Plasma Concentrations of Amikacin During Clinical Use in Kemp’s Ridley Sea Turtles (Lepidochelys kempii)

Introduction

Sea turtle populations are globally at risk due to fisheries interactions, habitat loss and alteration, watercraft trauma, pollution, climate change, and disease (Fuentes et al., 2023). Many populations are categorized as endangered or threatened by international conservation organizations and national agencies; therefore, substantial effort is put forth to rehabilitate injured or ill sea turtles (Innis et al., 2019). In Massachusetts, USA, hundreds of juvenile Kemp’s ridley sea turtles (Lepidochelys kempii) are hospitalized annually after being found stranded secondary to cold stunning, a hypothermia-like condition that causes substantial morbidity and mortality (Wyneken et al., 2006; Innis et al., 2009, 2014; Powell et al., 2021).

Infections caused by gram-negative bacteria, including pneumonia, sepsis, and osteomyelitis, are common in debilitated sea turtles (e.g., Cicarelli et al., 2020; Powell et al., 2021; Glassman and Zachariah, 2024); therefore, fluoroquinolone, extended spectrum beta-lactam, and aminoglycoside antibiotics are often prescribed. Pharmacologic data for enrofloxacin, ciprofloxacin, marbofloxacin, danofloxacin, ceftazidime, ceftriaxone, and ticarcillin have been reported for several sea turtle species (Stamper et al., 1999; Jacobson et al., 2005; Manire et al., 2005; Marín et al., 2008, 2009; Lai et al., 2009; Innis et al., 2011; Nardini et al., 2015; Mapongpeng et al., 2019; Poapolathep et al., 2020, 2021; Wanmad et al., 2022); however, there are no peer-reviewed aminoglycoside studies to date.

Amikacin is a broad-spectrum aminoglycoside that is commonly used to treat gram-negative bacterial infections. Aminoglycosides are bactericidal and concentration dependent and must be administered parenterally due to poor gastrointestinal absorption. Bioavailability of amikacin after injection by the IM route is >90% in some species (e.g., El-Gamma et al., 1992). Amikacin is eliminated via renal excretion and is potentially nephrotoxic (Yamada et al., 2021). Studies investigating the pharmacokinetics or measurement of plasma concentrations of amikacin have been performed for several reptile species, including tortoises and freshwater turtles (Caligiuri et al., 1990; Hiebert et al., 2024), snakes (Mader et al., 1985; Johnson et al., 1997; Sykes et al., 2006; Wijayanti et al., 2015), and alligators (Jacobson et al., 1988), but not sea turtles. Nonetheless, amikacin is used in sea turtles based on dosing protocols for other reptile species. For example, amikacin was used with some clinical success and no overt evidence of nephrotoxicity in Kemp’s ridley turtles at 5–10 mg/kg IV or IM every 3 days (q3d) for management of osteomyelitis, sepsis, and pneumonia (Innis et al., 2014; Powell et al., 2021).

Given its apparent clinical efficacy, safety, and ongoing use, the present study was conducted to assess plasma concentrations of amikacin after initial IM injection during clinical use in hospitalized Kemp’s ridley sea turtles. In addition, this study documented plasma concentrations of amikacin after the last of multiple doses to evaluate the potential for drug accumulation over time. We hypothesized that amikacin would be rapidly absorbed after IM injection and that the drug would not accumulate in the plasma after multiple doses. We also hypothesized that IM injection of high concentration amikacin in Kemp’s ridley turtles would achieve plasma levels consistent with plasma maximal plasma concentration (C_max): minimum inhibitory concentration (MIC) ≥ 8, which is associated with greater clinical efficacy in human patients.

Materials and Methods

During the time period of this study, pharmacologic research for sea turtles at New England Aquarium (NEAq; Boston, MA, USA) was conditionally authorized by U.S. Fish and Wildlife Service (USFWS) permit 69328D, Section VIID. Written approval of the present study was provided by the USFWS recovery permits coordinator, Division of Endangered Species, North Atlantic–Appalachian Region. This project was also approved by the NEAq Institutional Animal Care and Use Committee (proposal 2020-08).

Juvenile Kemp’s ridley turtles were admitted to NEAq during 2020 and 2021 after stranding secondary to cold stunning. Turtles were assessed and medically managed as previously described, including serial physical examination, radiographs, clinical pathologic analyses, gradual warming, cardiorespiratory support, fluid therapy, antimicrobials (initial treatment with oxytetracycline 42 mg/kg SC q6d; Innis et al., 2020), nutritional support, and additional treatments based on individual status (Wyneken et al., 2006; Innis and Staggs, 2017). After stabilization, turtles were maintained in pools with natural salt water, mechanical and biological filtration, and ozone disinfection. To complete the present study, amikacin was prescribed for individual turtles by the attending veterinarians based on clinical observations such as persistent radiographic evidence of pneumonia despite treatment with oxytetracycline, hyporexia, lethargy, and/or culture and susceptibility data. To avoid potential effects of initial dehydration, decreased glomerular filtration rate (Innis et al., 2016), and temperature-dependent variability of amikacin pharmacokinetics (Mader et al., 1985; Caligiuri et al., 1990; Johnson et al., 1997), only turtles that were hospitalized for at least 1 wk and maintained at 24–25°C (75–77°F) were included in this study.

Amikacin sulfate (250 mg/ml; Sagent Pharmaceuticals, Schaumburg, IL, USA) was administered at 5 mg/kg IM in the pectoral muscle q3d for as long as deemed necessary by the attending veterinarian, with volume rounded to the nearest 0.01 ml. This dosage was selected based on previous clinical experience, previous reptile amikacin pharmacokinetic studies, target maximal plasma concentration (C_max) of 25 µg/ml, estimated volume of distribution (V) of 0.2 L/kg, and MIC data for previous gram-negative bacterial isolates derived from NEAq Kemp’s ridley turtle patients (45/52 [87%] of isolates between 2017 and 2020 had amikacin MIC ≤ 2 µg/ml [NEAq, unpublished data]), and 2 µg/ml is in the “susceptible” range for bacteria isolated from other animals based on the current Clinical and Laboratory Standards Institute (CLSI) laboratory standards for testing (CLSI, 2024). Based on clinician discretion, during amikacin treatment, some individual turtles were concurrently treated with parenteral and/or enteral fluid therapy and nutritional support (see Results).

Upon prescription of amikacin, turtles were assigned to one of three sampling groups, with five turtles in each group. An additional turtle was enrolled due to death of one individual before acquisition of the desired sample suite. A sparse-sampling protocol was used to limit handling and volume of blood obtained from each animal. Samples were obtained at the following time points for each group: group 1: 0.5, 4, and 24 h after first dose and 24 h after last dose (post last dose [PLD]); group 2: 1, 6, and 48 h after first dose and 48 h PLD; and group 3: 2, 8, and 72 h after first dose and then 72 h PLD. For each sample, approximately 1.0 ml of blood was collected into a preheparinized syringe from the external jugular vein, alternating between right and left sides for each collection. Blood was placed into a lithium heparin green top tube and centrifuged using a LW Scientific ZIPcombo centrifuge (Jorgensen Laboratories, Loveland, CO, USA). The tubes were spun for 90 sec at 12,000 rpm. Plasma was harvested and stored frozen at −80°C (−112°F) until transfer to the analytical laboratory. Once all samples were collected, they were shipped overnight on dry ice to the University of Florida’s Infectious Disease Pharmacokinetics Laboratory (Gainesville, FL, USA) for analysis of plasma amikacin concentrations.

Amikacin was measured using an ultrahigh-performance liquid chromatography assay with a Thermo Endura triple-quadrupole mass spectrometry system (Thermo Fisher Scientific, Waltham, MA, USA). The assay was validated in accordance with the Guidance for Industry: Bioanalytical Method Validation (U.S. Department of Health and Human Services, 2001). The standard curve range was 1.00–100.00 µg/ml. The overall precision of validation quality control samples was 3.46–5.15% (coefficient of variation).

Normality of data distribution for turtle body weight, day of amikacin initiation, number of doses administered, and volume of each dose administered were assessed using the Shapiro–Wilk test, with P ≤ 0.05 considered significant (R 4.3.0; R Core Team, 2020).

Results

Turtle weights, number of doses administered, and volumes administered were normally distributed (P = 0.62, 0.98, and 0.41, respectively), whereas day of amikacin initiation was not (P < 0.001). Turtles weighed 2.1–5.1 kg (mean, 3.3 kg). Amikacin was initiated between day 7 and day 144 of hospitalization (median day, 27) and was administered for 2–26 doses (mean, 13 doses), resulting in a mean treatment duration of approximately 5–6 wk. Prescribed drug volumes ranged from 0.04 to 0.1 ml (mean, 0.07 ml), resulting in individual turtle theoretical doses of 4.5–5.1 mg/kg assuming exact volumes of administration (mean, 4.9 mg/kg). Two turtles were treated with amikacin based on the results of blood or bone culture, respectively. Other turtles were treated based on poor clinical condition (e.g., persistent anorexia, lethargy) and radiographic evidence of persistent pneumonia despite initial oxytetracycline treatment, two of which had a negative blood culture. Six turtles were treated with parenteral fluid therapy or enteral fluid and nutritional therapy during initiation of amikacin treatment, including subcutaneous lactated Ringer’s solution at 10 ml/kg q3d (n = 3, two of which also received enteral fish gruel q3d) or 20 ml/kg q3d (n = 2, one of which also received enteral fish gruel q3d), whereas one individual received only enteral fish gruel q3d. Parenteral and enteral fluid and nutritional support were typically discontinued before discontinuation of amikacin treatment due to general clinical improvement and voluntary food acceptance.

All scheduled samples were acquired except for two 24 h PLD (21-0648 died during treatment and 20-0946 was sampled up to 24 h after first dose for therapeutic drug monitoring, before initiation of full study design) and one 48 h PLD sample (21-0371 died during treatment). Plasma amikacin concentrations for individual turtles at each time point after initial injection are provided in Table 1 and Fig. 1.

View Figure

• Download as Powerpoint
• Download Figure

Table 1.Plasma amikacin concentrations (µg/ml) for 16 Kemp’s ridley turtles (Lepidochelys kempii) after administration of 5 mg/kg IM. Values in italics are below the assay low standard limit (1.0 µg/ml).

View Fullscreen

Thirteen of 16 turtles were successfully treated and were released to the wild. Two turtles died during amikacin treatment after receiving two and four doses, respectively. The former turtle (21-0648) was not evaluated histologically, but gross necropsy and radiography were consistent with severe pneumonia. Histopathologic evaluation of the latter turtle (21-0371) revealed disseminated bacterial and fungal infections. One turtle (21-0426) was euthanized 4 months after ending amikacin treatment. It had received 12 doses of amikacin before additional medical management, and histopathologic evaluation revealed encephalomalacia and chronic pneumonia with intralesional acid-fast bacilli. No renal lesions typical of aminoglycoside toxicity were observed in the two cases evaluated histologically, and cause of death in these cases was determined to be unrelated to amikacin administration.

Discussion

This study was planned with the purpose of collecting data for a pharmacokinetic analysis of amikacin in these turtles. However, upon review of the data, we observed that the plasma amikacin concentrations in this study were highly variable among turtles. Although several pharmacokinetic models were explored, the high between-subject variability (interindividual variability) precluded robust analysis. As a result, observed plasma concentration data are provided without detailed pharmacokinetic analyses. Nevertheless, these data provide useful clinical insight and a foundation on which future studies may be designed.

Half of the turtles had highest measured amikacin concentrations between 21 and 37 µg/ml. The C_max of amikacin is important because C_max:MIC ratios of 8–10 are associated with greater clinical efficacy in human patients; yet, even a ratio of 2 resulted in favorable treatment response in approximately 55% of human patients (Moore et al., 1987). Although such detailed studies have not been conducted for reptiles, a peak concentration corresponding to clinical efficacy may have been achieved in some of these turtles. Given our historical data commonly showing amikacin MIC ≤ 2 µg/ml for Kemp’s ridley turtle bacterial isolates, the amikacin dose used in this study may provide therapeutic plasma concentrations in many cases. The highest measured amikacin concentration occurred at the first sampling point for the majority of turtles, including three of five turtles sampled at 0.5 h. Thus, it is possible that actual C_max in some cases occurred before 0.5 h or between the first and second time points. Future studies of amikacin in this species should therefore include a greater number of earlier time points. For four turtles, the highest measured amikacin concentration occurred at the second time point, suggesting delayed absorption. Three turtles in this study (one of which died) had relatively low plasma amikacin concentrations (≤6 µg/ml) at every time point at which they were assessed, which may be subtherapeutic except for bacteria with very low MIC. Higher amikacin doses may be needed in some Kemp’s ridley turtle patients and doses up to 10 mg/kg IM or IV q3d have been used in some cases with no apparent adverse effects (Innis et al., 2014; Powell et al., 2021).

In addition to targeting a C_max to reach a therapeutic concentration, the trough concentrations (C_min) of amikacin are also important to avoid kidney injury. Amikacin treatment strategies in people often use a single daily dose to achieve high C_max, while also achieving C_min < 5–10 µg/ml to avoid nephrotoxicity (Duszynska et al., 2013; Jenkins et al., 2016; Logre et al., 2020), with the most recent meta-analysis supporting C_min < 10 µg/ml (Yamada et al., 2021). Although we do not know whether these targets are appropriate for turtles, in this study, all 19 samples collected at 48 or 72 h postinjection had plasma amikacin concentrations <10 µg/ml and 13 of them had concentrations <5 µg/ml.

Plasma amikacin concentrations in humans may be affected by the severity and type of illness, renal function, and other variables (Moore et al., 1987; Duszynska et al., 2013). This study was conducted in clinically ill patients that had already been managed for days to weeks with other therapies, yet clinicians remained concerned enough about their clinical status to prescribe amikacin. The extent of illness of turtles in this study may explain the high variability in amikacin plasma concentrations.

The concern for aminoglycoside-induced nephrotoxicity is widespread in herpetological medical education and practice despite quite limited evidence. Although there are reports of nephrotoxicity after administration of high doses of gentamicin in snakes (Montali et al., 1979), to the authors knowledge there are no published peer-reviewed reports of nephrotoxicity due to amikacin in any reptile species despite widespread use. In the present study, and in the authors’ clinical experience, evidence of nephrotoxicity associated with amikacin administration has not been observed in Kemp’s ridley turtles. Plasma uric acid concentration is considered the best routine biochemical measure of renal function. In Kemp’s ridley turtles, specifically, plasma uric acid correlates negatively with glomerular filtration rate, and positively with the presence of renal disease (Kennedy et al., 2012; Innis et al., 2016). Assessment of renal function was not part of the prospective design of the present study. Inclusion of serial plasma biochemical profiles and measurement of glomerular filtration rates (e.g., iohexol clearance) are recommended for future reptile aminoglycoside studies, when possible.

Another potential concern during aminoglycoside therapy is ototoxicity, including auditory and vestibular toxicity (Jenkins et al., 2016). Hearing is difficult to evaluate in turtles, generally requiring measurement of auditory evoked potential (Harms et al., 2009). Such measurements were beyond the scope of this project but could be considered for future studies. Diagnosis of vestibular dysfunction in sea turtles is not well established, but signs as seen in other species such as nystagmus, circling, and postural anomalies (Chrisman et al., 1997) were not seen in this study.

Recent antibiotic stewardship recommendations for reptiles support the use of “first line” or “tier 1” antimicrobial drugs to limit antimicrobial resistance unless susceptibility data indicate the need for other drugs (Hedley et al., 2021; Divers and Burgess, 2023). Although those recommendations include aminoglycosides as tier 1 drugs, the World Health Organization (WHO) lists aminoglycosides as “Critically Important Antimicrobials,” which is among the higher categories for antimicrobials that can be used for nonhuman patients, second only to “Highest Priority Critically Important Antimicrobials” (WHO, 2024). As such, the use of amikacin for sea turtle patients should be carefully considered and drugs of lower WHO categories should be used preferentially when possible (e.g., “Highly Important Antimicrobials” such as oxytetracycline). In the present study, clinicians selected amikacin based on culture results for two cases. Blood cultures were pursued before treatment for two other persistent pneumonia cases, but were negative. For other cases, clinicians prescribed amikacin to enhance gram-negative bacterial coverage based on historic microbiologic and antimicrobial susceptibility data for this species at this facility. Ideally, higher category antibiotics should not be prescribed without culture and susceptibility results for each patient. Cultures of blood, tracheal lavage, fine needle aspirate, bronchoscopic alveolar lavage, and tissue biopsies should be pursued when circumstances permit (Ciccarelli et al., 2020). Engineering controls such as ozonation, chemical scavenging, and chemical testing of wastewater are recommended for aquatic medical facilities as they may limit the discharge of pharmaceuticals to the environment (Yargeau and Leclair, 2008); and a withdrawal period (e.g., 1 month) is recommended before releasing pharmacologically treated sea turtles to the wild.

In this study, a concentrated U.S. Food and Drug Administration–approved dose formulation was used (250 mg/ml), resulting in quite small volumes of administration that approached the accuracy of syringe volume increments (rounded to 0.01 ml). If a volume error of 0.005cc occurred, the result would be ±1.25 mg, or ±0.59 mg/kg inaccuracy for the smallest turtle in this study. Although this possibility cannot be excluded, this maximal 12% dosing error is not likely clinically important. Another variable introduced by the high drug concentration, even with perfect volume measurement, was the dose range of 4.5–5.1 mg/kg. The lowest theoretical dose (4.5 mg/kg) was given to one of the turtles with very low plasma concentrations (most values <1 µg/ml), which could have been further compounded by potential volume error, resulting in an even lower delivered dose (e.g., 4 mg/kg). Although it is likely that a potential 20% lower dose would have affected concentrations, it does not seem severe enough to fully explain the very low observed concentrations. For future studies that involve animals with relatively low body mass, use of amikacin products of lower concentration (50 mg/ml) should be considered.

Results of this study indicate that 5 mg/kg amikacin IM q3d appears to be a safe and reasonable dosage for many Kemp’s ridley turtles, especially if supported by MIC data. However, given the individual variability seen in this study, clinicians may consider assessment of plasma amikacin concentrations for patients that fail to respond to treatment. The timing of such assessments is at the discretion of attending veterinarians but may be informed by results of this study.