Outcomes of Acute Renal Failure Patients Requiring Intermittent Hemodialysis

Abstract

The published studies on the prognosis of patients requiring intermittent hemodialysis (IHD) are scarce and have some conflicts. In this study, we retrospectively analyzed our data on ARF patients who were treated with IHD. A total of 192 (female: 85, 44.3%; male: 107, 55.7%) patients were included in the study. The mean age was 56.3 ± 17.1 years. In all, 48.9% of the patients were older than 60 years. The mean number of IHD sessions was 7.8 ± 8.0 per patient; 12.4% was due to prerenal causes, 76.8% was due to intrarenal causes, and 10.8% was due to postrenal causes. The leading indication of the IHD was uremic symptoms (46.8%). With the exclusion of hypertension, 72.4% of the patients had at least one systemic comorbidity. After treatment, 75.5% of the patients recovered, in contrast to 9.4% of patients who were transferred to chronic renal replacement programs and 15.1% who died during IHD period. Pre-dialytic serum creatinine (p = 0.003) and albumin levels (p = 0.016), total IHD session number per patient (p = 0.003), and age (p = 0.034) were the parameters that were related to high mortality in statistical analysis. Mortality was higher if the leading indication of IHD was biochemical disturbances (p = 0.013). Diabetes mellitus did not influence mortality. Consequently, predialytic serum creatinine and albumin levels may be very important predictors of mortality. Patients in high-risk groups (older age, female sex, and low pre-dialytic creatinine and albumin levels) should be considered to be treated with slow continuous renal replacement methods.

Keywords:

• acute renal failure
• intermittent hemodialysis
• mortality
• serum albumin
• serum creatinine

INTRODUCTION

Acute renal failure (ARF) has been defined as the loss of kidney functions within days or weeks. Most of the patients recover with supportive treatment, while some others need dialysis. Intermittent hemodialysis (IHD) is the most frequently used dialysis modality for these patients. Approximately one-third of the ARF patients need dialysis, and mortality of these patients is higher than the patients who do not need dialysis.[1] Some studies have been published on the prognosis of the patients suffering from ARF (with or without need of dialysis), especially from the intensive care units. On the other hand, data that have been particularly focused on the patients treated only with IHD is scarce. Due to different population characteristics, it is difficult to extrapolate present data to all ARF patients. Therefore there is a necessity for specific studies on the ARF patients who require IHD.

In this study, we retrospectively analyzed our data of ARF patients who were treated with IHD and tried to find out factors effective on mortality rate in this particular group of patients.

MATERIALS AND METHODS

This retrospective study included adult patients who suffered from ARF, and were treated with IHD from January 2002 to December 2005. The patients followed in intensive care units and treated with slow-continuous replacement therapies (due to multiple organ failure, severe respiratory insufficiency, severe sepsis, post-cardiac surgery, advanced liver failure, etc.), as well as end stage renal failure patients who were started on chronic dialysis program, were excluded. But “acute on chronic” patients were included, as were the patients with ARF who had transplanted kidneys.

All of the data were collected from the hospital records and registered in specific designed sheets. Age and sex of the patients, systemic comorbidities, urinary volume, type of ARF (prerenal, intrarenal, postrenal), causes of ARF, use of nephrotoxic agents, the leading indication for IHD treatment, type of vascular access, and the number of dialysis sessions were recorded. Hypertension was assumed as a sign of renal failure and thus was not categorized in systemic comorbidities.

In this study, to make more detailed analyses, ARF was classified as prerenal if it were related to long renal hypoperfusion (e.g., hypotension, hypovolemia, shock). In another word, the patients with extended prerenal azotemia were categorized to prerenal ARF group instead of ATN. If ARF were related to the obstruction of the urinary tract, it was accepted as postrenal ARF. All other renal failures were classified as intrarenal ARF. The drugs that are known as nephrotoxic were recorded, and nephrotoxic agent-related ARF was categorized as ATN; rapidly progressive glomerulonephritis was categorized as AGN.

In addition, some pre-dialytic biochemical parameters, which had been taken in the same day or one day before the initiation of IHD (i.e., BUN, creatinine, sodium, potassium, calcium, phosphorus, total protein, and albumin) were noted. The etiology of ARF was differentiated by means of history, physical examination, course, and laboratory assessment of the patients. Kidney biopsy was not performed in all patients. Only bicarbonate-based dialysis solutions were used in dialysis sessions.

In statistical analysis, SPSS program was used. Numeric data were expressed as mean ± standard deviation. Pearson and Spearman's correlation tests were used for correlation analyses.

RESULTS

A total of 192 patients (85, 44.3 % female; 107, 55.7 % male) were included in the study. The mean age was 56.3 ± 17.1 years; 48.9% of the patients were older than 60.

Considering the etiology (of the 192 patients, 88.7% of patients could be classified), 12.4% was due to prerenal causes, 76.8% was due to intrarenal causes, and 10.8% was due to postrenal causes. The diagnoses of the patients regarding the etiological factor are detailed in Table 1. Most of the patients were oliguric (77.1% versus 22.9%) at presentation.

Table 1 Classification of the intrarenal ARF patients according to the presumptive etiology

Diagnosis^*	%
Acute tubular necrosis (ATN)	44
Acute glomerulonephritis (AGN)	25
Transplant nephropathy^†	9
Acute uric acid nephropathy	6
Myeloma-lymphoma	5
Contrast nephropathy	5
Pregnancy-related nephropathies	3
Other	3
Total	100

^*Nephrotoxic agent related ARF was categorized as ATN; rapidly progressive glomerulonephritis was categorized as AGN.

^†Total of 12 patients: 11 chronic allograft nephropathy, 1 acute rejection.

With the exclusion of hypertension, 72.4% of the patients had at least one systemic comorbidity, including diabetes mellitus, ischemic heart disease, congestive heart failure, cirrhosis, renovascular disease, systemic infections, gastrointestinal hemorrhage, eclampsia, vasculitis, collagen tissue diseases, rhabdomyolysis, benign prostatic hypertrophy, and malignancies. In all, 46.8% of the patients had only one, 18.2% had two, and 7.4% had three or more systemic comorbidities.

Thirty patients (17.9%) were diabetic and 29 patients (17.3%) had malignancy, but 25 patients (13.0%) had no data related to diabetes or malignancy.

Thirty-eight patients (19.7%) were noted to have received nephrotoxic agents before the development of ARF (angiotensin-converting enzyme inhibitor or angiotensin-II receptor blocker, 26.3%; non-steroid anti-inflammatory drugs, 18.4%; radio contrast agent, 28.9%; aminoglycoside, 18.4%; other, 8.0%).

Mean catheter number was 1.22 per patient; 81.6 % of the patients were treated by means of only one catheter, while 14.0% and 4.5% needed two and three catheters, respectively. The mean number of hemodialysis sessions was 7.8 ± 8.0 per patient. Mean dialysis session per patients was statistically lower in dead patients than living patients (5.17 ± 5.1, 8.28 ± 8.3, respectively; p = 0.003)

The leading indications for initializing hemodialysis were uremic symptoms or signs in 46.8%, hypervolemia in 27.4%, electrolyte imbalances in 20.2%, and other in 5.6% of the patients. Uremic symptoms were the most frequent indication of the institution of IHD in men (55.8%) while hypervolemia was more common in women (37.5%). This difference was statistically significant (p = 0.011). Moreover, there was no significant effect of the presence of comorbidities on the principal indication of IHD. Hypervolemia was a more frequent indication in non-oliguric patients compared to oliguric patients and also in women compared to men (p = 0.035 and 0.032, respectively).

Biochemical parameters prior to the institution of the IHD are given in Table 2. Predialytic serum creatinine levels were also significantly lower in females (6.74 ± 3.7) than males (8.47 ± 4.1) (p = 0.013).

Table 2 Several important biochemical parameters of the patients before the institution of the IHD

	Mean	Standard deviation	Minimum	Maximum
BUN (mg/dL)	108.9	54.4	26.0	278.0
Creatinine (mg/dL)	7.7	4.04	1.40	21.8
Sodium (mmol/L)	132.7	8.60	103.00	152.00
Potassium (mmol/L)	4.8	1.46	2.30	8.90
Calcium (mg/dL)	7.8	1.09	4.50	9.50
Phosphorus (mg/dL)	6.5	2.32	2.10	11.50
Total protein (g/dL)	5.8	0.97	3.60	8.00
Albumin (g/dL)	2,8	0.81	1.40	4.60

At the end of hemodialysis sessions; 75.5% of the patients recovered without any further renal replacement treatment, while 9.4% of patients were transferred to chronic renal replacement programs. The remaining 29 of the patients (15.1%) died during IHD period. Several factors were found to be associated with survival probability (see Table 3). The mortality was higher in females (25%) than males (12.6%), but this difference did not reach statistical significance (p = 0.09). Likewise, mortality was higher in the patients who had more comorbidities (p = 0.053), but this finding also did not reach statistical significance.

Table 3 Some important parameters related to survival of the patients

	Survived	Died	p value
Age (years)	55.1 ± 16.7	62.5 ± 17.8	0.034^*
Female/male ratio	0.71	1.41	0.091
Predialytic serum levels of…
BUN (mg/dL)	111.57 ± 56.45	95.55 ± 41.16	0.23
Creatinine (mg/dL)	8.09 ± 4.14	5.4 ± 2.3	0.003^*
Sodium (mmol/L)	132.20 ± 8.02	135.42 ± 11.32	0.20
Potassium (mmol/L)	4.80 ± 1.51	4.83 ± 1.27	0.95
Albumin (g/dL)	2.99 ± 0.77	2.08 ± 0.51	0.016^*
Calcium (mg/dL)	7.91 ± 1.14	7.48 ± 0.52	0.41
Phosphorus (mg/dL)	6.59 ± 2.28	6.55 ± 2.56	0.97
Total IHD sessions/patient	8.28 ± 8.38	5.17 ± 5.19	0.003^*

^*statistically significant.

Pre-dialytic serum creatinine levels were weakly and negatively correlated with age (r:-0.20, p = 0.028). Pre-dialytic serum creatinine demonstrated positive correlation with pre-dialytic BUN, phosphorus, and potassium (r = 0.367, p = 0001; r = 0.518, p = 0,001; r = 0.285, p = 0.007; respectively). Pre-dialytic serum creatinine levels were much higher in the patients who started to IHD due to uremic symptoms and signs than the other patients (p = 0.001).

In mortality analyses, there was a relation between the mortality and the leading indication of the IHD. Mortality was higher if the leading indication of IHD was biochemical disturbances (i.e., electrolyte imbalance, hypercalcemia, hyperuricemia; p = 0.013). Diabetes mellitus did not influence mortality. There was no difference between diabetic and non-diabetic patients with respect to age but mean serum albumin level of the diabetic patients was statistically higher than their non-diabetic counterparts (p = 0.003 ). Furthermore, there was no statistical difference with respect to the etiology of the ARF and primary indication for the institution of IHD between these groups. Presence of malignancy, post-operative ARF, type of ARF (prerenal/renal/postrenal), catheter type, use of nephrotoxic agents, and oliguric-type ARF were not associated with mortality during the period of IHD.

DISCUSSION

Acute renal failure that necessitates IHD is a challenging problem in nephrology practice. Mortality of ARF is still high (30–50%) despite the technical improvements in dialysis practice.[2–4] Data on patient's demographics, causes of ARF, vascular access, and prognosis of these patients are still scarce, and existing data on these patients are controversial. This might be due to discrepancies in hospital and patient's characteristics, medical experiences, and presence of comorbidities. On the other hand, it seems to very difficult to arrange strong, universal, and randomized studies for the prediction of mortality in ARF patients who necessitate IHD.

In our study group, mortality was 15.1%. It must be emphasized that our department is a reference center; for example, all patients were hospitalized, almost half (48.9%) were older than 60, 17.9% were diabetic, and 11% had a malignancy. On the other hand, exclusion of the patients who underwent slow-continuous dialysis therapies might have contributed to relative low mortality ratio.

As it is well known, the most important cause of ARF is prerenal azotemia, which generally improves with appropriate treatment.[5] In this study, to make a detailed analysis, ARF was categorized as prerenal if it was related to any situation of prolonged renal hypoperfusion state. In other words, the patients with extended prerenal azotemia were categorized to prerenal ARF group, instead of ATN. In a study published by Feest et al.,[6] 72% of the patients were older than 70, and post-renal etiologies were the most important cause of ARF. However, in our study, 75% of the patients were diagnosed as having intrarenal ARF very probably due to lower mean age (56.3+17.1).

In our study, pre-dialytic serum creatinine and albumin levels, total IHD session number per patients, and age were the parameters that were related to high mortality in statistical analysis (see Table 3). As expected, total duration of catheter usage was positively correlated with total dialysis sessions (r = 0.58, p < 0.001).

In the literature, there is inconsistency regarding the relationship of mortality and serum creatinine level in ARF patients who underwent IHD. Some studies suggest that high mortality rate is associated with high predialytic serum creatinine levels, while others deny the existence of such an association.[7],[8] Furthermore, some imply an association between high mortality rate and low predialytic creatinine levels,[9],[10] like the present study. Additionally, mortality was higher in aged females (>60 years) and tended to be higher with the existence of comorbidities (p = 0.09 and 0.053, respectively) in our study. Hypervolemia was significantly a more common indication of IHD in females than males. Thus, older women who have comorbidities may need dialysis earlier than the other patients, especially due to hypervolemia. The lower serum creatinine levels of females than males might have delayed the initiation of dialysis in females and caused the time for development of hypervolemia. However, in our study, there was a positive correlation of serum creatinine with BUN, serum potassium, and phosphorus (r = 0.367, p = 0001; r = 0.518, p = 0.001; r = 0.285, p = 0.007; respectively), suggesting that low levels of creatinine were not due to its lower production from the less muscle mass. It means that patients who had low serum creatinine had also low serum potassium and phosphorus levels, indicating that these patients needed early dialysis generally due to non-biochemical reasons. According to our study, the ARF patients who needed early dialysis before the development of biochemical indication had high mortality. It may be speculated that old, female patients who had comorbidities were clinically the most ill patients, could not have tolerated the continuation of IHD, and died early.

As is well known, malnutrition is one of the important complications of ARF[14]: it has been shown that survival is better in well-nourished patients who require IHD during ARF.[15] Low serum albumin levels might be due to associated malnutrition, inflammation, and comorbidities, all of which might have a negative impact on mortality. There are many studies that support the relationship between mortality and hypoalbuminemia.[7],[8],[11–13] Our study also confirms these reports: in our study, mortality was significantly higher in patients with low albumin levels. In future studies, it may be worth investigating the hemodynamic effects of intradialytic or interdialytic albumin infusions in a low-albumin level group.

Oliguria has been suggested to be a bad prognostic indicator,[1],[7],[17–20] though not all authors agree with this hypothesis.[16] Furthermore, Liangos et al.[21] demonstrated that a higher hospital mortality rate was associated with higher urinary volume.[21] Our study did not show any considerable effect of urine volume on patient outcome. It must be stressed that the time of the patient referral to a nephrologists might affect the mortality.[9] In addition, the timing of the initiation of dialysis may differ in the patients who have oliguria (by nephrologists). In our study, hypervolemia was significantly a more frequent indication of IHD in non-oliguric patients compared to oliguric patients (p = 0.035), and there was no correlation between predialytic serum creatinine and urinary output. In oliguric patients, the physicians might have initiated the dialysis therapy earlier than the non-oliguric patients. In non-oliguric patients, the physicians might insist on volume replacement therapies in order to force diuresis, which might end with more frequent initiation of IHD due to hypervolemia. This might have obscured the possible outcome differences.

In the earlier days of dialysis therapy, it has been stated that no significant improvement was yielded in patients with severe pre-existing diseases or extensive extra-renal lesions. Such patients usually had died from non-uremic complications before or despite the dialysis.[22] In a retrospective study of 58 patients who underwent IHD, it has been demonstrated that survivors were significantly younger and less frequently malnourished.[15] On the other hand, in several studies of ARF patients (some requiring IHD and some not), older age and comorbidities were associated with higher mortality.[20],[23],[24] Our study also supports these results, as stated above.

There was no mortality difference in diabetics compared to the non-diabetic group in the present study. There was no age difference between these two groups either, but mean serum albumin level of the diabetic patients was statistically greater than non-diabetic counterparts (p = 0.003). As stated above, low albumin level is a bad prognostic parameter of survival during IHD. The difference of albumin levels might have obscured a potential outcome disparity in our study. Also, diabetic patients with low serum albumin levels might have a worse clinical condition than non-diabetic and needed other dialytic modalities not included in our study group. Thus, there is an evident need of studies targeted to find out outcomes in different subgroups of diabetic IHD patients.

There are many limitations of our study. First, it was a retrospective study, and some data were missing. In many cases, the etiology of the renal failure was predicted without any biopsy. In particular, causes of death, parameters related to effect of the membranes on outcomes, catheter-related problems, infections, use of diuretics, details of first dialyses, intra-dialytic hemodynamic parameters, and complications could not have been assessed adequately.

CONCLUSION

It is difficult to predict outcomes of ARF patients who necessitate IHD. However, predialytic serum creatinine and albumin levels may be very important predictors of mortality. Patients in high-risk groups (older age, female sex, and low predialytic creatinine and albumin levels) should be considered for treatment with slow continuous renal replacement methods.