ABO System: Biology and Forensic History

Karl Landsteiner noticed in 1901 that serum from one person could agglutinate the red cells of another, but not always. By systematically testing combinations among six colleagues he identified three patterns, which he labelled A, B, and C (later renamed O). A fourth group, AB, was added in 1902 by Decastello and Sturli. The underlying genetics was worked out over subsequent decades: the ABO locus on chromosome 9 encodes a glycosyltransferase enzyme, and the allele you inherit determines which terminal sugar residue that enzyme attaches to the H antigen on your red cells.

Group A individuals carry an enzyme that attaches N-acetylgalactosamine to the H chain, producing A antigen. Group B individuals attach galactose, producing B antigen. Group AB individuals inherit both functional alleles and carry both sugars. Group O individuals inherit two non-functional alleles and their enzyme adds nothing, leaving the H antigen intact. Because A and B are codominant, the four phenotypes arise from three alleles (IA, IB, and i) with six genotype combinations.

The naturally occurring antibodies that complete the system, called isohemagglutinins, appear in serum from about three to six months of age without any prior red-cell exposure. Their origin is thought to involve cross-reactive responses to environmental antigens on gut bacteria. A person carries antibodies against the ABO antigens they do not themselves possess: group A carries anti-B, group B carries anti-A, group O carries both, and group AB carries neither. This reciprocal relationship is what makes ABO typing work both forward (serum tested against known cells) and back (cells tested against known serum).

ABO group frequencies are not uniform across human populations. They vary substantially across ethnic and geographic groups, and those differences have direct forensic implications: a group B result in a northern European setting implicates roughly 10% of the local population, while in parts of South Asia that same group B spans perhaps 30-35% of the population, roughly three times the discriminating power lost.

These frequency differences mean that the statistical weight of a blood-group match depends on the reference population used. A correct population database was (and remains) essential to any legitimate interpretation. Using a European frequency table for a South Asian population would overstate the significance of a B or AB match by a factor of three or more. This problem predated DNA by decades: the earliest critiques of forensic blood-group evidence in court were often about whether the analyst had used the right population figures.

The first systematic forensic application of ABO typing came in Japan, where serologist Masaichi Kuhara published work on bloodstain grouping in the 1920s. Germany and Scandinavia followed. By the 1930s, several European jurisdictions were routinely submitting blood-group evidence in criminal cases. The technique spread to the United Kingdom and United States through the 1940s and 1950s, aided by the work of serologists such as Leon Lattes in Italy, who developed early absorption techniques for dried stains.

Paternity disputes were an early and steady source of casework. Blood-group testing could exclude an alleged father definitively: if a child's group B allele could only have come from the father and the alleged father was group O (carrying no B allele), the exclusion was absolute. Inclusion was always probabilistic, but exclusion was categorical, and that asymmetry made ABO testing legally reliable in parentage cases long before DNA.

Criminal casework was more complex. A bloodstain typed to group A at a murder scene, with a group A suspect and an group A victim, produced a circular problem the analyst had no genetic way to resolve. The value came when the stain typed to a group different from the victim's, suggesting a second individual was present, or when a suspect could be excluded entirely. Analysts learned to think in terms of contribution to population: what fraction of the local population could have been the stain donor, given this type result?

The 1954 Sam Sheppard case in Ohio is one of the earliest high-profile criminal trials in which forensic blood-group analysis was presented. The bloodstain evidence was incomplete and contested by the standards of even contemporaneous practice. Sheppard was convicted, his conviction was later overturned, and he was acquitted at retrial in 1966. The case is a documented example of both the evidentiary possibilities and the procedural failures that can attend physical evidence handling.

In the United Kingdom, the Metropolitan Police Forensic Science Laboratory developed multi-system blood-group typing through the 1970s and 1980s, layering ABO with secretor status, Rh, MNS, and enzyme polymorphisms to build profiles that could discriminate among a few percent of the population. The 1983 Colin Pitchfork investigation, which eventually became the first case solved by DNA profiling in 1986, was preceded by conventional serology that had already narrowed the suspect class substantially.

DNA short tandem repeat (STR) profiling, introduced to operational casework in the late 1980s and standardised during the 1990s, provided individualising power that blood-group typing could never approach. A full STR profile has a random match probability in the order of one in a billion or smaller; a blood group has a match probability of one in four to one in forty-five, depending on the group. The investigative advantage of DNA was decisive and rapid.

ABO typing survives in specific situations. When a stain is so degraded that DNA extraction yields nothing interpretable, a positive ABO result at least confirms the body-fluid type and narrows the pool. In mass-casualty events and disaster victim identification, ABO typing can provide a rapid triage screen before DNA workflows are applied. Some jurisdictions still include blood-group results in forensic reports as a consistency check, particularly for large stains where the result comes quickly and cheaply. In transfusion medicine and organ transplantation, ABO remains as critical as ever, but that is a clinical context, not a forensic one.

• Heavily degraded stains: when DNA is too fragmented for STR profiling, ABO typing may still be possible from the carbohydrate antigens, which can be more stable than nucleic acids under some degradation conditions.
• Rapid triage screening: in mass-casualty incidents, ABO grouping is fast and cheap and can direct DNA resources toward the most informative samples.
• Consistency checking: a DNA result and a blood-group result from the same stain should agree; a discrepancy flags a possible mixture, substrate interference, or error.
• Historical case review: revisiting older convictions often requires understanding what ABO evidence was presented at original trial, so analysts need to be fluent in the methodology even if they rarely perform it.

The A group is not a single entity. The two most common A subgroups are A1 and A2. A1 individuals produce about five times more A antigen per cell than A2 individuals. Around 20% of group A individuals are A2, while A2B accounts for approximately 10% of group AB individuals. Serologically, A2 cells react weakly with some anti-A reagents. If an analyst uses a reagent that doesn't distinguish, an A2 person could theoretically be misread as group O on a weak stain. Distinguishing A1 from A2 uses lectins such as Dolichos biflorus, which agglutinates A1 but not A2.

Weaker A subgroups such as A3, Ax, Am, and Aend exist at very low frequencies and are mainly clinical curiosities, but they illustrate a broader principle: antigen expression is a quantitative spectrum, not a binary flag. On a dried or degraded stain, reduced antigen density can produce a weak or equivocal typing result. Experienced serologists know to report such results as consistent with a given group rather than as a definitive typing, with documentation of the technical limitations.