Problems in the Therapy for Imported Malaria in the United States

Background  Errors in the diagnosis of imported malaria are increasingly recognized. However, there are few data on the treatment of malaria in the United States.

Methods  Medical records were reviewed for 83 patients with microscopically confirmed malaria at Cook County Hospital, Chicago, Ill, between 1991 and 1999.

Results  Errors in drug treatment occurred in 25% of patients in this study. The most common error in therapy was the failure to prescribe primaquine to eradicate the liver forms of Plasmodium vivax. Another 5 patients with P vivax received an inappropriate drug regimen. Errors in Plasmodium falciparum therapy occurred in 5 patients. All patients received an inappropriate drug regimen. While the clinical symptoms and signs do not help distinguish the infecting Plasmodium species, the travel history is extremely helpful in guiding drug selection. Non–infectious diseases specialists are more likely to make errors in therapy than are infectious diseases specialists.

Conclusions  Despite widely published guidelines on the treatment of malaria, there are frequent errors in the therapy for malaria. A detailed travel history emphasizing the duration and country of travel should be sought. Primaquine should be included in the primary prescription for the treatment of P vivax infection. Improvements in the therapy for malaria can be made with the aid of an infectious diseases specialist and/or a tropical medicine specialist.

MALARIA IS the most important parasitic disease of humans, affecting more than 500 million people and causing between 1 and 3 million deaths each year. Because of the increase in international travel and immigration, local hospitals and physicians are increasingly likely to encounter individuals at risk for malaria. In 2000, the Centers for Disease Control and Prevention received reports of 1402 cases of malaria with onset of symptoms in the United States.¹ This incidence is the third highest number of reported cases since 1980 and reflects a continuing trend in the growing numbers of cases imported into the United States.

While most malaria cases in the United States occur in persons who have traveled or immigrated from malaria-endemic parts of the world, the vector of malaria transmission, the Anopheles mosquito, exists in the United States. Thus there is the potential for local transmission of malaria, particularly via infected travelers who are inadequately treated or for whom chemoprophylaxis fails.²

While a number of studies have highlighted the problems of misdiagnosis of malaria,^3-8 there are few data on the drug treatment of imported malaria.⁹ Missed diagnosis and/or delays in diagnosis of malaria are frequently the result of physician failure to consider the diagnosis of malaria.³ However, as unaccustomed as physicians are to the presentation of malaria, they are even less familiar with the therapy for this disease. This is in part owing to the perplexing choices of drugs for malaria and the need to individualize drug selection based on (1) the Plasmodium species and (2) the severity of infection. Over the last decade, this problem has been further compounded by the emergence of drug-resistant malaria, requiring drug therapy to be also individualized based on (3) the geographic area of acquisition.¹⁰

The objectives of the present study were to (1) evaluate the travel patterns and clinical characteristics of imported malaria among recent immigrants and/or foreign-born US residents returning to their country of origin; (2) evaluate the drug therapy for malaria among infected individuals; and (3) identify useful tools to aid the physician in the selection of antimalarial drugs.

The site of our study, Cook County Hospital, Chicago, Ill, is a 591-bed inner-city hospital that serves the medical needs of more than 5 million people in the country's second largest county, including a large number of immigrants. It is also a major teaching hospital with an academic partnership with Rush Medical College.

All smear-positive malaria cases diagnosed between 1991 and 1999 by the microbiology laboratory were reviewed. Diagnoses and Plasmodium species identification were based on thin or thick malaria-positive smears stained with 5% Giemsa stain and examined by experienced laboratory personnel. Medical records were reviewed, and data were entered into a malaria chart review form. Information collected fell into the following categories: demographic information, travel history, immigration status, history of malaria, use of antimalarial chemoprophylaxis, onset of disease, symptoms and signs, laboratory indices, glucose-6-phosphate dehydrogenase (G6PD) testing in patients given primaquine, primary diagnosis, drug therapy, and fever clearance time. Those classified as semi-immune had either reported a history of malaria or been born in and immigrated from a malaria-endemic area.³

Where data were incomplete or missing, denominators for percentages are specified. Appropriateness of drug therapy was assessed based on published guidelines from the Centers for Disease Control and Prevention¹¹ and the World Health Organization.¹² Comparisons of variables of interest were performed using the t test of means and the χ² test or Fisher exact test for dichotomous variables. Significance was defined as P≤.05 (2-tailed). Odds ratios and confidence intervals were calculated using SPSS software (version 10.0; SPSS Inc, Chicago, Ill).

Between 1991 and 1999, 88 cases of malaria were diagnosed in semi-immune individuals. Eighty-three cases were available for review and were included in the study.

Patient characteristics are summarized in Table 1. All patients acquired malaria while traveling or living in malaria-endemic areas. Most patients had traveled from Africa (46/83; 55%) or the Indian subcontinent (26/83; 31%). Fifty-five patients (66%) were recent immigrants from malaria-endemic areas; the other 28 (33%) were foreign-born US residents traveling to their country of origin. None of the US residents reported using chemoprophylaxis despite a documented history of malaria in 52 (63%). Only 3 (5%) of 55 new immigrants used chemoprophylaxis; 2 of these 3 were pregnant women.

A microscopic diagnosis of malaria was made on analysis of the first smear in all but 2 cases (98%). Plasmodium vivax and Plasmodium falciparum were identified in blood smears from 46 (56%) and 37 (45%) persons, respectively. More than 1 species was present in 2 patients (Table 2).

Table 2 lists the areas of acquisition of the respective malaria species. Most cases of malaria in patients returning from the Indian subcontinent were due to P vivax (odds ratio, 9.7; 95% confidence interval, 2.4-38.2). Conversely, patients returning from Africa were more likely to have P falciparum malaria (odds ratio, 6.6; 95% confidence interval, 2.6-17.0). Although it is reported that P vivax malaria is rare in West Africans because most of the population lacks the Duffy antigen,¹³ 5 cases of P vivax malaria were identified in West Africans.

The mean duration from time of departure from the malaria-endemic area to the onset of symptoms was significantly shorter for P falciparum than for P vivax infection (11.2 days vs 47 days; P = .03). However, there was no significant difference in the duration of symptoms before patients sought medical attention (mean, 6.1 days) between those with P falciparum and P vivax infection or between US residents and recent immigrants.

The most common signs and symptoms reported at the time of initial positive smear are listed in Table 1. There were no significant differences in symptoms between patients with P falciparum and P vivax infection. Most patients (66/83; 80%) reported daily fevers; there was no significant correlation between fever periodicity and parasite species. Splenomegaly was detected in 7 patients, 6 of whom had P falciparum infection (odds ratio, 2.1; 95% confidence interval, 1.4-3.2).

The most common laboratory abnormalities were thrombocytopenia and anemia (Table 1). There were no significant differences in laboratory values between patients with P falciparum and P vivax infection or between US residents and recent immigrants.

Patients were admitted to the hospital in 72% of cases (n = 60); 5 patients were admitted to the intensive care unit. There were no differences in admission rates between patients with P falciparum and P vivax infection or between US residents and recent immigrants. However, the mean length of stay was longer for patients infected with P falciparum than for those infected with P vivax (3.4 days vs 2.5 days; P = .05). Three patients fulfilled the World Health Organization criteria for severe or complicated malaria.¹⁴ There were no case fatalities. There was no significant difference in the average time to defervescence (fever clearance time) between P vivax and P falciparum (mean, 1.9 days vs 2.0 days).

The antimalarial therapy prescribed for the treatment of malaria is summarized in Table 3. One pregnant patient was infected with P vivax. Overall, errors in antimalarial treatment occurred in 20 treated patients. The most common error in therapy (12 patients) was the failure to prescribe primaquine phosphate for P vivax to eradicate forms of the parasite (hypnozoites) that survive in the liver.¹⁵ The next most common error was the use of an inappropriate drug regimen for the infecting Plasmodium species: 5 patients with P vivax and 5 patients with P falciparum. In addition, 2 patients with P vivax infection received no documented therapy because of misdiagnoses.

Because primaquine can cause severe hemolysis in patients with G6PD deficiency, testing for G6PD was assessed in all patients prescribed primaquine for radical cure of P vivax or Plasmodium ovale infection.¹⁶ Twelve (41%) of 29 patients were prescribed primaquine without G6PD testing. Of note, unnecessary bone marrow biopsies were performed in 7 patients diagnosed with malaria because of a failure to properly recognize the associated hematologic complications of malaria.

In addition, we assessed a random sample of 32 of our 83 cases to see if the use of an infectious diseases (ID) specialist improved the management of malaria. We found that the use of an ID specialist decreased the risk of errors in malaria therapy (Fisher exact test, P = .03). Only 1 error in therapy occurred among 26 cases seen by an ID specialist compared with 3 errors in therapy among 6 cases without an ID consult.

This study examined the treatment of imported malaria among a unique population consisting of newly arrived immigrants or foreign-born US residents who were traveling to their country of origin. This is in part a reflection of the Cook County Hospital patient population, which consists mainly of patients without health insurance and those who have recently arrived in the country.

Malaria therapy is complex because of the need to individualize drug selection based on the infecting Plasmodium species, severity of infection, and the region of acquisition. Once intraerythrocytic forms of the parasite are identified on a peripheral blood smear, it is often necessary to begin empiric antimalarial therapy pending the final identification of the Plasmodium species. We found that while the clinical symptoms and signs are generally unhelpful in determining the infecting species, the travel history can provide useful clues. Patients from Africa were more likely to be infected with P falciparum (31 of the 35 P falciparum cases were acquired in Africa). Conversely, 24 of the 26 patients from the Indian subcontinent were infected with P vivax. In addition, the mean duration from time of departure from the malaria-endemic area to the onset of symptoms was significantly shorter for P falciparum than for P vivax infection (11.2 days vs 47 days; P = .03). These findings are worth emphasizing because previous studies on imported malaria in US-born travelers and semi-immune individuals have also confirmed these observations.^9,17

Despite widely published and available guidelines, errors in antimalarial therapy occurred in 25% of treated patients (20 patients). This figure is consistent with estimates from 2 previous studies in which errors in drug therapy occurred in 9% and 48% of patients.^9,17 Retrospective data from the Centers for Disease Control and Prevention indicate that errors in therapy occur more frequently in the life-threatening P falciparum malaria (approximately 30% of returning travelers).¹⁸ However, we observed fewer errors in the treatment of P falciparum infection (5 [14%] of 35 patients) than in the treatment of P vivax infection (15 [35%] of 43 patients).

The error in 10 of the patients infected with P vivax involved the failure to prescribe primaquine. These patients had no obvious contraindications for primaquine therapy (eg, G6PD deficiency, history of primaquine allergy, or pregnancy). It is common practice to withhold primaquine treatment until chloroquine therapy is completed. Hence, it is possible that the treating physicians expected primaquine to be prescribed on an outpatient basis. However, review of these patients' outpatient records revealed that most patients did not return for follow-up, and of those who did, none received further treatment. This is not surprising because patients often feel dramatically better after chloroquine therapy and prefer to avoid the long waiting lines at public outpatient clinics. Kain et al⁹ have similarly noted that the failure to prescribe primaquine was the most common error in P vivax therapy. We suggest that one measure to overcome this problem is to start inpatient therapy with primaquine even if the course of chloroquine therapy has not been completed. If the patient is treated as an outpatient, primaquine should be included in the prescription, and patients should be educated on the importance of taking primaquine to avoid a relapse.

The other 5 errors in P vivax therapy involved the use an inappropriate or alternative drug regimen: 1 patient received primaquine alone, and the other 4 received quinine or quinidine. Where there is doubt regarding the infecting Plasmodium species, it is not unreasonable to begin empiric treatment for the more serious and life-threatening P falciparum.¹⁹ However, the malaria species for the 4 patients treated with quinine or quinidine were clearly reported as P vivax by the microbiology laboratory. It is more likely that these mistakes resulted from inadvertent selection of the wrong drug for the infecting species.

Overall, errors in the therapy for P falciparum infection occurred in 5 patients. Serious errors involving the selection of a wrong drug regimen for P falciparum infection occurred in only 2 patients. Both patients were treated with chloroquine despite a history of travel to areas with chloroquine-resistant P falciparum (Nigeria and Sierra Leone). Again, it is likely that this was due to the inadvertent selection of the wrong drug regimen for the infecting Plasmodium species. We postulate that these individuals suffered no serious sequelae because they were both semi-immune newly arrived immigrants in whom malaria is frequently a milder disease than it is among foreign tourists.²⁰

In their study of imported malaria in Canada, Kain et al⁹ found that 48% of the patients treated at community hospitals received inappropriate therapy compared with 8% of the patients who were treated by expert physicians in tropical diseases at the Tropical Diseases Unit of Toronto Hospital. We similarly observed that the use of an ID specialist helped to significantly reduce the risk of errors in malaria therapy.

In summary, despite widely published and available guidelines on the treatment of malaria, there were significant errors made in the treatment of malaria. This is probably owing to a lack of familiarity with the disease and the apparent complicated antimalarial drug regimens that need to be individualized based on the infecting Plasmodium species. Knowledge of the region of travel and of the interval from leaving a malaria-endemic area to the onset of symptoms provides useful clues to help predict the Plasmodium species and the possibility of drug-resistant malaria. Chloroquine-sensitive P falciparum is only seen in Haiti, the Middle East, and regions of Central America that are west and north of the Panama Canal. Mefloquine-resistant P falciparum is present along the Thai-Cambodia border and chloroquine-resistant P vivax is increasingly reported in Papua New Guinea, Irian Jaya, and Sulawesi, with scattered reports from Myanmar, India, Guyana, Colombia, and the Solomon islands. Consultation with an ID specialist or a specialist in tropical medicine should be encouraged to assist in the therapy for malaria.

Corresponding author and reprints: Gordon M. Trenholme, MD, Section of Infectious Diseases, Rush Medical College, Rush-Presbyterian-St Luke's Medical Center, 600 S Paulina St, Suite 143, Armour Academic Facility, Chicago, IL 60612 (e-mail: gordon_m_trenholme@rush.edu).

Accepted for publication November 14, 2002.

Dr Singh is a medical research fellow funded by the National Medical Research Council of Singapore.