CKD Prevalence and Associated Factors in Jamaica

Abstract

Introduction

Jamaica has a high attributable burden of chronic kidney disease (CKD) but no population-based prevalence estimates. We aimed to estimate the prevalence of CKD and explore associated factors.

Methods

A secondary analysis of data from Jamaican residents aged ≥ 15 years from the nationally representative Jamaica Health and Lifestyle Survey-III was performed. CKD was defined as an estimated glomerular filtration rate (eGFR) < 60 ml/min per 1.73 m², using the CKD Epidemiology Collaboration 2021 or Schwartz-Lyon equations, and/or urine albumin-to-creatinine ratio (UACR) ≥ 30 mg/g. Associated factors included age, sex, socioeconomic status (SES), education level, body mass index (BMI), hypertension, diabetes mellitus, and sickle cell trait (SCT). Weighted prevalence estimates were determined, accounting for survey design. Multivariable logistic regression was used to evaluate CKD associations.

Results

Analyses included 583 individuals, 366 females, mean ± SD age was 49.0 ± 18.2 years. CKD prevalence was 14.8% (95% confidence interval [CI]: 11.5%–18.9%). Seven percent (7.2% (95% CI: 5.1%–10.1%) had CKD stage 3 or higher and 8.8% (95% CI: 6.3%–12.0%) had albuminuria. Individuals with CKD were older (mean age: 57 vs. 46.3 years, P < 0.001), had higher mean systolic blood pressure (140.3 mm Hg vs. 128.3 mm Hg, P < 0.001), and fasting glucose (6.7 vs. 5.8 mM/l, P < 0.001). In a multivariable regression model, hypertension (odds ratio [OR]: 2.14, 95% CI: 1.22–3.75), diabetes mellitus (OR: 2.39, 95% CI: 1.36–4.19), hemoglobin AC genotype (OR: 2.14, 95% CI: 0.64–7.13) were associated with higher odds of CKD, whereas higher education level with lower odds of CKD (OR: 0.47, 95% CI: 0.25–0.89) and (OR: 0.41, 95% CI: 0.18–0.96) for secondary and tertiary education, respectively.

Conclusion

CKD prevalence was estimated at 15%. This may translate to increased burden on the Jamaican health system.

Graphical abstract

CKD is an emerging global health problem and affects approximately 1 in 10 persons worldwide, resulting in > 1.2 million deaths annually.^1,2 Increased urinary albumin excretion (≥ 30 mg/d) or reduced glomerular filtration rate (< 60 ml/min per 1.73 m²) that lasts ≥ 3 months clinically defines CKD.³ End-stage-kidney disease may develop along the life course of CKD in which dialysis or transplantation is needed for survival.³ Excess cardiovascular mortality, health care costs, cognitive impairment, and poor quality of life are associated with CKD; with these risks increasing as eGFR declines and albuminuria increases.4, 5, 6, 7, 8

Two-thirds of the global CKD population reside in middle- and low-income countries,⁹ with Latin America and the Caribbean having among the highest mortality and disability-adjusted life years attributable to CKD worldwide,^1,10 yet sparse population-based prevalent data. Jamaica, a middle-income Caribbean country with gross domestic product of US$ 8960 per capita,¹⁰ shares the high burden of CKD described regionally.^10,11 CKD is the fourth leading cause of combined death and disability, representing a 20% increase in the past decade.¹² Estimates from the Global Burden of Diseases, CKD Study,¹ and the Global Kidney Health Atlas,¹³ indicate a 9% to 10% prevalence of CKD in Jamaica.^1,13 However, these estimates are derived from complex models based on geographic proximity from countries with robust CKD registry data or surveillance systems or modelled from systematic literature review of available data.¹

CKD risk factors are increasing in Jamaica, 1 in 3 Jamaicans have hypertension, 12% with diabetes mellitus and > 50% are overweight or obese.¹⁴ Furthermore, higher rates of systemic lupus erythematosus as well as sickle cell disease and SCT confer additional risk.^11,15 Previous reports of CKD epidemiology are based on data from the Caribbean renal registry, which are likely underestimates because of the reliance on voluntary reporting and the variability in CKD definitions.^11,16,17 Ferguson et al.¹⁸ in a cross-sectional analysis from a single-center study of 132 patients with diabetes attending a sub-specialty clinic, 22% had reduced eGFR, and 80% had albuminuria. From a cross-sectional analysis from the Jamaica Birth Cohort, an estimated 8% of persons aged 18 to 20 years had either albuminuria or reduced eGFR.¹⁹ This may suggest a higher than previously reported rate of CKD and highlights the need for population-based estimates of CKD prevalence. We aimed to estimate the prevalence of CKD in Jamaica and investigate the association with demographic, clinical, and socioeconomic factors in a nationally representative sample.

Methods

A cross-sectional secondary analysis of data from the Jamaica Health and Lifestyle Survey 2016–2017 was performed. The Jamaica Health and Lifestyle Survey 2016–2017 was a community-based national survey of noninstitutionalized persons resident in Jamaica, aged ≥ 15 years, conducted from 2016 to 2017. The details of the Jamaica Health and Lifestyle Survey 2016–2017 study protocol are published elsewhere.^14,20,21 The multistage sample resulted from randomly selected enumeration districts (EDs) stratified by parish, systematic sampling of 20 households from each ED, and selection of a household participant within EDs was done using the Kish method.^22,23 Ethical approval was obtained from the Ministry of Health and Wellness of Jamaica and The University of the West Indies Mona Campus Research Ethics Committees (Ethics Approval number: UWI - ECP 25, 15/16; MOHW- 2015/51). All participants provided written informed consent.

Data collectors were trained and certified in questionnaire administration and measurement techniques. Demographics, self-reported medical and medication history, education level, and smoking status were collected from questionnaire. Point-of-care blood samples for fasting blood glucose, cholesterol, and glycosylated hemoglobin levels; and a single early morning urine sample for UACR testing; blood pressure measurements; weight; and height were measured. Venous sampling for serum creatinine was performed. A total of 2807 Jamaicans completed the questionnaire of which 1189 provided data for ED, height, age, creatinine, and sex; and survey weighting for prevalence estimates. Six hundred ninety provided UACR testing. The final sample included in the analyses comprised 583 individuals.

Measurements

Blood pressure was measured using an oscillometric device (Omron 5 Series blood pressure monitor, Omrcon Healthcare, Lake Forest, IL). Three measurements were taken in the right arm after the participant had been seated for 5 minutes, and followed standardized procedures developed for the International Collaborative Study of Hypertension in Blacks.²⁴

Weight was measured using a portable digital scale (Tanita HD-351 Digital Weight Scale, Tnita Corportion, Tokyo, Japan). Height was measured using a portable stadiometer (Seca 213, Mobile Stadiometer, Seca GmbH & Co, Hamburg, Germany). Fasting glucose and total cholesterol were measured from a capillary blood sample using a point-of-care device (SD LipidoCare, Suwon, South Korea).

Serum creatinine from a single blood sample was measured using a Cobas c111 analyzer and urine creatinine from an early morning urinary sample was measured using a Cobas 6000 analyzer at the Tropical Metabolism Research Unit Laboratory in the Caribbean Institute for Health Research, Mona, using the Jaffe method calibrated using isotope dilution mass spectrometry, which is traceable to a Standard Reference Material. eGFR was calculated using the race-free 2021 CKD Epidemiology Collaboration equation if participant was aged > 20 years,²³ and the Schwartz-Lyon equation, if the participant was aged 15 to 20 years.²⁵ The Schwartz-Lyon equation was chosen because of its validity in estimating GFR in adolescents, and CKD Epidemiology Collaboration–based on Kidney Disease: Improving Global Outcomes 2023 recommendations.³ We stratified by risk categories (low to very high risk) using eGFR and UACR-based on Kidney Disease: Improving Global Outcomes guidelines³ (Supplementary Figure S1).

Urine albumin was measured at The University of the West Indies Chemical Pathology Laboratory from the early morning urine specimen, with concentration based on the turbidity reaction produced on addition of sulphosalicylic acid. Spot UACR was calculated by dividing the urine albumin concentration by the urine creatinine concentration. Hemoglobin genotyping was performed at the Tropical Metabolism Research Unit Laboratory using a plasma sample. These were analyzed using cellulose acetate electrophoresis and electrophoretic variants were assessed using agarose gel electrophoresis.

Inclusion and Exclusion Criteria

Inclusion criteria were age ≥ 15 years, availability of serum creatinine values, UACR, age, sex and height to calculate eGFR. Absence of serum creatinine, sex, height, age or ED for the survey weighting resulted in exclusion from this analysis.

The primary outcome was prevalent CKD, defined as reduced eGFR < 60 ml/min per 1.73 m² and/or UACR ≥ 30 mg/g.

Main Exposure Variables

Hypertension as defined as self-reported hypertension (if the respondent indicated medication use for hypertension) or if mean systolic pressure of three readings ≥ 140 mm Hg or diastolic blood pressure ≥ 90 mm Hg. Diabetes mellitus was defined as a fasting glucose > 7.0 mmol/l or self-reported diabetes (if the respondent indicated medication use for diabetes).

BMI was calculated and then categorized into World Health Organization weight categories.^14,20 SCT and disease were defined using hemoglobin genotyping. Responses to highest education level attained was used to determine education level. Responses were categorized as “less than secondary school” if the response included no education, primary, or junior secondary school (less than grade 10); “secondary school” if responded with high school education (grades 10–13) and “more than secondary school” for beyond a high school education. For household SES, participants’ responses to ownership of a list of 22 household assets were used. Terciles of the number of household assets were created as follows: low household SES persons with ≤ 9 household assets, middle household SES with 10 to 12 items, and high household SES with 13 to 22 items. This has been used in previous studies as a determinant of SES among Jamaicans.²⁶

Smoking status was determined by participant response to a question on tobacco use, and categorized as “never smoked” if response was never smoked, “former smoker” if response was former smoker, and “current smoker” if response was “yes, not every day” or “yes, daily.”

Sample Size and Power

Given the available sample size of 583, we estimated statistical power based on CKD prevalence of 10%,¹ and a margin of error of 5%. Given that data analysis accounted for study design, application of a design effect of 1.94 to the sample size of 583 yielded an effective sample size of 301.²¹ At the 95% confidence level, this effective sample of the study would have 78.3% power to estimate a 10% prevalence with 5% margin of error.

Statistical Analysis

Data were analyzed using STATA software (version 17BE; StataCorp LP). Sample-based estimates of means and SD for continuous variables, as well as proportions and frequencies for categorical variables were obtained. Associations for explanatory variables and the outcome of interest were tested using chi-square tests for proportions and t test, and analysis of variance.

Application of sampling weights yielded nationally representative CKD prevalence estimates. The sampling weights were based on the probability of selection of dwellings and EDs, and calibration using the Jamaican population distribution in parish- and sex-specific 5-year age bands. These base weights were multiplied by a nonresponse adjustment factor to produce survey weights adjusted for unit nonresponse. Bivariable logistic regression was performed on exposure variables and CKD as the outcome variable. Exposure variables with P-values ≤ 0.20 on bivariable analyses or a priori association with CKD were included in the multivariable logistic regression model. Hosmer–Lemeshow goodness-of-fit testing was done to assess differences between observed and expected results in the subgroups of the model population.

Missing Data

A complete case analysis was performed. To determine evidence of the missing at random mechanism, we compared the complete case analysis sample data with the remainder of the study sample using 2-sample unpaired t tests and Pearson’s chi-square tests, as appropriate, in unweighted analyses (Supplementary Table S1) and using Pearson’s chi-square tests corrected for survey weighting design (Supplementary Table S2).

Results

Sample Characteristics

Of 583 individuals, 366 were female (63%). The mean (±SD) age for the sample was 48.5 ± 17.8 years. More than half of the individuals (58.5%) were overweight or obese with mean ± SD BMI of 28.4 ± 9.5 kg/m². Forty-three percent had hypertension, with the mean ± SD systolic and diastolic blood pressures of 130.6 ± 21.0 mm Hg and 83.4 ± 12.1 mm Hg, respectively. Prevalence of diabetes mellitus was 20.2%, with mean ± SD fasting glucose of 6.0 ± 2.2 mmol/l. Most of the respondents were never-smokers (72.0%). Most had at least high school education (47%), whereas high household SES was noted in one-quarter of respondents (25.8%). Sex-based differences in the sample are presented in Table 1. Women had higher rates of diabetes mellitus (23.6 vs. 14.1%, P = 0.008) and hypertension (46.8% vs. 36.0%, P = 0.011) in the unweighted sample, with higher BMIs (30.1 vs. 25.0, P < 0.001) than men. Women had higher education, with a higher proportion with tertiary education (21.6 vs. 10.8%, P = 0.004) than men.

Table 1.

Sex differences in the study population using unweighted estimates

Characteristics	All (N = 583)	Male (n = 217)	Female (n = 366)	P-value
Age in yrs, mean ± SD	48.5 ± 17.8	49.2 ± 18.6	48.1 ± 17.3	0.448
BMI, (kg/m2) mean ± SD	28.4 ± 9.5	25.0 ± 8.0	30.4 ± 9.7	< 0.001a
eGFR (ml/min per 1.73 m2) mean ± SD	85.5 ± 21.5	84.8 ± 19.7	85.9 ± 22.5	0.565
CKD risk factors
Systolic blood pressure, mean ± SD	130.6 ± 21.0	131.2 ± 19.5	130.2 ± 21.9	0.614
Diastolic blood pressure, mean ± SD	83.4 ± 12.1	82.0 ± 12.9	84.2 ± 11.6	0.042a
Serum fasting glucose (mmol/l), mean ± SD	6.0 ± 2.2	6.2 ± 2.5	5.7 ± 1.4	0.022a
Glycosylated hemoglobin (%), mean ± SD	6.3 ± 1.3	6.1 ± 1.3	6.4 ± 1.3	0.011a
Fasting cholesterol (mmol/l) mean ± SD	4.5 ± 1.0	4.5 ± 1.1	4.6 ±1.0	0.300
Self-reported or measured hypertension, % (n)	42.8 (246)	36.0 (77)	46.8 (169)	0.011a
Self-reported or measured diabetes mellitus, % (n)	20.1 (111)	14.1 (28)	23.6 (83)	0.008a
Hemoglobin AS genotype, % (n)	9.4 (61)	10.8 (23)	10.6 (38)	0.903
Hemoglobin AC genotype, % (n)	2.3 (19)	3.8 (8)	3.1 (11)	0.903
Current smokers, % (n)	17.1 (69)	32.0 (49)	8 (40)	< 0.001a
Former smokers, % (n)	10.9 (44)	16.3 (25)	7.6 (19)	< 0.001a
Highest education level, % (n)
Less than high school % (n)	35.2 (200)	39.0 (83)	32.9 (117)	0.004a
High school, % (n)	47.3 (269)	50.2 (107)	45.5 (162)	0.004a
More than high school, % (n)	21.6 (100)	10.8 (23)	21.6 (77)	0.004a
Household possession categories (household SES)
Low % (n)	41.9 (244)	48.4 (105)	38.1 (139)	0.043a
Middle % (n)	32.3 (188)	30.0 (65)	33.7 (123)	0.043a
High (%)	25.8 (150)	21.7 (47)	28.2 (103)	0.043a

BMI, body mass index; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; SES, socioeconomic scale.

^aStatistical significance correspond to a P-value < 0.05.

Burden of CKD

Based on survey weighting, the prevalence of CKD was 14.8% (95% CI: 11.5%–18.9%). Just > 7% (7.2% [95% CI: 5.1%–10.1%]) had reduced GFR and 8.8% (95% CI: 6.3%–12.0%) had albuminuria. There were no sex differences in survey-weighted CKD prevalence (males: 14.5% vs. females 15.0%, P = 0.859). Over half of the individuals (58.5%) with CKD were aged > 64 years (Figure 1).

Figure 1.

Prevalence of CKD stages by Kidney Disease Improving Global Outcomes risk categories. Bold is percentages based on weighted estimates, with 95% confidence intervals in brackets. CKD, Chronic kidney disease; GFR, glomerular filtration rate; A1, urine albumin creatinine ratio (UACR) <30mg/g; A2, UACR 30-300mg/g; A3, UACR >300mg/g; G1, GFR ≧90mL/min/1.73m²; G2, 60-89 mL/min/1.72m²; G3a, GFR 45-59mL/min/1.73m²; G3b, GFR 30-44mL/min/1.73m2; G4, 15-29mL/min/1.73m²; G5, <15 mL/min/1.73m².

Of those with CKD, the majority had CKD stage 1 or 2 (46.1%) followed by CKD stage 3a (40.9%), CKD stage 3b (7.9%), CKD stage 4 (1.7%), CKD stage 5 (3.5%). For details of the weighted estimates by Kidney Disease: Improving Global Outcomes stage (Figure 2). The estimated prevalence of high risk or very high-risk CKD was 4.1% (95% CI: 2.4–6.9).

Figure 2.

Frequency of chronic kidney disease by 10-year age categories.

CKD Risk Factors

Compared with individuals without CKD, persons with CKD were older (mean age: 57.0 vs. 46.4 years, P < 0.001), had higher mean systolic blood pressure (140.3 vs. 128.2 mm Hg, P < 0.001), higher BMI (30.3 vs. 27.9 kg/m², P < 0.016), higher fasting glucose (6.7 vs. 5.8 mM/l, P < 0.001), and higher glycosylated hemoglobin (6.6 vs. 6.2%, P = 0.003). There were no statistically significant differences in diastolic blood pressure (85.2 vs. 82.9 mm Hg, P = 0.076) or fasting total cholesterol (4.7 vs. 4.5 mmol/l, P = 0.067) between CKD and non-CKD groups. Differences by CKD are presented in Table 2.

Table 2.

Differences by chronic kidney disease in the study population using unweighted estimates

Characteristics	All (N = 583)	CKD (n = 115)	No CKD (n = 468)	P-value
Age in yrs, mean ± SD	48.5 ± 17.8	57.0 ± 18.5	46.4 ± 17.0	< 0.001a
Male sex, % (n)	37.2 (217)	29.6 (34)	39.1 (183)	0.058
BMI, (kg/m2) mean ±SD	28.4 ± 9.5	30.3 ± 11.9	27.9 ± 8.7	0.142
eGFR (ml/min per 1.73 m2) mean ± SD	85.5 ± 21.5	66.8 ± 25.3	90.1 ± 17.6	< 0.001a
CKD risk factors
Systolic blood pressure, mean ± SD	130.6 ± 21.0	140.3 ± 26.1	128.2 ± 18.9	< 0.001a
Diastolic blood pressure, mean ± SD	83.4 ± 12.1	85.2 ± 13.7	82.9 ± 11.7	0.076
Serum fasting glucose (mmol/l), mean ± SD	6.0 ± 2.2	6.7 ± 2.8	5.8 ± 2.0	< 0.001a
Glycosylated hemoglobin (%), mean ± SD	6.3 ± 1.3	6.6 ± 1.7	6.2 ± 1.2	0.003
Fasting cholesterol (mmol/l) mean ± SD	4.5 ± 1.0	4.7 ± 1.0	4.5 ±1.1	0.067
Overweight, % (n)	24.8 (140)	20.4 (23)	25.9 (117)	0.360
Obese, % (n)	33.7 (190)	39.8 (45)	32.1 (145)	0.360
Self-reported or measured hypertension, % (n)	42.8 (246)	68.1 (77)	36.6 (169)	< 0.001a
Self-Reported or measured diabetes mellitus, % (n)	20.2 (111)	41.1 (44)	15.1 (67)	< 0.001a
Hemoglobin AS genotype, % (n)	10.7 (61)	9.8 (11)	10.9 (50)	0.154
Hemoglobin AC genotype, % (n)	3.3 (19)	6.3 (7)	2.6 (12)	0.154
Current smokers, % (n)	17.1 (69)	9.1 (7)	19.0 (62)	0.076
Former smokers, % (n)	10.9 (44)	9.1 (7)	11.4 (37)	0.076
Highest education level, % (n)
Less than high school % (n)	35.2 (200)	56.0 (61)	30.2 (139)	< 0.001a
High School, % (n)	47.3 (269)	33.0 (36)	50.7 (233)	< 0.001a
More than high school, %(n)	17.6 (100)	11.0 (12)	19.1 (88)	< 0.001a
Household possession categories (household SES)
Low % (n)	41.9 (244)	49.1 (56)	40.2 (188)	0.221
Middle, % (n)	32.3 (188)	28.1 (32)	33.3 (156)	0.221
High, % (n)	25.8 (150)	22.8 (26)	26.5 (124)	0.221

BMI, body mass index; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate.

^aStatistical significance corresponds to a P-value < 0.05.

Self-reported hypertension (68.1% vs. 36.5%, P < 0.001) and diabetes mellitus prevalence (41.1% vs. 15.1%, P < 0.001) were higher in individuals with CKD. Persons with CKD were less likely to have had secondary school education (56.0% vs. 30.2%, P < 0.001). There was no statistically significant difference in CKD by sex, household SES, or BMI category.

In Table 3, we present the final multivariable logistic model. Hypertension (OR: 2.14, 95% CI: 1.22–3.76) and diabetes mellitus (OR: 2.38, 95% CI: 1.36–4.19) were associated with higher odds of CKD. Hemoglobin AC genotype (OR: 2.14, 95% CI: 0.64–7.13) was associated with higher odds of CKD but did not reach statistical significance. Secondary education (OR: 0.47, 95% CI: 0.25–0.89) and tertiary education (OR: 0.41, 95% CI: 0.18–0.96) were associated with lower odds of CKD. Hosmer–Lemeshow goodness-of-fit chi-square statistic for the model was 3.6 (P = 0.731).

Table 3.

Bivariable and multivariable logistic regression for chronic kidney disease

Variable	Unadjusted odds ratio (95% CI)	P-value	Multivariable model odds ratio (95% CI)	P-value
BMI, kg/m2	1.02 (1.00–1.04)	0.020	1.01 (0.99–1.04)	0.333
Age (/1 yr)	1.04 (1.02–1.05)	< 0.001	1.01 (0.99–1.03)	0.249
Male sex	0.65 (0.42–1.02)	0.059	0.79 (0.46–1.35)	0.383
Hypertension	3.70 (2.39–5.75)	< 0.001	2.14 (1.22–3.76)	0.008a
Diabetes mellitus	3.91 (2.46–6.28)	0.001	2.38 (1.36–4.19)	0.003a
Hemoglobin genotype (reference: Hb AA genotype)
Genotype HbAS (sickle trait)	0.93 (0.47–1.85)	0.359	1.05 (0.46–2.36)	0.911
Genotype Hb AC (HbC carrier)	2.46 (0.94–6.43)	0.065	2.14 (0.65–7.13)	0.215
Education level: (reference: less than high school)
High school	0.35 (0.22–0.56)	< 0.001	0.47 (0.25–0.88)	0.019a
Post–high school	0.31 (0.16–0.61)	0.001	0.41 (0.18–0.96)	0.040a
Household possession category: (reference low household SES)
Middle	0.69 (0.43–1.12)	0.130	1.11 (0.61–2.03)	0.733
High	0.70 (0.42–1.18)	0.184	1.10 (0.56–2.16)	0.777

BMI, body mass index; CI, confidence interval.

^aStatistically significant P < 0.05 on the multivariable logistic regression.

We performed sex-specific logistic regression models, the results of which are shown in Supplementary Tables S3 and S4. In women, diabetes mellitus (OR: 1.83, 95% CI: 0.92–3.68), hypertension (OR: 4.04, 95% CI: 1.97–8.28), and sickle cell AC genotype (OR: 2.06, 95% CI: 0.41–10.45) were associated with higher odds of CKD. In men, diabetes (OR: 4.90, 95% CI: 1.70–14.2) was associated with higher odds of CKD but not hypertension (OR: 0.55, 95% CI: 0.19–1.67). Higher education was associated with lower odds of CKD in men and women.

Missing Data

Tables summarizing the characteristics of the sample with missing data on albuminuria and creatinine testing are available in Supplementary Tables 1 and 2. Individuals with nonmissing data were older (mean age: 48.5 ± 17.8 vs. 46.1 ± 18.9 years, P = 0.006), had higher mean glycosylated hemoglobin (6.3% vs. 5.9%, P < 0.001), higher mean fasting glucose (6.0 ± 2.2 vs. 5.8 ± 2.1, P = 0.015), were more likely to have a high household SES (31.1% vs. 25.8%, P = 0.040), and with a lower proportion of current smokers (11.6 % vs 17.1%). There were no differences by sex, proportion with diabetes mellitus, hypertension, obesity or overweight, education level, or hemoglobin genotype. Comparable results were observed in the weighted differences between missing and nonmissing groups. A higher proportion of individuals included in the analysis were in the 45 to 54 and 55 to 64 years age groups (18 vs. 14.4% and 13.3 vs. 6.9%, respectively, P = 0.03). A higher proportion were obese (32.7 vs. 27.3%, P = 0.053).

Discussion

In this cross-sectional analysis of Jamaicans, estimated prevalences of CKD, albuminuria, and reduced GFR were 15%, 9%, and 7 %, respectively. Diabetes and hypertension were associated with increased odds whereas higher education attainment reduced odds of CKD.

Our data suggest a higher rate of CKD than previous estimates from Global Burden of Disease and Global Kidney Health Atlas,^1,11 and is comparable with CKD rates among adults in the United States, Canada, and Haiti of 13% to 15%.27, 28, 29 Compared with Non-Hispanic Black Americans (19.5%), CKD prevalence was lower.²⁹ Seven percent had impaired kidney function, a rate higher than previous reports from other Caribbean islands of 3% and 4% from Haiti and St Kitts and Nevis, respectively,^27,30 but similar to high income countries.²⁹ Although these comparisons are less direct because of differences in study methodology (some using survey weighting or age-standardized prevalence rates vs. crude estimates) and differences in population characteristics.27, 28, 29, 30 Despite this, the observed rates of high risk CKD has public health care implications in middle-income countries such as Jamaica, in which treatment of advanced kidney disease (such as dialysis) are expensive and largely paid out-of-pocket, resulting in lack of access to care. Jamaica has one of the highest out-of-pocket payments (51%–75%) for dialysis in the Caribbean^10,13 and a high attributable CKD morbidity and mortality.¹²

The high prevalence of CKD, especially high-risk CKD in this study, suggest the need for early identification in health care systems in Jamaica. Initiation of renin-angiotensin-aldosterone system inhibitors, sodium-glucose cotransporter 2 inhibitors, and

nonsteroidal mineralocorticoid receptor antagonists reduce progression to kidney failure and cardiovascular mortality in CKD.31, 32, 33 Addition of sodium-glucose cotransporter 2 inhibitors to standard-of-care resulted in a projected 1.7 year longer time with eGFR > 15 ml/min per 1.73 m², 1.7 year longer life expectancy with cost-effective ratios of between $8280 and $17,623 per quality-adjusted life years in 3 European countries.³⁴ Data from South East Asia, and the United States also show cost effectiveness and reductions in health care costs and burden with the use of sodium-glucose cotransporter 2 inhibitors and renin-angiotensin-aldosterone system inhibitors therapy.^35,36 Focus on screening, at risk individuals, with initiation of early disease-modifying therapy may be needed to reduce CKD attributable health care burden. Sodium-glucose cotransporter 2 inhibitors and nonsteroidal mineralocorticoid receptor antagonists are available in Jamaica, but medication costs are prohibitive. CKD is not covered as one of the diseases in the National Health Fund (which subsidizes medications for chronic medical illnesses). This study may argue the need for inclusion to improve medication access.^10,13

Hypertension and diabetes were associated with a 2-fold risk of CKD; however, the higher than expected rates of advanced kidney disease may suggest environmental, genetic, or socioeconomic factors which affect progression. The presence of high-risk Apolipoprotein 1 (APOL1) genotypes in persons of African ancestry has been associated with incident CKD and its progression.^37,38 Homozygosity for APOL1 is associated with a 7- to 11-fold increased risk of hypertension attributable end-stage-kidney disease among African-Americans.³⁸ APOL1 is associated with lupus nephritis, HIV-associated nephropathy and sickle cell–related kidney disease.38, 39, 40, 41 And is a factor enhancing progression in diabetic kidney disease.³⁸ Jamaica has a predominant West African Ancestry but no data on APOL1 high-risk prevalence,⁴¹ which may warrant further study. Environmental factors act as a second hit in APOL1 risk alleles, or are independently associated with CKD.^38,41 Ambient temperature, air pollution (particulate matter 2.5), environmental toxins (e.g., heavy metals), and viral infections such as HIV and dengue virus are such potential mediators; and their role in CKD progression in Jamaica and/or the Caribbean could be examined in future longitudinal cohort studies.41, 42, 43

Surprisingly, despite a high rate of SCT (11%), there was no association with CKD on multivariable analysis. Data from 5 population-based cohort studies showed a 50% increased risk of incident CKD, and 75% risk of eGFR decline in SCT versus normal genotype, after adjusting for traditional CKD risk factors.⁴⁴ SCT also increased the risk of CKD in persons of African descent residing in the UK with well-controlled HIV infection.³⁹ We found a 2-fold higher odds of CKD in the hemoglobin carrier C status, and no association with SCT and CKD on multivariable analysis. This finding could be further explored in larger studies or meta-analyses.

Higher education attainment, a measure of individual SES, was protective for CKD, congruent with previous observational data showing an inverse association of incident CKD and education.⁴⁵ Although previous reports have attributed lower prevalence ratios of hypertension and diabetes in Jamaica than in a United States cohort because of differences in age, sex, and BMI rather than education attainment, this difference may be because Jamaican cohort had lower BMIs and had lower rates of hypertension and diabetes and were younger.⁴⁶ More recent analyses showed an inverse association of education attainment and multimorbidity independent of other covariates.⁴⁷ Education attainment was associated with 5 ideal cardiovascular health characteristics (e.g. physical activity, healthy diet, and lower BMI) among Jamaican women; these factors also reduce CKD rates.²¹ That higher education attainment was associated with lower CKD risk may reflect better health care access.

Women had higher rates of hypertension and diabetes, as well as higher BMIs than men. A higher proportion of women, had higher education attainment. Hypertension and diabetes were associated with higher odds of CKD in women, whereas diabetes was associated with higher odds of CKD in men. This is congruent with previous reports of higher rates of cardiovascular disease risk factors (diabetes, hypertension, obesity, and physical inactivity) among Jamaican women than among men.(see Supplmentary Tables 3 and 4)⁴⁸ Although a higher prevalence of cardiovascular disease risk factors was found, 10-year estimated cardiovascular disease risk was similar in men and women. SES modified this risk, with women of lower SES based on occupation and education having higher estimated cardiovascular disease risk.⁴⁸ In our study, despite higher rates of traditional CKD risk factors, weighted prevalence of CKD was similar in both sexes. This may be related to higher education attainment in women, resulting in better health care literacy and control of underlying medical illness. Hypertension was associated with CKD in women and not men, contrary to previous reports with higher risk of incident CKD and end-stage-kidney disease with hypertension in men.⁴⁹ Future cohort studies should explore sex differences in CKD risk, social determinants of health and health care outcomes. These findings highlight the need for sex-specific public health approaches to reduce CKD risk factors.

This is the first study estimating CKD prevalence in a large representative sample in Jamaica; however, there are several limitations. There was the possibility of selection bias limiting generalizability of results. Representativeness may be compromised by attrition from the primary sample because of missing data. We used single measurements of creatinine and UACR; therefore, chronicity could not be established, affecting prevalence estimates. We used creatinine to estimate GFR; the former can be affected by non-GFR determinants of creatinine (e.g., high protein diet and creatine supplements), whereas cystatin C improves the accuracy of eGFR in persons of African descent.^3,50

In conclusion, CKD prevalence in Jamaica is comparable to worldwide estimates, but with a high proportion with high-risk CKD. Accessible systems for renal replacement therapy coupled with public health measures to enhance early detection and management of CKD are needed.

Disclosure

L-AMF has received speakers fees from Dr Reddy’s International and Servier International, outside the submitted work. TSF reported grants from National Health Fund (Jamaica), during the conduct of the study and personal fees from Dr Reddy's Laboratories, outside the submitted work. RJW, MKT-R, AG, and KW-K have grants from the Ministry of Health and Wellness Jamaica, and the National Health Fund (Jamaica) during the conduct of the study. All the other authors declared no competing interests.

Supplementary Material

Supplementary File (PDF)

The data set and code book for data supporting this manuscript is available at Mendeley Data, https://data.mendeley.com/datasets/nnzfgs9j7p/1. Figure S1. Kidney Disease: Improving Global Outcomes: Prognosis of chronic kidney disease by GFR and albuminuria categories. Table S1. Table highlighting differences between chronic kidney disease subset and the Jamaica Health and Lifestyle III subset with missing data. Table S2. Table highlighting differences between chronic kidney disease subset and the Jamaica Health and Lifestyle III subset using with missing data using survey weighting. Table S3. Bivariable and multivariable logistic regression for chronic kidney disease for men. Table S4. Bivariable and multivariable logistic regression for chronic kidney disease for women.