Determinants of hyperuricemia in non-dialysed chronic kidney disease patients in three hospitals in Cameroon

Background Chronic kidney disease (CKD) poses a substantial health burden in sub-Saharan Africa, with risk factors ranging from communicable to non-communicable diseases. Hyperuricemia has been recently identified as a factor of progression of CKD. Identifying factors associated with hyperuricemia in CKD patients would help determine interventions to reduce CKD mortality, particularly in resources limited countries. We sought to determine the prevalence and factors associated with hyperuricemia in non-dialysed CKD adult patients in Cameroon. Methods This was a cross-sectional study of non-dialysed CKD patients, conducted in 3 referral nephrology units in Cameroon. Relevant clinical and laboratory data were collected using interviewer-administered questionnaires. Serum uric acid, spot urine protein and spot urine creatinine were assessed. Associations between variables were assessed using multivariate analysis. Level of statistical significance was set at α < 0.05. Results A sample of 103 participants was included. Mean age of study participants was 55.78 ± 12.58 years, and 59.3% were men. Sixty-nine (67%) had hyperuricemia. Patient’s age (OR: 1.08, 95% CI: 1.03–1.13), estimated glomerular filtration rate (OR: 0.94, 95% CI: 0.90–0.98), spot urine protein-creatinine ratio (OR: 1.83, 95% CI: 1.07–3.12), no hypertension (OR: 0.09, 95% CI: 0.02–0.46), urate lowering therapy (OR: 4.99, 95% CI: 1.54–16.16), loop diuretics (OR: 3.39, 95% CI: 1.01–11.42), obesity (OR: 6.12, 95% CI: 1.15–32.55) and no anaemia (OR: 0.04, 95% CI: 0.00–0.29) were independently significantly associated with hyperuricemia. Conclusions In this sample of non-dialysed CKD patients in Cameroon, about 7 out of 10 had hyperuricemia. Hyperuricemia was independently associated with patient’s age, estimated glomerular filtration rate, spot urine protein-creatinine ratio, hypertension, urate lowering therapy, loop diuretics, obesity and anaemia. More studies are required to establish causal relationships between these associations.


Background
Chronic kidney disease (CKD) poses a substantial health burden in sub-Saharan Africa (SSA), with an estimated prevalence of 13.9% [1]. Risk factors of CKD in SSA range from communicable to non-communicable diseases [1,2]. It is estimated that, by 2030, more than 70% of patients with end-stage renal disease (ESRD) will reside in low-income countries, such as countries of SSA [1]. In 2010, out of the 2.6 million people projected worldwide who accessed renal replacement therapy (RRT), only about 7% were from low-income and lower-middle income countries [3,4]. In Cameroon, the morbidity and mortality of CKD is high, especially amongst patients on haemodialysis [5][6][7]. Emphasis should therefore be on prevention, early detection and managing factors of progression of CKD, in order to curb incidence of ESRD, while improving access to RRT in SSA [3,8,9].
Hyperuricemia in CKD has been mostly attributed to declining renal function [28]. However, in clinical practice, raised serum uric acid level in CKD is associated with a combination of several other factors, as not every CKD patient presents with hyperuricemia. In non-CKD subjects, factors associated with hyperuricemia include obesity, dyslipidaemia, alcohol consumption, fructose-rich and purine-rich diets, genetics and physical inactivity [29][30][31]. In CKD patients, factors associated with hyperuricemia have not been fully elucidated.
We therefore sought to determine the prevalence and factors associated with hyperuricemia in non-dialysed adult CKD patients in Cameroon, in order to generate findings that would help determine interventions to reduce CKD mortality, particularly in resources limited countries.

Study population and sampling
This was a cross sectional study carried out in the nephrology units of 3  After prior ethical clearance from the Cameroon National Ethics Committee, we systematically sampled every non-dialysed CKD patients aged 20 years and above, followed up in the nephrology units of the 3 hospitals. We excluded patients with chronic liver diseases, haemoglobinopathies and/or cancers as reported in their medical charts, and patients who did not give consent.

Data collection
Socio-demographic data (age, gender) and relevant clinical information (aetiology of CKD, comorbidities, dietary regimens, use of medications known to influence serum uric acid and creatinine levels) were recorded from the patient's medical chart or obtained by interviewer-administered questionnaire. Anthropometric data including height and weight were obtained using a stadiometer (measured to the nearest 0.1 cm, without any foot or head wear) and a manual weighing scale (measured to the nearest 0.1 kg, with light clothing and no foot wear) respectively. Blood pressure was measured according to World Health Organisation (WHO) guidelines [32] using an automatic blood pressure machine (OMRON® M2, HEM-7121-E). The most recent (less than 3 months) values of haemoglobin, serum high-density lipoprotein cholesterol (HDL-c), serum low-density lipoprotein cholesterol (LDL-c), serum total cholesterol, serum triglycerides and plasma creatinine in patient's medical charts were recorded.
Uric acid in the serum specimen was measured using the direct kinetic uricase method [33]. A clean-catch mid-stream urine specimen from all participants was collected for spot protein and creatinine assessment using the Benzethonium Chloride [34] and Jaffé method [35] respectively.
Renal disease was Glomerular if manifesting with glomerular range proteinuria, and/or haematuria (deformed or red blood cells cast), associated with hypertension and/or oedema; Tubulo-interstitial if manifesting with sterile leucocyturia, and low urine specific gravity, associated with or without non-glomerular haematuria and proteinuria; Vascular if presence of hypertension, with moderate proteinuria and normal urine sediment. Mixed renal involvement was renal disease manifesting with a combination of 2 or more of glomerular, tubulo-interstitial and vascular involvement.
Hyperuricemia was defined as serum uric acid > 70 mg/l in males and > 60 mg/l in females [38]. Hypertension was defined as a systolic blood pressure ≥ 140 mmHg and/or a diastolic blood pressure ≥ 90 or current use of antihypertensive drugs. Diabetes was defined as history of diabetes confirmed in the patient's medical records and/or current use of blood glucose control medications. Gout was defined as history of gouty arthritis confirmed in the patient's medical records and/or current use of urate lowering therapy (ULT) against gout. Low protein, low-carb and low fat diet were defined as a dietary regimen poor in proteins (less than 30 g daily), sugars and lipids respectively, as prescribed by a dietician. Body mass index (BMI) was classified as; underweight (BMI < 18.5 kg/m 2 ), normal weight (BMI = 18.5-24.9 kg/m 2 ), overweight (BMI = 25-29.9 kg/m 2 ) and obesity (BMI ≥ 30 kg/m 2 ) [39]. Anaemia was defined as haemoglobin < 13.0 g/dl in males and < 12.0 g/dl in females [40]. Dyslipidaemia was defined as total cholesterol ≥240 mg/dl (6.21 mmol/l) or low-density lipoprotein cholesterol (LDL-c) > 160 mg/dl (4.14 mmol/l) or high-density lipoprotein cholesterol (HDL-c) < 40 mg/dl (1.03 mmol/l) or triglycerides ≥150 mg/dl (1.69 mmol/l) [41].

Statistical analysis
Statistical analyses were done using Statistical Package for Social Sciences (SPSS), version 23 Inc., Chicago, Illinois, USA). Normality of continuous data was assessed using Shapiro-Wilk test. Continuous variables were summarised as means and standard deviations (SD), and medians and interquartile ranges (IQR) where appropriate. Categorical variables were summarised as frequencies and percentages.
One-Way ANOVA was used to compare means between more than 2 independent groups. Bivariate analyses were done using generalized linear models (using serum uric acid as outcome variable) and logistic regression (using hyperuricemia as outcome variable). After bivariate analyses, variables with p < 0.02 were included as adjustment variables in multivariate analyses. Continuous data were selected over categorical data for variables with both continuous and categorical data. In order to avoid multicollinearity, bivariate correlations were assessed between the predictor variables prior to multivariate analyses. Predictor variables which were highly significantly (p < 0.01) on bivariate correlation were excluded from the multivariate analysis. Furthermore, collinearity diagnostics were done to ensure that none of the predictor variables had Tolerance < 0.3 and/or Variance Inflation Factor > 3. There were missing data in lipid profile indices (Total Cholesterol [10 missing], HDL-c [11 missing], LDL-c [11 missing], and Triglycerides [14 missing]. Missing data where excluded from the analysis pairwise. Level of statistical significance was set at α < 0.05.
Sensitivity analyses were done, defining hyperuricemia as serum uric acid > 70 mg/l in males and > 60 mg/l in females or current use of ULT. The results were very similar to our current results.

Results
Out of the 198 non-dialysed CKD patients aged 20 years and above who attended our study sites during the study period, 2 patients were excluded for chronic liver disease (liver cirrhosis), one patient excluded for haemoglobinopathy (sickle cell disease), and one patient excluded for cancer (prostate cancer). Of the 194 eligible CKD patients approached for participation in the study, 103 patients gave consent and were included into study; thus a participation rate of 53.1%.

Discussion
Our study showed that more than half of CKD patients had hyperuricemia. Among a wide range of demographic, clinical and laboratory parameters evaluated, patient's age, eGFR, spot urine PCR, hypertension, ULT, loop diuretics and anaemia were significantly associated with serum uric acid and hyperuricemia in CKD.
Hyperuricemia is common in CKD and mostly results from decreased uric acid excretion [10]. The kidneys play a major role in excreting uric acid, an end-product of purine metabolism [28]. We found a higher prevalence of hyperuricemia in CKD in our setting, compared to other SSA settings such as Nigeria [26] and Chad [27]. These divergent results can be largely attributed to study methodology. In the Nigerian study, the authors evaluated 120 pre-dialysis CKD patients and reported a prevalence of 47.5%. However, they excluded patients    with gouty arthritis and patients on ULT. In the Chadian study, after evaluating 712 CKD patients, an even lower prevalence of 15.2% was reported. However, in this Chadian study, no information was provided on whether patients were on maintenance dialysis or not. Our higher prevalence could also be due to the high proportion of advanced CKD (73.2% CKD stages 4 and 5) amongst our study participants. Our prevalence was similar to the 70% reported in a paediatric population of CKD patients in USA [42]. The high proportion of advanced CKD amongst our study participants is due to the fact that our study was conducted in referral nephrology units which receive advanced cases of CKD from peripheral nephrology units. We found significant association between eGFR, spot urine PCR and serum uric acid. Patients with CKD stages 4 or 5 and had 24.66 and 7.73 times higher significant odds respectively, of having hyperuricemia compared to those with CKD stage 3b. Also, patients with severe proteinuria had 6.27 significant higher odds of having hyperuricemia compared to those with no or mild proteinuria. These findings were similar to the results of Krishnan [43]. We also found a significant association between loop diuretics and serum uric acid in our sample of non-dialysed CKD. It is known that, loop diuretics predisposes individuals to hyperuricemia [44]. Diuretics cause hyperuricemia by increasing urate absorption and decreasing urate secretion in the kidneys [44]. Loop diuretics used in the management of hypertension and fluid retention in CKD should therefore be used judiciously.
We found a significant independent association between hypertension and hyperuricemia. Noone et al.. also reported a significant association between hypertension and hyperuricemia in paediatric patients with CKD [42]. A growing body of evidence suggest uric acid levels play a role in the development of hypertension [45,46]. Uric acid can cause hypertension by mediating pro inflammatory pathways in vascular smooth muscles, inhibition of endothelial nitric oxide and activation of the renin-angiotensin system [45,46]. On the other hand, hypertension increases renal vascular resistance, reducing renal flow, those increasing urate reabsorption [30]. Also, use of antihypertensive drugs like diuretics can increase serum uric acid levels [44]. It is however unlikely that in our study, the observed association between hypertension and hyperuricemia is due to use of diuretics as this was adjusted for in our analysis. Sedaghat et al.. suggested that hypertension mediates the decline in renal function caused by hyperuricemia [47]. Further research is needed to clarify the relationship between hyperuricemia and hypertension in CKD patients.
Subjects who were obese had 6.12 times significant higher odds of having hyperuricemia. This was in accordance with findings by Noone et al, in their paediatric CKD population [42]. Hyperuricemia seen in obesity has been attributed to insulin resistance and higher leptin production [30]. Insulin resistance in obese individuals causes larger amounts of insulin to be secreted in order to maintain glucose metabolism. The kidney responds to this hyperinsulinemia by decreasing urate clearance [48]. Obesity may therefore be a cause of hyperuricemia in our subjects, but a causal relationship between these two variables was not evaluated in this study.
It is possible that the observed association between hyperuricemia and anaemia is just an incidental finding. Both conditions can be caused by the underlying CKD, and there may not be a direct relationship between the two. However, in our analysis, after adjusting for eGFR and spot urine PCR, we still found a significant association between both variables. CKD patients with no anaemia displayed 0.04 times lower odds of having hyperuricemia after relevant confounder adjustment. To the best of our knowledge, there has been no study to depict the relationship between hyperuricemia and anaemia in CKD. In the Artherosclerosis Risk in the Communities (ARIC) study, McAdams-DeMarco et al. [49], after following up 10,791 individuals for a period of 9 years, reported that anaemia was associated with an approximately two-fold increased risk of 'self-reported' Losartan gout independent of renal function. The authors had no clear biological explanation for the link between anaemia and hyperuricemia, but hypothesised that this link may be mediated by oxidative stress. Increased oxidative stress seen in anaemia could cause hyperuricemia by increasing xanthine oxidase activity and increased cell death/turnover [50]. More research is needed to explicate this relationship. Our study had strengths and limitations. It was well designed to capture a wide range of demographic, clinical and biological variables that could influence hyperuricemia in CKD. Furthermore, important variables such as medications and diet that could influence our findings were taken into consideration and accounted for. However the small sample size may have limited statistical power of some analyses. Also, odds ratios on logistic regression overestimate relative risks when the prevalence is high. In addition, the cross-sectional design did not enable us carryout cause and effect analysis for the various associations seen. This was particularly true for the associations seen between hypertension, obesity and anaemia with hyperuricemia.

Conclusions
In this sample of non-dialysed CKD followed-up in referral centres in Cameroon, about 7 out of 10 had hyperuricemia. Hyperuricemia was independently associated with patient's age, eGFR, spot urine PCR, hypertension, ULT, loop diuretics, obesity and anaemia. More studies are required to establish causal relationships between these associations.