Epidemiology of Chronic Kidney Disease in Adults of Salvadoran Agricultural Communities

INTRODUCTION The global epidemic of chronic kidney disease (CKD) and consequent burgeoning demand for renal replacement therapy overwhelm health budgets of developing countries,[1] and only a few countries have economies robust enough to cope with growing CKD challenges.[2] The main traditional CKD causes reported worldwide are diabetes mellitus (DM) (responsible for 30%–40% of cases) and hypertension (HT) (25%–30%), mainly associated with cardiovascular risk factors and unhealthy lifestyles.[2–4] Glomerulonephritis and unknown causes are more frequent in Asia and sub-Saharan Africa than in the developed world.[5] Environmental toxins known to increase CKD risk include chronic exposure to: heavy metals (lead, cadmium, arsenic, mercury and uranium);[6–9] agrochemicals;[6,10–12] nephrotoxic substances such as aristolochic acid, found in starfruit (Averrhoa carambola L., associated with Balkan endemic nephropathy);[5,6,13] and some Chinese medicinal plants.[14,15] Other CKD risk factors are chronic use of drugs such as NSAIDs, nephrotoxic antibiotics ,[16] and sequelae of acute damage from poisoning, hypovo-lemia, obstruction or other causes.[17] An increase in CKD incidence has been described in poor agricultural communities of Egypt, India and Sri Lanka.[8,11,18– 21] Nontraditional risk factors, identifi ed in several studies in these countries, are toxic, environmental and occupational. There appears to be a complex interaction between poor working conditions , notably inadequate handling of agrochemical mixtures— many of which are banned—used in large amounts and without protection; and prolonged, intense physical effort at high temperatures , with insuffi cient hydration.[22,23] In Sri Lanka (2013), WHO recognized that CKD is associated with environmental contaminants, particularly heavy metals (cad-mium and arsenic) contained in agricultural pesticides and fertilizers .[24–26] In Central America and southern Mexico, an alarming CKD increase has been noted in the last two decades.[22,23,27–32] Epidemiological studies have found a high prevalence in agricultural areas, predominantly in men, mainly aged <60 years, exposed to agrochemicals in combination with other risk factors. [28–31] Several efforts have been made in the region to characterize CKD that is not associated with traditional risk factors. Two etiologic hypotheses—both multifactorial but emphasizing different primary triggers—have been posited: one related to heat stress with repeated episodes of rhabdomyolysis and dehydration;[27,32] the other related to toxic exposures at work and in the environment of agricultural communities, coupled with presence of the aforementioned factors (harmful by themselves), which potenti-ate effects of prolonged, intensive use of agrochemicals.[29] CKD constitutes a serious health problem in El Salvador. The Ministry of Health's (MINSAL, the Spanish acronym) annual report for 2011–2012 stated that …


INTRODUCTION
The global epidemic of chronic kidney disease (CKD) and consequent burgeoning demand for renal replacement therapy overwhelm health budgets of developing countries, [1] and only a few countries have economies robust enough to cope with growing CKD challenges.[2] The main traditional CKD causes reported worldwide are diabetes mellitus (DM) (responsible for 30%-40% of cases) and hypertension (HT) (25%-30%), mainly associated with cardiovascular risk factors and unhealthy lifestyles.[2][3][4] Glomerulonephritis and unknown causes are more frequent in Asia and sub-Saharan Africa than in the developed world.[5] Environmental toxins known to increase CKD risk include chronic exposure to: heavy metals (lead, cadmium, arsenic, mercury and uranium); [6][7][8][9] agrochemicals; [6,[10][11][12] nephrotoxic substances such as aristolochic acid, found in starfruit (Averrhoa carambola L., associated with Balkan endemic nephropathy); [5,6,13] and some Chinese medicinal plants.[14,15] Other CKD risk factors are chronic use of drugs such as NSAIDs, nephrotoxic antibiotics, [16] and sequelae of acute damage from poisoning, hypovolemia, obstruction or other causes.[17] An increase in CKD incidence has been described in poor agricultural communities of Egypt, India and Sri Lanka.[8,11,[18][19][20][21] Nontraditional risk factors, identifi ed in several studies in these countries, are toxic, environmental and occupational.There appears to be a complex interaction between poor working conditions, notably inadequate handling of agrochemical mixtures-many of which are banned-used in large amounts and without protection; and prolonged, intense physical effort at high temperatures, with insuffi cient hydration.[22,23] In Sri Lanka (2013), WHO recognized that CKD is associated with environmental contaminants, particularly heavy metals (cadmium and arsenic) contained in agricultural pesticides and fertilizers.[24][25][26] In Central America and southern Mexico, an alarming CKD increase has been noted in the last two decades.[22,23,[27][28][29][30][31][32] Epidemiological studies have found a high prevalence in agricultural areas, predominantly in men, mainly aged <60 years, exposed to agrochemicals in combination with other risk factors.[28][29][30][31] Several efforts have been made in the region to characterize CKD that is not associated with traditional risk factors.Two etiologic hypotheses-both multifactorial but emphasizing different primary triggers-have been posited: one related to heat stress with repeated episodes of rhabdomyolysis and dehydration; [27,32] the other related to toxic exposures at work and in the environment of agricultural communities, coupled with presence of the aforementioned factors (harmful by themselves), which potentiate effects of prolonged, intensive use of agrochemicals.[29] Original Research Peer Reviewed the third cause of hospital death in adults: the fi rst in men and fi fth in women, with a case fatality of 12.6%.[33] In the 1990s, what appeared to be a different kind of CKD began to emerge in Salvadoran agricultural communities.[29][30][31] However, knowledge is incomplete concerning CKD frequency and distribution in the general population, and also about the clinical, physiopathological, anatomopathological and toxicoepidemiological characteristics of the CKD of nontraditional etiology in Salvadoran rural communities.
These knowledge gaps motivated this study, NefroSalva Epidemiológico.Objectives were to determine prevalence of CKD, CKD risk factors (traditional and nontraditional) and renal damage markers in the adult population of specifi c rural areas in El Salvador; measure population distribution of renal function; and identify associated risk factors in CKD patients detected.

METHODS
NefroSalva was carried out by the Renal Health Research Unit of MINSAL's National Health Institute, supported by FOSALUD (El Salvador's Health Solidarity Fund, based on tax revenue from tobacco, alcohol and fi rearms).The research team included Salvadoran doctors, nurses, clinical laboratory technicians, epidemiologists, community health promoters and nephrologists, with collaboration by students from the University of El Salvador Faculty of Medicine and Cuba's Latin American Medical School, and active participation by community health committees in the areas studied.Experts from Cuba's Nephrology Institute served as PAHO advisors to the study.
An analytical cross-sectional study was carried out in 2009-2011, combining epidemiological and clinical elements, including active screening for CKD cases and risk factors in the population aged >18 years in three impoverished regions of El Salvador: Bajo Lempa (Usulután Department), Guayapa Abajo (Ahuachapán Department) and Las Brisas (San Miguel Department).
Agricultural communities studied Bajo Lempa is a rural region on the banks of the Lempa River in eastern El Salvador, [34] an area of elevated CKD prevalence where communities are mainly dependent on agriculture.[29,31] Guayapa Abajo is an agricultural region (primarily sugarcane) in western El Salvador, with no previous epidemiological studies of CKD.Las Brisas in eastern El Salvador comprises suburban communities around San Miguel (capital of the department of the same name), and is the site of a former storage depot for toxaphene, known to have severe toxic effects on the kidney, liver and nervous system (banned by USE-PA since 1990).[35] The abandoned depot was dismantled by Las Brisas residents.According to El Salvador's Ministry of Environment and Natural Resources, toxaphene residues were found in nine of ten wells tested in 2009.[36] Our study was carried out in two phases: Study variables Displayed in Table 1.
Procedures Registration and coding Each patient was assigned a registration code for subsequent clinical monitoring.
Clinical history and physical examination were done to obtain personal information, personal and family medical history, and occupational and behavioral risks; plus physical measurements (weight, height, blood pressure, waist circumference).
Laboratory analyses A fi rst morning urine sample was analyzed with reactive strips and strip-reading equipment, Multistix®10 SG Reagent Strip for Urinalysis (Bayer, USA).A 10-mL fasting venous blood sample was drawn to measure creatinine, glucose, cholesterol and triglycerides.Samples were processed in a laboratory installed in each region and equipped with a Cobas C111 spectrophotometer (Roche, Germany) and corresponding reagents.
Quality control, procedure standardization and data validation All measuring instruments and tools were calibrated to ensure data quality and consistency.Laboratory analyses were performed per manufacturers' specifi cations using appropriate controls.Measurements and analyses were done by trained and certifi ed personnel.
Data analysis Data were stored using Microsoft Excel for later export to SPSS 11.5 for Windows.Prevalence (%) was obtained for CKD, CRF and remaining study variables.
Contingency tables were created for bivariate analysis to explore CKD risk factors and eliminate those exceeding a signifi cance threshold (p <0.05).The independence test, Bartholomew test and chi-square test for regression were used, depending on whether the variable was nominal, ordinal or quantitative.To avoid collinearity in the multiple regression model, associations were fi rst assessed using the chi-square test for qualitative variable independence with a correlation coeffi cient (phi or Cramer's V, depending on whether the tables used 1 df or >1 df, respectively).The Pearson correlation coeffi cient (p) was used to assess associations for quantitative variables, a correlation >0.8 considered signifi cant.
Later, logistic regression was applied to binary response variables to determine probability of developing CKD related to each risk factor, as well as simultaneous estimation of the pure or absolute infl uence of each factor on CKD or CRF risk while controlling for the others.The Enter method was used (which introduces all variables at the same time, instead of step by step).Variables whose coeffi cients were signifi cantly different from 0 (p <0.05), were identifi ed using the Wald Test.Finally, the prevalence odds ratio (POR) was estimated for each variable and by intervals for each variable selected.
Model goodness of fi t was evaluated using the Hosmer-Lemeshow chi-square test.The model was considered to fi t if the probability of association with the statistical test was <0.05.

Ethical considerations
Written informed consent was obtained from all participants, who agreed with publication of their results, ensuring confi dentiality.All patients received clinical followup by health services.

Original Research
Peer Reviewed

RESULTS
CKD-related risk factor prevalences in the study population are displayed in Table 2.Only 14.8% of the population were aged ≥60 years.Family history of CKD was found in 16.5%, of DM in 18.5% and of HT in 30.6%.
Alcohol and tobacco use were more common in men (27.2% and 18.6%, respectively) than in women (6.7% and 3.9%, respectively).Use of NSAIDs was similar in both sexes (84% in men, 84.3% in women), likewise of medicinal plants (50.5% in men, 52.3% in women).Starfruit consumption was not reported.
HT was found in 17.7% of men and 23% of women.DM was not signifi cantly different between men and women (8.5% vs. 9.3%).High prevalence of overweight was reported (28.6% in men, 28% in women).Obesity and central obesity were more frequent in women (26.1% and 30.7%, respectively) than in men (16.1% and 16.5%, respectively) (Table 2).
Renal damage markers and CKD prevalence are displayed in Table 3.These markers were found in 12.5% of the study population, more frequently in men than in women (16.3% vs. 9.8%).The most frequent marker was microalbuminuria (6.9%).Some 5.6% of the sample had glomerular fi ltration rate (GFR) <60 mL/mL/min/ ) in women, for a prevalence ratio of 2.5 for male sex (Table 3).
In the population aged 18-59 years, CKD prevalence was 14.1% (17.6% in men, 11.8% in women).CKD prevalence in those aged ≥60 years was 40.7% (57% in men, 27.6% in women) (Table 3).The largest number of CKD cases was in the younger age group in all three communities.
CKD prevalence in people working in agriculture at the time of the study was 26.8% (31.3% in men and 15.8% in women).
More than half the population in Bajo Lempa and Guayapa Abajo communities worked in agriculture (51.9% and 72.1%, respec-

Original Research
Peer Reviewed tively), but only 13.5% did so in Las Brisas, the area of highest CKD (21.1%) prevalence, followed by Guayapa Abajo (20.5%) and Bajo Lempa (15.4%).In Guayapa Abajo and Bajo Lempa communities, CKD prevalence in men was more than twice that in women; while in Las Brisas, prevalence was similar, 21.5% in women and 20.2% in men (Table 3).
In general, prevalence decreases by CKD stage, but with an upward "bulge" from stage 2 to 3 (Figure 1).Most CKD patients (59.9%) had no DM, HT or proteinuria.
In Las Brisas communities, age ≥60, family history or presence of HT, family history of CKD, tobacco use, family history of DM, and contact with the agrochemicals methyl parathion and methomyl were signifi cantly associated with CKD in univariate analysis.The logistic model revealed that exposure to methyl parathion increased CKD probability by 2.6 times (POR 2.6, 95% CI 1.239-5.448).

DISCUSSION
Our results show a clear vulnerability in these poor agricultural communities to increased risk of CKD development, progression and complications.This is consistent with research on low socioeconomic status and CKD severity in 1657 patients in the United Kingdom, which demonstrated that patients with lower economic status had greater risk of decreased GFR, after adjusting for other sociodemographic, behavioral, and clinical variables.[40] The substantial prevalences we observed of family history of CKD, DM and HT could suggest increased genetic susceptibility to these conditions and/or predisposing environmental factors.They were similar to prevalences reported in other studies.For example, the KEEP-Japan study reported prevalence of family history of these conditions in the general population as 1.2%; 32.7% and 58.7%, respectively.[41] In Latin America, the KEEP-Mexico and KEEP-Jalisco studies reported a 52% prevalence of family history of CKD, and 23% prevalence DM and HT together.[42] We found higher population prevalence of DM than reported in most studies from other countries: USA (1999-2004) 10.3% and (2008) 7%, [39,43] Spain 9.2%, [44] Cuba 5.4%, [45] and Mexico 8%.[46] HT prevalence in this study was lower or similar to that seen elsewhere: USA (1999-2004) 29%, [43] Spain 24.1%; [44] Cuba 30%; [45] and Mexico 7.7%-10% and 31%.[42,46] Prevalence of current or previous tobacco use was close to that found in the Nefrolempa 2009 [29] study and in a rural community study in Mexico.[47] Prevalence of alcohol use was high in men, recalling the hypothesis from a Nicaragua study linking CKD with a homemade alcohol (known as lija).[48] Our study did not determine the most frequently consumed types of alcohol.
The prevalence we found of overweight and obesity (the latter particularly in women) is consistent with international rates reported, [40,[42][43][44][45] and indicates both need and scope for primary prevention, since obesity is associated with vascular and renal damage.

Original Research
The high prevalence of pesticide exposure in our study is not surprising.A 2007-2008 study on pesticide contamination in Bajo Lempa detected dieldrin in shrimp culture ponds at concentrations of 0.085-0.182ng/mL, 1.5 times higher than USEPA limits.Dieldrin was used for cotton crops until its agricultural use was banned in El Salvador in 1986.Six of the ten pesticides reported in our study were found in shrimp culture ponds: heptachloride, endrin, dieldrin and DDT metabolites-DDD and DDE.All of these are known to persist in soil as well.[49] In Nicaragua, increased CKD rates in farmers aged <60 years were associated with pesticide exposure, dehydration, alcohol consumption, exposure to heavy metals and living at low altitudes.[48] Studies in Sri Lanka have found association between bipyridylium (paraquat and diquat) and organophosphate poisoning with repeated episodes of acute renal failure.[10] In a recent study, Jayasumana described a chronic kidney disease of uncertain etiology affecting Sri Lankan agricultural communities, with predominance in men and onset in adolescence.He examined 125 patients and found skin hyperpigmentation on palms of the hands and soles of the feet; and toxicological studies demonstrated arsenic in biological samples.He concluded that the disease represented a new type of nephropathy, probably caused by arsenic poisoning with contributions by heavy metals and pesticide residues, and potentiated by heat stress, chronic repeated dehydration, poor quality drinking water and genetic susceptibility.[18] Also in Sri Lanka, Athuraliya detected histopathological fi ndings demonstrating tubulointerstitial nephritis with or without nonspecifi c interstitial lymphocytic infi ltration associated with glomerular atrophy and glomerular loss.[19] CKD in Sri Lanka is characterized by tubular proteinuria-usually alpha 1 and beta 2 microglobulin and elevated tubular proteinuria due to NGAL (>300 ng/mg creatinine/dL)-without edema.[19] In our study, the relative absence of urinary renal damage markers, even in many patients in advanced CKD stages, strengthens the possibility of primary tubulointerstitial damage rather than glomerular impairment.Together with the upward surge in prevalence at stage 3a, this suggests possible substantial underdetection of stages 1 and 2, in persons with CKD who cannot be diagnosed because screening at early stages depends on renal damage markers.If this is the case, there is an even larger CKD iceberg looming beneath the surface.CKD prevalence in both sexes was higher than rates commonly reported internationally, [41][42][43][44][45][50][51][52][53] but CKD was twice as frequent in men as in women from the earlier stages.The exception was Las Brisas, a community exposed to an agrochemical warehouse.Higher CKD prevalence in men than women in our study in part refl ects the fact that men had high prevalences of both traditional and nontraditional risk factors for vascular and renal damage, the latter including poor working conditions, and contact with agrochemicals-some of which are highly toxic and/or banned by international conventions-without adequate protection.Only in Las Brisas was CKD prevalence similar in the two sexes, and it was the only community where men did not vastly outnumber women as farmworkers.
Reports on the CKD pandemic emphasize the increased burden on health systems with rising numbers of cases because of population aging, since risk increases with age.[2,3] However, in all three communities, the absolute number of CKD cases was higher in the group aged 18-59 years than in those aged ≥60 years: Such early onset and high CKD frequency in the economically active population has disastrous consequences for communities.
In our study, multiple logistic regression did not demonstrate statistically signifi cant association of CKD with DM in any of the areas studied, contrary to the KEEP-Japan study, which reported a statistically signifi cant POR of 1.71 for DM; KEEP-Japan's statistically signifi cant POR of 3.42 for HT was similar to our fi nding.[41] HT prevalence is higher in patients with CKD and varies inversely with GFR; it is both a risk factor for cardiovascular morbidity and mortality and a progression factor for CKD.Some 80%-85% of CKD patients develop HT. [41] In our study, HT prevalence was higher in CKD patients, but the cross-sectional study design precludes establishing temporality, a limitation for interpreting the results.
Most CKD cases were not accompanied by DM, HT, or nephrotic proteinuria.CKD-associated factors revealed, according to the logistic model, that the possibility of developing CKD was strongly infl uenced by male sex, advanced age, agricultural occupation, HT, family history of HT and contact with the agrochemical methyl parathion.This reinforces the proposal of a new CKD with a different epidemiological pattern than hitherto predominating in most of the world, and that corresponds to the nephropathy described in agricultural communities of other Central American countries and Sri Lanka.[25,[27][28][29][30][31][32]48] Accumulating evidence for the hypothesis of nephrotoxic environmental and occupational factors supports the need for more thoroughgoing investigation.
Among the study's limitations is the equation used to estimate kidney function; the MDRD formula is the most widely used in epidemiologic studies, [38,39] but underestimates GFR in young adults and overestimates it in elderly patients.This should be considered when analyzing age-specifi c results.Also, the equation was designed for the US white and Afro-American populations and requires validation for the Salvadoran population.
Nevertheless, the study design is useful to estimate population prevalence, identify possible risk factors, and to generate research

CONCLUSION
The results of this study reinforce the hypothesis emerging from other research suggesting a new nephropathy, which could be called agricultural nephropathy.
kidney disease CRF: chronic renal failure GFR: glomerular fi ltration rate * Occupation not recorded for 44 Las Brisas residents Peer Reviewed