Epidemiologic malaria surveillance in French Armed Forces consists of continuous and systematic collection, analysis, interpretation, and feedback of epidemiologic data from all military physicians (Technical Appendix). Malaria is defined as any pathologic event or symptom associated with confirmed parasitologic evidence (Plasmodium spp. on a blood smear, a positive quantitative buffy coat malaria diagnosis test result, or a positive malaria rapid diagnosis test result) contracted in French Guiana. A case occurring in a person during or after a stay in French Guiana without a subsequent stay in another malaria-endemic area was assumed to be contracted in French Guiana. Each malaria attack was considered a separate case. Equal information was available for the entire 11-year study period. Data from weekly reports and malaria-specific forms were used for analysis.

Indicators are expressed as annual incidence and annual incidence rate. The denominator of the annual incidence rate is the average number of military personnel at risk for malaria during a given year.

Statistical analysis was performed by using Epi Info 6.04dfr (Centers for Disease Control and Prevention, Atlanta, GA, USA). Comparisons over time were made by using the χ² test for trend and between groups by using the Kruskal-Wallis test. A p value <0.05 was considered significant.

The incidence rate for malaria cases among French military personnel deployed to French Guiana has increased since 1998 (p<0.001). P. falciparum incidence has remained stable (p = 0.10), and P. vivax incidence has increased (p<0.001) (Figure). In 2007 and 2008, French military personnel in French Guiana represented only 23.0% and 22.2% of those deployed to malaria-endemic regions. However, most reported malaria cases were contracted in this region (50.0% and 62.9%, respectively, of all cases). P. vivax was responsible for most malaria attacks reported in French Guiana (Table). The proportion of malaria attacks caused by P. vivax increased from 44% to 84% (p<0.001) during the study period.

In 2008, among the 264 reported cases, 221 were temporarily unavailable for duty. Median lost work days were lower for attacks of P. vivax malaria than for P. falciparum malaria (5 days/attack, interquartile range 4–7 days/attack vs. 7 days/attack, interquartile range 5–10 days/attack; p = 0.006)

Among 264 malaria cases contracted in French Guiana, 39.4% were in persons who had reported >1 malaria attack in the previous 6 months. Among the 221 P. vivax malaria cases, 45.5% were in persons who had already reported >1 P. vivax malaria attack in the previous 6 months. In 2008, among those required to take chemoprophylaxis (i.e., during a mission to the Amazon basin and 28 days after the mission), 45.0% admitted not taking their chemoprophylaxis within 8 days before onset of symptoms.

P. vivax malaria attacks have resulted in a substantial number of lost work days and have adversely affected operational readiness of military personnel. Despite availability of appropriate chemoprophylaxis, since 1998, French Armed Forces have been affected by an increase in incidence of P. vivax malaria. Several causes of this increase have been hypothesized.

First, epidemiologic trends for all-cause malaria in French Guiana and overall reported malaria incidence has not changed substantially since the end of the 20th century: ≈4 000 cases were reported annually during the 1990s (5) and 3,500–4,500 were reported during 2008 (2). However, the proportion of P. vivax malaria has increased from 20% of cases in the 1990s (6) to 56.1% during September 2003–February 2004 among patients at the Cayenne Public Hospital (7). Furthermore, 70% of malaria cases diagnosed in 2006 among French travelers returning from French Guiana were caused by P. vivax (8). One explanation for this parasitologic evolution may involve immigration from Brazil and Suriname to illegal gold-mining areas in the Amazon basin of French Guiana (2,6). These immigrating populations brought P. vivax from high-prevalence regions to an area where an efficient vector for malaria, Anopheles darlingi mosquitoes, was present (9). Changes in weather patterns and regional infrastructure could also explain this increase.

Second, military missions have intensified. In 2008, a police and military operation to reduce illegal gold-panning activities in the Amazon basin occurred in French Guiana. This operation might explain the 2008 peak in the incidence rate. Since 2002, these operations have resulted in several outbreaks among forces in French Guiana, especially in 2003 and 2005 (2,3).

Third, deficiencies have occurred in implementing individual and collective protective measures. These military operations were conducted by personnel from French Guiana or France, few had any rainforest experience. Despite extensive training, instructions were clearly not followed, as demonstrated by a 45% noncompliance rate for chemoprophylaxis.

P. vivax has accounted for >80% of reported malaria cases in French Guiana for the past 3 years. In addition, relapses of P. vivax malaria occur in the absence of radical treatment. In 2008, 45.5% of persons with P. vivax malaria had already reported >1 P. vivax malaria attack in the past 6 months. Although the P. vivax malaria mortality rate is low, the effect of P. vivax malaria on force operational readiness is high because relapses decrease the availability of military personnel. In addition, P. vivax malaria can be severe, despite its reputation as a mild form of malaria (10). Since 2009, to reduce the number of relapses, a French Ministry of Defense circular has recommended treatment with primaquine for 2 or 3 weeks after a first attack of P. vivax malaria. Studies of the use of primaquine chemoprophylaxis are ongoing (11–13).

In conclusion, the incidence of P. vivax malaria is increasing in French Guiana, especially in French Armed Forces. The incidence of P. falciparum malaria remains stable. This new epidemiologic finding can affect the level of individual health and operational capabilities. Performance of vector evaluation studies and control in the regions could be another possible intervention.