Malaria incidence among indigenous communities in Amazonas, Bolivar and Sucre
A total of 116,097 new cases of malaria were registered in the 2014-2018 period among Amerindians. Of these, 62,267 (53.6%) were male and 53,830 (46.4%), female. Median age was 19 years (IQR: IQR: 9-31 years, additional file 1).
Studied patients belonged to 41 different ethnic groups. However, only nine accounted for 91% of cases. These are: Pemon (27.3%), Jivi (25%), Piaroa/Wotjuja (16.1%), Yekuana (8.9%), Kurripaco (3.9%), Eñepa (2.9%), Piapoko (2.9%), Yanomami (2.8%) and Baré (1.3%). For a full list, see additional file 1.
Most cases were registered in Amazonas (56.6%) followed by Bolivar (43.2%) and Sucre (0.25%). The year with the highest number of new malaria cases was 2017 with 30,976.
Amazonas state
Most patients belonged to ethnic groups located north of the state, namely, Jivi (39.8%), Wotjuja (25%) and Yekuana (9.1%). The number of new malaria cases in indigenous communities in 2018 (21,530), was 174.2% higher than compared to the 2014 baseline (7,852). An increasing trend in the number of monthly cases was observed between January 2016 and January 2018, when cases started to decrease (Figure 1A). Despite this, the relative malaria risk for indigenous people compared to the non-indigenous population fell from 2.44 in 2014 (p<0.01), to 0.89 (p<0.01) in 2018 (table 1). Cases in Amazonas were typically higher between January and June, but a clear seasonal pattern was not observed (Figure1B), and no significant differences were seen in the median number of monthly cases during the study period (p=0.57, additional file 5).
At municipal level, the Atures Municipality accounted for most cases during the five-year study period (55,28%). There was a significant difference in the median incidence of municipalities (p<0.01), with the highest median API-i registered in Manapiare: 249.05 (IQR:151.25-301.64, Figure 2, additional file 2).
Bolivar state
The Pemon, the most numerous ethnic group in the state, accounted for 63.2% of all cases in Bolivar, followed by the Yekuana (8.7%) and the Eñepa (6.6%). As in Amazonas, malaria cases increased during the study period, compared to the 2014 baseline (7,854), but particularly in 2016 (15,676). From this point on, incidence reduced, reaching 8,565 cases in 2018, 9.05% more than in 2014 (Figure 1C). Compared to that in the rest of the population, malaria risk for indigenous people was consistently higher in Bolivar: RR: 3.39 in 2014 (p<0.01), 2.47 in 2015 (p<0.01), 2.80 in 2016 (p<0.01), 1.11 in 2017 (p<0.01) and 1.02 in 2018 (p=0.04, table 1). A seasonal pattern seems more patent in this state, as cases peak between February and April, and start declining from June onwards (Figure 1C-D). There was a significant difference in the median number of monthly cases throughout the five-year period (p<0.01).
At municipal level, almost a quarter of all cases originated in the Gran Sabana municipality (24.2%), followed by Angostura (formerly known and noted in maps as Raul Leoni, 23.3%) and Cedeño (20.4%) municipalities. Incidence in the study period was significantly different between municipalities (p=0.012), with the highest median API-i registered in El Callao: 575.76 cases per 1,000 indigenous people (IQR: 457.14-1205.88). This however, is probably a result of the extremely small projected indigenous population in the municipality (additional file 3). Angostura had the highest median API-i of all municipalities with 100 indigenous inhabitants or more (310.84, IQR: 207.81-334.57, Figure 2, additional file 2).
Sucre state
The Warao ethnic group accounted for 228 cases (78.4%), followed by the Uruak (10%), and the Puinave (7.2%). Although the number of cases remained considerably lower than in Amazonas and Bolivar, it increased from zero in 2014 and 2015 to 140 in 2017, followed by a 72% reduction in 2018 (39 cases). Annual incidence peaked in 2017 (5.75 cases per 1,000 indigenous people), and then reduced to 1.58 in 2018 (Figure 1E, table 1).
At municipal level, Benítez registered the highest API-I of the state (174.52, IQR: 0-174.52), however, this difference was not significant (p=0.06) when municipalities with an indigenous population below 100 were excluded from the analysis. There is no evident seasonal pattern in Sucre (Figure 1F).
Table 1: Incidence of malaria among indigenous and non-indigenous groups in Amazonas, Bolivar and Sucre states, Venezuela, 2014-2018.
Amazonas
|
Year
|
Cases-i
|
Est. Pop-i
|
API-i
|
Cases non-i
|
Est. Pop non-i
|
API non-i
|
RR
|
2014
|
7852
|
82800
|
94.83
|
3512
|
90344
|
38.87
|
*2.44
|
2015
|
7164
|
85082
|
84.20
|
11477
|
92921
|
123.51
|
*0.68
|
2016
|
7942
|
87427
|
90.84
|
17009
|
95481
|
178.14
|
*0.51
|
2017
|
21187
|
89837
|
235.84
|
45113
|
98015
|
460.27
|
*0.51
|
2018
|
21530
|
92313
|
233.23
|
26283
|
100522
|
261.47
|
*0.89
|
Bolivar
|
2014
|
7854
|
57451
|
136.71
|
67089
|
1665910
|
40.27
|
*3.39
|
2015
|
8387
|
58404
|
143.60
|
98466
|
1693846
|
58.13
|
*2.47
|
2016
|
15676
|
59372
|
264.03
|
161943
|
1721527
|
94.07
|
*2.81
|
2017
|
9649
|
60357
|
159.87
|
251673
|
1748949
|
143.90
|
*1.11
|
2018
|
8565
|
61357
|
139.59
|
242801
|
1776128
|
136.70
|
*1.02
|
Sucre
|
2014
|
0
|
23256
|
0.00
|
922
|
988415
|
0.93
|
0
|
2015
|
0
|
23615
|
0.00
|
3208
|
1003939
|
3.20
|
0
|
2016
|
112
|
23978
|
4.67
|
20821
|
1019515
|
20.42
|
*0.23
|
2017
|
140
|
24348
|
5.75
|
61747
|
1035140
|
59.65
|
*0.10
|
2018
|
39
|
24723
|
1.58
|
67992
|
1050722
|
64.70
|
*0.02
|
Est. Pop: Estimated population. API: Annual parasite index. i: Indigenous people. Non-i: Non-Indigenous people (cases per 1,000 people). RR: Risk ratio indigenous/non-indigenous. Z test for significance: * p <0.05
Spatial autocorrelation and clustering
The origin of malaria cases registered in 2014 (2016 for Sucre) and 2018 was mapped to compare changes in geographic clustering. Maps were constructed to the parish level (fourth administrative level) in Bolivar and Sucre, and to the municipality level (third administrative level) in Amazonas, due to the lack of adequate parish-level data in this state.
The Atures municipality accounted for most cases in Amazonas (44.1% in 2014, and 67.5% in 2018). Yet, Moran’s I showed no significant clustering either year. The Getis-Ord analysis, however, revealed a low-significance hot spot (confidence 90%) in Atures (Figures 3A, and 3B).
In Bolivar, the Moran’s I was positive and statistically significant in 2014 (I=0.19, p=0.006) and 2018 (I=0.17, p=0.006) indicating stable clustering of cases. This was reflected in the hot spot analysis: In 2014, two highly significant (confidence 99%) hot spots were seen in Aripao (Sucre municipality) and Barceloneta (Angostura municipality). Lower significance ones (confidence 90%) were seen in Guaniamo (Cedeño), Gran Sabana and Ikabaru (Gran Sabana municipality, figure 3C and 3D). New high-significance clusters appeared in the entire Gran Sabana municipality in 2018. The central northern region of the state remained a high-significance cold spot during the entire study period.
The situation in Sucre state was only compared to 2016, due to the absence of cases the two previous years. No significant clusters were identified via Moran’s I analysis. The Getis-Ord’s Gi* revealed highly significant hot spots in Union in 2016 and 2018 (Benitez municipality), and Romulo Gallegos (Andrés Eloy Blanco municipality) in 2018.
The full results of the Getis-Ord’s analysis can be found in additional file 4.
Proportion of Plasmodium species and risk factors associated to P. falciparum infection
Most infections were caused by P. vivax: 85,124 (73.3%), P. falciparum: 25,201 (21.7%), or both: 5,726 (4.9%). P. malariae was only identified in 46 patients (0.04%), from eight ethnic groups, all of them from Amazonas, and mostly infected in Alto Orinoco (31 cases), and Manapiare (8 cases). The Yanomami accounted for 58.7% of all P. malariae cases. While P. vivax was the dominant species in all the states, the proportion of P. falciparum cases in Amazonas and Bolivar (22% and 21.4%, respectively) was more than twice as that in Sucre state (8.6%, X2= 35.82 df=2, p<0.01).
The proportion of patients with P. falciparum or mixed infection was also higher in two ethnic groups: the Hoti (42.9%) and the Yanomami (41.5.%), compared to the rest of the indigenous population (26.1%, X2=128.30, df=1, p<0.01 and X2=380.83, df=1, p<0.01, respectively).
To evaluate possible reasons behind this difference, the odds ratios (OR) for P. falciparum vs. P. vivax infection were calculated considering ethnic group, and other available variables such as occupation, gender, age group, and state of origin of patients, results are summarised in table 2. P. malariae and mixed infections were excluded from the analysis.
Table 2: Univariate and multivariable analysis of the odds ratios for P. falciparum infection among indigenous groups of Amazonas, Bolivar and Sucre states, Venezuela, 2014-2018
|
|
|
Univariate
|
|
|
|
Multivariable
|
|
|
Occupation
|
Pv
|
Pf
|
OR (Pf/Pv)
|
CI
|
|
p value
|
OR (Pf/Pv)
|
CI
|
|
p value
|
Mining
|
7763
|
2656
|
1.44
|
1.35
|
1.54
|
<0.01
|
1.09
|
1.01
|
1.18
|
0.03
|
Agriculture
|
2951
|
886
|
1.26
|
1.15
|
1.39
|
<0.01
|
0.82
|
0.74
|
0.91
|
<0.01
|
Commerce
|
8371
|
2873
|
1.45
|
1.35
|
1.55
|
<0.01
|
|
|
|
|
Student
|
8391
|
2257
|
1.13
|
1.06
|
1.21
|
<0.01
|
|
|
|
|
Others
|
7268
|
1726
|
1.00
|
|
|
|
|
|
|
|
Age group
|
|
|
|
|
|
|
|
|
|
|
5-14
|
23101
|
5619
|
1.02
|
0.97
|
1.07
|
0.50
|
1.08
|
1.02
|
1.15
|
0.01
|
15-29
|
28739
|
8893
|
1.30
|
1.23
|
1.36
|
<0.01
|
1.47
|
1.39
|
1.56
|
<0.01
|
30-64
|
21127
|
7565
|
1.50
|
1.43
|
1.58
|
<0.01
|
1.75
|
1.64
|
1.85
|
<0.01
|
65+
|
1307
|
550
|
1.76
|
1.58
|
1.96
|
<0.01
|
2.07
|
1.85
|
2.32
|
<0.01
|
0-4
|
10424
|
2489
|
1.00
|
|
|
|
|
|
|
|
Gender
|
|
|
|
|
|
|
|
|
|
|
Female
|
39443
|
11876
|
1.03
|
1.00
|
1.06
|
0.03
|
1.04
|
1.01
|
1.08
|
0.01
|
Male
|
45681
|
13325
|
1.00
|
|
|
|
|
|
|
|
Ethnic group
|
|
|
|
|
|
|
|
|
|
Pemon
|
22002
|
6633
|
1.19
|
1.12
|
1.26
|
<0.01
|
1.17
|
1.09
|
1.25
|
<0.01
|
Jivi
|
21332
|
6701
|
1.24
|
1.17
|
1.32
|
<0.01
|
1.29
|
1.22
|
1.37
|
<0.01
|
Yekuana
|
7804
|
2328
|
1.18
|
1.10
|
1.26
|
<0.01
|
1.21
|
1.13
|
1.30
|
<0.01
|
Wotjuja
|
14640
|
3429
|
0.93
|
0.87
|
0.99
|
0.02
|
|
|
|
|
Kurripaco
|
3591
|
812
|
0.89
|
0.82
|
0.98
|
0.02
|
0.87
|
0.79
|
0.96
|
<0.01
|
Eñepa
|
2335
|
946
|
1.60
|
1.46
|
1.75
|
<0.01
|
1.76
|
1.59
|
1.94
|
<0.01
|
Piapoko
|
2500
|
712
|
1.13
|
1.02
|
1.24
|
0.02
|
1.20
|
1.09
|
1.33
|
<0.01
|
Yanomami
|
1875
|
1248
|
2.63
|
2.41
|
2.87
|
<0.01
|
3.08
|
2.81
|
3.38
|
<0.01
|
Hoti
|
538
|
389
|
2.86
|
2.48
|
3.29
|
<0.01
|
3.33
|
2.89
|
3.84
|
<0.01
|
Kubeo
|
25
|
3
|
0.47
|
0.14
|
1.57
|
0.22
|
|
|
|
|
Warao
|
325
|
48
|
0.58
|
0.43
|
0.79
|
<0.01
|
|
|
|
|
Warekena
|
255
|
44
|
0.68
|
0.49
|
0.94
|
0.02
|
0.67
|
0.48
|
0.93
|
0.02
|
Akawayo
|
791
|
108
|
0.54
|
0.44
|
0.66
|
<0.01
|
0.46
|
0.38
|
0.57
|
<0.01
|
Other
|
7111
|
1800
|
1.00
|
|
|
|
|
|
|
|
State
|
|
|
|
|
|
|
|
|
|
|
Amazonas
|
48894
|
14452
|
2.92
|
1.93
|
4.41
|
<0.01
|
2.52
|
1.49
|
4.29
|
<0.01
|
Bolivar
|
35983
|
10724
|
2.94
|
1.95
|
4.45
|
<0.01
|
2.31
|
1.38
|
3.87
|
<0.01
|
Sucre
|
247
|
25
|
1.00
|
|
|
|
|
|
|
|
P.f: Plasmodium falciparum, P.v: Plasmodium vivax. OR: Odd ratios calculated via binomial logistic regression, CI: 95% confidence intervals, p values calculated via Fisher’s exact test