The Use of Serum Procalcitonin in the Setting of Infected Ureteral Stones: A Prospective Observational Study

Introduction

Urinary sepsis in the setting of an obstructing ureteral stone is a urologic emergency, requiring urgent decompression with either a retrograde ureteral stent or a percutaneous nephrostomy tube. The mortality rate caused by urinary sepsis has been reported to be as high as 16%, highlighting the importance of early identification and intervention. ¹ Accurately determining which patients have infected urine and thus are at risk for sepsis is essential in helping guide early intervention. In those who are infected, emergent decompression is indicated to prevent potentially life-threatening sepsis. However, if the urine is not infected, then decompression will potentially subject the patient to an unnecessary procedure with its associated risks and costs.

Unfortunately, symptoms do not always correlate with the severity of disease and many patients presenting to the emergency room (ER) have equivocal signs of infection. The gold standard for diagnosing a urinary tract infection (UTI) is a urine culture (UCx) that takes 24 to 48 hours to give a result. The decision to intervene vs observe these patients must be made before the culture results are available. Currently, initial clinical judgment is based on vital signs and laboratories including white blood cell count (WBC) and urinalysis (UA). ^2,3 Unfortunately, these tests are not always reliable in diagnosing a UTI. WBC is nonspecific as it can be elevated with any inflammatory process as is the case for a sterile obstructing ureteral stone. Furthermore, UA can be misleading in the setting of sterile pyuria, an infection with a non-nitrite producing organism, or if the upper urinary tract is completely obstructed, effectively blocking infected urine from reaching the bladder. ⁴ This places physicians in a diagnostic dilemma when patients with an obstructing ureteral stone presents with an equivocal work-up for infection. Hence, there is a need for better biomarkers to help guide early treatment in this patient population.

Procalcitonin (PCT), a precursor of the calcitonin protein, is an acute phase reactant that is often elevated during the systemic response to infection. ⁵ A variety of biologic functions of PCT have been described, including vasomotility and the expression of cytokines. ⁶ Assicot et al. first reported the association between elevated PCT and sepsis in 1993 (Ref. ⁷ ). Subsequent studies have shown that it is particularly useful in sepsis of bacterial origin. ^8,9 Bacterial infection leads to abundant secretion of PCT by mononuclear cells. ¹⁰ This has led to the Food and Drug Administration approving its use in assessing critically ill patients for risk of progressing to severe sepsis. It is an ideal biomarker as it is detected in blood within 2 hours of endotoxin exposure and has a half-life of 22 to 26 hours, which is more rapid than other infectious markers such as C-reactive protein (CRP). ^11,12 PCT value >0.5 ng/mL has a high positive predictive value (PPV) to rule in sepsis while a value <0.25 has a high negative predictive value (NPV) to rule out sepsis. The degree of PCT elevation has also been shown to correlate with clinical outcomes in septic patients, with higher levels predictive of worse outcomes. ⁶ PCT has an established role in septic patients and it has also been shown to be useful in other clinical situations. A study by Christ-Crain and coworkers looked at patients presenting to the ER with respiratory symptoms and showed that PCT was a useful marker in differentiating whether the symptoms were caused by a respiratory tract infection vs a noninfectious cause such as asthma. ¹³ A previous study showed PCT to be an effective marker of bacteremia in patients with febrile UTIs, with the absolute value being reflective of bacterial load in the bloodstream. ⁸ However, results on the accuracy of PCT in diagnosing bacterial infections have been inconsistent, leading many investigators to question its diagnostic accuracy. Furthermore, most of the current literature evaluates the role of PCT in critically ill patients. Herein we aim to determine whether PCT can be used as a biomarker for predicting UTIs in patients presenting to the ER with an obstructing ureteral stone with the goal of being able to more accurately determine which patients require urgent decompression.

Materials and Methods

After receiving institutional review board (IRB) approval, a prospective observational study was performed at a single institution from November 9, 2016, to November 10, 2018. Informed consent was deemed unnecessary by the IRB as the study did not include any additional blood draws or tests and thus did not involve any deviation from the standard of care.

During this time period, all patients presenting to the ER with a CT scan showing an obstructing ureteral stone and a urology consultation were included. Patients were excluded if they were <18 years old, had signs or symptoms of secondary source of infection, taking steroids or antibiotics at the time of presentation, or were otherwise immunocompromised. All patients underwent a routine work-up that included a complete blood count with differential, basic metabolic panel, UA, and UCx. Erythrocyte sedimentation rate (ESR) and CRP were added to existing laboratories so no additional blood draws were necessary. The remaining patient serum was then transferred to a −80°C freezer. These samples were then run for PCT with Roche assay material (Roche Diagnostics, Indianapolis, IN). All data were collected and stored in a password-protected database and a chart review was performed to identify pertinent patient demographics.

Statistical analysis was performed using SAS Studio 3.6 (Cary, NC) to compare patients with and without positive UCx. Sample size was determined by employing a PROC module to calculate receiver operating characteristic (ROC) curves. For an area under the curve (AUC) of 0.7, we determined that at least 30 patients with positive UCx needed to be enrolled and compared with 30 control patients with negative UCx. Based on data from an internal audit at our institution, we determined that the ratio of infected to noninfected stone patients presenting to our ER to be 1:7. Thus, at least 210 patients needed to be included in the study to obtain 30 positive UCx. An AUC of 0.7 was deemed appropriate based on a similar article that noted an AUC for CRP of 0.68 and ESR of 0.65 in predicting bacteremia in hospitalized patients. ¹⁴ Using an AUC of 0.7, we calculated a power of 80% and an alpha of 0.05. Covariates that were used included UA WBCs, nitrite, leukocyte esterase (LE), ESR, CRP, PCT, WBC, WBC differential (% neutrophils), and fever (defined as >101.5°F). We defined a negative UCx as those that resulted in no growth or mixed skin flora as this likely represents contamination. We performed a subgroup analysis of patients clinically equivocal for infection. This was defined as those having no fever, UA WBCs <50 per high powered field (hpf), and a WBC >10⁴/μL. This definition was determined by the authors based on their experience in the patients who cause the most uncertainty regarding their infection status. Categorical data are presented as percentage frequencies, with continuous data presented as mean values ± standard deviation. Categorical outcomes were analyzed using Chi-squared test. Continuous outcomes were analyzed by the Mann–Whitney U test. Two-tailed tests were used. For continuous variable, we developed an ROC curve and found the optimal cutoff point by using Youden index. Based on the cutoff point, we then created binary variables and performed a multivariate logistical model. Backward elimination was used to select variables to include in the multivariate model. Statistical significance was defined as p ≤ 0.05.

Results

A total of 231 patients met the inclusion criteria and were analyzed in the study. Fifty-six patients (24%) had a positive UCx. On comparing those with and without a positive UCx, a significant difference was seen in all the covariates (Table 1). Patients with positive UCx were also significantly more likely to have a fever (32% vs 2%), nitrite positive UA (30% vs 3%), and LE positive UA (38% vs 2%) (Table 1). In total, 103 patients (44.6%) underwent a procedure for urinary tract decompression during their initial presentation: 5 nephrostomy tube placements, 6 ureteroscopies with stone treatment, and 92 ureteral stent placements with delayed stone treatment. The rest of the patients were sent home on medical expulsive therapy with outpatient follow-up. A total of 74 patients met systemic inflammatory response syndrome (SIRS) criteria on presentation that was defined as having at least two of the following: temperature of >100.4°F or <96.8°F, heart rate >90 beats per minute, respiratory rate >20 breaths per minute, and WBC >12⁴/μL. ¹⁵ The median PCT value in this group was 0.079 ng/mL. A total of seven patients required intensive care unit care after urinary tract decompression for hemodynamic monitoring because of labile vital signs. All of these patients recovered without issue with resuscitation and antibiotics.

The univariate analysis results and the multivariable logistic regression model for the entire cohort are given in Tables 2 and 3, respectively. Of all the covariates, UA WBCs with a cutoff of 9 per hpf had the highest AUC. All covariates showed a statistically significant difference when comparing those with and without a positive UCx. While adjusting for other covariates, patients with UA WBCs >9 per hpf are more likely to have positive UCx than patients with UA WBCs ≤9 per hpf (odds ratio [OR] = 18.29, confidence interval [CI] 7.24–46.21). Patients with a PCT >0.08 ng/mL are also more likely to have a positive UCx than patients with PCT ≤0.08 ng/mL (OR = 4.13, CI 1.66–10.27) (Table 3). The sensitivity, specificity, PPV, and NPV for PCT with an ideal cutoff of 0.08 ng/mL were 70.6%, 73.9%, 34.3%, and 92.9%, respectively.

The univariate analysis results and the multivariable logistic regression model for the clinically equivocal cohort are given in Tables 2 and 4, respectively. In total, 105 patients made up this cohort with 17 having a positive UCx. Again, UA WBCs with a cutoff point of 6 per hpf had the greatest AUC of 0.72. Statistically significant differences in comparing those with and without positive UCx were only seen for UA WBCs >6 per hpf and WBC differential >86%. While adjusting for other covariates, patients with WBC differential >86% are more likely to have a positive UCx than patients with WBC differential ≤86% (OR = 12.5, CI 3.41–45.51), and patients with UA WBCs >6 per hpf are more likely to have positive UCx than those with UA WBCs ≤6 per hpf (OR = 6.8, CI 1.82–25.54). The sensitivity, specificity, PPV, and NPV for PCT with an ideal cutoff of 0.3 ng/mL in this cohort were 47.1%, 85.2%, 38.1%, and 89.3%, respectively. When performing serial and parallel testing, PCT combined with the other covariates did not yield superior sensitivity or specificity compared with combining other components of the routine work-up.

Discussion

An essential determinant to make in-patients presenting to the ER with an obstructing ureteral stone is whether they have a UTI as this can rapidly develop into sepsis along with its associated morbidities. Urgent decompression with either a Double-J stent or percutaneous nephrostomy is indicated in this situation. Blackwell et al. performed a large cross-sectional retrospective review of patients admitted to the hospital with a primary diagnosis of nephrolithiasis and found that there was an increased risk of mortality when decompressive interventions were delayed for 48 hours from admission. ¹⁶ This increased mortality risk with delayed intervention has been supported by other studies that reinforces the importance of early and accurate diagnosis. Urologists play a key role in determining the timing of upper tract decompression and are left in a diagnostic dilemma when the patient has equivocal signs and symptoms of UTI. This highlights the need for better biomarkers to aid clinicians in the diagnosis of infected ureteral stones as the stakes are high if the diagnosis is missed.

Based on the results of this study, PCT does not appear to be superior to other tests in predicting obstructing ureteral stone patients who have infected urinary systems. This held true for both the overall population and the clinically equivocal cohort. The other biomarkers evaluated, ESR and CRP, also did not outperform the laboratories that make up the standard work-up. These results are in contrast to the pilot study done by Papagiannopoulos and coworkers ¹⁷ in which PCT appeared to be a strong predictor of an eventual positive UCx when using a cutoff value of 0.1 ng/mL. Using similar methodology, this pilot study showed PCT to have an AUC of 0.812, sensitivity 80%, specificity 82%, PPV 57.1%, and NPV 93.3% (Ref. ¹⁷ ). PCT performed worse in this study with a lower sensitivity, specificity, PPV, and NPV. There are several possible reasons for this discrepancy. First, a different cutoff value for determining an elevated PCT was used in the two studies. In this study, the cutoff value was determined by developing an ROC curve and calculating the ideal cutoff by using Youden index (PCT of 0.08 ng/mL for entire study population and 0.3 ng/mL for clinically equivocal cohort), whereas the pilot study used the value of 0.1 ng/mL based on the reference range set by the study institutions laboratory. Of note, the cutoff for determining an elevated PCT value varies widely in the literature and depends on the clinical scenario in which PCT is being used. For example, if PCT is being used to rule out sepsis, then a cutoff of 0.2 ng/mL has been used, whereas if being used to rule out bacterial source of respiratory symptoms, then a cutoff of 0.1 ng/mL is used. ^6,13 One Japanese retrospective observational study looked at PCT values in all patients hospitalized at a single institution who had a serum PCT value obtained during admission. ¹⁸ The optimal cutoff value for PCT in predicting positive body fluid culture in this study was 1.32 ng/mL with a sensitivity of 64%, specificity of 63%, and AUC of 0.661. This study showed significantly different median PCT values in patients with positive cultures from different body fluids. For example, the median PCT value was 8.32 ng/mL in the positive ascites fluid culture group compared with 0.46 ng/mL in the positive cerebral spinal fluid group. ¹⁸ These results suggest that serum PCT values that are predictive of positive fluid cultures vary significantly based on the type of body fluid that is infected. Our data would suggest that serum PCT does not become significantly elevated during UTI compared with other body fluids as our median value was only 0.1 ng/mL for the positive UCx group. This may limit its utility as a biomarker for UTIs in patients with obstructing stones.

When looking at the total population, UA WBCs had the highest AUC as well as the highest OR for predicting a positive UCx compared with other variables. It appears that a UA WBCs >9 per hpf at our institution is the strongest predictor for infection when controlling for other variables. When looking specifically at the clinically equivocal group, UA WBCs >6 per hpf and WBC differential >86.37% met statistical significance in predicting a positive UCx with OR of 6.8 and 12.5, respectively. Based on these results, PCT is predictive in diagnosing an infected ureteral stone when looking at all patients presenting to the ER with a ureteral stone; however, it does not appear to be superior to the standard work-up of UA WBCs and WBC differential. The biomarkers PCT, CRP, and ESR do not outperform the components of the standard work-up when evaluating stone patients for concomitant infections. Based on these results, PCT, CRP, and ESR may have limited utility in this patient population. However, this brings to question whether there is utility for further study into a serum biomarker in the diagnosis of UTI in the setting of an obstructing ureteral stone. Our study would suggest so, as only a third of patients with positive UCx had signs of fevers (32%) and findings of nitrite positive (30%) and LE positive (38%) on UA. Thus, a strong biomarker to aid in diagnosing infected ureteral stones is needed and demands further investigation.

There are several notable limitations to this study. This was an observational study with data collected from chart review with inherent limitations when relying on the accuracy of the electronic medical record. Another limitation is our sample size. Although we exceeded our goal of 30 patients with positive UCx to compare with 30 controls with negative UCx, we only had 17 patients in the clinically equivocal cohort with a positive UCx. This low sample size decreases the statistical power of our subgroup analysis. We defined our clinically equivocal cohort as having no fever, a UA WBCs <50 per hpf, and a WBC >10⁴/μL. This definition was determined by the investigators based on their clinical experience. There are going to be patients in this study who did not meet these criteria, however, still had signs and symptoms equivocal for infection and thus pose a diagnostic dilemma. Another limitation is that PCT values can be elevated in certain noninfectious conditions such as surgery, cardiogenic shock, autoimmune disorders, and severe renal or liver dysfunction, and these conditions were not controlled for in this study. ⁶ Despite these limitations, this is a novel prospective study with real-world pertinence to the practicing urologist.

Conclusion

PCT does not appear to be a superior marker for diagnosing UTI in the setting of an obstructing ureteral stone when compared with components of the standard work-up. UA WBCs performed the best in both the overall and clinically equivocal cohorts. The entire clinical picture should be considered when determining the need for urgent upper tract decompression. None of the biomarkers evaluated, including PCT, CRP, and ESR, performed well enough to be routinely used in this patient population. Until better biomarkers are discovered, clinical intuition based on patient factors and the standard laboratory work-up will continue to be used to determine which patients require urgent decompression. Given the potential serious consequences of missing this diagnosis, continued efforts are needed to evaluate clinically useful biomarkers.