Hospitalization Criteria in Imported Falciparum Malaria

Transcription

1 306 Hospitalization Criteria in Imported Falciparum Malaria Valérie Briand, MD, MPH, * Olivier Bouchaud, MD, PhD, Jérôme Tourret, MD, Charlotte Behr, PhD, Sophie Abgrall, MD, PhD, Pascal Ralaimazava, MD, Jacques Le Bras, PharmD, PhD, and Arnaud Fontanet, MD, PhD * * Emerging Diseases Epidemiology Unit, Institut Pasteur, Paris, France ; Department of Infectious and Tropical Diseases, Hôpital Bichat Claude Bernard and Hôpital Avicenne, APHP, Paris, France ; CNRS UMR 5164 CIRID, Université Bordeaux 2, Bordeaux, France ; Department of Parasitology, Hôpital Bichat Claude Bernard and Hôpital Avicenne, APHP, Paris, France DOI: /j x Background. Controversy exists about the management of patients with imported Plasmodium falciparum malaria. We postulated that rapid parasite clearance supports ambulatory care, or, conversely, that factors associated with longer parasite clearance time (PCT) could be used as hospitalization criteria. Methods. Hospitalized patients with imported falciparum malaria recruited through one single travel clinic between 1993 and We used a linear regression to identify factors independently associated with PCT defined as the time in hours from antimalarial drug administration until the first negative malaria smear. Results. Among 400 patients hospitalized with falciparum malaria, mean (range) PCT was 58 (1 189) hours. In multivariate analysis, severe malaria, gastrointestinal signs, initial temperature greater than or equal to 40 C, parasitemia greater than or equal to 1%, and platelet counts less than 50,000/ L were associated with longer PCT. Offering ambulatory care to patients aged 15 to 64 years with none of the factors associated with longer PCT in the study would have resulted in 147 (37%) patients receiving outpatient care. Conclusion. Factors identified in this model may help physicians determine which P falciparum malaria patients can be treated on an ambulatory basis, and contribute to revisions of national guidelines for imported malaria management. Each year, an estimated 50 million travelers visit malaria-endemic areas, of whom around 30,000 contract malaria. 1 Controversy exists about the management of patients with imported Plasmodium falciparum malaria. In 1999, the 12th Consensus Conference of the French-Speaking Society of Infectious Diseases (FSSID) recommended the ambulatory management of uncomplicated falciparum malaria under certain conditions ( Table 1 ). 2 Similarly, in some clinics of Switzerland, patients with uncomplicated falciparum malaria are provided with ambulatory care, and only 5 (6%) of 82 patients in a prospective study needed readmission when treated as outpatients. 3 However, other experts insist on hospitalization of all cases of falciparum malaria, as complications can develop Corresponding Author: Valérie Briand, MD, MPH, Emerging Diseases Epidemiology Unit, Institut Pasteur, Bâtiment Laveran 3ème étage, 25, rue du Docteur Roux, F Paris, France. vale.briand@laposte.net quickly, 4 and initial assessment of patients can be misleading. 5 We postulated that rapid parasite clearance supports ambulatory care, or, conversely, that factors associated with longer parasite clearance time (PCT) could be used as hospitalization criteria. We have used data collected over 8 years among 400 patients hospitalized with falciparum malaria to identify factors associated with longer PCT. We have then examined the in-hospital outcome of patients who did not have any of these factors to see if they could have been treated as outpatients. Materials and Methods Population and Study Design Between January 1993 and December 2000, all patients visiting the travel clinic of Bichat-Claude Bernard University Hospital (Paris, France) for falciparum malaria have been considered for the study. On admission, the following information was collected: sociodemographic characteristics, 2007 International Society of Travel Medicine, Journal of Travel Medicine, Volume 14, Issue 5, 2007,

2 Hospitalization for Travel Malaria Table 1 Criteria for ambulatory treatment of uncomplicated falciparum malaria. French-speaking Society of Infectious Diseases, Results from the parasitological examination on the same day No signs of severity No gastrointestinal symptoms Parasitemia <5% No sociocultural factors compromising treatment compliance No at-risk conditions such as old age, splenectomy, pregnancy, underlying disease and particularly cardiovascular disease, and people living alone Nearby hospital Treatment available at the pharmacy for immediate intake Follow-up visits at days 3 and 7 Note: In children, the rapid changes in disease severity, and the high frequency of gastrointestinal symptoms, do not allow full ambulatory care. past history of malaria, travels characteristics (visited country, length of stay, type of accommodation), and malaria prophylaxis habits. The adequate use of chemoprophylaxis was defined as both taking a regimen adapted to the patterns of drug resistance in the region visited and being compliant with that regimen. Other information consisted in the delay between return from the endemic area and onset of symptoms, the delay between onset of symptoms and treatment, clinical data [including criteria used for the World Health Organization (WHO) definition of severity 6 and axillary temperature], and biological data (parasitemia expressed in percentage of parasitized red blood cells). WHO severity criteria, general condition, and expected compliance with treatment and follow-up were used to determine if patients required admission to hospital or could receive ambulatory treatment. Choice of therapy depended on the pattern of drug resistance in the area of acquisition, the history of chemoprophylaxis use, the severity of infection, the existence of gastrointestinal signs, and the general condition of the patient. Drugs were administered according to the recommendations of the FSSID. 2 Patients were retreated if any vomiting occurred within the first hour following treatment intake. In hospitalized patients, parasitemia was systematically measured on each day at 8 am during hospitalization, more often for severe patients, and at days 3, 7, and 28. PCT was defined as the time in hours from drug administration until no asexual parasites were detected in peripheral blood. 7 A smear was considered negative if no parasites were seen on the thick film within counting 1,000 leukocytes. For patients ( n = 10) whose smear turned positive again within less than 5 days of a first negative one, time to the second negative smear was used to Table 2 Baseline characteristics of patients hospitalized with P falciparum malaria ( n = 400). Hôpital Bichat, 1993 to N % Age categories (y) < Gender Female Male Birth and residence in a malaria-endemic country Not born, not resident Born, not resident Born, resident 25 6 Past malarial infection None or >1 y ago y ago 33 8 Unknown Self-treatment Yes No Severe malaria No Yes 29 7 Parasitemia on admission * < Platelet count on admission (/ L) <50, ,000 99, , Missing 23 6 Temperature on admission ( C) Median (IQR) 38.8 ( ) Gastrointestinal signs No Yes Treatment provided in hospital Quinine Halofantrine Co-artemether 12 5 Méfloquine 5 1 Others 12 3 IQR = interquartile range. * Parasitemia is expressed in percentage of parasitized red blood cells. Vomiting and/or diarrhea on admission. define PCT. These were transient repositivation, and none of the patients were retreated. Drug sensitivity of plasmodium strains was studied at the National Reference Center for Malaria Chemosusceptibility in Bichat-Claude Bernard Hospital, Paris. Data obtained by in vitro susceptibility tests of P falciparum to quinine and to halofanrine, the two most frequent drugs administered in this studied population, were used. In vitro assays were done

3 308 by isotopic tests [determination of the median inhibitory concentration (IC 50 )]. Patients were categorized as having been infected with susceptible, intermediate, or resistant strains. These categories were based on the following threshold values: 600 and 800 nm for decreased susceptibility and resistance to quinine, respectively; 4 and 6 nm for decreased susceptibility and resistance to halofantrine, respectively. Statistical Analysis The following patients were excluded from the analysis: those without parasite counts, those with malaria other than falciparum ( ovale, malariae, vivax ) or with a mixed plasmodial infection, and those treated on an ambulatory basis. For patients hospitalized twice during the study period ( n = 8), only the first infection was taken into account. We examined PCT in relation to the following variables: year of diagnosis, age, gender, birth and/or residence in a malaria-endemic country, recent history of malarial infection (<1 y), chemoprophylaxis, length of stay, self-antimalarial treatment before consulting, delay between return from the endemic area and onset of symptoms, delay between onset of symptoms and treatment, severity of malaria, parasitemia, platelet counts and temperature on admission, gastrointestinal signs, drug therapy, and in vitro drug sensitivity. Initial parasitemia was split into four categories: less than 1%, 1% to 1.9%, 2% to 4.9%, and greater than or equal to 5% of parasitized red blood cells; initial platelet count was divided in less than 50,000, 50,000 to 99,000, and greater than or equal to 100,000/ L; initial temperature was considered as less than 37.5 C, 37.5 C to 38.4 C, 38.5 C to 39.9 C, and greater than or equal to 40 C. PCT was considered as a continuous variable. Its normality was checked using graphical methods. Linear regression was performed for univariate and multivariate analysis using PCT as the dependant variable. The multivariate model considered all variables with p 0.20 in univariate analysis, together with age, gender, calendar year, and treatment, which were forced into the final model. Under the hypothesis that patients with none of the factors associated with longer PCT could have been treated as outpatients, the in-hospital outcome of such patients was examined to see whether they had developed any severe medical complications. Statistical analysis was performed by using the STATA 8.0 statistical package (Stata statistical software, Stata Corporation, College Station, TX, USA). Results Briand et al. Patients and Travels Characteristics A total of 745 malaria patients visited the travel clinic during the study period. Among them, the following patients were excluded: 30 having only a positive qualitative test (ICT malaria Pf, OptiMAL test, and QBC malaria test ), 71 with diagnosis of Plasmodium other than falciparum ( Plasmodium vivax, n = 28; Plasmodium ovale, n = 23; Plasmodium malariae, n = 9; and mixed infection, n = 11), 18 for whom plasmodial species was unknown, and 105 who received ambulatory care. Of the 521 remaining, 400 (77%) had data on PCT. The only significant difference between those with and without PCT was the proportion with recent (<1 y) history of malarial infection (8% vs 18%, respectively, p = 0.001). The baseline characteristics of the 400 hospitalized patients with PCT are summarized in Table 2. The median [interquartile range (IQR)] age of patients was 35 (27 43) years and the gender ratio (M/F) was Most of the patients (74%) were born in a malaria-endemic country (mostly in Africa) and lived in a nonendemic country (mostly in France). Past history of malaria was known for 309 patients, with 33 (11%) having had a malarial episode within the last year. Almost all patients visited endemic countries located in Africa: 63% ( n = 250) traveled to West Africa (mostly to Ivory Coast, Senegal, and Mali) and 26% ( n = 103) to Central Africa (mostly to Cameroon). Travels consisted mainly in family visit (87%) and the median (IQR) travel duration was 30 (18 60) days. More than half of the patients (55%, n = 218) did not use any chemoprophylaxis, only 9% ( n = 35) took an adequate one, defined as in Methods. Use of repellents and bed nets was 11% ( n = 44) and 2% ( n = 8), respectively. The median (IQR) delay between return from the endemic area and onset of symptoms and between onset of symptoms and treatment was 4 days (1 10 d; delay >60 d for four patients) and 4 (2 7) days, respectively. Most (93%) patients had uncomplicated malaria ( n = 371), 7% ( n = 29) had severe malaria, and no patient died. Temperature was normal (<37.5 C) in 14% of patients, while it was greater than or equal to 40 C in 16% of them. The median (IQR) initial parasitemia was 0.3% (0.05% 1.05%) and was higher among patients with severe malaria compared with those without severe malaria (5.8% vs 0.3%, p < ). Parasitemia greater than or equal to 5% accounted for 5% of all patients ( n = 19). The drugs most frequently used for treatment in hospital were quinine (49%, n = 196) after 1998, and halofantrine (42%, n = 166)

4 Hospitalization for Travel Malaria most often before Twenty-one patients received co-artemether ( ), 11 atovaquone plus proguanil, and 5 mefloquine. Predictors of PCT The mean PCT was 58 hours (median: 56 h, range: h). Factors associated with faster clearance were being born in a malaria-endemic country ( = 11.0 h for those still living in an endemic country and = 7.9 h for those living in a nonendemic country), having a recent history of malaria ( = 14.4 h), having already taken self-antimalarial therapy before consulting ( = 7.2 h), and having been treated with co-artemether or halofantrine as opposed to quinine ( = 19.0 and 11.1 h, respectively). In contrast, factors associated with longer PCT were calendar years (test for trend: p = 0.01), severe malaria ( = 34.1 h), temperature on admission greater than or equal to 40 C ( = 13.5 h), gastrointestinal signs ( = 9.1 h), higher initial parasitemia (test for trend: p < 0.001), and platelet count lower than 50,000/ L ( = 20.0 h). No association was found between PCT and age, gender, chemoprophylaxis, length of stay, delay between return from the endemic area and onset of symptoms or delay between onset of symptoms and treatment. In the subset of patients for whom data on in vitro drug sensitivity were available ( n = 115), there was no association between drug sensitivity and PCT (only 15 strains were drug intermediate and 4 were drug resistant). All factors associated with PCT in univariate analysis remained significantly associated with PCT in multivariate analysis except being born in a malaria-endemic country, self-antimalarial treatment, and antimalarial treatment used in hospital although marginally significant for the latter. Of note, interaction terms in the model showed that PCT was faster in Africans compared with non-africans in two circumstances: high parasitemia ( 1%) and severe malaria (data not shown) ( Table 3 ). Using factors associated with increased PCT (severe malaria, gastrointestinal signs, initial temperature 40 C, parasitemia >1%, platelet counts <50,000/ L), plus an age criteria (<15 or 65 y) as hospital admission criteria, 147 (37%) of the 400 hospitalized patients would have been eligible for outpatient care. Among these patients, the median (IQR) PCT was 48 (36 58) hours, and only three (2%) patients had PCT exceeding 4 days (from 120 to 128 h). None of them had any severe medical complication (ie, leading to transfer to intensive care or death) until hospital discharge. Increasing the parasitemia cutoff to 5% as in the FSSID consensus statement, 172 (43%) patients would have been eligible for 309 outpatient care, with mean (IQR) PCT of 48 (36 60) hours, and no severe medical complications during follow-up. Applying the model criteria to the entire population of falciparum malaria seen at the travel clinic during the study period (including those treated as outpatients or without data on PCT; see beginning of the Results), 45% of 621 patients would have been eligible for outpatient care (51% with the 5% parasitemia cutoff). Discussion In this study, we examined PCT and factors associated with it among 400 patients hospitalized with imported falciparum malaria. Mean (range) PCT was 58 (1 189) hours. Severe malaria, gastrointestinal signs, initial temperature greater than or equal to 40 C, parasitemia greater than or equal to 1%, and platelet counts less than 50,000/ L were all found significantly associated with longer PCT. Recent history of malarial infection was associated with shorter PCT. Those associations were found independently of each other and after having taken into account factors like history of malaria, self-antimalarial treatment, and birth/residence in a malaria-endemic area, which could reflect some level of protection against malaria. One limitation of our study is the exclusion of patients without PCT measurement ( n = 120). Such patients did not differ from those who were included in the analysis except for more common past history of malaria, a characteristic associated with shorter PCT. As a result, excluding them may have led to a slight overestimation of PCT. It has not, however, influenced the identification of factors associated with PCT, as we could not document any interaction between past history of malaria and any of the other factors associated with PCT. Thus, we are quite confident in the validity of the associations we found. While no study had yet addressed predictors of PCT as such, several of these factors had already been described elsewhere as predictive of falciparum malaria treatment failure: high admission parasitemia, 8 10 gastrointestinal symptoms, 9 11 and high temperature. 12 Factors associated with longer PCT in the model were the same as those used as clinical and biological hospitalization criteria in published guidelines. 2,3 Offering ambulatory care to patients aged 15 to 64 years with none of the factors associated with longer PCT in the study would have resulted in 147 (37%) patients receiving outpatient care. Although one cannot predict what would have happened, had these patients been treated as outpatients, it is comforting to know that none developed severe medical complications

5 310 Briand et al. Table 3 Factors associated with PCT in 400 patients with P falciparum malaria, in univariate and multivariate analysis. Hôpital Bichat, Paris, 1993 to 2000 Univariate analysis Multivariate analysis * n p Value CI p Value Birth and residence in malaria-endemic country Not born, not resident 117 (reference ) Born, not resident Born and resident Past malarial infection None, or >1 y ago (reference ) y ago ; Unknown ; 3.5 Self-antimalarial treatment No 280 Yes Severe malaria No 371 Yes < ; Temperature on admission ( C) < ; ; 12.5 > ; 18.2 Gastrointestinal signs No 275 Yes ; Parasitemia on admission # <1 (reference ) ; < ; 24.9 < Test for trend ; 32.6 p < Platelet count on admission (/ L) <50, ; ,000 99, < ; ,000 (reference ) 199 Missing ; 8.6 Curative treatment Quinine ± doxycycline 199 (reference ) Halofantrine ; 4.0 Co-artemether < ; Other drugs ** ; 18.1 PCT = parasite clearance time; CI = 95% confidence interval. * Multivariate analysis was performed on 400 patients infected with P falciparum ( R 2 = 31%, 0 = 40.2 h); other variables tested and not significantly associated with PCT were chemoprophylaxis, delay between return from the endemic area and onset of symptoms, delay between onset of symptoms and treatment, and in vitro drug sensitivity. Age, gender, and calendar years were forced into the multivariate model. Coefficient of regression ( ) is the mean variation of the PCT for each category displayed compared with the referent category (reference). For variables with more than two categories, the p value of the global test is given. To be born and/or to live in a malaria-endemic country. According to the World Health Organization criteria (1990 and 2000). Diarrhea and/or vomiting on admission. # Parasitemia is expressed in percentage of parasitized red blood cells. ** Atovaquone/proguanil, sulfadoxine/pyriméthamine, or mefloquine. during hospitalization. Applying the model criteria to the entire population of imported falciparum malaria seen during the study period (including those treated as outpatients or those with no PCT data) would have allowed outpatient care for 45% of patients. The model may further help to refine some hospitalization criteria. Indeed, two criteria present in the D Acremont s article, temperature greater than or equal to 40 C and low platelet counts, are absent from the FSSID guidelines. Both turned out

6 Hospitalization for Travel Malaria to be predictive of longer PCT independent of other severity criteria (eg, clinical severity and high parasitemia) and may therefore be worth considering in the list of hospitalization criteria. Recommendations also differ regarding the cutoff to use for parasitemia: 5% in the FSSID guidelines, 2% in the D Acremont article, while Whitty and Lockwood raised concerns about the risk of misleading low parasitemia in case of parasite sequestration. 5 The model identified longer PCT for high parasitemia as of 1% (see Table 3 ). We therefore suggest using a value of 1% and above as a criterion for hospitalization instead of 5% as currently recommended in the FSSID guidelines. Lowering the cutoff from 5% to 1% would have led to 36 additional hospitalizations during the study period (6% of the total falciparum malaria patients). Finally, we found, as widely demonstrated, 13 that patients treated with artemisinin derivatives had a significantly shorter PCT. Those patients should be particularly considered for ambulatory care if they do not meet severity criteria. One may wonder whether differences exist between immigrant and nonimmigrant populations for management of imported falciparum malaria. 14,15 In this study, once recent history of malaria was taken into account in multivariate analysis, PCT was no longer different between Africans and non-africans, except in two circumstances: high ( 1%) parasitemia and severe malaria, in which Africans had faster clearance compared with non-africans. In addition to the factors identified in this model, sociocultural characteristics and logistical issues need to be taken into consideration to decide on ambulatory care for imported falciparum malaria (see Table 1 ). Keeping them in mind, close to half of patients seen in this travel clinic could have received ambulatory care, with significant cost savings for the society, and improved comfort for the patients. Acknowledgments We are grateful to the travel clinic staff and the laboratory staff at Bichat-Claude Bernard Hospital who collaborated on this study. Declaration of Interests The authors state that they have no conflicts of interest. References Muentener P, Schlagenhauf P, Steffen R. Imported malaria ( ): trends and perspectives. Bull World Health Organ 1999 ; 77 : The 12th Consensus Conference of Anti-infectious Therapy of the French-speaking Society of Infectious Diseases, 14 April Management and prevention of imported Plasmodium falciparum malaria. Arch Pediatr 2000 ; 7 : D Acremont V, Landry P, Darioli R, et al. Treatment of imported malaria in an ambulatory setting: prospective study. BMJ 2002 ; 324 : Evans MR, Day JH, Behrens RH. Prevention and treatment of malaria in UK travellers. Hosp Med 2000 ; 61 : Whitty CJM, Lockwood DNJ. Commentary: should patients with imported malaria routinely be admitted? BMJ 2002 ; 324 : Warrell DA, Molyneux ME, Beaks PF. Severe and complicated malaria. Trans R Soc Trop Med Hyg 1990 ; 84 : Bruce-Chwatt LJ. Chemotherapy of malaria. World Health Organization. 2nd Ed. Geneva, Switzerland : WHO, Price RN, Nosten F, Luxemburger C, et al. Artesunate/mefloquine treatment of multi-drug resistant falciparum malaria. Trans R Soc Trop Med Hyg 1997 ; 91 : Fontanet AL, Walker AM. Predictors of treatment failure in multiple drug-resistant falciparum malaria: results from a 42-day follow-up of 224 patients in eastern Thailand. Am J Trop Med Hyg 1993 ; 49 : Ter Kuile FO, Luxemburger C, Nosten F, et al. Predictors of mefloquine treatment failure: a prospective study in 1590 patients with uncomplicated falciparum malaria. Trans R Soc Trop Med Hyg 1995 ; 89 : White NJ. Why is it that antimalarial drug treatments do not always work? Ann Trop Med Parasitol 1998 ; 92 : Dorsey G, Gasasira AF, Machekano R, et al. The impact of age, temperature, and parasite density on treatment outcomes from antimalarial clinical trials in Kampala, Uganda. Am J Trop Med Hyg 2004 ; 71 : McIntosh HM, Olliaro P. Artemisinin derivatives for treating uncomplicated malaria. Cochrane Database Syst Rev 2000 ; 2 : CD Bunn A, Escombe R, Armstrong M, et al. Falciparum malaria in malaria-naïve travellers and African visitors. QJM 2004 ; 97 : Bouchaud O, Cot M, Kony S, et al. Do African immigrants living in France have long-term malarial immunity. Am J Trop Med Hyg 2005 ; 72 :