In 1941, I published a psychiatric study of 69 same-sexed pairs of twins. The index cases were derived from 8596 consecutive admissions to three mental hospitals and one psychiatric unit of a general hospital in South Sweden, every patient being checked for twin birth in the official birth registers. There was no pair with more than one index case. Upon examination, 21 pairs were considered more or less certainly MZ, and 48 DZ. The degree of certainty of the zygosity diagnosis was expressed by a special formula (Essen-Möller, 1941b).
Of the 21 index cases, 7 (N. 1-7 of the monograph) were at that time judged schizophrenic. The later course revaled that one more index case (N. 12) was a schizophrenic and should have been included into this diagnostic group, which I shall have to concentrate upon in this brief presentation.
Out of the 8 cotwins of the schizophrenic index cases, 5 had presented symptoms of mental disorder up to the time of my investigation, which took place almost thirty years ago. This corresponds so far to a concordance rate of 62%. However, all of the cotwin disorders were relatively mild and transient in nature and consisted mainly of depressive or anxiety states. Even if some of the disordered cotwins had spent some time in a nursing home or in a psychiatric ward of a general hospital, none of them had been admitted to a mental hospital. And, although several of the clinical pictures contained some single trait suggestive of schizophrenia — such as ideas of reference (N. 2 and 7) or hallucinations (N. 5) — in no case a proper diagnosis of schizophrenia could be made. The cotwin who came closest to this diagnosis was a man aged 35 (N. 1), who gave much thought to telepathic phenomena and the like; yet, he was comparatively open-minded and accessible to discussion of his ideas, and he had never been incapacitated by them. Thus, at the time of the first investigation, the correct rate of concordance for schizophrenia in fact was zero. This finding was the more striking as the two other samples published up to that time, those of Luxenburger (1928) and Rosanoff (1934), showed rates of concordance at about 70%, although Luxenburger (1934) later corrected his figure to 33%, for diagnostic reasons. Personally, I was inclined to attribute my zero finding to chance, the number of pairs being small, and also to the relatively short time of observation.