The demographic data of the volunteers of urine samples (Table 2)
Overall, 15 spot urine samples were collected from 15 CKD patients previously diagnosed by doctors, and resided in CKDu-epidemic areas (local prevalence more than 10%) in Wilugamuwa and Anuradhapura. We could not confirm through a medical exam, whether the CKD patients were CKDu or not. Seventy-seven urine samples were collected from non-CKD participants including 15 CKD patients’ family members (CKD family members) and 62 healthy individuals (neighbors). Most of the CKD patients were male, 75 %, (in non-CKD participants 35.1%, p<0.001, Chi-square test) and the age was older, 54.9±13.1 years old (non-CKD participants 40.5±17.7 years old (mean±SE), p=0.009, t-test). 62 neighbors resided in Wilugamuwa or Anuradhapura, but whether their working place was CKDu-epidemic areas or not was not confirmed.
The status of pesticides applied onto the rice paddies
In summary, we could not find any concrete evidence of a large scale application of neonicotinoid insecticides on rice paddies in the studied area in May 2015 (Yala season) or in December 2015 (Maha season). In both season, rice seed sowing seems to be performed and pesticides applied to rice paddies, because significant crops were reported in Anuradhapura and Matale district by the national record [98, 99]. Whether any pesticide was applied to rice paddies or not was not confirmed. Unfortunately, we could not obtain a list of the registered pesticides in Wilgamuwa or in Anuradhapura. We could only obtain a list of the registered pesticides in other CKDu-epidemic districts, Kandy and in Negombo (Supplemental Table1). Among the category of neonicotinoids, imidacloprid was registered in Kandy, while imidacloprid and thiamethoxam in Negombo.
An interview with a clerk in shop A in Mihintale, Anuradapura revealed that glyphosate was commonly used from April to May, MCPA, 3-4 DPA (propanil) and Gulliver (azimslfuron) from June to July, and Avimavar (imidacloprid), Mospilan (acetamiprid) and Marshal (carbobulfan) in August. However, from September to next March no specific pesticide was sold (Supplemental Table 2). A clerk at shop B said “In Anuradhapura District, variations in the pesticide sales by month is not recorded. The reason is both vegetable farmers (throughout the year) and rice farmers buy different pesticides. But the most saleable pesticides are herbicides, such as Kiseki (bispyribac-sodium 40 g/L and metamifop 100g/L), Ceypectco (M.C.P.A), propanil, and Weed Lactor (a.i. is unknown), and organophosphate insecticide profenofos.”
Basic urinary findings
The result of the urine analysis is shown in Table 3. No remarkable difference between 15 CKD patients and 77 non-CKD participants was observed in urinary blood detection (40 %, 52.9%, p=0.41, Chi-square test); acidic dominant pH (5.7±0.3, 5.7±0.6, p=0.51, Chi-square test); low urinary creatinine concentration no more than 0.5 g/L (33.3%, 29.9 %, p=0.92, Chi-square test); low gravity less than 1.005 (55.6%, 20.8%, p=0.31, only performed in May); and UACR, creatinine adjusted albumin no less than 30 mg/g Cre (33.3 %, 3.6%, p=0.06, Chi-square test). Urinary glucose was not detected in CKD patients. Metal/metalloid was analyzed in 63 samples from 10 CKD patients, 11 CKD family members, and 43 neighbors in this study, and no significant difference of urine concentration of aluminum, arsenic, cadmium, chromium, copper, copper, iron, manganese, nickel, lead, selenium, uranium and zinc, between the three clinical categories was observed (one-way ANOVA, Supplemental Table 3).
Tubular biomarkers
Urine Cystatin-C analysis was performed in 92 samples but L-FABP analysis was in only 68 samples, because sample volume was limited. Distribution of Cystatin-C and L-FABP were both gamma distribution. In Cystatin-C, mean±SD was 153±935 μg/g Cre; geometric mean 4.95 μg/g Cre; median 6.1 μg/g Cre; and interquartile range (IQR) 10.0 μg/g Cre. In L-FABP, mean±SD was 7.82±23.5 μg/g Cre; geometric mean 1.08 μg/g Cre; median 1.66 μg/g Cre; and IQR 3.55 μg/g Cre, respectively. Linear correlation between Cystatin-C and L-FABP are shown in Figure 1, (Log10 (L-FABP) = 0.536 x
(Log10 (Cystatin-C) + 0.0193, r=0.71, p=0.005, Spearman’s rank correlation test. Interestingly, in seven cases of non-CKD participants, Cystatin-C concentrations were almost zero with low creatinine and L-FABP levels.
In 15 CKD patients and 77 non-CKD participants, L-FABP more than 8.4 μg/g Cre detected was 38.4 %, 7.3 %, respectively (p=0.19), mean±SD was 26.6±49.3, 3.2±6.1μg/gCre, respectively (p<0.001); Cystatin-C more than 70 μg/g Cre detected was 33.3%, 3.6%, respectively (p=0.097), mean±SD was 818±2175, 12.9±29.2 μg/gCre, respectively (p=0.0013). In 15 CKD family members and 62 neighbors, no significant difference of urine Cystatin-C was observed, mean±SD was 7.3±11.9, 14.3±31.9, respectively (p=0.41, t-test). In 12 CKD family members and 44 neighbors, no significant difference of L-FABP was observed, mean±SD was 1.5±1.5, 3.7±6.8, respectively (p=0.27, t-test).
Urine neonicotinoids and a metabolite analysis
The overall detection rates were highest in DMAP 92.4 %, following dinotefuran and thiamethoxam 17.4 %, clothianidin 9.8%, thiacloprid 3.3%, imidacloprid 2.2%, (Table 4). Dinotefuran and thiacloprid had not registered in 2015 in Sri Lanka. Distribution of neonicotinoids and DMAP concentration was gamma distribution.
Between creatinine-adjusted neonicotinoids concentration and metal/metalloid concentration in urine, significant positive relationship between clothianidin and aluminum (r=0.56, p<0.0001, n=63, Pearson’s correlation coefficient), clothianidin and manganese (r=0.29, p=0.019, n=63), imidacloprid and aluminum (r=0.35, p=0.005, n=63), and imidacloprid and manganese (r=0.25, p=0.047) (Supplemental Table 4) were detected.
There was no significant difference of neonicotinoids and DMAP concentration between 15 CKD patients, 15 CKD family members and 62 neighbors (one-way ANOVA, Table 5). Uncorrected concentration of dinotefuran and imidacloprid was significantly higher in 15 CKD patients than in 77 not CKD participants, but not creatinine adjusted concentration. Thiamethoxam and clothianidin was not detected in CKD patients.
The details of urine neonicotinoids and DMAP concentration in May 2015 and in December are shown in Supplemental Table 5. DMAP was significantly more detected in December 2015 than in May 2015 (Detection rate, in May 81.8 %, in December 98.3 %, p=0.0042, Chi-square test; mean±SD, in May 0.50±0.53 μg/gCre, in December 2.45±4.34 μg/gCre, p=0.012, t-test), but not significant dinotefuran, thiamethoxam, clothianidin, imidacloprid, or thiacloprid.
Between the concentration of urine Cystatin-C and N-desmethyl-acetamiprid (DMAP) concentration, weak negative correlation was observed, i.e. Log10(DMAP)=-0.138*Log10 (Cystatin-C)-0.139, r=-0.19, p=0.077. When all samples divided into two groups, i.e. urine Cystatin-C was less than 6.1 μg/g Cre (Low CysC group, n=46) and no less than 6.1 μg/g Cre (High CysC group, n=46), the mean value of DMAP was significantly lower in High CysC group than in Low CysC group (0.94±1.11 μg/gCre, 2.56±4.87 μg/gCre, respectively, p=0.030, t-test). This trend was also observed in dinotefuran, thiamethoxam, and clothianidin, but not significantly (Table 6).
Neonicotinoids related symptoms
The results are shown in Table 7 and Supplemental Table 6. Total 91 data including 15 data from CKD patients, and 76 data from non-CKD participants (15 CKD family members and 61 neighbors) were examined. A data from a neighbor in May 2015 was lost. The symptoms frequently complained of by 91 participants were recent memory loss with food diary 67.0%, muscle symptoms 60.4%, chest pains or palpitation 57.1%, general fatigue 52.7%, anger 51.6%, headache 49.5%, restlessness 34.1%, auditory hallucination 33.0%, and dizziness after stand up 31.9%.
Symptoms that 15 CKD patients significantly more complained of than 76 non-CKD participants were as follows: p<0.001 was high urine volume 66.7%, appetite loss 60.0 %, reduced body weight 53.3 %, finger tremor 46.7 %, fever 46.7%; p<0.05 was abnormal behavior 13.3% and constipation 13.3%. There was no significant detection rate difference of the symptoms between CKD patients in May and in December. In 15 CKD family members, symptoms significantly more complained of than 61 neighbors were as follows: p<0.05 was muscle symptoms 86.7% and abnormal behavior 6.7%.
As shown in Table 8, when all samples divided into Low CysC group, n=45 and High CysC group, n=46, only agitation and constipation was significantly higher in High CysC group than in Low CysC group (34.8% vs 10.9%, p=0.006; 8.7% vs 0%, p=0.04, Chi-square test). When all samples divided into two groups, i.e. urine DMAP was less than 0.597 μg/g Cre (Low DMAP group, n=45) and no less than 0.597 μg/g Cre (High DMAP group, n=46), only muscle symptoms and auditory hallucination was significantly higher in High DMAP group than in Low DMAP group (67.3% vs 45.7%, p=0.035; 43.4% vs 21.7%, p=0.026, Chi-square test). When all samples divided into two groups, i.e. urine aluminum concentration was less than 2.5μg/g Cre (Low Al group, n=31, mean±SD was 1.21±0.72 μg/g Cre) and no less than 2.5μg/g Cre (High Al group, n=32, 7.48±7.2), no significant difference of symptoms complained of was observed, but when divided into two groups, i.e. urine manganese concentration was less than 0.29 μg/g Cre (Low Mn Group, n=31, mean±SD was 0.09±2.58 μg/g Cre) and no less than 0.29 μg/g Cre (High Mn Group, n=32, 1.54±2.58 μg/g Cre), cough and reduced body weight were significantly more observed in High Mn Group (p=0.040, 0.038, respectively).