Aims:
- Primary aim (Aim 1):Evaluate the efficacy of immediate fluoxetine therapy versus calcium in improving PTSD symptomology and physical and mental recovery in the post-injury period for patients with musculoskeletal trauma.
- Hypothesis: Patients randomized to fluoxetine will have less severe depression, anxiety, and posttraumatic stress symptoms, and improved self-reported functional scores, than those receiving calcium.
- Methodology: Patients will be randomized to fluoxetine or calcium during hospitalization following their trauma. Serial patient reported outcomes measures and mental health surveys will be obtained prior to the initiation of treatment and during the first 12 months of their recovery to assess severity and duration of PTSD symptomology
- Secondary aim (Aim 2):Develop a safe fluoxetine treatment protocol with a goal to be therapeutic by 6 weeks and taper completely off treatment by 6-12 months.
- Hypothesis: We will be able to develop a standardized, time-limited protocol that could feasibly be utilized by non-mental health care providers and tolerated by patients.
- Methodology: We will measure medication adherence, medication toleration and side effects, fracture union rate, and other feasibility metrics in the 12-month recovery period.
Design: This is a single-center, repeated measures, nonblinded randomized controlled pilot study. A maximum of 200 participants will be involved in the study, ranging in age from 18-85 years old. This is an investigator-initiated trial that was registered with ClinicalTrials.gov (NCT04850222) on April 20, 2021 before patient enrollment began in September 2021. All study procedures have been approved by the University of Florida’s Institutional Review Board for the Protection of Human Subjects. All research team members have been trained in Good Clinical Practice for Clinical Trials. The financial sponsor of this study is the Orthopaedic Trauma Research Fund at the University of Florida.
- Design rationale:
- Fluoxetine: Fluoxetine, commonly known as Prozac, is an antidepressant that works as a selective serotonin reuptake inhibitor (SSRI) (4). It has typically been prescribed for depression and may also help with anxiety symptoms, with minor to no side effects. Our team chose fluoxetine versus other antidepressants due to its lack of severe side effects, positive outcomes particularly in PTSD patients, and its familiarity in the general population (7).
- Calcium: Calcium is a supplement that may provide some bone healing potential, but overall should have minor side effects on the patients (5). Our team chose calcium because we wanted both study arms to potentially receive a health benefit, whether mental (fluoxetine) or physical (calcium). Utilizing a placebo would not provide an additional health benefit. We selected elemental calcium in particular because most calcium formulations include Vitamin D to help with calcium absorption, but studies have shown Vitamin D may have mental health benefits (8). We want to avoid directly impacting the mental health outcomes of the “non-mental health drug” study arm.
- Nonblinded: Our team has decided to completely unblind for two reasons. First, for patient safety, we prefer our research and clinical team to be aware of a patient’s randomized medication to prevent and better prepare for any adverse side effects. Secondly, we believe that patients that may be biased or hesitant to take a mental health medication will feel more comfortable knowing what medication they are randomized to.
Locations: Our patients will only be recruited at UF Health at Shands Hospital located in Gainesville, Florida, USA. This is a level 1 trauma center serving 18 urban and rural counties comprising two million people. The hospital admits approximately 4500 patients with traumatic injuries annually. Informed consent and baseline surveys were collected at the level 1 trauma center. After discharge, patients are followed up with at the University of Florida Orthopaedic and Sports Medicine Institute. This facility houses the outpatient orthopaedic clinics and is where all patients receive their follow-up care and study subjects will receive their surveys unless phone or email follow-up is required.
Research team: Our research team includes our Principal Investigator Dr. Jennifer Hagen, Division Chief and Associate Professor of Orthopaedic Trauma. Co-Investigators include Dr. Carol Mathews, Professor and the Vice Chair for Strategic Development in the Department of Psychiatry, Dr. Ludmila Barbosa de Faria, Associate Professor in the Department of Psychiatry, Dr. MaryBeth Horodyski, Professor and Director of Research in the Department of Orthopaedic Surgery and Sports Medicine, and Dr. Terrie Vasilopoulos, Assistant Professor of Anesthesiology and Orthopaedic Surgery and Sports Medicine. Additionally, all five orthopaedic trauma surgeons and five Physician Assistants gained IRB approval and assisted in study procedures. Drs. Hagen and Barbosa de Faria review and report serious adverse events (SAEs), escalation of mood disturbance, and drug side effects in real-time. Dr. Carol Mathews reviews study conduct and data and provides recommendations as needed. Research study team members (including one Clinical Research Coordinator and one Research assistant) are involved in subject recruitment and consent administration, in-person and phone call study follow-ups, and day-to-day administration work.
Patients:
Inclusion criteria:
- persons aged 18-85 years old
- admitted for high energy musculoskeletal trauma (more than ground level fall) resulting in one or more operative extremity or pelvic fractures
- have granted informed consent and enrolled for the study in the hospital prior to discharge.
Exclusion criteria:
- presence of a traumatic brain injury
- past medical history or self-reported bipolar disorder or other mental health conditions on active pharmacologic treatment (including but not limited to suicidal ideation or depression)
- pregnancy
- incarceration
- expected injury survival of less than 90 days
- unlikely to maintain follow up
- any contra-indications to fluoxetine or calcium
- unable to provide informed consent due to language barriers or other barriers (e.g., intubation)
- fracture managed outside of the participating orthopaedic service (e.g., a hand fracture managed by plastics)
- transferred from an outside hospital
- have a medical or physical condition in opinion of the investigator that would preclude safe participation in the study.
Recruitment and consent: Study recruitment began in September 2021. Planned enrollment for race, ethnicity, and sex can be found in Table 1. After admission to UF Health Center and after at least one orthopaedic surgical intervention, patients are approached by a study team member who has pre-screened the patient for potential recruitment. Before initiating the informed consent process, the study team member assures that the patient is stable enough to proceed with consenting. Standard IRB enrollment guidelines are followed by team members to properly consent and enroll participants.
Personal subject data such as name, medical record number, age, home address, and phone number(s) will be de-identified as each subject is assigned a research identification number after consenting. This is to assure that at the completion of the study, no one will be able to associate the patient’s personal information with their study data. The master link between their research identification number and personal identifiers will be kept in a secure location and accessible to study personnel only. Additionally, this study team has obtained a Certificate of Confidentiality from the National Institutes of Health to further protect the subject’s privacy. Study data will be reviewed on a quarterly basis by Dr. Mathews and no external site monitor will be used for this study.
Table 1: Study Planned Enrollment Report (based on previous prospective study in this patient population) (1)
Racial Categories
|
Ethnic Categories
|
Not Hispanic or Latino
|
Hispanic or Latino
|
Total
|
Female
|
Male
|
Female
|
Male
|
American Indian/ Alaska Native
|
1
|
1
|
|
|
2
|
Asian
|
1
|
1
|
|
|
2
|
Native Hawaiian or Other Pacific Islander
|
1
|
1
|
|
|
2
|
Black or African American
|
4
|
10
|
|
|
14
|
White
|
30
|
43
|
4
|
3
|
80
|
More Than One Race
|
|
|
|
|
0
|
Total
|
0
|
0
|
0
|
0
|
100
|
Study processes:
Survey administration: The initiation of the study occurs at baseline, which is after the patient has been consented, but prior to hospital discharge and initiation of the randomized study medication. At baseline, patients begin with surveys regarding their depression, anxiety, a self-report of mental health concerns, and pain, enjoyment, and activity scales (Table 2). As it is the participant’s first time hearing the questions, these surveys may take up to 30 minutes. After hospital discharge, patients return to the orthopaedic clinic for post-surgical follow-ups. To minimize patient burden, study data is collected during their standard of care clinic follow-up visits. This means subjects will not need to visit the clinic for study purposes only and will come into the clinic as often as a non-study participant patient. Study personnel will collect patient data and administer surveys at follow-up visits at 2 weeks, 6 weeks, 3 months, 6 months, and 12 months (one year) post-surgery. To maintain quality follow-up rates, phone surveys will be planned as the type of follow-up for the 1-year visit, unless the patient reports surgery-related issues and is present in the clinic for a scheduled visit. The survey schedule for each visit can be found in Table 2. Total time to complete all the surveys at each timepoint is 15-20 minutes. Follow-up visits can be within a given range to meet the timeframe criteria, and this range for each follow-up is noted in Table 2.
In addition to collecting patient reported outcomes we will also record use of narcotics, cannabis supplements, illegal substances, and alcohol consumption. Prescription drug monitoring program (PDMP) will be queried for other sources of narcotic prescriptions. Patient reported pain scores (PEG score) will be recorded at each visit as one of the required surveys for completion. Any complications related to the surgery (e.g., infection, nonunion, unplanned return to operating room) will be documented. If patients do not return to clinic for their follow-up appointment, study personnel will call the patients to complete the surveys over the phone or send the participant the survey via email.
Table 2: Outcome assessment schedule
|
Baseline
|
2 weeks
|
6 weeks
|
3 months
|
6 months
|
1 year
|
ENROLLMENT
|
|
|
|
|
|
|
Eligibility
|
x
|
|
|
|
|
|
Consent
|
x
|
|
|
|
|
|
Demographics
|
x
|
|
|
|
|
|
AIM 1
|
|
|
|
|
|
|
Mental Health
|
|
|
|
|
|
|
BDI-II
|
x
|
x
|
x
|
x
|
x
|
x
|
BAI
|
x
|
x
|
x
|
x
|
x
|
x
|
PSS-SR5
|
x
|
x
|
x
|
|
x
|
x
|
Sleep Quality
|
|
|
|
|
|
|
PSQI
|
|
x
|
x
|
x
|
|
|
Pain and Pain Interference
|
|
|
|
|
|
|
PEG
|
x
|
x
|
x
|
x
|
x
|
x
|
Social Participation
|
|
|
|
|
|
|
PROMIS
|
|
|
x
|
x
|
x
|
x
|
AIM 2
|
|
|
|
|
|
|
Medication Adherence, Safety, and Acceptability
|
|
|
|
|
|
|
MMAS-8
|
|
x
|
x
|
x
|
x
|
x
|
ASEC
|
|
|
x
|
|
|
|
[BDI-II- Beck Depression Inventory-II; BAI- Beck Anxiety Inventory; PSS-SR5- PTSD Scale-Self report for DSM-5; PSQI- Pittsburgh Sleep Quality Index; PEG- Pain, Enjoyment, General Activity Scale; PROMIS- Physical function; MMAS-8- Morisky Medication Adherence Scale; ASEC- Antidepressant Side-Effect Checklist].
Drug administration and cessation: Patients will be made aware of what drug they are randomized to after the informed consent process and baseline surveys are completed. They will begin their randomized study medication before hospital discharge. While the patient is still admitted in the hospital, they will be prescribed a 90-day supply of the randomized medication to take post-hospital discharge. Those patients that have been randomized to fluoxetine will be started on 10mg daily and have their dose increased to 20mg PO/day at their 2-week visit, while the calcium group will continue their 1000 mg original dose. Patients will be provided with subsequent 90-day prescriptions by our outpatient pharmacy at 3 months and prescriptions will be billed to the study. At the 6-month visit, one last 90-day prescription of the participant’s randomized medication will be given. If patients are unable to come into the clinic for this visit, their medication will be mailed to them. If patients are unable to come into the clinic for any of their visits within the required timeframe for survey completion, they will be contacted via phone call or email to complete the required surveys. Study personnel will use the phone numbers and emails provided by patients at baseline, and any phone or email changes will be updated at each clinic visit to assist with subject follow-up success.
Although there is no survey requirement at 9 months post-surgery, patients will receive a follow-up phone call or email around this timeframe to check in on their post-medication cessation symptoms. Since their last medication dose should have been completed or stopped by this timepoint, patients will be monitored for symptomology changes. Patients in the fluoxetine study arm reporting worsening of mental health symptoms following cessation will be referred to their primary care physician for further treatment. A visual of the study timeline is shown in Figure 1, and a visual of the drug administration timeline is provided in Figure 2.
Study retention: Retaining patients in all 12 months of the study period will be a priority of the study team. To avoid loss to follow-up, patients that are unable or unwilling to return to the clinic for in-person appointments will be contacted via phone call. When feasible and appropriate, patients may also be offered telemedicine appointments. Phone surveys will be planned as the type of follow-up for the 1-year visit, unless the patient reports surgery-related issues and is present in the clinic for a scheduled visit.
Safety monitoring: A monthly review will be conducted for participant side effects, mood scores, and medication adherence by Dr. Hagen and Dr. Barbosa de Faria. Any serious adverse event (SAE) will be identified and addressed, and properly adjusted through dose titration, medication stoppage, and/or mental health evaluation. If any participant answers a survey question that exceeds a threshold where they may be experiencing harmful thoughts, questionnaires have been set up in REDCap to alert Dr. Barbosa de Faria and Dr. Hagen. The thresholds for alerting the study team vary by survey and are established as follows: BDI II more than or equal to 28 (severe range); BAI = more than or equal to 26 (severe range); PSS SR: anything above 34 (moderate range). With this notification, Dr. Barbosa De Faria and Dr. Mathews will triage issues, and appropriate care will be given. Participants may be referred to UF Health mental health services or a local community mental health organization.
Potential discomfort and risks: All participants for the study have been admitted to the hospital through orthopaedic surgery to treat their injury. There is no additional anesthetic or complication risk that is subjected to study participants that is not already a typical risk for patients following orthopaedic surgery. These risks are discussed with the patient prior to surgery and are not directly related to the study. Many of the surveys presented in the study may be uncomfortable or even distressing for some patients. The questions in the surveys regard their pain, mood, and feelings and include distressing themes such as self-harm and suicide. All answers are confidential and only viewed by the IRB approved study staff. If patients state that they are expressing self-harm or thoughts of suicide, study personnel will address them as specified prior. This study may include risks that are unknown at this time.
- Risks related to study drugs are documented below:
- Fluoxetine: Some patients taking fluoxetine might experience GI distress (nausea, vomiting, diarrhea), headaches, insomnia, nervousness, anxiety, sweating, abnormal dreams, and dizziness. If patients experience these symptoms, then Dr. Barbosa de Faria or Dr. Mathews will decrease the dose or stop the medication. It is possible patients could have improvement in depressive, anxious, and other negative mood symptoms if they are randomized to fluoxetine. As fluoxetine has a decades long safety record in the general population, we do not anticipate having to stop the trial due to overwhelming participant side effects. If an individual patient is unable to tolerate fluoxetine, they will no longer be given the medication but will be followed for the duration of the study (12 months) and analyzed as intent-to-treat.
- Calcium: Some patients taking calcium might experience some minor side effects such as belching or gas. Calcium is potentially unsafe when taken by mouth in doses above the daily tolerable upper intake level (UL). The UL is 2000 mg for adults ages 19 and older and 1300 mg for adults aged 18. Our dosage of calcium for subjects in the calcium group is 1000 mg by mouth per day, well within the safety margins. Taking more than this amount of calcium daily can increase the chance of having serious side effects, such as milk-alkali syndrome, a condition that can lead to renal stones, kidney failure, and death. It is possible patients in the calcium group could benefit from calcium’s bone healing potential.
Randomization
- Allocation and concealment:
- Once consented, patients will be randomized to fluoxetine (10mg by mouth per day) or calcium (1000mg by mouth per day). The randomization will be stratified by gender and initial BDI-II score (no/mild symptomology BDI 0-13 or moderate severe symptomology BDI>14). Both BDI-II groups will be randomized to treatment because although the initial BDI-II was correlated with long term symptomology, the prior study also had patients with no initial symptomology develop it over time (4)
- Medication randomization was done before beginning the consenting process, and stringent measures have been taken to prevent bias. The established order of medication randomization will be stored in two places: a secured online drive and within sealed envelopes in a secure location. To prevent bias, the envelopes are used to hide the next randomized medication from the view of the orthopaedic team member who administered the consent. The envelopes are separated by BDI score and gender (BDI <13 or >14, male or female). After consent has been given and the baseline BDI survey administered, the orthopaedic team member will select the envelope that corresponds to the patient to reveal the randomization group. The randomized medication will be prescribed by the orthopaedic team on the day of randomization to enhance monitoring of the patient for side effects during the remainder of their hospitalization. The patient will be prescribed the randomized medication on the day of discharge with a 90-day supply provided by the inpatient research pharmacy.
Outcome Measures:
Aim 1: Improvement in symptomology
We will be utilizing the Beck Depression Inventory-II (BDI-II, Primary outcome) (5) Beck Anxiety Inventory (BAI) (6) PTSD Scale-Self report for DSM-5 (PSS-SR5-PTSD) (7); Pittsburgh Sleep Quality Index (PSQI) (8); Pain, Enjoyment, General Activity Scale (Peg) (9); and Patient-Reported Outcomes Measurement Information System-Physical function surveys (PROMIS-PF) (10). All are validated patient reported surveys. Survey administration will be facilitated by the study team via data entry into portable tablet devices.
Aim 2: Safety and feasibility
We will be utilizing the Morisky Medication Adherence Scale (MMAS-8, Primary outcome) and Antidepressant Side-Effect Checklist (ASEC) to assess both the acceptability and safety of the use of fluoxetine in the patient population (11,12). Desire to continue medication/treatment past the 9th month point will also be recorded as will incidence of escalation in psychiatric distress in both patient arms.
Analysis
- Sample size determination: Sample size calculations focused on 1) efficacy of intervention and 2) patient acceptability of intervention (fluoxetine adherence). Calculations assumed 80% power, alpha=0.05, and a 10% drop rate after baseline; we observed a loss to follow up of approximately 10% at each timepoint with our previous prospective clinical trial (13). Our goal is to follow the patients for 12 months and with five follow up time points scheduled; we estimate a dropout rate of 50% by 12 months. Our target sample size is n=90 (n=45 per group); we will aim to recruit n=100 at baseline. For longitudinal efficacy analyses (BDI-II, Aim 1), the proposed linear mixed model analysis is able to handle missing data and uneven data structure for patients; all patients will be included in these analyses as long as they have one follow-up timepoint of completed data; GLIMMPSE online program was used for sample size calculation for Aim 1 and estimates were used from our previous study from our group.(13) For Aim 1, n=90 will be able to detect between group differences of 4 points in BDI-II (SD=9); these analyses included five time points for repeated measures, with assumed correlation of 0.75 for adjacent timepoints that decreases to 0.5 for farthest apart timepoints. Baseline BDI-II was also included in the model, with assumed correlation of 0.55 with 2-week BDI-II that decreases to 0.2 at 1-year BDI-II. For medical adherence (MMAS-8, Aim 2), if we observe 75% of sample with medium or high adherence in n=45 of the fluoxetine group, the 95% confidence interval would be (63%, 88%); with n=23 (potential at 1-year), confidence interval would be (56%, 92%).
- Data analysis plan: Analyses will be performed in JMP Pro 15 (SAS Institute, Inc, Cary, NC). Continuous measures will be summarized as means and standard deviations and categorical measures will be summarized as counts and percentages. For intervention efficacy analyses (Aim 1), linear mixed models for repeated measures will be used for both primary (BDI-II) and secondary (BAI, PSQI, PSS-SR5, PEG, PROMIS-PF scales) outcomes. These models account for within-person correlations across time points and can handle uneven data structure. These models utilize maximum likelihood estimation so that patients with incomplete data at some timepoints will still be included in analyses. Baseline measures (if applicable) will be included as independent variables in these models along with intervention group. Tukey’s post hoc test (which adjusts for multiple comparisons), will be used to test for pairwise group differences at each timepoint, with mean differences reported with 95% confidence intervals. A significant group x time interaction will be modeled to test if changes in outcomes differed between the intervention groups. For Aim 2, counts and percentages (with 95% confidence intervals) of adverse outcomes will be recorded. Medicine adherence (primary outcome) will be measured at every time point starting at week six, with patients classified as high adherence (8), medium adherence (6-7), and low adherence (< 6), which will be reported as percentages with 95% confidence intervals. (11) P < 0.05 will be considered statistically significant, with only one primary outcome, no correction for multiple comparisons will be performed; raw p-values for all analyses will be reported (14,15).
- Data management and confidentiality: Data will be collected using electronic case report forms (CRFs) and immediately entered into the Research Electronic Data Capture (REDCap) database. REDCap is a secure, Web-based software platform designed to support data capture for research studies, providing 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources. (2,3) All surveys utilized are validated tools.
- Dissemination policy: During the study, the research team will notify participants of new information that may become available and might affect their decision to remain in the study. They may not be allowed to see the research information collected about them for this research study, including the research information in their medical record, until after the study is completed. When this research study is over, participants will be allowed to see any research information collected and placed in their medical record. A deidentified dataset can also be made available to participants after completion of the study, upon reasonable request to the corresponding author.