You are here

Brexanolone: A Novel Treatment Option For Postpartum Depression

Clayton English PharmD, BCPS, BCPP, BCGP
Heather Goodwin PharmD, BA, MS
Jaden A. Dickinson BSBC, PharmD candidate, Class of 2020
Jose A. Rey MS, PharmD, BCPP

INTRODUCTION

Maternal mental health is an area of health care that continues to pose substantial challenges. Lack of clinical studies, difficulties initiating pharmacotherapeutic interventions, and concerns regarding the safety of medications during the perinatal period limit treatment options for many patients in need. It is estimated by the World Health Organization that 10% of pregnant women and 13% of postnatal women are impacted by a mental health disorder, primarily a mood disorder.1 Major postpartum depression (PPD) is a serious mood disorder affecting 13% of women, with the clinical onset of depressive symptoms occurring during the perinatal period or four weeks following delivery.2 The Diagnostic and Statistical Manual of Disorders, Fifth Edition: DSM-5, identifies PPD as a clinical specifier under the category of depressive disorders entitled “peripartum onset.”3 Diagnostic criteria for PPD are similar to those for major depressive disorder (MDD) (Table 1). Timing of the onset of symptoms is the major difference in diagnostic criteria between PPD and MDD. A majority of women with PPD will begin experiencing symptoms within the first month of delivery and up to 90 days after. 

Table 1 DSM-5 Diagnostic Criteria for Major Depressive Disorder3
For a diagnosis of major depressive disorder, 5 or more of the following symptoms must be present nearly every day for at least a 2-week period (at least one of the symptoms is either depressed mood or anhedonia). Symptoms must impair functioning and cannot be secondary to another drug or illness.
  • Depressed mood most of the time and on most days (e.g., feels sad, appears tearful)
  • Anhedonia (e.g., decreased interest or pleasure in activities)
  • Significant change in weight (e.g., weight loss, weight gain)
  • Sleep disturbances (e.g., insomnia, hypersomnia)
  • Psychomotor fluctuations (e.g., psychomotor agitation or retardation)
  • Changes in energy levels (e.g., fatigue, loss of energy)
  • Excessive guilt or feelings of worthlessness
  • Difficulties concentrating
  • Recurrent thoughts of death or suicidal thoughts (e.g., active or passive suicidal ideation with or without a plan)
Peripartum onset specifier: Onset of symptoms occurs during pregnancy or within 4 weeks after childbirth. Episodes can present with psychotic features.

PPD is often confused with “baby blues.” Unlike postpartum “baby blues,” which typically presents with mood lability and generally resolves with self-care after two weeks post-delivery, symptoms of PPD are more severe.2 A small minority of patients with “baby blues” progress to PPD. The severe symptoms of PPD can leave a woman with feelings of extreme sadness, produce feelings of obsessiveness, potentially lead to psychotic symptoms, and impair ability to care for family. Risk factors for PPD include stress, previous occurrences of depression unrelated to childbirth, and family history.2 Symptoms most likely stem from the rapid fluctuation of peripartum reproductive hormones estrogen and progesterone post-delivery.4 Changes in sleep patterns brought on by motherhood could lead to continuous sleep deprivation and further worsen symptoms. Untreated symptoms of PPD may lead to maternal self-neglect, which can impact infant care, maternal-infant bonding, and lead to poor developmental outcomes for the infant.1 Additionally, the acute risk of suicidality and potentially infanticide stresses the importance of appropriate screening so the condition can be appropriately treated to reduce long-term sequelae to the infant and mother.1-4 

Historically, antidepressants including selective serotonin reuptake inhibitors (SSRIs) and cognitive behavioral therapy have been utilized to treat PPD.4 Pharmacological therapies are warranted for those experiencing severe symptoms, patients desiring antidepressants, or where cognitive behavioral therapy alone is ineffective. SSRIs that pose low breastfeeding risks (e.g., sertraline) have classically been the treatment of choice for depression occurring during the perinatal time frame. A few small studies have demonstrated positive efficacy and safety outcomes of using SSRIs for the treatment of PPD; however, SSRIs are not currently approved for this indication.5,6 On March 19, 2019, brexanolone (Zulresso, Sage Therapeutics) became the first drug approved by the Food and Drug Administration specifically indicated to treat PPD in adults. Acting as a neuroactive steroid on the gamma-aminobutyric acid (GABA) system rather than modulating monoamines, brexanolone represents a novel option for the treatment of depression with peripartum onset. 

Pharmacology

Brexanolone is an intravenous (IV) formulation of the endogenous steroid allopregnanolone, a metabolite of progesterone, and betadex sulfobutyl ether sodium, a solubilizing agent.7 Although the mechanism of action in PPD is not completely understood, brexanolone acts as a positive allosteric modulator of the GABAA receptor. It is not well understood how the receptor affects mood symptoms; however, other mechanisms have been proposed for its clinical effects. For example, endogenous allopregnanolone levels increase throughout pregnancy, with levels reaching maximal concentrations during the third trimester. After delivery of the infant, levels of allopregnanolone drop rapidly.8 The rapid decline in serum endogenous allopregnanolone has been hypothesized to contribute to the development of PPD. 

Pharmacokinetics

Brexanolone is a continuous IV infusion administered over 60 hours demonstrating complete bioavailability and dose-proportional pharmacokinetics. The area under the curve of brexanolone when administered at 60 and 90 mcg/kg/hour, was 52 ng/mL and 79 ng/mL, respectively. There is 99% protein binding independent of plasma concentrations and a sizable volume of distribution of 3 L/kg. No cytochrome P450 metabolism occurs, but three inactive metabolites are produced through metabolism by keto-reduction, sulfation, and glucuronidation. These metabolites make up 47% of the drug that is eliminated via the feces, while 42% is excreted practically unchanged by the kidneys. The half-life of the medication is approximately nine hours. Pharmacokinetic parameters in special populations are provided in Table 2.7

Table 2 Use in Special Populations7
Pregnancy No available human data regarding risks. Doses above recommended doses in animals resulted in developmental toxicities. Not recommended in pregnancy.
Lactation Brexanolone is excreted in breast milk at small concentrations. The relative infant dose is 1%-2% and infant exposure is expected to be low. Available data are limited.
Pediatric Use Efficacy and safety not established.
Geriatric Use Efficacy and safety not established.
Hepatic Impairment No dose modifications necessary.
Renal Impairment No dose adjustments needed for patients with mild, moderate, or severe renal impairment. Avoid use in patients with end-stage renal disease due to substantial accumulation of the solubilizing agent, betadex sulfobutyl ether sodium, increasing the Cmax 1.7-fold.
Cmax = Maximal concentration
Chemical Properties

Brexanolone is undistinguishable from allopregnanolone, an endogenous steroid metabolite of progesterone. Its chemical and empirical formulas are 3-hydroxy-5-pregnan-20-one and C21H34O2, respectively. It is supplied as a sterile, colorless solution for injection and includes a solubilizing agent, betadex sulfobutyl ether sodium. Other inactive ingredients include citric acid monohydrate, sodium citrate dihydrate, and water. Hydrochloric acid and sodium hydroxide are occasionally in the product for pH balancing.7 The medication’s structural formula is illustrated in Figure 1.

Figure 1

CLINICAL EFFICACY AND SAFETY

Efficacy

The efficacy and safety of brexanolone were assessed in three double-blind randomized placebo-controlled trials. Studies 202A, 202B, and 202C were conducted independently; however, as the studies were nearly identical in design, the data were pooled prior to analysis and submission to the FDA.9 Study participants were included if they were between the ages of 18 and 45, were within six months postpartum, had a DSM-5 verified major depressive episode at the onset of the third trimester through four weeks postpartum, and were agreeable to temporarily discontinuing breastfeeding.10,11 A total Hamilton Rating Scale for Depression (HAM-D) ≥26 was required prior to brexanolone dosing in studies 202A and 202B,11 and scores between 20 and 25 were required for study 202C.10 Patients with dialysis requirements, anemia, a known allergy to allopregnanolone or progesterone, schizophrenia, bipolar disorder, or schizoaffective disorder were excluded from participation.10,11

In all three trials, study participants were randomly assigned to the recommended 60-hour continuous IV infusion of the 90 mcg dose per kg regimen of brexanolone (BRX90) or placebo.10,11 Study 202B also included a 60 mcg dose regimen (BRX60), in which the 60 mcg/kg/hour rate was continued through hour 52 instead of increasing to 90 mcg/kg/hour at hour 24.11 The primary efficacy endpoint was the change from baseline HAM-D total score at the end of the infusion (hour 60). For studies 202B and 202C, the key secondary outcome was durability of response, defined as the change from baseline HAM-D total score at day 30.9-11 Additional secondary endpoints included change in HAM-D individual item scores and HAM-D scores over time, HAM-D response (50% reduction from baseline), HAM-D remission (total scores ≤7), and clinical global impression improvement (CGI-I) responses.9 Studies 202A, 202B, and 202C enrolled 21, 122, and 104 participants respectively, for a total of 247 participants. Twenty additional patients were randomized but later withdrawn due to personal reasons (n = 10) or inability to meet the necessary timing requirements between randomization and dosing (n = 10).9

In the pooled analysis, 35 patients were randomized to BRX60, 94 to BRX90, and 105 to placebo. The demographic characteristics were similar between the brexanolone and placebo treatment arms, in both the individual studies and pooled data set. Study participants were racially and ethnically diverse with an average age of 28 years.9 The mean baseline HAM-D score was 26 for both the pooled BRX90 and placebo groups. The mean Edinburgh Postnatal Depression Scale (EPDS) score was 20 for both treatment arms.9 The onset of postpartum depression occurred within four weeks of delivery in most study participants (78.4% BRX90 versus 72% placebo), with participants entering the study around 3.5 months postpartum, on average. Of note, 22 (21.6%) of the women in the BRX90 arm and 25 (23.4%) in the placebo arm utilized antidepressants at baseline.9 

For the primary efficacy endpoint, each individual study demonstrated statistically significantly greater mean reductions in baseline HAM-D total score at hour 60 for the BRX90 treatment arm compared to placebo. The mean differences between BRX90 and the placebo treatment were –12.2 (P = .008), –3.7 P = .02), and –2.5 (P = .02) for studies 202A, 202B, and 202C, respectively.9-11 Interestingly, study participants had variable responses to placebo treatments, with mean HAM-D reductions of –8.8, –14.0, and –12.1, respectively.9 The BRX60 treatment arm also demonstrated statistically significantly greater reductions in mean HAM-D total scores than placebo, with a mean difference of –5.5 (P = .001).11 

For the key secondary endpoint, statistically significantly greater mean reductions in HAM-D total scores were sustained at day 30 from the BRX90 treatment arm compared with placebo in two of the clinical trials. The mean difference in HAM-D total scores between the two treatment arms was 11.9 (P = .01) and –3.8 (P = .048) for studies 202A and 202B, respectively.9,11 For study 202C, the mean difference was 0.5, although not statistically significant (P = .67).10 The placebo response was also sustained at day 30 with mean HAM-D reductions of –8.8, –13.8, and –15.23 in studies 202A, 202B, and 202C, respectively.9-11 For the other clinically relevant secondary endpoints, a greater percentage of patients in the BRX90 treatment arm achieved HAM-D remission at hour 60 (70% vs. 9%; 31% vs. 16%; 61% vs. 39%).9 Additionally, a larger percentage of participants in the BRX90 treatment arms achieved CGI-I response scores of “very much improved” or “much improved” when compared to placebo (80% vs. 36%, study 202A; 82% vs. 56%, study 202B; 80% vs. 56%, study 202C).9

Safety and Adverse Events

The safety population included 140 patients who received brexanolone and 107 patients who received placebo.9 The incidence of adverse effects was nearly identical between the two treatment arms (brexanolone, 50.0%; placebo, 50.5%).9
A dose interruption or reduction was required in 10 participants receiving brexanolone compared to three in the placebo group. Two (1.4%) patients in the brexanolone treatment arm experienced serious adverse events, which included suicidal ideation, syncope, and loss of consciousness.9 Sedation-related events occurred in 27.1% of participants who received brexanolone compared to 14.0% who received placebo.9 Dizziness and somnolence were the sedation-related adverse events that were reported at least 5% more frequently than placebo.11 Sedation occurred in 15%, dizziness in 13.6%, fatigue in 3.6%, loss of consciousness in 4.3%, and amnesia in 0.7%.9 Concurrent use of antidepressants and benzodiazepines was associated with an additive risk of sedation-related adverse events.9 Among the six patients who lost consciousness, one case was attributed to fear of needles and two to infusion pump malfunctions, resulting in medication overdose.9 

WARNINGS AND PRECAUTIONS

Brexanolone has a black box warning for excessive sedation and sudden loss of consciousness and, because of these risks, is available only through a restricted program entitled Zulresso REMS (see Formulary Considerations below for details). Based on the clinical trials, alteration or loss of consciousness was seen in 4% of patients treated with brexanolone compared to 0% treated with placebo. Interrupting the infusion led to recovery in all patients within an hour after stopping the infusion. No correlations were identified with changes in sensorium and timing of dose. Additionally, not all patients who lost consciousness conveyed feeling somnolent beforehand. Due to this risk, patients require staff monitoring while interacting with their children and must be monitored every two hours during prearranged, non-sleep periods. Continuous pulse oximetry must be utilized during the course of treatment to alert staff to hypoxic episodes. For excessive sedation, interrupting the infusion can ameliorate the symptoms, and once the symptoms subside the infusion can be restarted at the same or lower dose. If hypoxia develops, the infusion should be interrupted and brexanolone should not be rechallenged.7 

All antidepressants, including brexanolone, have a warning for an increased risk of suicidal thoughts and behaviors. The risk of suicidality has not been fully elucidated based on the current clinical trial data with brexanolone; however, due to the difference in mechanism of action it cannot be assumed that brexanolone impacts the development of suicidal thoughts like traditional antidepressants (e.g., SSRIs). Suicidal ideation was evaluated in the clinical trial series using the Columbia Suicide Severity Rating Scale (C-SSRS). When compared to placebo, there was no clinical worsening of suicidal ideation in patients receiving brexanolone.9 One study participant in the brexanolone treatment arm reported suicidal behavior after completion of the infusion and one reported non-suicidal self-injurious behavior.9 It is recommended by the manufacturer to consider discontinuing or altering therapy if suicidal thoughts emerge during the course of treatment.7

DRUG INTERACTIONS

Brexanolone’s ability to modulate the GABAA receptor makes it susceptible to pharmacodynamic drug-drug interactions with other sedatives. Benzodiazepines, barbiturates, and selective GABAA positive allosteric modulators (e.g., zolpidem, eszopiclone) all modulate the GABAA receptor. The sedative effects of these agents can be additive when concomitantly administered with brexanolone. Opioids and other CNS depressants should be used cautiously when combined with brexanolone due to the risk of excessive sedation associated with these drugs. Patients receiving antidepressants concurrently with brexanolone had higher rates of sedation-related adverse effects.7 The mechanism behind this interaction is not understood. Brexanolone is metabolized by multiple non-CYP pathways; therefore, pharmacokinetic interactions are less likely
to occur. 

DOSAGE AND ADMINISTRATION

Brexanolone is administered as a continuous infusion over 2.5 days under the supervision of a health care provider to ensure appropriate monitoring for excessive sedation. The delivery setting for brexanolone requires the ability to continuously monitor patients for excessive sedation and hypoxia and to intervene as necessary. Patients should be monitored for excessive sedation every two hours during awake periods while receiving brexanolone, and continuous pulse oximetry should be used with alarm settings to monitor hypoxia. Due to the potential for excessive sedation, the administration of brexanolone should begin early in the day to provide ample opportunities to assess sedation and modify the dosing
regimen.

The dose of brexanolone is titrated to a maximal rate of 90 mcg/kg/hour and then deescalated prior to discontinuation. The recommended dose initiation and titration for brexanolone is provided in Table 3.7 For patients who are unable to tolerate the maximal infusion rate, a dose decrease to 60 mcg/kg/hour should be considered. The infusion should be stopped if excessive sedation occurs at any point and can be resumed at a lower dose. 

Table 3 Dose Initiation and Titration of Brexanolone7
Hours: 0 to 4 30 mcg/kg/hour
Hours: 4 to 24 60 mcg/kg/hour
Hours: 24 to 52 90 mcg/kg/hour
Hours: 52 to 56 60 mcg/kg/hour
Hours: 56 to 60 30 mcg/kg/hour

Brexanolone is supplied in a 100 mg/20 mL single-dose vial requiring dilution and is prepared in an infusion bag at a concentration of 1 mg/mL. Infusion bags can be stored in the refrigerator for up to 4 days; however, infusion bags will expire after 12 hours at room temperature. Due to the faster expiration at room temperature, preparation of multiple infusion bags is required. For the majority of patients, five infusion bags will be required for the 60-hour infusion and additional bags will be needed for patients weighing over 90 kg. Dose modifications and treatment parameters for special populations are provided in Table 2. Detailed instructions guiding the preparation and delivery of brexanolone can be found in the package
insert.

Cost

The exact cost of administering brexanolone is unknown and will depend on a number of factors including insurance coverage, inpatient reimbursement models, health-care system contracts, delivery setting, and the weight of the patient. The cost of five vials of brexanolone is $34,000.12 Requirements for delivery in a registered health care facility and staffing needed to ensure monitoring and supervision will contribute to additional indirect costs on top of the direct cost of the medication. The administration of other antidepressants used for PPD are less expensive, with the average cost of an SSRI ranging from $2.90 to $6.50 for a 30-day supply.13 The low cost of traditional antidepressants is offset by their slower onset of action. Pharmacoeconomic studies and head-to-head studies with other antidepressants demonstrating superior efficacy and cost-effectiveness are needed to help further delineate brexanolone’s role in the treatment of depression compared to other agents due to the high cost of administration. 

Formulary Considerations

Staffing requirements, available space within the health care facility, and cost should all be considered prior to the addition of brexanolone to formulary. Patients must be hospitalized for the duration of the 60-hour infusion, which may be challenging depending on the available space within the facility.9 If patient rooms within the mother–baby units are not available, and the organization cannot accommodate children outside of these units, disruptions in family life and bonding may occur. Should health care organizations choose to expand their facilities to accommodate brexanolone administration, they need to consider the time and budget required for construction.

Due to the risk of sedation-related events, brexanolone was approved with a Risk Evaluation and Mitigation Strategy (REMS).7,12 This restricted access program ensures that brexanolone is available only at certified health care facilities that can provide continuous monitoring. In order to become certified, the dispensing location must demonstrate the ability to monitor patients continuously for the duration of the infusion, have access to continuous pulse oximetry and programmable infusion pumps with alarm abilities, develop a fall precaution protocol, and designate an authorized representative to oversee the implementation and compliance with the REMS, as well as patient registration within the program.7,12 Aside from the continuous pulse oximetry monitoring, patients must be assessed for signs and symptoms of excessive sedation and loss of consciousness every two hours during treatment.7 The staffing requirements for both the dispensing certification and patient monitoring may be a barrier to health care facilities with limited employee resources. 

Last, the cost of brexanolone may be prohibitive for patients who require treatment. The listed cost of brexanolone is approximately $34,000 for a standard-size patient and it is unclear if insurance companies will cover the infusion and required hospitalization.12 Although clinical trial data have demonstrated efficacy 30-days post-infusion, no data are available to assess durability of response beyond this. If subsequent infusions are required, access to brexanolone will likely be restricted due to the cost and logistics of administration.

Regardless of these formulary considerations, it is important to recognize that PPD is a serious medical condition that can have lasting consequences for mothers and children. The available data for brexanolone demonstrated rapid efficacy at 60 hours, with continued response approximately one month after administration. While it is unknown if brexanolone therapy should replace the current standards of care (e.g., SSRIs, nonpharmacological management), it is important to consider the severity of depression and the consequences if it is not treated urgently when deciding if brexanolone is an appropriate treatment option.

CONCLUSION

Rapid-acting antidepressants are a heavily desired treatment option for both patients and providers. Prior to 2019, medications for the treatment of depressive disorders have taken weeks to demonstrate clinical improvement and have required a level of patience from all parties. Depression occurring during the perinatal time frame has been an area where research has fallen behind, due to the difficulties of performing clinical trials within this special population. 

Brexanolone appears to provide clinically important short-term symptom improvement for patients with PPD. The rapid onset of action of brexanolone makes it an attractive therapy for patients with severe PPD, especially those requiring hospitalization. Although well-tolerated, brexanolone comes with substantial monitoring requirements for the small, albeit, serious risk of causing excessive sedation potentially leading to unconsciousness. The REMS will mitigate this risk, but implementing the requirements for the REMS will pose challenges to health care facilities. Finding appropriate space and staff to execute the requirements of the REMS is just one barrier in addition to the high direct cost of the drug. Additionally, many questions regarding the therapy remain unanswered. There are currently no data available to assess durability of response beyond 30 days of treatment, and clinical management for patients who have recurrent symptoms post-infusion is a gray area. It is also important to note that SSRIs have clinical evidence for treating PPD, although the evidence is small.5,6 Additionally, SSRIs have demonstrated efficacy in preventing relapses of depressed episodes in the non-perinatal population and are less expensive. Although their onset of action is slower, SSRIs will still play a role in the management of PPD, and more data are needed to determine preference of treatment by both clinicians and patients. Comparative studies assessing cost and clinical outcomes between monoaminergic antidepressants and brexanolone will further delineate the place in therapy of both drug classes. 

With the paucity of evidence for treating PPD, brexanolone clearly demonstrates some advantages in treating the condition, especially in regard to its onset of action. The finer details regarding placement of therapy will be better defined with time and post-marketing studies. Due to the level of oversight and monitoring required for brexanolone, discussions regarding the utility of using brexanolone may be beyond the scope of a hospital’s P&T committee and will likely require coordination at higher administrative levels to ensure adequate space and staff are available for delivery. If institutions can meet the REMS requirements, brexanolone is a viable option for patients requiring urgent treatment of PPD.

Author bio: 

Clayton English is an Associate Professor, Department of Pharmacy Practice, at the Albany College of Pharmacy and Health Sciences, Vermont campus, in Colchester, Vermont. Heather Goodwin is a Psychiatric Pharmacy Resident at Yale New Haven Hospital in New Haven, Connecticut. Jaden A. Dickinson is a PharmD candidate, Class of 2020, Albany College of Pharmacy and Health Sciences, Vermont campus; and Jose A. Rey is a Professor in the Department of Pharmacy Practice at Nova Southeastern University College of Pharmacy in Davie, Florida.

Clayton English is the corresponding author (Clayton.English@acphs.edu).

References

  1. Maternal Mental Health (online). World Health Organization. Available at: https://www.who.int/mental_health/maternal-child/maternal_mental_health/en/. Accessed April 25, 2019.
  2. Roberts LW. The American Psychiatric Association Publishing Textbook of Psychiatry. 7th ed. Washington, D.C.: American Psychiatric Association Publishing; 2019.
  3. American Psychiatric Association. American Psychiatric Association DSM-5 Task Force. Diagnostic and Statistical Manual of Mental Disorders DSM-5: 5th ed. Washington, DC: American Psychiatric Association; 2013.
  4. Stewart DE, Vigod S. Postpartum depression. N Engl J Med. 2016;375(22):2177-2186.
  5. Bloch M, Meiboom H, Lorberblatt M, et al. The effect of sertraline add-on to brief dynamic psychotherapy for the treatment of postpartum depression: a randomized, double-blind, placebo-controlled study. J Clin Psychiatry. 2012;73(2):235-241.
  6. Yonkers KA, Lin H, Howell HB, et al. Pharmacologic treatment of postpartum women with new-onset major depressive disorder: a randomized controlled trial with paroxetine. J Clin Psychiatry. 2008;69(4):659-665.
  7. Zulresso (brexanolone) [package insert]. Cambridge, MA: Sage Therapeutics, 2019. 
  8. Wilkinson ST, Sanacora G. A new generation of antidepressants: an update on the pharmaceutical pipeline for novel and rapid-acting therapeutics in mood disorders based on glutamate/GABA neurotransmitter systems. Drug Discov Today. 2019;24(2):606-615.
  9. Therapeutics S. Brexanolone Injection, IV Briefing Document. In Psychopharmacologic Drug Advisory Committee and Drug Safety and Drug Safety and Risk Management Meeting; 2018:1-111. https://www.fda.gov/media/121348/download. 
  10. Kanes S, Colquhoun H, Gunduz-Bruce H, et al. Brexanolone (SAGE-547 injection) in post-partum depression: a randomised controlled trial. Lancet. 2017;390(10093):480-489.
  11. Meltzer-Brody S, Colquhoun H, Riesenberg R, et al. Brexanolone injection in post-partum depression: two multicentre, double-blind, randomised, placebo-controlled, phase 3 trials. Lancet. 2018;392(10152):1058-1070.
  12. Brexanolone (Zulresso) for postpartum depression. Med Lett Drugs Ther. 2019;61(1571):68-70.
  13. Truven Health Analytics-Redbook Online. Available from www.micromedex.com. Subscription required. Accessed June 17, 2019.