Background: Aromatase inhibitors (AI) are recommended for at least five years to reduce risk of breast cancer recurrence in postmenopausal women with hormone receptor-positive disease. Despite the potential benefit, up to 50% of patients discontinue AI early, primarily because of musculoskeletal side effects. Some patients, including those with fibromyalgia and other chronic pain conditions, have preexisting pain that arises from altered nociception without clear evidence of tissue or peripheral nerve damage, called nociplastic or centralized pain. We hypothesized that preexisting nociplastic pain is associated with early discontinuation of AI therapy due to toxicity. Methods: Patients diagnosed with breast cancer between 2012-19 and who enrolled in the Michigan Genomics Initiative (MGI) were identified. Patients who were female, had hormone receptor-positive breast cancer, had not received neoadjuvant therapy, and who initiated adjuvant AI therapy were included in the analysis. Prior to breast cancer surgery patients completed surveys about pain in the affected breast and overall worst pain (Brief Pain Inventory [BPI]), nociplastic pain (2011 Fibromyalgia Survey [FS], which assesses widespread pain plus somatic symptom severity), and anxiety (PROMIS Anxiety 4a v.1). Demographics, cancer characteristics, and treatment history, including date and reason for discontinuation of AI therapy, were abstracted from the medical record. Patients were censored at date of last oncology follow-up. Univariate and multivariable analyses were conducted to assess relationship between age, BMI, surgical site pain, overall worst pain, FS score, anxiety, and time to discontinuation of initial AI medication. Results: Of 207 analyzed patients, the average age was 61 years, average BMI was 30.3 kg/m², 181 (87%) were postmenopausal at the time of breast cancer diagnosis, and 31% received chemotherapy. The majority were prescribed anastrozole (n=196); 6 took letrozole and 5 took exemestane. Average scores prior to surgery were 1.5/10 for worst pain, 0.2/10 for surgical site pain, 4.2/31 for FS score, and 54.1/100 for anxiety. On univariate analysis, only use of exemestane (HR 5.55, p<.001) and use of letrozole (HR 3.11, p<.001) were associated with shorter time to ET discontinuation; age, BMI, surgical site pain, worst pain, FS score, and anxiety were not statistically significantly associated with time to AI discontinuation. On multivariable analysis, only higher FS score was statistically significantly associated with time to AI discontinuation (HR 1.11, p=.021). Conclusions: Greater pre-existing nociplastic pain was associated with early discontinuation of AI therapy. For these patients, early interventions that target centralized pain, or treatment with tamoxifen instead of AI therapy, may be appropriate in order to reduce risk of nonpersistence with AI therapy.