Diagnostic Difficulties with Plasmodium knowlesi Infection in Humans

To the Editor: Studies conducted in Malaysia have raised questions about Plasmodium knowlesi as the fifth human pathogenic malaria parasite (1,2); additional cases of P. knowlesi malaria have subsequently been reported from other Asian countries (3–5). Microscopic diagnosis is hindered because P. knowlesi morphologically resembles P. falciparum or P. malariae, depending on blood stage (6). Singh et al. has designed a nested PCR assay for identification of P. knowlesi infections (1). 
 
As part of an ongoing research project focusing on characterizing genes from malaria isolates in Indonesia (E. Sulistyanisih, unpub. data), during December 2008–February 2009, blood samples from 22 gold miners with uncomplicated malaria were collected in South Kalimantan Province in Indonesia. Ring forms typical for P. falciparum were seen during microscopy. DNA was extracted and species were identified by nested PCR by using Plasmodium genus- and species-specific primers derived from the small subunit ribosomal RNA gene described elsewhere (1). PCR products were directly sequenced and verified by 2 independent amplifications of the same DNA sample. PCR using P. knowlesi–specific primers yielded a 153-bp product in samples from 4 of the 22 malaria cases. Sequencing showed perfect matching with the recently published P. knowlesi S-type from Malaysian Borneo for 1 of the 4 samples. The other sequences were repeatedly consistent with the small subunit ribosomal RNA gene of sporozoite P. vivax (S-type), and random blasting (http://blast.ncbi.nlm.nih.gov) showed higher homology (93%–100%) with various P. vivax strains than with different P. knowlesi (<84%) or other Plasmodium strains. The vivax-specific PCR showed the expected bands in each case, and sequencing confirmed P. vivax A-type DNA that matched perfectly with a strain from Thailand. Of the miners with malaria, 3 case-patients were coinfected with P. falciparum. All 22 samples from the case-patients were negative for P. malariae. One case-patient (P 15) infected with P. knowlesi (4,000 parasite ring forms/μL) had a mixed infection with P. vivax and was successfully treated with chloroquine-primaquine (Table). 
 
 
 
Table 
 
Profile of Plasmodium knowlesi–positive patients, South Kalimantan Province, Indonesia, December 2008–February 2009* 
 
 
 
The results of this study indicate the geographic distribution of natural P. knowlesi human infections includes Indonesian Borneo, although this detection is no surprise because many P. knowlesi isolates are found in Malaysian Borneo (1,2). However, the diagnosis would have been unrecognized without molecular techniques, and even those techniques posed a problem. 
 
The species-specific nested PCR assay repeatedly showed bands of 153 bp, indicating 4 P. knowlesi cases, but sequencing confirmed P. knowlesi in only 1 sample. There was no indication of contamination of the samples tested by PCR, and the other 18 samples and the negative control remained negative for P. knowlesi. All 3 samples showed molecularly confirmed mixed infections with P. falciparum and P. vivax in the case-patients. As P. vivax was only molecularly detected, low parasitemia was assumed. 
 
The reverse primer sequence (pmkr 9) is found in P. vivax S-type strains and other Plasmodium spp., especially those related to P. vivax, thus, amplification from this site should be theoretically possible. The forward primer pmk 8, on the other hand, seemed to be highly specific. 
 
One Plasmodium strain (GenBank accession no. {"type":"entrez-nucleotide","attrs":{"text":"DQ660817","term_id":"110084549","term_text":"DQ660817"}}DQ660817) found in orangutans in Kalimantan, Indonesia, and classified as P. vivax, seemed to be more likely to bind to pmk 8 (7). However, this classification was recently disproved by Singh and Divis (8), and the parasite was categorized as probably being P. pitheci or P. silvaticum, where human infections are not described. Other primate malaria parasites, such as P. hylobati, P. inui, P. cynomolgi, P. simium, P. fieldi, and P. simiovale, showed better binding sites for pmk 8 than P. vivax S- or A-strains. Regarding the theory of P. vivax originating in macaques in Southeast Asia and the close relationship to other primate malaria parasites (9), one might imagine that P. vivax strains in Indonesia differ slightly from the strains described so far. A P. vivax isolate from Indonesia, recently sequenced in cooperation with the University of Heidelberg (GenBank accession no. {"type":"entrez-nucleotide","attrs":{"text":"GU233452","term_id":"283443763","term_text":"GU233452"}}GU233452), showed 2 point mutations; the patient had been in Flores, Bali, and Lembata. However, the 3 P. vivax samples presented no mutations at the pmk 8 binding sites. Notably, pmk 8 and pmkr 9 seem always to amplify the S-type and the rVIV 1 and rVIV 2 primers, the A-type DNA, respectively. The genus-specific DNA amplified both types at random. 
 
Some colleagues have experienced similar difficulties with the primers pmk 8 and pmkr 9 in samples from Vietnam (5); 2 of 5 samples gave false positive results for P. knowlesi. Unfortunately, their report did not mention which species was actually amplified (5). 
 
Until recently, we had no satisfying explanation for the 3 assumed false-positive results. Then, in 2009, Imwong et al. reported that these P. knowlesi primers stochastically cross-react with P. vivax genomic DNA. No polymorphisms alleviating the binding of pmk8 were found; however, a new PCR for P. knowlesi was introduced (10). 
 
Given the large distribution of the vector and the natural host of P. knowlesi in Southeast Asia, it is likely that P. knowlesi will be found in other parts of Indonesia. As microscopic and molecular diagnosis of this parasite seems difficult, the underestimation of its distribution and clinical relevance can be assumed.

with the recently published P. knowlesi S-type from Malaysian Borneo for 1 of the 4 samples. The other sequences were repeatedly consistent with the small subunit RNA gene of sporozoite P. vivax (S-type), and random blasting (http://blast.ncbi.nlm.nih.gov) showed higher homology (93%-100%) with various P. vivax strains than with different P. knowlesi (<84%) or other Plasmodium strains. The vivax-specific PCR showed the expected bands in each case, and sequencing confi rmed P. vivax A-type DNA that matched perfectly with a strain from Thailand. Of the miners with malaria, 3 case-patients were coinfected with P. falciparum. All 22 samples from the case-patients were negative for P. malariae. One casepatient (P 15) infected with P. knowlesi (4,000 parasite ring forms/μL) had a mixed infection with P. vivax and was successfully treated with chloroquineprimaquine (Table).
The results of this study indicate the geographic distribution of natural P. knowlesi human infections includes Indonesian Borneo, although this detection is no surprise because many P. knowlesi isolates are found in Malaysian Borneo (1,2). However, the diagnosis would have been unrecognized without molecular techniques, and even those techniques posed a problem.
The species-specifi c nested PCR assay repeatedly showed bands of 153 bp, indicating 4 P. knowlesi cases, but sequencing confi rmed P. knowlesi in only 1 sample. There was no indication of contamination of the samples tested by PCR, and the other 18 samples and the negative control remained negative for P. knowlesi. All 3 samples showed molecularly confi rmed mixed infections with P. falciparum and P. vivax in the case-patients. As P. vivax was only molecularly detected, low parasitemia was assumed.
The reverse primer sequence (pmkr 9) is found in P. vivax S-type strains and other Plasmodium spp., especially those related to P. vivax, thus, amplifi cation from this site should be theoretically possible. The forward primer pmk 8, on the other hand, seemed to be highly specifi c.
One Plasmodium strain (Gen-Bank accession no. DQ660817) found in orangutans in Kalimantan, Indonesia, and classifi ed as P. vivax, seemed to be more likely to bind to pmk 8 (7). However, this classifi cation was recently disproved by Singh and Divis (8), and the parasite was categorized as probably being P. pitheci or P. silvaticum, where human infections are not described. Other primate malaria parasites, such as P. hylobati, P. inui, P. cynomolgi, P. simium, P. fi eldi, and P. simiovale, showed better binding sites for pmk 8 than P. vivax S-or A-strains. Regarding the theory of P. vivax originating in macaques in Southeast Asia and the close relationship to other primate malaria parasites (9), one might imagine that P. vivax strains in Indonesia differ slightly from the strains described so far. A P. vivax isolate from Indonesia, recently sequenced in cooperation with the University of Heidelberg (GenBank accession no. GU233452), showed 2 point mutations; the patient had been in Flores, Bali, and Lembata. However, the 3 P. vivax samples presented no mutations at the pmk 8 binding sites. Notably, pmk 8 and pmkr 9 seem always to amplify the S-type and the rVIV 1 and rVIV 2 primers, the A-type DNA, respectively. The genus-specifi c DNA amplifi ed both types at random. Some colleagues have experienced similar diffi culties with the primers pmk 8 and pmkr 9 in samples from Vietnam (5); 2 of 5 samples gave false positive results for P. knowlesi. Unfortunately, their report did not mention which species was actually amplifi ed (5).
Until recently, we had no satisfying explanation for the 3 assumed falsepositive results. Then, in 2009, Imwong et al. reported that these P. knowlesi primers stochastically cross-react with P. vivax genomic DNA. No polymorphisms alleviating the binding of pmk8 were found; however, a new PCR for P. knowlesi was introduced (10).
Given the large distribution of the vector and the natural host of P. knowlesi in Southeast Asia, it is likely that P. knowlesi will be found in other parts of Indonesia. As microscopic and molecular diagnosis of this parasite seems diffi cult, the underestimation of its distribution and clinical relevance can be assumed.

Toscana Virus Infection Imported from Elba into Switzerland
To the Editor: Toscana virus (TOSV) is a serotype of Sandfl y fever Naples virus (SFNV) within the family Bunyaviridae and the genus Phlebovirus. TOSV is transmitted to humans by sandfl ies (Phlebotomus spp.) and is a prominent cause of aseptic meningitis in Mediterranean countries (1). In Italy, for populations living in rural areas and persons engaging in outdoor activities, the highest risk for acquiring TOSV is from August through October (1). TOSV infections should therefore be considered in travelers returning from the Mediterranean area who have fever and signs of meningitis. Laboratory diagnosis of TOSV infections is often limited to the detection of immunoglobulin (Ig) M and IgG because of the short period of viremia and the low amount of virus in the cerebrospinal fl uid (CSF) during the acute phase (2). We report a reverse transcription-PCR (RT-PCR)-confi rmed TOSV infection acquired on the island of Elba that was then imported into Switzerland.
A 17-year-old man was referred to Basel University Medical Clinic, Liestal, Switzerland, in August 2009 with headache, recurrent episodes of vomiting, photophobia and phonophobia, and an elevated temperature of 38.1°C. The patient had returned to Switzerland from a vacation on the island of Elba, Italy, 14 days before. He recalled that he had received multiple insect bites on the beach. Cardiopulmonary and neurologic examination showed tachycardia and nuchal rigidity. Results of a complete blood count and liver and kidney function tests showed no abnormalities. CSF analysis showed lymphocytic pleocytosis (47 cells/μL), and aseptic meningitis of viral origin was suspected. Empirical treatment with acyclovir (2.3 g/day) was started for the fi rst 48 hours. The results of a