Quick Sequential Organ Failure Assessment Score Is a Better Predictor of Septic Shock After Percutaneous Nephrolithotomy: A Secondary Analysis of Two Multicenter Prospective Trials

Introduction

Percutaneous nephrolithotomy (PCNL) is the first-line treatment for nephrolithiasis in symptomatic patients with a stone burden >20 mm. ¹ Furthermore, PCNL is frequently the recommended treatment for lower pole stones >10 mm, as it is highly efficacious with stone-free rates of 81.6%–95%. ^{2 –4} However, given its more invasive nature compared with alternative treatments such as ureteroscopy with lithotripsy or extracorporeal shockwave lithotripsy, PCNL carries a higher risk of morbidity and mortality. A thorough review by Michel et al. cites complication rates upward of 83%.

Most of these complications are minor, with infection being the most common. Although sepsis is a relatively uncommon major complication, it still carries a risk of 0.8%–4.7%. ⁵ Given the risk of adverse outcomes in a setting of decreasing post-PCNL length of stays, it becomes all the more important for the surgeon to identify complications (such as sepsis) early. ^6,7

Per the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3), the proposed definition of sepsis is, “life-threatening organ dysfunction due to dysregulated host response to infection.” ⁸ The Sepsis-3 proceedings noted that sepsis carries a mortality rate in excess of 10%, but qualified that, “septic shock should be defined as a subset of sepsis in which particularly profound circulatory, cellular, and metabolic abnormalities are associated with a greater risk of mortality than with sepsis alone” (citing mortality rates >40% in cases of septic shock). ⁸ Importantly, Sepsis-3 cites the Surviving Sepsis Campaign to highlight the importance of early recognition of sepsis and septic shock—early recognition allowing for early intervention in an effort to improve mortality outcomes. ^8,9

There is no gold-standard diagnostic test for sepsis or septic shock, but rather clinical signs and symptoms that develop in a patient with infection. Previously, systemic inflammatory response syndrome (SIRS) criteria scoring was often recommended as a quick tool to aid in detecting patients with sepsis. ¹⁰ However, quick sequential organ failure assessment (qSOFA) has become the recommended tool in more recent years as retrospective studies have shown it to be superior to SIRS for predicting sepsis and septic shock. ⁸

Although Yaghoubian and coworkers applied qSOFA to a retrospective review in urology, to our knowledge, there are no studies using prospective data to compare SIRS and qSOFA after urologic procedures. ¹¹ We, therefore, sought to determine the accuracy of qSOFA compared with SIRS to detect septic shock with prospectively collected data.

Materials and Methods

We performed a secondary analysis on prospectively collected data sets from two randomized controlled multicenter trials including PCNL patients across nine institutions in the EDGE (Endourologic Disease Group for Excellence) Research Consortium. ^12,13 Participating centers included Indiana University (39 patients), University of California San Diego (50 patients), Dartmouth-Hitchcock Medical College (10 patients), Vanderbilt University Medical Center (36 patients), University of British Columbia (40 patients), The Ohio State University (18 patients), Mayo Clinic Arizona (15 patients), Cleveland Clinic (6 patients), and Columbia University Medical Center (4 patients). This study was conducted under UCSD IRB #160158.

A total of 218 PCNL cases were evaluated between 2015 and 2020. Summary of criteria is as follows: (1) inclusion criteria from index studies included patients aged 18 years or older who were planned to undergo a PCNL; (2) exclusion criteria included an estimated glomerular filtration rate <60, liver disease, pregnancy, a positive urine culture at time of study enrollment, a febrile urinary tract infection in the past year, indwelling ureteral stent or nephrostomy tube, recent antibiotic use, and severe hydronephrosis—a detailed description is available on published articles of index studies. ^12,13

Methods were similar to those described previously in a study comparing qSOFA and SIRS for PCNL patients. ¹ Clinical signs of SIRS and qSOFA were collected postoperatively. Septic shock was defined as patients with suspected infection from urinary tract source who required intensive care unit (ICU) level care for initiation of vasopressors to maintain mean arterial pressure >65 mm Hg postoperatively. The primary outcome was the sensitivity and specificity of SIRS and qSOFA in predicting septic shock.

Patients met SIRS criteria if two or more of the following were present: heart rate >90 bpm, respiratory rate >20 breaths/min, temperature <36°C or >38°C, or white blood cell count <4000 or >12,000 cells/mm³. Patients met qSOFA criteria if two or more of the following were present: systolic blood pressure ≤100 mm Hg, respiratory rate ≥22 breaths/min, or noted to display altered mental status.

IBM SPSS v28.0 (Armonk, New York, USA) was utilized for applying statistical tests (including Pearson chi-square, independent sample t-test, McNemar's test, and binary logistic regression—where applicable).

Results

A total of 218 patients from 9 institutions were analyzed. Postoperatively, 46 (17.35%) patients met the SIRS criteria, and 30 (11.32%) patients met the qSOFA criteria. One patient (0.46%) developed septic shock. This patient met the criteria for both SIRS and qSOFA, resulting in a sensitivity and negative predictive value of 100% for both tools; however, the specificity and positive predictive value (PPV) for qSOFA were slightly higher than SIRS: 90.78% and 4.78% vs 72.4% and 1.66%, respectively (Table 1). On multivariate analysis, neither tool appeared predictive of severe surgical complications (defined as Clavien–Dindo class III or greater) (Table 2).

To differentiate between those patients who were “positive” for either SIRS or qSOFA criteria, we evaluated demographics and perioperative variables. Both groups were similar with exception of the SIRS group having a higher estimated blood loss (EBL) and larger stone size (Table 3).

Discussion

Early identification of septic shock is vital in the clinical timeline to prevent profound morbidity and even mortality. Urolithiasis-related sepsis is a known problem for which clinicians search for means to identify even in the nonoperative setting. ¹⁴ In the operative setting, preoperative urine testing has been observed to be less reliable than analysis of intraoperative specimens (though such testing takes days to result). ¹⁵ Moreover, sepsis presentation can be occult with various presenting manifestations, making diagnosis challenging. There is no single diagnostic test for sepsis. Instead, sepsis is a constellation of signs and symptoms that place a patient at high risk for infectious complications.

However, our analysis of a prospectively collected multicenter data sets confirms earlier retrospective findings that qSOFA appears as sensitive but more specific than SIRS as a tool in identifying septic shock. The clinical take-home message: If your patient has positive qSOFA, the likelihood of developing sepsis is much higher than just positive SIRS alone—and yet both these tests have overall low PPV (4.78% vs 1.66%, respectively). Although qSOFA is not meant as a binary decision tool for a urologist to disposition a post-PCNL patient, it may aid in postoperative decision making (especially in an era where more PCNL procedures are completed in an outpatient setting).

In this analysis of prior prospective studies, we report and confirm the prior retrospective evidence (Yaghoubian et al.) that qSOFA is a more specific test than SIRS in predicting septic shock. The clinical significance of this study is highly relevant as qSOFA provides an extremely simple nonlaboratory-based bedside evaluation to diagnose septic shock. As septic shock is associated with mortality upward up to 40%, delays in diagnosis and subsequently treatment are extremely relevant.

The ability to mobilize transfer of care to higher acuity level potentially portends less morbidity and mortality for patients in early septic shock. And although septic shock rates among the PCNL population are in single digits, these proportions still translate into an opportunity to avoid serious morbidity and even mortality in endourology.

Also consistent with the findings of Yaghoubian and coworkers, qSOFA was not predictive of postoperative complications. However, as described in Table 3, there were perioperative variables (EBL and stone size) observed more frequently in the SIRS positivity (vice qSOFA). The reasoning for the higher EBL and stone size associated with SIRS is not entirely clear; however, it is conceivable that these variables evoke a greater inflammatory response without evoking septic shock.

Other studies have been done to confirm the utility of qSOFA in urologic patients, however, these have been limited using retrospective data. Yaghoubian et al. looked at 320 post-PCNL patients from 8 institutions and found qSOFA and SIRS to have 100% sensitivity in predicting septic shock—and, in line with our findings, the specificity of qSOFA was significantly greater than that of SIRS (93.3% vs 68.4%). ¹¹ Similarly, Peng et al. retrospectively looked at 431 post-PCNL patients and compared qSOFA and SIRS in predicting septic shock.

Again, qSOFA was found to be significantly more specific than SIRS (92% vs 81%); however, qSOFA had lower sensitivity than SIRS (92% vs 100%). ¹⁶ The lower sensitivity of qSOFA has been a criticism in other nonurologic studies as well. Askim and coworkers performed a prospective study of 1535 patients with suspected infection admitted to the emergency department at a single center in Norway. ¹⁷ It sought to compare the ability of these tools in predicting severe sepsis, defined as a patient with suspected infection and an SIRS score of two or more, plus one additional sign of organ failure. The qSOFA was more specific (98% vs SIRS 72%), but less sensitive (32% vs SIRS 74%).

Shi and coworkers compared both SOFA and qSOFA with SIRS in predicting in-hospital mortality and prolonged ICU stay among patients with urolithiasis-associated sepsis—the authors' noted that SOFA provided the greatest prognostic performance (though required the greatest amount of clinical data). ¹⁸ The comparability with other studies may be limited given this study's outcome was severe sepsis rather than septic shock, and the SIRS sensitivity and specificity data were calculated without considering leukocyte count. It is also important to note that the qSOFA score was calculated retrospectively.

Our study is important as it represents higher level evidence (prospective), validating retrospective studies with respect to qSOFA value in PCNL prediction of septic shock. One limitation of this study is the sample size that included only one patient who met the septic shock criteria. While it is reassuring that only one patient suffered septic shock clinically, it is problematic from a statistical perspective as the effective sample size is so small. Furthermore, we did not have Glasgow Coma Scale scores to incorporate into the qSOFA criteria (but rather relied on reports of altered mental status). Similarly, we did not have specific lactate levels recorded, which is part of the septic shock criteria (serum lactate >2 mmol/L)—rather we utilized the treatment of pressors as a surrogate for septic shock diagnosis.

Also, although most data were collected within the initial hours of the postoperative period, from the index studies it may have been collected as late as postoperative day 1. Thus, although the predictive value qSOFA appears supported by our findings, it is not clear how early these signs may be expected and, therefore, concern for sepsis should not be limited to qSOFA findings alone. Considering the limited sample size of patients who met septic shock criteria, further research is needed. Nevertheless, we hope this study and the study of prior authors emphasize the need to diffuse qSOFA into routine clinical care among urologists performing PCNL. This quick screening tool will become even more important as same-day discharge PCNL becomes more prevalent. ^6,7

Conclusions

Although PPV for both qSOFA and SIRS in prediction of post-PCNL septic shock is low, prospectively collected data demonstrate use of qSOFA may offer greater specificity than SIRS criteria when predicting post-PCNL septic shock.