10

Improving the Quality of Research on the Long-Term Health Effects of Antimalarial Drugs

The committee was charged with reviewing the available scientific evidence regarding the prophylactic use of Food and Drug Administration (FDA)-approved antimalarial drugs, particularly those that were used by U.S. service members or were of interest to the Department of Veterans Affairs (VA), and their persistent or latent adverse health effects, with a focus on neurologic and psychiatric conditions. This report is an assessment of the evidence with a focus on the published research, supplemented with other evidence as available (such as national and foreign government reports, responses to committee-generated information requests, and information submitted by the public through invited presentations, comments, and data submissions) relating the use of antimalarial drugs to adverse health effects, with specific consideration of the quality and quantity of studies and their findings.

The previous six chapters provide comprehensive assessments of the literature pertaining to each of the individual antimalarial drugs of interest (mefloquine, tafenoquine, atovaquone/proguanil [A/P], doxycycline, primaquine, and chloroquine) and their association with adverse events that might occur in any organ system. Those chapters offer integrated summaries and assessments of the evidence for each drug and each type of adverse event, organized by body system (neurologic, psychiatric, gastrointestinal, eye, cardiovascular, and other outcomes and disorders), and those specific assessments will not be repeated here. In this chapter the committee reflects more broadly on the current overall state of scientific knowledge regarding persistent and latent adverse events of the antimalarial drugs of interest when used for malaria prophylaxis and how to best advance the understanding of possible persistent events of antimalarial drugs. The committee was not asked to design the “gold standard” epidemiologic study for future research on this topic. However, following its review of the studies, the committee was able to identify

areas where the methodologic rigor could be strengthened in order to guide future research efforts that would then allow researchers to make stronger inferences and conclusions. This is in response to the penultimate sentence of its Statement of Task, “Additionally, the committee will consider approaches for identifying short-term, long-term, and persistent adverse health effects of antimalarials.”

As noted in the drug-specific chapters, there is a sharp contrast between the abundant amount of literature pertaining to concurrent adverse events that are experienced while a drug is being used or shortly following its cessation and the dearth of information, especially high-quality information, pertaining to adverse experiences after the use of that drug has ended. To assess the persistent effects of exposure to antimalarial drugs the committee opted to use a conservative cutoff time of ≥28 days (which was considered equivalent to expressions of 4 weeks or 1 month) post-cessation of drug intake to differentiate between events that are concurrent (outside the committee’s scope) and those that are persistent or latent (within the committee’s scope). Because some terms are used interchangeably in the literature and may have very different connotations depending on their context, the committee endeavored to use language that allowed it to be as precise as possible. As such, instead of “long term,” which may refer to the duration of drug use for prophylaxis or to the duration or timing of symptoms, the committee uses “persistent” and “latent” to describe associations with adverse events after the use of a drug has ended. The committee defined persistent adverse events as those adverse events that began during the period in which the drug was used and continued after its cessation beyond the period that the drug would still be present, which is defined as ≥28 days post-cessation. Latent adverse events are those adverse events that were not apparent during the period the drug was in use but that were present at any time (i.e., ≥28 days, after the cessation of antimalarial prophylaxis). The focus of the committee’s assessment was research that examined persistent or latent adverse effects, both of which indicate adverse health outcomes that extend beyond user experience while taking the drug.

ATTRIBUTES OF AVAILABLE RESEARCH

The currently available body of high-quality research addressing the use of antimalarial drugs for malaria prophylaxis (some of which have been in use for more than 70 years) and persistent or latent adverse effects is quite limited, even when combined across all the drugs of interest and all organ systems and types of possible adverse events. There appears to be a disconnect between the level of concern raised—millions of people have used the drugs, and there are many known concurrent events and case reports of adverse events—and the systematic research that has been conducted, particularly in areas such as the use of mefloquine and persistent neurologic or psychiatric outcomes. As reflected in the chapter syntheses, only a small subset of studies, mostly conducted in military populations or

travelers, provide the most relevant and informative evidence regarding persistent adverse events. A few of those studies compared the occurrence of adverse events across several antimalarial drugs of interest. From all of the studies considered and assessed by the committee, only about half (Ackert et al., 2019; Eick-Cost et al., 2017; Green et al., 2014; Leary et al., 2009; Meier et al., 2004; Nasveld et al., 2010; Schneider et al., 2013, 2014; Schneiderman et al., 2018; Wells et al., 2006) of the 21 epidemiologic studies that met the committee’s inclusion criteria were considered to be the most informative due to their methodologic attributes of having sufficient quality of exposure and outcome information, being of sufficient size to potentially provide adequate statistical power (especially for rare outcomes), and presenting data on adverse events that occurred or persisted 28 days or more post-drug-cessation. Each of these studies has its own limitations, but they were determined to be the most informative for addressing the question of whether there is evidence of persistent or latent adverse health outcomes associated with the prophylactic use of antimalarial drugs.

A number of randomized controlled trials were identified that provide rigor in their control of confounding, but most did not extend the follow-up period to at least 28 days beyond the duration of the antimalarial drug use. For those that did, they were generally were too small to yield information on any but the most common adverse events, or the published study had poor documentation of adverse events. In addition, the committee notes that the randomized trials were primarily designed to study tolerability or efficacy, and, as a result, they were generally non-informative about persistent or latent adverse events. Although many of the randomized controlled trials and a much larger body of observational studies did not meet the committee’s inclusion criteria because they did not report on or distinguish between outcomes at least 28 days post-drug-cessation, these studies do contribute to the evidence base regarding concurrent adverse events. The research base for concurrent adverse events is substantial and shows a consistent pattern of outcomes that are associated with the tolerability and safety of these drugs, as detailed in each chapter. While concurrent adverse events are only indirectly relevant to the charge of examining persistent adverse events, they do provide an indication of particular body systems, symptoms, and diagnoses to focus on, assuming that the problems most likely to be persistent are those that initially manifest during drug use. That scenario is more commonly observed and plausible in the committee’s view than true latent effects which arise de novo at some time after exposure to the drug of interest has ceased. Although it would be ideal to have the entire time course available, beginning from exposure to an antimalarial drug to years post-cessation, and symptom manifestations that occurred at multiple time points, the literature is so limited that the committee was not able to be more specific about adverse events that were likely to persist in different time intervals (e.g., 1–6 months, 6–12 months, >12 months). As a result, all adverse events that persisted for at least 28 days were presented. Several published case reports and case series were identified that provided supportive information on adverse events,

some of these being rare outcomes, that arose in conjunction with taking an antimalarial drug and persisted for some period after the cessation of drug use. The case reports varied in quality and detail. While these reports cannot contribute to causal inference in part because of a lack of comparison groups, they can direct attention to and inform areas or health outcomes that merit more methodologically rigorous evaluation for specific drugs.

The biologic effects of the various antimalarial drugs are relatively well understood with regard to their effectiveness in preventing clinical malaria and aspects of acute toxicity, but there is a very limited body of research that directly addresses the pathways by which these drugs might result in persistent changes that produce adverse events that may or may not be reversible. In general, while the animal and in vitro studies support different biologic actions of the antimalarials, the published experimental research has not rigorously tested biologic plausibility in its fullest sense with regard to the impact of prolonged treatment (as would occur in prophylaxis) of relevant doses on well-defined behavioral and neurologic endpoints. Most studies reviewed involve acute drug treatment, which may or may not be relevant to long-term administration in the setting of prophylaxis; involve supratherapeutic dosing in laboratory-based animal in vivo studies; or involve the use of in vitro systems that do not duplicate the full context of prophylaxis regimens. While the data provide hints of processes that may be relevant to the central question at hand (the plausibility of pathology following prolonged treatment), relatively few address it directly (which is highlighted in the drug-specific chapters where relevant). The pathways by which drug use for a defined period leads to irreversible biological changes that manifest as clinically recognizable symptoms or diagnoses have simply not been pursued. Many of the available basic science experimental studies have examined outcomes that do not directly link to recognizable clinical symptoms or manifestations in humans. As a result, there is very little information that can be gleaned from these types of basic science studies to provide information about the mechanisms of the adverse events observed in the epidemiologic studies.

Cumulatively, while the available research to date points toward possible avenues to pursue based on acute adverse events and case reports, the small number of directly pertinent studies precludes drawing firm conclusions. The committee has attempted to glean the most information that the scientific literature offers but, at most, only tentative inferences are possible.

QUALITY OF METHODS OF REVIEWED STUDIES

Several methodologic considerations were introduced in Chapter 3 that the committee used to assess the quality of individual studies. Providing a detailed appraisal of the methodologic quality of each of the identified epidemiologic studies allowed the committee to offer tentative inferences from very limited evidence with a clear appreciation of how fragile those inferences were. For the epide-

miologic studies, those principles included study design (population, sample size, comparison groups), exposure assessment, outcome assessment, and confounding.

Study Design

As this work is at the request of the Department of Veterans Affairs (VA), the specific population of greatest interest is U.S. military service members and veterans. However, antimalarial prophylaxis is not limited to these two groups, and it is reasonable to assume that research conducted in other populations may provide relevant information regarding the persistence of adverse events following the prophylactic use of antimalarial drugs, and thus, studies of non-military populations were included in the committee’s assessment. An important consideration in incorporating any evidence from non-military populations is the evaluation of differences between these groups and the military population of interest and assessing whether these differences may influence the interpretation of study results. In earlier chapters, results are reported of studies conducted using Peace Corps volunteers, pre-screened research volunteers (particularly those recruited for randomized clinical trials), travelers, and endemic populations. Recruitment into military service includes thorough health screenings and assessments that are likely very different from those experienced by, for instance, Peace Corps volunteers, and this premilitary screening may result in the exclusion of individuals with specific physical or psychological characteristics. If these health characteristics are contraindications to the use of specific prophylactic antimalarials, for example, then one may expect a lower incidence of the adverse outcomes in military and veteran populations as a result of screening out individuals at higher risk of adverse outcomes from the antimalarials of interest. Moreover, over the duration of their military careers, which may include multiple deployments or temporary duty assignments, service members may use a drug multiple times. People who experienced adverse outcomes while using a specific drug may elect (if given a choice) to use a different drug. This would result in a “depletion of susceptibles,” which means that people who may have adverse events are no longer in the risk pool, which could result in an observed lower incidence of adverse outcomes for a certain drug. In the studies of populations of military and veterans evaluated by the committee, prior users were either excluded or the use of multiple drugs was censored.

A further consideration of the populations studied is the location and intent of the travel. For military populations, travel is usually occupation related (i.e., deployment), and it likely results in stress prior to travel and, depending on the circumstances, additional stress and possibly trauma and combat while in the deployment location. Other characteristics of the deployment location may also confound the association between antimalarial prophylaxis and adverse outcomes, including environmental or other exposures (see Concurrent Exposures of Military Service in Chapter 2) that may adversely affect health. In studies of travelers, most often the purpose of the travel is for leisure, the location of the travel is chosen (rather than

mandated), the time is typically short term (weeks to few months), and the location is very unlikely to be an area of civil unrest or active conflict. There are limits to how informative studies based on leisure travelers or other populations, such as people living in malaria-endemic areas who may have naturally acquired immunity to malaria, can be to persistent adverse health effects in military or veteran populations.

Comparison Groups

The committee was asked to focus its assessment on the potential association between the use of any of the six FDA-approved antimalarial drugs for prophylaxis and persistent or latent adverse events. It was not asked to assess the efficacy of the antimalarials of interest, nor was it asked to compare the antimalarial drugs on the basis of toxicity. Ideally, research should be conducted to enable the comparison of each antimalarial separately against a meaningful comparison group. In the context of antimalarial prophylaxis, this is a difficult task, given that the indication for antimalarial use is travel to a malaria-endemic area. Antimalarials are highly recommended for such travel, so that it is difficult to identify, for comparison purposes, population subgroups who travel to malaria-endemic areas but do not take antimalarials. Furthermore, in the case of travelers to malaria-endemic areas who do not take antimalarials, the reasons for that choice could confound the associations reported from such studies unless the authors have collected information on the reasons for nonuse. For the most part, the epidemiologic studies have examined the adverse outcomes in groups that differ on which antimalarial they have used for malaria prophylaxis. Thus, the comparisons are necessarily relative (e.g., how does mefloquine compare to chloroquine?) rather than absolute (e.g., how does mefloquine compare to no antimalarial use?). The answers to both questions are relevant, but the conclusions that can be drawn about a specific antimalarial are limited by these designs: when comparing two antimalarial drugs, the inferences will be limited to statements that the two drugs are equivalent in their adverse event profile or that one drug is “more harmful” than the other drug. The latter conclusion implies that the other drug will appear as “less harmful,” but the absolute impact of the drug is unknown. Within studies of military personnel there are also comparisons of deployed versus nondeployed personnel (Eick-Cost et al., 2017; Schneiderman et al., 2018; Wells et al., 2006), but study designs could also conceivably compare deployment location (endemic area versus not, as in Wells et al., 2006). The committee did not find any studies that made comparisons across different types of antimalarial drug users (e.g., leisure travelers versus military personnel).

Exposure Assessment

Assessing exposure outside of a clinical trial (i.e., where the drug is assigned to an individual by a process of randomization and is often monitored for adherence) is also challenging. Medications are prescribed by a health professional,

and it is often up to the individual to fill the prescription and then to take the medication as prescribed. In research conducted using pharmacy or prescription databases, often the only available information is about the prescription dispensed (dose, regimen, number of tablets, date of prescription or dispensing), not about how well the individual adhered to the medication regimen. Details concerning medication adherence are often obtained through self-report, and individuals may be asked to recall medication details from the recent or even distant past (e.g., Schneiderman et al., 2018; Tan et al., 2017). In studies of people who are employed by or participate in organizations in which the use of antimalarial drugs for prophylaxis is required (e.g., military, Peace Corps, Department of State), reported adherence rates may be inflated. All these limitations add to the difficulty in being able to evaluate the role of the duration of medication use and the specifics of the doses taken. An additional challenge when studying adverse events of drugs is that the occurrence of adverse events may cause an individual to decide to modify the dose, or even stop the drug completely, without consulting a health professional. Individuals may decide to change their regimen when the adverse effect is minor and not serious enough to report but still is bothersome to the individual. Such changes are rarely recorded, and people may be reluctant to admit that they did not take the medication as prescribed. These individuals would still be counted in the exposed categories, but there would be no drug-associated adverse events. Another concern related to exposure assessment is the phenomenon known as the “depletion of susceptibles” (introduced under Study Design). This phenomenon can occur when the initiation of a drug is associated with acute effects early on, followed by a decrease in the frequency of the effects as time goes on. If individuals who stop taking the drug because of these early events are no longer followed, they do not contribute to the follow-up time. This is a form of selection bias in studies that do not use a new-user design and count all person-time in follow-up equally.

Along the same lines, most of the epidemiologic studies that met the committee’s inclusion criteria did not collect information on any previous use of antimalarials prior to the time of the study. For example, Eick-Cost et al. (2017) identified service members filling a prescription for mefloquine, doxycycline, or A/P between January 1, 2008, and June 30, 2013, but service members could have used one of these (or other) antimalarials prior to 2008. Those individuals who previously took an antimalarial and experienced no adverse effects would be more likely to take an antimalarial again and to adhere with the regimen and also less likely to experience adverse events. Without specific information on adherence to the antimalarials, the committee has to assume good adherence when assessing the studies unless the authors specifically noted otherwise. Thus there could have been unmeasured misclassification of drug use, with nonusers or suboptimal users being classified as users, potentially attenuating the associations between antimalarial use and adverse health outcomes. Studies that used drug dispensing records as their sole way of ascertaining drug exposures have certain limitations, including

uncertainty as to whether the dispensed medications were actually ingested (Eick-Cost et al., 2017; Schneider et al., 2013, 2014; Wells et al., 2006). Furthermore, as noted in Chapter 3, relying solely on dispensing records for determining exposure to medicines used to prevent a disease may lead to an overestimation of peoples’ exposure to a given medicine, particularly if there is reason to believe that the drug is associated with concurrent adverse events.

Outcome Assessment

The gold standard for outcome assessment is the clinical assessment of individuals at multiple time points. This is rarely feasible in epidemiologic research, but there are a variety of both active and passive sources of information that can be used when conducting research to identify persistent or latent adverse events. The first step, however, is to specify and define the outcomes of interest, preferably using standardized diagnoses or definitions of outcomes. Given the dearth of available and informative literature on the persistent or latent adverse events associated with the use of antimalarial drugs, this area of research is in its infancy, and often the specific outcome is not defined, but instead broad classes of outcomes are included, such as “gastrointestinal effects” or “neuropsychiatric disorders.” In the research that forms the basis of this report, even with broad classes of outcomes, definitions vary considerably. It is difficult to make comparisons across studies when not only the definitions or methods of adverse-event collection vary but also the sources of data are often different (i.e., self-report, medical records, administrative databases, etc.). A variety of methods were used by the different studies to elicit the occurrence of adverse events, including nonspecific questions (such as overall satisfaction of using a drug [Andersson et al., 2008]), checklists of symptoms or symptom diaries (e.g., Davis et al., 1996; Jaspers et al., 1996; Korhonen et al., 2007; Petersen et al., 2000; Rendi-Wagner et al., 2002; Tan et al., 2017), standardized instruments and tests (Boudreau et al.,1993; Schlagenhauf et al., 1996; Schneiderman et al., 2018), and International Classification of Diseases or other administrative coded diagnoses (e.g., Eick-Cost et al., 2017; Schneider et al., 2013, 2014; Wells et al., 2006).

A particular challenge in outcome assessment is that there needs to be clarity regarding the temporal sequence of the outcomes of interest (e.g., incident during prophylaxis, incident after cessation of prophylaxis, incident during prophylaxis and continuing after cessation, etc.). Without a known temporal sequence, the associations generated are difficult to interpret with regard to the timing of the drug use and the adverse events. In particular, measuring antimalarial exposure and outcomes at the same time (as in a cross-sectional survey) may lead to bias if respondents misremember which happened first (exposure or outcome), and, in fact, previous psychiatric conditions may actually lead to different antimalarial exposures as they are contraindicated for certain drugs.

Approaches to Assessing Neurologic and Psychiatric Outcomes

The assessment of neurologic and psychiatric outcomes, especially posttraumatic stress disorder (PTSD), which is included in the Statement of Task, may be challenging for a number of reasons. First, for a diagnosis of PTSD, the person should have been exposed to an identified traumatic event by directly experiencing it, witnessing it in person, learning that the traumatic event occurred to a close family member or close friend (with the actual or threatened death being either violent or accidental), or experiencing firsthand repeated or extreme exposure to aversive details of the traumatic event (not through media, pictures, television, or movies unless work related). The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria explicitly exclude medication from being the potentially traumatic event. Therefore, while some patients may experience symptoms of PTSD following exposure to a drug, PTSD symptoms must be related to experiencing trauma (e.g., combat). This may make a clear relationship between taking a medication and a resultant diagnosis of PTSD difficult to ascertain because the symptoms involved the exposure to trauma and their onset relative to taking a medication is hypothetical at best. It may be that exposure to a specific medication or several medications taken together confers an elevated risk for PTSD in the context of a different traumatic experience. It may be that concurrent adverse events associated with a medication may themselves be traumatic, but the current empirical literature and classification systems do not allow for such an assertion as the basis for a PTSD diagnosis. Second, for PTSD the reported symptoms should be directly related to the specific traumatic event reported by the person. Third, because there are currently no diagnostically valid and reliable biomarkers of PTSD, the diagnosis of any psychiatric outcomes is based on the patient’s self-reported experiences, which may be biased or distorted by memory processes known to be influenced by stress and fear. For certain groups (e.g., service members), there may be also incentives to minimize or deny the experience of stigmatized neuropsychiatric symptoms, such as when acknowledging such symptoms may result in the individual’s removal from duties, or there may be an incentive to over-report these symptoms, such as in the case of financial compensation related to disability. Symptoms or experiences such as nightmares, hallucinations, or paranoia may be particularly stigmatized and, as a result, possibly under-reported.

When PTSD was assessed in the reviewed studies (e.g., Eick-Cost et al., 2017; Schneiderman et al., 2018; Wells et al., 2006), the potentially traumatic events experienced previous to drug exposure and concurrently with drug exposure were not systematically assessed. Therefore, any associations between use of an antimalarial drug and PTSD symptoms were generally insufficiently ascertained. Confounders, for example, may influence such associations (e.g., PTSD could be wrongly diagnosed because the drug’s symptoms mimic PTSD).

A systematic assessment of DSM-5 or International Classification of Diseases, Tenth Revision criteria by a trained clinician can improve diagnostic accu-

racy, as can the use of psychometrically sound self-report assessment tools. The use of standardized methods of assessment can reduce, although not completely eliminate, potential reporting biases. Assessing the timing of a potentially traumatic event, medication exposure, and symptom onset and content is critical to answering the scientific question of whether there are persistent psychiatric outcomes associated with antimalarial use. Additionally, it is important to recognize that it may be normal to experience some symptoms of depression, PTSD, anxiety, etc., but meeting the full diagnostic criteria is rarer, further making the measurement of these outcomes challenging.

Drug-Associated Neurologic and Psychiatric Adverse Effects

The committee was charged with assessing the evidence for persistent adverse events, with an emphasis on neurologic or psychiatric events, that are associated with the use of antimalarial drugs when used for prophylaxis. The concurrent use of many prescription drugs has been associated with neuropsychiatric adverse events, and therefore the manifestations associated with antimalarials are not unique. Such outcomes include depression, agitation, confusion, psychosis, seizures, a change in the level of consciousness, and nightmares (Ruha and Levine, 2014). Although the mechanisms of some drug-associated neuropsychiatric effects have been elucidated, for many drugs the mechanisms remain unknown. In addition, there is little information about persistent effects for these types of events.

There is a body of evidence concerning persistent effects following the cessation of drugs associated with addiction (Korpi et al., 2015) as well as on tardive dykinesia, a movement disorder, following the cessation of antipsychotic drugs (Macaluso et al., 2016). However, these types of studies appear to be the exception. For other drugs with well-recognized and common acute neuropsychiatric effects, there is no information about persistence. For example, glucocorticoids, which are used for a wide variety of inflammatory conditions, and efavirenz, a very effective HIV-1 antiviral, are associated with a high frequency of such effects. Judd et al. (2014) reviewed neuropsychiatric effects associated with glucocorticoids. In one large study, patients taking glucocorticoids were four to seven times more likely to develop suicide/suicide attempt, delirium/confusion, and mania than age-, gender-, practitioner-, and disease-matched controls. The incidence of such outcomes approached 20% for those on high doses. Dalwadi et al. (2018) reviewed neuropsychiatric events associated with efavirenz, including abnormal dreams, sleep disturbance, anxiety, depression, and dizziness. The incidence of such adverse outcomes exceeds 50% in most studies. Symptoms improve over time for many, but not all, and the trajectory of symptoms after withdrawal is unknown.

Thus, neuropsychiatric symptoms are associated with many prescription drugs, and for some, like mefloquine, there is good evidence that the acute events are causally related to drugs used as prescribed. It is plausible that a prolonged

continuation of drugs that continue to produce disturbing neuropsychiatric signs and symptoms might lead to persistent effects after drug cessation. It is equally, if not more, plausible that drug-related signs and symptoms go away with drug withdrawal and that the persistence or recurrence of neuropsychiatric events after drug cessation would have occurred regardless of drug exposure.

Confounders

As noted above, there are medical contraindications for some of the antimalarials. While these contraindications (e.g., a previous psychiatric history) are not always followed either because this information is not available or because, when asked, the individual does not provide this information to the medical care provider, any tendency to preferentially give one drug versus another because of a history of health problems has the potential to introduce substantial confounding if not addressed explicitly in the analysis. Depending on the analysis methods used, such contraindications could possibly lead to a depletion of susceptibles, as discussed under the heading of Study Design, and result in findings of decreased risk of certain adverse events, such as specific psychiatric diagnoses, among mefloquine users compared with doxycycline or A/P users. Without the information on prior psychiatric history from the groups being studied, for example, the observed results may be confounded by the contraindication for use, and it is not possible to stratify by psychiatric history or restrict the analysis to those without a history of psychiatric disorders. Depending on the frequency of the contraindication and the difference in frequency across the comparison groups, the magnitude of this potential confounding bias will vary and thus be unknown in any given situation. Furthermore, as contraindications for selected antimalarials are introduced over time, studies will differ in their susceptibility to this bias in relation to the altered prescribing practices applicable at the time that drugs are being prescribed.

In addition, as described in earlier chapters, there are other exposures (e.g., concomitant drug exposures, combat, etc.) that may place individuals at a higher risk of experiencing adverse events. In military populations, there is a particular concern with the many other challenges associated with deployment in addition to any impact of antimalarial drug use. These would include the exposure to and threat of combat and trauma, social isolation resulting from separation from family and friends, and cultural dislocation from living in an environment notably different from home. If this information is not available for the drug groups being compared, then, again, confounding bias may result if the prevalence of these conditions differs by the type of drug used. At a minimum, it is important that the groups being compared are equally likely to have been deployed to a location potentially involved in combat in order to avoid substantial confounding. For example, in their analyses, Schneiderman et al. (2018) assessed exposure to combat as both a dichotomous variable (yes/no) and as a multiple-level variable to assess combat intensity.

COMPARISONS OF FINDINGS ACROSS ALL ANTIMALARIAL DRUGS OF INTEREST

A total of 21 primary epidemiologic studies (Ackert et al., 2019; Andersen et al., 1998; DeSouza, 1983; Eick-Cost et al., 2017; Green et al., 2014; Laothavorn et al., 1992; Leary et al., 2009; Lee et al., 2013; Lege-Oguntoye et al., 1990; Meier et al., 2004; Miller et al., 2013; Nasveld et al., 2010; Rueangweerayut et al., 2017; Schlagenhauf et al., 1996; Schneider et al., 2013, 2014; Schneiderman et al., 2018; Schwartz and Regev-Yochay, 1999; Tan et al., 2017; Walsh et al., 2004; Wells et al., 2006) were identified that met the committee’s inclusion criteria and were assessed in detail for the information they provided regarding persistent or latent adverse effects. Nine of these studies included multiple drugs of interest, and they contribute to the evidence in multiple chapters. A table that gives a high-level comparison (study design, population, exposure groups, and outcomes examined by body system) of each of these epidemiologic studies is presented in Appendix C. Just over half of the identified studies (11 primary studies) examined exposure to mefloquine. Fewer primary epidemiologic studies met inclusion for the other drugs of interest: tafenoquine, 7; A/P, 4; doxycycline, 7; primaquine, 4; and chloroquine, 3. From the perspective of biologic plausibility, the mechanistic links between antimalarial drugs and persistent or latent adverse outcomes have yet to be systematically and definitively explored through experimental studies, and the current literature in that area is not strong.

In general, five outcome categories emerged as the areas of greatest interest in the literature: neurologic, psychiatric, gastrointestinal, eye, and cardiovascular. Although for the majority of outcomes in this entire body of literature, no firm conclusions were warranted, in some cases suggestive patterns were apparent and useful to note. In its examination and assessment of the available evidence, the committee was looking for signals of associations and it endeavored to be sensitive rather than specific, so that even isolated findings that may well reflect random error from making multiple comparisons or those that have not been corroborated are reported. Ultimately, replications of results were considered indications of stronger evidence for an association that the committee considered in its weighting but in assessing the rather limited literature, some of the indications may not be confirmed with further research. The concern about neurologic and psychiatric effects was most apparent for mefloquine compared with the other antimalarials. For tafenoquine and chloroquine there was emphasis placed on adverse events associated with eye disorders. Doxycycline’s known concurrent gastrointestinal adverse events, which are commonly experienced, have led to some concern about the development of chronic gastrointestinal diseases, but the only study focused on this issue had significant methodologic limitations. The major issue with primaquine and tafenoquine is the risk of hemolysis associated with glucose-6-phosphate dehydrogenase (G6PD) deficiency. A/P has the fewest concurrent adverse events reported, and there is insufficient evidence for any persistent event associated with

A/P; for that reason it is often used as a comparator in studies of other antimalarial drugs. Several of the studies (Eick-Cost et al., 2017; Nasveld et al., 2010; Schneider et al., 2013, 2014; Tan et al., 2017) used comparisons with different antimalarials rather than a placebo or nonuse. As described earlier in the report, the difficulty with conducting such comparisons is that results from analyses that compare one drug with another may result in an observed lack of difference in effect because both drugs cause the adverse events. This is concerning because the use of antimalarial drugs in malaria-endemic areas is recommended, and a user’s choice of drug may be informed by the frequency and type of adverse events reported.

Of the 31 conclusions that the committee considered across all drugs and outcome categories, in all but one case the evidence of an association between the drug of interest and persistent or latent adverse events was deemed inadequate or insufficient. The committee concluded that there was sufficient evidence of an association between the use of tafenoquine and vortex keratopathy, which although it was found to persist beyond 28 days, was also found to resolve within 3 to 12 months and did not have a clinical implication, such as loss of vision. There was no convincing evidence of latent effects that did not manifest in individuals while they were taking the antimalarial and only emerged later, after drug cessation, with the exception of some eye disorders observed for A/P users. Individuals with past exposure (i.e., more than 90 days after the last day of use) to A/P were more likely to develop eye disorders than nonusers. This association was not present for current use. Based on information from studies of short-term follow-up, case reports, and biologic plausibility, the committee considers the existence of some persistent events, such as vertigo and tinnitus, to be highly plausible for certain antimalarials. For this reason, in its conclusion for each outcome category, the committee specifies whether the existing evidence warrants additional research in a specific area. For those health outcomes for which the committee concluded there is not a clear justification for additional research, the intention was to distinguish those issues for which there is presently an empirical basis for looking more closely and those for which such a basis is not present based on the currently available evidence. The committee’s intention is not to dismiss any issue or outcome but rather to highlight those in which a signal has been detected that warrants further study of a potential association. As more research accumulates, the outcomes that warrant further research may change or new ones that were not previously reported may become recognized, such as from additional case reports or mechanistic studies.

ADVANCING RESEARCH ON ANTIMALARIAL DRUGS

Given the seriousness of malaria and the billions of people at risk for it, and given that the parasite continues to develop resistance to currently available prophylactic drugs, there will be a continuing need for the currently available antimalarial drugs as well as new ones. Studying the persistent and latent effects of expo-

sures is challenging, and therefore it is important to recognize that seeking perfect or complete understanding is likely unrealistic. That said, in order to establish causal links between antimalarial exposure and persistent adverse events, it would be important to have a series of randomized trials designed to answer the specific safety questions ethically, with a long-term follow-up of participants and multiple well-designed observational studies of varying design with well-documented drug exposures and adverse event outcomes that control for confounding in rigorous ways. These studies would ideally have explicit documentation of the timing of antimalarial drug use and symptom occurrence (with clear temporal ordering), an extended follow-up that includes assessments at multiple time points, and a validated collection of information regarding potential confounders, antimalarial exposure (dose and timing), and the outcomes of potential interest, including a careful collection of neurologic and psychiatric outcomes using validated instruments. Given that some of the outcomes of concern are or may be rare, it will also be important to have sample sizes that are sufficient to detect associations if they do exist. While carrying out a large set of studies that has all of those components may not be realistic, there are strong designs that take advantage of existing data sets that would be possible. In addition, a series of well-designed studies that each has a number of (but perhaps not all) these components could be quite informative, and it could be used to triangulate the evidence to develop an understanding of the potential mechanisms and persistent adverse events. For example, studies that use large-scale electronic health records (including drug dispensing or prescribing records) with long-term follow-up of individuals could be used with strong non-experimental study designs and be complemented with studies of the biologic pathways that evaluate the link between the pharmacologic effects of drugs and the clinical conditions of interest. Additionally, if a sufficient number of studies are available, meta-analyses may be possible and very informative.

A key limitation of the existing literature is that very few studies were designed specifically to examine latent or persistent adverse events. However, more recently there has been more interest in assessing potential persistent or latent adverse events of antimalarial drugs, as compared with when the first of these drugs were approved in the 1940s and 1950s. The market authorization holder of tafenoquine is pursuing two required Phase IV trials to evaluate long-term tafenoquine safety. The first, NCT03320174, which is currently recruiting participants, is a randomized double-blind placebo-controlled study that will enroll 600 healthy G6PD-normal volunteers. Participants who meet the eligibility criteria will be randomized to receive a loading dose of either tafenoquine 200 mg (2 × 100 mg tablets) or placebo daily for 3 consecutive days, followed by study treatment (tafenoquine 200 mg or placebo) once per week for 51 weeks, with safety followup visits at weeks 4, 12, 24, and 52. All participants will return to the clinic at week 64 for an end-of-study visit. A participant who has an ongoing adverse event at the week 64 visit will be assessed up to three more times at approximately 12-week intervals, or until the resolution or stabilization of the adverse event, whichever

comes first. In addition, a large observational study to compare the rates of rare adverse events of tafenoquine relative to A/P in travelers (>10,000 participants) is in the planning stages.¹ These studies, and others like them, offer an excellent opportunity to study a broader set of persistent or latent outcomes.

With regard to mefloquine specifically, several factors may influence whether additional studies of its use for malaria prophylaxis are conducted and how informative those results will be. Although mefloquine is still recommended for civilian use, the numbers of prescriptions for it have declined substantially, likely in part due to the 2013 FDA boxed warning, media reports of adverse events, and the availability of similarly efficacious drugs with comparatively fewer adverse events or different adverse event profiles. Since 2009, Department of Defense (DoD) policy has been to restrict the use of mefloquine for service members to people who cannot take the other available antimalarials and do not have a history of the contraindications; in those who cannot take the other available antimalarials and have a history of neurobehavioral disorders, it is to be used very cautiously with clinical follow-up. In 2017, the latest year for which current prescription information is available, mefloquine was prescribed to a total of 52 individuals on active duty (Wiesen, 2019). Therefore, any prospective or retrospective studies conducted using service members since the policies went into effect will lack generalizability, and the channeling of persons who are healthier or who have previously tolerated mefloquine may account for some of the findings of no difference in frequency of most outcomes compared with other antimalarials in the literature.

Administrative Databases

Some of the most informative studies thus far have used health care databases or other data sources that cover large populations. Therefore, a logical place to look for additional opportunities is in other large databases that include a sufficiently large number of individuals who used antimalarial drugs and provide documentation of their subsequent health experience, or else by obtaining data needed for both exposure and outcome assessment by linking several large databases. Before embarking on such studies, it will be essential to first ensure that there is sufficient information on exposure and outcomes for a population large enough to generate meaningful results.

Information on U.S. military populations has been valuable, as reflected in studies of veterans participating in the National Health Study for a New Generation of U.S. Veterans (referred to as the “NewGen Study”) (Schneiderman et al., 2018), hospitalization data for active-duty service personnel (Wells et al., 2006), and a study with potentially greater value (limited because the timing of adverse events was not distinguished in the presented results) on medical encounters among

___________________

¹ Personal communication to the committee, Geoffrey Dow, M.B.A., Ph.D., Chairman and Chief Executive Officer of 60 Degrees Pharmaceuticals, LLC, January 28, 2019.

active-duty personnel (Eick-Cost et al., 2017). There may be value in revisiting other data resources generated for the study of military personnel to assess the feasibility of conducting informative research. Such sources might include general VA health care databases, registries such as the one developed for exposure to open burn pits, and cohorts assembled previously. Other countries, particularly those with national health care systems, may also have sufficient numbers of personnel deployed to areas in which malaria is endemic to learn from their experiences. For example, data could include which antimalarials were prescribed or used for different deployments and, for people who had multiple deployments, the health care experience during the deployed and nondeployed intervals as opposed to whether a condition or diagnosis was or was not present following the cessation of an antimalarial. However, as previously discussed, when using large databases for research, good adherence is assumed when in practice that has not been demonstrated to be the case.

Beyond the value of further observational research on military populations, the potential for randomized trials warrants serious consideration. Active-duty military and veterans currently participate in two distinct but overlapping health care systems administered by DoD and VA, respectively, which could facilitate examination of potential long-term health outcomes of exposures that occurred during active-duty service. Both DoD and VA collect a vast amount of health data on their personnel. VA has been a champion of the concept of a learning health care system, in which system factors are used to help incorporate established evidence into practice and reciprocally, in which new evidence is generated from a combination of rigorous analysis of quality improvement efforts and independent research investigations (Atkins et al., 2017). Although current ethical frameworks may distinguish the practice of quality improvement from research, some argue that research and practice should be viewed together in a learning health care environment. Furthermore, a case can be made that there is a moral obligation to incorporate important research questions into routine clinical practice so that such practice can be improved (Faden et al., 2013). An aspirational goal of both military and civilian health care systems might include transparently and prospectively incorporating large clinical trials (including pragmatic trials) and/or observational studies into routine practice. While there are an array of logistic and ethical considerations—especially because the people using and served by these systems may be considered populations with limited decision-making autonomy—the potential value of addressing possible health consequences of antimalarial prophylaxis in the population of interest using DoD and VA health information systems offers tremendous potential to advance knowledge and should be considered for future studies.

General population databases also have merit, as illustrated by the studies based on the UK General Practice Research Database (Meier et al., 2004; Schneider et al., 2013, 2014). Although a detailed evaluation would be required to assess the potential value of databases, large administrative data resources such as

Medicare, Sentinel, Kaiser-Permanente health system database, or commercially available claims databases, such as Optum, might be suitable for examining these issues, making up for a very low prevalence of exposure (and sometimes rare outcomes) with extremely large numbers of enrollees. These data sources—which are intended only to be illustrative of potential data sources and not exhaustive or directive—along with advanced statistical methods, are increasingly being used to study medication-related adverse events, and those methods could be applied to the question of persistent or latent health effects associated with the use of antimalarial drugs. However, an examination of psychiatric disorders, and the associated measurement difficulties, bring particular challenges that will need to be carefully considered, as discussed earlier in this chapter. For example, it may be particularly important to have longitudinal data on individuals so as to be able to account for pre-antimalarial exposure to mental health conditions and other exposures. One limitation of standard data sources such as these will be a lack of detailed information on non-medical factors, such as socioeconomic factors or military service. As such, large DoD and VA administrative databases that contain military health care records may be of particular use for studying antimalarial use and other exposures among service members over multiple time points.

FDA and Identifying and Evaluating Postmarketing Safety Concerns

FDA requirements for drug licensure have evolved over time. New requirements have often been driven by major safety events (Avorn, 2012; IOM, 2007). Thus, the amount and quality of data available about drugs prior to and post-licensure has improved over time. Many antimalarials were licensed decades ago, which may account in part for the relatively weak pre-licensure safety data for older drugs, such as chloroquine, which was licensed in the United States in 1949, and primaquine, which was licensed in the United States in 1952. Although FDA has the authority to require additional studies when safety concerns arise postmarketing, this authority was historically weak, and even now is used infrequently.

In addition, generic drug companies do not have the same requirements as the company responsible for the original labeled drug, leaving the responsibility for addressing concerns about older drugs unclear. Legislation in the 1980s encouraged the development of generic drugs, which are often less costly to consumers. Companies are required to demonstrate to FDA that their generic drugs are equivalent to the brand name drugs in terms of therapeutic effect, but they do not have to repeat the time-consuming and expensive clinical trials that have already been performed by brand companies to show safety and effectiveness. They are, however, required to establish and maintain records and make reports to FDA of all serious, unexpected adverse drug experiences associated with the use of their drug products (FDA, 2019a). Also, the original manufacturer is the “steward” of new safety issues, not the generic manufacturer, who is often not equipped to perform postmarketing safety studies. However, after generics are introduced, the

brand company may stop manufacturing the product or radically reduce resources focused on that product, leaving a gap in responsibility for addressing new safety concerns and new labeling (Kesselheim et al., 2012).

The FDA adverse event reporting system (FAERS) is an important source for potential safety signals. However, adverse events are under-reported, these reports lack complete information, there is uncertainty about whether or not the reported event was caused by the product, and FAERS data cannot be used to calculate the incidence of an adverse event. Additionally, FAERS has little ability to detect relatively common events, such as heart attacks, when the background rate in the population using the drug is relatively high. For example, the drug rofecoxib, a non-steroidal anti-inflammatory drug (NSAID), was on the market for 5 years prior to its withdrawal after a large postmarketing safety study identified an association with cardiovascular events. The trial was designed to show that rofecoxib had superior gastrointestinal safety compared to an older NSAID. The cardiovascular signal was unexpected, and it is unclear when or if this association would have been detected without this study. An estimated 88,000 to 140,000 excess cases of serious coronary heart disease occurred over the time rofecoxib was marketed in the United States (Graham et al., 2005). This event helped stimulate development of the FDA Sentinel Initiative, a network of administrative data and electronic record system data from insurance organizations and health plans that have been transformed into a standardized format. After a pilot period (mini Sentinel), Sentinel was officially launched in 2014. It now maintains a database of medical information on more than 200 million people that includes prescription drug use and health outcomes (FDA, 2019b). FDA’s Sentinel system has been designed to address gaps in knowledge about drug safety. Among those gaps is the lack of information on the possible persistent effects of antimalarials.

Collaborations

In its evaluation of the available evidence to address persistent adverse events, the committee identified studies that clearly had collected data that could be informative, but the analyses were either not conducted or not presented in a way that was informative for the committee’s purposes. The clearest example of this was Eick-Cost et al. (2017), which referred only in passing to the pattern of health outcomes over time following the cessation of drug use. The investigators collected information that could have been quite helpful had they restricted the analysis to ≥28 days post-drug-cessation. For a substantial proportion of the other studies that qualified for consideration based on having collected health data pertaining to the time period of interest, the committee was unable to draw inferences regarding persistent adverse events because the experience during and after drug use was aggregated or not clearly distinguished. A possible contribution to advancing the literature might come from embarking on selected re-analyses of studies that collected data on adverse events for 28 days post-drug-cessation but did not

analyze these data or report them. A re-analysis of individual studies with notable potential value would make the temporal course of drug use and health experience explicit and enable inferences regarding concomitant versus persistent adverse events. Re-analyses could also allow for the discovery of symptoms or diagnoses that covary. For example, if certain symptoms or diagnoses occur together in the same patients, there may be reason to consider a syndrome of “neuropsychiatric” symptoms that co-occur, rather than looking individually at separate neurologic or psychiatric experiences.

A pooled data analysis effort, where there is sufficient compatibility across studies and the ability to apply a standardized approach to classifying exposure and outcome and control for potential confounders, may also move this area of scientific inquiry forward. Many small studies have potentially informative data, but the researchers analyzed or presented the data in uninformative or simply different ways; therefore, assembling and re-analyzing data from these studies could also be beneficial. However, the committee recognizes that a meaningful summary estimate cannot be generated when study methods are fundamentally incompatible. By using standardized methods for making definitions of exposure, outcome, and covariates as compatible as possible and by conducting parallel or unified analyses, inferences may be drawn that go beyond the published results of the component studies.

Approaches to Research That Are Unlikely to Contribute to the Evidence Base

Based on the questions of concern and past experience, there are a number of approaches that are unlikely to provide much insight regarding persistent or latent adverse events of antimalarial drugs. Cross-sectional studies that attempt to correlate drug use and symptoms or diagnoses without the ability to explicitly consider the temporal course of events will not make it possible to separate acute from persistent or latent adverse events. The data need to lend themselves to analyses that can address the temporal sequence of drug use, cessation of drug use, and health experience.

Small clinical trials often contain detailed health information but rarely include sufficiently long follow-up periods to assess persistence, and they almost never have sufficient numbers of participants to provide the needed statistical precision to address clinically significant outcomes. Perhaps with some effort they could address common, relatively mild symptoms of interest longitudinally, but this generally is beyond the scope of what is conventionally done.

Adverse event registries and individual reports of suspected adverse events to medications, such as that used by FDA, provide limited information with regard to rigorous research. While the experiences reported may serve as signals to indicate reactions and events of concern that were not necessarily identified during clinical trials, they also serve to inform when changes to label warnings or precautions

may be merited. Because adverse event registries do not provide comparative data on people with varying exposures and do not offer any quantitative data on the frequency with which side effects are experienced, at most they are case reports that offer hints of areas of outcomes to guide subsequent research that is more methodologically rigorous.

Different Strategies and Approaches for Advancing Research

For many complex issues there is no single research approach that can provide definitive answers since all strategies have varying strengths and limitations. The need for convergent evidence and triangulation (integrating results from several different approaches that have different and unrelated key sources of potential bias; see Lawlor et al., 2016) is clearly applicable to the assessment of persistent adverse events of antimalarial drugs. Beyond the continued exploitation of large administrative databases, some other designs warrant consideration for complementing such studies.

Conducting studies of “medium-term” adverse events that continue beyond the events that occur while taking a drug (such as those up to 3 or 6 months post-cessation) would be informative if focused and validated assessments of health status were performed over the subsequent weeks or months. This might involve extending clinical trials or systematically following returning travelers with such examinations as clinical evaluations that are sufficiently sensitive to discern even subclinical health status or even carefully constructed questionnaires. While such studies would not likely be large enough to identify rare, clinically significant events, such evidence would complement larger, less detailed studies. To the extent that there are hypotheses regarding which individuals are especially likely to be vulnerable based on genetics, pre-existing health conditions, or other factors, these smaller, more intensive evaluations could target such high-risk groups.

Large case–control studies of specific adverse events or health outcomes of interest could, provided there was sufficient prevalence of exposure, generate additional evidence on associations of antimalarial drugs. Such studies might be best conducted in populations that have more than background rates of exposure, enriched with military personnel, international travelers, or those whose work requires spending time in settings where malaria is endemic, such as Peace Corps volunteers, missionaries, and Department of State employees.

Finally, well-conceived studies of experimental systems, in vitro or in vivo, could provide meaningful information to help in interpreting the evidence from human populations. Attempts to establish correlations between the effects of an antimalarial drug on experimental systems and their effects on humans are particularly difficult because there are well-known species-, sex-, and outcome-specific differences in susceptibility to drug toxicity. Even in humans the data on the adverse effects of the drugs are not consistent across studies. Building on model systems for studying irreversible neurobehavioral effects of drugs—for example,

a comparison across antimalarial drugs—would add to the constellation of data to help refine interpretation. Other examples of research that would be required for suitable rigor include (1) testing of the impact of prolonged exposure to biologically relevant antimalarial dosing (e.g., human dose adjusted to lab animal drug clearance/metabolism) across a battery of behavioral tests with face validity for persistent or latent psychiatric, neurologic, or other disorders and (2) in vivo testing of lasting antimalarial-induced cell loss and toxicity using contemporary standards of assessment, such as a stereologic assessment of cell loss, microglioisis, astrocytosis, and white matter loss in multiple brain regions and tissues of interest.

REFERENCES

Graham, D. J., D. Campen, R. Hui, M. Spence, C. Cheetham, G. Levy, S. Shoor, and W. A. Ray. 2005. Risk of acute myocardial infarction and sudden cardiac death in patients treated with cyclo-oxygenase 2 selective and non-selective non-steroidal anti-inflammatory drugs: Nested case-control study. Lancet 365(9458):475-481.

Green, J. A., A. K. Patel, B. R. Patel, A. Hussaini, E. J. Harrell, M. J. McDonald, N. Carter, K. Mohamed, S. Duparc, and A. K. Miller. 2014. Tafenoquine at therapeutic concentrations does not prolong Fridericia-corrected QT interval in healthy subjects. J Clin Pharmacol 54:995-1005.

IOM (Institute of Medicine). 2007. The future of drug safety: Promoting and protecting the health of the public. Washington, DC: The National Academies Press.

Jaspers, C. A., A. P. Hopperus Buma, P. P. van Thiel, R. A. van Hulst, and P. A. Kager. 1996. Tolerance of mefloquine chemoprophylaxis in Dutch military personnel. Am J Trop Med Hyg 55(2):230-234.

Judd, L. L., P. J. Schettler, E. S. Brown, O. M. Wolkowitz, E. M. Sternberg, B. G. Bender, K. Bulloch, J. A. Cidlowski, E. R. de Kloet, L. Fardet, M. Joëls, D. Y. Leung, B. S. McEwen, B. Roozendaal, E. F. Van Rossum, J. Ahn, D. W. Brown, A. Plitt, and G. Singh. 2014. Adverse consequences of glucocorticoid medication: Psychological, cognitive, and behavioral effects. Am J Psychiatry 171(10):1045-1051. Erratum in Am J Psychiatry. 2014. 171(11):1224.

Kesselheim, A. S., J. Avorn, and J. A. Greene. 2012. Risk, responsibility, and generic drugs. N Engl J Med 367(18):1679-1681.

Korhonen, C., K. Peterson, C. Bruder, and P. Jung. 2007. Self-reported adverse events associated with antimalarial chemoprophylaxis in Peace Corps volunteers. Am J Prev Med 33(3):194-199.

Korpi, E. R., B. den Hollander, U. Farooq, E. Vashchinkina, R. Rajkumar, D. J. Nutt, P. Hyytiä, and G. S. Dawe. 2015. Mechanisms of action and persistent neuroplasticity by drugs of abuse. Pharmacol Rev 67(4):872-1004.

Laothavorn, P., J. Karbwang, K. Na Bangchang, D. Bunnag, and T. Harinasuta. 1992. Effect of mefloquine on electrocardiographic changes in uncomplicated falciparum malaria patients. Southeast Asian J Trop Med Public Health 23(1):51-54.

Lawlor, D. A., K. Tilling, and G. Davey Smith. 2016. Triangulation in aetiological epidemiology. Int J Epidemiol 45(6):1866-1886.

Leary, K. J., M. A. Riel, M. J. Roy, L. R. Cantilena, D. Bi, D. C. Brater, C. van de Pol, K. Pruett, C. Kerr, J. M. Veazey, Jr., R. Beboso, and C. Ohrt. 2009. A randomized, double-blind, safety and tolerability study to assess the ophthalmic and renal effects of tafenoquine 200 mg weekly versus placebo for 6 months in healthy volunteers. Am J Trop Med Hyg 81:356-362.

Lee, T. W., L. Russell, M. Deng, and P. R. Gibson. 2013. Association of doxycycline use with the development of gastroenteritis, irritable bowel syndrome and inflammatory bowel disease in Australians deployed abroad. Intern Med J 43(8):919-926.

Macaluso, M., A. Flynn, and S. Preskorn. 2016. Determining whether a definitive causal relationship exists between aripiprazole and tardive dyskinesia and/or dystonia in patients with major depressive disorder, part 4: Case report data. J Psychiatr Pract 22(3):203-220.

Meier, C. R., K. Wilcock, and S. S. Jick. 2004. The risk of severe depression, psychosis or panic attacks with prophylactic antimalarials. Drug Saf 27(3):203-213.

Miller, A. K., E. Harrell, L. Ye, S. Baptiste-Brown, J. P. Kleim, C. Ohrt, S. Duparc, J. J. Möhrle, A. Webster, S. Stinnett, A. Hughes, S. Griffith, and A. P. Beelen. 2013. Pharmacokinetic interactions and safety evaluations of coadministered tafenoquine and chloroquine in healthy subjects. Br J Clin Pharmacol 76:858-867.

Nasveld, P. E., M. D. Edstein, M. Reid, L. Brennan, I. E. Harris, S. J. Kitchener, P. A. Leggat, P. Pickford, C. Kerr, C. Ohrt, W. Prescott, and the Tafenoquine Study Team. 2010. Randomized, double-blind study of the safety, tolerability, and efficacy of tafenoquine versus mefloquine for malaria prophylaxis in nonimmune subjects. Antimicrob Agents Chemother 54:792-798.

Petersen, E., T. Ronne, A. Ronn, I. Bygbjerg, and S. O. Larsen. 2000. Reported side effects to chloroquine, chloroquine plus proguanil, and mefloquine as chemoprophylaxis against malaria in Danish travelers. J Travel Med 7(2):79-84.

Rendi-Wagner, P., H. Noedl, W. H. Wernsdorfer, G. Wiedermann, A. Mikolasek, and H. Kollaritsch. 2002. Unexpected frequency, duration and spectrum of adverse events after therapeutic dose of mefloquine in healthy adults. Acta Trop 81(2):167-173.

Ruha, A. M., and M. Levine. 2014. Central nervous system toxicity. Emerg Med Clin North Am 32(1):205-221.

Rueangweerayut, R., G. Bancone, E. J. Harrell, A. P. Beelen, S. Kongpatanakul, J. J. Möhrle, V. Rousell, K. Mohamed, A. Qureshi, S. Narayan, N. Yubon, A. Miller, F. H. Nosten, L. Luzzatto, S. Duparc, J.-P. Kleim, and J. A. Green. 2017. Hemolytic potential of tafenoquine in female volunteers heterozygous for glucose-6-phosphate dehydrogenase (G6PD) deficiency (G6PD Mahidol variant) versus G6PD normal volunteers. Am J Trop Med Hyg 97(3):702-711.

Schlagenhauf, P., R. Steffen, H. Lobel, R. Johnson, R. Letz, A. Tschopp, N. Vranjes, Y. Bergqvist, O. Ericsson, U. Hellgren, L. Rombo, S. Mannino, J. Handschin, and D. Sturchler. 1996. Mefloquine tolerability during chemoprophylaxis: Focus on adverse event assessments, stereochemistry and compliance. Trop Med Int Health 1(4):485-494.

Schneider, C., M. Adamcova, S. S. Jick, P. Schlagenhauf, M. K. Miller, H. G. Rhein, and C. R. Meier. 2013. Antimalarial chemoprophylaxis and the risk of neuropsychiatric disorders. Travel Med Infect Dis 11(2):71-80.

Schneider, C., M. Adamcova, S. S. Jick, P. Schlagenhauf, M. K. Miller, H. G. Rhein, and C. R. Meier. 2014. Use of anti-malarial drugs and the risk of developing eye disorders. Travel Med Infect Dis 12(1):40-47.

Schneiderman, A. I., Y. S. Cypel, E. K. Dursa, and R. Bossarte. 2018. Associations between use of antimalarial medications and health among U. S. veterans of the wars in Iraq and Afghanistan. Am J Trop Med Hyg 99(3):638-648.

Schwartz, E., and G. Regev-Yochay. 1999. Primaquine as prophylaxis for malaria for nonimmune travelers: A comparison with mefloquine and doxycycline. Clin Infect Dis 29(6):1502-1506.

Tan, K. R., S. J. Henderson, J. Williamson, R. W. Ferguson, T. M. Wilkinson, P. Jung, and P. M. Arguin. 2017. Long term health outcomes among returned Peace Corps volunteers after malaria prophylaxis, 1995–2014. Travel Med Infect Dis 17:50-55.

Walsh, D. S., C. Eamsila, T. Sasiprapha, S. Sangkharomya, P. Khaewsathien, P. Supakalin, D. B. Tang, P. Jarasrumgsichol, C. Cherdchu, M. D. Edstein, K. H. Rieckmann, and T. G. Brewer. 2004. Efficacy of monthly tafenoquine for prophylaxis of Plasmodium vivax and multidrug-resistant P. falciparum malaria. J Infect Dis 190(8):1456-1463.

Wells, T. S., T. C. Smith, B. Smith, L. Z. Wang, C. J. Hansen, R. J. Reed, W. E. Goldfinger, T. E. Corbeil, C. N. Spooner, and M. A. Ryan. 2006. Mefloquine use and hospitalizations among US service members, 2002-2004. Am J Trop Med Hyg 74(5):744-749.

Wiesen, A. 2019. Overview of DoD antimalarial use policies. Presentation to the Committee to Review Long-Term Health Effects of Antimalarial Drugs, January 28, 2019.

This page intentionally left blank.