Published ahead of print on October 31, 2002, doi:10.1164/rccm.200206-631OC
© 2003 American Thoracic Society
The Clinical Spectrum of Severe Imported Falciparum Malaria in the Intensive Care UnitReport of 188 Cases in AdultsClinique de Réanimation des Maladies Infectieuses et Tropicales, Hôpital Bichat-Claude Bernard, Assistance Publique-Hôpitaux de Paris, Paris; Service de Réanimation, Hôpital André Mignot, Centre Hospitalier de Versailles, Le Chesnay; Service des Maladies Infectieuses et Tropicales, Centre Hospitalier d'Orléans-La Source, Orléans Cedex 2; Département de Biostatistique et Informatique Médicale, INSERM U-444, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris; and Service de Parasitologie, Hôpital Bichat-Claude Bernard, Paris, France Correspondence and requests for reprints should be addressed to Michel Wolff, Service de Réanimation des Maladies Infectieuses, Hôpital Bichat Claude-Bernard, AP/HP, 40 rue Henri Huchard 75018, Paris, France. E-mail: michel.wolff{at}bch.ap-hop-paris.fr
Little is known about severe imported malaria in nonendemic industrialized countries. The purpose of this retrospective study was to describe the clinical spectrum of severe imported malaria in adults and to determine factors that were present at admission and were associated with in–intensive care unit mortality. This retrospective study evaluated the 188 patients who were admitted to our intensive care unit in 1988–1999 with severe and/or complicated imported malaria. Among them, 93 had strictly defined severe malaria, and 95 had less severe malaria. The mean age was 38 years, 51% of patients were nonimmune whites, 94% acquired Plasmodium falciparum in sub-Saharan Africa, and 96% had taken inadequate antimalarial chemoprophylaxis. Mortality was 11% (10 patients) in the severe malaria group, whereas no patients died in the less severe malaria group (p = 0.002). In the bivariable analysis, the main factors associated with death in the severe malaria group were the Simplified Acute Physiology Score, shock, acidosis, coma, pulmonary edema (p < 0.001 for each), and coagulation disorders (p = 0.002). Bacterial coinfection is not infrequent and may contribute to death. Severe imported malaria remains a major threat to travelers. In our population, the most relevant World Health Organization major defining criteria were coma, shock, pulmonary edema, and acidosis.
Key Words: coma • shock • acidosis • respiratory distress syndrome • infection Plasmodium falciparum malaria remains a major public health problem in endemic areas, with approximately 1.5 to 2.7 million deaths each year. Tropical African countries are estimated to contribute more than 90% of the total malaria incidence and the great majority of malaria deaths, especially in children who are younger than 5 years (1). In Europe, as well as in nonendemic industrialized areas, imported malaria is generally diagnosed in travelers or military personnel, as well as in immigrants from endemic countries (2, 3). Such imported cases have increased worldwide, with approximately 16,000 diagnoses in Europe each year, of which more than 5,000 occur in France, and approximately 1,300 in the United States (2–4). Severe P. falciparum malaria has been extensively studied in endemic areas. The defining criteria of severe and complicated falciparum malaria established by the World Health Organization (WHO) in 1990 (5) and revised in 2000 (6, 7) were developed mainly on the basis of studies performed in tropical areas. In addition, few studies have focused on severe imported falciparum malaria in patients admitted to the intensive care unit (ICU), and little is known about the relevance of the WHO criteria in populations living in nonendemic industrialized areas. The objective of this retrospective study was to assist clinicians in diagnosing and managing imported falciparum malaria by (1) describing the clinical spectrum of severe imported malaria in adults, (2) identifying predictors of mortality in patients admitted to the ICU with major WHO criteria, and (3) assessing the relevance of WHO definitions in this population.
The study was performed in the ICU of a 1,200-bed university-affiliated teaching hospital. Patients who admitted to the ICU with a diagnosis of malaria between 1988 and 1999 were retrospectively evaluated. Information abstracted from the ICU records was used to complete a standardized case report form for each malaria episode (F.B., L.H.). The data were entered into a database at the Département de Biostatistique et Informatique Médicale (4) and were reviewed for inconsistencies and data entry errors (C.A., F.B.). Imported malaria was defined as malarial infection acquired in an endemic country and treated in France (8). Patients with asexual forms of P. falciparum in the blood were classified into two mutually exclusive groups, severe malaria and less severe malaria, according to the 1990 WHO definition of severe and complicated malaria (5). Severe malaria was strictly defined as the presence of one or more of the following WHO major criteria at admission or within the first 3 days: (1) unrousable coma with a Glasgow Coma Scale score of nine or less (9); (2) anemia with hemoglobin of less than 5 g/L; (3) renal failure with serum creatinine of more than 265 µmol/L; (4) pulmonary edema with the presence of criteria for acute respiratory distress syndrome or acute lung injury (10); (5) hypoglycemia with blood glucose of less than 2.2 mM; (6) circulatory collapse with systolic blood pressure of less than 80 mm Hg despite adequate volume repletion; (7) spontaneous bleeding and/or disseminated intravascular coagulation; (8) repeated generalized seizures; (9) acidemia (pH of less than 7.25) or acidosis (serum bicarbonate of less than 15 mM); and (10) macroscopic hemoglobinuria if definitively related to acute malaria. All patients admitted to the ICU with falciparum malaria but lacking the criteria for severe malaria were classified as having less severe malaria. These patients had one of the WHO minor criteria (impaired consciousness but rousable, parasitemia of more than 5%, extreme weakness, temperature of more than 40°C, and jaundice or total bilirubin of more than 50 µmol/L) and/or belonged to a high-risk population (5). Patients with blackwater fever (5, 11) were not included. Community-acquired coinfection was defined as any infection diagnosed within the first 2 days of hospitalization. Infections occurring later were considered nosocomial. From the medical records, we abstracted information on demographics, current and previous medical history, area where the malaria was acquired, and chemoprophylaxis. Chemoprophylaxis was considered inappropriate when either absent or inadequate (12). The degree of immunity to malaria was estimated as follows: (1) adult immigrant Africans living in France were considered partially immune, as they had been exposed to P. falciparum during childhood, (2) patients who had been living in an endemic area for at least 2 years at the time of the diagnosis were presumed semi-immune, (3) Europeans who traveled occasionally to endemic areas were considered nonimmune. Clinical, biologic, and imaging findings were recorded, as were the treatments and the outcome at ICU discharge. Severity at ICU admission was assessed using the Simplified Acute Physiology Score II (13). When P. falciparum cultures were available, previously described in vitro techniques were used to test susceptibility to antimalarial drugs (14, 15). Neuroimaging was performed in patients with suspected severe cerebral edema or with focal neurologic deficits. In the group with severe malaria, intravenous quinine was given as recommended, with a loading dose in patients who had not received quinine, mefloquine, or halofantrine before ICU admission. In our ICU, symptomatic treatment of comatose patients includes 30° semirecumbent position, maintenance of serum sodium concentration of approximately 145 mM and of PaCO2 of approximately 35–40 mm Hg (if mechanically ventilated), serum glucose optimization, and sedation if necessary. The use of hypertonic mannitol to treat severe cerebral edema is at the physician's discretion. In our unit, intracranial monitoring devices are not used to monitor patients with cerebral edema during cerebral malaria, one of the main reasons being the high risk of bleeding. In the less severe group, antimalarial treatment was left to physician's discretion. When feasible and with the consent of the family, an autopsy was performed, with a gross examination and a histologic evaluation. Statistical analysis was performed with SAS version 8.0 (SAS Inc., Cary, NC). Comparisons were done using the Fisher exact test or chi-square test for categorical data and Wilcoxon tests or Kruskal-Wallis tests for continuous data, as appropriate. When studying factors related to ICU mortality, multivariable analysis was not performed because the number of nonsurvivors was small (16). For multiple comparisons, Holm's procedure, as modified by Hochberg and colleagues, was used to adjust p values (17).
During the study period, 5,400 patients were admitted to our ICU, including 199 (3.7%) with P. falciparum malaria. Ten patients with blackwater fever (who have been reported in part elsewhere [11]) were excluded, as was a patient with uncomplicated malaria who was referred to the ICU in a moribund state after a cardiac arrest ascribed to halofantrine-induced toxicity occurring on a background of previously unsuspected advanced rheumatic cardiomyopathy. This left 188 patients for our study, including 93 with strictly defined severe malaria and 95 with less severe malaria. The median age in the overall population was 38 years (range of 14 to 74), and there was a predominance of males. Twenty-seven percent of the patients were partially immune Africans. Seventeen percent were presumed semi-immune patients of any ethnicity, and 51% were nonimmune whites; in nine patients, malaria immune status was not evaluable. Patients who were presumed semi-immune included 14 Africans who were born and lived in endemic areas, one Asian, and 17 Europeans who lived in sub-Saharan Africa for a median time of 87 months (interquartile range of 36–168). None of these 18 patients took prophylactic therapy, and all but three of them lived in high-transmission areas. All but 12 of the 188 patients acquired P. falciparum malaria in sub-Saharan Africa. Antimalarial chemoprophylaxis was inappropriate in 180 of 188 cases (96%). Overall, 59 and 45% of the tested strains were resistant to chloroquine and cycloguanil/proguanil, respectively. In contrast, all strains were susceptible to quinine, and only one strain (2%) was resistant to halofantrine. The median time from symptom onset to quinine therapy initiation was 4 days. Mortality in the severe malaria group was 11% (10 patients), whereas no patients died in the less severe group (p < 0.0001). Consequently, the two groups were analyzed separately.
Severe Malaria Group
The main characteristics and the pattern of WHO defining criteria at ICU admission are given in Tables 2 and 3 , respectively. In 22 patients, 27 WHO major defining criteria were met for the first time on Day 2 or Day 3 of treatment: coma (n = 5), bleeding (n = 5), severe anemia (n = 5), pulmonary edema (n = 4), shock (n = 4), renal failure (n = 3), and acidosis (n = 1). Median parasitemia at ICU admission in the overall group of patients with severe malaria, in the survivors, and in the nonsurvivors was 4.1, 3.5, and 18.2%, respectively (in the same groups, mean ± SD parasitemia at ICU admission was 10.4 ± 14.5, 9.4 ± 14.2, and 19.5 ± 18.2%, respectively).
Community-acquired bacterial coinfections were diagnosed in 13 patients (14%), 9 among the survivors (11%) and 4 among the nonsurvivors (40%) (p = 0.03). Pneumonia and bacteremia accounted for 35 and 29% of the 17 involved sites, respectively (Table 4) . There were 34 episodes of nosocomial infection in 23 (25%) patients, including 18 (22%) in survivors and 5 (50%) in nonsurvivors (p = 0.1). The main sites of nosocomial infections were pneumonia (n = 19), primary bacteremia (n = 4, including Escherichia coli in two patients and Staphylococcus aureus in one and Candida albicans in one), and sinusitis (n = 4).
Unrousable coma was present in 39 patients (34 at admission), with a mean duration of 4 days (interquartile range of 3–5). In addition, an unrousable coma developed secondarily on ICU Day 2 or Day 3, that is, 5.3% of the overall group of 93 patients and 8.4% of the 59 patients whose GCS score was more than nine at ICU admission. Eight patients had focal neurological abnormalities. Twenty-eight patients underwent cerebral imaging (computed tomography and/or magnetic resonance imaging), which was abnormal in 10 (computed tomography, n = 7; magnetic resonance imaging, n = 5). Bilateral abnormalities of the deep white matter where seen in three patients. In two patients with diffuse and massive cerebral edema, computed tomography showed either a hematoma of the vermis or cortical infarcts. In one patient, computed tomography performed immediately after ICU admission showed cerebral edema of moderate severity, but cerebral magnetic resonance imaging performed later was normal. One patient had a frontal hematoma. Another had meningeal enhancement, and two had residual changes from ischemic brain lesions. Five patients had neurologic abnormalities at ICU discharge: behavioral disturbances in three (none of whom took mefloquine); vegetative state related to severe anemia, sustained hypotension, and prolonged hypoxemia (during air lifting) in one; and ICU-induced peripheral polyneuropathy in one. Shock occurred in 24 patients (20 at admission), of whom 16 received dopamine (more than 5 µg/kg/min) and 8 received epinephrine and/or norepinephrine. Right heart catheterization with a Swan-Ganz catheter was performed in 11 patients and was consistently disclosed a hyperdynamic pattern. At the time of shock, bacterial coinfection was diagnosed in 8 of the 24 patients (four cases each of community-acquired and early nosocomial infection). Forty patients (43%) required mechanical ventilation for a median duration of 6 days (interquartile range of 5–10). Pulmonary edema was present in 14 patients (10 on admission), of whom 12 had acute respiratory distress syndrome, and 2 had acute lung injury. At the time of pulmonary edema, bacterial coinfection was diagnosed in three patients (community-acquired pneumonia in two and early-onset nosocomial pneumonia in one). Acute renal failure occurred in 50 patients (47 at admission), of whom 29 required renal replacement therapy (intermittent hemodialysis, continuous venovenous hemodiafiltration, and peritoneal dialysis in 27, 3, and 2 patients, respectively). Twenty-six patients left the ICU with a plasma creatinine level greater than 265 µM, but all recovered gradually. Significant bleeding, which occurred in nine patients, included gastrointestinal bleeding (n = 5), cerebral hematoma (n = 2), diffuse alveolar hemorrhage (n = 1), and uterine bleeding (n = 1). Blood (median of three units), platelet (median of one unit), and fresh plasma (median of two units) transfusions were required in 51, 14, and 9 patients, respectively. Most transfusions were performed after Day 3. All patients received intravenous quinine. Only 20 (21.5%) patients received a quinine loading dose; the remaining patients had received quinine, mefloquine, or halofantrine within hours before ICU admission. Tetracycline was given also in 16 patients (17%) before the results of quinine susceptibility testing. No patients underwent exchange transfusion. Ten patients died (11%). The main characteristics of these patients and causes of death are shown in Table E1 of the online supplement. Eight patients died within 1 week after ICU admission. Among them, three died from cerebral edema and five from septic shock, which was caused by concomitant bacterial infection in all but one patient. Two patients died later from disseminated aspergillosis on Day 10 and from splenic rupture on Day 57, respectively. The median ICU stay duration in survivors was 7.5 days (interquartile range of 4–13). In the bivariable analysis, after adjustment for multiple comparisons, four WHO major criteria present within 24 hours after admission were significantly associated with in-ICU mortality, namely unrousable coma, pulmonary edema, shock, and metabolic acidosis (p < 0.001 for all four) (Tables 2 and 3).
Less Severe Malaria Group
Imported malaria is a growing problem in many nonendemic areas throughout the world. However, most studies focusing on severe imported malaria are case reports or included fewer than 30 patients (18–26). Moreover, the definition of severe malaria seems to have varied, with failure to use recognized criteria or to distinguish clearly between major and minor 1990 WHO criteria. We report a single-center study of 188 consecutive adults admitted to an ICU for imported malaria, including 93 with strictly defined severe malaria and 95 with less severe disease. To the best of our knowledge, this is the largest study of severe imported malaria in a nonendemic industrialized country. Almost all the patients acquired malaria in sub-Saharan Africa, in accordance with the epidemiology of malaria in France (2, 3). Consequently, it is of interest to note that all tested P. falciparum strains were susceptible to quinine, as reported in Africa by the WHO (12). The largest patient group in the present series was composed of occasional travelers of European descent who were considered nonimmune. Among these patients, severe malaria was significantly more common than less severe malaria. Moreover, all of the deaths occurred in the Europeans. Adult African patients who were either long-term residents of France or who lived in Africa and traveled to France had a lower risk of severe malaria and a better prognosis (3, 25). The main epidemiologic feature was the very high proportion of patients with inappropriate chemoprophylaxis, which may have contributed to the severity of the disease (27). It should be emphasized that none of the patients who died had taken chemoprophylaxis. Using the 1990 WHO definition of severe malaria (5), we divided our ICU patients with P. falciparum malaria into two groups. Those with at least one major criterion were the more severely ill, required a high level of intensive care, and included all deaths. The others either had less severe disease with only minor WHO criteria or were at risk for complications. Whether patients with only minor criteria require ICU admission remains controversial (20). However, it should be kept in mind that rapid neurologic deterioration can occur within 24 hours of quinine therapy initiation. In 2000, the WHO revised its 1990 definition of severe and complicated malaria and provided additional information about the frequency and prognostic value of each criterion in adults and children (6, 7). However, little is known about the frequency and prognostic value of these criteria in adults with severe imported malaria. In our study, which included only adults, severe anemia, repeated seizures, hypoglycemia, and hemoglobinuria were very infrequent and of questionable prognostic value. The first three of these complications are far more common in children with severe malaria (5–7, 28). In contrast with the WHO evaluation (6, 7) and studies in endemic areas (6, 29–31), acute renal failure in this study was not significantly different in survivors and nonsurvivors. This finding is probably explained by differences in the availability and quality of dialysis between endemic areas and industrialized European countries. In a study performed in Vietnam, mortality in patients with malaria and renal failure was 75% without dialysis and 26% when dialysis was available (32). Moreover, the authors stated that more effective dialysis or diafiltration might further reduce the mortality rate (32). Furthermore, the same team showed recently that hemofiltration was superior to peritoneal dialysis in the treatment of infection-associated acute renal failure (including 69% of patients with severe malaria) (33). Our prognostic study was limited because the number of nonsurvivors was too small for a valid multivariable analysis. Nevertheless, in our bivariable analysis, four WHO major criteria present within 24 hours of admission were strongly correlated with in-ICU mortality, namely, unrousable coma, pulmonary edema, shock, and metabolic acidosis, even after adjustment for multiple comparisons. Unrousable coma is a common clinical presentation of severe malaria, with frequencies ranging from 17–52% in studies performed in endemic areas (6, 29, 31). A previous study of severe imported malaria reported an incidence of 39% (18), which is very similar to the incidence in our patients (37%). Cerebral malaria was both common and strongly correlated with mortality. Thus, a higher survival rate could perhaps be achieved with a more standardized and more aggressive treatment approach to coma. Pulmonary edema, septic shock, and metabolic acidosis are most likely to occur in patients with severe multisystemic P. falciparum malaria and appear to contribute to mortality. Although complicating and associated bacterial infections have been reported in patients with severe malaria (5, 6, 28), most studies were performed in endemic areas and did not report detailed bacteriologic data or distinguish between community-acquired and hospital-acquired infections. In the control groups of the two steroid studies in cerebral malaria, pneumonia or bacteremia developed in 10 and 36% of patients (34, 35). In our study, 10% of patients had community-acquired bacterial coinfection, and 19% had nosocomial infections. As expected, most of these infections occurred in the group with severe malaria. The most common community-acquired infections in patients with severe malaria are aspiration pneumonia favored by impaired consciousness and primary Gram-negative bacteremia possibly due to impaired splanchnic perfusion (6, 36, 37). Moreover, transient malaria-induced immunosuppression related to multiple factors may contribute to the occurrence and severity of early and nosocomial infections (6, 28, 38). Therefore, in the most severely ill patients, bacteremia and/or pneumonia should be suspected and empirically treated without delay as they may contribute to death (6, 37). Hyperparasitemia (more than 5%) and jaundice, two minor criteria in the 1990 WHO definition (5), are major criteria in the 2000 definition (6, 7). In our study, these two criteria were common in both the severe and the less severe groups, suggesting that in themselves they may not be of strong prognostic value.
In this study, in-ICU mortality was 5% overall but was 11% among the 93 adults with strictly defined severe imported malaria. According to large studies performed in endemic areas, mortality among adults with severe malaria treated by highly trained teams ranges from 10–25% (5, 6). Moreover, mortality has not decreased during the last 20 years, despite numerous new therapeutic approaches, including dexamethasone, anti–tumor necrosis factor- In conclusion, severe imported malaria still carries a relatively high mortality rate, even when treated under optimal conditions in a highly experienced ICU. Although WHO criteria are not all relevant to imported malaria in adults, the presence in the emergency room of any degree of neurologic, acid-base, circulatory, or pulmonary failure should lead to ICU admission. Bacterial coinfection is not infrequent and may contribute to death. Finally, it should be kept in mind that most of our patients did not take appropriate malaria chemoprophylaxis. Thus, the best way to reduce the number of deaths caused by imported malaria is to improve the quality of prevention.
The authors thank the nurses and the physicians of the ICU for their excellent care of the patients and Antoinette Wolfe for helping to prepare the manuscript. Presented in part at the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington DC: American Society for Microbiology; 1999. p. 730.
This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form June 28, 2002; accepted in final form October 27, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
monoclonal antibodies, hyperimmune IgG, deferoxamine, and pentoxifylline, none of which has been found clearly effective (
