Gestational Diabetes: History and Discovery

Gestational diabetes mellitus (GDM) is high blood sugar that is first recognized during pregnancy. For most of medical history it was barely a category at all: before insulin was discovered in 1921, women with what we now call type 1 diabetes rarely lived long enough to conceive, and the few who became pregnant faced maternal and fetal death rates near 50 percent. Insulin changed that overnight — and as diabetic women began to survive pregnancy, physicians started to notice a milder, second pattern: glucose intolerance that surfaced only during pregnancy, often with unusually large babies and unexplained stillbirths. The story below traces that recognition from an 1824 Berlin doctoral thesis, through the mid-century observations of J. P. Hoet, to the 1964 oral-glucose-tolerance-test criteria of John B. O'Sullivan and Claire M. Mahan, Norbert Freinkel's metabolic insights, and the landmark 2008 HAPO study that reshaped how the condition is diagnosed today. Throughout, names, dates, and "firsts" are stated only where confirmed by published sources, and uncertainties are flagged as such.

Table of Contents

  1. What Gestational Diabetes Is — and Why Its History Is Unusual
  2. Before Insulin: Diabetes and Pregnancy as a Near-Death Sentence
  3. Bennewitz 1824: The First Recorded Diabetic Pregnancy
  4. The Insulin Revolution of 1921–1922
  5. Mid-Century Recognition: Hoet, Obstetric Risk, and a New Name
  6. O'Sullivan and Mahan, 1964: Defining the Modern Entity
  7. Norbert Freinkel and Fuel-Mediated Teratogenesis
  8. HAPO 2008 and the IADPSG Criteria
  9. Legacy: Macrosomia, Screening, and the Road to Type 2 Diabetes
  10. Research Papers and References
  11. Connections

What Gestational Diabetes Is — and Why Its History Is Unusual

Gestational diabetes mellitus is defined as glucose intolerance that is first detected during pregnancy. Pregnancy is naturally a state of rising insulin resistance: the placenta secretes hormones that blunt the mother's response to her own insulin, ensuring a steady supply of glucose crosses to the growing fetus. In most women the pancreas simply makes more insulin to compensate. In gestational diabetes that compensation falls short, blood sugar rises, and the excess glucose crossing the placenta drives the fetus to grow large (macrosomia) and to over-produce its own insulin — the central mechanism behind most of the condition's complications.

The history of GDM is unusual among diseases because, unlike a fever or a tumour that humans have always been able to see, gestational diabetes is largely invisible without a blood test. A woman with mild GDM feels well; the diagnosis lives in laboratory numbers, not in symptoms. As a result, the "discovery" of gestational diabetes was less the discovery of a new illness than the gradual invention of the tools and thresholds needed to name a metabolic state that had presumably always existed. Much of this page is therefore a history of measurement — of glucose tolerance tests, cut-off values, and the studies that justified them.

That measurement story has a sharp historical boundary. Severe diabetes in pregnancy and mild gestational diabetes are two different problems, and the first could not be properly distinguished from the second until medicine had a way to keep severely diabetic women alive long enough to be pregnant at all. That tool arrived in 1922, and everything in the modern history of gestational diabetes flows from it.

Back to Table of Contents

Before Insulin: Diabetes and Pregnancy as a Near-Death Sentence

Before the introduction of insulin in 1922, pregnancy in a woman with established diabetes was a rarity, and when it occurred it was catastrophic. Severe diabetes — what we now recognize as type 1 — typically killed reproductive-age patients within months to a few years of diagnosis, and the disease itself suppressed fertility. Contemporary accounts attributed this infertility to amenorrhoea, atrophy of the ovaries and uterus, and the general wasting of uncontrolled diabetes. The number of recorded diabetic pregnancies in the nineteenth century is correspondingly tiny.

The grim statistics of the era are preserved in the historical literature. The Scottish obstetrician James Matthews Duncan, in an often-cited 1882 report on 22 pregnancies in 16 diabetic women, recorded maternal mortality exceeding 60 percent and fetal (perinatal) mortality of roughly 47 percent. Stillbirth was the dominant cause of fetal death. These figures, summarized in modern historical reviews such as Negrato and Gomes (2013), give a sense of why diabetes and pregnancy together were regarded as one of the most dangerous combinations in all of medicine.

It is important to be clear that these pre-insulin cases were overwhelmingly women with severe, pre-existing diabetes — not the milder, pregnancy-limited condition we now call gestational diabetes. The very concept of a transient, pregnancy-induced glucose intolerance could not come into focus while the field was dominated by the immediate problem of keeping severely diabetic mothers alive. The earliest descriptions, however, already hinted that pregnancy itself could unmask or worsen disordered blood sugar.

Back to Table of Contents

Bennewitz 1824: The First Recorded Diabetic Pregnancy

The case most often cited as the first detailed medical description of diabetes in pregnancy is that of Heinrich Gottlieb Bennewitz, set out in his 1824 doctoral thesis De diabete mellito, submitted at the University of Berlin. The thesis describes a 22-year-old Berlin woman whose symptoms of intense, unquenchable thirst and heavy, sugary urine (glycosuria) appeared during pregnancy, resolved or eased after delivery, and then returned — more severely — in a subsequent pregnancy. One of her infants was reported to be very large, with a weight given in modern summaries as around 5.5 kg.

This pattern — diabetic symptoms surfacing in pregnancy, easing afterward, and recurring with the next pregnancy, accompanied by an oversized baby — is remarkably close to the modern clinical picture of gestational diabetes, which is why Bennewitz's thesis is so frequently invoked as an early case description. The original Latin thesis was translated and analyzed by D. R. Hadden, whose 1989 paper in Diabetologia presents it as "the first recorded case of diabetic pregnancy"; an English translation has been deposited in the library of the Wellcome Institute for the History of Medicine in London.

A note of historical caution is warranted. Whether Bennewitz's patient had what we would now classify as gestational diabetes, or instead had underlying diabetes that worsened in pregnancy, cannot be determined from a nineteenth-century account that predates any blood-glucose measurement; the distinction is, to a degree, a retrospective one. What the case reliably establishes is that careful clinicians recognized a link between pregnancy and disordered sugar metabolism more than a century before the condition was formally defined — a genuinely early observation, presented here as such rather than as a modern diagnosis.

Back to Table of Contents

The Insulin Revolution of 1921–1922

The single most important event in the history of diabetic pregnancy was not a discovery about pregnancy at all. In the summer of 1921, at the University of Toronto, Frederick Banting and Charles Best — working in the laboratory of John Macleod, with the later crucial involvement of biochemist James Collip — succeeded in isolating the pancreatic hormone insulin. On 11 January 1922, a 14-year-old patient named Leonard Thompson became the first person treated with their insulin; after a refined preparation, the effect was dramatic, and a disease that had been uniformly fatal became survivable.

For women with diabetes, insulin did two things at once: it kept them alive, and it restored their fertility. Women who would previously have died young, or who could not conceive because of the metabolic chaos of untreated diabetes, were now able to become pregnant and to survive that pregnancy. Maternal mortality in diabetic pregnancy fell sharply over the following decades. This shift is captured in the title of a well-known historical paper describing insulin's effect on diabetic pregnancy as "a story of two miracles."

The downstream consequence for the history of gestational diabetes was profound. As more and more diabetic and pre-diabetic women carried pregnancies to term, obstetricians accumulated, for the first time, a large population in which to observe how pregnancy and glucose metabolism interact. It became possible to notice women who were not classically diabetic before pregnancy yet showed glucose intolerance during it — and who tended to deliver large babies or to suffer unexplained late stillbirths. The stage was set for the mid-century clinicians who would define that milder pattern.

Back to Table of Contents

Mid-Century Recognition: Hoet, Obstetric Risk, and a New Name

Through the 1940s and 1950s, clinicians increasingly described a category of women whose blood sugar was abnormal during pregnancy but who were not overtly diabetic outside it. The Belgian investigator J. P. Hoet is credited in historical reviews with being among the first, in 1954, to describe the increased obstetrical risk associated with this pregnancy-related glucose intolerance — helping to establish that the milder, transient form was a clinically meaningful entity in its own right, not merely a faint version of classic diabetes.

The term "gestational diabetes" itself entered the medical vocabulary in this period. Historical accounts generally place its acceptance in the 1950s, and some sources specifically credit the term to Carrington and colleagues around 1957. The naming is not attributable to a single undisputed author or a single exact date, and this page deliberately does not credit O'Sullivan with coining the phrase, a misconception that sometimes appears; what O'Sullivan and Mahan supplied was not the name but the first rigorous diagnostic criteria. The honest historical summary is that "gestational diabetes" emerged as an accepted label during the 1950s, with the Carrington attribution noted but not treated here as certain.

By the end of the 1950s, then, the field had a name and a recognized clinical risk, but it still lacked an agreed, statistically grounded way to decide who had the condition. Two different women given a sugary drink would show two different blood-sugar curves; without defined cut-offs, "gestational diabetes" remained a matter of clinical impression. Supplying those cut-offs was the achievement of the next decade.

Back to Table of Contents

O'Sullivan and Mahan, 1964: Defining the Modern Entity

The modern, operational definition of gestational diabetes was established in 1964 by John B. O'Sullivan, a physician working in Boston, together with the statistician Claire M. Mahan. Their paper, "Criteria for the oral glucose tolerance test in pregnancy," published in the journal Diabetes (1964;13:278–285), gave a 100-gram oral glucose tolerance test (OGTT) to a large series of pregnant women — about 752 of them — and, with Mahan's statistical analysis, derived threshold blood-glucose values (set in terms of standard deviations above the mean) for the fasting, one-, two-, and three-hour time points. A woman who met or exceeded two of the thresholds was classified as having gestational diabetes. These were the first statistically based criteria for the upper limit of normal blood sugar in pregnancy, and they anchored diagnosis for roughly the next 40 years.

A historically crucial and often-overlooked point is what the O'Sullivan and Mahan criteria were actually designed to predict. They were not calibrated to identify which pregnancies would suffer complications such as macrosomia or stillbirth. Instead, they were constructed to identify which women were at high risk of developing overt (type 2) diabetes in the years after pregnancy. In other words, the founding definition of gestational diabetes treated pregnancy primarily as a metabolic "stress test" that revealed a woman's long-term diabetic tendency — a framing that still echoes in today's understanding of GDM as a powerful predictor of future type 2 diabetes.

Because O'Sullivan's original thresholds were measured on whole blood using older laboratory methods, they were later re-expressed for plasma and modern assays — most notably the National Diabetes Data Group conversion in 1979 and a further modification by Carpenter and Coustan in 1982. O'Sullivan and Mahan also proposed the now-familiar two-step approach: a 50-gram glucose challenge as an initial screen, followed by the full 100-gram OGTT in women who screened positive. John O'Sullivan's central role is recognized in the diabetes literature, including a Diabetes Care appreciation titled "John B. O'Sullivan: A Pioneer in the Study of Gestational Diabetes."

Back to Table of Contents

Norbert Freinkel and Fuel-Mediated Teratogenesis

If O'Sullivan and Mahan defined how to detect gestational diabetes, Norbert Freinkel did much to explain why it matters for the baby. Freinkel, a diabetes researcher at Northwestern University in Chicago, delivered the American Diabetes Association's prestigious Banting Lecture in 1980 under the title "Of Pregnancy and Progeny" (published in Diabetes, 1980;29:1023–1035). In it he advanced the concept he called fuel-mediated teratogenesis: the idea that the developing fetus is shaped not only by genes but by the mixture of metabolic "fuels" reaching it across the placenta.

Freinkel's insight was that maternal hyperglycemia is only part of the picture. Pregnancy alters the handling of every class of nutrient, and the fetus is exposed to elevated glucose and to altered levels of lipids and amino acids. Excess fuel delivery, he argued, could permanently influence the differentiation and later function of fetal tissues — producing not just an oversized newborn but long-range effects on the child's metabolism, body composition, and risk of disease in later life. This was an early articulation of what later became the broad field of developmental "programming" of adult disease.

Freinkel's framework reframed gestational diabetes from a maternal laboratory abnormality into a question about the lifelong health of the offspring. The annual Norbert Freinkel Award lecture of the American Diabetes Association, named in his honour, continues to mark work in diabetes and pregnancy — a measure of how central his ideas became. His emphasis on the fetal consequences of even modest maternal fuel excess pointed directly toward the question that the next major study would try to answer empirically: does mild maternal hyperglycemia, below the diabetic range, actually harm the baby?

Back to Table of Contents

HAPO 2008 and the IADPSG Criteria

The question of whether mild maternal hyperglycemia matters was answered, on a large scale, by the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study, whose principal results were published in The New England Journal of Medicine in 2008 (HAPO Study Cooperative Research Group, led by Boyd Metzger; 2008;358:1991–2002). HAPO was an international, multi-centre study of roughly 25,000 pregnant women, each given a standardized 75-gram, two-hour OGTT, with the results blinded so that mild abnormalities did not trigger treatment. Researchers then tracked pregnancy outcomes against the mothers' glucose levels.

The central finding was a continuous, graded relationship: as maternal blood glucose rose — even within ranges previously considered normal — so did the risk of important outcomes such as a large-for-gestational-age baby, elevated cord-blood C-peptide (a marker of fetal insulin output), neonatal hypoglycemia, and cesarean delivery. Crucially, there was no obvious threshold below which risk vanished; harm increased smoothly with glucose. This empirically vindicated Freinkel's concern that even modest excess fuel reaching the fetus carries consequences, and it undermined the idea that gestational diabetes is an all-or-nothing diagnosis.

Because risk was continuous, deciding where to draw the diagnostic line became a matter of consensus rather than a natural boundary. In 2010 the International Association of Diabetes and Pregnancy Study Groups (IADPSG) used the HAPO data to propose new diagnostic criteria (Diabetes Care, 2010;33:676–682), based on a single 75-gram, two-hour OGTT with thresholds set at the glucose values associated with a defined increase in adverse-outcome risk: fasting 92 mg/dL, one-hour 180 mg/dL, and two-hour 153 mg/dL, with any one value sufficient for diagnosis. These criteria moved the field toward a "one-step" 75-gram approach. It should be noted, honestly, that the IADPSG criteria were not universally adopted: many centres, including under long-standing guidance in the United States, continued to use the older "two-step" 50-gram-then-100-gram method, and the choice between approaches remains an area of legitimate debate.

Back to Table of Contents

Legacy: Macrosomia, Screening, and the Road to Type 2 Diabetes

Three enduring themes run through the whole history of gestational diabetes, and all three remain central to care today. The first is macrosomia — the large baby. From Bennewitz's 5.5-kilogram infant in 1824 to the large-for-gestational-age findings of HAPO, the oversized newborn has been the most visible signature of maternal hyperglycemia, driven by the fetus over-producing insulin in response to excess maternal glucose. Excess fetal growth raises the risk of birth injury and cesarean delivery, which is why controlling maternal blood sugar in the third trimester is a core goal of treatment.

The second theme is screening. Because gestational diabetes is usually silent, it must be actively looked for. The two-step screen-then-confirm logic that O'Sullivan and Mahan proposed in the 1960s, and the one-step 75-gram test that emerged from HAPO, are both answers to the same problem: how to find a symptomless metabolic abnormality in the window of pregnancy when intervention can still help. Routine glucose testing in the late second trimester is now standard obstetric practice in much of the world, a direct descendant of that 1964 work.

The third theme is the link to future type 2 diabetes. It is fitting that the founding criteria of O'Sullivan and Mahan were designed precisely to flag women at risk of later diabetes, because that prediction has proven robust: a history of gestational diabetes is one of the strongest known risk factors for developing type 2 diabetes in the years and decades after pregnancy. This has turned GDM into something more than an obstetric concern — it is now understood as an early, modifiable warning sign in a woman's lifelong metabolic health, opening a window for prevention. For the practical, present-day clinical picture — diagnosis, blood-sugar targets, diet, and management — see the main Gestational Diabetes article. The thread that connects an 1824 Berlin thesis to a 25,000-woman international study is a single, slowly answered question: what does a mother's blood sugar mean for her child, and for herself?

Back to Table of Contents

Research Papers and References

The references below combine the landmark primary papers in the history of gestational diabetes — given with real DOIs or PMIDs where confidently identified — with curated PubMed topic-search links into the historical and clinical literature. Bennewitz's 1824 thesis De diabete mellito is named in the article as a historical primary source; the modern analysis of it is cited below. Each external link opens in a new tab.

  1. O'Sullivan JB, Mahan CM. Criteria for the oral glucose tolerance test in pregnancy. Diabetes. 1964;13:278–285. — PMID: 14166677
  2. HAPO Study Cooperative Research Group (Metzger BE, et al.). Hyperglycemia and adverse pregnancy outcomes. New England Journal of Medicine. 2008;358(19):1991–2002. — doi:10.1056/NEJMoa0707943
  3. International Association of Diabetes and Pregnancy Study Groups (IADPSG) Consensus Panel. Recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care. 2010;33(3):676–682. — doi:10.2337/dc09-1848
  4. Freinkel N. Banting Lecture 1980: Of pregnancy and progeny. Diabetes. 1980;29(12):1023–1035. — doi:10.2337/diab.29.12.1023 (PMID: 7002669)
  5. Negrato CA, Gomes MB. Historical facts of screening and diagnosing diabetes in pregnancy. Diabetology & Metabolic Syndrome. 2013;5:22. — doi:10.1186/1758-5996-5-22
  6. Hadden DR. The first recorded case of diabetic pregnancy (Bennewitz HG, 1824, University of Berlin). Diabetologia. 1989;32(8):625–626. — doi:10.1007/BF00285339
  7. O'Sullivan JB — A pioneer in the study of gestational diabetes (historical appreciation). Diabetes Care. — PubMed: O'Sullivan pioneer gestational diabetes history
  8. Gestational diabetes: an update 60 years after O'Sullivan and Mahan. Journal of Clinical Endocrinology & Metabolism. 2025. — PMID: 39389786
  9. The discovery of insulin and its impact on diabetic pregnancy (Banting and Best, 1921–1922) — PubMed: insulin discovery and diabetic pregnancy
  10. Hoet JP and the early recognition of obstetric risk in gestational glucose intolerance — PubMed: Hoet gestational diabetes glucose intolerance
  11. Origin and history of the term "gestational diabetes" (Carrington and 1950s usage) — PubMed: history of the term gestational diabetes
  12. Carpenter MW, Coustan DR — criteria for screening tests for gestational diabetes (plasma conversion of O'Sullivan thresholds) — PubMed: Carpenter and Coustan screening criteria
  13. Gestational diabetes and the subsequent risk of type 2 diabetes — PubMed: gestational diabetes and future type 2 diabetes risk
  14. Macrosomia, fetal hyperinsulinemia, and maternal hyperglycemia — PubMed: gestational diabetes macrosomia fetal hyperinsulinemia

External Authoritative Resources

Back to Table of Contents

Connections

Back to Table of Contents