Chapter 54: Aminoglycosides

History and Source. Aminoglycosides are natural products or semisynthetic derivatives of compounds produced by a variety of soil actinomycetes. Streptomycin was first isolated from a strain of Streptomyces griseus. Gentamicin and netilmicin are derived from species of the actinomycete Micromonospora. The difference in spelling (-micin) compared with the other aminoglycoside antibiotics (-mycin) reflects this difference in origin. Tobramycin is one of several components of an aminoglycoside complex (nebramycin) that is produced by S. tenebrarius. It is most similar in antimicrobial activity and toxicity to gentamicin. In contrast to the other aminoglycosides, amikacin, a derivative of kanamycin, and netilmicin, a derivative of sisomicin, are semisynthetic products. Other aminoglycoside antibiotics have been developed (e.g., arbekacin, isepamicin, and sisomicin), but they have not been introduced into clinical practice in the U.S. because numerous potent, less toxic alternatives (e.g., broad-spectrum β-lactam antibiotics and quinolones) are available.

The aminoglycoside families are distinguished by the amino sugars attached to the aminocyclitol. In the neomycin family, which includes neomycin B and paromomycin, three amino sugars are attached to the central 2-deoxystreptamine. The kanamycin and gentamicin families have only two such amino sugars.

In the kanamycin family, which includes kanamycins A and B, amikacin, and tobramycin, two amino sugars are linked to a centrally located 2-deoxystreptamine moiety; one of these is a 3-aminohexose (Figure 54–1).

Amikacin is a semisynthetic derivative prepared from kanamycin A by acylation of the 1-amino group of the 2-deoxystreptamine moiety with 2-hydroxy-4-aminobutyric acid.

The gentamicin family, which includes gentamicins C₁, C_1a, and C₂, sisomicin, and netilmicin (the 1-N-ethyl derivative of sisomicin), contains a different 3-amino sugar (garosamine). Variations in methylation of the other amino sugar result in the different components of gentamicin (Figure 54–1). These modifications appear to have little effect on biological activity.

Streptomycin differs from the other aminoglycoside antibiotics in that it contains streptidine rather than 2-deoxystreptamine, and the aminocyclitol is not in a central position, as shown here.

The primary intracellular site of action of the aminoglycosides is the 30S ribosomal subunit, which consists of 21 proteins and a single 16S molecule of RNA. At least three of these ribosomal proteins, and perhaps the 16S ribosomal RNA as well, contribute to the streptomycin-binding site, and alterations of these molecules markedly affect the binding and subsequent action of streptomycin. For example, a single amino acid substitution of asparagine for lysine at position 42 of the S₁₂ ribosomal protein prevents binding of the drug; the resulting mutant is totally resistant to streptomycin. Substitution of glutamine for lysine creates a mutant that actually requires streptomycin for survival. The other aminoglycosides also bind to the 30S ribosomal subunit; however, they also appear to bind to several sites on the 50S ribosomal subunit (Davis, 1988).

Aminoglycosides disrupt the normal cycle of ribosomal function by interfering, at least in part, with the initiation of protein synthesis, leading to the accumulation of abnormal initiation complexes, or streptomycin monosomes, shown schematically in Figure 54–2B (Luzzatto et al., 1969). Aminoglycosides also cause misreading of the mRNA template and incorporation of incorrect amino acids into the growing polypeptide chains. Aminoglycosides vary in their capacity to cause misreading, presumably owing to differences in their affinities for specific ribosomal proteins. Although there appears to be a strong correlation between bactericidal activity and the ability to induce misreading (Hummel and Böck, 1989), it is not clear whether misreading is the primary mechanism of aminoglycoside-induced cell death.

The resulting aberrant proteins may be inserted into the cell membrane, leading to altered permeability and further stimulation of aminoglycoside transport (Busse et al., 1992). This phase of aminoglycoside transport, termed energy-dependent phase II (EDP₂), is poorly understood; however, EDP₂ may link to disruption of the structure of the cytoplasmic membrane, perhaps by the aberrant proteins. This concept is consistent with the observed progression of the leakage of small ions, followed by larger molecules and, eventually, by proteins from the bacterial cell prior to aminoglycoside-induced death. This progressive disruption of the cell envelope, as well as other vital cell processes, may help to explain the lethal action of aminoglycosides (Bryan, 1989).

Resistance owing to mutations that alter ribosomal structure and reduce aminoglycoside binding is relatively uncommon. Missense mutations in Escherichia coli that substitute a single amino acid in a crucial ribosomal protein may prevent binding of streptomycin. Although highly resistant to streptomycin, these strains are not widespread in nature. Similarly, only 5% of strains of Pseudomonas aeruginosa exhibit such ribosomal resistance to streptomycin. Resistance in ~50% of streptomycin-resistant strains of enterococci is attributable to ribosomal mutations (Eliopoulos et al., 1984). Because ribosomal resistance usually is specific for streptomycin, these strains of enterococci remain sensitive to a combination of penicillin and gentamicin in vitro.

Absorption. The aminoglycosides are highly polar cations and therefore are very poorly absorbed from the GI tract. Less than 1% of a dose is absorbed after either oral or rectal administration. The drugs are not inactivated in the intestine and are eliminated quantitatively in the feces. Long-term oral or rectal administration of aminoglycosides may result in accumulation to toxic concentrations in patients with renal impairment. Absorption of gentamicin from the GI tract may be increased by GI disease (e.g., ulcers or inflammatory bowel disease). Instillation of these drugs into body cavities with serosal surfaces also may result in rapid absorption and unexpected toxicity (i.e., neuromuscular blockade). Similarly, intoxication may occur when aminoglycosides are applied topically for long periods to large wounds, burns, or cutaneous ulcers, particularly if there is renal insufficiency.

All the aminoglycosides are absorbed rapidly from intramuscular sites of injection. Peak concentrations in plasma occur after 30-90 minutes and are similar to those observed 30 minutes after completion of an intravenous infusion of an equal dose over a 30-minute period. These concentrations typically range from 4-12 μg/mL following a 1.5-2 mg/kg dose of gentamicin, tobramycin, or netilmicin and from 20 to 35 μg/mL following a 7.5 mg/kg dose of amikacin or kanamycin. In critically ill patients, especially those in shock, absorption of drug may be reduced from intramuscular sites because of poor perfusion.

There is increasing use of aminoglycosides administered via inhalation, primarily for the management of patients with cystic fibrosis who have chronic Pseudomonas aeruginosa pulmonary infections. Amikacin and tobramycin solutions for injection have been used, as well as a commercial formulation of tobramycin designed for inhalation (TOBI, others). Studies of this formulation indicate that high sputum concentrations are obtained (mean of 1200 μg/g), but serum concentrations remain low (mean peak concentration of 0.95 μg/mL) (Geller et al., 2002).

Distribution. Because of their polar nature, the aminoglycosides do not penetrate into most cells, the CNS, or the eye. Except for streptomycin, there is negligible binding of aminoglycosides to plasma albumin. The apparent volume of distribution of these drugs is 25% of lean body weight and approximates the volume of extracellular fluid. The aminoglycosides distribute poorly into adipose tissue, which must be considered when using weight-based dosing regimens in obese patients. Approaches using ideal or adjusted body weight are recommended in conjunction with drug-level monitoring to avoid excessive serum concentrations.

Concentrations of aminoglycosides in secretions and tissues are low. High concentrations are found only in the renal cortex and the endolymph and perilymph of the inner ear; the high concentration in these sites likely contribute to the nephrotoxicity and ototoxicity caused by these drugs. As a result of active hepatic secretion, concentrations in bile approach 30% of those found in plasma, but this represents a very minor excretory route for the aminoglycosides. Penetration into respiratory secretions is poor (Panidis et al., 2005). Diffusion into pleural and synovial fluid is relatively slow, but concentrations that approximate those in the plasma may be achieved after repeated administration. Inflammation increases the penetration of aminoglycosides into peritoneal and pericardial cavities.

Concentrations of aminoglycosides achieved in CSF with parenteral administration usually are subtherapeutic. In patients, concentrations in CSF in the absence of inflammation are <10% of those in plasma; this value may approach 25% when there is meningitis (Kearney and Aweeka, 1999). Given the limit on dosage escalation due to the toxicity of aminoglycosides, treatment of meningitis with intravenous administration is generally suboptimal. Intrathecal or intraventricular administration of aminoglycosides has been used to achieve therapeutic levels, but the availability of third- and fourth-generation cephalosporins has made this unnecessary in most cases. Penetration of aminoglycosides into ocular fluids is so poor that effective therapy of bacterial endophthalmitis requires periocular and intraocular injections of the drugs.

Administration of aminoglycosides to women late in pregnancy may result in accumulation of drug in fetal plasma and amniotic fluid. Streptomycin and tobramycin can cause hearing loss in children born to women who receive the drug during pregnancy. Insufficient data are available regarding the other aminoglycosides; it is therefore recommended that they be used with caution during pregnancy and only for strong clinical indications in the absence of suitable alternatives.

Elimination. The aminoglycosides are excreted almost entirely by glomerular filtration, and urine concentrations of 50-200 μg/mL are achieved. A large fraction of a parenterally administered dose is excreted unchanged during the first 24 hours, with most of this appearing in the first 12 hours. The half-lives of the aminoglycosides in plasma are similar, 2-3 hours in patients with normal renal function. Renal clearance of aminoglycosides is approximately two-thirds of the simultaneous creatinine clearance; this observation suggests some tubular reabsorption of these drugs.

After a single dose of an aminoglycoside, disappearance from the plasma exceeds renal excretion by 10-20%; however, after 1-2 days of therapy, nearly 100% of subsequent doses eventually is recovered in the urine. This lag period probably represents saturation of binding sites in tissues. The rate of elimination of drug from these sites is considerably longer than from plasma; the t_1/2 for tissue-bound aminoglycoside has been estimated to range from 30 to 700 hours. For this reason, small amounts of aminoglycosides can be detected in the urine for 10-20 days after drug administration is discontinued. Aminoglycoside bound to renal tissue exhibits antibacterial activity and protects experimental animals against bacterial infections of the kidney even when the drug no longer can be detected in serum (Bergeron et al., 1982).

The concentration of aminoglycoside in plasma produced by the initial dose depends only on the volume of distribution of the drug. Because the elimination of aminoglycosides depends almost entirely on the kidney, a linear relationship exists between the concentration of creatinine in plasma and the t_1/2 of all aminoglycosides in patients with moderately compromised renal function. In anephric patients, the t_1/2 varies from 20-40 times that is determined in normal individuals. Because the incidence of nephrotoxicity and ototoxicity is likely related to the overall drug exposure to aminoglycosides, it is critical to reduce the maintenance dosage of these drugs in patients with impaired renal function.

Aminoglycosides can be removed from the body by either hemodialysis or peritoneal dialysis. Approximately 50% of the administered dose is removed in 12 hours by hemodialysis, which has been used for the treatment of overdosage. As a general rule, a dose equal to half the loading dose administered after each hemodialysis should maintain the plasma concentration in the desired range; however, a number of variables make this a rough approximation at best. Continuous arteriovenous hemofiltration (CAVH) and continuous venovenous hemofiltration (CVVH) will result in aminoglycoside clearances approximately equivalent to 15 and 15-30 mL/min of creatinine clearance, respectively, depending on the flow rate. The amount of aminoglycoside removed can be replaced by administering ~15-30% of the maximum daily dose (Table 54–2) each day. Frequent monitoring of plasma drug concentrations is again crucial.

Peritoneal dialysis is less effective than hemodialysis in removing aminoglycosides. Clearance rates are ~5-10 mL/min for the various drugs but are highly variable. If a patient who requires dialysis has bacterial peritonitis, a therapeutic concentration of the aminoglycoside probably will not be achieved in the peritoneal fluid because the ratio of the concentration in plasma to that in peritoneal fluid may be 10:1 (Smithivas et al., 1971). Thus, it is recommended that antibiotic be added to the dialysate to achieve concentrations equal to those desired in plasma.

Aminoglycosides can be inactivated by various penicillins in vitro and thus should not be admixed in solution. Some reports indicate that this inactivation may occur in vivo in patients with end-stage renal failure (Blair et al., 1982), thus making monitoring of aminoglycoside plasma concentrations even more necessary in such patients. Amikacin appears to be the aminoglycoside least affected by this interaction, and penicillins with more nonrenal elimination (such as piperacillin) may be less prone to cause this interaction.

Although excretion of aminoglycosides is similar in adults and children >6 months of age, half-lives of the drugs may be prolonged significantly in the newborn: 8-11 hours in the first week of life in newborns weighing <2 kg and ~5 hours in those weighing >2 kg (Yow, 1977). Thus, it is critically important to monitor plasma concentrations of aminoglycosides during treatment of neonates (Philips et al., 1982). For unknown reasons, aminoglycoside clearances are increased and half-lives are reduced in patients with cystic fibrosis compared to subjects without cystic fibrosis, after controlling for age and weight (Mann et al., 1985). Larger doses of aminoglycosides may likewise be required in burn patients because of more rapid drug clearance, possibly because of drug loss through burn tissue.

Dosing. Recommended doses of individual aminoglycosides in the treatment of specific infections are given in later sections of this chapter. Historically, aminoglycosides have been administered as two or three equally divided doses, based on the short t_1/2 of the drugs. However, studies of the pharmacokinetic/pharmacodynamic properties of aminoglycosides demonstrate that administering higher doses at extended intervals (typically once daily in patients with normal renal function) is likely to be at least equally efficacious and potentially less toxic than administration of divided doses. A comparison of this high-dose, extended-interval dosing method to traditional divided-dose methods is illustrated in Figure 54–3. Because of the post-antibiotic effect of aminoglycosides, good therapeutic response can be attained even when concentrations of aminoglycosides fall below inhibitory concentrations for a substantial fraction of the dosing interval. High-dose, extended-interval dosing schemes for aminoglycosides may also reduce the characteristic oto- and nephrotoxicity of these drugs. This diminished toxicity is probably due to a threshold effect from accumulation of drug in the inner ear or in the kidney. More drug accumulates with higher plasma concentrations, particularly at trough, and with prolonged periods of exposure. Net elimination of aminoglycoside from these organs occurs more slowly when plasma concentrations are relatively high. High-dose, extended-interval regimens, despite the higher peak concentration, provide a longer period when concentrations fall below the threshold for toxicity than does a multiple-dose regimen (12 hours versus <3 hours total in the example shown in Figure 54–3), potentially accounting for the lower toxicity of this approach.

Numerous studies and meta-analyses demonstrate that administration of the total dose once daily is associated with less nephrotoxicity and is just as effective as multiple-dose regimens (Bailey et al., 1997; Buijk et al., 2002). Extended-interval dosing also costs less and is administered more easily. For these reasons, high-dose, extended-interval administration of aminoglycosides is the preferred means of administering aminoglycosides for most indications and patient populations. Although the use of extended-interval dosing has been controversial in pregnancy, neonatal, and pediatric infections (Knoderer et al., 2003; Rastogi et al., 2002), data from meta-analyses now support this mode of administration in appropriately selected patients from these populations (Contopoulos-Ioannidis et al., 2004; Nestaas et al., 2005; Ward and Theiler, 2008). One key exception to the use of extended-interval dosing is for aminoglycoside use as combination therapy with a cell wall–active agent in the treatment of gram-positive infections, such as endocarditis. In these infections, administration of multiple daily doses (with a lower total daily dose) is preferred because data documenting equivalent safety and efficacy of extended-interval dosing are inadequate. Although schemes exist for adjusting dosages of aminoglycosides dosed by extended-interval methods in patients with significant renal dysfunction (i.e., creatinine clearance <25 mL/minute), some clinicians prefer to use the traditional multiple-dose regimen in such patients.

Nomograms may be helpful in selecting initial doses, but variability in aminoglycoside clearance among patients is too great for these to be relied on for more than a few days (Bartal et al., 2003). If it is anticipated that the patient will be treated with an aminoglycoside for >3-4 days, then plasma concentrations should be monitored to avoid drug accumulation. Whether extended-interval or multiple-daily dosing is chosen, the dose must be adjusted for patients with creatinine clearances of <80-100 mL/minute (Table 54–2), and plasma concentrations must be monitored. Determination of the concentration of drug in plasma is an essential guide to the proper administration of aminoglycosides. In patients with life-threatening systemic infections, aminoglycoside concentrations should be determined several times per week (more frequently if renal function is changing) and should be determined within 24-48 hours of a change in dosage. The size of the individual dose, the interval between doses, or both can be altered based on the results of monitoring of drug levels in plasma. Methods for calculation of dosage are described in Appendix II. There are obvious difficulties in using any of these approaches for ill patients with rapidly changing renal function. In addition, even when known factors are taken into consideration, concentrations of aminoglycosides achieved in plasma after a given dose vary widely among patients. If the extracellular volume is expanded, the volume of distribution is increased, and concentrations will be reduced.

For twice- or thrice-daily dosing regimens, both peak and trough plasma concentrations are determined. The trough sample is obtained just before a dose, and the peak sample is obtained 60 minutes after intramuscular injection or 30 minutes after an intravenous infusion given over 30 minutes. The peak concentration documents that the dose produces therapeutic concentrations, generally accepted to be 4-10 μg/mL for gentamicin, netilmicin, and tobramycin and 15-30 μg/mL for amikacin and streptomycin. The trough concentration is used to avoid toxicity by monitoring for accumulation of drug. Trough concentrations should be <1-2 μg/mL for gentamicin, netilmicin, and tobramycin and <10 μg/mL for amikacin and streptomycin.

Monitoring of aminoglycoside plasma concentrations also is important when using an extended-interval dosing regimen, although peak concentrations are not determined routinely (these will be three to four times higher than the peak achieved with a multiple-daily-dosing regimen). Several approaches may be used to determine that drug is being cleared and not accumulating.

The most accurate method for monitoring plasma levels for dose adjustment is to measure the concentration in two plasma samples drawn several hours apart (e.g., at 2 and 12 hours after a dose). The clearance then can be calculated and the dose adjusted to achieve the desired target range.

Another approach relies on nomograms to target a range of concentrations in a sample obtained earlier in the dosing interval. For example, if the plasma concentration from a sample obtained 8 hours after a dose of gentamicin is between 1.5 and 6 μg/mL, then the concentration at 18 hours will be <1 μg/mL (Chambers et al., 1998). Target ranges of 1-1.5 μg/mL for gentamicin at 18 hours for patients with creatinine clearances >50 mL/min and 1-2.5 μg/mL for those with clearances <50 mL/min also have been used. This method also tends to be inaccurate, particularly when conditions that alter aminoglycoside clearance are present (Bartal et al., 2003; Toschlog et al., 2003).

The simplest method is to obtain a trough sample 24 hours after dosing and adjust the dose to achieve the recommended plasma concentration, (e.g., <1-2 μg/mL in the case of gentamicin or tobramycin). This approach probably is the least desirable. An undetectable trough concentration could reflect grossly inadequate dosing in patients who clear the drug rapidly with prolonged periods (perhaps well over half the dosing interval) during which concentrations are subtherapeutic. In contrast, a 24-hour trough concentration target of 1-2 μg/mL actually would increase aminoglycoside exposure compared with a multiple-daily-dosing regimen (Barclay et al., 1999), which defeats the goal of providing a washout with concentrations of 0-1 μg/mL between 18 and 24 hours after a dose.

Accumulation within the perilymph and endolymph occurs predominantly when aminoglycoside concentrations in plasma are high. Diffusion back into the bloodstream is slow; the half-lives of the aminoglycosides are five to six times longer in the otic fluids than in plasma. Back diffusion is concentration dependent and facilitated at the trough concentration of drug in plasma Thus, it is generally thought that ototoxicity is more likely to occur in patients with persistently elevated concentrations of drug in plasma. However, studies have not consistently shown an association between ototoxicity and putative risk factors such as aminoglycoside serum levels, total dose and duration of aminoglycoside exposure, and renal dysfunction (including aminoglycoside-induced nephrotoxicity) (de Jager and van Altena, 2002; Peloquin et al., 2004). Drugs such as ethacrynic acid and furosemide potentiate the ototoxic effects of the aminoglycosides in animals, but data from humans implicating furosemide are less convincing (Moore et al., 1984). Hearing loss following exposure to these agents also is more likely to develop in patients with pre-existing auditory impairment.

The description of families where several members experienced profound deafness from as little as a single dose of aminoglycoside led to the investigation of a potential genetic predisposition for this toxicity. Subsequent studies have identified several predisposing mutations in mitochondrial ribosomal RNA genes (Fischel-Ghodsian, 2005). The contribution of genetic susceptibility to the overall incidence of aminoglycoside-induced ototoxicity is unclear, but surveys of subjects of European origin put the prevalence of one of these mutations at 1 in 500 (Vandebona et al., 2009).

Although all aminoglycosides can affect cochlear and vestibular function, there is some preferential toxicity. Streptomycin and gentamicin produce predominantly vestibular effects, whereas amikacin, kanamycin, and neomycin primarily affect auditory function; tobramycin affects both equally. The incidence of ototoxicity is difficult to determine. Audiometric data suggest that the incidence may be as high as 25% (de Jager and van Altena, 2002); however, these data need to be interpreted in the context of the limitations of audiometric testing in acutely ill patients and day-to-day variability of audiometric testing (Brummett and Morrison, 1990). The relative incidence appears to be equal for tobramycin, gentamicin, and amikacin. Initial studies suggested that netilmicin is less ototoxic than other aminoglycosides (Lerner et al., 1983); however, the incidence of ototoxicity from netilmicin is not negligible—such complications developed in 10% of patients in one clinical trial of netilmicin. The incidence of vestibular toxicity is particularly high in patients receiving streptomycin; nearly 20% of individuals who received 500 mg twice daily for 4 weeks for enterococcal endocarditis developed clinically detectable irreversible vestibular damage (Wilson et al., 1984). In addition, up to 75% of patients who received 2 g streptomycin for >60 days showed evidence of nystagmus or postural imbalance.

Because the initial symptoms may be reversible, patients receiving high doses and/or prolonged courses of aminoglycosides should be monitored carefully for ototoxicity; however, deafness may occur several weeks after therapy is discontinued. Screening patients for a family history of aminoglycoside-induced deafness seems reasonable, but the cost effectiveness of genetic screening for mutations predisposing to aminoglycoside ototoxicity is not yet clear.

Clinical Symptoms of Cochlear Toxicity. A high-pitched tinnitus often is the first symptom of toxicity. If the drug is not discontinued, auditory impairment may develop after a few days. The tinnitus may persist for several days to 2 weeks after therapy is stopped. Because perception of sound in the high-frequency range (outside the conversational range) is lost first, the affected individual is not always aware of the difficulty, and it will not be detected except by careful audiometric examination. If the hearing loss progresses, the lower sound ranges are affected, and conversation becomes difficult.

Clinical Symptoms of Vestibular Toxicity. Moderately intense headache lasting 1-2 days may precede the onset of labyrinthine dysfunction. This is followed immediately by an acute stage in which nausea, vomiting, and difficulty with equilibrium develop and persist for 1-2 weeks. Prominent symptoms include vertigo in the upright position, inability to perceive termination of movement ("mental past-pointing"), and difficulty in sitting or standing without visual cues. Drifting of the eyes at the end of a movement so that both focusing and reading are difficult, a positive Romberg test, and rarely, pendular trunk movement and spontaneous nystagmus, are prominent signs. The acute stage ends suddenly and is followed by the appearance of manifestations of chronic labyrinthitis, in which, although symptomless while in bed, the patient has difficulty when attempting to walk or make sudden movements; ataxia is the most prominent feature. The chronic phase persists for ~2 months; it is gradually superseded by a compensatory stage in which symptoms are latent and appear only when the eyes are closed. Adaptation to the impairment of labyrinthine function is accomplished by the use of visual cues and deep proprioceptive sensation for determining movement and position. It is more adequate in the young than in the old but may not be sufficient to permit the high degree of coordination required in many special trades. Recovery from this phase may require 12-18 months, and most patients have some permanent residual damage. Although there is no specific treatment for the vestibular deficiency, early discontinuation of the drug may permit recovery before irreversible damage of the hair cells.

Nephrotoxicity. Approximately 8-26% of patients who receive an aminoglycoside for several days develop mild renal impairment that is almost always reversible. The toxicity results from accumulation and retention of aminoglycoside in the proximal tubular cells (Lietman and Smith, 1983). The initial manifestation of damage at this site is excretion of enzymes of the renal tubular brush border (Banday et al., 2008). After several days, there is a defect in renal concentrating ability, mild proteinuria, and the appearance of hyaline and granular casts. The glomerular filtration rate is reduced after several additional days (Schentag et al., 1979). The non-oliguric phase of renal insufficiency is thought to be due to the effects of aminoglycosides on the distal portion of the nephron with a reduced sensitivity of the collecting-duct epithelium to endogenous antidiuretic hormone (Appel, 1982). Although severe acute tubular necrosis may occur rarely, the most common significant finding is a mild rise in plasma creatinine (5-20 μg/mL; 40-175 μM). Hypokalemia, hypocalcemia, and hypophosphatemia are seen very infrequently. The impairment in renal function is almost always reversible because the proximal tubular cells have the capacity to regenerate.

The biochemical events leading to tubular cell damage and glomerular dysfunction are poorly understood but may involve perturbations of the structure of cellular membranes. Aminoglycosides inhibit various phospholipases, sphingomyelinases, and ATPases, and they alter the function of mitochondria and ribosomes (Humes et al., 1984; Queener et al., 1983). Because of the ability of cationic aminoglycosides to interact with anionic phospholipids, these drugs may impair the synthesis of membrane-derived autacoids and intracellular second messengers such as prostaglandins, inositol phosphates, and diacylglycerol.

Several variables appear to influence nephrotoxicity from aminoglycosides. Toxicity correlates with the total amount of drug administered. Consequently, toxicity is more likely to be encountered with longer courses of therapy. Continuous infusion is more nephrotoxic in animals than intermittent dosing (Powell et al., 1983). High-dose, extended-interval dosing approaches lead to less nephrotoxicity at the same level of total drug exposure (as measured by area under the curve) than divided-dose approaches (Figure 54-3). Advanced age, liver disease, diabetes mellitus, and septic shock have been suggested as risk factors for the development of nephrotoxicity from aminoglycosides, but data are not convincing. Note, however, that renal function in the elderly patient is overestimated by measurement of creatinine concentration in plasma, and overdosing will occur if this value is used as the only guide in this patient population (Baciewicz et al., 2003).

The nephrotoxic potential varies among individual aminoglycosides. The relative toxicity correlates with the concentration of drug found in the renal cortex in experimental animals. Neomycin, which concentrates to the greatest degree, is highly nephrotoxic in human beings and should not be administered systemically. Streptomycin does not concentrate in the renal cortex and is the least nephrotoxic. Most of the controversy has concerned the relative toxicities of gentamicin and tobramycin. If differences between the renal toxicity of these two aminoglycosides do exist in human beings, they appear to be slight. Comparative studies with amikacin, sisomicin, and netilmicin are not conclusive. Other drugs, such as amphotericin B, vancomycin, angiotensin-converting enzyme inhibitors, cisplatin, and cyclosporine, may potentiate aminoglycoside-induced nephrotoxicity (Wood et al., 1986). Clinical studies have not proven conclusively that furosemide potentiates aminoglycoside nephrotoxicity (Smith and Lietman, 1983), but volume depletion and wasting of K⁺ that accompany the use of furosemide may predispose to aminoglycoside toxicity.

Neuromuscular Blockade. An unusual toxic reaction of acute neuromuscular blockade and apnea has been attributed to the aminoglycosides. The order of decreasing potency for blockade is neomycin, kanamycin, amikacin, gentamicin, and tobramycin. In humans, neuromuscular blockade generally has occurred after intrapleural or intraperitoneal instillation of large doses of an aminoglycoside; however, the reaction can follow intravenous, intramuscular, and even oral administration of these agents. Most episodes have occurred in association with anesthesia or the administration of other neuromuscular blocking agents. Patients with myasthenia gravis are particularly susceptible to neuromuscular blockade by aminoglycosides (Chapter 11).

Aminoglycosides may inhibit prejunctional release of acetylcholine while also reducing postsynaptic sensitivity to the transmitter, but Ca²⁺ can overcome this effect, and the intravenous administration of a calcium salt is the preferred treatment for this toxicity (Sarkar et al., 1992). Inhibitors of acetylcholinesterase (e.g., edrophonium and neostigmine) also have been used with varying degrees of success.

Other Untoward Effects. In general, the aminoglycosides have little allergenic potential; anaphylaxis and rash are unusual. Rare hypersensitivity reactions—including skin rashes, eosinophilia, fever, blood dyscrasias, angioedema, exfoliative dermatitis, stomatitis, and anaphylactic shock—have been reported as cross-hypersensitivity among drugs in this class. Parenterally administered aminoglycosides are not associated with pseudomembranous colitis, probably because they do not disrupt the normal anaerobic flora. Other reactions that have been attributed to individual drugs are discussed next.

The typical recommended intramuscular or intravenous dose of gentamicin sulfate when used for the treatment of known or suspected gram-negative organisms as a single agent or in combination therapy for adults with normal renal function is 5-7 mg/kg daily given over 30-60 minutes. For patients with renal dysfunction, the interval may be extended. For patients who are not candidates for extended-interval dosing, a typical dosing regimen for gram-negative coverage is a loading dose of 2 mg/kg and then 3-5 mg/kg per day, one-third given every 8 hours when administered as a multiple-daily-dosing regimen. Dosages at the upper end of this range may be required to achieve therapeutic levels for trauma or burn patients, those with septic shock, patients with cystic fibrosis, and others in whom drug clearance is more rapid or volume of distribution is larger than normal. Several dosage schedules have been suggested for newborns and infants: 3 mg/kg once daily for preterm newborns <35 weeks of gestation (Hansen et al., 2003; Rastogi et al., 2002); 4 mg/kg once daily for newborns >35 weeks of gestation; 5 mg/kg daily in two divided doses for neonates with severe infections; and 2-2.5 mg/kg every 8 hours for children up to 2 years of age. Peak plasma concentrations range from 4-10 mg/mL (dosing: 1.7 mg/kg every 8 hours) and 16-24 mg/ mL (dosing: 5.1 mg/kg once daily). It should be emphasized that the recommended doses of gentamicin do not always yield desired concentrations. Periodic determinations of the plasma concentration of aminoglycosides are recommended strongly, especially in seriously ill patients, to confirm that drug concentrations are in the desired range (see earlier sections on dosing for more details).

Urinary Tract Infections. Aminoglycosides usually are not indicated for the treatment of uncomplicated urinary tract infections, although a single intramuscular dose of gentamicin (5 mg/kg) has been effective in uncomplicated infections of the lower urinary tract. However, as strains of E. coli have acquired resistance to β-lactams, trimethoprim- sulfamethoxazole, and fluoroquinolones, use of aminoglycosides may increase. In the seriously ill patient with pyelonephritis, an aminoglycoside alone or in combination with a β-lactam antibiotic offers broad and effective initial coverage. Once the microorganism is isolated and its sensitivities to antibiotics are determined, the aminoglycoside should be discontinued if the infecting microorganism is sensitive to less toxic antibiotics.

Pneumonia. The organisms that cause community-acquired pneumonia are susceptible to broad-spectrum β-lactam antibiotics, macrolides, or a fluoroquinolone, and usually it is not necessary to add an aminoglycoside. Therapy with an aminoglycoside alone is likely to be ineffective; therapeutic concentrations are difficult to achieve owing to relatively poor penetration of drug into inflamed tissues and the associated conditions of low O₂ tension and low pH—both of which interfere with aminoglycoside antibacterial activity. Aminoglycosides are ineffective for the treatment of pneumonia due to anaerobes or S. pneumoniae, which are common causes of community-acquired pneumonia. They should not be considered as effective single-drug therapy for any aerobic gram-positive cocci (including S. aureus or streptococci), the microorganisms commonly responsible for suppurative pneumonia or lung abscess. An aminoglycoside in combination with a β-lactam antibiotic is recommended as standard therapy for hospital-acquired pneumonia in which a multiple-drug-resistant gram-negative aerobe is a likely causative agent (American Thoracic Society, 2005). Once it is established that the β-lactam is active against the causative agent, there is generally no benefit from continuing the aminoglycoside. Patients presenting with pulmonary exacerbations of cystic fibrosis often receive aminoglycosides as a component of therapy. Due to the altered pharmacokinetics of aminoglycosides in cystic fibrosis patients, higher daily doses (up to 10 mg/kg/day) may be necessary.

Meningitis. Availability of third-generation cephalosporins, especially cefotaxime and ceftriaxone, has reduced the need for treatment with aminoglycosides in most cases of meningitis, except for infections caused by gram-negative organisms resistant to β-lactam antibiotics (e.g., species of Pseudomonas and Acinetobacter). If therapy with an aminoglycoside is necessary, in adults, 5 mg of a preservative-free formulation of gentamicin (or equivalent dose of another aminoglycoside) is administered directly intrathecally or intraventricularly once daily (Barnes et al., 2003).

Peritonitis Associated with Peritoneal Dialysis. Patients who develop peritonitis as a result of peritoneal dialysis may be treated with aminoglycoside diluted into the dialysis fluid to a concentration of 4-8 mg/L for gentamicin, netilmicin, or tobramycin or 6-12 mg/L for amikacin. Intravenous or intramuscular administration of drug is unnecessary because serum and peritoneal fluid will equilibrate rapidly.

Bacterial Endocarditis. "Synergistic" or low-dose gentamicin (3 mg/kg per day in three divided doses) in combination with a penicillin or vancomycin has been recommended in certain circumstances for treatment of infections due to gram-positive organisms, primarily bacterial endocarditis. Penicillin and gentamicin in combination are effective as a short-course (i.e., 2-week) regimen for uncomplicated native-valve streptococcal endocarditis. In cases of enterococcal endocarditis, concomitant administration of penicillin and gentamicin for 4-6 weeks has been recommended because of an unacceptably high relapse rate with penicillin alone. A large case series from Sweden, however, found that cure rates were not substantially affected by shortening the duration of aminoglycoside therapy to a median of 15 days (Olaison and Schadewitz, 2002). A 2-week regimen of gentamicin or tobramycin in combination with nafcillin is effective for the treatment of selected cases of staphylococcal tricuspid native-valve endocarditis in intravenous drug users (Chambers et al., 1988), although the need to include the aminoglycoside is not established (Le and Bayer, 2003). For patients with native mitral or aortic valve staphylococcal endocarditis, the risks of aminoglycoside administration likely outweigh the benefits (Cosgrove et al., 2009). Administration of an aminoglycoside in combination with a cell wall–active agent and rifampin is recommended for treatment of staphylococcal prosthetic valve endocarditis, but clinical studies supporting this practice are limited.

Sepsis. Inclusion of an aminoglycoside in an empirical regimen is commonly recommended for the febrile patient with granulocytopenia and for sepsis when P. aeruginosa is a potential pathogen. Most studies that demonstrate a benefit for combination aminoglycoside therapy in these infections compared weak β-lactams (e.g., carbenicillin, no longer marketed in the U.S.) alone versus these agents plus an aminoglycoside (Paul et al., 2003, 2004). More recent studies using potent broad-spectrum β-lactams (e.g., carbapenems and anti-pseudomonal cephalosporins) have demonstrated no benefit from adding an aminoglycoside to the regimen. If there is concern that an infection may be caused by a multiple-drug-resistant organism that may be susceptible only to an aminoglycoside, then adding this antibiotic to the regimen is reasonable. Evidence that aminoglycosides are beneficial for other gram-negative infections is weak if the isolate is susceptible to other antibiotics. To avoid toxicity, aminoglycosides should be used briefly and sparingly as long as other alternatives are available.

Topical Applications. Gentamicin is absorbed slowly when it is applied topically in an ointment and somewhat more rapidly when it is applied as a cream. When the antibiotic is applied to large areas of denuded body surface, as may be the case in burned patients, plasma concentrations can reach 4 μg/mL, and 2-5% of the drug used may appear in the urine.

Untoward Effects. Like other aminoglycosides, the most important and serious side effects of the use of gentamicin are nephrotoxicity and irreversible ototoxicity. Intrathecal or intraventricular administration is used rarely because it may cause local inflammation and can result in radiculitis and other complications.

Therapeutic Uses. Indications for the use of tobramycin are the same as those for gentamicin. The superior activity of tobramycin against P. aeruginosa makes it the preferred aminoglycoside for treatment of serious infections known or suspected to be caused by this organism; the drug has been usefully administered by inhalation to combat P. aeruginosa infections (LoBue, 2005). Tobramycin usually is used with an anti-pseudomonal β-lactam antibiotic. In contrast to gentamicin, tobramycin shows poor activity in combination with penicillin against many strains of enterococci. Most strains of E. faecium are highly resistant. Tobramycin is ineffective against mycobacteria. Dosages and serum concentrations are identical with those for gentamicin.

Untoward Effects. Tobramycin, like other aminoglycosides, causes nephrotoxicity and ototoxicity. Studies in experimental animals suggest that tobramycin may be less toxic to hair cells in the cochlear and vestibular end organs and cause less renal tubular damage than gentamicin. However, clinical data are less convincing.

Therapeutic Uses. Amikacin is the preferred agent for the initial treatment of serious nosocomial gram-negative bacillary infections in hospitals where resistance to gentamicin and tobramycin has become a significant problem. Amikacin is active against the vast majority of aerobic gram-negative bacilli in the community and the hospital. This includes most strains of Serratia, Proteus, and P. aeruginosa. It is active against nearly all strains of Klebsiella, Enterobacter, and E. coli that are resistant to gentamicin and tobramycin. Most resistance to amikacin is found among strains of Acinetobacter, Providencia, and Flavobacter and strains of Pseudomonas other than P. aeruginosa; these all are unusual pathogens. Like tobramycin, amikacin is less active than gentamicin against enterococci and should not be used for this organism. Amikacin is not active against the majority of gram-positive anaerobic bacteria. It is active against M. tuberculosis (99% of strains inhibited by 4 μg/mL), including streptomycin-resistant strains and atypical mycobacteria. It has been used in the treatment of disseminated atypical mycobacterial infections in patients with acquired immunodeficiency syndrome (AIDS).

The recommended dose of amikacin is 15 mg/kg per day as a single daily dose or divided into two or three equal portions, which must be reduced for patients with renal failure. The drug is absorbed rapidly after intramuscular injection, and peak concentrations in plasma approximate 20 μg/mL after injection of 7.5 mg/kg. An intravenous infusion of the same dose over a 30-minute period produces a peak concentration in plasma of nearly 40 μg/mL at the end of the infusion, which falls to ~20 μg/mL 30 minutes later. The concentration 12 hours after a 7.5 mg/kg dose typically is 5-10 μg/mL. A 15 mg/kg once-daily dose produces peak concentrations 50-60 μg/mL and a trough of <1 μg/mL. For treatment of mycobacterial infections, thrice-weekly dosing schedules of amikacin are often used, with doses up to 25 mg/kg.

Untoward Effects. As with the other aminoglycosides, amikacin causes ototoxicity, hearing loss, and nephrotoxicity.

Therapeutic Uses. Netilmicin is useful for the treatment of serious infections owing to susceptible Enterobacteriaceae and other aerobic gram-negative bacilli. It is effective against certain gentamicin-resistant pathogens, with the exception of enterococci (Panwalker et al., 1978).

The recommended dose of netilmicin for complicated urinary tract infections in adults is 1.5-2 mg/kg every 12 hours. For other serious systemic infections, a total daily dose of 4-7 mg/kg is administered as a single dose or two to three divided doses. Children should receive 3-7 mg/kg per day in two to three divided doses; neonates receive 3.5-5 mg/kg per day as a single daily dose (Gosden et al., 2001). The distribution and elimination of netilmicin, gentamicin, and tobramycin are very similar. The t_1/2 for elimination is usually 2–2.5 hours in adults and increases with renal insufficiency.

Untoward Effects. As with other aminoglycosides, netilmicin also may produce ototoxicity and nephrotoxicity. Animal models suggest that netilmicin may be less toxic than other aminoglycosides, but this remains to be proven in humans (Tange et al., 1995).

Therapeutic Uses. Bacterial Endocarditis. Streptomycin and penicillin in combination are synergistically bactericidal in vitro and in animal models of infection against strains of enterococci, group D streptococci, and the various oral streptococci of the viridans group. The combination of penicillin G, which by itself is only bacteriostatic against enterococci, and streptomycin is effective as bactericidal therapy for enterococcal endocarditis. Gentamicin is generally preferred for its lesser toxicity; also, gentamicin should be used when the strain of enterococcus is resistant to streptomycin (MIC >2 mg/mL). Streptomycin should be used instead of gentamicin when the strain is resistant to the latter and susceptibility has been demonstrated to streptomycin, which may occur because the enzymes that inactivate these two aminoglycosides are different.

Streptomycin may be administered by deep intramuscular injection or intravenously. Intramuscular injection may be painful, with a hot tender mass developing at the site of injection. The dose of streptomycin is 15 mg/kg per day for patients with creatinine clearances >80 mL/minute. It typically is administered as a 1000-mg single daily dose for tuberculosis or 500 mg twice daily, resulting in peak serum concentrations of ~50-60 and 15-30 μg/mL and trough concentrations of <1 and 5-10 μg/mL, respectively. The total daily dose should be reduced in direct proportion to the reduction in creatinine clearance for creatinine clearances >30 mL/minute (Table 54–2).

Tularemia. Streptomycin (or gentamicin) is the drug of choice for the treatment of tularemia. Most cases respond to the administration of 1-2 g (15-25 mg/kg) streptomycin per day (in divided doses) for 10-14 days. Fluoroquinolones and tetracyclines also are effective, although the failure rate may be higher with tetracyclines.

Plague. Streptomycin is effective agent for the treatment of all forms of plague. The recommended dose is 2 g/day in two divided doses for 10 days. Gentamicin is probably as efficacious (Boulanger et al., 2004).

Tuberculosis. Streptomycin is a second-line agent for the treatment of active tuberculosis, and streptomycin always should be used in combination with at least one or two other drugs to which the causative strain is susceptible. The dose for patients with normal renal function is 15 mg/kg per day as a single intramuscular injection for 2-3 months and then two or three times a week thereafter.

Untoward Effects. Streptomycin has been replaced by gentamicin for most indications because the toxicity of gentamicin is primarily renal and reversible, whereas that of streptomycin is vestibular and irreversible. The administration of streptomycin may produce dysfunction of the optic nerve, including scotomas, presenting as enlargement of the blind spot. Among the less common toxic reactions to streptomycin is peripheral neuritis. This may be due either to accidental injection of a nerve during the course of parenteral therapy or to toxicity involving nerves remote from the site of antibiotic administration.

Therapeutic Uses. Kanamycin is all but obsolete, and there are few indications for its use. Kanamycin has been employed to treat tuberculosis in combination with other effective drugs. It has no therapeutic advantage over streptomycin or amikacin and probably is more toxic; either should be used instead, depending on susceptibility of the isolate.

Prophylactic Uses. Kanamycin can be administered orally as adjunctive therapy in cases of hepatic encephalopathy. The dose is 4-6 g/day for 36-72 hours; quantities as large as 12 g/day (in divided doses) have been given.

Untoward Effects. Kanamycin is ototoxic and nephrotoxic. Like neomycin, its oral administration can cause malabsorption and superinfection. The untoward effects of the oral administration of aminoglycosides are considered in the section on neomycin.

Therapeutic Uses. Neomycin has been used widely for topical application in a variety of infections of the skin and mucous membranes caused by microorganisms susceptible to the drug. These include infections associated with burns, wounds, ulcers, and infected dermatoses. However, such treatment does not eradicate bacteria from the lesions.

The oral administration of neomycin (usually in combination with erythromycin base) has been employed primarily for "preparation" of the bowel for surgery. For therapy of hepatic encephalopathy, an oral daily dose of 4-12 g (in divided doses) is given, provided that renal function is normal. Because renal insufficiency is a complication of hepatic failure and neomycin is nephrotoxic, it is used rarely for this indication.

Neomycin and polymyxin B have been used for irrigation of the bladder. For this purpose, 1 mL of a preparation (neosporin g.u. irrigant) containing 40 mg neomycin and 200,000 units polymyxin B per milliliter is diluted in 1 L of 0.9% sodium chloride solution and is used for continuous irrigation of the urinary bladder through appropriate catheter systems. The goal is to prevent bacteriuria and bacteremia associated with indwelling catheters. The bladder is irrigated at the rate of 1 L every 24 hours.

Absorption and Excretion. Neomycin is poorly absorbed from the GI tract and is excreted by the kidney, as are the other aminoglycosides. An oral dose of 3 g produces a peak plasma concentration of 1-4 μg/mL; a total daily intake of 10 g for 3 days yields a blood concentration below that associated with systemic toxicity if renal function is normal. Patients with renal insufficiency may accumulate the drug. About 97% of an oral dose of neomycin is not absorbed and is eliminated unchanged in the feces.

Untoward Effects. Hypersensitivity reactions, primarily skin rashes, occur in 6-8% of patients when neomycin is applied topically. Individuals sensitive to this agent may develop cross-reactions when exposed to other aminoglycosides. The most important toxic effects of neomycin are renal damage and nerve deafness; as a consequence, the drug is no longer available for parenteral administration. Toxicity has been reported in patients with normal renal function after topical application or irrigation of wounds with 0.5% neomycin solution. Neuromuscular blockade with respiratory paralysis also has occurred after irrigation of wounds or serosal cavities. Individuals treated with 4-6 g/day of the drug by mouth sometimes develop a sprue-like syndrome with diarrhea, steatorrhea, and azotorrhea. Overgrowth of yeasts in the intestine also may occur; this is not associated with diarrhea or other symptoms in most cases.