|
|
||||||||
| ABSTRACT |
|
|
|---|
| INTRODUCTION |
|
|
|---|
Observational studies and randomized clinical trials have found that those who took mefloquine were at increased risk for a wide variety of neuropsychiatric events, including depression, dizziness, panic attacks, strange thoughts, altered spatial perception, fatigue, headache, and vivid dreams.1928 It has also been observed that women report the greatest number of adverse events, as well as first-time users.29 In contrast, several observational epidemiologic studies and a few clinical trials have reported that mefloquine is safe and well tolerated, especially among military populations.3034
Safe and effective antimalarials are of great importance to the health of US service members. Studies of civilian populations may not apply to military forces because this group represents a unique population with differences in demographic composition, health, environmental exposures, and purpose of travel in comparison with most civilian cohorts studied to date. The objective of this study was to describe serious health outcomes after mefloquine use among US service members.
| MATERIALS AND METHODS |
|
|
|---|
We used two reference groups. The first reference group was comprised of service members who had duty zip codes for either Europe or Japan at some time during 2002 and had no evidence of having been deployed from October 1, 2001 through the individuals period of observation. This group was chosen to represent those healthy enough to be stationed overseas and healthy enough to have taken mefloquine during the study period. The second reference group consisted of US service members who were identified as having been deployed for a minimum of 1 month during 2002. Both reference groups were restricted to individuals who had no evidence of having received a prescription for mefloquine or chloroquine or a doxycycline prescription for more than 14 tablets. Excluding all antimalarials was done to ensure the reference groups did not include mefloquine-prescribed people not represented in the prescription database.
Hospitalization diagnoses. The study cohort was electronically linked to the Standardized Inpatient Data Record (SIDR) and the Health Care Service Record (HCSR) to identify hospitalizations. The SIDR record may contain up to eight International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) coded36 diagnoses per admission, while the HCSR inpatient record contains an admission, a primary, and up to eight secondary ICD-9-CMcoded diagnoses. For this study, each unique diagnosis was included in the analysis. We analyzed any-cause hospitalization (excluding complications of pregnancy, childbirth, and the puerperium [630676], congenital anomalies [740759], and certain conditions originating in the perinatal period [760779]). Other a priori outcomes included hospitalizations among 14 broad ICD-9-CMcoded categories, including infectious and parasitic diseases (001139); neoplasms (140239); endocrine, nutritional, and metabolic diseases and immunity disorders (240279); diseases of the blood and blood-forming organs (280289); mental disorders (290319); diseases of the skin and subcutaneous tissue (680709); symptoms, signs, and ill-defined conditions (780799); injury and poisoning (800999); and diseases of the following systems: nervous and sense organs (320389), circulatory (390459), respiratory (460519), digestive (520579), genitourinary (580629), and musculoskeletal and connective tissue (710739). Mental disorder diagnoses were stratified into five categories, including somatoform, mood, anxiety, substance use, and other disorders.37 In addition, hospitalizations caused by migraine (346), nystagmus associated with disorders of the vestibular system (379.54), vertiginous syndromes, and other disorders of vestibular system (386), and dizziness and giddiness (780.4) were analyzed.38
Methods.
Statistical analyses included both descriptive and multivariable methods. After descriptive investigation of population characteristics, analyses were performed to assess the significance of associations between mefloquine prescription and sex, age, race/ethnicity, service, marital status, rank, occupation, and history of hospitalization during 2001 in comparison with both reference groups. Age was grouped into three categories (
24, 2534, and
35 years). Service branches included Army, Navy, Marines, and Air Force. Race/ethnicity included black non-Hispanic, Hispanic, other, and white non-Hispanic. Marital status was categorized as married and nonmarried. Pay grade was dichotomized to enlisted and officers, and occupations included combat and non-combat.
Cox proportional hazards time-to-event modeling was used to compare the hospitalization experience among mefloquine-prescribed US service members with the hospitalization experience among those stationed in Europe/Japan and the deployed reference groups, while accounting for attrition from active duty service during the follow-up period. Follow-up time began on return from deployment for mefloquine-prescribed members, and for the deployed reference group, on assignment to Europe or Japan, or January 1, 2002, whichever occurred last for the Europe/Japan reference group. Follow-up continued for 12 months or until date of separation from active-duty service, date of next deployment, date of next antimalarial prescription, or end of the study period, March 31, 2004, whichever occurred first. Individuals who could not be followed a minimum of 2 months were excluded from the study. Statistical analyses including adjusted hazard ratios (HRs) and associated 95% confidence intervals (CIs) were performed using SAS software (Version 9.1; SAS Institute, Cary, NC).39 This research was conducted in compliance with all applicable federal regulations governing the protection of human subjects in research (NHRC.2004.0025).
| RESULTS |
|
|
|---|
0.05), and these covariates were entered into the Cox proportional hazards models.
|
|
|
| DISCUSSION |
|
|
|---|
In comparison with military members stationed in Europe or Japan, mefloquine-prescribed individuals were at a significantly decreased risk for hospitalizations because of any cause, diseases of the respiratory and digestive systems, and hospitalizations caused by musculoskeletal and connective tissue disorders, ill-defined conditions, injuries and poisonings, and mood disorders. It is possible, but quite unlikely, that mefloquine use provides some protective effect for these outcomes. More likely, these decreased HRs are the result of a selection, or reporting, bias. We chose service members living in Europe or Japan because they undergo a medical clearance process before being assigned overseas, where health care services may be limited. However, we restricted this group to those with no evidence of having been deployed, so members of this group may not have been as healthy as the mefloquine-prescribed group, all of whom had deployed. Another possible explanation for these findings is service members who reside in Europe or Japan might be hospitalized for less-severe conditions than are those in the United States, or perhaps overseas hospitalizations are more likely to be reported.
Finding an elevated, but not statistically significant, hazard for vertiginous syndromes compared with both reference groups is interesting. Recently, there has been much interest in the relationship between mefloquine and vertiginous syndromes in the media and among federal legislators.38 Further studies are required to better assess whether an association between mefloquine and vertiginous syndromes exists.
The unique nature of this study makes comparisons with previously published results difficult. Of those published, a study by Meier and others was most similar in methodology using electronically recorded prescriptions and diagnoses to define exposure and outcome, respectively.22 However, remaining differences in study design and the coding of neuropsychiatric outcomes between the two studies makes comparisons hard. Although different in methodologies, it might be reassuring that the absence of significant risk for severe neuropsychiatric illnesses among service members taking mefloquine was observed in this study, three randomized clinical trials, and one open-label prospective study using military populations.27,28,31,34
The results of this study should be considered within its limitations. Using a prescription database as a surrogate for mefloquine exposure created unique challenges, including potentially low sensitivity for identifying exposure. We attempted to minimize exposure misclassification by requiring a minimum pill count of at least seven tablets per mefloquine prescription to qualify as an exposure, yet we acknowledge this serves only as a proxy for having taken mefloquine. Among the deployed reference group, there may have been poor specificity in mefloquine exposure assessment because an unknown percentage of individuals in this population may have actually taken mefloquine while deployed. We attempted to improve specificity of mefloquine exposure by using a reference group containing only nondeployed service members who resided in Europe or Japan. Although we chose to assess hospitalizations as the outcome measure for this study, this choice restricted analyses to those medical conditions that were of ample severity to require hospitalization, and it does not represent the entire spectrum of morbidity that may be associated with mefloquine. Finally, the study design called for a large number of analyses, which increases the likelihood of finding a statistically significant, but spurious, association.
This study has a number of strengths. It was specifically designed to assess the association between mefloquine prescriptions and hospitalizations among US military personnel, an otherwise healthy population of younger adults. We identified an adequate number of mefloquine-prescribed individuals and two large reference groups to allow adequate power to explore associations, with the exception of very rare outcomes. The use of two reference groups allowed for comparisons with a population that had high specificity for mefloquine exposure, the Europe/Japan reference group. While the deployed reference group had some potential for misclassification of mefloquine exposure, this group was more homogeneous in comparison with the mefloquine-prescribed group. Although the use of hospitalization data limited the number of outcomes available for analysis, the use of these objective data eliminated the possibility of recall bias.
This study was the first to describe the relations between mefloquine-prescribed US service members and a wide range of health outcomes using objective data. We found that mefloquine-prescribed service members were not at a statistically significant increased risk for hospitalization over a wide range of broad and specific disease categories. Future studies should explore additional data sources that complement those used in this study to define mefloquine exposure and neuropsychiatric outcomes.
Received September 8, 2005. Accepted for publication December 27, 2005.
Acknowledgments: The authors thank Dr. M. David Rudd, Baylor University, for assistance in the design of this study, Scott Seggerman and his team of professionals at the Defense Manpower Data Center, for providing the necessary demographic data; Dr. David Guerin and his team of professionals for providing access to the Military Health Systems Management Analysis and Reporting Tool for both pharmacy data and hospitalization data; and Dr. Roger Gibson and the Armed Forces Epidemiological Board and Dr. Steven Phillips, Director, Deployment Medicine and Surveillance, Office of the Assistant Secretary of Defense, Health Affairs, for providing expertise in study design and for critical review of the manuscript. Dr. Wells had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
This represents report 0505, supported by the Department of Defense, under work unit no. 60002. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Navy, Department of Defense, or the US Government. Approved for public release; distribution is unlimited.
* Address correspondence to Margaret A. K. Ryan, DoD Center for Deployment Health Research, PO Box 85122, Naval Health Research Center, San Diego, CA 92186-5122. E-mail: ryan{at}nhrc.navy.mil ![]()
Authors addresses: Timothy S. Wells, Building 824, Room 206, 2800 Q Street, Wright-Patterson AFB, OH, 45433, E-mail: timothy.wells{at}wpafb.af.mil. Tyler C. Smith, Besa Smith, Linda Z. Wang, Christian J. Hansen, Robert J. Reed, Wendy E. Goldfinger, Thomas E. Corbeil, Christina N. Spooner, and Margaret A. K. Ryan, DoD Center for Deployment Health Research, PO Box 85122, Naval Health Research Center, San Diego, CA 92186-5122, E-mails: Smith{at}nhrc.navy.mil, Besa{at}nhrc.navy.mil, Wang{at}nhrc.navy.mil, Hansen{at}nhrc.navy.mil, Reed{at}nhrc.navy.mil, Wgoldfinger{at}hotmail.com, Tecorbeil{at}hotmail.com, Spooner{at}nhrc.navy.mil, and Ryan{at}nhrc.navy.mil.
| REFERENCES |
|
|
|---|
This article has been cited by other articles:
![]() |
P. Schlagenhauf and E. Petersen Malaria Chemoprophylaxis: Strategies for Risk Groups Clin. Microbiol. Rev., July 1, 2008; 21(3): 466 - 472. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. H. Chen, M. E. Wilson, and P. Schlagenhauf Controversies and Misconceptions in Malaria Chemoprophylaxis for Travelers JAMA, May 23, 2007; 297(20): 2251 - 2263. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. H. Chen, M. E. Wilson, and P. Schlagenhauf Prevention of malaria in long-term travelers. JAMA, November 8, 2006; 296(18): 2234 - 2244. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |