Malaria and Rome: A History of Malaria in Ancient Italy

by Robert Sallares

  Malnutrition of the host adversely affects the parasite as well as the host. At the most extreme level, it has been suggested that severe malaria is rare in children suffering from the worst forms of protein-energy malnutrition, but the most recent research indicates that protein-energy malnutrition is indeed associated with increased morbidity and mortality from malaria.⁸⁴ A considerable body of research has found that deficiencies of several different vitamins reduce the reproduction rate of malarial parasites, leading to lower parasite counts in the blood of patients. For example, a ⁸³ Ricci et al. (1995); Soranus, Gynaecology 2.43–4; Lapi (1749: 75): il tenere i bambini a marcire nelle camere lontanissimi dal sentire l’alito dell’aria esterna; Levi (1945: 34, 76) on the association of rickets, general malnutrition, malaria and trachoma in Lucania; Davidson et al. (1979: 121–4).

  ⁸⁴ Shankar (2000) gives a detailed survey of all facets of the malaria-nutrition problem.

  142

  Demography of malaria

  shortage of vitamin E, an antioxidant, predisposes erythrocytes to oxidant-induced premature lysis, rupturing the red blood cells before the parasites have completed their development inside.

  Deficiency of vitamin C has been reported to have the same effect in experiments on monkeys. Other experiments have shown that a diet deficient in a nutrient essential for the malaria parasite, para-aminobenzoic acid, suppresses infections of rats with the malaria species Plasmodium berghei. Para-aminobenzoic acid is a precursor of the important coenzyme folic acid. It has been suggested that a diet of pure breast milk, which contains a very small proportion of this chemical, may increase resistance to malaria among very young human infants, although there is no experimental proof, and human malaria parasites do appear to have some capacity to produce it themselves. Iron deficiency is inimical to the parasites, although its effects on the chances of the host’s developing malaria are unclear; different studies have yielded conflicting results in both humans and animals. Nevertheless iron supplements to the diet do appear to increase the risk and severity of attacks of malaria.

  Studies in India have shown that a deficiency of riboflavin (vitamin B2) in vivo results in a reduced parasite reproduction rate. The best sources of riboflavin are milk, eggs, and liver. It is relatively scarce in cereals, especially if they are highly processed.⁸⁵

  These are all deficiencies in the diet of the human hosts. Natural selection for resistance to malaria can also favour human genetic mutations that produce an inherited reduction in the uptake of essential nutrients, even if the nutrients are abundant in the diet. Some recent research in biochemistry at the hospital in Grosseto has produced very interesting results. It has been shown that there is a high frequency of familial flavin-deficient erythrocytes in Grosseto and in Ferrara in the Po delta (which had endemic malaria in the late medieval and early modern periods), but not in Florence, a city with no history of P. falciparum malaria. This inherited genetic condition causes a reduced uptake of riboflavin by red blood cells even if it is abundant in the diet. It is a product of natural selection for resistance to malaria under conditions of extreme pressure and illustrates the importance of malaria in the past as an agent of natural selection.⁸⁶

  ⁸⁵ Das et al. (1988), with Davidson et al. (1979: 140–2); Gilles and Warrell (1993: 64); Har-El and Chevion (1997); Dobson (1997: 336–8).

  ⁸⁶ Anderson, et al. (1994), cf. Anderson et al. (1995).

  Demography of malaria

  143

  Nevertheless, even with the help of this trait conferring a degree of resistance to malaria, life expectancy at birth in Grosseto in the nineteenth century was still only twenty (see Ch. 5. 4 below). In other parts of Italy, especially in the Mezzogiorno, Sicily, and Sardinia, where there was also intense pressure from malaria, other genetic traits conferring degrees of resistance to it attained high frequencies, notably glucose-6-phosphate dehydrogenase (G6PD) deficiency and thalassaemia. Research in Sardinia by Siniscalco and colleagues first provided strong support for Haldane’s hypothesis, proposed in 1949, that heterozygotes for thalassaemia have increased resistance to P. falciparum malaria. Although that research was later criticized because of problems concerning the geographical origin of some of the populations of the villages which were studied, the new techniques of molecular biology have not only confirmed Haldane’s idea over the last few years, but have also shown that the statistical correlation between high frequencies of thalassaemia mutations and the presence (past or present) of endemic malaria is valid not just in the Mediterranean, but all over the world wherever P. falciparum malaria occurs. Grmek has comprehensively studied the evidence for the history of thalassaemia, G6PD deficiency, and favism (which is associated with G6PD

  deficiency) in antiquity.⁸⁷ G6PD deficiency reduces the activity of a gene that plays a critical role in a biochemical cycle that eliminates oxidizing agents from erythrocytes. It is thought, in general terms, that this results in the premature lysis of red blood cells before malaria parasites inside have completed their development, although the precise mechanism remains unclear. Thalassaemia (predominantly b-thalassaemia in Mediterranean populations) is caused by mutations in the globin genes which lead to an imbalance in the correct synthesis of the a-and b-globin chains of the haemoglobin molecule. Malarial parasites digest haemoglobin to obtain amino acids for protein synthesis.

  Thalassaemia and G6PD deficiency are very common in the modern populations of areas in the Mediterranean which were colonized by Greeks and Phoenicians in antiquity. They also occur ⁸⁷ Siniscalco et al. (1961); Weatherall (1997), Grmek (1983: 355–407), Ruwende and Hill (1998), Astolfi et al. (1999), and Vezzoso (1946) on thalassaemia; Greene and Danubio (1997) and Battin (1998) on G6PD deficiency. Lieber (1973) suggested that the Pythagorean communities in the cities of Magna Graecia arose from ‘a society of favism sufferers’. This hypothesis is not impossible, but cannot be proved or disproved given the lack of contemporary evidence for the origin of these communities.

  144

  Demography of malaria

  but are not so frequent in Latium and Tuscany. The problem arises of whether thalassaemia was introduced into Italy by Greek colonists in the eighth century  or alternatively had been present in human populations in Italy since the Neolithic period.⁸⁸ These two possibilities are not necessarily mutually exclusive, since it is now known that numerous independent mutations for thalassaemia (and G6PD deficiency) in response to malaria have occurred in many different parts of the world.⁸⁹ Nevertheless, since high frequencies of thalassaemia in mainland Italy are mainly concentrated in regions that were colonized by Greeks, it is likely that Greek colonization did indeed play a critical role in the spread of thalassaemia in Italy. In view of Ampolo’s convincing arguments about the free movement of individuals (e.g. Demaratus of Corinth) and groups from state to state in archaic central Italy, there is little doubt that there were plenty of opportunities for gene flow between populations in the archaic period.⁹⁰ Nevertheless the fact that different genotypes conferring degrees of resistance to malaria attained high frequencies in different regions demonstrates that Latium and Tuscany had a different population history in antiquity from the other areas in question, never having been colonized by Greeks or Phoenicians. Presumably the original Greek colonists of Magna Graecia brought common mutations for thalassaemia and G6PD deficiency with them from Greece. It has been suggested that the two commonest mutations for b-thalassaemia in Mediterranean populations were both spread by colonization.

  According to this hypothesis the b+IVS nt 110 mutation occurred in Greece and was then spread to southern Italy by Greek colonists, while the b°39 mutation originated in the Levant and was carried westwards by Phoenician colonists to North Africa, Sardinia, and the Iberian peninsula. McCormick provides an interesting discussion of alternative hypotheses. The suggestion that thalassaemia was brought to the Mediterranean by the Mongols or Huns is in fact quite impossible because the thalassaemia mutations of East Asian populati
ons are quite different from those of Mediterranean populations (this was not known at the time the idea was proposed because DNA sequencing was not yet possible at that time). How-

  ⁸⁸ Ascenzi and Balistreri (1977) discussed this question inconclusively.

  ⁸⁹ Weatherall (1997). About 200 different thalassaemia and 120 different G6PD mutations are now known.

  ⁹⁰ Ampolo (1976).

  Demography of malaria

  145

  ever, the idea that thalassaemia might have been brought to Italy during the Byzantine period, as part of the internal diaspora of the Byzantine Empire which is the subject of the book containing McCormick’s article, is more plausible. Nevertheless it remains the case that there were substantial population movements in Europe before the Byzantine period, and malaria was active in both Greece and Italy before the Byzantine period. Consequently it is virtually certain that the human genetic response to malaria commenced before the Byzantine period.⁹¹

  A skeleton with a probable diagnosis of heterozygosity for b-thalassaemia was excavated at the Roman villa of Settefinestre.

  This skeleton might have belonged to an imported slave and is not a direct proof of the presence of malaria there, since thalassaemia is an inherited condition. Nevertheless it is a sign that the slave labour force of the Roman villas in western central Italy was in contact with malaria during the time of the Roman Empire. Marsh vegetation first appears in the palaeobotanical record at Settefinestre in the third century . However this alone cannot be used to date the spread of malaria at Settefinestre because Anopheles mosquitoes do not necessarily require large marshes for breeding, as has already been seen (Ch. 4. 2 above).⁹² The commonest DNA mutation for G6PD deficiency in Mediterranean populations has been identified in the human skeletal remains from the infant cemetery at Lugnano in Teverina in Umbria, dating to the fifth century , using ancient DNA (see Ch. 4. 2 above).⁹³ Undoubtedly over the next few years the application of the techniques of molecular biology to human skeletal remains excavated on archaeological sites will greatly increase our knowledge of the history of these human genetic adaptations to malaria in antiquity.

  Both dietary and genetically determined nutritional deficiencies occurred in the past in western central Italy. Both interacted with malaria. The aforementioned work on dietary deficiencies and ⁹¹ Cao et al. (1989); McCormick (1998: 25–31). We can invisage three phases in the spread of thalassaemia and G6PD mutations in Italy: 1. archaic Greek colonization (particularly important in southern Italy, less important in northeastern Italy); 2. further immigration from the eastern Mediterranean in the Byzantine period (probably particularly important in northeastern Italy, around Ravenna); 3. the indigenous spread of antimalarial mutations owing to the pressure of natural selection by P. falciparum malaria in situ (in southern Italy since the fifth century  at least, in northeastern Italy since the medieval period).

  ⁹² Mallegni and Fornaciari (1985) on the skeleton number 26.203 from Settefinestre, a young woman nearly seventeen years old, diagnosed as suffering from thalassaemia; Celuzza (1993: 25–6, 230).

  ⁹³ Sallares, et al. (2002).

  146

  Demography of malaria

  malaria was principally directed at the effects of malnutrition on the malaria parasite, without considering its effects on the human host. Thus there was a tendency in some medical literature to argue that malnutrition in the host reduced the severity of malaria infections. It may well be true that the severity of clinical symptoms is reduced if the host is significantly malnourished, because the reproduction rate of the parasites is reduced. However, it is undeniable that malnutrition is bad for the host, and any malaria infection is also bad for the host.⁹⁴ Recent re-evaluations of this problem suggest that the combination of malnutrition and malaria does increase human morbidity and mortality, even if the parasites suffer as well from malnutrition. Zurbrigg argued that acute hunger (as indicated by elevated grain prices) was statistically significantly associated with recurrent severe malaria epidemics in the Punjab between 1868 and 1940. The correlation was strongest in areas of the Punjab where crop failure was mainly caused by drought, thus bearing the closest resemblance to the conditions of semi-arid Mediterranean climate regions. She also argued that a link between severe malaria and starvation was frequently observed in the Punjab. It is likely, even if there is no specific evidence for it, that a similar link occurred in the famines that certainly occurred sometimes in Italy in antiquity, for example the great famine during the Gothic Wars in the sixth century  described by Procopius, which forced people to eat bread made from acorns and to resort to cannibalism.⁹⁵

  Tognotti studied the diet of people who lived in areas of holoendemic malaria on Sardinia. The diet of such people was frequently short of meat, fish and dairy products, which often ended up in the towns even where animal husbandry was important. The fundamental component of the diet was bread, accompanied by prickly pears (not available in antiquity), other fruit, legumes, and mushrooms. Researchers on Sardinia felt that malnutrition was positively, not negatively, correlated with malaria.⁹⁶ Similarly researchers who worked in Rome and the Roman Campagna

  generally reckoned that poor malnutrition was associated with ⁹⁴ I. A. McGregor in Wernsdorfer and McGregor (1988: i. 753–67).

  ⁹⁵ Zurbrigg (1994) and (1997); Procopius, BG 2.20.15–33.

  ⁹⁶ Tognotti (1996: 106) quoted a doctor who said that: ad Orosei…la frequenza dei casi di malaria era in ‘ragione diretta della scarsa alimentazione’. Levi (1945: 19) described the diet of the inhabitants of a region with endemic malaria in Lucania.

  Demography of malaria

  147

  malaria. For example, Celli thought that peasants in the Roman Campagna in the nineteenth century had a very poor diet, eating a lot of maize (not available in antiquity), but not much else, although he thought that shepherds had a somewhat better diet, including both wheat and milk.⁹⁷ Other researchers and travellers reached similar conclusions. Gregorovius wrote as follows:

  If you live among them, you will see, too often, hunger-stricken human beings coming out of this paradise to meet you . . . [the peasant] would starve if it were not for the meal of the Indian corn [sc. maize], which is his sole nutriment.⁹⁸

  North carried out a short experiment which suggested to him that ‘a few ounces of well-cooked red meat, and a liberal allowance of good red wine, will have an effect, equal, if not superior, to a large dose of quinine’.⁹⁹ North’s experiment was not a double blind test, as is generally expected in modern medical research, and his conclusion undoubtedly understates the seriousness as a disease of malaria in non-immune individuals. In fact, the belief in the value of red meat and red wine against malaria, a disease of red blood (cells), was ultimately yet another instance of sympathetic magic.

  Other similar experiments yielded completely different results.

  For example, in 1897 Cirio, who had made his fortune in the fruit-canning industry, brought a colony of non-immune farming families from the Veneto in the north of Italy to the Pontine territory and provided them with very large rations of meat, wine, and other commodities to test this theory.¹⁰⁰ Although the results of the experiment were disastrous, they do not necessarily exclude the possibility that malnutrition might exacerbate the severity of malarial infections, perhaps particularly in individuals with some degree of immunity. The increasing use of quinine, especially after the Italian government made it available to all free of charge by ⁹⁷ Celli (1900: 170–2); Mallegni and Fornaciari (1985) found that the skeletons from Settefinestre had a high zinc content and a low strontium content, indicative of a diet rich in meat and short of vegetables. They rightly interpreted these results in terms of a pastoral economy in the Maremma in late antiquity. Ciuffoletti and Guerrini (1989: 97–100) described the traditional diet of the inhabitants of the Maremma. Similarly Arlacchi (1983: 176–9) commented on the very poor nutritional state in the nineteenth and early
twentieth centuries of the inhabitants of the Crotonese, a population severely affected by malaria. He noted the consequence that the average height of these people was 10 cm less than the regional average, cf. Douglas (1955: 130).

  ⁹⁸ Gregorovius (1902: 82).

  ⁹⁹ North (1896: 161).

  ¹⁰⁰ Snowden (1999: 33–4); Hackett (1937: 28).

  148

  Demography of malaria

  a series of laws passed in the years 1900–4, a few years after North’s research, was probably the most important single factor in the reduction of mortality from malaria in Italy.¹⁰¹ Nevertheless North’s primitive experiment and the general experience of observers at the time indicate how important malnutrition may have been in the past. De Felice, after considering comments on the matter in the writings of the early modern Roman agronomists, concluded that animals were better fed, on the lush pastures of Latium, than most of the local inhabitants!¹⁰²

  Of course the early modern period had its own peculiar problems, which were not necessarily shared by antiquity. Gregorovius attributed the poor diet of peasants in Latium to the economic effects of high taxation. Reliance on maize, imported from the western hemisphere after Columbus, introduced the possibility of pellagra. This was not a problem in antiquity.¹⁰³ Nevertheless the general impression given, namely that malnutrition did significantly aggravate the effects of malaria in the Roman Campagna in the early modern period, is very important. The balance of probability is that the sort of malnutrition now being revealed by the Italian studies of the Vallerano skeletal population in the second century  increased both mortality and morbidity from malaria in western central Italy in Roman times. One vital but imponderable question about antiquity is the question of how well fed were slaves.