Dual-score framework: National Early Warning Score 2 and quick Sequential Organ Failure Assessment scores in acute pyelonephritis

Abstract

BACKGROUND

Acute pyelonephritis (APN) is a significant part of the healthcare burden globally. Early risk stratification in APN is crucial to guide intensive care unit (ICU) admission and anticipate in-hospital mortality. The relevance of scoring tools like quick Sequential Organ Failure Assessment (qSOFA) and National Early Warning Score 2 (NEWS2) in urology-specific infections has been unclear. Our study is among the first to prospectively compare the qSOFA and NEWS2 scores exclusively in non-emphysematous APN, a real-world cohort previously overlooked in research.

AIM

To assess the accuracy of NEWS2 and qSOFA scores in predicting ICU admission, in-hospital mortality and their role in guiding urological interventions among adults with APN.

METHODS

A prospective observational study, conducted at a tertiary care centre in South India over two years, on adult patients diagnosed clinically and radiologically with APN. Comprehensive clinical, laboratory, and imaging parameters were recorded. The cohort was stratified into patients requiring ICU and not requiring ICU to assess differences across parameters. The discriminative power of the NEWS2 and qSOFA scores for predicting expedited care and mortality was analyzed.

RESULTS

The NEWS2 score ≥ 5 emerged as a robust tool, identifying nearly all patients requiring ICU admission [sensitivity, 98.1%; area under the receiver operating characteristic curve (AUC): 0.977; P < 0.05] and predicting intervention with high accuracy (AUC-0.823; P < 0.05). However, the qSOFA ≥ 2 score proved to be the silent sentinel of mortality (AUC-0.839; P < 0.05), outperforming the NEWS2 score in prognostication. Interestingly, patients who later required ICU care had elevated NEWS2 scores at baseline, suggesting the missed early red flags. To the best of our knowledge, this is the first and largest prospective study to compare NEWS2 and qSOFA scores exclusively in non-obstructive, non-emphysematous APN.

CONCLUSION

In APN, NEWS2 (≥ 5) proved to be a pragmatic trigger for ICU triage, while qSOFA (≥ 2) better flagged in-hospital mortality. Together, they form a complementary framework as a simple, objective, and lifesaving bedside tool.

Key Words: Acute pyelonephritis; Triage; Intervention; In-hospital mortality; Quick Sequential Organ Failure Assessment; National Early Warning Score 2; Early warning scores

Core Tip: Acute pyelonephritis (APN) is common in urology, but existing studies on early warning scores often mix heterogeneous sepsis cohorts or emphasize stone-related or emphysematous pyelonephritis. Our prospective study is the first to focus exclusively on non-obstructive, non-emphysematous APN, a group frequently overlooked yet clinically challenging. We demonstrate that the National Early Warning Score 2 score reliably identifies patients who require intensive care unit care or urgent intervention, while the quick Sequential Organ Failure Assessment score more accurately predicts mortality. Used together, they provide a complementary dual-score framework for bedside triage and prognostication. This simple approach offers urologists and nephrologists a practical, objective tool to enhance decision-making and improve patient outcomes.

INTRODUCTION

Acute pyelonephritis (APN) remains one of the most common urological emergencies encountered in clinical practice. APN is a significant part of the healthcare burden globally. While many patients improve with antibiotics and supportive care, a subset deteriorates quickly, needing intensive care unit (ICU) admission or even succumbing to sepsis. The challenge lies in identifying those patients early, often complicated by overlapping clinical presentations, nonspecific early imaging findings, and delays in culture results[1,2]. The relevance of scoring tools like quick Sequential Organ Failure Assessment (qSOFA) and National Early Warning Score 2 (NEWS2) in urology-specific infections has been unclear. By contrast, this prospective study addresses this gap to directly compare the qSOFA and NEWS2 scores exclusively in non-stone, non-emphysematous APN, a real-world cohort previously overlooked in research.

Several early warning tools have been proposed to guide frontline decision making[3], with the qSOFA and the NEWS2 scores being among the most widely used[4,5]. To date, prior reports evaluating early warning scores predominantly assess the sepsis cohorts presenting to the emergency department in which urinary infection like APN is only a subset or included heterogeneous sepsis populations, such as emphysematous pyelonephritis (EPN), or focused predominantly on severe obstruction secondary to stones[6-8]. These conditions inherently carry high morbidity and inflate apparent performance, making its applicability in generalized APN population uncertain[9,10]. As a result, the predictive utility of qSOFA and NEWS2 in general APN remains poorly defined.

This study is among the first prospective effort to directly compare these two scoring systems exclusively in a real-world cohort, designed to bridge that gap by eliminating confounding factors and better reflecting the nuanced presentations of APN seen in routine urological practice. By narrowing the focus to this understudied population, our objective was to evaluate the true clinical value of these scores in early triage, need for intervention and mortality prediction.

MATERIALS AND METHODS

Study area

This study was conducted at a tertiary care institute in South India.

Study design

A prospective observational study.

Study population and period

Adult patients (aged 18 years or older) diagnosed clinically and/or radiologically with APN from January 2023 to January 2025.

Inclusion criteria

(1) Age ≥ 18 years; and (2) APN suspected at presentation with fever and/or flank pain and/or pyuria and/or costovertebral angle tenderness and/or sepsis, with supportive laboratory and radiological evidence obtained after stabilization.

Exclusion criteria

(1) Age < 18 years; (2) Pregnancy; (3) Missing core variables required to compute NEWS2 or qSOFA at triage; and (4) EPN, stone related obstructive uropathy, congenital urinary tract abnormalities evidence documented on imaging.

Data collection

After obtaining ethical clearance from the institutional research ethics committee, a prospective data collection protocol was established. The following data were collected from patients meeting the eligibility criteria using a standardized case record form designed specifically for this study. All enrolled patients were assessed within 24 hours of hospital admission or at first contact.

Demography and comorbidities, signs and symptoms, physiological vital parameters, laboratory and radiological reports, admission records including ICU stay, treatment records with response to treatment (recovery/mortality) were collected. NEWS2 and qSOFA scores were calculated simultaneously at first contact and thereafter as per the triggers reflected for each patient. The scoring was performed by the principal investigator, a resident of the department of urology, with assistance from colleagues of the same grade. Daily review of all entries was conducted by the principal investigator to ensure completeness, accuracy, and internal consistency. Any discrepancies or missing data were immediately rectified through patient chart review or direct patient assessment. The consolidated data sheets were thoroughly audited weekly by the supervising faculty member to maintain data quality and adherence to the study protocol.

Data collection methodology

All clinical assessments, laboratory investigations, and radiological studies were performed as part of routine clinical care. Data were extracted from direct patient assessment, patient chart reviews, electronic medical records, laboratory information systems, and picture archiving and communication systems. For patients requiring ICU admission, additional data on vasopressor use, mechanical ventilation, and length of stay were recorded. Follow-up data on clinical outcomes (recovery, mortality, duration of hospitalization) were collected until hospital discharge or patient death.

Study variables

Demographics: Age, sex, laterality of renal involvement.

Symptomatology and its duration: Fever, flank pain, pyuria, hematuria.

Clinical findings: Consciousness, Glasgow Coma Scale (GCS), temperature, pulse, shock [systolic blood pressure (SBP) < 90 mmHg or requiring vasopressors], respiratory rate, oxygen saturation at room air and need for oxygen, costovertebral angle tenderness/tender palpable kidney.

Comorbidities: Diabetes mellitus, hypertension, chronic kidney disease (CKD), and immunocompromised status.

Laboratory parameters: Haemoglobin, total leukocyte count, platelet count, serum creatinine, blood urea nitrogen, urine and blood culture with antibiotic sensitivity pattern.

Imaging: Ultrasonography kidney, ureter, and bladder (KUB) and computed tomography KUB to confirm the diagnosis of APN and exclude cases with obstructive renal or ureteric calculi and EPN. Findings including degree of renal parenchymal involvement, and presence of perinephric fat stranding or collection were documented.

Outcomes definition

ICU Admission was defined as admission to a monitored intensive care setting at any time during the hospital stay, as per the institutional decision of the treating consultant based on the requirement of one or more organ support, sepsis care. Need for intervention, such as Placement of DJ stent or percutaneous nephrostomy (PCN) performed as part of source control during hospital stay, was at the discretion of the treating consultant based on clinical, laboratory and radiological parameters. Prolonged Hospital Stay was defined as patients requiring ICU admission for 3 days or more, and an in-hospital stay for 10 days or more[11].

Mortality was defined as death occurring during the same admission for the APN. Readmission within 30 days was defined as any unplanned readmission with urinary symptoms or sepsis within 30 days of discharge[12].

Sampling technique and procedure

Scoring was applied at the time of presentation. The decision to stent, admit to the ICU, or escalate care was not influenced by the score but rather based on clinical judgment. Patients were stratified into Group 1 (ICU admission needed) and Group 2 (ICU admission not required). Later, score performance was analyzed against the outcomes [primarily the need for expedited care (ICU) and in-hospital mortality]. Secondarily to the need for the placement of a DJ Stent and/or PCN. All patients presenting with APN underwent clinical and imaging assessment, followed by application of NEWS2 and qSOFA scores at admission. Patients were subsequently categorized into Group 1 and Group 2 (as described above) cohorts based on clinical need, independent of scoring. Final outcomes were then correlated with the initial scoring to assess predictive performance. The workflow of our study is detailed in Figure 1.

Figure 1 Study methodology (tertiary care institute in South India, 2023-2025). The algorithm of patient inclusion, assessment, and outcome analysis. APN: Acute pyelonephritis, ABC: Airway breathing circulation; ICU: Intensive care unit; NEWS2: National Early Warning Score 2; qSOFA: Quick sequential organ failure assessment score; PPV: Positive predictive value; NPV: Negative predictive value.

Statistical analysis

Based on these, the total score for the two scoring systems was calculated for each patient.

The NEWS2 score: Scores derived from 6 physiological parameters plus consciousness and oxygen requirement, yielding a total score from 0 to 20.

The qSOFA score: Score based on altered mental status (GCS < 15), SBP ≤ 100 mmHg, and respiratory rate ≥ 22/min, with a total score ranging from 0 to 3.

Data were entered into Microsoft Excel and analyses were conducted using IBM SPSS Statistics for Windows, Version 25.0 (IBM Corp., Armonk, NY, United States). Continuous variables are expressed as mean ± SD, while categorical variables are presented as n (%). Comparisons between groups were performed using independent t-tests for continuous data and χ² or Fisher’s exact tests for categorical variables. Two-tailed P values < 0.05 were considered statistically significant, and 95% confidence intervals (CI) were reported. Statistical review was performed by a biomedical statistician.

Univariate analyses were first conducted to identify individual clinical and laboratory features significantly associated with ICU admission and in-hospital mortality. Multivariate logistic regression models to determine independent predictors were performed with cues taken from the univariate analysis. Adjusted odds ratios (ORs) along with their corresponding 95%CI and P values were reported. In addition, subset analyses were performed to assess patient-level differences, where one scoring system (NEWS2 or qSOFA) demonstrated superior performance. Diagnostic accuracy for predicting ICU admission and mortality was evaluated for both scores using sensitivity, specificity, positive predictive value, negative predictive value, and AUC curve. Results were supported by tabular representations, infographic visual summaries, and receiver operating characteristic (ROC) curve analyses.

RESULTS

A total of 220 patients with APN were evaluated, with a median age of 67 years. ICU admission was required in 47.7% of cases. On comparison of baseline demographic, clinical, and laboratory parameters, the median age was 56.3 for ICU-admitted patients and 56.2 for non-ICU admitted patients. Female patients constituted 55.9% of the total cohort, with a slightly higher proportion in the ICU group (58.1%). A total of 70 patients (32%) demonstrated the classical triad of fever, flank pain, and pyuria at presentation, reinforcing the observation that APN may not always follow textbook patterns, and that a high index of suspicion remains essential. Fever was the consistent symptom, present in 75% (n = 162) of the cohort. Hematuria and palpable kidney were relatively uncommon. Shock at presentation was observed in 18.18% overall but was far more common in ICU patients (36.19%) than in non-ICU patients (1.73%). Diabetes mellitus was a significant comorbidity, present in 71.6% of the overall cohort. Notably, this included patients with either acute kidney injury (AKI) (44%) or CKD (11.8%), suggesting that diabetic patients may be at a higher risk of renal compromise during APN episodes. Patients requiring ICU admission showed more deranged physiological and laboratory parameters like higher pulse rate (106 bpm vs 82 bpm, P < 0.0001), lower SBP (92 mmHg vs 119 mmHg, P < 0.0001), higher respiratory rate (28 breaths/min vs 18 breaths/min, P < 0.0001), marked leukocytosis (21016 vs 11986, P < 0.0001) and thrombocytopenia (136735 vs 278605, P < 0.0001). These derangements highlight the systemic impact of APN, especially in a population with pre-existing vulnerabilities (Table 1).

Table 1 Baseline patient characteristics, key predictors and clinical differences stratified by intensive care unit admission in acute pyelonephritis (tertiary care institute in South India, 2023-2025), n (%).

Parameter	Group 1 (n = 105)	Group 2 (n = 115)	95%CI	P value
Age in years (mean ± SD)	56.3 ± 14	56.2 ± 14	53.6-59; 53.6-58.8	0.96
Female sex	61 (58.1)	62 (53.9)	48.7-67.5; 44.7-63.1	0.53
Fever	105 (100)	57 (49.6)	96.5-100; 40.6-58.7	< 0.001
Flank pain	67 (63.8)	22 (19.1)	54.4-72.5; 12.6-27.1	< 0.001
Pyuria	60 (57.1)	22 (19.1)	47.4-66.3; 12.6-27.1	< 0.001
Diabetes	93 (88.6)	64 (55.7)	81.2-93.9; 46.3-64.9	< 0.001
CKD	19 (18.1)	12 (10.4)	11.5-26.6; 5.5-17.5	0.11
AKI	60 (57.1)	25 (21.7)	47.4-66.3; 15-30.2	< 0.001
Temperature in degrees celsius (mean ± SD)	38.8 ± 0.6	37.7 ± 1.05	38.6-38.9; 37.5-37.9	< 0.001
Pulse in beats/minute (mean ± SD)	109 ± 15	82 ± 7	106-112; 81-83	< 0.001
SBP in mmHg (mean ± SD)	91 ± 10	118 ± 8	89-93; 117-120	< 0.001
RR in breaths/minute (mean ± SD)	28.3 ± 3.5	17.8 ± 2.2	27.6-29; 17.4-18.2	< 0.001
Total WBC count in × 109/L (mean ± SD)	21 ± 2.3	14.8 ± 5.4	20.6-21.4; 13.8-15.8	< 0.001
Platelets in × 109/L (mean ± SD)	141 ± 39	283 ± 77	191.2-222.8; 269-297	< 0.001
BUN in mmol/L (mean ± SD)	17.3 ± 8.9	13.6 ± 7.5	16.6-21.7; 9.2-11.3	< 0.001
Serum creatinine in μmol/L (mean ± SD)	230 ± 106	150 ± 71	188-271; 95-117	< 0.001
NEWS2 score (mean ± SD)	14.7 ± 3.1	1.7 ± 1.3	14.1-15.3; 1.46-1.94	< 0.001
Intervention	105 (100)	51 (44.3)	96.5-100; 35.3-53.7	< 0.001
Duration of stay in days (mean ± SD)	9.2 ± 3.1	3.5 ± 1.3	8.6-9.8; 3.3-3.7	< 0.001
Mortality	13 (12.4)	1 (0.9)	6.8-20; 0.02-5	0.01

CI: Confidence interval: CKD: Chronic kidney disease; AKI: Acute kidney injury; SBP: Systolic blood pressure; RR: Respiratory rate; WBC: White blood cell; BUN: Blood urea nitrogen; NEWS2: National Early Warning Score 2.

The patients who underwent DJ stenting had significantly higher white blood cell (WBC) counts at admission. The distribution, median, and interquartile ranges of WBC counts among patients who underwent and did not undergo intervention, reflecting an apparent upward shift in total counts among stented patients, with a median value of 21146/mm³. In contrast, the non-stented group had WBC counts predominantly under 12000/mm³. Urine culture revealed predominantly gram-negative colonization (Escherichia coli) in 109 patients across the cohort. Still, Klebsiella spp., Pseudomonas spp., and fungal pathogens, such as Candida species, were predominantly associated with ICU admissions (35%). Suggesting a potential association with more complicated or resistant infections (Figure 2).

Figure 2 Clinical and microbiological correlates of disease severity in acute pyelonephritis (tertiary care institute in South India, 2023-2025). A: Violin plot comparing total white blood cell count in patients with and without intervention; B: Distribution of intensive care unit admissions stratified by urine culture organisms. WBC: White blood cell; ICU: Intensive care unit.

The intervention in the form of the DJ stenting or PCN was performed in 70.9% of patients, often in those with obstructive features on imaging or worsening renal parameters. On comparison of the NEWS2 score and the qSOFA score in predicting the need for the intervention, the NEWS2 score ≥ 5 had higher sensitivity (67.9% vs 55.8%) and a better area under the ROC curve (AUC) (0.839 vs 0.779) than the qSOFA score ≥ 2. This suggests that NEWS2 can identify at-risk patients for urological intervention more reliably, reducing the chances of missed critical interventions. A distinct strength in predicting ICU admission was noted between the scores, as NEWS2 ≥ 5 showed increased sensitivity (98.1% vs 89.1%) with an AUC of 0.977 (excellent discrimination) compared to the qSOFA ≥ 2 score (AUC = 0.905). The qSOFA score ≥ 2 demonstrated perfect sensitivity for mortality prediction, while the NEWS2 score ≥ 7 had lower sensitivity (50%). The qSOFA score of 2 or more flagged the three patients that were missed by the NEWS2 score of 7 or more. Whereas the NEWS2 score lacks granularity in the high-risk group, due to significant clustering with a wide range, it is a comparatively weak predictor of mortality. The qSOFA score demonstrated good discriminative power with an AUC of 0.823, compared to the NEWS2 score, which had an AUC of 0.674. This supports the qSOFA score of 2 or higher as a stronger “prognostic flag” for predicting mortality in APN (Figure 3).

Figure 3 Comparison of National Early Warning Score 2 and quick Sequential Organ Failure Assessment scores in predicting critical outcomes in acute pyelonephritis (tertiary care institute in South India, 2023-2025). A: Box plot demonstrating that National Early Warning Score 2 (NEWS2) score showing a broader distribution among the intervention group; B: Receiver operating characteristic (ROC) curves for NEWS2 and quick Sequential Organ Failure Assessment score (qSOFA) in predicting intensive care unit admission; C: Venn diagram shows the distribution of high-risk patients (NEWS2 score ≥ 7 and qSOFA score ≥ 2); D: ROC curves for predicting in-hospital mortality. NEWS2: National Early Warning Score 2; qSOFA: Quick Sequential Organ Failure Assessment score; AUC: Area under the receiver operating characteristic.

Patients who were admitted to the ICU tended to have significantly higher NEWS2 scores, specifically those with a score of 5 and above, which were consistent with progressive physiological compromise, affirming the need for ICU care. The difference in what each score captured became evident during analysis, showing the distribution of NEWS2 scores in patients who were either not admitted to the ICU or shifted later during the course. Those who required late ICU admission had higher baseline NEWS2 scores (median: 14.2, interquartile range: 12.6-16.8). These patients didn’t meet traditional ICU triggers then but ended up needing escalation.

The NEWS2 score appeared to identify early red flags and maintained a better balance of sensitivity and specificity across various thresholds. This makes it more practical in triage, especially in a setting like APN, where a patient can look okay and then deteriorate at a fast pace. Thus, making it the preferred tool for triage and escalation. Variables with clinical relevance and those significant in univariate analysis were considered for multivariate analysis.

The multivariate logistic regression analysis that was performed to identify independent predictors of ICU admission and in-hospital mortality in patients with APN. Notably, classical symptoms such as fever, flank pain, and pyuria were excluded from the final model due to either lack of discriminatory power or collinearity with more robust physiological predictors. For ICU admission, the presence of diabetes mellitus (adjusted OR: 2.75; 95%CI: 1.40-5.38; P = 0.003), AKI at presentation (adjusted OR: 3.21; 95%CI: 1.62-6.38; P = 0.001), shock (SBP < 90 mmHg) at presentation (adjusted OR: 4.68; 95%CI: 2.12-10.33; P < 0.001), respiratory rate ≥ 22/min (adjusted OR: 3.89; 95%CI: 1.91-7.91; P < 0.001), thrombocytopenia with platelet count ≤ 150000/mm³ (adjusted OR: 2.77; 95%CI: 1.34-5.73; P = 0.006), and a NEWS2 score ≥ 5 (adjusted OR: 6.91; 95%CI: 2.84-16.83; P < 0.001) emerged as significant independent predictors. For mortality, shock at admission (adjusted OR: 5.96; 95%CI: 2.03-17.54; P = 0.001), thrombocytopenia (adjusted OR: 3.45; 95%CI: 1.18-10.07; P = 0.024), serum creatinine ≥ 2 mg/dL (adjusted OR: 2.64; 95%CI: 1.01-6.91; P = 0.048), and qSOFA score ≥ 2 (adjusted OR: 7.83; 95%CI: 2.16-28.39; P = 0.002) were independently associated with increased odds of death as depicted in Table 2.

Table 2 Multivariate logistic regression analysis for predictors of intensive care unit admission and mortality in acute pyelonephritis (tertiary care institute in South India, 2023-2025).

Parameter	Adjusted OR for ICU admission (95%CI)	P value	Adjusted OR for mortality (95%CI)	P value
Age ≥ 60 years	1.24 (0.78-1.96)	0.35	1.49 (0.72-3.11)	0.28
Female sex	1.13 (0.68-1.87)	0.62	0.98 (0.42-2.31)	0.97
Diabetes mellitus	2.75 (1.40-5.38)	0.003	1.96 (0.84-4.54)	0.12
AKI at admission	3.21 (1.62-6.38)	0.001	2.18 (0.91-5.21)	0.08
Shock (SBP < 90 mmHg)	4.68 (2.12-10.33)	< 0.001	5.96 (2.03-17.54)	0.001
RR ≥ 22 breaths/min	3.89 (1.91-7.91)	< 0.001	2.94 (1.02-8.50)	0.047
Total WBC count ≥ 15 × 109/L	1.62 (0.89-2.94)	0.11	2.01 (0.78-5.16)	0.15
Platelets ≤ 150 × 109/L	2.77 (1.34-5.73)	0.006	3.45 (1.18-10.07)	0.024
Creatinine ≥ 176.8 μmol/L	1.98 (0.95-4.14)	0.07	2.64 (1.01-6.91)	0.048
NEWS2 score ≥ 5	6.91 (2.84-16.83)	< 0.001	1.65 (0.61-4.48)	0.32
qSOFA score ≥ 2	2.14 (0.88-5.20)	0.09	7.83 (2.16-28.39)	0.002

OR: Odds ratio; ICU: Intensive care unit; CI: Confidence interval; AKI: Acute kidney injury; SBP: Systolic blood pressure; RR: Respiratory rate; WBC: White blood cells; NEWS2: National Early Warning Score 2; qSOFA: Quick Sequential Organ Failure Assessment score.

These findings affirm the utility of integrating score-based and objective clinical parameters for early risk stratification and prognostication in APN. Overall, Trends in scoring, intervention, and outcomes suggest that the NEWS2 score shows a more prominent difference in the lower threshold, supporting its use in triage decisions. Although it lacks granularity in the high-risk group, due to significant clustering with a wide range, it is a comparatively weak predictor of mortality. Hence, both the NEWS2 and the qSOFA score offer value, but in different clinical domains, with the NEWS2 score being better aligned with ICU triage decisions, and the qSOFA score being more predictive of mortality risk.

DISCUSSION

APN continues to present a clinical challenge, particularly in identifying patients who may progress to severe illness requiring ICU care or those at risk of in-hospital mortality. In the acute setting, especially in urology, where intervention decisions like DJ stenting may be time sensitive, having a reliable, quick bedside scoring system becomes not just helpful but essential. In this study, two rapid scoring systems, NEWS2 and qSOFA were evaluated in patients with APN. The NEWS2 scoring tool was better suited to signal early physiological deterioration and prompt escalation towards ICU admission, consistent with previous literature reports that the NEWS2 scoring tool is preferable for early escalation in acute care settings[13,14]. On the other hand, the qSOFA scoring tool although more specific, lacked the sensitivity required to flag patients early, especially those who presented without overt signs of systemic decompensation.

Importantly, our analysis showed that NEWS2 alone in several patients captured subtle physiological drift and triggered early ICU admission and intervention[15]. These cases highlight the value of a graded scoring system like NEWS2, which can capture mild-to-moderate shifts in physiology and prompt early review by urology or the ICU. On the contrary, qSOFA’s limited scope made it less sensitive to this window of opportunity, even though it aligned well with poor outcomes in patients who eventually deteriorated or were already in advanced stages of systemic illness.

Our multivariate analysis identified that along with NEWS2 score ≥ 5 few other parameters like diabetes mellitus, AKI, shock at presentation, tachypnea (respiratory rate ≥ 22/min), and thrombocytopenia as key independent predictors of ICU admission in patients with APN. These findings underscore the multifactorial nature of disease severity in APN, wherein both systemic inflammatory response and baseline comorbidities contribute to clinical deterioration. Interestingly, classical symptoms were non-discriminatory, although statistically significant in univariate analysis, did not retain significance in the multivariate model. This likely reflects their limited discriminatory power in stratifying disease severity, as they are commonly present across the entire spectrum of APN severity. Additionally, their physiological impact may already be captured through composite scoring systems like NEWS2, rendering them redundant in adjusted analyses.

The differing nature of these scores likely explains their performance divergence. The NEWS2 score is continuous, graded, and sensitive to early shifts in physiology-traits ideal for triage and dynamic monitoring. In contrast, the qSOFA score is simpler and binary, focusing on late physiological decompensation markers like hypotension, altered mentation, and tachypnea, which explains its specificity for poor outcomes[16]. For in-hospital mortality, the most salient predictors were shock, elevated serum creatinine (≥ 2 mg/dL), thrombocytopenia, and a qSOFA score ≥ 2. This highlights the prognostic relevance of qSOFA in identifying patients at higher risk of fatal outcomes, aligning with prior literature emphasizing its utility in mortality prediction, particularly in sepsis-related contexts. Collectively, these findings support a dual-score strategy, wherein NEWS2 facilitates early triage and ICU escalation, while qSOFA offers enhanced specificity for predicting adverse outcomes. The independent associations observed affirm that combining clinical judgment with objective scoring metrics can significantly refine risk stratification and guide timely intervention in APN management.

By focusing this study on APN with non-stone obstruction or no anatomical cause, we attempted to reduce the confounding influence of obvious obstructive uropathy[17,18]. Many of these patients still required DJ stenting, often on clinical grounds. We observed that patients with higher NEWS2 scores were more likely to receive early stenting, often even before definitive culture results or radiology reports were available. This supports the idea that NEWS2 can act as a valuable adjunct in the decision to intervene. Accordingly, the scores differed in their functional utility and their optimal use is complementary with the NEWS2 score aiding in early escalation and triage, and the qSOFA score serving as a focused prognostic tool[19,20].

Leveraging NEWS2 for escalation and qSOFA for outcome prediction may offer a balanced strategy for managing APN. Thus, integration of both scores, interpreted in the appropriate clinical context, provides a structured and objective framework for decision-making in APN. While numerous studies have validated NEWS2[21] and qSOFA[22] in the general sepsis population, their application to APN has remained superficial. These studies were often diluted by the inclusion of emphysematous cases or stone-induced obstruction that inherently signal clinical severity. These high-risk phenotypes tend to inflate the predictive accuracy of early warning scores, limiting their generalizability.

What sets our study apart is its deliberate exclusion of these extremes to isolate a more clinically ambiguous, yet highly prevalent subset: Non-obstructive, non-emphysematous APN. This “middle-ground” cohort is where the real diagnostic challenge lies. The patients who appear deceptively stable but harbour evolving systemic compromise. Prior literature has largely overlooked this group[23], despite representing a significant proportion of APN admissions. By examining scoring systems in this context, our findings offer a more pragmatic view: NEWS2 proved invaluable in detecting early physiological shifts and prompting escalation before overt decompensation, while qSOFA, although less sensitive upfront, emerged as a potent predictor of mortality once deterioration was underway.

This divergence in performance highlights the importance of aligning the scoring system with the clinical question, specifically triage vs prognosis, rather than treating these tools as interchangeable. In doing so, our study not only fills a critical gap in the current literature but also reframes how clinicians can strategically deploy these scores in the nuanced terrain of urological infections.

Strengths and limitations of the study

To the best of our knowledge, this is the first and largest prospective study in PubMed-indexed literature to systematically evaluate and compare the prognostic performance of the NEWS2 and qSOFA scores exclusively in adult patients with non-obstructive, non-emphysematous APN. Prior research has primarily been limited to retrospective or heterogeneous cohorts, often confounded by emphysematous infections or obstructive uropathy, thereby diminishing the generalizability of findings to the broader APN population.

By deliberately excluding these high-risk phenotypes, our study isolates a clinically challenging yet underrepresented subset of patients who frequently present with deceptively mild symptoms but are at genuine risk of rapid physiological deterioration. Through this focused lens, we demonstrate that the NEWS2 score offers superior sensitivity for early triage and escalation decisions, while the qSOFA score exhibits higher specificity for mortality prediction. Together, these findings fill a critical evidence gap and support a dual-score framework that is both pragmatic and transformative for risk stratification in urological sepsis.

Despite its strengths, this study has certain limitations. First, as a single-centre investigation conducted in a tertiary care setting, the findings may not be fully generalizable to other healthcare environments with differing resource availability and clinical practices. Second, the absence of an external validation cohort limits the ability to confirm the reproducibility of these results across diverse patient populations. Third, while the exclusion of emphysematous and obstructive pyelonephritis was intentional to reduce confounding, it may limit the applicability of the findings to more complex or high-risk cases often encountered in urological practice.

CONCLUSION

This study highlights that the NEWS2 score, with its broader physiological coverage, more reliably identifies patients with APN who require early ICU admission or intervention, offering a practical edge in triage at an acute care setting. In contrast, the qSOFA score correlates more closely with mortality, reinforcing its role as a specific prognostic indicator. Employing both scores in a complementary manner may enhance clinical decision-making and improve patient outcomes in APN.