1. SSRIs/SNRI remain the pharmacological backbone
The current VA/DoD evidence assessment gives its strongest medication support to only three drugs:
sertraline, paroxetine, and venlafaxine. Sertraline and paroxetine are the two drugs FDA-approved for PTSD; venlafaxine has similarly good efficacy evidence.
Typical evidence-based PTSD ranges are:
- Sertraline 50–200 mg/day
- Paroxetine 20–60 mg/day
- Venlafaxine XR roughly 75–300 mg/day in the VA/DoD dosing table.
For C-PTSD, I would expect these drugs principally to reduce intrusive symptoms, hyperarousal, generalized anxiety and depressive affect. I would be much less optimistic that increasing the dose will somehow repair the characteristic C-PTSD phenomena of chronic shame, identity disturbance, attachment pathology, dissociation, or inability to trust another person.
That distinction strikes me as clinically fundamental:
Medication may turn down the volume of the alarm system; it does not reconstruct the damaged world of self-and-other.
And C-PTSD is disproportionately about the latter.
2. Prazosin: useful if the target is nightmares—not “C-PTSD”
Prazosin occupies an interestingly precise position. The 2023 VA/DoD guideline suggests it for PTSD-associated nightmares, while suggesting against using it for global PTSD symptoms.
The VA dosing appendix starts at 1 mg at bedtime, with slow titration, emphasizing orthostatic hypotension and first-dose syncope; its listed range is approximately 3–20 mg at bedtime.
So I would think of it as:
nightmare → prazosin, rather than
C-PTSD → prazosin.
This phenotype-oriented way of prescribing probably makes more sense throughout C-PTSD.
3. What about mirtazapine, trazodone, quetiapine, etc.?
Here everyday psychiatric practice and evidence-based PTSD guidelines separate somewhat.
Mirtazapine or trazodone may obviously be useful when depression + marked insomnia coexist, and quetiapine sometimes makes a dramatic symptomatic difference to insomnia, agitation or affective instability. But that doesn’t mean they are established C-PTSD therapies.
VA/DoD currently regards evidence as insufficient to recommend for or against mirtazapine, escitalopram, fluoxetine, duloxetine, pregabalin, quetiapine, olanzapine, lamotrigine, topiramate, and many others as PTSD monotherapy. Risperidone is specifically suggested against for PTSD; quetiapine and olanzapine have weak evidence with significant metabolic cost.
This suggests a useful distinction between:
treating the PTSD/C-PTSD, versus
treating something troublesome occurring in a person with C-PTSD.
Quetiapine for severe insomnia may accomplish the second without having demonstrated the first.
4. Mood stabilizers are less convincing than one might intuitively expect
This is particularly interesting in C-PTSD because affect dysregulation can look as though it ought to respond to anticonvulsant mood stabilizers.
But the evidence doesn’t really support that inference. Divalproex is suggested against for PTSD, and evidence is insufficient for lamotrigine and several others.
Obviously, if there is genuine comorbid bipolar disorder, that becomes a different matter. Indeed, VA/DoD explicitly cautions that antidepressants can destabilize bipolar illness and that this may necessitate mood stabilization.
This is another place where C-PTSD easily generates diagnostic confusion:
affect dysregulation ≠ bipolarity.
5. Dissociation is the pharmacological desert
For me, this is one of the biggest limitations of the whole pharmacological approach.
C-PTSD often includes derealization, depersonalization, emotional numbing, discontinuity of experience, body alienation, and sometimes striking fragmentation. There is no reliably established medication for dissociation itself.
Attempts involving naltrexone, lamotrigine and various serotonergic/glutamatergic strategies have produced scattered signals, but nothing remotely analogous to the SSRI evidence for depression.
And this perhaps tells us something about the level at which the pathology exists: dissociation is not simply “too much anxiety.” It is often an organization—or disorganization—of experience.
6. Benzodiazepines: particularly unattractive in complex trauma
VA/DoD strongly recommends against benzodiazepines for PTSD. There is no convincing evidence of sustained PTSD benefit, while concerns include dependence, cognitive impairment, disinhibition and potentially poorer engagement with trauma-focused psychological processes.
For C-PTSD, I think the objection is even more interesting than merely “addiction risk.”
Patients with chronic developmental trauma can understandably discover that benzodiazepines provide an extremely efficient way of achieving:
I don’t have to experience this.
That is precisely why they can become phenomenologically attractive—and therapeutically problematic.
Short-term, highly circumscribed use for another indication is of course a different clinical question, but I would avoid making benzodiazepines the organizing principle of C-PTSD treatment.
7. What about ketamine, MDMA and psychedelics?
This field is active, but we should distinguish excitement from established treatment.
The 2023 VA/DoD guideline suggests against ketamine as PTSD monotherapy because the evidence considered was inadequate, and does not currently recommend MDMA augmentation. A 2026 review of investigational PTSD drugs emphasizes how little pharmacological progress has occurred despite extensive work; only sertraline and paroxetine remain FDA-approved.
There is now interesting thinking about therapy-aligned entactogens specifically for complex PTSD, precisely because a drug might transiently modify fear, trust, social threat and plasticity sufficiently to alter what becomes possible inside psychotherapy. But this remains an emerging research program rather than routine C-PTSD pharmacotherapy.
And that distinction is conceptually fascinating: perhaps the future drug will not treat trauma, but temporarily alter the conditions under which another relationship to trauma can become possible.
8. My practical formulation
If I were reducing the pharmacotherapy of C-PTSD to a clinical algorithm, it would be something like this:
First, identify the pharmacologically tractable layer.
Intrusion/hyperarousal + anxiety/depression → sertraline / paroxetine / venlafaxine.
Dominant nightmares → consider prazosin.
Major depressive episode, bipolar illness, ADHD, substance-use disorder, psychosis, severe insomnia, etc. → treat the genuine comorbidity on its own terms, rather than pretending every symptom is C-PTSD.
Then stop asking medication to do things medication has very little evidence of doing.
In particular:
chronic shame → not an SSRI deficiency
fragmented identity → not a serotonin deficiency
inability to trust → not a norepinephrine disorder
relational reenactment → not something quetiapine repairs
dissociation → not simply insufficient antidepressant dosage
The 2026 C-PTSD psychotherapy literature makes essentially the complementary observation: established treatments can reduce conventional PTSD symptoms, while the DSO component is often harder to change.
So I would describe pharmacotherapy in C-PTSD as supporting treatment from underneath, rather than constituting the treatment itself.
And perhaps the most interesting implication is that C-PTSD exposes the limit of biological psychiatry unusually clearly: the further one moves from fear/intrusion/hyperarousal toward shame, selfhood, dissociation and relationship, the more rapidly the pharmacological signal fades. That does not make medication unimportant; it tells us which ontological level medication is acting upon.
