided to treat at that level.

Physicians deal with uncertainty on a daily basis. We rarely know everything about a patient (not until the autopsy, anyway), and our patients rarely fit the textbook completely. Recognizing patterns and responding to them makes practice possible, although numerous examples in Jerome Groopman’s book illustrate the dangers of these thought processes.

Lots of doctors proclaim that they do not practice “cookbook medicine.” In other words, they treat each patient individually rather than relying on rules, flowcharts, and other “cookbooks.” Of course, they all do rely on rules to some extent- the important thing is to realize when the pattern and response to treatment vary in some important way that will influence the patient’s outcome.

As I noted at the end of my earlier post, my rule seems to be serving me well. I spent far more time considering why I formulated it than I took to create it.

Hyper-what? In the NICU!

I get consulted a lot for neonatal hypertension.

In actuality, most cases I see are not really neonates but premature infants who are nearing their due dates. They often have chronic lung problems or other major health issues and a history of umbilical vessel catheterizations.

Because of issues with measuring and confirming blood pressures in this age group, our practice is to rely on the systolic measurement (SBp). A variety of studies address normal blood pressure in neonates, with most showing a value of approximately 110 mmHg as the 95th percentile after the first 2 weeks of life. Elevated blood pressures in these infants rarely cause detectable end-organ damage such as cardiac enlargement. Virtually all infants with blood pressures above the 95th percentile have complete resolution of this condition by 2 years of age. Risks and benefits of treatment or non-treatment remain unclear.

Values repeated greater than the 95th percentile for age warrant a work-up that includes estimated glomerular filtration rate via serum creatinine measurement; a urinalysis and culture; and a renal ultrasound with doppler bloodflow studies. Echocardiogram may show a cause for the SBp elevation, or may indicate end-organ damage. We rarely find a secondary cause for the blood pressure elevation. 

The question remains whether to treat or not to treat. No studies provide concrete evidence of benefit of treatment. Major risks of medications include dropping the blood pressure too low. Occasional reports of death, heart failure, and encephalopathy associated with high blood pressure warrant careful consideration of treatment benefits, though.

About 10 years ago, I decided to make a “rule” about how I would handle these babies. SBp of 110 seemed the critical value. Since most of these infants were still in neonatal units, we generally had access to multiple readings over several days. If the SBp were greater than 110 on more than half of the readings, treatment with an angiotensin-coverting enzyme inhibitor was recommended. Otherwise, ongoing observation of SBp occurred.

My last call week included several infants with SBp elevations, so I pulled more recent articles on this topic. While no definitive randomized trials or other evidence-based studies have been published, the overall recommendations remain consistent with those I created a decade ago, as far as I can tell.

I feel better now. But I wish we knew more.

References:

Arch Dis Child Fetal Neonatal Ed 2002;86:F78-F81 doi:10.1136/fn.86.2.F78

Pediatr Nephrol. 2009 Jan;24(1):141-6. Epub 2008 Jul 9 doi:10.1007/s00467-008-0916-9

del.icio.us Tags: Clinical,medicine