INTRODUCTION

National government support for Renal Replacement Therapy (RRT), which started after the 1970s in the US, has led to a tendency to early initiation of dialysis in chronic kidney disease (CKD) patients until the 2010s⁽1^,2⁾.

Several studies have shown that initiating early dialysis did not provide a meaningful survival advantage in patients with CKD⁽3^,4⁾.

It is known that initiating early hemodialysis treatment may worsen the patient's residual kidney function more rapidly, lead to adverse cardiovascular events or sudden death, and also affect patients' survival through infections and catheter-related complications⁽5^,6^,7^,8⁾. On the other hand, some studies claim that starting early dialysis in patients with CKD provides at least a modest improvement in mortality⁽2^,9^,10⁾. In addition, volume overload, fluid-electrolyte, and potassium homeostasis disturbances are frequently present in patients not hastened for dialysis treatment. Cardiac arrhythmias and cardiovascular diseases caused by these disturbances can be considered as the causes that increase mortality in delayed initiation of dialysis⁽11⁾.

In the context of mortality associated with dialysis treatment, acute kidney injury requiring dialysis (AKI-D), which represents the more extreme version of acute kidney injury (AKI), is known to be at high risk for mortality, morbidity, and end-stage renal disease⁽12^,13^,14⁾.

The relationship between AKI requiring dialysis and short or long-term mortality in hospitalized patients or critically ill Intensive Care Unit (ICU) patients has been demonstrated in many studies⁽15^,16^,17⁾. At the same time, there are conflicting data on the long-term outcomes and mortality of CA-AKI⁽18^,19⁾.

In this study, in patients who started on hemodialysis for the first time, we aimed to investigate 1-) the short (90 days) and long-term (one year) mortality, 2-) the relationship between clinical-biochemical variables at admission and mortality, 3-) kidney disease outcomes of the patients after one year from the first dialysis.

METHODS

In this study, we analyzed retrospectively patients who had started intermittent hemodialysis for the first time in the hemodialysis unit of the tertiary care hospital. The study population consisted of patients starting dialysis while on ambulatory patient follow-up due to CKD, those starting on urgent dialysis with CA-AKI diagnosis, and general ward patients requiring hemodialysis on admission (HA-AKI).

We excluded patients with the following criteria from the study: 1-) Patients who had previously received dialysis treatment for any reason. 2-) Patients who received hemodialysis due to drug intoxication (digoxin, lithium, methyl alcohol, etc.). 3-) patients who underwent hemoperfusion due to mushroom poisoning, and 4-) patients who started on dialysis in the intensive care unit.

In addition, patients with missing clinical-biochemical data were also excluded from the study.

The clinic and demographic characteristics of the patients: age, gender, type, and cause of kidney disease, indication for dialysis, duration of dialysis, vascular access for dialysis, concomitant diseases and medications, and complications were recorded.

If the patients had kidney disease before dialysis or if the serum creatinine values of patients were high within three months before the date of dialysis, or if the kidney sizes of patients were small on renal ultrasonography, were recorded as having previous kidney disease (Group 1).

Patients who did not meet these criteria were categorized as having no previous kidney disease (Group 2). Systolic and diastolic blood pressures during dialysis and ejection fractions (EF) of patients who underwent echocardiography were obtained from the dialysis charts.

The biochemical data of the patients on the day of dialysis: serum glucose, urea, creatinine, calcium, phosphorus, parathormone, sodium, potassium, and albumin levels), hemogram parameters, C-reactive protein (CRP), hepatitis serologies (HBsAg, anti-HBs, AntiHCV) and blood gas values (pH, PCO2, HCO3) were obtained from the hospital electronic data record system.

We determined and recorded the patient's survival at 90 days and one year and renal disease outcomes (recovery, CKD without dialysis, and Renal Replacement Therapy) in the living patients at the end of one year.

STATISTICS

Statistical analyzes were performed using SPSS Statistic software 24 (SPSS Inc., Chicago, III).

Continuous variables are summarized as mean or median (interquartile range).

The categorical measurements were summarized as numbers (percentage of the diagnostic group). All continuous variables were categorized according to optimal or international cutoffs. Optimal cutoff values were determined by the receiver operating characteristic (ROC) curve and the area under the curve (AUC).

Univariate and multivariate logistic regression analyses were used to identify predictive factors for short-term and long-term mortality-dependent variables.

The Odds Ratio (OR) was reported with the corresponding 95% confidence intervals (95% CI). The receiver operating characteristic curve (ROC curve) and ROC-AUC were calculated to compare the independent prognostic factors. Statistical significance was accepted as p < 0.05.

RESULTS

A total of 229 patients were included in the study. 133 (58.8%) of the patients were male, 96 (41.9%) were female, and the median age was 64 years. 91.8% of the patients had at least one comorbid disease. While 166 patients (72.5%) had pre-existing renal disease (Group ¹), 63 patients (27.5%) had no prior renal disease (Group 2). The most common reasons for dialysis were hypervolemia (41%) and uremic symptoms (24%). A non-tunneled jugular catheter was used in 60.3% of patients as the vascular access route for hemodialysis, while a femoral catheter was used in 35.4%. Catheter-related complications (infection or thrombosis) were detected in 19 patients (Table 1).

Forty-nine (21.40%) patients died within the first 90, and 73 (31.9%) patients died within the first year (Table 1). At the end of one year, 38% of the patients continued to receive renal replacement therapy. In contrast, 10% of all patients had CKD not requiring dialysis, and only 9.17% had renal recovery (normal renal functioning).

Ideal cut-off values for predicting 90-day and one-year mortality.

ROC-AUC curves determined the ideal cut-off values of laboratory variables for predicting short-term (90 days) and long-term (one year) mortality. For predicting mortality within 90 days, the ideal cut-off value was 63.5 years (AUC: 0.812) for age, 114 mmHg (AUC: 0.724) for systolic blood pressure, 60 mmHg for diastolic blood pressure (AUC: 0.747), 7,15 mg/dL (AUC: 0.681) for creatinine, 291500 (AUC: 0.584) for platelet, 1.55 mg/dL (AUC: 0.659) for CRP, and 29 mmHg (AUC: 0.659) for pCO2 (Figure 1 A).

In terms of predicting mortality within one year, the ideal cut-off value was 63.5 years (AUC: 0.768) for age, 120 mmHg (AUC: 0.692) for systolic blood pressure, 60 mmHg (AUC: 0.684) for diastolic blood pressure, 7.35 mg/dL (AUC: 0.615) for creatinine, 256 500 (AUC: 0.602) for platelets, 2.26 mg/dL (AUC: 0.622) for CRP, and 29.5 mmHg (AUC: 0.587 ) for pCO2, 20.5 (AUC: 0.592) for bicarbonate.

The International cut-off values, for ejection fraction, potassium, and albumin were considered (Figure 1 B).

Figura: 1 A y B

Univariate and multivariate analysis of factors predicting mortality within 90 days:

In the univariate analysis the age (OR [odds ratio] = 0.56, 95% CI: 0.02-0.16, p <0.001), etiology of AKI (OR = 1.07, 95% CI: 1.01-1.13, p=0.013), vascular access for dialysis (OR=3.18, 95% CI: 1.71-5.92, p< 0.001), EF (OR=3.55, 95% CI: 1.53-8.27, p= 0.003), systolic blood pressure (OR=0.12, 95% CI: 0.06-0.27, p< 0.001), diastolic blood pressure (OR=0.09, 95% CI: 0.04-0.20, p< 0.001), creatinine (OR=0.22, 95% CI: 0.11 -0.44, p<0.001), potassium (OR=0.38, 95% CI: 0.20-0.73, p=0.004), albumin (OR=4.67, 95% CI: 2.20-9.99, p<0.001), platelet (OR=0.44, 95% CI: 0.22-0.87, p=0.018), CRP (OR=0.15, 95% CI: 0.06-0.39, p<0.001), pCO2 (OR=2.07, 95% CI: 1.06-4.05, p<0.001) showed to be significant for predicting short-term mortality (Table 2).

Table 2. Univariate analyses of factors for patients died within 90 days and 1 year

s Significant values are indicated in bold.

AKI: Acute Kidney Injury, CRP, C-reactive protein, * A: Hypervolemia, B: Uremic Findings, C: Hyperpotasemia, D: Anuria, E: Acidosis

In the multivariate model established with significant variables in the univariate model; EF (OR = 3.80, 95% CI: 1.05-13.72, p=0.042), CRP (OR = 0.20, 95% CI: 0.04-0.92, p= 0.039), age (OR = 0.21, 95% CI: 0.05-0.91, p= 0.038), and diastolic blood pressure (OR = 0.08, 95% CI: 0.02-0.28, p< 0.001) remained independent predictive for short-term mortality (Table 3).

Table 3: Multivariate analyses of factors for patients died within 90 days and 1 year

s Significant values are indicated in bold, Pf.: Forward: LR method, CRP: C-reactive protein

Univariate and multivariate analysis of factors predicting mortality within One year:

Univariate analysis for mortality within one year showed that age (OR = 0.16, 95% CI: 0.08-0.31, p< 0.001), etiology of AKI (OR = 1.06, 95% CI: 1.01-1.10, p= 0.015), indication for dialysis (OR = 1.21, 95%) CI: 1.02-1.44, p= 0.026), vascular access for dialysis (OR = 2.38, 95% CI: 1.40-4.05, p= 0.001), EF (OR = 2.50, 95% CI: 1.24-5.03, p = 0.010), systolic blood pressure (OR = 0.18, 95% CI: 0.09-0.35, p< 0.001), diastolic blood pressure (OR = 0.09, 95% CI: 0.04-0.21, p< 0.001), creatinine (OR = 0.001) 0.36, 95% CI: 0.20-0.65, p= 0.001), potassium (OR = 0.38, 95% CI: 0.22-0.68, p= 0.001), albumin (OR = 4.69, 95% CI: 2.19-9.99, p< 0.001), platelets (OR = 0.46, 95% CI: 0.26-0.82, p= 0.009), CRP (OR = 0.39, 95% CI: 0.21-0.71, p= 0.002), pCO2 (OR = 2.51, 95% CI) : 1.36-4.65, p= 0.003), and bicarbonate (OR = 2.49, 95% CI: 1.25-4.98, p< 0.001) were predictive (Table 2).

Established multivariate analysis for one-year mortality indicated that systolic blood pressure (OR = 0.26, 95% CI: 0.08-0.82, p= 0.022), diastolic blood pressure (OR = 0.21, 95% CI: 0.68-0.66, p= 0.008), and potassium (OR = 0.27, 95% CI: 0.10-0.70, p= 0.007) were independent predictive (Table 3).

DISCUSSION

In this study, we investigated the 90-day and 1-year mortality and mortality-affecting factors of patients having started hemodialysis for the first time.

In the multivariate model, we established: age (≥63.5 years), diastolic blood pressure (<60.5mmHg), EF (<50), and CRP (≥1.55) together could predict mortality within 90 days.

Regarding mortality within one year: systolic blood pressure (<120 mmHg), diastolic blood pressure (<60.5 mmHg), and potassium (≥ 5.1 mg/dL) showed to be independent predictive in multivariate analysis. Moreover, we found the mortality rates within 90 days and one year to be approximately 20% and 30%, respectively.

At the end of one year, we detected renal recovery in 9% of the patients; approximately 40% continued renal replacement therapy, and 10% had chronic kidney disease not requiring dialysis.

Despite our study including both HA-AKI and CA-AKI patients, it is noteworthy that it represents the short and long-term outcomes of patients who started dialysis for the first time in the outpatient hemodialysis unit.

Previous studies indicated that hypotension increases mortality in patients with AKI⁽20⁾. Lins R L et al. showed that hypotension increases mortality in hospitalized patients with episodes of AKI⁽21⁾. Sepsis, as the possible reason for hypotension, can be claimed as the main culprit of this increase in mortality. Hypotension due to sepsis has been shown to increase mortality consistently in AKI, particularly in intensive care unit patients⁽22⁾.

On the other hand, in addition to sepsis, hemodynamic instability due to cardiorenal syndrome or hypovolemia may also be associated with mortality. Considering that more than 60% of the patients in our study had pre-existing CKD, it can be put forward that these patients have an increased risk of death related to hemodynamic instability caused by volume overload. There is evidence that the prognosis is poor in patients with volume overload who start dialysis later. In a study by Bouchard et al., patients who started dialysis due to hypervolemia had a higher risk of death in their 60-day follow-up than patients who underwent hemodialysis for other indications⁽23⁾. Similarly, in a study comparing continuous and conventional dialysis, patients who started on dialysis for electrolyte dysregulation had better survival outcomes than patients on dialysis due to volume overload⁽24⁾.

In our study, systolic or diastolic hypotension was associated with higher mortality in our patient group who started on hemodialysis for the first time.

Bell JS et al. reported higher CRP, diabetes, and lower systolic blood pressure as high-risk factors for CA-AKI⁽25⁾. Duarte, Inês, et al. found hypoalbuminemia and higher CRP as predictors for mortality in AKI-D patients⁽26⁾. Consistent with the literature, we found high levels of CRP, hypoalbuminemia, and hypotension to be factors predicting mortality in our study.

Most studies in the literature report that age is a factor that increases in-hospital mortality in AKI. Lu, R. H. et al. reported that in-hospital mortality in hospitalized patients with AKI increased in older people⁽27⁾. Ejaz, Taymmia, et al. reported that sepsis and age increased the risk for mortality in HA-AKI⁽28⁾. The present study determined that a higher age at hospital admission was a factor in predicting mortality in our patient population.

With a decrease in GFR, potassium tends to rise due to deficiencies in potassium homeostasis.

It is known that increasing hyperkalemia is directly related to oxidative stress, cardiac disease development, and mortality by stimulating the arrhythmogenic effect and aldosterone release. In a study that included 234 AKI patients by Abebe et al., they reported that those with hyperkalemia were significantly associated with higher mortality⁽29⁾. In a study based on electronic health records, Nakhoul, Georges, et al. reported the effects of hyperkalemia on mortality in approximately 35 000 patients; the overall mortality of patients with potassium levels >5.5 mmol was higher⁽30⁾. In our study, hyperkalemia was an independent predictor of mortality at one-year mortality.

It has been reported in previous studies that the risk of all-cause death due to tunneled catheters increases in dialysis patients⁽31^,32⁾. Femoral vascular accesses cause more infection-related complications than jugular vascular accesses, probably due to their location.

It has been reported that the risk of bacteremia increases weekly for femoral catheters but does not increase until the third week for jugular catheters⁽33⁾. In our center, approximately 60% of the patients entered hemodialysis with a jugular catheter, and catheter infection was detected in 6.55%. According to the literature, dialysis via a femoral catheter was associated with mortality in our study.

This study has some limitations. The retrospective design is the main limitation of the study. In addition, the included patients with CKD were in different stages of the disease, and the cause of death of those who died is unknown, adding additional limitations. On the other hand, the strength of the present study is that the patients included were a particular group, representing the severe form of AKI requiring dialysis.

Very few studies investigated the factors that predict and affect prognosis in patients starting dialysis for the first time outside the ICU. We found hyperkalemia, hypotension, EF, and age to be predictors of death. We recommend dealing with early management of clinical conditions such as hemodynamic instability or sepsis that may cause hypotension in patients starting hemodialysis treatment and that the fluid-electrolyte balance should be optimized carefully in the patients at risk that we have determined. In addition, patients with CKD who have not yet started hemodialysis treatment should be followed closely, as hypervolemia, hypotension, and hemodynamic instability increase the risk of death, according to our study.

Recibido: 25-12-2022

corregido:27-01-2023

Aceptado: 12-05-2023