Optimal inclusion levels of palm kernel cake in diets for large and small ruminants: a meta-analysis

2.1 Literature search

This meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) protocol [34], 35]. A comprehensive search of the research papers was conducted using the keywords ‘palm kernel cake’ and ‘ruminant’ to identify relevant studies related to the effects of PKC on ruminant performance, nutrient utilization, and production. The keywords were entered into three international scientific databases: Google Scholar (https://scholar.google.com/), ScienceDirect (https://www.sciencedirect.com/), and Springer Link (https://link.springer.com/). This process resulted in 980 papers from Google Scholar, 495 papers from Science Direct, and 633 papers from Springer Link. ScienceDirect allows the screening process to be conducted directly on its platform; thus, 303 papers were retained for further selection.

2.2 Selection process and inclusion criteria

This process was conducted to obtain papers aligned with the research objectives and carried out objectively and transparently according to the PRISMA guidelines. Papers retrieved using the specified keywords were entered into a Microsoft Excel^® Sheet 2021 (Microsoft Corp, Redmond, WA, USA). Duplicate papers were identified using the PivotTable function in Microsoft Excel^® Sheet 2021. A total of 106 papers were excluded during the screening process. Subsequently, the selection continued by assessing the relevance of each study based on its title and abstract. Through this process, 132 potential papers were subjected to full-text eligibility assessment using Mendeley Desktop 1.19.8 software. The full-text screening adhered to the following inclusion criteria; (i) papers reporting in vivo trials of PKC inclusion in ruminants, (ii) papers that reported or allowed calculation of PKC inclusion levels in ruminant diets, (iii) papers published in English, (iv) peer-reviewed papers published in reputable journals, (v) no restriction on publication year, and (vi) papers with full-text availability. A total of 51 papers met the eligibility criteria and were included in the meta-analysis. The PRISMA flow diagram illustrating the selection process is presented in Figure 1.

Figure 1:

The preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram of the selection process.

The PICO elements of this meta-analysis were defined as follows. The population included ruminant species such as cattle, buffalo, goats, and sheep. The intervention was the PKC inclusion level. The comparator was a control diet without PKC or, when unavailable, the diet with the lowest inclusion level within each study. The outcomes evaluated animal performance, nutrient intake, nutrient digestibility, rumen fermentation, carcass traits, and milk production.

2.3 Data extraction

The dataset was compiled from a total of 51 peer-reviewed research papers. Among these, 26 studies investigated the inclusion of PKC in diets of large ruminants (cattle and buffaloes), whereas the remaining 25 studies focused on small ruminants (goats and sheep). Considering that large and small ruminants differ in body size and physiological performance, the current meta-analysis was conducted separately for each group. A summary of the 26 studies on large ruminants and 25 on small ruminants is presented in Table 1, which includes the authors’ names, year of publication, basal diet, animal type and breed, and country. Descriptive statistics of the dataset, including mean, standard deviation, minimum, and maximum values, for large and small ruminants are presented in Table 2 and Table 3, respectively. Moreover, the chemical composition of PKC used in the included studies is summarized in Table 4.

Detailed outcome data were also extracted, including nutrient intake, body weight gain, ADG, feed conversion ratio, feed efficiency, carcass production, and milk production as the primary outcomes, as they represent the direct production response to PKC inclusion. The secondary outcomes included nutrient digestibility, rumen fermentation, carcass composition, and milk composition, which describe physiological or metabolic responses. Some variables were reported in different units across studies. To ensure consistency, all values were converted into common units. PKC inclusion levels reported as percentages were converted to g/kg of dietary DM.

2.4 Statistical analysis

The database was analyzed using the mixed model methodology [68] in SAS^® OnDemand for Academics. The different study was treated as a random effect, while the PKC inclusion level was treated as a fixed effect. The statistical model used was:

where Y _ij = dependent variable; B ₀ = overall intercept; B ₁ = linear regression coefficient of Y on X; X _ij = value of the continuous predictor variable; s _i  = random effect of study i; b _i  = random effect of study on the regression coefficient of Y on X in study i; and e _ij = the unexplained residual error. The CLASS statement was used to specify the study variable as it contained no quantitative information. The RANDOM statement was declared based on different studies. These models were weighted based on the number of replicates in each study, as described by Jayanegara et al. [69].

Model fit and accuracy were evaluated using p-values, root mean square error (RMSE), and Akaike Information Criterion (AIC). A significance threshold of p < 0.05 was used. When 0.05 < and ≤0.10, the trend was considered a tendency toward significance. AIC was used to evaluate model fit, where lower AIC values indicated better goodness of fit, reflecting the model’s balance and accuracy. Model accuracy and precision were further assessed by calculating RMSE using PROC GLM [68].

Linear and quadratic regression models were used to describe the response of all parameters to increasing PKC inclusion levels. The linear model was applied to identify a proportional relationship, while the quadratic model was used to detect curvilinear responses and determine the optimal inclusion level [70]. Optimal PKC inclusion level for parameters exhibiting a significant quadratic effect (p < 0.05) was determined at the vertex of the quadratic curve, which represents the peak performance response. The optimal value was calculated using the first derivative of the quadratic regression function [70]:

where X = optimal PKC inclusion level.

3.1 Large ruminants

The current meta-analysis shows that the variation of PKC inclusion has a quadratic effect (p < 0.05) on CPI, ether extract intake (EEI), non-fiber carbohydrate intake (NFCI), total digestible nutrient intake (TDNI), and NDFI. Nonetheless, DMI, OMI, and ADFI remained unaffected by the inclusion of PKC in large ruminants’ diets (Table 5).

The result demonstrates a quadratic relationship (p < 0.05) in DMD, CPD, NDFD, and hemicellulose digestibility with the various levels of PKC inclusion in the diets. It also showed a significant linear reduction (p < 0.05) in NFC digestibility (NFCD) and TDN. However, no significant change (p > 0.05) was identified in OM digestibility (OMD), EE digestibility (EED), and ADF digestibility (ADFD) (Table 6).

The diverse levels of PKC inclusion in large ruminants’ diets did not significantly (p > 0.05) influence the average daily gain (ADG), feed conversion ratio (FCR), weight gain (WG), and feed efficiency (FE) (Table 7).

Feeding PKC in large ruminants did not result in significant (p > 0.05) change in all the rumen fermentation characteristics such as pH value, ammonia concentration (N–NH₃), acetate (C₂), propionate (C₃), butyrate (C₄), iso-butyrate (iso-C₄), valerate (C₅), iso-valerate (iso-C₅), and C₂:C₃ ratio, excluding total volatile fatty acids (VFA). The total VFA was decreased by a linear response (p < 0.05) (Table 8).

The results of the meta-analysis show that there is no significant (p > 0.05) influence on milk yield (MY) and milk composition (fat, protein, lactose, and total solids) of large ruminants consuming PKC (Table 9). The results demonstrate that all the carcass production and composition variables remained unchanged (p > 0.05). Large ruminants consuming PKC did not vary in hot carcass weight (HCW), hot carcass yield (HCY), subcutaneous fat thickness (SFT), ribeye area (RA), leg length (LL), and carcass length (CL) (Table 9).

3.2 Small ruminants

It is revealed that no significant (p > 0.05) result was found for DMI, OMI, CPI, EEI, NFCI, TDNI, and NDFI of small ruminants fed PKC (Table 10). The levels of PKC inclusion in small ruminants’ diets exhibit a quadratic influence (p < 0.05) on ADG, FCR, and FE (Table 11). The rumen fermentation characteristics (pH value, N–NH₃, total VFA, C₂, C₃, and C₄) remained unchanged (p > 0.05) by the various levels of PKC inclusion in the small ruminants’ diet (Table 12).

Apart from MY, which followed a quadratic response (p < 0.05) to PKC inclusion, other milk composition (fat, protein, lactose, total solids, and ash) remained unaffected (p > 0.05) (Table 13). PKC inclusion did not significantly (p > 0.05) affect HCW, cold carcass weight (CCW), HCY, cold carcass yield (CCY), SFT, marbling, loin eye area (LA), carcass conformation (CC), external length, and LL. However, SW exhibited a quadratic response (p < 0.05) (Table 14).

4.1 Large ruminants

This meta-analysis found that PKC inclusion did not significantly affect DMI in large ruminants. Elevated levels of NDF and silica in the diet can reduce the acceptability and intake [5]. High-NDF diets may lead to earlier satiety due to rumen filling, thereby reducing feed intake [11]. Silica also contributes to feed stiffness and low palatability, which limits intake and digestion due to its indigestible nature.

These findings align with Soares et al. [12], who reported no change in DMI in heifers fed PKC, but noted a reduction in NFCI and TDNI. High EE content can also affect the ruminant intake rate. Diets containing over 4.5 % EE can restrict DMI, indirectly limiting other nutrient intake [71]. PKC contains 9.74 % EE, with lauric acid and myristic acid comprising 39.39 % and 18.8 %, respectively, limiting nutrient intake due to their amphiphilic nature. The fatty acid content of PKC may reduce palatability despite being less prone to rancidity [39].

The decline in nutrient intake due to PKC inclusion affected the nutrient degradation process in the rumen. All nutrient intakes declined, except for EED, which increased at higher PKC inclusion levels. Crude fiber (CF) provides a physical filling effect, even though the other nutrients have not been optimally digested. In this study, ADFD showed a slight decreasing trend with increasing PKC inclusion, reflecting the complexity of ADF digestion. This suggests that while NDF is digestible at certain levels, ADF remains a major inhibitor. Increased levels of NDF and ADF render the diet more indigestible by rumen microorganisms [21]. Although fiber affects DMD and CPD, the fat content of the diet contributes as an alternative energy source, as the fat composition in the PKC is degradable. Notably, PKC contains high short and medium-chain fatty acids that are better digested and have higher uptake rates than long-chain fatty acids [72].

Xylan and mannan, the main non-starch polysaccharides (NSP) in PKC, are complex fibers that might reduce PKC’s overall digestibility and utilization. PKC contains 13.91 % xylan and 28.78 % mannan of DM [73]. Xylan increases chyme viscosity, thereby interfering with enzyme access to substrates and reducing nutrient digestibility efficiency. Mannan forms a complex fiber structure that is tougher to digest by digestive enzymes as compared to other hemicelluloses, limiting nutrient digestibility. Although they generate VFA (acetate and propionate) during fermentation, the absence of enzymes like xylanase and mannanase can reduce fermentation efficiency [74], 75].

Furthermore, the high amount of neutral detergent insoluble protein (NDIP) and acid detergent insoluble protein (ADIP) in PKC also restricts crude protein (CP) degradation [76]. The heating process of palm oil before mechanical extraction promotes protein denaturation and complex reactions between peptides and carbohydrates. This contributes to an increase in NDIP and ADIP in PKC, making it less digestible in the rumen [21].

Nutrient digestibility resulted in a noticeable impact on rumen fermentability characteristics. This meta-analysis found that greater PKC inclusion reduced total VFA concentrations. Since VFA is the main energy source for ruminants, this decline may affect productivity. In contrast, Kumar et al. [5] reported no change in total VFA in buffaloes fed PKC. Fortunately, PKC inclusion was also found to stabilize rumen pH. The reduced CPI might contribute to lower ruminal N–NH₃ concentration. Although CPD increased quadratically, N–NH₃ remained unaffected, consistent with previous findings [5], 40].

Increasing dietary PKC levels significantly reduced total VFA by 12.6 mmol/L per 100 g/kg DM PKC, whereas pH and N–NH₃ showed only slight, non-significant changes, with pH decreasing by about 0.09 units and N–NH₃ increasing by 2.7 mg/dL. These findings suggest that higher PKC inclusion may lower rumen fermentation activity, mainly reflected by the decline in total VFA, likely due to its high fat and fiber contents that limit the fermentation of rapidly degradable carbohydrates.

The unchanged WG and ADG suggest that even with reduced intake, animals can maintain growth by improving nutrient utilization. Enhanced digestibility may compensate for lower nutrient intake. This implies that performance is not compromised if digestibility improves. DMI is essential for growth in beef cattle [13]. However, in this case, improvement in digestibility helped maintain an optimal growth rate amid a decrease in intake. Our results are in agreement with Lisboa et al. [20], who found that reduced DMI in bulls fed PKC did not impair ADG or FCR. In fact, growth performance remained stable with up to 240 g/kg DM PKC inclusion. This suggests that a decrease in intake is not necessarily associated with a decrease in performance when digestibility is improved.

As a further observation, PKC inclusion showed a negligible effect on carcass and milk production. HCW and HCY remained stable, consistent with previous findings [37]. Since HCW is positively correlated with SW, this indicates no adverse effect. Similarly, MY was unaffected in cattle fed PKC. This is likely due to the lower DMI. The diet-to-milk conversion process relies on the amount and quality of nutrients digested by the animals [39]. Kumar et al. [5] supported this finding. They discovered that the PKC inclusion was unlikely to modify MY in buffaloes. This is partially due to the limited protein content in PKC.

PKC is low in lysine and methionine compared to other oil cakes. Both are crucial for milk synthesis in dairy buffaloes [5]. Despite this, milk composition remained unchanged. Although some studies report reductions in milk protein and lactose with PKC inclusion, this variation likely depends on diet composition and management [39], as milk fat is influenced by the acetate:propionate ratio [5], while milk protein and lactose depend on dietary CP and TDN [39].

4.2 Small ruminants

This meta-analysis highlights that PKC is a promising ingredient for small ruminants. Despite its low acceptance, it can still be utilized effectively at appropriate inclusion levels. Nutrient intake showed similar patterns to large ruminants, with DMI remaining stable. This suggests that PKC is less palatable and poorly accepted, likely due to its high CF content. Elevated CF content in the diet due to the inclusion of PKC restricts nutrient intake [15]. Higher NDF and lignin levels in PKC further limit utilization by increasing digestion time and reducing nutrient intake [59], 77]. High-fiber diets cause delayed digestion and contribute to reduced DMI [17]. Previous meta-analysis reported that PKC reduces ADF digestibility quadratically in small ruminants, whereas CF and NDF digestibility remain unaffected [31].

Nutrient intake is influenced not only by feed type but also by animal species. Among domestic ruminants, goats are the most selective [78]. This explains why no effect was shown for DMI, since PKC is difficult for goats to tolerate [14]. Contrary to this meta-analysis, Ferreira et al. [14], 54] and Silva et al. [79] found that DMI, CPI, and TDNI in lactating goats decreased significantly with PKC inclusion. This rejection implies that PKC is simply unpalatable to goats. This suggests that even if total intake appears stable, rejection can still occur at higher inclusion rates. Since DMI affects other nutrient intakes, it also influences milk yield and composition [23], 54].

In ruminants, DMI is regulated by a chemotactic mechanism, where VFA absorption from the rumen triggers feeding signals, while nutritional imbalance can suppress microbial activity and digestion efficiency, lowering DMI [17]. Rumen fermentation produces acetate, propionate, and butyrate, primary energy sources for ruminants. In this study, total and partial VFA were not affected by PKC inclusion, consistent with Chanjula et al. [7]. Easily fermented substrates tend to increase propionate, while fibrous substrates increase acetate [80]. Lauric and myristic acids in PKC can interfere with rumen microorganisms. Both are main PKC fat components, which are amphipathic and toxic to rumen microorganisms [22], 81].

PKC inclusion also did not destabilize the rumen pH. It remained within the optimal range (6.22–6.61), supporting fermentation activity. Microbial growth peaks when rumen pH is between 6 and 7 [82]. This appropriate pH condition creates a conducive environment for microbial development. Rumen fermentation that occurs optimally is also signaled by the production of N–NH₃. The concentration of N–NH₃ in the rumen highly depends on the protein content of the feed. Protein promotes microbial growth and replication, which ultimately leads to protein synthesis. The values obtained in this study (14.14–16.71 mg/dL) fall within the optimal range of 8.5–30 mg/dL McDonald et al. [83]. Although protein intake was unchanged, the 15.67 % CP in PKC (Table 4) appears sufficient to maintain microbial activity and N–NH₃ concentration.

Optimizing production in animal farming is essential to maximizing the economic value. This meta-analysis revealed a quadratic decrease in ADG, with a gradual decrease initially and a more significant decrease at higher levels of PKC. Conversely, FCR increased quadratically, indicating reduced feed efficiency as inclusion levels rose. At low levels, changes were minor, but higher inclusion resulted in more pronounced inefficiencies in converting feed to body weight.

The simultaneous decrease in ADG and increase in FCR indicate reduced growth efficiency, as the additional nutrient intake does not result in proportional weight gain. Rodrigues et al. [22] reported similar results, identifying limited DMI as a primary factor. DMI will ensure adequate nutrient intake for animal growth [23]. DMI directly influences CPI and TDNI, which provide essential nutrients for growth and microbial fermentation.

Increased NDIP, ADIP, and lignin from PKC reduce dietary quality and nutrient availability, ultimately impairing animal performance. PKC inclusion did not significantly affect carcass traits, morphometrics, or subjective carcass evaluation. However, SW declined quadratically, consistent with the ADG trend. High intake of feed rich in soluble carbohydrates enhances starch digestion in the small intestine. Free glucose obtained from the digestion process will be stored in adipose tissue [15].

MY increased at low PKC inclusion, reaching a maximum of ∼134 mL/d at 140 g/kg PKC (Table 13), then declined at higher PKC levels, despite stable DMI and nutrient intake, indicating insufficient energy or protein supply for optimal milk production. At low inclusion levels, nutrient intake may still support milk synthesis. However, higher PKC levels reduce diet quality due to increased fiber and NDIP, limiting energy and protein availability for milk production.

These cumulative effects, particularly at high inclusion levels, reduce nutrient efficiency for milk production, which is evident in the quadratic decreasing pattern in MY, as shown by the declining MY trend. This suggests that even with unchanged total intake, shifts in nutrient profile due to PKC affect milk production efficiency [14].

4.3 Optimal PKC inclusion level in large and small ruminants

The findings of this meta-analysis suggest that the optimal level of PKC inclusion in large ruminants is strongly influenced by the type of animal and its production purpose. For dairy cattle, the best DMI was reached at 65 g/kg DM of PKC inclusion (Figure 2). This is due to the high fat and fiber content in PKC, which may affect feed efficiency in dairy cattle, making lower levels more appropriate. The saturated fatty acid content in PKC, such as lauric and myristic acid, reduces palatability and nutrient intake. Excessive inclusion of PKC in dairy cattle may lead to lower TDNI and CPI, which negatively affects milk production, as it can cause milk fat depression [84].

Figure 2:

Estimating the optimal level of palm kernel cake inclusion for dairy cattle.

In contrast, beef cattle tend to be less feeder-selective, making it relatively acceptable for them to tolerate PKC in their diets with less impact on production. The saturated fatty acid content of PKC is also safer for beef cattle, as they do not encounter the risk of milk fat depression that often occurs in dairy cattle. Vargas and Mezzomo [33], through a meta-analysis, determined that the optimal PKC inclusion level for confined cattle was 110.6 g/kg DM without affecting DMI.

For small ruminants and goats (Figure 3), the optimal PKC inclusion levels to support ADG are 106 and 115 g/kg DM, respectively. Regarding FCR, the optimal levels are 139 g/kg DM for small ruminants and 101 g/kg DM for sheep. These findings emphasize the importance of adjusting PKC inclusion levels according to species and production objectives, whether for improving growth performance, enhancing feed efficiency, or ensuring better feed quality. Such adjustments support the effective use of PKC to achieve optimal production outcomes.

Figure 3:

Estimating the optimal level of palm kernel cake inclusion for small ruminants.

Based on previous studies, PKC inclusion in small ruminants has shown variable results according to animal type. Ferreira et al. [14] recommended the use of PKC up to 80 g/kg DM for dairy goats, while in feedlot goats, the recommended levels of PKC inclusion range from 107.7 to 120 g/kg DM to improve growth and meat quality [15], 23], 79]. The range of optimal levels of PKC inclusion for goats from previous studies ranges from 80 to 120 g/kg DM, depending on the production objective.

The buffalo dataset included treatment means from five studies, but not all response variables were consistently reported. For example, only three studies provided data on ADG and FCR at different inclusion levels. This limited coverage reduces statistical power and may not fully capture variability in responses to PKC levels. Additional studies examining growth performance and nutrient utilization at varying PKC inclusion levels in buffaloes are therefore needed to improve model reliability.

It should be noted that the calculated optimal PKC inclusion levels are based solely on intake and performance responses. Economic aspects, such as feed cost or profitability, were not considered because the included studies did not provide such data. Therefore, these optima represent physiological rather than economic recommendations, and future studies should incorporate cost-benefit analysis to determine economically optimal inclusion levels.