The study sample was representative of the larger community for age, sex, education and employment status [23
]. That is, the sample had similar proportions of the previously listed characteristics as the study population. Nine subjects were ineligible due to incomplete data and 6 were excluded from the albuminuria analysis due to persistent hematuria, leaving 468 subjects for albuminuria analysis. Participants with persistent hematuria were excluded due to the inability to obtain an ACR and we could not determine the cause of the hematuria. These individuals were referred for clinical evaluation. However, most individuals whose first test was positive for hematuria were found not to be hematuric on subsequent tests and were included in analysis. The majority of these participants were women who were tested close to their menstrual period.
Demographic and health risk characteristics of the study population are described in Table . Both sexes were equally represented. The population was young and the mean age was similar for men and women (37 for men; 38 for women). Less than half the study sample had completed grade 9. The study population had high rates of cardiovascular risk factors; three-quarters of participants were current smokers, 43% had hypertension, 35% had diabetes or prediabetes and obesity was highly prevalent.
Demographic and health risk characteristics of the study sample
Twenty-five percent of the sample returned for at least a second random urine test (Table ). However, of those who had abnormal samples (ie. hematuria or proteinuria) on their first test, 79% returned for a second or third test. Albuminuria was present in 95/468 or 20% of participants. Twenty-five participants (5%) had proteinuria and 70 (15%) had microalbuminuria. In univariate analysis [OR (95% CI) p-value], age [1.039 (1.008, 1.072) p
0.014], diabetes [11.474 (4.209, 31.279) p
0.001], systolic blood pressure [1.039 (1.019, 1.061) p
0.001], diastolic blood pressure [1.051 (1.018, 1.086) p
0.002], fasting glucose [1.257 (1.155, 1.367) p
0.001] and hypertension [4.390 (1.709, 11.277) p
0.002] were associated with proteinuria. There was a non-significant trend for lower odds of proteinuria among females [0.428 (0.181, 1.012) p
0.053]. BMI, ever smoker, and years diagnosed with diabetes were not associated with proteinuria. For all other analysis, those with proteinuria and microalbuminuria were combined for analysis due to small numbers and equivalence on the following characteristics: age, sex, smoking status and mean years smoked. However, the proportion of those with diabetes was greater among the proteinuria group compared to the microalbuminuria group. Eighty percent (20/25) of those with proteinuria had diabetes.
Albuminuria category by number of tests done
Participants with and without albuminuria are compared in Table . Those with albuminuria were older (p
0.001) and albuminuria prevalence progressively increased with age (Figure ). Men were more likely to have albuminuria than women (p
0.01). Interestingly, those with albuminuria were less likely to report ever having smoked (p
0.007). However, neither years of smoking nor pack-years smoked were significantly different between groups. Participants with albuminuria were more likely to be hypertensive by any measure. Seventy percent (70%) of those with albuminuria had hypertension compared with 36% of those without albuminuria (p
0.0001). In this regard, mean blood pressures were significantly higher in the group with albuminuria compared to those without. Target blood pressure in those with albuminuria is generally recommended to be <130/80
mm Hg, however, this was achieved in only 34% of those with albuminuria.
Comparison of those with and without albuminuria
Figure 1 Percent albuminuria by age categories. There is a significant linear association between age group and proportion of those with albuminuria (p<0.001). ‘n’ refers to the total number of participants in respective (more ...)
Participants with albuminuria were significantly heavier and had a higher BMI (Figure ). This increased weight and BMI may be related to the greater proportion of participants in the albuminuric group with diabetes and impaired fasting glucose. Fasting glucose was also higher among those with albuminuria compared to those without (p
0.001). While almost 60% of those with any degree of albuminuria had diabetes, 42% of those with diabetes were albuminuric.
Figure 2 Percent albuminuria by BMI categories. There is a significant linear association between BMI category and proportion of those with albuminuria (p<0.001). ‘n’ refers to the total number of participants in respective (more ...)
Odds ratios and 95% CI are listed in Table for both univariate and multivariate logistic regression. Backwards stepwise multivariate logistic regression was performed. Univariate analysis indicated the odds of having microalbuminuria increased 4% for each year increase in age. However, age was not significant in the final multivariate model. When both fasting glucose and years of diabetes were removed from the model and age was included, age became significant suggesting high multicolinearity between the diabetes and age (model not shown). Females were less likely to have microalbuminuria than males, which remained significant in the multivariate model. Smokers were less likely to have albuminuria than non-smokers and while smoking appeared protective in univariate analysis, it did not achieve significance in the multivariate model.
Univariate and multivariate logistic regression with albuminuria as an outcome
Those with diabetes were seven times more likely to have albuminuria than those without diabetes. In addition, for each mmol/L increase in average fasting glucose, albuminuria increased by 27% in the univariate model, and 20% when controlling for sex, years diagnosed with diabetes, systolic blood pressure and BMI. Length of time with diabetes was also a significant predictor of albuminuria and confirms the natural history of diabetic nephropathy. Systolic blood pressure was independently associated with albuminuria in the multivariate model and risk of albuminuria increased by 2% for every mm Hg increase in systolic blood pressure. Lastly, weight and BMI were both significant alone, but BMI was an independent predictor of albuminuria in the multivariate analysis, after accounting for average fasting glucose, gender, years diagnosed with diabetes mellitus and systolic blood pressure [1.042 (1.001, 1.085), p
In contrast to participants with diabetes and hypertension, very few with albuminuria were aware of their condition. Participant awareness of kidney disease increased with severity such that 17% of those with proteinuria were aware of kidney disease while awareness kidney disease in those with microalbuminuria was only 1% (Table ).