Presumptive nonthyroidal illness syndrome in critically ill foals



Reasons for performing the study: Hypothalamic-pituitary-thyroid (HPT) axis dysfunction is associated with morbidity and mortality in critically ill people. To date, investigations of HPT axis in critically ill foals are limited.

Objectives: To document the occurrence of low thyroid hormone concentrations (presumptive nonthyroidal illness syndrome; NTIS) in critically ill newborn foals and investigate whether NTIS is associated with severity of disease and outcome.

Hypothesis: NTIS occurs frequently in foals with sepsis and is associated with sepsis score and outcome. Reverse T3 (rT3) concentrations will be increased in septic foals and highest in nonsurvivors.

Methods: Thyroid hormones (total and free thyroxine [TT4 and fT4], total and free tri-iodothyronine [TT3 and fT3], reverse T3 [rT3]) were prospectively measured in healthy, sick nonseptic and septic foals. Clinical and laboratory information was retrieved from the medical records. Hormones were measured by validated radioimmunoassays.

Results: Concentrations of all thyroid hormones except rT3 (P = 0.69) were decreased in septic and sick nonseptic foals (P<0.01). Reductions in hormone concentrations were associated with an increased sepsis score (P<0.01). Nonsurviving septic foals had lower TT4, fT4, TT3 and fT3 concentrations than surviving septic foals (P<0.01). rT3 concentrations were higher in nonsurviving septic premature foals than surviving septic premature foals (P<0.05).

Conclusions: NTIS (euthyroid sick syndrome) is frequently observed in critically ill and premature foals, and associated with severity of disease and mortality.

Potential relevance: More research is needed to better understand the mechanism of this finding and determine whether manipulation of the HPT axis or thyroid replacement therapy could be beneficial.


Neonatal sepsis is the leading cause mortality in foals during the first week of life (Cohen 1994). Endocrine maturation in foals occurs in late gestation and the early post natal period (Silver et al. 1984), with adaptations of the hypothalamus–pituitary–adrenal (HPA) axis, energy metabolism and vasomotor system to withstand stressful conditions. Recent studies have documented that dysfunction of various endocrine systems is associated with sepsis and mortality in newborn foals (Hurcombe et al. 2008, 2009; Hart et al. 2009; Barsnick et al. 2011). Despite the essential functions of thyroid hormones in development, as is evident in foals with congenital hypothyroidism (Allen et al. 1996), information on the hypothalamic–pituitary–thyroid (HPT) axis and thyroid hormone (TH) concentrations in critically ill foals is minimal.

Thyroid hormones are important in the development and maturation of the nervous, respiratory and musculoskeletal systems during the pre- and post natal periods. They also regulate metabolic activities. In man, several factors are known to affect TH synthesis, including hormones (cortisol, ACTH), drugs (phenylbutazone, dexamethasone), calorie intake and disease (Adler and Wartofsky 2007). In other species, and probably in equids, starvation (Powell et al. 2000), stress and illness suppress the HPT axis, leading to low TH concentrations (Toribio 2010). This condition has been known as the ‘euthyroid sick syndrome’, but was recently named the ‘nonthyroidal illness syndrome (NTIS)’ because it does not presume the metabolic status of the patient, which would normally drive changes in thyroid hormone concentration (De Groot 2006).

In man, NTIS, also referred to as ‘low T3 syndrome’, is characterised by decreased concentrations of total tri-iodothyronine (TT3), in addition to decreases in free tri-iodothyronine (fT3), and altered thyroxine metabolism. Increases in reverse T3 (rT3) concentration are also observed in NTIS as thyroxine (T4) conversion to rT3 through 5-deiodinase activity of the so called ‘inactivating pathway’ is upregulated (Adler and Wartofsky 2007). Moreover, thyrotropin stimulating hormone (TSH) concentrations are often normal, differentiating NTIS from hypothyroidism. Without knowing the concurrent TSH concentration, a presumptive diagnosis of NTIS can be made as thyroid axis function is not fully evaluated. However, NTIS as documented by thyroid hormone concentrations alone or in combination with TSH concentrations is a common finding in critically ill human patients, in particular in association with sepsis (Raymond and LaFranchi 2010). Furthermore, NTIS has been associated with nonsurvival in septic children (Yildizdas et al. 2004; Meyer et al. 2010).

In mature horses, thyroid dysfunction has been associated with anhydrosis (Breuhaus 2009) and equine metabolic syndrome, and sparse case reports have detailed hypothyroidism and hyperthyroidism (Ramirez et al. 1998; Schwarz et al. 2008; Tan et al. 2008). Conversely, in foals, thyroid gland dysfunction has been documented in foals born to dams grazing on endophyte-infested fescue grass (Boosinger et al. 1995) and foals with congenital hypothyroidism/goitre (McLaughlin et al. 1986; Allen et al. 1996). Decreased TT3 concentrations were found in few premature foals (Silver et al. 1991); however, peer-reviewed information on thyroid hormone concentrations in critically ill foals remains minimal.

The purpose of this study was to determine the prevalence of presumptive NTIS in sick foals by measuring TT4, fT4, TT3, fT3 and rT3 concentrations in newborn septic, sick nonseptic and premature foals presented to 2 referral care facilities. We hypothesised that presumptive NTIS occurs frequently in septic and premature foals, and that the magnitude of TH dysregulation will be associated with the severity of disease and survival. We also proposed that TH concentrations will be lower in premature than in full-term foals and will be further decreased in response to sepsis.

Materials and methods


A total of 179 full-term and premature foals aged ≤7 days of both sexes and any breed admitted to 2 referral care facilities (The Ohio State University, Columbus, Ohio, USA and Hagyard Equine Medical Institute, Lexington, Kentucky, USA) were included in a prospective study design. All foals were admitted during 2 northern hemisphere foaling seasons (2009 and 2010) during the months of January–April.

Foals were classified into 3 main groups: septic, sick nonseptic and healthy.

Foals with a positive blood culture and/or sepsis score ≥12 were classified as septic (Brewer and Koterba 1988). Sick nonseptic foals were those foals with no evidence of local or systemic infections (e.g. meconium impaction, perinatal asphyxia syndrome etc.), and that had a sepsis score ≤10 and negative blood culture results. Healthy foals were all aged 12–36 h at the time of examination and evaluated by field practitioners on local breeding farms in central Kentucky, USA. Foals were considered healthy based on a normal physical examination, normal complete blood count (CBC), normal serum biochemistry, serum immunoglobulin G (IgG) consistent with adequate transfer of passive immunity (>8.0 g/l) and had a sepsis score ≤5.

Premature foals were classified into 2 separate groups: premature septic foals and premature nonseptic foals. Prematurity was determined based on gestation age (<320 days) from known breeding dates and phenotypical appearance of the foal, including 2 or more of the following: silky hair coat, domed head, curled ears, low birthweight.

Foals with a history of administration of nonsteroidal anti-inflammatory drugs, glucocorticoids, plasma or maternal exposure to endophyte-infested fescue were excluded. Foals that had received antimicrobials and/or crystalloid solutions prior to admission were included.

Survival was defined as being discharged alive from the treatment facility. Nonsurvival was defined as death from progressively worsening disease or euthanasia based on grave medical prognosis. Foals subjected to euthanasia due to financial limitations were excluded from the study.

This study was approved by the clinical research advisory committee and adhered to the principles for the humane treatment of animals in veterinary clinical investigations as stated by the National Institutes of Health guidelines. Owner consent was obtained prior to foal inclusion in the study.

Data collection

A complete medical history including expected foaling date, maternal health and drug administration was obtained. Categorical variables assessed included age at presentation, breed and sex. Physical examination variables included heart rate, respiratory rate, rectal temperature, mucous membrane colour, capillary refill time, presence of cold extremities, peripheral pulse quality, mental status, thoracic and abdominal auscultation, ocular examination, and umbilical examination. The presence of a septic focus such as septic arthritis/physitis, pneumonia, enteritis, colitis, uveitis or omphalitis was recorded. All foals had a sepsis score calculated (Brewer and Koterba 1988). Clinicopathological variables evaluated included a complete blood count, serum biochemistry, fibrinogen concentration, serum IgG concentration (IgG turbidoimmunometric assay)a, blood culture, and serum hormone concentrations (TT4, fT4, TT3, fT3, rT3).


Blood samples were collected by aseptic venipuncture within 30 min of admission and prior to any medical therapy, for laboratory tests and hormone concentration analyses. Blood (20 ml) for TH concentrations was placed into plain serum clot tubes. Samples were centrifuged within 12 h at 4°C at 2000 g for 15 min, aliquoted and stored at -80°C until analysed. All hormone assays were completed within 6 months of collection.


Hormone concentrations for TT4, fT4, TT3, fT3 and rT3 were determined in serum using previously validated direct radioimmunoassays (Coat-A-Count Total T4, free T4, Total T3, free T3 assaysb; Adaltis Reverse T3c) (Meredith and Dobrinski 2004; Sommardahl et al. 2005).

Inter- and intra-assay coefficients of variations for samples of low, mid-range and high concentrations of TT4, fT4, TT3, fT3 and rT3 were all below 8.5%. The ability of these assays to detect TH was also assessed by running, in the same assays, samples from 2 horses that received 1 mg of thyrotropin releasing hormone (TRH) i.v. In these horses, fT3, fT4, TT3 and TT4 concentrations were 2–3 times baseline values 3 h after TRH stimulation. There was no difference in rT3 concentrations in these horses between 0 and 3 h after TRH administration.

Data analysis

Data are presented as median and range unless otherwise stipulated. Normality was assessed with the Shapiro-Wilk method. Normal ranges were generated with data of healthy foals from this study. Values below the lowest confidence interval of healthy foals were considered below the normal range. TH comparisons between foal groups were carried out with one-way ANOVA or the Kruskal-Wallis ANOVA on ranks and the Dunn's post test was used to compare groups individually. Student's t test and the Mann-Whitney U test were used to compare TH concentrations between survivors and nonsurvivors. Spearman's rank order (ρ) was used to determine correlations between TH concentrations and sepsis score. Significance was set at P<0.05. Commercial software was used for graph generation (Excel)d and statistical analysis (SigmaStat 3.5)e.


Study population

A total of 179 neonatal foals were included in the study, 53 of which were septic (39 full-term and 14 premature), 100 sick nonseptic (88 full-term and 12 premature); and 26 full-term healthy foals. Of these, 102 sick foals were admitted to Hagyard Equine Medical Institute; 51 sick foals were admitted to The Ohio State University; and all 26 healthy foals were evaluated on local breeding farms in central Kentucky.

Table 1 shows the age, gender distribution and outcome for each group of foals. Breeds in the septic and sick nonseptic groups included Thoroughbreds (109), Standardbreds (20), Quarter Horses (13), Paint horses (3), Appaloosa (2), Arabian (2) and one each of Warmblood, Friesian, Saddlebred and Haflinger. Healthy foals were all Thoroughbreds.

Table 1. Demographic data on all foals at admission or initial examination. Values expressed as median and (range)
VariableHealthy (n = 26)Full-term sick nonseptic (n = 88)Premature sick nonseptic (n = 12)Full-term septic(n = 39)Premature septic (n = 14)
Age (hours)16 (12–36)18 (1–192)12 (1–48)24 (1–168)36 (5–120)
 Male14/26 (54%)47/88 (54%)6/12 (50%)27/39 (69%)8/14 (57%)
 Female10/26 (38%)34/88 (38%)6/12 (50%)10/39 (26%)5/14 (36%)
Not recorded2/26 (8%)7/88 (8%) 2/39 (5%)1/14 (7%)
Survivor26/26 (100%)77/88 (88%)9/12 (75%)28/39 (72%)7/14 (50%)

Serum TT4, fT4, TT3, fT3 and rT3 concentrations for each group of foals are presented in Table 2. Hormone concentrations in nonsurviving foals and surviving septic foals are presented in Table 3.

Table 2. Thyroid hormone concentrations in all foals. Values expressed as median and (range)
HormoneHealthy (n = 26)Full-term sick nonseptic (n = 88)Premature sick nonseptic (n = 12)Full-term septic (n = 39)Premature septic (n = 14)
  1. TT4 = total thyroxine; fT4 = free thyroxine; TT3 = total tri-iodothyronine; fT3 = free tri-iodothyronine; rT3 = reverse tri-iodothyronine. *P<0.01 compared–healthy foals. ∧P<0.05 compared–full-term sick nonseptic foals.

TT4 (nmol/l)712 (295–1012)577 (18–956)*592 (78.5–860)578 (73.3–1120)*596 (78.5–689)
fT4 (pmol/l)50.2 (27–70)32.2 (1.3–67)*29.6 (11.6–56.7)25.7 (3.5–125)*24.5 (10.3–56.6)
TT3 (nmol/l)7.9 (3.2–9.5)5.7 (1.1–11)*5.2 (1.1–8.5)3.6 (0.3–10.6)*∧4.1 (1.1–7.0)
fT3 (pmol/l)21.2 (6.9–34.3)9.4 (1.4–25.8)*9.1 (1.1–21.2)5.1 (0.46–25)*∧5.8 (1.0–14.9)
rT3 (pmol/l)6.6 (5.2–8.1)6.8 (4.1–10.1)6.9 (4.5–9.2)7.3 (5.0–9.0)5.7 (4.5–9.2)
Table 3. Thyroid hormone concentrations in septic foals. Values expressed as median and (range)
HormoneFull-term septic (n = 39)Premature septic (n = 14)
Survivor (n = 28)Nonsurvivor (n = 11)Survivor (n = 7)Nonsurvivor (n = 7)
  1. TT4 = total thyroxine; fT4 = free thyroxine; TT3 = total tri-iodothyronine; fT3 = free tri-iodothyronine; rT3 = reverse tri-iodothyronine. *P<0.05 between full-term surviving and full-term nonsurviving foals. #P<0.05 between premature surviving and premature nonsurviving foals.

TT4 (nmol/l)624 (160–1120)538 (73.4–701.4)602.3 (79–682)544.4 (362–695)
fT4 (pmol/l)32.3 (9–125)*18 (3.9–25.7)37.2 (11.6–56.6)19.3 (10.7–33.7)
TT3 (nmol/l)6.2 (0.4–10.6)*2.4 (0.3–5.1)4.99 (1.74–7.0)3.2 (1.1–5.8)
fT3 (pmol/l)8.8 (0.6–24.9)*2.2 (0.5–6.0)7.1 (3.0–14.9)3.0 (1.1–9.1)
rT3 (pmol/l)6.8 (5.5–8.9)7.8 (4.9–8.8)6.1 (5.1–7.2)#7.8 (6.0–9.5)

Hormone concentrations in full-term foals

Median TT4 concentrations were significantly lower in septic and sick nonseptic than in healthy foals (P<0.001), but there was no difference between sick nonseptic and septic foals, or within septic survivors and nonsurvivors (P = 0.28). TT4 concentrations in septic foals were below the normal range in 88% of nonsurvivors and 60% of survivors.

Median fT4 was lower in the septic and sick nonseptic foals than in the healthy foals (P = 0.001). There were no differences between sick nonseptic and septic foals (P = 0.3). Nonsurviving septic foals had the lowest fT4 concentrations, which were significantly decreased compared to surviving septic foals (P = 0.016). All nonsurviving and 78% of surviving septic foals had fT4 concentrations below the normal range.

Median TT3 concentrations were decreased in sick hospitalised foals. Septic foals had significantly lower TT3 concentrations than healthy foals (P<0.01) and sick nonseptic foals (P = 0.02). Sick nonseptic foals had lower TT3 concentrations than healthy foals (P<0.01). Nonsurviving septic foals had the lowest TT3 concentration of any group, significantly lower than in surviving septic foals (P = 0.003). Serum TT3 concentrations were below the normal range in all (100%) nonsurviving septic foals compared to 68% of surviving septic foals.

Median fT3 concentrations were significantly lower in sick foals. Serum fT3 concentrations were lower in septic foals than in healthy and sick nonseptic foals (P<0.001). Sick nonseptic foals had lower fT3 concentrations than healthy foals (P<0.001). Nonsurviving septic foals had the lowest fT3 concentration of any group, significantly lower than in surviving septic foals (P = 0.013). Serum fT3 concentrations were below the normal range in all (100%) septic nonsurvivors compared to 84% of septic survivors.

Median rT3 concentrations were not different between groups of foals (P = 0.69; Table 2), or between surviving and nonsurviving septic foals (P = 0.66). However, 69% of nonsurviving septic foals had a rT3 value above the normal range compared to 40% of surviving septic foals.

There was no statistical difference between survivors and nonsurvivors for any TH in the sick nonseptic group.

Hormone concentrations in premature foals

Serum TT4, fT4, TT3 and fT3 concentrations were lower in premature foals than in healthy foals (P<0.001; Table 2). There were no significant differences between TT4 (P = 0.8), fT4 (P = 0.35), TT3 (P = 0.1), fT3 (P = 0.12) and rT3 (P = 0.66) concentrations between foals that were premature and septic, and foals that were premature and not septic (Table 3). Serum fT3 concentrations were lower in nonsurviving than in surviving premature septic foals but not statistically different (P = 0.09), whereas rT3 concentrations were significantly higher in nonsurviving premature septic foals than in surviving premature septic foals (P = 0.026; Table 3).

Hormone concentrations and correlations

All TH concentrations were negatively correlated with the sepsis score (P<0.01) in septic foals (TT4, ρ= -0.4; fT4, ρ= -0.35; TT3, ρ= -0.4; fT3, ρ= -0.38). Serum rT3 was positively correlated with sepsis score in septic foals (ρ= 0.3; P = 0.05). Sepsis score was not correlated with TH concentrations in sick nonseptic or healthy foals (P>0.05). There was no correlation between foal age or gender and any TH concentration for any group of foals (P>0.05).


Critically ill foals had decreased concentrations of all TH except reverse T3. These changes in TH were also correlated to the sepsis score in septic foals. We found that nonsurviving septic foals had lower concentrations of fT3, TT3, fT4 and TT4 than surviving septic foals. Results of this study suggest that presumptive NTIS is frequent in critically ill foals and that low TH concentrations are associated with mortality. Similar results were recently shown in a metanalysis evaluating NTIS in critically ill children (Angelousi et al. 2011) and in dogs with severe sepsis and parvoviral enteritis, in which T4 concentrations were low and associated with mortality (Schoeman and Herrtage 2008).

Nonthyroidal illness syndrome is thought to represent an adaptive response to systemic inflammation in order to reduce metabolic rate and prevent organ failure and death (Singer et al. 2004). NTIS has been characterised as having low TT3, increased rT3 and/or low TT4 with normal fT4 and no evidence of thyroid disease (Warner and Beckett 2010). Increased rT3 levels suggest that affected cases do not have true hypothyroidism, as low rT3 and increased TSH would be expected with tertiary (thyroid dependent) hypothyroidism.

Several mechanisms have been proposed for low TH concentrations during illness, including alterations of the HPT axis, which may begin in peripheral tissues by decreased TH conversion to T3 and progress to decreased pituitary TSH secretion (Meyer et al. 2010). Leptin-induced hypothalamic inhibition resulting in decreased TRH production, reduced plasma thyroid hormone binding capacity and alterations in thyroid hormone receptor expression are also contributing events associated with human NTIS (Warner and Beckett 2010). Recently, Barsnick et al. (2011) showed that critically ill foals had decreased leptin concentrations. These data would suggest that probable HPT suppression observed in the foals reported here is unlikely to be mediated through leptin responses.

Increased T4 deiodination to rT3 instead of fT3 has also been documented, possibly due to a decrease in 5′-monodeiodinase activity (Adler and Wartofsky 2007). Other mechanisms for decreased thyroid hormone concentrations include increased T3 catabolism to 3,3-diiodothyronine (T2) and decreased T4-binding globulin or transthyretin production (Angelousi et al. 2011). We did not assess thyroid binding proteins or T2 concentrations; however measurement of these analytes might provide more information regarding the pathogenesis of NTIS in severely ill foals.

Proinflammatory cytokines have been shown to inhibit thyroid function through direct and indirect pathways (Van der Poll et al. 1990). In septic rats, decreased thyroid hormones were associated with increased interleukin (IL)-6, tumour necrosis factor-α and IL-1β production and consequent inhibition of 5′-deiodinase activity. Moreover, increased iodine utilisation by activated neutrophils has also been observed and purported to contribute to decreased TH concentrations through a relative iodine deficiency (Inan et al. 2003).

In foals of the study reported here an increase in rT3 was not observed, which is contrary to what was expected. Reverse T3 is the inactive form of T3 and is typically produced in small amounts. Reverse T3 tends to increase with illness and malnutrition, related to increased T4 conversion to rT3 and/or decreased hepatic rT3 clearance (De Groot 2006). It is possible that an increase was not seen because blood samples were taken only on admission and prior to a rise in rT3 concentration.

We found that premature foals had decreased concentrations of all thyroid hormones except rT3, supporting HPT axis hypofunction. Silver et al. (1991) showed that premature foals had decreased TT3 (and cortisol) concentrations compared to full-term foals. These findings are also consistent with transient hypothyroxinaemia of prematurity (THOP) in preterm human babies (Dilli et al. 2010). Other studies have shown an association of THOP with visual deficits (Rovet and Simic 2008) and neurocognitive disabilities in childhood (Delahunty et al. 2010). We only reported on the concentration of TH at admission in premature foals; however, following these foals into maturity may be useful to determine the clinical significance of perinatal thyroid hypofunction in horses.

Premature septic foals had a lower survival rate than full-term septic foals (50% vs. 72%). Similar results are reported in preterm human infants with sepsis, which were 5.5 times more likely to die than full-term septic infants (Ogunlesi and Ogunfowora 2010). Similarly, all TH concentrations were lower and rT3 was higher in nonsurviving premature and septic foals than in surviving premature and septic foals. Compared to full-term septic foals, premature septic foals had lower TH concentrations, with the exception of rT3. These data suggest that prematurity, as previously described (Silver et al. 1991), in addition to NTIS associated with critical illness, is an important influence on HPT function in this population.

We did not have a standardised population of critically ill foals with regard to age, breed, maturity, disease severity or treatment protocol; several of which can allow for bias. Previous studies evaluating endocrine variables in foals have shown the effect of foal age and hormone concentrations (Irvine and Evans 1975; Hart et al. 2011). Ideally we would have age-matched populations of foals to compare TH concentrations between septic, sick nonseptic and healthy foals to more accurately describe the effect of illness on TH concentrations. Furthermore, we cannot exclude the effect of goitrogen exposure in mares of the foals reported here to account for changes in TH concentration observed. However, the region where the study was performed has a low incidence of congenital hypothyroidism in foals seen in the clinics.

In the study reported here, we stratified foals based on disease severity as determined by the clinical findings and sepsis score (Brewer and Koterba 1988) as well as maturity. However, the utility of scoring systems to select patients with sepsis has inherent flaws, with a variable sensitivity of 67–93% in neonatal foals (Brewer and Koterba 1988; Corley and Furr 2003) and might have led to incorrect grouping of foals in this study and misrepresentation of results as evidenced by only modest correlations between sepsis score and TH concentrations.

We did not measure hormones of the hypothalamus (TRH) or adenohypophysis (TSH) in conjunction with TH or concurrent proinflammatory cytokine concentrations. Measuring these analytes would have been useful to determine overall HPT axis (dys)function in relation to critical illness. The data we report are similar to those generated in the evaluation of thyroid dysfunction in critically ill human patients and offer similar value; however, sequential sampling or HPT axis stimulation testing (i.e. TSH stimulation test) would be a useful adjunct in HPT axis assessment in age-matched critically ill foals.

Critically ill foals have decreased concentrations of thyroid hormones compared to healthy foals except for rT3. However, one must be cautious to make definitive conclusions based on admission TH concentrations and outcome in septic foals, as other systemic influences, disease states and treatment protocols can affect the prognosis for survival. As such, a recommendation regarding hormone supplementation in the treatment of critically ill foals remains unknown.

Conflicts of interests

No conflicts of interest have been declared.

Sources of funding

Funding for M. Himler was provided by the Morris Animal Foundation.


We thank all of the technical staff and veterinarians at the Galbreath Equine Center of The Ohio State University and Hagyard Equine Medical Institute (Lexington, Kentucky) for their assistance with this project. Special mention to Brandy Marlow, Krista Hernon and Dr Eason Hildreth for assistance with laboratory techniques.

Manufacturers' addresses

a Boehringer Mannheim/Hitachi 911 system, Mannheim, Germany.

b Siemens Medical Solutions Diagnostics, Los Angeles, California, USA.

c Adaltis Italia S.p.A., Rome, Italy.

d Excel, Microsoft Corporation, Mountain View, California, USA.

e Systat, Chicago, Illinois, USA.