1st trimester miscarriage: four decades of study

摘要

Miscarriage is a very common occurrence in humans. This paper sets out to present published data on research that has provided increased understanding of pregnancy failure. Clarification of definitions, exploring the range of failures from preclinical to later pregnancy losses, and the scientific tools employed to find information on the losses have been documented. What is now understood, which tools work best, and the associated limitations are all discussed. Early studies used cytogenetic methods and tissue culture to obtain results. Improvements in laboratory tools such as better tissue culture incubators, inverted microscopes, laminar flow hoods, improvements in culture media, all contributed to obtaining more results for patients. These studies demonstrated the significant contribution of unbalanced chromosomal karyotypes to pregnancy failure. Maternal age as a contributing factor in trisomy was clearly demonstrated. First trimester miscarriage exhibits very high cytogenetic abnormality; in contrast to very low rates in later losses. Combining data across all time periods of pregnancy will affect the significance of chromosomal error in the early pregnancy failures. Cytogenetic methods investigate whole genomes, and are considered to represent the standard against which new methods must be validated. New molecular genetic methods provide the opportunity to examine samples without the necessity of tissue culture. Techniques may be site-specific or whole genome. Fluorescent in situ hybridisation (FISH), comparative genomic hybridisation (CGH), array-based CGH, single nucleotide polymorphism (SNP) detection, quantitative polymerase chain reaction (qPCR), and quantitative fluorescent PCR (QF-PCR), have all been utilised. In comparison studies with classical/conventional cytogenetics, each newer method offers advantages and limitations. At the present time, a combined approach using conventional and molecular methods will elucidate the cause of miscarriage for almost all samples. In a clinical setting this would be optimum.