Asymptomatic bacteriuria in sickle cell disease: a cross-sectional study

The definition of asymptomatic bacteriuria (ASB) is controversial as some have defined it as the quantitative growth of bacteria, greater than or equal to 10⁵ colony forming units per milliliter urine of the same organism, on aseptically collected midstream urine specimens, in the absence of symptoms of urinary tract infection on two or more consecutive occasions [1, 2] while for others a single occasion is sufficient [3‐5]. The prognostic significance of ASB resides in the observation that persons with ASB in certain medical conditions, such as diabetes mellitus and pregnancy, are at increased risk of pyelonephritis and renal impairment [5‐7].

In sickle cell disease renal disease and dysfunction are common and increase morbidity and mortality [8, 9]. Additionally, symptomatic urinary tract infection (UTI) is associated with painful crisis, bacteraemia, pneumonia, and osteomyelitis in SCD [10‐13]. Whether ASB is a prelude to more serious renal disease and dysfunction in SCD is unclear. This present study was undertaken to estimate the prevalence and possible risk factors of ASB in the Jamaican sickle cell cohort.

A total of 267 patients attended the 2004 cohort review. One person refused to participate in the study. Thus 266 patients had screening clean-catch MSU (Figure 1), of which 50% were male. One woman was pregnant at the cohort review visit; and another became pregnant during follow-up of the first abnormal MSU.

As expected there were significant differences in the haematological values by genotype with the mean steady state haemoglobin levels of HbSS group being significantly lower than sickle cell-haemoglobin C group, sickle cell-β⁺ thalassaemia, and sickle variants but having higher steady state nucleated cell counts (Table 1). The steady state serum creatinine values were significantly lower in the HbSS group compared with other genotypes except the sickle variant group. In addition the mean uric acid values were lower in the HbSS group compared with sickle cell-β⁺ thalassaemia group (Table 1).

There were 32 patients with abnormal first MSUs: 29 were positive MSU samples growing identifiable pathogens (Figure 1), 2 were grossly contaminated and one grew bacteria of doubtful significance (growth of a urinary tract pathogen between 10⁴ and 10⁵ per mL). Of the 29 positive cultures 15 re-grew the same organisms in the next culture but one of these grew the same organism (E. coli) but with a different sensitivity pattern and was therefore classified as a probable ASB. Thus 14 samples satisfied the criteria for a confirmed episode of asymptomatic bacteriuria (Figure 1) yielding a prevalence of 5.3% (95% CI 0.8% to 29.4%). Fifteen subjects were classified as probable ASB. The probable ASB group consisted of seven samples which grew organisms that were different from the first MSU sample, 2 grew bacteria in insignificant quantities, four were no growth and two were grossly contaminated specimens (Figure 1). None of the patients were on antibiotic therapy during this study and all of the patients with identified ASB had at least one follow-up visit.

Sixteen of the 29 positive first MSU samples (55%) grew E. coli in significant quantity. This was the commonest organism cultured in this study. Other urinary tract pathogens identified in this study included: group B Streptococcus, Enterobacter aerogenes, coagulase negative Staphylococcus, Enterobacter cloacae, Proteus mirabilis, Microaerophilic streptococcus and Bacteroides (mixed), group B Streptococcus and Streptococcus viridans (mixed), Enterobacter species, group D Streptococcus, Klebsiella pneumoniae, and Enterobacter koseri. All of these organisms were grown in pure cultures with only the two exceptions listed above. Fifteen of the patients with probable ASB (52%) had evidence of a host response, as evidenced by pyuria. Of those with confirmed ASB, 13 of 14 patients (93%) had a host response as evidenced by pyuria in the second midstream urinalysis.

There were no significant differences for mean haemoglobin and nucleated blood cell counts between the confirmed ASB group and sterile urine group (86.4 ± 15.2 g/L vs. 92.6 ± 17.4 g/L and 14.0 ± 4.5 × 10⁹/L vs. 13.4 ± 4.1 × 10⁹/L) respectively. Similarly there were no differences in mean serum creatinine and uric acid values between the confirmed ASB group and sterile urine group (Table 2). Twelve and a half percent (1/8) of patients with ASB had proteinuria of trace or greater on dipstick urinalysis. However there was no association between having proteinuria measured by dipstick and ASB (Table 2).

The proportion of subjects with HbSS in the confirmed ASB group (64%) was not statistically different from the proportion of HbSS in the sterile group (55%). However controlling for genotype, females were 14.7 times more likely to have confirmed ASB compared to males (95%CI 1.8 to 121.0) but having a recorded history of ever having a symptomatic UTI was not associated with current ASB status (Odds ratio 2.5, 95%CI 0.79 to 7.76). The numbers of recorded episodes of gram negative sepsis were few. Notwithstanding, there were 148 clinic visits for symptomatic urinary tract infections (Table 2). Compared to the sterile urine group the expected number of visits for symptomatic UTI in the confirmed ASB group is increased by a factor of 2.5 (95% CI 1.4 to 4.5, p < 0.005).

The relationship between the various predictors and probability of having ASB were explored with a series of logistic regression models. Likelihood ratio tests were used to determine which variables should be incorporated into a final regression model with ASB as the response variable. The results indicated that the absence of the variables HbSS genotype (coded as 1 for homozygous β^s, 0 otherwise), steady state haematology, and serum creatinine from the full model which included these variables along with age, gender (coded as 1 for female, 0 otherwise) and steady state uric acid concentrations as predictors did not significantly reduce the log likelihood. Thus the model which best predicted ASB from the data included the predictors age OR 1.37 (95 % CI: 0.99 1.64), female gender OR 15.3 (95% CI: 1.94, 120.1), and uric acid concentration OR 21.5 (95% CI: 0.64, 722.2) (Table 3).

The prevalence of probable ASB was 10.9% and the prevalence of confirmed ASB was 5.3% with 95% CI 0.8% to 29.4% in this sample of Sickle Cell patients. These prevalence rates are comparable to other prevalence estimates of ASB in non-sickle cell healthy populations [5, 16]. The major determinants of ASB status were gender, age and steady state serum uric acid concentrations. Additionally there were increased numbers of recorded clinical visits for symptomatic UTI in the confirmed ASB group. The effect of genotype status on these relationships was not statistically significant.

The strength of this study is related to the cohort design which resulted in the selection of a representative population based sample. This reduces the ascertainment bias associated with symptomatic recruitment from a clinic based population. In addition, the participation rates at this cohort review were >99%, thus strengthening our confidence in extrapolating our findings to the wider sickle cell population. However the small number of patients identified with confirmed ASB represents a limitation of the present study. This was partly due to our conservative definition of confirmed ASB. The consequence of having a small numerator is that the precision of our prevalence estimate is affected and this is reflected in a wide 95% confidence interval.

The sexual dimorphism in the prevalence of ASB has been reported before in healthy adults [6]. It is thought that this may be related a relative deficiency of secretory IgA antibody response from the mucosal surface in the urogenital tract of females compared with males [6].

In non pregnant females the presence of confirmed ASB has a high sensitivity and specificity of predicting a future symptomatic episode of UTI [5, 6]. However in the absence of anatomical or functional abnormalities of the urinary tract, ASB per se is not associated with renal scarring, renal dysfunction or hypertension [17]. In adults with SCD, anatomical abnormalities such as renal cortical scarring, enlarged glomeruli, vascular disorganization of the medulla and renal functional abnormality such as increased effective renal blood flow and hyperfiltration are common [8]. These pathophysiological changes are thought to contribute to sickle cell nephropathy and the high prevalence of renal failure observed in subjects with sickle cell disease [8]. In this context, one might expect persons with ASB to have more UTI and worse renal function. Thus the finding of an increased number of expected visits for symptomatic UTI in subjects with ASB compared to those without as well a trend for a significant association between ASB status and a history of UTI is not surprising. On the other hand, the absence of significant differences in measures of renal function between those with ASB and those without ASB was unexpected. However, serum creatinine and uric acid are not sensitive measures of renal function as glomerular filtration rate has to fall by up to 50% before there are changes in serum creatinine [18, 19]. Additionally variation in creatinine production due to differences in muscle mass may alter serum creatinine levels independent of renal functional status [20, 21]. Notwithstanding, further prospective studies are needed to define the renal risk posed by ASB in sickle cell disease and to inform therapeutic strategies.

In sickle cell disease, the increased purine turnover associated with accelerated erythropoesis results in the greater production of uric acid [22]. This is counterbalanced by increased renal clearance of urate. It is possible that the increased renal clearance may play a predisposing role for ASB by promoting the formation of microcrystals which damage the uroepithelium thereby facilitating colonization.

The infecting organisms identified in this study are in keeping with commonly isolated bacteria in other studies [1‐3, 5, 10, 11]. The presence of bacteria in the urine in the absence of a major inflammatory response and hence symptoms, suggest that there are alterations in the host-pathogen interaction. For example it has been reported that uropathogenic strains of microbes have evolved mechanisms to promote survival in the urinary space. Features of these adaptive mechanisms include presence of adhesion which promote attachment to the urogenital tract, and the production of factors such as α-haemolysin, and cytotoxic necrotizing factor 1 (CNF1) [6]. On the other hand the host factors in SCD that increases susceptibility to bacterial colonization are unclear but in Diabetes Mellitus altered leukocyte function contribute to the susceptibility to ASB in this disease [23].

In summary ASB is a significant problem in individuals with SCD and is associated with UTI. However, further research is needed to determine the clinical and renal significance of ASB in SCD.