Biopsy-Proven Acute Tubular Necrosis due to Vancomycin Toxicity

Abstract

Vancomycin (VAN) has been associated with acute kidney injury (AKI) since it has been put into clinical use in the 1950's. Early reports of AKI were likely linked to the impurities of the VAN preparation. With the advent of the more purified forms of VAN, the incidence of AKI related to VAN were limited to acute interstitial nephritis (AIN) or as a potentiating agent to other nephrotoxins such as Aminoglycosides. VAN as the sole etiologic factor for nephrotoxic acute tubular necrosis (ATN) has not been described. Here, we report a case of biopsy-proven ATN resulting from VAN.

1. Case

R. H. is a 23-year-old male with a history of childhood acute lymphocytic leukemia in remission, neuroectodermal tumor status postresection, and gamma knife therapy, who presented to our Emergency Department (ED) with complaints of fever and chills. He also noticed yellowish drainage around the site of a peripherally inserted central catheter (PICC). In the ED, patient was febrile to 101.7° Fahrenheit with a blood pressure of 113/83 mm Hg and pulse of 134 bpm. The PICC line was removed, and antibiotic therapy initiated with piperacillin-tazobactam and VAN. RH received 1 gm IV of VAN in the ED, and upon admission to the floor, received another 2 gm IV. Piperacillin-tazobactam was discontinued. Eight hours later, another 2 grams of VAN were given, with cumulative dose of 5 gm in 24 hrs (50 mg/kg in 24 hours). Patient's admission creatinine was 0.97 mg/dL. The patient was 103 kg and 72′′ tall. Attempting to use a loading dose of 15 mg/kg based on total body weight [1], the patient was initiated on 2000 mg of VAN per hospital protocol of using VAN doses in 1000 mg aliquots. Standard pharmacokinetic equations with simplified one-compartment models employing log-linear drug removal were used for estimations [2]. With a volume of distribution estimated at approximately 0.6 L/kg of actual body weight and a VAN clearance estimated to be at least 120 mL/min [3], the patient was expected to achieve an initial VAN trough of approximately 8 mg/L with steady state troughs above 10 mg/L with a dosing scheme of 2000 mg given every 12 hours. As pharmacokinetic estimates in patients above their average body weight exhibit substantial variability, the patient was initiated on the aforementioned dose with a measured VAN trough planned antecedent to the fourth dose (to capture steady-state concentrations). Goal troughs and monitoring schemes were performed according to national VAN guidelines [4].

The following day, the serum creatinine was 3.62 mg/dL. He became oligo-anuric with a urine output of 50 cc in 24 hours. Urinalysis revealed a bland urine sediment and a urine sodium of 75 meq/L and a fractional excretion of sodium of 2.77%. He had not been exposed to IV contrast, received an aminoglycoside or exposed to any other potential nephrotoxins.

Blood cultures were negative, and the PICC line tip cultures were positive for Serratia marcescens. As the patient experienced acute nephrotoxicity, VAN was ceased after the third dose and ciprofloxacin initiated. The creatinine continued to rise (Table 1). A renal ultrasound ruled out obstruction. On day 3, a renal biopsy was performed. The main finding was ATN (Figure 1). Many tubules showed moderate degree of vacuolization of their cytoplasm (green arrows). Some of tubules contained hyaline or epithelial casts in their lumina (black arrows). The VAN serum concentration done on day 4 was 64.7 mg/L, and hemodialysis was initiated for volume overload. By day 5, his urine output increased to 1 L/day. On day 10, he was taken off hemodialysis. His creatinine improved to 1.2 mg/dL over the next few weeks.

Table 1.

Laboratory and clinical parameters.

Day	Scr mg/dl	UO 24 hrs	VAN dose	VAN S. Conc.	HD
0	.97	NR	5 gm	ND	No
1	3.62–4.26	50 cc		ND	No
2	6.25	<50 cc		ND	No
3	8.41	<50 cc		ND	No
4	9.96	<50 cc		64.7	Yes
5	9.36	1000 cc			Yes
9	9.21	2000 cc		ND	Yes
10	5.88	2500 cc		ND	No
30	1.24	NR		ND	No

Scr—serum creatinine; UO—urine output; VAN—Vancomycin; S. Conc—Serum concentration; HD—hemodialysis; ND—not done; NR—not recorded.

Figure 1.

Representative photographs of the renal biopsy showing tubular damage secondary to drug toxicity. In both panels (a) and (b) a number of tubules show moderate degree of acute tubular necrosis (asterisks). Some of tubules contain hyaline or epithelial casts in their lumina (green arrows), while several tubules show vacuolization of their cytoplasm (green arrowheads). One of the glomerular afferent arteriole shows swollen endothelia and occlusive change (black arrow).

2. Discussion

VAN, a parenteral glycopeptide antibiotic, has been used clinically since 1956. The original preparation contained a number of impurities and was brown in color, hence, the nickname “Mississippi Mud”. It is these impurities that were thought to be responsible for certain toxicities such as anaphylaxis, toxic epidermal necrolysis, erythema multiforme, ototoxicity, and nephrotoxicity [5, 6]. Between 1956 and 1986, 57 cases of VAN-associated nephrotoxicity were described. More than 50% of the cases were identified within the first six years of VAN use when the product was relatively impure [7]. As manufacturing processes improved and VAN was purified, nephrotoxicity became less common [8], and contemporary preparations seemed to nearly eliminate the original adverse events.

In 1983, Farber andMoelleringreported the incidence of nephrotoxicity with VAN to be 5% [5]. This, however, was a retrospective study, and it failed to evaluate whether the AKI was due to AIN or ATN. VAN in association with an aminoglycoside is known to be associated with an increased incidence of nephrotoxicity presumed to be ATN [9, 10].

Other risk factors associated with nephrotoxicity include prolonged therapy for >21 days, [9] higher APACHE scores, [11] loop diuretics, and steady-state VAN concentration > or = 28 mg/L [12]. Recent studies have shown that doses greater than or equal to 4 grams per day [13] and trough concentrations greater than 15 mg/L are associated with nephrotoxicity [11, 14]. More specifically, when trough concentrations are analyzed as a linear variable with classification and regression tree modeling, a threshold of 9.9 mg/L has been identified [15]. These studies, while instructive for the association of VAN and toxicity, are limited by a failure to control for weight-adjusted dose, and the retrospective nature of the studies do not determine causality [11, 13–15]. Instead, elevated VAN concentrations could be merely an intermediate variable in the pathway to the ultimate causal event of renal failure by other mechanisms.

As VAN safety profile improved and with the emerging prevalence of methicillin-resistant Staphylococcus aureus (MRSA), VAN became one of the most frequently prescribed intravenous anti-infective for treatment of these infections [16]. With increasing the minimum inhibitory concentrations for VAN and proper testing methodologies for completing minimum inhibitory concentrations under continual debate [17–20], the dire outcomes associated with MRSA infections have lead many to seek novel treatment strategies including increasing VAN doses and exposure. Specifically, various national guidelines for serious infections have recommended that VAN troughs be maintained between 15–20 mg/L to ensure an AUC/MIC of ~400 [21–23]. Such a position has been substantiated on the basis of appropriate pharmacodynamic targets, limited clinical data, and the supposition that such a strategy is safe [21, 22, 24]. However, no study to date has prospectively evaluated the impact of VAN exposure on renal endpoints and that elevated doses (such as a 5 gram dose given in 24 hours in our patient) could predispose patients to kidney injury.

The potential mechanism of action of VAN-associated nephropathy has been studied in both humans and animals. The energy-dependent transport mechanisms found in the proximal tubular epithelium render the kidneys highly susceptible to toxicant-induced renal injury. VAN exposure in renal proximal tubule epithelial cells results in increased cell proliferation as evidenced by increased number of cells, total protein, and DNA synthesis. VAN enhances cellular ATP concentration and stimulates oxygen consumption, supporting its role as a stimulant of oxidative phosphorylation [25]. The beneficial effect of some antioxidants like DL-α lipoic acid, Melatonin, Ginkgo biloba and milrinone have been shown to reduce the renal damage, suggesting the involvement of free radicals in renal damage [26].

While it is unclear if VAN-associated acute tubular necrosis is preventable, these data argue for prompt serum VAN monitoring given the rapid functional decline of patient nephrologic function even before the patient was scheduled to receive the 4th dose of therapeutically dosed VAN. Additionally, these data may support previously published literature suggesting that VAN doses >4 g per day, and patient body weights in excess of 101.4 kg may predispose to nephrotoxicity, [27] in this case, acute tubular necrosis.

3. Conclusion

The most common VAN-associated nephrotoxicity is AIN [28–30]. Although an association with ATN has been described, to our knowledge, VAN as the sole etiologic cause of ATN has never been reported in adult patients [31, 32]. Our findings support these retrospective associations and provide evidence for contemporary VAN-associated ATN. Further studies are needed to determine the safety profile of targeting higher VAN trough levels for serious infectious in light of the potential for AKI.