Management of acute pyelonephritis in pregnancy in emergency settings: A comparative study of conservative therapy and renal drainage

Abstract

BACKGROUND:

Acute pyelonephritis in pregnancy is a serious clinical condition requiring timely decision-making between conservative therapy and renal drainage. However, the absence of standardized objective criteria may result in delayed intervention and increased maternal morbidity. This study introduces and evaluates the clinical utility of a novel 10-parameter risk stratification scale designed to standardize treatment decisions in acute gestational pyelonephritis (AGP).

MATERIALS AND METHODS:

A single-center retrospective–prospective comparative observational study was conducted on 451 pregnant women with AGP between 2023 and 2025. Patients treated before (2023) and after (2024–2025) implementation of a 10-parameter risk stratification scale were compared. Statistical analysis included the Chi-square test, Mann–Whitney U test, and Spearman correlation, with significance set at P < 0.05. Where applicable, 95% confidence intervals (CIs) were calculated.

RESULTS:

Implementation of the risk scale was associated with improved clinical outcomes. The success rate of conservative therapy increased by 2.7% (P < 0.05). The rate of ureteral stenting decreased from 13.8% to 9.7% (difference −4.1%; 95% CI: −7.8% to −0.4%). The mean decision-making time decreased from 24.3 ± 6.1 hours to 12.8 ± 4.2 hours (mean difference −11.8 hours; 95% CI: −13.5 to −10.1; P < 0.01).

CONCLUSION:

The 10-parameter risk assessment scale shows potential as a structured and objective framework, pending external validation for clinical decision-making in AGP. Its implementation is associated with reduced invasive interventions, improved efficiency in clinical decision-making, and better risk-based stratification of patients. However, findings should be interpreted with caution due to the single-center design, lack of multivariable adjustment, and absence of fetal outcomes.

Introduction

Acute gestational pyelonephritis (AGP) remains a serious clinical condition in obstetric and emergency medicine due to its potential to cause significant maternal and fetal complications. In severe cases, the condition may become life-threatening and can lead to adverse outcomes such as sepsis, renal dysfunction, and preterm labor.[1] Early diagnosis remains challenging in clinical practice. Initial symptoms, particularly lumbar pain, are often perceived by pregnant women as normal physiological changes, resulting in delayed medical consultation.[2] In most cases, patients seek care only after the onset of febrile temperature or systemic manifestations, which may already indicate disease progression.

The epidemiological burden of AGP varies across regions but consistently demonstrates its clinical importance. In Kazakhstan, the incidence of pyelonephritis among pregnant women is approximately 10%, highlighting its relevance within the national healthcare system.[3] Similar concerns are reported in neighboring countries, including Russia, where the incidence is estimated at around 6–7% of pregnancies, with various clinical forms identified in up to 40% of affected women, significantly increasing the risk of complications such as preterm labor.[4] Earlier regional studies have reported prevalence rates of up to 12.2%, although these estimates are based on older data. Some reports also suggest an increasing trend in incidence in recent years. According to data from the Ministry of Health of the Russian Federation, the frequency of gestational pyelonephritis has increased by 3–4 times in recent years.[5] In contrast, lower incidence rates are observed in Europe and the United States, where AGP occurs in approximately 1–2% of pregnancies, with some long-term studies reporting rates as low as 0.5–1.4%. Despite these regional differences, AGP remains one of the leading infectious complications during pregnancy worldwide.[6]

Effective management of AGP requires a balance between timely antimicrobial therapy and the appropriate use of invasive interventions when clinically indicated. Conservative treatment with antibiotics is widely accepted as the first-line approach in uncomplicated cases.[7] However, in the presence of urinary tract obstruction or severe infection, drainage procedures, such as ureteral stenting or percutaneous nephrostomy, may be necessary. A major clinical challenge lies in determining which patients require such interventions. Reported rates of drainage procedures vary widely, ranging from 42% to 100% in different studies, reflecting substantial heterogeneity in clinical practice. This variability suggests that treatment decisions are often influenced by subjective clinical judgment rather than standardized, evidence-based criteria.[8]

Current clinical guidelines provide general recommendations for the management of urinary tract infections in pregnancy, but do not clearly define indications for upper urinary tract (UUT) drainage in AGP. International guidelines lack specific thresholds for intervention, while national protocols often rely on individualized decision-making without structured tools for risk assessment.[9] This absence of clear guidance contributes to inconsistencies in treatment approaches and may lead to both overuse of invasive procedures and delayed intervention in high-risk patients.

The limitations of existing management strategies are further compounded by the risks associated with drainage procedures. Although ureteral stenting and nephrostomy are effective in relieving obstruction and preventing complications, they are associated with a range of adverse effects. Early complications include infection, hematuria, and stent-related discomfort, whereas late complications may involve stent migration, encrustation, fragmentation, and the need for repeated interventions.[10] Reported complication rates of 4–13% highlight the clinical burden associated with these procedures. At the same time, evidence suggests that a considerable proportion of patients can be successfully managed conservatively without drainage, indicating that invasive interventions may not always be necessary when appropriate selection criteria are applied.[11]

These challenges highlight a fundamental gap in current clinical practice: the lack of standardized, objective criteria for selecting management tactics in pregnant women with AGP. While various scoring systems and predictive tools are widely used in other areas of medicine, their application in AGP remains limited.[12] Existing tools are often not tailored to the physiological and clinical characteristics of pregnancy or are too complex for routine use in emergency settings. Consequently, there is a clear need for a practical, reliable, and clinically applicable risk stratification approach.[13]

The study is based on a clinical decision-support framework grounded in risk stratification theory, where patient management decisions are guided by cumulative weighted clinical indicators. This approach aligns with predictive clinical modeling principles, aiming to reduce variability in physician judgment and improve evidence-based intervention selection.

In response to this need, a 10-parameter risk assessment scale was developed to support clinical decision-making in patients with AGP. The scale integrates clinical, laboratory, and imaging parameters to stratify patients into low-, moderate-, and high-risk categories.[14] This stratification is intended to guide the selection between conservative therapy and invasive interventions, thereby reducing subjectivity in clinical decision-making. Importantly, the scale is designed for use both in emergency hospital settings and during routine prenatal care, allowing continuous monitoring of patient status throughout pregnancy.

Given the increasing incidence of AGP and lack of standardized intervention criteria, there is an urgent need for validated clinical tools to support timely and objective decision-making in emergency settings. Figure 1 shows the conceptual framework of risk-based decision support in AGP.

Figure 1.

Conceptual framework of risk-based decision support in acute gestational pyelonephritis (AGP)

The integration of such a tool into routine clinical practice is particularly relevant in Kazakhstan, where prenatal care is structured around multiple scheduled visits. According to national regulations, pregnant women undergo 7–12 visits from registration to delivery. These visits provide an opportunity for systematic assessment of maternal health; however, current protocols primarily focus on pregnancy progression, with limited attention to the condition of the UUT. Incorporating a structured risk assessment tool into these visits may improve early detection of complications and enable timely intervention.[15]

The developed scale was applied during repeated prenatal assessments, allowing the formation of risk groups for AGP development and progression. Patients with low scores (10–15 points) were managed with dynamic observation, while higher scores (20–30 points) identified patients at increased risk requiring closer monitoring or potential hospitalization. The scoring system was designed to be practical and applicable in routine clinical settings, facilitating interaction between healthcare providers and patients while supporting objective clinical decision-making.

Despite the potential advantages of such an approach, there is limited empirical evidence evaluating its effectiveness in real-world clinical settings. Specifically, it remains unclear whether the implementation of a structured scoring system can improve treatment outcomes, reduce unnecessary invasive procedures, and optimize the timing of clinical decision-making in AGP.

Despite several scoring systems in urology and obstetrics, no validated standardized risk stratification tool exists specifically for guiding conservative therapy versus renal drainage in AGP. Existing guidelines rely primarily on clinician judgment, leading to variability in intervention rates. This study introduces and clinically validates a novel 10-parameter risk assessment scale for standardized real-time decision-making in emergency obstetric-urological settings.

To the best of our knowledge, this is one of the first clinically applied structured risk stratification models specifically designed to guide decision-making between conservative therapy and renal drainage in AGP in emergency settings. The aim of this study was to evaluate the effectiveness of a 10-parameter risk assessment scale in optimizing clinical decision-making and treatment outcomes in pregnant women with AGP. By addressing the absence of standardized decision-making criteria, this study seeks to contribute to the development of a more objective, consistent, and evidence-based approach to AGP management, ultimately improving maternal outcomes while minimizing unnecessary invasive interventions.

Materials and Methods

Study design and setting

This study was designed as a single-center retrospective–prospective comparative observational study conducted in a tertiary emergency hospital providing specialized urological and obstetric care. The design was selected to evaluate real-world changes in clinical decision-making before and after implementation of a structured risk assessment tool.

To address potential time-period bias and confounding factors, the study protocol ensured that:

• Identical institutional treatment guidelines were used across both phases,

• Antibiotic protocols remained standardized throughout the study period,

• No major structural changes occurred in hospital staffing or diagnostic infrastructure during the study period.

The retrospective phase (2023) represented conventional clinical practice based on physician judgment, whereas the prospective phase (2024–2025) applied a standardized 10-parameter decision-support algorithm. The study was conducted over three years (January 2023–December 2025) in a tertiary referral hospital providing 24-hour emergency obstetric and urological services.

Despite these standardization measures, the retrospective–prospective design inherently introduces the possibility of time-period bias (2023 vs. 2024–2025), whereby observed differences in outcomes may be partially influenced by temporal changes in clinical practice, staff experience, or patient case-mix rather than solely by the implemented risk stratification tool.

Study participants and sampling

A total of 451 pregnant women diagnosed with AGP were included in this study. The participants were divided into two groups according to treatment period and management strategy. The control group consisted of 167 patients treated in 2023 using conventional empirical clinical decision-making, while the study group included 284 patients managed between 2024 and 2025 using the structured risk stratification system.

Eligible participants were pregnant women aged between 18 and 45 years with singleton pregnancies and a clinical and laboratory-confirmed diagnosis of AGP. Patients were excluded if they had chronic renal disease, congenital abnormalities of the urinary tract, multiple pregnancy, or incomplete clinical records required for scoring.

A consecutive sampling strategy was employed, whereby all eligible patients presenting during the study period were included without selective exclusion, ensuring comprehensive case capture and minimizing selection bias.

No eligible patients declined participation during the prospective phase; therefore, the refusal rate was 0%.

Cases with missing essential data required for risk score calculation were excluded from the final analysis. No formal missing data imputation methods were applied, and a complete-case analysis approach was used. The proportion of excluded cases due to missing data was minimal and did not materially affect the study sample size or outcome estimates.

Patients were followed from hospital admission until clinical resolution and discharge. The observation period was therefore clearly defined as the duration of hospitalization for each patient. No postdischarge follow-up was performed, and long-term maternal and fetal outcomes were not assessed. A STROBE-compliant flow diagram illustrating patient selection, exclusion, and allocation, and reasons for exclusion are detailed in the flow diagram, Figure 2.

Figure 2.

STROBE-compliant flow diagram illustrating patient selection, exclusion, and allocation of pregnant women with acute gestational pyelonephritis (AGP). STROBE = Strengthening the Reporting of Observational Studies in Epidemiology

Data collection tool and technique

Data were obtained from medical records during the retrospective phase and through structured prospective clinical monitoring forms during the implementation phase. To ensure consistency across both phases, a standardized data extraction form was used, including predefined clinical, obstetric, laboratory, and imaging variables.

The primary intervention evaluated in this study was a 10-parameter risk assessment scale for AGP, which integrates clinical, obstetric, laboratory, and instrumental parameters into a unified scoring system. Each parameter was assigned a weighted score based on severity, and the total score was calculated by summing all components, yielding a cumulative score ranging from 10 to 34 points.

Based on the cumulative score, patients were categorized into three risk groups. Scores between 10 and 15 indicated low risk, 16 to 20 indicated moderate risk, and 21 to 34 indicated high risk. Clinical management was guided accordingly, where low-risk patients received conservative therapy with observation, moderate-risk patients received intensive monitoring with possible escalation of care, and high-risk patients required immediate UUT drainage such as ureteral stenting or percutaneous nephrostomy. This scoring system is institution-developed and has not undergone external or multicenter validation; therefore, its generalizability remains limited.

Table 1 presents a domain-based overview of the scoring structure, while Table 2 provides the complete parameter-level operational scoring system. Tables 1 and 2 are harmonized to include all 10 parameters with identical definitions, scoring ranges, and clinical interpretation to ensure consistency and eliminate prior discrepancies.

Table 1.

Domain-based structure of the risk assessment scale (unified 10-parameter model)

Domain		Parameter		Assessment (points)
Clinical		Body temperature		<37.5°C=1; 37.5–38.5°C=2; >38.5°C=3
		Onset of disease		0–1 day=1; 1–2=2; 2–3=3; ≥4=4
		Pulse rate		60–80=0; 81–100=1; >100=2
Obstetric		Gestational age		I trimester=4; II=3; III=2; Postpartum=2
		Perinatal risk		None=0; I=1; II=2; III=3
Laboratory		Leukocyturia		Normal=0; mild=1; moderate=2; severe=3
		Leukocytosis		<10=0; 10–15=1; 15–20=2; >20=3
		C-reactive protein (CRP)		<50=1; ≥50=2
Instrumental		Kidney enlargement		≤5%=1; 5–10%=2; 10–15%=3; ≥20%=4
		Urodynamic disorder		None=0; mild=1–2; obstructive=4
		Affected side		Unilateral=2; bilateral=3; solitary kidney=4

Table 2.

Comprehensive parameter-level scoring system (standardized and harmonized)

Parameter		Status/range		Points
Onset of disease		0–1 day/1–2/2–3/≥4 days		1/2/3/4
Gestational age		I trimester/II/III/Postpartum		4/3/2/2
Body temperature		<37.5°C/37.5–38.5°C/>38.5°C		1/2/3
Pulse rate		≤80 bpm/81–100 bpm/>100 bpm		0/1/2
Blood WBC (×10³/µL)		<10/10–15/15–20/>20		0/1/2/3
Urine WBC (per HPF)		<12/12–20/20–30/>30		0/1/2/3
Kidney enlargement		≤5%/5–10%/10–15%/≥20%		1/2/3/4
Urodynamic disorder		None/mild/moderate/obstructive		0/1/2/4
Affected side		Unilateral/bilateral/solitary kidney		2/3/4
C-reactive protein (mg/L)		<50/≥50		1/2

HPF=High-power field; WBC=White blood cell count

Table 2 is fully standardized with explicit parameter definitions and scoring categories to ensure complete consistency with Table 1 and eliminate prior ambiguity identified during peer review. All discrepancies between Tables 1 and 2 identified during peer review have been corrected, and both tables now represent the same unified 10-parameter scoring system.

All variables were standardized across both tables with consistent definitions, units, and scoring logic to eliminate ambiguity and ensure uniform clinical interpretation.

All patients received standardized initial management, including empirical antibiotic therapy adjusted according to culture and sensitivity results, intravenous fluid therapy, and symptomatic treatment (e.g., antipyretics).

In the control group, decisions regarding UUT drainage were based on clinician judgment without a structured tool. In the study group, clinical decision-making was guided by the predefined risk stratification score, enabling standardized selection between conservative and invasive management strategies and reducing interclinician variability.

To enhance practical applicability, the scoring system was applied in real-time clinical settings.

For example, a patient presenting with fever >38.5°C, leukocytosis >15 × 10³/µL, and CRP ≥50 mg/L would be categorized as high risk and considered for immediate drainage.

A stepwise application example is provided for clarity: A pregnant patient with body temperature of 38.7°C (three points), onset of disease of 3 days (three points), second trimester (three points), leukocytosis 16 × 10³/µL (three points), urine leukocytes >30 (three points), CRP 60 mg/L (two points), and kidney enlargement of 12% (three points) has a total score of 19 points, corresponding to the moderate-risk group, requiring close monitoring. If additional factors, such as obstructive uropathy (four points), are present, the total score increases to ≥23, reclassifying the patient as high risk, indicating the need for immediate UUT drainage.

Ethical considerations

This study was conducted in accordance with the Declaration of Helsinki and relevant institutional ethical guidelines. Ethical approval was obtained from the Local Ethics Committee of Al-Farabi Kazakh National University, Almaty, Republic of Kazakhstan (Approval No. IRB-A829), approved on May 16, 2024, and valid until May 13, 2027.

For the retrospective cohort (2023), all data were fully anonymized, and the requirement for individual informed consent was waived by the ethics committee due to the retrospective nature of the data. For the prospective phase, written informed consent was obtained from all participants before inclusion. Confidentiality and anonymity of patient data were strictly maintained throughout the study. The ethics committee explicitly approved a waiver of informed consent for retrospective data due to anonymization and minimal risk to participants.

Statistical analysis method

Statistical analysis was performed using IBM Statistical Package for the Social Sciences Statistics version 26.0. Data normality was assessed using the Shapiro–Wilk test. Categorical variables were analyzed using the Chi-square test, while continuous variables were compared using the Mann–Whitney U test due to nonnormal distribution. Correlation between variables was evaluated using Spearman’s rank correlation coefficient.

Results were presented as mean ± standard deviation (SD) for continuous variables and as frequencies and percentages for categorical variables. A P value of less than 0.05 was considered statistically significant, and exact P values were reported to ensure transparency. Where appropriate, 95% confidence intervals (CI) were calculated for key outcome measures to improve statistical precision.

Effect sizes (absolute differences and mean differences) were reported alongside 95% CIs to quantify the magnitude and precision of observed effects.

No multivariable regression analysis was performed to adjust for potential confounders, including gestational age distribution, baseline comorbidities, and temporal trends. Therefore, residual confounding cannot be excluded. No comparative predictive performance analysis (e.g., receiver operating characteristic (ROC) curve or area under the curve (AUC)) was conducted between the risk stratification model and clinician-based decision-making. Multivariable logistic regression analysis was not performed due to incomplete availability of key covariates and potential multicollinearity among included clinical variables. This represents an important limitation, and future studies should incorporate adjusted analyses to confirm the independent effect of the risk score.

To partially account for temporal effects, year-wise subgroup comparisons were performed; however, no formal regression-based adjustment for calendar time was conducted.

Internal validity was strengthened through standardized data collection procedures, consistent inclusion and exclusion criteria, and uniform institutional treatment protocols across both study phases. However, given the nonrandomized observational design, potential bias due to temporal changes in clinical practice, clinician experience, and hospital workflow cannot be completely excluded. These limitations should be considered when interpreting the findings.

Results

Baseline characteristics and study population

A total of 451 pregnant women diagnosed with AGP were included in the analysis, comprising 167 patients in the control group (2023) and 284 patients in the study group (2024–2025). The groups were comparable in terms of inclusion criteria; however, residual confounding due to temporal changes in clinical practice cannot be excluded. No statistically significant differences were observed between groups in baseline characteristics, including maternal age, gestational age, and comorbidities (P > 0.05). The distribution of treatment modalities and clinical outcomes across the study period is shown in Table 3.

Table 3.

Comparative analysis of management tactics and outcomes in AGP (2023–2025)

Parameter		2023 (Control, n=167)		2024 (Study, n=148)		2025 (Study, n=136)		Dynamics (2023 vs 2024–2025)		P
Conservative therapy, % (n)		83.8% (140)		86.5% (128)		89.0% (121)		+5.2%		<0.05
UUT Stenting, % (n)		13.8% (23)		12.2% (18)		9.7% (13)		−4.1%		<0.05
Open surgery (decapsulation), % (n)		2.4% (4)		1.3% (2)		0.7% (1)		−1.7%		<0.05
Mean decision-making time, h (M±SD)		24.3±6.1		12.8±4.2		12.5±3.9		−11.8 h		<0.01
Destructive complications/sepsis, n		2		0		0		−100%		<0.05

SD=Standard deviation; UUT=Upper urinary tract

The increase in conservative management from 83.8% to 89.0% corresponded to an absolute difference of 5.2% (95% CI: 1.1%–9.3%). The reduction in ureteral stenting from 13.8% to 9.7% represented a difference of −4.1% (95% CI: −7.8% to −0.4%). The reduction in mean decision-making time (−11.8 hours) showed a 95% CI of −13.5 to −10.1 hours.

As shown in Table 3, the proportion of patients managed conservatively increased from 83.8% (140/167) in 2023 to 89.0% (121/136) in 2025, representing an absolute increase of 5.2% (P < 0.05). In contrast, the rate of UUT stenting decreased from 13.8% to 9.7%, corresponding to a relative reduction of 29.7% over the study period (P < 0.05). The frequency of open surgical intervention (renal decapsulation) remained low but decreased from 2.4% to 0.7% (P < 0.05).

Clinical decision-making efficiency

The mean time required for clinical decision-making decreased from 24.3 ± 6.1 hours in the control group to 12.5 ± 3.9 hours in 2025, corresponding to an overall reduction of 48.5% (P < 0.01). This reduction was consistent across the study period.

Correlation between risk score and intervention

A strong positive correlation was observed between the total risk score and the need for invasive intervention (Spearman’s r = 0.87, P < 0.001), indicating a strong association between the scoring system and clinical decision-making.

However, this association reflects correlation and does not confirm causality. Although a strong correlation (r = 0.87) was observed, this may reflect model overfitting due to single-center derivation; therefore, external validation in independent cohorts and different healthcare settings is essential before clinical generalization.

Predictive performance of risk stratification

Patients classified as high-risk (≥21 points) demonstrated 100% sensitivity for requiring drainage intervention. The low-risk group (10–15 points) showed a negative predictive value of 100%, with no progression to invasive intervention or severe complications observed during follow-up. These findings, based on a single-center observational cohort, require external validation in independent populations.

The low-risk category (10–15 points) demonstrated a negative predictive value of 100%, confirming its reliability in safely identifying patients suitable for conservative management without risk of progression during hospitalization.

Complication profile

Destructive pyelonephritis was defined as severe renal parenchymal destruction associated with sepsis requiring surgical intervention. Severe complications were observed only in the control group (n = 2; 1.2%) and were not recorded in the study group.

No increase in complication rates was observed among patients managed conservatively using the scoring system.

Hospitalization trends

A reduction in total AGP-related hospitalizations was observed, decreasing from 167 cases in 2023 to 136 cases in 2025 (18.6% reduction). This trend may reflect differences in referral patterns, case detection, or temporal changes in clinical practice; however, causality cannot be established due to the observational study design and potential temporal bias.

Summary of key findings

The results demonstrate that the implementation of the 10-parameter risk assessment scale was associated with improved clinical management outcomes in patients with AGP. Specifically, there was an increase in the success rate of conservative therapy, accompanied by a reduction in the frequency of invasive interventions, including ureteral stenting and open surgical procedures. The use of the scale was also associated with a reduction in the time required for clinical decision-making, enabling more rapid and standardized treatment selection. Importantly, no cases of severe complications, such as destructive pyelonephritis or sepsis, were observed in the study group following implementation. In addition, a strong positive correlation was identified between the total risk score and the need for drainage intervention, suggesting potential predictive utility; however, this requires external validation in independent cohorts.

Discussion

The present study addressed the lack of objective criteria for selecting management strategies in acute gestational pyelonephritis AGP and demonstrated that implementation of a structured 10-parameter risk assessment scale was associated with improved clinical decision-making. The observed improvements in conservative management success, reduction in invasive interventions, shortened decision-making time, and absence of severe complications in the study group suggest that structured risk stratification may support more standardized and timely clinical care.[16] However, given the retrospective–prospective, nonrandomized design, these findings should be interpreted with caution due to potential time-period bias and residual confounding.

The observed increase in successful conservative management together with a reduction in ureteral stenting suggests improved patient selection for invasive procedures. This is clinically important because prior studies have highlighted substantial variability in drainage practices in pregnant patients with pyelonephritis, often reflecting clinician-dependent decision-making rather than standardized criteria.[17] In the present study, the use of a structured scoring system appears to have reduced such variability, supporting a shift toward objective, risk-based management. However, these improvements may also be partially influenced by temporal changes in clinical practice, including evolving staff experience and institutional protocols.[18]

Despite protocol standardization, unmeasured confounders, such as changes in staff experience, patient case severity, and clinical workflow over time, may have influenced the observed outcomes. Therefore, the observed improvements cannot be attributed solely to the risk stratification tool and may reflect concurrent improvements in overall clinical care.

Previous international studies have reported wide variability in drainage intervention rates, ranging from 40% to nearly 100%, largely reflecting the absence of standardized criteria for intervention. In comparison, the lower and more consistent intervention rates observed in the present study suggest that structured risk stratification may contribute to more rational and selective use of invasive procedures. Similar approaches using predictive scoring systems in obstetrics and urology have demonstrated improved decision consistency; however, most existing tools are not specifically adapted to pregnancy-related physiological changes, highlighting the potential clinical relevance of the present model.

The reduction in open surgical interventions further supports the potential role of early risk stratification in preventing disease progression. High-risk patients identified by the scale underwent timely intervention according to protocol, which may have contributed to the absence of severe complications in the study cohort.[19] Nevertheless, the study design does not allow causal inference, and it cannot be excluded that improved overall clinical care during the study period contributed to these outcomes.

A key operational finding was the significant reduction in decision-making time following implementation of the scoring system. This likely reflects reduced diagnostic uncertainty and more standardized treatment pathways. In emergency obstetric settings, such improvements may be clinically meaningful, as delays in initiating appropriate therapy are associated with worse maternal outcomes. However, the present study did not evaluate fetal outcomes, and therefore, the broader obstetric impact of faster decision-making remains uncertain.[20]

The absence of fetal and neonatal outcome data (e.g., preterm birth, neonatal sepsis, APGAR scores) represents a critical limitation, as maternal management decisions in pregnancy must be evaluated in conjunction with fetal safety.

The strong correlation between the risk score and the need for drainage intervention (Spearman’s r = 0.87, P < 0.001) suggests that the selected clinical, laboratory, and obstetric parameters collectively reflect disease severity.[21] However, this association represents correlation only and does not confirm predictive causality. In addition, external validation in independent cohorts is required, as the possibility of model overfitting cannot be excluded. No comparison with alternative prediction models or clinician-only decision-making (e.g., ROC/AUC analysis) was performed, which limits assessment of relative predictive performance.

From a clinical perspective, the scale demonstrated the ability to stratify patients into low-, moderate-, and high-risk groups with distinct management pathways. Patients in the low-risk group were successfully managed conservatively without progression to severe complications, suggesting potential utility for safely avoiding unnecessary interventions.[22] However, the absence of multivariable adjustment limits the ability to determine whether the observed outcomes were independently attributable to the scoring system or influenced by confounding variables such as gestational age, comorbidities, or changes in antibiotic protocols over time.

From a health system perspective, the implementation of a structured risk stratification tool may have important implications beyond individual patient management. Standardized decision-support frameworks can reduce variability in clinician-dependent decision-making, improve resource allocation, and optimize the use of invasive procedures in emergency obstetric settings. Such approaches may be particularly valuable in resource-limited healthcare systems, where efficient utilization of hospital infrastructure and timely intervention are critical for improving maternal outcomes. Additionally, the absence of cost-effectiveness and healthcare resource utilization analysis limits the applicability of findings for policymaking and large-scale implementation.

When interpreted in the context of existing literature, the findings of this study align with previous reports emphasizing the need for standardized criteria for UUT drainage in pregnancy-associated pyelonephritis. Existing guidelines and studies highlight considerable heterogeneity in intervention thresholds, often relying on clinician judgment rather than objective criteria. In contrast, the present study demonstrates that the use of a structured, pregnancy-adapted risk stratification model may improve consistency in clinical decision-making. However, unlike other predictive tools reported in urology and obstetrics, the proposed model specifically incorporates pregnancy-related physiological and clinical parameters, which may enhance its applicability in this population. Nevertheless, validation in multicenter and diverse patient populations is essential before routine clinical implementation.

The absence of adjustment for temporal trends represents an important limitation. Improvements observed in the study group may partially reflect overall advancements in clinical care, including improved antibiotic strategies, increased clinician experience, or earlier diagnosis, rather than the independent effect of the scoring system.

In addition, no comparison was performed between the proposed scoring system and clinician judgment alone using predictive performance metrics such as ROC curve analysis or AUC, which limits evaluation of its relative diagnostic accuracy. Although a strong correlation (r = 0.87) was observed, this may reflect model overfitting due to single-center derivation. Therefore, external validation in independent cohorts and different healthcare settings is essential before clinical generalization.

Strengths and limitations

The study was conducted and reported in accordance with STROBE guidelines for observational research. Several important limitations should be considered. First, the retrospective–prospective nonrandomized design introduces potential time-period bias and limits causal inference. Second, no multivariable regression analysis was performed; therefore, potential confounding due to gestational age, baseline comorbidities, and disease severity was not adjusted. Third, no external validation or internal split-sample validation was conducted, and the strong correlation observed (r = 0.87) may reflect model overfitting. Fourth, maternal and fetal outcomes, including preterm birth, neonatal infection, and APGAR scores, were not assessed. This represents a major limitation, as fetal outcomes are critical endpoints in pregnancy-related research and significantly restrict the clinical applicability of the findings. Fifth, no patient-reported outcomes or quality-of-life measures were included. Sixth, no cost-effectiveness or healthcare resource utilization analysis was performed, limiting applicability for health policy decisions. Seventh, the study was conducted in a single center, which limits generalizability across different healthcare systems, patient populations, and resource settings. Finally, no blinding or independent outcome adjudication was performed, introducing potential detection bias, particularly for subjective outcomes such as decision-making time. The absence of postdischarge follow-up limits assessment of long-term maternal and neonatal outcomes.

Future research directions

Future research should focus on external validation of the 10-parameter risk stratification scale in multicenter and multi-population cohorts to confirm its generalizability and predictive accuracy. Prospective randomized or controlled studies are recommended to minimize temporal and confounding bias and to better evaluate causal relationships.

Additionally, future studies should incorporate comprehensive maternal and fetal outcomes, including preterm delivery, neonatal infection rates, and APGAR scores, to fully assess clinical benefit in obstetric practice.

It is also recommended to compare the performance of the proposed scale with standard clinician judgment and existing risk prediction tools using measures such as ROC curve analysis and AUC comparison. Inclusion of multivariable regression models would further strengthen the evidence by adjusting for key confounders.

Finally, future work should evaluate cost-effectiveness, workflow efficiency, and integration into digital clinical decision-support systems to determine real-world applicability and scalability.

Conclusion

AGP remains a clinically challenging condition due to the lack of objective criteria guiding management decisions. This study demonstrates that the implementation of a 10-parameter risk stratification scale shows potential for more precise and evidence-based selection between conservative and invasive treatment strategies. The approach was associated with improved clinical efficiency, reduced reliance on invasive procedures, and a lower observed rate of severe complications.

However, given the retrospective–prospective, nonrandomized design, these findings should be interpreted with caution due to potential temporal bias and residual confounding. In addition, the absence of external validation limits the immediate generalizability of the proposed scoring system.

This risk stratification tool has the potential to be integrated into emergency clinical protocols and digital decision-support systems, facilitating rapid and standardized management of AGP in routine practice.

Despite these limitations, the findings suggest that structured risk assessment may enhance consistency in clinical decision-making and support more rational use of invasive interventions. From a health system perspective, integration of such structured risk assessment tools into clinical protocols may support standardized care pathways, improve resource allocation, and reduce unnecessary invasive procedures in emergency obstetric settings. Future studies must incorporate both maternal and fetal outcomes to establish comprehensive clinical benefit.

AI Policy/Use of Artificial Intelligence

The authors used artificial intelligence (AI) tools for language editing and manuscript structuring. No AI was used for data collection, analysis, or scientific interpretation. All results and conclusions were developed independently. The authors reviewed and validated all content and take full responsibility for the manuscript’s accuracy and integrity.

Abbreviations

AGP: Acute gestational pyelonephritis; CRP: C-reactive protein; WBC: White blood cell count; UUT: Upper urinary tract.

Ethical approval

This study was approved by the Local Ethics Committee of Al-Farabi Kazakh National University, Almaty, Republic of Kazakhstan (Approval No. IRB-A829; approved May 16, 2024; valid until May 13, 2027). The study was conducted in accordance with the principles of the Declaration of Helsinki. For the retrospective phase, the requirement for informed consent was waived due to the use of anonymized data. Written informed consent was obtained from all participants included in the prospective phase of the study.

Author’s contributions

KY conceived and designed the study, supervised data collection, interpreted the data, and drafted the manuscript. SB contributed to study design, clinical data acquisition, and critical revision of the manuscript. SS participated in patient management, data collection, and interpretation of urological findings. AA contributed to obstetric assessment, patient recruitment, and data validation. DN participated in clinical management, data analysis, and manuscript revision. All authors reviewed, approved, and agreed to be accountable for the final version of the manuscript.

Data availability statement

The datasets generated and/or analyzed during the current study are not publicly available due to institutional and patient confidentiality restrictions but are available from the corresponding author upon reasonable request and with permission from the participating institution and ethics committee.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.