Medication articles
When the antidepressant that worked stops working
It is one of the most disheartening moments in treatment: the medication that gave you your life back seems to run out of effect. The symptoms creep back, first at the edges, then into the middle of ordinary days, and the conclusion writes itself: the drug failed, or you did. Neither conclusion is earned yet. Before anyone switches, augments, or gives up on a medication that once worked well, there is a checklist to run. This article is that checklist, in the order I run it.
PsychMed Care is the medication management practice of Solutions Psychiatry. If you are looking for therapy, or therapy and medication together, the full practice is at psych.us.
First, name what is happening
Three different stories wear the same disguise
Clinicians have a dry nickname for the classic version of this problem: the "poop out," or tachyphylaxis, where a drug that held for months or years gradually loses its grip for no obvious reason. It is real, and it is common enough that nobody in this practice treats it as a surprise. But the fade has at least three different causes, and they do not share a treatment.
The first story is the true poop out: same dose, same habits, same life, and the effect still thins. The second is a partial response wearing a good disguise: the medication helped, never fully, and what looks like a relapse is the untreated remainder finally getting louder. The third story is the most common of all: the medication is doing exactly what it always did, and the load changed underneath it. A grief, a brutal work season, a new illness, a drinking habit that crept up. The drug did not get weaker; the weight got heavier. Sorting out which story you are in is the first clinical job, because story three does not get fixed by a higher dose, and story two was never going to be fixed by waiting.
This workup runs inside ordinary care here: the labs are drawn at a LabCorp near you, the visit time is the standard follow-up length, and the costs are the published ones. How plans are verified before a first visit, and what visits cost, lives at the practice hub, under insurance and coverage. Nothing on this page requires a special program. It requires a prescriber willing to ask why before reaching for the prescription pad.
One safety note before the checklist. If the returning symptoms include thoughts of harming yourself, stop reading and get help now: call or text 988, call 911, text HOME to 741741 for Crisis Text Line, or call NAMI at 1-800-950-6264. This practice is not an emergency service, and a fading antidepressant is never a reason to wait out a crisis alone. The full list lives on the crisis resources page.
The checklist
Eight questions before anyone touches the prescription
Was the dose ever optimized?
Many antidepressants are started low, work partly, and are never raised to the dose where they do their full job. The dose that carried the first hard month may simply not be the dose that holds a life together. This is the cheapest question on the list and the one skipped most often.
Is the dose actually arriving?
Missed days, refill gaps, a pill taken at midnight instead of morning: none of this is a character flaw, and all of it thins the blood level the brain actually sees. An honest adherence conversation, without blame, comes before any conclusion about the molecule.
What is the thyroid doing?
An underactive thyroid imitates depression almost perfectly: the flatness, the fog, the weight, the cold. Treating that picture with a second antidepressant, while the thyroid drifts unmeasured, is how people collect prescriptions instead of answers.
B12 and folate
Low B12 and low folate drain energy, memory, and mood, and both are common in people eating ordinary modern diets or taking ordinary acid-blocking medication. They are also cheap to check and straightforward to correct, which makes skipping them hard to defend.
Vitamin D
Low vitamin D tracks with low mood closely enough that it belongs on the panel whenever a working drug fades. It is not a magic lever, and this practice does not sell it as one. It is a correctable contributor, checked so it stops being a mystery variable.
Iron
Iron deficiency, with or without anemia, produces exhaustion and cognitive drag that antidepressants cannot out-prescribe. It hides in plain sight in menstruating patients, in donors, in anyone whose diet narrowed during the depression itself.
How is sleep, really?
Sleep loss dismantles what any antidepressant builds. Undiagnosed sleep apnea is the classic saboteur: the patient snores, wakes unrefreshed, and concludes the medication died. Sometimes the medication is fine and the airway is not. Insomnia that returned for its own reasons gets the same honest look.
Alcohol, and everything else on board
Alcohol works against an antidepressant every single night it is used, and cannabis and sedating over-the-counter habits can blunt one quietly. So can a new prescription from another clinician that speeds up or slows down how fast your body clears the drug. The full list, supplements included, is part of this question.
Five of those eight questions are answered by one blood draw. The panel used here, thyroid, B12, folate, vitamin D, iron, and metabolic and inflammatory markers around them, is the same workup described on the functional psychiatry page, drawn through LabCorp and read against your whole picture rather than against a reference range alone. A result inside the normal band can still matter when it sits at the bottom of the band in a person whose symptoms say otherwise; that judgment call is what a prescriber is for.
One more discipline belongs in this section, because it decides whether the checklist gets run at all: timing. A single bad week is weather, and depression moves in weather. What earns the full workup is a trend: several weeks of scores or symptoms drifting the wrong way on the same measures that documented the recovery, or a drop sharp enough that nobody needs a graph to see it. Patients who report a fade early, while it is still a drift, hand their prescriber a solvable problem. Patients who wait, hoping to be wrong, often arrive after the drift has become a relapse, and then the same checklist has to run under much heavier conditions. If you take one habit from this article, take that one: say it at the first follow-up where you notice it, not the fourth.
When the checklist comes back clean
Optimize, switch, or augment: the order matters
Optimize first
If the dose was never truly tested, test it: a real dose, held for a real duration, with measures taken before and after. It is the fastest honest answer available, and it rescues more "failed" medications than any other move on this page.
Switch with intent
When a full trial has genuinely failed, change the mechanism, not just the brand. The next drug is chosen against your history, your side-effect ledger, and what the first drug taught us about your system, not against a formulary's preference.
Augment a partial win
A medication that delivered half a recovery has earned an ally rather than a dismissal. Adding a targeted second agent, or a supplement with real evidence behind it, builds on measured ground instead of restarting the search from zero.
Underneath all three moves sits one discipline: measure. The same rating scales used at intake get repeated at each decision point, so "better" and "worse" are numbers on a page rather than two impressions arguing in a tired person's memory. A fade you can graph is a fade you can reason about. A fade you cannot graph turns every follow-up into a coin flip, and coin flips are how patients end up on four medications with no idea which one does what.
When guessing has cost enough
Where GeneSight fits in the sequence
Pharmacogenomic testing is not the first move, and this practice does not order it as a ritual. It earns its place after a failure or two, when the question has sharpened from "which drug" to "why does this body keep mishandling reasonable drugs." A cheek swab shows how your genes metabolize the common psychiatric medications: which ones your system clears too fast to hold, which ones build up and cause side effects at ordinary doses, and which are likely to behave as the label expects.
That map does not choose your next medication; your history does that. What it does is clear the underbrush. If the drug that "stopped working" turns out to be one you clear twice as fast as the average patient, the fade stops being a mystery and becomes arithmetic. The cost and coverage picture, including why most patients pay $330 or less and why it is typically $0 on Medicare Part B, is laid out on the GeneSight testing page.
It is equally worth saying what the test does not do. It does not measure your thyroid, your iron, or your sleep, which is why it comes after the blood draw in this sequence and not instead of it. It does not grade how hard you tried, and a "normal" result does not mean the fade was imaginary. And it does not retire clinical judgment: a gene panel can rank the likely behavior of a list of drugs, but it cannot know that one of them left you wired at 3 a.m. in 2019. That knowledge lives in your history, and the history outranks the printout every time they disagree.
Two warnings belong in any honest version of this article. First: do not stop the medication abruptly while you wait for answers. Stopping suddenly can produce discontinuation symptoms that imitate the original illness and scramble every conclusion; if coming off is ever the right move, it is done as a planned taper, the subject of the companion guide on tapering off antidepressants. Second: do not quietly double the dose on your own, and do not add a shelf of supplements without telling your prescriber. Both feel like initiative. Both can interact, and both destroy the clean read the checklist is designed to give.
After the win comes back
Keeping a recovered medication recovered
When the workup finds its answer, a low thyroid treated, a dose finally optimized, a drug switched to one your metabolism handles honestly, the relief can tempt everyone to declare the story over. It is not quite over. The months after a fade are when the maintenance habits get built, because a brain that has slipped once has shown you its preferred direction of travel.
The habits are unglamorous. Refills requested before the bottle is empty, so a pharmacy delay never becomes an accidental discontinuation. The rating scales repeated on schedule even while things are good, so the baseline is documented by the good months and not reconstructed from memory in a bad one. Sleep defended like a prescription, because it behaves like one. And a standing agreement with your prescriber about what a drift looks like for you specifically, which symptoms showed up first last time, so the early warning sign has a name before it has a season. None of this is heroic. All of it is cheaper, in every currency, than running the checklist again from the bottom of a relapse.
Depression care here
The service behind this article: evaluation that asks why the last medication failed before choosing the next one.
The test, priced plainly
What GeneSight costs, what Medicare and commercial plans do with it, and what the report actually says.
The lab half of the question
Thyroid, B12, folate, vitamin D, iron: the physical workup that runs beside every medication decision here.
“A fading medication is a question, not a verdict.”
Bring the question to a full medication review