Microalgae offer a potential solution to the growing need for more sustainable alternatives to fishmeal and fish oils in aquafeeds, and for healthier, more nutritional substitutes to plant oils 16, but high production costs and wide variation in the purported benefits have so far hampered a greater uptake by industry 48–50. The potential of microalgae to serve as sustainable replacement of animal or plant based protein and oils in aquafeeds has been extensively reviewed in recent years 21,51−59, but surprisingly there is no quantitative global assessment of their nutritional benefits. Without a statistical analysis, it is difficult to determine to what extent the nutritional benefits of microalgae can be extrapolated across species or depend on inclusion levels. For example, some authors have reported negative impacts of Spirulina at high inclusion levels in some species, while others have found no such constraints 51. To address these issues, we conducted a rigorous meta-analysis on the nutritional benefits of incorporating two of the most important microalgae, Spirulina and Schizochytrium, into aquafeeds for use in fish farming, assessed the extent and sources of variation, and critically examined various potential sources of bias.
Benefits of Spirulina replacement on fish growth
The results of our meta-analysis show that partial replacement of fish meal with Spirulina can have a significant positive effect on fish growth, with benefits being apparent from very modest inclusion levels, 1% and less 45. However, growth benefits are dose-dependent and higher inclusion levels of Spirulina results in better growth, 45% being the maximum Spirulina replacement considered. Growth was improved in 71% of the 17 species examined, but the best results occurred among the Cichlidae (tilapias), Clariidae (airbreathing catfishes), and Mugilidae (mullets), species which are all herbivorous.
Negative results were also found, although these instances were rare. Loss of weight compared to controls following replacement with Spirulina was reported in 5% of studies (Fig. 10) and involved three species: Nile tilapia at 2-2.7%% replacement 60,61, mullet at 3.9% replacement 62 and rainbow trout at 0.1-4% replacement 45,63. In most cases (95%), however, Spirulina either improved growth or had no negative effect compared to controls, and replacements of up to 40–45% have been used without detrimental impacts in several species 62,64−66.
Benefits of Schizochytrium replacement on omega-3 fillet content
Ingestion of suitable sources of omega-3 PUFA is essential for proper egg development and offspring survival 67 and Schizochytrium represents a sustainable and rich source of DHA for maturing fish 68. Moreover, given the importance of the early environmental conditions for subsequent development 69,70, the essential fatty acids provided by Schizochytrium and other similar thraustochytrids can have long-term beneficial effects fish health and growth, as seen in Siberian sturgeon, 71, Nile tilapia 26,27,46,72, red sea bream 73, channel catfish 74, and jade perch 47.
We did not find a positive global increase in omega-3 in the fish fillet compared to controls, but the mean SMD was not statistically different from zero, indicating that replacement of fish or plant oil with Schizochytrium oil is possible without a significant loss of omega-3 content. Indeed, positive or neutral (i.e., zero-effect) results were reported in 74% of the trials (Fig. 10). The 26% of cases where the omega-3 content of the fish fillet deteriorated after Schizochytrium inclusion refer to studies involving five species: red drum 75, hybrid striped bass 76, Atlantic salmon 77, red seabream 78, and gilt-head bream 79. The absence of a dose effect means that 100% substitution of animal or plant oils with Schizochytrium oil is possible and should not decrease the nutritional value of the fish fillet, as demonstrated for Nile tilapia 26, although variability is very high and the prediction interval wide, and this introduces considerable uncertainty on the expected results.
Heterogeneity between studies and sources of variation
We found substantial heterogeneity in the results of fish feeding studies using Spirulina (I2 = 96%) and Schizochytrium (I2 = 89%). The relative frequency of different outcomes (positive, neutral, and negative results) differs significantly between Spirulina and Schizochytrium studies (χ2 = 11.197, df = 2, P = 0.004; Fig. 10). Non-negative results (i.e. positive plus neutral) were more common for Spirulina effects on growth (94%) than for Schizochytrium effects on omega-3 fillet content (74%), confirming the results of the two meta-analyses, which yielded a significant non-zero global effect for Spirulina (95%CI SMD = 0.71–1.70) but included zero in the case of Schizochytrium (95%CI SMD = -0.51-1.76). Highly variable outcomes are common in microalgal studies. For example, Ahmad, et al. 80 reported 36% significant improvements in 11 studies that examined changes in growth or fillet quality following inclusion of Chlorella vulgaris in aquafeeds, 36% with no discernible benefit, and 27% negative effects, which were apparently exacerbated at high inclusion levels.
High heterogeneity in meta-analysis is problematic because it makes it difficult to generalize across contexts 35,81. Heterogeneity can be caused by clinical (or structural) differences between subjects and how they respond to treatments, but also by methodological differences in study design, and by statistical variation in intervention effects 82,83. We dealt with high heterogeneity by performing meta-regression and by conducting subgroup analysis 84. We found that family effects were the main source of heterogeneity, but this only explained a small part of the observed variation (~ 24–27%). Most of the variation could not be explained by differences in the way different fish families responded to microalgae replacement, or by variation in microalgae inclusion levels, differences in fish size, habitat, feeding guild or the way the data were recorded.
It is likely that other, unaccounted, biotic and abiotic sources of variation contributed to the high observed level of heterogeneity 85. For example, fish growth can vary enormously depending on sex and stocking density 86, water temperature 87, photoperiod 88, light intensity 89, tank size 90, tank colour 89, social status 91, trial duration, seasonality and feeding rates 92. These are likely to differ between studies but are seldom reported. Likewise, substantial variation has also been reported in the fatty acid composition of fish fed identical diets under communal rearing conditions 93, suggesting that individual differences in deposition of omega-3 can be substantial. The nutritional value of micro-algae also differs between strains and producers 94, depending on culture conditions 95, geographic location 96, and post-harvest treatment 97–99 adding additional sources of unaccounted variation.
Publication bias
We found no clear evidence of systematic publication bias. Plotting effect sizes against standard error of the estimates resulted in asymmetric funnel plots for both Spirulina and Schizochytrium which can be indicative of publication bias 100. However, asymmetry could not be confirmed by the more explicit Egger’s tests 40 in the case of Schizochytrium and the results of p-curve analysis 42 indicated that there was sufficient evidential value for both micro-algae, suggesting there was an underlying true effect. Publication bias could have been masked by high study heterogeneity which may have diminished the power of the p-curve method 101, but our sensitivity analysis indicates that the pooled effect sizes calculated for Spirulina and Schizochytrium were robust to the exclusion of outliers and overly influential points.
Wider benefits of using microalgae in aquafeeds
There are over 40 different species of micro-algae used in fish farming, but these
are mostly used to feed rotifers and copepods to wean fish larvae, or are administered live directly to fish reared in ‘green waters’ 30,102. Only ~ 19 microalgae are used as part of formulated aquafeeds 16,103, the production being dominated by freshwater species such as Spirulina, which is the dominant species with 41% of the global market due its ease of culture, nutrient profile, and high yield 104.
Although live microalgae are a staple feed in many fish hatcheries 105, ingestion rates are difficult to control in ‘green waters’ and their use is typically restricted to larval stages. In contrast microalgae-based aquafeeds can be used at all stages of fish development, offering superior control over feeding, necessary for precision aquaculture 106. Also, unlike plant-based aquafeeds that are difficult to be accepted by carnivorous species 107, microalgae incorporated into aquafeeds can be used to feed both carnivorous and herbivorous species 59. Many microalgae have rigid cell walls which results in low digestibility 108, but new technical solutions are being developed to overcome this challenge 59,109,110.
Not all species are as rich in omega-3 PUFA as Schizochytrium 26, or have the high protein content of Spirulina (~ 63–65%) to replace fish meal 98 but combining different microalgae can overcome this limitation. For example, Schizochytrium represents a good source of DHA for maturing fish, but is poor in EPA 68, but by combining it with oil from Nannochloropsis which is rich in EPA 111 an appropriate balance of omega-3 fatty acids can be ensured, necessary for the production of high quality gametes 112. Likewise, while Schizochytrium oil possess a nutritional profile comparable to fish oil 26,27,96, Spirulina lacks essential amino-acids compared to fish meal, which can potentially reduce growth at high inclusion levels for some species 21,113. Thus, different combinations of microalgae may be required to meet the nutritional needs of different fish species 114. Yet, few studies have compared the benefits of combining different proportions of microalgae and this is an area where more research is clearly needed.
One advantage of microalgae over plant-based aquafeeds is that their benefits are not limited to enhanced growth or nutritional value, but can also extend to fish health 16,115. Microalgae are increasingly being considered for their therapeutic properties, in addition to their nutritional aspects 116. For example, Spirulina and Chlorella can boost the immune system of fish 80,117, and Spirulina may also have anti-viral properties 118. Incorporation of Spirulina in the fish diet was reported to enhance hepatic antioxidant function and disease resistance in coral trout, Plectropomus leopardus 119, great sturgeon, Huso huso 120, Nile tilapia 60,121,122, African catfish 123, mullet 124, as well as in several cyprinids 24,125 and salmonids 45,126. Inclusion of Spirulina at 8–10% was also found to increase fecundity in three-spot gourami 127.
Maximizing the value of feeding studies using microalgae
Microalgae can provide substantial benefits to aquaculture nutrition but only if results can be replicated and can be used by the aquafeed industry 8,128. In common with other meta-analysis in aquaculture 129, we found it difficult to extract the necessary information from fish feeding trials to ascertain effect sizes. A surprisingly large number of studies do not provide enough information to replicate the work, or to ascertain the experimental validity of the results. Of 1,474 studies we screened, only 3% were eligible for analysis. We urge authors to adhere to accepted guidelines for reporting results of fish studies, including mean effects, sample sizes and measures of variability 130, as well as ethical considerations 131.
In the studies reviewed, 14% of trials involved batch measurements in the case of Spirulina and 23% in the case of Schizochytrium, and this may have also introduced some biases. Batch measurements are not recommended as they can mask important sources of variation, reduce sample size (and thus statistical power) and may result in inflated effect sizes, which can be misleading. It might be beneficial for future meta-analysis to weigh studies by some measure of reliability 41,81.
The unit of replication should also take into account the nested nature of the data and the statistical power to detect differences, particularly in growth studies 132,133. For example, there is little benefit in using triplicate tanks if tank effects are ignored and data are pooled. Fish can now be individually marked since a young age 134, which is essential for precision fish farming 106, and tank effects can be accounted for using linear mixed effects models 132.
All results we reviewed were based on feeding trials typically carried out in comparatively small tanks or enclosures under relatively low densities, which are unlikely to be representative of commercial conditions. Given the high heterogeneity found in effect sizes, there is some uncertainty about the wider applicability of the reported results. There is clearly a need to examine the performance of algae-enriched aquafeeds under commercially relevant conditions that extend over longer time periods than the average 60-day feeding trial to ascertain the validity and potential limitations of upscaling 135.
Outlook and Conclusions
Our meta-analyses examined the nutritional benefits of only two species of microalgae, Spirulina and Schizochytrium, but these represent the main ones, and also the ones with enough quantitative data to conduct a statistical analysis. The results indicate that inclusion of Spirulina in the fish diet improves standard growth rate overall, while replacement of fish or plant oil with Schizochytrium oil is possible without loss of omega-3 content in the fish fillet in 71% of the cases involving 91% of the species examined. However, the results are very heterogenous and depend not only on fish species, but in the case of Spirulina also on inclusion level, as well as on unaccounted sources of variation likely related to differences in the way feeding trials were conducted. We found no clear evidence of publication bias, and the results were generally robust to exclusion of extreme values.
The Aquaculture industry will be worth $50.6 billion by 2026 136, the main cost of which will continue to be the cost of aquafeeds 137 . The use of microalgae in aquafeeds is still more expensive than using fishmeal, fish oils or plant crops 48 but the price of fish meal has increased more than 200% over the last two decades 138 . As microalgae production becomes cheaper and more efficient 20 , microalgal-based aquafeeds will become more competitive 139 . Production of Spirulina is expected to be worth $4.6 billion by 2027 140, mostly driven by the nutraceutical, food and beverage segment, but also by aquaculture 141 . To speed the transition towards more sustainable, zero-catch aquafeeds, rigorous comparative analyses of novel microalgal diets are needed, but results need to be directly relevant to the feed manufacturers 142 . To achieve this, we recommend that feeding trials using microalgae are conducted under commercially relevant conditions that consider the challenges of upscaling, and that the raw data and full rearing details are reported, along with units of replication, mean effects, sample sizes and measures of variability.