- Email: [email protected]
Paradoxical associations between soil-transmitted helminths and Plasmodium falciparum infection
Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
Contents lists available at SciVerse ScienceDirect
Transactions of the Royal Society of Tropical Medicine and Hygiene journal homepage: http://www.elsevier.com/locate/trstmh
Paradoxical associations between soil-transmitted helminths and Plasmodium falciparum infection ˜ a,∗ , Alvaro J. Idrovo b , Zulma M. Cucunubá c , Julián A. Fernández-Nino a Patricia Reyes-Harker , Ángela P. Guerra c , Ligia I. Moncada a , Myriam C. López a , Sandra M. Barrera c , Liliana J. Cortés c , Mario Olivera a , Rubén S. Nicholls a,c a Grupo de Infecciones y Salud en el Trópico, Unidad de Parasitología, Departamento de Salud Pública, Facultad de Medicina, Ciudad Universitaria, Universidad Nacional de Colombia, Bogotá D.C, Colombia b Centro de Investigación en Sistemas de Salud, Instituto Nacional de Salud Pública de México, Cuernavaca, México c Grupo de Bioquímica, Grupo de Parasitología, Instituto Nacional de Salud, Bogotá, Colombia
a r t i c l e
i n f o
Article history: Received 7 December 2011 Received in revised form 19 July 2012 Accepted 19 July 2012 Available online 11 August 2012 Keywords: Helminths Hookworm Malaria Ascaris lumbricoides Epidemiology Colombia
a b s t r a c t Evidence on the comorbidity between soil-transmitted helminth infections and malaria is scarce and divergent. This study explored the interactions between soil-transmitted helminth infections and uncomplicated falciparum malaria in an endemic area of Colombia. A paired case-control study matched by sex, age and location in Tierralta, Cordoba, was done between January and September 2010. The incident cases were 68 patients with falciparum malaria and 178 asymptomatic controls. A questionnaire was used to gather information on sociodemographic variables. Additionally physical examinations were carried out, stool samples were analysed for intestinal parasites and blood samples for Ig E concentrations. We found associations between infection with hookworm (OR: 4.21; 95% CI: 1.68–11.31) and Ascaris lumbricoides (OR 0.43; 95% CI: 0.18–1.04) and the occurrence of falciparum malaria. The effects of soil-transmitted helminths on the occurrence of malaria were found to be paradoxical. While hookworm is a risk factor, A. lumbricoides has a protective effect. The findings suggest that, in addition to the comorbidity, the presence of common determinants of soil-transmitted helminth infections and malaria could also exist. While the biological mechanisms involved are not clear, public health policies aimed at the control of their common social and environmental determinants are suggested. © 2012 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.
1. Introduction Malaria is considered to be the world’s most important tropical infection; the disease affects 243 million people annually, causing 900 000 deaths each year, and contributing approximately 30% of the burden of disease attributable to tropical diseases.1 Soil-transmitted helminth (STH) infections are chronic intestinal infections ∗ Corresponding author. Tel.: +57 1 316 5000x15032; fax: +57 1 316 5000x15033. ˜ E-mail address: [email protected] (J.A. Fernández-Nino).
that are widely distributed globally and considered to be among the neglected tropical diseases.2,3 Malaria and STH infection tend to occur in the same regions and affect the same individuals; they share the same risk factors (overlapping phenomena) and their effects on morbidity, especially among pregnant women and children, are well known.4,5 In the late 1970s the existence of a direct cause and effect relationship between STH infections and the occurrence of malaria was suggested.6 It was then proposed that STH infections could be a risk factor or a protecting factor for the development of malaria. These hypotheses have been explored many times in the
0035-9203/$ – see front matter © 2012 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.trstmh.2012.07.012
702
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
past 10 years,7 being supported by contradictory immunological models with some experimental evidence in animal models; this is primarily based on the potential shift to a Th2-type immune response as a result of STH infection, which would reduce the Th1-type response against malaria.8 The results of studies on the association between STH infections and the incidence of malaria have not been consistent. Four analytical studies, two cohort studies, two cross-sectional studies, one ecological study and a randomised controlled clinical trial found STH infections, or infection with a particular STH species, to be a risk factor for malaria.9–14 In contrast, an ecological study6 and two randomised clinical trials found a protective relationship.15,16 A recent cross-sectional study in pregnant women found that while Ascaris lumbricoides had a protective association, hookworm was positively associated with the occurrence of malaria.17 A case-control study reported no association.18 Several authors have drawn attention to the methodological limitations of these investigations.7,19 These limitations include temporal ambiguity, lack of information on specific explored parasites and the potential role of confounding. The importance of considering the places where the participants in these studies live (location bias) has also been emphasised, because the observed associations could be explained through the geographical distribution of parasites.12,20 The present study was explicitly designed to assess the association between STH infections and the occurrence of uncomplicated falciparum malaria, and to explore the possible interactions between parasites by controlling for potentially confounding variables and for location bias. 2. Materials and methods 2.1. Design and study site A case-control study matched by sex, age (range: 5–15, 16–45 and 46–60 years) and neighbourhood of usual residence was conducted between January and September 2010 in Tierralta (Cordoba, Colombia). Tierralta is the municipality with the highest number of reported cases of malaria in Cordoba, which in turn is one of the three departments with the highest incidence in Colombia. In 2010, 6204 malaria cases were reported for Tierralta, of which 25.53% were caused by Plasmodium falciparum. The people of Tierralta make their living from agriculture and rearing livestock; serious problems in accessing healthcare services are further complicated by the presence and substantial influence of illegal armed groups and drug traffickers. 2.2. Recruitment and selection Each participant included in the study met the following criteria: had lived for at least 1 year in a rural area of Tierralta; had not received anthelmintic drugs within the previous 3 months; was aged between 5 and 60 years old; was not pregnant. Cases were defined as patients with
a diagnosis of uncomplicated P. falciparum malaria confirmed by a thick blood smear, detected as incident cases during the study period at the centre for malaria diagnosis in the urban area of the municipality. All the included cases lived in one of 11 of the 19 rural areas of the municipality. Eight rural areas were excluded because of problems of access and concern for the security of the fieldwork team. Controls were defined as individuals who lived within 500 m of the residence of the case with whom they were paired. Controls were required not to have had a malaria infection within the previous 3 months, nor to have reported having fever in the previous 15 days. All controls were subjected to a blood smear to rule out asymptomatic malaria infection. The selection and recruitment of controls required visiting the house of each participant during the week of his or her inclusion in the study. While at a participant’s house we identified neighbours who could qualify as controls, and in order to recruit controls walked in a clockwise direction from the participant’s house. One to three controls per case were recruited, depending on the possible candidates’ availability and willingness to participate. To control for a possible location bias, controls were required not to have left their neighbourhood for more than 3 consecutive days in the previous 3 weeks. This condition was intended to increase the likelihood that each case would have controls who may have acquired malaria in the same place of recruitment,20 and who therefore shared the same ecological transmission risk. 2.3. Data collection Participants responded to a socioeconomic survey and their place of residence was observed to measure relevant covariates that could act as confounding variables, on the basis of a recent systematic review.7 The questionnaires were administered and the observations made by a physician trained in data collection instruments. The instruments used were previously validated with 70 individuals of the region in a pilot study with a good reproducibility (kappa: 0.82; 95% CI: 0.45–0.98). On the day after their recruitment, at their place of residence, all participants underwent a general physical examination (height, weight, blood pressure, heart rate, respiratory rate, temperature, cardiovascular and respiratory auscultation, abdominal palpation) and were asked to provide a stool sample. Stool samples were preserved in 10% formalin and shipped by air for analysis at the parasitological laboratory of the National University of Colombia School of Medicine in Bogota. All the faecal samples underwent direct examination and the Ritchie–Frick modified concentration technique for diagnosis of STH and protozoa.21 The samples were analysed independently by two experienced observers who were unaware of the status as cases or controls of the study subjects. Disagreements were discussed between both observers; the interobserver final concordance was very good (kappa 0.86; 95% CI: 0.63–0.91). The thick blood smears from cases and controls were read out at the study site by a certified microscopy
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
technician who was retrained by the research team before beginning the study. Additionally all slides were subsequently subject to quality control by an expert at the parasitological laboratory of the Colombian National Institute of Health (NIH) in Bogota. This expert knew neither data about the patient nor his or her case or control status. When there was a disagreement (<2%), the slides were reviewed by another expert observer. At the time of recruitment, blood samples were drawn from all the participants; the sera obtained were stored at −20 ◦ C in the biochemistry laboratory at the NIH in Bogota for measurement of Ig E. To measure Ig E levels, we used the immunoglobulin kit known as AccuBind ELISA (Monobind Inc., Lake Forest, CA, USA), which allows the quantitative determination of total Ig E in serum. Samples were processed strictly according to the manufacturer’s recommendations. Calibration curves were prepared using known standards of concentration. Samples were processed in batches and each trial had its own calibration curve and internal controls. Reference values for Ig E were reported taking into account the age of patients as follows: age 3–16 years, up to 280 IU/ml; age 17 years or over, range 0–200 IU/ml. 2.4. Statistical analysis The main demographic and socioeconomic characteristics for both cases and controls were initially described. Then associations between malaria and geohelminths and other variables were explored using simple conditional logistic regression models. We examined frequency of infection by intestinal parasites for the parasites present in the participants. For all the infections found, an estimate of their agreement per parasite pair was made and Cohen’s Kappa index was calculated. Variables showing statistically significant associations (p<0.20) were included for the multivariable analysis. In addition, an exploration was made for interactions between the three STH species, A. lumbricoides, Trichuris trichiura (whipworm) and hookworm. We evaluated several models, including one or two STH species as independent variables at the same time, until a parsimonious model with A. lumbricoides and hookworm was observed, which was adjusted for sociodemographic variables. Analyses were performed using the statistical program Stata 11 (Stata Corporation, College Station, TX, USA). 2.5. Ethical considerations Participants were informed about the study objectives and were guaranteed confidentiality of the results and the data collected. All participants gave informed consent to taking part in the study. When STH or other intestinal parasites were diagnosed, free treatment was provided according to the currently recommended and available treatment schemes in Colombia.22 All patients with malaria, symptomatic or asymptomatic, were treated immediately in accordance with existing national guidelines23 and clinically monitored by staff of the research team.
703
3. Results 3.1. General characteristics of participants Initially 74 cases and 182 controls were recruited; seven cases and four controls were later excluded from the study. The five cases were excluded because it was not possible to find their respective controls (no neighbours meeting inclusion criteria or neighbours not willing to give informed consent) and two cases because during quality control their thick blood smears were found to be positive for either P. vivax malaria or mixed infection. The four controls were excluded because quality control showed that they were actually positive for P. vivax. Finally 67 cases and 178 controls were included for further analyses. Of these, three individuals were initially recruited as controls but found to be positive for P. falciparum in new thick blood smears within the first week after recruitment. Therefore they were included in the study twice, both as cases and as controls,24 and controls were sought for them. The socioeconomic characteristics and practices relevant for malaria and their crude associations with P. falciparum infection are presented in Table 1. Of the total of 245 subjects, 143 (58.37%) were men, 103 (57.86%) of the control group and 40 (59.79%) of the cases. The mean age of the controls was 31 years (minimum 5 and maximum 60 years) and that of the cases 27 years (minimum 6 and maximum 59 years); 35 participants (14.28%) were under 15 years of age. Table 1 shows the crude associations between P. falciparum and STH infections. The presence of hookworm was associated with malaria, while T. trichiura and A. lumbricoides infections were not. The frequencies of other parasites in both cases and controls are also presented in Table 1, but no statistically significant differences were found between the groups. No significant differences were found between cases and controls in most socioeconomic characteristics explored, such as type of social security affiliation, occupation, persons per room or weekly family income (p>0.25). Of these conditions only house wall material (wood or other plant material vs block or brick) showed a statistically significant association with the occurrence of falciparum malaria. The time since last deworming did not show statistically significant differences between cases and controls, demonstrating the lack of widespread access to some anthelmintics in the study population, 64.08% of whom did not receive anthelmintic drugs during the previous year. For the analysis of body mass index (BMI), this variable was categorised for children as high (higher than percentile 97), normal (percentiles 3–97) and low (below percentile 3), using international reference curves;25 for adults, we used the standard classification proposed by the WHO, re-classifying the classic categories into three: low (BMI<18), normal (BMI≥18 to <25) and high (including overweight and obesity, BMI≥25), and therefore were able to analyse jointly this variable in children and adults. The results showed that most of the population (212, 86.53%) had normal BMI values. A low BMI was associated with the occurrence of falciparum malaria (p<0.05). Agricultural
704
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
Table 1 Bivariate association between soil-transmitted helminths, other relevant variables and Plasmodium falciparum infection Controls Variable
Cases %
n
%
OR
56 37 31
31.64 20.90 17.42
17 18 24
25.37 26.87 35.82
0.65 1.39 2.95
Age group <15 years 15-45 years >45 years
35 117 26
19.66 65.73 14.61
10 42 15
14.92 62.69 22.39
Social security regimen Subsidised affiliated Social security payer State subsidiseda None
128 2 17 31
71.91 1.12 9.55 17.41
55 2 3 7
82.09 2.99 4.48 10.45
1 2.59 0.31 0.44
0.21–31.31 0.83–1.13 0.17–1.13
52 55 36 35
29.91 30.90 20.22 19.66
15 20 18 15
22.06 29.41 26.47 22.06
1 4.03 5.56 1.38
0.81–20.05 1.40–22.11 0.54–3.53
Malaria episodes in previous year Zero One Two Three or more
139 34 2 3
78.09 19.10 1.12 1.69
44 1 14 8
65.67 1.49 20.90 11.95
1 0.93 15.53 7.38
0.12–0.72 3.28–73.53 1.18–45.97
Last deworming 3–6 months ago 6 months–1 year ago >1 year ago
35 23 120
19.66 12.92 67.42
10 6 51
14.93 8.96 76.12
1 1.12 1.85
Body mass indexb High Normal Low
15 157 6
8.43 88.20 3.37
4 55 8
5.97 82.02 11.94
0.69 1 3.37
House wall material Brick or block Crude wood/vegetation No walls
56 119 3
31.46 65.89 1.69
6 56 5
8.96 83.54 7.46
1 7.23 37.81
87 91
48.88 51.12
34 33
50.75 49.25
1 0.99
0.47–2.11
143
80.34
54
80.60
1.10
0.52–2.32
47 37 94
26.4 20.79 52.81
12 20 35
17.91 29.85 52.24
1 2.53 1.40
1.10–5.86 0.66–2.98
34 47 134
19.10 26.40 75.28
5 16 52
7.46 23.88 77.61
0.33 0.75 0.99
0.12–0.95 0.36–1.54 0.49–2.00
12 33 61 55 15 11
6.74 18.54 34.27 30.90 8.43 6.18
4 12 19 17 2 3
5.97 17.91 28.36 25.37 2.99 4.48
0.99 1.03 0.87 0.77 0.34 0.73
0.31–3.17 0.50–2.10 0.46–1.66 0.40–1.47 0.74–1.55 0.20–2.65
160
89.89
61
91.04
0.98
0.35–2.75
Geohelminths Ascaris lumbricoides Trichuris trichiura Hookworm
Occupation Farmer Housework Student Other
Place of defaecation Outdoors Toilet Literacy Bednet use Net not available Net available but not in use Net used House spraying the previous month Use of protective clothing Agriculture practice Other parasites Giardia duodenalis Entamoeba histolytica/dispar Entamoeba coli Entamoeba nana Iodoameba sp. Blastocystis hominis Ig Eb High
Weekly income by resident (US$) People per room Age (continuous)
n
Median
Median
2.60 3.46 31.72
2.80 3.42 29.45
95% CI 0.31–1.33 0.67–2.85 1.429–6.16
0.97 0.99 0.96
All estimations were made with simple conditional logistic models. a In Colombia, people not yet affiliated to the new health system but who have been identified and receive health care provided by the government. b For reference values, see text (Results).
0.34–3.72 0.74–4.76
0.19–2.50 1.20–11.86
2.44–21.45 5.24–272.97
0.82–1.23 0.84–1.16 0.92–0.99
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
705
Table 2 Proportion of agreement in infection status between parasite pairs (%) observed in the whole study groupa
Trichuris trichiura Hookworm Giardia duodenalis Entamoeba histolytica/dispar Entamoeba coli Endolimax nana Iodoamoeba sp. Blastocystis hominis
Ascaris lumbricoides
Trichuris trichiura
72.24b 67.35b 66.94 68.98 65.31 57.14 67.35 68.57
73.06b 75.92b 67.35 58.78 58.78 72.24 75.92
Hookworm
72.65 71.43b 62.86 57.96 71.43 74.29
Giardia duodenalis
Entamoeba histolytica/dispar
Entamoeba coli
Endolimax nana
76.73 63.27 69.80 86.53 89.39
71.84b 66.94b 81.22b 78.37
66.53b 71.02b 67.35
71.84b 70.61b
Iodameba sp.
90.61b
a
This table reports the agreement in infection status between parasite pairs (%). The values were evaluated with the p-value of the respective Cohen’s Kappa to determine agreement regardless of random occurrence. b p value <0.05, Cohen’s Kappa.
labour, literacy, use of mosquito nets and wearing protective clothing during work were not significantly associated with the occurrence of falciparum malaria (p>0.25). In contrast, spraying of the house with insecticides by the municipal government in the past month was a protective factor for the outcome (p<0.05). Of the samples analysed, 221 (90.20%) had high Ig E levels, with no significant differences between cases and controls (89.89% vs 91.04%, p=0.97). The analysis of 245 faecal samples tested showed that 73 (29.80%) were positive for A. lumbricoides, 55 (22.44%) for T. trichiura and 55 (22.45%) for hookworm. Exclusive coinfections between pairs of these STH were present with frequencies of 4.08% for Ascaris and hookworm, 6.53% for Ascaris and Trichuris, and 3.30% for Trichuris and hookworm. Finally, 5.71% had simultaneous infection with the three soil-transmitted helminth species. Table 2 shows the percentage of agreement in infection status between parasite pairs. Concurrent infections were found (p<0.05) between some of the soil-transmitted helminths and protozoa, such as between T. trichiura and Giardia duodenalis, and between hookworm and the Entamoeba histolytica/E. dispar complex. When exploring the interactions between soiltransmitted helminths and the occurrence of falciparum malaria, only those between hookworm and Ascaris and hookworm and Trichuris were statistically significant (defined as p<0.20) (Table 3). The multiple regression model confirmed the effect of hookworm as a risk factor and showed a protective effect of A. lumbricoides for the occurrence of uncomplicated falciparum malaria (Table 4).
4. Discussion The results of this study show that infection with A. lumbricoides is a protective factor, while hookworm infection is a risk factor, for the occurrence of uncomplicated P. falciparum malaria. The existence of a protective effect of A. lumbricoides for the occurrence of uncomplicated malaria has been described previously in two clinical trials.15,16 Our results are also consistent with studies reporting hookworm infection as a risk factor for uncomplicated malaria,11–13 although none of these studies found any effect of A. lumbricoides. This paradoxical association has been pointed out by Nacher and can be explained by different biological hypotheses.26 The prevalence of infection appears to be important in the observed associations. In the only study that has described the simultaneous effects of Ascaris and hookworm, the prevalences were 42.82% (95% CI: 39.42–46.27) and 34.38% (95% CI: 31.15–37.72), respectively.17 In the study by Kirwan et al.,14 Kenyan preschool age children from a region with high STH infection, especially by A. lumbricoides (45.63%, 95% CI: 40.07–51.26), reported A. lumbricoides as a risk factor. The prevalence of hookworm in this population was only 4.06% (95% CI: 2.18–6.85).14 These results are in contrast to the reported prevalences in two clinical trials with findings similar to those of our study. In the first of these trials, the prevalence of A. lumbricoides was 26.86% (95% CI: 22.28–31.83) and that of Necator americanus was 9.43% (95% CI: 6.58–12.99).15 In the second, the prevalence of A. lumbricoides was 55.66%
Table 3 Interactions between soil-transmitted helminths and Plasmodium falciparum infectiona
a
Interaction
Cases n (%)
Controls n (%)
OR
95% CI
p value
Ascaris lumbricoides and Trichuris trichiura A. lumbricoides T. trichiura A. lumbricoides + T. trichiura
17 (25.37) 18 (26.87) 8 (11.94)
56 (31.46) 37 (20.79) 22 (12.36)
0.72 1.75 0.73
0.31–1.67 0.68–4.49 0.16–3.39
0.44 0.24 0.69
T. trichiura and hookworm T. trichiura Hookworm T. trichiura + hookworm
18 (26.87) 24 (35.82) 12 (5.62)
37 (20.79) 31 (17.42) 10 (17.91)
0.63 1.19 6.27
0.23–1.72 0.45–3.13 1.08–36.38
0.37 0.73 0.04
Hookworm and A. lumbricoides A. lumbricoides Hookworm A. lumbricoides + hookworm
17 (25.37) 24 (35.82) 8 (11.94)
56 (31.46) 31 (17.42) 16 (8.99)
0.79 4.38 0.31
0.34–1.83 1.48–12.95 0.59–1.66
0.58 0.00 0.17
This table shows the interactions between the soil-transmitted helminths calculated adding multiplicative terms in conditional logistic models.
706
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
Table 4 Adjusted odds ratios (aOR) between Plasmodium falciparum infection and soil-transmitted helminthsa
a
Variable
aOR
95% CI
p-value
Ascaris lumbricoides Hookworm sp. Age (years) Wooden walls, walls of other plant material, or no walls Spraying of home with insecticide
0.43 4.36 0.95 6.11 0.26
0.18–1.04 1.68–11.31 0.91–0.99 2.02–18.46 0.77–0.85
0.06 0.00 0.03 0.00 0.02
Multiple conditional logistic model.
(95% CI: 48.70–62.46) and that of hookworm was not reported.16 The possible biological explanation for the paradoxical associations between Ascaris or hookworm and the occurrence of uncomplicated falciparum malaria was not established in this study. From animal models nematode parasites are known to induce a Th2 immune response in their hosts, while protozoa usually induce a Th1 response. Each of these responses facilitates the expulsion or death of the respective parasites, and indeed it has been suggested that the presence of nematode infections and a Th2polarised immune response decreases the Th1 response to stimuli that usually elicit it, such as infections with Plasmodium or Toxoplasma parasites.27–30 Some authors suggest that the immunological mechanisms that may explain the contradictory results (protection or increased susceptibility) of the interactions between helminths and malaria may depend upon the intensity of infection (helminth parasite load), the age of the study population and the helminth species.14,31 The study found no differences in the frequency distribution of total Ig E between cases and controls that would allow the exploration of possible biological explanations for the associations found. One of the potential biological mechanisms to explain the positive association between Ascaris infection and malaria is that STH can decrease the acquisition of immunity to Plasmodium.32 The paradoxical results for Ascaris, as found in this and other studies where this species has also been identified as a protective factor,15,16 could also be explained by age differences, with a trend to be a risk factor for children of preschool age and a protective factor for school-age children, pregnant women and adults.27,30 As for hookworm, both the current study and others have found mainly a positive association with the occurrence of malaria. The biological explanation for this association would be similar to that found in animal models.27,32 Additionally, a recent study found that this association is mainly present in groups of preschool-age children and adults, but not in school-age children.11 A major difference in the epidemiology of hookworm compared to other soil-transmitted helminths is that its prevalence increases with age, which suggests an occupational explanation. Unfortunately, the effect of age on the association could not be completely explored in the present study given the limited number of children under 15 years included (35, 14%). Although the study was explicitly designed to assess the association between STH and malaria, some limitations should be considered in interpreting the findings. First, the measurement of most variables of exposure was based
on self-reporting, which can have a measurement error. However, we have no reason to believe that this potential error was differential between cases and controls. Second, it is important to remember that, despite the attempt to isolate the effects of each parasite, it is possible that the observed associations are a reflection of co-infection or of its effects on the immune response. The findings reported on the differences in prevalences required to demonstrate the associations support this idea, which could also be seen with other soil-transmitted helminths and protozoa. It is possible that some confounding variables were not included in the analysis. Those that seem most relevant are those related to the mobility of individuals to areas where the mosquito bites occur, which also correspond to areas where hookworm infection is more likely, given the association of hookworm infection with occupational activities.13 Though partial control of this was sought through matching by location, significant differences between the individuals involved may remain. Unfortunately, the social conditions and on-going conflict in the study region did not allow this to be more accurately explored. However: 1. In this region, most of the existing malaria vectors (the most prevalent being Anopheles nuneztovari) bite during the night, inside and outside human dwellings.33 2. A protective effect of residential spraying for the occurrence of malaria was observed. 3. The absence of walls, or walls of inappropriate material, is a risk factor for malaria. 4. No difference in time of going to bed was found between cases and controls (p>0.05). Considering the above observations, it can be assumed that malaria is transmitted mainly during the night and within the house.33 Finally, it is interesting to recognise that the alleged associations found could not be explained, or at least not entirely, as real biological relationships between soiltransmitted helminths and malaria; rather, it is evident that the risk factors and social determinants predisposing individuals and populations to infection with STH tend to be the same as those associated with the occurrence of malaria.13 Therefore it is advisable that further studies be carried out to explore the common determinants of morbid complexes, taking a multilevel approach and including social factors in new and more complex aetiological models. The true significance of these findings from the standpoint of public health has yet to be established. However, the consistency and strength of the association found so far strongly support the recognition of hookworm
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
infection as a risk factor for malaria, which would support the instigation of an integrated control programme for both diseases. And rather than being understood as only co-infections, as is presently the case, the diseases could be considered as ‘co-determined’,13 and therefore requiring to be addressed from their common structural determinants and risk factors. The protective effect of A. lumbricoides in no way contradicts these potential implications for public health, because in this case the task would be to implement integrated control programmes for these infections. Given the known morbidity of STH alone and when overlapping with other infections,3,4 a potentially beneficial effect would be overshadowed by the well known deleterious effects of STH infections and, far from disputing their control, this constitutes an additional argument in favour of undertaking integrated control programmes based on their determinants. A cross-cutting perspective is required to increase the impact of isolated efforts through interprogramme coordination, especially in regions where resources are limited, and through intersectoral cooperation to address the common social determinants. Author’s contributions: JAF, ZMC and AJI designed the study. JAF, APG and PR coordinated the field work. JAF and AJI analysed data and prepared the first version of the manuscript. LIM, MCL, SB and APG analysed the biological samples, laboratory results, and reviewed the literature. All the authors participated completely in the discussion of the results and the writing of this manuscript. RSN was the coordinator and participated in all activities. JAF and RSN are guarantors of the paper. Acknowledgements: The authors thank especially the urban and rural community of Tierralta, Cordoba, and also the health and military authorities. In addition, this work would not have been possible without the valuable support of the microscopy technicians and motorcycle drivers. Funding: This project was funded by the National Program for Health Science and Technology, National Administrative Department of Science and Technology, Colciencias, project code: 210445921590, by the Colombian National Institute of Health (Instituto Nacional de Salud) and by the National University of Colombia School of Medicine. Competing interests: None declared. Ethical approval: The study was approved by the Institutional Review Boards of both the National University of Colombia School of Medicine and the Colombian National Institute of Health. Participants were informed about the study objectives and were guaranteed confidentiality of the results and the data collected. All gave informed consent to taking part in the study. References 1. WHO. World malaria report 2009. Geneva: WHO; 2009. 2. WHO. Report of the first meeting of WHO Strategic and Technical Advisory Group on Neglected Tropical Diseases. Geneva: WHO; 2007.
707
3. Crompton DWT, editor. Accelerating work to overcome the impact of neglected tropical diseases. A roadmap for implementation. Geneva: WHO; 2012. 4. Petney TN, Andrews RH. Communities of multiple parasites in animals and humans: frequency, structure and pathogenic significance. Int J Parasitol 1998;28:377–93. 5. Brooker S, Clarke S, Njagi JK, et al. Spatial clustering of malaria and associated risk factors during an epidemic in a highland area of western Kenya. Trop Med Int Health 2004;9:757–66. 6. Murray J, Murray A, Murray M, Murray C. The biological suppression of malaria: an ecological and nutritional interrelationship of a host and two parasites. Am J Clin Nutr 1978;31:1363–6. 7. Fernandez JA, Idrovo AJ, Cucunubá ZM, Reyes P. Validity of the studies of the association between malaria incidence and geohelminths: Should it impact the health policy? Rev Bras Epidemiol 2008;11:365–78. 8. Basavaraju S, Schantz P. Soil-transmitted helminths and Plasmodium falciparum malaria: epidemiology, clinical manifestations, and the role of nitric oxide in malaria co-infection and geohelminth. Do worms have a protective role in P. falciparum infection? Mt Sinai J Med 2006;73:1098–104. 9. Nacher M, Singhasivanon P, Yimsamran S, et al. Intestinal helminth infections are associated with increased incidence of Plasmodium falciparum malaria in Thailand. J Parasitol 2002;88:55–8. 10. Spiegel A, Tall A, Raphenon G, Trape JF, Druilhe P. Increased frequency of malaria attacks in subjects co-infected by intestinal worms and Plasmodium falciparum malaria. Trans R Soc Trop Med and Hyg 2003;97:198–9. 11. Pullan RL, Katabereine NB, Bukirwa H, Staedke SG, Brooker S. Heterogeneities and consequences of Plasmodium species and hookworm infection: A population based study in Uganda. J Infect Dis 2011;203:406–17. 12. Hilliguier SD, Booth M, Muhangi L, et al. Malaria and helminth coinfection in a semi-urban population of pregnant women in Uganda. J Infect Dis 2008;198:920–7. 13. Valencia CA, Fernández AJ, Cucunubá ZM, Reyes P, López MC. Correlation between malaria incidence and prevalence of soiltransmitted helminths in Colombia: an ecological approach. Biomédica 2010;30:501–8. 14. Kirwan P, Jackson AL, Asaolu SO, et al. Impact of repeated four-monthly anthelmintic treatment on Plasmodium infection in preschool children: a double-blind placebo-controlled randomized trial. BMC Infect Dis 2010;10:277. 15. Brutus L, Watier L, Briand V, Hanitrasoamampionona V, Razanatsoarilala H, Cot M. Parasitic co-infections: Does Ascaris lumbricoides protect against Plasmodium falciparum infection? Am J Trop Med Hyg 2006;75:194–8. 16. Brutus L, Watier L, Briand V, Hanitrasoamampionona V, Razanatsoarilala H, Cot M. Confirmation of the protective effect of Ascaris lumbricoides on Plasmodium falciparum infection: results of a randomized trial in Madagascar. Am J Trop Med Hyg 2007;77:1091–3. 17. Boel M, Carrara VI, Rijken M, et al. Complex interactions between soiltransmitted helminths and malaria in pregnant women on the ThaiBurmese border. PLoS Negl Trop Dis 2010;4:e887. 18. Shapiro AE, Tukahebwa EM, Kasten J, et al. Epidemiology of helminth infections and their relationship to clinical malaria in southwest Uganda. Trans R Soc Trop Med Hyg 2005;99:18–24. 19. Nacher M. Worms and malaria: resisting the temptation to generalize. Trends Parasitol 2006;22:350–1. 20. Booth M. The role of residential location in apparent helminth and malaria associations. Trends Parasitol 2006;22:359–62. 21. Beck JW, García A, Jartog EM, Shaner AL. Use of the Ritchie-Frick egg counting technic in the study of the effectiveness of the anthelmintic Monopar (stilbazium iodide) [in Spanish]. Rev Fac Med 1965;14:33–6. 22. Botero D, Restrepo M. Parasitosis humanas. 4th ed. Medellín, Colombia: Corporación para Investigaciones Biológicas; 2005. p. 93–139. 23. República de Colombia, Ministerio de Protección Social. Instituto Nacional de Salud. Guía para la Atención Clínica Integral del paciente con malaria. Bogata: Organización Panamericana de la Salud; 2010, p. 26-36. 24. Rothman KJ, Greenland S, Lash TL. Modern epidemiology. 3rd ed. Philadelphia, PA: Lippincott Williams; 2008. p. 116. 25. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: International survey. BMJ 2000;320:1240–3. 26. Nacher M. Interactions between worms and malaria: Good worms or bad worms? Malar J 2011;10:259. 27. Hanitrasoamampionona V, Brutus L, Hébrard G, et al. Epidemiologic study of the main human intestinal nematodes in the middle west of Madagascar [in French]. Bull Soc Pathol Exot 1998;91:77–80.
708
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
28. Jackson JA, Friberg IM, Little S, Bradley JE. Review series on helminths, immune modulation and the hygiene hypothesis: immunity against helminths and immunological phenomena in modern human populations: co evolutionary legacies? Immunology 2009;126: 18–27. 29. Roussilhon C, Brasseur P, Agnamey P, Pérignon JL, Druilhe P. Understanding human-Plasmodium falciparum immune interactions uncovers the immunological role of worms. PLoS One 2010;195: e9309. 30. Helmby H. Gastro-intestinal nematode infection exacerbates malaria induced liver pathology. J Immunol 2009;182:5663–71.
31. Hartgers FC, Obeng BB, Boakye D, Yazdanbakhsh M. Immune responses during helminth-malaria co-infection: a pilot study in Ghanaian school children. Parasitology 2008;135:855–60. 32. Jankovic D, Kullberg MC, Caspar P, Sher A. Parasite-induced Th2 polarization is associated with down-regulated dendritic cell responsiveness to Th1 stimuli and a transient delay in T lymphocyte cycling. J Immunol 2004;173:2419–27. 33. Gutiérrez LA, González JJ, Gómez GF, et al. Species composition and natural infectivity of anthropophilic Anopheles (Diptera: Culicidae) in the states of Córdoba and Antioquia, Northwestern Colombia. Mem Inst Oswaldo Cruz 2009;104:1117–24.
Contents lists available at SciVerse ScienceDirect
Transactions of the Royal Society of Tropical Medicine and Hygiene journal homepage: http://www.elsevier.com/locate/trstmh
Paradoxical associations between soil-transmitted helminths and Plasmodium falciparum infection ˜ a,∗ , Alvaro J. Idrovo b , Zulma M. Cucunubá c , Julián A. Fernández-Nino a Patricia Reyes-Harker , Ángela P. Guerra c , Ligia I. Moncada a , Myriam C. López a , Sandra M. Barrera c , Liliana J. Cortés c , Mario Olivera a , Rubén S. Nicholls a,c a Grupo de Infecciones y Salud en el Trópico, Unidad de Parasitología, Departamento de Salud Pública, Facultad de Medicina, Ciudad Universitaria, Universidad Nacional de Colombia, Bogotá D.C, Colombia b Centro de Investigación en Sistemas de Salud, Instituto Nacional de Salud Pública de México, Cuernavaca, México c Grupo de Bioquímica, Grupo de Parasitología, Instituto Nacional de Salud, Bogotá, Colombia
a r t i c l e
i n f o
Article history: Received 7 December 2011 Received in revised form 19 July 2012 Accepted 19 July 2012 Available online 11 August 2012 Keywords: Helminths Hookworm Malaria Ascaris lumbricoides Epidemiology Colombia
a b s t r a c t Evidence on the comorbidity between soil-transmitted helminth infections and malaria is scarce and divergent. This study explored the interactions between soil-transmitted helminth infections and uncomplicated falciparum malaria in an endemic area of Colombia. A paired case-control study matched by sex, age and location in Tierralta, Cordoba, was done between January and September 2010. The incident cases were 68 patients with falciparum malaria and 178 asymptomatic controls. A questionnaire was used to gather information on sociodemographic variables. Additionally physical examinations were carried out, stool samples were analysed for intestinal parasites and blood samples for Ig E concentrations. We found associations between infection with hookworm (OR: 4.21; 95% CI: 1.68–11.31) and Ascaris lumbricoides (OR 0.43; 95% CI: 0.18–1.04) and the occurrence of falciparum malaria. The effects of soil-transmitted helminths on the occurrence of malaria were found to be paradoxical. While hookworm is a risk factor, A. lumbricoides has a protective effect. The findings suggest that, in addition to the comorbidity, the presence of common determinants of soil-transmitted helminth infections and malaria could also exist. While the biological mechanisms involved are not clear, public health policies aimed at the control of their common social and environmental determinants are suggested. © 2012 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.
1. Introduction Malaria is considered to be the world’s most important tropical infection; the disease affects 243 million people annually, causing 900 000 deaths each year, and contributing approximately 30% of the burden of disease attributable to tropical diseases.1 Soil-transmitted helminth (STH) infections are chronic intestinal infections ∗ Corresponding author. Tel.: +57 1 316 5000x15032; fax: +57 1 316 5000x15033. ˜ E-mail address: [email protected] (J.A. Fernández-Nino).
that are widely distributed globally and considered to be among the neglected tropical diseases.2,3 Malaria and STH infection tend to occur in the same regions and affect the same individuals; they share the same risk factors (overlapping phenomena) and their effects on morbidity, especially among pregnant women and children, are well known.4,5 In the late 1970s the existence of a direct cause and effect relationship between STH infections and the occurrence of malaria was suggested.6 It was then proposed that STH infections could be a risk factor or a protecting factor for the development of malaria. These hypotheses have been explored many times in the
0035-9203/$ – see front matter © 2012 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.trstmh.2012.07.012
702
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
past 10 years,7 being supported by contradictory immunological models with some experimental evidence in animal models; this is primarily based on the potential shift to a Th2-type immune response as a result of STH infection, which would reduce the Th1-type response against malaria.8 The results of studies on the association between STH infections and the incidence of malaria have not been consistent. Four analytical studies, two cohort studies, two cross-sectional studies, one ecological study and a randomised controlled clinical trial found STH infections, or infection with a particular STH species, to be a risk factor for malaria.9–14 In contrast, an ecological study6 and two randomised clinical trials found a protective relationship.15,16 A recent cross-sectional study in pregnant women found that while Ascaris lumbricoides had a protective association, hookworm was positively associated with the occurrence of malaria.17 A case-control study reported no association.18 Several authors have drawn attention to the methodological limitations of these investigations.7,19 These limitations include temporal ambiguity, lack of information on specific explored parasites and the potential role of confounding. The importance of considering the places where the participants in these studies live (location bias) has also been emphasised, because the observed associations could be explained through the geographical distribution of parasites.12,20 The present study was explicitly designed to assess the association between STH infections and the occurrence of uncomplicated falciparum malaria, and to explore the possible interactions between parasites by controlling for potentially confounding variables and for location bias. 2. Materials and methods 2.1. Design and study site A case-control study matched by sex, age (range: 5–15, 16–45 and 46–60 years) and neighbourhood of usual residence was conducted between January and September 2010 in Tierralta (Cordoba, Colombia). Tierralta is the municipality with the highest number of reported cases of malaria in Cordoba, which in turn is one of the three departments with the highest incidence in Colombia. In 2010, 6204 malaria cases were reported for Tierralta, of which 25.53% were caused by Plasmodium falciparum. The people of Tierralta make their living from agriculture and rearing livestock; serious problems in accessing healthcare services are further complicated by the presence and substantial influence of illegal armed groups and drug traffickers. 2.2. Recruitment and selection Each participant included in the study met the following criteria: had lived for at least 1 year in a rural area of Tierralta; had not received anthelmintic drugs within the previous 3 months; was aged between 5 and 60 years old; was not pregnant. Cases were defined as patients with
a diagnosis of uncomplicated P. falciparum malaria confirmed by a thick blood smear, detected as incident cases during the study period at the centre for malaria diagnosis in the urban area of the municipality. All the included cases lived in one of 11 of the 19 rural areas of the municipality. Eight rural areas were excluded because of problems of access and concern for the security of the fieldwork team. Controls were defined as individuals who lived within 500 m of the residence of the case with whom they were paired. Controls were required not to have had a malaria infection within the previous 3 months, nor to have reported having fever in the previous 15 days. All controls were subjected to a blood smear to rule out asymptomatic malaria infection. The selection and recruitment of controls required visiting the house of each participant during the week of his or her inclusion in the study. While at a participant’s house we identified neighbours who could qualify as controls, and in order to recruit controls walked in a clockwise direction from the participant’s house. One to three controls per case were recruited, depending on the possible candidates’ availability and willingness to participate. To control for a possible location bias, controls were required not to have left their neighbourhood for more than 3 consecutive days in the previous 3 weeks. This condition was intended to increase the likelihood that each case would have controls who may have acquired malaria in the same place of recruitment,20 and who therefore shared the same ecological transmission risk. 2.3. Data collection Participants responded to a socioeconomic survey and their place of residence was observed to measure relevant covariates that could act as confounding variables, on the basis of a recent systematic review.7 The questionnaires were administered and the observations made by a physician trained in data collection instruments. The instruments used were previously validated with 70 individuals of the region in a pilot study with a good reproducibility (kappa: 0.82; 95% CI: 0.45–0.98). On the day after their recruitment, at their place of residence, all participants underwent a general physical examination (height, weight, blood pressure, heart rate, respiratory rate, temperature, cardiovascular and respiratory auscultation, abdominal palpation) and were asked to provide a stool sample. Stool samples were preserved in 10% formalin and shipped by air for analysis at the parasitological laboratory of the National University of Colombia School of Medicine in Bogota. All the faecal samples underwent direct examination and the Ritchie–Frick modified concentration technique for diagnosis of STH and protozoa.21 The samples were analysed independently by two experienced observers who were unaware of the status as cases or controls of the study subjects. Disagreements were discussed between both observers; the interobserver final concordance was very good (kappa 0.86; 95% CI: 0.63–0.91). The thick blood smears from cases and controls were read out at the study site by a certified microscopy
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
technician who was retrained by the research team before beginning the study. Additionally all slides were subsequently subject to quality control by an expert at the parasitological laboratory of the Colombian National Institute of Health (NIH) in Bogota. This expert knew neither data about the patient nor his or her case or control status. When there was a disagreement (<2%), the slides were reviewed by another expert observer. At the time of recruitment, blood samples were drawn from all the participants; the sera obtained were stored at −20 ◦ C in the biochemistry laboratory at the NIH in Bogota for measurement of Ig E. To measure Ig E levels, we used the immunoglobulin kit known as AccuBind ELISA (Monobind Inc., Lake Forest, CA, USA), which allows the quantitative determination of total Ig E in serum. Samples were processed strictly according to the manufacturer’s recommendations. Calibration curves were prepared using known standards of concentration. Samples were processed in batches and each trial had its own calibration curve and internal controls. Reference values for Ig E were reported taking into account the age of patients as follows: age 3–16 years, up to 280 IU/ml; age 17 years or over, range 0–200 IU/ml. 2.4. Statistical analysis The main demographic and socioeconomic characteristics for both cases and controls were initially described. Then associations between malaria and geohelminths and other variables were explored using simple conditional logistic regression models. We examined frequency of infection by intestinal parasites for the parasites present in the participants. For all the infections found, an estimate of their agreement per parasite pair was made and Cohen’s Kappa index was calculated. Variables showing statistically significant associations (p<0.20) were included for the multivariable analysis. In addition, an exploration was made for interactions between the three STH species, A. lumbricoides, Trichuris trichiura (whipworm) and hookworm. We evaluated several models, including one or two STH species as independent variables at the same time, until a parsimonious model with A. lumbricoides and hookworm was observed, which was adjusted for sociodemographic variables. Analyses were performed using the statistical program Stata 11 (Stata Corporation, College Station, TX, USA). 2.5. Ethical considerations Participants were informed about the study objectives and were guaranteed confidentiality of the results and the data collected. All participants gave informed consent to taking part in the study. When STH or other intestinal parasites were diagnosed, free treatment was provided according to the currently recommended and available treatment schemes in Colombia.22 All patients with malaria, symptomatic or asymptomatic, were treated immediately in accordance with existing national guidelines23 and clinically monitored by staff of the research team.
703
3. Results 3.1. General characteristics of participants Initially 74 cases and 182 controls were recruited; seven cases and four controls were later excluded from the study. The five cases were excluded because it was not possible to find their respective controls (no neighbours meeting inclusion criteria or neighbours not willing to give informed consent) and two cases because during quality control their thick blood smears were found to be positive for either P. vivax malaria or mixed infection. The four controls were excluded because quality control showed that they were actually positive for P. vivax. Finally 67 cases and 178 controls were included for further analyses. Of these, three individuals were initially recruited as controls but found to be positive for P. falciparum in new thick blood smears within the first week after recruitment. Therefore they were included in the study twice, both as cases and as controls,24 and controls were sought for them. The socioeconomic characteristics and practices relevant for malaria and their crude associations with P. falciparum infection are presented in Table 1. Of the total of 245 subjects, 143 (58.37%) were men, 103 (57.86%) of the control group and 40 (59.79%) of the cases. The mean age of the controls was 31 years (minimum 5 and maximum 60 years) and that of the cases 27 years (minimum 6 and maximum 59 years); 35 participants (14.28%) were under 15 years of age. Table 1 shows the crude associations between P. falciparum and STH infections. The presence of hookworm was associated with malaria, while T. trichiura and A. lumbricoides infections were not. The frequencies of other parasites in both cases and controls are also presented in Table 1, but no statistically significant differences were found between the groups. No significant differences were found between cases and controls in most socioeconomic characteristics explored, such as type of social security affiliation, occupation, persons per room or weekly family income (p>0.25). Of these conditions only house wall material (wood or other plant material vs block or brick) showed a statistically significant association with the occurrence of falciparum malaria. The time since last deworming did not show statistically significant differences between cases and controls, demonstrating the lack of widespread access to some anthelmintics in the study population, 64.08% of whom did not receive anthelmintic drugs during the previous year. For the analysis of body mass index (BMI), this variable was categorised for children as high (higher than percentile 97), normal (percentiles 3–97) and low (below percentile 3), using international reference curves;25 for adults, we used the standard classification proposed by the WHO, re-classifying the classic categories into three: low (BMI<18), normal (BMI≥18 to <25) and high (including overweight and obesity, BMI≥25), and therefore were able to analyse jointly this variable in children and adults. The results showed that most of the population (212, 86.53%) had normal BMI values. A low BMI was associated with the occurrence of falciparum malaria (p<0.05). Agricultural
704
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
Table 1 Bivariate association between soil-transmitted helminths, other relevant variables and Plasmodium falciparum infection Controls Variable
Cases %
n
%
OR
56 37 31
31.64 20.90 17.42
17 18 24
25.37 26.87 35.82
0.65 1.39 2.95
Age group <15 years 15-45 years >45 years
35 117 26
19.66 65.73 14.61
10 42 15
14.92 62.69 22.39
Social security regimen Subsidised affiliated Social security payer State subsidiseda None
128 2 17 31
71.91 1.12 9.55 17.41
55 2 3 7
82.09 2.99 4.48 10.45
1 2.59 0.31 0.44
0.21–31.31 0.83–1.13 0.17–1.13
52 55 36 35
29.91 30.90 20.22 19.66
15 20 18 15
22.06 29.41 26.47 22.06
1 4.03 5.56 1.38
0.81–20.05 1.40–22.11 0.54–3.53
Malaria episodes in previous year Zero One Two Three or more
139 34 2 3
78.09 19.10 1.12 1.69
44 1 14 8
65.67 1.49 20.90 11.95
1 0.93 15.53 7.38
0.12–0.72 3.28–73.53 1.18–45.97
Last deworming 3–6 months ago 6 months–1 year ago >1 year ago
35 23 120
19.66 12.92 67.42
10 6 51
14.93 8.96 76.12
1 1.12 1.85
Body mass indexb High Normal Low
15 157 6
8.43 88.20 3.37
4 55 8
5.97 82.02 11.94
0.69 1 3.37
House wall material Brick or block Crude wood/vegetation No walls
56 119 3
31.46 65.89 1.69
6 56 5
8.96 83.54 7.46
1 7.23 37.81
87 91
48.88 51.12
34 33
50.75 49.25
1 0.99
0.47–2.11
143
80.34
54
80.60
1.10
0.52–2.32
47 37 94
26.4 20.79 52.81
12 20 35
17.91 29.85 52.24
1 2.53 1.40
1.10–5.86 0.66–2.98
34 47 134
19.10 26.40 75.28
5 16 52
7.46 23.88 77.61
0.33 0.75 0.99
0.12–0.95 0.36–1.54 0.49–2.00
12 33 61 55 15 11
6.74 18.54 34.27 30.90 8.43 6.18
4 12 19 17 2 3
5.97 17.91 28.36 25.37 2.99 4.48
0.99 1.03 0.87 0.77 0.34 0.73
0.31–3.17 0.50–2.10 0.46–1.66 0.40–1.47 0.74–1.55 0.20–2.65
160
89.89
61
91.04
0.98
0.35–2.75
Geohelminths Ascaris lumbricoides Trichuris trichiura Hookworm
Occupation Farmer Housework Student Other
Place of defaecation Outdoors Toilet Literacy Bednet use Net not available Net available but not in use Net used House spraying the previous month Use of protective clothing Agriculture practice Other parasites Giardia duodenalis Entamoeba histolytica/dispar Entamoeba coli Entamoeba nana Iodoameba sp. Blastocystis hominis Ig Eb High
Weekly income by resident (US$) People per room Age (continuous)
n
Median
Median
2.60 3.46 31.72
2.80 3.42 29.45
95% CI 0.31–1.33 0.67–2.85 1.429–6.16
0.97 0.99 0.96
All estimations were made with simple conditional logistic models. a In Colombia, people not yet affiliated to the new health system but who have been identified and receive health care provided by the government. b For reference values, see text (Results).
0.34–3.72 0.74–4.76
0.19–2.50 1.20–11.86
2.44–21.45 5.24–272.97
0.82–1.23 0.84–1.16 0.92–0.99
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
705
Table 2 Proportion of agreement in infection status between parasite pairs (%) observed in the whole study groupa
Trichuris trichiura Hookworm Giardia duodenalis Entamoeba histolytica/dispar Entamoeba coli Endolimax nana Iodoamoeba sp. Blastocystis hominis
Ascaris lumbricoides
Trichuris trichiura
72.24b 67.35b 66.94 68.98 65.31 57.14 67.35 68.57
73.06b 75.92b 67.35 58.78 58.78 72.24 75.92
Hookworm
72.65 71.43b 62.86 57.96 71.43 74.29
Giardia duodenalis
Entamoeba histolytica/dispar
Entamoeba coli
Endolimax nana
76.73 63.27 69.80 86.53 89.39
71.84b 66.94b 81.22b 78.37
66.53b 71.02b 67.35
71.84b 70.61b
Iodameba sp.
90.61b
a
This table reports the agreement in infection status between parasite pairs (%). The values were evaluated with the p-value of the respective Cohen’s Kappa to determine agreement regardless of random occurrence. b p value <0.05, Cohen’s Kappa.
labour, literacy, use of mosquito nets and wearing protective clothing during work were not significantly associated with the occurrence of falciparum malaria (p>0.25). In contrast, spraying of the house with insecticides by the municipal government in the past month was a protective factor for the outcome (p<0.05). Of the samples analysed, 221 (90.20%) had high Ig E levels, with no significant differences between cases and controls (89.89% vs 91.04%, p=0.97). The analysis of 245 faecal samples tested showed that 73 (29.80%) were positive for A. lumbricoides, 55 (22.44%) for T. trichiura and 55 (22.45%) for hookworm. Exclusive coinfections between pairs of these STH were present with frequencies of 4.08% for Ascaris and hookworm, 6.53% for Ascaris and Trichuris, and 3.30% for Trichuris and hookworm. Finally, 5.71% had simultaneous infection with the three soil-transmitted helminth species. Table 2 shows the percentage of agreement in infection status between parasite pairs. Concurrent infections were found (p<0.05) between some of the soil-transmitted helminths and protozoa, such as between T. trichiura and Giardia duodenalis, and between hookworm and the Entamoeba histolytica/E. dispar complex. When exploring the interactions between soiltransmitted helminths and the occurrence of falciparum malaria, only those between hookworm and Ascaris and hookworm and Trichuris were statistically significant (defined as p<0.20) (Table 3). The multiple regression model confirmed the effect of hookworm as a risk factor and showed a protective effect of A. lumbricoides for the occurrence of uncomplicated falciparum malaria (Table 4).
4. Discussion The results of this study show that infection with A. lumbricoides is a protective factor, while hookworm infection is a risk factor, for the occurrence of uncomplicated P. falciparum malaria. The existence of a protective effect of A. lumbricoides for the occurrence of uncomplicated malaria has been described previously in two clinical trials.15,16 Our results are also consistent with studies reporting hookworm infection as a risk factor for uncomplicated malaria,11–13 although none of these studies found any effect of A. lumbricoides. This paradoxical association has been pointed out by Nacher and can be explained by different biological hypotheses.26 The prevalence of infection appears to be important in the observed associations. In the only study that has described the simultaneous effects of Ascaris and hookworm, the prevalences were 42.82% (95% CI: 39.42–46.27) and 34.38% (95% CI: 31.15–37.72), respectively.17 In the study by Kirwan et al.,14 Kenyan preschool age children from a region with high STH infection, especially by A. lumbricoides (45.63%, 95% CI: 40.07–51.26), reported A. lumbricoides as a risk factor. The prevalence of hookworm in this population was only 4.06% (95% CI: 2.18–6.85).14 These results are in contrast to the reported prevalences in two clinical trials with findings similar to those of our study. In the first of these trials, the prevalence of A. lumbricoides was 26.86% (95% CI: 22.28–31.83) and that of Necator americanus was 9.43% (95% CI: 6.58–12.99).15 In the second, the prevalence of A. lumbricoides was 55.66%
Table 3 Interactions between soil-transmitted helminths and Plasmodium falciparum infectiona
a
Interaction
Cases n (%)
Controls n (%)
OR
95% CI
p value
Ascaris lumbricoides and Trichuris trichiura A. lumbricoides T. trichiura A. lumbricoides + T. trichiura
17 (25.37) 18 (26.87) 8 (11.94)
56 (31.46) 37 (20.79) 22 (12.36)
0.72 1.75 0.73
0.31–1.67 0.68–4.49 0.16–3.39
0.44 0.24 0.69
T. trichiura and hookworm T. trichiura Hookworm T. trichiura + hookworm
18 (26.87) 24 (35.82) 12 (5.62)
37 (20.79) 31 (17.42) 10 (17.91)
0.63 1.19 6.27
0.23–1.72 0.45–3.13 1.08–36.38
0.37 0.73 0.04
Hookworm and A. lumbricoides A. lumbricoides Hookworm A. lumbricoides + hookworm
17 (25.37) 24 (35.82) 8 (11.94)
56 (31.46) 31 (17.42) 16 (8.99)
0.79 4.38 0.31
0.34–1.83 1.48–12.95 0.59–1.66
0.58 0.00 0.17
This table shows the interactions between the soil-transmitted helminths calculated adding multiplicative terms in conditional logistic models.
706
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
Table 4 Adjusted odds ratios (aOR) between Plasmodium falciparum infection and soil-transmitted helminthsa
a
Variable
aOR
95% CI
p-value
Ascaris lumbricoides Hookworm sp. Age (years) Wooden walls, walls of other plant material, or no walls Spraying of home with insecticide
0.43 4.36 0.95 6.11 0.26
0.18–1.04 1.68–11.31 0.91–0.99 2.02–18.46 0.77–0.85
0.06 0.00 0.03 0.00 0.02
Multiple conditional logistic model.
(95% CI: 48.70–62.46) and that of hookworm was not reported.16 The possible biological explanation for the paradoxical associations between Ascaris or hookworm and the occurrence of uncomplicated falciparum malaria was not established in this study. From animal models nematode parasites are known to induce a Th2 immune response in their hosts, while protozoa usually induce a Th1 response. Each of these responses facilitates the expulsion or death of the respective parasites, and indeed it has been suggested that the presence of nematode infections and a Th2polarised immune response decreases the Th1 response to stimuli that usually elicit it, such as infections with Plasmodium or Toxoplasma parasites.27–30 Some authors suggest that the immunological mechanisms that may explain the contradictory results (protection or increased susceptibility) of the interactions between helminths and malaria may depend upon the intensity of infection (helminth parasite load), the age of the study population and the helminth species.14,31 The study found no differences in the frequency distribution of total Ig E between cases and controls that would allow the exploration of possible biological explanations for the associations found. One of the potential biological mechanisms to explain the positive association between Ascaris infection and malaria is that STH can decrease the acquisition of immunity to Plasmodium.32 The paradoxical results for Ascaris, as found in this and other studies where this species has also been identified as a protective factor,15,16 could also be explained by age differences, with a trend to be a risk factor for children of preschool age and a protective factor for school-age children, pregnant women and adults.27,30 As for hookworm, both the current study and others have found mainly a positive association with the occurrence of malaria. The biological explanation for this association would be similar to that found in animal models.27,32 Additionally, a recent study found that this association is mainly present in groups of preschool-age children and adults, but not in school-age children.11 A major difference in the epidemiology of hookworm compared to other soil-transmitted helminths is that its prevalence increases with age, which suggests an occupational explanation. Unfortunately, the effect of age on the association could not be completely explored in the present study given the limited number of children under 15 years included (35, 14%). Although the study was explicitly designed to assess the association between STH and malaria, some limitations should be considered in interpreting the findings. First, the measurement of most variables of exposure was based
on self-reporting, which can have a measurement error. However, we have no reason to believe that this potential error was differential between cases and controls. Second, it is important to remember that, despite the attempt to isolate the effects of each parasite, it is possible that the observed associations are a reflection of co-infection or of its effects on the immune response. The findings reported on the differences in prevalences required to demonstrate the associations support this idea, which could also be seen with other soil-transmitted helminths and protozoa. It is possible that some confounding variables were not included in the analysis. Those that seem most relevant are those related to the mobility of individuals to areas where the mosquito bites occur, which also correspond to areas where hookworm infection is more likely, given the association of hookworm infection with occupational activities.13 Though partial control of this was sought through matching by location, significant differences between the individuals involved may remain. Unfortunately, the social conditions and on-going conflict in the study region did not allow this to be more accurately explored. However: 1. In this region, most of the existing malaria vectors (the most prevalent being Anopheles nuneztovari) bite during the night, inside and outside human dwellings.33 2. A protective effect of residential spraying for the occurrence of malaria was observed. 3. The absence of walls, or walls of inappropriate material, is a risk factor for malaria. 4. No difference in time of going to bed was found between cases and controls (p>0.05). Considering the above observations, it can be assumed that malaria is transmitted mainly during the night and within the house.33 Finally, it is interesting to recognise that the alleged associations found could not be explained, or at least not entirely, as real biological relationships between soiltransmitted helminths and malaria; rather, it is evident that the risk factors and social determinants predisposing individuals and populations to infection with STH tend to be the same as those associated with the occurrence of malaria.13 Therefore it is advisable that further studies be carried out to explore the common determinants of morbid complexes, taking a multilevel approach and including social factors in new and more complex aetiological models. The true significance of these findings from the standpoint of public health has yet to be established. However, the consistency and strength of the association found so far strongly support the recognition of hookworm
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
infection as a risk factor for malaria, which would support the instigation of an integrated control programme for both diseases. And rather than being understood as only co-infections, as is presently the case, the diseases could be considered as ‘co-determined’,13 and therefore requiring to be addressed from their common structural determinants and risk factors. The protective effect of A. lumbricoides in no way contradicts these potential implications for public health, because in this case the task would be to implement integrated control programmes for these infections. Given the known morbidity of STH alone and when overlapping with other infections,3,4 a potentially beneficial effect would be overshadowed by the well known deleterious effects of STH infections and, far from disputing their control, this constitutes an additional argument in favour of undertaking integrated control programmes based on their determinants. A cross-cutting perspective is required to increase the impact of isolated efforts through interprogramme coordination, especially in regions where resources are limited, and through intersectoral cooperation to address the common social determinants. Author’s contributions: JAF, ZMC and AJI designed the study. JAF, APG and PR coordinated the field work. JAF and AJI analysed data and prepared the first version of the manuscript. LIM, MCL, SB and APG analysed the biological samples, laboratory results, and reviewed the literature. All the authors participated completely in the discussion of the results and the writing of this manuscript. RSN was the coordinator and participated in all activities. JAF and RSN are guarantors of the paper. Acknowledgements: The authors thank especially the urban and rural community of Tierralta, Cordoba, and also the health and military authorities. In addition, this work would not have been possible without the valuable support of the microscopy technicians and motorcycle drivers. Funding: This project was funded by the National Program for Health Science and Technology, National Administrative Department of Science and Technology, Colciencias, project code: 210445921590, by the Colombian National Institute of Health (Instituto Nacional de Salud) and by the National University of Colombia School of Medicine. Competing interests: None declared. Ethical approval: The study was approved by the Institutional Review Boards of both the National University of Colombia School of Medicine and the Colombian National Institute of Health. Participants were informed about the study objectives and were guaranteed confidentiality of the results and the data collected. All gave informed consent to taking part in the study. References 1. WHO. World malaria report 2009. Geneva: WHO; 2009. 2. WHO. Report of the first meeting of WHO Strategic and Technical Advisory Group on Neglected Tropical Diseases. Geneva: WHO; 2007.
707
3. Crompton DWT, editor. Accelerating work to overcome the impact of neglected tropical diseases. A roadmap for implementation. Geneva: WHO; 2012. 4. Petney TN, Andrews RH. Communities of multiple parasites in animals and humans: frequency, structure and pathogenic significance. Int J Parasitol 1998;28:377–93. 5. Brooker S, Clarke S, Njagi JK, et al. Spatial clustering of malaria and associated risk factors during an epidemic in a highland area of western Kenya. Trop Med Int Health 2004;9:757–66. 6. Murray J, Murray A, Murray M, Murray C. The biological suppression of malaria: an ecological and nutritional interrelationship of a host and two parasites. Am J Clin Nutr 1978;31:1363–6. 7. Fernandez JA, Idrovo AJ, Cucunubá ZM, Reyes P. Validity of the studies of the association between malaria incidence and geohelminths: Should it impact the health policy? Rev Bras Epidemiol 2008;11:365–78. 8. Basavaraju S, Schantz P. Soil-transmitted helminths and Plasmodium falciparum malaria: epidemiology, clinical manifestations, and the role of nitric oxide in malaria co-infection and geohelminth. Do worms have a protective role in P. falciparum infection? Mt Sinai J Med 2006;73:1098–104. 9. Nacher M, Singhasivanon P, Yimsamran S, et al. Intestinal helminth infections are associated with increased incidence of Plasmodium falciparum malaria in Thailand. J Parasitol 2002;88:55–8. 10. Spiegel A, Tall A, Raphenon G, Trape JF, Druilhe P. Increased frequency of malaria attacks in subjects co-infected by intestinal worms and Plasmodium falciparum malaria. Trans R Soc Trop Med and Hyg 2003;97:198–9. 11. Pullan RL, Katabereine NB, Bukirwa H, Staedke SG, Brooker S. Heterogeneities and consequences of Plasmodium species and hookworm infection: A population based study in Uganda. J Infect Dis 2011;203:406–17. 12. Hilliguier SD, Booth M, Muhangi L, et al. Malaria and helminth coinfection in a semi-urban population of pregnant women in Uganda. J Infect Dis 2008;198:920–7. 13. Valencia CA, Fernández AJ, Cucunubá ZM, Reyes P, López MC. Correlation between malaria incidence and prevalence of soiltransmitted helminths in Colombia: an ecological approach. Biomédica 2010;30:501–8. 14. Kirwan P, Jackson AL, Asaolu SO, et al. Impact of repeated four-monthly anthelmintic treatment on Plasmodium infection in preschool children: a double-blind placebo-controlled randomized trial. BMC Infect Dis 2010;10:277. 15. Brutus L, Watier L, Briand V, Hanitrasoamampionona V, Razanatsoarilala H, Cot M. Parasitic co-infections: Does Ascaris lumbricoides protect against Plasmodium falciparum infection? Am J Trop Med Hyg 2006;75:194–8. 16. Brutus L, Watier L, Briand V, Hanitrasoamampionona V, Razanatsoarilala H, Cot M. Confirmation of the protective effect of Ascaris lumbricoides on Plasmodium falciparum infection: results of a randomized trial in Madagascar. Am J Trop Med Hyg 2007;77:1091–3. 17. Boel M, Carrara VI, Rijken M, et al. Complex interactions between soiltransmitted helminths and malaria in pregnant women on the ThaiBurmese border. PLoS Negl Trop Dis 2010;4:e887. 18. Shapiro AE, Tukahebwa EM, Kasten J, et al. Epidemiology of helminth infections and their relationship to clinical malaria in southwest Uganda. Trans R Soc Trop Med Hyg 2005;99:18–24. 19. Nacher M. Worms and malaria: resisting the temptation to generalize. Trends Parasitol 2006;22:350–1. 20. Booth M. The role of residential location in apparent helminth and malaria associations. Trends Parasitol 2006;22:359–62. 21. Beck JW, García A, Jartog EM, Shaner AL. Use of the Ritchie-Frick egg counting technic in the study of the effectiveness of the anthelmintic Monopar (stilbazium iodide) [in Spanish]. Rev Fac Med 1965;14:33–6. 22. Botero D, Restrepo M. Parasitosis humanas. 4th ed. Medellín, Colombia: Corporación para Investigaciones Biológicas; 2005. p. 93–139. 23. República de Colombia, Ministerio de Protección Social. Instituto Nacional de Salud. Guía para la Atención Clínica Integral del paciente con malaria. Bogata: Organización Panamericana de la Salud; 2010, p. 26-36. 24. Rothman KJ, Greenland S, Lash TL. Modern epidemiology. 3rd ed. Philadelphia, PA: Lippincott Williams; 2008. p. 116. 25. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: International survey. BMJ 2000;320:1240–3. 26. Nacher M. Interactions between worms and malaria: Good worms or bad worms? Malar J 2011;10:259. 27. Hanitrasoamampionona V, Brutus L, Hébrard G, et al. Epidemiologic study of the main human intestinal nematodes in the middle west of Madagascar [in French]. Bull Soc Pathol Exot 1998;91:77–80.
708
J.A. Fernández-Ni˜ no et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 701–708
28. Jackson JA, Friberg IM, Little S, Bradley JE. Review series on helminths, immune modulation and the hygiene hypothesis: immunity against helminths and immunological phenomena in modern human populations: co evolutionary legacies? Immunology 2009;126: 18–27. 29. Roussilhon C, Brasseur P, Agnamey P, Pérignon JL, Druilhe P. Understanding human-Plasmodium falciparum immune interactions uncovers the immunological role of worms. PLoS One 2010;195: e9309. 30. Helmby H. Gastro-intestinal nematode infection exacerbates malaria induced liver pathology. J Immunol 2009;182:5663–71.
31. Hartgers FC, Obeng BB, Boakye D, Yazdanbakhsh M. Immune responses during helminth-malaria co-infection: a pilot study in Ghanaian school children. Parasitology 2008;135:855–60. 32. Jankovic D, Kullberg MC, Caspar P, Sher A. Parasite-induced Th2 polarization is associated with down-regulated dendritic cell responsiveness to Th1 stimuli and a transient delay in T lymphocyte cycling. J Immunol 2004;173:2419–27. 33. Gutiérrez LA, González JJ, Gómez GF, et al. Species composition and natural infectivity of anthropophilic Anopheles (Diptera: Culicidae) in the states of Córdoba and Antioquia, Northwestern Colombia. Mem Inst Oswaldo Cruz 2009;104:1117–24.