Establishment of biochemical reference values for backyard chickens in Colorado (Gallus gallus domesticus)

Abstract

Background

Previous literature pertaining to biochemical RIs of domestic chickens has primarily focused on commercial production flocks and not backyard birds.

Objective

We aimed to establish biochemistry RIs for privately‐owned backyard chickens (Gallus gallus domesticus) using reference laboratory equipment.

Methods

Samples were collected from 123 presumably healthy adult chickens between 2017 and 2019 from 22 different flocks in Colorado. Heparinized blood was obtained, and a biochemistry profile was evaluated, including sodium, potassium, chloride, calcium, phosphorous, uric acid, AST, CK, glucose, cholesterol, and total protein. Reference values were created according to current American Society for Veterinary Clinical Pathology recommendations.

Results

Differences in measurand intervals compared with previous literature were found for sodium, calcium, total protein, potassium, phosphorus, uric acid, and glucose. Hens were found to have higher median calcium (17.9 mg/dL vs 11.2 mg/dL [P = .0001]), total protein (5.2 g/dL vs 4.8 g/dL [P = .0046]), and potassium (3.80 mEq/L vs 3.48 mEq/L [P = .0267]) concentrations, as well as lower sodium (155 mEq/L vs 158 mEq/L [P = .0046]) concentrations, calculated osmolalities (310 mOsm/kg vs 314 mOsm/kg [P = .0249]), and AST (165 U/L vs 194 U/L [P = .0121]) activities, than roosters. Seasonal variation was found between summer and winter samples for median sodium (144 mEq/L vs 148 mEq/L [P = .0008]), chloride (111.8 mEq/L vs 113.5 mEq/L [P = .0033]) concentrations, calculated osmolalities (306 mOsm/kg vs 311 mOsm/kg [P = <.0001]), and AST (185 U/L vs 159 U/L [P = .0053]) and CK (1098 U/L vs 770 U/L [P = .0007]) activities.

Conclusions

This study presents biochemical reference values for backyard chickens in Colorado that can be a basis for evaluations in similar settings.

1. INTRODUCTION

Traditionally, poultry medicine has focused on flock health and pursued diagnostic testing that includes post‐mortem evaluation of a few, representative birds. However, chickens are more commonly kept in privately‐owned backyard flocks of domestic poultry and viewed as pets, not livestock. ¹ This shift requires that veterinarians are equipped to perform ante mortem diagnostic evaluations. Biochemical analysis is a simple but useful ante mortem diagnostic technique for assessing the health of poultry patients.

There are an estimated 500 chicken breeds worldwide. Of these, 53 breeds are recognized by the American Poultry Association, excluding the subclassification of large vs bantam varieties. ² Most published studies evaluating biochemical analyses have used selectively bred and genetically similar commercial production bird flocks and targeted production‐related outcomes, such as the effect of dietary changes. ³ , ⁴ , ⁵ These studies are likely not reflective of the physiologic and biochemical parameters of animals in backyard flocks. Backyard flocks are highly diverse, with many different breeds of chickens and other poultry commingling in a single flock. Biochemical RIs that reflect the diversity of privately owned, small backyard flocks using reference laboratory equipment are currently not available.

The goal of this study was to establish biochemical RIs for chickens in these non‐commercial flocks that could be used as a basis for transference studies and to provide insight into the health status of individual chickens.

2. MATERIALS AND METHODS

This direct a priori study was approved by the animal care and use committee at Colorado State University.

2.1. Chickens

The birds used in this study were from privately‐owned backyard flocks throughout Colorado. Most flock sizes ranged from four to 100 birds (mean = 33, maximum flock size was 800). The birds were fed commercially available layer feed, as well as other possible feed sources commonly fed to backyard flocks (eg, forage, table scraps, or other supplements). Samples were collected from a randomly selected subset of the visited flocks during regular flock health checks by the Colorado State University Avian Health Program. A priori inclusion criteria were sexually mature birds, older than 4 months of age, of both sexes (roosters and laying hens), any breed residing in northern Colorado (<250 miles of Fort Collins, CO), and birds with an unremarkable physical examination at the time samples were collected. Location, flock size, breed, sex, individual animal condition, overall flock health, amount of blood collected, and any additional flock notes were recorded at the time of sample collection.

Physical examination and blood collection were performed between June 1 and August 1, 2017‐2019, and November 1 and February 1, 2017‐2019, respectively. The birds were manually restrained while performing physical evaluations and venipunctures. The physical examination included palpation of the keel to determine body condition, examination of the mouth, and external examination of the head and body. One to 2 mL of blood were collected from the ulnar vein using a 1 or 3 mL syringe and a 22 or 25 gauge needle. Blood was transferred to a heparin microtainer in the field that approximated the appropriate recommended ratio of heparin (30 USP heparin/mL blood) and gently mixed prior to transport. All samples were put into a cooler immediately after sample collection and transported in insulated coolers on ice packs to maintain refrigerated temperatures between 2° and 8°C. Blood samples were transported to the Colorado State University Clinical Pathology Laboratory within 20 minutes typically and within a maximum of 6 hours after being obtained. On arrival at the laboratory, samples were centrifuged at 1500g for 10 minutes; samples with hemolysis were excluded. Plasma biochemistry profiles were performed on a commercially available biochemistry analyzer using commercially available reagents (Cobas c501; Roche Diagnostics). Measurands and analytic methods are listed in Table 1; all tests were performed according to manufacturer recommendations for mammalian samples. Prior to each analytical run, the analyzer was shown to be performing within laboratory established quality assurance standards that include the use of standard operating procedures and quality control procedures to evaluate commercially available quality control material (Bio‐Rad Multiqual 1 and Multiqual 3; Bio‐Rad) using consensus total allowable error, in compliance with current guidelines. ⁶ , ⁷ Samples were evaluated a single time. These methods have not been specifically validated in our laboratory for use in chickens but have been used previously for chickens and other avian species within the laboratory. ⁸ , ⁹

TABLE 1.

Analytical methods used to produce chicken biochemical reference intervals

Measurand	Method	Manufacturer
Sodium	Indirect Ion Specific Electrode (ISE)	Roche
Potassium	Indirect ISE	Roche
Chloride	Indirect ISE	Roche
Bicarb	Phosphoenolpyruvate carboxylase	Roche
Anion gap	Calculated
Osmolality (calculated)	Calculated
Calcium	BAPTA (2,2′,2′′,2′′′‐[Ethane‐1,2‐diylbis(oxy‐2,1‐phenylenenitrilo)]tetraacetic acid)	Roche
Phosphorus	Phosphomolybdate	Roche
Uric acid	Uricase	Roche
Aspartate aminotransferase (AST)	Ultraviolet (UV) without P5P (pyridoxal‐5‐phosphate)	Roche
Creatine kinase (CK)	N‐acetyl‐cystein (NAC) activated	Roche
Glucose	Hexokinase	Roche
Cholesterol	Cholesterol oxidase, esterase, peroxidase	Roche
Total protein	Biuret	Roche

Outliers were identified using the Dixon‐Reed and Tukey's tests with emphasis on retaining results; values identified as outliers using the statistical tests as performed by the Reference Value Advisor software were removed. ¹⁰ Descriptive statistics, including median, minimum, maximum, mean, and standard deviation, were calculated for each measurand individually. Distribution was classified using both statistical methods (Anderson‐Darling test, statistical significance P < .05 for n > 65, P < .25 for n < 65) and by visual evaluation of histograms. ¹¹ When appropriate, the 95% RIs were calculated using nonparametric statistical methods with 90% CI calculated using bootstrap methods. Current recommendations suggest that the width of the 90% CI should be <0.2 times the width of the RI; if this relationship is not present, it is recommended that additional samples should be obtained. ¹² Because of disparity in sample size and distribution of the datasets, the Mann‐Whitney test was used to compare results between hens and roosters and between seasons within hens. The need to partition by season was evaluated using the criteria described by Lahti; specifically, partitioning is recommended if >4.1% or <0.9% of a subgroup falls outside the upper or lower limits of a combined RI. ¹³ For the purposes of this study, samples collected from June 1 to August 1 were grouped as summer samples and samples collected from November 1 to February 1 were grouped as winter samples. The statistical methods used in this study followed the current best practices. ¹²

Statistical analyses were performed using Microsoft Excel (Microsoft Office 2016; Microsoft) and the Reference Value Advisor Software or GraphPad Prism 8 (GraphPad Software Inc). ¹⁰ , ¹²

3. RESULTS

There were 123 adult chickens from 19 flocks included in this study. This included 18 (15%) roosters and 105 (85%) hens, representing 46 distinct and mixed breeds. The breed was unknown for seven individuals. Breeds with three or more sampled individuals are recorded in Table 2. Samples were collected from 69 chickens during winter and 54 chickens during summer.

TABLE 2.

Summary of breed variation from the sampled population

Breed	Count
Rhode Island Red	10
Silkie	10
Ameraucana	8
Barred Rock	7
Mix	7
Buff Orpington	6
Brahma	5
Easter Egger Mix	4
Old English Standard	4
Bantam Brahma	3
Whiting Blue	3
Wyandotte	3

The 95% RIs were developed for all samples and are reported in Table 3.

TABLE 3.

Plasma biochemistry 95% RIs for backyard chicken in Northern Colorado (N = 123)

Measurand	Conventional units	n	Mean	SD	Median	Min	Max	P‐value	Distribution a	Method b	LRL of RI	URL of RI	LRL 90% CI	URL 90% CI
Sodium	mEq/L	119	155	3.8	155	148	165	.100	G	NP	148	163	148‐149	162‐165
Potassium	mEq/L	121	3.77	0.555	3.77	2.46	5.18	.795	G	NP	2.64	5.03	2.46‐2.84	4.65‐5.18
Chloride	mEq/L	116	112.8	3.22	113.1	104.0	120.5	.786	G	NP	105.4	119.5	105‐107.9	118.4‐120.5
Bicarb	mEq/L	114	25.0	2.62	25.05	18.1	31.0	.074	G	NP	19.1	29.7	18.1‐20.4	29.3‐31.0
Anion gap	mEq/L	118	21	3.60	20.0	11	30	.029	NG	NP	15	29	11‐15	27‐30
Osmolality (calc)	mOsm/Kg	119	310	7.5	310	290	327	.510	G	NP	296	325	290‐298	323‐327
Calcium	mg/dL	123	17.4	7.04	13.3	9.9	34.3	<.000	NG	NP	10.4	32.7	9.9‐10.5	30.5‐34.3
Phosphorus	mg/dL	118	4.01	1.29	4.0	1.7	7.3	.002	NG	NP	1.9	6.5	1.7‐2.3	6.3‐7.3
Uric acid	mg/dL	118	5.9	2.04	5.6	1.5	11.9	.003	NG	NP	2.9	10.5	1.5‐3.2	9.9‐11.9
AST	U/L	113	171	31.8	166	110	247	.032	NG	NP	116	241	110‐128	234‐247
CK	U/L	106	909	398.2	833	236	1857	<.000	NG	NP	316	1792	236‐367	1645‐1857
Glucose	mg/dL	116	246	20.8	245	195	301	.211	NG	NP	210	295	195‐216	283‐301
Cholesterol	mg/dL	117	118	34.6	115	47	200	.658	G	NP	61	186	47‐69	175‐200
Total protein	g/dL	116	5.1	0.66	5.1	3.5	6.7	.106	G	NP	3.6	6.5	3.5‐4.0	6.1‐6.7

^a G, Gaussian; NG, non‐Gaussian.

^b P, parametric; NP, nonparametric.

The Mann‐Whitney test documented that hens had higher median calcium (median difference 6.5 mg/dL, 95.08% CI 1.3‐10.8 mg/dL), total protein (median difference 0.35 g/dL, 95.07% CI 0.1‐0.8 g/dL), and potassium (median difference 0.36 mEq/L, 95.04% CI 0.06‐0.73 mEq/L) concentrations, and lower sodium (median difference 3 mEq/L, 95.03% CI 1‐4 mEq/L) concentrations, calculated osmolalities (cOsm; median difference 4.5 mOsm/kg, 95.07% CI 1‐7 mOsm/kg), and AST (median difference 29 U/L, 95.09% CI 5‐48 U/L) activities. Because the low number of roosters prevented the composition of RIs or an accurate evaluation of partitioning, RIs for hens were produced and compared with summary statistics for roosters, Table 4.

TABLE 4.

Comparison of plasma biochemical measurands in backyard hens and roosters and 95% RIs for hens

Measurand	Units	Gender	n	Mean	SD	Median	Min	Max	Mann‐Whitney P‐value	Anderson‐Darling P‐value	Distribution a	Method b	LRL of RI	URL of RI	LRL 90% CI	URL 90% CI
Sodium	mEq/L	Hen	101	155	3.8	155	145	163	.0046	.090	G	NP	148	163	145‐149	162‐163
		Rooster	18	157	3.0	158	152	163
Potassium	mEq/L	Hen	104	3.81	0.523	3.80	2.51	5.18	.0267	.976	G	NP	2.72	5.02	2.51‐2.92	4.59‐5.18
		Rooster	18	3.44	0.729	3.48	2.22	5.05
Chloride	mEq/L	Hen	98	112.7	3.31	112.9	105	120.5	.12	.534	G	NP	105.3	119.6	105.0‐107.5	118.6‐120.5
		Rooster	18	113.8	2.54	115.1	109.9	117.8
Bicarb	mEq/L	Hen	98	25.0	2.61	25.2	18.1	31	.955	.067	G	NP	18.9	29.9	18.1‐20.6	29.3‐31.0
		Rooster	18	24.8	3.78	24.8	15.8	31.6
Anion gap	mEq/L	Hen	102	21	3.7	20.0	11	30	.6	.032	NG	NP	15	29	11‐15	27‐30
		Rooster	16	21	2.9	21	15	26
Osmolality (calc)	mOsm/kg	Hen	102	310	7.6	310	290	327	.0249	.267	G	NP	296	325	290‐298	323‐327
		Rooster	17	314	6.1	314	302	326
Calcium	mg/dL	Hen	105	18.5	7.07	17.9	9.9	34.3	<.0001	<.000	NG	NP	10.55	32.9	9.9‐10.7	30.6‐34.3
		Rooster	18	11.1	0.56	11.2	10.0	12.1
Phosphorus	mg/dL	Hen	101	4.2	1.36	4.2	1.7	8.2	.0848	.008	NG	NP	2.01	7.3	1.7‐2.4	6.4‐8.2
		Rooster	18	3.6	1.05	3.4	1.9	5.9
Uric acid	mg/dL	Hen	100	5.9	2.07	5.7	1.5	11.9	.458	.019	NG	NP	2.8	10.8	1.5‐3.1	10.0‐11.9
		Rooster	18	5.6	1.93	4.9	3.5	9.9
AST	U/L	Hen	98	168	30.5	165	110	247	.0121	.123	G	NP	116	238	110‐128	226‐247
		Rooster	17	197	43.2	194	131	270
CK	U/L	Hen	92	894	397.6	833	236	1857	.371	.006	NG	NP	310	1824	236‐367	1660‐1857
		Rooster	14	1010	402	825	599	1655
Glucose	mg/dL	Hen	102	246	20.4	244	195	296	.178	.484	G	NP	210	294	195‐216	283‐296
		Rooster	16	260	32.2	254	220	323
Cholesterol	mg/dL	Hen	99	119	32.6	117	47	200	.197	.923	G	NP	59	188	47‐68	175‐200
		Rooster	17	109	20.6	110	72	151
Total protein	g/dL	Hen	93	5.2	0.56	5.2	3.9	6.6	.0046	.159	G	NP	4.0	6.4	3.9‐4.3	6.1‐6.6
		Rooster	18	4.7	0.54	4.8	3.7	5.6

^a G, Gaussian; NG, non‐Gaussian.

^b NP, nonparametric.

The use of the Mann‐Whitney test documented that summer samples had lower median sodium (median difference 3 mEq/L, 95.03% CI 1‐4 Eq/L), and chloride (median difference 1.7 mEq/L, 95.06% CI 0.7‐3.4 mEq/L) concentrations, and cOsm (median difference 5.5 mOsm/kg, 95.05% CI 3‐9 mOsm/kg), and higher AST (median difference 26 U/L, 95.05% CI 7‐36 U/L) and CK (median difference 328.5 U/L, 95.06% CI 151‐582 U/L) activities than winter samples, see Table 5. Partitioning was not warranted, and RIs were not produced for hens partitioned by season.

TABLE 5.

Comparison of plasma biochemical measurands in backyard hens between summer and winter

Measurand	Units	Season	n	Mean	SD	Median	Min	Max	Mann‐Whitney P‐value
Sodium	mEq/L	Summer	42	153	3.7	153	144	162	.0008
		Winter	59	156	3.6	156	148	163
Potassium	mEq/L	Summer	42	3.70	0.542	3.71	2.51	5.03	.1276
		Winter	61	3.86	0.474	3.87	2.82	5.02
Chloride	mEq/L	Summer	41	111.4	3.11	111.8	103.1	119.0	.0033
		Winter	59	113.5	3.68	113.5	105	122.4
Bicarb	mEq/L	Summer	40	25.3	2.70	25.8	20.6	31	.6377
		Winter	50	25.1	1.68	25.1	21.5	28.4
Anion Gap	mEq/L	Summer	42	21	4.1	20	14	31	.3936
		Winter	59	21	3.3	21	15	29
Osmolality (calc)	mOsm/kg	Summer	42	306	7.1	306	288	323	<.0001
		Winter	60	312	7.1	311	296	327
Calcium	mg/dL	Summer	42	184	6.38	18.9	10.7	32.8	.9468
		Winter	63	186	7.55	17.1	9.9	34.3
Phosphorus	mg/dL	Summer	41	4.1	1.35	3.8	1.7	6.5	.8002
		Winter	59	4.2	1.29	4.2	2.1	7.3
Uric Acid	mg/dL	Summer	37	5.9	1.42	5.9	3.5	8.6	.5118
		Winter	63	5.9	2.48	5.3	1.5	13.0
AST	U/L	Summer	40	187	43.7	185	110	280	.0053
		Winter	61	162	24.6	159	115	222
CK	U/L	Summer	39	1269	720	1098	320	2880	.0007
		Winter	59	823	357	770	236	1703
Glucose	mg/dL	Summer	42	242	24.4	238	195	307	.126
		Winter	58	247	15.3	246	216	288
Cholesterol	mg/dL	Summer	41	117	43.2	113	47	229	.1213
		Winter	60	125	28.8	124	74	200
Total protein	g/dL	Summer	39	5.3	0.86	5.2	3.5	7.2	.4756
		Winter	60	5.2	0.57	5.1	3.9	6.7

4. DISCUSSION

This work composed plasma biochemical reference values for privately owned chickens in a backyard, non‐commercial setting in normal Colorado. Clinicians working with backyard chickens can use this data as a starting point for the evaluation of clinical data to help characterize the physiologic status of their patients (Table 6).

TABLE 6.

Comparison of biochemical values with previously reported reference values from production (Bowes), backyard (Board), and unspecified (Carpenter) flocks

Measurand	Units	Current study	Carpenter	Bowes (mean female broiler)	Board
Sodium	mEq/L	148‐163	131‐171	151	133‐151
Potassium	mEq/L	2.64‐5.03	3‐7.3	4.83	3.2‐6.1
Chloride	mEq/L	105.4‐118.5	—	112	—
Calcium	mg/dL	10.4‐32.7	13.2‐23.7	11.4	>10.9
Phosphorus	mg/dL	1.9‐6.5	6.2‐7.9	8.2	1.6‐7.2
Uric acid	mg/dL	2.9‐10.5	2.5‐8.1	6.6	0.9‐8.9
AST	U/L	116‐241	—	199	118‐298
CK	U/L	316‐1792	—	—	107‐1780
Glucose	mg/dL	210‐295	—	264	174‐239
Cholesterol	mg/dL	61‐186	86‐211	121	—
Total protein	g/dL	3.6‐6.5	—	3.42	3.9‐7.0

The analytical methods used in this study are not generally considered the gold standard for the evaluation of mammal samples and have not been specifically validated for use in chickens. Indeed, some methods used in this study are known to be a poor reflection of the methods deemed to be the gold standard but have been used when evaluating the performance of point of care testing assays. ¹⁴ , ¹⁵ As these methods have not been validated, we feel it is wise to view the results not as an accurate reflection of the measurand concentration in these samples but as the results obtained by using the mammalian assay on chicken samples. The production of RIs using these methods should allow for further evaluation of the clinical utility of these methods in chickens.

Variables that might influence biochemistry values include but are not limited to breed, season, environment, nutritional status, reproductive status, age, and sex. This study design produced a reference population that reflects the breadth of individuals within backyard flocks, with representatives of a variety of husbandry settings, representatives of both sexes, and a sampling of the encountered breeds. This design should allow for the evaluation of a wider variety of individuals; however, narrower RIs than could have been obtained if more partitioning was possible.

Breed‐specific RIs have been documented in other species, and it is likely that backyard chickens may warrant breed‐specific RIs, but these could not be assessed in this study. ¹⁶ , ¹⁷ The reference population consisted of 46 different breeds; therefore, the generation of breed‐specific reference values should be pursued.

The data were compatible with differences between hens and roosters, which can be explained by egg‐laying. Hens had significantly different electrolyte, calcium, and total protein concentrations; therefore, partitioned data was justified for hens. Ideally, RIs partitioning for roosters would be produced. The limited number of roosters allowed for the presentation of reference values but not RIs. Reference values are the numerical results derived from the reference population and can appropriately be presented when a small reference population prevents reasonably accurate statistical calculation of the reference limits that define a RI. The production of rooster‐specific RIs is recommended.

The data were compatible with differences between samples collected in summer vs winter. Partitioning for the season within this population of laying hens was evaluated and was found to not be warranted based on this data. An evaluation of the need for reference values partitioned by egg‐laying and non‐egg‐laying hens might help to provide a better representation of non‐egg‐laying hens than the evaluation of season alone.

These results were generated on commercially available reference laboratory equipment. The width of the 90% CI was <20% of the width of the RI in over half of the measurands; while this study presents the largest reference population, the accuracy of the limits of RIs can be strengthened by the evaluation of a larger population. Use of these reference values in other geographic regions, laboratories, or with other equipment using the same methods, or the same measurands assayed using different methods, requires at least a transference study to confirm that the values are representative of the population in the new setting. ¹²

Currently, the literature for domestic chickens (Gallus gallus domesticus) is limited to that of commercial flocks and the Exotic Animal Formulary ¹⁸ and a recent study of backyard chickens in Washington state. ⁸ While the Exotic Animal Formulary is a helpful guide, its values are derived from Johnson‐Delaney (1996), ¹⁹ and the origin of these values could not be traced. A study ⁸ in 2019 was the first study to produce results that described the backyard chicken demographic using flocks located in the Pacific Northwest and conducted using the VetScan VS2 (Abaxis, Union City, CA, USA) with Avian/Reptilian Profile Plus. When compared with our study, there is a slight difference in every measurand RI, but this variation might have been a result of either geographic variation of backyard poultry, variation between benchtop and laboratory biochemical analyzers, or a combination of both. Our biochemical profile included chloride, cholesterol, and a full RI for calcium, which was not included in the 2019 study.

The reference population used in this work was intended to encompass the effects of breed and husbandry differences. This will produce wider RIs that may not reflect all populations perfectly but should produce biochemical reference values representative of the general population of healthy backyard chickens in Colorado. They can be used as a basis for transference studies to generate RIs in other laboratories and settings.

DISCLOSURE

The authors have indicated that they have no affiliations or financial involvement with any organization or entity with a financial interest in, or in financial competition with, the subject matter or materials discussed in this article.

Kaiser JC, Reider H, Pabilonia KL, Moore AR. Establishment of biochemical reference values for backyard chickens in Colorado (Gallus gallus domesticus). Vet Clin Pathol. 2022;51:577‐584. doi: 10.1111/vcp.13136