Effects of oral administration of metronidazole and doxycycline on olfactory capabilities of explosives detection dogs (original) (raw)

Olfactory acuity is the ability to detect and recognize odorants. In canids, it is expressed as the detection threshold, which is defined as the lowest concentration of a given odorant that can be consistently detected.1 Olfactory dysfunction in up to approximately 24% of affected humans is caused by metabolic derangement or induced by pharmaceuticals,[2–4](#ref2 ref3 ref4) but the prevalence of olfactory dysfunction in dogs is unknown. Diminished olfaction in dogs reportedly can be caused by canine distemper virus infection,5 parainfluenza virus infection,6 administration of dexamethasone or hydrocortisone,7 and feeding of a diet high in saturated fat.8

In humans, generalized decreased olfactory acuity (hyposmia) is characterized as type I, II, or III. Type I hyposmia is defined as the inability to correctly recognize and identify odors. Type II hyposmia is a quantitative decrease in the ability to detect odorants, and type III hyposmia is a decrease in estimation of the magnitude of odors. Types I and III cannot be assessed in dogs. Type II hyposmia is the most common form of hyposmia detected in humans following pharmaceutical treatment,1 and it can be evaluated in dogs via a number of laboratory methods, including use of behavioral olfactometry,[7–9](#ref7 ref8 ref9) electroencephalographic olfactometry,10 and scent-wheel techniques.[11–14](#ref11 ref12 ref13 ref14)

Hyposmia is of concern in working dogs, particularly detection dogs. Consequences of hyposmia could be catastrophic, especially for dogs trained to detect explosives. Pharmaceutical-induced hyposmia often may remain undetected in dogs because affected animals cannot report hyposmia. Therefore, it is important to elucidate whether pharmaceuticals cause olfactory deficits in working dogs so that clinicians can avoid the use of those medications. To the authors' knowledge, there is only 1 published study7 on the olfactory effects of pharmaceuticals in dogs. Investigators in that study7 found that a high dosage of dexamethasone (2 mg/kg/d) or a combination of hydrocortisone and deoxycorticosterone (0.25 mg/kg/d) decreased olfactory acuity (with or without observable clinical signs) in laboratory dogs after 7 and 18 days of administration, respectively.

Most clinically relevant information on the effects of pharmaceuticals on canine olfaction is extrapolated from human medicine. There is an extensive list of pharmaceuticals that can induce hyposmia in humans. This list includes anesthetics, antihistamines, antimicrobials, antineoplastics, cardiovascular drugs, endocrine drugs, gastrointestinal drugs, neurologic drugs, and NSAIDs.[1,15–18](#ref1 ref15 ref16 ref17 ref18) Of particular interest are metronidazole and doxycycline, which are 2 antimicrobials commonly administered to working dogs. Metronidazole is used to treat diarrhea, especially that caused by Giardia spp, which can be transmitted rapidly through working dog kennels. Metronidazole has been associated with taste disorders, specifically a bitter metallic taste, in humans.19 It has also been linked to decreased cognitive function in elderly humans,20 which suggests that it may impact neurologic function. Doxycycline is also commonly administered to working dogs for the prevention and treatment of vector-borne diseases, such as those caused by Ehrlichia spp, Babesia spp, and Rickettsia spp. Doxycycline can cause hyposmia in humans.21 The objective of the study reported here was to evaluate the effects of metronidazole and doxycycline on olfactory capability in dogs. Our hypothesis was that metronidazole and doxycycline would cause degraded olfactory detection capabilities in ED dogs.

Materials and Methods

Animals

Eighteen adult Labrador Retrievers currently enrolled in an ED training program were included in the study. Before the dogs were enrolled in the ED training program, they passed government health and temperament standards for detection dogs. The dogs were exercised at least 5 d/wk by use of traditional methods of easy-paced trotting for moderate distances, free playing, running on roads, running on treadmills, and swimming to improve their physical condition. All dogs were fed the same commercially available diet.a Health status of dogs was evaluated by means of physical examination, ECG, a CBC, biochemical analysis, and urinalysis; only healthy dogs (without substantial abnormalities) were included in the study. The study was conducted with the approval of the Auburn University Institutional Animal Care and Use Committee.

ED training

Dogs were trained for ED by use of standard odor detection training techniques, as described elsewhere.[12–14](#ref12 ref13 ref14) Dogs were trained to detect odors of 3 typical explosives (eg, target odors): AN, TNT, and SP. Standard odors and concentrations were used for training ED dogs22 for US law enforcement agencies. The amount of material serving as a source of vapor ranged from 1 to 500 mg, independent of explosive type. There were no target materials used that produced excessive or noxious amounts of vapor. Negative control samples consisted of empty Petri dishes and distractors (nontarget odors commonly encountered in a search environment and that are strong enough to potentially distract dogs when searching for target odors). Distractors included sugar, various teas, various nuts, mulch, and bark. The use of empty Petri dishes is important for testing false-positive results because the lower the weight of a target material, the more the target material is similar to an empty Petri dish.

The scent wheel is a common method for testing odor detection.[11–14,23](#ref11 ref12 ref13 ref14 ref23) Briefly, a climate-controlled room contained a device with 8 arms arranged like spokes of a wheel (Figure 1). At the end of each arm was a small metal basket that held a glass Petri dish covered by a mesh screen. Each arm was connected to a central sealed vacuum unit that allowed air to be evacuated from around each basket and prevented dispersal of odors. A target odor was placed in 1 Petri dish; the other 7 Petri dishes contained nontarget odors or were empty. The amount of time between placement of the target odor and detection by a dog was ≤ 90 seconds. The Petri dish in which the target odor was placed was randomly assigned by use of a random number generator. The Petri dish and basket with the target odor were changed and sanitized between subsequent trials, and at least half of the Petri dishes and baskets with nontarget odors were changed and sanitized between subsequent trials. This ensured that a dog could not mark (ie, identify on the basis of a characteristic other than odor) a positive target.

Figure 1—

Figure 1—

Figure 1—

Photographs of a scent wheel used to test olfactory abilities of ED dogs (A) and the basket at the end of one of the arms of the scent wheel (B). The distance from the central hub to each basket was 0.75 m. Notice the small metal basket holds a glass Petri dish covered by a mesh screen. The pipe above the basket is a central sealed vacuum unit that allows air to be evacuated from around the basket to prevent dispersal of odors.

Citation: American Journal of Veterinary Research 77, 8; 10.2460/ajvr.77.8.906

Other industry standard methods were used to prevent odor cross-contamination.12 A dog was brought into the room by its handler, the dog's leash was removed, and the dog was commanded to search, starting at the first basket and continuing around the scent wheel to the eighth basket. Handlers were not aware of the basket that contained the target odor. Handlers were instructed to stand still, to look forward and not at the dog, and to not give the dog any verbal commands. Each dog searched the scent wheel independently and without the assistance of the handler.

A positive response (ie, alert) was recorded when the dog stopped and sat at a basket. A negative response was recorded when the dog did not pause at a basket. Dog behavior was categorized. Find meant the dog had a positive response (alert) to a target odor, search past target meant the dog did not react to a target odor, and false meant the dog had a positive response (alert) to a nontarget odor. An intermittent reward system was used whereby the dog was randomly rewarded with a toy, rewarded with praise, or not rewarded for finds. Dogs were not rewarded if they searched past a target or had a false-positive response.

Odor detection threshold was defined as the lowest repeatable amount of each material that a dog was capable of detecting. Detection threshold was measured through sequential testing on the scent wheel. Initially, dogs were tested with 500 mg of material for each odor; when a dog successfully detected the odor, it was then tested with sequentially smaller amounts of material (250, 100, 50, 25, 12, 6, 3, and 1 mg). When a dog failed to successfully detect an odor, it was then retested by use of the sequentially larger amount of material (ie, the last amount for which that dog successfully detected the odor). Repeatability of detection threshold was determined by at least 2 successful detections at a specific amount of material with an inability to detect at the sequentially lower amount.

Experimental design

Eighteen Labrador Retrievers that had been trained by use of a scent wheel to detect odors of explosives were enrolled in a controlled, blinded, clinical trial. Each dog served as its own control animal. The study consisted of two 10-day drug administration periods separated by a 10-day washout period, which was used to ensure at least 7 drug serum half-lives elapsed between treatment periods. The day prior to drug administration was designated as day 0.

Metronidazole was administered (25 mg/kg, PO, q 12 h for 10 days) during the first drug administration period. Doxycycline was administered (5 mg/kg, PO, q 12 h for 10 days) during the second drug administration period. Dog handlers and the test administrator were not aware of the drug administered to each dog.

Odor testing for 3 explosives (AN, TNT, and SP) by use of the scent wheel was performed on days 0, 5, and 10 of each drug administration period. Each dog performed 28 trials/test day (14 trials in the morning and 14 trials in the evening). Odor detection threshold measured on day 0 of each drug administration period was recorded as the baseline value for each odor. Detection threshold on days 5 and 10 of each drug administration period was compared with the baseline value for that drug administration period. Degradation in performance was defined as an alteration in the detection threshold for a specific odor, such that the threshold amount detected was higher than the baseline value (eg, 25 mg on day 10 vs 12 mg on day 0). For the 28 trials on each test day, 22 (78.6%) were conducted by use of 1 target odor and 7 nontarget odors on the scent wheel. For the remaining 6 (21.4%) trials, blank testing was performed during which the scent wheel contained no target odors and 8 nontarget odors.

Statistical analysis

Degradation in the threshold amount detected by a dog between days 0, 5, and 10 of drug administration was assessed by use of the McNemar test of paired proportions based on an exact binomial probability distribution. The 95% CI for proportions was also determined by use of calculations based on an exact binomial probability distribution. Geometric mean values were calculated because of noncontinuous ordinal data. Continuous variables (body weight and age) were assessed for normality by use of the Shapiro-Wilk test. Values were considered significant at P < 0.05.

Results

Animals

All 18 dogs completed the study. Mean ± SD body weight was 28.49 ± 3.87 kg, and mean age was 3.36 ± 1.23 years. Dogs consisted of 10 neutered males, 5 sexually intact males, 2 sexually intact females, and 1 spayed female. A total of 3,024 odor trials were conducted during the 6 days of testing (3 test days/drug); each dog completed 168 trials and discriminated > 650 distractors.

Metronidazole administration

Metronidazole was orally administered at a median dose of 25.1 mg/kg (range, 22.5 to 26.9 mg/kg) every 12 hours for 10 days. During AN odor trials, 2 of 18 (11.1%; 95% CI, 1.4% to 34.7%) dogs had a degradation in performance on day 5 and 3 of 18 (16.7%; 95% CI, 3.6% to 41.4%) dogs had a degradation in performance on day 10, compared with baseline results (Table 1). A total of 4 of 18 (22.2%; 95% CI, 6.4% to 47.6%) dogs had diminished olfactory capability for AN; one of these dogs had a degradation in performance on both days 5 and 10.

Table 1—

Olfactory detection threshold of AN, TNT, and SP during metronidazole administration for 9 ED-trained dogs with evidence of metronidazole-induced hyposmia.

Dog AN (mg) TNT (mg) SP (mg)
Day 0 Day 5 Day 10 Day 0 Day 5 Day 10 Day 0 Day 5 Day 10
1 1 1 1 1 6* 1 1 1 1
2 12 25* 12 25 25 50* 3 1 1
3 1 1 12* 25 25 100* 3 1 1
4 12 6 3 6 25* 12* 1 1 1
5 25 12 12 12 25* 12 3 1 1
6 12 50* 25* 50 50 50 3 1 1
7 25 12 50* 25 25 100* 6 6 6
8 25 6 1 3 25* 6* 1 1 1
9 25 12 12 12 50* 25* 6 3 3

During TNT odor trials, 5 of 18 (27.8%; 95% CI, 9.6% to 53.5%) dogs had a degradation in performance on day 5 and 6 of 18 (33.3%; 95% CI, 13.3% to 59.0%) dogs had a degradation in performance on day 10, compared with baseline results (Table 1). A total of 8 of 18 (44.4%; 95% CI, 21.5% to 69.2%) dogs had diminished olfactory capability for TNT, with 3 of these dogs having a degradation in performance on both days 5 and 10.

During SP odor trials, none of the 18 dogs had a degradation in performance on day 5 or 10, compared with baseline results.

Thus, a significant (P = 0.004) number of dogs (9/18) had a degradation in performance, compared with baseline results, for 1 or more odors while receiving metronidazole. Six dogs had a degradation in performance for 1 odor, and 3 dogs had a degradation in performance for > 1 odor.

Doxycycline administration

Doxycycline was orally administered at a median dose of 5.0 mg/kg (range, 4.5 to 5.4 mg/kg) every 12 hours for 10 days. During AN and TNT odor trials, none of the dogs had a degradation in performance on day 5 or 10, compared with baseline results. During SP odor trials, 1 of 18 (5.6%; 95% CI, 0.1% to 27.3%) dogs had a degradation in odor detection threshold on day 5. Thus, degradation in detection threshold during doxycycline administration was detected in 1 dog for 1 odor; this result was not significant (P = 1.000).

Dog performance

A learning effect was evident within and between drug administration periods for each of the 3 odors (Table 2). Mean odor detection threshold for AN decreased sequentially from 7.7 mg on day 0 of metronidazole administration to 2.9 mg on day 0 of doxycycline administration. Mean odor detection threshold for TNT decreased sequentially from 9.6 mg on day 0 of metronidazole administration to 4.5 mg on day 0 of doxycycline administration. Mean odor detection threshold for SP decreased from 2.2 mg on day 0 of metronidazole administration to 1.3 mg on day 0 of doxycycline administration.

Table 2—

Geometric mean values for odor detection threshold for 18 ED-trained dog on the basis of explosive and antimicrobial administered.

Antimicrobial AN (mg) TNT (mg) SP (mg)
Day 0 Day 5 Day 10 Day 0 Day 5 Day 10 Day 0 Day 5 Day 10
Metronidazole 7.70 4.62 4.34 9.61 11.73 8.67 2.43 1.41 1.33
Doxycycline 3.93 2.92 2.92 7.15 5.21 4.53 1.43 1.47 1.33

Of the 3,024 trials, there were false-positive results for 230 (7.6%; 95% CI, 6.7% to 8.6%). There was a significant (P < 0.001) difference in false-positive results among the dogs, with false-positive results more commonly recorded for 8 of the 18 dogs. There was no association between the frequency of false-positive results and odor type, odor threshold, or effects of metronidazole. Of the 8 dogs that more commonly had false-positive results, 7 (87.5%; 95% CI, 47.3% to 99.7%) had more false-positive results during metronidazole administration than during doxycycline administration, whereas the other dog (12.5%; 95% CI, 0.3% to 52.7%) had more false-positive results during doxycycline administration than during metronidazole administration.

Discussion

In the study reported here, metronidazole administration resulted in degradation of detection ability in some ED dogs, but doxycycline administration did not significantly affect detection ability. Metronidazole is a commonly used antimicrobial; therefore, it is important to consider that administration may impact detection abilities of ED dogs.

Metronidazole is a concentration-dependent bactericidal antimicrobial widely used to combat diarrhea in dogs, especially that caused by Giardia spp. It is highly lipophilic and widely distributed, with a half-life of 4 to 6 hours in dogs.24 Metronidazole has been associated with alterations in taste, specifically association with a bitter metallic taste,19 in humans and has been linked to decreased cognitive function in elderly humans.20 Metronidazole-induced CNS toxicosis has been reported in cats25 and dogs[25–28](#ref25 ref26 ref27 ref28); therefore, we hypothesized that it could cause hyposmia in dogs. In the present study, 9 of 18 dogs had a degradation in performance for 1 or more odors of explosives during metronidazole administration. Because of the small number of dogs and limited instances of degradation, an association between performance degradation and body weight, age, sex, or neuter status could not be assessed. Metronidazole was orally administered at a mean dose of 25.1 mg/kg every 12 hours on the basis of an antigiardial protocol used at Auburn University. It is possible that there would not be olfactory dysfunction at lower doses, but other doses of metronidazole were not evaluated in the present study. Degradation in olfaction threshold was measured as the number of milligrams of raw material; degradation was detected only for AN and TNT and not for SP. In contrast to the odor of AN and TNT, SP has a distinct strong odor detectable even by humans, so it is possible that dogs could detect SP even in the face of metronidazole-induced hyposmia. Metronidazole-induced hyposmia in the 9 dogs did not result in clinical signs, such as a decrease in or absence of appetite or altered social behavior. Metronidazole-induced hyposmia also appeared to be transient, similar to the situation in humans, because the baseline value on day 0 of the second drug administration period was the same or better than the baseline value for the first drug administration period.

Doxycycline is a time-dependent bacteriostatic tetracycline antimicrobial administered to dogs for the treatment of a wide variety of infectious diseases, especially tick-borne diseases. A case of a human with doxycycline-associated reversible hyposmia has been reported,21 and the most likely mechanism appears to be interference with the binding of odorants to their specific receptor.1 In the present study, there was no detectable degradation in olfaction in ED dogs during doxycycline administration.

Analysis of means for each odor revealed a learning effect within and between drug administration periods. A learning effect is to be expected with repeated training on an odor device, but it may have reduced the magnitude of the drug effect we detected. Therefore, it is possible that > 50% of dogs receiving metronidazole will have diminished olfactory performance and that some dogs receiving doxycycline could have diminished olfactory performance.

The tendency for dogs to have a false-positive result more often during the metronidazole administration period than the doxycycline administration period may have reflected the learning effect, or it may have been attributable to the effect of metronidazole on olfaction. Four of 9 dogs that had a degradation in performance (ie, increase in odor threshold) on days 5 or 10 (or both) of the metronidazole administration period had a false-positive rate higher than the mean threshold value for all dogs. However, these 4 dogs all exhibited a learning effect with regard to false-positive results (ie, rate of false-positive results decreased from day 0 to day 5 to day 10). These 4 dogs had improvement in overall performance as the study progressed (rate of false-positive results decreased), but their odor detection threshold for AN or TNT was worse when receiving metronidazole, which suggested the change in odor detection threshold was attributable to the effect of the drug.

In the study reported here, we did not evaluate potential mechanisms by which metronidazole could cause olfactory dysfunction, but it was likely a sensory or neurologic dysfunction, similar to the situation in humans.[1,15,17,19,29](#ref1 ref15 ref17 ref19 ref29) Olfactory dysfunction is broadly categorized in humans as conductive or sensory. Conductive dysfunction results from the failure of odorant molecules to access olfactory mucosa, as in the case of nasal polyps, infection, inflammation, or neoplasia. Sensory dysfunction results from damage to the olfactory mucosa or olfactory nerves caused by pharmaceuticals, viruses, chemical exposure, or central or peripheral neuropathy. Pharmaceuticals cause olfactory dysfunction through impairment of odorant binding to the olfactory receptor, injury to the olfactory receptor,1 or neurologic impairment.17 Diminished olfaction in dogs has been associated with canine distemper virus infection,5 parainfluenza virus infection,6 administration of dexamethasone or a combination of hydrocortisone and deoxycorticosterone,7 feeding a diet high in saturated fat,8 endocrine disease (eg, hyperadrenocorticism, diabetes mellitus, or hypothyroidism), granulomatous meningoencephalitis, and nasal tumors,9 but few studies have been conducted to elucidate the mechanism for olfactory dysfunction. The decreased olfactory acuity in laboratory dogs receiving high doses of dexamethasone (2 mg/kg/d) or hydrocortisone combined with deoxycorticosterone (0.25 mg/kg/d) that were evident after 7 and 18 days of treatment, respectively, was attributed to elevations in the olfactory detection threshold,7 which was another possible mechanism for the hyposmia detected in the dogs of the present study. To the authors' knowledge, the present study was the first in which hyposmia was detected in dogs secondary to antimicrobial administration.

The present study had several limitations. Only Labrador Retrievers were included because they were one of the breeds most commonly used for detection dogs; it is possible that results may not be applicable to the entire population of working dogs. Additionally, we used only ED dogs (and did not include dogs trained for the detection of narcotics, cadavers, or other factors), but it would be reasonable to suspect that olfactory dysfunction would impact all types of detection dogs. The duration of metronidazole-induced hyposmia was not specifically determined in the present study, but there was no detectable hyposmia at the completion of the 10-day washout period. No placebo-controlled treatment was used to evaluate the effect of time, and order of drug administration was not randomized; nonetheless, the study design revealed a significant effect of metronidazole on olfaction and a lack of a significant effect of doxycycline on olfaction. This study was conducted in a controlled room, rather than a field setting, but the study involved the use of explosive odorants and distractors commonly encountered in a field or working environment. Pharmaceutical-induced hyposmia in dogs should be investigated by olfactory testing of multiple types of detection dogs in a field environment and examining effects of other pharmaceuticals.

In the study reported here, metronidazole administration caused a significant degradation in the ability to detect odor of explosives in 9 of 18 trained Labrador Retrievers. It does not appear possible to predict which dogs would be affected; thus, metronidazole should be used with caution in ED dogs as well as other detection dogs. Hyposmia induced by metronidazole was temporary, with olfaction returning to baseline values by 10 days after discontinuation of drug administration. Doxycycline did not have a significant effect on olfaction in ED dogs; thus, it likely can be considered safe for use at this dosage and for this duration of administration.

Acknowledgments

Supported in part by the Interdepartmental Research Grants Program; the Scott-Richey Research Center, College of Veterinary Medicine; and the Animal Health Disease Research Intramural Grants Program, Auburn University.

The authors declare that there were no conflicts of interest.

This manuscript was submitted by Dr. Jenkins to the Department of Clinical Sciences at the Auburn University College of Veterinary Medicine as partial fulfillment for a Master of Science degree.

Presented in abstract form at the American College of Veterinary Internal Medicine Forum, Nashville, Tenn, June 2014; and the Canine Science and Technology Conference, Raleigh, NC, July 2014.

The authors thank Dr. Jay Barrett for medical and technical assistance and Terrence Fischer and Pamela Haney for technical assistance.

ABBREVIATIONS

AN Ammonium nitrate
CI Confidence interval
ED Explosives detection
SP Smokeless powder
TNT Trinitrotoluene

References

  1. Henkin RI. Drug-induced taste and smell disorders: incidence, mechanisms and management related primarily to treatment of sensory receptor dysfunction. Drug Saf 1994; 11: 318–377.10.2165/00002018-199411050-00004
    )| false
  1. Kharoubi S. Olfaction disorders: preliminary findings and results in 45 cases. J Laryngol Otol 2001; 122: 43–49.
    )| false
  1. Henkin RI, Levy LM, Fordyce A. Taste and smell function in chronic disease: a review of clinical and biochemical evaluations of taste and smell dysfunction in over 5000 patients at The Taste and Smell Clinic in Washington, DC. Am J Otolaryngol 2013; 34: 477–489.
    )| false
  1. Fonteyn S, Huart C, Deggouj N, et al. Non-sinonasal-related olfactory dysfunction: a cohort of 496 patients. Eur Ann Otorhinolaryngol Head Neck Dis 2014; 131: 87–91.10.1016/j.anorl.2013.03.006
    )| false
  1. Myers LJ, Hanrahan LA, Swango LJ, et al. Anosmia associated with canine distemper. Am J Vet Res 1988; 49: 1295–1297.
    )| false
  1. Myers LJ, Nusbaum KE, Swango LJ, et al. Dysfunction of sense of smell caused by canine parainfluenza virus infection in dogs. Am J Vet Res 1988; 49: 188–190.
    )| false
  1. Ezeh PI, Myers LJ, Hanrahan LA, et al. Effects of steroids on the olfactory function of the dog. Physiol Behav 1992; 51: 1183–1187.10.1016/0031-9384(92)90306-M
    )| false
  1. Altom EK, Davenport GM, Myers LJ, et al. Effect of dietary fat source and exercise on odorant-detecting ability of canine athletes. Res Vet Sci 2003; 75: 149–155.10.1016/S0034-5288(03)00071-7
    )| false
  1. Myers LJ. Dysosmia of the dog in clinical veterinary medicine. Prog Vet Neurol 1990; 1: 171–179.
    )| false
  1. Hirano Y, Oosawa T, Tonosaki K. Electroencephalographic olfactometry (EEGO) analysis of odor responses in dogs. Res Vet Sci 2000; 69: 263–265.10.1053/rvsc.2000.0420
    )| false
  1. Fjellaner R, Andersen EK, McLean IG. A training program for filter-search mine detection dogs. Int J Comp Psychol 2002; 15: 278–287.
    )| false
  1. Angle TC, Wakshlag JJ, Gillette RL, et al. The effects of exercise and diet on olfactory capability in detection dogs. J Nutr Sci 2014; 3: 1–5.
    )| false
  1. Sargisson RJ, McLean IG. The effect of reinforcement rate variations on hits and false alarms in remote explosive scent tracing with dogs. J ERW Mine Action 2010; 14: 64–68.
    )| false
  1. Bromley SM. Smell and taste disorders: a primary care approach. Am Fam Physician 2000; 61: 427–436.
    )| false
  1. Medlink neurology clinical summaries. Drug-induced disturbances of smell and taste. Available at: www.medlink.com/medlinkcontent.asp. Accessed Oct 10, 2011.
    )| false
  1. Ship JA, Chavez EM. Special senses: disorders of taste and smell. In Silverman S, Eversole L, Truelove E, eds. Essentials of oral medicine. London: BC Booker, 2001;277–288.
    )| false
  1. Ackerman BH, Kasbekar N. Disturbances of taste and smell induced by drugs. Pharmacotherapy 1997; 17: 482–496.10.1002/j.1875-9114.1997.tb03058.x
    )| false
  1. Doty RL, Shah M, Bromley SM. Drug-induced taste disorders. Drug Saf 2008; 31: 199–215.10.2165/00002018-200831030-00002
    )| false
  1. Rogers J, Wiese BS, Rabheru K. The older brain on drugs: substances that may cause cognitive impairment. Geriatr Aging 2008; 11: 284–289.
    )| false
  1. Bleasel AF, McLeod JG, Lane-Brown M. Anosmia after doxycycline use. Med J Aust 1990; 152: 440.10.5694/j.1326-5377.1990.tb125275.x
    )| false
  1. Oxley J, Waggoner P. Detection of explosives by dogs. In: Marshall M, Oxley J, eds. Aspects of explosives detection. Amsterdam: Elsevier, 2009;27–40.
    )| false
  1. Williams M, Johnston JM. Training and maintaining the performance of dogs (Canis familiaris) on an increasing number of odor discriminations in a controlled setting. Appl Anim Behav Sci 2002; 78: 55–65.10.1016/S0168-1591(02)00081-3
    )| false
  1. Neff-Davis CA, Davis LE, Gillette EL. Metronidazole: a method for its determination in biological fluids and its disposition kinetics in the dog. J Vet Pharmacol Ther 1981; 4: 121–127.10.1111/j.1365-2885.1981.tb00720.x
    )| false
  1. Caylor KB, Cassimatis MK. Metronidazole neurotoxicosis in two cats. J Am Anim Hosp Assoc 2001; 37: 258–262.10.5326/15473317-37-3-258
    )| false
  1. Bradley WG, Karlsson IJ, Rassol CG. Metronidazole neuropathy. Br Med J (Clin Res Ed) 1977; 2: 610–611.10.1136/bmj.2.6087.610
    )| false
  1. Halloran TJ. Convulsions associated with high cumulative doses of metronidazole. Drug Intell Clin Pharm 1982; 16: 409.
    )| false
  1. Ray RE, Holton JL, Lister T, et al. The glio-vascular toxicity of m-dinitrobenzene and related agents: modulation of toxicity by neuronal activation. Arch Toxicol Suppl 1996; 18: 140–148.10.1007/978-3-642-61105-6_15
    )| false
  1. Hawkes CH, Doty RL. General disorders of olfaction. In: The neurology of olfaction. Cambridge, England: Cambridge University Press, 2009;111–152.
    )| false