Visible signs of illness from the 14th to the 20th century: systematic review of portraits

Abstract

Objectives

To see whether a collection of portraits depicting inhabitants of a defined geographical region and covering several centuries is a useful source for studying the sociocultural significance and epidemiology of particular visible diseases, such as goitre, which is known to have been common in this region.

Design

Systematic review of portraits and description of visible signs of illness.

Setting

The Burgerbibliothek (archives of the burghers' community) in Berne, Switzerland.

Data sources

3615 portraits; 2989 of individuals whose identity is known and 626 of individuals whose identity is unknown.

Main outcome measures

Visible signs of illness evaluated by means of a standardised visual assessment.

Results

Visible signs of illness in portraits were common and appeared in up to 82% (451/553) of paintings from the 17th and 18th centuries. The most common findings were signs of goitre in women and overweight in men. In only the portraits where the neck region could be evaluated, 41% of women with known identities (139/343) had goitre compared with 24% of men with known identities (21/86). The prevalence of goitre was even higher in sitters whose identities were unknown: 63% in men (5/8) and 68% in women (82/121). Overweight in people with known identities was more common in men than in women (30%, 346/1145 v 44%, 811/1844). Overweight was most common in sitters aged >40 than in those aged 40 or younger. Other conditions, such as missing teeth, amputated limbs, or osteoarthritic deformations were surprisingly rare in the portraits under evaluation.

Conclusions

Goitre and other diseases are under-represented in the people depicted in these portraits. Artistic idealisation is a likely explanation for this observation: what was reproduced depended on what was considered pathological or shameful at the time, and therefore depended on age and sex. Stigmatising details may have been omitted. Further, artistic skills and contemporary fashion may have influenced the way in which people were reproduced. People depicted are possibly not representative of the general Bernese population as they constituted a socioeconomically advantaged group.

What is already know on this topic

Reviews of individual portraits from the past have found clinical signs of illness that have led to discussions of underlying diseases

Goitre probably affected in excess of 80% of the population of the canton of Berne up to the beginning of the 20th century

What this study adds

In a large series of portraits from the Bernese region, goitre and other diseases are under-represented

Findings of age dependent overweight (a survival advantage in times of potential famine) were probably more realistic

Likely explanations for this include idealisation depending on sex and age, artistic skills, fashion, and sociocultural significance of illness

A decline in depicted signs of illness from the 19th century may indicate progress of preventive medicine and hygiene

Introduction

Images or pictures, particularly portraits, are important sources of information in the field of medical history. In case reports, clinical signs of illness in portraits from times past have resulted in vivid discussions about the underlying diseases.^1–4 Particular attention has been paid to overweight,² transfusion between neonatal twins,³ facial palsy,⁵ and rheumatoid polyarthritis.¹ Representations of disfiguring diseases such as alopecia areata,⁶ goitre, and cretinism⁷ sometimes prompted moral judgment in the spectators. Historians have tried to deduce the socioprofessional standing of scientists from the way they were portrayed.⁸,⁹ The analyses were often based on artworks created by famous and skilled artists, whose intention was to describe not a pathological condition but rather an individual personality in its social context. Only few artists—for example, the physician (and draughtsman) Jean-Martin Charcot (1825-93)—documented visible clinical signs of diseases for teaching purposes.¹⁰

In landlocked, mountainous countries, iodine deficiency prevailed for many centuries, probably since the last glacial period. The Bernese region in Switzerland was particularly known for severe endemic goitre as a result of iodine deficiency.⁷ Since the 1920s, the prevalence of goitre has decreased as iodised table salt became widely used.¹¹

We thought that a large number of portraits of inhabitants of a defined geographical region painted over several centuries could be a valuable resource for studying the epidemiology of particular diseases—for example, goitre. Therefore we evaluated signs of illness shown in portraits.

Methods

Portrait material

We evaluated 5493 photos of historical portraits of people living in the canton of Berne. The photographic documentation, mostly black and white pictures, has been gathered since 1962 in the Burgerbibliothek Bern (the archives of the burghers' community).¹²,¹³ Our study period extended from the Renaissance to the 20th century, with most portraits dating from the 17th (934, 17%), 18th (2252, 41%), and 19th (2142, 39%) centuries. Photos of paintings were collected irrespective of the artistic quality of the painting. The depicted individuals were of high and middle social rank but were not always wealthy. Their portraits usually aimed at documenting genealogy and representing rank. Unmarried women were generally not portrayed, as their social rank was not considered to be of importance. As all sitters were dressed in the fashions of their times, analysis of their body features was occasionally difficult. We excluded paintings for which sociodemographic data were incomplete, paintings that showed several people, and miniatures and shadow portraits.

One observer (YS) was trained in a pilot study with 500 portraits and systematically evaluated all portrait photos over four months in 1998. Where people had been depicted more than once we considered only the picture with the highest number of apparent clinical signs.

Evaluation criteria

We evaluated sociodemographic data, signs of disease, and represented body parts. Sociodemographic data were usually noted on the back of the portrait and in the archives of the Burgerbibliothek. For each picture we extracted the name, year and place of birth of the sitter, the approximate age (where the year was not given), the date of the portrait, and the name of the artist.

We extracted signs of goitre (characteristic swelling of anterior neck), overweight (fat neck or apparent obesity), anomalies of eyes and vision (glasses, strabism, exophthalmos, conjunctivitis, lid retraction, lacrimal cyst, xanthelasma, Hertoghe sign (lateral thinning of eyebrows), epicanthal fold, or ptosis), appearance of face and neck (cretinoid aspect, coarse or expressionless face, missing teeth, concave nose saddle, rhinophyma, parotid hypertrophy, and deformed neck), appearance of skin (pallor, flush, eczema, alopecia, and tumour), and appearance of the whole body (disproportionate physique, osseous deformations, thickened joints of hands or wrists, and whether or not the sitter used a cane).

We extracted whether only the head and neck or whether also the trunk and extremities were depicted. We also extracted how well the neck could be evaluated (it was often concealed by clothes, jewels, laces, ribbons, or painting artefacts). We registered symptoms of disease and technical data after performing a standardised visual inspection of each portrait.¹⁴

Statistical analysis

We compared clinical signs visible in the 3615 portraits between men and women, age groups (⩽40 years v >40 years), and people whose identities were known and those whose identities were unknown. We also described the occurrence of clinical signs over the centuries. We used χ² tests to calculate differences and their 95% confidence intervals.

Results

We excluded 1878 of the 5493 portraits because they depicted body contours only, more than one person, or non-Bernese people (n=108); people whose age was not given or whose sex could not be defined (n=281); and 870 people who were depicted more than once, usually well known people (n=1489). A total of 3615 pictures remained for analysis, 2989 of people whose identities were known and 626 of people whose identities were not known. A total of 1844 (62%) of the identified people and 316 (50%) of the unidentified people were male. The age of the identified people ranged from newborn to 90 years; 10% (304) were 10 years or younger, and 18% (541) were between 11 and 20.

The greatest proportion of clinical signs of illness occurred in portraits from the 17th and 18th centuries (see table A on bmj.com), and more signs were visible in people over the age of 40 than in those aged 40 or younger. The most common findings were signs of goitre in women and overweight in men.

Goitre

We first analysed only portraits in which the neck region could be evaluated (see table B on bmj.com), and overall we found that 41% (139/343) of identified women had goitre compared with 24% (21/86) of identified men (difference 17%, 95% confidence interval 6% to 27%). Comparison between the sexes was difficult as the necks of all identified male sitters were hidden by collars in 95% (1750/1844) of cases, and the necks of all identified female sitters were not visible in 70% (801/1145) of portraits because of clothes, ribbons, jewels, embroideries, or shadow artefacts.

Among all people whose identities were known, more women than men showed signs of goitre (24%, 279/1145 v 2%, 35/1844). The signs were more common in unidentified women than in unidentified men (42%, 129/310 v 3%, 10/316). Goitre was most noticeable in portraits from the 18th century (32%, 153/483), when décolletage was fashionable.

The neck was visible in 129 unidentified sitters; goitre occurred in 63% (5/8) of men and 68% (82/121) of women. In 10 women the goitres were so large that they were visible even though their necks were hidden. Overall the prevalence of goitre was higher in identified women with a visible neck who were aged over 40 than in women aged 40 years or younger (64%, 18/28 v 38%, 121/315; 26%, 7% to 44%).

Overweight

In the identified group, overweight was more common in men than in women (44%, 811/1844 v 30%, 346/1145; 14%, 10% to 17%). In the unidentified group the difference in overweight between men and women was smaller (43%, 136/316 v 34%, 106/310; 9%, 1% to 16%). Overall, overweight in identified sitters was higher in people aged over 40 than in people aged 40 or younger (53%, 394/742 v 34%, 763/2247). The association between overweight and higher age was most pronounced during the 17th and 18th centuries (see table C on bmj.com). During the 18th century among people over 40, 79% (166/209) of identified men were overweight compared with 51% (30/59) of identified women. From the 18th to the 20th century the prevalence of overweight decreased in men over 40.

Other visible signs

We also looked for signs that were potentially related to thyroid dysfunction, such as hypothyroidism. In the identified group we found a cretinoid aspect in 8% (251/2989), epicanthal fold in 7% (197), Hertoghe sign in 5% (159), ptosis in 5% (140), a concave nose saddle in 3% (93), and a coarse face in 2% (70) of portraits. Most other signs that we evaluated systematically, including pathological conditions of bones and joints, had a low prevalence, with low frequencies of 0-1% (0-48/2989). Medical aids such as glasses or a cane were depicted in 40 people.

Discussion

This systematic review shows that historical portrait paintings are conditionally relevant for studying the epidemiology of particular visible diseases for a geographically defined region. Although visible signs of illness were common in the paintings, diseases were under-represented, possibly owing to age and sex dependent artistic idealisation, skills, and fashion, or to a sample that was not representative of the general population.

Strengths of the study

In many European countries major political upheavals and wars since the 18th century caused dispersion of cultural artefacts such as portraits throughout the world. This occurred to a lesser degree in Switzerland, which was sociopolitically more stable. We were therefore able, for the first time, to obtain photographic collections documenting a large number of portraits covering seven centuries and a territory of about 4500 km^2 12,¹⁵ and to assess them systematically. As scarcely any data exist on occurrence of illness in past centuries except some mortality data in the case of epidemics of infectious diseases, this study shows for the first time that a large series of portraits may provide relevant findings.

Limitations of the study

Paintings originated from talented as well as less talented artists. Accordingly, signs of illness may often not have been reproduced appropriately because of a lack of artistic competence rather than because of artistic idealisation. Signs of illness might not be detectable because body parts may have been presented in a non-standard way. A well known 18th century Swiss portraitist, for example, charged double if the hands were represented.¹⁶ Clothing, determined largely by fashion, also often made evaluation difficult, particularly of the neck. We tried to overcome this problem by stratifying for visibility of the neck.

Another important limitation is that artists did not reproduce signs of illness because they were stigmatising. For many centuries goitre and “cretinism” were considered to be linked.¹⁷ Therefore artists may have tried to maintain or even increase the magnificence and dignity of their sitters by not representing their goitre realistically. The idealised representation of the clinically well documented goitre of a famous Bernese novelist, Jeremias Gotthelf (1797-1854), is exemplary. Gotthelf used doses of Lugol's solution that were toxic by current standards. This led to iodine induced hyperthyroidism and finally death from cardiac failure.¹⁸ In portraits or busts his neck is carefully hidden by large and elegant collars (fig 1). Maybe artistic idealisation also explains why unknown people are represented with goitre more often than famous ones (additional portraits showing people with goitre appear on bmj.com). A more obvious example of visible goitre is shown in the portrait of a young boy (fig 2).

Figure 1.

Portrait of Swiss novelist Jeremias Gotthelf (1797-1854), with goitre (WHO grade III) well concealed. Portrait painted by Friedrich Dietler (1804-74). Photo: G Howald, Kirchlindach, Switzerland

Figure 2.

Diffuse (WHO grade II) goitre in a 6 year old Bernese boy. Portrait painted in 1797 by French artist N N Dutot. Photo: G Howald, Kirchlindach, Switzerland

Interpretation of findings

The estimated prevalence of goitre ranged from 80% to 90% in Bernese adults in 1889¹⁹ and had probably been similar in the preceding centuries. Accordingly, goitre is under-represented in our study population. As a region is defined as goitre-free if the illness occurs in less than 3% of the population, findings may be considered conditionally relevant for a region with endemic goitre. In addition to artistic idealisation, the people depicted in portraits (our study population) may not be representative for the target population as it is probably a particularly wealthy and healthy selection compared with the majority of the contemporary population.

The prevalence of overweight (which was more common in older people than in younger ones) as seen in the pictures seems plausible, but we cannot rule out that artistic idealisation played a part. Overweight may have been considered a survival advantage in times of potential famine in Europe. To the end of the 19th century, moderate overweight was not considered pathological but indicated determination, health, and wealth. Only severe obesity was considered pitiable.²⁰

Other conditions that were undoubtedly common are also under-represented in the paintings because of artistic idealisation. These include missing teeth, amputation, or osteoarthritic deformations. Pathological findings became less common from the 19th century onwards, which may represent the progress of preventive medicine and hygiene.