Abstract
Objective:
Fluoxetine, paroxetine, sertraline, topiramate, and venlafaxine have previously shown efficacy for PTSD. One prior study using VA medical records data to compare these agents found no differences in symptom reduction in clinical practice. We address several weaknesses in that study, including limited standardization of treatment duration, inability to account for prior treatment receipt, use of an outdated symptomatic assessment for PTSD, and lack of functional outcome.
Method:
We identified 834 VA outpatients with DSM-5 clinical diagnoses of PTSD between October 2016 and March 2018 who initiated one of the medications and met pre-specified criteria for treatment duration and dose, combined with baseline and endpoint PTSD checklist-5 (PCL-5) measurements. We compared 12-week acute phase changes in PCL score and remission of PTSD symptoms. We compared use of acute psychiatric services use in the subsequent 6-month continuation phase.
Results:
In the acute phase, patients improved by a mean of 6.8–10.1 points on the PCL-5 and 0.0%−10.9% achieved remission of PTSD symptoms. Those taking venlafaxine were significantly more likely to achieve remission (p<0.0001). In the continuation phase, there were no differences in acute psychiatric care use between medications. Those who continued their medication were less likely to use acute psychiatric services (HR=0.55; p=0.03).
Conclusion:
There may be an advantage to venlafaxine over other agents in achieving acute-phase remission for DSM-5 PTSD routine clinical practice, but this requires further study. Regardless of the agent chosen, medication cessation during the continuation phase is associated with a higher risk of acute psychiatric care use.
INTRODUCTION
Posttraumatic stress disorder (PTSD) is a serious condition that can follow exposure to a traumatic event, characterized by intrusive re-experiencing of the trauma, avoidance of trauma reminders, negative alterations in cognitions and mood, and increased arousal and reactivity.1 PTSD has a lifetime prevalence of 6.1% in the United States.2 Over 10% of Veterans receiving care in the Department of Veterans Affairs (VA) health care system have PTSD, comprising an active caseload of approximately 600,000 in 2016.3
Randomized controlled trials (RCTs) show that effective treatments for PTSD include both pharmacologic and psychotherapeutic approaches.4,5 Several individual medications have shown efficacy as PTSD treatments in placebo-controlled RCTs.4,5 Because there is limited data comparing medications that are individually superior to placebo to one another in a single population, one prior VA study used electronic medical record (EMR) data from 2008–2013 to compare the real-world clinical effectiveness of five efficacious medications.6 While that study found no differences in symptom reduction between fluoxetine, sertraline, paroxetine, topiramate, and venlafaxine, there were several weaknesses. These included limited standardization of treatment duration, inability to account for prior treatment receipt, use of an outdated patient-reported outcome measure (PROM) for PTSD that aligned with the fourth version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV),7 and lack of a functional outcome. We sought to improve upon this study by addressing these limitations.
Therefore, we conducted a retrospective comparative effectiveness study of the same five medications for PTSD using contemporary VA EMR data. We accounted for prior receipt of evidence-based pharmacologic and psychotherapeutic approaches for PTSD dating back almost 20 years, standardized acute-phase treatment duration at 12 weeks, and aligned acute-phase treatment with administration of the PROM for PTSD that is updated for the fifth version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5).1 Additionally, we compared the functional outcome of acute psychiatric services use in the subsequent 6-month continuation phase among the five agents. This replication and extension is important both to ensure that the prior null finding is not due to type II error and because different treatments may be effective under the DSM-5 case conceptualization of PTSD,8–10 which was implemented in 2013 and has an increased emphasis on negative alterations in cognitions and mood compared to DSM-IV.1,7
Method
Data Sources
This was a retrospective chart review. We used the VA Corporate Data Warehouse (CDW) to identify all VA users with a DSM-5 clinical diagnosis of PTSD (F43.1x) from 10/1/2016–3/7/2018. While DSM-5 was published in 2013, development and EMR-based implementation of related diagnostic and outcomes assessment tools in the VA occurred slowly, thus we chose 10/1/16 as the start date. We obtained information on services use, clinical diagnoses, pharmacy data, and standardized PTSD symptom measures from the CDW for these patients. This study was approved by the Veterans Institutional Review Board of Northern New England.
Cohort Selection
We identified patients who initiated a course of fluoxetine, sertraline, paroxetine, topiramate, or venlafaxine. The study sample was further restricted to those who met our criteria for adequate acute phase medication management. Patients receiving continuous treatment of sertraline, fluoxetine, paroxetine, venlafaxine, or topiramate daily for ≥12 weeks at an adequate dose were considered to have received an adequate medication trial (AMT). Adequate doses, which were required for the final 8 weeks only to allow for titration, were as follows: fluoxetine ≥20 mg, paroxetine ≥20 mg, sertraline ≥100 mg, topiramate ≥100 mg, and venlafaxine ≥150 mg.
For our outcomes analysis, we further restricted to those who received baseline PTSD symptom measurement within 2 weeks of treatment initiation, as well as follow-up symptom measurement within 2 weeks of the 12-week point, and met our symptomatic criteria for PTSD at baseline (defined below).
PTSD Symptoms
We measured PTSD symptoms using the DSM-5 PTSD Checklist (PCL-5),11 which is administered in routine VA clinical practice. We used a baseline cutoff score of ≥31 out of 80 due to optimal efficiency for diagnosing PTSD in Veterans, compared to the gold-standard Clinician Administered PTSD Scale for DSM-5.12 Our minimal symptomatic criteria required a score of “moderate” or higher on one avoidance symptom, two negative alterations symptoms, and two increased arousal symptoms.
While a threshold for clinically meaningful change had not yet been established when we implemented our coding rules, the largest prospective trial using the PCL-5 at that time used a severity score of ≤18 as a cutoff for remission.13 Therefore, we considered a score of ≤18 plus no longer meeting symptomatic criteria to be consistent with remission at follow-up. In addition to examining overall change in symptoms, we evaluated change in sub-scores for PTSD symptoms clusters as well as sleep difficulties using the sum of two items: nightmares and insomnia.
Acute Psychiatric Services Use
We determined whether patients were admitted to a VA psychiatry unit (acute inpatient or observation) or visited a VA emergency room for a primary psychiatric indication during the six-month continuation phase, which followed the initial 12-week acute phase.
Independent Variables
We measured six groups of covariates that could plausibly affect the relationship between treatment and outcome. See Table 1 for details.
Table 1:
Explanation of covariates
| Trial Characteristics | |
|---|---|
| Number of Adequate Medication Trials (AMTs) Aligned with PTSD Checklist (PCL) Measurement | Adequate trials of fluoxetine, paroxetine, sertraline, topiramate, or venlafaxine aligned with PCL measurement that each patient contributed to the outcomes analysis. Trials of different agents could overlap or dovetail, but we required a one-year gap in prescriptions to count as a new trial of the same agent. |
| PCL Severity and Timing | Baseline PCL score, number of days between first available PTSD diagnosis and baseline PCL, number of days from baseline PCL to follow-up PCL for each trial included in the outcomes analysis. |
| Number of Prior AMTs | AMTs with or without PCL measurement between 1999 and the start of each trial included in the outcomes analysis. |
| Number of Prior Adequate Prolonged Exposure (PE) or Cognitive Processing Therapy (CPT) Trialsa | Episodes where patients received ≥8 sessions PE or CPT over the course of one year between 1999 and the start of each medication trial included in the outcomes analysis. |
| Concurrent Treatments | Additional treatments received at the same time as an AMT associated with PCL measurement |
| Psychotherapy | Categorical receipt and number of sessions |
| PEa | Individual only |
| CPTa | Group and individual |
| Other psychotherapy | Group and individual |
| Medications | Categorical receipt of other antidepressants, sedative hypnotics, opioids, atypical antipsychotics, prazosin, medications for alcohol abuse including naltrexone or acamprosate, and opioid replacement medications including buprenorphine or methadone prescribed within the context of methadone treatment clinic |
| Primary Prescribing Clinician Characteristics | Clinician who wrote the plurality of each patient’s psychotropic prescriptions during the 12-week treatment period |
| Age | Continuous |
| Gender | Categorical male or female |
| Professional background | e.g. psychiatrist or nurse practitioner |
| Percent of time spent seeing PTSD patients in various settings | e.g. specialized PTSD clinic or primary care clinic, based on assumption that prescribing clinicians who spend a higher percentage of their time in specialized PTSD settings may bring increased knowledge and experience in treating PTSD, even when seeing patients in non-specialized settings. |
| Baseline Patient Characteristics | Demographics, military service characteristics |
| VA Health Service use Characteristics | Assessed in the year preceding baseline PCL. |
| Outpatient visits | e.g. visits to specialized PTSD clinics or to primary care clinics |
| Acute psychiatric care use | e.g. emergency department visits for psychiatric indications or psychiatric hospitalizations |
| Residential treatment | e.g. stays in residential PTSD or substance abuse programs |
| Psychiatric Comorbidities | Psychiatric diagnoses in the two years preceding the baseline PCL measurement |
EBP use was measured with a natural language processing algorithm that classifies psychotherapy notes in individual and group delivery formats.14
Analysis
To understand how AMTs with aligned PCL measurement differed from AMTs initiated without aligned PCL measurement from 10/1/2016–3/7/2018, we compared covariates describing concurrent treatment, primary prescribing clinicians, patient characteristics, VA service use characteristics, and comorbidities using χ2 analysis and t-tests, as appropriate.
To account for differences in covariate profile among trials of each of the five medications, we used the RAND Toolkit for Weighting and Analysis of Nonequivalent Groups (TWANG).15 The TWANG package supports causal modeling of observational data through the estimation and evaluation of propensity scores and associated weights. In our application, the propensity score represented the probability that a particular trial would be of each medication.16 We estimated propensity scores with multinomial logistic regression using generalized booster effects,17 in which the dependent variable is an indicator for each of the five medications and the independent variables are an anti-parsimonious specification of variables that have a plausible correlation with the outcome (i.e. our six groups of covariates).16,17 Using these propensity scores, we weighted participants in order to balance the covariate distributions across medications.
We compared continuous and categorical outcomes among the five groups with regression analyses, using medication received as the sole independent variable. In general, weighted means can have greater sampling variance than unweighted means. Therefore, we used survey commands, which account for the weights, to perform the outcomes analyses when comparing the weighted groups. These weighted groups were defined by the inverse of the propensity scores and adjusted covariates unbalanced at the p<0.01 level after TWANG weighting. In balancing over 50 covariates, a Bonferroni correction would indicate a corrected alpha of p<0.001. However, we conservatively maintained an alpha threshold of p<0.01 for significant differences to avoid type II error. For acute-phase continuous outcomes of pre/post change in total PCL score and sub-scores, we used linear regression analysis, whereby the coefficient of the variable tests the hypothesis that each of the five psychotropic medications has the same mean change from baseline to follow-up. For our categorical outcome of remission, we used logistic regression analysis, whereby the coefficient of the variable tests the hypothesis that each of the five psychotropic medications results in the same percentage of patients achieving remission. We assessed the potential contribution of unmeasured confounding on significant baseline to follow-up comparisons by calculating E-values, which indicate the minimum strength of association on the risk ratio scale that an unmeasured confounder would need to have with both the exposure and the outcome, conditional on the measured covariates, to fully explain away a specific exposure-outcome association.18,19
Finally, for the continuation phase we used a weighted proportional hazard models to measure differences in acute psychiatric services use in the subsequent 6-month continuation phase, controlling for symptom change during the acute phase as well as whether there was prescription fill evidence that patients continued to take each medication. We performed data management in SAS version 9.4 (SAS Institute), and developed causal models in R version 3.5.0 (R core team). This included IPTW models created using the R TWANG package,15 and models to detect unmeasured confounding using the R EVALUE package.20
Results
There were 834 AMTs aligned with PCL measurement and 38,089 AMTs that were not aligned with PCL measurement. Patients who had AMTs aligned with PCL measurement generally contributed only one trial (Table 2), had received few adequate prior trials of evidence-based treatments for PTSD, and had severe baseline PTSD symptoms (M=57.8; SD=11.1). Inclusion of data from the early implementation of the EMR-based PCL-5 tool (10/1/2014–9/30/2016) would have yielded a maximum of 21 additional AMTs aligned with PCL measurement while making the analytic cohort less representative of the overall population receiving AMTs during the period of examination. There were 16 cases where AMTs aligned with PCL measurement overlapped, and all of these cases involved concurrent prescribing of topiramate with one of the four antidepressants. AMTs associated with PCL measurement in our analytic cohort differed from contemporaneous AMTs without PCL measurement in many ways (Table 3). Most notably, AMTs with measurement coincided with higher levels of all forms of individual and group psychotherapy, including PE, CPT-I, and CPT-G.
Table 2:
Characteristics of New Trials of Adequate Dose and Duration Evidence-Based Medications for PTSD, with aligned PCL Measurement, including start dates from October 1, 2016 through March 7, 2018
| Number of AMTs Aligned with PCL Measurement Patients Contribute | |
| 1, %(n) | 94.2 (786) |
| 2, %(n) | 5.8 (48) |
| Number of Prior AMTs since October 1, 1999 (with or without PCL) | |
| 0, %(n) | 81.1 (676) |
| 1, %(n) | 15.0 (125) |
| 2+, %(n) | 4.0 (33) |
| Number of Prior Adequate PE or CPT Trials since October 1, 1999 (with or without PCL) | |
| 0, %(n) | 93.2 (777) |
| 1, %(n) | 6.2 (52) |
| 2+, %(n) | 0.6 (5) |
| Days from First Available PTSD Diagnosis to Baseline PCL, M (SD) | 1,074.6 (1,378.7) |
| Days from Baseline PCL to Follow-Up PCL, M (SD) | 80.8 (10.3) |
| Baseline PCL Score, M (SD) | 57.8 (11.1) |
Abbreviations. PTSD=posttraumatic stress disorder, PCL=PTSD Checklist, FY=Fiscal Year, AMT=Adequate Medication Trial (12 or more weeks of fluoxetine, sertraline, topiramate, paroxetine, or venlafaxine at required dose at a minimally adequate dose), PE=Prolonged Exposure, CPT=Cognitive Processing Therapy
Table 3:
Comparison of New Trials of Adequate Dose and Duration Evidence-Based Medications for PTSD, with and without aligned PCL Measurement, including start dates from October 1, 2016 through March 7, 2018
| 0.8 (319) | 8.2 (68) |
| 3.6 (2.8) | 4.5 (3.1) |
| 2.7 (1,018) | 29.6 (247) |
| 3.6 (2.9) | 5.2 (3.3) |
| 1.1 (421) | 8.4 (70) |
| 4.6 (4.3) | 5.8 (4.2) |
| 32.2 (12,264) | 62.2 (519) |
| 15.2 (5,792) | 30.9 (258) |
| 51.9 (19,749) | 56.0 (467) |
| 28.7 (10,924) | 27.2 (227) |
| 21.9 (8,334) | 16.1 (134) |
| 13.5 (5,149) | 9.0 (75) |
| 16.8 (6,380) | 15.2 (127) |
| 29.3 (11,143) | 40.5 (338) |
| 2.9 (1,097) | 4.8 (40) |
| 1.3 (483) | 1.3 (11) |
| 51.0 (12.1) | 49.6 (12.3) |
| 37.9 (14,449) | 37.1 (309) |
| 40.9 (15,588) | 45.2 (377) |
| 31.8 (12,103) | 22.2 (185) |
| 4.3 (1,651) | 4.9 (41) |
| 18.0 (6,858) | 18.6 (155) |
| 2.9 (1,120) | 8.6 (72) |
| 5.7 (19.7) | 9.4 (24.1) |
| 2.5 (11.4) | 2.1 (9.0) |
| 76.7 (39.7) | 82.7 (34.3) |
| 5.6 (17.3) | 8.1 (21.1) |
| 18.3 (37.3) | 12.0 (31.1) |
| 46.5 (14.6) | 40.9 (11.1) |
| 16.9 (6,434) | 16.1 (134) |
| 54.9 (20,925) | 55.8 (465) |
| 33.6 (12,789) | 33.0 (275) |
| 63.2 (24,079) | 60.8 (507) |
| 20.6 (7,835) | 18.1 (151) |
| 9.2 (3,491) | 13.3 (111) |
| 48.7 (18,533) | 68.9 (575) |
| 11.9 (4,535) | 4.0 (33) |
| 42.2 (16,069) | 49.9 (416) |
| 14.9 (5,688) | 15.1 (126) |
| 49.7 (18,933) | 50.8 (424) |
| 19.1 (7,260) | 33.7 (281) |
| 9.6 (15.7) | 9.6 (11.2) |
| 88.3 (33,612) | 88.1 (735) |
| 20.6 (42.1) | 26.9 (45.5) |
| 11.0 (4,205) | 15.4 (128) |
| 22.9 (37.7) | 21.9 (42.7) |
| 87.9 (33,495) | 82.1 (685) |
| 7.1 (7.8) | 6.0 (5.9) |
| 13.1 (4,979) | 13.2 (110) |
| 2.1 (2.2) | 1.8 (1.6) |
| 8.6 (3,257) | 11.3 (94) |
| 16.1 (24.8) | 17.8 (23.3) |
| 1.0 (364) | 2.9 (24) |
| 44.8 (40.7) | 28.8 (21.0) |
| 1.9 (728) | 2.5 (21) |
| 43.2 (43.3) | 39.3 (33.6) |
| 26.5 (10,110) | 35.3 (294) |
| 3.9 (7.9) | 3.1 (4.3) |
| 12.1 (4,596) | 10.1 (84) |
| 2.4 (2.3) | 2.3 (1.8) |
| 14.7 (5,597) | 14.6 (122) |
| 2.2 (1.6) | 2.3 (1.8) |
| 6.3 (2,379) | 11.5 (96) |
| 4.7 (13.7) | 3.3 (4.7) |
| 80.7 (30,749) | 80.9 (675) |
| 31.4 (11,974) | 36.9 (308) |
| 3.8 (1,449) | 1.6 (13) |
| 8.2 (3,103) | 6.0 (50) |
| 73.0 (27,802) | 80.8 (674) |
| 46.8 (17,840) | 50.0 (417) |
| 8.6 (3,267) | 14.8 (123) |
| 27.6 (10,522) | 33.3 (278) |
| 5.7 (2,176) | 7.3 (61) |
| 16.6 (6,337) | 21.0 (175) |
p<0.05,
p<0.01,
p<0.001
Abbreivations. PTSD=posttraumatic stress disorder, PCL=PTSD Checklist, FY=Fiscal Year, PE=Prolonged Exposure, CPT=Cognitive Processing Therapy, F/S/P/V=Fluoxetine/Sertraline/Paroxetine/Venlafaxine, PCT=PTSD Care Team, OEF/OIF/OND=Operations Enduring Freedom/Iraqi Freedom/New Dawn, VA=Department of Veterans Affairs.
The number of participants in the analytic cohort receiving each medication ranged from 307 who received sertraline to 87 who received paroxetine. While there were differences among the medication treatment groups (Supplementary Table 1), our weighting procedure allowed us to balance almost all covariates (Supplementary Table 2). The exceptions were the percentage of time the primary prescribing clinician spent working in the integrated care service section and the percentage of Vietnam veterans receiving each medication, with both being significantly lower in the topiramate group. Therefore, these variables were retained as covariates, along with a covariate for concurrent antidepressant and topiramate prescribing, in subsequent analyses. PTSD symptom measurement was well-aligned to acute-phase medication treatment, with participants’ baseline PCLs administered at 1.1 days (SD=6.4) days after the start of the medication and end-point PCLs were administered at 80.8 days (SD=10.3) later. Mean baseline PCL scores indicated a high burden of symptoms, ranging tightly from 57.5 (SD=10.9) for the sertraline group to 58.6 (SD=11.8) for the venlafaxine group.
All five of the medications were associated with moderate acute phase improvements in PTSD symptoms (Table 4). The mean improvement in total PCL score ranged from 6.8 points for the paroxetine and topiramate groups to 10.1 points for the venlafaxine group; acute phase remission rates ranged from 0% for the paroxetine group to 10.9% for the venlafaxine group. While there was no difference in total PCL change between the agents, there was a significant overall difference in achievement of remission (p<0.0001). Pairwise comparisons indicated superior achievement of remission between the venlafaxine group compared to other groups (venlafaxine versus fluoxetine p=0.008, venlafaxine versus paroxetine, sertraline, and topiramate p<0.0001). We could not calculate E-values for comparisons involving paroxetine as there were no remissions in the paroxetine group. However, where they could be calculated, E-values indicated the superior achievement in acute phase for the venlafaxine group to be robust (venlafaxine versus fluoxetine 7.3, versus sertraline 4.2, versus topiramate 15.2). Similarly, pairwise comparisons indicated inferior achievement of remission for paroxetine compared to all other groups (p<0.0001). Finally, sertraline was significantly superior to topiramate in achievement of acute phase remission (p<0.0001; E=6.0).
Table 4:
Weighted Outcomes for Patients with an Adequate Trial of an Effective Medication for PTSD plus PCL Measurement
| Fluoxetine (n=228) | Paroxetine (n=87) | Sertraline (n=307) | Topiramate (n=96) | Venlafaxine (n=116) |
|---|---|---|---|---|
| 58.5 (11.7) | 58.4 (11.3) | 57.5 (10.9) | 58.3 (15.0) | 58.6 (11.8) |
| −8.1 (15.0) | −6.8 (16.0) | −7.5 (12.2) | −6.8 (14.0) | −10.1 (19.5) |
| 2.9 (8) | 0.0 (0) | 4.7 (14) | 1.3 (2) | 10.9 (9) |
| 14.4 (3.6) | 14.4 (3.6) | 14.3 (3.8) | 15.1 (3.5) | 14.3 (3.9) |
| −1.8 (4.2) | −1.6 (4.9) | −1.8 (3.6) | −1.8 (4.5) | −2.1 (5.2) |
| 6.5 (1.5) | 6.3 (1.8) | 6.3 (1.6) | 6.4 (1.8) | 6.2 (1.9) |
| −1.1 (2.4) | −0.8 (2.2) | −0.7 (1.8) | −0.7 (2.9) | −1.0 (2.6) |
| 19.9 (4.6) | 19.5 (5.6) | 19.4 (5.2) | 19.5 (6.0) | 20.6 (5.3) |
| −2.8 (6.2) | −2.4 (6.2) | −2.5 (5.1) | −2.2 (5.5) | −4.2 (7.6) |
| 17.4 (3.9) | 17.9 (3.8) | 17.3 (3.9) | 17.2 (7.1) | 17.2 (4.7) |
| −2.5 (5.2) | −2.1 (5.6) | −2.5 (4.0) | −2.1 (5.1) | −2.8 (6.2) |
| 6.0 (1.8) | 6.1 (1.6) | 6.2 (1.7) | 6.3 (1.9) | 6.1 (1.8) |
| −0.7 (2.1) | −0.8 (2.3) | −1.1 (1.8) | −0.9 (2.3) | −0.9 (2.7) |
Significant Differences are assessed at p<0.05 for the Omnibus comparison, with pair-wise testing where indicated.
Abbreviations. PTSD=Posttraumatic Stress Disorder, PCL=PTSD Checklist, NACM=Negative Alterations in Cognitions and Mood
There was a very limited range of change in PTSD symptom clusters and sleep item scores, with the greatest differences being between change in negative alterations in cognitions and mood for the venlafaxine group versus the other groups (−4.2 for venlafaxine versus −2.2 to −2.8 for the other groups). However, these differences were not statistically significant.
In our weighted survival analysis examining acute psychiatric care use in the six-month continuation phase, there were 57 events. We added a time-varying co-variate for whether patients stayed on their medication in the continuation phase and controlled for change in total PCL during the acute phase in addition to the unbalanced covariates. While we found no difference between medications in acute psychiatric care use during the continuation phase, there was a significant protective effect for medication adherence (HR=0.55; p=0.03).
Discussion
We compared the effectiveness of five evidence-based medications for DSM-5 PTSD and found that they all appear to be effective in routine clinical practice. Furthermore, we found evidence of possible superiority of venlafaxine in achieving acute phase remission. Though there were no between-groups differences in the continuation phase, our findings indicate that medication continuation beyond the initial 12 weeks of treatment is associated with lower risk of acute psychiatric care use such as psychiatric admission. This finding, combined with our finding that patients in all groups experienced a modest level of symptomatic improvement during the acute phase, even after controlling for other important patient and concurrent treatment factors, supports our assertion that these five agents are effective in clinical practice. Our findings are consistent with meta-analytic findings that have suggested that fluoxetine, paroxetine, sertraline, topiramate, and venlafaxine are efficacious treatments for PTSD in RCTs.4,5
We believe this study achieved our goal of improving upon the prior VA study comparing these five agents in routine practice,6 and these changes may have accounted for the differences in findings. Firstly, we achieved far better standardization of treatment duration. While there was a wide range of time between baseline and follow-up symptomatic measurements in the prior VA study (mean length of 254.1 days; SD=119.5 days), we standardized the trial length and achieved a treatment duration that better approximated the typical acute phase clinical trial (mean length of 80.8 days; SD=10.3 days). Thus, we have significantly decreased the heterogeneity of exposure and improved the comparability of our retrospective results with those of prospective studies. Standardizing the acute phase treatment period also allowed us to add a continuation phase and related functional outcome (acute psychiatric services use). Second, we have accounted for prior evidence-based PTSD treatment receipt, including both psychotherapy and medication. We found that 19.0% of patients had previously received an adequate evidence-based medication trial and 6.8% had previously received an adequate evidence-based psychotherapy trial. Measuring this allowed us to account for differing levels of treatment resistance, as patients in the paroxetine, topiramate, and venlafaxine groups were more likely to have received prior adequate evidence-based medication trials. Third, this study used the PCL-5. The PCL-5 represents the current case conceptualization of PTSD. Importantly in this version, the avoidance and numbing items were split into separate clusters, and additional items have been added to the prior numbing items to make the new negative alterations in cognitions and mood cluster. While the finding was not statistically significant, it was notable that patients in the venlafaxine group had the greatest magnitude of change in the negative alterations in cognitions and mood cluster. These symptoms were emphasized in the transition between DSM-IV and DSM-5 definitions of PTSD. Thus, better standardization of treatment duration, an ability to account for prior treatment resistance, and changes in the PTSD case definition may have all contributed to our finding of possible acute phase superiority for venlafaxine over other agents.
While we found that all the medication treatments for PTSD that we studied were effective in clinical practice, their effect seemed somewhat reduced compared to that seen in the clinical trials. Such comparisons are difficult to make precisely in all cases because various studies use different measures and allowed various concurrent treatments. However, as an example, Rauch et al.’s recent four-site VA and private sector study of PTSD treatment approaches for Iraq and Afghanistan Veterans included a sertraline plus enhanced medication management arm.21 Enhanced medication management consisted of eight manualized 30-minute appointments over the first 12 weeks for those randomized to sertraline.22 Sessions included psychoeducation and support from prescribing clinicians. Participants experienced a decrease PTSD symptom severity from 56.2 to 42.8 on the version of the PCL corresponding to DSM-IV over the first 12 weeks. This translates to an approximately 15-point improvement on the PCL-5,23 and compares to a 7.5-point PCL-5 improvement in our sertraline group. The reasons for possible reduction in effectiveness are unknown. One possibility is that that enhanced medication management practices are uncommon in routine practice, but as in psychotherapy, manualization may be required to obtain maximum benefit from psychopharmacologic approaches to treat PTSD.
There are several major limitations to our study, all of which are inherent to our uncontrolled, retrospective cohort design. First, participants meeting PCL-based inclusion criteria for our analytic cohort differed significantly from those receiving adequate medication trials without PCL measurement in several ways. Most notably, those with aligned PCL measurement received far more psychotherapy. The limited availability of PCL data indicates low use of measurement-based care (MBC) in routine psychopharmacology practice, despite a VA initiative to promote MBC starting in 2016.24 Low use of MBC indicates an emerging quality problem,25 as proactive measurement-driven approaches to psychotropic prescribing are associated with superior clinical outcomes.26 Moreover, we have no clear understanding of whether these findings would apply to non-veterans with PTSD. Second, we were unable to measure all related aspects of care. As an example, we could not measure medication adherence or psychotherapy protocols that are less frequently in the VA such as EMDR. However, while patients could have met our prescribing standard with a single 90-day initial supply, 77.3% of patients finished their initial supply and requested refills. Lastly, we only considered PTSD outcomes and acute psychiatric care use, with PTSD outcomes based on a self-report measure. Depression and quality of life measures were not available, but they may have enriched our exclusive focus on PTSD outcomes.
We conclude that there may be an advantage to venlafaxine over other established medications in achieving acute-phase remission for DSM-5 PTSD routine clinical practice. However, additional prospective research is needed to confirm this result. Regardless of the agent chosen, medication cessation during the continuation phase is associated with a higher risk of acute psychiatric care use. Our study lacks adequate sample size to adequately address issues regarding either specific medication effects on specific symptoms or patient characteristics that predict response with a particular medication. These are both fertile areas for future research.
Supplementary Material
Clinical Points:
Five medications for PTSD with consistent efficacy in metaanalyses of randomized controlled trials—including fluoxetine, paroxetine, sertraline, topiramate, and venlafaxine—are also effective in routine clinical practice.
It appears that venlafaxine may have superior effects in helping patients achieve acute phase (12-week) remission, but this requires further study.
Regardless of which agent is used, medication adherence in the continuation phase (subsequent 6 months) is associated with superior functioning, as indicated by less use of acute psychiatric services.
Additional Information:
The VA Corporate Data Warehouse (CDW) contains electronic medical record data compiled from individual VA facilities and is described at http://www.hsrd.research.va.gov/for_researchers/vinci/cdw.cfm. Data are stored on geographically dispersed server farms. To access the CDW, researchers generally need to have an employment relationship with the VA. After local institutional review board approval, requests for data are submitted to VA National Data Systems using the Data Access Request Tracker. Datasets are then built and analyzed in secure virtual project workspaces within the VA Informatics and Computing Infrastructure environment. Researchers with VA network access can obtain descriptions of CDW data at http://vaww.virec.research.va.gov/.
Podcast Text:
In this study, which used the treatment records of all patients treated in the Veterans Health Administration, authors determined that five medications that had been shown effective to treat PTSD in research studies also appear to work in real world clinical use. The medications fluoxetine, paroxetine, sertraline, topiramate, and venlafaxine all led to improvements in PTSD symptoms during the 12-week acute treatment phase and improved functioning in the 6-month continuation phase. Patients in the venlafaxine group were more likely to experience remission of their PTSD symptoms during the acute phase, compared to patients in the other treatment groups. However, the possible superiority of venlafaxine over other established medications in achieving acute-phase remission for DSM-5 PTSD should be evaluated with additional prospective research.
Disclosures and acknowledgements:
The authors report no conflicts of interest.
This study was funded by the Department of Defense Peer Reviewed Medical Research Program (PR160203), Congressionally Directed Medical Research Program, Fort Dietrich, MD (Dr. Shiner) as well as a Department of Veterans Affairs Health Services Research and Development Career Development Award (CDA11-263), VA Office of Research and Development, Washington, DC (Dr. Shiner). The sponsors did not have any role in the study design, methods, analysis, and interpretation of results, or in preparation of the manuscript and the decision to submit it for publication. The views expressed in this article are those of the authors and do not necessarily represent the position or policy of the US Department of Veterans Affairs or US Department of Defense.
Footnotes
Previous Presentations: None
Contributor Information
Brian R. Shiner, National Center for PTSD and Veterans Affairs Medical Center, White River Junction, VT;; Departments of Psychiatry and The Dartmouth Institute for Health Policy & Clinical Practice, Geisel School of Medicine at Dartmouth, Hanover, NH
Christine E. Leonard, Veterans Affairs Medical Center, White River Junction, VT.
Jiang Gui, Veterans Affairs Medical Center, White River Junction, VT;; Departments of Biomedical Data Science, Community & Family Medicine, and The Dartmouth Institute for Health Policy & Clinical Practice, Geisel School of Medicine at Dartmouth, Hanover, NH
Sarah Cornelius, Veterans Affairs Medical Center, White River Junction VT..
Paula P. Schnurr, National Center for PTSD, White River Junction, VT; Department of Psychiatry, Geisel School of Medicine at Dartmouth, Hanover, NH.
Jessica E. Hoyt, Veterans Affairs Medical Center, White River Junction, VT.
Yinong Young-Xu, Clinical Epidemiology Program and National Center for Patient Safety, White River Junction VT;; Departments of Epidemiology and Psychiatry, Geisel School of Medicine at Dartmouth, Hanover, NH
Bradley V. Watts, Office of Systems Redesign and Veterans Affairs Medical Center, White River Junction, VT,; Department of Psychiatry, Geisel School of Medicine at Dartmouth, Hanover, NH
REFERENCES
- 1.American Psychiatric Association. Diagnostic and statistical manual of mental disorders : DSM-5. 5th ed Arlington, VA: American Psychiatric Association; 2013. [Google Scholar]
- 2.Goldstein RB, Smith SM, Chou SP, et al. The epidemiology of DSM-5 posttraumatic stress disorder in the United States: results from the National Epidemiologic Survey on Alcohol and Related Conditions-III. Soc Psychiatry Psychiatr Epidemiol. 2016;51(8):1137–1148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Greenberg G, Hoff R. 2016 Veterans with PTSD data sheet: National, VISN, and VAMC tables. West Haven, CT: Northeast Program Evaluation Center; 2016. [Google Scholar]
- 4.Jonas DE, Cusack K, Forneris CA, et al. In: Psychological and Pharmacological Treatments for Adults With Posttraumatic Stress Disorder (PTSD). Rockville (MD): Agency for Healthcare Research and Quality; 2013. [PubMed] [Google Scholar]
- 5.Watts BV, Schnurr PP, Mayo L, Young-Xu Y, Weeks WB, Friedman MJ. Meta-analysis of the efficacy of treatments for posttraumatic stress disorder. J Clin Psychiatry. 2013;74(6):e541–550. [DOI] [PubMed] [Google Scholar]
- 6.Shiner B, Westgate CL, Gui J, et al. A Retrospective Comparative Effectiveness Study of Medications for Posttraumatic Stress Disorder in Routine Practice. J Clin Psychiatry. 2018;79(5). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision: DSM-IV-TR. Arlington, VA: American Psychiatric Association; 2000. [Google Scholar]
- 8.Friedman MJ, Kilpatrick DG, Schnurr PP, Weathers FW. Correcting Misconceptions About the Diagnostic Criteria for Posttraumatic Stress Disorder in DSM-5. JAMA Psychiatry. 2016;73(7):753–754. [DOI] [PubMed] [Google Scholar]
- 9.Guina J Changes to the Definition of Posttraumatic Stress Disorder in the DSM-5. JAMA Psychiatry. 2016;73(11):1201–1202. [DOI] [PubMed] [Google Scholar]
- 10.Hoge CW, Yehuda R, Castro CA, et al. Unintended Consequences of Changing the Definition of Posttraumatic Stress Disorder in DSM-5: Critique and Call for Action. JAMA Psychiatry. 2016;73(7):750–752. [DOI] [PubMed] [Google Scholar]
- 11.Blevins CA, Weathers FW, Davis MT, Witte TK, Domino JL. The PTSD Checklist for DSM-5 (PCL-5): Development and Initial Psychometric Evaluation. J Trauma Stress. 2015;28(6):489–498. [DOI] [PubMed] [Google Scholar]
- 12.Bovin MJ, Marx BP, Weathers FW, et al. Psychometric properties of the PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders-Fifth Edition (PCL-5) in veterans. Psychol Assess. 2016;28(11):1379–1391. [DOI] [PubMed] [Google Scholar]
- 13.Schnurr PP, Chard KM, Ruzek JI, et al. Design of VA Cooperative Study #591: CERV-PTSD, comparative effectiveness research in veterans with PTSD. Contemp Clin Trials. 2015;41:75–84. [DOI] [PubMed] [Google Scholar]
- 14.Maguen S, Madden E, Patterson OV, et al. Measuring Use of Evidence Based Psychotherapy for Posttraumatic Stress Disorder in a Large National Healthcare System. Adm Policy Ment Health. 2018;45(4):519–529. [DOI] [PubMed] [Google Scholar]
- 15.Ridgeway G, McCaffrey DF, Morral AR, Burgette LF, Griffin BA. Toolkit for Weghting and Analysis of Nonequivalent Groups: A Tutorial for the TWANG Package. Santa Monica, CA: RAND Corporation; 2017. [Google Scholar]
- 16.Stuart EA. Matching methods for casual inference: a review and a look forward. Statistical Science. 2010;25(1):1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.McCaffrey DF, Griffin BA, Almirall D, Slaughter ME, Ramchand R, Burgette LF. A tutorial on propensity score estimation for multiple treatments using generalized boosted models. Stat Med. 2013;32(19):3388–3414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Haneuse S, VanderWeele TJ, Arterburn D. Using the E-Value to Assess the Potential Effect of Unmeasured Confounding in Observational Studies. Jama. 2019;321(6):602–603. [DOI] [PubMed] [Google Scholar]
- 19.VanderWeele TJ, Ding P. Sensitivity Analysis in Observational Research: Introducing the E-Value. Ann Intern Med. 2017;167(4):268–274. [DOI] [PubMed] [Google Scholar]
- 20.Mathur MB, Ding P, VanderWeele TJ. Package ‘EValue’: Sensitivity Analyses for Unmeasured Confounding in Observational Studies and Meta-Analyses. https://cran.r-project.org/web/packages/EValue/EValue.pdf. Published 2018. Accessed March 11, 2019.
- 21.Rauch SAM, Kim HM, Powell C, et al. Efficacy of Prolonged Exposure Therapy, Sertraline Hydrochloride, and Their Combination Among Combat Veterans With Posttraumatic Stress Disorder: A Randomized Clinical Trial. JAMA Psychiatry. 2019;76(2):117–126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Rauch SAM, Simon NM, Kim HM, et al. Integrating biological treatment mechanisms into randomized clinical trials: Design of PROGrESS (PROlonGed ExpoSure and Sertraline Trial). Contemp Clin Trials. 2018;64:128–138. [DOI] [PubMed] [Google Scholar]
- 23.Moshier SJ, Lee DJ, Bovin MJ, et al. An Empirical Crosswalk for the PTSD Checklist: Translating DSM-IV to DSM-5 Using a Veteran Sample. J Trauma Stress. 2019;32(5):799–805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Resnick SG, Hoff RA. Observations from the national implementation of Measurement Based Care in Mental Health in the Department of Veterans Affairs. Psychol Serv. 2019. [DOI] [PubMed] [Google Scholar]
- 25.Muir HJ, Coyne AE, Morrison NR, Boswell JF, Constantino MJ. Ethical implications of routine outcomes monitoring for patients, psychotherapists, and mental health care systems. Psychotherapy (Chic). 2019;56(4):459–469. [DOI] [PubMed] [Google Scholar]
- 26.Fortney JC, Unutzer J, Wrenn G, et al. A Tipping Point for Measurement-Based Care. Psychiatr Serv. 2017;68(2):179–188. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.