The genetic selection to the fast growth of broilers resulted in a proportional increase of feed intake by broiler breeders, causing a greater intake of nutrients and energy above the energy requirements for production and maintenance [1]. The high feed intake increases the fat deposition, carrying to body weight (BW) increase [2–4]. Due to higher feed intake, several metabolic disturbances can occur and reduce egg production, fertility, and hatchability [2–5]. According to Renema and Robinson [6], to reduce the excessive weight gain and incidence of reproductive disorders, broiler breeders must be submitted to restricted feeding programs. This feeding system aims to meet birds' nutritional requirements by providing reduced feed to control the weight variation and favour sexual development [2,7]. For this, the calculation of dietary nutrients must consider the amount of feed allocated, obtaining an adequate balance of nutrients and energy [8]. Also, adequate body composition should be achieved so the birds can withstand high reproductive rates [9]. In this way, limiting the regulation of feed intake according to their nutritional needs becomes critical. It depends exclusively on the feeding strategies adopted by specialists.

On the other hand, ad libitum feeding can be an alternative, where the diets' energy and nutritional density must be adjusted (qualitative restriction) to prevent over-feeding. Still, this feeding system presents several disadvantages, once the weight and body composition variations lead to alteration in the dynamic of energy metabolism and, consequently, in maintenance utilization [10,11]. Therefore, it was hypothesized that the feed regimen, depending on the previous feed intake level, affects the maintenance requirement measured throughout fasting heat production (FHP).

The determination of energy requirement for maintenance through the FHP [12] was barely described in poultry, and principally it was estimated with ad libitum fed birds. Thus, there is little information available to broiler breeders. Based on this, the broad understanding of bird's metabolic variations on feed restriction will favour the correct determination of maintenance, supporting future studies in the dynamic of energy metabolism studies in broiler breeders.

We aimed to evaluate the effect of feed regimen (restricted and ad libitum) on the FHP in broiler breeders in the production phase and describe their body components variation (protein, lipids, and minerals).

The performance parameters are shown in Table 1. At the beginning of the adaptation period to the treatments, the initial BW was similar between the birds distributed for each treatment (p>0.05), with an average weight of 4.003±0.230 kg for all birds. Also, the BW after the fasting period showed a tendency (p<0.10) to have a higher weight of birds fed ad libitum (4.153±0.349 kg) than birds under restricted feed (3.839±0.160). On the other hand, weight variation between the beginning and the final fasting period was significantly different (p<0.01) between treatments. The ad libitum fed birds lost 53 g more than restricted treatment before 24 h of fasting. Additionally, the birds under ad libitum treatment significantly (p<0.01) increased daily feed intake by 17%. Because ad libitum groups have free access to feed and water, a higher variation in feed intake on ad libitum treatments (SD = ±13.12 g) was observed compared to restricted groups (148±2.97). Also, the feed intake under restricted regimen treatment was close to the amount of feed allocated.

The VO₂, VCO₂, and FHP were different (p<0.01) between treatments. Ad libitum birds have higher VO₂ (+3 VO₂/kg^0.75/d) and VCO₂ (+2 VCO₂/kg^0.75/d) consumption, and consequently higher FHP values (+63 kJ/kg^0.75/d). However, the RQ was not different between treatments (p = 0.490) (Table 2).

The body composition is presented in Table 3. There was no statistical difference in lipid and lean mass (p>0.05). However, bone mineral content was slightly different (p = 0.045), with +2 g/kg of bone mineral content on the birds submitted to restricted compared to ad libitum feeding regimen.

The broken line model was fit for each treatment, and the parameters are presented in Table 4. Also, the HP (kJ/kg^0.75/d) as a function of the time is illustrated in Figure 3. It was shown that the rate of energy expenditure declining after starting the fasting was higher on the ad libitum feeding, while the time to meet the broken point was the same, around 11 h 47 min in both treatments. As a result, the plateau of FHP was 64 kJ/kg^0.75/d higher when the birds were fed ad libitum.

The thirteen days that the birds were submitted to ad libitum feeding resulted in a BW gain of 314 g per bird, representing 7.5% more from the initial BW. An important characteristic to be considered in broiler breeders, compared to other poultry categories, is that the birds present a growth even when they are adults and in production [2,3]. Thus, these birds are characterized as having multiphasic growth [16], explaining why they had different growth rates during the adaptation period on both feeding regimes due to the different amounts of nutrient intake. While the feed restriction induced a slight BW loss, as Caldas et al [17] observed between 33 to 37 weeks of age, the group fed ad libitum increased BW by 132 g. Therefore, no control on feed intake and the BW can be carried to reproductive and sexual disorders affecting birds performance [1]. However, this depends on the length of the feed restriction period as hens can use body sources to recover nutrient needs for a certain period depending on the restriction level [2]. Also, under ad libitum feeding, with nutrient and energy intake above the recommendation [18], the individual feed intake variation leads to high variability in BW, making it difficult to get the flock's uniformity weight [19].

Gas exchange parameters have been scarcely described for broiler breeders during the production phase. For example, Caldas et al [17] reported recent results of 16.9 to 23 VO₂ and 15.4 to 21.9 L/kg^0.75/d for VCO₂ measured in birds fed restricted programs. Also, the heat production ranged from 351 to 480 kJ/kg^0.75/d, higher than we observed since the authors measures were conducted in fed birds, contrasted by the high RQ of 0.9 to 0.96. On the other hand, Anderson and Kulp [20], who made measurements 18 to 20 h after the feed was withdrawn, reported similar values to our study, VO₂ (18.7 L/kg^0.75/d), VCO₂ (14.7 L/kg^0.75/d), and RQ (0.82); which was close to the FHP (377 kJ/kg^0.75/d) of birds on ad libitum treatment.

Besides, Misson et al [21] reported the same values of heat production (318 kJ/kg^0.75/d), VO₂ (16.1 L/kg^0.75/d), and VCO₂ (11.5 L/kg^0.75/d) as those observed in the present study in birds on feeding restriction. However, the RQ was different (0.72) for a 24 h period of fasting. This could be due to the another strain (Light Sussex) and productive level used by the mentioned authors, which was different from the present study [22]. According to Salas et al [23] and Attia et al [2], broiler breeders during the peak of production have a high protein catabolism rate to prioritize the egg formation. Even under fasting conditions, this period was not enough to interrupt the egg production; the birds mobilize body reserve to supply nutrients, mainly energy, to form the next egg. Thus, this explains why the oxidation of proteins occurs, which is used to synthesize the egg and induce a higher RQ value [24].

A feed restriction induced a basal metabolic rate reduction in growing swine due to the body mass reduction [25], in growing chicks [10] and, calves [26]. It was explained by Spratt et al [27], who mentioned that the FHP variation due to the visceral mass is only 4%, and the significant fraction can be attributed to the body composition variation of around 30%. However, the lean and fat mass was not affected by the feed regimen treatments. These results were similar to those reported by Caldas et al [17], where birds of 30 weeks old were described with tissue contents of 186 g/kg for lean and 818 g/kg for fat mass. While the bone mineral content reported was 14 g/kg lower than those of our study. It is important to note that the methodology for determining body composition by DEXA has some limitations; lean mass is a conjunction between the protein and the water content that includes some mineral solutes. Therefore, it prevents a precise description of the actual value of body protein content since the water fraction is the factor that has more significant variability [28]. Additionally, DEXA is widely described by precision for determining mineral content. However, if there is a presence of an egg inside the animal, the mineral content would be influenced.

After feed deprivation in both groups, the metabolic rate behaviour was similar to reported by Zubair and Leeson [10] in 11 days-old broiler chicks, expressing a decline phases up to reach a minimum value described as plateau-FHP. Meanwhile, the broken point followed by the stable phases was reached at around 12 h of fasting in broiler breeders was not shown in growing birds. This metabolic phenomenon depends on the feed passage rate and gastrointestinal tract retention capacity, nutrient catabolism related to the body storage available, and demand for other metabolic functions. A significant reduction in HP is observed two hours after food deprivation due to the digestive tract is emptying, absorption of remaining nutrients, and metabolic adaptation to fasting conditions [26].

It was more evident in the ad libitum treatment, where birds presented a higher reduction of the heat production until the broken point. The time to reach the minimum stable phases of the heat production (or plateau-FHP) was the same between treatments. It is related to the feed intake amount and the bulk capacity of the broiler breeders. The broiler breeders bulk capacity is associated with the physic-chemical composition of feed and water retention that affects the passage rate [29]. As the same feed was offered, the bulk capacity was similar between treatments. The plateau value of HP along the fasting time has shown a difference of 64 kJ/kg^0.75/d lower on the restricted birds group when compared with birds fed ad libitum. Similar results were observed by Zubair and Leeson [10], who reported a variation of 67 kJ/kg^0.75/d on fasting. It corresponds to the energy expenditure of the body mass metabolic activity [30].

Broiler breeders fed ad libitum had a high cost of basal energy expenditure and maintenance energy requirement in the production peak, without a variation in the body composition compared to those on restricted feeding.