Main results

Many of the findings of the Canqué study were either new, poorly understood, controversial, or unbelievable when communicated to the scientific community. Some found it difficult to accept that calories were more limiting than protein in the village diet, that 40% of village newborns had low birth weight (LBW), that the high concentration of S-IgA in human colostrum and milk protected against enteric infection, that children averaged seven or more attacks of diarrhea per year, and that children had enteric pathogens during more than one-half of their first three years of life. The most difficult postulate to understand was that the force of infection and infectious disease was the main determinant of poor health, growth retardation, malnutrition, and death of infants and young children. Such ideas conflicted with prevailing dogma, despite previous studies that pioneered the field (Scrimshaw et al. 1959, 1968). The main findings of the Cauqué, study follow.

Maternal Health and Nutrition

Courtship began at 14-17 years for girls and at 17-20 years for boys. Stable couples by common-law union or marriage were the rule; the few unaccompanied pregnant women were engaged. Motherhood started two or three years after union of the couple. Pregnancies usually occurred before age 20 and were spaced by two-year intervals as a result of prolonged breast-feeding and sexual restraint. The number of pregnancies averaged 12 for women 40-43 years old; the total span of fertility was 25 years. Birth intervals strongly correlated with gestational age: short birth intervals (917 months) led to more premature births than longer birth intervals (18 months or more), a highly significant difference. The association probably reflected a more limiting maternal nutrition and health when there was shorter spacing between pregnancies. It may also reflect the social practices in this group. Child-rearing expertise was acquired from older women.

Most young women were slim and stunted, although weight for height was adequate for most when they approached reproductive age; some became stocky at the end of their obstetric experience (Table 1). Stunting was the result of early synergism between infectious diseases and nutritional deprivation. Almost all pregnant women were barefoot, did not observe adequate personal hygiene, and lived in an unsatisfactory sanitary environment. Therefore, a high proportion harbored enteric pathogenic parasites, bacteria, and viruses (Table 2).

TABLE 1 Anthropometric Values of 171 Women^a of Santa María Cauqué, 1963-1971.

Age (years)	Number of women	Weight (kg)	Height (cm)	Tricipital Skinfold (mm)
14	1	58.4	142.2	14.0
15-19	80	52.2 (0.7)b	142.5 (0.5)	10.7 (0.4)
20-24	117	52.7 (0.6)	143.5 (0.4)	9.5 (0.3)
25-29	66	52.8 (0.7)	144.3 (0.6)	8.5 (0.4)
30-34	75	52.8 (0.8)	142.8 (0.5)	9.4 (0.4)
35-39	50	53.1 (0.9)	141.9 (0.6)	10.0 (0.5)
40-44	21	56.8 (1.6)	143.5 (0.9)	9.6 (0.8)
Mean		52.9	143.1	9.6

^a Contributing 410 pregnancies to the study
^b Mean (standard error)
Source: Mata (1978a)

The diet consisted of maize-the staple food-prepared as tortillas (flat pancakes of ground lime-treated kernels), black beans, and greens, with little animal protein and an occasional fruit. It was deficient in calories, animal protein, iron, and some vitamins. Food intake during pregnancy did not meet the increased need. Women did not gain enough weight during gestation (mean 6.8 kg) and experienced several disease episodes during pregnancy. Caloric intake, but not protein intake, during pregnancy correlated with birth weight and length. Thus, mothers had been affected by infectious diseases and poor diets and were stunted. Stunting correlated with fetal growth retardation and premature birth.

Birth at Home

Data were obtained by nurses for 297 natural births occurring from February 1964 through 1973; additional data were furnished by relatives for another 111 deliveries (Mate, 1978a). The midwife, relatives, and friends were present at the deliveries. Women gave birth on their knees (90%), squatting (8%), or in the supine position (2%). Traditional obstetrics did not include shaving, disinfection, analgesics, episiotomy, or forceps. Enemas were not given, and in all but two deliveries, passage of maternal feces occurred with soiling of the newborn. External maneuvers were the rule, with an occasional vaginal or rectal exploration, without handwashing or gloves. The main role of the midwife was to support the delivering woman with words, prayers, and massage of the back, abdomen, and limbs. The nurse assumed obstetric roles on few occasions, by specific invitation of the midwife or a relative.

TABLE 2 Pathogenic Viruses, Bacteria, and Parasites and Antibodies to Selected Pathogens, Santa María Canqué Mothers of Cohort Children, 1964-1965

	Number	Infectious Agents Diagnosed	Prevalence (%)
Enteric Infections	32	Enteroviruses	25
		Adenoviruses	3
	116	Shigella spp.	9
		Salmonella spp.	5
	24	Entamoeba histolytica	5 4
		Dientamoeba fragilis	8
		Giardia intestinalis	8
		Ascaris lumbricoides	83
		Trichuris trichiura	58
Significant Antibodies	50	Poliovirus type 1	100
	93	Coxsackievirus type B 1	55
	93	Echovirus type 3	27
	93	Reovirus type 1	14
	48	Rubella agent	50
	200	Brucella spp.	0.5
	200	Treponema pallidum	0
	22	Toxoplasma gondii	50

Source: Mata (1978a)

Rupture of membranes was spontaneous in all deliveries, and birth ensued within 10 to 20 minutes (85.5%), 20 to 60 minutes (4.4%), one to 2 hours (3.4%), or 3 to 98 hours (6.6%). The duration of labor ranged from a few minutes to 19 hours (mode, four hours). The midwife cut and cauterized the umbilical cord with a hot sickle or machete, and in 72% of the cases dressed the stump with gauze and treated it with alcohol. These practices likely accounted for the virtual absence of tetanus neonatorum. Oxytocin, antibiotics, and other drugs were unknown or not available. Over 10 years, births were uncomplicated, 99.3% cephalic and 0.7% with breech presentation. No maternal deaths in puerperium were recorded. Puerperal fever and other diseases were not observed. One mother had cellulitis in the elbow and one developed psychosis-both while nursing —and they died within six months postpartum. Their infants were nursed by foster mothers and survived.

The success of natural birth was remarkable when compared with hospital delivery before the sanitary revolution initiated by Semmelweis in Europe in the nineteenth century or the advent of asepsis, antisepsis, and modern obstetrics (Table 3). There were no cesarean sections or nosocomial infections and no recorded puerperal fever or similar infections, despite the lack of asepsis in childbirth. In contrast, abnormal delivery and puerperal infections are common in modern societies. In Costa Rica, more than 95% of the births are in clinics and hospitals, with 36% spontaneous deliveries, 15% cesarean sections, and 15% maternal infections (Mate, 1982a). Cesarean section was found to be associated with shorter periods of breast-feeding (Mate et al., 1988) and an increased risk of neonatal death (Bobadilla et al., 1991). More complications and deaths would occur with such sections if it were not for the wide use of broad-spectrum antibiotics and other medical resources.

Mother-Infant Interaction

The newborn baby was immediately placed naked on a mat or cloth on the dirt floor, without removing the vernix caseosa. After a few minutes, the baby was wrapped in clothes. Maternal colostrum was given to only one-third of the babies. If the mother had been nursing an older child throughout the present pregnancy, she offered the breast to the newborn without delay. In many instances the mother gives colostrum to her baby. Under other circumstances a foster mother, generally a friend or relative, the grandmother of the infant as recorded in a few instances, serves in this capacity. In the first two days postpartum, the mother and baby had a 30 to 45-minute bath in the temascal (traditional sauna) accompanied by the midwife and perhaps a friend. She resumed her domestic, agricultural, and market duties within three days postpartum, carrying the baby wrapped on her front or back in intimate contact with her body. She slept with the baby to keep him warm and nursed him on demand (day and night) for months or years.

Optimal mother-infant contact promoted exchange of pheromones and odor, eye-to-eye contact, and verbal communication (Klaus and Kennell, 1976). It also stimulated breast-feeding, which is crucial for optimal intake of nutrients, maternal hormones, and immune factors. The bifidus factor of human milk, along with S-IgA and lactoferrin, favored colonization of the child's intestine with bifidobacteria. The peaceful environment of home and village was supportive of such mother-infant interaction. Bonding contributed to the universal success of efficient breast-feeding for months and years, the absence of child abandonment, and the very low rate of corporal punishment of children in the village. The quietness of the home was occasionally altered by a quarrel, perhaps preceded by a fiesta accompanied by guaro (local hard liquor).

TABLE 3 Delivering Babies in Contrasting Ecosystems Vienna (1850), USA (1926), Santa María Cauqué, (1964-72), and Puriscal (1979-81)

	Place and Date
Variable	Vienna 1858 (%)	USA 1979 (%)	Canqué 1964-72 (%)	Puriscal 1979-81 (%)
Type of Delivery
Spontaneous	90	-	100	36
Conduced	0	-	0	45
Cesarean	0	15	0	14
Othera	10	-	0	5
Puerperal Complications
Maternal infection	High	7	0	8
Maternal mortality	12b	0.04c	0	0.04
Source	Haggard, 1929	Cohen and Estner, 1983	Mata, 1982b	Mata, 1983c

^a Induced, forceps, manual maneuvers, combinations
^b Reduced to 1.2% by hand washing of attendants
^c 40% by puerperal fever

Breast-Feeding and Weaning

Breast-feeding was considered a natural event. Approximately one-half the mothers continued to nurse while pregnant. Among women not lactating during gestation, one-third offered colostrum to their infants. Although 33% of the families owned cows (the 1971 census showed 45 cattle), there was no bottle-feeding in the village. Few families had refrigerators, and milk formula was not available. An absence of urban life, low levels of modernity, and no organized labor for women outside the home were contributing elements. Soiling of the babies with maternal feces during birth resulted in infections with pathogenic viruses and bacteria during the first days of life, which generally were asymptomatic. About one-half the mothers gave sweetened water and infusions to their newborns with the aid of a piece of cloth, but this practice did not result in diarrhea. Colostrum and milk had large concentrations of S-IgA, which, together with the bifidus flora and lactoferrin, provided an effective defense against infection (Wyatt et al., 1972; Yolken et al., 1978). Antibodies in S-IgA are resistant to digestive enzymes and remain active after transit through the alimentary canal (Hanson et al., 1975).

Most women produced sufficient milk (an average of 686 g/day at three months) to nurse satisfactorily for three to six months, but quality and quantity were lower when compared to values for well-nourished Swedish women (776 ml/day) (WHO, 1985). This volume, however, may not be enough to satisfy the needs in the second trimester (Waterlow, 1981). Women experienced infectious diseases and low food intake since childhood, but most had adequate weight for height, and all lactated for two to four years. The intimate contact between the bodies of the mother and child, as it is carried constantly, the heat of the temascal baths, and frequent exposure to sunshine were factors counteracting the hypothermia of small newborns in this cool climate.

The weaning foods, normally introduced at three to five months of age, were tortillas, broths, beans, and other foods of the adult diet. They were bulky, of low caloric density, and a source of microbes (Mate et al., 1976a). For instance, tortillas, relatively sterile soon after cooking, became contaminated with Escherichia coli, Bacillus cereus and Clostridium perfringens within 24 hours of storage under a moist cloth, a traditional practice (Capparelli and Mata, 1975). Lack of refrigeration and limited firewood were contributing factors.

Recurrent infections during and after weaning contributed to the marked reduction in calorie consumption (Table 4). The main nutritional deficit was in calories and not in protein, as was thought in the past (Gopalan et al., 1974; Valverde et al., 1975). Dietary deficiencies were assumed to be due to low food availability in the community, until it was shown that Cauqué, children had marked caloric deficits ("dips") during attacks of infectious disease, and adequate caloric intake during good health (Mate et al., 1977; Mata, 1978b, 1979).

No change in onset of definitive weaning was noticed in five yearly cohorts, reflecting the stability of this deeply rooted tradition (Figure 2) (Mate et al., 1982). Weaning age was conditioned by the interplay of maternal factors: a new pregnancy (main cause of abrupt interruption of breast-feeding), prepregnancy nutrition, and working, economic, and emotional conditions. Such child factors as size and maturity at birth, growth velocity, and health status also have an influence. Infants who grew faster in the first months of life were weaned earlier; those who grew more slowly were weaned at a later date (Mate, 1 978a) (Figure 3).

TABLE 4 Nutrient Value of Food Supplements Consumed by Fully Weaned Children One to Three Years Old.

	26 months (14 children)	36 months (31 children)	International Recommendation, 1-to 3-year-oldsa
Energy (kcal)	619 (176)b	992 (272)	1300
Protein (g)	17 (4.9)	27 (6.6)	16
Iron (mg)	6(1.8)	10 (3.6)	10
Retinol (mg)	90 (95)	111 (77)	250
Thiamin (mg)	0.4 (0.2)	0.6 (0.1)	0.6
Riboflavin (mg)	0.3 (0.1)	0.4 (0.2)	0.9
Niacin (mg)	3.4 (1.1)	5.8 (1.5)	10.2
Ascorbic acid (mg)	16 (8.0)	24 (10)	20

^a NAS/NRC (1989)
^b Mean (SD)
Source: Mata (1978a), adapted

FIGURE 2 Percent distribution of weaning age, in months, of five cohorts by year of birth, Santa María Canqué, 1964-1972. Figures above bars indicate the number of girls and boys weaned at given ages. Most children were weaned late, from 21 to 32 months of age. The slight increase in length of breast-feeding for the 1966 and 1968 cohorts may not be significant in view of the relatively small number of children in each cohort (Mate et al., 1982).

Force of Infection and Infectious Disease

Infectious diseases in the 1940s were very common and devastating, according to the Registry kept in the village Town Hall since 1936. Simultaneous epidemics of dysentery, respiratory disease, whooping cough, and measles ravaged the village, killed as many as 7% of the total village population in a given year (Mate, 1978a). The Canqué study revealed a formidable pressure of infection on pregnant women and children. Most women of reproductive age harbored intestinal pathogens, which infected newborns at delivery (see Table 2). In pregnancy, 31% oft he women had one episode of infectious disease, 22% had two, 16% had three, and 2% had four or five (Table 5) (Mate, 1978a). Diarrheal disease, lower respiratory infection, cystitis, pyelonephritis, and febrile bacterial infections often had deleterious effects on maternal health. Infection in the mother likely contributed to the frequency of elevated fetal IgM (Mate, 1970a; Mata and Villatoro, 1977), but no correlation was found between this and fetal growth retardation.

FIGURE 3 Weight curves of two cohorts of children defined by different weaning ages, from birth to age 42 weeks, the stage of growth deceleration. Top curve: children weaned earlier, at 12-29 months. Lower curve: children weaned later, at 30-47 months. Children who grew faster in the first months of life tended to be weaned later. A similar relation occurred with body length, Santa María Canqué, 1964-1972 (Mate, 1978a).

Intestinal colonization of newborns with indigenous microorganisms was rapid and efficient and originated in perinatal contamination with microflora of maternal feces and tegumenta (Mate and Urrutia, 1971). The bifidus flora, S-IgA, and lactoferrin provide a natural barrier against many intestinal pathogens that invade the neonate (Table 6). Enteroviruses and pathogenic bacteria were found in the intestinal tract of several neonates, but such infections often were asymptomatic, reflecting the effective defense conferred by immune principles in maternal colostrum and milk (Mate and Wyatt, 1971; Cruz et al., 1977).

Breast-fed infants were markedly resistant to Giardia and Shigella during the first semester or longer (Figure 4). As they grew older, however, they became infected with pathogenic enteric viruses (Table 7), shigellae (Table 8), and intestinal parasites (Table 9). The indigenous flora were less protective at older ages when children were exposed to larger infectious doses acquired from weaning foods, water, and attendants (Mate, 1983a).

TABLE 5 Incidence of Infectious Diseases and Syndromes 365 Pregnancies, by Trimester

	Trimester
Illness or Syndrome	1st	2nd	3rd
Upper respiratory disease	18.9a	26.3	30.9
Bronchitis	1.9	2.5	3.0
Laryngotracheobronchitis	0.3	1.6	1.6
Otitis media	0	0.8	0.8
Pneumonia	0	0.6	0.3
Conjunctivitis	0.3	0	0.3
Fever of unknown origin	0.3	0	0
Diarrhea	4.9	4.9	8.2
Dysentery	1.4	1.1	1.1
Cystitis, pyelonephritis	1.9	1.9	1.4
Hepatitis	0	0.5	1.1
Skin infectionb	3.0	0.8	3.0
Tenosynovitis	0	0.3	0

^a Percentage of pregnancies
^b Impetigo, cellulitis, abscess

Source: Mata (1978a), adapted

TABLE 6 Early Neonatal Infection with Pathogenic Viruses, Bacteria, and Parasites

Age of Child	Number of Children Tested (% positive)	Pathogenic Agent (Number of Children)
1 day	79 (1.3)	Echo 7 (1)
2 day	54 (7.4)	Polio 1 (1) Echo 6 (3)
3 days	61 (8.2)	Polio 1 + echo 6 (1) Echo 6 (1) Echo 7 (1) Echo 9 (1) Echo 11 (1)
3 weeks	109 (3.6)	Sh.f!exnerit type 1
4 weeks		Sh. flexneri type 3
12 days		Sh. flexneri type 2
28 days		Sh. sonnei
1 day	191 (2.6)	E. histolytica ca (1)
3 days		G. intestinalis c t (1)
4 days		G. intestinalis t (1)
4 days		G. intestinalis c (2)

^a c = cysts; t = trophozoites
Source: Mata ( I 978a); Mata et al. (1969)

FIGURE 4 Incidence and prevalence of Giardia intestinalis and Shigella spp., by age of child, cohort children from Santa María Cauqué, examined weekly from birth to age three years. Marked differences m incidence and prevalence were noted, reflecting the chronicity of infection with both pathogens (Mate et al., 1971; Mata, 1982a,b). Note the significantly lower rates of infection during the period of exclusive breast-feeding. With weaning, children became more readily infected, and if they were malnourished, they often had persistent infection.

(Giardia)

(Shigella)

Infection with rotavirus was investigated by an antigen-capture enzyme-linked immunosorbent assay (ELISA) in 5,891 weekly specimens from the 45 cohort children, which had been kept frozen since 1964-1969 (Mate et al., 1983). Rotaviruses were uncommon during exclusive breast-feeding. When they entered the village, however, they rapidly infected as many as 50% of the children, particularly those 6 to 18 months old (Figure 5). Rotavirus diarrhea often was accompanied by fever, dehydration, and weight loss (Wyatt et al., 1979; Mata 1983a).

TABLE 7 Prevalence of Infection with Pathogenic Enteric Viruses 45 Cohort Children of Santa María Cauqué, Birth to Age Three

Age (months)	Number of specimens	Enteroviruses	Adenoviruses
0-5	1,116	230 (20.6)a	34 (3.1)
6-11	1,162	483 (41.6)	46 (3.9)
12-17	917	481 (52.5)	33 (3.6)
18-23	953	438 (45.9)	60 (6.3)
24-29	908	446 (49.1)	58 (6.4)
30-35	867	530 (61.1)	48 (5.5)

^a Number of fecal specimens (percentage positive)
Source: Mata (1978a)

TABLE 8 Incidence of Shigella Excreted in Feces 45 Cohort Children, Birth to Age Three Years

Age (weeks)	Number of children	Weeks at riska	Prevalence
0-25	81	1,783	1 (0.06)b
26-51	65	1,546	11 (0.7)
52-77	52	1,192	20 (1.7)
78-103	46	1,096	37 (3.4)
104-129	44	1,100	39 (3.6)
130-155	43	1,075	40 (3.7)
Total		7,792	148 (1.9)

^a Weeks e' which children were examined clinically and bacteriologically b Number of new infections (rate per 100 child-weeks)
Source Mata (1978a)

Enteroviral and rotaviral infections were short-lived. In contrast, many adenoviral, bacterial, and parasitic infections were prolonged, either due to failure of the host defense to repel them or due to repeated infections. One striking example were the shigellae, with more than 50% of the infections lasting two weeks or more, and 10% lasting two months or more (Table 10) (Mate et al., 1984). The use of the immunomagnetic separation/polymerase chain reaction might increase these rates by two to threefold, as shown for Costa Rican children (Achí et al., 1995). Prolonged infections with Giardia also had a deleterious effect on nutrition and growth of Cauqué, children (Farthing et al., 1986).

The 45 cohort children harbored one or more intestinal pathogens in one-half of their infant period. These were found in two-thirds during the second and third years of life. Consequently, most children had recurrent diarrhea during most of their early years of life. With regard to etiology, more than 50% of the diarrheas were associated with known pathogens (Table 11), an underestimate because in that epoch many agents had not been discovered or rediscovered. For instance, by retrospective testing, rotaviruses were found associated with 10% of the diarrheas of the 45 cohort children (Mate et al., 1983) (Table 12). Also, in the Mayan village of Santa María de Jesús, located in the same geographic region and similar in many respects to Cauqué, (see Figure 1), adenovirus types 40 and 41 were incriminated in the acute diarrheas of childhood (Cruz et al., 1990).

Diarrhea morbidity, which is low during exclusive breast-feeding, increased with weaning to attain the highest rates from the second to the fifth semesters (Mate, 1975). There were, on the average, seven attacks of diarrhea per child per year in the first three years of life. The incidence was greater during weaning ("weanling diarrhea"; Gordon et al., 1963), from 6 to 26 months of age (Figure 6). Other infections were less frequent than diarrhea, but added together they were very important as causes of disease, disability, and death (Table 13): namely, tonsillo-pharyngitis; bronchitis and bronchopneumonia; thrush, herpes simplex, glossitis, and cheilitis; impetigo, abscess, and furuncle; measles, rubella, and varicella; and undifferentiated febrile illness. These illnesses were found in greater frequency during weaning ("weanling morbidity"). Complications of measles (Urrutia and Mata, 1974) and whooping cough (Mate 1978a) were common.

In sum, children were ill with infectious diseases during much of their infancy and preschool years. Such a force of infection contributed to an early increase in serum immunoglobulins. By the age of one year, 95% of children had acquired the village adult level of IgG, and 99% had reached the adult value of IgM. With regard to IgA, 73% of the children had adult levels of this immunoglobulin by the age of five to nine years (Cáceres and Mata, 1974).

TABLE 9 Incidence of Pathogenic Parasites 45 Cohort Infants, Birth to Age Three Years, by Six-Month Intervals

	Months
Parasite	0-5	6-11	12-17	18-23	24-29	30-35
E. histolytica	5a	0	6	9	12	5
		(5)b	(11)	(20)	(32)	(37)
D. fragilis	0	3	1	1	1	5
		(3)	(4)	(5)	(6)	(11)
G. intestinalis	8	13	14	4	1	2
		(21)	(35)	(39)	(40)	(42)
A. Iumbricoides	4	9	14	14	1	0
		(13)	(27)	(41)	(42)	(42)
T trichiura	2	2	4	6	3	3
		(4)	(8)	(14)	(17)	(20)
H. nana	0	0	0	1	3	2
					(4)	(6)

^a Number of children with parasite
^b (Accumulated number of children with parasite)
Source: Mata (1978a); Melvin and Mata (1979)

FIGURE 5 Incidence of rotavirus infection in 45 cohort children sampled weekly from birth to age three years. Each dot represents a week positive for rotavirus antigen by ELISA. The vertical shaded bands represent the periods in which rotaviruses spread across all age groups. Such events immunized most of the susceptible population, until new individuals accumulated to face new entrances of these viruses (Mate et al., 1983).

TABLE 10 Duration of Shigella Infection (Cases and Carriers) 45 Cohort Infants, Birth to Age Three Years

		Weeks of Duration
Shigella Subgroup	No. of Episodesa	1	2-4	5-8	9-12	13-16	17-38
dysenteriae	29	7 (24)b	10 (34)	4 (14)	3 (10)	2 (7)	3 (10)
flexneri	5	18 (24)	21 (28)	18 (24)	9 (12)	5 (7)	4 (5)
boydii	21	14 (67)	6 (28)	1 (5)
sonnet	7	7(100)
Total	132	46 (35)	37 (28)	23 (14)	12 (9)	7 (5)	7 (5)

^a Two isolations were considered independent if separated by more than 2 weeks.
^b Number of events (percentage within Shigella subgroup)
Source Mata (1978a)

TABLE 11 Enteric Pathogens in 381 Cases of Diarrhea Experienced by 22 Cohort Children from Birth to Age Three Years.

Agenta	Number Positive (%)
Giardia intestinalis	92 (24.1)
Shigella spp.	89 (23 3)
Rotavirusesb	44 (11.5)
Entamoeba histolytica	44 (11.5)
Adenoviruses (cultivatable)	29 (7.6)
Salmonella spp.	11 (2.9)
Enteropathogenic Escherichia coli	6 (1.6)
Dientamoeba fragilis	3 (0.8)
One or more of the above	227 (59.6)

^a Strains of Escherichia cold [enterotoxigenic (ETEC), enteroinvasive (EIEC), enterohemorrhagic (EHEC), and enreroaggregative (EAggEC), Campylobacter, Cryptosporidium, diarrhea adenoviruses, and small round structured viruses were not investigated at the time of the Cauqué, Study
^b Rotaviruses were studied in stored frozen specimens (Mate et al., 1983).
Source: Mata et al. (1984)

TABLE 12 Incidence of Rotaviruses in Diarrhea of Children Living in Contrasting Ecosystems

Population	Child-Years	All Diarrhea	Rotavirus Diarrhea	Rotavirus Infection	% Diarrhea Due to Rotavirus
Canquéa 1964-1969	132.5	1,050 (7.9)	109 (0.8)	166 (1.2)	10.4
Matlaba 1978-1979	120	727 (6.1)	34 (0.3)	-	4.7
Winnipegb 1976-1979c	139	165 (1.2)	40 (0.3)	50 (0.4)	80.0

^a Rural poor, adjusted to account for uncollected or unavailable specimens (Mate et al., l 983)
^b Recalculated from Black et al. (1982a)
^c Adapted from Gurwith et al. (1981)

Effect of Infection on Food Intake and Growth

Frequent symptoms and signs in children were despondency, prostration, fever, anxiety, and anorexia. These are common manifestations of stress triggered by interleukins and tumor necrosis factor released by macrophages in response to infection (Beisel 1977; Dinarello, 1984). The main nutritional consequences of infection are reduced consumption of calories; loss of ingested foods; increased transit of food through the alimentary canal; altered digestion and absorption; protein-losing enteropathy; loss of electrolytes, vitamins, and other nutrients; altered metabolism; sequestration of trace elements; and nutrient diversion (Beisel, 1977).

Canqué children ate less food during episodes of infectious diseases, regardless of etiology, severity, or target organ. The effect was more pronounced with fever, sepsis, diarrhea, and lower respiratory infection. Thirty-three percent of children with whooping cough consumed only one-half of the customary amount of tortilla in the first month of illness; maize intake was also significantly depressed in the second and third months of the disease (Table 14).

FIGURE 6 Incidence of infectious diseases, per 100 person-months, in a cohort of 45 children observed from birth to age three years, Santa María Cauqué, 1964-1972. Diarrheal diseases predominated, with rates as high as 80/100, about eight cases per child per year. Upper and lower respiratory infections, combined, were almost as frequent as diarrhea. Note that infectious morbidity tended to accumulate in the weaning months (Mata, 1978a).

Fully weaned children consumed an adequate amount of food when they were free of disease. However, intake fell below recommendations when they acquired enteric, skin, and lower respiratory infections (Figure 7) (Mate et al., 1977; Mata, 1979). The average decrease attributed to diarrhea alone was 24% for total protein and 21% for calories (Mate, 1983b). In Uganda, the mean reduction of intake in children consequent to infection was 48% (Whitehead, 1901). Food restriction is worsened by the traditional custom of withholding food during illness (Scrimshaw et al., 1968). Fortunately, much of the absorptive capacity of the mucosa remains intact during enteric infection, permitting feeding during illness and convalescence (Molla et al., 1982).

TABLE 13 Incidence of Selected Infectious Diseases and Symptoms Experienced by 45 Cohort Children, Birth to Age Three Years

Diseases, Symptomsa		Cases	Percent of All Cases	Rate per 100 Person-Yearsb
Diseases	Respiratory
	Bronchitis	256	10.5	193.9
	Bronchopneumonia	70	2.9	52.0
	Tonsillo-pharyngitis	10	0 4	7.6
	Enteric
	Diarrhea	640	26.1	484.8
	Diarrhea with mucus	313	12.8	237.1
	Dysentery	97	4.0	73.5
	Mouth
	Herpes simplex, primary	20	0.8	15.1
	Stomatitis	23	0.9	17.4
	Thrush	28	1.1	21.2
	Glossitis, cheilitis	14	0.6	10.6
	Skin, scalp
	Impetigo	31	1.3	23.5
	Abscess, furuncle	16	0.7	12.1
	Common communicable
	Measles	32	1.3	24.2
	Rubella	15	0.6	11.4
	Chicken pox	26	1.1	19.7
	Febrile exanthem	25	1.0	18.9
	Whooping cough	10	0.4	7.6
	Other
	Fever of unknown origin	14	0.6	10.6
Symptoms	Fever, 37.5+ C	1,148	25.7	869.7
	Diarrhea, all	1,205	27.0	912.7
	Vomiting	235	5.3	178.0
	Anorexia	823	18.4	623.5
	Despondent, irritable	1,008	22.6	763.6

^a Twenty-four diagnoses had fewer than 10 cases each and were omitted from the table.
^b 132 person-years of experience. taking attrition into account

The deleterious effect of infection on nutrition is reflected in lower body weight and height. Male child No. 12, born with some weight deficit, had normal growth velocity during exclusive breast-feeding (Figure 8). Enteroviral infections and mild respiratory illnesses in that period had no apparent clinical impact. But with increasing morbidity from 6 to 27 months (weaning), the curve flattened, particularly from the second to the fifth semesters. Most cohort children behaved like child No. 12: they became malnourished in the second semester until the end of the second year or longer. Weight increments improved in the third year of life for some, but not all, children.

TABLE 14 Consumption of Tortillas During Whooping Cough by 21 Cauqué Children Two to Three Years Old

		Percent Consumed by Month Ill
Child Number	Age (months)	1st	2nd	3rd
22	41	43b	84	108
54	36	93	79	82
37	38	41	94	94
52	35	108	109	149
59	35	130c	200	200
76	31	94	90	92
11	44	133	111	102
16	44	110	122	142
35	39	98	45	105
24	41	52	109	105
23	41	95	54	90
31	41	96	122	117
69	31	99	91	97
83	29	55	39	103
91	26	100	100	60
49	36	78	80	69
34	38	139	57	80
88	27	98	89	98
80	29	100	91	100
94	27	113	154	224
18	42	33	36	74

^a Percent consumption in relation to intake before onset, considered as 100%
^b Reduction by more than 50% of usual intake
^c Increase by at least 25% of usual intake
Source: Mata (1978a)

FIGURE 7 Calorie consumption and infectious diseases after cessation of breast-feeding in child No. 37. Intakes were measured at weekly intervals, and adequacy was expressed as percentage of recommended calories per day. Low caloric intakes were observed during illnesses, and adequate intakes during periods of good health. A considerable calorie deficit ("dip") of approximately 40% of the recommended calories was noted at about 33 months of age, coinciding with diarrhea with anorexia. A similar response was observed in the remaining fully weaned children (Mate et al., 1977).

The age of the child and the type of feeding influenced the effect of infection on growth. For instance, whooping cough was of less nutritional consequence during exclusive breast-feeding (Figure 9, children No. 214 and 216) than in the protracted weaning period (Figure 9, children No. 172, 177, and 181). Girl No. 177 lost almost 2 kg in two weeks, which she could not recover until 16 weeks later. She caught up to preinfection weight after about 36 weeks, but was still further below the mean weight for village children by this time.

The nutritional effect of infection was worse if there was underlying fetal growth retardation or prematurity, illustrated for two children with different birth weights (Figure 10). All diarrhea episodes in the first two years of life were associated with some stagnation of linear growth, and this was greater and more prolonged in the child who had been born with greater fetal growth deficit (Mate, 1982b).

Infectious diseases were efficient killers of infants who were either born small-for-gestational age or prematurely. For instance, female child No. 19, born with deficient weight but growing satisfactorily under exclusive breast-feeding (Figure 11), developed respiratory infection and anorexia at 9 weeks of age and meningitis at 13 weeks, which lasted 6 weeks. These events were associated with 12 weeks of weight faltering. The child continued at the breast, with supplementary food, and caught up somewhat (oddly, after diarrhea and varicella). However, growth deteriorated shortly thereafter, coinciding with attacks of stomatitis, diarrhea, and respiratory infection. The child developed severe psychomotor retardation and malnutrition, and died at the end of her second year of life. The infection-malnutrition complex can kill children who have survived the perilous weaning period. For instance, child No.44, discussed above (see Figure 8), reached school age but succumbed to typhoid fever at seven years of age. Other examples of infection-nutrition interactions in the 45 cohort children were described elsewhere (Mate, 1975, 1978a, 1982b, 1983a,b,c, 1985, 1990, 1992; Mata et al., 1971, 1972a,b, 1975, 1976b, 1977, 1982, 1984).

FIGURE 8 Natural history of enteric infections, infectious diseases, and growth (body weight and increments) of village child No. 12 during his first three years of life. Good protection against infection was observed during exclusive breast-feeding (7 months in this case). Infections appeared early in life but were most common after the first year. Diarrhea increased with weaning, to reach maximum rates in the second year of life. Growth was adequate during exclusive breastfeeding, but thereafter, there were recurrent periods of acute weight loss or growth faltering in conjunction with emerging infectious diseases. The child became overtly malnourished by age one year (Mate et al., 1971; Mata, 1978a).

FIGURE 9 Effect of whooping cough on the weight curve of children at the breast or undergoing weaning. This disease had a negative effect on growth. Exclusively breast-fed infants showed moderate weight loss and better recuperation (children Nos. 214 and 216) than weanlings and fully weaned children. These lost more weight and took several months to make up the loss (children Nos. 177, 172, 181, and 186) (Mate et al., 1972a,b,c; Mata, 1978a).

FIGURE 10 Growth in body length during the first two years of life, two typical village boys with different grades of intrauterine growth (IUGR). The child at the left (No. 40) had moderate IUGR; the one at the right (No. 2) had marked IUGR. Most episodes of diarrhea in both children were associated with growth retardation of varying duration, regardless of the etiology of the diarrhea. Stunting (deficit in length) was noticeable in both children, but it was more evident in the child born with the lower birth weight.

FIGURE 11 Elimination of village child No. 19 by the interaction of infectious diseases and malnutrition. Growth velocity of this low-birth-weight girl was adequate during the first nine weeks of life when she was fed exclusively at the breast. At that point she developed bronchitis with growth faltering and, soon after, an episode of Streptococcus meningitis, followed by diarrhea and varicella. For 12-14 weeks, the child hardly grew. Later, she caught up somewhat but still could not attain the village average growth curve. A new series of infections affected the girl, who already had evident psychomotor retardation. Eventually, the child died prematurely at the end of the second year of life (Mate et al., 1975; Mata, 1978a).

Growth and Development of Under-Fives

The distribution of 430 Cauqué, singletons, by birth weight and gestational age in the Lubchenco grid, revealed fetal growth retardation for all types of newborns (Figure 12). The large shaded area in Figure 12 identifies small-for-dates infants (33% of the total). The smaller area (7%) identifies premature babies (Mate, 1978a). The total of these subpopulations equaled 42% of he singletons. The remaining infants were adequate for gestational age, although of smaller size than newborns from well nourished women. The relative distribution of low birth weight and small for-dates infants did not vary over the nine years of study.

This population of small and malnourished neonates evolved into small and stunted children, as seen in the mean weight curve of Canqué girls and boys, during the first three years of life, in comparison with the standard (Figure 13) (Mate, 1 978a). Weight gain was stereotyped: it was adequate during exclusive breast-feeding, departed from the reference curve there after, and showed marked deficits from one to three years of age. The same behavior was seen for height and for circumference of head and thorax. Children tended to remain within growth tracks defined at birth, with the largest children growing better postnatally, and the smallest ones growing more slowly during infancy and preschool age (Figure 14) (Mate, 1 978a). The same was seen with the other anthropometric variables.

FIGURE 12 Distribution of 430 singleton infants by birth weight and gestational age, in the Lubchenco grid. Figures are the number of infants within each compartment. Three shaded areas are marked: two of infants less than 37 weeks of gestation (premature, of which the largest (less than 2 kg) were the most immature. One large area brackets the babies born at term, with less than 2.5 kg, namely, the "small-for-gestational age." These two groups of babies accounted for 42% of the total delivered during the study period (Mate, 1978a).

There was a significant correlation between birth parameters and postnatal growth, but only a small part of the variance of physical growth was explained by weight and height at birth. Small girls grow into stunted women, who deliver small babies. The interruption of the intergeneration cycle would require breakage of the infection-malnutrition interaction in childhood, through control and prevention of infections.

FIGURE 13 Average weight curve for Cauqué, children (boys and girls together), compared with the Jackson-Kelly standard. There were individual variations in growth, but children in general experienced some degree of wasting and stunting after they entered into the protracted weaning period. By one year of age, all had nutritional deficit of some sort. On average, boys were slightly larger than girls, but malnutrition was similar in both sexes. Surviving children were adapted to the deficient environment, and wasting tended to disappear after preschool age. In contrast, stunting persisted through adolescence (Mate, 1978a).

Individual growth curves were fitted by the regression equation y = a + bx + c log x used to fit cross-sectional data (Malcolm, 1970), where a is the intercept, an estimate for age 0; b is the linear parameter, reflecting most of the growth after growth deceleration; and c is the parameter of growth in the first months. The fitting gave R² values as high as 0.99. Stepwise regression analysis of fitted physical growth showed that higher maternal weight, lower maternal age, fewer deliveries, lower weaning age, and smaller number of days with respiratory infections correlated with faster growth in the first months of life (Mate, 1978a).

By stepwise regression for parameter c (child growth in early months), days of respiratory disease, maternal weight, young maternal age, number of deliveries, weaning age, and fewer days of respiratory infection were positively correlated. In a multivariate analysis, days of respiratory infection in infancy were negatively correlated with growth in the first months, explaining 19% of the variance. Maternal age was negatively correlated, adding 14.5%. Birth length and maternal weight also explained part of the variance. These four variables accounted for 48.7% of the total variance of early child growth (Mate, 1978a). Young women had infants who grew faster in the first months, whereas older mothers had infants who grew more slowly. Premature infants, as a whole, had adequate growth velocities while they remained at the breast.

FIGURE 14 Mean weight curves and standard errors for all village children separated into four subcohorts defined by birth weight. Numbers on the curves represent the number of children measured. On the average, infants with higher birth weight grew better, and those with the lowest birth weight grew poorly. Children tended to remain within their growth tracks. A similar behavior was noted for body length and for circumferences of head and thorax (Mate, 1978a).

Regarding parameter b weaning age accounted for 16.3% of the variance of linear growth by stepwise multivariate analysis. Other variables contributing significantly to explaining delayed linear growth were short birth interval, maternal height, rate of Entamoeba histolytica infection, fewer deliveries, and rate of infection with Shigella flexneri. This evidence indicated the need to improve the condition of the mother and the sanitary environment, in order to boost nutrition of young girls, pregnant women, and other women of reproductive age.

With regard to psychomotor development, most Cauqué, newborn infants appeared adequate at birth, as noted for other Cakchiquel infants (Wug de León et al., 1964). With adapted scales (Cravioto et al., 1966), Canqué infants showed adequate psychomotor development in the first semester of life. Thereafter, deficits became evident, and an evaluation of six- to seven-year-olds of the 1964 and 1965 cohorts, aided by the bilingual teacher of the village, showed a lower performance as compared to urban children. The lack of marked individual variation among children suggested that deficits probably reflected the influence of the "village environment" on all children. As adolescents and adults, Canqué Indians used a rich vocabulary to describe and interpret the weather, agriculture, handicrafts, health, and other aspects of village life, a vocabulary in some ways richer than that of Spanish-speaking mestizo and caucasians of Central America.

Malnutrition and Adaptation

The main deficiencies of nutrients in the village were calories, good quality protein, iron, and some vitamins and micronutrients. Iron deficiency in the absence of malaria and hookworm is explained by exceedingly high rates of infection that are associated with sequestering of free serum iron, coupled with the low bioavailability of iron in the predominantly vegetarian diet. Interestingly, most of the anemia disappeared without specific treatment by the fourth year of life, probably when children became immune to the infections prevailing in the village, and when their food consumption improved (Mate, 1978a). No hypoproteinemia was found in Cauqué, children except after severe infections (Viteri et al., 1973). Xerophthalmia, keratomalacia, and blindness attributable to vitamin A deficiency were not found during the study. Fruits and vegetables supplied sufficient carotenes and ascorbic acid. There was sufficient exposure to sunlight to prevent vitamin D deficiency. Iodine deficiency had been corrected by the iodination of table salt (Ascoli and Arroyave, 1970). No longitudinal studies of biochemical and endocrine functions were carried out.

Severe infectious diseases diminished food intake, caused nutrient losses, and altered metabolism, aggravating chronic malnutrition or precipitating its acute forms. There were 32 cases of edematous protein energy malnutrition in children under five during the study period (9 prekwashiorkor, 10 kwashiorkor, and 13 marasmic-kwashiorkor), which appeared a few weeks after the occurrence of diarrhea and other infectious diseases (Figure 15). Most cases appeared from May to September-the rainy season-the period of greater incidence of diarrhea, measles, and other communicable diseases. Paradoxically, this was the period of better supplies of local foods (maize, green leaves, fruits), greater cash income from selling produce, and increased availability of firewood for cooking. There were few or no cases of edematous malnutrition in the drier months (February, March, April, November, and December).

Infant Survival and Mortality

Maternal factors were the main determinants of infant growth and size at birth, and correlated strongly with child survival. Low birth weight infants had lower infant survival than those of normal birth weight. A birth weight greater than 2,750 g correlated with absolute infant survival, regardless of postnatal events, such as measles, dehydrating diarrhea, or deficient diet (Table 15). Exclusively breast-fed infants who weighed at least 2,000 g at birth survived the first week of life; if they weighed at least 2,750 g at birth, they survived the first three months of life; if they weighed at least 3,000 g at birth, they survived the first semester (see Table 16). Similar behavior was seen in cohorts defined by gestational age. LBW accounted for more than 70% of the total infant mortality in Cauqué, just as in modern societies with low infant mortality (Mata,1982b). Prematurity was the main predictor of low infant survival, but such risk was not evident after the first year of life; that is, premature infants who survived infancy did not die thereafter (Mata, 1978a). Low birth weight infants also had low survival, an effect carried on into the second, third, and fourth years of life (Table 16).

These data show the urgent need to reduce the incidence of low birth weight infants as a necessary step to reduce mortality in children under five. Almost all 18 neonatal deaths occurred in premature babies and were attributed to antenatal causes that are also typical of modern societies. With postneonatal fatalities, all but two "crib" deaths and one undetermined death were attributed to infectious diseases (Table 17). Lower respiratory disease accounted for almost half the deaths in the postneonatal infant period, followed by diarrhea and measles. Whooping cough, pneumonia, measles, and diarrhea were related to two-thirds of the deaths in the second year. The enhanced risk of death due to infection in the second year had been recognized (Gordon et al., 1967). Had the population been vaccinated-a utopian hope at that time-19 of the 58 deaths (33%) would have been spared. Medical treatment at the Health Clinic most likely averted many deaths from diarrhea and other infectious diseases, as evidenced by the progressive reduction of mortality through the study period.

Paradigm of Infection-Malnutrition

The evidence furnished by the Cauqué, study (Mate, 1 978a), complemented by observations in Costa Rica (Mate, 1982b, 1983c), identified infections as the main cause of malnutrition, growth retardation, and premature death (Figure 16). This is quite evident for diarrhea! disease, unless proper therapy is aggressively implemented, meaning rehydration, proper feeding, and adequate drugs, when indicated. The effect of infection on nutrition seems stronger than that of the nutritional state on the outcome of infection. In Figure 16, the directions and widths of the arrows illustrate the causality and intensity of the associations.

FIGURE 15 Accumulated incidence of edematous energy-protein malnutrition, by month of year, Santa María Canqué, 1964-1972. There were 32 cases, 72% of which appeared from May through September. This is the rainy season, and the months of infectious diseases. Also, they are months of relatively greater availability of food, firewood, and money. Edematous malnutrition was not observed or was infrequent from October through February. Cases were similar for both sexes, and the case fatality ratio was 15.6% (Mate, 1978a).

TABLE 15 Survival During Infancy, by Birth Weight, Among 430 Canqué Singletons, 1964-1973

		Birth Weight, g
Time Survived	No.of lnfants	1,000-	1,500-	2,000-	2,250-	2,500-	2,750-	3,000-	3,250-	3,500-	Total
24 hours	430	4/5a (80)	25/28 (89)	47/47 (100)	99/99 (100)	125/125 (100)	82/82 (100)	31/31 (100)	11/11 (100)	2/2 (100)	426/43 (99)
7 days	430	4/5 (80)	23/28 (82)	47/47 (100)	99/99 (100)	125/125 (100)	82/82 (100)	31/31 (100)	11/11 (100)	2/2 (100)	424/430 (99)
28 days	429	2/5 (40)	22/28 (79)	45/47 (96)	96/99 (97)	123/125 (98)	82/82 (100)	30/30 (100)	11/11 (100)	2/2 (100)	413/429 (96)
3 months	429	2/5 (40)	18/28 (64)	44/47 (94)	96/99 (97)	123/125 (98)	82/82 (100)	30/30 (100)	11/11 (100)	2/2 (100)	408/429 (95)
6 months	429	1/5 (20)	15/28 (54)	43/47 (91)	96/99 (97)	120/125 (96)	80/82 (98)	30/30 (100)	11/11 (100)	2/2 (100)	398/429 (93)
1 year	428	1/5 (20)	13/28 (46)	41/47 (87)	95/99 (96)	117/124 (94)	78/82 (95)	30/30 (100)	10/11 (91)	2/2 (100)	387/428 (90)

^a Number of survivors/total cases; below, (percentage surviving)
Source: Mata (1978a)

TABLE 16 Relationship Between Growth Retardation and One- to Four Year-Old Mortality, Santa María Canqué, 1964-1972

	Mortality, Year of Life
Newborn Classa	1st	2nd	3rd	4th
Normal	12 (50) [242]	9 (44)a [204]	5 (33) [153]	1 (8) [122]
Small-for-Gestational Age	12 (84) [143]	8 (76)b [105]	3 (39) [78]	3 (50) [60]
Premature	16 (516) [31]	0 [15]a	0 [13]	0 [8]
Total	40 (96) [416]	17 (52) [324]	8 (33) [244]	4 (21) [190]

^a By birth weight and gestational age combined (see Figure 11)
^b Number of deaths (rate per 1,000 children in the class alive at the beginning of the period); below [population at the beginning of the period]. Note: attrition in numbers with age reflects the lower age of younger cohort children. The 1964 cohort contributed more children to the table than cohorts born later on.
Source: Mata (1978a), modified

TABLE 17 Numbers and Relative Percentages of Child Deaths Attributed to Infectious Diseases, Santa María Cauqué, 1964-1972

Age (Years)	Population at Riska	Total Deathsa	Acute Diarrhea	Measles, Whooping Cough	Measles, Whooping Cough, Diarrhea
<1	458	41	4 (10)	9 (22)b	13 (32)
1	400	19	3 (16)	13 (68)	16 (84)
2	323	8	2 (25)	3 (38)	5 (63)
3	258	4	1 (25)	3 (75)	4(100)
4	198	0
5	148	0
6	110	0

^a-Accumulated during the Cauqué, Study, 1964-1972
^b Number of deaths (relative percentage)
Source: Mata (1978a), modified

Recurrent infections progressively lead to wastage and stunting. Severe infectious episodes precipitate acute malnutrition and cause death. In turn, some infections impair delayed hypersensitivity and other immune functions (Scrimshaw et al., 1968). Immunosuppression favors infection, closing the vicious circle. These findings support the conclusion that to improve nutrition it is necessary to control and prevent infections.

The control of environment (water supply, sanitation, personal hygiene) is fundamental to reducing diarrhea! disease, and this is a sine qua non to decrease infant mortality (Mate, 1982b).

FIGURE 16 Infection-Malnutrition Interaction Cycle of infection-nutrition interactions, as seen from the Cauqué, experience. Recurrent infectious disease (and asymptomatic infection) reduce consumption of the village diet. Infections alter digestion and absorption and cause nutrient losses and metabolic alterations. The net result is progressive wasting and eventual stunting. Infectious disease precipitates acute energy-protein malnutrition, causes disability, and ends in premature death. Malnourished individuals have an altered capacity to respond to infection and to heal. The cycle can be interrupted by control and prevention of infections acting on the childhood population, to diminish the negative effects of infection and improve nutrition, growth, and survival (Mate, 1990, 1992).