Iron deficiency anemia, population health and frailty in a modern Portuguese skeletal sample
Iron deficiency anemia, population health and frailty in a modern Portuguese skeletal sample
Samantha M. HensID 0 1
Kanya Godde 1
Kristin M. MacakID 0 1
0 Department of Anthropology, California State University Sacramento, Sacramento, California, United States of America, 2 Department of Sociology and Anthropology, University of La Verne, La Verne, California, United States of America, 3 Department of Sociology and Criminal Justice, Old Dominion University , Norfolk, Virginia , United States of America
1 Editor: Christopher D. Lynn, University of Alabama , UNITED STATES
Portugal underwent significant political, demographic and epidemiological transitions during the 20th century resulting in migration to urban areas with subsequent overcrowding and issues with water sanitation. This study investigates population health during these transitions and interprets results within a framework of recent history and present-day public health information. We investigate skeletal evidence for anemia (cribra orbitalia and porotic hyperostosis) as indicators of stress and frailty-i.e., whether the lesions contribute to susceptibility for disease or increased risk of death.
Data Availability Statement: All relevant data are
within the manuscript and Supporting Information
Funding: The author(s) received no specific
funding for this work.
Competing interests: The authors have declared
that no competing interests exist.
The presence and severity of skeletal lesions were compared against known sex and cause
of death data to investigate potential heterogeneity in frailty and the relationship between
lesions and risk of dying over time. Additionally, we tested for the presence of selective
mortality in our data (i.e., whether or not the sample is biased for individuals with higher frailty).
Our sample derives from a large, documented, modern Portuguese collection from Lisbon
and is the first study of its kind using a documented collection. The collection represents
primarily middle-class individuals.
Results and conclusions
Analyses indicated that porotic hyperostosis became more common and severe over time,
while cribra orbitalia severity increased over time. Neither process was linked to cause of
death. However, there was a significant relationship to sex; males exhibited a higher
prevalence and severity of lesions and increased mortality. A Gompertz function showed
decreased survivorship in early life but increased survivorship over age 60. Using
comorbidities of anemia, we were unable to detect selective mortality?i.e., in our sample, lesions do
not represent a sign of poor health or increased frailty and are not significantly linked with a
decreased mean age-at-death. However, lesion prevalence and severity do reflect the
socioeconomic processes in urban Lisbon during the 1800s and 1900s and the possibility of
water-borne parasites as the contributing factor for iron deficiency anemia.
Skeletal samples provide information about the age-at-death, health, and lifestyles of past
populations and may be useful in illuminating trends in morbidity and mortality across
sociopolitical or socioeconomic transitions throughout history. The 20th century in Portugal was
characterized by significant changes to the political-economy resulting in migration into
urban centers with subsequent overcrowding and water contamination issues. Thus, an
analysis of skeletal indicators of disease (e.g., iron deficiency anemia) coupled with age-at-death
data provide an opportunity to investigate whether skeletal lesions are a sign of overall frailty,
susceptibility to disease and/or an increased risk of death during this critical period in
Portuguese history. However, over twenty-five years ago, the landmark paper, The Osteological
Paradox: Problems of Inferring Prehistoric Health from Skeletal Samples , outlined three
fundamental problems underlying interpretations of health and disease in past populations.
These problems include: (1) demographic nonstationarity, wherein skeletal collections derived
from cemetery contexts may reflect populations that were subject to migration or changes in
fertility rates; (2) hidden heterogeneity in frailty, wherein individuals exhibit unequal (and
ultimately unknown) susceptibility to disease and risk of death due to genetically determined
biological differences, differential exposure to disease vectors, or differences in nutritional status
owing to behavioral, cultural, or environmental factors; and (3) selective mortality, wherein
the skeletal samples represent biased representatives of the living populations from which they
are drawn because individuals with the highest frailty at a particular age are more likely to die
and enter the skeletal assemblage. While nonstationary populations are a concern for
paleodemographers and population-level studies within paleopathology, hidden heterogeneity in risk
and selective mortality affect ancient health research more directly [
]. Wood and coworkers
 argue that researchers must exercise caution when interpreting skeletal lesions and their
link to health. Skeletal lesions have traditionally been viewed as a sign of poor health; however,
because some skeletal lesions take considerable time to form, they may actually indicate a
healthier individual, one who has survived for an extended length of time [
], an idea that is
essential to Wood and coworkers?  arguments.
A small number of papers successfully incorporate a research design that directly address
these issues [
], including selecting an appropriate archaeological context by focusing on
simple sites or simple societies, emphasizing subadults as non-survivors, linking frailty and
demography, and highlighting lesion formation processes. Appropriate research design should
include focusing on severity data over simple prevalence data, identifying diseases that
contribute to mortality, and examining disease patterns against age structure [
Anemia is a pathological condition characterized by a lack of iron in the blood, usually due
to insufficient dietary intake or blood loss due to parasitic infections (e.g., worms). Iron
deficiency affects more than 2 billion people worldwide [
]. Iron metabolism is a complex
process with a number of contributing factors including the ability of the intestinal cells to adjust
to available stores of dietary iron as well as physiological demands. The hormone hepcidin,
synthesized in the liver, adjusts iron levels. Hepcidin expression increases with systematic
inflammation or infection and tends to be low in girls and young women [
2 / 20
during pregnancy, infancy, and childhood, the body requires higher levels of iron to support
physiological changes. Blood loss due to heavy menses is a well-known cause of iron-deficiency
]. The prevalence of iron deficiency in menstruating and pregnant women between
1990 and 2010 was approximately 38% [
] globally, indicating that women of
reproductive age are most at risk of the condition. Furthermore, because iron deficiency has been
associated with inflammatory conditions , it may predispose an individual to infections and
heart failure [
]. Anemic patients show greater than average comorbidities in a modern
clinical office [
], with simultaneous presence of anemia with hypertension, hypothyroidism,
chronic kidney disease, malignancies, rheumatologic disorders, congestive heart failure, and
coronary artery disease.
When iron levels fall below what the body requires, the red marrow is stimulated in an
effort to generate a greater number of red blood cells and replenish oxygen levels to tissue [
]. This process causes an increase in bone marrow resulting in expansion of the cranial
diploe?, especially apparent in the cranial vault and the orbital roofs. The compact bone of the
outer table eventually resorbs if red blood cell levels are corrected, creating porotic lesions
apparent on the ectocranial surface [
Skeletally, porotic hyperostosis (PH) manifests as numerous porous lesions in the outer
table of the cranial vault, especially the parietal and occipital bones, along with expansion of
the marrow cavity (i.e., marrow hypertrophy); while cribra orbitalia (CO) similarly affects the
orbital roofs [
]. Modern clinical studies of iron-deficiency anemia , epidemiological
studies of porotic hyperostosis and cribra orbitalia [
], and radiographic evidence for cranial
vault hypertrophy support the link between cranial lesions and anemia [
However, some research suggests these cranial lesions have a more complicated etiology
beyond simple dietary iron deficiency, to include megaloblastic anemia, environmental insult,
parasitism, genetic predisposition, inflammation, and scurvy [
18, 21, 22, 25?28
infections are closely linked with sedentism, agriculture, animal domestication, as well as
aspects of environment, sanitation, personal hygiene, education, poverty, and economic
]. Emerging research in parasitism is challenging earlier views of the role of iron in
health and infection and has implications for the significance of understanding porotic
hyperostosis and cribra orbitalia in the balance of diet and disease [
]. However, Oxenham and
Cavill  and McIlvaine [
] argue that iron-deficiency anemia remains a plausible candidate
for lesions attributed to PH and CO and any dismissal of iron deficiency as an underlying
cause for these lesions is premature.
Regardless of a unifying catalyst, porotic hyperostosis and cribra orbitalia appear as
correlated indicators in the skeletons of individuals who have endured compromised conditions,
whether they be tied to nutrition, sanitation, or infectious disease. Thus, linking back to The
Osteological Paradox , these cranial lesions may be a sign of higher frailty and related to an
increased risk of death for affected individuals. Bioarchaeologists have documented PH and
CO in prehistoric and historic contexts worldwide and commonly use both conditions to
assess health and nutritional status on a population scale. Studies cover an array of causal
mechanisms including dietary-based deficiencies [
], weaning , parasitic causes
], and pathogen load , across an array of geographic regions (e.g., Australia and the
Pacific Islands [
], Greece [
], China [
] and Mexico [
]). However, no PH or CO studies
have directly addressed the issues of hidden heterogeneity or selective mortality and no
archaeological skeletal samples had documentation on the individuals comprising the collection.
Additionally, insight into modern public health issues may be gained from the examination of
skeletal lesions in documented historic populations, which may lead to a better understanding
of health trends across historic and modern contexts.
3 / 20
This study utilizes a research design intended to maximize understanding of the potential
relationship between PH and CO lesion incidence and severity with sex, known cause of
death, and over time (as measured by year of birth and year of death). Using known
comorbidities to anemia, we also test for the presence of selective mortality in our sample by examining
whether the presence of lesions (i.e., the assumption that more lesions indicates increased
frailty) contributed to the risk of death when compared to individuals without lesions. Using a
large, documented cemetery of modern Portuguese comprising both subadult and adult
individuals, an age-specific survivorship function is applied to find corresponding parameters in
subpopulations that inform the impact of selective mortality and individual heterogeneity in
the risk of death and disease. Then, we can evaluate whether (1) there are more lesions or
more severe lesions in certain subgroups that might be traditionally expected to show higher
levels of anemia and subsequently higher frailty (e.g., adult females), (2) lesions are linked to
known cause of death, and (3) year of birth or year of death affects lesion prevalence. The use
of a documented collection provides a level of detail unobtainable from an archaeological
collection. Finally, we interpret our lesion data within the demographic, socioeconomic, and
public health context of the time represented by the sample and to documented anemia prevalence
in living adult Portuguese [
Background of the sample and associated portuguese history
This study used the Lu??s Lopes Collection of identified modern Portuguese skeletons curated
at the National Museum of Natural History and Science in Lisbon, Portugal (formally known
as the Boc?age Museum). The collection comes from three cemeteries in Lisbon (Alto de S?o
Jo?o, Benfica, and Prazeres) dating to the 19th and 20th centuries. Documentation for interred
individuals includes sex, age-at-death, birth and death dates, occupation, and cause of death.
Tradition predicates the exhumation of individuals from temporary graves after five years
or complete skeletonization to allow for reuse of gravesites. If family fail to claim the remains
or to pay the fee for ossorio storage, remains are destroyed via incinerator and included in a
communal grave. Beginning in 1981, the National Museum stepped in to collect and curate
unclaimed remains and individuals for whom fees were not paid [
]. All available remains
were acquired until 1991, which resulted in a collection composed of predominantly adult
remains as many subadults were directly interred in communal graves [
]. Cardoso [
reports more than half of the infants and children were interred in the communal grave at Alto
de S?o Jo?o from at least 1875 [
] and possibly into the 20th century, particularly those from a
lower socioeconomic status (SES).
Most of the individuals represented are of middle SES from the city of Lisbon, a conclusion
supported by occupation data and the method of acquisition of the remains [
many individuals may have been born outside Lisbon and immigrated to the city where they
eventually passed [
]. The middle class is predominantly represented as the cost of temporary
interment is prohibitive for the lowest SES and the wealthy are usually able to pay for
permanent interments [
]. While the collection provides a cross-section of middle SES individuals,
some higher SES individuals are in the collection, as evidenced by their occupation and
interment in the cemetery associated with wealthier residents (Prazeres) [
]. Moreover, in the
collection, earlier dates of death are associated with a higher SES [
Common male occupations include service and sales workers, skilled workers, craftsmen,
and similar jobs. Female occupations are most frequently recorded as housewife, maid,
teacher, or student. The socioeconomic context at the time of burial was dominated by
agriculture until 1900, when industrial growth emerged in Lisbon. The expansion the domestic
market occurred slowly, with low productivity in all branches of textiles, metalworking, and food
4 / 20
production. As the main port city, Lisbon increased more rapidly in industrial production [
The first half of the 20th century experienced accelerated urban growth due to large
migrations of rural farmers into cities in search of work, resulting in overcrowding and poor living
conditions for the working class and highly impoverished [
]. For the most underprivileged
members of society, such as the sick, poor, widows, and orphans, the family was still the basis
for support and survival. As a result of urbanization, migration increased the disease load of
city inhabitants wherein overcrowding coincided with water sanitation issues. Stratification of
classes punctuated the period, whereby the middle class was subsumed into the elite and the
working class morphed into a peasantry. Nutritional disparities are evident between port cities
and inland towns and the negative effects of urbanization have been documented in various
Portuguese skeletal collections from this time period, including increased mortality rates and
decreased stature [
Materials and methods
This study examined the skeletal remains of 540 individuals from the Luis Lopez Collection of
identified, modern Portuguese individuals curated at the National Museum of Natural History
and Science in Lisbon, Portugal. Remains are available for study to qualified individuals with
permission. Specimen catalog numbers, all demographic information and lesions scores for
individuals used in this research are listed in the S1 Appendix. In our sample, year of birth
ranges from 1806 to 1950, and year of death ranges from 1880 to 1970. Sexes were
approximately equal in representation and include 281 females (52% of total population) and 259
males (48% of total population). The demographic profile by age and sex is presented in
The presence and severity of porotic hyperostosis and cribra orbitalia were visually assessed
by the third author (KMM). All skulls showing preservation of the relevant regions (i.e., cranial
vault and superior eye orbits) were scored for lesions. Presence/absence and severity of
expression were recorded on a scale of 0?3 according to the procedures for data collection outlined
by the codebook for the Global History of Health Project, modified from the Western
Hemisphere Project [
]. Use of the codebook provides uniformity in scoring lesions and
standardization in method to support replicability of results, objective identification of lesions, and is
the foundation for making an evidence-based diagnosis of lesions. Lesions need only be
present in one orbit or one portion of the cranial vault in order to generate a cumulative
porosity score for the individual [
]. If only one orbit or one parietal exhibited lesions and the
other did not, the highest score assigned was used for this study. These cases represent a small
percent of the total sample and are retained in the study. A small number of individuals had
only one side of the cranium present for observation; the present side was scored and included
in the analysis. There were 59 separate causes of death listed in the formal death registration
records. These records are reliable and have a great degree of accuracy, as they were generated
from death certificates [
]. Cause of death (COD) categories contained as few as 1 individual
(asthma), to a maximum of 82 individuals (tuberculosis). These were grouped into four
categories representing degenerative, infectious, neoplastic, and other conditions following other
published studies examining population health [
] (Table 2). Some of these may have no
association with anemia, while others may be comorbidities (discussed below). The ?other?
category was designed to capture causes of death that were not degenerative, infectious, or
neoplastic , and is not expected to be related to frailty or anemia; but is included for
comparison. These four COD categories were used for subsequent analyses comparing lesion
presence and severity to COD in order to more broadly evaluate the relationship of anemia to
In a direct test of the issue of selective mortality, comorbidities with anemia were identified
from the medical literature [
] and cross-referenced with documented cause of death in
the Portuguese sample. Thus, we test if the comorbidity mortality was selective for individuals
with anemia, i.e., whether individuals with lesions showed higher frailty and were more likely
to die from comorbid conditions than those without lesions. Comorbidities of anemia [
identified in this sample include: cancer, cirrhosis of the liver, congenital heart failure,
hypertension, chronic kidney disease, diabetes mellitus, chronic obstructive pulmonary disease, and
coronary artery disease.
Statistical analyses investigated the interaction between lesion presence and severity against
sex and COD in relation to mortality. The time frame represented by this sample spans
important epidemiological and demographic shifts in Portugal?s history (see Discussion). Therefore,
statistical analyses were conducted to model for changes over time by each parameter (e.g.,
sex, COD). Both year of death and year of birth were available in the collection?s
documentation and so they were both used as variables to represent change over time in the sample.
Binomial and ordinal (the latter using a proportional odds ratio) logistic regressions were invoked
where CO and PH presence were dependent variables for the binomial logit and CO and PH
severity for the ordinal logit. Sex and COD were main independent variables that were tested
separately against the dependent variables. Year of death and year of birth were included
6 / 20
individually and their effects were held constant as they are confounders in relation to the
main independent variables of interest. To summarize, the statistical models were built with
one dependent variable (CO or PH presence or severity) and two independent variables (sex
or COD as the main independent variable of interest and year of death or year of birth as a
confounding variable for time). P-values are provided (from applying a z test to the t values in
ordinal regression), which reflect associated confidence intervals, and allows for easier
comparisons to other studies who use p-values generated from statistical tests.
A Gompertz function plotted with a Kaplan-Meier survivorship described the survivorship
in the sample. R scripts were modified by the second author (KG) from coding written by Dr.
Lyle Konigsberg available on his website (http://faculty.las.illinois.edu/lylek/) to conduct the
survivorship analysis. A Gompertz model distribution is typical when demographic sample
age cohorts are differentially represented in the sample, such as the Lu??s Lopes Collection.
Selective mortality and frailty were tested by comparing the frail and non-frail samples to
one another, separated by presence of comorbidity. Our selective mortality protocol
incorporated a binomial logistic regression in order to test whether the presence of lesions in the
comorbidity sample led to frailty, while also controlling for age-at-death and time. The sample
type (comorbidity present or not) was the dependent variable and the independent variables
were lesion presence (main independent variable), age-at-death (confounder for age-related
changes), and year of death or year of birth (confounder for time). Age-at-death, year of death,
and year of birth were treated as continuous variables. All statistical analyses were performed
using Microsoft Excel (2008) and R Statistical Computing software, version 3.0.2. [
Statistical significance was judged using the Bonferroni corrected p-value cutoff of 0.025 (calculated
as 0.05 divided by the number of statistical tests per hypothesis, in this case two), which mostly
eliminates the impact of multiplicity of p-values from running multiple tests to explore the
same hypothesis (Type I error).
The mean age-at-death differed between males (53.980 years) and females (61.953 years). A
one-way ANOVA showed a statistically significant difference in age-at-death for both sexes
(F-value = 8.9512, p-value = 0.0029), indicating differences in survivorship in favor of female
longevity by approximately eight years.
The prevalence of PH and CO by sex and age group is shown in Table 3. Additionally, Figs
1?6 depict the prevalence and severity of PH and CO by decade. CO was documented with
greater frequency than PH. Some form of lesion was present in approximately half of the
sample. Of the 540 individuals scored, a total of 289 individuals displayed no lesions, 160
Male % PH
Female % PH
7 / 20
Fig 1. Prevalence of CO by year of birth and year of death.
individuals displayed only CO, and 35 individuals displayed only PH. The highest rates of CO
were found in young males, ages 11?24, while young adult and middle-aged adult males
exhibited the highest prevalence of PH at ages 18?24 and 35?44.
The positive coefficients for time (year of death or birth) were significant in all analyses of
sex and COD differences except CO presence, indicating PH presence and severity and CO
severity increased over time in our sample (Table 4). Lesion presence/severity varied across
sex, when year of death was held constant, in only one test: PH presence, whose coefficient
(0.6206) shows PH prevalence in males increased over time as compared to female prevalence
(Table 4). Lesion presence/severity did not differ among COD, when time was held constant
(Table 5). For selective mortality, the presence of lesions was not increased in either the sample
with comorbidities or without (Table 6). Age-at-death and time were both significant in the
selective mortality models and the positive coefficients revealed that age-at-death increased in
the comorbidity sample, as did year of death/birth. The older ages-at-death were likely more
prevalent in the comorbidity sample as it excluded accidental deaths not related to
healthinduced causes. In sum then, selective mortality was not present in these results,
demonstrating lesions did not contribute to frailty for individuals dying of anemia comorbidities.
A Gompertz hazard model produced parameters of ? = 0.001449602 and ? = 0.053574250
and appeared to have a good fit to the Kaplan-Meier survivorship (see Fig 7). The survivorship
curve produced for the Portuguese collection by the Gompertz function indicates moderate
population decline from birth to early adulthood. Survivorship declines steadily from birth to
middle age. After age 80, the hazard of dying increases exponentially as survivorship declines.
8 / 20
Fig 2. Prevalence of PH by year of birth and year of death.
We explored the relationship between lesions attributed to anemia and sex, cause of death,
frailty, and mortality in a modern, documented Portuguese skeletal sample. Because the
skeletal sample represents people living in a time of significant socioeconomic change in Portugal,
we examined our results across time. Additionally, we formally addressed the issue of selective
mortality by testing the differences in mortality patterns between frail and non-frail
individuals. Our research design linked frailty to demography, emphasized lesion severity and
prevalence, and considered diseases that may contribute to mortality (i.e., comorbidities), while
utilizing age-structured data. Numerous authors [
2, 10, 53?55
], state that this rigor is required
to appropriately address the issues raised by Wood and coworkers  in The Osteological
At the outset of this analysis, we expected that females would show more lesions and/or
more severe lesions when compared to males. The blood loss due to menstruation, along with
pregnancy, places females at a higher risk of anemia worldwide [
13?15, 41, 47
] and has been
documented archaeologically [
]. However, while lesions were common in our sample
(>50%), the highest rates of both PH and CO were found in males; but only the presence of
PH in males significantly increased over time. Klein  discusses the effects of hormones on
sex differences in infection and finds that males are more susceptible than females to infections
from parasites, fungi, bacteria, and viruses?an issue that may affect parasite-derived iron
deficiency. If the anemia in our sample was parasite-based rather than dietary-based, it may
explain the higher prevalence of anemic lesions in the males. This male bias is upheld by
9 / 20
Fig 3. Severity of CO by year of birth. A score of 2 indicates a cluster of fine foramina covering a small area, while a score of 3 indicates substantial areas covered by
small or large foramina [
] study of infectious pathogens and Bernin and Lotter?s [
] study of the effects of
steroid hormones on infectious disease. However, Klein and coworkers [
] report females at
higher risk of morbidity and mortality in influenza outbreaks, especially women of
reproductive age. The NIH recommends including sex as a criterion in studies to better understand
sexrelated differences in infection and disease [
]. While young adult males are particularly
affected by lesions in this sample, this may be an artifact of younger females with lesions
already being removed from the population at younger ages (see Table 1).
Our second expectation was that if anemic lesions were a sign of poor health and increased
frailty, then they may show a relationship to COD. After grouping the myriad causes of death
in our sample into four broad categories representing infection, degenerative conditions,
cancers, and miscellaneous, we were unable to detect a significant link between lesions and COD
groups. Results indicated no relationship between lesion presence or severity and COD in our
Finally, we expected that if selective mortality was present in our sample, individuals with
the highest frailty at any given age would be more likely to die compared to their non-frail
counterparts (i.e., selective mortality would be present) akin to that reported in other studies
]. After linking CO and PH with comorbidities identified in the medical literature, we
grouped our sample into frail and non-frail subsets, based on the presence of lesions and
associated comorbidity. This hypothesis was not supported by our analysis, indicating that selective
mortality is not present in our sample; the presence of anemia in individuals whose COD was
a comorbidity did not contribute to frailty in our sample. This result may be due to the source
of the comorbidities being cause of death; the length of time the comorbidity affected the
10 / 20
Fig 4. Severity of CO by year of death. A score of 2 indicates a cluster of fine foramina covering a small area, while a score of 3 indicates substantial areas covered by
small or large foramina .
individual while they also had anemia is unknown. Thus, tests for the presence of selective
mortality indicated that the presence of lesions was irrelevant; i.e., lesions did not contribute to
frailty for individuals with associated comorbidities nor with people dying from other causes
of death. However, lesions did increase with older ages-at-death and became more common
Starting at the beginning of the 20th century and continuing to mid-century, there was
significant migration into the city resulting in overcrowding and decreased living conditions in
urban environments [
]. This overcrowding led to water sanitation issues [
waterborne parasites are a probable cause of iron deficiency anemia?explaining the rise in lesions
over time in our sample and exposing the sexes equally. However, as mentioned above, males
are at a higher risk of parasitic infection [
] and not surprisingly show higher prevalence
of anemic lesions in our sample. Many individuals in our sample likely migrated into urban
Lisbon at some point in their life, exposing them to overcrowding and poor water quality as
they aged?and lesions did increase with older age-at-death in our sample. Additionally, while
our sample is comprised of primarily middle-class individuals, many of the higher-class
individuals (i.e., Prazares cemetery) were born to the earliest time periods in our collection when
anemia prevalence was lowest. It is entirely possible that the increase in prevalence of anemia
might also be an artifact of SES as later births represented primarily middle-class individuals.
11 / 20
Fig 5. Severity of PH by year of birth. A score of 2 indicates slight pitting or severe porosity, while a score of 3 indicates gross parietal lesion with excessive enlargement
of bone .
A more nuanced discussion of the broader sociopolitical and economic changes occurring
across Portugal may be warranted to place our results in historical context. During the time
period represented by the Lu??s Lopes Collection, the country was transitioning from a
monarchy to a Republic, followed by a dictatorship, which would have contributed to compromised
economic viability of working-class individuals, compromised nutrition, and limited health
care options. According to Moreira and Filipa [
], Portugal began epidemiological and
demographic shifts starting 1900?1920, with a change point in the 1940s, and persisting until around
the 1970s [
]. This shift is evidenced in the overall pattern of disease prevalence in the
] and in the decrease in infectious and parasitic diseases as a COD, among others. The
prevalence of infectious and parasitic disease decreased over the period from 1930?1970
(1930 = 21.1%, 1940 = 20.9%, 1950 = 13.5%, 1960 = 1.9%, 1970 = 1%) [
]. Mortality was
higher in Portugal (21%) in 1900 compared to Western Europe (17%). The life expectancy for
women in Portugal in 1900 was 35.2 years (men = 33.3 years) [
], 60.7 years in 1950
(men = 55.6 years) [
], and just under 83.6 years (men = 77.3 years) in 2012 [
The recent decrease in mortality in Portugal has been attributed to sanitation, nutrition,
and healthcare changes in the 1960s and 1970s [
]. At this time the number of homes with
tap water, showers and/or bathtubs, indoor toilets, electricity, and sewage systems began to
], leading to an increase in sanitation and better hygiene. Nutrition also changed
with the incorporation of more milk, meat, fats, sugar, eggs, and alcohol, which led to an
increase in caloric consumption [
]. An increase in the number of doctors also began at this
]. Heights increased greatly after these changes were implemented [
reflects better health of the Portuguese. Stature is often used as a suitable proxy for growth, as
it is highly dependent on appropriate nutrition and lack of disease [
]. However, our
12 / 20
Fig 6. Severity of PH by year of death. A score of 2 indicates slight pitting or severe porosity, while a score of 3 indicates gross parietal lesion with excessive
enlargement of bone .
Significant value (p<0.025)
DY or BY
DY or BY
DY or BY
DY or BY
13 / 20
z or t value
Type of COD
DY or BY
DY or BY
DY or BY
DY or BY
DY or BY
DY or BY
sample predates these modern improvements and therefore would be expected to show higher
prevalence of disease.
Recent research from the EMPIRE study [
] in public health on nearly 8,000 living
residents of Portugal found the overall prevalence of anemia was 19.9%, and that anemia was
more common in women (20.8%) and especially pregnant women (54.2%). Additionally, the
authors report anemia was more frequent in young adult (18?34) and much older adults (80+)
and that there were marked regional asymmetries with anemia more prevalent in Lisbon
(39%). Our results were not consistent with the higher rates of anemia in women and older
adults reported in the EMPIRE studies; but did find higher rates of anemic lesions in young
adults and was consistent with higher prevalence for an urban environment. We found much
Significant value (p<0.025)
14 / 20
Fig 7. A Gompertz function (solid line) fit to a Kaplan-Meier survivorship (stepped lines) curve in historic Portuguese.
higher levels of anemic lesions overall in our sample than reported for modern Portuguese. In
the EMPIRE study [
] biomarker and survey data from residents were randomly selected so
that each inhabitant would have an equal probability of being sampled, regardless of adult age,
gender, ethnicity, SES, or nutrition. The Lu??s Lopes Collection represents a cross-section of
middle-class individuals in early historic urban Lisbon where diet and environment-driven
stressors were expected to have affected these families.
Before the shift in the 1960s and 1970s, children exhibited stunting due to the living
conditions at the time [
43, 46, 72
] and a lack of antibiotics [
]. The high levels of cribra orbitalia
15 / 20
and the increasing prevalence of anemic lesions prior to 1970 coincides with this decreased
growth potential and may be linked to the lack of available antibiotics to counter water-borne
The Lu??s Lopes Collection yields complex data that provide a glimpse into past health in
Portugal. Selective mortality was not detected in our sample, so lesions did not contribute to frailty
for individuals dying from comorbidities. Indeed, individuals with lesions died at older ages
than those without lesions, suggesting lesions are a sign of survivorship. There were significant
differences in lesion severity and incidence over time and between males and females. Males in
the Lu??s Lopes Collection showed a higher incidence of porotic hyperostosis, while also
exhibiting a significantly younger mean age-at-death when compared to females.
The mortality pattern has important implications for our understanding of the living
population that produced the study sample. For this study, demographic records denote place of
birth for individuals in this sample, which indicate regional movement occurred in the lifetime
of this population. City of birth spans the country of Portugal, but all individuals died in
Lisbon, which characterizes a population that was neither insular nor stationary. Regional
movement of a population over a lifetime impacts both the distribution of pathology in a skeletal
assemblage and heterogeneity of risk. Extra-regional migration into Portugal was low at this
time and demographic and genetic indicators do not suggest appreciable differences among
the regions in Portugal and the population is relatively homogeneous . Therefore, this is
not likely a bias reflected in the results here.
If lesions signified poor health and increased frailty, then lesion expression would reflect an
elevated mortality hazard. However, if lesions represent survivorship, then lesions may be a
sign of lower frailty relative to people without lesions and would not show an elevated
mortality hazard. Our results do not support the traditional model for disease?i.e., in our sample,
lesions do not represent a sign of poor health or increased frailty and are not significantly
linked with a decreased mean age-at-death. But lesion prevalence and severity do reflect the
rich history of urban Lisbon during the 1800s and 1900s and the possibility of water-borne
parasites as the contributing factor for iron deficiency anemia. We concur with other
researchers who stress the importance of sex- and age-structured data as a critical tool for evaluating
lesion and mortality data [
2, 53, 54
]. This study was the first to use a documented skeletal
collection to address the issues of hidden heterogeneity in frailty and selective mortality as well as
providing important contextual information on local and regional history in our analysis of
morbidity and mortality.
S1 Appendix. Catalog numbers, demographic information and all data scores for all
specimens used in this study from the Luis Lopez collection at the National Museum of Natural
History, Lisbon, Portugal (n = 540).
Thank you to Dr. Susan Garcia and the curatorial staff at the National Museum of National
History and Science (Museu Boc?age) for permission to use the Lu??s Lopes Collection. We are
grateful to Dr. Margaret Gough who provided input on the demographic shift component in
our paper. Hugo Cardoso and two anonymous reviewers contributed extensive and thoughtful
16 / 20
comments on an earlier version of this manuscript and have our deepest thanks?the end result
was significantly improved by their guidance.
Conceptualization: Samantha M. Hens, Kanya Godde, Kristin M. Macak.
Data curation: Kanya Godde.
Formal analysis: Kanya Godde, Kristin M. Macak.
Investigation: Kristin M. Macak.
Methodology: Kanya Godde.
Project administration: Samantha M. Hens.
Software: Kanya Godde.
Supervision: Samantha M. Hens.
Validation: Kanya Godde, Kristin M. Macak.
Visualization: Samantha M. Hens, Kanya Godde, Kristin M. Macak.
Writing ? original draft: Samantha M. Hens, Kristin M. Macak.
Writing ? review & editing: Samantha M. Hens, Kanya Godde.
17 / 20
18 / 20
35. El-Najjar MY, Robertson AL. Spongy Bones in Prehistoric America. Science. 1976; 193(4248):141?3.
19 / 20
Wood JW , Milner GR , Harpending HC , Weiss KM . The Osteological Paradox. Current Anthropology . 1992 ; 33 ( 4 ): 343 - 70 .
2. DeWitte SN , Stojanowski CM . The Osteological Paradox 20 Years Later: Past Perspectives, Future Directions . Journal of Archaeological Research . 2015 ; 23 ( 4 ): 397 - 450 .
3. Ortner DJ . Theoretical and methodological issues in paleopathology . In: Ortner DJ , Aufderheide AC , editors. Human Paleopathology: Current Syntheses and Future Options . Washington D.C.: Smithsonian Institution; 1991 . p. 5 - 11 .
4. Boldsen JL . Estimating patterns of disease and mortality in a medieval Danish village . In: Paine RR, editor. Integrating Archaeological Demography: Multidisciplinary Approaches to Prehistoric Population. 24 . Carbondale , Ill.: Center for Archaeological Investigations , Southern Illinois University at Carbondale; 1997 . p. 229 - 41 .
5. Boldsen JL . Analysis of dental attrition and mortality in the medieval village of Tirup, Denmark . American Journal of Physical Anthropology . 2005 ; 126 ( 2 ): 169 - 76 . https://doi.org/10.1002/ajpa.20057 PMID: 15386219
6. Boldsen JL . Leprosy and mortality in the Medieval Danish village of Tirup . American Journal of Physical Anthropology . 2005 ; 126 ( 2 ): 159 - 68 . https://doi.org/10.1002/ajpa.20085 PMID: 15386293
7. Boldsen JL . Early childhood stress and adult age mortality-A study of dental enamel hypoplasia in the medieval Danish village of Tirup . American Journal of Physical Anthropology . 2007 ; 132 ( 1 ): 59 - 66 . https://doi.org/10.1002/ajpa.20467 PMID: 16897774
8. Boldsen JL , Milner GR , Weise S. Cranial vault trauma and selective mortality in medieval to early modern Denmark . Proceedings of the National Academy of Sciences . 2015 ; 112 ( 6 ): 1721 - 6 .
9. DeWitte SN , Wood JW . Selectivity of Black Death mortality with respect to preexisting health . Proceedings of the National Academy of Sciences . 2008 ; 105 ( 5 ): 1436 - 41 .
10. Lukacs JR . The osteological paradox and the Indus civilization: problems inferring health from human skeletons at Harappa . In: Kenoyer J, editor. From Sumer to Meluhha: Contributions to the Archaeology of South and West Asia in Memory of George F Dales , Jr . 3. Madison: University of Wisconsin; 1994 . p. 143 - 55 .
11. Storey R . Individual frailty, children of privilege, and stress in Late Classic Copa?n . In: Whittington S , Reed D , editors. Bones of the Maya. Tuscaloosa , Alabama: University of Alabama Press; 1997 . p. 116 - 26 .
12. Wright LE , Chew F . Porotic Hyperostosis and Paleoepidemiology: A Forensic Perspective on Anemia among the Ancient Maya . American Anthropologist . 1998 ; 100 ( 4 ): 924 - 39 .
13. Camaschella C . Iron-deficiency anemia . New England Journal of Medicine . 2015 ; 372 ( 19 ): 1832 - 43 . https://doi.org/10.1056/NEJMra1401038 PMID: 25946282
14. Kassebaum NJ , Jasrasaria R , Naghavi M , Wulf SK , Johns N , Lozano R , et al. A systematic analysis of global anemia burden from 1990 to 2010 . Blood. 2014 ; 123 ( 5 ): 615 - 24 . https://doi.org/10.1182/blood2013-06-508325 PMID: 24297872
15. Stevens GA , Finucane MM , De-Regil LM , Paciorek CJ , Flaxman SR , Branca F , et al. Global, regional, and national trends in haemoglobin concentration and prevalence of total and severe anaemia in children and pregnant and non-pregnant women for 1995-2011: a systematic analysis of population-representative data . Lancet of Global Health . 2013 ; 1 ( 1 ): e16 - 25 . https://doi.org/10.1016/ S2214 -109X( 13 ) 70001 - 9 PMID: 25103581
16. Anker SD , Comin Colet J , Filippatos G , Willenheimer R , Dickstein K , Drexler H , et al. Ferric Carboxymaltose in Patients with Heart Failure and Iron Deficiency . New England Journal of Medicine . 2009 ; 361 ( 25 ): 2436 - 48 . https://doi.org/10.1056/NEJMoa0908355 PMID: 19920054
17. Ganghi SJ , Hagans I , Nathan K , Hunter K , Roy S . Prevalence, comorbidity and investigation of anemia in the primary care office . Journal of Clinical Medicine Research . 2017 ; 9 ( 12 ): 970 - 80 . https://doi.org/ 10.14740/jocmr3221w PMID: 29163729
18. Holland TD , O'Brien MJ . Parasites, Porotic Hyperostosis, and the Implications of Changing Perspectives . American Antiquity . 1997 ; 62 ( 2 ): 183 - 93 .
19. Kent S , Weinberg ED , Stuart-Macadam P . The etiology of the anemia of chronic disease and infection . Journal of Clinical Epidemiology . 1994 ; 47 ( 1 ): 23 - 33 . PMID: 8283191
20. Stuart-Macadam P . Porotic hyperostosis: New evidence to support the anemia theory . American Journal of Physical Anthropology . 1987 ; 74 ( 4 ): 521 - 6 . https://doi.org/10.1002/ajpa.1330740410 PMID: 3327384
21. Stuart-Macadam P . Porotic hyperostosis: a new perspective . American Journal of Physical Anthropology . 1992 ; 87 ( 1 ): 39 - 47 . https://doi.org/10.1002/ajpa.1330870105 PMID: 1736673
22. Walker PL , Bathurst RR , Richman R , Gjerdrum T , Andrushko VA . The causes of porotic hyperostosis and cribra orbitalia: A reappraisal of the iron-deficiency-anemia hypothesis . American Journal of Physical Anthropology . 2009 ; 139 ( 2 ): 109 - 25 . https://doi.org/10.1002/ajpa.21031 PMID: 19280675
23. Cook JD . Adaptation in Iron Metabolism. The American Journal of Clinical Nutrition . 1990 ; 51 ( 2 ): 301 - 8 . https://doi.org/10.1093/ajcn/51.2.301 PMID: 2407101
24. Stuart-Macadam P . A radiographic study of porotic hyperostosis . American Journal of Physical Anthropology . 1987 ; 74 ( 4 ): 511 - 20 . https://doi.org/10.1002/ajpa.1330740409 PMID: 3327383
25. Blom DE , Buikstra JE , Keng L , Tomczak PD , Shoreman E , Stevens-Tuttle D . Anemia and childhood mortality: Latitudinal patterning along the coast of pre-Columbian Peru . American Journal of Physical Anthropology . 2005 ; 127 ( 2 ): 152 - 69 . https://doi.org/10.1002/ajpa.10431 PMID: 15558829
26. Stuart-Macadam P . Porotic hyperostosis: relationship between orbital and vault lesions . American Journal of Physical Anthropology . 1989 ; 80 ( 2 ): 187 - 93 . https://doi.org/10.1002/ajpa.1330800206 PMID: 2679121
27. Walker PL . Anemia Among Prehistoric Indians of the American Southwest . In: Merbs CF , Miller RJ , editors. Health and Disease in the Prehistoric Southwest . 34 : Arizona State University; 1985 . p. 139 - 64 .
28. Walker PL . Porotic hyperostosis in a marine-dependent California Indian population . American Journal of Physical Anthropology . 1986 ; 69 ( 3 ): 345 - 54 . https://doi.org/10.1002/ajpa.1330690307 PMID: 3518481
29. Stuart-Macadam P. Integrative Anthropology: A Focus on Iron-Deficiency Anemia . Archeological Papers of the American Anthropological Association . 2006 ; 16 ( 1 ): 129 - 37 .
30. Aufderheide AC . The Scientific Study of Mummies . Cambridge, UK; New York: Cambridge University Press; 2003 .
31. Oxenham MF , Cavill I. Porotic hyperostosis and cribra orbitalia: the erythropoietic response to iron-deficiency anemia . Anthropological Science . 2010 ; 118 ( 3 ): 199 - 200 .
32. McIlvaine BK . Implications of Reappraising the Iron-Deficiency Anemia Hypothesis . International Journal of Osteoarchaeology . 2015 ; 25 ( 6 ): 997 - 1000 .
33. El-Najjar MY , Lozoff B , Ryan DJ . The Paleoepidemiology of Porotic Hyperostosis in the American Southwest: Radiological and Ecological Considerations . American Journal of Roentgenology . 1975 ; 125 ( 4 ): 918 - 24 .
34. El-Najjar MY , Ryan DJ , Turner CG , Lozoff B. The etiology of porotic hyperostosis among the prehistoric and historic Anasazi Indians of Southwestern United States . American Journal of Physical Anthropology . 1976 ; 44 ( 3 ): 477 - 87 . https://doi.org/10.1002/ajpa.1330440311 PMID: 937525
36. Pechenkina EA , Benfer RA , Zhijun W. Diet and health changes at the end of the Chinese Neolithic: The Yangshao/Longshan transition in Shaanxi province . American Journal of Physical Anthropology . 2002 ; 117 ( 1 ): 15 - 36 . https://doi.org/10.1002/ajpa.10014 PMID: 11748560
37. Pearson JA , Hedges REM , Molleson TI , O ?zbek M. Exploring the relationship between weaning and infant mortality: An isotope case study from A??kl? Ho?yu?k and C?ayo?nu? Tepesi . American Journal of Physical Anthropology . 2010 ; 143 ( 3 ): 448 - 57 . https://doi.org/10.1002/ajpa.21335 PMID: 20623678
38. Zaino DE , Zaino EC . Cribra orbitalia in the aborigines of Hawaii and Australia . American Journal of Physical Anthropology . 1975 ; 42 ( 1 ): 91 - 3 . https://doi.org/10.1002/ajpa.1330420112 PMID: 1090189
39. Keenleyside A , Panayotova K. Cribra orbitalia and porotic hyperostosis in a Greek colonial population (5th to 3rd centuries BC) from the Black Sea . International Journal of Osteoarchaeology . 2006 ; 16 ( 5 ): 373 - 84 .
40. Piperata BA , Hubbe M , Schmeer KK . Intra-population variation in anemia status and its relationship to economic status and self-perceived health in the Mexican Family Life Survey: Implications for bioarchaeology . American Journal of Physical Anthropology . 2014 ; 155 ( 2 ): 210 - 20 . https://doi.org/10.1002/ ajpa.22543 PMID: 24862211
41. Fonseca C , Marques F , Robalo Nunes A , Belo A , Brilhante D , Cortez J . Prevalence of anaemia and iron deficiency in Portugal: the EMPIRE study . Internal Medicine Journal . 2016 ; 46 ( 4 ): 470 - 8 . https://doi.org/ 10.1111/imj.13020 PMID: 26841337
42. Cardoso HFV . Brief Communication: The Collection of Identified Human Skeletons Housed at the Bocage Museum (National Museum of Natural History) , Lisbon, Portugal. American Journal of Physical Anthropology . 2006 ; 129 ( 2 ): 173 - 6 . https://doi.org/10.1002/ajpa.20228 PMID: 16323180
43. Cardoso HFV , Saunders SR . Patterns of Growth and Development of the Human Skeleton and Dentition in Relation to Environmental Quality: A Biocultural Analysis of a Sample of 20th Century Portuguese Subadult Documented Skeletons . McMaster University. Department of Anthropology: Library and Archives Canada = Biblioth?que et Archives Canada; 2007 .
44. Lopes L. Subs ??dio para o Estudo da Importa?ncia Social da Crianc?a ate? ao Ano de Idade em 1875 . Garcia de Orta, Se?rie de Antropobiologia. 1995 ; 8 : 15 - 8 .
45. Birmingham D. A Concise History of Portugal . Cambridge: Cambridge University Press; 2003 .
46. Cardoso HFV . Environmental effects on skeletal versus dental development: Using a documented subadult skeletal sample to test a basic assumption in human osteological research . American Journal of Physical Anthropology . 2007 ; 132 ( 2 ): 223 - 33 . https://doi.org/10.1002/ajpa.20482 PMID: 17078036
47. Nunes AR , Fonseca C , Marques F , Belo A , Brilhante D , Cortez J . Prevalence of anemia and iron deficiency in older Portuguese adults: An EMPIRE substudy . Geriatrics & Gerontology International . 2017 ; 17 ( 11 ): 1814 - 22 .
48. Steckel RH , Larsen CS , Sciulli PW , Walker PL . The Global History of Health Project: Data Collection Codebook: Ohio State University; 2006 . Available from: http://global.sbs.ohio-state.edu/docs/ Codebook-12-12-05.pdf.
49. Gage TB . Are Modern Environments Really Bad for Us?: Revisiting the Demographic and Epidemiologic Transitions . American Journal of Physical Anthropology . 2005 ; 128 ( S41 ): 96 - 117 .
50. Weisensee KE . Assessing the relationship between fluctuating asymmetry and cause of death in skeletal remains: a test of the developmental origins of health and disease hypothesis . American Journal of Human Biology . 2013 ; 25 ( 3 ): 411 - 7 . https://doi.org/10.1002/ajhb.22390 PMID: 23559481
51. Zaninetti C , Klersy C , Scavariello C , Bastia R , Balduini CL , Invernizzi R . Prevalence of anemia in hospitalized internal medicine patients: Correlations with comorbidities and length of hospital stay . European Journal of Internal Medicine . 2018 ; 51 : 11 - 7 . https://doi.org/10.1016/j.ejim. 2017 . 11 .001 PMID: 29153343
52. R Development Core Team. R: A language and environment for statistical computing . 3 . 0 .2 ed. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from http://www.R-project.org/; 2013 .
53. Wood JW , Holman DJ , Weiss KM , Buchanan AV , LeFor B . Hazards models for human population biology . American Journal of Physical Anthropology . 1992 ; 35 ( S15 ): 43 - 87 .
54. Wright LE , Yoder CJ . Recent Progress in Bioarchaeology: Approaches to the Osteological Paradox . Journal of Archaeological Research . 2003 ; 11 ( 1 ): 43 - 70 .
55. Zuckerman MK , Harper KN , Armelagos GJ . Adapt or Die: Three Case Studies in Which the Failure to Adopt Advances from Other Fields has Compromised Paleopathology . International Journal of Osteoarchaeology [Internet] . 2016 ; 26 ( 3 ):[ 375 - 83 pp.]. Available from: http://dx.doi.org/10.1002/oa. 2426.
56. Sullivan A. Prevalence and Etiology of Acquired Anemia in Medieval York, England. American Journal of Physical Anthropology . 2005 ; 128 ( 2 ): 252 - 72 . https://doi.org/10.1002/ajpa.20026 PMID: 15795886
57. Klein SL . The effects of hormones on sex differences in infection: from genes to behavior . Neuroscience and Biobehavioral Reviews . 2000 ; 24 ( 6 ): 627 - 38 . PMID: 10940438
58. Fischer J , Jung N , Robinson N , Lehmann C . Sex differences in immune responses to infectious diseases . Infection . 2015 ; 43 ( 4 ): 399 - 403 . https://doi.org/10.1007/s15010-015 -0791-9 PMID: 25956991
59. Bernin H , Lotter H . Sex bias in the outcome of human tropical infectious diseases: influence of steroid hormones . The Journal of Infectious Diseases . 2014 ; 209 ( S3 ): S107 - 13 .
60. Klein SL , Hodgson A , Robinson DP . Mechanisms of sex disparities in influenza pathogenesis . Journal of Leukocyte Biology . 2011 ; 92 : 67 - 73 . https://doi.org/10.1189/jlb.0811427 PMID: 22131346
61. Clayton JA , Collins FS . Policy: NIH to balance sex in cell and animal studies . Nature . 2014 ; 509 ( 7500 ): 282 - 3 . PMID: 24834516
62. Moreira MJ , Filipa H. Demographic and health changes in Portugal ( 1900 - 2013 ). Hygiea Internationalis . 2016 ; 12 : 9 - 39 .
63. Kinsella KG . Changes in life expectancy 1900-1990 . The American Journal of Clinical Nutrition . 1992 ; 55 ( 6 ): 1196S - 202S .
64. Alcoforado MJ , Marques D , Garcia RAC , Cana?rio P, Nunes MdF , Nogueira H , et al. Weather and climate versus mortality in Lisbon (Portugal) since the 19th century . Applied Geography . 2015 ; 57 : 133 - 41 .
65. Padez C . Stature and stature distribution in Portuguese male adults 1904-1998: The role of environmental factors . American Journal of Human Biology . 2003 ; 14 : 39 - 49 .
66. Cardoso HFV . Secular Changes in Body Height and Weight of Portuguese Boys Over One Century . American Journal of Human Biology . 2008 ; 20 ( 3 ): 270 - 7 . https://doi.org/10.1002/ajhb.20710 PMID: 18172870
67. Cardoso HFV , Gomes JEA . Trends in adult stature of peoples who inhabited the modern Portuguese territory from the Mesolithic to the late 20th century . International Journal of Osteoarchaeology . 2009 ; 19 : 711 - 25 .
68. Bennike P , Lewis ME , Schutkowski H , Valentin F . Comparison of child morbidity in two contrasting medieval cemeteries from Denmark . American Journal of Physical Anthropology . 2005 ; 128 ( 4 ): 734 - 46 . https://doi.org/10.1002/ajpa.20233 PMID: 16044468
69. Kemkes-Grottenthaler A . The short die young: The interrelationship between stature and longevityevidence from skeletal remains . American Journal of Physical Anthropology . 2005 ; 128 ( 2 ): 340 - 7 . https://doi.org/10.1002/ajpa.20146 PMID: 15861421
70. Pinhasi R , Teschler-Nicola M , Knaus A , Shaw P . Cross-population analysis of the growth of long bones and the os coxae of three Early Medieval Austrian populations . American Journal of Human Biology . 2005 ; 17 ( 4 ): 470 - 88 . https://doi.org/10.1002/ajhb.20406 PMID: 15981184
71. Ribot I , Roberts C. A Study of Non-specific Stress Indicators and Skeletal Growth in Two Mediaeval Subadult Populations . Journal of Archaeological Science . 1996 ; 23 ( 1 ): 67 - 9 .
72. Heuze Y , and Cardoso H.F.V. Testing the quality of nonadult Bayesian dental age assessment methods to juvenile skeletal remains: the Lisbon collection children and secular trend effects . American Journal of Physical Anthropology . 2008 ; 135 : 275 - 83 https://doi.org/10.1002/ajpa.20741 PMID: 18000887 73 .
Weisensee K , Jantz RL . Secular changes in craniofacial morphology of the Portuguese using geometric morphometrics . American Journal of Physical Anthropology . 2011 ; 145 : 548 - 59 . https://doi.org/10.
1002/ajpa.21531 PMID: 21541933