The evaluation of the efficiency of Atellica UAS800 in detecting pathogens (rod, cocci) causing urinary tract infection

Introduction

Urinary tract infections (UTIs) are one of the most common infections among outpatients and hospitalized patients [1]. Gram-negative bacteria, especially Escherichia coli and Klebsiella spp., are the most common causative pathogens [2]. Urine culture is the gold standard method in the diagnosis of UTI. The causative bacteria are detected and the appropriate antibiotic to be used in the treatment is determined with the antibiogram with the urine culture. However, this method requires more time for diagnosis and starting of the treatment. For this reason, complete urine analysis (CUA), in which chemical and microscopic sediment analysis of the urine is performed, is widely used as a screening test in the diagnosis of UTIs. The results can be given to the patients on the day of urine collection because the turn around time (TAT) is quite short in CUA. In this way, negative samples for UTI can be excluded and prevent extra urine culture test requests. When waiting for the culture results, choosing an effective treatment according to the CUA result will allow the patient to be treated quickly, providing social benefits in terms of preventing resistance to improper antibiotic use. However, in manual microscopy, steps such as centrifugation, decantation, and resuspension cause cell fragmentation and cell losses [3]. Interindividual variation is also quite wide in manual microscopy. CUA devices, in which urine microscopic sediment analysis is performed with automatic devices, have been developed by manufacturers in the last two decades. Clinical laboratories use CUA analyzers, which use different technical systems, widely [4]. This has brought improvements in efficiency, standardization, and accuracy in most laboratories [5]. Atellica UAS800 Device which was developed by Siemens Healthineers is a fully automatic urine microscopy device in which sediment analysis is performed with a digital camera. Atellica UAS800 evaluates 14 items, including erythrocytes, leukocytes, total bacteria, hyaline cylinders, squamous epithelial cells, non-squamous epithelial cells, crystals, leukocyte clusters, pathological cylinders, yeasts, mucus, and sperm, as well as rods and cocci. It is one of the advanced devices in differentiating urinary tract pathogens into rods and cocci, but its performance has not been evaluated yet.

For this reason, the present study aims to evaluate the bacteria detection and bacterial morphological discrimination ability of the Atellica UAS800 automated microscopy analysis system.

Materials and methods

Results

Cut-off values ​​for urinary bacterial infection and negative urine culture

Diagnostic performance evaluation of leukocyte, bacteria, rod and cocci results for different cut-offs on the Atellica UAS800 microscopy analyzer was performed by ROC analysis and the AUC was calculated as 0.783, 0.743, 0.865, and 0.652 respectively (Figures 1–3). Bacterial growth was detected in the culture of 121 (21.9 %) urine samples in 551 male patients. The AUC in the ROC analysis for Atellica UAS800 leukocytes, total bacteria, rods, and cocci was calculated as 0.779, 0.693, 0.859, and 0.651 respectively in male urine samples. There was growth in 171 (33.8 %) urine samples in 505 female patients, and in the AUC, leukocytes, total bacteria, rods, and cocci were calculated as 0.775, 0.772, 0.867, and 0.655 respectively in the ROC analysis in Atellica UAS800.

Figure 1:

ROC curve of Atellica UAS800 for leukocytes and total bacteria. The AUC for leukocytes and total bacteria are 0.783 and 0.743 respectively.

Figure 2:

ROC curve of Atellica UAS800 for rod in 1,056 urine samples. The AUC for rod is 0.865.

Figure 3:

ROC curve of Atellica UAS800 for cocci in 1,056 urine samples. The AUC for cocci is 0.652.

The bacterial results were compared with urine culture results with different cut-off values in Atellica UAS800, and sensitivity, specificity, PPV, and NPV are given in Tables 2. Sensitivity, specificity, PPV, and NPV values obtained by comparing the results of Atellica UAS800 rods with the results of bacteria grown in urine culture are given in Table 3.

Table 2:

Diagnostic performance of Atellica UAS800 at different cut-off values for total bacteria compared to urine culture results.

Bacteria cut-off (cells/HPF)	%Sensitivity	%Specificity	%PPV	%NPV	TP (n)	FP (n)	TN (n)	FN (n)
Total
65	67.5	74.2	50	85.6	197	197	567	95
70	65.4	76.8	51.9	85.3	191	177	587	101
75	64	79.2	54	85.2	187	159	605	105
Female
75	73,1	72.5	57,6	84.1	125	92	243	46
80	71,3	76.7	61	84	122	78	257	49
85	68.4	78.5	61.9	83	117	72	263	54
Male
75	51.2	84.4	48.1	86	62	67	363	59
80	48.8	86.5	50.4	85.7	59	58	372	62
85	47.1	88.4	53.3	85.6	57	50	380	64

1. HPF, high power field; PPV, positive predictive value; NPV, negative predictive value; TP, true positive; FP, false positive; TN, true negative; FN, false negative.

Table 3:

Diagnostic performance of Atellica UAS800 at different cut-off values for rod compared to urine culture results.

Rod cut-off (cells/HPF)	%Sensitivity	%Specificity	%PPV	%NPV	TP (n)	FP (n)	TN (n)	FN (n)
Total
15	70.4	91.1	71.8	90.5	181	71	728	76
20	67.3	93.1	75.9	89.9	173	55	744	84
25	63.4	94.4	88.9	81	163	45	754	94
Female
25	72	91.1	78.5	87.8	113	31	318	44
30	70.1	93.7	83.3	87.4	110	22	327	47
35	67.5	94.6	84.8	86.6	106	19	330	51
Male
5	71	86.9	54.6	93.1	71	59	392	29
10	64	93.6	68.8	92.1	64	29	422	36
15	57	95.3	73.1	90.9	57	21	430	43

1. HPF, high power field; PPV, positive predictive value; NPV, negative predictive value; TP, true positive; FP, false positive; TN, true negative; FN, false negative.

The compatibility of the culture results with the Atellica UAS800 was compared, showing good agreement at 70 cells/HPF for bacteria and 20 cells/HPF for rods, with Cohen’s kappa coefficients of 0.419 and 0.628, respectively.

Discussion

Although urine culture is the gold standard method in the diagnosis of UTIs, it takes time as long as 24–48 h. For this reason, the general approach in the diagnosis and treatment of UTIs is to start antibiotics treatment according to the results of the CUA as a screening test and to change the treatment according to the culture result. Improper use of antibiotics has led to increased antibiotic resistance. Especially in recent years, clinicians would like to know the discrimination of the causative agent at least as cocci or bacilli in order to choose the appropriate empirical antibiotic treatment until the culture report is released.

The results of the study showed that UTIs can be differentiated from the results of leukocytes, total bacteria, and rods evaluated on the Atellica UAS800 microscopy analyzer. According to Hosner et al., if the AUC obtained in ROC curve analysis is between 0.7 and 0.8, the discrimination is acceptable, if the AUC is >0.8–09, this indicates excellent discrimination, and AUC >0.9 indicates exceptional discrimination [7]. In the present study, the order of AUC for leukocytes, total bacteria, and rods in ROC curve analysis were 0.783, 0.743 and 0.865. Similar results were obtained when male and female urine samples were evaluated. The AUC for leukocytes, total bacteria, and rods for male were 0.779, 0.693, and 0.859 respectively, and AUC for female were 0.775, 0.772, and 0.867 respectively.

In our study, regardless of gender, the cut-off value for bacteria was determined as 70 cells/HPF (sensitivity was 65.4 %, specificity was 76.8 % and NPV was 85.3 %) in 1,056 urine samples and we found that this value was most compatible with culture (Cohen’s kappa coefficient: 0.419). When urine samples were evaluated separately according to gender, it was seen that the threshold values were similar in both genders. The cut-off value was 80 cells/HPF for females (sensitivity: 71.3 %, specificity: 76.7 %, NPV: 84 %, and Cohen’s kappa coefficient: 0.461), and it was 80 cells/HPF for male patients (sensitivity: 48.8 %, specificity: 86.5 %, NPV: 85.7 %, and Cohen’s kappa coefficient 0.357). In previous studies that were conducted with different urine sediment analyzers, it was reported that cut-off values for leukocytes and bacteria should be determined according to different subgroups, in accordance with our study [8].

In the literature, few studies evaluate the diagnostic performance of the Atellica UAS800 microscopy analyzer. Nilker et al. reported sensitivity and specificity for bacteria as 91 and 76 %, respectively in their study with Atellica UAS800 and urine culture [9]. The sensitivity and specificity obtained in the study of Nilker et al. were different from our study which may be related to the number of urine samples evaluated. Although 1,056 urine samples were compared in the present study, Nilker et al. compared a total of 65 urine samples obtained from patients. In their study, Stijn et al. found sensitivity for total bacteria to be 73 %, specificity to be 71 %, sensitivity for leukocytes to be 56 %, and specificity to be 84 % on the Atellica UAS800 Microscopy Analyzer [10]. Similar sensitivity and specificity values were obtained for the total number of bacteria in the present study. Unlike the studies of Nilker et al. and Stijn et al., the results of the rods evaluated on the Atellica UAS800 microscopy analyzer were also compared with the results of the bacterial species grown in the urine culture in the present study. The sensitivity and specificity results at the cut-off value for total bacteria in the present study were similar to those of other analyzers that used sediment analysis by digital microscopy [11], 12].

Lezzi and colleagues evaluated the diagnostic performance of the Atellica1500 microscopy analyzer in their study. They compared bacteria, leukocyte count, leukocyte esterase and nitrite results with urine culture results in 5,490 patients. In the ROC analysis, while their leukocyte results closely resembled those of our study, they reported bacterial outcomes as superior to ours (AUC for leukocytes and bacteria were 0.861 and 0.911, respectively) [13]. This may be due to differences in the number of patients between studies. Additionally, as stated by Lezzi et al., it may be due to the fact that the presence of amorphous crystals affects the number of bacteria both positively and negatively depending on the instrumentation used. In fact, digital microscopy software may recognize small crystals as bacteria (false positive results), or the presence of crystals may cause the device to fail to count bacteria (false negative results) [13]. Our study’s scope was limited by the lack of crystal evaluation in urine microscopic analysis.

When the Atellica UAS800 microscopy analyzer rod results were compared with the urine culture results, the sensitivity, specificity, and NPV were found to be 67.3 %, 93.1%, and 89.9 %, respectively, regardless of gender, when the cut-off value was taken as 20 cells/HPF. When the cut-off value of male urine samples was determined as 10 cells/HPF, the sensitivity was 64 %, the specificity was 93.6 %, the NPV was 92.1 %, and Cohen’s kappa coefficient was 0.592. When the cut-off value for female urine samples was 30 cells/HPF the sensitivity was 70.1 %, the specificity was 93.7 %, the NPV was 87.4 % and Cohen’s kappa coefficient was 0.667. We found that when the cut-off value for rods was determined as 20 cells/HPF, the best compatibility with the culture was achieved (Cohen’s kappa coefficient: 0.628). As expected, the cut-off value that was determined for leukocytes, and rods in female patients was higher than for male patients in our study because of anatomical differences.

In the present study, when compared with total bacteria and leukocytes, it was found that the NPV values obtained for the rod were higher, which shows that the evaluation of the rod status in the urine with the Atellica UAS800 microscopy analyzer will be more helpful than the evaluation of total bacteria and leukocytes in the selection of the antibiotic to be administered to patients. In this way, patients with UTIs can be treated more effectively with the right antibiotic before the culture test results. Increasing antibiotic resistance in recent years complicates the treatment of infectious agents. Improper use of antibiotics is one of the most important reasons for the spread of antibiotic resistance. It was shown in the present study that the prevalence of resistance to antibiotics that are commonly prescribed in primary care is high in children with urinary tract infections caused by E. coli, and it was also found that this would render some antibiotics ineffective in the first-line treatment for urinary tract infections [14].

The limitations of the study were that a microscopic examination was not performed to detect the bacteria in the urine samples due to a high number of samples processed in our microbiology laboratory. There is a need for future studies that include especially large numbers of contaminated urine samples. Diluted urine samples and alkaline urine pH may cause the destruction of leukocytes and make it difficult to identify leukocytes in microscopic analysis. For this reason, the fact that urine specific gravity and pH were not analyzed in our study constitutes another limitation. In our study, the Atellica UAS800 bacteria, rod and cocci counts and the clinical symptoms and culture results of the patients were not evaluated together. Therefore, the number of patients with asymptomatic bacteriuria and the cut-off value in these patients could not be determined, which is another limitation of our study. Atellica UAS800 urine microscopy device was compared with urine culture in the diagnosis of UTI in this study. However, in recent years, many new generation devices using different technologies for urine microscopy analysis have been developed. One of the limitations of the study is that urine culture was compared with a single urine microscopy analyzer.

As a result, the separation of total bacteria, leukocytes, and rods can be made at an acceptable level with the Atellica UAS800 automatic microscopy analyzer. The findings of the present study showed that different cut-off values must be determined for total bacteria and rods according to gender. As we know this is the first study in which bacterial species causing UTIs were compared with urine cultures. With these aspects, we believe that the study will guide clinicians in the diagnosis and treatment of UTIs. The effectiveness of empirical antibiotic treatment that will be given according to rod and cocci results of the Atellica UAS800 urine microscopy analyzer should be investigated in future studies.