Contraindications

Mefloquine use is contraindicated in persons with a known hypersensitivity to mefloquine or related compounds (e.g., quinine and quinidine) and to drug-formulation excipients (FDA, 2016). It is also contraindicated for people with current depression, a recent history of depression, generalized anxiety disorder, psychosis, schizophrenia or other major psychiatric disorders, or with a history of convulsions (FDA, 2016).

Although policies are in place to prevent those with a contraindication from being prescribed mefloquine, in practice it still happens. For example, according to an analysis using the UK-based Clinical Practice Research Datalink, from January 2001 through June 2012, 165,218 people had a recorded prescription for an antimalarial for prophylaxis, of whom 25,294 (15.3%) were prescribed mefloquine. People with contraindications to mefloquine were twice as likely to be prescribed a different antimalarial drug, but occasionally people with contraindications were prescribed mefloquine (Bloechliger et al., 2014). However, no additional follow-up or analyses of any reported adverse events were conducted to determine whether those with contraindications were at higher risk or experienced more severe adverse events.

Drug Interactions

The Warnings section of the package insert alerts against using halofantrine or ketoconazole concomitantly or within 15 weeks of the last dose of mefloquine due to risk of sudden cardiac death that can result from prolongation of the QTc interval (FDA, 2016). Co-administration of other drugs that affect cardiac conduction (e.g., anti-arrhythmic or beta-adrenergic blocking agents, calcium channel blockers, antihistamines or H₁-blocking agents, tricyclic antidepressants, and phenothiazines) might also contribute to a prolongation of the QTc interval (FDA, 2016). Administration of mefloquine with related antimalarials (e.g., quinine, quinidine, chloroquine) may produce electrocardiographic abnormalities and increase the risk of convulsions (FDA, 2016).

Taking mefloquine with an anticonvulsant (e.g., valproic acid, carbamazepine, phenobarbital, or phenytoin) may reduce seizure control, and the blood level of anti-seizure medication should be monitored. Moreover, concomitant administration of mefloquine with quinine or chloroquine in addition to an anticonvulsant can further increase the risk of seizures. Taking mefloquine concurrently with oral live typhoid vaccines may make the immunization ineffective. It is recommended that vaccination with live attenuated bacteria be completed at least 3 days before beginning mefloquine. Taking rifampin with mefloquine can decrease mefloquine concentration and elimination time. Mefloquine is metabolized by CYP3A4, and CYP3A4 inhibitors may modify the pharmacokinetics and metabolism of mefloquine and thus increase mefloquine plasma concentrations and the risk of adverse reactions. Similarly, CYP3A4 inducers may decrease mefloquine plasma concentrations and reduce mefloquine efficacy. Mefloquine is a substrate and an inhibitor of P-glycoprotein; thus drug–drug interactions could occur with drugs that are substrates or are known to modify the expression of this transporter, although the clinical relevance of these interactions is not known to date.

Changes to the Mefloquine Package Insert Over Time

There have been multiple important changes to the mefloquine package insert since the drug was first approved for prophylaxis and treatment of malaria in 1989. According to the FDA Center for Drug Evaluation and Research, for a drug to be approved by FDA, at a minimum it must be shown through submitted clinical trials, animal toxicology studies, and other evidence that the drug works as intended and that the health benefits outweigh the known risks (FDA, 2019a). Label changes may indicate that FDA has recognized potential problems with a drug but there may be other reasons for label changes, including approval for a new indication and expansion of the population for which the initial approval was obtained. Most safety-related label changes are the result of spontaneous adverse event reports that have been received during the postmarketing surveillance period, rather than well-designed epidemiologic studies, although if such epidemiologic studies are available they are considered along with new results from pharmacokinetic studies (Sekine et al., 2016). Moreover, the adverse event reports describe events that follow the reported use of the drug, and causality has not necessarily been proven.

Many of the labeling-update letters and package inserts for mefloquine listed on the Drugs@FDA Search website for the period 1989–2002 were unavailable for download. (The downloadable package insert listed with a May 1989 action date is actually a July 2002 revision.) In response to a request for the unavailable information, FDA provided a PDF of the original 1989 package insert as well as abbreviated extractions from editions of the Physicians’Desk Reference but noted that the committee might want to confirm the summary information.¹ In response to the committee’s request for the specific information upon which FDA based mefloquine-labeling changes, FDA stated that it had performed “safety analyses” in 2007, 2013, 2015, and 2016 that supported labeling changes and that the committee could request redacted versions of these reviews via the Freedom of Information Act.² In response to the committee’s request for the information that underlay the addition of the boxed warning to the mefloquine label, FDA referred the committee to the 2013 drug safety communication, a public announcement regarding the boxed warning (FDA, 2013a). The committee had quoted this document in its request to FDA, explaining that it sought more detail than the document provided. The 2013 drug safety communication states: “In conducting its assessment of vestibular adverse reactions associated with mefloquine use, FDA reviewed adverse event reports from the FDA Adverse Event Reporting System and the published literature, identifying patients that reported one or more vestibular symptoms such as dizziness, loss of balance, tinnitus, and vertigo.” It notes further that “Patients who experienced vestibular symptoms usually had concomitant psychiatric symptoms such as anxiety, confusion, paranoia, and depression. Some of the psychiatric symptoms persisted for months to years after mefloquine was discontinued.” As desired details were not provided about the evidence base for the labeling changes (e.g., quantification of adverse reactions reported, epidemiologic data), it was difficult for the committee to assess the implications of the changes.

A comparison of the 1989 package insert and the 2002 package inserts (July and December) showed numerous additions, many pertaining to neurologic and psychiatric adverse events, which were often grouped as “neuropsychiatric” (see Table 4-1 for a summary of the major changes to the package insert over time regarding neuropsychiatric adverse events). The 1989 version stated that “neuropsychiatric reactions have been reported during the use of Lariam” and warned that “if signs of unexplained anxiety, depression, restlessness or confusion are noticed, these may be considered prodromal to a more serious event” and the drug must be discontinued (FDA, 1989). The 2002 package insert added information on neuropsychiatric symptoms to the Contraindications and Warnings sections (FDA, 2002). Mefloquine used as prophylaxis was now contraindicated in persons with current psychiatric problems or a history of psychiatric disorders or convulsions. Symptoms that in 1989 had been listed in the Adverse Reactions’ postmarketing surveillance section as “additional adverse reactions”—vertigo, visual disturbances, and central nervous system disturbances (e.g., psychotic manifestations, hallucinations, confusion, anxiety, and depression)—now appeared in the Warnings section. The Warnings section also stated that psychiatric symptoms “ranging from anxiety, paranoia, and depression to hallucinations and psychotic behavior” had been “reported to continue long after mefloquine has been stopped” (FDA, 1989, 2002). In addition, it noted, “Rare cases of suicidal ideation and suicide have been reported” (FDA, 2002). Cautions were expanded for mefloquine use while performing certain activities, specifically actions requiring alertness and fine motor coordination, “as dizziness, a loss of balance, or other disorders of the central or peripheral nervous system have been reported during and following the use of Lariam” (FDA, 2002). The Adverse Reactions’ postmarketing surveillance section listed as among the most frequently reported adverse events dizziness or vertigo, loss of balance, and neuropsychiatric events such as headache, somnolence, and sleep disorders (insomnia, abnormal dreams) (FDA, 2002). This section also added a lengthy list of “more severe neuropsychiatric disorders” that had “occasionally” been reported (FDA, 2002).

TABLE 4-1

Evolution of Neuropsychiatric Safety-Related Information in the FDA Mefloquine Package Insert and Medication Guide.

The Precautions section now warned, “Hypersensitivity reactions ranging from mild cutaneous events to anaphylaxis cannot be predicted” (FDA, 2002). Users were also advised that contraception should be practiced for up to 3 months after drug cessation and that mefloquine use should be weighed carefully in patients aged ≥65 years since electrocardiographic abnormalities had been observed and cardiac disease is more prevalent in older patients (FDA, 2002).

Updates to information about other body systems included alerts that the concomitant administration of mefloquine and quinine or chloroquine may increase the risk of convulsions. Taking halofantrine after mefloquine might cause potentially fatal prolongation of the QTc interval on electrocardiograms (ECGs); theoretically, the co-administration of other drugs affecting cardiac conduction might also have that effect (FDA, 2002). Previously, users had been informed that if the drug was administered for a prolonged period, periodic evaluations, including liver function tests, should be performed; this language was strengthened to note that in those with impaired liver function, elimination of mefloquine may be prolonged, leading to higher plasma levels (FDA, 2002). The postmarketing surveillance section listed among “infrequent adverse events” cardiovascular, skin, and musculoskeletal disorders as well as “visual disturbances, vestibular disorders including tinnitus and hearing impairment, dyspnea, asthenia, malaise, fatigue, fever, sweating, chills, dyspepsia and loss of appetite.” The two serious adverse reactions reported were cardiopulmonary arrest in one patient shortly after ingesting a single prophylactic dose of mefloquine while using propranolol and, second, encephalopathy of unknown etiology during prophylactic mefloquine administration.

In 2003 FDA required that pharmacists provide a medication guide—a paper handout that conveys risk information that is specific to a particular drug or drug class—to persons to whom mefloquine was dispensed (FDA, 2003a,b). The medication guide included labeled cautions and contraindications and advised users to consult a health care provider in the case of a sudden onset of anxiety, depression, restlessness, or confusion (FDA, 2003b). In 2008 the package insert’s Precautions section added vertigo as a side effect and stated that “in a small number of patients, dizziness and loss of balance have been reported to continue for months after mefloquine has been stopped” (FDA, 2008). The medication guide warned users that they might suddenly experience severe anxiety, paranoia, hallucinations, depression, unusual behavior, and disorientation; “feeling restless” was added to possible side effects. The Adverse Reactions postmarketing surveillance section added respiratory disorders to “infrequent adverse events.” In the 2009 package insert, the Warnings section alerted users against co-administration of halofantrine or ketoconazole with mefloquine; several additions were also made to the Drug Interactions section (FDA, 2009).

In 2013 FDA strengthened and updated warnings of previously included neurologic and psychiatric side effects, and it added a boxed warning, sometimes informally referred to as a “black box” (FDA, 2013b). This is FDA’s most serious type of warning, and it appears on a prescription drug’s label to call attention to serious or life-threatening risks (FDA, 2012). The boxed warning stated, “Mefloquine may cause neuropsychiatric adverse reactions that can persist after mefloquine has been discontinued,” and it added that mefloquine should not be prescribed in patients with major psychiatric disorders and that if psychiatric or neurologic symptoms occur during prophylactic use, drug use should be halted (FDA, 2013b). The Warnings section now informed users that psychiatric symptoms “may occur early in the course of mefloquine use and that in some cases, symptoms have been reported to continue for months or years after mefloquine has been stopped” (FDA, 2013b). It also warned that neurologic effects, including dizziness, vertigo, loss of balance, and ringing in the ears, could occur soon after starting the drug and that they could persist or become permanent. It recommended that evaluations for “neuropsychiatric” effects be performed in persons using the drug long term. Prior language that stated that no relationship had been established between mefloquine and suicide or suicidal thoughts was deleted. Users with impaired liver function were now warned that this placed them at a higher risk of adverse reactions. The Toxicology section included a study in which rats given mefloquine daily for 22 days at levels equivalent to human therapeutic levels showed that mefloquine penetrated the central nervous system, with a 30- to 50-fold greater brain/plasma drug ratio up to 10 days after drug cessation. In the Adverse Reactions postmarketing surveillance section, hepatobiliary disorders, and blood and lymphatic system disorders were added to the “less frequently reported adverse reactions.”

The most recent update to the package insert was made in 2016, and added ocular effects to the Warnings section (FDA, 2016). Regarding adverse events, the package insert states that the most frequently observed adverse event in clinical trials of malaria prophylaxis was vomiting (3%). Dizziness, syncope, extrasystoles, and other complaints were reported in less than 1% of users. Postmarketing surveillance has found that the most frequently reported adverse events are nausea, vomiting, loose stools or diarrhea, abdominal pain, dizziness or vertigo, loss of balance, and “neuropsychiatric” events such as headache, somnolence, and sleep disorders (insomnia, abnormal dreams). These adverse events are often reported without reference to a comparison group, and their duration is rarely detailed.

The Warnings section of the package insert includes several adverse events. It warns that psychiatric symptoms such as acute anxiety, depression, restlessness, or confusion should be viewed as potential precursors to more serious psychiatric or neurologic adverse reactions and that when they occur, mefloquine should be discontinued. More severe neurologic and psychiatric disorders have been reported, including sensory and motor neuropathies (including paresthesia, tremor, and ataxia), convulsions, agitation or restlessness, anxiety, depression, mood swings, panic attacks, memory impairment, confusion, hallucinations, aggression, psychotic or paranoid reactions, and encephalopathy. Suicidal thoughts and suicide have been also been reported. Neurologic symptoms including dizziness or vertigo, tinnitus, hearing loss, and loss of balance have been reported to occur after beginning the drug regimen and in some cases have continued for months, years, or even permanently after discontinuing mefloquine. Users are instructed to discontinue the drug if neurologic symptoms occur, and to use caution when performing activities requiring alertness and fine motor coordination (e.g., driving, piloting aircraft, operating machinery, and deep-sea diving) (FDA, 2016). Other short-term adverse events reported with the use of mefloquine have included transitory and clinically silent ECG alterations such as sinus bradycardia, sinus arrhythmia, first degree atrial–ventricular (AV) block, prolongation of the QTc interval, and abnormal T waves. Eye disorders, including optic neuropathy and retinal disorders, have also been reported during mefloquine use.

United States

Mefloquine was possibly used by the U.S. military as early as 1990³ and by other military forces as early as 1986 (Croft and Geary, 2001). It was used as a first-line prophylactic agent only for deployments to high-malaria-risk areas in sub-Saharan Africa, such as for the Liberian Task Force in 2003, and it was used as a second-line agent in Operation Enduring Freedom (OEF; 2001–2014), Operation Iraqi Freedom (OIF; 2003–2010), and Operation New Dawn (OND; 2010–2011).

In 2003 a Department of Defense (DoD) memorandum on antimalarials was issued by the Armed Forces Epidemiological Board (DoD, 2003). The authors note first that DoD is subject to Section 1107 of Title 10, United States Code, regarding the off-label use of force health protection medications. It then states that this would limit the prescription of Centers for Disease Control and Prevention (CDC)recommended off-label prophylactic regimens (e.g., a loading dose of mefloquine for persons being deployed on short notice) to the context of a doctor–patient relationship or an investigational new-drug protocol, both of which could be problematic in a military setting. In its findings and recommendations, the board stated that it found the CDC consensus guidelines for malaria prevention “appropriate” for use by DoD and listed three options (A/P, mefloquine, and doxycycline) for areas with chloroquine-resistant P. falciparum. It noted that the contraindications for mefloquine were active depression and a history of psychosis or seizures and that it should be used cautiously in those with psychiatric disturbances. The board stated that mefloquine should continue to be available in the military drug armamentarium for malaria prophylaxis.

The committee reviewed a June 2004 “health information letter” that the U.S. Veterans Health Administration issued to clinicians caring for veterans who may have taken mefloquine as prophylaxis during OEF or OIF (VA, 2004). The Department of Veterans Affairs (VA) noted that mefloquine causes adverse events, possibly affecting adherence, and that anecdotal and media reports had suggested that the drug may cause serious neuropsychiatric effects. The letter also cited a DoD mefloquine “warning label” for clinicians that stated mefloquine should not be prescribed to persons “with a history of psychiatric or alcohol problems.” The VA letter described a literature review that had been performed and noted that the literature (based on case reports, clinical trials, and epidemiologic studies with no separation of the timing of adverse events) suggested that “certain health effects of mefloquine may persist after the drug is stopped.” It also stated that “clinical trials and epidemiological studies suggest that reported side effects are not common and are self-limiting” and that they included depression, panic attacks, anxiety, insomnia, vertigo, nausea and headache, and strange or vivid dreams. VA told the committee that such health information letters were used to provide information to VA staff and are not policy and that no record of additional information letters on the subject of mefloquine had been found.⁴ The VA letter listed all of the published sources that were used in drawing its conclusions. The committee considered all of those case reports and studies captured by VA for its own assessment but found that most did not meet its criteria of reporting empirical data on adverse events that persist or occur at least 28 days post-cessation of mefloquine.

In response to the committee’s request for further information on the DoD “warning label,” VA could not provide it, and DoD responded that it does not issue warning labels.⁵ DoD provided copies of information sheets for service members and their families (dated 2004) and for leaders (dated 2005) that had been available on the health.mil website (DoD, 2004, 2005a). These guides, in addition to warning against mefloquine use in those with a current or past history of psychiatric disorders, repeatedly warned against drinking alcohol while taking the drug because “alcohol may interfere with the medicine’s effectiveness and cause more serious side effects.” FDA mefloquine package inserts, including the most recent 2016 version, do not provide warnings or guidance on concurrent alcohol use (FDA, 2016).

In 2005 a DoD issuance outlined the U.S. Central Command deployment health protection policy (DoD, 2005b). It noted that component Combined Joint Task Force surgeons were authorized to modify malaria prophylaxis guidance for subordinate units based on latest intelligence, ground truth, and medical-risk assessment. The issuance stated that a mefloquine or doxycycline regimen must be used by personnel deploying to Central Asia, the Arabian Peninsula, and Africa; it stated further that mefloquine was not authorized for people on flight status.

A 2006 DoD policy memorandum directed that mefloquine be used by Coalition Forces Land Component Command personnel traveling to the Combined Joint Task Force–Horn of Africa area of operations; it instructed aviators and individuals unable to take mefloquine to take doxycycline (DoD, 2006).

The National Defense Authorization Act for Fiscal Year 2006 required the Secretary of Defense to conduct a study of adverse health events (including mental health) that may be associated with the use of antimalarial drugs, including mefloquine.⁶ In response, the assistant secretary of defense for health affairs commissioned four scientific studies to assess the comparative rates of adverse events resulting from the use of antimalarial medications, including mefloquine, chloroquine, doxycycline, and A/P, in deployed service members (DoD, 2009a). One study associated with this charge was published (Wells et al., 2006), and it is summarized in the Post-Cessation Adverse Events section of this chapter. Another study was reported to have been completed but not published. This committee has no information on other studies that may have been commissioned in association with this charge.

In 2009 a DoD memorandum advised that in chloroquine-resistant areas where doxycycline and mefloquine are equally efficacious and when personnel have a history of neurobehavioral disorders, doxycycline should be the first-line agent, A/P should be the second-line agent, and, in those who cannot take doxycycline or A/P, mefloquine should be used very cautiously and with clinical follow-up (DoD, 2009b). The memo also stated, presumably regarding personnel with no history of neurobehavioral disorders, that mefloquine should only be used by those with contraindications to doxycycline and without contraindications to mefloquine. In a retrospective analysis of 11,725 active-duty U.S. military personnel who were assigned in support of combat and reconstruction operations in Afghanistan around 2007, DoD administrative databases were used to determine the number of personnel with medical or pharmacologic contraindications to mefloquine prior to their deployment (Nevin, 2010; Nevin et al., 2008). In this cohort, 4,505 (38.4%) service members received a prescription for mefloquine, including 155 (13.7%) of the 1,127 service members with an identified medical or pharmacologic contraindication to mefloquine. A 2013 DoD memo stated that doxycycline and A/P were to be considered first-line agents in chloroquine-resistant areas, reserving mefloquine for use by those intolerant to or with contraindications to both doxycycline and A/P (DoD, 2013a). The same year, a DoD issuance stated that U.S. Special Operations Command medical personnel were to immediately cease prescribing and using mefloquine for prophylaxis and that personnel currently taking mefloquine were to transition to one of three alternative medications (DoD, 2013b). Total DoD mefloquine prescriptions fell from 23,889 in 2008 (18,942 active duty) to 263 (52 active duty) in 2017, representing a 99% reduction (99.8% among active duty) (Wiesen, 2019).

Australia

Although mefloquine continues to be recommended by WHO and CDC for malaria prophylaxis in civilians, several militaries have issued policies regarding its use in their members. Since 2015 the governments of Australia, Canada, and the United Kingdom have performed inquiries or investigations into the possible association of mefloquine with adverse effects, particularly neurologic and psychiatric effects, when used for malaria prophylaxis by their military forces (Australia, 2018; Canada, 2017; UK, 2016). Both the Canadian and Australian governments performed a literature review as a part of their inquiry process (Australia, 2018; Canada, 2017). Concerns raised by veterans and commentary in the media contributed to the initiation of these inquiries. Military veterans of Canada, Ireland, and the United Kingdom have filed lawsuits against their governments, holding them responsible for adverse events they state were caused by mefloquine use during their military service, and a U.S. veteran has filed a lawsuit against Hoffman-La Roche, the manufacturer of Lariam® (BBC, 2016; Connolly, 2019; O’Faolain, 2019).

As part of its inquiry, the Australian Senate commissioned a literature review of mefloquine and a research study that involved a re-analysis of health study data on antimalarial use from the 2007–2008 Centre for Military and Veterans’ Health deployment health studies (Australia, 2018). It heard or reviewed submitted testimony from government agencies (Department of Defence, Department of Health, Department of Veterans’ Affairs, Australian Defence Force Malaria and Infectious Disease Institute, Indo-Pacific Centre for Health Security, Department of Foreign Affairs and Trade Repatriation Medical Authority), a malaria-control organization (Asia Pacific Leaders Malaria Alliance), professional medical associations (Australasian Society for Infectious Diseases, Australasian College of Tropical Medicine, Royal Australian College of General Practitioners), advocate organizations (Australian Quinoline Veterans and Families Association, Quinism Foundation, Defence Force Welfare Association, Royal Australian Regiment Corporation, RSL National), and product-development partnerships and pharmaceutical manufacturers (Medicines for Malaria Venture, National Health and Medical Research Council, GlaxoSmithKline, Biocelect, 60 Degrees Pharmaceuticals, Roche), as well as from roughly 25 individuals, includingphysicians, academics, and veterans. In submitted testimony, symptoms attributed to mefloquine use were referred to as “mefloquine poisoning” or an “acquired brain injury” by the Australian Quinoline Veterans and Families Association and as “chronic quinoline encephalopathy” or “neuropsychiatric quinism” by the U.S.-based Quinism Foundation (Australia, 2018). Some veterans attributed their symptoms to mefloquine use 15 or more years earlier. In the report summary, while the Australian Senate committee acknowledged that its members were not medical experts, it stated, “The weight of prevailing medical evidence provided to the committee in response to these claims is that long term problems as a result of taking mefloquine are rare,” and it added that the committee had been informed that there was no definitive evidence to support the claim that mefloquine use results in acquired brain injury. It stated that while it believed that symptoms were being experienced by individuals, assigning a single cause to these illnesses did not take into account the multiple potential contributors to their health while they took the drug and in the years after. The committee recommended that the Australian Department of Veterans’ Affairs expedite its investigation into antimalarial claims logged since September 2016 and that it offer assistance to claimants and facilitate their access to legal representation. That committee also made recommendations to ensure better access to care for sick veterans, including that the Australian Department of Veterans’Affairs prioritize developing a neurocognitive health program. It did not recommend that changes be made to military policy on antimalarial use, which currently allows mefloquine to be prescribed as a “third line agent” only when doxycycline or A/P are contraindicated. Few Australian Defence Force members have been prescribed mefloquine since 2010; in 2017 only two prescriptions were made (Australian Department of Defence, n.d.).

Canada

The Canadian Armed Forces recommends the use of A/P, doxycycline, and mefloquine for malaria prophylaxis in addition to other measures to prevent mosquito bites. The Canadian Armed Forces follows the guidance set forth by the Public Health Agency of Canada. Individual armed forces members, in consultation with their health care providers, make a personal and informed decision on which antimalarial drug they want to be prescribed (Canada, 2017). Similar to the U.S. and Australian military experiences, the number of mefloquine prescriptions has decreased since 2010; 20 prescriptions were made in 2016 (Canadian Forces Health Services Group, 2017). The Canadian Surgeon General report, which was developed by a task force of Canadian Armed Forces personnel and civilians from the Department of National Defence, examined the Canadian Armed Forces experience with mefloquine and conducted a systematic review and assessment of military-specific safety information compared with other available antimalarial drugs (Canada, 2017). Although the report concluded that mefloquine was not associated with an overall excess risk of adverse effects in force personnel and its use did not prevent personnel from being able to perform their occupational duties, the quality of the evidence of the available published literature on the long-term health effects of mefloquine compared with other available antimalarial drugs was weak and itself did not support a change to policy. However, based on a consideration of other factors, such as most members showing a preference for the other available agents (A/P and doxycycline), the fact that screening for potential contraindications was lacking (a medical chart audit showed that 12% of mefloquine prescriptions had been made to service members who had contraindications), the lack of evidence on long-term safety, a desire for consistency with allied militaries (such as the United States), and the desire to be responsive to defence force member and societal concerns, the report recommended that the military change its policy to limit mefloquine use to (1) persons for whom use of A/P, doxycycline, and chloroquine are inappropriate (e.g., due to contraindications or intolerance); and (2) persons who have previously used and tolerated mefloquine, indicate a preference for it, and do not have contraindications. The report further recommended that the Canadian Armed Forces develop policies or procedures to enhance screening (and screening documentation) of service members for contraindications to mefloquine and other antimalarials and that a formal audit process be implemented to enable monitoring of antimalarial screening and prescription practices (Canada, 2017).

United Kingdom

The UK Ministry of Defence amended its policy regarding the use of mefloquine and other antimalarials on September 12, 2016, and it was further revised in June 2017 in response to recommendations from the UK House of Commons Defence Committee’s report on mefloquine (UK, 2016). The inquiry by the House of Commons Defence Committee was more limited in scope than those undertaken by the Australian and Canadian governments. The committee did not perform a literature search, but testimony was heard and reviewed from government agencies (Surgeon General; Ministry for Defence Personnel, Welfare and Veterans; Defence Medical Services), the pharmaceutical manufacturer Roche Products Ltd., and roughly 15 individuals, including a research scientist, physicians, and veterans. That committee concluded that mefloquine should be considered as a drug of last resort in defense forces. The new policy restricts the use of mefloquine even more narrowly to military personnel who are unable to tolerate available alternatives, have been screened for safe use via a face-to-face assessment, and have been informed of and provided the option to take alternative agents. The other prophylactic drugs available to armed forces members are doxycycline, chloroquine, and A/P. Consistent with Australian and Canadian defense forces, few prescriptions for mefloquine are made; from April 2018 to March 2019, there were 31 mefloquine prescriptions (UK, 2019).

Concurrent Adverse Events

Concurrent adverse events are well characterized for mefloquine. In general, mefloquine has a poorer reputation among the public and in military populations than the other available drugs for malaria prophylaxis. This is due mainly to the neurologic and psychiatric events associated with mefloquine, which are dose related, but that may occur at prophylactic doses (Stürchler et al., 1990; Weinke et al., 1991) and at a greater frequency than with other antimalarial prophylactics (Schlagenhauf et al., 2003, 2010). However, mefloquine-associated serious adverse events—defined as those that constitute a threat to life, require hospitalization, or result in severe disability—are rare, with estimated occurrences ranging from 1 in 10,000 to 1 in 20,000 depending on the population examined (Björkman et al., 1991; Schlagenhauf et al., 2003, 2010; Stürchler et al., 1990; Weinke et al., 1991; Wells et al., 2006). Instead of detailing every study that has reported concurrent adverse events that have been reported with use of mefloquine, the following paragraphs summarize the most common adverse events as well as those that are less commonly reported but still recognized as possibly related to the use of mefloquine for malaria prophylaxis using two identified Cochrane systematic reviews of the literature (Croft and Garner, 2000; Tickell-Painter et al., 2017a). Results from analyses that compared mefloquine with placebo or no drug (as opposed to comparisons with another antimalarial drug) were of greatest interest to the committee because an observed lack of difference in effect between two drugs could occur because both drugs cause the (same) adverse events. Use of a placebo-controlled design helps provide information about the “base rate” of the adverse events, to understand if the rates observed among individuals taking the drug are higher than would be expected with no drug exposure.

The aim of the first published Cochrane review (Croft and Garner, 2000) was to determine the effects of mefloquine in nonimmune adult travelers compared with other antimalarial regimens in relation to episodes of malaria, withdrawal from prophylaxis, and adverse events. Ten randomized trials of adult travelers and non-traveling volunteers were considered as well as 516 case reports for adverse events analyses. More recently, Tickell-Painter et al. (2017a) conducted a systematic review to summarize the efficacy and safety of mefloquine used as prophylaxis for malaria in adults, children, and pregnant women travelers. This review included 20 randomized controlled trials (totaling 11,470 participants), 35 cohort studies (totaling 198,493 participants), and 4 large retrospective analyses of health records (800,652 participants). Although the aims of these two large reviews were slightly different, Tickell-Painter et al. included nearly all of the same randomized controlled trials as Croft and Garner.

Croft and Garner (2000) assessed the use of mefloquine in nonimmune adult travelers compared with other regimens; the analysis included a total of 2,750 participants. To compare tolerability, the authors reviewed data on neurologic and psychiatric symptoms (depression, abnormal dreams, fatigue, headache, insomnia), gastrointestinal symptoms (abdominal discomfort, anorexia, diarrhea, nausea, vomiting), and fever and pruritus; data were to have been collected “at first assessment.” The authors identified five trials that compared outcomes with mefloquine versus placebo, and they reported that the tolerability outcome measures showed no statistically significant pattern relative to mefloquine or placebo, but that the numbers of study participants were generally small. Six trials were identified that compared mefloquine with other malaria-prophylactic drugs, but the comparator drugs were not named, except incidentally when specific comparisons were made. The authors calculated Peto odds ratios (used when pooling odds ratios) and found the overall incidence of adverse events with mefloquine to be no different from that of other antimalarials (OR = 1.00, 95%CI 0.80–1.27; 4 studies, 1,344 participants). There was no consistent pattern across the five neurologic and psychiatric symptoms analyzed (depression, dreams, fatigue, headache, insomnia), but mefloquine was shown to be more likely than other agents to cause insomnia (OR = 1.64, 95%CI 1.18–2.28; 4 studies, 1,344 participants) and fatigue (OR = 1.57, 95%CI 1.01–2.45; 4 studies, 1,344 participants). No consistent pattern was seen for the gastrointestinal symptoms analyzed, but abdominal discomfort was reported less frequently among users of mefloquine than among users of other antimalarials (OR = 0.57, 95%CI 0.42–0.77; 5 studies, 1,464 participants), as was the case with anorexia (OR = 0.64, 95%CI 0.43–0.95; 4 studies, 1,444 participants) and nausea (OR = 0.74, 95%CI 0.57–0.96; 6 studies, 1,717 participants). The authors noted the heterogeneity of the studies and stated that the overall effect regarding gastrointestinal symptoms appeared to be due to one study in which participants reported symptoms in the chloroquine-proguanil group more frequently than in the mefloquine group. Reports of fever and pruritus were similar in the mefloquine and comparator arms. The authors also noted that they had identified 328 case reports that involved mefloquine prophylaxis and adverse events (discussed later in this chapter under Case Reports and Case Series).

Tickell-Painter et al. (2017a) prespecified adverse events of interest to include these disorders: psychiatric (abnormal dreams, insomnia, anxiety, depression, psychosis); nervous system (dizziness, headache); ear and labyrinth (vertigo); eye (visual impairment); gastrointestinal tract (nausea, vomiting, abdominal pain, diarrhea, dyspepsia); and skin and subcutaneous tissues (pruritus, photosensitivity, vaginal candida). The assessment comparing the use of mefloquine for malaria prophylaxis with placebo or no treatment included 13 randomized controlled trials and 5 cohort studies. Dosages varied, as did methods of collecting adverse event data; eight of the trials were considered to be at high risk of bias from selective outcome reporting. The authors applied categories of certainty to the results based on the five GRADE considerations (risk of bias, consistency of effect, imprecision, indirectness, and publication bias) (Higgins et al., 2019).

Overall, among the six randomized controlled trials only one serious adverse event (death from septic shock after an emergency cesarean section for obstructed labor) was reported among study participants who used mefloquine (n = 592) compared with two that occurred among people using placebo (n = 629); none of these events were attributed to the drug regimen (Tickell-Painter et al., 2017a). In the cohort studies, seven serious adverse events (five were depression and two were dizziness, and all were attributed to the drug regimen) were reported among 913 mefloquine users, and none were reported in 254 travelers who did not use antimalarials. When analyses were performed to compare mefloquine with doxycycline (4 trials and 20 cohort studies), A/P (3 trials and 16 cohort studies), and chloroquine (6 trials and 15 cohort studies), no difference in the incidence of serious adverse events was found between mefloquine and doxycycline, A/P, or chloroquine. Participants receiving mefloquine were no more likely to discontinue their medication due to adverse events than were doxycycline users (RR = 1.08, 95%CI 0.41–2.87; 4 trials, 763 participants; low-certainty evidence), but mefloquine users were more likely to discontinue their medication due to adverse events than A/P users (RR = 2.86, 95%CI 1.53–5.31; 3 trials, 1,438 participants; high-certainty evidence) (Tickell-Painter et al., 2017a).

Regarding neurologic outcomes, people taking mefloquine were less likely than those taking placebo in trials to experience headache (RR = 0.84, 95%CI 0.71–0.99; 5 trials, 791 participants), but this was not observed in the one cohort study that reported on headache. Whereas mefloquine users in trials were no more likely than recipients who took a placebo or no drug to experience dizziness (RR = 1.03, 95%CI 0.90–1.17; 3 trials, 452 participants), in the cohort studies, participants who used mefloquine were statistically significantly more likely to experience dizziness (RR = 1.80, 95%CI 1.29–2.49; 3 studies, 1,901 participants) than those who used placebo or no drug. No differences were observed between mefloquine and placebo groups for vertigo in either trials or cohort studies. Cohort study comparisons between mefloquine and doxycycline users found no differences for headache or dizziness.

None of the randomized controlled trials reported on the psychiatric symptoms of abnormal dreams, insomnia, anxiety, depressed mood, or abnormal thoughts and perceptions (psychosis). Participants in cohort studies who received mefloquine were more likely than participants who did not take prophylaxis to experience abnormal dreams (RR = 2.35, 95%CI 1.15–4.80; 2 studies, 931 participants) and insomnia (RR = 1.46, 95%CI 1.06–2.02; 2 studies, 931 participants). Effects on anxiety (RR = 1.21, 95%CI 0.67–2.21; 2 studies, 931 participants), depressed mood (RR = 2.43, 95%CI 0.65–9.07; 3 studies, 1,901 participants), and abnormal thoughts or perceptions (RR = 5.77, 95%CI 0.79–42.06; 1 study, 970 participants) were not consistent across studies and did not reach standard levels of statistical significance. Findings from trials and cohort studies that used A/P as a comparator were similar, with mefloquine users statistically significantly more likely to report abnormal dreams, insomnia, anxiety, and depressed mood, although it should be noted that all of the effect estimates were quite imprecise. Using the six cohort studies that used doxycycline as a comparator, mefloquine users were more likely to report abnormal dreams (RR = 10.49, 95%CI 3.79–29.10; 4 studies, 2,588 participants), insomnia (RR = 4.14, 95%CI 1.19–14.44; 4 studies, 3,212 participants), anxiety (RR = 18.04, 95%CI 9.32–34.93; 3 studies, 2,559 participants), and depressed mood (RR = 11.43, 95%CI 5.21–25.07; 2 studies, 2,445 participants), but the pooled effect estimates were very imprecise. Additionally, 15 episodes of abnormal thoughts and perceptions were reported among mefloquine users and none among doxycycline users in the cohort studies reporting adverse events. In the single trial included and the large retrospective health care record analyses, there were either no differences between groups, or doxycycline users were more likely to experience psychiatric symptoms. Overall, the authors concluded that people taking mefloquine are more likely to have abnormal dreams, insomnia, anxiety, and depressed mood during travel than people who take A/P (moderate-certainty evidence) or doxycycline (very low-certainty evidence).

Mefloquine recipients were more likely to experience nausea than placebo recipients for both trials (RR = 1.35, 95%CI 1.05–1.73; 2 trials, 244 participants) and cohort studies (RR = 1.85, 95%CI 1.42–2.43; 3 studies, 1,901 participants), but there was no difference between groups for vomiting, abdominal pain, or diarrhea. For both trials and cohort studies, when mefloquine users were compared with A/P users, mefloquine users were statistically significantly more likely to experience nausea, but there was no statistically significant difference for vomiting, abdominal pain, or diarrhea. Based on data from cohort studies, mefloquine users were less likely than doxycycline users to report dyspepsia (RR = 0.26, 95%CI 0.09–0.74; 5 studies, 5,104 participants), vomiting (RR = 0.18, 95%CI 0.12–0.27; 4 studies, 5,071 participants), nausea (RR = 0.37, 95%CI 0.30–0.45; 5 studies, 2,683 participants), and diarrhea (RR = 0.28, 95%CI 0.11–0.73; 5 studies, 5,104 participants). No difference between mefloquine users and doxycycline users was found for abdominal pain (RR = 0.30, 95%CI 0.09–1.07; 4 studies, 2,569 participants). The authors stated that the estimates for dyspepsia and vomiting were given low or very low certainty of evidence. Other symptoms were also included when available. Based on one cohort study of 197 participants, mefloquine users were more likely than those who were given placebo to experience pruritus (RR = 6.71, 95%CI 1.58–28.55), although the estimate was imprecise. Pruritus was not statistically different between mefloquine users and placebo or non-drug users in trials (RR = 0.86, 95%CI 0.60–1.24; 3 trials, 609 participants). Based on the data from cohort studies, mefloquine users were less likely than doxycycline users to report photosensitivity (RR = 0.08, 95%CI 0.05–0.11) and vaginal thrush (RR = 0.10, 95%CI 0.06–0.16), but for both of these results the evidence was considered to be very low certainty. No differences were observed between mefloquine and placebo groups for visual impairment in either trials or cohort studies. Authors noted that comparisons of mefloquine with chloroquine added no new information and that subgroup analysis by study design, duration of travel, and military versus non-military participants provided no conclusive findings.

Post-Cessation Adverse Events

A total of 1,577 abstracts or titles were identified by the committee for inclusion for mefloquine. After screening, 489 abstracts and titles remained, and the full text for each was retrieved and reviewed to determine whether it met the committee’s inclusion criteria, as defined in Chapter 3. The committee reviewed each article and identified 11 epidemiologic studies that included some mention of adverse events that occurred ≥28 days post-cessation of mefloquine (DeSouza, 1983; Eick-Cost et al., 2017; Laothavorn et al., 1992; Meier et al., 2004; Schlagenhauf et al., 1996; Schneider et al., 2013, 2014; Schneiderman et al., 2018; Schwartz and Regev-Yochay, 1999; Tan et al., 2017; Wells et al., 2006). A table that gives a high-level comparison (study design, population, exposure groups, and outcomes examined by body system) of each of the 11 epidemiologic studies that examined the use of mefloquine and that met the committee’s inclusion criteria is presented in Appendix C. Other identified articles are cited in the background, case reports and selected subpopulations, and biologic plausibility sections as relevant.

Case Reports and Case Series

Published case reports can offer detailed information about symptoms and their course, such as the timing of onset in relation to exposure to the drug, treatment, remission, and persistence of symptoms, but they rarely generate information for causative inference. To be considered, published case reports and case series had to report on a follow-up of at least 28 days post-mefloquine-cessation. Of the 56 case reports identified, many reported only acute symptoms that resolved within 28 days post-cessation. The committee closely reviewed the remaining 20 case reports (totaling 25 patients) that had been identified (Baker, 1996; Borruat et al., 2001; Chester and Sandroni, 2011; Dietz and Frolich, 2002; Eaton, 1996; Even et al., 2001; Jain et al., 2016; Javorsky et al., 2001; Jha et al., 2006; Katsenos et al., 2007; Lobel et al., 1998; McEvoy et al., 2015; Meszaros and Kasper, 1996; Nevin, 2012; Potasman and Seligmann, 1998; Tran et al., 2006; Udry et al., 2001; Walker and Colleaux, 2007; Watt-Smith et al., 2001; Whitworth and Aichhorn, 2005) as well as eight case series papers (Adamcova et al., 2015; Bem et al., 1992; Beny et al., 2001; Croft and Garner, 2000; Croft and Herxheimer, 2002; Ringqvist et al., 2015; Smith et al., 1999; van Riemsdijk et al., 1997b). Among the case reports, all patients had acute effects, and 16 patients had persistent neurologic or psychiatric effects for more than 28 days following their last dose of mefloquine. These symptoms included dizziness, anxiety, depression, insomnia/exhaustion, paranoia, hallucinations, visual illusions, mania, depersonalization, and suicidal ideation. Nevin (2012) published a detailed case of a patient who took mefloquine and acutely experienced anxiety, then developed fatigue, confusion, psychosis, dissociation, personality change, tinnitus, vertigo, dizziness, disequilibrium, and cognitive deficits, and he exhibited parasuicidal behavior. Objective testing discovered central vestibulopathy. The resultant diagnoses included vertigo of central origin, toxic encephalopathy, various psychiatric disorders, ataxic gait, and memory loss. Persistent findings (follow-up ended after 10 months of first symptom onset) following the resolution of symptoms of psychosis, were fatigue, vertigo, disequilibrium, visual illusions, photosensitivity, memory impairment, and personality changes.

One case of persistent retinopathy (Walker and Colleaux, 2007) and other ocular disturbances (Jain et al., 2016), one case of tinnitus resulting in hearing loss (Lobel et al., 1998), and one case of worsening psoriasis (Potasman and Seligmann, 1998) following mefloquine administration were reported. Additional cases reported neuropathy (Chester and Sandroni, 2011; Jha et al., 2006; Watt-Smith et al., 2001); paralysis, trouble breathing, heart palpitations (Eaton, 1996); eosinophilic pneumonia (Katsenos et al., 2007); weakness (Jha et al., 2006; Whitworth and Aichhorn, 2005); skin rash (Eaton, 1996; Jha et al., 2006); and pain in the face and extremities (Chester and Sandroni, 2011).

Five case series reported similar symptoms to those in the individual case reports. Beny et al. (2001) reported on 15 travelers who prematurely terminated their travel because of neurologic or psychiatric symptoms, and 7 of those had taken mefloquine. Of the mefloquine users, three had persistent anxiety and depression, although the timing of these symptoms relative to mefloquine use was unclear. In a review of adverse events reports submitted to the drug manufacturer, Bem et al. (1992) found 430 cases of adverse events when mefloquine was used prophylactically. More than half (56%) of these events were considered neurologic or psychiatric, as defined by WHO, and 59 of these individuals required hospitalization or resulted in severe disability. There were 26 reported cases of convulsions, and half of these cases had a neurologic or psychiatric history. All but one of the cases of convulsions resolved within 1 month of the last dose of mefloquine. Additionally, Bem reported 12 cases of depression or “manic-depression,” and 9 of those cases had suicidal ideation or attempts, or both. Psychosis was reported in 20 cases, and 11 of those individuals recovered within 40 days (mean 21 days), while 2 recovered within 4–7 months. There was one case of toxic encephalopathy reported, but that person recovered within 3 months. Using postmarketing surveillance data of mefloquine in the Netherlands, van Riemsdijk et al. (1997b) reported on 132 cases with a range of symptoms including depression, anxiety, agitation, nightmares, insomnia, concentration impairment, psychosis, hallucinations, depersonalization, and paranoia. Of the 132 cases, 36 had persistent symptoms, 74 had complete recoveries following cessation of mefloquine, and the disposition of the remaining 22 people was unknown. Using reports of adverse events to the manufacturer’s drug safety database between February 1984 and January 2011, Adamcova et al. (2015) performed an analysis of eye disorders associated with the prophylactic use of mefloquine. A total of 591 individuals were identified who experienced 695 eye disorder events, 223 of which were considered serious, that were subsequently categorized into visual acuity (to include blindness, reduced visual acuity, visual impairment, and blurred vison), events affecting the anatomical parts of the eye (retina, vitreous, lens, cornea, optic nerve and glaucoma, and other disorders), and neuro-ophthalmic disorders. The temporal relationship of mefloquine use to adverse events was considered. When available, risk factors such as relevant medical history, comedication, and associated conditions were also considered. The time of onset, which was available for only 70 of the events, ranged from 1 hour to 1,095 days (median 16.5 days). The duration of adverse events was known for only 5% of reports and, among those, ranged from 30 minutes to 270 days (median 10.5 days). Symptoms of optic neuropathy were reported for 48 individuals (53 events); 8 individuals (reporting 10 events) recovered, with sequelae that continued to affect visual acuity; and 3 individuals reported no complete recovery. Six events involving the cornea and five events involving the lens were reported, but eight of these had explanations other than mefloquine exposure and three of the reports did not contain sufficient information for a medical assessment. Of the 23 events involving retinal disorders, 9 were maculopathy, and most of these events were considered to be due to factors other than mefloquine.

Ringqvist et al. (2015) reported on 73 adults with mefloquine-associated adverse events (67 of them had used mefloquine for malaria prophylaxis) based on 95 reports to the Danish National Drug Authority Committee of Adverse Drug Reactions. Each person was contacted, and standardized instruments or interviews were used to elicit and categorize symptoms; these measures were completed 270–2,010 days following the adverse event. For 77% of cases, the individuals reported their symptoms as beginning in the first 3 weeks of mefloquine use, while 15% reported an onset of symptoms after 1–2 months of use, and 8% reported a symptom onset more than 8 weeks after the initiation of mefloquine. Of the 73 people, 45 reported physical symptoms, 27 reported signs of anxiety, 26 reported sleep disturbances including nightmares, 18 reported depression or feeling low, 11 reported signs of possible psychotic states (delusions/hallucinations), 9 reported cognitive problems, 3 reported confusion, and 1 reported mania; 40 individuals reported more than one complaint. Perceptual disturbances/hallucinations or delusional experiences were reported by 17 individuals following mefloquine use; all of these resolved within 9 months, and most within 3 weeks. Recurring nightmares were reported for 43 cases, and 9 individuals continued to have recurring nightmares for more than 3 years after mefloquine cessation. Cognitive dysfunction was reported in 42 cases and persisted for more than 3 years for 14 people. Included individuals reported significantly worse psychiatric symptoms than the matched controls in the Danish normative sample. Of the participants, 41% reported that they had obtained some treatment for their psychiatric adverse event. Although this case series provides some evidence supporting the development of persistent psychiatric problems after mefloquine use, the series was limited in that it was based on 73 cases of adverse events deemed severe enough to be reported to the Danish national registry. It is not known how complete the reporting to the Danish registry is, and there was no appropriate comparison group, only a comparison to Danish national norms for the self-report questionnaires administered after cases were reported to the registry. The investigators estimated that adverse events occurred at a rate of about 2 per 10,000 doses, suggesting that serious persistent events, if related to mefloquine, are rare.

A Cochrane review by Croft and Garner (2000) identified 136 published case studies totaling 516 nonimmune travelers who had experienced adverse events while using mefloquine. Of those 516 individuals, 328 were using mefloquine as malaria prophylaxis. Four case reports involved fatal reactions to mefloquine, but it is unclear whether the deaths were reported in cases involving mefloquine prophylaxis or treatment, and further details were not provided. The authors discussed the best measures of tolerability and the possible influence of differences among groups (e.g., gender, weight, age, ethnicity) on the occurrence of adverse events, but they did not provide analyses or conclusions regarding the case reports. It was not clear how many of these people had persistent symptoms, and, other than a listing of the citations, additional information was sparse. Croft and Herxheimer (2002) elaborated on these 516 cases, reporting that 328 of the individuals had taken mefloquine prophylactically, and the median duration of adverse events symptoms was 16 days (range 1–550). The authors postulated that the symptoms associated with taking mefloquine were primarily related to liver or thyroid pathology.

Smith et al. (1999) reviewed 74 published case reports of mefloquine use (prophylaxis or treatment) specific to dermatologic adverse events. Some of these cases were collected from outcomes of clinical trials, and nearly half had used mefloquine as treatment for malaria. The onset of the dermatologic effects was recorded in only 11 of the cases. The most common symptoms were pruritus and itching, which were reported in more than 40% of the cases, followed by rashes. The majority of effects were reported as mild or moderate in intensity and were usually self-limiting, although the timing was not specified. Other dermatologic adverse events included two reports of cutaneous vasculitis and one report each of Stevens-Johnson syndrome and toxic epidermal necrolysis.

Tickell-Painter et al. (2017b) performed a systematic review of reports of death or parasuicide (a suicide attempt not resulting in death) associated with mefloquine when used at various dosages for malaria prophylaxis. The literature search included all forms of prospective and retrospective studies of individual case reports or reviews of case reports that reported deaths or parasuicide up to July 11, 2017. Each case was reviewed using a formal causality assessment based on a causality assessment by WHO’s Uppsala Monitoring Centre. When information was poor or conclusions could not be drawn, the event was categorized as “unclassifiable.” Of the 527 articles identified and reviewed, 17 reported deaths or parasuicide, and only 8 had sufficient detail to be included in a causality assessment. Two deaths were identified as having a probable association with mefloquine. Both were in children and were characterized as “idiosyncratic drug reactions” (one involved pulmonary fibrosis and interstitial pneumonia; the other involved erythema, blistering, other complications, and eventually cardiac asystole). In the first case, symptoms began during mefloquine use, and death occurred 5 weeks after drug cessation; in the second, symptoms began during mefloquine use, but it is unclear when or even if the mefloquine use was stopped. Eight deaths were deemed “unlikely” to have been related to mefloquine or “unclassifiable” because of insufficient information. The authors identified one parasuicide with a “possible” causal association. A 22-year-old woman experienced episodes of crying, emotional detachment, and low mood 1 day after taking mefloquine 250 mg; her symptoms decreased on days 5 and 6; after an additional dose 1 week later, she experienced a relapse of symptoms, with ideas of guilt and death, and feelings of body transformation; and 5 days later she was hospitalized after a suicide attempt by drowning. Authors noted that the original source provided no information regarding the individual’s past medical history, including her use of any other medications.

The authors concluded that the number of deaths that could be reliably attributed to the prophylactic use of mefloquine is very low (Tickell-Painter et al., 2017b). In their discussion, however, the authors stated that a limiting factor in their review was poor reporting in the literature; few reports, including those deriving from spontaneous adverse event reporting databases, provided sufficient detail to perform a critical assessment. Additionally, the cases represented different time points for the outcomes: some were concurrent and some were longer term, further limiting the contribution of this paper to this report.

In summary, there are published cases of persistent neurologic, psychiatric, and other adverse events following exposure to mefloquine. The majority of these case reports and case series presented individuals whose symptoms eventually resolved, even if they initially persisted beyond 28 days following the last dose of the medication. Although the case reports are compelling, without larger samples or comparison groups to establish base rates of disorders, it is difficult to establish a causal role for mefloquine in these cases.

Selected Subpopulations

In the course of its review of the literature on mefloquine, the committee identified and reviewed available studies that reported results stratified by demographic, medical, or behavioral factors to assess whether the risk for adverse events when using mefloquine for prophylaxis is associated with being part of or affiliated with a specific group. This was not done exhaustively, and the evidence included in this section is generally limited to adverse events observed with concurrent use of mefloquine. Many of these studies did not meet the inclusion criteria of following their population for at least 28 days post-mefloquine-cessation, but the committee considers these findings to be important indicators when considering the evidence as a whole. The following risk groups were specifically considered: females in general and pregnant women in particular, people with low weight or BMI, people with allergies and chronic conditions who may be taking concurrent medications, and people who use alcohol, marijuana, or illicit drugs.

Epidemiologic Studies Presenting Contributory Evidence

Eick-Cost used DoD administrative databases to perform a large retrospective cohort study among active-duty service members who filled at least one prescription for mefloquine, doxycycline, or A/P between 2008 and 2013. The primary study objective was to assess and compare the risk of incident and recurrent ICD-9-CM-coded neurologic and psychiatric outcomes that were reported at medical care visits during concurrent antimalarial use plus 365 days after the end of a prescription. This was a well-designed study and included several important factors that increased its methodologic quality: a large sample size, an administrative data source for both exposure and outcomes, and careful consideration of potential confounders including demographics, psychiatric history, and the military characteristics of deployment and combat exposure. Because the neurologic and psychiatric diagnoses occurring during current and recent use were analyzed together without distinguishing between events that occurred within 28 days of antimalarial use and those that occurred ≥28 days post-cessation, it provides no quantitative information regarding the persistence of most events other than the notation in the text that the results did not change when restricted to a post-cessation period of 30 days. Whereas the use of medical diagnoses is likely to be more reliable for the outcomes than self-report, there was no validation of the diagnoses recorded in the administrative databases, and symptoms or events that did not result in a medical visit or diagnosis would have been missed. For PTSD diagnoses, there was no information concerning when the index trauma occurred.

Schneiderman et al. (2018) conducted an analysis of self-reported health outcomes associated with the use of antimalarials in a population-based cohort of deployed and nondeployed U.S. veterans, using information collected as part of the NewGen Study. Exposure and outcomes were systematically obtained, and psychiatric outcomes were measured by standardized assessment instruments. Antimalarial medication use was grouped by mefloquine, chloroquine, doxycycline, primaquine, mefloquine in combination with other drugs, other antimalarials, and not specified or no antimalarial drug exposures. Health outcomes were self-reported using standardized instruments: the SF-12 for general physical health status, PCL-C for PTSD, and the Patient Health Questionnaire. The overall sample was large, and the researchers used a reasonably thorough set of covariates in models estimating drug–outcome associations, including deployment and combat exposure. Although the time period of drug use and the timing of health outcomes were not directly addressed, given that the population consisted of veterans who had served between 2001 and 2008 and the survey was not administered until 2009–2011, it is reasonable to assume that antimalarial drug use had ceased some time before. The methodology and response rate (34% total; weighted 4.4% of deployed and weighted 2.2% of nondeployed individuals used mefloquine) for this study may have led to the introduction of non-response, recall, or selection biases; however, the committee believed that investigators used appropriate data analysis techniques to mitigate the effects of any biases that were present.

Wells et al. (2006) was a large, well-designed study that used DoD administrative databases to examine incident hospitalizations by body system among active-duty service members who had been prescribed mefloquine and deployed at some time in calendar year 2002. Because the follow-up of mefloquine users began at the time of their return from deployment, it is reasonable to assume that these results largely reflect experience following the cessation of exposures of varying duration. Nonetheless, the results for varying time intervals following the cessation of use (or time since return from deployment) were not presented. Two comparison groups who were not prescribed antimalarials (service members assigned to Europe or Japan and service members who were deployed for 1 month or longer) were used in the analysis, but the demographic and military characteristics of the Europe- or Japan-assigned individuals differed substantially from those of the deployed individuals, suggesting that this was not an appropriate comparison group. Several attributes of its design increase its methodologic quality: a large sample size, the use of an administrative data source for both exposure and ICD-9-CM-based outcomes, and the inclusion of a reasonable set of sociodemographic, psychiatric history, and military characteristic covariates in the analyses. However, combat exposure was not specifically addressed, and although deployment may have been assumed to be a surrogate for combat, the lack of control for combat exposure itself is a limitation. The health outcomes were systematically and objectively ascertained but would reflect only the most severe experiences requiring hospitalization, which would likely exclude most people who experienced mental health symptoms or disorders. The small number of specific diagnoses for certain outcomes further limits the generalizability of these results.

Three large, retrospective studies of travelers (Meier et al., 2004; Schneider et al., 2013, 2014) were conducted using data from the UK-based GPRD to assess the incidence and compare the odds of developing first-time neurologic, psychiatric, or eye disorders in individuals using mefloquine compared with other antimalarial drugs for malaria prophylaxis. While the specific outcomes examined differed by study, the general design and methodology were the same for all three. The use of GPRD data (a well-established platform designed for both clinical practice and research) allowed for adequate power to detect differences in outcomes and for a uniform collection of exposures (although recorded drug prescriptions do not equate to use or adherence) and outcomes (based on clinical diagnoses coded from medical care visits) that were not subject to recall bias. Events that did not result in a medical care visit or that occurred outside of the national health care system would have been missed; however, it is unlikely that this would result in differential selection bias. Diagnoses were defined a priori, which excluded other outcomes, including the potential to identify rare outcomes. The antimalarial-exposed populations were large, an appropriate comparison group of travelers not using any form of malaria prophylaxis was included, and health outcomes were reported in defined time periods, including current use through 90 days after a prescription ended (termed current use or recent use in analyses that included both irrelevant [7–27 days] and relevant [28–90 days] time periods) and 91–540 days following cessation of use (termed past use in analyses). Adjustments were made for several confounders, including age, sex, calendar time, practice, smoking status, and BMI using appropriate study design or analytic methods. Each study included a nested case–control component that allowed for control of important covariates.

The primary aim of Tan et al. (2017) was to assess the prevalence of several health conditions experienced by returned Peace Corps volunteers associated with the use of prophylactic antimalarial drugs. The number of participants was large (8,931 participants), and of those who used an antimalarial drug a majority (59%) had used mefloquine. A number of important covariates, such as psychiatric history and alcohol use, were collected, but the study had several methodologic issues. These limitations included its study design (self-report Internet-based survey), exposure characterization based on self-report (which introduces several potential biases such as recall bias, sampling bias, and confounding), outcome assessment (based on self-report of health provider-diagnosed conditions up to 20 years post-service), the use of mixed comparison groups, a lack of detail regarding the analysis methods, and a poor response rate (11%, which likely introduced selection bias). The evidence generated by this study was thus considered to contribute only weakly to the inferences of mefloquine and persistent adverse events or disorders.

Neurologic Disorders

There are recognized concurrent adverse neurologic events associated with mefloquine use, including dizziness, vertigo, loss of balance, headache, memory impairment, confusion, encephalopathy, sensory or motor neuropathies, convulsions, tinnitus, and hearing loss (FDA, 2016; Tickell-Painter et al., 2017a). In the epidemiologic studies examining persistent neurologic outcomes, these effects were not observed to occur at statistically different rates for mefloquine users compared with people who used other antimalarial drugs or who did not use any prophylaxis. However, persistent dizziness was found in a few case reports, and studies of mefloquine use in pregnant women showed an increased risk for dizziness that resolved spontaneously within a few days (Briand et al., 2009; Nosten et al., 1994). A recent Cochrane review of mefloquine use for prophylaxis in travelers also reported that current mefloquine use was associated with statistically significantly higher risks of dizziness than placebo or no prophylaxis in cohort studies, but in clinical trials no difference in experiencing dizziness was found between mefloquine users and those given a placebo. Other persistent neurologic symptoms and conditions of neuropathy, weakness, paralysis, convulsions, and concentration impairment were described in the case reports.

In addition to the data on neurologic outcomes in humans, animal and cell culture studies lend some support for plausible biologic mechanisms through which mefloquine may contribute to neurotoxic processes. These include the modulation of calcium homeostasis, the induction of oxidative stress, the inhibition of connexin signaling, and the modulation of neurotransmitter receptor binding.

The committee reviewed five epidemiologic studies that examined neurologic outcomes that occurred at least 28 days following the cessation of mefloquine (Eick-Cost et al., 2017; Schlagenhauf et al., 1996; Schneider et al., 2013; Tan et al., 2017; Wells et al., 2006). These outcomes were inconsistently identified and measured across studies: ICD-9-CM-coded disorders of the nervous system as a category, and specific neurologic outcomes of nystagmus, vertiginous syndromes, dizziness and giddiness, and migraine (Wells et al., 2006); ICD-9-CM-coded outcomes of confusion, tinnitus, vertigo, and convulsions (Eick-Cost et al., 2017); epilepsy and peripheral neuropathy (Schneider et al., 2013); “neuropsychologic” as a category and that separately included dementia, migraines, seizures, tinnitus, vestibular disorder, and “other” neurologic disorder (Tan et al., 2017); and “neuropsychiatric” adverse events that included dizziness/vertigo, headache, and decreased concentration (Schlagenhauf et al., 1996). While all five of these studies have methodologic limitations, the three that provided the most evidence for potential persistent or latent neurologic outcomes based on the strength of the methods used were Eick-Cost et al. (2017), Wells et al. (2006), and Schneider et al. (2013).

In their analysis of data from DoD administrative databases, Eick-Cost et al. (2017) examined neurologic outcomes, and analyses were stratified by deployment and, separately, by psychiatric history. Adjusted incident rates of tinnitus, convulsions, and confusion were higher among the nondeployed than among the deployed groups who used mefloquine. There were no statistically significant differences for any of the neurologic outcomes among the deployed mefloquine users compared with the doxycycline users. Among the nondeployed, only vertigo was statistically significantly different (decreased) for mefloquine versus doxycycline users. Adjusted IRRs comparing mefloquine with A/P by deployment status found that the risk of tinnitus was statistically significantly increased among both the deployed and the nondeployed groups. No other outcomes were statistically significantly different between deployed mefloquine and A/P users. For both the mefloquine and doxycycline groups, there were no statistically significant differences between drugs when adjusting for history of psychiatric disorder. In a second study of U.S. service members, Wells et al. (2006) presented hospitalizations from nervous system disorders as a single category. Only six hospitalizations due to nervous system disorders were reported for mefloquine users, and comparisons with both reference groups showed that mefloquine users had no statistically significant different risk for nervous system disorders as a group. When hospitalizations due to specific neurologic outcomes were considered, among those receiving mefloquine there were no cases of nystagmus or dizziness and giddiness, one case of vertiginous syndromes, and three cases of migraine, which resulted in wide, imprecise, and null effect estimates when these rates were compared with those of the two reference groups of service members. Schneider et al. (2013) assessed incident diagnoses of epilepsy and peripheral neuropathy among travelers who had been prescribed mefloquine and compared them with those given another antimalarial and, separately, with travelers who had a travel consult but were not prescribed antimalarial drugs; the analysis was stratified by time since cessation. Over the approximately 8.5-year period of data examined, a total of 14 mefloquine users were diagnosed with incident epilepsy, 6 of whom were current use and 8 of whom were past use. Among the eight mefloquine users with incident neuropathies, five were current users and three were past users. In the nested case–control analysis, after adjusting for smoking and BMI, the odds of developing epilepsy were decreased, and the odds of developing peripheral neuropathies were elevated for mefloquine users, but neither of these results reached statistical significance. Similarly, when stratified by current use or past use, the adjusted odds of epilepsy for mefloquine users compared with non-antimalarial users were not statistically significantly different. The FDA package insert warns individuals with epilepsy that taking mefloquine may increase the risk for convulsions, and people who had previously been diagnosed with epilepsy were excluded from the study. Compared with nonusers of antimalarials, current users of mefloquine had increased odds of neuropathy, while past users had decreased odds of neuropathy but neither of these estimates was statistically significant. In sum, this was a well-designed study, and the stratification of past use in particular provides some evidence for an absence of increased persistent neurologic effects of epilepsy and peripheral neuropathy following the use of mefloquine. Overall, these three well-designed studies provide some evidence for an absence of persistent neurologic events following the use of mefloquine, but the number of neurologic disorders was small, making these results far from definitive.

The two other epidemiologic studies with post-cessation follow-up (Schlagenhauf et al., 1996; Tan et al., 2017) that presented some information on neurologic outcomes were not as methodologically robust as Eick-Cost et al. (2017), Wells et al. (2006), or Schneider et al. (2013), and their results lend additional weak support for an absence of increased persistent neurologic effects.

Based on the available evidence, the committee concludes that there is insufficient or inadequate evidence of an association between the use of mefloquine for malaria prophylaxis and persistent or latent neurologic events. Current evidence suggests further study of such an association is warranted, given the evidence regarding biologic plausibility, adverse events associated with concurrent use, or data from the existing epidemiologic studies.

Psychiatric Disorders

The evidence supporting concurrent adverse psychiatric effects of mefloquine is compelling. These effects include anxiety, depression, mood swings, panic attacks, abnormal dreams, insomnia, hallucinations, aggression, and psychotic or paranoid reactions. Suicidal thoughts and death by suicide have been also been reported with concurrent use of mefloquine. While the charge to the committee was to address persistent or latent adverse events, the occurrence of concurrent adverse events enhances the plausibility that problems may persist beyond the period of drug use. The FDA package insert warns that psychiatric symptoms such as acute anxiety, depression, and restlessness should be viewed as potential precursors to more serious psychiatric or neurologic adverse events and that mefloquine should be discontinued if they occur. In addition, the FDA labeling has increasingly invoked the potential for persistent adverse psychiatric events, suggesting reports received warranted these changes, although no research is cited as the basis for these changes. Two Cochrane reviews examined concurrent adverse events of mefloquine use in travelers. Croft and Garner (2000) reported on psychiatric symptoms in six trials comparing the tolerability of mefloquine to other antimalarials and found that the only outcomes with increased odds associated with mefloquine use were insomnia and fatigue. Tickell-Painter et al. (2017a) found that mefloquine users were more likely than users of doxycycline and users of A/P to experience insomnia, anxiety, abnormal dreams, and depressed mood. In cohort studies mefloquine users were more likely than participants who did not take prophylaxis to experience abnormal dreams and insomnia. However, this review included concurrent adverse events, and several outcomes had imprecise effect estimates because of the small numbers of adverse events (e.g., serious adverse effects, depressed mood, abnormal thoughts or perceptions). Additional studies of selected subpopulations of concurrent mefloquine use lend some evidence for a relationship with psychiatric outcomes. Mefloquine users who use another medication for a chronic illness (Handschin et al., 1997; Huzly et al., 1996) or who drink alcohol while taking mefloquine (van Riemsdijk et al., 1997a) appear to have an increased risk for adverse events. Furthermore, women (Rendi-Wagner et al., 2002; Schlagenhauf et al., 1996; Wernsdorfer et al., 2013) and individuals with low BMI (van Riemsdijk et al., 2002b, 2003) may be at increased risk for adverse psychiatric symptoms when taking mefloquine. A number of published case reports suggest that persistent psychiatric symptoms (including anxiety, depression, insomnia/exhaustion, paranoia, hallucinations, visual illusions, mania, depersonalization, or suicidal ideation) may be associated with mefloquine use and continue beyond the period after drug exposure has ended. Again, these findings support the plausibility of persistent adverse events, but they are inherently limited in the quality of scientific evidence that they can provide. However, a recent Cochrane review concluded that mefloquine is safe during pregnancy (González et al., 2018).

Animal and in vitro studies indicate that mefloquine may negatively affect processes relevant to psychiatric conditions. Mefloquine can affect processes that may in turn interfere with brain circuits regulating mood and cognition, including calcium homestasis (synaptic signaling), oxidative stress (managing energetic challenge), and connexins (intercellular communication). In particular, mefloquine’s binding to serotonin receptors suggests possible interactions with signaling processes relevant to mood regulation. However, the data from these experimental studies do not definitively explore mefloquine exposures relevant to prophylaxis doses or use. Moreover, the studies have measured endpoints immediately after mefloquine administration, making it difficult to address persistent or latent pathologic changes.

The most weight for evidence of an association between use of mefloquine and persistent or latent psychiatric adverse events comes from the seven epidemiologic studies that examined psychiatric outcomes that occurred at least 28 days following cessation of mefloquine (Eick-Cost et al., 2017; Meier et al., 2004; Schlagenhauf et al., 1996; Schneider et al., 2013; Schneiderman et al., 2018; Tan et al., 2017; Wells et al., 2006). The seven studies each used different methods for measuring outcomes, and the psychiatric outcomes of interest varied across studies. Considering the studies of U.S. military or veteran populations, Eick-Cost et al. (2017) examined adjustment disorder, anxiety disorder, depressive disorder, PTSD, psychoses, suicide ideation, paranoia, hallucinations, insomnia, and death by suicide using clinical diagnoses coded in DoD administrative databases. Wells et al. (2006) also used clinical diagnoses coded in DoD administrative databases to examine “mental disorders” as a diagnostic category and specific psychiatric diagnoses of somatoform disorder, mood disorder, anxiety disorder, PTSD, substance use disorders, personality disorder, adjustment reaction, “mixed syndromes,” and “other disorders.” And Schneiderman et al. (2018) used standardized self-report instruments to examine outcomes of PTSD, thoughts of death or self-harm, other anxiety disorders, and major depression. Both studies of UK travelers used clinical diagnoses coded in a health care administrative database to examine incident psychiatric outcomes. Meier et al. (2004) included depression, psychoses, panic attacks, and death by suicide among people aged 17–79 years, and Schneider et al. (2013) examined depression and anxiety, stress-related disorders, and psychoses as a group in individuals aged ≥1 year (Schneider et al., 2013). Schlagenhauf et al. (1996) used standardized self-report instruments to examine sleep disturbances, mood changes, unusual or vivid dreams, and phobias in travelers who had taken mefloquine for short-term malaria prophylaxis after visiting a Swiss vaccination center.

In their analysis of returned U.S. Peace Corps volunteers, Tan et al. (2017) used unverified self-reported symptoms of depression, anxiety, and insomnia to derive clinical diagnoses of major depressive disorder, bipolar disorder, anxiety disorder, schizophrenia, obsessive-compulsive disorder, and “other.” Findings related to PTSD are considered separately, below.

While all seven of these studies have methodologic limitations, the five that, based on their methodologic quality, provided the strongest evidence for examining the presence of persistent psychiatric outcomes are Eick-Cost et al. (2017), Meier et al. (2004), Schneider et al. (2013), Schneiderman et al. (2018), and Wells et al. (2006). Four of the studies (Eick-Cost et al., 2017; Meier et al., 2004; Schneider et al., 2013; Wells et al., 2006) evaluated data from administrative databases with clinically diagnosed outcomes, included at least two comparison groups in the analyses, applied a reasonably thorough set of covariates to the analyses of effect estimates, and measured the psychiatric outcomes of interest systematically and objectively, based on medical care visits and coded in the database. Although both Eick-Cost et al. and Wells et al. used data from DoD administrative databases, they used different years, and Wells et al. limited diagnoses to hospitalizations, which would suggest that the outcomes reported in Wells et al. were of greater severity than those in the Eick-Cost et al. sample, limiting the cross-study. The Schneiderman et al. (2018) study was somewhat less rigorous as the researchers based their exposure and outcome assessments on self-report. Both exposure and outcomes were systematically obtained, and psychiatric outcomes were measured by standardized psychometric instruments. The sample was large and adequately powered, and the investigators used a reasonably thorough set of covariates in analyses of effect estimates. Again, the difference in ascertainment of data limits comparison of data across studies.

In their analysis of active-duty service members, Eick-Cost et al. (2017) found that with the exception of psychoses and death by suicide, the adjusted incident rates for psychiatric outcomes were higher among the deployed groups who used mefloquine than among the nondeployed groups who used mefloquine. When comparisons between mefloquine and doxycycline use were stratified by deployment, the only statistically significant difference for any of the psychiatric outcomes for the deployed was a slight increased risk for anxiety disorders among mefloquine users. Among the nondeployed, mefloquine users had statistically significantly decreased risks of adjustment disorder, insomnia, anxiety disorder, depressive disorder, and PTSD compared with doxycycline users, but no differences were found for the other five psychiatric outcomes. In comparisons of mefloquine users and A/P users by deployment status, no outcomes were statistically significantly different for the deployed, but in the nondeployed group, mefloquine users had an increased risk of PTSD, although no other psychiatric outcomes showed differences in risk between mefloquine and A/P users. For both the mefloquine and doxycycline groups, individuals with a psychiatric diagnosis in the year preceding the prescription had statistically significantly elevated risks for a subsequent diagnosis of the same condition for all conditions reported (adjustment disorder, anxiety, insomnia, depressive disorder, and PTSD) compared with individuals without a diagnosis in the prior year. However, when the IRRs comparing mefloquine and doxycycline users were stratified by those with and without prior psychiatric diagnoses, there were no statistically significant differences between mefloquine and doxycycline for any of the conditions. The results of a sensitivity analysis in which the risk period was restricted to 30 days post-prescription were not reported, although the authors stated that the results were similar to the primary analyses. Similarly, in their analysis of service member hospitalizations Wells et al. (2006) reported a total of 37 hospitalizations for mental disorders as a category for mefloquine users, and the rate of hospitalizations was not statistically significantly different from the two comparison groups. When hospitalizations due to specific psychiatric outcomes were considered, there were no cases of somatoform disorders, 6 cases each of mood disorders and anxiety disorders, 1 case of PTSD, 19 cases of substance use disorders, 7 cases of personality disorders, 13 cases of adjustment reactions, 4 cases of mixed syndromes, and 20 cases of “other disorders” among mefloquine users, which resulted in imprecise and null effect estimates when these rates were compared with those of the two reference groups of service members. Using a large population-based cohort of deployed and nondeployed U.S. veterans, Schneiderman et al. (2018) found that, like Eick-Cost et al., deployed mefloquine users had higher frequencies of mental health diagnoses than nondeployed mefloquine users for the four psychiatric outcomes examined. However, in the adjusted logistic regression models with all covariates considered (including demographics, deployment, and combat exposure), mefloquine was not associated with any of the psychiatric outcomes examined: composite mental health score, thoughts of death or self-harm, other anxiety, and major depression. It is noteworthy that adjusting for combat exposure consistently reduced the measures of association, but when combat exposure intensity was specifically considered, the weighted prevalence estimates indicated that the prevalence of disorders increased with greater combat exposure intensity. This study could not address explicitly the health experiences during use and in specific time intervals following the cessation of use. Overall, the studies in military service members and veterans were well designed and provide some evidence for an absence of increased risk of persistent or latent psychiatric outcomes in mefloquine users.

Factors that may be present in groups of military or veterans that may confound associations between the use of mefloquine and adverse psychiatric events, such as deployment and combat exposure, are rarely encountered with leisure travelers. The results of Meier et al. (2004) and Schneider et al. (2013), who used UK travelers and stratified by time post-cessation corroborated the findings of Eick-Cost et al. (2017) and Schneiderman et al. (2018) in that the use of mefloquine was not associated with an increased risk of depression diagnoses in either the cohort analysis or the nested case–control studies. Schneider et al. (2013) found that in the adjusted analyses, the odds of developing an incident diagnosis of depression was statistically significantly decreased in both current and past mefloquine users compared with nonusers. Meier et al. (2004) also found no difference in the risk of developing depression for recent mefloquine users versus all past users of other antimalarials. Schneider et al. (2013) found that when the data were stratified by current use or past use, the adjusted odds of anxiety, stress-related disorders, or psychosis as a group were no different in current users but were statistically significantly reduced in the past users of mefloquine compared with nonusers. When anxiety, psychosis, phobia, and panic attack were analyzed as separate outcomes with no timing stratifications, compared with nonusers of antimalarials, only the odds of anxiety were statistically significantly decreased for mefloquine users (which was consistent with the findings of Eick-Cost et al.). Meier et al. (2004) found that first-time diagnoses of panic attacks and psychosis were not statistically significantly different for recent users of mefloquine compared with all past users of antimalarials, but the odds of panic attacks were statistically significantly increased in the adjusted nested case–control analysis. Both Meier et al. and Schneider et al. excluded people who had previously been diagnosed with the psychiatric outcomes of interest from their study populations. In sum, the studies of travelers corroborate the findings of studies of service members and veterans, and the use of stratification of post-cessation time, particularly past use, provides some evidence for an absence in—and possibly even a reduction in—persistent psychiatric effects of anxiety, stress-related disorders, or psychoses as a group, depression, and panic disorder following the use of mefloquine, but the small number of incident diagnoses for these psychiatric disorders does not provide definitive evidence of no effect.

The two other studies considered by the committee that presented some information on psychiatric outcomes (Schlagenhauf et al., 1996; Tan et al., 2017) were not as methodologically robust as Eick-Cost et al. (2017), Wells et al. (2006), Schneider et al. (2013), Meier et al. (2004), or Schneiderman et al. (2018), and therefore their findings were given less weight. However, the results of these two studies overall lend additional weak support for an absence of persistent or latent psychiatric events.

Gastrointestinal Disorders

The most recent FDA package insert for mefloquine states that the most frequently observed adverse event in clinical trials of malaria prophylaxis was vomiting (3%), while postmarketing surveillance found the most frequently reported gastrointestinal adverse events to be nausea, vomiting, loose stools or diarrhea, and abdominal pain, but the duration of such symptoms is not detailed. Systematic reviews of adverse events in travelers who used mefloquine compared with other regimens, placebo, or no antimalarial drug included concurrent gastrointestinal symptoms (abdominal discomfort or pain, anorexia, diarrhea, nausea, vomiting, dyspepsia). In the systematic review by Croft and Garner (2000), no consistent pattern was seen for the gastrointestinal symptoms analyzed, due in part to the heterogeneity of studies, but abdominal discomfort was reported statistically less frequently with other antimalarial drugs, as were anorexia and nausea. In a second systematic review examining the adverse events of mefloquine prophylaxis among travelers that included two randomized controlled trials and three cohort studies, mefloquine recipients were statistically significantly more likely to experience nausea than placebo recipients, but there was no difference between groups for vomiting, abdominal pain, or diarrhea. Based on cohort studies that compared mefloquine users with doxycycline users, mefloquine users were statistically significantly less likely to report dyspepsia and vomiting, but these results were given low or very low certainty of evidence, respectively. However, among pregnant women using mefloquine for intermittent preventive treatment in pregnancy, for which the dosage used is substantially higher than the dosage used for malaria prophylaxis, mefloquine was associated with a statistically significantly higher risk of drug-related vomiting and higher rates of nausea compared with use of sulfadoxine-pyrimethamine, but these symptoms all were reported to resolve spontaneously within 3 days.

Published individual case reports that had follow-up of at least 28 days post-mefloquine-cessation did not report on gastrointestinal disorders. The FDA package insert warns that mefloquine elimination may be prolonged in people who have impaired liver function, which may lead to higher plasma levels and a higher risk of adverse events. In a small study, DeSouza (1983) found that liver and spleen enlargement was reduced among the mefloquine participants (and sulfadoxine/pyrimethamine participants) over the course of follow-up. In one case series (Croft and Herxheimer, 2002) that reviewed case reports of adverse event reports associated with the use of mefloquine, the researchers hypothesized that adverse events may be due to liver or thyroid pathology; however, no objective validation of the adverse events reported by the cases or other follow-up was conducted, among other limitations of this analysis.

Biologic plausibility data on gastrointestinal effects are lacking. While there is some evidence of mefloquine action on β-cells, no experimental studies have provided data on mechanisms to support the potential for observed gastrointestinal disorders to become persistent.

The committee reviewed several epidemiologic studies that examined gastrointestinal disorders and outcomes that occurred during or immediately after (within 28 days of) mefloquine use, but because they did not follow or report on these adverse events 28 days post-cessation, the results are not considered to contribute to the evidence base of persistent gastrointestinal events post-mefloquine-use. Only Wells et al. (2006), based on the strength of the methods used in that analysis, was considered to provide robust evidence for gastrointestinal disorders that occurred or persisted at least 28 days following the cessation of mefloquine. Using ICD-9-CM codes, Wells et al. grouped disorders of the digestive system and found that 23 mefloquine users were hospitalized for these disorders. When mefloquine users were compared with deployed service members who were not prescribed an antimalarial, there was no difference in the risk of digestive system disorders, but compared with those service members who were assigned to Europe or Japan, deployed mefloquine users had a statistically significantly lower risk of hospitalization for digestive system disorders. This study provides some evidence for an absence of increased risk for serious persistent digestive system disorders following the use of mefloquine, but it is unclear if some of these concurrent adverse events persisted or if concurrent events preceded persistent outcomes that may not resolve without additional treatment. Tan et al. (2017) lends additional weak support (given its serious methodologic limitations) to an absence of increased risk of persistent gastrointestinal disorders following use of mefloquine.

Based on the available evidence, the committee concludes that there is insufficient or inadequate evidence of an association between the use of mefloquine for malaria prophylaxis and persistent or latent gastrointestinal events. Current evidence does not suggest further study of such an association is warranted, given the lack of evidence regarding biologic plausibility, adverse events associated with concurrent use, or findings from the existing epidemiologic studies.

Eye Disorders

Although there are reports of concurrent visual disturbance including optic neuropathy and retinal disorders associated with mefloquine use (FDA, 2016; Tickell-Painter et al., 2017a), in the epidemiologic studies that examined persistent eye disorders, these effects were not observed to occur at statistically different rates for mefloquine users than for people who used other antimalarial drugs or who did not use any prophylaxis. Among the case reports, concurrent adverse events included visual illusions and one case of persistent retinopathy. A large analysis of eye disorders associated with mefloquine use reported to the manufacturer’s drug safety database provides additional indirect support for adverse events of visual acuity and disorders affecting the retina or cornea (Adamcova et al., 2015). In addition to the available data on eye disorders in humans, experimental data may support plausible biologic mechanisms for mefloquine affecting ocular components, acting via a disruption of connexin signaling in the lens and possible phototoxic changes in the retina.

Of the 11 epidemiologic studies on persistent adverse events, 2 made a mention of eye disorders that occurred at least 28 days following the cessation of mefloquine (Schneider et al., 2014; Tan et al., 2017). Given the serious methodologic limitations of Tan et al. (2017), only Schneider et al. (2014) was considered, based on the strength of the methods used in that analysis, to provide robust evidence for persistent ophthalmic outcomes. Schneider et al. (2014) assessed incident diagnoses of eye disorders among travelers aged ≥1 year who had been prescribed mefloquine and compared them with two other groups of travelers: travelers who had been prescribed another antimalarial and travelers who had a travel consult but were not prescribed antimalarial drugs. Eye disorders were grouped into eight categories, some specific (such as cataract, glaucoma, and vitreous detachment) and others a compilation of disorders of the cornea, retina, visual acuity, uvea, and neuro-ophthamology. The timing of incident diagnoses was stratified into “current use,” which mixed irrelevant (7–28 days post-cessation) and relevant (28–90 days post-cessation) time periods, and “past use” (91–540 days post-cessation), all of which was relevant. Over the approximately 8.5-year period of data examined, a total of 85 people who had used mefloquine were diagnosed with an incident eye disorder of interest; 23 incident eye disorders were found for current users, and 62 were found for past users. A nested case–control analysis found that the odds of developing any of the eye disorders of interest were statistically significantly elevated for mefloquine users compared with travelers who did not use any antimalarial drugs. However, when mefloquine use was stratified by current use and past use and the users compared with the nonusers, there was no statistically significant difference for current users, although past users had statistically significantly increased odds of experiencing any eye disorder when all were grouped as a single category. When each of the individual eye disorders was examined without timing stratification, only cataract was statistically significantly related to mefloquine use compared with no use of antimalarials. Other risk factors for cataracts, such as occupation and sun exposure, were not included in the analysis and may have differed between the groups. Overall, this was a well-designed study, and the stratification of past use in particular provides some evidence for an absence of increased risk of persistent eye disorder diagnoses following the use of mefloquine. The findings of no differences in risk of ophthalmologic disorders of macular degeneration, retinopathy, and “any” ophthalmologic disorder by Tan et al. (2017) provide additional weak supportive evidence of an absence of increased risk of eye disorders. However, the finding of increased risk of cataracts with mefloquine use in Schneider et al. (2014) requires confirmatory evidence.

Based on the available evidence, the committee concludes that there is insufficient or inadequate evidence of an association between the use of mefloquine for malaria prophylaxis and persistent or latent eye disorders, including cataract. Current evidence suggests further study of such an association is warranted, given the evidence regarding biologic plausibility, adverse events associated with concurrent use, or findings from the existing epidemiologic studies.

Cardiovascular Disorders

The most recent FDA package insert for mefloquine states that syncope and extrasystoles were reported in less than 1% of mefloquine users participating in clinical trials of malaria prophylaxis. Other concurrent adverse events reported with the use of mefloquine have included transitory and clinically silent ECG alterations such as sinus bradycardia, sinus arrhythmia, first degree AV block, prolongation of the QTc interval, and abnormal T-waves. Among the case reports that followed outcomes at least 28 days post-cessation of mefloquine, heart palpitations were reported in one case in which concurrent symptoms of paralysis and trouble breathing were also reported (Eaton, 1996). The available biologic plausibility data on cardiovascular effects are limited, but some data suggest that mefloquine may induce cardiovascular effects through the inhibition of several cardiac potassium channels. Mefloquine may also affect intracellular calcium homeostasis in cardiac myocytes, suggesting some potential for cardiac indications, although this was not tested in the context of persistent or latent actions.

The committee reviewed four epidemiologic studies that examined cardiovascular or circulatory system outcomes that occurred at least 28 days following the cessation of mefloquine (DeSouza, 1983; Laothavorn et al., 1992; Tan et al., 2017; Wells et al., 2006). Similar to the other body system outcome categories, cardiovascular and circulatory system outcomes were inconsistently identified and measured across studies. DeSouza (1983) used ECGs and measured blood pressure and pulse rate, as well as hematologic parameters of red blood cell count, hemoglobin erythrocyte volume fraction, total and differential white blood cell counts, reticulocyte count, and platelet count. Measurements of other biochemical parameters (including cholesterol, triglycerides, glucose, urea, creatinine, etc.) in sera were also performed. Laothavorn et al. (1992) also used ECGs to measure heart rate and different cardiac intervals and to diagnose abnormalities of sinus bradycardia, sinus arrhythmia, ventricular ectopic beats, atrial ectopic beats, atrial–ventricular block, and heart rate; they also performed weekly blood count tests. The two other studies (Tan et al., 2017; Wells et al., 2006) grouped cardiovascular outcomes. In Tan et al. the cardiac category included arrhythmia, congestive heart failure, myocardial infarction, and “any” cardiac disorder, while Wells et al. grouped outcomes by ICD-9-CM code into disorders of the blood and blood-forming organs and a separate category of disorders of the circulatory system.

While none of these studies is without methodologic limitations, Wells et al. (2006) provided the most robust evidence regarding persistent cardiovascular and circulatory system outcomes. In short, only four hospitalizations related to blood and blood-forming organs (ICD-9-CM: 280–289) and nine hospitalizations from circulatory system disorders (ICD-9-CM: 390–459) were reported for mefloquine users. Comparisons with both reference groups showed that mefloquine users had no difference in risk for both groups of disorders, providing some evidence for an absence of increased risk of persistent disorders of blood or blood-forming organs or the cardiovascular system following use of mefloquine. The results from the three other epidemiologic studies lend additional support, although of less weight, for an absence of increased persistent cardiovascular events. Tan et al. (2017) reported that there were no statistically significant differences in cardiac outcomes between users of mefloquine and of the other antimalarial drugs for prophylaxis, but they did not provide frequencies of the events or effect estimates. Although both DeSouza (1983) and Laothavorn et al. (1992) used objective tests (ECGs) and standard hematologic and laboratory measures in their investigations, the presented results are not readily comparable between studies and were sometimes vague. DeSouza stated that blood pressure, pulse rate, and ECG remained normal throughout the study period (63 days after mefloquine administration), but no other details regarding the ECG results were provided. Hematologic tests were conducted several times throughout the study, but only those taken on days 28 and 63 post-administration were relevant to the committee’s work. No significant adverse changes were reported for any of the collected parameters for the group administered mefloquine. Laothavorn et al. performed ECGs on healthy volunteers prior to mefloquine administration, daily for 1 week post-administration, and then weekly until day 42 post-administration. All ECG parameters were reported to be within normal limits, and no changes in biochemical or hematologic measures were found following mefloquine administration. Although the results from the DeSouza and Laothavorn studies appear to be consistent with an absence in increased persistent events of cardiovascular or circulatory disorders following use of mefloquine—especially considering that the administered doses of mefloquine were 3–4 times higher than the dose used for prophylaxis—both of these studies were small and underpowered and were limited in the information reported. The concurrent events listed in the FDA package insert were not found to occur in the epidemiologic studies that measured them.

Based on the available evidence, the committee concludes that there is insufficient or inadequate evidence of an association between the use of mefloquine for malaria prophylaxis and persistent or latent cardiovascular events. Current evidence does not suggest further study of such an association is warranted, given the lack of evidence regarding biologic plausibility, adverse events associated with concurrent use, or findings from the existing epidemiologic studies.

Other Outcomes and Disorders

In addition to those outcomes synthesized above, two of the epidemiologic studies examined other outcomes and disorders that occurred at least 28 days following the cessation of mefloquine. Tan et al. (2017) included dermatologic outcomes as a group that included allergic dermatitis, eczema, psoriasis, “other” and “any” dermatologic conditions and reported that there were no statistically significant differences between users of mefloquine and those of the other antimalarial drugs for prophylaxis, but neither the frequencies of such events nor effect estimates were provided. Wells et al. (2006) reported nine hospitalizations from skin and subcutaneous tissues (ICD-9-CM: 680–709) among mefloquine users in their study of U.S. service members, but no difference in risk was found between deployed service members who were prescribed mefloquine and those who were not prescribed an antimalarial. In one systematic review, reports of fever and pruritus were similar in the mefloquine and comparator arms (Croft and Garner, 2000). In a second systematic review, skin and subcutaneous tissues outcomes (pruritus, photosensitivity, vaginal candida) were examined, and based on data from cohort studies, mefloquine users were statistically significantly less likely than doxycycline users to report photosensitivity or vaginal thrush, but both findings were based on very low-certainty evidence. In the case reports, one case of worsening psoriasis was reported (Potasman and Seligmann, 1998). In a case series (Smith et al., 1999) of 74 published case reports of mefloquine use (prophylaxis or treatment) specific to dermatologic adverse events, the timing of onset of dermatologic effects was only recorded in 11 of the cases; pruritus and itching were reported in more than 40% of all the cases. Most effects were reported as mild or moderate in intensity and usually self-limiting, although the timing was not specified. Other dermatologic adverse events in this case series included two reports of cutaneous vasculitis and one report each of Stevens-Johnson syndrome and toxic epidermal necrolysis. In sum, several studies of varying quality have examined skin disorders associated with the use of mefloquine, but taken as a whole there is some evidence for an absence of increased risk of persistent skin and subcutaneous tissue disorders following use of mefloquine.

Wells et al. (2006) also reported hospitalizations for other system disorders among active-duty U.S. service members. In total, 135 hospitalizations for any cause were reported among mefloquine users, but there was no statistical difference in the risk compared with deployed service members who were not prescribed an antimalarial. Hospitalizations related to categories of infections; neoplasms; disorders of endocrine, nutritional, or metabolism; disorders of the respiratory system; disorders of the genitourinary system; disorders of musculoskeletal and connective tissue; ill-defined conditions; and injury and poisoning were also examined and compared between mefloquine users and the two reference groups. Comparisons of mefloquine users with deployed service members who were not prescribed an antimalarial resulted in a mix of increased and decreased effect estimates for categories of neoplasms; disorders of endocrine, nutritional, or metabolism; disorders of the respiratory system; disorders of the genitourinary system; disorders of musculoskeletal and connective tissue; ill-defined conditions; and injury and poisoning, but none reached statistical significance. Although methodologically limited, Tan et al. (2017) reported that reproductive outcomes (miscarriage), infections (amebiasis, giardia, “other” and “any” gastrointestinal infection), and hematologic/oncologic disorders (breast cancer, gastric cancer, leukemia, liver cancer, lymphoma, prostate cancer, “other” and “any” cancers) were not statistically significantly different between users of mefloquine and the other antimalarial drugs for prophylaxis, but frequencies of such events or effect estimates were not provided. In sum, several categories of other outcomes were examined for differences in risk associated with use of mefloquine, and there is some limited evidence for an absence of increased risk of persistent adverse events for any of those categories of disorders following the use of mefloquine.