Acute kidney injury (AKI) is easily defined as a syndrome characterized by a sudden decrease in GFR accompanied by azotemia [4]. However; the reported incidence of AKI varies depending on a number of independent variables. For example, was the patient population surveyed derived from a community wide database or was it restricted to hospitalized patients? What definition was adopted to designate acute kidney injury (AKI)? The lack of a universally agreed upon definition of acute kidney injury (AKI), makes it difficult to compare clinical reports as to the incidence, severity and outcome. Recently, the Acute Dialysis Quality Initiative [5] has attempted to address this issue involving both nephrologists and critical care physicians in the discussion. Success of this project is critical for it will allow the sharing of information regarding interventions which, in turn, will improve the dismal outcomes that currently exist for patients with acute kidney injury. This dismal outcome is especially true if the patient suffers from the constellation of multiple organ failure which is becoming common place in ICUs. Encouragement comes from the success of the KDOQI classification of Chronic Kidney Disease (CKD) [6] that is being adopted world wide and allows consistent stratification based on glomerular filtration rate (GFR) [7].

Additional variables exist, for example, in-hospital surveys enroll both post-surgical and medical patients, and it is important to isolate the contribution from ICU patients with multi-organ failure! With what precision was the AKI diagnoses established? Were multiple centers involved in providing the information? These are the often unanswered questions that complicate meaningful estimates of the incidence of AKI. In addition, as detailed by Turney et al. [8] and Nash and co-workers [3], significant changes have occurred in both the age of the AKI patients and also the etiologies, with older and sicker patients being admitted for treatment.


The incidence of in-hospital AKI attributed to drug nephrotoxicity is estimated at between 18 and 40% of cases [9-16], while earlier reports, derived from community-based statistics set the incidence from 0 to 7% of cases [10, 11]; however, more recent series report incidence from 17 to 29% [17,18] which is still lower than the 37% reported by Baraldi et al. [14]. In both series that included hospital acquired AKI [17,18], the contribution of nephrotoxic drugs was slightly higher, 22 and 35%. The recently reported increase in the contribution of nephrotoxic drugs to community acquired AKI is particularly important since the total number of all cause cases of AKI, as derived from community based studies, is 2 to 3.5 times greater that reported from in-hospital statistics [17,19]. This suggested that patients who are hospitalized are either exposed to more nephrotoxic agents and/or they are more vulnerable to the drugs inducing an adverse renal effect. Community acquired AKI is associated with a significantly reduced mortality as compared to hospital acquired AKI [21], (41% vs. 59%, p<0.001 [18])and for elderly patients this equated to a risk of mortality that was 2.2 times greater for the hospital acquired AKI group. Irrespective of which aspect of the drug interaction is more important, it has been observed that hospital-acquired AKI is usually associated with one of three renal insults, either a pre-renal event, exposure to nephrotoxins, or sepsis [1], and that nephrotoxins, alone or in combination, contribute to at least 25% of all cases of hospital acquired AKI [2] and have an inhospital mortality rate similar to ATN [20]

A one-year survey of 2,175 cases of AKI, 398 (18.3%) were considered to be drug-induced [11]. Antibiotics were the most frequently cited drug followed by analgesics, NSAIDs and contrast media and this relationship persists [3]. More than half of the patients had non-oliguric AKI. The mortality rate of 12.6% is much lower than for patients who develop AKI following surgery or trauma [22]. At 6-month post-AKI, 47.7% were fully recovered, 15.3% had regained previous renal function, and 23.1% had some degree of residual renal impairment. Chronic hemodialysis was required in only 2 patients (0.5%). This is a better outcome than reported for a group of patients post-AKI due to multiple causes [20,23]. Of the 39% who survived AKI, 41% had residual renal insufficiency and 10% required chronic dialysis. Residual renal impairment was more frequent in both older and oliguric patients, in those with previous chronic renal insufficiency, those who received antibiotics, and those whose duration of AKI was prolonged. The percentage of residual renal impairment is higher than that reported in the series of Davidman et al. [24] or Pru et al. [25], but is in accordance with that found 5 years later in the same country [26] and is supported by an earlier report from the European Dialysis and Transplant Association [27].

Table 1 summarizes the incidence of drug-induced AKI reported for the last two decades. As can be seen, the incidence of AKI due to contrast media and antibiotics is variable depending on the population included. If the population is drawn from hospital acquired AKI, then contrast media and antibiotics are prominent; however, if the population is mixed hospital and community acquired, as in the case of Sesso et al [18], the NSAID's are major contributors. Since the 1990's two

Table 1. Incidence of drug-induced AKI reported for the last two decades.

Author Year N % acute renal failure due to

_Antibiotic Contrast Analgesic NSAIDs ACEI Total

Kleinknecht et al [11] 1986 2175 6% 2% 4% 3% 0.5% 18%

new categories of offending agents have appeared, e.g. NSAIDs and ACE inhibitors. This trend has been confirmed in the survey conducted by Ronco et al. [28]. Recently, Nash et al have repeated a survey of hospital acquired renal insufficiency 17 years after the first published report from this group of investigators [3]. Due to the more wide-spread use of antibiotics and contrast media, while the percentage contribution has fallen, the total number has increased. The class of antibiotic has changed dramatically. In the original report aminoglycosides nephrotoxicity accounted for nearly 80% of the drug induced renal insufficiency, while in the more recent analysis aminoglycoside accounted for less than 30% of the cases with significant contributions from amphotericin and pentamidine. In the series reported by Sesso et al [18], antibiotics( aminoglycosides, vancomycin, cephalosporines, qui-nolones), and NSAID's (diclofenac, ibuprofen, ketopro-fen and indomethacin) were the two dominate classes of drugs leading to AKI for all patients irrespective of whether community or hospital acquired, while in the hospital acquired AKI group, 6.5% of the patients had contrast nephropathy [18].

The estimated incidence of 18-33% drug-induced AKI in hospitalized patients contrasts with the extremely low incidence of drug-induced renal disease in outpatients as reported by Beard et al. [30], i.e., 1:300,000 person/year. This low incidence is in part due to the author's exclusion of chronic renal disease. On the other hand, acute iatrogenic renal disease developed in 1% of all patients admitted to a Canadian hospital and in as many as 5.6% of those admitted directly to the nephrology unit of the same institution [25]; the AKI was due to multiple etiologies in 50% of these patients. Outcome

Despite improved dialysis techniques and more aggressive supportive treatment, conventional wisdom is that mortality from AKI has not improved in the last decade. Support for this belief comes from a systematic review of mortality rates in nearly 16,000 patients with AKI, reported in 80 clinical studies, which concluded that mortality rates were unchanged over the 3+ decades covered by the review [31] However, 2 recent retrospective studies using nation wide databases have reported on the secular trends regarding both the incidence and mortality rates for AKI over the 10 years from 1992 to 2001 [32] or the 15 years between 1988 and 2002 [33] Both studies are unique for they are the first to use multi-year data to determine trends both in incidence and mortality for AKI. Xue et al [32] used a Medicare 5% sample beneficiary standard analytical file and collected data both for community-acquired AKI, eg AKI as the primary diagnosis at time of admission, and hospital-acquired AKI, eg AKI as a secondary diagnosis occurring during hospitalization. During the 10year period 2.4% of hospitalization involved AKI, 24% were community-acquired and 75% were hospital-acquired which is a reversal of the findings reported in 1997 [19]. Importantly, the over all incidence of AKI more than doubled over the 10 year period, with an annual increase of 11% (p<0.0001). Paralleling this increase in incidence was an increase in sepsis, ICU stay, and multiple organ failure [32]. Patients with AKI were older, more often male, and African -American. In hospital death rates declined as did death rates within 90 days of hospital discharge. However, using logistic regression analysis in cases of multi-organ failure, AKI was still a significant cause of death. While this represents a retrospective study which is complicated by coding changes which occurred during the study interval, there is reason to believe that the frequency of AKI is underreported in Medicare claims. The recent report by Ali et al [32a] would support this concept of underreporting. They reported, based on a comprehensive population-based study, an AKI prevalence of 1811 per million population. This estimate exceeds that of Waikar et al [33] by six fold. The second report by Waikar et al [33] uses the Nationwide Inpatient Sample which is the largest all-payer administrative database of hospitalizations in the United States. ICD-9 codes were used to identify AKI subjects from 1988 thru 2002. The incidence of annual discharges with AKI rose from 0.4% in 1988 to 2.1% in 2002, while in-hospital mortality declined from 40.4% in 1988 to 20.3% in 2002 (p<0.001). This decline was observed across groups stratified for age, gender, race/ethnicity, co morbidity index and a broad array of concomitant conditions. Interestingly, in hospital mortality was lower from AKI with CKD than AKI without CKD although rates declined in both groups over time. A significant decline in length of stay was also documented, decreasing from 10.3 days to 7.0 days. So both studies confirm an increased incidence in AKI with a predominance of in hospital acquired AKI , but a significant decline in annual mortality indicating that newer dialysis techniques coupled with aggressive supportive treatment are improving overall survival of AKI patients.

One of the significant changes in management strategies for patients with AKI is the introduction of severity score systems to address issues of treatment effectiveness, quality of care, and allocation of limited resources. Two classification used in ICU patients that have gained wide acceptance are APACHE-II [ 34] and the system introduced by Liano [35]. However, neither system is useful for non-dialyzed AKI patients. The Stuivenberg Hospital Acute Renal Failure (SHARF) score, has proven to be a predictive model for in-hos-pital mortality in cases of AKI [36] More recently, this same scoring system has be evaluated as a predictor of mortality over the first post-AKI year [37]. The 11% additional mortality 1 year after the episode of AKI occurred in patients who had significantly higher

SHAKI scores as compared to either APACHE II or Liano scores. At the time of hospital discharge 32% of patients were CKD I-II, 58% CKD III-IV, and 10% CKD V. Interestingly, the degree of renal impairment had an inverse correlation with the 1 year death rate being 33%, 18% and 14%, respectively.

Although the search for an ideal definition of AKI continues, Chertow et al [38], recently evaluated the effect of AKI on mortality, hospital stay and cost using changes in serum creatinine as the marker of renal injury. The surprising result was that changes in serum creatinine of > 0.3 mg/dl occurred in 31% of the study population and was associated with a 4.1 multivariable odds ratio of mortality (3,1-5,5), and the OD rose to 6.5 at serum creatinines > 0.5 mg/dl, 9.7 at serum creatinines > 1.0 mg/dl, and 16.4 at serum creatinines > 2.0mg/dl. Patients with de novo AKI and serum creatinine >0.5 mg/dl had a significantly higher risk of death than patients with AKI superimposed in CKD. Even small increases in serum creatinine (0.3-0.4 mg/ dl) had a multivariable OR of 1.7 [1.2-2.6]. In addition to the increased mortality risk, both hospital length of stay and cost had a direct linear correlation with the increase in serum creatinine. This association between serum creatinine and hospitalization outcomes occurred over a wide spectrum of clinical conditions. Similar conclusion regarding utilization of hospital resources for individuals with acute kidney injury was reported by Fischer et al [39].

Mechanisms of drug induced acute kidney injury

Because of the rich blood flow to the kidney (25% of the resting cardiac output), plus the enormous oxygen supply required to support active ion and solute transport, the kidneys are vulnerable to any change in blood flow and/or oxygen deprivation. In particular, acute tubular necrosis involving thick ascending limb (TAL) is a prominent manifestation of a sudden reduction in renal blood flow with accompanying hypoxia. This anatomic site is especially vulnerable to oxygen deprivation due to the marginal oxygen balance that results from high oxygen consumption related to the active NaCl reabsorption and the limited blood supply due to the anatomic structure of the vasa recti [40]. A second important contributor to AKI occurs when the tubulo-glomerular feedback system fails. Tubulo-

glomerular feedback is an auto regulatory mechanism that reduces glomerular filtration rate (GFR) and decreases the sodium load that is delivered to the TAL. The net result of the diminished sodium load is to minimize the oxygen required for active NaCl reabsorption [40,41].

When considering the mechanism by which drug causes nephrotoxicity, two components of renal function are decisive. The first are the renal transport processes which are critical to recovering essential minerals and nutrients from the glomerular filtrate and the second are the renal enzyme systems which are essential to both detoxification of xenobiotics and maintaining the body's acid/base homeostasis [41,42]. Recently, these two components have been merged into a single scheme of vectorial drug transport in both liver and kidney cells [43]. As can be appreciated, in addition to the traditional phase 1 and 2 metabolism steps, three new phases have been added (figure 1). The new phases are: phase 0 (entry via transporter), transcellular translocation (phase 3) and phase 4 (cell elimination via transporter). Solute carriers (SLC) are responsible for cellular entry of the xenobiotic, while ATP-dependent carriers (ABC) are responsible for cellular elimination or recycling. This concept allows for both intracellular metabolism and recycling or transcellular transport to eliminate xenobiotics.

The principle renal transport systems, which con-

Transport Metabolism Transport Phase 0 d!^ Phase 1 Phase 2 Phase 4

xenobiotic I

Ir" Transport œ-|«C^>Phase3 ^

0 0

Post a comment