**Results for objective 1: Evidence for the existence of Van-AKI**

Towards objective 1, the findings of our current studies reported here belong to two sections. In the first section (A), we have performed and will present a comprehensive, systematic, and an up-to-date literature review to draw on all described indirect and circumstantial evidence cited to support the concept and the existence of Van-AKI. In the second section (B), we shall describe the 6 patients we personally saw and helped manage who were consulted for acute renal failure and in whom we found compelling evidence for the diagnosis of Van-AKI. We will detail their presentation and the clinical course of their ARF. We shall provide serial laboratory findings to document the causality of vancomycin, including their recovery course following the discontinuation of the offending agent.

### **(A) Evidence for Van-AKI based on literature review**

The literature has provided two independent sources of indirect evidence in support of the issue of Van-AKI. The first body of evidence (1) (Table 1) comes from several epidemiologic

Table 5. For Intra-renal insults (N = 78 or 77% of the 101 cases of Acute Kidney Injury)

from AKI (Fig 7).

**4. Findings** 

**minimize vancomycin nephrotoxicity** 

the days of nadir vancomycin levels, and finally to the days of nadir serum creatinine at maximal recovery 60 days after the initiation of vancomycin. We also plotted serial vancomycin levels against the renal functional profile to evaluate and define the temporal relationships between drug levels and kidney function during the evolution of and recovery

**Objective 3: To generate and provide simple practical guidelines and recommendations to** 

Inferences from the analyses performed for Objective 2 will provide the basis for us to formulate the proposed guidelines designed to prevent and /or ameliorate the emergence of Van-AKI. These will be elaborated as a narrative in the Results section and presented in a

Towards objective 1, the findings of our current studies reported here belong to two sections. In the first section (A), we have performed and will present a comprehensive, systematic, and an up-to-date literature review to draw on all described indirect and circumstantial evidence cited to support the concept and the existence of Van-AKI. In the second section (B), we shall describe the 6 patients we personally saw and helped manage who were consulted for acute renal failure and in whom we found compelling evidence for the diagnosis of Van-AKI. We will detail their presentation and the clinical course of their ARF. We shall provide serial laboratory findings to document the causality of vancomycin,

including their recovery course following the discontinuation of the offending agent.

The literature has provided two independent sources of indirect evidence in support of the issue of Van-AKI. The first body of evidence (1) (Table 1) comes from several epidemiologic

tabulated format (Table 7) in the final Conclusion and Recommendations.

**Results for objective 1: Evidence for the existence of Van-AKI** 

**(A) Evidence for Van-AKI based on literature review** 

surveys and drug toxicity monitoring studies performed in sizable patient cohorts taking vancomycin, which suggested an association between the drug and acute elevation of serum creatinine (Farber et al. (1983), Sorrell et al. (1985), Bailie et al. (1988), Rybak et al. (1990), Goetz et al. (1993), Vance-Bryan et al. (1994), Hidayat et al. (2006), Lodise et al, (2008), Pritchard et al. (2008), Ingram et al. (2008), Pertel et al. (2009), Kalil et al. (2010), Rodriguez Colomo et al. (2010). The second body of evidence (2) (Table 2) is based on the growing number of case reports describing the association between acute nephrotoxicity and vancomycin (Dutton & Elmes et al.(1959), Dangerfield et al. (1960), Odio et al. (1984), Frimat et al. (1995), Sokol et al. (2004), Barraclough et al. (2007), Ladino et al. (2008), Psevdos et al. (2009)).

## **(A) (1): Epidemiologic and drug toxicity monitoring studies**

These studies have collectively provided four lines of evidence implicating vancomycin in the pathophysiology of AKI: (a) Correlation between acute rise in serum creatinine and high serum vancomycin levels (Rybak et al 1990, Hidayat et al 2006, Ingram et al 2008, Lodise et al 2008, Pritchard et al 2008); (b) Increased incidence of acute renal failure (or potentiation of nephrotoxicity) when vancomycin was also administered concurrent with aminoglycosides (Farber et al 1983, Rybak et al 1990, Goetz et al 1993); (c) Increased incidence of AKI with prolonged duration of vancomycin therapy (Hidayat et al, 2006, Pitchard et al 2008); (d) Increased incidence of AKI with vancomycin compared to linezolid in comparable cohorts with similar patient characteristics (Lodise et al 2008, Colomo et al, 2010). The studies providing these four lines of evidence will be presented in the same order.

a. ARF was more often associated with a higher steady-state or trough serum vancomycin levels and linked to higher daily doses.

Rybak et al reported in 1990 that higher serum trough vancomycin levels were associated with the development of elevated serum creatinine (Rybak et al., 1990). In the ensuing two decades, this observation was not only confirmed but also extended by the studies of Hidayat et al (2006), Ingram et al (2008), and Lodise et al (2008).

Since the new millennium, the widespread use of vancomycin has led to the expected emergence of strains of methicillin resistant staphylococcus aureus (MRSA) that have only intermediate sensitivity to vancomycin, based on higher than the classical minimum inhibitory concentration (MIC) of 1 mg/L. Accordingly, the Infectious Disease (ID) guidelines have recommended higher trough concentrations like between 15-20 mg/L (Rybak et al., 2009) in order to maximize the chances of eradicating such infections. One unintended consequence was the apparent rise in the incidence of AKI by following such guidelines too rigidly but without closer vigilance of the level of renal function.

Thus, in a prospective study on the efficacy and toxicity of vancomycin during treatment of these relatively resistant MRSA strains by targeting and achieving the higher trough level of 15-20 mg/L, Hidayat et al. (2006) not only noted a higher mortality rate and a poorer end-oftreatment response**,** but also the development of nephrotoxicity in the subset of patients with demonstrably higher trough levels.

In 2008, Ingram et al. performed a retrospective cohort study of 102 adults to identify risk factors for nephrotoxicity during continuous outpatient vancomycin administration between 2004 and 2007. The incidence of nephrotoxicity, defined as ≥ 50% increase in baseline serum creatinine, was about 15.7%. Based on their analyses, a steady-state serum vancomycin concentration of ≥ 28 mg/L was thought to be an independent risk factor for developing nephrotoxicity.

Since the published new ID guidelines to keep trough level between 15-20 mg/L for resistant strains of MRSA, a good number of clinicians have increased the dose to >4 g/day

Vancomycin-Induced Nephrotoxicity 191

the prospective studies by Hidayat et al on targeting higher trough vancomycin levels for MRSA strains with high MIC, they not only confirmed the previous association between high trough levels and nephrotoxicity, but also a link between prolonged treatment

The retrospective review by Pritchard et al in 2008 also identified the duration of vancomycin administration as an independent risk factor. In their studies of ~ 3,000 courses of vancomycin given between 2003 and 2007, they observed that therapy over 7 days was associated with AKI. They also suggested baseline serum creatinine > 1.7 mg/dl as another independent risk factor. It is however unclear if this association merely reflects a heightened sensitivity of the clinicians to AKI, an enhanced detection of renal failure with an already elevated baseline creatinine, and/or intrinsically greater susceptibility of chronically

d. Increased incidence of AKI or slower recovery from pre-existing ARF if treated with

In the retrospective review by Lodise et al (2008) on the impact of ≥ 4 g/day of vancomycin, they also found a significantly lower incidence of renal failure in patients on linezolid (6.7% vs. 34.6% in patients on > 4 g /day or 10.9 % in those on < 4 g /day) (*P=*0.001). In the treatment of nosocomial pneumonia, Kalil et al (2010) performed a meta-analysis to test the hypothesis of the superiority of linezolid over vancomycin. But they found no significant difference in either vancomycin efficacy or risks of renal dysfunctions, although the study

Rodriguez Colomo et al. (2010) conducted a retrospective, multicenter observational study in patients in intensive care unit with pre-existing renal failure. They found that those patients treated with linezolid had a better renal recovery than those treated with vancomycin, implying either continued nephrotoxic susceptibility or superimposed injury

In brief, despite the mounting body of indirect evidence summarized above suggesting a role of vancomycin in AKI, firm and unambiguous proofs for a cause-and-effect linkage remain elusive. Virtually all of the studies cited and reviewed above did not offer sufficient details on those individual patients with presumed Van-AKI to allow independent and objective confirmation of a causal and unequivocal relationship. Inherent in the nature of these large-cohort surveys and drug toxicity monitoring studies, despite showing statistical significance among different cohorts, none of the other known and potential etiologic factors for the ARF could be readily evaluated in individual affected patients, let alone vigorously excluded. For instance, the evolution of their serial renal function and the subsequent clinical course after cessation of vancomycin were either not provided or extractable from those individuals afflicted with AKI. A larger prospective study with sufficient clinical details is therefore needed to objectively eliminate all other confounding variables and to

Between 1956 and 1986, 57 cases of ARF were described in the course of vancomycin administration and had been attributed to vancomycin. However, over a half of them were reported within the first 6 years of vancomycin use, when impurities were considered to be

Although they spanned out over 5 decades in the medical literature, there have been but fewer than two dozen well documented cases of ARF which can be confidently and

was not powered to compare the nephrotoxic potential between the two drugs.

and AKI.

diseased kidneys to new and acute insults.

vancomycin versus linezolid.

with vancomycin in these cohorts.

prove the implied cause-and-effect relationship.

the most likely culprit (Bailie & Neal 1988)

**(A) (2): Evidence for Van-AKI based on published case reports** 

to achieve the recommended required trough levels. This approach has afforded an opportunity for Lodise et al. (2008) to conduct a retrospective cohort study to describe the impact of ≥ 4 g/day of vancomycin on renal function. They found a 3-fold greater incidence of nephrotoxicity in patients receiving >4 g vancomycin/day (34.6 %) versus those receiving <4 g /day (10.9%). In the same review, these investigators also found a much lower incidence of renal failure in similar patients who received only linezolid ( 6.7%) as opposed to either > 4 or < 4 g /day of vancomycin (*P=*0.001).

Pritchard et al. (2008) conducted a retrospective analysis of ~ 3,000 courses of vancomycin given between 2003 and 2007. The aim of their study was to determine the relationship between vancomycin trough concentrations and nephrotoxicity. They noted that trough levels >14 mg/L was an independent risk factor for renal injury among others to be elaborated below.

In contrast, when serum trough vancomycin levels were prospectively limited to the lower range of 5 to 10 mg/L and if peak levels were kept ~ 28 mg/L, in 1985, Sorrel et al found no AKI with vancomycin (when used alone in two patients) and < 8% incidence of AKI even if combined with an aminoglycoside among the 54 patients studied. Taken together, these findings indicate that vancomycin must be considered nephrotoxic, especially at high serum levels, although it was found to be relatively safe at low trough or steady-state levels.

Parenthetically, Bailie et al. (1988) had reviewed the utility of peak serum levels as an indicator of vancomycin induced nephrotoxicity and ototoxicity. They determined that peak vancomycin concentration per se may be relatively minor in producing and predicting nephrotoxicity as opposed to the total area under the serum concentration-time curve (AUC).

b. Increased incidence of AKI when vancomycin was concurrently administered with an aminoglycoside.

At least three to four studies have found the synergistic nephrotoxic potential between vancomycin and aminoglycosides. In a retrospective study, Farber et al in 1983 found that more patients who received both vancomycin and gentamicin (12 of 34) had suffered from nephrotoxicity as compared to those getting vancomycin alone (3 of 60) with a p value of <0.001. In the studies by Sorrel et al (1985), AKI was found in 4 of 54 vancomycin-treated patients, but all 4 had also received aminoglycosides. In contrast, no AKI was found in the two on vancomycin alone.

These findings, however, were not uniformly observed (Downs et al. in 1989, Cimino et al. in 1987, Mellor et al. in 1985), perhaps due to intrinsic differences in their patient characteristics, definitions of acute renal failure, and the divergence in their study methods. In contrast, in 1990 Rybak et al. confirmed that the combination of aminoglycosides and vancomycin was more nephrotoxic than either drug alone.

The prospective studies by Goetz & Sayer published in 1993 provided corroboration for the additive nephrotoxic potential between vancomycin and aminoglycosides. The incidence of nephrotoxicity was 19% in patients receiving vancomycin alone, 12% in patients receiving an aminoglycoside alone and 24% in patients receiving combined vancomycin and an aminoglycoside.

c. Increased incidence of AKI with prolonged vancomycin administration.

Several studies have led to the conclusion that prolonged therapy with vancomycin was a risk factor for AKI (Goetz & Sayer 1993, Hidayat et al 2006, Pritchard et al 2008). Besides showing the synergism between aminoglycosides and vancomycin in causing AKI, Goetz & Sayer observed that a duration of >21 days posed greater risk for renal toxicity. In

to achieve the recommended required trough levels. This approach has afforded an opportunity for Lodise et al. (2008) to conduct a retrospective cohort study to describe the impact of ≥ 4 g/day of vancomycin on renal function. They found a 3-fold greater incidence of nephrotoxicity in patients receiving >4 g vancomycin/day (34.6 %) versus those receiving <4 g /day (10.9%). In the same review, these investigators also found a much lower incidence of renal failure in similar patients who received only linezolid ( 6.7%) as opposed

Pritchard et al. (2008) conducted a retrospective analysis of ~ 3,000 courses of vancomycin given between 2003 and 2007. The aim of their study was to determine the relationship between vancomycin trough concentrations and nephrotoxicity. They noted that trough levels >14 mg/L was an independent risk factor for renal injury among others to be

In contrast, when serum trough vancomycin levels were prospectively limited to the lower range of 5 to 10 mg/L and if peak levels were kept ~ 28 mg/L, in 1985, Sorrel et al found no AKI with vancomycin (when used alone in two patients) and < 8% incidence of AKI even if combined with an aminoglycoside among the 54 patients studied. Taken together, these findings indicate that vancomycin must be considered nephrotoxic, especially at high serum levels, although it was found to be relatively safe at low trough or steady-state levels. Parenthetically, Bailie et al. (1988) had reviewed the utility of peak serum levels as an indicator of vancomycin induced nephrotoxicity and ototoxicity. They determined that peak vancomycin concentration per se may be relatively minor in producing and predicting nephrotoxicity as opposed to the total area under the serum concentration-time curve

b. Increased incidence of AKI when vancomycin was concurrently administered with an

At least three to four studies have found the synergistic nephrotoxic potential between vancomycin and aminoglycosides. In a retrospective study, Farber et al in 1983 found that more patients who received both vancomycin and gentamicin (12 of 34) had suffered from nephrotoxicity as compared to those getting vancomycin alone (3 of 60) with a p value of <0.001. In the studies by Sorrel et al (1985), AKI was found in 4 of 54 vancomycin-treated patients, but all 4 had also received aminoglycosides. In contrast, no AKI was found in the

These findings, however, were not uniformly observed (Downs et al. in 1989, Cimino et al. in 1987, Mellor et al. in 1985), perhaps due to intrinsic differences in their patient characteristics, definitions of acute renal failure, and the divergence in their study methods. In contrast, in 1990 Rybak et al. confirmed that the combination of aminoglycosides and

The prospective studies by Goetz & Sayer published in 1993 provided corroboration for the additive nephrotoxic potential between vancomycin and aminoglycosides. The incidence of nephrotoxicity was 19% in patients receiving vancomycin alone, 12% in patients receiving an aminoglycoside alone and 24% in patients receiving combined vancomycin and an

Several studies have led to the conclusion that prolonged therapy with vancomycin was a risk factor for AKI (Goetz & Sayer 1993, Hidayat et al 2006, Pritchard et al 2008). Besides showing the synergism between aminoglycosides and vancomycin in causing AKI, Goetz & Sayer observed that a duration of >21 days posed greater risk for renal toxicity. In

c. Increased incidence of AKI with prolonged vancomycin administration.

to either > 4 or < 4 g /day of vancomycin (*P=*0.001).

elaborated below.

(AUC).

aminoglycoside.

two on vancomycin alone.

aminoglycoside.

vancomycin was more nephrotoxic than either drug alone.

the prospective studies by Hidayat et al on targeting higher trough vancomycin levels for MRSA strains with high MIC, they not only confirmed the previous association between high trough levels and nephrotoxicity, but also a link between prolonged treatment and AKI.

The retrospective review by Pritchard et al in 2008 also identified the duration of vancomycin administration as an independent risk factor. In their studies of ~ 3,000 courses of vancomycin given between 2003 and 2007, they observed that therapy over 7 days was associated with AKI. They also suggested baseline serum creatinine > 1.7 mg/dl as another independent risk factor. It is however unclear if this association merely reflects a heightened sensitivity of the clinicians to AKI, an enhanced detection of renal failure with an already elevated baseline creatinine, and/or intrinsically greater susceptibility of chronically diseased kidneys to new and acute insults.

d. Increased incidence of AKI or slower recovery from pre-existing ARF if treated with vancomycin versus linezolid.

In the retrospective review by Lodise et al (2008) on the impact of ≥ 4 g/day of vancomycin, they also found a significantly lower incidence of renal failure in patients on linezolid (6.7% vs. 34.6% in patients on > 4 g /day or 10.9 % in those on < 4 g /day) (*P=*0.001). In the treatment of nosocomial pneumonia, Kalil et al (2010) performed a meta-analysis to test the hypothesis of the superiority of linezolid over vancomycin. But they found no significant difference in either vancomycin efficacy or risks of renal dysfunctions, although the study was not powered to compare the nephrotoxic potential between the two drugs.

Rodriguez Colomo et al. (2010) conducted a retrospective, multicenter observational study in patients in intensive care unit with pre-existing renal failure. They found that those patients treated with linezolid had a better renal recovery than those treated with vancomycin, implying either continued nephrotoxic susceptibility or superimposed injury with vancomycin in these cohorts.

In brief, despite the mounting body of indirect evidence summarized above suggesting a role of vancomycin in AKI, firm and unambiguous proofs for a cause-and-effect linkage remain elusive. Virtually all of the studies cited and reviewed above did not offer sufficient details on those individual patients with presumed Van-AKI to allow independent and objective confirmation of a causal and unequivocal relationship. Inherent in the nature of these large-cohort surveys and drug toxicity monitoring studies, despite showing statistical significance among different cohorts, none of the other known and potential etiologic factors for the ARF could be readily evaluated in individual affected patients, let alone vigorously excluded. For instance, the evolution of their serial renal function and the subsequent clinical course after cessation of vancomycin were either not provided or extractable from those individuals afflicted with AKI. A larger prospective study with sufficient clinical details is therefore needed to objectively eliminate all other confounding variables and to prove the implied cause-and-effect relationship.

#### **(A) (2): Evidence for Van-AKI based on published case reports**

Between 1956 and 1986, 57 cases of ARF were described in the course of vancomycin administration and had been attributed to vancomycin. However, over a half of them were reported within the first 6 years of vancomycin use, when impurities were considered to be the most likely culprit (Bailie & Neal 1988)

Although they spanned out over 5 decades in the medical literature, there have been but fewer than two dozen well documented cases of ARF which can be confidently and

Vancomycin-Induced Nephrotoxicity 193

young man with chills, high fever, tachycardia and catheter infected with Serratia (Shah-Khan et al (2011)). He appeared to be septic from a PICC line and exit site infection although peripheral blood culture was negative and there was no frank hypotension. Serum creatinine rose from 0.97 to 4.26 mg/dl in a day and required three hemodialysis treatments for several days of severe oliguria, a serum vancomycin of 64.7 mg/L on day 4, and a sustained elevation of creatinine > 9 mg/dl from days 4 to 9. Although urine output rose to 1-2.5 liters a day since day 5, serum creatinine remained elevated at 1.24 mg/dl even by day 30. The AKI in this man confirmed the observation and caution by Lodise et al (2008) that >4 g of vancomycin /day posed extra nephrotoxic risks. The rapidity of his functional recovery, albeit incomplete, might be related to the single day of brief exposure to

The difficulty of identifying in 5 decades even 2 dozen cases of definite or probable Van-AKI serve to explain the uncertainty and continued controversy regarding the nephrotoxicity of vancomycin. Although there have been many other reports of vancomycin-associated nephrotoxicity, most of them turned out to have been very poorly documented. In most of them, some other renal insults could be easily identified to explain their ARF if only the clinical details were more meticulously, comprehensively, and/or objectively analyzed. In general, often overlooked and/or frequently missed were the concomitant aminoglycosides or nephrotoxic medications, coexisting sepsis or bacteremia, hypotension, hemodynamic factors, pre-existing renal diseases, radio-contrast dye insults, and/or allergic interstitial nephritis. In all objectivity, these factors proved to be the more reasonable and probable etiologies for the ARF without necessarily invoking vancomycin. To offer more vigorous evidence for Van-AKI, we will describe in the following section B our 6 patients. We shall provide sufficient details to demonstrate the causal role of vancomycin, having carefully considered and then excluded most if not all described confounding factors or other potential etiologies. In all six patients, we shall also provide complete information about their entire clinical course showing the temporal evolution of the AKI (in an individual set of three figures per patient as well as a separate case report for each). High-lighted will be the initial renal dysfunctions and the subsequent recovery upon cessation of vancomycin, against the temporal profile of the rising and falling serial

vancomycin although excessive in total quantity.

vancomycin levels (Figs 1-6). **Results for objective 1:** 

**(B) Evidence for VAN-AKI derived from 6 cases observed at OUHSC** 

Of the 101 cases referred for acute renal failure (ARF) (Table 4), 78 (77%) were attributable to intra-renal causes (Table 5), as opposed to pre-renal factors like hemodynamic etiologies or volume depletion (19 or 19%) , or post-renal causes like obstructive nephropathy (4 or 4%). Among the 78 patients with AKI due to intra-renal etiologies (Table 5), 45 (or 58%) could be attributed to sepsis or septic shock, 11 (or 14%) to multiple or unidentifiable factors, 7 (or 9%) to allergic interstitial nephritis, 3 (or 4%) to radio-contrast dye insults, 2 (or 2.6 %) to rhabdomyolysis, and 10 (or 13%) to nephrotoxic antibiotics. Of the 10 patients with antibiotic-induced AKI, one was linked to colistin, another to amphotericin B, 4 solely caused by vancomycin and 4 principally due to vancomycin. We shall focus on 6 of these 8 (4 solely due to vancomycin and two others with vancomycin as the uncontested primary etiology). Demographic details and clinical characteristics at baseline for the entire group are tabulated in Table 6A. The usual pre-renal and obstructive etiologies were excluded by conventional clinical and laboratory studies. None exhibited signs of hypotension, sepsis or

objectively attributed to vancomycin (Dutton & Elmes et al (1959), Dangerfield et al (1960), Odio et al (1984), Frimat et al (1995), Sokol et al (2004), Barraclough et al (2007), Ladino et al (2008), Psevdos et al (2009), Shah-Khan et al (2011)) (Table 2).

A very early case series was described by Dutton & Elmes (1959) who reported that 4 out of 9 vancomycin-treated patients developed renal failure (Table 2). The authors did not measure vancomycin or report drug levels. Unfortunately, all 4 affected patients had suffered from pre-existing renal diseases. Most remarkably, they had all received relatively high doses of vancomycin (between 6-13 grams over 2-5 days). One described method of administration involved rather rapid direct injection in 20 ml saline over only 5 minutes. In retrospect, the high dose and the bolus injection might have resulted in excessive blood and renal tissue concentrations and contributed to the high rates of acute nephrotoxicity, as clearly demonstrated by the dosage comparison studies of Lodise et al (2008).

Dangerfield et al (1960) described nephrotoxicity in 11 out of 85 patients in their series. They defined nephrotoxicity as an otherwise unexplained elevation in serum creatinine ≥ 0.5mg/dl. Eight of these patients had no pre-existing renal disease. Follow up demonstrated a return to baseline renal function in 3-4 weeks. No serum vancomycin concentrations were reported. More importantly, no details on these 11 patients were provided for an objective review or an independent confirmation that no other factors could have contributed to the ARF. Since patients with serious infections requiring vancomycin therapy could have concurrent sepsis, hemodynamic derangements, volume depletion and/or dehydration, it remains unclear if the well known etiologies for AKI had been systematically and definitively excluded, especially if the criterion for ARF was a simply fixed serum creatinine elevation of 0.5 mg/dl irrespective of the starting baseline levels. The actual drop in glomerular filtration rate (GFR) is relatively minor (~ 10 ml/min or ~ 10 %) if serum creatinine rose from a baseline of 2 to 2.5 mg/dl, which could be easily explained by many pre-renal factors. In contrast, a rise in serum creatinine from a baseline of 1 to 1.5 mg/dl could easily reflect ~33 ml/min or ~33 % drop in GFR. Thus at the very best, these generalized descriptions offer no stronger evidence for Van-AKI than that inferred from large efficacy and toxicity studies reviewed and commented above [(A) (1); Table 1].

Each of the report by Frimat et al (1995) and by Barraclough et al (2007) described a case in which they found no alternative explanations for the ARF except vancomycin (Table 2). In the former report, the patient received 39 grams of vancomycin over 17 days, with a peak drug level of 50 mg/L. The patient needed 2 sessions of hemodialysis before the renal function very slowly recovered over > 2 months. In the latter report, the patient had a peak vancomycin level of 66 mg/L and had no other plausible explanation for ARF. Kidney function recovered to baseline in 5-6 weeks. In our opinion, these 2 cases demonstrated rather convincingly the causal relationship between vancomycin and the associated ARF, very similar to our 6 patients to be described below (see Fig 1-6 and Table 6).

In 2008 Ladino et al. presented a case series of 5 patients believed to be Van-AKI. Vancomycin levels were reported to be between 42-86 mg/L and renal function recovered in 3-4 weeks after stopping vancomycin. These authors believed that they had ruled out other causes of ARF. However, in three of the five, there were equally viable etiologic explanations. Thus, one patient had full-blown sepsis, another had bacteremia, and the third had evidence for acute interstitial nephritis (AIN), making it difficult to accept vancomycin as the principal or sole culprit.

In 2011 a case of renal biopsy-proven acute tubular necrosis (ATN) was reported. The authors attributed the ARF to 5 g of intravenous vancomycin given in < 24 h to a 103-kg

objectively attributed to vancomycin (Dutton & Elmes et al (1959), Dangerfield et al (1960), Odio et al (1984), Frimat et al (1995), Sokol et al (2004), Barraclough et al (2007), Ladino et al

A very early case series was described by Dutton & Elmes (1959) who reported that 4 out of 9 vancomycin-treated patients developed renal failure (Table 2). The authors did not measure vancomycin or report drug levels. Unfortunately, all 4 affected patients had suffered from pre-existing renal diseases. Most remarkably, they had all received relatively high doses of vancomycin (between 6-13 grams over 2-5 days). One described method of administration involved rather rapid direct injection in 20 ml saline over only 5 minutes. In retrospect, the high dose and the bolus injection might have resulted in excessive blood and renal tissue concentrations and contributed to the high rates of acute nephrotoxicity, as

Dangerfield et al (1960) described nephrotoxicity in 11 out of 85 patients in their series. They defined nephrotoxicity as an otherwise unexplained elevation in serum creatinine ≥ 0.5mg/dl. Eight of these patients had no pre-existing renal disease. Follow up demonstrated a return to baseline renal function in 3-4 weeks. No serum vancomycin concentrations were reported. More importantly, no details on these 11 patients were provided for an objective review or an independent confirmation that no other factors could have contributed to the ARF. Since patients with serious infections requiring vancomycin therapy could have concurrent sepsis, hemodynamic derangements, volume depletion and/or dehydration, it remains unclear if the well known etiologies for AKI had been systematically and definitively excluded, especially if the criterion for ARF was a simply fixed serum creatinine elevation of 0.5 mg/dl irrespective of the starting baseline levels. The actual drop in glomerular filtration rate (GFR) is relatively minor (~ 10 ml/min or ~ 10 %) if serum creatinine rose from a baseline of 2 to 2.5 mg/dl, which could be easily explained by many pre-renal factors. In contrast, a rise in serum creatinine from a baseline of 1 to 1.5 mg/dl could easily reflect ~33 ml/min or ~33 % drop in GFR. Thus at the very best, these generalized descriptions offer no stronger evidence for Van-AKI than that inferred from

clearly demonstrated by the dosage comparison studies of Lodise et al (2008).

large efficacy and toxicity studies reviewed and commented above [(A) (1); Table 1].

very similar to our 6 patients to be described below (see Fig 1-6 and Table 6).

vancomycin as the principal or sole culprit.

Each of the report by Frimat et al (1995) and by Barraclough et al (2007) described a case in which they found no alternative explanations for the ARF except vancomycin (Table 2). In the former report, the patient received 39 grams of vancomycin over 17 days, with a peak drug level of 50 mg/L. The patient needed 2 sessions of hemodialysis before the renal function very slowly recovered over > 2 months. In the latter report, the patient had a peak vancomycin level of 66 mg/L and had no other plausible explanation for ARF. Kidney function recovered to baseline in 5-6 weeks. In our opinion, these 2 cases demonstrated rather convincingly the causal relationship between vancomycin and the associated ARF,

In 2008 Ladino et al. presented a case series of 5 patients believed to be Van-AKI. Vancomycin levels were reported to be between 42-86 mg/L and renal function recovered in 3-4 weeks after stopping vancomycin. These authors believed that they had ruled out other causes of ARF. However, in three of the five, there were equally viable etiologic explanations. Thus, one patient had full-blown sepsis, another had bacteremia, and the third had evidence for acute interstitial nephritis (AIN), making it difficult to accept

In 2011 a case of renal biopsy-proven acute tubular necrosis (ATN) was reported. The authors attributed the ARF to 5 g of intravenous vancomycin given in < 24 h to a 103-kg

(2008), Psevdos et al (2009), Shah-Khan et al (2011)) (Table 2).

young man with chills, high fever, tachycardia and catheter infected with Serratia (Shah-Khan et al (2011)). He appeared to be septic from a PICC line and exit site infection although peripheral blood culture was negative and there was no frank hypotension. Serum creatinine rose from 0.97 to 4.26 mg/dl in a day and required three hemodialysis treatments for several days of severe oliguria, a serum vancomycin of 64.7 mg/L on day 4, and a sustained elevation of creatinine > 9 mg/dl from days 4 to 9. Although urine output rose to 1-2.5 liters a day since day 5, serum creatinine remained elevated at 1.24 mg/dl even by day 30. The AKI in this man confirmed the observation and caution by Lodise et al (2008) that >4 g of vancomycin /day posed extra nephrotoxic risks. The rapidity of his functional recovery, albeit incomplete, might be related to the single day of brief exposure to vancomycin although excessive in total quantity.

The difficulty of identifying in 5 decades even 2 dozen cases of definite or probable Van-AKI serve to explain the uncertainty and continued controversy regarding the nephrotoxicity of vancomycin. Although there have been many other reports of vancomycin-associated nephrotoxicity, most of them turned out to have been very poorly documented. In most of them, some other renal insults could be easily identified to explain their ARF if only the clinical details were more meticulously, comprehensively, and/or objectively analyzed. In general, often overlooked and/or frequently missed were the concomitant aminoglycosides or nephrotoxic medications, coexisting sepsis or bacteremia, hypotension, hemodynamic factors, pre-existing renal diseases, radio-contrast dye insults, and/or allergic interstitial nephritis. In all objectivity, these factors proved to be the more reasonable and probable etiologies for the ARF without necessarily invoking vancomycin.

To offer more vigorous evidence for Van-AKI, we will describe in the following section B our 6 patients. We shall provide sufficient details to demonstrate the causal role of vancomycin, having carefully considered and then excluded most if not all described confounding factors or other potential etiologies. In all six patients, we shall also provide complete information about their entire clinical course showing the temporal evolution of the AKI (in an individual set of three figures per patient as well as a separate case report for each). High-lighted will be the initial renal dysfunctions and the subsequent recovery upon cessation of vancomycin, against the temporal profile of the rising and falling serial vancomycin levels (Figs 1-6).

### **Results for objective 1:**

## **(B) Evidence for VAN-AKI derived from 6 cases observed at OUHSC**

Of the 101 cases referred for acute renal failure (ARF) (Table 4), 78 (77%) were attributable to intra-renal causes (Table 5), as opposed to pre-renal factors like hemodynamic etiologies or volume depletion (19 or 19%) , or post-renal causes like obstructive nephropathy (4 or 4%). Among the 78 patients with AKI due to intra-renal etiologies (Table 5), 45 (or 58%) could be attributed to sepsis or septic shock, 11 (or 14%) to multiple or unidentifiable factors, 7 (or 9%) to allergic interstitial nephritis, 3 (or 4%) to radio-contrast dye insults, 2 (or 2.6 %) to rhabdomyolysis, and 10 (or 13%) to nephrotoxic antibiotics. Of the 10 patients with antibiotic-induced AKI, one was linked to colistin, another to amphotericin B, 4 solely caused by vancomycin and 4 principally due to vancomycin. We shall focus on 6 of these 8 (4 solely due to vancomycin and two others with vancomycin as the uncontested primary etiology). Demographic details and clinical characteristics at baseline for the entire group are tabulated in Table 6A. The usual pre-renal and obstructive etiologies were excluded by conventional clinical and laboratory studies. None exhibited signs of hypotension, sepsis or

Vancomycin-Induced Nephrotoxicity 195

with resolution of his pneumonia and improvement of the diverticulitis. He was discharged with follow up in the general medicine clinic after completing his course of antibiotics.

Fig. 1. A

Fig. 1. B

bacteremia despite mild leukocytosis. There were no physical, hematologic or urinary evidence to suggest allergic interstitial nephritis. Although two patients had received radiocontrast dye injection, these were temporally unrelated to the AKI.

These 6 cases will also be individually presented in narrative form, along with an accompanying 3-part figure per patient. Three of them were treated for MRSA or Health Care Associated Pneumonia (HCAP) (Cases 1, 4, 5) and three for osteomyelitis from proven or presumed MRSA (Cases 2, 3, 6) (Table 6A). The individual figure serves to illustrate the changes in serum creatinine (A), changes in 100/serum creatinine, as an estimate of CrCl (B), and changes in the levels of serum vancomycin (C) as a function of time from the first day of vancomycin therapy through day 80 since the initiation or the last day of follow-up whichever was longer (Fig 1-6). Thus individually and collectively these 6 cases offer the strongest support for the concept and diagnosis of Van-AKI, especially in the context of the previously reviewed literature.

#### **Case I:**

A 48-year-old white man was admitted to the general internal medicine ward with delirium tremens, diverticulitis, and community acquired pneumonia. He had a past medical history significant for rheumatic fever, rheumatic heart disease, history of infective endocarditis 6 years earlier with septic emboli. There was also a history of diverticulitis and alcoholism. His heart rate was 125 beats/minute and his blood pressure was 143/85 mmHg. The patient was agitated and hallucinating, but otherwise physical examination was unremarkable. Initially white blood cell count was 11.1 K/mm3 and hemoglobin of 12.4 g/dL. Blood chemistries were significant for Na of 130 mEq/L, K of 2.6 mEq/L, and Cl of 85 mEq/L. His serum creatinine was 0.93 mg/dl. He had elevated liver enzymes and bilirubin (aspartate aminotransferase 306 units/L, alanine aminotransferase126 units/L, total alkaline phosphatase 203 units/L, and total bilirubin 2.7 mg/dl), which all eventually resolved in the course of general and specific therapy during his hospitalization. Serum alcohol level was < 10 mg/dl. Chest X ray revealed perihilar right lower lobe and left lower lobe pneumonia as well as right middle and left lingular pneumonia. Computed tomography (CT) scan of the abdomen and pelvis with intravenous and oral contrast showed findings consistent with sigmoid diverticulitis. Blood and urine cultures were negative.

The patient was given lorazepam as needed for alcohol withdrawal symptoms. Regarding his antibiotic regimen, he was initially started on moxifloxacin 400 mg intravenous once daily; the antibiotic regimen was changed on day 3 to vancomycin, piperacillin/tazobactam, and ciprofloxacin given his poor clinical response. Later on day 5, levofloxacin substituted ciprofloxacin for the same reason. Vancomycin was initially started at a dose of 1 g intravenously q12 h (from hospital days 3 through 5). The dose was increased to 1 g intravenously q8 h on day 6 because of a low vancomycin trough level of < 5 mg/L. The dose was further increased on day 7 when trough level was 8 mg/L.

On hospital day 10, vancomycin trough level was found to be 68 mg/L. Vancomycin was therefore discontinued. Creatinine level ranged between 0.57 and 0.93 mg/dl during the first 9 hospital days, but it increased to 2.34 mg/dl on day 10 and continued to rise to a peak of 4.69 mg/dl on day 15 (Fig 1A). Urinalysis done on day 10 of hospital stay was normal with no urinary sediment. Renal ultrasound was unremarkable. Serum creatinine started to decline after that and reached 1.07 mg/dl on day 33 (two days prior to his discharge). Random vancomycin levels were checked periodically after stopping the drug. Level declined to 5 mg/L on day 20. The patient improved clinically throughout his hospital stay

bacteremia despite mild leukocytosis. There were no physical, hematologic or urinary evidence to suggest allergic interstitial nephritis. Although two patients had received radio-

These 6 cases will also be individually presented in narrative form, along with an accompanying 3-part figure per patient. Three of them were treated for MRSA or Health Care Associated Pneumonia (HCAP) (Cases 1, 4, 5) and three for osteomyelitis from proven or presumed MRSA (Cases 2, 3, 6) (Table 6A). The individual figure serves to illustrate the changes in serum creatinine (A), changes in 100/serum creatinine, as an estimate of CrCl (B), and changes in the levels of serum vancomycin (C) as a function of time from the first day of vancomycin therapy through day 80 since the initiation or the last day of follow-up whichever was longer (Fig 1-6). Thus individually and collectively these 6 cases offer the strongest support for the concept and diagnosis of Van-AKI, especially in the context of the

A 48-year-old white man was admitted to the general internal medicine ward with delirium tremens, diverticulitis, and community acquired pneumonia. He had a past medical history significant for rheumatic fever, rheumatic heart disease, history of infective endocarditis 6 years earlier with septic emboli. There was also a history of diverticulitis and alcoholism. His heart rate was 125 beats/minute and his blood pressure was 143/85 mmHg. The patient was agitated and hallucinating, but otherwise physical examination was unremarkable. Initially white blood cell count was 11.1 K/mm3 and hemoglobin of 12.4 g/dL. Blood chemistries were significant for Na of 130 mEq/L, K of 2.6 mEq/L, and Cl of 85 mEq/L. His serum creatinine was 0.93 mg/dl. He had elevated liver enzymes and bilirubin (aspartate aminotransferase 306 units/L, alanine aminotransferase126 units/L, total alkaline phosphatase 203 units/L, and total bilirubin 2.7 mg/dl), which all eventually resolved in the course of general and specific therapy during his hospitalization. Serum alcohol level was < 10 mg/dl. Chest X ray revealed perihilar right lower lobe and left lower lobe pneumonia as well as right middle and left lingular pneumonia. Computed tomography (CT) scan of the abdomen and pelvis with intravenous and oral contrast showed findings consistent with

The patient was given lorazepam as needed for alcohol withdrawal symptoms. Regarding his antibiotic regimen, he was initially started on moxifloxacin 400 mg intravenous once daily; the antibiotic regimen was changed on day 3 to vancomycin, piperacillin/tazobactam, and ciprofloxacin given his poor clinical response. Later on day 5, levofloxacin substituted ciprofloxacin for the same reason. Vancomycin was initially started at a dose of 1 g intravenously q12 h (from hospital days 3 through 5). The dose was increased to 1 g intravenously q8 h on day 6 because of a low vancomycin trough level of < 5 mg/L. The

On hospital day 10, vancomycin trough level was found to be 68 mg/L. Vancomycin was therefore discontinued. Creatinine level ranged between 0.57 and 0.93 mg/dl during the first 9 hospital days, but it increased to 2.34 mg/dl on day 10 and continued to rise to a peak of 4.69 mg/dl on day 15 (Fig 1A). Urinalysis done on day 10 of hospital stay was normal with no urinary sediment. Renal ultrasound was unremarkable. Serum creatinine started to decline after that and reached 1.07 mg/dl on day 33 (two days prior to his discharge). Random vancomycin levels were checked periodically after stopping the drug. Level declined to 5 mg/L on day 20. The patient improved clinically throughout his hospital stay

contrast dye injection, these were temporally unrelated to the AKI.

sigmoid diverticulitis. Blood and urine cultures were negative.

dose was further increased on day 7 when trough level was 8 mg/L.

previously reviewed literature.

**Case I:** 

with resolution of his pneumonia and improvement of the diverticulitis. He was discharged with follow up in the general medicine clinic after completing his course of antibiotics.

Fig. 1. B

Vancomycin-Induced Nephrotoxicity 197

was started on intravenous vancomycin at 1 g every 12 h with an intended duration of

On a regular out-patient follow up towards the end of antibiotic therapy, the patient was found to have a creatinine of 6.9 mg/dl from a baseline of 1.1 mg/dl (Fig 2A). On admission, he did not have any significant complaints. Outpatient medications included amlodipine, clonidine, glipizide, hydrochlorothiazide, lisinopril, hydrocodone, insulin, omeprazole, tramadol, and naproxen. Lisinopril dose had been constant for at least two years prior to admission. The patient admitted to taking two tablets of naproxen 500 mg

Examination was significant for right ankle pitting edema and for a sinus tract over the lateral malleolus draining serous fluid. Initial laboratory revealed a white blood cell count of 7.3 K/mm3 and hemoglobin of 10.0 g/dl. Chemistry was significant for BUN of 63 mg/dl, bicarbonate of 16 mEq/L, and creatinine 6.87 mg/dl. Previously, his serum trough vancomycin levels ranged between 11.5 mg/L and 20.9 mg/L since the initiating the antibiotics, with a level of 11.5 mg/L measured two weeks prior to admission (Fig 2C). On admission, random vancomycin level was however found to be 67 mg/L. Urine was positive for eosinophils. Urine creatinine was 87.6 mg/dl and urine protein 37 mg/dl, yielding a ratio of 0.42. Renal ultrasound revealed horseshoe kidneys with dimensions of

11.3 x 5.3 x 5.1 cm and 11.1 x 5.1 x 4.7 cm respectively for the right and left kidneys.

On day 1, vancomycin was discontinued along with stopping lisinopril and naproxen. Creatinine started to trend down reaching 2.15 mg/dl about 10 weeks later (Fig 2A). Since all medications except vancomycin had previously been taken without producing any renal toxicity, the temporal relationship between the high vancomycin level and elevated creatinine strongly suggests Vancomycin-AKI. His subsequent clinical course of a slow but steady recovery upon cessation of vancomycin lends further support to this formulation.

treatment for eight weeks.

Fig. 2. A

daily for about seven months previously for pain relief.

#### Fig. 1. C

We believed his acute kidney injury (AKI) was secondary to direct vancomycin nephrotoxicity based on the temporal relationship between the continually escalating dosage and documented excessive trough vancomycin levels on the one hand and the worsening kidney function on the other hand. Although he had received IV contrast on day 1, his serum creatinine did not rise until day 8. All pre-renal hemodynamic factors and postrenal causes were excluded, as were the absence of other intrinsic nephrotoxins. His recovery upon stoppage of vancomycin gave additional credence to our formulation. Although he had pneumonia, at no times did he have bacteremia or any signs of sepsis or hypotension. There were also no signs of allergic interstitial nephritis by serial exam, blood or urine eosinophilia. The patient was on multiple medications when he developed his AKI, including vancomycin, levofloxacin, piperacillin/tazobactam, ondansetron, enoxaparin, lorazepam, morphine sulfate, ompeprazole, sucralfate, and thiamine. But all these medications (except for vancomycin and enoxaparin, the latter replaced by unfractionated heparin) were continued during his subsequent renal recovery, arguing against any possible pathogenic role in the AKI.

#### **Case 2:**

This was a 53-year-old man with recurrent and recalcitrant osteomyelitis admitted for acute renal failure. He was known to suffer from diabetes mellitus, hypertension, chronic hepatitis C infection, alcohol abuse, and cocaine dependence. The patient sustained a right ankle fracture secondary to a fall and status post intramedullary nailing for fusion of right tibiotalar and subtalar joints. The patient's course was complicated by two episodes of right ankle osteomyelitis post surgery; the first episode happened about four months after the surgery which was treated with intravenous vancomycin and piperacillin/tazobactam in addition to the removal of 2 screws from the right foot. The second episode took place five months thereafter. At that time he underwent removal of the remaining screws and nail and

We believed his acute kidney injury (AKI) was secondary to direct vancomycin nephrotoxicity based on the temporal relationship between the continually escalating dosage and documented excessive trough vancomycin levels on the one hand and the worsening kidney function on the other hand. Although he had received IV contrast on day 1, his serum creatinine did not rise until day 8. All pre-renal hemodynamic factors and postrenal causes were excluded, as were the absence of other intrinsic nephrotoxins. His recovery upon stoppage of vancomycin gave additional credence to our formulation. Although he had pneumonia, at no times did he have bacteremia or any signs of sepsis or hypotension. There were also no signs of allergic interstitial nephritis by serial exam, blood or urine eosinophilia. The patient was on multiple medications when he developed his AKI, including vancomycin, levofloxacin, piperacillin/tazobactam, ondansetron, enoxaparin, lorazepam, morphine sulfate, ompeprazole, sucralfate, and thiamine. But all these medications (except for vancomycin and enoxaparin, the latter replaced by unfractionated heparin) were continued during his subsequent renal recovery, arguing against any possible

This was a 53-year-old man with recurrent and recalcitrant osteomyelitis admitted for acute renal failure. He was known to suffer from diabetes mellitus, hypertension, chronic hepatitis C infection, alcohol abuse, and cocaine dependence. The patient sustained a right ankle fracture secondary to a fall and status post intramedullary nailing for fusion of right tibiotalar and subtalar joints. The patient's course was complicated by two episodes of right ankle osteomyelitis post surgery; the first episode happened about four months after the surgery which was treated with intravenous vancomycin and piperacillin/tazobactam in addition to the removal of 2 screws from the right foot. The second episode took place five months thereafter. At that time he underwent removal of the remaining screws and nail and

Fig. 1. C

**Case 2:** 

pathogenic role in the AKI.

was started on intravenous vancomycin at 1 g every 12 h with an intended duration of treatment for eight weeks.

On a regular out-patient follow up towards the end of antibiotic therapy, the patient was found to have a creatinine of 6.9 mg/dl from a baseline of 1.1 mg/dl (Fig 2A). On admission, he did not have any significant complaints. Outpatient medications included amlodipine, clonidine, glipizide, hydrochlorothiazide, lisinopril, hydrocodone, insulin, omeprazole, tramadol, and naproxen. Lisinopril dose had been constant for at least two years prior to admission. The patient admitted to taking two tablets of naproxen 500 mg daily for about seven months previously for pain relief.

#### Fig. 2. A

Examination was significant for right ankle pitting edema and for a sinus tract over the lateral malleolus draining serous fluid. Initial laboratory revealed a white blood cell count of 7.3 K/mm3 and hemoglobin of 10.0 g/dl. Chemistry was significant for BUN of 63 mg/dl, bicarbonate of 16 mEq/L, and creatinine 6.87 mg/dl. Previously, his serum trough vancomycin levels ranged between 11.5 mg/L and 20.9 mg/L since the initiating the antibiotics, with a level of 11.5 mg/L measured two weeks prior to admission (Fig 2C). On admission, random vancomycin level was however found to be 67 mg/L. Urine was positive for eosinophils. Urine creatinine was 87.6 mg/dl and urine protein 37 mg/dl, yielding a ratio of 0.42. Renal ultrasound revealed horseshoe kidneys with dimensions of 11.3 x 5.3 x 5.1 cm and 11.1 x 5.1 x 4.7 cm respectively for the right and left kidneys.

On day 1, vancomycin was discontinued along with stopping lisinopril and naproxen. Creatinine started to trend down reaching 2.15 mg/dl about 10 weeks later (Fig 2A). Since all medications except vancomycin had previously been taken without producing any renal toxicity, the temporal relationship between the high vancomycin level and elevated creatinine strongly suggests Vancomycin-AKI. His subsequent clinical course of a slow but steady recovery upon cessation of vancomycin lends further support to this formulation.

Vancomycin-Induced Nephrotoxicity 199

This patient was a 56 year old man admitted for chronic open draining wound on right foot. He had a history of diabetes mellitus of unknown duration, although he was not taking any medications for diabetes. He reported chronic drainage from his right foot with worsening pain. Otherwise, the review of system was negative, notably for the absence of fever, chills, vomiting, diarrhea, dyspnea, and chest pain. He denied taking any NSAID or recent

On admission he was normotensive and afebrile. He had no orthostatic hypotension. The big toe on his right foot had a large ulcer with purulent drainage and surrounding cellulitis. Nuclear scan confirmed osteomyelitis. Blood cultures and wound cultures were negative. His serum creatinine was 0.9 mg/dl on admission. He was treated with 1 g vancomycin q 12 h. On hospital day 3 his creatinine was 2.56 mg/dl and rose to a peak of 7.4 on hospital

Throughout his hospital stay, he was normotensive and received no other potential nephrotoxic insults including radio-contrast dyes. Due to persistent though mild leukocytosis and a low grade fever, he had undergone above knee amputation on his right side on hospital day 9. This was also prompted by the consideration that he had failed medical treatment and the wound was deemed to have very poor chance of healing based on vascular studies and transcutaneous oxygen tension gradients. He had received 2 g of vancomycin daily for the first 5 hospital days and given his extremely high serum trough or random vancomycin levels (Fig 3C) and the temporal relationship with the acute rise in serum creatinine, his AKI was best explained by vancomycin. In addition, all other etiologic factors, both pre-renal and post-renal causes, had been vigorously excluded. Three weeks after discharge, his amputation wound was healing well and his serum creatinine fell to 1.6 mg/dl, towards his normal baseline

although still significantly elevated considering the loss of his right leg (Fig 3 A).

day 12, falling down to 4.85 at the time of discharge (Fig 3A).

**Case 3:** 

hospitalizations.

Fig. 3. A

Fig. 2. B

#### Fig. 2. C

We found no other stigmata of allergic interstitial nephritis (to either vancomycin) or other potential offending agents. Similar to the other patients, all pre-renal and post-renal factors had been carefully considered and excluded, including the absence of other known intrarenal insults in our patient.

### **Case 3:**

198 Basic Nephrology and Acute Kidney Injury

We found no other stigmata of allergic interstitial nephritis (to either vancomycin) or other potential offending agents. Similar to the other patients, all pre-renal and post-renal factors had been carefully considered and excluded, including the absence of other known intra-

Fig. 2. B

Fig. 2. C

renal insults in our patient.

This patient was a 56 year old man admitted for chronic open draining wound on right foot. He had a history of diabetes mellitus of unknown duration, although he was not taking any medications for diabetes. He reported chronic drainage from his right foot with worsening pain. Otherwise, the review of system was negative, notably for the absence of fever, chills, vomiting, diarrhea, dyspnea, and chest pain. He denied taking any NSAID or recent hospitalizations.

On admission he was normotensive and afebrile. He had no orthostatic hypotension. The big toe on his right foot had a large ulcer with purulent drainage and surrounding cellulitis. Nuclear scan confirmed osteomyelitis. Blood cultures and wound cultures were negative. His serum creatinine was 0.9 mg/dl on admission. He was treated with 1 g vancomycin q 12 h. On hospital day 3 his creatinine was 2.56 mg/dl and rose to a peak of 7.4 on hospital day 12, falling down to 4.85 at the time of discharge (Fig 3A).

#### Fig. 3. A

Throughout his hospital stay, he was normotensive and received no other potential nephrotoxic insults including radio-contrast dyes. Due to persistent though mild leukocytosis and a low grade fever, he had undergone above knee amputation on his right side on hospital day 9. This was also prompted by the consideration that he had failed medical treatment and the wound was deemed to have very poor chance of healing based on vascular studies and transcutaneous oxygen tension gradients. He had received 2 g of vancomycin daily for the first 5 hospital days and given his extremely high serum trough or random vancomycin levels (Fig 3C) and the temporal relationship with the acute rise in serum creatinine, his AKI was best explained by vancomycin. In addition, all other etiologic factors, both pre-renal and post-renal causes, had been vigorously excluded. Three weeks after discharge, his amputation wound was healing well and his serum creatinine fell to 1.6 mg/dl, towards his normal baseline although still significantly elevated considering the loss of his right leg (Fig 3 A).

Vancomycin-Induced Nephrotoxicity 201

having developed left lower lobe pneumonia attributed to MRSA cultured from the tracheal aspirate on the fourth hospital day. This was treated initially with vancomycin 1.5 g every 8 h. After 3 doses, trough vancomycin level was 10 mg/L. Thus vancomycin was increased to 2 g every 8 h, a regimen which was continued for the ensuing 10 days. His vancomycin trough levels on days 4, 5 and 9 of administration were respectively 15,

His serum creatinine was 1.3 mg/dl on admission. After repletion of his extracellular fluid volume, it dropped to 0.6 and stayed in that range for a week (Fig 4A). On days 9 to 10 of vancomycin therapy, his serum creatinine began to climb slightly to 0.9 mg/dl. It rose to 1.2 on day 11 and to 2.8 mg/dl on day 12 of vancomycin administration. It peaked and plateaued at 3.5 to 3.6 mg/dl two weeks after the initiation of vancomycin (Fig 4A). Of note, his serum trough vancomycin level was found to be 110 mg/L eight hours after the last dose of vancomycin. Although there was a peripheral eosinophilia of 12.5% with a peak absolute count of 1,400 ten days after the last dose of vancomycin, his serum creatinine level then was already trending down, arguing against an allergic interstitial nephritis. There was no significant granulocytosis despite an intermittent low-grade fever and mild leukocytosis. All blood cultures drawn throughout his hospital course were negative. Hemophilus influenza grew out from his tracheal aspirate on hospital day 9 and treated for 9 days withpiperacillin/ tazobactam. The patient was hemodynamically stable throughout his hospital stay and he made a slow but steady and significant physical recovery to be able to transfer to a full rehabilitation center on hospital day 36. At that time his serum creatinine

Although he had received IV contrast on day 1, his serum creatinine had remained in the normal range and stable over the first 2 weeks of his hospitalization. All known nephrotoxic insults, pre-renal and post-renal factors were excluded as potential explanation for his AKI.

had also returned to 0.84 mg/dl, very close to his normal baseline.

17 and 20 mg/L (Fig 4C).

Fig. 4. A

#### **Case 4:**

This patient was a 33 year old man with a past medical history of Hirschsprung disease as a child admitted to the trauma service of our Medical Center after an alleged assault. He was intubated at the scene and subsequently treated for multiple facial fractures, right orbital fractures and intracranial hemorrhage. His hospital course was significant for

This patient was a 33 year old man with a past medical history of Hirschsprung disease as a child admitted to the trauma service of our Medical Center after an alleged assault. He was intubated at the scene and subsequently treated for multiple facial fractures, right orbital fractures and intracranial hemorrhage. His hospital course was significant for

Fig. 3. B

Fig. 3. C **Case 4:** 

having developed left lower lobe pneumonia attributed to MRSA cultured from the tracheal aspirate on the fourth hospital day. This was treated initially with vancomycin 1.5 g every 8 h. After 3 doses, trough vancomycin level was 10 mg/L. Thus vancomycin was increased to 2 g every 8 h, a regimen which was continued for the ensuing 10 days. His vancomycin trough levels on days 4, 5 and 9 of administration were respectively 15, 17 and 20 mg/L (Fig 4C).

His serum creatinine was 1.3 mg/dl on admission. After repletion of his extracellular fluid volume, it dropped to 0.6 and stayed in that range for a week (Fig 4A). On days 9 to 10 of vancomycin therapy, his serum creatinine began to climb slightly to 0.9 mg/dl. It rose to 1.2 on day 11 and to 2.8 mg/dl on day 12 of vancomycin administration. It peaked and plateaued at 3.5 to 3.6 mg/dl two weeks after the initiation of vancomycin (Fig 4A). Of note, his serum trough vancomycin level was found to be 110 mg/L eight hours after the last dose of vancomycin. Although there was a peripheral eosinophilia of 12.5% with a peak absolute count of 1,400 ten days after the last dose of vancomycin, his serum creatinine level then was already trending down, arguing against an allergic interstitial nephritis. There was no significant granulocytosis despite an intermittent low-grade fever and mild leukocytosis. All blood cultures drawn throughout his hospital course were negative. Hemophilus influenza grew out from his tracheal aspirate on hospital day 9 and treated for 9 days withpiperacillin/ tazobactam. The patient was hemodynamically stable throughout his hospital stay and he made a slow but steady and significant physical recovery to be able to transfer to a full rehabilitation center on hospital day 36. At that time his serum creatinine had also returned to 0.84 mg/dl, very close to his normal baseline.

#### Fig. 4. A

Although he had received IV contrast on day 1, his serum creatinine had remained in the normal range and stable over the first 2 weeks of his hospitalization. All known nephrotoxic insults, pre-renal and post-renal factors were excluded as potential explanation for his AKI.

Vancomycin-Induced Nephrotoxicity 203

This man was a 75 year old resident of a skilled nursing facility admitted to our medical center because of altered mental status. He had a history of dementia and old cerebrovascular accidents and his outside medications included no nephrotoxic

On examination, he appeared to be confused and disoriented, responsive only to painful stimuli. His blood pressure was 126/70 mm Hg. His pulse rate was 67. Temperature was 36.1 C and his respiratory rate was 18. On room air, his pulse oxygen saturation was 95%. He had coarse crackles in left lower lobe with decreased air entry. No other sources of

The white blood cell count was 11.3 K/mm3. A chest X ray showed left lower lobe consolidation and a small pleural effusion. A CT scan of the head revealedno acute intracranial process. His serum creatinine was 2.42 mg/dl (versus a baseline of 1.5mg/dl). He was thought to be volume depleted. After receiving intravenous fluids, his serum creatinine returned to normal and on day 5, it was 1.11 mg/dl. In the mean time he was given vancomycin 1g q12 h and piperacillin/tazobactam 2.25 g q 6 h (adjusted dose for his renal function) for the treatment of his HCAP. On day 5 of his admission, he was

Six days later, he was re-admitted to the hospital, again with altered mental status and decreased oral intake. His serum creatinine was elevated to 3.45 mg/dl (Fig 5 A). He was hemodynamically stable with blood pressure of 147/96 mm Hg and a pulse rate of 89. White blood cell count was 8.4 K/mm3. Serum Na was 153 mEq/L and K was 3.8 mEq/L. BUN was 15 mg/dl. The urine fractional excretion of Na (FENa) was 13.8%, suggestive of intrinsic or intra-renal disease. Despite intravenous fluids, his serum creatinine continued

discharged back to nursing home to complete a 2 week course of HCAP treatment.

**Case 5:** 

medications.

Fig. 5. A

infection were found on physical examination.

to rise during the first few days (Fig 5A).

Fig. 4. B

#### Fig. 4. C

Thus we believe his clinical course and renal function profile were best explained by acute vancomycin nephrotoxicity. His kidney recovery 3 weeks after stopping vancomycin was also consistent with the typical picture of improved serum creatinine over this time frame as in classical Van-AkI shown here and in the few documented cases published in the literature.

#### **Case 5:**

202 Basic Nephrology and Acute Kidney Injury

Thus we believe his clinical course and renal function profile were best explained by acute vancomycin nephrotoxicity. His kidney recovery 3 weeks after stopping vancomycin was also consistent with the typical picture of improved serum creatinine over this time frame as in classical Van-AkI shown here and in the few documented cases published in the

Fig. 4. B

Fig. 4. C

literature.

This man was a 75 year old resident of a skilled nursing facility admitted to our medical center because of altered mental status. He had a history of dementia and old cerebrovascular accidents and his outside medications included no nephrotoxic medications.

On examination, he appeared to be confused and disoriented, responsive only to painful stimuli. His blood pressure was 126/70 mm Hg. His pulse rate was 67. Temperature was 36.1 C and his respiratory rate was 18. On room air, his pulse oxygen saturation was 95%. He had coarse crackles in left lower lobe with decreased air entry. No other sources of infection were found on physical examination.

The white blood cell count was 11.3 K/mm3. A chest X ray showed left lower lobe consolidation and a small pleural effusion. A CT scan of the head revealedno acute intracranial process. His serum creatinine was 2.42 mg/dl (versus a baseline of 1.5mg/dl). He was thought to be volume depleted. After receiving intravenous fluids, his serum creatinine returned to normal and on day 5, it was 1.11 mg/dl. In the mean time he was given vancomycin 1g q12 h and piperacillin/tazobactam 2.25 g q 6 h (adjusted dose for his renal function) for the treatment of his HCAP. On day 5 of his admission, he was discharged back to nursing home to complete a 2 week course of HCAP treatment.

Six days later, he was re-admitted to the hospital, again with altered mental status and decreased oral intake. His serum creatinine was elevated to 3.45 mg/dl (Fig 5 A). He was hemodynamically stable with blood pressure of 147/96 mm Hg and a pulse rate of 89. White blood cell count was 8.4 K/mm3. Serum Na was 153 mEq/L and K was 3.8 mEq/L. BUN was 15 mg/dl. The urine fractional excretion of Na (FENa) was 13.8%, suggestive of intrinsic or intra-renal disease. Despite intravenous fluids, his serum creatinine continued to rise during the first few days (Fig 5A).

Vancomycin-Induced Nephrotoxicity 205

Patient appeared clinically stable and euvolemic. A kidney ultrasound did not show any obstruction. Of note other causes of acute renal failure were ruled out. He did not have acute interstitial nephritis as there was no rash, peripheral eosinophilia or eosinophiluria. He showed no signs of sepsis and his blood cultures remained negative. He did not receive any nephrotoxic agents or radio-contrast dyes. His antibiotics were switched to

On days 7 and 9 of his second hospitalization, he underwent two sessions of hemodialysis to help manage his oliguria and to help remove the cumulated vancomycin. After the hemodialysis his serum creatinine and serum vancomycin levels both trended down. As vancomycin disappeared from his system, his kidney function improved significantly. Although he required furosemide drip to help manage his oliguria, he became relatively polyuric in the recovery phase of his AKI. Four weeks after discontinuation of vancomycin, his serum creatinine was 1.64 mg/dl close to though still higher than his best baseline value. But he was vastly improved and able to be discharged. At that time vancomycin level was

This patient was a 65 year old man hospitalized for hand osteomyelitis. He had a significant and complicated past and ongoing medical history due to uncontrolled type 2 diabetes mellitus, hypertension, hyperlipidemia, atrial fibrillation, previous stroke, degenerative joint disease of his left hip and knee, status-post knee replacement, gastroesophageal reflux disease, diabetic neuropathy, and a chronic but recently resolved MRSA

His present illness related to his left thumb pain that was initially treated with local steroid injections by his outside doctor. Subsequently, he had a draining ulcer at the first metacarpal joint of his left hand. Four days prior to his transfer from a local hospital to our medical center, MRI showed first metacarpal osteomyelitis and tendonitis. He was started on vancomycin, initially at a dose of 1.5 g every 18 hours three days before the transfer. On the 2nd hospital day with us, gram stain and culture from the left thumb wound showed MRSA. MRSA was also confirmed by intra-operative bone biopsy culture on the 3rd hospital day.

The patient was discharged on the 5th hospital day to complete a prolonged course of vancomycin at a dose of 1.5 g daily at the recommendation of ID consultants. Blood for vancomycin levels and basic metabolic profile was drawn and checked by home health nurse once weekly. After two weeks, his vancomycin dose was increased to 2.0 g daily. Four more weeks later, it was further raised to 2.5 g daily to keep level >15 mg/L (Fig 6C). Three weeks after the last dose increase, although the 24-h trough levels finally reached 17-19 mg/L (Fig 6C), his serum creatinine had also risen from 1.3 to 2.3 mg/dl (Fig 6 A). This represented a further hike from initial baseline of 0.8 at the start of vancomycin

Meloxicam, an NSAID, and lisinopril, which he had taken for years, were temporarily stopped, along with holding his vancomycin for 3 days. When his serum creatinine appeared to stop rising and seemed to stabilize at ~ 2.1 mg/dl, vancomycin was resumed albeit at a reduced dose of 1 g daily. This was however stopped completely due to the persistent elevation of his serum creatinine at 2 mg/dl (Fig 6A). By having excluded obstruction with renal ultrasound, pre-renal or hemodynamic factors, bacteremic or septic etiologies and other intra-renal insults, we believe his subacute decline in renal function (Fig

Vancomycin was continued targeting 24-h trough levels ≥ 15 mg/L.

ciprofloxacin 400 mg IV q 24 h and cefepime 1g q12 h.

9 mg/L. **Case 6:** 

therapy.

diabetic left foot ulcer.

Fig. 5. B

#### Fig. 5. C

In the nursing home his vancomycin level was not monitored. On re-admission, the vancomycin level (77 mg/L) was found to be in toxic range. Vancomycin was thus discontinued. 5 days later serum creatinine was 6.6mg/dl and it continued to climb to a peak level of 10.3 mg/dl on hospital day 7 (Fig 5A).

Patient appeared clinically stable and euvolemic. A kidney ultrasound did not show any obstruction. Of note other causes of acute renal failure were ruled out. He did not have acute interstitial nephritis as there was no rash, peripheral eosinophilia or eosinophiluria. He showed no signs of sepsis and his blood cultures remained negative. He did not receive any nephrotoxic agents or radio-contrast dyes. His antibiotics were switched to ciprofloxacin 400 mg IV q 24 h and cefepime 1g q12 h.

On days 7 and 9 of his second hospitalization, he underwent two sessions of hemodialysis to help manage his oliguria and to help remove the cumulated vancomycin. After the hemodialysis his serum creatinine and serum vancomycin levels both trended down. As vancomycin disappeared from his system, his kidney function improved significantly. Although he required furosemide drip to help manage his oliguria, he became relatively polyuric in the recovery phase of his AKI. Four weeks after discontinuation of vancomycin, his serum creatinine was 1.64 mg/dl close to though still higher than his best baseline value. But he was vastly improved and able to be discharged. At that time vancomycin level was 9 mg/L.
