Trazodone Cancer Risk Signal Review: What the Evidence Actually Shows

What Is the Trazodone Cancer Risk Signal?

The cancer risk signal associated with trazodone originates almost entirely from spontaneous adverse-event reporting systems and retrospective pharmacoepidemiologic studies, not from randomized controlled trials designed to detect oncologic outcomes. Signal detection in pharmacovigilance assigns a reporting odds ratio (ROR) or proportional reporting ratio (PRR) to a drug-event pair; an ROR above 2.0 with a lower 95% confidence interval above 1.0 typically triggers review. Trazodone has crossed that threshold for certain neoplasm categories in the FDA Adverse Event Reporting System (FAERS) database, but FAERS data are hypothesis-generating, not confirmatory.

No prospective randomized trial has assigned trazodone cancer incidence as a primary or secondary endpoint. The signal, as it currently stands, requires independent replication in well-controlled cohort studies before any clinical guidance changes.

How Pharmacovigilance Signals Are Generated

The FDA uses disproportionality analysis on FAERS submissions to flag unexpected drug-event combinations [1]. A signal means a drug-event pair is reported more often than expected relative to all other drugs in the database. It does not mean causation, it does not control for confounders such as obesity, smoking, or comorbid psychiatric illness, and it does not establish incidence rates in the general population.

For trazodone, the specific neoplasm subcategories flagged have varied across analyses and have not consistently pointed to a single organ system, which further weakens biological plausibility.

What the EMA and FDA Have Said

As of January 2025, neither the FDA nor the European Medicines Agency has issued a safety communication, label change, or Risk Evaluation and Mitigation Strategy (REMS) related to trazodone and cancer [2]. The current FDA prescribing label for trazodone hydrochloride lists priapism, cardiac arrhythmia, and serotonin syndrome as the most clinically significant risks, with no oncology-related black-box warning [2].

Trazodone's Pharmacology and Why a Cancer Mechanism Has Been Proposed

Trazodone is a serotonin antagonist and reuptake inhibitor. It blocks the serotonin transporter (SERT), antagonizes 5-HT2A and 5-HT2C receptors, and has significant alpha-1 adrenergic and histamine H1 antagonism, which accounts for its sedating properties [3]. Its primary active metabolite, meta-chlorophenylpiperazine (mCPP), is a partial 5-HT1B/1D agonist and can itself produce anxiogenic effects at higher plasma concentrations.

Serotonin Signaling and Tumor Biology

Serotonin is not exclusively a neurotransmitter. Peripheral serotonin receptors are expressed on multiple epithelial cell lines, and preclinical data suggest that 5-HT2A receptor agonism or antagonism may modulate cell proliferation pathways in vitro [4]. However, in-vitro receptor modulation rarely translates cleanly to human carcinogenesis at therapeutic drug exposures. The plasma 5-HT2A receptor occupancy achieved with antidepressant doses of trazodone (roughly 150 to 400 mg per day) does not necessarily reflect tumor-tissue receptor occupancy, and no pharmacokinetic-pharmacodynamic model has established a meaningful link.

The mCPP Metabolite Question

MCPP has drawn separate scrutiny because it activates 5-HT1B receptors expressed in some gastrointestinal and hepatic cell lines [5]. Preclinical rodent studies at supratherapeutic doses have shown mitogenic effects in hepatocyte cultures, but the doses required (typically 10x to 50x the human equivalent dose) make direct clinical extrapolation unreliable. No human hepatocellular carcinoma registry study has identified trazodone or mCPP as an independent risk factor after controlling for alcohol use, hepatitis B/C status, and metabolic syndrome.

CYP3A4 Inhibitors as a Confounding Variable

Trazodone is metabolized almost entirely through CYP3A4 [3]. Co-prescription of CYP3A4 inhibitors such as fluconazole, ritonavir, or clarithromycin substantially raises trazodone and mCPP plasma levels. Any pharmacoepidemiologic study that does not adjust for CYP3A4 co-medication will systematically misattribute metabolite-related effects to trazodone itself. This confounding has not been fully addressed in the database studies that generated the original signal.

The Mendelson 2005 Trial and Off-Label Sleep Use

Trazodone's off-label use for insomnia is extremely common. Mendelson (J Clin Psychiatry 2005) remains one of the most-cited controlled investigations of trazodone for primary insomnia in non-depressed adults [6]. The trial enrolled adults with DSM-IV primary insomnia and used polysomnography to assess sleep architecture.

Key Findings From Mendelson 2005

Trazodone 50 mg produced statistically significant improvements in total sleep time and wake after sleep onset compared with placebo at week 2, but these benefits were not sustained at week 6, suggesting tolerance development with continuous use [6]. The study did not track cancer-related outcomes, was not designed or powered to do so, and had a duration of only 6 weeks.

Implications for Long-Term Sleep Prescribing

The common clinical pattern of using trazodone 25 mg to 100 mg nightly for months or years has never been evaluated in a trial longer than a few months. Long-duration use means longer cumulative drug exposure, which is precisely the exposure window relevant to any carcinogenesis hypothesis. The absence of long-term trial data is itself a data gap, not evidence of safety or harm.

The table below summarizes a practical framework for evaluating trazodone duration-of-use decisions in the context of the current signal uncertainty:

| Duration of Use | Cancer Signal Relevance | Clinical Recommendation | |---|---|---| | Acute (under 4 weeks) | Negligible; exposure window too short | Standard prescribing; monitor for priapism and orthostasis | | Short-term (4 to 12 weeks) | Very low; no cohort data showing harm | Reassess need at 8 weeks; document indication | | Chronic (over 6 months) | Theoretically relevant but unconfirmed | Annual medication review; document risk-benefit discussion | | Very long-term (over 2 years) | Unknown; no prospective data exist | Consider tapering trial if insomnia indication; reassess annually |

Pharmacoepidemiologic Studies: What the Data Show

FAERS Disproportionality Analyses

Multiple independent analyses of FAERS have calculated RORs for trazodone and neoplasm-related preferred terms. One published disproportionality analysis found an ROR of 2.14 (95% CI 1.43 to 3.19) for a broad "neoplasms benign, malignant, and unspecified" category, but this analysis did not adjust for age, sex, smoking, or alcohol use [1]. Age alone is the single strongest predictor of cancer incidence in any population; psychiatric medication users skew older than the general population, which inflates unadjusted RORs for any cancer outcome.

Cohort and Case-Control Studies

A 2019 population-based cohort study using the UK Clinical Practice Research Datalink (CPRD), examining antidepressant exposure and colorectal cancer risk, found no statistically significant association for trazodone specifically (adjusted hazard ratio 1.07, 95% CI 0.88 to 1.31) after controlling for BMI, smoking status, and prior colonoscopy [7]. The study was not powered to detect site-specific rare cancers, and trazodone users represented a minority of the antidepressant cohort.

A separate Taiwanese National Health Insurance database analysis of over 200,000 insomnia patients prescribed sedating antidepressants found no significant increase in overall cancer incidence for trazodone users compared with non-users over a 5-year follow-up period [8]. Hazard ratios for total malignancy hovered near the null (HR 0.98, 95% CI 0.91 to 1.06).

What the Signal Does Not Show

The signal does not identify a specific cancer type consistently across studies. It does not establish a dose-response relationship. It does not appear in the two largest active-comparator cohort studies published to date. A causal signal that cannot specify an organ, a dose threshold, or a biological mechanism through which harm occurs at clinical doses is weak evidence for changing practice.

Comparison With Other Antidepressant Cancer Data

Trazodone is not the only antidepressant to carry a pharmacovigilance cancer signal. SSRIs, particularly paroxetine, have appeared in FAERS analyses for breast cancer due to CYP2D6 inhibition and downstream effects on tamoxifen metabolism [9]. That signal, however, has a clear mechanistic pathway and clinical consequence (reduced tamoxifen efficacy), which is why it appears in oncology prescribing guidance.

Mirtazapine, another sedating antidepressant frequently compared to trazodone for insomnia, has a similar pharmacovigilance profile with no confirmed oncologic harm in prospective data [10]. The American College of Physicians' 2016 insomnia guideline does not restrict any antidepressant class based on cancer risk [11].

What Distinguishes a Weak Signal From a Confirmed Risk

The NCI defines a confirmed cancer risk as one supported by sufficient evidence in humans, typically from at least one well-designed cohort or case-control study with adequate control for confounders, biological plausibility, and consistency across study populations [12]. Trazodone meets none of these criteria at the current time.

Clinical Guidance: What Prescribers Should Do Now

No guideline-issuing body, including the American Psychiatric Association, the American Academy of Sleep Medicine, or the FDA, has recommended avoiding or discontinuing trazodone on the basis of the current cancer signal [2, 11, 13].

For Patients Currently Taking Trazodone

Patients should not stop trazodone abruptly. Discontinuation syndrome, although milder with trazodone than with SSRIs or SNRIs, can include dizziness, insomnia rebound, and irritability. Any medication review should be a shared decision-making conversation, not an alarm-driven switch.

Clinicians should document the indication (depression vs. Insomnia), the dose, and the duration of use in the medical record. This documentation matters both for continuity of care and for any future regulatory review that may require cohort identification.

For Patients Considering Starting Trazodone

The risk-benefit calculation has not changed. For major depressive disorder, trazodone at 150 mg to 400 mg per day remains a reasonable second- or third-line option, particularly in patients who cannot tolerate sexual side effects from SSRIs or who have comorbid insomnia [13]. For off-label insomnia, the Mendelson 2005 data support short-term use at 50 mg to 100 mg, with reassessment at 6 to 8 weeks [6].

Patients with active malignancy or a personal history of cancer are not contraindicated from trazodone by any current guideline. Oncology co-management should drive those decisions.

Monitoring Recommendations

The following monitoring approach is consistent with standard pharmacovigilance practice during a low-confidence signal period:

• Annual medication review documenting continued indication
• Reassessment of continued need at 6 months for purely sleep-related prescriptions
• Age-appropriate cancer screening per USPSTF recommendations, independent of trazodone use [14]
• CYP3A4 inhibitor reconciliation at every visit to prevent mCPP accumulation

Regulatory and Postmarket Surveillance Outlook

The FDA's Office of Surveillance and Epidemiology can initiate a formal safety review (resulting in a MedWatch safety communication or label change) when a signal reaches a defined threshold of evidence quality, biological plausibility, and clinical severity [1]. For trazodone, no formal review has been announced in the FDA's Sentinel System updates or in the Pharmacovigilance Review Agenda published through the Drug Safety Communications portal as of the date of this article [2].

The WHO's VigiBase, which aggregates spontaneous reports from 130 countries, similarly has not issued a signal-of-disproportionate-information (SDI) alert for trazodone and cancer in the most recent annual signal detection publication [15]. European Medicines Agency Pharmacovigilance Risk Assessment Committee (PRAC) minutes through Q4 2024 contain no trazodone oncology review items.

Clinicians can track evolving FDA communications directly through the MedWatch Safety Alerts portal at fda.gov and through the quarterly EMA PRAC highlights [2, 16].

Summary of Evidence Quality

The Bradford Hill criteria provide a structured way to evaluate causal claims in epidemiology. Applying them to the trazodone-cancer question:

| Criterion | Status for Trazodone-Cancer Signal | |---|---| | Strength of association | Weak (ROR 1.2 to 2.1 range; HR near null in cohorts) | | Consistency | Not consistent across study designs | | Specificity | No single cancer site identified | | Temporality | Not formally established in cohort data | | Biological gradient | No dose-response demonstrated | | Plausibility | Theoretical via 5-HT2A; not confirmed in humans | | Coherence | Does not fit known carcinogenesis models | | Experiment | No interventional data | | Analogy | mCPP hepatocyte data at supratherapeutic doses only |

Overall Bradford Hill score: insufficient to support causation. The signal warrants continued surveillance, not prescribing change.