1. Ecology
  2. Epidemiology and Global Health

Landscape drives zoonotic malaria prevalence in non-human primates

  1. Emilia Johnson  Is a corresponding author
  2. Reuben Sunil Kumar Sharma
  3. Pablo Ruiz Cuenca
  4. Isabel Byrne
  5. Milena Salgado-Lynn
  6. Zarith Suraya Shahar
  7. Lee Col Lin
  8. Norhadila Zulkifli
  9. Nor Dilaila Mohd Saidi
  10. Chris Drakeley
  11. Jason Matthiopoulos
  12. Luca Nelli
  13. Kimberly Fornace
  1. School of Biodiversity, One Health and Veterinary Medicine, University of Glasgow, United Kingdom
  2. Department of Disease Control, London School of Hygiene & Tropical Medicine, United Kingdom
  3. Centre on Climate Change and Planetary Health, London School of Hygiene & Tropical Medicine, United Kingdom
  4. Faculty of Veterinary Medicine, Universiti Putra Malaysia, Malaysia
  5. Lancaster University, Bailrigg, United Kingdom
  6. Liverpool School of Tropical Medicine, Pembroke Place Liverpool, United Kingdom
  7. School of Biosciences, Cardiff University, United Kingdom
  8. Wildlife Health, Genetic and Forensic Laboratory, Sabah Wildlife Department, Wisma Muis, Malaysia
  9. Danau Girang Field Centre, Sabah Wildlife Department, Malaysia
  10. Department of Infection Biology, London School of Hygiene & Tropical Medicine, United Kingdom
  11. Saw Swee Hock School of Public Health, National University of Singapore, Singapore
https://doi.org/10.7554/eLife.88616.4
Share this article
Cite this article
  1. Emilia Johnson
  2. Reuben Sunil Kumar Sharma
  3. Pablo Ruiz Cuenca
  4. Isabel Byrne
  5. Milena Salgado-Lynn
  6. Zarith Suraya Shahar
  7. Lee Col Lin
  8. Norhadila Zulkifli
  9. Nor Dilaila Mohd Saidi
  10. Chris Drakeley
  11. Jason Matthiopoulos
  12. Luca Nelli
  13. Kimberly Fornace
(2024)
Landscape drives zoonotic malaria prevalence in non-human primates
eLife 12:RP88616.
https://doi.org/10.7554/eLife.88616.4
  2. Download BibTeX
  3. Download .RIS
20 figures, 20 tables and 1 additional file

Figures

Figure 1
Download asset Open asset
Sampling sites and primate species sampled across Southeast Asia.

‘Other’ includes Trachypithecus obscurus and undefined species from the genus Presbytis. Total surveys (n) = 148.

Figure 2
Download asset Open asset
Random-effects meta-analysis of P. knowlesi prevalence across Southeast Asia.

(A) Forest plot of pooled estimates for P. knowlesi prevalence (%) in all non-human primates tested (n=6322) across Southeast Asia, disaggregated by species and sampling site (k=148). Random-effects meta-analysis sub-grouped by region, with 95% confidence intervals and prediction intervals. (B) Map of regional prevalence estimates for P. knowlesi prevalence in NHP in Southeast Asia from meta-analysis. Point colour denotes pooled estimate (%). Size denotes total primates tested per region (n). Shading indicates data availability.

Figure 3
Download asset Open asset
Multivariable regression results.

Spatial scale denoted in square bracket. Canopy cover = %. Adjusted odds ratios (OR, dots) and 95% confidence intervals (CI95%, whiskers) for factors associated with P. knowlesi in NHPs at significant spatial scales. N=1354, accounting for replicate pseudo-sampling.

Appendix 1—figure 1
Download asset Open asset
Search strategy for background research.
Appendix 1—figure 2
Download asset Open asset
WHO Report primate data extraction form.
Appendix 1—figure 3
Download asset Open asset
Flow chart illustrating study selection process.
Appendix 3—figure 1
Download asset Open asset
Assessment of small study effects in meta-analysis.

(A) Funnel plot of transformed prevalence (%) against standard error (SE) for study sites (B) Funnel plot with imputed data to illustrate asymmetry using trim-and-fill method.

Appendix 3—figure 2
Download asset Open asset
Forest plot of P. knowlesi prevalence (%) in all species of NHPs in Southeast Asia, disaggregated by species and sampling site, including 95% confidence intervals and individual study weighting.

Random-effects analysis, sub-grouped by region. N=148. Zarith et al. refers to personal correspondence derived from the reference Shahar, 2019.

Appendix 4—figure 1
Download asset Open asset
Recent forest loss in Peninsular Malaysia (first row) and Malaysian Borneo (secoond row), shown for the years (A) 2006 (B) 2012 and (C) 2019.
Appendix 4—figure 2
Download asset Open asset
Flowchart of data processing.

Details of pseudo-sampling and environmental covariate extraction at multiple spatial scales to create final weighted dataset (N=1354).

Appendix 4—figure 3
Download asset Open asset
Example covariate resolutions in Peninsular Malaysia.

(A) Data point and (B) 20 km buffer over population density layer, 1 km resolution. (C) Data point and (D) 20 km buffer over SRTM elevation layer, 1 km resolution.

Appendix 4—figure 4
Download asset Open asset
Boxplots showing distribution and interquartile range (IQR) of proportional forest cover (0–1) for sampling sites within 5, 10 and 20 km circular buffers across all sites (N=1354).
Appendix 4—figure 5
Download asset Open asset
Examples of buffer zones around macaque sample sites.

Shown over forest cover for 2019 (Hansen et al., 2013).

Appendix 5—figure 1
Download asset Open asset
Sensitivity analysis comparing centroid forest cover to 10 randomly generated points, shown per radius for the largest polygon at each GADM level.
Appendix 5—figure 2
Download asset Open asset
Boxplots of standard deviation in repeat sampling of covariates at multiple buffer and boundary sizes.

Standard deviation of environmental covariates across 10 sampling site realisations within 5/10/20km buffers, grouped by administrative boundary size or GPS coordinates. Shown for (A) canopy cover (%) (B) forest fragmentation (P: A ratio) and (C) human population density (p/km2).

Appendix 6—figure 1
Download asset Open asset
Spearman’s correlation matrix for all candidate covariates at all spatial scales (n=1354).
Appendix 6—figure 2
Download asset Open asset
Spearman’s correlation matrix for covariates at selected spatial scales for final model inclusion (n=1354).

Percentage forest cover (5 km) and forest fragmentation (PARA, 5 km) show strong negative correlation (ρ=–0.75).

Appendix 6—figure 3
Download asset Open asset
Distribution and habitat range of dominant macaque species (M. fascicularis, M. nemestrina, M. leonina) according to predicted probability of occurrence ≥0.5 (on a scale of 0–1.0) per 5x5 km pixel.
Appendix 6—figure 4
Download asset Open asset
Distribution and habitat range of dominant macaque species (M. fascicularis, M. nemestrina, M. leonina) according to predicted probability of occurrence ≥0.75 (on a scale of 0–1.0) per 5x5 km pixel.
Appendix 6—figure 5
Download asset Open asset
Predicted distribution and habitat range of all macaque species (M. fascicularis, M. nemestrina, M. leonina) according to predicted probability of occurrence ≥0.9 (on a scale of 0–1.0) per 5x5 km pixel.

Tables

Table 1
Spatial and temporal resolution (res.) and sources for environmental covariates.

Summary metrics extracted within 5, 10 and 20 km circular buffers.

CovariateSpatial res.Temporal res.Source
Human density (p/km2)1 km2012WorldPop, 2018
Elevation (m)1 km2003SRTM 90 m Digital Elevation v4.1 Jarvis et al., 2008
Tree cover (1/0)*30 mAnnualHansen’s Global Forest Watch Hansen et al., 2013
  1. *

    Derivatives: Proportion canopy cover (%), Perimeter: area ratio (PARA >0)

Appendix 1—table 1
Standardised definitions for qualitative primate characteristics.
VariableCategoryDefinition
SamplingRoutineAnimals collected for surveillance purposes or extracted from human–conflict zones; data collected opportunistically Shearer et al., 2016
StudyAnimals captured and sampled specifically for a study of Plasmodium spp and/or P. knowlesi
StatusCaptiveAnimal resident in sanctuary or conservation park
WildFree-living animal, not registered/resident in any sanctuary
SanctuaryA wildlife sanctuary/rehabilitation centre housing key primate species Gamalo et al., 2019
AreaForestAreas that are uninhabited with extensive tree cover
Peri–domesticAs defined by the author. Example definitions as follows:
AgriculturalAnimal located in agricultural areas, predominantly monoculture (e.g. orchard, plantation) Shahar, 2019
UrbanAs defined by the author. Generally, areas with high human population density Li et al., 2021b
Appendix 1—table 2
Characteristics of the included studies.
AuthorYear(s)Country/regionN*Sample†DiagnosticTarget gene(s)Primer
Lee et al., 20112004–2008Malaysia/Borneo108StudyNested PCRSSU-rRNA/csp/mtDNAKn1f/Kn3r
Seethamchai et al., 20082006Thailand99StudySequencingA-type-SSU-rRNA/cytb•
Vythilingam et al., 20082007Malaysia/Peninsular145StudyPCR/SequencingSSU-rRNA/cspPmk8/Pmkr9
Zhang et al., 20162007Singapore40StudyPCR••
2007–2010Indonesia/Sumatra70StudyPCR••
2011Cambodia54StudyPCR••
2012Philippines68StudyPCR••
2015Laos44StudyPCR/SequencingSSU-rRNAPK18SF/PK18SRc
Jeslyn et al., 20112008Singapore13RoutinePCR/SequencingSSU-rRNA /cspPmk8/Pmkr9
Ho et al., 20102008Malaysia/Peninsular107RoutineNested PCRSSU-rRNAPmk8/Pmkr9
Li et al., 2021b2008–2017Singapore1039RoutineNested PCRSSU-rRNAPmk8/Pmkr9
Putaporntip et al., 20102009Thailand655StudySequencingcytb•
Chang et al., 20112010Myanmar45StudyPCRSSU-rRNA•
Muehlenbein et al., 20152010Malaysia/Borneo41StudyPCRmtDNA/AMA-1/MSP-1•
Shahar, 2019 ‡2010–2017Malaysia/Peninsular1587RoutineNested PCRSSU-rRNA•
§2013Malaysia/Peninsular15StudyPCR••
§2013–2016Malaysia/Borneo25StudyNested PCRcytbPKCBF/PKCBR
Saleh Huddin et al., 20192014Malaysia/Peninsular415StudyPCR/SequencingSSU-rRNAPmk8/Pmkr9
Akter et al., 20152015Malaysia/Peninsular70RoutinePCR/SequencingA-type-SSU-rRNAPmk8/Pmkr9
Amir et al., 20202016Malaysia/Peninsular103RoutineNested PCRSSU-rRNAPkF1140/PkR1550
Gamalo et al., 20192017Philippines95StudyNested PCRSSU-rRNAKn1f/Kn3r
Fungfuang et al., 20202017–2019Thailand93StudyNested PCRSSU-rRNAKn1f/Kn3r
Nada-Raja et al., 20222018Malaysia/Borneo73StudyNested PCRSSU-rRNA/csp/mtDNAKn1f/Kn3r
Yusuf et al., 20222016–2019Malaysia419StudyNested PCRSSU-rRNAKn1f/Kn3r
Zamzuri et al., 20202018Malaysia/Peninsular212RoutinePCR••
Kaewchot et al., 20222019Thailand649StudyNested PCRSSU-rRNAPmk8/Pmkr9
Salwati et al., 20172015Indonesia/Sumatra38StudyPCR/Sequencing••
  1. *

    N=number of primates sampled.

  2. †

    Animal trapped either on routine or study basis.

  3. ‡

    Unpublished, personal correspondence (p/c).

  4. §

    Danau Girang Field Centre, p/c from Dr Salgado Lynn.

Appendix 1—table 3
Published studies of P. knowlesi infections (n) in non-human primate species collected (N) in Southeast Asia, grouped by region and author.
RegionSpeciesTotalRef
M. fascicularisM. nemestrinaM. arctoidesOther
Peninsular25/107•••473/3069Ho et al., 2010
Malaysia48/415•••Saleh Huddin et al., 2019
21/70•••Akter et al., 2015
11/980/5••Amir et al., 2020
0/15•••
10/145•••Vythilingam et al., 2008
215/1587•••Shahar, 2019
66/415•••Yusuf et al., 2022
74/2073/5••Zamzuri et al., 2020
Borneo4/262/15••119/251Muehlenbein et al., 2015
71/8213/26••Lee et al., 2011
18/25•••
7/452/28••Nada-Raja et al., 2022
2/4•••Yusuf et al., 2022
Sumatra0/70•••6/108Zhang et al., 2016
5/321/4•0/2†Salwati et al., 2017
Thailand1/1955/4490/41/7 ‡8/1496Putaporntip et al., 2010
0/21•••Seethamchai et al., 2008
0/4•••Fungfuang et al., 2020
0/320/25*1/32•Fungfuang et al., 2020
0/78•••Seethamchai et al., 2008
0/649•••Kaewchot et al., 2022
Philippines18/95•••18/163Gamalo et al., 2019
0/68•••Zhang et al., 2016
Singapore3/13•••148/1092Jeslyn et al., 2011
145/1039•••Li et al., 2021b
0/40•••Zhang et al., 2016
Laos1/44•••1/44Zhang et al., 2016
Cambodia0/54•••0/54Zhang et al., 2016
Myanmar0/45•••0/45Chang et al., 2011
Total743/572026/5571/361/9773/6322
  1. *

    Macaca-leonina (Northern Pig-tailed macaque, recently classified as separate species).

  2. †

    Presbytis spp.

  3. ‡

    Trachypithecus obscurus (Dusky leaf monkey).

Appendix 1—table 4
Characteristics of primates tested and number/percentage of confirmed P. knowlesi infections (Pk+).
N(%)*Pk+Pk+ (%)†CI95% ‡
SpeciesM. fascicularis5720(90.5%)74513.0%(12.2–13.9)
M. nemestrina532(8.4%)264.9%(3.4–7.1)
M. leonina25(0.4%)00.0%(0.0–13.3)
M. arctoides36(0.6%)12.8%(0.5–14.2)
T. obscurus7(0.1%)114.3%(2.6–51.3)
Presbytis spp.2(0.03%)00.0%(0.0–65.8)
AreaForest1740(27.5%)25314.5%(13.0–16.3)
Agriculture491(7.8%)7214.7%(11.8–18.1)
Peri-domestic2192(34.7%)34115.6%(14.1–17.1)
Urban1143(18.1%)564.9%(3.8–6.3)
Sanctuary109(1.7%)54.6%(2.0–10.3)
Unspecified647(10.2%)467.1%(5.4–9.4)
StatusWild6183(97.8%)76812.4%(11.6–13.3)
Captive139(2.2 %)53.6%(1.5–8.1)
RegionPen. Malaysia3069(48.5%)47315.4%(14.2–16.7)
Borneo251(4.0%)11947.4%(41.3–53.6)
Sumatra108(1.7%)65.5%(2.6–11.6)
Thailand1496(23.7%)80.5%(0.3–1.1)
Philippines163(2.6 %)1811.0%(7.1–16.8)
Singapore1092(17.3%)14813.6%(11.7–15.7)
Cambodia54(0.9%)00.0%(0.0–6.6)
Laos44(0.7%)12.3%(0.4–11.8)
Myanmar45(0.7%)00.0%(0.0–7.9)
Total6322(100%)77312.2%(11.4–13.1)
  1. *

    Percentage of total number of primates tested (column %).

  2. †

    Proportion of N positive for P. knowlesi (row %).

  3. ‡

    95% confidence interval (CI95%) calculated in R using count and sample size (binomial distribution).

Appendix 2—table 1
JBI criteria for assessing bias in meta-analyses of prevalence studies.
CriteriaYesNoUnclearN/A
Q1Was the sample frame appropriate to address the target population?
Q2Were study participants sampled in an appropriate way?
Q3Was the sample size adequate?
Q4Were the study subjects and the setting described in detail?
Q5Was the data analysis conducted with sufficient coverage of the identified sample?
Q6Were valid methods used for the identification of the condition?
Q7Was the condition measured in a standard, reliable way for all participants?
Q8Was there appropriate statistical analysis?
Q9Was the response rate adequate, and if not, was the low response rate managed appropriately?
Appendix 2—table 2
Example rationale for quality appraisal.
Sample questionExampleAssessment
Q1Was the sample frame appropriate to address the targetWild animalYes
population?Captive animalNo
Not specifiedUncertain
Q2Were study participants sampled in an appropriate way?Trapped for studyYes
Routine collectionNo
Not specifiedUncertain
Q9Was the response rate adequate?Primate dataN/A
Appendix 3—table 1
Sensitivity analysis for transformation of P. knowlesi prevalence estimate under random-effects model, shown overall and for Thailand subgroup analysis.
MethodOverall (k=148)Subgroup (Thailand, k=21)
P*CI95%PCI95%
Freeman-Turkey double arcsine0.0943(0.0641–0.1284)0.0000(0.0000–0.0000)
Logit0.1199(0.0935–0.1526)0.0199(0.0113–0.0346)
Untransformed0.1415(0.1101–0.1730)0.0022(0.0000–0.0059)
  1. *

    Estimated proportion

Appendix 4—table 1
Environmental covariates assembled for regression analysis.

Summary values extracted for each covariate within 5, 10, and 20 km circular buffers during processing.

CovariateDescriptionMetricResolutionProcessingSource
SpatialTemporal
PopulationUN-adjusted gridded posterior population model estimates at 30 arc-seconds resolutionPerson count/1 km21 km2000 2012 2019Population density reclassified as high/low (≤300 persons/km2) in QGISWorldPop, 2018 Downloaded as tiff files per country in AOI for years 2000/2012/2019
ElevationMean height above sea levelm1 km2003Mean and SD of continuous elevation per radii. Mean-centred and scaled. Categorised into discrete classifications: low (≤200 m), moderate (200–500 m) or high elevation (>500 m)NASA SRTM 90 m Digital Elevation Database v4.1 (CGIAR-CSI) Jarvis et al., 2008. Downloaded as a tiff file at 1 km resampled resolution
ForestPercentage canopy cover per grid cell. Derived from tree cover (vegetation >5 m) and loss (forested to non-forested)0–130 mAnnual 2006–2020Tree cover classified as ≥50% crown density per raster cell, generating binary raster (1=forest, 0=non-forest). Annual cover calculated by subtracting cumulative loss per year 2006–2019. Data records matched to reclassified tile by geolocation and year. Posterior proportions categorised as high (>50%) medium (20–50%) or low (≤20%)Hansen’s Global Forest Watch, 30 m resolution Landsat imagery Hansen et al., 2013. Tiles downloaded as tiff files for each year 2006–2019 to cover AOI
Fragmentation (perimeter: area ratio, PARA)Perimeter length (m) to patch area (m2) ratio for contiguous forest cover McGarigal et al., 2021 within bufferPARA >030 mAnnual 2006–2020Extracted from annual reclassified tree cover rasters within 5, 10 and 20 km circular buffers Output categorised into quartilesHansen’s Global Forest Watch Hansen et al., 2013 (as above)
Appendix 4—table 2
Summary of forest cover data (N=1480).
MeanSDRange
Forest cover (5 km)50.20%±29.29%0.00–100.00%
Forest cover (10 km)49.68%±27.30%0.00–99.96%
Forest cover (20 km)48.29%±25.35%0.00–99.64%
Total1480 (100%)
Appendix 5—table 1
Geo-positioning of available primate survey data.
Resolution/GADM*Records/nPrimates/NMin. (km2)†Max. (km2)
PolygonCountry/GID06 (4.9%)853 (17.3%)70077,650
State/GID140 (22.0%)2699 (32.2%)13087,860
District/GID288 (61.8%)2433 (43.6%)27015,890
PointGCS ‡14 (11.4%)337 (6.8%)––
Total148 (100%)4931 (100%)
  1. *

    Administrative boundaries, as classified by GADM (v3.6).

  2. †

    Minimum and maximum size (km2) of polygons containing P. knowlesi data at each admin level.

  3. ‡

    Geographic Coordinate System.

Appendix 6—table 1
Bivariable analysis for P. knowlesi in NHP against all covariates at all spatial scales (N=1354).
Bivariable analysis
VariableCrude ORCI95%p value†
Elevation (m) *
≤5 km1.18(1.07–1.28)0.000562
≤10 km1.20(1.09–1.31)0.0001246
≤20 km1.22(1.11–1.33)7.23E-05
Human density (p/km2) *
≤5 km0.84(0.77–0.92)4.72E-05
≤10 km0.75(0.68–0.82)2.70E-12
≤20 km0.71(0.63–0.79)1.37E-12
Forest cover (%) *
≤5 km1.34(1.21–1.49)1.86E-08
≤10 km1.41(1.26–1.57)6.51E-10
≤20 km1.47(1.30–1.67)8.66E-10
Fragmentation (PARA) *
≤5 km0.85(0.76–0.95)0.003944
≤10 km0.69(0.60–0.80)1.80E-07
≤20 km0.67(0.57–0.79)5.14E-07
PARA2 *Quadratic term
≤5 km0.69(0.60–0.80)0.102.06E-06
≤10 km0.64(0.55–0.74)0.082.65E-08
≤20 km0.67(0.57–0.78)0.032.78E-06
Host species
OtherRef
M. fascicularis2.37(1.25–4.60)0.007971
  1. *

    Continuous variable, mean-centred and scaled.

  2. †

    p value derived from Likelihood ratio test (LRT).

Appendix 6—table 2
Multivariable binomial regression analysis of P. knowlesi prevalence in NHP with environmental covariates at influential spatial scales, full dataset (N=1354).

AIC = 1229.8.

Multivariable analysis
VariableRadiusaOR *CI95%p value †
Human density (p/km2) ‡
≤5km1.36(1.16–1.58)1.082E-04
≤20 km0.56(0.46–0.67)1.311E-10
Forest cover (%) ‡
≤5 km1.38(1.19–1.60)2.046E-05
Fragmentation (PARA) ‡
≤5 km1.17(1.02–1.34)0.0281
Host species
OtherRef
M. fascicularis2.50(1.31–4.85)0.005121
  1. *

    Odds Ratios adjusted for all other variables in the table (aOR). Radius calculated as distance from sample point.

  2. †

    p value derived from Likelihood ratio test (LRT).

  3. ‡

    Continuous variable, mean-centred and scaled. OR shown per 1 SD increase.

Appendix 6—table 3
Admin boundary sensitivity analysis.

Binomial regression analysis of P. knowlesi prevalence in NHP for datapoints assigned to GPS or small sized administrative boundaries (excluding country data) (N=1324).

AIC = 1221.9Multivariable analysis
aORCI 95%p value (Wald test)*
Human density [5 km]1.36(1.16–1.58)***
Human density [20 km]0.56(0.46–0.67)***
Forest cover (%) [5 km]1.38(1.19–1.60)***
Fragmentation (PARA) [5 km]1.18(1.02–1.34)*
Host group Other
M. fascicularis		REF
2.51		(1.–.31–4.85)**
  1. *

    Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

Appendix 6—table 4
Distribution of standard deviations across 10 environmental covariates per prevalence data point for landscape variables at all spatial scales (N=1354).
CovariateMeanRangeMedianIQR
Canopy [5 km]*0.15880.0000–0.42370.15340.1039–0.2277
Canopy [10 km]0.13190.0000–0.41240.11770.0827–0.1891
Canopy [20 km]0.10510.0000–0.39990.09210.0541–0.1635
Fragmentation [5 km]*0.00830.0000–0.03750.00630.0041–0.0094
Fragmentation [10 km]0.00610.0000–0.04550.00430.0025–0.0071
Fragmentation [20 km]0.00410.0000–0.03160.00270.0017–0.0047
  1. *

    Spatial scales selected in final variables.

Appendix 6—table 5
Tree canopy cover sensitivity analysis.

Binomial regression of P. knowlesi prevalence in NHP for datapoints, with data where SD < ½ the maximum for tree canopy within 5 km (N=814).

AIC = 771.1Multivariable analysis
aORCI 95%p value (Wald test)*
Human density [5 km]0.90(0.67–1.20)-
Human density [20 km]0.72(0.50–1.01).
Forest cover (%) [5 km]1.70(1.30–2.24)***
Fragmentation (PARA) [5 km]1.38(1.01–1.88)*
Host group Other
M. fascicularis		REF
2.63		(1.35–5.21)**
  1. *

    Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘-’ 1.

Appendix 6—table 6
Landscape fragmentation sensitivity analysis.

Binomial regression of P. knowlesi prevalence in NHP for datapoints, with datapoints where SD < ½ the maximum for fragmentation within 5 km (N=1134).

AIC = 982.9Multivariable analysis
aORCI 95%p value (Wald test)*
Human density [5 km]0.91(0.69–1.18)-
Human density [20 km]0.69(0.52–0.93)*
Forest cover (%) [5 km]1.31(1.08–1.60)**
Fragmentation (PARA) [5 km]1.18(0.90–1.54)-
Host group Other
M. fascicularis		REF
2.52		(1.33–4.87)**
  1. *

    Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

Appendix 6—table 7
Macaque habitat suitability sensitivity analysis.

Binomial regression of P. knowlesi prevalence in NHP for datapoints, including only datapoints with 5 km buffers that intersect with areas with ≥0.5 probability of predicted macaque occurrence (N=1331).

AIC = 1197.2Multivariable analysis
aORCI 95%p value (Wald test)*
Human density [5 km]1.32(1.13–1.54)***
Human density [20 km]0.55(0.45–0.66)***
Forest cover (%) [5 km]1.30(1.12–1.52)***
Fragmentation (PARA) [5 km]1.12(0.97–1.29)-
Host group Other
M. fascicularis		REF
2.48		(1.31–4.82)**
  1. *

    Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

Appendix 6—table 8
Macaque habitat suitability sensitivity analysis.

Binomial regression of P. knowlesi prevalence in NHP for datapoints, including only datapoints with 5 km buffers that intersect with areas with ≥0.75 probability of predicted macaque occurrence (N=1177).

AIC = 1115.5Multivariable analysis
aORCI 95%p value (Wald test)*
Human density [5 km]1.34(1.14–1.58)***
Human density [20 km]0.57(0.47–0.69)***
Forest cover (%) [5 km]1.23(1.04–1.47)*
Fragmentation (PARA) [5 km]1.04(0.86–1.24)-
Host group Other
M. fascicularis		REF
2.69		(1.38–5.38)**
  1. *

    Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

Appendix 6—table 9
Macaque habitat suitability sensitivity analysis.

Binomial regression of P. knowlesi prevalence in NHP for datapoints, including only datapoints with 5 km buffers that intersect with areas with ≥0.9 probability of predicted macaque occurrence (N=567).

AIC = 685.2Multivariable analysis
aORCI 95%p value (Wald test)*
Human density [5 km]1.86(1.49–2.32)***
Human density [20 km]0.36(0.26–0.49)***
Forest cover (%) [5 km]1.47(1.14–1.90)**
Fragmentation (PARA) [5 km]1.35(1.02–1.77)*
Host group Other
M. fascicularis		REF
3.13		(1.50–6.75)**
  1. *

    Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

Additional files

MDAR checklist
https://cdn.elifesciences.org/articles/88616/elife-88616-mdarchecklist1-v1.docx

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Emilia Johnson
  2. Reuben Sunil Kumar Sharma
  3. Pablo Ruiz Cuenca
  4. Isabel Byrne
  5. Milena Salgado-Lynn
  6. Zarith Suraya Shahar
  7. Lee Col Lin
  8. Norhadila Zulkifli
  9. Nor Dilaila Mohd Saidi
  10. Chris Drakeley
  11. Jason Matthiopoulos
  12. Luca Nelli
  13. Kimberly Fornace
(2024)
Landscape drives zoonotic malaria prevalence in non-human primates
eLife 12:RP88616.
https://doi.org/10.7554/eLife.88616.4