Sudden renal function deterioration in an elderly patient on vancomycin therapy for endocarditis

Overview

Vancomycin is currently the drug of choice in meticillin-resistant Staphylococcus aureus (MRSA) infection. It is also used prophylactically in some situations in which the patient is at risk for endocarditis. It is often used in combination with other antibacterials in the treatment of endocarditis and is a potential nephrotoxin. Various consensus guidelines differ in their interpretation of vancomycin plasma concentrations. This paper describes a case of a 72-years old Caucasian female patient, who developed significant renal impairment when prescribed vancomycin in combination with penicillin for the treatment of endocarditis, caused by Streptococcus pneumoniae.

Key words:
vancomycin, nephrotoxicity, elderly, endocarditis.

Introduction

Vancomycin is a glycopeptide antibiotic that has been available for 50 years. Its use has increased considerably in the past two decades because it is effective in the treatment of methicillin-resistant Staphylococcus aureus (MRSA) infections which have increased significantly in both community and health care settings (1). Vancomycin is renally excreted unchanged and blood levels can accumulate with decreased renal clearance and its there are case reports of it causing serious nephrotoxicity. Algorithms exist for dosing patients with renal impairment, but these rely on estimation of glomerular filtration rate (GFR) using prediction equations. Vancomycin nephrotoxicity is reported in 5–7% of patients when the drug is used alone (2). However, the data is less clear when it is used in combination with an aminoglycoside, because it is difficult to confirm causality. Some studies found an increase of nephrotoxicity (up to 35% of patients) when vancomycin is used in combination with an aminoglycoside, whereas others report no change (3). Nevertheless, vancomycin is undoubtably a nephrotoxin even as monotherapy (4). The determination of peak concentration is less useful, because levels as high as 50 mg/l did not predict changes in renal or hearing function (5). There are differences regarding interpretation of plasma vancomycin concentrations with a consensus paper in 2002 advocating that trough levels less than 20 mg/L are safe and effective (6). Elderly patients in particular are prone to drug toxicity due to deterioration in organ function with age and polymorbidity. In the present case the clinical fate of an elderly patient, who developed significant renal function impairment on vancomycin therapy is described.

Case description

A 72-years old Caucasian female with past history of myocardial infarction (MI) was admitted to the department of internal medicine with a working diagnosis of endocarditis based on ultrasound examination. Physical examination revealed: body weight = 77 kg, height = 170 cm, body mass index (BMI) = 26.64, blood pressure (BP) 80/50 mm Hg with regular heart rhythms. Initial laboratory tests including blood urea nitrogen (BUN) 3 mmol/L, C-reactive protein (CRP): 145 mg/L. Haemoculture proved positive for Streptococcus pneumonia, and trans-oesophageal echocardiogram revealed tricuspid valve involvement indicating endocarditis. Penicillin 5 mil units I.V. every 6 hours was given initially as the treatment of choice. The Streptococcus pneumonia was shown in vitro to be sensitive to penicillin and vancomycin, but resistant to gentamicin. Initially the patient was treated with penicillin alone, however since CRP remained high (120 mg/L) after seven days penicillin therapy, vancomycin 1g in a one hour infusion was administered at 12 hours intervals. The serum creatinine level on the day vancomycin was started was 35 ∝mol/L. The drug level was determined by a fluorescent polarization immunoassay (FPIA) method, which was validated for routine use in our laboratory. The vancomycin trough serum level taken just before the 5^th dose was 19 mg/L which is within the recommended therapeutic range (4). Meanwhile the serum creatinine measured from the same sample had increased to 284 ∝mol/L from the initial value of 35 after three days of vancomycin therapy, warranting further dosage reassessment. The vancomycin dosing interval was prolonged by 24 hours, but serum creatinine continued to rise (Fig. 1). To manage changes of BP from the initial 80/50 to 170/90 mm Hg, and weight gain (81 kg compared to 77 kg at admission), furosemide 40 mg i.v. bolus was administered and then maintained on 20 mg/day. Following vancomycin discontinuation for 50 hours, the serum creatinine was still elevated at 326 ∝mol/L. Blood urea nitrogen also increased from 3 mmol/L to 11 mmol/L and peaked at 17 mmol/L. CRP decreased from 145 mg/L on penicillin monotherapy to 120, 118, 49 and finally to 34 mg/L after vancomycin co-administration while body temperature normalized. The big challenge was that, despite the reduction of vancomycin dose frequency, the trough serum level of 20 mg/L indicated severe slowing of the elimination process and the continued high serum creatinine (350 ∝mol/L compared to the previous value of 335 ∝mol/L), was sufficiently concerning that withdrawal of vancomycin was considered. Because of the life threatening nature of the patient’s condition a computer assisted therapeutic drug monitoring modeling pharmacokinetic software MW/PHARM (Madiware) version 3.60 was used to calculate the vancomycin dose required to avoid therapeutic failure with an acceptable risk for renal dysfunction. The glomerular filtration rate was estimated to be 0.19 ml/sec using the modification of diet in renal disease (MDRD) equation and the vancomycin dose calculated to maintain the antimicrobial effect was 500 mg every other day (Fig 2).

Discussion

Vancomycin continues to be the antibiotic of choice for the treatment of methicillin-resistant Staphylococcus aureus (MRSA) infection (1, 4). Current recommendations for vancomycin therapy also include the treatment of serious infections of β-lactam- or metronidazole-resistant organisms or when other antibiotic is contraindicated. Vancomycin may be also used prophylactically alone in some situations (7), particularly when the patient is at risk of endocarditis. Vancomycin is also used in combination with other antibacterials to treat endocarditis and similar serious clinical situations. However, it carries the risk of nephrotoxicity since the drug undergoes renal excretion and its blood levels will increase with decreased renal clearance. Algorithms are published for dosing patients with renal impairment, but these rely on estimation of glomerular filtration rate (GFR) using prediction equations (3), which are of limited accuracy. Nephrotoxicity of vancomycin when administered alone is reported to occur in 5–7% of cases (4) but there is more concern when it is administered to elderly patients, who are more prone to drug toxicity due to physiological decreased organ function and polymorbidity. Reports have been emerging that indicate an increased incidence of nephrotoxicity with high-dose vancomycin therapy (8); however, the data are less clear when the drug is used in combination with an aminoglycoside. Some studies report an increase in nephrotoxicity (up to 35% of patients), whereas others report no change when vancomycin is used in combination with an aminoglycoside (3). The patient described here was prescribed penicillin monotherapy before vancomycin addition. The patient also had a loop diuretic administered and the concurrent administration of loop diuretics has been associated with vancomycin nephrotoxicity in elderly patients (9). Furthermore, it is well documented that elderly patients are at a significantly greater risk of vancomycin-associated nephrotoxicity than are younger patients (5) and the relationship between high trough vancomycin levels and nephrotoxicity was recently published (10). In the present case the trough level of vancomycin was 19 mg/L which lies within the reported therapeutic range (1, 4). Endocarditis is a life threatening clinical condition that needs appropriate resolution without delay. In this case based on causative agent confirmation by haemoculture and reportedly well sensitive to penicillin; monotherapy with penicillin was initiated. However, there was a poor clinical response with persistently high CRP level seven days after starting penicillin. Therefore, dual therapy (penicillin with vancomycin) was initiated to gain a synergic effect with the objective of eradicating the infection while maintaining an acceptable renal function since persistent infection and renal function impairment are reported to be among the risk factors which increase mortality in patients with endocarditis (11). A relationship probably exists between vancomycin toxicity and serum concentration, but the information is limited. There is some evidence that nephrotoxicity occurs more frequently when trough serum concentrations exceed 10 mg/L (12). The frequency of ototoxicity varies from < 2% to 5.5% and is more commonly reported in patients with decreased renal function with serum concentrations above 80 mg/L (3, 4). Temporary tinnitus in association with serum vancomycin concentrations near to 40 mg/L or above has been also reported (1). Although it is not clear whether the risk of ototoxicity increases when the drug is co-administered with another non-ototoxic or potentially nephrotoxic drug, experimental study results suggest that the nephrotoxicity of β-lactam antibiotics, where penicillin belongs to may be dependent on organic anion transporter (OAT1) expression (13). In the present case furosemide was also administered on daily basis and the role this may have played in augmenting the renal deterioration is unclear. Although routinely provided, the benefit of the therapeutic drug monitoring in minimizing the risk of toxicity or therapeutic failure with vancomycin is not as clear as for the aminoglycoside antibiotics. Currently it is controversial whether vancomycin should be monitored in all patients and if monitoring is best practice then which specimens should be collected and what serum concentrations should be targeted. The recent increases in vancomycin minimal inhibitory concentrations (MICs) among clinical MRSA isolates (14–16), have prompted some clinicians to use higher-dose therapy, although data supporting positive outcomes with higher doses are lacking (17) and there are recent reports which indicate an increased incidence of nephrotoxicity with high-dose vancomycin therapy (8). Ototoxicity is less common with vancomycin therapy (18) and there are no recent evaluations of ototoxicity with aggressive vancomycin dosing, although a recent study evaluated vancomycin ototoxicity by using audiometry (19). Elderly patients are more prone to drug toxicity because of physiologic deterioration in their organ function. Glycopeptide antibiotics are known nephrotoxins and elderly patients will be particularly at risk. Therapeutic drug monitoring may be helpful for guiding dose adjustments in complex situations (20); however differences exist between the various guidelines regarding interpretation of the drug concentrations. Recent consensus and previous conventional guidelines differ, where the recent consensus considers the trough levels not exceeding 20 mg/L safe and more effective. Many consider determining peak concentration as useless, because levels as high as 50 mg/L have not been associated with any change of renal or hearing states. However, it is now accepted that the vancomycin trough value is an indication of exposure and is the variable well associated with nephrotoxicity which indicates that an exposure-toxicity response relationship exists and trough level may be the best pharmacodynamic index for describing this association (21). An initial vancomycin dose regimen of 1 g every 12 hours has been recommended provided that a trough serum level is determined not later than before the 5^th dose to guide dose adjustment and avoid not only therapeutic failure but also a high risk of toxicity. In the present case the trough level after the fourth dose was only 19 mg/L, which is within the target range recommended in one consensus document (4) that is applied in many clinical settings. Unfortunately, the theoretical base that vancomycin is time dependent rather than high concentration is ignored by those, who advocate higher concentration without any explanation on the benefit of such interventions. However, there are no unequivocal recommendations regarding the maximum vancomycin trough level which will be therapeutically effective and at the same time safe e.g. differences were apparent among institutional experts from the results of the United Kingdom National External Quality Assessment Scheme (UK NEQAS) antibiotic assay questionnaire (6). In the present case the patient was of advanced years with a history of ischemic heart disease and endocarditis, which together increased the risk of renal function decline. Altered renal blood flow may also play role in sudden GFR reduction which would cause an elevation in serum creatinine. Hypotension related low kidney perfusion could have also reduced GFR. The persistence of relatively high trough vancomycin level three days after withdrawal of the drug in our patient may be explained partly by drug leaching back from the peripheral tissue coupled with a decrease in renal elimination. Renal failure may also be a manifestation related to the severity of endocarditis, and this probably explains why it is also associated with worse prognosis (10). This fact also explains the deterioration in renal function with a large increase of serum creatinine in this case following the introduction of vancomycin. Furthermore some recent evidence indicates that high failure rates were observed in patients with MRSA bacteraemia treated with vancomycin, despite high vancomycin trough levels, where predictors of vancomycin failure included endocarditis (22). In our case, therapy with vancomycin for endocarditis proved effective even after significant dose reduction to accommodate the decreased renal elimination in this elderly patient. Therefore, TDM with appropriate dose reduction as predicted from the computer assisted model provided a better outcome with lower risk of toxicity. However, the patient’s age, gender, and disease with certain intervention as well as past disease history may dictate the dose adjustment by careful interpretation in contexts of the risk benefit rather than hanging only on the drug concentration alone. Therapeutic drug monitoring may minimize risk of toxicity or therapeutic failure, if timely provided. However, the interpretation of determined drug levels and its relationship to the adverse event or causality confirmation is a very difficult task. What does surprise us is that after a decrease the serum creatinine increased once again without any more doses of the potentially toxic suspected drug (vancomycin).

Conclusions

This case illustrates that elderly patients with pre-existing risk may be exposed to unpredictable renal function deterioration when vancomycin is prescribed. Therefore nephrotoxic antibiotics should be restricted in the elderly whenever possible. In cases where vancomycin is the drug of choice and there is no alternative it is vital to obtain serum levels immediately before the second dose (never later than the 5^th dose) and adjust the dosing regimen accordingly. Timely TDM as a clinical tool can ensure safer, more effective therapy which results in cost-savings and more favourable overall outcomes.

Abbreviations

• BMI – body mass index
• BP – blood pressure
• BUN – blood urea nitrogen
• CRP – C-reactive protein
• FPIA – fluorescent polarization immunoassay
• GFR – glomerular filtration rate
• MDRD – modification of diet in renal disease
• MI – myocardial infarction
• MICs – minimal inhibition concentrations
• MRSA – meticilli-resistant Staphylococcus aureus
• MW/PHARM – Madiware/Pharmacokinetics
• OAT – organic anion transporter
• TDM – therapeutic drug monitoring

The corresponding author expresses appreciations and thanks to co-authors Dr. Jirina Lukaskova from former KlinLab Division of Biochemistry, Prague, Czech Republic and Dr. Jana Horinkovoa from the Department of Internal Medicine Neratovice Hospital, Czech Republic for cooperation and useful information for data analysis and comments on interpretations. Overall, Dr. Dave Berry of St. Tomas Hospital, London, UK, deserves much appreciation for revising the manuscript both language wise and as an expert in the field of clinical toxicology.

ADRESSE FOR CORRESPONDENCE:

Hundie Tesfaye, MD, Ph.D.

Department of Medical Chemistry and Clinical Biochemistry Division of Clinical Pharmacology, Faculty Hospital in Motol, 2nd Faculty of Medicine, Charles University

V Úvalu 84, 150 06 Prague 5, Czech Republic

fax: +420 224 435 320, e-mail: hundie.tesfaye@fnmotol.cz