Prevalence and effect of Plasmodium spp. and hookworm co-infection on malaria parasite density and haemoglobin level: a meta-analysis

The dual effects of co-infection of Plasmodium spp. and hookworm on malaria remain under debate. This study investigated prevalence, prevalence odds ratio (POR) of co-infection and impact of co-infection on malaria parasite density and haemoglobin levels in comparison to Plasmodium mono-infection. The protocol for this systematic review and meta-analysis is registered at PROPERO under ID: CRD42020202156. Relevant literatures were obtained from PubMed, ISI Web of Science, and Scopus on 25 December 2020. Mean difference (MD) and confidence interval (CI) of malaria parasite density and haemoglobin were compared using a random effect model. Heterogeneity was assessed using Cochrane Q and I2 statistics. Publication bias was determined by visualising funnel plot asymmetry. Of 1756 articles examined, 22,191 malaria cases across 37 studies included 6096 cases of co-infection of Plasmodium spp. and hookworm. The pooled prevalence was 20% (95% CI 15–26%, I2 99.6%, 37 studies) and was varied in terms of geographical region. Co-infection occurred by chance (OR 0.97, p 0.97, 95% CI 0.73–1.27, I2 95%, 30 studies). The mean malaria parasite density for co-infection (478 cases) was similar to Plasmodium mono-infection (920 cases) (p 0.24, MD 0.86, 95% CI − 0.58–2.29, I2 100%, 7 studies). The mean haemoglobin level for co-infection (90 cases) was similar to Plasmodium mono-infection (415 cases) (p 0.15, MD − 0.63, 95% CI − 1.49–0.23, I2 98%, 4 studies). Co-infection was common and occurred by chance but varied by geographic region. Further studies are required to investigate the mechanism of hookworm infection on malaria severity. Additionally, detection of hookworm infections among patients with malaria in endemic areas of both diseases is recommended to prevent severe malaria.

Study selection and data extraction. Two independent authors (AM, MK) selected the potentially relevant studies based on eligibility criteria. Any discrepancy between the two reviewers was resolved by discussion or request of the second author (FM) for the conclusion. Data extraction was performed by the same authors. The following data were extracted for the pilot standardised datasheet: the name of the first author, publication year, study location, year that the study was conducted, study design, characteristics and number of participants, number of malaria cases, number of hookworm cases, number of co-infection, number of single infections, malaria parasite density and haemoglobin level.
Quality of included studies. The quality of the included studies was assessed using the Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses (Supplementary Table S2). NOS provided a star system for judging the included studies based on the selected study groups, comparability of the groups and ascertainment of either exposure or outcome of interest 22 . The comparability criteria that are not applicable to most study types were defined as "not applicable" since no data from control groups were unavailable. For this study, any included study rated more than 6 out of 7 stars indicated a high-quality study. Any included study rated 4-5 stars indicated moderate quality, whereas any study rated below 3 stars indicated poor quality.
Outcomes. The outcomes of the study were (i) magnitudes of co-infection, (ii) magnitude of parasitaemia, (iii) magnitude of malaria severity (anaemia).

Results
Search results. The searches retrieved 522, 769 and 465 articles from PubMed, Scopus and ISI Web of Science, respectively. Of the 1756 studies screened, 712 were duplicates and were removed. Of the 1044 articles screened for titles and abstracts, 796 articles were excluded due to irrelevant articles. Of the 248 articles examined for full texts, 211 articles were excluded for the following reasons: 55 showed no report on co-infection, 42 were review articles, 22 showed co-infection but the data could not be extracted, 18 were in vitro, 16 were intervention studies/randomised control trials, 13 were published in local languages, 12 were not full-text, 6 were not malarial case, 5 were in vivo, 4 were co-infection with other nematodes, 4 were model prediction, 4 were case-control studies, 3 were books, 3 were studies on the same group of participants, 2 were case reports/case series, 1 was protocol and 1 was a questionnaire. Finally, 37 studies met the inclusion criteria and were included in the analysis (Fig. 1).

Characteristics of the included studies.
Characteristics of the included studies are shown in Table 1.
Quality of the included studies. Ten studies 8,9,11,13,24,28,38,40,50,53 were rated as high-quality studies, since they reported the outcomes of interest, whereas the rest of the studies were rated as moderate quality, since they www.nature.com/scientificreports/ than in Plasmodium mono-infection, whereas one study 38 demonstrated a higher mean haemoglobin level in coinfection than in Plasmodium mono-infection (Fig. 6). The pooled MD of haemoglobin between patients with co-infection (79 cases) and without any infection (645 cases) was estimated from three studies 8,9,38 . The results demonstrated no difference in the mean haemoglobin level between patients with co-infection and without any infection when a random effects model was used (p 0.062, MD − 1.4, 95% CI − 2.87 to 0.07, I 2 98.8%) (Fig. 7).
Publication bias. Publication bias was assessed using the funnel plot demonstrating the effect size (pooled OR) and selogES from 30 studies (Fig. 8). Egger's test demonstrated that no small-study effect was found (p 0.128, coefficiency 6.77, standard error 4.31). Visualisation of the funnel plot and the result of Egger's test demonstrated asymmetrical distribution of the funnel plot, and no small-study effect was found among the included studies, respectively.          51 . A possible explanation for the high rate of co-infection in these participants may be attributed to the impairment of immunity during pregnancy 63 . A study demonstrated that that pregnant women are more susceptible to Plasmodium spp. and hookworm infections in their first pregnancy, which might cause nutrient deficiency in subjects, which would lead to poor pregnancy outcomes 33 .
The co-infection of Plasmodium spp. and hookworm might occur by chance. The pooled analysis of this study suggested that underlying infection by hookworms may not increase the chance of being infected with malaria. Nevertheless, the meta-analysis for each subgroup of participants demonstrated that hookworm infections of people within communities increased the risk of malaria infection. Previous studies showed that rural communities are associated with poverty, poor sanitation and personal hygiene, and in turn, are related to  , previous studies demonstrated that co-infection was more common among boys, less common with increasing age and highest among children from poor households 29,48,68,69 . Therefore, the risk of Plasmodium spp. and hookworm might be associated with access to sanitation and clean water, recent deworming and living in urban settings 69 , as in these areas, children are exposed to open defecation grounds, which is a major source of hookworm transmission infection 70 .
The present meta-analysis demonstrated that there was no difference in malaria parasite density among patients with co-infection when compared to patients with Plasmodium spp. mono-infection. Only two studies demonstrated that a higher hookworm intensity was positively correlated with higher malaria parasite density 13,50 . Previous studies demonstrated that a higher hookworm intensity was positively correlated with a higher malaria parasite density, whereas it was negatively correlated with a lower malaria parasite density when malaria was co-infected with A. lumbricoides 12,13 . The possible association of hookworm with protection from severe malaria was that infection of helminth modulates inflammatory factors and immunoglobulin E-induced nitric oxide (NO) production and is related to reduced parasite sequestration, which protects against severe malaria 71,72 . In addition, helminth may decrease cytophilic IgG1 and IgG3 antibodies, which protect the host from malaria disease. Moreover, helminth infection can increase the non-cytophilic IgG2, IgG4 and IgM antibodies, thereby accelerating the severity of malaria 73 .
Anaemia caused by malaria and hookworm is attributable to a combination of chronic blood loss, haemolysis, and haemopoietic suppression 52 . In addition, children with asymptomatic Plasmodium infection demonstrated impaired intestinal iron absorption, which may play an important role in the development of anaemia 74 . Previous studies demonstrated that hookworm infection is more prevalent in older children than in young children, and is associated with chronic intestinal blood loss 62,75 . In areas where co-infection was low, co-infection was related to anaemia and its effect on the child's health and development 62 . The mechanisms by which Plasmodium spp. causes anaemia involve decreased red blood cell production by bone marrow suppression and increased red blood cell destruction through rupturing, phagocytosis and hypersplenism 76,77 , while hookworm infection contributes to anaemia through chronic blood loss in the intestine 78 . The present study found no association between co-infection and malaria parasite density or haemoglobin level among the included studies. A possible explanation is that the impact of co-infection on these parameters might be due to an increase in the number of intestinal helminths species than only hookworm co-infection 12,71 . In addition, when compared to other intestinal helminth infections, the mean Plasmodium density in co-infected individuals with hookworm was lower than in malarial patients co-infected with A. lumbricoides alone, S. mansoni alone or T. trichiura alone 31 . Another explanation for the difference in contradicting reports about parasite density among co-infected patients is the variation among the included studies, such as the difference in participants, study design, methodology, level of parasite endemicity and local climate. These variations impact the differences in Plasmodium spp. and hookworm interactions and influence the heterogeneity among the includes studies. Another explanation suggested by a previous study was that hookworm infection was not associated with anaemia if low infection intensities were detected in the studied population 38,79 . The non-impact of Plasmodium spp. and hookworm co-infection www.nature.com/scientificreports/ on anaemia in the present study suggested that the anaemia was most likely due to dietary deficiency. Therefore, more studies are needed to explore the impact of co-infection on anaemia. This study has several limitations. First, several important sources of databases, such as ScienceDirect and OVID, were not included in the search. Therefore, some relevant studies may have been missed from the search. Second, the source of heterogeneity across the included studies in the pooled prevalence analysis cannot be explored due to the limited data on the included studies. Therefore, a pooled analysis needed to be interpreted with caution. Third, the number of included studies that reported the mean or median haemoglobin and mean malarial parasite density was limited, which caused imprecision in the estimate for a pooled analysis of the mean haemoglobin and mean parasite density between patients with Plasmodium spp. and hookworm infection. Third, only studies published in English were included. For this reason, studies in Latin America are absent in this review, although this region has high malaria and hookworm endemicity.
In conclusion, co-infection of Plasmodium spp. and hookworm was common and it most likely occurred by chance. The meta-analysis demonstrated no difference in the malaria parasite density and haemoglobin level in patients with co-infection compared to Plasmodium monoinfection. However, these results were based on the limited number of studies included for meta-analysis. Therefore, for a more comprehensive review, further meta-analysis studies should include non-English literature or case-control studies. Additionally, further studies are needed to investigate the mechanism of hookworm infection on malaria severity. Finally, the detection of hookworm infections among patients with malaria in endemic areas of both diseases is recommended to prevent severe malaria.

Data availability
All data supporting the findings of this study are available within the article and its supplementary files.