Skip to main content

Table 2 Studies from Sri Lanka and other non-Mesoamerican countries assessing the role of pesticides in CKD

From: Pesticide exposures and chronic kidney disease of unknown etiology: an epidemiologic review

Reference & country Study design Population Exposure assessment Case definition/outcome(s) Main findings Pesticide association
Validity and explanation valuea
Sri Lanka
Peiris-John et al., 2006 [87]
Sri Lanka
Cross-sectional 4 groups: 23 OP-exposed farmers with chronic renal failure (CRF) vs 18 unexposed patients with CRF vs 239 OP-exposed farmers without CRF vs 50 unexposed fishermen without CRF Red blood cell acetyl cholinesterase (AChE) levels (U/g) as proxy of organophosphate exposures CRF (not further specified) Significant differences in AChE levels: exposed CRF (18.6 U/g) < unexposed CRF (26.6) < exposed non-CRF (29.1) < non-exposed non-CRF (32.6) Possible association between long-term low-level OP-exposures, cholinesterase levels and CKD
Exploratory aim with unconventional cross-sectional design; inadequate selection of study participants; high risk of bias from exposure misclassification; high risk of confounding
Explanation value: low
Wanigasuriya et al., 2007 [36]
Sri Lanka
Hospital- based case – control (prevalent cases) 183 CKDu cases (136 M, 47 F), 200 controls among HT and DM patients (139 M, 61 F), age 36-67 Questionnaire:
Farmer yes/no
Pesticide exposure yes/no
Drinking water source (well-water home, well-water field, pipe born)
SCr > 2 mg/dL Bivariate analyses:
OR farmer = 4.68 [2.50- 8.82)
OR pesticides = 2.94 [1.73-5.01]
OR drinking well-water field = 1.72 [0.92-3.22]
OR farmer = 1.28 [0.55-2.99)
OR pesticides: 0 cases
OR drinking well-water home = 4.24 [1.51-12.32]
Multivariate logistic regressions: NO associations for farming, pesticide use and drinking well-water
No associations in multivariate analyses with farming, pesticide use and well-water
Prevalent cases; high risk of bias from exposure misclassification; inadequate reporting of statistical analyses and pesticide results
Explanation value: low
Athuraliya et al., 2011 [19]
Sri Lanka
Cross-sectional population-based survey with case –control analyses 6153 (2889 M, 3264 F): age >19

CKDu endemic area Medawachchiya 2600
Two non-endemica areas Yatinuwara708
Hambantota 2844

109 CKDu patients in Medawachchiya (66 M, 43 F)
-Farmer yes/no
-Spraying or handling agrochemicals yes/no
Proteinuric chronic kidney disease Entire study population:
Adj OR farmer 2.6 (1.9–3.4)
Adj OR agrochemical exposure 2.3 (1.4–3.9)
Medawachchiya (CKDu region)
Adj OR farmer 2.1 (1.4–3.3)
Adj OR agrochemical exposure 1.1 (0.7–1.9)

Yatinuwara (non-CKDu region)
Adj OR farmer 1.5 (0.5–3.9)
Adj OR agrochemical exposure 1.6 (0.8–3.2)

Hambantota (non-CKDu region)
Adj OR farmer 1.6 (1.0–2.7)
Adj OR agrochemical exposure 5.6 (2.3–13.2)
Pesticide use was not associated to proteinuric CKD in the CKDu region, but it was associated to CKD of known causes in one of the two non-CKDu regions.
Cross-sectional, crude pesticide exposure assessment, misclassification of disease
Explanation value: medium
Wanigasuriya et al., 2011 [92]
Sri Lanka
Cross-sectional population-based survey with case –control analyses 886 (461 M, 425 F) household members aged ≥18 Questionnaire:
Farmer yes/no
Pesticide spraying yes/no
Drinking water source
Micro-proteinuria Bivariate analyses:
OR farmer = 1.38 [0.71- 2.70)
OR pesticides = 1.01 [0.60-1.72]
OR well-water in the field = 1.79 [1.07-3.01]

Multivariate logistic regression:
OR pesticides = 0.43 [0.21-0.90]
OR well-water in the field = 1.92 [1.04-3.53]
Positive association with drinking from well-water in the field
Negative association with pesticide spraying
Cross-sectional, crude pesticide exposure assessment, misclassification of disease
Explanation value: medium
Jayasumana et al., 2015 [95]
Sri Lanka
Hospital-based case-control (prevalent cases) 125 cases (89 M, 36 F), 180 controls (98 M, 82 F) Questionnaire:
Usual occupation last 10 years, farming yes/no
Use of fertilizer and specific pesticides over last 10 years yes/no
(organophosphates, paraquat, MCPA, glyphosate, bispyribac, carbofuran, mancozeb and other common pesticides)
Glyphosate, metals and hardness measured in water of serving and abandoned wells
CKDu Bivariate logistic regression with significantly increased ORs for farming, use of fertilizers, and use of organophosphates, paraquat, MCPA, glyphosate, bispyribac and mancozeb

Multivariate logistic regression:
OR drinking well water = 2.52 [1.12-5.70]
OR history drinking water from abandoned well = 5.43 [2.88-10.26]
OR pesticide application = 2.34 [0.97- 5.57]
OR use of glyphosate = 5.12 [2.33-11.26]

Water hardness: abandoned wells: very high; serving wells: moderate to hard; reservoir and pipeline: soft
Glyphosate concentration in water from abandoned well significantly higher than in serving wells (median 3.2 μg/L and 0.6 μg/L, respectively).
Positive association with pesticide applications
Positive association with use of glyphosate
Positive association with drinking well-water and, especially, with history of drinking water from abandoned (with hardest water and highest glyphosate levels)
Prevalent cases; relatively good case ascertainment; specific, unquantified pesticide exposure assessment
Exposure response for glyphosate in water; control of potential confounders
Explanation value: high
Other countries
Kamel & El-Minshawy, 2010 [6]
Hospital-based case-control (prevalent cases) 216 ESRD cases (141 M, 75 F) from unknown cause
220 random controls (152 M, 68 F) from other patients
Rural residency yes/no
Drinking unsafe (non-pipe) water yes/no
Farming occupation yes/no
Pesticide exposures by any mean yes/no
ESRD of unknown cause (clinical exams) Bivariate analyses: rural living, drinking unsafe water, being a farmer and pesticide exposure associated with ESRD (p < 0.001)
Multivariate analyses (model not specified):
OR pesticide exposure 2.08 [1.42 – 3.06]
Possible association with pesticide exposures
Prevalent cases; no data to evaluate potential selection bias; crude exposure assessment, statistical methods not well described
Explanation value: low
Siddharth et al., 2012 [75]
Note: this study is an interim report of Siddarth et al., 2014 [76]
Hospital-based case-control (prevalent cases) 150 CKD cases (77 M, 73 F): patients attending nephrology departments
96 controls (51 M, 45 F): staff or persons accompanying CKD patients in the hospital
Age 30-50
Levels of organochlorine (OC) pesticides in blood CKDu: eGFr <60 ml/min/1.73m2 for >3 months
Oxidative stress markers
Significantly higher blood levels in cases for α–HCH, γ-HCH, total HCH, α-endosulfan, β-endosulfan, aldrin, p,p’-DDE, and TPL.
Among cases, adjusted Spearman correlations between eGFR and different pesticide analytes varied between −0.07 and −0.23 (significant for γ-HCH, total HCH and aldrin). When adjusting additionally for levels of other analytes, the association with eGFR remained significant only for aldrin. In addition, significant correlation between eGFR and TPL (r = −0.26).
Association of blood levels of OCs (from environmental exposures) with CKDu, mediated partially through genotype
Prevalent cases; specific and quantitative assessment for non-occupational exposures to OCs, study in a non-CKDu setting; some potential for inverse causation; low risk for confounding
Explanation value: high
Siddarth et al., 2014 [76]
Hospital-based case-control (prevalent cases) 270 cases (140 M, 130 F): patients attending nephrology departments
270 age and sex matched controls: staff or persons accompanying CKD patients in the hospital
Concentrations of organochlorine pesticides in blood
GST genotyping
CKDu: eGFR <90 ml/min/1.73m2 with or without proteinuria, for 3 months Cases had significantly higher blood concentrations of α–HCH, γ-HCH, total HCH, α-endosulfan, β-endosulfan, aldrin, p,p’-DDE, and total pesticides
Significant associations with CKDu for 3rd versus 1st tertile for α-HCH (OR = 2.52), γ-HCH (OR = 2.70), total-HCH (OR = 3.18), aldrin (OR = 3.07), α-endosulfan (OR = 2.99), and β-endosulfan (OR = 3.06). Total pesticides 3rd to 1st tertile OR = 2.73 [(1.46–9.47).
CKDu patients having either one null or two null genotypes tend to accumulate majority of pesticides, whereas in healthy controls only in the subset with both null genotypes for some pesticides.
Lebov et al., 2016 [97]
Cohort (follow-up since 1993-1997) 55,580 licensed pesticide applicators (320 ESRD) Self-administered questionnaires:
Ordinal categories of intensity-weighted lifetime days for 39 specific pesticides
Pesticide exposure resulting in medical visit or hospitalization
Diagnosed pesticide poisoning
High level pesticide exposure event
ESRD Significantly increased HR for highest category of use vs non-users and significant exposure-response trends:
Alachlor HR = 1.51 [1.08-2.13], p for trend 0.015
Atrazine HR = 1.52 [1.11-2.09], p for trend 0.008
Metolachlor HR = 1.53 [1.08-2.13], p for trend 0.008
Paraquat HR = 2.15 [1.11-4.15], p for trend 0.016
Pendimethalin HR = 2.13 [1.20-3.78], p for trend 0.006
Permethrin HR = 2.00 [1.08-3.68], p for trend 0.031
More than one medical visit due to pesticide use HR = 2.13 [1.17 - 3.89], p for trend for increasing number of doctor visits 0.04.
Hospitalization due to pesticide use HR = 3.05 [1.67 to 5.58]
Association between use of specific pesticides and ESRD
Association between ESRD and exposures resulting in medical visits or hospitalization and ESRD
Large cohort with long follow-up; study in non-CKDu endemic regions; specific and quantitative exposure assessment; multiple comparisons; low risk for confounding
Explanation value: high
Lebov et al., 2015 [96]
Cohort (follow-up since 1993-1997) 31,142 wives of licensed pesticide applicators (98 ESRD) Self-administered questionnaires or telephone interview
-direct exposures (n = 17,425): ordinal categories of intensity weighted lifetime use of any pesticide, 10 specific pesticides and 6 chemical classes
-Indirect pesticide exposures (husband’s pesticide use) among wives without personal use (n = 13,717)
-Indicators of residential pesticide exposure
ERSD Highest category of cumulative lifetime-days of pesticide use in general vs never personal use: HR 4.22 [1.26-14.2]
Exposure-response trends for husband’s use of paraquat HR 1.99 [1.14-3.47] and butylate HR 1.71 [1.00-2.95]
No excess risk for indicators of residential exposures
Association between direct general pesticide use and husband’s use of paraquat and ESRD in women
No associations with residential exposures
Large cohort with long follow-up; study in non-CKDu endemic regions; specific and quantitative exposure assessment; multiple comparisons; low risk for confounding
Explanation value: high
Aroonvilairat et al., 2015 [98]
Cross-sectional 64 workers of orchids (30 M, 34 F) and 60 controls (33 M, 27 F) Mixing and spraying pesticides during work at orchard for at least three months Difference in BUN and SCr BUN (mg/dL) exposed 12.64 ± 3.7 (3.7% abnormal) vs BUN unexposed 12.43 ± 2.9 (1.7% abnormal), p = 0.76
SCr (mg/dL) exposed females 0.86 ± 0.11 (3.7% abnormal) vs unexposed females 0.82 ± 0.11 (2.9% abnormal), p = 0.11
SCr exposed males 1.09 ± 0.11 (0% abnormal) vs unexposed males 1.09 ± 0.10 (0% abnormal), p = 0.95
No association between occupation in highly pesticide exposed farming and decreased kidney function
Cross-sectional; crude exposure assessment, selection of study population not well described; no confounding adjustment
Explanation value: low
  1. Abbreviations: AChE red blood cell acetylcholinesterase, ACR albumin to creatinine ratio, ANOVA analysis of variance, CKD chronic kidney disease (u, of unknown etiology), BUN blood urea nitrogen, CRF chronic renal failure, DB diabetes, DW drinking water, DDE dichlorodiphenyldichloroethylene, eGFR glomerular filtration rate, ESRD end-stage renal disease, F female, GST glutathione-S-transferase, HCH hexachlorocyclohexane, HT hypertension, M male, MVLR multivariate logistic regression, OP organophosphate pesticides, SCr serum creatinine
  2. aExplanation value: The study’s ability to address potential associations between pesticdes and CKD or CKDu. For details see Additional file 2: Table S1 and the main text