Feed Fortification: Final Findings
Summary
Since 2024, we have been exploring if a custom-formulated fish feed improves fish welfare compared to the feed commonly utilized by farmers of Indian major carp in Andhra Pradesh. We have conducted a series of studies to provide the evidence we needed to determine whether or not to take forward a program based around this “fortified feed”.
Unfortunately, our results across different study types show that the evidence does not support moving forward with a program. While our fortified feed, compared to the regularly-used feed type (DORB), consistently showed improved growth rates of fishes (good for farmers, and perhaps good for the fishes from a nutritional/welfare perspective), it offered no discernible direct (as assessed by health, behavior, and physiological indicators of fish stress) or indirect (as assessed by water quality) welfare advantages. Aside from a lack of clear welfare benefits, the cost of using fortified feed as a replacement for DORB would likely be a deterrent to farmers.
Overall, we have concluded that there is not a viable path forward for a program built around our current formulation of fortified feed.
Background
Feed quality and management are critical determinants of fish welfare. Most farmers of Indian major carp (IMC) in Andhra Pradesh use de-oiled rice bran (DORB), a by-product of rice production, as the primary supplemental feed during the grow-out stage (“supplemental feed”, as IMC also feed on naturally occurring phytoplankton). DORB is the feed type of choice for IMC farmers in Andhra Pradesh due to its low cost. Cost is a critical consideration for IMC farmers, as margins are so low. There are other feed options available that have improved nutritional value, but these “premium” feeds are not considered cost-effective for widespread use by IMC farmers (they are sometimes used towards the end of the grow-out phase, but they are not routinely used throughout the farming process). FWI believes that DORB’s poor nutritional value, coupled with the presence of mycotoxins and other “antinutritional factors” (components that reduce nutrient absorption and digestibility), likely impacts the fishes’ growth and development, increases their susceptibility to disease, and ultimately impacts their overall health and welfare. Apart from its poor nutritional value, feed management practices that farmers typically follow for DORB—which is formulated as a powder-like mash—contribute to poor water quality, compounding welfare issues for the fishes.
Since 2024, we have been exploring if a modified version of DORB, supplemented with a small number of nutrients and formulated as a pellet, can improve fish welfare while being a viable option for IMC grow-out farmers to utilize in place of conventional DORB.
In February of this year we published a blog post announcing a set of experiments that we were initiating to provide the evidence we needed to determine whether or not to take forward a program based around “fortified feed”. In this current blog post, we provide a summary of the key findings from the complete set of experiments we have run as part of this overall initiative, and describe how those findings have informed our decision about a future program.
What We Tested
Developed by our Fish Welfare Specialist, Dr. Cerqueira, to be both nutritionally superior and cost-effective compared to DORB, our fortified feed was produced at a local feed mill in Andhra Pradesh. DORB was selected as the basis for our customized feed (the final product is approximately 50% DORB), as our ultimate aim was to present farmers with a cost-effective product that they are likely to accept. Given the low cost and widespread use of DORB, we reasoned that a fortified version of DORB—rather than formulating a whole new bespoke feed—was the best strategy to pursue. The formulation of our fortified feed is shown in Table 1.
1 The formulation of fortified feed used for our first study (a pond-based study) differed slightly from the formulation used for later studies (tank-based and field-based studies). Due to issues—supply and quality—with mustard cake, we partially replaced mustard cake with groundnut cake in the formulation used for the later studies. This helped to reduce antinutritional factors, but also helped to increase the overall protein content of the feed. We slightly boosted the levels of vitamins and minerals, as well as mycotoxin binders, to compensate for adding groundnut cake. Additionally, we took the opportunity to further improve the formulation by adding phytase (see note 6). This means that the pond-based study tested an earlier version of the feed, with lower overall protein content than the version tested in the tank-based and field-based studies.
2 Protein levels can vary for DORB produced by different rice mills, and from batch to batch. The DORB used as the control (and the basis of our fortified feed) for our tank-based and field-based studies had an average protein content of approximately 17%. The DORB used as the basis for our earlier pond-based study had an average protein content of approximately 15%.
3 Several ingredients were incorporated to increase the protein content and diversify the protein sources of the final feed.
4 Included to provide the essential micronutrients that are typically deficient in plant-based diets, including vitamins A, D₃, E, K, water-soluble vitamins (B-complex, C), and essential minerals (calcium, phosphorus, magnesium, iron, zinc, manganese, copper, selenium).
5 Included to mitigate the effects of naturally occurring mycotoxins (notably aflatoxins and ochratoxins) frequently present in rice and groundnut products, which can impair immune function, reduce growth performance, and compromise liver health in fishes. Aflatoxin levels in all batches of fortified feed we manufactured were below the limits of detection for the analytical methods used.
6 Included to improve the availability of phosphorus and minerals trapped in phytic acid, which is present in high concentrations in DORB.
While improved welfare is FWI’s primary concern, a future program built around provision of fortified feed needs to consider the economics from the perspective of the fish farmers. We were not aiming to position our fortified feed as a premium feed; rather, our intention was to create a feed type that provided significant welfare improvements for fishes—compared to DORB—at a price likely to be viable for farmers to use throughout their grow-out cycle. We knew from the beginning that upfront costs (cost to procure each ton) for fortified feed would be more expensive than DORB, but the hope was that long-term costs (overall feed costs in relation to the biomass of fish at the end of the grow-out cycle) would be lower compared to DORB, due to less feed being needed.
Our Research Journey
Our research question was as follows: Does a modified formulation of DORB—supplemented with selected nutrients and formulated as a pellet—improve the welfare of fishes through improving their nutritional requirements and quality of the water in which they live, while resulting in fishes that are at least the same size as those fed on conventional DORB?
To answer this research question, we conducted a series of experiments in different types of settings: experimental ponds, a laboratory, and real fish farms. Over the course of our studies, the key metrics for assessing the performance of fortified feed were:
Weight and feed-related indicators: Do fishes eat the fortified feed, and if yes, how does their weight/size gain compare to fishes fed on conventional DORB? (Note: while growth is not necessarily an indicator of welfare, if the feed is sufficiently nutritious, fishes should grow well.)
Water quality indicators: Does the use of fortified feed result in improved water quality of ponds compared to ponds where conventional DORB is used?
Health and behavioral indicators: Is the welfare of fishes, as measured by the health, condition, and behavior of fishes, improved when fortified feed is used? (Measured in the tank-based and field-based studies only.)
Physiological indicators: Is the welfare of fishes, as measured by physiological indicators of the fish stress response, improved when fortified feed is used?
We conducted two types of “efficacy” studies: one pond-based study, and a tank (laboratory)-based study. As efficacy studies, they were 100% conducted by our research team with no involvement of farmers, and were conducted under more controlled conditions than are possible with a study conducted at real farms.
The single most important study we conducted—vis-à-vis providing the most critical evidence for proceeding, or not, with a program—was a field-based “effectiveness” study of our fortified feed formulation. As an effectiveness study, this was designed to assess fortified feed when used by real farmers at real farms under real-world conditions (i.e., the study was conducted in a real-world setting involving the stakeholders who would ultimately be targeted by a program, with the research team only involved in coordination and data collection).
Below we summarize each of these studies: what we did and what we found.
Pond-Based Efficacy Study
What We Did
We started off on our journey of investigating the potential for fortified feed by conducting a study in test ponds at our research facility in Rajahmundry, Andhra Pradesh, in the second half of 2024 (see our earlier blog post announcing the start of this study in June 2024). This research facility, which we have since decommissioned, included six earthen test ponds, each of which was approximately 1/6th of an acre. While these test ponds were intended to replicate real-world conditions (i.e., real fish farms) as much as possible, it should be noted that studies in a research facility such as this are not fully representative of real fish farms. The advantage of test ponds for study purposes is that they allow for more controlled conditions than are possible in real working fish farms.
This study involved approximately 2,700 fishes in total, with each of the six test ponds stocked with approximately 450 fishes. Fishes in two ponds were fed with conventional DORB for a period of four months, while fishes in two ponds were fed with fortified feed. We also included a “positive control”: a commercially available pellet feed, considered nutritionally superior to both conventional DORB and our fortified feed, was used to feed fishes in two other ponds. Every month, we collected a sample of fishes from each pond and weighed and measured them (all fishes were returned to the ponds immediately afterwards). Blood samples were also taken at the end of the study (day 120) for analysis of physiological indicators of fish stress response. Additionally, we regularly tracked water quality.
What We Found
After four months, we found a statistical difference between fortified feed and conventional DORB for weight/feed-related indicators. Fishes fed fortified feed gained on average 32% more weight than those fed conventional DORB. Fortified feed had a better “feed conversion ratio” (FCR) than DORB. This is a critical metric used for comparing feed types in aquaculture, measuring how efficiently fishes convert food into body mass: a lower FCR indicates that food is more efficiently utilized by the fishes (less feed but more weight gain). While a lower (better) FCR is not a direct indicator of a feed being more nutritious than a feed with a higher FCR, it can be an indirect indicator that the feed is more nutritious. The study revealed an average FCR for fortified feed of 2.1, compared to an average of 3.0 for DORB. This means that, on average, it took 2.1 kg of fortified feed—0.9 kg less than DORB—for every kilogram of fish biomass gained.
However, there were no statistical differences for water quality. Of the five physiological indicators of stress we had planned to measure, three (cortisol, lactate and glutathione) could not be reliably analyzed because the laboratory results did not pass our quality checks. The two that could be analyzed (plasma protein and triglycerides) showed no significant differences between diets.
Surprisingly, fishes fed fortified feed grew more than those fed with the commercial “positive control” feed. This was unexpected, and likely explained by issues we encountered during the study: a significant invasion of tilapia in one pond (a positive control pond), a lice outbreak across all ponds, and a significant overgrowth of vegetation in the pond bottom of another. For a feed-based study, these issues were problematic. Indeed, we had a high degree of variability in the data across the ponds/diets, which was likely—fully or partially—due to these issues we encountered during the study. This left us cautious about how to interpret the results. Of the various studies we conducted to test fortified feed, we have the least confidence in the data from this study.
Overall, this study did not provide the evidence we were hoping for to inform moving directly to a resource-intensive field-based study. However, recognizing that the issues we encountered may have hidden any benefits of fortified feed, and encouraged by the superior growth performance, we decided to conduct additional experiments.
Tank-Based Efficacy Study
What We Did
Our pond-based study was complicated by issues we encountered during the study. This prompted us to run a laboratory-based version of the study, with the rationale being that the issues that plagued us during that pond-based study—which complicated how we interpreted the findings—are less likely to occur in a controlled laboratory setting. Accordingly, we decided to conduct a study in tanks, similar to how we conducted our Dissolved Oxygen Tolerance study in 2025. This study was conducted at our test facility in Bikkavolu in Andhra Pradesh in early 2026.
This study involved 120 fishes in total across six experimental tanks (450 L in size). The fishes in three tanks were fed with fortified feed, and the fishes in the other three tanks were fed with conventional DORB. Every month, all fishes were weighed, measured and directly observed for health and condition (note: fishes were observed throughout the study for any signs of distress). Blood samples were also taken at these sampling points for analysis of physiological indicators of fish stress response. Additionally, we regularly tracked water quality and behavioral indicators. The study was originally planned to last 60 days, but we extended it to 90 days because fishes were slow to accept the pellet in the first month.
What We Found
After three months, fishes fed fortified feed gained 17% more weight and grew 13% faster on average than fishes fed conventional DORB. In the first month, fishes fed with fortified feed were slow to accept the unfamiliar pellet and ate about 35% less than those fed DORB (during the first month, uneaten pellets were removed after an hour, but we subsequently modified the protocol to give fishes more time to eat the fortified feed). This represents a short-term nutritional welfare cost of switching feed. As we found with the earlier pond-based study, there were no discernible differences in welfare-related indicators (health, behavior or stress responses) across the two groups of fishes fed the different diets. The study design meant we couldn’t assess differences in water quality, so we conducted a separate experiment specifically to assess this issue (see below).
Tank-Based Study to Assess Water Quality
What We Did
Shortly before the tank-based efficacy study, we conducted a quick screening study in the same tanks that specifically focused on assessing if fortified feed is less polluting and better for water quality than conventional DORB. Comparing any differential impact on water quality from the two feeds in the tank-based efficacy study (described above) would have been difficult—if not impossible—because that study necessitated constant water exchange to prevent the buildup of toxic levels of ammonia and to maintain favorable levels of dissolved oxygen in the study tanks.
For this “nutrient-leaching” study, we used the same tanks/facility as used for the tank-based efficacy study, but no fishes were involved. Rather, we allowed both feeds (in three tanks each) to dissolve in stagnated water over 24 hours, and collected samples at predefined intervals (0, 1, 4, 8, 12, 24 hours) and assessed the following water quality parameters: total nitrogen, ammonia, total phosphorus, dissolved phosphorus, pH, and turbidity.
What We Found
There were no statistical differences between the two groups for the amounts of nutrients leached into the water, or the pH. The key difference was turbidity. DORB led to a rapid increase in the turbidity of the water, whereas for fortified feed, turbidity slowly increased over time (albeit by the final time point—24 hours—it exceeded the peak turbidity levels observed in the DORB tanks; Figure 1). This difference is due to the mash-formulation of DORB allowing it to disperse almost instantly, causing a sharp spike in turbidity, while the pellet formulation of fortified feed allows it to remain largely intact initially, requiring time to break down.
While the turbidity of the water in the tanks with fortified feed eventually exceeded the peak levels in tanks with DORB, the findings suggest a potential advantage to using fortified feed in real ponds. In a real-world environment, the slower breakdown of the fortified pellet and the concomitant slower increase in turbidity should give the fishes a longer window to consume the feed before turbidity becomes an issue and nutrients are lost to the water (in the tank-based nutrient-leaching study there were no fishes to consume the feed). However, we did not test this with fishes present, and in our field study (see below) we found no differences in water quality between the two feeds.
Figure 1: Profile of water turbidity over the course of 24 hours. When conventional DORB (orange line) is added to water, turbidity instantly spikes, but decreases over time as DORB settles to the bottom. Fortified feed (blue line) shows a different temporal pattern. The pellet formulation means that fortified feed floats on the water surface, so there is no immediate effect on turbidity. Turbidity increases over time as the pellet slowly disintegrates, reaching a peak at 24 hours (the final sampling point during the study), slightly exceeding the peak levels of conventional DORB, but much later.
Field-Based Effectiveness Study
What We Did
This study was conducted at twelve working fish farms in the Nellore region of Andhra Pradesh, between February and August 2026. Half of the farms were provided with fortified feed, and the other half were provided with conventional DORB. Farmers were tasked with feeding their fishes, and FWI staff members collected data throughout the study period (four months at each study farm). Note that all farmers involved in this study were farming fishes regardless of this study, so we did not procure fishes specifically for this study. Every month, we collected a sample of fishes at each farm and weighed and measured them, and directly observed their health and condition (fishes were returned to the ponds after the sampling process). Blood samples were taken on days 0, 60 and 120 for analysis of physiological indicators of fish stress response. Water quality indicators and indicators of behavior were recorded every two days during the study period.
What We Found
Compared to DORB, our fortified feed improved growth rates of fishes (good for farmers, and perhaps good for the fishes from a nutritional/welfare perspective). On average, fishes gained 73% more weight after four months. Fortified feed had a better FCR than DORB: an average FCR for fortified feed of 2.8 compared to an average of 4.3 for DORB. This means that, on average, it took 2.8 kg of fortified feed—1.5 kg less than DORB—for every kilogram of fish biomass gained. This suggests—but does not confirm—that fortified feed may be more nutritious than DORB. Plasma protein was higher in fishes fed fortified feed (22.9 vs 20.5 g/L, p = 0.013). This is consistent with better nutritional status, although it is a single indicator.
However, there were no discernible differences in welfare-related indicators across the two groups of fishes fed the two different diets:
Water quality over the course of the study showed no differences.
Health indicators (i.e., direct examination of fishes) showed no differences.
Behavioral indicators showed no differences, apart from anticipatory behavior, which was worse in fishes fed fortified feed—a difference driven by the first month, likely as fishes adjusted to the new feed (speed of feeding and time to finish feeding also differed, but these reflect the physical form of each feed rather than the welfare of the fishes).
Physiological indicators of stress—cortisol and glucose—showed no differences.
Low dissolved oxygen was common at all farms, regardless of the feed used. Across both groups, about 45% of early-morning measurements of dissolved oxygen were below 3 mg/L, the welfare threshold for dissolved oxygen that FWI applies. Some ponds had weeks of consecutive low readings. This was the main welfare problem we observed in the field study.
Aside from assessing growth and welfare, another key purpose of the study was to assess if fortified feed was an economically viable option for farmers, compared to DORB.
The cost of fortified feed is higher than anticipated. The plan had been to price fortified feed in-between DORB and premium feeds, but current market conditions mean we have not achieved this goal, unfortunately: since we first formulated our feed (in 2024), the cost of fortified feed has crept closer to the cost of premium feeds. These costs have been impacted by recent increases in the cost of DORB and other ingredients needed for fortified feed (our formulation is about 50% DORB), so these costs could come down in future (or with scaled-up production).
Lower quantities of fortified feed were needed compared to DORB over the course of the study—as indicated by the lower FCR—but the projected overall cost to a farmer over the course of a whole farming cycle is higher for fortified feed.
Overall, the price differential—DORB vs fortified feed vs premium feed variants—is a concern from the perspective of enticing DORB-using farmers to switch to fortified feed. While we managed to lower the FCR (a key success), unfortunately the costs of fortified feed coupled with its FCR means that the economics from the farmer’s perspective still favor DORB.
Similar to the pond-based efficacy study, we had some confounding factors to deal with during the field-based study. Most notably, lice outbreaks occurred at some ponds. Because this was an effectiveness study, farmers managed their own ponds—including treating for lice—as per their normal management practices. At nine of the twelve farms (five using fortified feed and four using DORB) farmers treated fishes for lice, with treatment regimens differing across affected farms. Aside from treating for lice, general management practices differed somewhat across farms. Farms using DORB more often applied products such as jaggery, zeolite, probiotics and molasses. Most farms also recorded some days without feeding. These practices reflect real farming conditions—which is exactly why we conducted an effectiveness study at real fish farms—but they add variability and could affect how results are interpreted.
What The Findings Mean
Across our studies, we consistently showed that fortified feed—compared to DORB—improved the growth of fishes. However, the studies did not reveal any discernible direct (as assessed by health, behavior, and physiological indicators of stress) or indirect (as assessed by water quality) welfare advantages. Fortified feed may improve the nutritional domain of welfare, as assessed by increased weight (all studies) and plasma protein levels (field-based study). However, the strength of evidence for welfare improvement is not sufficiently strong to allow us to be confident that there is a sufficiently meaningful welfare benefit to fortified feed compared to DORB. For us to move forward with a program based around fortified feed, we want to be sure it offers clear welfare benefits to the fishes.
Aside from a lack of clear welfare benefits, the cost of using fortified feed as a replacement for DORB would likely be a deterrent to farmers.
The original rationale for this project was predicated on two things: (1) FWI being confident that there were welfare benefits to using fortified feed, and (2) using fortified feed would not negatively impact farmers’ margins. The findings from our studies show that we have not met these requirements.
Loss of Fishes During the Studies
Conducting studies with live fishes always comes with a risk of harming or losing some of them. This is a risk we simply must be willing to accept to make meaningful changes for fishes in the long term, but we mitigate that risk as much as we can. Our study protocols carefully consider the risk of suffering that fishes will endure, and our teams are trained to act swiftly in response to any issues that arise. Importantly, we never purposely sacrifice fishes as part of a study.
Unfortunately, during the course of the pond-based study, we lost eleven fishes: six fishes died as a result of bird predation, four fishes died as a result of the lice outbreak, and one fish died when being transferred to our facility. We took swift actions in response to issues that arose, including installing bird netting to mitigate against bird predation and applying treatments quickly to respond to lice outbreaks. At the end of the study, all fishes were released into a river in Andhra Pradesh, with the approval of relevant authorities (see our earlier blog post).
In the tank-based study, we are happy to report that there were no mortalities and no issues of concern. As with the pond-based study, all fishes were released into a natural water body at the end of the study.
The fishes in the field-based study belonged to the farmers and stayed on their farms after the study. No mortalities were recorded in ten of the twelve ponds during the study, although deaths in earthen ponds can go unnoticed. In two ponds (both using DORB), the field notes record nine fishes in total were lost.
Path Forward
Based on the findings from our various studies, we have concluded that there is not a viable path forward for a program built around our current formulation of fortified feed.
While this is a disappointing outcome after all our work on this project—the team was optimistic throughout that fortified feed would offer significant benefits to fishes and farmers—this is exactly why we use evidence to inform our programmatic decision-making. The study may not have given us the results we had hoped for, but we achieved what we had set out to do: to gather evidence to make an informed go/no-go programmatic decision. Suffice it to say, we would have preferred a more positive result so that we could proceed with a new program. However, the findings from the studies, and our “no-go” decision, give us clarity that we need to keep pushing on so that we can find something that does work—we need to direct resources elsewhere, and try again!
We are now directing resources to other projects in the hope of identifying suitable interventions that will improve fish welfare. Two such projects we have recently embarked on are a study focused on early identification of lice, and an initiative whereby we are opening up the process of identifying and testing potential interventions to third parties.
We have learned lessons during our research journey to test fortified feed. We intend to share these lessons in an upcoming blog post, as we think these lessons would be useful for other organizations embarking on studies to inform interventions to improve animal welfare.
