CP and intake

There are well established positive responses in milk yield that accompany increases in dry matter intake (DMI) of dairy cows (Hristov et al., Reference Hristov, Price and Shafii2004) and, given that DMI responds positively to increased dietary CP concentration (Oldham, Reference Oldham1984; Newbold, Reference Newbold1994), it follows that a significant component of the milk yield response to dietary CP (Ipharraguerre and Clark, Reference Ipharraguerre and Clark2005) resides in improvements in voluntary intake. In turn, this arises in part through improvements to digestibility (in particular that of fibre) in the rumen (Faverdin, Reference Faverdin1999), but also due to direct and indirect effects on intermediary metabolism, nutrient partitioning to the mammary gland and metabolic signalling of intake regulation (Ingvartsen and Andersen, Reference Ingvartsen and Andersen2000; Huhtanen et al., Reference Huhtanen, Rinne and Nousiainen2008). It should be noted, however, that while such effects are quite well characterised in non-ruminants (particularly in rodents), this is not the case in ruminant species, and effects of dietary protein on intake regulation in dairy cows has not been a hot topic of investigation in recent years.

Faverdin (Reference Faverdin1999) usefully identified nitrogenous nutrients as having both putative short- and long-term effects on feed-intake regulation. The former has been ascribed to the actions of ammonia (NH₄) which, when absorbed rapidly across the rumen wall at levels that exceed the liver’s capacity to detoxify can, as we have shown in cattle, act on central mechanisms to depress feed intake (Sinclair et al., Reference Sinclair, Sinclair and Robinson2000a). It is unlikely, however, that such levels of peripheral NH₄ would be attained with contemporary total mixed ration (TMR) diets of 180 g CP/kg DM or less, although blood NH₄ may come into play at higher levels of protein feeding particularly with high effective (E)RDP: fermentable metabolisable energy silages. Longer-term effects of CP on intake are equally difficult to characterise fully but are dependent on the digestibility and nature (i.e. grass v. grass silage v. maize silage) of the basal forage, as well as the level of concentrate supplementation. In general, intake responses to increased CP are greatest for diets of low digestibility, for maize rather than grass silages, and where high levels of concentrates are fed (Faverdin, Reference Faverdin1999). Newbold (Reference Newbold1994) proposed that intake response to MP for grass-based silages is probably due to MP supply whereas, for maize based silages, responses are because of the effects of both ERDP and MP supply. However, post-ruminal effects of amino acid supply on feed intake in dairy cows have been difficult to quantify, varying as it does by the proportion and source of ERDP and the degradability and amino acid content of DUP (Ipharraguerre and Clark, Reference Ipharraguerre and Clark2005). Furthermore, intake responses to increased MP are often greater when animals are deficient in ME (Newbold, Reference Newbold1994).

In an era of reducing dietary protein levels, and in order to minimise feed costs and N excretion to the environment, the challenge is how to mitigate effects associated with reduced DMI. Contemporary evidence for high-yielding dairy cows, however, indicates that reductions in intake are likely to be modest at CP levels at or above 140 g/kg DM (e.g. Broderick, Reference Broderick2003; Colmenero and Broderick, Reference Colmenero and Broderick2006; Law et al., Reference Law, Young, Patterson, Kilpatrick, Wylie and Mayne2009a; Lee et al., Reference Lee, Hristov, Heyler, Cassidy, Long, Corl and Karnati2011) and could be at least partially offset by either improving the digestibility of the UDP component of the diet (Noftsger and St-Pierre, Reference Noftsger and St-Pierre2003), by increasing fermentable metabolisable energy (FME) (Herrera-Saldana et al., Reference Herrera-Saldana, Gomez-Alarcon, Torabi and Huber1990), and/or by increasing the overall ME of the diet (thereby altering the ME : MP ratio).

Physiological status (stage of lactation)

Dietary requirements to support milk production vary during lactation reflecting changes in milk yield, appetite and endocrine regulation of nutrient partitioning between tissues and the mammary gland. Scope, therefore, may exist to modify dietary protein levels at different stages during lactation in order to increase the efficiency of N utilisation and to maximise milk yields. Working with contemporary high-yielding Holstein-Friesians, Law et al. (Reference Law, Young, Patterson, Kilpatrick, Wylie and Mayne2009a) investigated the effects of three different dietary CP concentrations on milk yield and N retention during either early or late lactation using a partial change-over design. The efficiency of N use for milk production (NUE; defined as kg of milk N output per kg dietary N intake) was 0.42, 0.39 and 0.35 for dietary CP concentrations of 114, 144 and 173 g/kg DM respectively during early (i.e. the first 150 days) lactation. The NUE for milk production was reduced when dietary CP increased during the latter 155 days lactation (i.e. from 144 to 173 g/kg DM; 0.37 to 0.30) and increased when dietary CP was reduced (i.e. from 173 to 144 g/kg; 0.30 to 0.36). Collectively, these results reveal that the efficiency of N capture for milk production is proportionately greater at lower CP concentrations. In this respect these results agree with earlier observations (Huhtanen and Hristov, Reference Huhtanen and Hristov2009). Whole lactation milk yields for cows on the lowest (i.e. 114 g CP/kg DM) level of protein were 30% below that for cows on the highest (i.e. 173 g CP/kg DM) level of protein. In contrast, whole-lactation milk yields were reduced by only 10% and 7% for cows on the 144/144 and 173/144 treatment groups, respectively, suggesting that there may be scope to mitigate losses associated with reduced dietary CP concentrations by manipulating diet composition (discussed later). This is important for strategies that aim to reduce dietary CP concentrations throughout lactation, as they offer the greatest potential to reduce feed costs and N excretion in the environment.

Diet composition and milk yield responses to CP

In a meta-analysis of the effects of dietary protein concentration on milk protein yield and milk N efficiency, Huhtanen and Hristov (Reference Huhtanen and Hristov2009) reported an improvement in efficiency of N use as dietary CP decreased or milk yield increased; the latter in the absence of an increase in dietary CP. In a survey of 372 dairy farms by Jonker et al. (Reference Jonker, Kohn and High2002), cows in protein balance had a NUE of 0.30, and NUE changed by 0.05 per 100 g N/day above or below requirements; at all levels of supply, however, NUE was highly variable among farms.

Broderick (Reference Broderick2003) demonstrated the importance of energy source in determining N efficiency in lactating dairy cows. Increasing dietary energy, by reducing proportion of forage (thereby NDF) and increasing the proportion of shelled corn (mostly starch), increased yields of total and true milk protein (Figure 2), and improved the efficiency of N use (defined earlier) from 0.25 to 0.30. In contrast, reducing dietary CP concentrations from 184 to 151 g/kg DM had little effect on total milk protein yield and no effect on true milk protein yield, although efficiency of N use was improved. The proportion of forage in the study was reduced from 0.75 to 0.50. It is uncertain by how much further efficiency of N use can be improved by altering concentrate carbohydrate source for levels of forage inclusion <0.50. Indeed, such responses are likely to vary with forage digestibility and DMI. However, level and balance of intestinally absorbable amino acids, in addition to energy-yielding metabolites, is central to sustaining milk yield and composition with minimal dietary protein.

Figure 2

Effect of dietary CP (g/kg DM) and NDF (NDF; % DM) on milk true-protein yield (data derived from Broderick, Reference Broderick2003; GA Broderick, personal communication). Non-fibre carbohydrates (CHO; % DM) would have consisted mainly starch and a small amount of sugars. Diets contained high levels of maize grain.

It is generally accepted that of the 10 essential amino acids, Met and Lys are normally the first two rate-limiting for growth and milk production. Diets high in NDF reduce the proportion of Met and Lys in duodenal proteins, perhaps as a consequence of reduced flow of microbial proteins (which are richer in Met and Lys) to the duodenum (Robinson, Reference Robinson2010). In general the amino acid composition of MCP is superior to that of DUP as it more closely resembles the amino acid composition of milk and lean tissues (National Research Council (NRC), 2001). In contrast, the amino acid composition of DUP more closely reflects that of the various vegetable feedstuffs from which it is derived. Consequently, dietary strategies that aim to maximise the intestinal delivery of MCP may go some way to mitigating the effects of reduced supply of dietary CP. Such strategies are inextricably linked to energy metabolism within the rumen. Increased MCP probably accounted for much of the positive milk protein yield responses to dietary energy observed by Broderick (Reference Broderick2003).

As the efficiency of conversion of MP and that of metabolisable amino acids into milk TP declines with both increasing supply of MP (Metcalf et al., Reference Metcalf, Mansbridge, Blake, Oldham and Newbold2008) and with Met and Lys (Vyas and Erdman, Reference Vyas and Erdman2009), it follows that production responses to rumen protected forms of specific amino acids may increase at lower dietary protein levels so long as dietary energy intake and other amino acids do not become rate limiting. Most major protein sources contain less Met and Lys than does MCP, so that the proportion of these limiting amino acids in duodenal digesta decline at higher levels of dietary CP (Robinson, Reference Robinson2010). Sometimes, however, responses to added Met and Lys are influenced by the supply of other potentially limiting amino acids (e.g. histidine) and energy (e.g. gluconeogenic) metabolites. In particular for grass-silage diets histidine, rather than Met or Lys, is often the first limiting amino acid (Vanhatalo et al., Reference Vanhatalo, Huhtanen, Toivonen and Varvikko1999).

In a systematic review and analysis of dietary protein sources and responses to metabolisable Lys and/or Met, Robinson (Reference Robinson2010) found that milk yield responses to Met and Met+Lys generally increased as CP level of the basal diet increased although, as previously stated, marginal efficiencies of metabolisable Met and Lys use for milk protein yield decreased (Vyas and Erdman, Reference Vyas and Erdman2009). Looking to the future, in an era of reduced dietary protein use, it will be important to establish precise responses to rate-limiting amino acids at low dietary CP levels. With this in mind a comprehensive systematic-review and meta-analysis of the pertinent literature relating to production responses to specific amino acids in lactating cows fed low-protein diets was conducted.

Meta-analysis of responses to rumen protected amino acids for low CP diets

This meta-analysis used peer-reviewed publications accessible via Thompson Reuters Web of Science and NCBI PubMed. For the latter database the following MeSH terms included ‘cattle’, ‘metabolism’, ‘lactation’, ‘milk’, ‘protein’, ‘methionine’, ‘lysine’. To be eligible only studies conducted since 1987 were considered with lactating Holstein cows offered diets with a CP content of ⩽150 g/kg DM during either early or mid lactation. Analysis was also limited to those studies and experiments within studies where a combination of rumen protected Met and Lys were incorporated into the diet, and measurements of milk yield and composition reported. This is because few studies have considered Lys alone, and it is widely recognised that a balance of around 7.2% Lys and 2.4% Met within MP favours milk production (NRC, 2001). Only one study was included in the final analysis where these two amino acids were infused directly into the abomasum (Robinson et al., Reference Robinson, Chalupa, Sniffen, Julien, Sato, Fujieda, Ueda and Suzuki2000).

Source and level of protected Met and Lys varied between studies, but this was ignored in the final analysis based on the premise that their use and level of incorporation was predicted by the reporting authors to be sufficient to evoke a metabolic and/or production response. In a systematic analysis of the literature, however, Robinson (Reference Robinson2010) noted that few studies made any attempt to actually measure or even predict delivery of absorbable amino acids to the duodenum. The final analysis, therefore, involved 13 peer-reviewed studies with 16 independent experiments (Supplementary Table S1). Responses analysed were fat-corrected milk yield (kg/day) and protein yield (kg/day). For protein yield it was not always clear whether studies reported crude or true milk protein, and so no attempt was made to adjust these values. Data were analysed using a fixed-effects model (Comprehensive Meta-Analysis software, Biostat, Englewood, NJ, USA; http://www.meta-analysis.com). Heterogeneity (i.e. variation in study outcomes between studies) was assessed using the χ ² test on Cochran’s Q-statistic and by calculating I ² (i.e. percentage of variation across studies that is due to heterogeneity rather than chance). As heterogeneity was present for both analysis of milk yield and milk protein yield, a random-effects meta-analysis was carried out for each. These analyses, however, agreed with that for the fixed model and so only outputs of fixed-model analyses are presented.

Results from these analyses revealed no overall response in FCM yield (Figure 3) but an overall increase (P=0.002), albeit small, in milk protein yield (Figure 4) when low CP diets were supplemented with rumen protected Met+Lys. Furthermore, in the absence of publication bias both sets of analyses revealed significant (P<0.001) heterogeneity between studies indicating that biological factors rather than sampling error accounted for variation in responses between studies. Also, the outcome of this meta-analysis matched the general responses of the largest study in this cohort (i.e. Polan et al., Reference Polan, Cummins, Sniffen, Muscato, Vicini, Crooker, Clark, Johnson, Otterby and Guillaume1991) which was replicated across several institutes and involved on average 55 cows per treatment.

Figure 3

Effect of rumen protected methionine plus lysine on fat-corrected milk yield (kg/day) for diets of ⩽150 g CP/kg DM. A meta-analysis of 13 studies, involving 16 experiments, where cows were fed either a ‘Control’ diet or a diet of similar composition but with added rumen protected methionine and lysine. The Forest plot presents standardized differences in means (s.d. in means) with 95% CI. Heterogeneity (Q-statistic, χ ²) between studies=62.1 (15 df), P<0.001 (I ²=75.8%). Mean values (■) proportional in size to each study’s weight. Overall treatment effect: Z=1.107 (P=0.103).

Figure 4

Effect of rumen protected methionine plus lysine on milk protein yield (kg/day) for diets of ⩽150 g CP/kg DM. A meta-analysis of 13 studies, involving 16 experiments, where cows were fed either a ‘Control’ diet or a diet of similar composition but with added rumen protected methionine and lysine. The Forest plot presents standardized differences in means (s.d. in means) with 95% CI. Heterogeneity (Q-statistic, χ ²) between studies=63.5 (15 df), P<0.001 (I ²=76.4%). Mean values (■) proportional in size to each study’s weight. Overall treatment effect: Z=3.141 (P=0.002).

The greatest positive response to supplemental rumen-protected Met+Lys was observed in the study of Robinson et al. (Reference Robinson, Fredeen, Chalupa, Julien, Sato, Fujieda and Suzuki1995), which was conducted at two sites with largely similar diets (based on a mixture of maize, lucerne and grass silages) offered to mid-lactation Holstein cows averaging around 31 kg/day FCM. Although CP content of the diets was around 140 to 150 g/kg DM they were calculated, based on chemical composition and measurements of intake, to meet but not exceed protein requirements for microbial growth and milk production. Under these conditions supplementation of 19 g/day ruminally protected l-Lys.HCl and 6.5 g/day ruminally protected dl-Met evoked a positive response but by different mechanisms at the two sites; at one site this was due to apparent increases in gross efficiency of utilisation of both protein and energy, whilst at the other site the response arose largely through improvements in DMI. In the other study that showed significant improvements to rumen protected AA, Christensen et al. (Reference Christensen, Cameron, Clark, Drackley, Lynch and Barbano1994) offered fat (ground high-oil shelled corn) to cows fed a diet containing 142 g CP/kg DM and reported increases in FCM yield and milk true-protein yield relative to non-fat supplemented cows fed a diet containing 175 g CP/kg DM. The feeding of ruminally protected AA and post-ruminal infusion of casein into cows fed fat-supplemented diets containing 160 to 170 g CP/kg DM was previously found to increase milk protein yields (Canale et al., Reference Canale, Muller, McCahon, Whitsel, Varga and Lormore1990; Chow et al., Reference Chow, DePeters and Baldwin1990; Cant et al., Reference Cant, Depeters and Baldwin1991).

In contrast, the greatest negative effect of supplemental Met+Lys in low (144 g/kg DM) CP diets was found in the study of Robinson et al. (Reference Robinson, Chalupa, Sniffen, Julien, Sato, Fujieda, Ueda and Suzuki2000), where these two AA were directly infused into the abomasum of mid/late-lactation cows receiving 140% of calculated intestinally absorbable requirements. Such was the magnitude of this effect that removal of this study from the meta-analysis led to a small but significant (P=0.05) positive increase in overall FCM yield. The reduction in performance in this study was due to inhibitory effects on DMI. In the two studies of Lee et al. (Reference Lee, Hristov, Heyler, Cassidy, Lapierre, Varga and Parys2012a and b), the CP content of Control diets was around 157 g/kg DM and calculated to be adequate in MP. In contrast the CP content of treatment diet was ⩽140 g/kg DM and deemed to be deficient in MP. The inclusion of rumen protected Met+Lys failed to increase FCM and milk protein yields to levels comparable to the MP-adequate Control groups in these studies but it did, in the study of Lee et al. (Reference Lee, Hristov, Cassidy, Heyler, Lapierre, Varga, de Veth, Patton and Parys2012b), lead to small but significant improvements in milk protein yield relative to a second CP and MP deficient group, where no rumen protected AA were added.

The variability in milk yield and compositional responses for diets of ⩽150 g CP/kg DM highlighted by these analyses agree with the general findings of previous systematic reviews which considered a wider range of dietary protein levels (Vyas and Erdman, Reference Vyas and Erdman2009; Patton, Reference Patton2010; Robinson, Reference Robinson2010). The general consensus is that it is possible to manipulate the proportions of Met and Lys in duodenal protein, but predicting this and production responses that may arise are difficult. There is also a consensus that responses to added Met and Lys are generally superior to either amino acid offered alone, particularly during early lactation, but that the response to one amino acid is not necessarily dependent on the addition of the other (Patton, Reference Patton2010). Responses to added amino acids do, however, depend on diet composition (e.g. level and nature of forage) and animal factors such as stage of lactation, body composition, negative energy balance and milk yield; all of which will determine nutrient priorities and partitioning to the mammary gland. Variability in responses also exists between commercially available sources of rumen-protected Met and Lys, which is beyond the scope of the current article but is the subject of related reviews (e.g. Robert, Reference Robert2005; Vyas and Erdman, Reference Vyas and Erdman2009); and it is evident from Robinson et al. (Reference Robinson, Chalupa, Sniffen, Julien, Sato, Fujieda, Ueda and Suzuki2000) that excessive levels of rumen-protected amino acids may be detrimental.

Tissue, body fat/protein mobilisation

A decrease in dietary protein supply in early lactation is associated with an increase in the responsiveness of adipose tissue to anti-lipolytic stimuli such as insulin, and decrease in the responsiveness to lipolytic stimuli such as epinephrine (Cadórniga and López Díaz, Reference Cadórniga and López Díaz1995; Schor and Gagliostro, Reference Schor and Gagliostro2001). It would therefore be anticipated that decreasing tissue amino acid supply should be translated into a net decrease in body fat mobilisation. The relationship between dietary protein concentration and live weight change and body fat mobilisation (as represented by plasma non-esterified fatty acids (NEFA); Figure 5a and b), does not support this. An analysis of data from studies (i.e. Jones and Garnsworthy, Reference Jones and Garnsworthy1988; Cadórniga and López Díaz, Reference Cadórniga and López Díaz1995; Garcia-Bojalil et al., Reference Garcia-Bojalil, Staples, Risco, Savio and Thatcher1998; Hongerholt and Muller Reference Hongerholt and Muller1998; Bach et al., Reference Bach, Calsamiglia, Huntington and Stern2000; Westwood et al., Reference Westwood, Lean, Garvin and Wynn2000; Schor and Gagliostro, Reference Schor and Gagliostro2001; Davidson et al., Reference Davidson, Hopkins, Diaz, Bolt, Brownie, Fellner and Whitlow2003) that have examined the influence of DUP level within the same dietary protein concentration have also failed to find a consistent relationship (geometric means for plasma NEFA: 255 v. 310 µeq/l for Low v. High DUP respectively; P=0.158). The effects of protein supply on body tissue mobilisation are, however, confounded by dietary factors (e.g. amino acid profile, dietary energy concentration and source), and animal factors (e.g. body condition score (BCS), stage of lactation and DM intake). For example, Jones and Garnsworthy (Reference Jones and Garnsworthy1988) and Cadórniga and López Díaz, (Reference Cadórniga and López Díaz1995) demonstrated a greater negative energy balance in cows that had a high BCS when provided with diets with a high by-pass protein content, but not in cows with a low condition score. Additionally, increased dietary protein intake or DUP supply is often associated with an elevated DM intake, reducing the requirement for body tissue mobilisation.

Figure 5

Relationship between dietary CP concentration and plasma non-esterified fatty acids (a) and live weight change (b) in early lactation dairy cows (Data from Nachtomi et al., Reference Nachtomi, Halem, Bruckental and Amir1991; Hongerholt and Muller Reference Hongerholt and Muller1998; Bach et al., Reference Bach, Calsamiglia, Huntington and Stern2000; Wu and Satter Reference Wu and Satter2000; Kokkonen et al., Reference Kokkonen, Syrjälä-Qvist, Tsehai Tesfa, Tuori and Yrjänen2002; Davidson et al., Reference Davidson, Hopkins, Diaz, Bolt, Brownie, Fellner and Whitlow2003; Schei et al., Reference Schei, Volden and Baevre2005; Law et al., Reference Law, Young, Patterson, Kilpatrick, Wylie and Mayne2009a). R ² for NEFA=0.10 and LWG=0.07.

Liver steatosis (or fatty liver; FLS), occurs during the periparturient period in dairy cows when excess body fat is mobilised, and it has been estimated that ~45% of dairy cows in early lactation have a moderate steatosis (Jorritsma et al., Reference Jorritsma, Jorritsma, Schukken and Wentink2000; Bobe et al., Reference Bobe, Young and Beitz2004). Dairy cows have been shown to have a particularly low rate of secretion of very low density lipoproteins (VLDL) from the liver (Grummer, Reference Grummer1995), and enhancing the rate of lipid transport from the liver offers one option to reduce the onset of FLS. Met serves as a methyl donor for the synthesis of apolipoprotein B100 and phosphotidyl choline in the liver. These metabolites are required for the hepatic production of VLDL, and are necessary for the synthesis of carnitine, required for the uptake of long-chain fatty acids across mitochondrial membranes (Grummer, Reference Grummer1995). It has therefore been suggested that a deficiency of Met may restrict hepatic VLDL synthesis, and surveys of dairy cows that have experienced severe liver steatosis in the periparturient period indicate a strong association with low plasma Met concentrations (Shibano and Kawamura, Reference Shibano and Kawamura2006). However, controlled studies that have investigated the response to supplementary Met have not produced a conclusive benefit (Durand et al., Reference Durand, Chilliard and Bauchart1992; Bertics and Grummer, Reference Bertics and Grummer1999; Piepenbrink et al., Reference Piepenbrink, Marr, Waldron, Butler, Overton, Vázquez-Añón and Holt2004). Indeed, some have reported an increase in hepatic lipid concentrations (Preynat et al., Reference Preynat, Lapierre, Thivierge, Palin, Cardinault, Matte, Desrochers and Girard2010), and therefore the practical role of Met on energy metabolism in early lactation is unclear.

An alternative source of amino acids for metabolism in early lactation is the mobilisation of body protein reserves, which can be equivalent to an extra daily dietary protein supply over the first 100 days of lactation of ~250 g, or 5% to 10% of protein intake (Paquay et al., Reference Paquay, DeBaere and Bull1972; Botts et al., Reference Botts, Hemken and Bull1979). The mobilised amino acids, particularly Met, may be used to enhance the rate of lipid transport from the liver and reduce the risk of FLS in early lactation, as described previously. Although some, but not all studies have shown a benefit to milk yield and protein output from feeding higher protein during the dry period (see review of Friggens et al., Reference Friggens, Andersen, Larsen, Aaes and Dewhurst2004), there is little benefit on reducing liver triglyceride concentration or the incidence of health disorders in early lactation (e.g. Huyler et al., Reference Huyler, Kincaid and Dostal1999; Vandehaar et al., Reference Vandehaar, Yousif, Sharma, Herdt, Emery, Allen and Liesman1999; Greenfield et al., Reference Greenfield, Cecava, Hohnson and Donkin2000). Interestingly, cows that are most susceptible to the development of metabolic disorders (i.e. above a BCS of 3.25 units) respond better to increased dietary protein supply in late gestation as evidenced by lower blood NEFA levels (Putnam and Varga, Reference Putnam and Varga1998), a situation similar to that found when supplying additional protein in early lactation. A confounding effect of feeding elevated protein levels in late gestation is the association with an increase in intake. This can lead to an increase in BCS, which in itself depresses intake in early lactation and promotes fat mobilisation, predisposing the cow to health issues such as downer cows (Grummer, Reference Grummer1995) and metabolic disease (Friggens et al., Reference Friggens, Andersen, Larsen, Aaes and Dewhurst2004).

Reductions in dietary CP: effects on cow fertility

Although numerous studies have tested the hypothesis that high dietary CP is bad for fertility, far less emphasis has been placed on the effects of low CP diets on reproductive parameters. The results of six studies in which balanced groups of lactating dairy cows were fed either high or low levels of dietary CP are summarised in Table 1. Reductions in dietary protein can lead to a fall in milk yield. In some studies this has been associated with a reduction in calculated negative energy balance (e.g. Law et al., Reference Law, Young, Patterson, Kilpatrick, Wylie and Mayne2009b), a key factor driving poor fertility. In the Law et al. (Reference Law, Young, Patterson, Kilpatrick, Wylie and Mayne2009b) study there was no associated increase in either live weight or BCS, nor were there any changes in blood metabolites related to energy balance. In the results presented here, while milk yield was consistently reduced, low CP diets did not appear to have any major effects on reproductive performance. However, once again, it should be noted that most of these studies were carried out some time ago and the milk yields recorded bear little comparison to the much higher yields achieved in the modern dairy cow. Thus, although these studies suggest that a reduction in dietary protein may be possible without impacting on fertility traits, it is essential that the effects of such an approach are assessed in modern high-yielding cows where metabolic demands are far in excess of animals studied previously.

Table 1

Mean (weighted for study size) milk yield and reproductive parameters measured in studies (n=6) in which balanced groups of lactating dairy cows (n=12 to 146 per study) were fed diets containing either high (mean 19.2%, range 16.9% to 20.0%; n=181 cows) or low (mean 13.8%, range 12.7% to 14.5%; n=186 cows) levels of CP (based on data from; Jordan and Swanson Reference Jordan and Swanson1979; Edwards et al.,Reference Edwards, Bartley and Dayton1980; Howard et al., Reference Howard, Aalseth, Adams and Bush1987; Carroll et al., Reference Carroll, Barton, Anderson and Smith1988; Barton et al., Reference Barton, Rosario, Anderson, Grindle and Carroll1996; Law et al., Reference Law, Young, Patterson, Kilpatrick, Wylie and Mayne2009b)

PUN=plasma urea nitrogen; AI=artificial insemination; CR=conception rate; PR=pregnancy rate.

Nitrogenous metabolites: urea and ammonia

Dietary CP intake can be reflected in concentrations of plasma or milk urea. Urea measurements have been used in a number of studies to assess the relationships between levels of the metabolite and fertility parameters. Plasma levels of urea in excess of 6.8 mmol/l have for many years been considered indicative of negative influences on reproduction (Butler et al., Reference Butler, Calaman and Beam1996). However, the numerous studies looking at the association between circulating urea levels and fertility within dairy herds have produced variable results. In the UK, studies have reported reduced conception rates in cows with elevated plasma urea levels (>7.5 mmol/l; Wathes et al., Reference Wathes, Bourne, Cheng, Mann, Taylor and Coffey2007) while others have failed to find any relationship (Cottrill et al., Reference Cottrill, Biggadike, Collins and Laven2002; Mann et al., Reference Mann, Mann, Blache and Webb2005). In a recent large study in Poland (Sawa et al., Reference Sawa, Bogucki and Krezel-Czopek2011), while significant relationships were reported between milk urea and a number of fertility parameters, the associated R values were low (typically 0.01 to 0.07) emphasising the tenuous nature of these associations. Indeed, in a recent and comprehensive meta-analysis on the topic, Lean et al. (Reference Lean, Celi, Raadsma, McNamara and Rabiee2012) found no influence of blood urea levels on the effects of dietary protein on conception rate.

That notwithstanding, various correlates of dietary protein, including ammonia and urea, have been shown to act at both the ovarian and uterine level to influence reproductive function (Butler, Reference Butler1998). In vitro studies have established the ability of high ammonia/urea to impair oocyte development (Sinclair et al., Reference Sinclair, Kuran, Gebbie, Webb and McEvoy2000b; De Wit et al., Reference De Wit, Cesar and Kruip2001) though, in parallel with the effects of feeding high protein diets, the effects are inconsistent. For example, Ocon and Hansen (Reference Ocon and Hansen2003) reported negative effects of 7.5 mM but not 10 mM urea during in vitro maturation on subsequent blastocyst formation. What does appear more consistent is that the oocyte/blastocyst is more susceptible than the developing embryo as Rhoads et al. (Reference Rhoads, Rhoads, Gilbert, Toole and Butler2006) reported adverse effects of high urea in donor but not recipient animals following day 7 embryo transfers in lactating dairy cows. It has been suggested that the main site of action of the elevated ammonia/urea associated with the feeding of high protein diets may be the uterus where elevated urea is associated with lowered uterine pH (Elrod and Butler, Reference Elrod and Butler1993). However, in vitro studies would suggest that the period of reproductive susceptibility may occur before the embryo reaches the uterus, suggesting the oviductal environment may be more important. Furthermore, Carroll et al. (Reference Carroll, Barton, Anderson and Smith1988) reported similar conception rates in cows fed low and high protein diets resulting in widely differing uterine urea levels. Thus while a number of potential sites of action through which dietary protein can directly influence reproductive function have been identified, no conclusive mechanism has been established.