Spread of multidrug-resistant Plasmodium falciparum malaria and predictors of treatment failures in Vietnam

Background Drug-resistant falciparum malaria is an increasing public health burden. We examined the magnitude of Plasmodium falciparum infection and the patterns and predictors of treatment failure in Vietnam. Methods Medical records of all 443 patients with malaria infection admitted to the Hospital for Tropical Diseases between January 2015 and December 2018 were used to extract information on demographics, risk factors, symptoms, laboratory tests, treatment, and outcome. Results More than half (59.8%, 95%CI 55.1%-64.4%) of patient acquired P. falciparum infection of whom 21.9% (95%CI 17.1%-27.4%) had severe malaria, while 10.2% (95%CI 6.8%-14.5%) and 19.2% (95%CI 14.7%-24.5%) developed early treatment failure (ETF) and late treatment failure (LTF) respectively. ETF was 6.8% among patients referred from Binh Phuoc province and Central Highland, 11.3% from other areas in Vietnam, and 6.9% from Africa. LTF was 16.2% among patients from Binh Phuoc province and Central Highland, 22.6% from other areas in Vietnam, and 27.6% from Africa. Most patients (98.5%) recovered completely. Having severe malaria was a predictor of ETF (AOR 4.42, 95%CI 1.85-10.61, P = 0.001). No predictor of LTF was identified. using molecular technique in these regions is required to completely understand the magnitude of drug-resistant malaria and to design appropriate control strategies.

recommended in the clinical management of patients with malaria and the surveillance of anti-malarial drug resistance (12). The resistance to DHA-PPQ combination therapy causes delayed parasite clearance time (13). The susceptibility of P. falciparum to DHA-PPQ combination therapy in southern Vietnam has declined rapidly with the increase in the proportion of patients with parasite clearance time of more than 72 hours (from 38% in 2012 to 57% in 2015) (5). To  Medical records of all patients with malaria admitted to the HTD between January 2015 and December 2018 were retrieved for review. Information derived from medical records of patients with P. falciparum infection was extracted and included demographics (age, sex, job, and residential address), risk factors (blood transfusion, injecting drug use (IDU), travelling to malaria endemic areas domestically and internationally within 14 days before the onset of illness (14), current health conditions (pregnancy, end stage renal disease, cirrhosis, and HIV infection), malaria disease and treatment. It is noted that if patients have any risk factor for HIV infection, they will be consulted to undertake HIV testing in accordance with the HTD guideline. Information on malaria disease and treatment included admission time, number of days of illness at the time of admission, symptoms and signs (fever, anemia, splenomegaly, and hepatomegaly), laboratory tests (malaria microscopy, parasite counts, and aminotransaminases (AST and ALT)), antimalarial and other supportive treatments, response to treatment (number of inpatient days, fever clearance time, parasite clearance time, and early (ETF) and late treatment failure (LTF)) (15), outcomes (recovery and death), and having malaria previously. All laboratory tests including microscopy were performed at the HTD and in line with the national laboratory performance standards. According to the HTD policy, all cases were diagnosed using microscopy. Based on the course of the disease, P. falciparum infection was classified into different types of severe malaria in accordance with the WHO guideline for the treatment of malaria and included shock, acute kidney failure, impaired consciousness, jaundice, anemia, hemoglobiuria, acidosis, hyperparasitemia, prostration, convulsion, hypoglycemia and bleeding (7). ETF and LTF were defined according to the WHO recommendation (15) and the distribution of these treatment failures was also examined. Given the updated guideline that recommends increasing the DHA-PPQ dosage in treating uncomplicated P. falciparum infection has been utilized in Vietnam since September 2016 (16,17), the change in the magnitude of malaria and severe malaria before and after 2017 was also examined. The study protocol was approved by the HTD's Ethics Committee (approval number 65/QD-BVBND) and the Human Research Ethics Committee at UNSW Australia (approval number HC180340).

Statistical analysis
Data were managed and analyzed using SPSS version 22 (IBM  (Figure 1).

Predictors for early and late treatment failures
Patients with severe malaria were significantly more likely to develop ETF (P < 0.001, OR 4.3, 95% CI 1.9-9.9) and less likely to develop LTF (P = 0.044, OR 0.44, 95% CI 0.19-0.99) compared with those who did not have severe malaria (Table 4).
Gender, age, living in or travelling to malaria endemic areas 14 days prior to the onset of disease, number of days of illness > 7 days at the time of admission, acquiring malaria before 2017 and hyperparasitemia (P > 0.05) were not a predictor of ETF and LTF. Similarly, having parasitemia 72 hours of treatment and fever 3 days of treatment (P > 0.05) were not predictors of LTF.

Model for the prediction of early and late treatment failures
No predictor for ETF was identified other than having severe malaria (AOR 4.42, 95% CI 1.85-10.61, P = 0.001) (  (18) which is probably due to the effectiveness of ACTs (19). We found that most patients were male in labor age and more than half (56.3%) of them had jobs related to the forest or worked in areas which are known as malaria endemic areas in Vietnam, Cambodia and Africa (20,21). Additionally, more than 80% of patients lived in or traveled to malaria endemic areas within 14 days before the onset of disease which is an established risk for malaria infection (14). These findings were not surprised since according to Vietnamese culture, breadwinner is perceived as the man's role in the family, and therefore more men tend to be exposed to malaria infection than women when they work in or travel to malaria endemic areas (22). At this stage, most malaria morbidities and mortalities occur in 21 out of 58 provinces, of which forested areas of provinces located in Central and Central-Southern Vietnam account for the highest malaria burden in Vietnam (23). In our study, P. falciparum infection was recorded throughout the year and peaked during the months from November to March of the following year. In Vietnam, Lunar New Year holiday which is the most important holiday usually occurs in the second half of this period. Based on our experience, to financially prepare for the holiday many people travel to malaria endemic areas to work in this period in response to the increase in the number of seasonal jobs in these areas.
Blood transfusion and IDU are considered as malaria risk factors (24, 25). We found only one patient who was suspected to have malaria infection from blood transfusion. Two other cases were injecting drug users who may have shared needles and syringes in the days before their illness developed. Pregnancy is the high risk for developing severe malaria (26). Fortunately, two pregnant patients in our study had uncomplicated malaria. HIV infection has been reported as an underlying disease that may delays parasite clearance time (27). We did not detect any HIV infection in our study.
According to the WHO definition of severe falciparum malaria (28) and the Vietnam Ministry of Health guidelines for the management of malaria infection (16), patients were classified into uncomplicated and severe malaria groups. In our study, severe malaria accounted for 21.9% (95%CI 17.1% -27.4%) which is higher than the rate of 9.1% (277/3053, 95%CI 8.1% -10.1%) reported in 1990 in Vietnam (18). Although our rate is comparable to the rate in Africa which is estimated to be between 10% and 70% based on a simulation-based study (29), our rate is higher than the rate of 6.6% (43/650, 95%CI 4.8% -8.8%) from a study conducted in low and unstable malaria transmission settings of Colombia (30). This is probably because our hospital is a leading referral hospital that receives more severe cases. Our rate is also higher compared with the results of a study conducted on international travelers who were mostly exposed in sub-Saharan Africa (7.8%, 444/5689, 95%CI 7.1% -8.5%) (31) because patients infected with all 4 Plasmodium species were recorded in this study.
Regarding the clinical symptoms, similar to previous reports (30, 32), fever was frequent (99.6%), while anemia was rare (0.8%) in our study. However, splenomegaly was reported to be a common symptom among P. falciparum infected patients, while hepatomegaly was not mentioned in another study (32). We only detected splenomegaly in only 3% and hepatomegaly in up to 10.2% of study participants. Among 58 cases with severe malaria, the most common manifestations were impaired consciousness (20.7%), jaundice (24.1%) and mixed manifestations (22.4%). Other manifestations of severe malaria included anemia, renal failure, hyperparasitemia and shock. Our findings are in line with the other reports (18,30).
Mild elevations of transaminases are common in Plamodium infection, particularly in P. falciparum infection because of hemolysis that affects the liver function (33).
About 30% of patients in our study had transaminase higher than 40 U/L. Treatment delay of seven days or more is documented to be a risk for severe malaria with multi-organ failure (34). We noted that 25% of study population admitted to our hospital after seven days of illness and this included some patients receiving treatment from previous hospitals. In our study, more than one fifth of study participants receiving treatment from previous hospitals had been misdiagnosed as having dengue infection or septicemia. An inappropriate therapy provided in previous hospitals could aggravate patient's condition (35). This could explain why some patients admitted to the HTD late, after 7 days of disease and developed severe malaria. Therefore, clinicians must be more alerted to malaria infection when patients have a travel history to endemic areas (35). In addition, severe malaria should be suspected if patients have been under inappropriate treatment (36).
DHA-PPQ combination has been widely used as a first-choice therapy for P. falciparum infection in many countries including Vietnam (44). However, the spread of artemisinin and partner drug resistance have caused high treatment failure rates to this combination (45) and subsequently threatens the success of malaria control and elimination (9). In Vietnam, before 2015 DHA-PPQ treatment dose depended on patient's age (46). However, many studies have showed that PPQ under-dosing (<48mg/kg) is an important factor for recrudescent parasitemia (44,47). In order to optimize the effectiveness of DHA-PPQ, the WHO recommends treatment with 3 days of ATCs to cover at least two asexual life circles of P. falciparum and DHA-PPQ weight-based dosing (7). In September 2016 the Vietnam Ministry of Health adopted the WHO guideline which has been widely used since 2017 (16). It is documented that the increased dose of DHA-PPQ reduces recrudescent parasitemia and sequentially decreases treatment failure (44). However, we did not find any difference in the proportion of ETF and LFT before and after 2017 in our study. The A study on African children with uncomplicated falciparum malaria demonstrated the association between age less than 2 years and delayed parasite clearance (48).
In addition, hyperparasitemia (>50.000/ul) was also reported as a predictor of delayed parasite clearance (48). We did not find any association between both two treatment failure types and age and hyperparasitemia. However, we found that having severe malaria was a significant predictor of ETF (P < 0.001, OR 4.3, 95% CI 1.9-9.9) which has not been reported elsewhere. Similar to other countries, in Vietnam, treatment of severe malaria is at least 24 hours of parenteral artesunate single-therapy plus a three-day oral DHA-PPQ combination therapy when the patient can tolerate oral therapy (16). We wonder if the variation in time to receive DHA-PPQ combination therapy and the above-mentioned DHA-PPQ resistance may facilitate ETF among patients with severe malaria. In light of this, early initiation of combination therapy including parenteral artesunate and an oral partner drug for severe malaria may reduce the risk for ETF. It is important to re-evaluate the effectiveness of the current WHO recommended antimalarial therapy for both uncomplicated and severe malaria as well as to develop new intramuscular or parenteral antimalarial drugs in the context of antimalarial drug resistance in the GMS.
Our study had some limitations. First, this is a single center study, and thus we may have missed patients with falciparum infection receiving treatment at other hospitals but were not referred to the HTD. However, the HTD is the only major tertiary teaching hospital for infectious diseases including malaria in southern Vietnam and receives not only local patients, but also patients from the other countries. This would enhance the generalizability of the study findings. Our study interval included the period when the WHO's new treatment guideline has been adopted. This allowed us to examine the change in the pattern of treatment failure in response to the utilization of the new guideline Second, we were unable to perform molecular tests to further examine the magnitude of parasites carrying mutations due to the nature of a retrospective study. Third, we were unable to examine the presence of parasitemia on day 28 for all patients due to the same reason. We may have missed some cases who had parasitemia on day 28 but did not exhibit any clinical symptom. Therefore, the burden of LTF that we have identified may be underestimated.    (14) 19.2 (5) 27.6 (8)   Figure 1 Frequency distribution of P. falciparum infected cases between January 2015 and December