Instructor Notes — Chapter 28

Natriuretic Peptides: The Heart Failure Biomarkers That Became Treatments Part 5 · intermediate · prerequisites: Chapters 3, 5, 27

Why this chapter is worth extra class time

Chapter 5 introduces the rating system. Chapters 6 through 27 apply it. Chapter 28 is where the system stops being a convention and starts being an argument.

Every prior chapter has demonstrated "a rating attaches to a claim, not a molecule" using a molecule that works for one indication and not another — semaglutide for obesity versus for Alzheimer's, and so on. Students absorb that pattern easily and often absorb it shallowly, as a rule about indications.

This chapter breaks that reading. Here the same gene product gets different ratings in different roles: measured, administered, preserved. The claims are not "does it work for X versus Y." They are "does knowing this number help," "does adding this molecule help," and "does keeping this molecule around longer help" — three questions requiring three study designs, with answers yes, no, and yes.

If a student leaves this chapter still able to say "so, natriuretic peptides — do they work?", the chapter has not landed. That sentence is the diagnostic.

The single most important thing to teach

§28.4. Everything else is scaffolding.

The specific skill is claim typing: given a sentence, identify the claim, its population, its endpoint, and the study design that could settle it. Students who can do this can evaluate any peptide claim in the book; students who cannot will keep asking whether molecules "work."

A useful drill that takes ten minutes: put four sentences on the board about a single molecule from an earlier chapter and ask students to sort them by required study design before saying anything about whether each is true. Most students will initially try to answer them and have to be stopped. That resistance is the lesson.

Common misconceptions, in order of frequency

1. "Nesiritide was pulled because it killed people." By far the most common, and it is wrong. The pooled re-analyses raised mortality and renal signals; the definitive trial did not confirm them. What it showed was absence of benefit. Quiz question 16 is designed to surface this misconception, and it usually does. Correct it explicitly, because the distinction between "we tested it and it didn't help" and "we tested it and it hurt people" is the same distinction students need for Part III compounds, where usually neither has been established.

2. Treating a biomarker's prognostic strength as a reason to target it. "NT-proBNP predicts death so strongly that lowering it must help." This is the surrogate fallacy from Chapter 5 §5.6 in a new costume, and the GUIDE-IT result is the cleanest refutation in the book. Students find it genuinely counterintuitive that a number can be an excellent predictor and a useless target.

3. Assuming the ARNI combination is a commercial convenience. Students who have been taught to be skeptical of combination products — correctly, in general — over-apply that skepticism here. The substrate list settles it: neprilysin degrades angiotensin II, and inhibiting neprilysin alone would raise it. Have them derive the necessity from the substrate table rather than telling them.

4. Believing BNP and NT-proBNP are interchangeable. Reference ranges differ roughly ten-fold. This is the most practically consequential fact in §28.3 and the easiest to skim past.

5. Expecting mechanism to predict trial results. The whole chapter is an argument against this, and yet in discussion students routinely reconstruct the mechanistic case for nesiritide and find it so persuasive that they look for a flaw in the trial. That impulse is worth naming out loud when it appears — it is exactly what rule 3 is protecting against, and watching yourself do it is more instructive than being told about it.

6. Extending the PARADIGM-HF result to HFpEF. Students hear "proven in heart failure." The later trial in preserved ejection fraction did not meet its primary endpoint. Rule 1 requires the population.

Suggested session plan (90 minutes)

• 0–10 Open with de Bold's experiment and the ventricular control. Ask why the control mattered. Establishes both the biology and the habit of looking at design first.
• 10–25 ANP/BNP/CNP, receptors as cyclases, cGMP, and the resistance paradox. Do not skip the paradox; §28.9 needs it.
• 25–40 The biomarker: proBNP processing, why the inactive fragment is the better test, rule-out logic, the four confounders and their directions. Ask which confounder is most dangerous and why — this reliably produces good argument.
• 40–65 §28.4, the core. Put the three-claims diagram up and work each row. Then introduce the fourth claim (titration) and let students predict the answer before revealing it. Most predict yes.
• 65–80 Nesiritide and PARADIGM-HF as a paired case. Do not present them separately.
• 80–90 The §28.9 lesson. Close on "targeting a system is not a strategy," and on the fact that the successful drug is not a peptide.

If you have only 50 minutes: biomarker (10), §28.4 (25), the paired trials (15). Cut the receptor biology and the frontier section.

Assessment guidance

• The quiz is weighted toward recall in items 1–19 and toward synthesis in 20–22. Items 20–22 are where differentiation happens; consider weighting them double.
• In the exercises, the † items are the ones worth grading. AE, AG, and AH are the strongest single indicators of whether the chapter landed.
• Exercise V (write a rating block for an unrated claim) is the best short assessment in the chapter. Do not mark on the verdict reached; mark on whether the claim was correctly typed and whether the "what would change it" line is genuinely falsifiable.
• Case Study 28.1 Q4 and Case Study 28.2 Q6 have substantially constrained correct answers; the rest are open. See _scratch/answers/ch28.md.

Two things not to do

Do not let the chapter become a cardiology lecture. The clinical detail is a vehicle. Students in the cosmetic and performance tracks need §28.4 and will not need ejection fraction categories again.

Do not editorialize on achondroplasia. The chapter states, accurately, that whether the condition should be treated is genuinely contested by serious people on both sides, and that a positive efficacy rating is not an endorsement. Some students will want you to resolve it. Decline. The teachable point is that an evidence rating answers an empirical question and leaves the normative one entirely untouched — which is a general property of the whole rating system and worth stating as such.