Impact of acute kidney injury on long-term adverse outcomes in obstructive uropathy

Obstructive uropathy is known to be associated with acute kidney injury (AKI). This study aimed to investigate the etiologies, clinical characteristics, consequences and also assess the impact of AKI on long-term outcomes. This multicenter, retrospective study of 1683 patients with obstructive uropathy who underwent percutaneous nephrostomy (PCN) analyzed clinical characteristics, outcomes including progression to end-stage kidney disease (ESKD), overall mortality, and the impact of AKI on long-term outcomes. Obstructive uropathy in adults was most commonly caused by malignancy, urolithiasis, and other causes. AKI was present in 78% of the patients and was independently associated with preexisting chronic kidney disease (CKD). Short-term recovery was achieved in 56.78% after the relief of obstruction. ESKD progression rate was 4.4% in urolithiasis and 6.8% in other causes and older age, preexisting CKD, and stage 3 AKI were independent factors of progression. The mortality rate (34%) was highly attributed to malignant obstruction (52%) stage 3 AKI was also an independent predictor of mortality in non-malignant obstruction. AKI is a frequent complication of adult obstructive uropathy. AKI negatively affects long-term kidney outcomes and survival in non-malignant obstructions. A better understanding of the epidemiology and prognostic factors is needed for adult obstructive uropathy.

Etiologies of obstructive uropathy in adults. The patients were divided into three groups according to the etiology of obstruction. The most common cause was malignancy (n = 936, 55.6%), followed by urolithiasis (n = 481, 28.6%), and other causes (n = 266, 15.8%) ( Table 1). Other causes included ureteral stricture, intraabdominal abscess, retroperitoneal fibrosis, vesicoureteral reflux, and renal tuberculosis. Among the malignancies, metastatic colorectal cancer (n = 245, 26.2%) was the most common, followed by gastric cancer (n = 188, 20.1%), bladder cancer (n = 143, 15.3%), uterine cervical cancer (n = 97, 10.4%), and prostate cancer (n = 68, 7.3%) (Supplementary Table S1). Patients with malignancy-associated obstruction were more likely to be older, male, have lower hemoglobin levels, and have AKI. In contrast, patients with urolithiasis had a higher prevalence of diabetes, hypertension, ischemic heart disease, hemiparesis, and dementia. The prevalence of underlying CKD was the highest in the other causes group, with a significantly lower baseline eGFR. Inflammatory markers, including levels of C-reactive protein (CRP), procalcitonin, and leukocyte counts, were also significantly higher in patients with urolithiasis or other causes, suggesting that these patients might have superimposed urinary tract infection (Table 1).
Obstructive uropathy associated with AKI. Among the 1683 patients who underwent PCN, 1313 (78.0%) initially presented with AKI. A total of 216 patients (16.4%) required temporary dialysis. Patients with AKI were more likely to be older, male, and have preexisting CKD with a significantly lower baseline eGFR, and lower hemoglobin, albumin, and total CO 2 levels. However, patients with AKI had a significantly higher leukocyte count, CRP, procalcitonin, and uric acid levels than those without. Furthermore, the percentage of malignancy as the etiology of obstruction was significantly higher in patients with AKI (Table 2).
Using the decision tree model, we identified some relevant risk factors for AKI and performed a logistic regression analysis. Preexisting CKD, presence of hyponatremia (serum Na ≤ 135 mEq/L), and higher uric acid level at initial presentation were found to be independent risk factors for AKI in patients with obstructive uropathy (Table 3).
Kidney outcomes of obstructive uropathy. Of the 1313 patients who developed AKI, 746 (56.8%) showed short-term functional recovery within 7 days after PCN, and there was no difference in the recovery rates across the different etiologies of obstruction (Table 4). Logistic regression analysis showed that only the presence of preexisting CKD was an independent factor for the failure of short-term recovery (Supplementary  Table S2). www.nature.com/scientificreports/ Regarding long-term kidney outcomes, 45 patients (2.67%) progressed to ESKD. The progression rates in patients with urolithiasis and other causes were 4.4% and 6.8%, respectively, while that of malignant obstruction was only 0.6% (Table 4). This may be related to a higher cancer-related death rate. Due to huge differences, we separately analyzed the risk factors of ESKD progression between malignant and non-malignant causes of obstruction.
Logistic regression analysis revealed that older age, preexisting CKD, and stage 3 AKI were independent factors leading to the progression to ESKD. The presence of CKD and stage 3 AKI showed a 19-fold and fivefold increased risk of ESKD, respectively ( Table 5).
Mortality of obstructive uropathy. The overall mortality rate during the median follow-up period was 34%, with the highest rate observed in malignancy-associated obstruction (52%), followed by other causes (15.4%) and urolithiasis (8.9%) (p < 0.001). The mortality rate was significantly higher in patients with AKI, regardless of the etiology of the obstruction (Table 6).

Discussion
This multicenter, retrospective study demonstrated that the presence of AKI negatively affects the long-term outcomes of patients with who underwent PCN due to obstruction. The major findings of our study were as follows: (1) the most common cause of obstructive uropathy in adults was malignancy, followed by urolithiasis and other causes; (2) a substantial portion (78%) of patients presented with AKI, concomitant infection and short-term recovery after PCN insertion was achieved in 56.8% of these patients; (3) 4.4% of patients in the urolithiasis group and 6.8% in other causes group progressed to ESKD; (4) the overall mortality rate was 34%, with the highest mortality attributed to malignancy-associated obstruction (52%); and (5) stage 3 AKI was independently associated with both ESKD progression and mortality only in non-malignant causes of obstruction. Obstruction to urine flow can occur at any site of the urinary tract. Clinical manifestations depend on the etiology, location, and duration of obstruction, as well as the presence of underlying comorbidities; thus, they can vary widely, ranging from asymptomatic cases to renal colic, urosepsis, and dialysis requiring AKI or ESKD. In the pediatric population, obstructive uropathy from congenital anomalies of the urinary tract is responsible Table 2. Demographic and clinical characteristics of the study cohort according to the presence of AKI. DM diabetes mellitus, HTN hypertension, IHD ischemic heart disease, PAOD peripheral arterial occlusive disease, COPD chronic obstructive pulmonary disease, CKD chronic kidney disease, Hb hemoglobin, eGFR estimated glomerular filtration rate, CRP C-reactive protein, AKI acute kidney injury.     Table 6. Mortality of obstructive uropathy. AKI acute kidney injury.    This study aimed to investigate the clinical characteristics and outcomes of obstructive uropathy. PCN or pyelonephrostomy is an interventional procedure for the decompression of collecting systems and has become the primary option for temporary relief of obstruction in upper urinary tract obstruction since its first introduction in 1955 7 . We first demonstrated that the most common causes of obstructive uropathy in adults undergoing PCN were malignancy (55.6%), followed by urolithiasis (28.7%), and other causes (15.2%). In contrast with a previous study that demonstrated that uterine cervical and bladder cancer are the most common malignancy types, metastatic colorectal and gastric cancers were responsible for 46.3% of cases in our study 8 . This might be due to the exclusion of patients with urogenital cancer-associated lower tract obstruction, which was likely to be relieved with bladder decompression, or due to geographical differences in prevalent cancer types. Obstructive uropathy from direct tumor invasion or metastatic retroperitoneal lymphadenopathy is usually considered an ominous sign, and as expected, the overall mortality from malignancy-associated obstructive uropathy was higher (52%) than that from urolithiasis or other causes.

Malignancy (n-936) Urolithiasis (n = 481) Others (n = 266) p-value NonAKI (n = 162) AKI (n = 774) NonAKI (n = 134) AKI (n = 347) NonAKI (n = 72) AKI (n = 194)
Patient characteristics varied significantly according to the different etiologies of obstruction. Patients with malignant obstruction were significantly older, male, had lower hemoglobin levels, and were more likely to be associated with AKI despite a preserved baseline renal function. In contrast, patients with urolithiasis had a higher prevalence of diabetes, hypertension, ischemic heart disease, dementia, or hemiparesis, reflecting the close association between metabolic syndrome and urolithiasis. The prevalence of urolithiasis has been shown to have doubled in the United States in the recent years. Experimental studies have shown that type 2 diabetes and obesity facilitate lithogenesis by altering the proximal tubule handling of sodium and uric acid, and by lowering the urine pH [9][10][11] . The association between hypertension and urolithiasis has also been demonstrated in a population-based study 12 . Our study confirmed the association of urolithiasis not only with diabetes and hypertension but also with various cardiovascular and cerebrovascular diseases. Significantly higher CRP, procalcitonin levels, and leukocyte counts in this group suggested that stone-related obstructive uropathy might be associated with superimposed infection of the urinary tract. In contrast, obstructive uropathy from other causes was characterized by a female predominance, young age, and a higher prevalence of preexisting CKD.
The present study demonstrated that 78.1% of patients initially presented with AKI, suggesting that obstructive uropathy is frequently complicated by AKI. However, given that PCN is usually indicated upon renal functional deterioration or pyonephrosis, it could simply indicate that AKI is one of the most common indication of PCN in obstructive uropathy. In general, obstruction of a single kidney does not lead to AKI if the contralateral kidney function is normal. However, because we did not analyze the dual obstruction, it was difficult to classify the exact causes of AKI as predominantly postrenal or multifactorial. Given that obstructive uropathy is frequently complicated by urinary tract infection, sepsis, or combined volume depletion, the mechanisms underlying AKI development in this setting seem to be more complex. Among the many factors with significant differences, preexisting CKD, hyponatremia (≤ 135 mEq/L), and higher uric acid levels were independently associated with AKI.
Kidney outcomes in obstructive uropathy primarily depend on the prompt relief of obstruction. Therefore, delayed diagnosis due to the absence of specific signs and symptoms or underlying comorbidities might affect recovery. An estimated GFR of ≥ 60 ml/min/1.73 m 2 , recovery to < 25% of baseline eGFR, or dialysis independence within 7 days after PCN were defined as short-term recovery. Functional recovery was achieved in 56.8% of the patients, and the underlying etiologies of obstruction did not affect the short-term recovery rates. The presence of CKD was an independent factor associated with the non-recovery of short-term renal function. This is comparable with data by Khalaf et al. that preoperative GFR is an independent factor affecting the functional recovery in a prospective study of 91-patients with unilateral obstructive uropathy 13 . However, the lack of information about the exact duration of obstruction, which is the single-most important factor for recovery, was one of the limitations of our study, resulting in a possibility that failed short-term recovery in 43.2% of patients is related to delayed timing of relief of obstruction. Failure to determine the regression of hydronephrosis was another limitation of this study.
Because only a few studies have determined the long-term outcome of obstructive uropathy in adults, we examined the ESKD progression and mortality. Using data from both the ESKD registry of the Korean Society of Nephrology and the hospitals, we found that 45 (2.67%) patients progressed to ESKD during the median followup period of 91 [range 12-219] months. The progression rate in patients with urolithiasis/other causes was 4.4% and 6.8%, respectively, while that from malignancy was only 0.6%. The very low progression rate might be due to the high premature death rate from malignancy-related complications. According to the United States Renal Data System (USRDS), urolithiasis accounts for only 0.2% of the incidents of ESRD 14 . However, El-Zoghby et al. demonstrated that the attributable risk of ESKD from symptomatic urolithiasis was 5.1% in a population-based study conducted in Olmsted County, Mn, USA, which is comparable with our data 15 . As patients with urolithiasis were more likely to be diabetic, hypertensive, or have various cardio-cerebro-vascular diseases in our study, the mechanisms underlying the progression of urolithiasis seem to be complex, involving multiple comorbidities, superimposed infection, or obstruction-related mechanisms. We found that older age, pre-existing CKD, and stage 3 AKI were independently associated with ESKD. Although AKI is known to contribute to the increased burden of progressive CKD/ESKD 16,17 , most of the data came from the study of patients with intrinsic causes of AKI, such as ischemia, sepsis, or nephrotoxicity, but not from obstructive uropathy. Therefore, this is the first study to demonstrate that superimposed AKI also plays an important role in progressive CKD in obstructive uropathy. In contrast, none of these factors, such as stage 3 AKI, affect the progression of malignant obstruction. Stage 3 AKI complicated by obstructive uropathy in our study was independently associated with overall mortality in patients with non-malignant obstruction. Older age, male sex, and hypoalbuminemia were found www.nature.com/scientificreports/ to predict mortality. However, similar to ESKD progression, none of these factors were associated with mortality in malignant obstruction, in which patients are likely to die from cancer-related complications.
Despite several novel findings, our study had some limitations. This was a retrospective study of patients with predominantly upper urinary tract obstruction who received PCN. Another limitation is the exclusion of common causes of obstructive uropathy, such as benign prostatic hypertrophy. In addition, the complex mechanisms that lead to AKI were not determined. However, to our knowledge, this is the first large-scale clinical study to analyze various etiologies, clinical manifestations, short-and long-term outcomes, and related risk factors in adult obstructive uropathy. With the increase in the aging population with multiple comorbidities, the incidence and outcome of adult obstructive uropathy are likely to change. A better understanding of the epidemiology and prognostic factors is needed for the development of proper management strategies for obstructive uropathy.

Methods
Study subjects. We retrospectively screened the medical records of 2127 patients who underwent PCN for obstructive uropathy between January 1, 2002 and December 31, 2018, in three tertiary hospitals in Korea. Four hundred forty four patients who were under age of 18 years, diagnosed with ESKD or died within 48 h of admission or obstruction due to BPH were excluded. Because lower urinary tract obstruction is frequently relieved by bladder decompression, we excluded those with obstruction due to BPH or other urethral pathologies. The study was approved by the Institutional Review Board of Korea University Anam Hospital (IRB number 2019AN0403) and conducted according to the criteria set by the Declaration of Helsinki. This study was a retrospective medical chart review, so informed consent was waived by Institutional Review Board of Korea University Anam Hospital.
Baseline demographics and clinical characteristics. Demographic features, and clinical and laboratory data of the patients at the time of admission were reviewed. The presence of comorbidities was identified using the International Classification of Disease-9 (ICD-9) code. Estimated GFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, and a prior history of eGFR < 60 ml/ min/1.73 m 2 for ≥ 3 months was considered as CKD. Baseline kidney function was defined as eGFR determined within 3 months before admission or the highest eGFR recorded during the follow-up period when baseline serum creatinine was not available. AKI was defined and staged according to the Kidney Disease Improving Global Outcomes (KDIGO) 18 .
Outcomes. Patients were followed up until December 2019. The median follow-up period for mortality and ESKD progression was 83 months (range 0-219 months) and 91 months (range 12-219 months), respectively. Short-term recovery was defined as the recovery of kidney function to either eGFR ≥ 60 ml/min/1.73 m 2 , ≤ 25% of baseline eGFR, or attainment of dialysis independence within 7 days after PCN insertion. Patients were classified as progressing to ESKD if they needed maintenance hemodialysis, peritoneal dialysis, or kidney transplantation. ESKD data were obtained from the Korean Society of Nephrology ESKD Registry. Mortality data were obtained from hospital records and the National Database of Statistics, Korea.

Statistical analysis.
Continuous variables with standard distributions were described as the mean ± standard deviation using Student's t-test. If the variable distributions were skewed, they were described as medians (minimum-maximum). Categorical variables were analyzed using the Chi-square test. One-way analysis of variance with Bonferroni post-hoc analysis was used to compare two or more groups. A decision tree model was used to identify the risk factors for AKI and performed logistic regression analyses to assess the risk factors for AKI or recovery from AKI. Cox regression analysis was used to explore the risk factors of mortality. Statistical analyses were performed using IBM SPSS Statistics for Windows, version 28.0. Statistical significance was set at p < 0.05.