High Prevalence of viral and bacterial coinfections in malaria in Venezuela

Malaria remains a signicant public health problem worldwide. Simultaneous infections with other pathogens complicate its diagnosis and can also change the clinical course of the disease. The similarities in the clinical presentation of malaria and other infections and the superimposed endemicity result in underdiagnosis of coinfections and increase mortality. No studies have focused on the presence of coinfections in patients with malaria in Venezuela. study

cases are clustered in the Sifontes municipality, an area where gold mining is the main economic activity [3,13].

Study design and participants
We conducted a cross-sectional study that included patients with malaria, con rmed by microscopic examination of thin and thick blood smears (TBS) and who presented to type II outpatient centers: "El Perú", "La Sabanita" and the "Complejo Hospitalario Universitario Ruiz y Páez", during the period June to November 2018. A trained clinician performed the standard clinical evaluation and a detailed physical examination of all study participants. All patients were treated by the local health provider as soon as the infections were con rmed, using the national antimalarial therapeutic protocol approved by the health authorities [14].

Coinfection evaluation
Two blood samples were collected by venipuncture from each study participant: 3 mL in a tube with EDTA, which was used for hematology evaluation, and 2 mL in a tube without anticoagulant, which was used for the analysis of blood chemistry (urea, creatinine, glycemia, electrolytes, transaminases, lactate dehydrogenase). Coinfections were diagnosed by the presence of speci c antibodies detected in serum by enzyme-linked immunosorbent assay (ELISA). We used, Dengue (IgG/IgM, B.Q. Kits, Inc. USA), Chikungunya (IgM/IgG, Abcam USA), Hepatitis A (IgG/IgM, Abcam USA), Hepatitis B (Surface Ag/Anti-Core, Abcam USA), Hepatitis C (IgG/IgM, Abcam USA), and Leptospirosis (IgG/IgM Serion ELISA, USA) following the manufactures' guidelines. These serological tests were selected based on sensitivity and speci city, both higher than 92%.

Data analysis and interpretation
The data were processed using IBM SPSS Statistics for Windows v. 25 The distribution of the parameters was statistically evaluated using Kolmogorov-Smirnov test and comparison tests were applied as required (Chi 2 Pearson, Chi 2 with Yates correction, Fisher's exact test and t-Student) The Odds Ratio (OR) for complications was determined according to the presence of coinfection with a 95% CI. A p-value < 0.05 was considered statistically signi cant.

Laboratory ndings
Hemoglobin levels were similar between coinfected and non-coinfected patients. The patients coinfected with CHKV had lower leucocyte counts (p= 0.010).

Discussion
Several studies, mainly from sub-Saharan Africa and Southeast Asia, report malaria coinfections with other pathogens such as DENV [16], CHIKV [17,18], HAV [9], HBV [10,19], LEP [20,21], HIV [22], helminths [23] and other febrile illnesses [24]. In Latin America, however, reports of coinfections in malaria patients are limited [10,25,26,27,28]. To the best of our knowledge, there are no reports about the interactions of these pathogens in coinfections in Venezuela, despite multiple infections may complicate malaria and lead to failure when it comes to treatment responsiveness. High prevalence of malaria coinfection was found in this study (34.2%), even higher than reported in Brazil (20%) [25]; but lower than that found in a recent study in India (60%) [29]. Thus, physicians should be suspicious of coinfection in malaria cases with inadequate treatment response or atypical manifestations.
The prevalence of malaria coinfection with DENV (24/161; 14.9%) was much higher than the found in a cross-sectional study in hospitalized patients with the acute febrile syndrome in the Brazilian Amazon (44/1578; 2.8%) [25] or in another study in Mumbai (16/156; 10.25%) [16], or during a dengue outbreak in India (27/367; 7.4%) [30]. In contrast, in Pakistan, the prevalence found was higher (26/78; 33.3%) [31], as well as in India (29/66; 44%), [29]. Thus, the prevalence of coinfection may uctuate, depending on local endemicity and the sensitivity of the diagnostic methods used. In these studies, the prevalence was estimated based on hospitalized and non-hospitalized patients; therefore, it could not be extrapolated to the community level. We found that DENV coinfection was signi cantly associated with somnolence and splenomegaly. In agreement with our ndings, a study in French Guiana showed worse clinical outcome, with a higher risk of severe thrombocytopenia and anemia in DENV coinfection than in patients with only malaria [28]. Other studies have reported a markedly low platelet count [31] or high elevation of transaminases in the DENV coinfection group [16], however, we did not nd these paraclinical alterations in our study. A study in Peru indicated Plasmodium/DENV coinfection was not associated with worse disease [26], similar to another study in India were the coinfection with DENV serotype 4 (DENV-4), even was associated with mild malaria. [29]. Differences in DENV serotypes or Plasmodium spp may explain the differences of the results.
The second most frequent coinfection was HAV (19/161, 11.8%), higher than that found by Klein et al. (10/222, 1.7%) in children from sub-Saharan Africa [9]. This high incidence could be due to a deteriorated water system in Venezuela [32] added to low vaccination rates [33]. In this study, neurological manifestations (stupor and seizures) were associated with HAV coinfection. In contrast, no signi cant alterations were found in the clinic and in the liver function of the coinfected patients, as in the study carried out in sub-Saharan Africa [9]. The age group studied could explain this; however, limited information is available on this coinfection. On the other hand, the prevalence of coinfection with HBV (10/161, 6.2%) in our study was similar to that found in Nigeria (11/166, 6.6%) [19] but higher than documented by Braga et al. [27] in western Brazilian Amazon (4.2%). In the same study, patients with coinfection presented no clinical differences from those with malaria only, and similar to our ndings, nor showed any association with classic signs of a hepatic disorder. In another study, HBV coinfection was more likely to be asymptomatic (OR: 120.13, p<0.0001), even Plasmodium parasitemia inversely correlated with plasma HBV DNA levels (R = −0.6; p = 0.0003) [10]. In contrast, other studies revealed that coinfection amongst individuals signi cantly affected the hematological and liver parameters [34,35]. Our result should be interpreted with caution due to the limited number of coinfected patients evaluated. We found no malaria/HCV coinfection cases, although coinfection is possible [36], our nding may be explained by the low prevalence of HCV previously reported in Venezuela [37].
CHKV coinfection was found in 9/161 patients (5.5%), a prevalence lower to that found in Tanzania (8/112, 7.14%) [17], and reported in Kenyan Children (15/158, 9.4%) [38]. In contrast, two extensive studies in India [39] and Senegal [40] found low coinfection prevalence (15/1564, 1.3%) and (3/13845, 0.02%), respectively. The observed variations in the prevalence of CHKV between different studies may be attributed to study site location including seasonal variations, targeted age groups, agricultural activities, and time [41][42][43]. Our ndings on LEP coinfection (6/161, 3.7%) contrasts with those reported in South India (48/222, 22%) [20] and Thailand (15/193, 7.7%) [44], incidence of LEP in these regions could explain these differences. Although a previous study of Leptospira found a high prevalence (80.6 %) of Leptospirosis in Bolivar city, this was documented in a group of febrile patients highly suspected of leptospirosis [45]. Other LEP coinfection cases have been documented [21] and have even been associated with severe sepsis [46]. We found an association between LEP coinfection with aminotransferases elevation and thrombocytopenia previously described [47,48]. In Latin America, LEP/malaria coinfections are rarely reported, but high clinical suspicion must prevail since a late diagnosis could increase morbidity and mortality. Thus, for patients with severe malaria presenting with fever, thrombocytopenia, and altered liver and kidney function diagnosis [49,50] and empirical treatment for this coinfection should be considered.
Interestingly, simultaneous coinfection with DENV/HAV was found in 4/11 (36.4%) patients with malaria. To date, there are few case reports of this concurrent mixed infection [51]. Thus, we consider they likely occur more frequently than reported in the available literature mainly in developing countries. Other coinfections with two or three pathogens could be explained by overlapping breeding sites for mosquito vector species, especially in malaria, DENV, and CHKV [52,53]. Additionally, outbreaks of febrile illnesses are often associated with rainy seasons in the tropics [8].
A high frequency (42.9%) of complicated disease was found and complications were more likely in coinfected patients compared to patients without coinfections, suggesting that coinfection with another pathogen could exacerbate the clinical course of malaria. Nevertheless, due to the small sample size, further investigation is needed to con rm this observation. Similar results have been found for patients with P. vivax and DENV coinfection who had a higher chance of presenting severe disease than those mono-infected with dengue [25]. In contrast, Andrade et al found (among 636 Brazilian patients) that HBV infection was associated with a decreased intensity of malaria infection among individuals in the study [10]. In order to determine an appropriate correlation between coinfections, prospective studies should be designed that include a larger number of patients, however controlling real life variables remains a challenge in Venezuela. This study has some limitations. Although we evaluated both speci c IgG and IgM antibodies of possible coinfections, cross-reactivity cannot be ruled out due to chemical similarities of the antibodies investigated as a consequence of the polyclonal activation induced by Plasmodium spp infection [54,55], just as it happens with other highly prevalent infectious including that caused by Epstein Barr virus [56]. Another limitation was the absence of comparison between acute and convalescent sera from the same patient, and the inability to perform molecular tests to evaluate the coinfection. Nonetheless, this a frequent reallife situation regarding resources and poor settings where tests for follow-up of recovered patients are usually not collected and molecular diagnostic studies are limited. Another limitation is that all the enrolled individuals were febrile patients, then studies enrolling asymptomatic individuals should be performed in the future to evaluate the real burden of coinfections in malaria. Finally, the small number of coinfected patients along with the even smaller frequencies for some confections and a number of highly prevalent diseases that were not explored (Chagas disease, Tuberculosis, leishmaniasis, HIV, Syphilis), also represented a limitation.

Conclusions
To the best of our knowledge, this is the rst malaria coinfection study in Venezuela. The high prevalence of coinfections found in the main Venezuelan endemic state should contribute to the understanding of the clinical and paraclinical behavior, in order to subsequently develop guidelines and protocols aimed to optimize early diagnosis and target treatments in patients with acute febrile illness. Delay in either diagnosis or start of therapy for any of these infections could have fatal outcomes. Our results should be interpreted with caution due to the limited number of coinfected patients and to the possibility of cross-reactivity due to polyclonal activation induced by malaria infection. Prospective studies with samples at different points of infection and the use of molecular tools are needed to clarify these ndings. The study protocol was approved by the Ethics Committee "Complejo Hospitalario Universitario Ruiz y Páez" (CHRRP-CBBS-001-2018). The written informed consents were administered to adult patients who met the recruitment criteria and were procuring treatment. Patients agreed on allowing genetic studies on the parasite.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare they have no competing interests.