Impact of Heat Stress on the In-Line Registered Milk Fat-to-Protein Ratio and Metabolic Profile in Dairy Cows

Abstract

The aim of our study was to investigate and quantify the impact of heat stress on the milk fat-to-protein ratio (F/P) and the metabolic profile in dairy cows, utilizing in-line registration methods for accurate and real-time data collection. This study was carried out in Lithuania at coordinates 55.819156, 23.773541, from 1 June 2023 until 31 August 2023. Milk composition, including fat and protein, was measured using a BROLIS HerdLine in-line milk analyzer from Brolis Sensor Technology, Vilnius, Lithuania. During the general clinical examinations (twice per week), blood samples were collected and determined for GLU—blood glucose concentration; BHB—blood β-hydroxybutyrate concentration; AST—aspartate transaminase activity; GGT—gamma-glutamyltransferase activity; and NEFAs—non-esterified fatty acids. The parameters based on the Temperature–Humidity Index (THI) were categorized into two groups: group 1, consisting of THI values less than 72, representing the comfort zone, and group 2, with THI values of 72 or higher, indicating a greater risk of thermal stress. Specifically, group 2 exhibited an 8.6% increase in the F/P ratio compared to group 1 (p = 0.049). Additionally, there was a 4.2% decrease in glucose levels in group 2 (p = 0.056) and a notable 5.8% decrease in albumin levels compared to group 1 (p < 0.001). We found a very weak, non-significant correlation between humidity and the milk fat-to-protein ratio (r = 0.043, p = 0.447) and a similarly negligible correlation with Beta-Hydroxybutyrate (BHB; r = 0.046, p = 0.417). We observed significant changes in milk composition, particularly an increase in the milk fat-to-protein ratio, and alterations in metabolic indicators like glucose, albumin, and liver enzymes. These changes, indicative of a negative energy balance and altered metabolic processes such as gluconeogenesis and lipolysis, correspond to previous research. The adoption of advanced tools, such as the BROLIS HerdLine analyzer, is recommended for the real-time monitoring of milk composition, which assists in the early detection of negative energy balances and metabolic issues. It is also crucial to adjust feeding practices to maintain energy balance during periods of high THI and to conduct regular health checks with a special focus on cows in early lactation.

1. Introduction

Recently, the rapid pace of global warming has had harmful impacts on agricultural regions, particularly affecting the livestock sector. Elevated temperatures and humidity levels are harmful to the performance of domestic animals [1]. Livestock exposed to heat stress from hot and humid conditions experience negative effects on their performance, including reduced body weight, average daily gain, and growth rate [2]. Dairy cows, among various livestock species, are particularly vulnerable to environments with high temperatures and humidity, which significantly affect their overall productivity, most notably in milk production [3]. Heat stress (HS) poses a significant challenge to the dairy industry in various regions globally due to its detrimental impact on productivity and profitability [4].

Heat stress occurs when animals are unable to eliminate the heat produced by their metabolism and the surrounding environment, resulting in a disruption of their body’s thermal equilibrium [3]. This condition causes an elevation in several physiological indicators, such as rectal temperature, respiration rate, core body temperature, panting score, pulse rate, sweating rate, and heart rate [5]. The impact of heat stress is notably detrimental to milk production characteristics, reproductive efficiency, and overall health in dairy cattle [6], particularly in tropical regions where the climate is consistently hot and humid [7]. Heat stress negatively impacts animal well-being and productivity, posing a significant economic challenge to the worldwide dairy industry. A common reaction among animals experiencing heat stress is reduced feed consumption, which is likely an effort to lower metabolic heat production [1,8,9]. Heat stress leads to changes in the metabolites within the mammary glands of lactating dairy cattle, affecting processes such as glycolysis, lactose synthesis, ketone body formation, the tricarboxylic acid cycle, and the metabolism of amino acids and nucleotides. These alterations hinder the provision of necessary components for milk production in lactating Holstein cows [10]. As a result, heat stress modifies both the synthesis and the composition of milk by impacting the metabolic processes in the mammary gland tissues of lactating dairy cows [4]. Heat stress in dairy cows results in elevated core body temperatures, impacting fat production in the mammary gland. This condition leads to decreased dry matter intake and milk yield and modifies milk components, including fat, protein, lactose, and the percentages of solids-not-fat, among others [4]. The composition of milk is significantly changed during hyperthermia, showing that HS affects not only the total production of milk but also the synthesis of its components [3]. Heat stress is known to affect dairy cows by impacting milk production and composition, as well as by altering the overall amounts of saturated and unsaturated fatty acids present in the milk fat [11]. Heat stress evidently has a more pronounced direct impact on the production and composition of milk protein and casein than the indirect effect caused by reduced intake [12]. However, there are limited studies on how heat stress affects milk protein composition or protein fractions, with existing findings being somewhat inconsistent.

The Temperature–Humidity Index (THI), which serves as a measure of heat stress conditions, was initially formulated by Thom [13] and subsequently revised by Kibler [14], Yousef [15], and Mader et al. [16]. Additionally, milk and blood metabolites undergo changes during heat stress in dairy cows [17]. The acclimation process to HS induces a range of physiological, endocrine, and biochemical alterations in dairy cattle [18].

Milk serves as an ideal medium for diagnostic purposes due to its non-invasive and straightforward collection process. It is commonly utilized for detecting ketosis and other production issues [19]. The frequent and non-invasive acquisition of milk during regular milking sessions, coupled with established standard analysis methods, underscores its diagnostic utility [20]. Therefore, further exploration into the use of early milk samples and the potential integration of new data sources is warranted to enhance early warning systems. To ensure the model’s applicability and adaptability to different management practices and diets, it is essential to incorporate and evaluate additional data from various farms before its practical implementation [21]. Recently, technologies that record behavior have become commercially accessible, paving the way for advancements in precision cattle husbandry [22]. The daily milk fat-to-protein ratios of cows can be measured using a BROLIS HerdLine in-line milk analyzer (Brolis Sensor Technology, Vilnius, Lithuania). This device utilizes an innovative GaSb widely tunable external cavity laser-based in-line spectrometer operating within the 2100–2400 nm spectral range [23]. Our past findings suggest that, by comparing metabolic data from blood and milk, the milk fat-to-protein ratio emerges as a reliable metric for evaluating the metabolic health of cows. The properties of milk serve as a crucial indicator of metabolic stress in cows, correlating with signs of lipolysis and ketogenesis in their blood. The non-invasive nature of milk sampling makes it an ideal method for regularly assessing the metabolic status of cows [23]. Incorporating this approach could significantly enhance herd health programs on dairy farms. By monitoring the energy status of individual cows through this method, farmers can identify cows that are susceptible to metabolic stress [23]. Contemporary dairy farming often involves intense milk production, leading to health complications in cows [24].

The hypothesis of this study is that heat stress significantly affects the milk fat-to-protein ratio in dairy cows, as measured through in-line registered data, and alters their metabolic profile. According to this hypothesis, the aim of our study was to investigate and quantify the impact of heat stress on the milk fat-to-protein ratio and the metabolic profile in dairy cows, utilizing in-line registration methods for accurate and real-time data collection.

2. Materials and Methods

2.1. Facilities, Cows, and Research Methodology

This study was carried out in compliance with the Lithuanian Law on Animal Welfare and Protection, under the approval number PK012858. It took place in Lithuania at coordinates 55.819156, 23.773541, from 2023.06.01 until 2023.08.31. All dairy cows (n = 1200) were housed in a free-stall barn and fed a Total Mixed Ration (TMR) tailored to their physiological requirements throughout the year. Feeding times were set at 06:00 and 18:00 daily, offering a TMR primarily composed of 50% grain concentrate mash, 6% alfalfa hay, 10% grass silage, sugar beet pulp silage, 30% corn silage, 4% grass hay wheat straw, and compound feed, suitable for high-producing, multiparous cows. The diets were designed to fulfill or surpass the nutritional needs of a 600 kg Holstein cow producing on average 37 kg of milk per day. The ration’s chemical composition included 48.8% dry matter (DM), 28.2% neutral detergent fiber (of DM), 19.8% acid detergent fiber (of DM), 38.7% nonfiber carbohydrates (of DM), 15.8% crude protein (of DM), and a net lactation energy of 1.6 Mcal/kg. Milking occurred twice daily at 05:00 and 17:00 using a parlor system.

Milking was conducted twice daily using a DeLaval (DeLaval International AB., Tumba, Sweden) milking parlor, and the cows were housed in a free-stall barn. The average body weight of the cows was 550 kg, with a variance of ±45 kg. They were kept in well-ventilated free-stall barns. The average yield per cow per lactation, with 4.2% fat and 3.6% protein, was 12,500 kg. The average number of inseminations was 2.2, the pregnancy rate was 85%, and the interval between calving was 395 days.

2.2. Measurements

Milk composition, including fat and protein, was measured using a BROLIS HerdLine in-line milk analyzer from Brolis Sensor Technology, Vilnius, Lithuania. This system incorporates a GaSb widely tunable external cavity laser-based in-line spectrometer operating within the 2100–2400 nm spectral range. The milk flow is continuously monitored in transmission mode during each milking cycle. Molecular absorption spectra are analyzed to determine the levels of primary constituents, effectively turning it into a compact on-site laboratory for the dairy farm. The analyzer continuously assesses the milk composition of each cow during milking. This compact “mini-spectroscope” is installed in the milking stalls or on the milking robot within the milk line (Figure 1).

After a general clinical examination (twice per week), blood samples were collected using evacuated tubes without anticoagulants (BD Vacutainer^®, Eysin, Switzerland) and centrifuged for 10–15 min at 3500 RPM. A total of 319 blood samples were collected. These samples were then sent to the Laboratory of Clinical Tests at the Large Animal Clinic of the Lithuanian University of Health Sciences Veterinary Academy. Blood serum analyses were conducted using a Hitachi 705 analyzer (Hitachi, Tokyo, Japan) and DiaSys reagents (Diagnostic Systems GmbH, Berlin, Germany) to measure activities of GGT and AST and albumin concentrations. NEFA levels were assessed with an Rx Daytona automated wet chemistry analyzer (Randox Laboratories Ltd., London, UK) and corresponding reagents.

Plasma levels of BHB and glucose were determined using the Medi Sense and Free Style Optium H systems (Abbott, Great Britain), with capillary blood samples taken from the ear. BHB concentrations in the blood samples were checked daily, with all the samples obtained as part of a clinical evaluation. Whole-blood BHB concentrations were used. To ascertain the highest possible BHB levels, samples were collected twice weekly, consistently timed in relation to feeding on each farm, and within 2 to 4 h after the delivery of new feed [25]. For each sample, the cows were positioned in a resting stall or headlock to facilitate the collection of a small blood sample from the coccygeal vein using a syringe with a needle.

Milk composition, including fat, protein, and blood parameters, was registered once for each cow during the period from 1 June 2023 to 31 August 2023 at different THI levels. Apart from THI, other conditions (feeding, housing, milking, etc.) remained the same throughout the investigated period.

2.3. Group Creation and THI Calculation

Out of 1200 cows, 319 were selected. The research involved cows in their second or more lactation, producing an average of 35 kg of milk daily per cow. These cows consumed an average of 19 kg of DM as feed each day. The milk characteristics included 4.3% fat, 3.55% protein, an average somatic cell count of 190,000/mL, and an average milk urea nitrogen content of 22%. According to the results of clinical examinations, all cows were categorized as clinically healthy.

The THI was calculated according to the formula THI = (0.8 × T_db) + [(RH/100) × (T_db − 14.4)] + 46.4 [16]. The parameters based on the THI were categorized into two groups: group 1, consisting of THI values less than 72, representing the comfort zone, and group 2, with THI values of 72 or higher, indicating a greater risk of thermal stress. THI groups were created according to Gantner et al. [26].

2.4. Statistical Analysis

All statistical evaluations were conducted using SPSS 25.0 software (IBM Corp., 2017), specifically version 25.0 of IBM SPSS Statistics for Windows (Armonk, NY, USA). The Shapiro–Wilk test was applied to verify the normal distribution of the data. Results were presented as the mean standard error of the mean (M S.E.M.). A significance level of 0.05 (p < 0.05) was established for probability. To establish a statistical relationship between the variables under study, Pearson correlation analysis was performed.

3. Results

3.1. Summary Statistics for the Examined Parameters

We observed significant differences between the groups in terms of the F/P ratio and the glucose and albumin concentrations. Specifically, group 2 exhibited an 8.6% increase in the F/P ratio compared to group 1 (p = 0.049) (Figure 2). Additionally, there was a 4.2% decrease in glucose levels in group 2 (p = 0.056) (Figure 3) and a notable 5.8% decrease in albumin levels compared to group 1 (p < 0.001) (Figure 4). Non-significant results were seen in body temperature, BHB, AST, GGT, and NEFAs. Specifically, BHB levels in group 2 increased by 4.8% (p = 0.685), body temperature decreased by 2.6% (p = 0.752), AST levels decreased by 11.1% (p = 0.110), GGT levels decreased by 20.3% (p = 0.056), and NEFA levels increased by 11.5% (p = 0.337) (

Table 1. Summary statistics for milk fat-to-protein ratio and biochemical indicators.

Descriptives
		N	Mean	Std. Deviation	Std. Error	95% Confidence Interval for Mean		Minimum	Maximum
						Lower Bound	Upper Bound
F/P	1	276	1.35	0.33	0.02	1.31	1.39	0.61	2.75
	2	43	1.47	0.42	0.06	1.33	1.60	0.74	2.63
	Total	319	1.37	0.34	0.01	1.33	1.41	0.61	2.75
GLU	1	276	2.74	0.37	0.02	2.70	2.79	1.1	4.1
	2	43	2.63	0.33	0.05	2.527	2.73	2.0	3.8
	Total	319	2.73	0.36	0.02	2.69	2.77	1.1	4.1
BHB	1	276	0.41	0.29	0.01	0.38	0.45	0.1	2.5
	2	43	0.43	0.33	0.05	0.33	0.54	0.1	1.9
	Total	319	0.41	0.29	0.01	0.38	0.45	0.1	2.5
AST	1	276	106.01	42.32	2.54	100.99	111.02	50.9	478.2
	2	43	95.40	23.62	3.60	88.13	102.67	61.6	177.4
	Total	319	104.58	40.44	2.26	100.12	109.03	50.9	478.2
GGT	1	276	32.82	18.47	1.11	30.63	35.01	13	151
	2	43	27.28	10.30	1.57	24.11	30.45	10	65
	Total	319	32.07	17.68	0.99	30.12	34.02	10	151
NEFAs	1	276	0.40	0.28	0.01	0.36	0.43	0.08	1.65
	2	43	0.44	0.31	0.04	0.35	0.54	0.11	1.45
	Total	319	0.40	0.29	0.01	0.37	0.44	0.08	1.65
Albumin	1	276	35.76	2.29	0.13	35.48	36.03	25.3	40.4
	2	43	33.79	2.60	0.39	32.98	34.59	24.9	38.4
	Total	319	35.49	2.43	0.13	35.22	35.76	24.9	40.4

F/P—in-line milk fat-to-protein ratio; GLU—blood glucose concentration; BHB—blood β-hydroxybutyrate concentration; AST—aspartate transaminase activity; GGT—gamma-glutamyltransferase activity; NEFAs—non-esterified fatty acids.

).

3.2. Correlation of the Examined Parameters

In the results section of this paper, we explored correlations between the THI and various parameters. We found a very weak, non-significant correlation between humidity and the milk fat-to-protein ratio (r = 0.043, p = 0.447) and a similarly negligible correlation with BHB (r = −0.046, p = 0.417). However, there was a weak negative correlation with glucose (GLU) (r = −0.160, p = 0.005), suggesting an inverse relationship.

No significant correlations were observed with body temperature (r = 0.026, p = 0.652), AST (r = −0.035, p = 0.538), GGT (r = −0.060, p = 0.286), and non-esterified fatty acids (NEFAs) (r = 0.054, p = 0.333). Notably, a moderate negative correlation was found with albumin (r = −0.284, p < 0.001), indicating a stronger relationship (

Table 2. Correlation of the examined parameters.

		THI	T	H	F/P	GLU	BHB	AST	GGT	NEFAs	Albumin
THI	Pearson correlation	1	0.977 **	0.199 **	0.043	−0.160 **	−0.046	−0.035	−0.060	0.054	−0.284 **
	sig. (2-tailed)		<0.001	<0.001	0.447	0.005	0.417	0.538	0.286	0.333	<0.001
	N	319	319	319	312	308	309	319	318	319	319
T	Pearson correlation	0.977 **	1	−0.004	0.044	−0.121 *	−0.077	0.005	−0.040	0.045	−0.263 **
	sig. (2-tailed)	<0.001		0.946	0.440	0.034	0.176	0.936	0.480	0.423	<0.001
	N	319	319	319	312	308	309	319	318	319	319
H	Pearson correlation	0.199 **	−0.004	1	0.012	−0.116 *	0.168 **	−0.181 **	−0.086	0.081	−0.143 *
	sig. (2-tailed)	<0.001	0.946		0.836	0.041	0.003	0.001	0.124	0.151	0.011
	N	319	319	319	312	308	309	319	318	319	319
F/P	Pearson correlation	0.043	0.044	0.012	1	0.150 **	0.206 **	−0.057	−0.038	0.583 **	−0.220 **
	sig. (2-tailed)	0.447	0.440	0.836		0.009	<0.001	0.318	0.505	<0.001	<0.001
	N	312	312	312	312	302	303	312	311	312	312
GLU	Pearson correlation	−0.160 **	−0.121 *	−0.116 *	0.150 **	1	−0.294 **	−0.028	−0.016	0.048	−0.102
	sig. (2-tailed)	0.005	0.034	0.041	0.009		<0.001	0.622	0.782	0.400	0.075
	N	308	308	308	302	308	308	308	307	308	308
BHB	Pearson correlation	−0.046	−0.077	0.168 **	0.206 **	−0.294 **	1	0.062	−0.037	0.235 **	0.037
	sig. (2-tailed)	0.417	0.176	0.003	<0.001	<0.001		0.278	0.516	<0.001	0.520
	N	309	309	309	303	308	309	309	308	309	309
AST	Pearson correlation	−0.035	0.005	−0.181 **	−0.057	−0.028	0.062	1	0.457 **	0.080	0.129 *
	sig. (2-tailed)	0.538	0.936	0.001	0.318	0.622	0.278		<0.001	0.156	0.021
	N	319	319	319	312	308	309	319	318	319	319
GGT	Pearson correlation	−0.060	−0.040	−0.086	−0.038	−0.016	−0.037	0.457 **	1	0.039	0.091
	sig. (2-tailed)	0.286	0.480	0.124	0.505	0.782	0.516	<0.001		0.486	0.105
	N	318	318	318	311	307	308	318	318	318	318
NEFAs	Pearson correlation	0.054	0.045	0.081	0.583 **	0.048	0.235 **	0.080	0.039	1	−0.066
	sig. (2-tailed)	0.333	0.423	0.151	<0.001	0.400	<0.001	0.156	0.486		0.240
	N	319	319	319	312	308	309	319	318	319	319
Albumin	Pearson correlation	−0.284 **	−0.263 **	−0.143 *	−0.220 **	−0.102	0.037	0.129 *	0.091	−0.066	1
	sig. (2-tailed)	<0.001	<0.001	0.011	<0.001	0.075	0.520	0.021	0.105	0.240
	N	319	319	319	312	308	309	319	318	319	319

F/P—in-line milk fat-to-protein ratio; GLU—blood glucose concentration; BHB—blood β-hydroxybutyrate concentration; AST—aspartate transaminase activity; GGT—gamma-glutamyltransferase activity; T—ambient temperature; H ambient humidity; NEFAs—non-esterified fatty acids. ** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

).

4. Discussion

The THI is commonly employed for forecasting heat stress occurrences in cattle. It was initially established that a THI score of 72 marks the onset of heat stress [27]. Our aim for this study was to investigate and quantify the impact of heat stress on the milk fat-to-protein ratio and the metabolic profile in dairy cows, utilizing in-line registration methods for accurate and real-time data collection. Comparisons were made in the milk fat-to-protein ratio, as registered in-line, using a BROLIS HerdLine analyzer (Brolis Sensor Technology, Vilnius, Lithuania). The properties of milk play a vital role in predicting metabolic stress in cows as they align with signs of lipolysis and ketogenesis found in cow blood. The non-invasive nature of milk sampling renders it an appropriate tool for the regular assessment of metabolic health [23].

We observed a higher F/P ratio in cows experiencing a higher risk of heat stress. Rhoads et al. [28] found an increase in milk fat concentration as a result of heat stress. The literature indicates that the milk fat-to-protein ratio is frequently used for identifying energy deficiencies [28] or subclinical ketosis [21]. An increase in this ratio is associated with a negative energy balance (NEB) and the mobilization of adipose tissue. Milk F/P ratios exceeding 1.35 to 1.50 are indicative of cows experiencing an energy deficit [29]. In early-lactation cows, blood non-esterified fatty acid (NEFA) levels are significantly higher compared to cows in mid-, full, and late lactation, making it an effective indicator of an NEB and lipid mobilization during lactation [30]. Furthermore, blood and milk serum levels of Beta-Hydroxybutyrate, additional markers of energy metabolism, are notably higher in early-lactation cows, signifying the intense mobilization of fat reserves [31]. During hyperthermia, milk composition undergoes significant changes, suggesting that heat stress (HS) influences the synthesis of various components, in addition to its known impact on milk production [3]. Specifically, HS leads to a reduction in the content and yield of milk protein, although the exact processes behind this remain largely unexplored. Rhoads et al. [32] proposed that slight alterations in the somatotropic axis might account for a small part of the decrease in milk protein yield during HS. Additionally, Cowley et al. [12] found that the decrease in milk protein in cows under heat stress is due to a targeted reduction in the synthesis of mammary proteins, rather than merely a side effect of the overall decline in milk production. We found that during a higher risk of heat stress, there was a decrease in the glucose level.

According to the literature, heat-stressed cows exhibited a 7% reduction in glucose concentrations, mirroring the gradual decline observed over time [32]. This pattern closely aligns with the findings from previous climate-controlled heat stress studies, such as those by Shwartz et al. [33], and is consistent with the results from earlier ruminant heat stress experiments [34]. Heat stress has been shown to decrease the total rumen Volatile Fatty Acid (VFA) content [35] and specifically alter the molar ratio of propionate to acetate [36]. Nonetheless, the similar glucose pattern in cows suggests that this reduction is likely due to decreased hepatic propionate delivery. Propionate is a key precursor in ruminant gluconeogenesis [37]. According to the literature, the numerous post-absorptive metabolic alterations observed in high-stride cows are significant. These include increased insulin activity and a diminished ability to mobilize adipose tissue. Consequently, this leads to an impaired capacity to recruit glucose-sparing mechanisms [38]. Production losses are caused by these changes, and they are greater than those incurred by cows with inadequate nutrition. Similar causes and effects are also demonstrated for growth metrics, with the HS-induced decrease in feed intake accounting for a significant portion of the explanation [39]. It has been demonstrated that HS cows respond well to insulin in a glucose tolerance test and have higher basal insulin concentrations [38]. Heat stress led to a mild decrease in the levels of serum glucose, cholesterol, and albumin as a result of a persistent negative energy balance [40]. Ronchi et al. [34] observed that cattle under heat stress exhibit a decrease in albumin secretion and liver enzyme production. Our findings indicate a reduction in albumin levels during the risk of HS.

Our findings suggest an influence of the environment on the negative energy balance. We observed a very weak and non-significant correlation between the Heat Index (HI) and the milk fat-to-protein ratio (r = 0.043, p = 0.447) and a similarly minimal correlation with Beta-Hydroxybutyrate (BHB) (r = −0.046, p = 0.417). However, a weak negative correlation was noted with glucose (GLU) (r = −0.160, p = 0.005), indicating a potential inverse relationship. The weak correlations observed may be attributed to the limited number of animals included in this study.

5. Conclusions

We observed significant changes in milk composition, particularly an increase in the milk fat-to-protein ratio, and alterations in metabolic indicators like glucose, albumin, and liver enzymes. These changes, indicative of a negative energy balance and altered metabolic processes such as gluconeogenesis and lipolysis, correspond to previous research. Overall, our findings underscore the critical role of managing heat stress in dairy cattle to maintain health and productivity and demonstrate the value of using innovative technologies in such research.

Our recommendations for practice are as follows: The adoption of advanced tools, such as the BROLIS HerdLine analyzer, is recommended for the real-time monitoring of milk composition, which assists in the early detection of negative energy balances and metabolic issues. It is also crucial to adjust feeding practices to maintain an energy balance during periods of high THI and to conduct regular health checks.