Salmonellosis and Subsequent Abortion in Two African Elephants (Loxodonta africanus)
American Association of Zoo Veterinarians Conference 2000
Karen A. Emanuelson, DVM; Colleen E. Kinzley
Oakland Zoo, Oakland, CA, USA


At the Oakland Zoo, one male and three female African elephants (Loxodonta africanus) are managed intensively within a protected contact system and are trained for a variety of medical procedures. In March of 1998 an outbreak of salmonellosis occurred in the herd and affected two young cows that were in their second year of pregnancy. The three cows tested positive for Salmonella and were affected by the disease; one bull was not affected. The source of the Salmonella that infected the elephants in the outbreak described in this report has not been identified.

Case Report

Salmonellosis has been reported previously in African and Asian elephants of both sexes and of various ages in several countries.8 Salmonella infections have been associated with a high mortality rate in elephants, but, to date, a source of a Salmonella infection in an elephant has not been positively identified or reported. There may be an increase in the number of cases in recent years, in elephants as in domestic animals.1,4,8

Elephant 1

On 27 March 1998, a 14-months-old pregnant female elephant presented with sudden onset of anorexia and lethargy. A reoccurrence of a previous illness, cholangitis, was suspected initially.2 Results of a CBC, serum biochemistry analysis, and bile acids evaluation were unremarkable, except for a relative increase in neutrophil bands (11%). On day 2 the anorexia and lethargy were still present, stool volume and size were decreased, and signs of abdominal discomfort were evident. Treatment was initiated with flunixin meglumine (0.5–1.0 mg/kg IM SID to BID PRN; Banamine, Schering-Plough Corp., Union, NJ, USA) and penicillin G procaine (5670 IU/kg IM SID; Penicillin G Procaine, Butler Co., Columbus, OH, USA), and injections were administered in an elephant restraint chute.

Two days later clinical signs of colic had worsened and stool quality had deteriorated to frequent bouts of mucoid loose diarrhea. Feces and rectal mucus were submitted for Salmonella/Campylobacter culture and fecal floatation. CBC revealed leukopenia (3400 cells/µl; normal range 8000–13000) and neutropenia (952 cells/µl; normal range 1200–6200). Ceftiofur (2.5 mg/kg IM SID; Naxcel, Smithkline Beecham Corp., Philadelphia, PA, USA) was added to the treatments to broaden spectrum and protect against secondary infection. Serum progesterone level decreased from a normal pregnancy mean level of 2000 pg/ml to 970 pg/ml, which was the first indication of ill effects on the pregnancy of the cow. Prolactin remained at elevated pregnant levels (215 ng/ml).

A significant decline in condition was noted by the sixth day with the continuation of severe clinical signs (complete inanition, fresh blood admixed with mucus and feces, profound weakness, markedly decreased urination, and clear mucoid vaginal discharge). Intensive intravenous therapy was planned; the hydraulic restraint chute was prepared with supplemental heat, protective tarps, and rubber floor mats, to more comfortably maintain the cow within the confines of the chute continuously. For placement of the catheter, the cow was further restrained in the chute with a cargo strap on the left rear leg. An over-the-needle catheter (10 ga 15 cm Angiocath, Becton Dickinson, Sandy, UT, USA) with the stylet in place was placed in the left rear medial saphenous vein after lidocaine (10 ml ID; 2% Lidocaine, Butler Co., Columbus, OH, USA) infusion and a vertical 2-inch cut-down incision. The catheter was secured, a betadine pack was placed, and the IV line was wrapped with tape for long-term use. Intravenous fluids (10 L/h IV; lactated Ringer’s solution [LRS], Abbott Laboratories, North Chicago, IL, USA) were administered by attaching four 5 L bags to a large animal administration set. This administration rate represents 75 ml/kg/day at the estimated weight of 3200 kg, or 1.5 times maintenance.3

Leukopenia (3000/µl) persisted; progesterone declined, while prolactin remained elevated. Results of the initial fecal culture were received and were positive for Salmonella (later serotyped as Salmonella typhimurium4,5 one of the more virulent forms that has proven lethal in elephants1,9). Sensitivity testing indicated resistance to the penicillin group and sensitivity to amikacin and prompted substitution of amikacin for penicillin G procaine. The initial amikacin dose (3.77 mg/kg IV q12h; Amikacin, Butler Co., Columbus, OH, USA) was extrapolated from the single intravenous dose (8 mg/kg IV) listed in an elephant pharmacokinetic study,6 and the equine dose range (3.7–7.0 mg/kg)9. Serum amikacin levels were monitored to ensure that peak levels were within the therapeutic range (15–25 µg/ml preferred range in human therapy; >35 µg/ml toxic), and trough levels were not in the toxic range (1–4 µg/ml preferred range in human therapy; >8–10 µg/ml toxic). Because a case of fatal herpesviral infection had occurred in a calf at the zoo 2 years previously, heparinized whole blood was submitted for herpesvirus PCR and was later confirmed negative.

Fresh blood and pieces of necrotic tissue were passed with frequent small amounts of mucoid diarrhea over the next few days. Peripheral edema developed in the ventral area surrounding the vulva. Differential diagnoses included the following etiologies: endotoxemia, as a complication of Salmonella enteritis;11 fluid overload, though the fluid rate was well within the therapeutic range of 35–95 ml/kg/day;3 and salmonellosis, as it is a known cause of vasculitis,4 has been associated with ventral edema in horses with Salmonella enteritis (Gary Magdesian, pers. comm.) and has been reported previously in an elephant with salmonellosis8. Feces were submitted for Clostridium difficile and perfringens toxin ELISA. Polymixin B (3000 IU/kg IV q12h; Bedford Laboratories, Bedford, OH, USA) was added to the treatment regimen as a binder of endotoxins. Due to the seriousness of the infection and persistent clinical signs of colic, diarrhea, and lethargy, the 1.5 times maintenance fluid rate was not decreased and resulted in improved hydration and urination. The amikacin dose was increased to 5 mg/kg q12h since serum levels (peak 12.6 µg/ml, trough 2.4 µg/ml) were slightly below the therapeutic range. Potassium chloride (KCl; 20 mEq/L LRS fluids IV; Butler Co., Columbus, OH, USA) and vitamin B complex (2 ml/L LRS fluids IV; Butler Co., Columbus, OH, USA) were added to the fluid therapy, which approximated 240 L/day. The leukocyte count (4000/µl) was stable, while a mild anemia (27% PCV) was evident, secondary to gastrointestinal hemorrhage and unmasked by improved hydration. Suspicion arose that the fetus had died when both the serum progesterone and prolactin levels declined.

On day 12 the neutrophil count dropped to a low level (459 cells/µl, normal range 1200–6200/µl) and peripheral edema of the ventrum, legs, and vulva progressed. The ventral vulva became severely swollen, then blistered and began seeping edematous fluid. The fluid administration rate was decreased from 10 L/h to 5 L/h. Furosemide (0.26 mg/kg IV q8h; Furosemide, Butler Co., Columbus, OH, USA) was initiated the next day in efforts to reduce the edema. Cimetidine (4.3 mg/kg IV q8h; Tagamet, Mova Pharmaceuticals, Caquas, Puerto Rico) was added to combat potential gastrointestinal ulcers, another complication of severe enteritis in horses.11 At this time, Clostridium difficile endotoxin ELISA results were positive, suggesting, along with clinical signs, that the elephant was indeed endotoxemic.

The elephant’s cachexia, with an estimated a weight loss of 350–450 kg, was striking by the end of the second week. Despite the poor outlook and possible loss of venous access, the decision was made to release her from the chute and move her into the barn, hoping that the opportunity for movement, proximity to the other elephants, and the change of location might be helpful. The catheter was flushed with a heparin lock and bandaged, leaving access to the end of an IV extension set for administration of medications. The cow moved slowly into the barn, and though weak and exhausted, she remained standing and responsive to voice commands so that medications and catheter care could be delivered. After release to the bull yard with two other cows the next morning, the cow slowly but deliberately made her way to the mud wallow and did a great job of covering herself with mud. Shortly after being brought back into the barn, about 20 L of blood was passed from the vagina and abdominal discomfort was noted, but no fetus was passed.

All medications with the exception of amikacin had been discontinued by day 16 and the amikacin dose (5 mg/kg) interval was reduced to SID as serum amikacin levels (peak 19.1 µg/ml, trough 9.0 µg/ml) indicated that the dose interval of 18–24 hours would be most appropriate. Repeat fecal cultures for Salmonella continued to be positive, however, a second Clostridium endotoxin test and Clostridium culture were negative. The edema was slowly resolving so that a second IV catheter was placed in the medial saphenous vein of the right leg and the first catheter was removed. The elephant was brought to barrier bars and chained twice daily for administration of 25 L of LRS and other treatments. A 26 by 15 cm wound at the site where the vulva had blistered and split was treated topically with antiseptic scrubs.

Improvement in appetite, activity, comfort level, and stool quality occurred toward the end of the third week, so that IV fluids and antibiotics were completely discontinued by day 20 and day 21, respectively. A large flap of necrotic tissue fell off the vulvar wound at the end of week 5, at the same time appetite had returned to 80% of normal. A low serum prolactin level (27 ng/ml) supported presumed fetal death, despite a rebound in serum progesterone level (>2000 pg/ml). It was postulated that either the continued presence of the dead fetus in the uterus, or one or more retained corpus luteum, was causing the continued elevation of progesterone. Progesterone remained elevated for several weeks before declining. By week 9, appetite and CBC values had returned to normal. An accurate body weight of 3727 kg was obtained, at which time it was estimated that she had recovered approximately 75% of her original weight. Nine months after presentation, the cow continued to have small amounts of mucoid/hemorrhagic vaginal discharge almost daily, but had returned completely to normal in weight, appetite, and activity level. A 30 kg intact fetus was passed 17.5 months after the presumed time of fetal death. Minimal vaginal discharge and the passage of one piece of placenta occurred in the 6-day period surrounding the abortion. Behavior, appetite, CBC, and chemistry panels remained normal. Salmonella was not cultured from fetal tissues, placental tissues, nor vaginal discharge. There have been reports of elephant fetuses that had been retained for 3.5 months to 4 years after fetal death, with no known deleterious effects on the dam.7,10 Progesterone levels started cycling normally immediately after the abortion. Normal estrous cycles have followed with interest from the bull. A reproductive ultrasound examination that was performed nearly 2 years after presentation yielded normal findings consistent with estrous cycle stage.

Elephant 2

Another female African elephant that was also in the second year of pregnancy showed clinical signs of lethargy and sudden anorexia 7 days after those of the previously discussed cow. Although feces submitted that week was negative on enteric culture for Salmonella, infection was assumed and the cow was immediately started on amikacin (6 mg/kg IM q24h). Reduced quantities of mucoid feces proceeded a brief period of severe diarrhea. Clinical signs of extreme lethargy and colic were also noted. Fecal culture was positive for Salmonella after a relatively extended period of 5 days. Serotyping later confirmed Salmonella typhimurium. Although the elephant’s appetite was very depressed (1–5% of normal) for 17 days, it remained fairly consistent and was accompanied by good intake of liquids. By the end of the fourth week, appetite and stool quality were normal, and the lethargy was slowly resolving. Hormone analysis indicated that progesterone and prolactin levels remained within the normal range consistent with pregnancy. CBC and chemistry values did not deviate from normal limits during the episode. Repeated fecal cultures continued to be positive for Salmonella.

Six weeks after presentation, a second bout of anorexia and lethargy with abdominal discomfort and weakness developed. Fecal culture for Salmonella was negative. A CBC indicated monocytosis (4515 cells/µl) and hyperfibrinogenemia (700 mg/dl). Large amounts of vaginal blood were discharged 2 days later in the early morning (0300 hours). Active labor was noted at 07:00 and a full-term male fetus weighing 99 kg was aborted 1 hour later. A torn abnormally-thickened and discolored placenta was passed soon after the calf. A necropsy was performed on the fetus and placenta at the University of California, Davis. Histology and culture revealed severe Salmonella placentitis and disseminated Salmonella infection of the fetal tissues. Seventeen days after the abortion, large quantities of mucoid, dark pink and red material were passed from the vagina. Over the following three nights, approximately 24 L of material were passed. Repeated cultures of the vaginal discharges were positive for Salmonella typhimurium. Vaginal discharge in varying amounts persisted for the next 2 months. Due to the passage of a very large piece (10 cm diameter, 15 cm thick) of placenta-like material 3 months after the abortion and persistent Salmonella-positive cultures, treatment with trimethoprim-sulfadiazide (4.5 mg/kg [based on trimethoprim portion] PO SID for 12 days; Uniprim, Macleod Pharmaceuticals Inc., Fort Collins, CO, USA) was initiated. Following treatment, the vaginal discharge was culture negative for Salmonella, and the volume of discharge tapered off rapidly until resolution 3 days after cessation of medications. CBC values returned to normal after treatment.

Normal cycling resumed 17 weeks after the resolution of the uterine infection. Four normal estrous cycles have occurred, each with breeding by the bull. A reproductive ultrasound examination that was conducted 9 months after the abortion showed normal findings. Ongoing hormonal analysis determined a 20-week pregnancy by progesterone levels at the time of this writing.

Elephant 3

A non-pregnant third female African elephant became ill 10 days after Elephant 1. Though clinical signs of reduced appetite and lethargy were present, this cow maintained a consistent food intake that allowed treatment with oral antibiotics. Enrofloxacin (1.2 mg/kg PO q12h; Baytril, Bayer Corp., Shawnee Mission, KS, USA) was administered for 8 days. This elephant experienced a short period of loose stools with mild to moderate lethargy on day 3–5 of her illness. Within 7 days, appetite and activity level had returned to normal. Fecal culture on the first day of illness was positive for Salmonella typhimurium. This cow has experienced no complications following her Salmonella infection.

Elephant 4

A 24-year-old breeding bull experienced a brief period of loose stool at the time of the outbreak. However, the elephant never tested positive for Salmonella.


The source of the S. typhimurium was never found. A Salmonella culture protocol, consisting of weekly cultures of elephant feces (each individual) and food (produce, hay, and pellets), as well as, monthly cultures of indoor and outdoor elephant areas and water drinkers has been carried out for 2 years. Fecal cultures on the cows were positive for a few weeks after the outbreak; selected area cultures were positive for a few months; and reproductive discharge cultures were positive only in the second cow (Lisa) until a few days after the abortion. Shedding of Salmonella has not been seen in times of stress; no positive cultures have occurred since September 1998, 6 months after presentation. The routine Salmonella culture protocol will continue.

This outbreak of salmonellosis highlights the need for more investigation into the prevention and treatment of this devastating disease in elephants. Salmonella has now been shown to be a cause of abortion in elephants, as it is in domestic animals. Pregnancy may increase susceptibility to Salmonella infection and may increase the likelihood of development of serious disease in an elephant that contracts Salmonella. There are preliminary indications that reproductive recovery can occur after abortion due to Salmonella placentitis/endometritis and may possibly occur after prolonged fetal retention. Intensive medical care can be successfully administered to elephants in a protected contact system, as illustrated in this case report.


The authors would like to thank the animal care staff at the Oakland Zoo for their efforts in caring for the elephants in this report.

Literature Cited

1.  Crawshaw GJ. Salmonellosis in elephants. In: Proceedings of the 15th Annual Elephant Managers Association Workshop. Toronto, Canada. 1994.

2.  Emanuelson KA. Clinical update. J Elephant Manag Assoc. 1998;9(1):52–54.

3.  Fowler ME. Elephants (Proboscidae). In: Zoo and Wild Animal Medicine. Philadelphia, PA: WB Saunders Co.; 1986:884–923.

4.  Howard JL, ed. Salmonellosis. In: Current Veterinary Therapy, Food Animal Practice. Philadelphia, PA: WB Saunders Co.; 1993:562–565.

5.  Kenny DE. Salmonella spp. survey of captive rhinoceroses in U.S. zoological institutions and private ranches. J Zoo Wildl Med. 1999;30:383–388.

6.  Lodwick LJ, Dubach JM, Phillips LG, Brown CS, Jandreski MA. Pharmacokinetics of amikacin in African elephants (Loxodonta africanus). J Zoo Wildl Med. 1994;25:367–375.

7.  Lung NP, Murray S, Zachariah NY. Management of an Asian elephant pregnancy: the unexpected twins. In: Proceedings of the 18th annual Elephant Managers Workshop. Fort Worth, Texas. 1997:120–123.

8.  Mikota SK, Sargent EL, Ranglack GS. Medical Management of the Elephant. West Bloomfield, MI: Indira Publishing; 1994:97–99, 147.

9.  Robinson NE, ed. Antimicrobial chemotherapy: Acute colitis. In: Current Therapy in Equine Medicine. Philadelphia, PA: WB Saunders Co.; 1992:6, 247–248.

10.  Schilfarth J. World News, Copenhagen Zoo, Denmark. J Elephant Manag Assoc. 1997;8:6.

11.  Smith BS, ed. Diseases of the alimentary system. In: Large Animal Internal Medicine. St. Louis, MO: CV Mosby Co.; 1990:660–714.


Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Karen A. Emanuelson, DVM
Oakland Zoo
Oakland, CA, USA

MAIN : All : Salmonellosis & Abortion in African Elephants
Powered By VIN