5. Protein fiasco

In 1933, Cicely Williams, a doctor in the British Colonial Medical Service posted to the Gold Coast (modern-day Ghana), had published a paper describing a disease afflicting children in her care. The disease caused oedema (swelling), wasting, diarrhoea, sores, discolouration and loss of patches of skin, and eventually death. It typically affected children a short while after weaning, particularly in cases where the child was mainly living on maize. Williams suspected the cause was dietary and had been giving an inpatient treatment diet including Marmite, fruit, cod-liver oil, tinned milk and malt. Her impression was that “the most important elements in the treatment” were the cod-liver oil and milk, and she recommended treatment with Nestlé’s sweetened condensed milk.¹⁶⁸ Commenting on the association with a maize diet she cautiously stated only that “some amino acid or protein deficiency cannot be excluded as a cause.”¹⁶⁹ Two years later she published a further paper rebuffing criticism¹⁷⁰ that she was just misdiagnosing pellagra (vitamin B3 deficiency) and naming the condition kwashiorkor, a term borrowed from the Gã language which she explained “indicates the disease the deposed baby gets when the next one is born.”¹⁷¹

Williams’ work didn’t get much attention at first. Few other scholars cited her papers or used the term kwashiorkor, and in 1936 she was transferred to Malaya.¹⁷² Other scholars continued to study the disease but accepted neither Williams’ name for it nor the suggestion that it was due to an amino acid or protein deficiency — instead, they explored the idea that this was a peculiar infant presentation of pellagra or that malnutrition was just one component of a complex of causes.¹⁷³ The choice of name may have undermined her credibility — as her identities as a woman and as a medical practitioner rather than a researcher certainly did. But after the war, this was to change.

In 1949, the newly-founded World Health Organization (WHO) and Food and Agriculture Organization (FAO), picking up the role of the LNHO, met to co-ordinate their work on nutrition. One item on the agenda was “a syndrome at present ill-defined” but “[o]ne of the most widespread nutritional disorders in tropical and sub-tropical areas”, particularly affecting infants and young children in “some parts of Africa”.¹⁷⁴ The committee listed a series of names used for this syndrome but gave the agenda item the heading “Kwashiorkor” (in quotation marks). They offered no definitive account of its aetiology, mentioning cirrhosis of the liver, the diets of mothers and children, and tropical parasitism as points on which to focus further research. In their next agenda item, however, they discussed problems of malnutrition in young children after weaning “in some countries” more broadly, and emphasised that they believed these were particularly due to lack of (cows’) milk.¹⁷⁵

The inverted commas around the word kwashiorkor, the list of other possible names and lack of a singular diagnosis show that expert opinion had yet to settle: the reader gets the impression that the choice of the name ‘kwashiorkor’ (and not, for example, ‘polydeficiency disease’) for the heading was arbitrary. However, a focus on this particular presentation of poor infant health and a set of associations with milk and inferred specific nutrient deficiencies were clearly developing. Over the following few years, various WHO and FAO studies added to the impression that kwashiorkor was a widespread problem — not only in children but also in adults, and not only in Africa but also in South America and India.¹⁷⁶ Just as Cicely William’s work had before the war, many of these studies focused on the ‘exotic’ diets of the populations under study, with high consumption of maize and starchy roots and low consumption of animal proteins. Quickly, the consensus shifted towards seeing kwashiorkor as the result of protein deficiency. Particularly influential was a 1952 study by J. F. Brock (of the WHO) and M. Autret (of the FAO) entitled ‘Kwashiorkor in Africa’, which concluded that none of the traditional foodstuffs of the areas they had studied contained enough protein to meet the needs of newly-weaned infants,^177,178 and that the only ethnic groups that did not suffer from the disease were those who consumed large quantities of cows’ milk.¹⁷⁹ The sense of alarm around kwashiorkor grew, and in 1955, the WHO’s Nutrition Section stated that “kwashiorkor is without doubt the most important nutritional public health problem of the present time”.¹⁸⁰ In the same year, the UN established a special Protein Advisory Group.¹⁸¹ The stage was thus set for nearly two decades of myopic focus on protein in international development.

An important concept informing discourse through this period was that of the ‘protein gap’: the gap between the amount of protein needed to maintain the world population in good health and the amount actually being produced. Part of this was a projection into the future: current agriculture would not be able to produce enough protein to feed everyone as the world population expanded. But it was rooted in beliefs about current, endemic undernutrition. Even if it was acknowledged that some communities were straightforwardly short of food rather than eating the wrong foods, it was argued that the best way to solve both problems concurrently was by producing more protein: after all, this would also entail producing more calories.

This was world hunger understood through the frame of nutritional science — not as a social problem in need of a political response, but as a scientific problem in need of a technological solution. Initially, it seemed that this technology might be milk. Williams had identified tinned milk as a possible cure for kwashiorkor from the beginning, and even those who believed there was more to the disease than simple protein deficiency recommended skimmed milk powder as treatment.¹⁸² Interest developed in protein quality, and milk was proposed as the reference protein for determining amino acid requirements in young children at the second conference on protein malnutrition in 1955.¹⁸³ UNICEF had already been distributing skimmed milk powder as part of work to alleviate postwar hunger in Europe, so increasing consumption of skimmed milk seemed like the ideal cure.¹⁸⁴ In countries like the UK and US, milk still had great symbolic power as the vehicle of feminine domestic care (Milk Marketing Board slogans at the time included “is your man getting enough?”¹⁸⁵) — given that kwashiorkor was often framed as a maternal failing, it seemed an apt solution. However, WHO and FAO policymakers quickly concluded that increasing dairy production in Africa was impractical, and the population in need was too poor to afford imports. What was needed was a protein-rich food that could be made cheaply, from local ingredients.¹⁸⁶ A founding aim of the Protein Advisory Group was to help find “new protein foods”,¹⁸⁷ and in 1960 it established the Committee on Protein Malnutrition (funded to the tune of $880k by UNICEF and the Rockefeller Foundation) to focus on furthering this goal.¹⁸⁸ The Group put out calls for research into unconventional sources of protein for human consumption during the 1960s,¹⁸⁹ but national governments, UN agencies and private companies had started exploring possible new technologies from an even earlier point.

In the late 1950s, the Chilean and South African governments pursued a project to create a flavourless protein flour from fish. This involved drying and powdering huge quantities of small fish, and removing the fat by centrifuge and by extraction using ethanol and hexane. The resulting meal could be added in small quantities to bread flour. This was done in South Africa at government expense from 1956 to 1959, but stopped when it was realised that the population at risk from kwashiorkor ate very little bread. The Chilean government project was then abandoned in 1961 due to engineering problems, without ever reaching the distribution stage.¹⁹⁰ The US Bureau of Commercial Fisheries pursued a similar project throughout the 1960s, even building a pilot plant and then a production plant — but eventually abandoned it, both due to regulatory barriers and because the fish stocks the project relied on collapsed.¹⁹¹ None of the US-produced fish flour was ever distributed.

Some of these projects did achieve a significant scale, both in the West and the USSR — by the 1980s, the Soviet Union was reporting production of over a million tonnes of SCP per year.¹⁹⁷ What they did not do was solve world hunger or the protein gap. Many SCPs have a higher proportion of nucleic acids than other proteins, and it turned out that humans lack the digestive enzyme possessed by other mammals needed to process nucleic acids in large quantities.¹⁹⁸ This puts a limit on the proportion of dietary protein that can come from SCPs. On top of this, many SCPs were found to produce allergic reactions in a substantial minority of human test subjects. Aside from the health concerns, SCP production required a level of sterilisation and contamination control that challenged contemporary engineering — these were not technologies suited to low-cost expansion to poor areas.¹⁹⁹ As a result, the vast majority of SCP was used as an additive in animal feed in developed countries, not as food in developing countries.

One partial exception was Quorn, a British brand of mycoprotein derived from the fusarium venenatum fungus²⁰⁰ and developed in a collaboration between Rank Hovis McDougall and Imperial Chemical Industries building on infrastructure from ICI’s ‘Pruteen’ process.²⁰¹ Development of Quorn started in 1964 in response to concerns about the protein gap and it remains a highly successful brand of vegetarian meat substitute today, especially in the EU and UK. However, in two senses, this was the exception that proved the rule: it was not grown on cheap waste products but on wheat;²⁰² and it did not come to market until 1985, after international interest in protein had waned.²⁰³

Other techno-protein projects included oilseed flours, peanut flour, peanut milk, soybean flour, fortified chocolate soy milk, cottonseed flour, leaf protein concentrate, synthetic amino acids and lysine-enhanced wheat, maize and rice flours, as well as various projects involving skimmed milk.²⁰⁴ Not all of these ventures were completely without results, but none had any substantial effect on hunger in developing countries or on infant malnutrition. A post-mortem co-written by one of the scientists involved in the US work on fish flour²⁰⁵ suggests several reasons for these failures. For one thing, the incentives were not well aligned, with US and European entrepreneurs favouring projects with the potential to be profit-making and US and European government agencies favouring projects which made for good publicity. Both of these groups were keen to find profitable uses for surplus Western produce — particularly fish and milk.²⁰⁶ In the case of fish and petroleum derivatives, projects were built on assumptions about the abundance of inputs that turned out to be wrong. In almost all of these cases, foods were being designed with no reference to the food cultures of the people they were supposed to help, and little reference to economics or the availability of infrastructure for delivery. It is important to note that none of this is to imply that these projects were carried out under false pretences. Scientists and engineers, aid workers and their institutional backers, entrepreneurs and those in corporate governance — all were urgently trying to respond to real and affecting humanitarian need. Rather, systemic and cultural forces channelled these efforts into ineffectual and even irrelevant projects.

Some protein technologies developed in this period did achieve lasting success, even if they did little to solve the problem of world hunger. In 1965, the Archer Daniels Midland company filed a patent for a dried, textured protein food made from soybeans²⁰⁷ which they would later trademark TVP (‘textured vegetable protein’).²⁰⁸ In the same year, the company took on new leadership in the person of Dwayne Andreas, previously an executive in soy oil processing. Andreas had the “zeal of a missionary for bringing cheap protein to consumers” and liked to “talk about protein shortages and world food problems”.²⁰⁹ He sold Archer Daniels Midland’s $65m chemical division to free up capital and concentrate on soy processing, a decision that would prove lucrative; importantly, the USDA’s Agricultural Research Service would approve the use of TVP to bulk out or replace meat in US school lunches in 1971, creating a huge domestic market.²¹⁰ Another soy derivative, ‘spun soy protein’ from General Mills, had been trademarked ‘Bontrae’ a few years earlier,²¹¹ although it would not achieve the same degree of lasting commercial success.

These companies presented new protein foods not merely as commercial ventures but as vehicles of social good. If nothing was done then “many millions yet unborn [would] die from the consequences of malnutrition”, but since “even starving people are funny about food," inventions were needed to make cheap proteins resemble familiar foods²¹² — specifically, to create analogues of meat.²¹³ Private enterprises of this type were lauded by government actors, operating on the assumption that domestic commercial success was the necessary precursor to wider impact.²¹⁴ A huge range of textured vegetable proteins are still produced for use in meat substitutes and as meat extenders; indeed, Archer Daniels Midland (now styled ADM) is still a major producer. But these products did not play a major role in food aid or efforts to address world food security.

The focus on protein and the story of the impending protein gap had cultural resonance beyond the politics of international development. One event of particular relevance for current debates about protein and sustainability was the publication of Diet for a small planet by Frances Moore Lappé in 1971.²¹⁵ This book — an international bestseller at the time whose influence remains important half a century later — was one of the first to make an environmental argument against eating meat. Lappé highlighted the land use inefficiency of animal agriculture as a means of producing calories and protein, arguing that contemporary world hunger and the protein gap created by expected population growth could be solved by shifting to vegetarian diets. In retrospect, it is this argument that represents the book’s lasting contribution — but much of the text was focused on ‘protein combining’ and the supposed difficulty of getting enough protein with the right balance of amino acids without eating meat. In her later writing Lappé would note that the suggestion that it was possible to eat enough protein without eating meat seemed like “heresy” in 1971,²¹⁶ an observation that 19th century proponents of vegetarianism might have found familiar.

However, by the mid-1960s, new scientific findings had started to complicate the protein gap story: the best estimates of infant dietary protein requirements had come down. Where in 1936 the League of Nations had estimated that a one-year-old infant needed 3.5g of protein per 1kg of body weight per day for healthy growth, and Brock & Autret’s report in 1952 had suggested a rate as high as 3.8g/kg/day, official FAO/WHO publications from 1965 onwards consistently estimated rates below 1.5g/kg/day²¹⁷ (see Figure 7). If the ratio of protein needed to calories needed was lower, this challenged the medical understanding of kwashiorkor: was it really caused by simple protein deficiency? Even if it was, the structural account of the problem was even harder to maintain. The argument of the ‘protein gap’ was that, from a population perspective, kwashiorkor was caused by an insufficient supply of protein even though enough calories were available. Lower protein requirements undermined this, implying that protein deficiency might be a consequence of general lack of food—if enough calories were supplied, enough protein would be also. At the same time, evidence was mounting that, in clinical practice, it was hard to distinguish kwashiorkor from other forms of malnutrition. In a population of children of the same age eating the same diet, some might show symptoms of kwashiorkor and some more general starvation (marasmus).²¹⁸

This initially led to little more than a change of terminology: where previously kwashiorkor and “protein malnutrition” had been identified as the problem, now the phrases “protein-calorie malnutrition” or “protein- energy malnutrition” were used. Yet the rhetoric and the policies remained broadly similar for some time,²¹⁹ with UN reports entitled “Feeding the Expanding World Population: Action to Avert the Impending Protein Crisis” (1967) and “The Protein Problem” (1968). As late as 1972, the UN General Assembly adopted a resolution stating that “protein-calorie malnutrition is the primary cause of high infant and child mortality, reaching from 25 to 30 per cent in many developing countries.”²²⁰

But, just as with falling estimates of adult protein requirements in the run-up to the First World War, the gap between the scientific evidence and the general wisdom could not last forever — and just as before, when the consensus changed, it changed quickly. In 1974, Donald McLaren, a nutritionist who had worked with starving children in India, Tanzania, Lebanon and Jordon,²²¹ published an article in the Lancet under the provocative title “The Great Protein Fiasco”.²²² In this he argued that the importance of protein as both problem and solution had been wildly overestimated for 20 years. It simply wasn’t the case that all or most childhood malnutrition was kwashiorkor, or that kwashiorkor was straightforwardly protein deficiency; there was no longer any good reason to believe that there was a gap between protein production and demand above and beyond the gap between general food production and demand;²²³ and projects to create innovative sources of protein had not and were never going to offer effective solutions to world hunger. The World Food Conference in November of that year focused on an ongoing famine in Bangladesh that had killed 1.5 million people: the Protein Advisory Group (at this point already renamed the Protein-Calorie Advisory Group) was not consulted for the conference and protein featured little in the proceedings.²²⁴ One year later, in spite of strident criticism of McLaren,²²⁵ a review article in Nature concluded that the concept of a worldwide protein gap was no longer viable.²²⁶ Three years on, the Protein-Calorie Advisory Group was disbanded.

There had been voices speaking out against the focus on protein and kwashiorkor throughout these two decades. In McLaren’s paper, Cicely Williams, the ‘discoverer’ of kwashiorkor as protein deficiency, is quoted: “For the last 20 years I’ve been spending my time trying to debunk kwashiorkor.”²²⁷ McLaren later called out other researchers whose work had pushed back against the consensus, including Indian statistician P. V. Sukhatme.²²⁸ So why had the focus of international bodies including the WHO and FAO been so narrow?

Widespread protein deficiency had been an appealing diagnosis for the problem of world hunger. A more complex understanding of the situation, allowing very different stories in different places, would have required reliance more on subjective forms of evidence: appeals to personal experience, culture, context or politics. The protein gap and prevalent protein deficiency, by contrast, had the ring of objectivity: an explanation with scientific authority, backed by quantitative and experimental evidence, suited the cultures of science and government²²⁹ and fitted well into an optimistic, technocratic approach to public health led by nutritional scientists.²³⁰ Although no-one had designed it to be so, the profession of nutritional science was structured to favour single-nutrient explanations of problems, since individual scientists tended to be experts in a specific nutrient, and because such explanations were easier to explain to policy makers._²³¹ Kwashiorkor itself had the advantages of a visually striking presentation and of mostly affecting babies and children, the most sympathetic victims when building public support for aid. The structures of corporate and (inter)national power that impoverished half the world were manifold, and a diagnosis that made the problem of hunger solvable without the need to challenge them head-on could recruit the largest number of stakeholders in support. In retrospect this period was characterised by a huge waste of effort and money by the organisations trying to tackle world hunger, with systemic factors pulling the focus of many well-intentioned actors towards an explanation and solutions that were later revealed to be not just illusory but arguably counterproductive, since they distracted attention from the real causes of hunger and malnutrition.

At the same time, that it was again protein that had risen to the place of most important nutrient appears to be part of a longer-term historical pattern. Beliefs about protein, meat and race first developed in the 19th century had never really left the public imagination, even though scientific consensus had moved on — this made a scapegoat of the ignorance of poor black and brown mothers about the importance of protein in their children’s diets. The assumption that child growth was the single best indicator of a healthy diet was methodologically defensible (it is a uniquely convenient, object measure), yet it was bound to put special emphasis on the dietary role of protein. Add to this the beliefs about cow’s milk as the ‘perfect food’ that were current among nutritionists at the time and had been reinforced by successful nutritional policies during the Second World War, and the idea that the solution to world hunger was more protein seemed intuitively plausible. Institutions and infrastructures were then shaped by this understanding, creating a network of vested interests that sustained the same approaches long after the evidence had changed.

Figure 7: Recommendations for daily protein intake for 1yo infants by body weight by year; the shaded area shows the Second World War²³²

< Previous: Section 4                                                                                              Next: Section 6 >