Emphysematous Pyelonephritis: A Correlation of Clinical Characteristics With Serum Parameters and Its Contribution in Determination of Prognostic Factors

Background: Emphysematous pyelonephritis (EPN) is a severe life-threatening infection of renal parenchyma, peri-renal tissue and collecting system primarily caused by gas forming bacteria and carries a significant mortality of 80% when primarily treated non-surgically. Early identification of the disease process along with timely surgical management is of paramount significance. In the last 2 decades evolution of surgical management has lead to a substantial decrease in mortality of up to 21%. The focus of the present study was to investigate optimal serum panel parameters and to propagate prognostic model for the optimal evaluation management of EPN.

Methods: The present study was conducted at the Department of Urology at GMCH, Gauhati. Authors retrospectively analyzed all clinical, laboratory and radiological data along with outcome of 44 patients with EPN from January 2015 to August 2017. Chi-square analysis and Spearman tests were used to carve out significant association and correlation.

Results: The overall survival rate was 86.37% (38/44). Escherichia coli (E.Coli) was the most common bacterial pathogen isolated from urine and pus culture. Uncontrolled type 2 diabetes mellitus (DM), polymicrobial infection, shock, thrombocytopenia, bacteremia, hypoalbuminemia, and hemodialysis were significantly more common in cases of death and correlated with adverse serum profile. The AUC of the ROC curve for the seven prognostic factors was 0.59 and only thrombocytopenia (TCP) (AUC = 0.78, P = 0 .033) along with liver cirrhosis (AUC = 0.76, P = 0.018) came out to be significant post ROC analysis.

Conclusions: Upfront clinical factors such as uncontrolled DM type 2, polymicrobial infection, shock, thrombocytopenia, bacteremia, hypoalbuminemia, hemodialysis along with altered serum parameters are significantly correlated with outcome in patients of EPN.

Keywords: Polymicrobial infection; Thrombocytopenia; Bacteremia; Hypoalbuminemia; Hemodialysis; Emphysematous pyelonephritis

Introduction

Emphysematous pyelonephritis (EPN) is an infrequent type of necrotizing infection, and is lethal if not managed promptly. EPN is defined as a subtle and serious infection of the renal parenchyma and peri-renal tissue, which results in gas within the renal parenchyma, collecting system or perinephric tissue. Although it was first described in 1898, the term EPN was first used in 1962. It has a female: male predominance of 4:1 and a mean (range) age of 57 (24 - 83) years [1]. Diabetes mellitus (DM) is a major source for these infections and is also associated with an increased risk of asymptomatic bacteriuria and certain symptomatic UTIs (urinary tract infections) such as renal and perinephric abscess, cystitis and Candida infections [2, 3]. The exact etiopathogenesis of emphysematous UTIs is poorly understood. Elevated tissue glycemic levels in diabetic patients may provide a more favorable microenvironment for gas-forming microbes. Nonetheless, bacterial gas production does not fully explain the pathologic and clinical manifestations of emphysematous UTI [4, 5]. Uncontrolled DM and urinary tract obstruction are the major risk factors for emphysematous UTIs. In past studies, diabetes was present in more than 80% of patients with EPN, at least 50% of patients with emphysematous pyelitis, and 60-70% of patients with emphysematous cystitis [6-8].

Though it is a serious and life-endangering disease with a mortality rate of up to 50%, there is no clear consensus on the optimal treatment of these patients [9]. The mainstay of management is an emergency nephrectomy (EN), percutaneous drainage (PCD), or medical management (MM), with or without stenting of the urinary tract. Few centers recommend that timely EN should be the primary treatment [10], but other studies proposed a more conservative approach with PCD [11, 12], and a conservative mode with IV antibiotics and other supportive analgesics (non- NSAID) alone has also been advocated [13, 14], which further enhances the management scenario. To standardize the management protocol, a few studies have observed mortality rates stratified by risk variables, separating which patients need more urgent action [15, 16-18]. However, these studies were only based on few patients but also had different results for the same risk variable. Nevertheless, because there is no formidable evidence, these studies remain the only guide for managing EPN. Therefore the aim of this study was to systematically review past reports and to ascertain the mortality rate of each treatment from the available series. Furthermore, the present study aimed to risk-stratify the prognostic factors across the published studies to determine the specific factors determining poor outcome of EPN and to throw light on the clinical and laboratory presentation of patients with EPN.

Methods

The present study was conducted at the Department of Urology and Renal Transplantation at GMCH, Gauhati. Authors retrospectively analyzed all clinical, laboratory and radiological data along with outcome of 44 patients with EPN from January 2015 to August 2017. The main objective was to determine the mortality rate associated with EPN, depending on the various treatments offered. The three main treatments reviewed were EN, PCD and conservative MM. The secondary aim was to assess the risk factors associated with death along with the most common presenting symptoms and most common causative organisms. Where available, patients with risk factors for death were compared to those with no or minimal risk factors. Furthermore, the risk of death was assessed based on the different types of classification of EPN.

According to the Huang system, class 1 EPN is defined as gas in the collecting system only; class 2 as gas in the renal parenchyma with no extension to the extrarenal space; class 3A as extension of gas or abscess to the perinephric space; class 3B as extension of gas or abscess to the pararenal space; and class 4 as bilateral EPN or EPN in a solitary kidney [19]. Clinical and biochemical variables were extracted from patient records from different departments depicting information on demographics, length of hospitalization, presenting symptoms, diagnostic tests, culture results, laboratory findings, treatment, and mortality rate. For continuous data, a Mantel-Haenszel Chi-square test was used and expressed as the mean difference with 95% CI, and for dichotomous data an inverse variance was used and expressed as the odds ratio (OR) with 95% CI. In both groups P < 0.05 was considered to indicate significance. Thrombocytopenia (TCP) was defined as a platelets less than 120,000/mL, hyponatremia as serum Na < 135 mEq/L, hypoalbuminemia as serum albumin < 3.0 g/dL. Septic shock was defined as persisting hypotension requiring vasopressors to maintain a mean arterial pressure (MAP) of 65 mm Hg or higher and serum lactate > 18 mg/dL. Patients with an absolute increase in serum creatinine of ≥ 0.3 mg/dL after admission compared with nadir serum creatinine level were diagnosed with AKI (acute kidney injury). Recurrent EPN was diagnosed when both clinical presentation of sepsis and progressive lesions on the imaging study were noted within 3 months after adequate treatment. A poor outcome was declared in cases of death or recurrent EPN. The management modalities consisted of antibiotics alone; PCD with antibiotics; and nephrectomy, if progressive or persistent lesions were detected on imaging studies with clinical scenario of unstable hemodynamics or prolonged fever on MM.

Results

From a period of January 2015 to August 2017, 44 patients were diagnosed with EPN based on clinical, biochemical suspicion along with contrast enhanced computed tomography (CECT) W/A report. There were 28 female and 16 male out of which five female (17.85%) and two male (12.5%) patients were eventually lost during the course of their treatment (Fig. 1). The median age of the patients in the study was 60 years (SD = 11.253) (Fig. 2).

Figure 1. Showing gender wise distribution with EPN frequency.

Figure 2. Showing histogram depicting median age with frequency of disease.

A total of 33 patients (75%) had uncontrolled type 2 DM. Obstructive uropathy and urolithiasis occurred in 38.6% (n = 17) and 43.2% (n = 19) patients respectively. Other clinical characteristics which were assessed were hypertension (22 patients, 50%), hyponatremia (38 cases, 86.4%), liver cirrhosis (25 patients, 56.8%), thrombocytopenia (23 patients, 52.3%), hypoalbuminemia (28 cases, 63.6%), and cerebrovascular accident (12 patients, 27.3%). Three patients presented with septic shock out of which two patients eventually succumbed to their disease. Bacteremia diagnosed with positive blood cultures taken during fever spikes was present in 14.3% (n = 18) cases and one patient suffering from persistent bacteremia died in spite of all measures available at hand. AKI was diagnosed subsequently in 34.1 % (n = 15) cases. Emergency hemodialysis was done in 36.4% (n = 16) cases after consultation with nephrology department. Altered mental state was observed in 17 cases (38.6%). Polymicrobial infection was present in 45.5% (n = 20) and its incidence in the non-survival group was 71.4 % (n = 5). The left kidney was more frequently involved (54.5%, n = 24) than the right side (45.5%, n = 20). Table 1 shows the clinical and epidemiological characteristics of the survivors and non-survivors.

Table 1. Comparison of the Epidemiological and Clinical Profile Between Survivors and the Non-Survivors

The median leukocyte count with interquartile range at admission was 16,220 ± 8,697/µL (7,875 - 20,008) (Fig. 3). Similarly the mean platelet and median albumin count was 1.10 ± 0.23 lakhs/µL and 2.3 ± 0.65 mg/dL (1.82 - 2.76) (Fig. 4, 5). Three patients had class 1, 11 had class 2, 15 each had class 3A and class 3B CECT imaging findings based on the Huang and Tseng classification. E. coli was the most common organism cultured from specimens of urine, blood and pus, and was seen in 63.64 % patients. The other organisms included Klebsiella pneumoniae (22.72%), Proteus mirabilis (18.18%), Enterococcus species (12.5%), and Pseudomonas aeruginosa (11.36%). Urine culture was positive in 36 patients out of which 16 patients had polymicrobial infection. Bacteremia occurred in 18 patients, with polymicrobial bacteremia in 12 and E. coli in six patients. Culture of pus from PCD was positive in 24 of 44 patients, with polymicrobial culture in five, Proteus mirabilis in six, E. coli in 10 and Pseudomonas aeruginosa in three patients (Table 2). The overall survival rate was 84.09% (37/44). Uncontrolled DM type 2 (P = 0.032), hyponatremia (P = 0.013), TCP (P = 0.006), hypoalbuminemia (P = 0.041), bacteremia (P = 0.028), shock on initial presentation (P = 0.013) and polymicrobial infection (P = 0.011) were significantly associated in the non-survivors than in the survivors. There were no statistically significant differences between the non-survivors and survivors with respect to AKI, CVA, AMS, urinary tract obstruction, and urolithiasis.

Figure 3. Showing histogram depicting median leukocyte with frequency of disease.

Figure 4. Showing histogram depicting median platelet with frequency of disease.

Figure 5. Showing histogram depicting median albumin with frequency of disease.

Table 2. Causative Organisms From Blood, Urine and Pus Culture With Their Distribution

Recurrent EPN occurred in seven patients (7/44, 15.91%) and overall seven patients died during the investigation duration. Tables 3 and 4 show the results of multivariate analysis of three significant independent variables in association with mortality and poor outcome based on the information which we gathered from Table 5 as it suggested the important parameters to work with. We also compared the group treated with antibiotics alone with the group that required urological interventions, including PCD and EN (Table 6). Treatment with IV antibiotics alone was successful in 43.18 % (n = 19) of patients and four patients succumbed in IV antibiotic group without giving any time for any surgical drainage or EN. Treatment with PCD and IV antibiotics was successful in 15 of 16 patients and one patient died due to severe sepsis. Five patients underwent EN and three of them survived. In the present analysis, three patients did not survive in the Tseng class 2 with mortality rate up to 27.27% (n = 3) which was maximum for the entire cohort while only one patient succumbed in the class 3A. Two patients were lost while still on MM and one each died just after PCD and EN in the respective group (Table 6). The casualty figure for the class 3B was at 20% (n = 3) with two patients lost during MM and one after EN. There were no significant alterations in the biochemical profile of the patients who did or did not undergo EN. All patients having the triple H condition consisting of hyponatremia (P = 0.014), hypoalbuminemia (P = 0.031), and need for emergency hemodialysis (P = 0.011) required urgent intervention as per the statistical parameters obtained after analyzing their procedural records. Seven independent variables (hypoalbuminemia, shock, bacteremia, need for hemodialysis, polymicrobial infection, hypoalbuminemia, TCP) that were significantly different between survivors and non-survivors along with liver cirrhosis, were evaluated by ROC analysis and the AUC. The AUC of the ROC curve for the seven prognostic factors was 0.59 and only TCP (AUC = 0.78, P = 0.033) along with liver cirrhosis (AUC = 0.76, P = 0.018) came out to be significant post ROC analysis (Fig. 6).

Table 3. Multivariate Analysis Predicting Significant Variables for Mortality

Table 4. Multivariate Analysis Predicting Significant Variables for Morbidity

Table 5. The Odds Ratio and Correlation of Clinical Parameters Between Survivors and Non-Survivors

Table 6. Comparison of the Clinical Profile Between Procedure (PCD/EN) and the Medical Management Group

Figure 6. Showing ROC analysis of two significant variables: TCP and liver cirrhosis.

Discussion

EPN demands urgent management because of the life-threatening nature of the septic complications. The present retrospective study sought to chalk out the parameters that were marker of high mortality and unfavorable outcome. From the available past literature, there is an ascendancy of EPN in females, owing to their increased perceptivity to UTI [20, 21]. The female to male ratio of our cohort is in conjunction to that reported by one of the study, but much higher ratios have been reported previously [22, 23]. As per other series, our maximum patients presented with fever with or without flank pain. On an average, our patients had symptoms for 3 days or more before hospital admission which further explains why in few patients the infective etiology had already lapsed into septic shock. Laboratory parameters in our study conformed to those published in the literature, with TCP, leukocytosis and renal impairment being the most common findings [8, 24, 25].

After examination of the entire panel of biochemical and clinical parameters it was observed that DM type 2, hyponatremia, liver cirrhosis, TCP, hypoalbuminemia, shock, bacteremia, need for HD and polymicrobial (PMI) as significantly associated with the outcome of EPN. DM type 2 appeared to be the most important risk factor for EPN resulting in serious complications (Table 5). Since a high tissue dextrose level provides a congenial environment for the growth of gas-forming micro-organism, DM type 2 was associated with increased mortality in the present study (P = 0.032, OR = 0.18), with significant correlation with the outcome (r = -0.523, P = 0.035). The present outcome is in agreement with most of the past literature but goes against one of the published past study [26].

In ASA grade 3, four patients’ albumin concentration of < 3.5 g/dL has been correlated with increased morbidity which reflects the low synthesis, prompt degradation and loss of albumin, in addition to faulty exchange between the intravascular and extravascular compartments. In an analysis of over 15,000 diseased patients, an admission serum albumin < 3.4 g/dL was seen as a strong forecaster of mortality [27]. It has been observed that survival in AKI is significantly affected by the level of serum albumin [28]. In our retrospective cohort, hypoalbuminemia (< 3.0 g/dL) was associated with an increased mortality rate (2/7, 28.6%, P = 0.041). Though accelerated breakdown of protein is a precursor of metabolic alterations in patients with AKI, we could not find any such significant association or correlation with the same independently in our study. Generally albumin production is suppressed in response to inflammatory scenario but hypoalbuminemia was probably associated with the AKI in our cohort. When the patients with liver cirrhosis were also included along with hypoalbuminemia (Chi-square coefficient 6.33, P = 0.012), there was a significant correlation between survival and the same (r = 0.379, P = 0.011). Therefore, hypoalbuminemia along with liver cirrhosis is a factor that should be evaluated as a predictor of outcome in patients with EPN.

TCP (platelet count < 120,000/mL) at admission has been projected to be an independent prognostic factor for poor outcome. In our present study, TCP was significantly associated with mortality (P = 0.01). All the seven patients in the non-surviving group had TCP along with severe liver cirrhosis, which might result in TCP.

Need for hemodialysis, hyponatremia and hypoalbuminemia during management was identified as a marker for urgent intervention in the intervention vs. MM group in multivariate analysis. Septic shock at time of admission, bacteremia, hypoalbuminemia, DM type 2 and polymicrobial infection were found to be predictor of poor outcome on multivariate Cox regression analysis. It was further observed that no significant association exist between higher mortality in EPN and overall age and sex or other comorbidities, such as HTN, CVA, obstructive uropathy, urolithiasis, AKI and AMS.

E. coli was the most common organism and present in 63.64% patients followed by PMI which was present in 45.50% evenly distributed in the 2 year period. Regarding the bacterial resistance, one patient grew methicillin-resistant Staphylococcus aureus (as a constituent of PMI) in his blood, and another patient had extended-spectrum beta-lactamase Klebsiella pneumonia (1/8) from their urine. Pertaining to the management policy of EPN, the usual management of EPN until 1980s has been EN and/or open surgical drainage in addition with IV antibiotics, with a mortality rate of 40-50%. The advent of PCD techniques led to maximum nephron sparing and return of renal function [29]. The last 2 decades has seen EPN mortality declining to 21% owing to the widespread use of PCD. Majority of our cases received IV antibiotics alone or PCD in company with antibiotics in this 2-year investigation period. In our analysis management with IV antibiotics alone was successful in majority of patients and four patients died without giving any time for PCD or EN. Treatment with PCD and IV antibiotics was successful in 93.75% patients and one patient died due to severe sepsis. Five patients underwent EN and 60% of them survived. EN was done out for prolonged fever and sepsis in these five patients. In the present analysis, three patients did not survive in the Tseng class 2 with mortality rate up to 27.27% which was the maximum for the entire cohort while only one patient succumbed in the class 3A. Two patients were lost while still on MM and one each died just after PCD and EN in the respective group. The casualty figure for the class 3B was at 20% with two patients lost during MM and one after EN. These observations strongly indicate that the management of septic shock and appropriate IV antibiotic therapy are priorities during initial EPN treatment. A judicial diagnosis is pivotal, and aggressive treatment should be pursued in these cases. Early studies noted that early and prompt surgical treatment resulted in a successful outcome, as compared with medical treatment alone [7, 30]. In 1986, one study proposed that USG guided PCD treatment of EPN with good clinical outcome [31]. Since then, many EPN series have reported favorable outcomes with PCD treatment, which has now been regarded as the initial mode of treatment in the majority of EPN patient [32]. It has been postulated that the aim of EPN management should not be mere survival but also to increase the chance of renal recovery. In contrast to early analysis, recent EPN series have found a higher mortality resulting from EN vis a vis conservative approach consisting of IV antibiotics and PCD [15, 33]. Nonetheless, investigators have warned against the universal use of PCD in all EPN cases, with subsequent EN only if required. One study stated that of the 19 patients who had PCD as management protocol, the procedure failed in 11 of them, five of whom eventually died [34]. In Huang and Tseng’s series, 92% of the patients with poor risk CECT class 3 and class 4 failed PCD miserably, with a mortality rate of 15% [35]. In our analysis, we found a higher mortality among conservatively treated patients as compared with those who received any form of early intervention. When we analyzed the fate of patients who were conservatively managed to those who received EN, although a larger proportion of patients collapsed from conservative treatment alone, the difference did not reach statistical significance (P = 0.845) which is most likely related to our small sample size. In one large analysis of EPN series, conservative treatment, type 1 EPN (Wan system), thrombocytopenia, and bilateral EPN were predictors for mortality [36]. In a recent series, the researchers found AMS, TCP, renal failure and severe hyponatremia to be associated with higher mortality in EPN [37].

In 2000, Huang and Tseng described four distinct radiological classifications of EPN based on the extent of gas seen on CECT. There was a trend towards higher mortality in patients with protracted disease, and patients with type 1 EPN had the best outcome. Nonetheless, in the present analysis, the fallout of patients with class 3 EPN was not statistically different from those with class 2 disease. Consistent with past studies, there was also no statistically significant difference in our study of the rate of class 2 and 3B CT images in patients requiring EN because of poorly controlled infection versus those not requiring EN. It has been postulated that the degree of renal parenchyma destruction rather than the extent of gas predicted the need for EN. When equal significance was applied to each risk factor identified by univariate analysis, trend analysis reported that the risk of death was directly associated with the number of prognostic factors (P = 0.002). There were significant digression in risk of mortality between patients who had ≥ five prognostic factors (P = 0.02), which is in disjunction to the results of Khaira et al [26]. As per our knowledge this is the first sojourn to identify the number of prognostic factors that were significantly associated with mortality and poor outcome in this part of the country.

Present study had several limitations. First, it was a retrospective study and the number of cases was too few to analyze other risk factors which led to lack of significance of some of the factors analyzed. Second, we did not designate a specific follow-up period for discharged patients irrespective of the management offered. However, we believe that a 6-month follow-up period is sufficient to demonstrate the trends of the disease outcomes

Conclusions

The authors in the present retrospective analysis reported overall mortality of 15.91% in present study vis a vis 18% of several previous meta-analysis. The AUC of ROC for TCP and liver cirrhosis had statistically significant power to predict mortality. Patients without DM type 2 had urinary albumin acting as a substrate for the gas-producing organism. PMI acts a milestone for sudden deterioration. Admission serum albumin < 3.32 g/dL was a strong predictor of mortality. Both PCD and MM were associated with significant lower mortality rates than with EN alone. If both PCD and MM did not result in enhancement of the recovery, bypassing of obstruction was needed before contemplating EN. Last but not the least the amount of renal destruction rather than the extent of gas predicted the need for EN.

References