Microalbuminuria among Type 1 and Type 2 diabetic patients of African origin in Dar Es Salaam, Tanzania

Background The prevalences and risk factors of microalbuminuria are not full described among black African diabetic patients. This study aimed at determining the prevalence of microalbuminuria among African diabetes patients in Dar es Salaam, Tanzania, and relate to socio-demographic features as well as clinical parameters. Methods Cross sectional study on 91 Type 1 and 153 Type 2 diabetic patients. Two overnight urine samples per patient were analysed. Albumin concentration was measured by an automated immunoturbidity assay. Average albumin excretion rate (AER) was used and were categorised as normalbuminuria (AER < 20 ug/min), microalbuminuria (AER 20–200 ug/min), and macroalbuminuria (AER > 200 ug/min). Information obtained also included age, diabetes duration, sex, body mass index, blood pressure, serum total cholesterol, high-density and low-density lipoprotein cholesterol, triglycerides, serum creatinine, and glycated hemoglobin A1c. Results Overall prevalence of microalbuminuria was 10.7% and macroalbuminuria 4.9%. In Type 1 patients microalbuminuria was 12% and macroalbuminuria 1%. Among Type 2 patients, 9.8% had microalbuminuria, and 7.2% had macroalbuminuria. Type 2 patients with abnormal albumin excretion rate had significantly longer diabetes duration 7.5 (0.2–24 yrs) than those with normal albumin excretion rate 3 (0–25 yrs), p < 0.001. Systolic and diastolic blood pressure among Type 2 patients with abnormal albumin excretion rate were significantly higher than in those with normal albumin excretion rate, (p < 0.001). No significant differences in body mass index, glycaemic control, and cholesterol levels was found among patients with normal compared with those with elevated albumin excretion rate either in Type 1 or Type 2 patients. A stepwise multiple linear regression analysis among Type 2 patients, revealed AER (natural log AER) as the dependent variable to be predicted by [odds ratio (95% confidence interval)] diabetes duration 0.090 (0.049, 0.131), p < 0.0001, systolic blood pressure 0.012 (0.003–0.021), p < 0.010 and serum creatinine 0.021 (0.012, 0.030). Conclusion The prevalence of micro and macroalbuminuria is higher among African Type 1 patients with relatively short diabetes duration compared with prevalences among Caucasians. In Type 2 patients, the prevalence is in accordance with findings in Caucasians. The present study detects, however, a much lower prevalence than previously demonstrated in studies from sub-Saharan Africa. Abnormal AER was significantly related to diabetes duration and systolic blood pressure.


Background
Diabetic nephropathy accounts for a significant reduction in life expectancy of diabetic patients. It is the leading cause of end-stage renal disease in the western world [1,2]. In Caucasians, the cumulative incidence of diabetic nephropathy in Type 1 and Type 2 diabetic patients has been estimated to range from 20% to 27% after a diabetes duration of 20 years [3,4].
Extensive studies in the Western world have demonstrated that diabetic patients with microalbuminuria have increased risk of progression to overt proteinuria, and after some time, renal failure. The progression of diabetic nephropathy from the appearance of clinical proteinuria to end stage renal failure is usually irreversible. Without any intervention, approximately 80% Type 1 patients with persistent microalbuminuria develop overt nephropathy after 10-15 years. Eventually 50% of these develop end stage renal failure within 10 years and 75% by 20 years [5]. In Type 2 diabetic patients, 20-40% with microalbuminuria progress to overt nephropathy and 20 years later, approximately 20% develop end stage renal failure [6].
There is a racial difference in the prevalence of diabetic nephropathy and end stage renal failure. African American patients for instance have been reported to suffer greater diabetic nephropathy and kidney damage than Caucasian Americans [7,8].
Information on nephropathy in African diabetic populations is scarce. Available data between 1971 and 2002 shows the prevalence of microalbuminuria in Africa to vary between 26% and 57% in diabetic patients, with variable duration of the disease [9][10][11][12], [ Table 1] Early medical treatment and lifestyle adjustments have been shown to halt the progression from micro-to macroalbuminuria and eventually end stage renal failure [13,14]. Therefore, detection of microalbuminuria as early as possible in the course of the disease is very impor-tant. In the developing countries, this is even more so because of the economical constraints, kidney replacement therapy is seldom an option.
The aim of this study was to determine the prevalence of microalbuminuria and diabetic nephropathy among Type 1 and Type 2 African diabetic patients in Dar es Salaam. We also wanted to assess the interrelation of microalbuminuria with basic patient socio-demographic features as well as clinical and biochemical parameters.

Methods
Patients attending the Muhimbili outpatient diabetic clinic for their regular visits and who accepted to participate in the study were included. The study was done from July 2003 to March 2004. Informed consent verbally and in writing was obtained from all patients and were interviewed using a standardized questionnaire. The history including date of diabetes diagnosis, family history of diabetes, history of cigarette smoking, age, gender and date of presentation were obtained.
Patients were enrolled in two steps. In step 1, 204 consecutive patients were included. Secondly, to enlarge the group of Type 1 patients, 68 consecutive patients aged 30 years or younger and on insulin therapy were enrolled. Patients were classified into Type 1 and Type 2 according to the WHO clinical stage criteria [15] using information from each patient's diabetes sheet and data collected at enrolment. Patients not fitting into the clinical features of either type were classified as undetermined.

Clinical assessment
Weight was measured without shoes or heavy clothing with a SECCA ® scale. It was recorded to the nearest 0.5 kilogram. Height was determined to the nearest 0.5 cm with a rigid measure against a vertical wall.
Blood pressure was measured with a suitable mercury sphygmomanometer. After a 10 minutes rest with the patient in the sitting position, blood pressure was measured two times at 5 minutes interval. The 1 st and the 5 th Korotokoff's sound were used to determine the systolic (SBP) and diastolic blood pressure (DBP) measurements respectively. The second blood pressure measurement was used as the blood pressure for the individual. Hypertension was defined as SBP ≥140 and/or DBP of ≥90 mmHg or use of hypertensive medication [16].

Blood sample collection and storage
A morning sample of 6-9 mls of venous blood was drawn from each subject and centrifuged within 2 hours of collection. The sera and urine samples were frozen at -80°C and transported to Bergen using a dry ice-filled container.
Capillary blood glucose was measured using a HemoCue AB glucose analyzer (Angelholm, Sweden). HbA 1 c was measured immunochemically at the clinic using DCA 2000 ® + (Bayer Corporation) [17]. The instrument is standardized against the DCCT method [18]. Quality was controlled using standard solutions.

Biochemical assays
Assessment of serum potassium, sodium, serum cholesterol, high-density lipoproteins, and low-density lipoprotein, serum triglycerides, serum creatinine were done at the Haukeland University hospital, Norway. Biochemical tests were analyzed on a BM/Hitachi 917 Auto Analyzer.

Urine collection
Two overnight urine specimens were collected on consecutive nights. All patients received clear oral and written instruction on how to collect a timed overnight urine sample in a 2-liter plastic container with no preservatives. Patients were instructed to empty the bladder and discard the urine just before going to bed and to note the time. All the urine collected after going to bed (sleeping time) and first morning urine sample on waking up was collected into the container and the time of the first morning sample recorded. Sample were delivered to the clinic the same morning. Patients who were unable to comply with the instructions or lived far from the hospital were admitted and samples collected in the hospital. Collected urine were measured and screened for signs of infection using a dipstick (Multistix ® 10SG reagent strips, Bayer Diagnostics, UK) and microscopically. Samples found to be infected were discarded; the patient treated accordingly and after a week of cure allowed to re-submit another sample. Samples were then frozen at -80°C until analysis for albuminuria and urine creatinine in Norway.

Assessment of urinary albumin excretion
Urine albumin concentrations were determined using an automated immunoturbidity method, (anti-human serum albumin obtained from Dako, Oslo) with sensitivity of 2.3 mg/L and inter-and intra assay coefficients of variation of 4.4% and 4.3% respectively (Dade Behring Marburg Gmbh.USA). We used the average of the two readings to determine the final urinary excretion rate (AER). AER was then categorized as normoalbuminuria (AER of < 20 μg/min), microalbuminuria (20-200 μg/ min) and macroalbuminuria (> 200 μg/min). Albumin to creatinine ratio (ACR) was calculated and values ≥ 2.5 mg/ mmol for men, ≥3.5 mg/mmol for women defined as microalbuminuria. Macroalbuminuria was defined as ACR ≥30 mg/mmol for men and women.

Renal function
Creatinine clearance calculations using the Cockcroft Gault Formula were used as a correlate of eGFR [19]. Findings were classified according to stages of renal dysfunction as proposed by the National Kidney Foundation [20].

Ethics
The study was done in accordance with the amended 2000 Helsinki declaration. The Scientific and Publication committee at the Muhimbili University College of Health Sciences (MUCHS) and the Norwegian Regional Medical Research Ethics Committees of Western Norway approved the study protocol.

Statistical analysis
Distribution within groups was tested for skewness and statistically significant differences of standard deviations. Parametric and nonparametric tests were then used as appropriate. Differences between proportions were assessed by Fisher's Exact test or Chi-square test. Multiple linear regression was performed with ln AER as the dependent variable. Data were analyzed using SPSS version 12 for Windows and GraphPad INSTAT-version 3.00 supplied by the GraphPad Software, Inc San Diego, USA. P < 0.05 was considered statistically significant.

Albuminuria and the associated risk factors
The median diabetes duration, SBP, DBP and proportion of patients on anti-hypertensive drugs among Type 2 patients with abnormal AER was significantly higher than for those patients with normal AER were (p < 0.001), [ Table 3].
The differences in age at inclusion, diabetes duration, SBP, DBP, HbA1c, BMI, and creatinine clearance distribution between patients with normalbuminuria and abnormal AER levels were not statistically significant in Type 1 patients.
Type 2 patients with known diabetes duration of more than 5 years had significantly larger proportion of patients with abnormal AER levels (p < 0.05), higher DBP (p < 0.05), and poorer glycaemic control (p < 0.001) than those patients with diabetes duration of 5 years or less. No significant differences were noted in all the variables in both duration ranges among Type 1 patients [ Table 4].
Only six subjects (2.4%) were current smokers, 11.5% had stopped smoking and 86.1% had never smoked. All patients who were current smokers had normal albumin excretion rates and only three of those who had stopped smoking had abnormal AER.

Regression analysis
Linear regression analysis was performed with log(n) AER as the dependent variable and diabetes duration, age at inclusion, gender, BMI, HbA1c, SBP and DBP, serum creatinine and cholesterol as covarites. Among Type 1 diabetic patients a simple linear regression analysis demonstrated a significant association between diabetes duration (p = 0.045). Among Type 2 patients, a simple linear regression analysis revealed diabetes duration (p < 0.001), SBP (p < 0.001), DBP (p = 0.007) and serum creatinine (p < 0.001) to be correlated with log AER [ Table  5].
A backward stepwise multiple linear regression analysis by elimination procedure was used to identify factors that might independently be associated with development of albuminuria. Diabetes duration (p < 0.001) was picked as the strong predictor for the development of AER and SBP (p = 0.010) and serum creatinine (p = 0.016) for Type 2 diabetes but no variable for Type 1 diabetes was picked on multiple regression analysis.

Discussion
This cross sectional study reveals a prevalence of microalbuminuria and its determinants in a selected population from a diabetic clinic in Dar es Salaam, Tanzania. Prevalence for microalbuminuria (12.1%) among Type 1 patients in this study who had median diabetes duration of 3.0 years was relatively high when compared with prevalences found among Type 1 Caucasians diabetic patients with similar diabetes duration. In a longitudinal study, a cumulative incidence of about 4% was found in 148 Type 1 diabetic patients after diabetes duration of three years [21]. Other studies in Caucasians have reported prevalences of microalbuminuria in Type 1 patients from 7-22% [22][23][24] however, in these studies disease duration was much longer than in the present study. Among the Type 2 patients in our study microalbuminuria was 9.8%, and median diabetes duration 4 (0-25) years. This is comparable with findings in Caucasian Type 2 patients (8-32%) [22,23,[25][26][27].
This study is one of relatively few from sub-Saharan Africa dealing with microalbuminuria. A Medline search on microalbuminuria in diabetic patients revealed four comparable studies from sub-Saharan Africa [ Table 1] which display relatively high prevalences compared with our findings. Screening for bacteriuria was done in these studies too. They used mainly the "MICRAL test" for urinary albumin assessment. The MICRAL test is less sensitive  Tests - ‡ ‡Fischer's exact test#, ‡Mann Whitney test, # Unpaired t test P value -*p < 0.05, **p < 0.01, ***p < 0.001  than the immunoturbidimetry method employed in our study. The way of assessing albuminuria in this study should not have caused lower prevalence of microalbuminuria but rather the opposite. Some studies have shown decay in albumin concentration when urine samples were kept frozen at -20°C, but not with storage temperature of -80°C as was applied in this study [28][29][30].
Even if freezing and thawing should have caused an albumin decay in our urine samples as well, this reduction could hardly account for the big difference in albumin excretion rates demonstrated between the present study and the former studies from sub-Saharan Africa. Differences in design, setting, methods for collection of urine and assessment of microalbuminuria, and patient populations may contribute to the differences found between these African studies.
From table 1, one might get the impression of a decreasing trend in the prevalence of microalbuminuria in Africa. Decreasing prevalences of microalbuminuria and microvascular complications have been reported from western countries, which is attributed to improvement in metabolic control, better treatment of hypertension, the use of angiotensin converting enzyme inhibitors and angiotensin receptor blockers [31,32]. Even though tempting, it is not appropriate to draw any firm conclusion, from the studies listed in table 1, as to whether prevalence of microalbuminuria really is on a decrease in sub-Saharan Africa. This is especially so since the use of angiotensin receptor blockers and converting enzyme inhibitors is limited.
It is uncommon in Caucasian Type 1 patients to diagnose microalbuminuria during the first five year [33,34]. In the study by Marshall et al, only one patient with Type 1 diabetes of less than 5 years duration had an AER in the microalbuminuria range [23]. Among Type 2 Caucasian patients, the association of diabetes duration and microalbuminuria is equivocal [35,36]. In the current study, patients with abnormal AER had longer known diabetes duration than those with normal AER (4.2 versus 3 years), for Type 1 and (7.5 versus 3 years, p < 0.001) for Type 2.
On regression analysis, known diabetes duration was identified as a strong predictor for development of abnormal albuminuria in Type 2 patients.
Poor glycaemic control is well defined contributor to the development and progression of microalbuminuria among Type 2 patients [35][36][37][38], as well as Type 1 patients [22,39,40]. In contrast, most of the available studies from sub-Saharan African including the present study fail to demonstrate any significant relationship between the level of glycaemic control and microalbuminuria. Goldschmid et al demonstrated a high prevalence of microalbuminuria and nephropathy among African-American diabetic patients, they also showed that metabolic control was not associated with the disease [7].
Elevated blood pressure is documented as the most significant contributing factor in the pathogenesis and the progression of abnormal AER and eventually development of diabetic nephropathy in both Type 1 and Type 2 diabetic patients [41][42][43]. In the present study, the median SBP and DBP were lower among Type 1 than among Type 2 patients. When SBP and DBP of Type 1 patients with normal AER were compared with those with abnormal AER there was no significant differences. In contrast however, Type 2 patients with abnormal AER had statistically significant higher SBP and DBP than those with normal AER, (p < 0.0001).
The number of Type 1 diabetic patients in this study was limited and few had microalbuminuria. These limitations probably account for the lack of statistical correlation to the commonly acknowledged risk factors for microalbuminuria. Other limitations of the current study include those inherent to patient reliability or compliance in com- plete urine collection and recording correctly the timed urine collections. Incomplete urine collection will have an effect of lowering the calculated AER and eventually underestimating the actual prevalence of microalbuminuria. This is however not supported by the calculated ACR from the same sample that showed an agreement between ACR and AER, the kappa coefficiency was 0.85, 95%CI of (0.76, 0.95) which is considered to be a very good level of agreement.

Conclusion
The occurrences of microalbuminuria detected in this study are much lower than previously reported in studies from sub-Saharan Africa and similar to what have been found in Caucasian diabetic patients. In Type 2 diabetic patients the duration of diabetes was the strongest predictor; SBP and serum creatinine as well predicted increased albumin excretion rate.