20.2 Treatment Adjustment and Deprescribing

Starting a drug is Implementation. Changing it because the outcome is wrong is Evaluation. Domain V asks whether you intensify, simplify, hold, taper, or stop and rethink the diagnosis. Candidates lose these items by doing more of the same: a higher dose of a drug that never matched the disease, a third agent when the first two were never taken, or a refill of a proton-pump inhibitor that healed nothing 7 years ago.

Adjustment has a sequence. Confirm adherence, technique, and the diagnosis. Measure the outcome that matters (control, harm, a patient-reported score). Then step up, step down, switch class, or deprescribe. “The guideline says max the dose” is not a reason to ignore edema, hypoglycemia, or a peak flow that never moved.

Asthma: GINA step-up and step-down

Global Initiative for Asthma (GINA) tracks are built to be moved. Current preferred adult and adolescent Track 1 uses inhaled corticosteroid–formoterol as both reliever and, at higher steps, maintenance-and-reliever therapy (MART). Steps rise when the patient has daytime symptoms, night waking, activity limitation, reliever use more than twice a week, or exacerbations. Before you step up, watch the patient use the inhaler, ask how many canisters they actually bought, and ask whether this is asthma. A preschooler with a sudden unilateral wheeze may have a foreign body. An older adult with “new asthma” may have heart failure. A teenager with throat tightness and a normal exam between attacks may have inducible laryngeal obstruction.

Step down after about 3 months of good control. Leaving someone on high-dose inhaled corticosteroid plus a long-acting beta agonist plus a daily oral corticosteroid “because they were sick last winter” is how you create thrush, dysphonia, osteopenia, and adrenal suppression. Step down one increment, keep a written action plan, and schedule the review. Rescue-only short-acting beta agonist as the sole adult plan is outdated on current GINA language — that is a step-up problem, not a character problem.

Hypertension: add a class, do not worship one molecule

If blood pressure is not at goal on a moderate dose of one well-chosen agent, add a complementary second class — typically an ACE inhibitor or ARB, a dihydropyridine calcium-channel blocker, and a thiazide-like diuretic in combination — rather than pushing amlodipine until the ankles are balloons or hydrochlorothiazide to 50 mg “to keep the pill count at one.” Outcome trials and resistant-hypertension pathways (including PATHWAY-2 thinking: add spironolactone when potassium and kidney function allow) reward combinations. Maxing one drug forever is how you get adverse effects without getting control.

Resistant hypertension is blood pressure above goal on three agents at appropriate doses including a diuretic. Before you add a fourth, prove the patient is taking the three, exclude white-coat effect with home or ambulatory readings, use a correct cuff, and think secondary causes: primary aldosteronism, renal-artery stenosis, obstructive sleep apnea, pheochromocytoma in the right story, exogenous steroids or sympathomimetics. That workup is Evaluation. Another 10 mg of the same calcium-channel blocker is not.

Diabetes: organ protection is an adjustment, not a luxury

Type 2 diabetes adjustment is no longer “add whatever lowers A1C.” If atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease is present, add an SGLT2 inhibitor and/or a GLP-1 receptor agonist with outcome evidence — often independent of the current A1C. Metformin stays if estimated glomerular filtration rate allows. What you do not do is max a sulfonylurea into hypoglycemia to “save” a brand-name agent, or add a fourth glycemic drug while the patient with reduced-ejection-fraction heart failure is still not on an SGLT2 inhibitor. Chapter 12 taught the classes. This section scores the moment you change the list because the organ, not only the A1C, is the outcome.

If A1C is already at an individualized goal (tighter in a healthy 48-year-old, looser in a frail 88-year-old), do not intensify just because a single-disease flowchart is still green. Hypoglycemia, weight loss you did not intend, and treatment burden are outcomes that justify step-down.

Deprescribing is a plan, not a failure

Deprescribing is a supervised dose reduction or stop of a medicine whose harm now outweighs benefit. It has an indication (“no ulcer, no Barrett esophagus, no dual antiplatelet therapy”), a taper when the drug causes withdrawal, a monitoring plan, and a follow-up date. It is not “the patient is old, stop everything.”

American Geriatrics Society Beers Criteria list drugs that are potentially inappropriate in older adults. They are a risk list to individualize, not an automatic ban and not an ANCC statute. High-yield primary-care cuts: first-generation antihistamines, benzodiazepines and nonbenzodiazepine hypnotics, sliding-scale insulin as the sole insulin plan, glyburide, chronic systemic nonsteroidal anti-inflammatory drugs, strongly anticholinergic bladder and antidepressant agents, and proton-pump inhibitors beyond 8 weeks without a durable indication.

Proton-pump inhibitors. An 8-week course is enough for most uncomplicated gastroesophageal reflux. Then stop, step down to an H2-receptor antagonist, or use on-demand therapy unless the patient has Barrett esophagus, severe erosive disease, or high gastrointestinal-bleed risk (for example dual antiplatelet therapy or a steroid-plus-NSAID combination you cannot stop). “They have been on omeprazole since 2018” is not an indication. Rebound acid can last days to weeks — warn the patient so they do not restart at 40 mg on day three and call the taper a failure.

Benzodiazepines. Taper. Do not stop a long-term alprazolam or clonazepam on Friday and wish the patient luck. Seizure, rebound anxiety, and insomnia are expected if you crash the dose. Substitute cognitive behavioral therapy for insomnia when the original indication was sleep. Do not add a second sedative “for the taper.” Document the schedule, the next contact, and what to do if withdrawal symptoms appear.

Anticholinergics. Diphenhydramine for sleep, oxybutynin for urgency, amitriptyline for pain, paroxetine, and first-generation antihistamines stack confusion, constipation, urinary retention, blurred vision, and falls in older adults. Two moderate anticholinergics can be worse than one strong one. The adjustment is subtraction.

Duplicate therapy. Two ACE inhibitors, an ACE inhibitor plus an ARB for garden-variety hypertension or albuminuria (combination is generally avoided), two NSAIDs, two proton-pump inhibitors, two benzodiazepines, or an inhaled corticosteroid from two devices the patient thinks are different drugs. Medication reconciliation is how you find the pair. Deprescribing is how you end it.

Sick-day and sick-week holds

When a patient with diabetes or chronic kidney disease is vomiting, has profuse diarrhea, or cannot keep fluids down, holding the drugs that precipitate acute kidney injury, hypoglycemia, lactic acidosis, or euglycemic diabetic ketoacidosis is the intervention. A common teaching cluster is SADMANS: sulfonylureas, ACE inhibitors, diuretics (and direct renin inhibitors), metformin, ARBs, NSAIDs, and SGLT2 inhibitors. Teach the patient before the gastroenteritis. Restart when eating and drinking recover. That is a sick-day hold.

A sick-week hold is longer: bowel preparation, a planned fast, perioperative SGLT2 inhibitor interruption (often about 3–4 days before surgery to reduce ketoacidosis risk), or a week of poor intake in a frail adult. Write the restart date. Insulin is the exception that candidates miss: do not stop basal insulin in type 1 diabetes because the patient is sick — that is how diabetic ketoacidosis starts. Reduce prandial insulin if the patient is not eating; keep a basal backbone.

NSAIDs during a dehydrating illness are a primary-care own-goal. So is “just take your lisinopril and furosemide with sips of water” in a patient who has not kept down a meal in 36 hours.

When failure means the diagnosis is wrong

Stacking drugs on a wrong problem is not persistence. It is harm.

• “Resistant hypertension” that is nonadherence, a 12-cm cuff on a 40-cm arm, white-coat effect, or aldosteronism.
• “Asthma” that is cardiac wheeze, vocal-cord dysfunction, or a peanut in a toddler’s bronchus.
• “Depression” that is hypothyroidism, untreated apnea, bipolar disorder, grief, or substance use.
• “Recurrent cystitis” that is chlamydia, incomplete emptying, or interstitial cystitis.
• “Reflux” that is eosinophilic esophagitis or cardiac ischemia.

The Evaluation move is to stop escalating and reopen Assessment and Diagnosis. Chapter 5 taught differentials and traps. This section is the moment the failed treatment is your clue.

Walk three charts. A 46-year-old using albuterol five nights a week with a terrible inhaler technique: teach, then step up to an inhaled-corticosteroid-containing plan — do not start chronic oral prednisone as the first adjustment. A 74-year-old on omeprazole 40 mg for 8 years, no ulcer, no Barrett esophagus, no antiplatelet therapy: deprescribe with a taper and a reflux plan. A 58-year-old with A1C 7.2% and heart failure with reduced ejection fraction who is still not on an SGLT2 inhibitor: the missing adjustment is organ-protective therapy, not a higher glipizide dose.