Chromosome Abnormalities and Genetic Counseling , 3rd Edition


Insertions are a type of translocation: sometimes the expression “insertional translocation” is used. In the common, simple insertion, three breaks are required. The first two breaks release an interstitial segment of chromosome, which is then inserted into the gap created by the third break. In the interchromosomal insertion, a segment from one chromosome is intercalated into another chromosome; there is no reciprocal exchange. In the intrachromosomal insertion, a segment is intercalated into another part of the same chromosome. The segment may be inserted “right way around”—that is, with the same orientation to the centromere as before; this is a direct insertion (dir ins). Or it may be reversed—an inverted insertion (inv ins). With a very small segment, its orientation toward the centromere may not be distinguishable. Since there is no reciprocal segment involved, the imbalances that result from mis-segregation will be a pure trisomy or a pure monosomy (not mixed trisomy/monosomy, as with reciprocal translocations and inversions).

More complicated scenarios are possible. Wang et al. (1994) describe the single recorded case of a reciprocal exchange of interstitial segments, following a four-break rearrangement. A five-break rearrangement involving the reciprocal movement of two interstitial and two terminal segments (Gibson et al., 1997) could equally well rate mention as a complex rearrangement, and might more appropriately be dealt with in Chapter 11.

In this chapter, we consider the case of the phenotypically normal heterozygote, in whom the rearrangement is assumed to be balanced.

Interchromosomal Insertion


The interchromosomal insertion is the most common form of this uncommon rearrangement: Van Hemel and Eussen (2000) estimate a prevalence on the order of 1 in 80,000. The format of the rearrangement is depicted in Figure 9-1. The recipient chromosome now carries the insertional segment, and the donor chromosome lacks it. (In the ISCN nomenclature, the recipient chromosome is noted first, followed by the donor chromosome.) In theory, two categories of meiotic behavior are possible.


Independent Synapsing of Homologous Pairs

Meiosis could proceed in the usual fashion, with homologs pairing independently as bivalents.

In essence, we can suppose that the insertional segment is disregarded and that the homologs synapse, with segments matching for as much of their length as they are able. The insertional segment could be thrown into a loop1 to accommodate this requirement (Fig. 9-2). (Some crossing-over will presumably occur between synapsed regions, but this would not alter segregation outcomes.) Alternatively, homologs may pair along their full lengths, which would bring some nonmatching segments “incorrectly” alongside each other (heterosynapsis). Then, with normal segregation of the two bivalents, independently of each other, two alternative pairs of gametes are possible. Overall, there would be gametes of four possible segregant types, in the ratio 1:1:1:1—two with a correct amount of genetic material, and two with an incorrect amount. The former two combinations are 46,N and the balanced insertion carrier. The two unbalanced combinations would produce conceptuses with a partial trisomy (duplication) and those with a partial monosomy (deletion) for the insertional segment (Fig. 9-2). It makes no difference whether the insertion is direct or inverted. The foregoing scenario of independent synapsing is more likely to apply when the insertional segment is of small size, as the following cases may exemplify.

Figure 9-1. Formation of an interchromosomal insertion. Single and double asterisks indicate orientation of the inserted segment. The direct insertion has the same orientation to the centromere; the inverted insertion has the opposite orientation.

The viability of the conceptuses—in other words, the risk to the heterozygote of having an abnormal child—depends on the degree of the aneuploid states. Consider the example illustrated in Figure 9-3. A small segment from the middle of chromosome 8 long arm (8q21.2–q22) has been removed and is inserted within the chromosome 10 long arm. This segment comprises about 0.4% of haploid autosomal length (HAL). The heterozygote for this rearrangement could produce two types of unbalanced conceptus: one with a duplication of the segment 8q21.2–q22, and one with this segment deleted. Figure 9-2 depicts these combinations (striped chromosome is no. 8, open chromosome is no. 10, black segment is 8q21.2–q22). In this family (Fig. 9-4), only the duplication was observed. These individuals had mild to moderate mental retardation and minor physical anomalies (Bowen et al., 1983). A segregation analysis of the family was done, and the segregation ratio was close to 1:1:1:0 for normal/balanced/partial trisomy:partial monosomy. This implies a normal viability for the partially trisomic conceptus and complete nonviability for the partially monosomic state. Thus, in this family, the risk for having an aneuploid child is 1/1110, or 33%. This assessment is an example of a private segregation analysis (see Chapter 3).

Figure 9-2. Gamete production following independent pairing of the two sets of homologs. The insertional segment is shown in black, both in its original and in its translocated positions. The horizontal line marks the site from which the segment from the donor (cross-hatched) chromosome came, and the site of its destination on the recipient (open) chromosome.

Figure 9-3. An insertion from chromosome 8 to chromosome 10, ins(10;8)(q21;q21.2q22), showing (a) the balanced carrier, (b) the duplication, and (c) the deletion states. In this family, the duplication was the only unbalanced form to be observed. (Case of P. A. Bowen; Bowen et al., 1983.)

A genetically smaller insertional segment has the potential to be viable in both the duplicated and deleted states. For example, Doheny et al. (1997) describe two first cousins, one with a duplication, the other with a deletion. The connecting relatives carried an insertion, 46,ins(12;10)(q15;q21.2q22.1). The insertional segment, 10q21.2–q22.1, was small, comprising about 0.5% HAL. The child with the duplication was identified with learning difficulty in first grade, and her IQ measured 74; the physical phenotype was rather mild. Her cousin with the deletion had considerable lag in neurodevelopmental progress as an infant, which would lead one to anticipate a more serious mental defect when the child is older, and she had a more obviously dysmorphic appearance.

Figure 9-4. Pedigree of the family in which the insertion illustrated in Figure 9-3 was segregating.

An insertion of a very small segment may be difficult to detect. Löffler et al. (2000) were presented with an adult male suspected to have fragile X syndrome. In the event, he had an abnormal chromosome 14, with additional material at band 14q13. His retarded brother and normal mother had the same chromosome. Was this an insertion, an inversion, or what? FISH using microdissection from the abnormal 14 showed a very small hybridizing segment on chromosome 7, as well as on the whole of chromosome 14. Both no. 7 chromosomes in the brothers showed this spot of hybridization, but just one of the mother's did. Going back to the G-banded preparations, and now knowing exactly where to look, a deletion at 7q32–q34 was discerned on the mother's other no. 7 chromosome, and the definitive interpretation could be made. She had the karyotype 46,XX,ins(14;7)(q13;q32q34), and the two sons were 46,XY,der(14)ins(14;7)(q13; q32q34)mat. In a similar vein, consider the insertion in Figure 9-5, in which two small sub-bands from 2q (2q33.2 and 33.3) and adjoining parts of q33.1 and q34 are inserted into chromosome 4. This is only about 0.3% of HAL. This rearrangement was at the limit of detection of high-resolution G-banding, and required FISH for its clear recognition. In this family, three of five children had a duplication of the insertion, inheriting from the carrier parent the normal no. 2 along with the derivative no. 4 containing the insertional segment (2q33.1–q34). The children with this very short duplication had a clinical picture of poor speech development, distractable and aggressive behavior, and subtle facial dysmorphism. The opposite situation, a parent with a very small insertion having children with the deletion state, is illustrated by Campbell et al. (2002), the mother having the karyotype 46,XX,ins(4;1)(q35;p32.1p32.3). The karyotype of one of the children, originally examined in 1986, had been considered at that time to be normal.

Figure 9-5. A very small insertion, needing FISH to be seen clearly. The karyotype of the carrier parent (upper) is 46,inv ins(4;2)(q32;q34q33.1). The child is duplicated for the segment 2q33.1–q34, but this is difficult to appreciate on the G-banded karyotype (middle). Through FISH, using chromosome in situ suppression (CISS) with chromosome 2–specific paint (lower), the small insertion segment from no. 2 is clearly seen present in the der(4). (Case of M. Curtis.)

Rieger syndrome is a dominantly inherited condition including a particular eye malformation, dental maldevelopment, characteristic facies, and an unusual umbilicus. It is typically due to mutation in the RIEG homeobox gene. Schinzel et al. (1997) studied a boy with Rieger syndrome and severe mental defect who had a deletion in 4q25–q27 and whose normal father carrier a balanced insertional translocation, 46,ins(6;4)(q26;q24q26). Thus, with this boy, it was presumably haploinsufficiency of the RIEG gene that led to the Rieger phenotype. The loss of other genes in the 4q25–q27 segment contributed to a wider phenotype, including the mental defect.

Formation of a Quadrivalent

Probably only in exceptional cases, with larger insertional segments, a quadrivalent forms, and this would enable recombination within the insertional segments. In the review of Van Hemel and Eussen (2000), the mean size of the inserted segment in recombining cases was 1.5% HAL, compared with 1% and 0.5% HAL in non-recombining families in which the imbalances were due, respectively, to duplication and to deletion. With the direct insertion, a recombinant chromosome would be monocentric and therefore functional. Inverted insertions, by contrast, would be associated with dicentric or acentric recombinant chromosomes, with the resulting gametes predicted to be nonviable.

Consider the large direct insertion depicted in Figures 9-6 and 9-7. Most of the material within the chromosome 5 long arm (q11–q22) has been removed and inserted into the distal long arm of chromosome 1. A pachytene configuration at meiosis I such as that depicted in Figure 9-6 would allow for complete synapsis of homologous segments. If no crossover occurred in the insertional loop (and assuming 2:2 disjunction with symmetric segregation ofcentromeres), the same four outcomes noted in the preceding section would eventuate. The gametic combination (a,c) would produce a del(5)(q11q22), and the combination (b,d) would produce a duplication for this segment. But if a crossover did occur, two recombinant chromosomes would be formed, and now three further unbalanced outcomes from symmetric 2:2 disjunction would be possible: gametes(b,d), (b,c), and (a,d) in Figure 9-6. The du-plication/deletion combinations, (b,c) and (a,d), are judged to be nonviable, although they might cause miscarriage. The least imbalanced, least monosomic combination is the dup ins (b,d), which leads to a partial trisomy for the insertional segment, 5q11–q22. This was, in fact, the karyotype of the proposita in this family (Fig. 9-7). Actually, this karyotype endows the same genetic imbalance as would the non-recombinant (b,d) gamete; so in practical terms, it made no difference that this recombination did happen.

Figure 9-6. Gamete production following formation of a quadrivalent in the interchromosomal insertion, with a single crossover having occurred in the insertion loop. Only one of each sister chromatid is shown. Recombinant chromosomes are noted as b′ and d′. Based on the case shown in Figure 9-7.

Figure 9-7. Interchromosomal insertion with recombinant chromosomes in phenotypically abnormal offspring. Partial karyotypes of 46,ins(1;5)(q32;q11q22) carrier parent (upper) and her recombinant child with 46,rec(1)rec(5)dup(5q)ins(1;5)(q32;q11q22) (lower). The latter is the (b′,d′) combination in Figure 9-6. The child is trisomic for the segment 5q11–q22. Cartoon karyotype: open, chromosome 1; criss-cross–hatched, 5q11–q22; cross-hatched, remainder of 5. (Case of P. Jalbert; Jalbert et al., 1975.)

Although the range of abnormality may thus be greater in the case of the heterozygote for a large direct insertion because of the additional risks for gametes having recombinant chromosomes, the outlook is not so discouraging in practice. Often the extent of imbalance associated with a recombinant chromosome is so substantial that nonviability is very likely. In other words, these abnormal pregnancies are lost at any early stage, and do not produce an abnormal child. Each family needs to be assessed individually. The counselor may find it useful to follow the format outlined in Figure 9-6 in making this assessment, in terms of what combinations of imbalanced segments might arise.

As for the inverted insertion, if a quadrivalent did form and recombination happened, an acentric or a dicentric chromosome would result. Such a chromosome, in addition to being genetically unbalanced, would be unstable and compromise the progress of further cell cycles. Thus, the conceptus would be presumed to be nonviable very early post-conception.

Gametogenesis Studies

Gametic analysis has been reported in two insertion heterozygotes. Goldman and Hultén (1992) examined testicular material from an ins(6;7) heterozygote and demonstrated independent synapsis of the no. 6 and no. 7 homologous pairs at diakinesis, with the two bivalents occupying quite separate parts of the nucleus. This is a direct demonstration that the segregation scenario set out in Figure 9-2 does happen. Testicular tissue and sperm were studied from one ins(3;10) carrier in whom a very small segment of no. 10 (p13–p14) was inserted into chromosome 3 at q13.2 (Goldman et al., 1992). In meiosis I, the pairing chromosomes did not loop out the nonhomologous segments, but in fact, the normal chromosome 3 appeared to pair fully with the der(3) and, likewise, the chromosome 10 with the der(10). This may be heterosynapsis. As expected from the theoretical considerations noted above, sperm karyotyping showed similar proportions of gametes with normal, balanced, duplication, and deletion chromosomes: the actual figures were 22%, 32%, 24%, and 22%, respectively. No recombinant forms were seen. Possibly, small insertions may show similar meiotic behavior, with absence of looping out and no quadrivalent formation. Spermatogenesis may be compromised in some carriers: only half as many index cases have carrier fathers as they do carrier mothers (Van Hemel and Eussen, 2000).

Rare Complexities

Most nucleolar organizing region (NOR) translocations are harmless (see p. 240). But a single example exists of an NOR insertion into the X chromosome associated with a familial X-linked spastic paraplegia (Tamagaki et al., 2000). Two brothers and their maternal uncle had the disease, and the carrier mother was unaffected. Plausibly, the inserted material disrupts a “spinal motor neuron gene” in this region, at Xq11.2. It cannot yet be excluded that there is an X-linked Mendelian disorder whose locus resides in Xq11.2, cosegregating in the family by chance, and the NOR insertion is simply serving as a cytogenetic marker. The discovery of the gene would prove the point.

One family with an ins(2;6) enabled some clarification of an imprinting question (Temple et al., 1996). Transient neonatal diabetes (TND) is associated with paternal UPD 6. In a three-generation study, grandmother, father, and daughter had the same unbalanced der(2)ins(2;6)(p22.2;q22.33q23.3), each thus being duplicated for the region 6q22.33–q23.3, but only the daughter had TND. In the father, the duplicated 6q region, being of maternal origin, would have been subject to silencing; while in his daughter it would be active. These observations suggest that the TND locus/region lies within the 6q22–q23 segment. It remains unclear why grandmother was intellectually normal, the father had a mild learning difficulty, and the daughter was microcephalic and retarded.

There can be a link with cancer if a tumor suppressor gene is located in the insertional segment (Barber et al., 1994). An extraordinary case is seen in a father who had had Wilms' tumor as a child and whose daughter had retinoblastoma, due to an insertion that was apparently balanced in him and unbalanced in his child. A segment from 13q14 including the retinoblastoma (RB) gene was inserted into 11p13, this being the site on chromosome 11 of the WT1 Wilms' tumor locus (Punnett et al., 2003).


Insertions are among rearrangements implying the highest reproductive risk. Families like the one in Abuelo et al. (1988), presenting through multiple miscarriage, are exceptional; usually, families come to attention through having had an abnormal child. Pooled data from a number of insertion families (Van Hemel and Eussen, 2000) indicate the average risk of having an abnormal child as being 32% for the male carrier and 36% for the female. It may reach 50%. The risk is greater in the small-segment insertion and smaller in the large-segment insertion. Offering prenatal testing should certainly be the rule. Of the phenotypically normal offspring, approximately half will have normal chromosomes and half will be insertion heterozygotes. A more detailed discussion follows.

For the short insertion (e.g., <1% HAL), the segregation ratio at conception would be expected to be 1:1:1:1 for normal/balanced/du-plication/deletion (meiosis proceeding as described in the first subsection under Biology). If the insertional segment is not only short but also genetically small, both trisomically and monosomically, the maximum risk of having a liveborn aneuploid child would approach 50% (1 + 1/1 + 1 + 1 + 1). The segment 18q11–q21 (HAL = 0.8%), for example, meets these criteria, as seen in the insertion family presented by Chudley et al. (1974). Carriers for this insertion had the four karyotypic classes of off-spring—insertion heterozygotes, karyotypically normal individuals, individuals with a duplication of a small segment of 18q, and individuals with the same segment deleted—in approximately equal numbers. A similar scenario is presented by Marinescu et al. (1999b), with a family segregating an insertion ins(16;5) (q22;p14p15.3). Here, the small segment comprised 5p14–p15.3. In two generations from a heterozygous grandparent, there were two children with 5p, two with 5p, four normals, and three carriers. If viability is reduced or impossible for the trisomic or monosomic conceptuses, the risk would be correspondingly less. Trisomic lethality presumably increases with an increasing fraction of HAL, with monosomic imbalances being more lethal.

It may not be possible to make a clear judgment, based on the literature, about the qualitative content of the imbalance, because the insertion involves an interstitial segment of chromosome, whereas most data on record relate to distal segments. A review of the insertional data on record up to 2000, taken from nearly 90 families, is provided in Van Hemel and Eussen (2000), and Figure 9-8 is taken from their study. Any insertion involving the same open-bar (deletion) or filled-bar (duplication) segment, or a part thereof, will have a significant risk. Schinzel's (2001) cytogenetic database may also be consulted; some more distal insertional segments may well be bounded within terminal duplications and deletions that have been described. An insertion segment of this sort would presumably be at least as viable as these duplications/dele-tions; the phenotype it would produce might be similar or, since the extent of the insertion segment would be less, it could be somewhat less severe. Of course, any unbalanced child in the counselee's family will provide proof of viability and an illustration of that particular phenotype. A study of the wider family may provide a guide to the recurrence risk—a private segregation analysis, as illustrated above in the Biology section. But in any case, the starting point with a patient having a short insertion is that the risk for an abnormal child is high, in the range 10%–50%.

Figure 9-8. Presentation of chromosome segments in which recombinant imbalances have been recorded, in the child of a parent heterozygous for an interchromosomal insertion. Segments seen only as duplications are shown in black bars, those seen only as deletions are in open bars, and filled and open bars connected show segments observed in either state. Insertions seen only in the balanced state are identified with striped bars. (From Van Hemel and Eussen, 2000, courtesy J. O. Van Hemel, and with the permission of Springer-Verlag.)

For the direct insertion involving a longer segment (e.g., <1.5% HAL), there is theoretically an additional risk for the formation of recombinant duplication and deletion chromosomes (an insertion loop may form, as described earlier in the section Formation of a Quadrivalent). But in fact the deletion for a long segment (whether the result of a non-recombinant or recombinant chromosome) would usually impose a nonviable degree of partial monosomy. The dup/del combinations (see Fig. 9-6) are even more unbalanced, leading to spontaneous abortion. Thus, only the duplication (whether non-recombinant or recombinant) is likely to allow for viability. In the great majority of cases, therefore, the segregation ratio for pregnancies going to term is 1:1:x:0 for normal/balanced/partial trisomy/other imbalances, where x is less than 1, and probably very much less than 1. The risk of having an abnormal liveborn may be only a few per cent. In the family of Jalbert et al. (1975) discussed above, the insertional segment 5q11–q22 comprised 2.2% HAL, and this duplication did allow survival, although the child was dysmorphic and severely mentally retarded. This case is the sole example of dup(5)(q11q22) in Schinzel's database. In a family such as that in Abuelo et al.'s (1988) study, with an insertional segment comprising most of 3p (p13–p26, 2.5% HAL), one could be rather confident that any imbalanced conception would miscarry. The closest viable segment in Schinzel's database is 3pter–p14, and there are only two cases of this listed. A risk of probably 0% for an abnormal child could be offered.

Intermediate-length segments (1%–1.5% HAL) might imply a risk in the range 5%–10%. But each segment needs to be judged on its merits, both according to the reproductive history in the family and with reference to the cytogenetic databases.

Intrachromosomal Insertion


Intrachromosomal insertions, also known as “centromere shifts,” are very rare. In their review, Madan and Menko (1992) list only 27 cases. They emphasize that the cytogenetic recognition can be difficult, with 4 out of the 27 having originally been interpreted as paracentric inversions with unbalanced meiotic products, a point subsequently revisited (Madan and Nieuwint, 2002). The formation of the intra-chromosomal insertion is outlined in Figure 9-9. These insertions can be within-arm or be-tween-arm, and direct or inverted, and they may undergo incomplete or complete synapsis. These differences can have practical consequences, and we need to consider each in turn.

Within-Arm Insertion

A shift of chromatin within the same arm is called, logically enough, a “within-arm”2 insertion. Since both segments shift, essentially switching positions, each could be called an “inserted segment.” If both segments maintain the same orientation toward the centromere, it is a direct insertion. If the orientation of one segment is reversed, it is an inverted insertion.3 In the case of the inverted inversion, we can distinguish one segment from the other by referring to respective inverted and noninverted segments. In the direct insertion, the shorter of the two segments can be arbitrarily labeled as the “inserted” segment, and the longer as the “noninserted” or “interstitial” segment (Madan and Menko, 1992; Barber et al., 1994). Since they are both really insertion segments, we can also speak of the “shorter inserted” and the “longer inserted” segments.

Figure 9-9. Formation of the intrachromosomal insertion. Left, the within-arm insertion, with the inserted segments cross-hatched. Right, the between-arm insertion, with the inserted segment in black. Short arrows indicate breakpoints. Compare with the ins(5) shown in Figure 9-13 and the ins(5) in Figure 9-14, respectively.

Between-Arm Insertion

The other type is the between-arm4 insertion, with a segment of chromatin from one arm inserted into a point in the other arm. If we consider the centromere as the fixed reference point of a chromosome, we can regard the centromeric segment as staying still, while the insertion segment shifts from one arm to the other. This somewhat arbitrary point of view allows us to use the term inserted segment unambiguously, in the context of the between-arm insertion.

Incomplete Synapsis

Meiosis perforce proceeds in a modified fashion. Consider the between-arm shift. In most cases, perhaps, the inserted segments fold out so as to allow a good degree of synapsis of the bivalent. This synapsis would include that part of the chromosome between the two inserted segments, that is to say, the centromeric segment. There would be no difference, at least in theory, regardless of whether the insertion is direct or inverted. One (or any odd number) crossover within the centromeric segment will produce recombinant chromosomes: one with a duplication of the insertion segment, and the other with a deletion (Fig. 9-10). The centromeric segment may be quite long, as a proportion of the whole chromosome, and provide considerable opportunity for crossover. Thus, the genetic risk is expected to be high; and in theory would approach 50%. In other words, the segregation ratio for the four possible segregant outcomes of normal/balanced inser-tion/duplication/deletion would be close to 1:1:1:1.

Figure 9-10. Gamete production following a recombination between the sites of rearrangement in the between-arm intrachromosomal insertion. There is incomplete synapsis, with ballooning out. Based on the ins(5) shown in Figure 9-14.

The within-arm shift, in the case of the direct insertion, can have a similar folding out of one inserted segment and its homolog on the normal chromosome to enable synapsis of the other inserted segment and its homologous region. In Figure 9-11, we depict the shorter insertion segment folded out, with synapsis of the larger inserted segment (it could have been drawn the other way around). Recombination within the larger segment will lead, respectively, to duplication and deletion of the shorter segment in the recombinant products passed on to the two resulting gametocytes. Vice versa, if there is synapsis of the shorter inserted segments, followed by recombination, there would be duplication of the larger inserted segment in one gametocyte and deletion of this segment in the other. In theory, the longer the larger segment is, the more likely it is that recombination will happen. One might suppose a lesser risk if both segments are short, possibly making crossing-over less likely, but there are insufficient data to be sure of this. Certainly, cases are on record of crossing-over taking place in very short inserted segments, in both direct and inverted insertions (Webb et al., 1988; Rethoré et al., 1989; Barber et al., 1994). A greater likelihood for crossing-over in segments more distal from the centromere may also affect the risk. Theoretically, if this recombination were to happen with an inverted insertion, dicentric and acentric products, almost certainly nonviable, would result.

Figure 9-11. Gamete production following a recombination within one of the insertion segments (the longer segments) of a direct within-arm intrachromosomal insertion. There is incomplete synapsis. There are four possible gametic outcomes. Compare with the ins(5) shown in Figure 9-13, although note the subtle difference that in the latter the recombination took place between the shorter inserted segments.

Complete Synapsis, Direct Insertion

Alternatively, complete synapsis may be achieved. The insertion and the centromeric segments (between-arm shift) or the two insertion segments (within-arm shift), and their matching segments on the normal homolog, would need to loop back and forth into each other, forming a double loop (Fig. 9-12). Various outcomes are possible from crossing-overs within one or other loop. In the direct between-arm shift, crossing-over within the centromeric segment will lead to recombinant chromosomes deficient or duplicated for the inserted segment (Fig. 9-12, ab). If, however, following complete synapsis, there is crossing-over in the inserted segment, this will lead to the generation of new recombinant forms: chromosomes that are duplicated for terminal p and deleted for terminal q, or vice versa (Fig. 9-12, cd) (e.g., Vekemans and Morichon-Delvallez, 1990).

If complete synapsis is achieved in the direct within-arm shift, there is no new category of recombinant form beyond the four that could be generated from incomplete synapsis with folding out of one of the segments. Cross-ing-over within the longer inserted segment will lead to recombinant chromosomes deficient or duplicated for the shorter inserted segment (Fig. 9-12, ij). Vice versa, crossing-over within the shorter inserted segment will lead to recombinant chromosomes deficient or duplicated for the longer inserted segment (Fig. 9-12, kl). We illustrate such a case from Webb et al. (1988) in Figure 9-13; equally, this outcome could have arisen from incomplete synapsis, with the longer segments folded out.

Complete Synapsis, Inverted Insertion

Recombination in the inverted between-arm insertion, in the setting of complete synapsis, has the same consequences as for the direct insertion discussed above, when crossovers take place within the centromeric segment (Fig. 9-12, ef). The family illustrated in Figure 9-14 demonstrates this. The recombinant child with a dup(5) could equally have arisen from recombination in a partial synapsis (Fig. 9-10) or in a complete synapsis (Fig. 9-12, f), but in either event the cross-over is within the centromeric segment. The duplication comprises the inverted insertion segment. If, however, the crossover is in the insertedsegment, dicentric and acentric products will result, and if a zygote were to result from such a gamete, the compromised conceptus would probably  degenerate very early, and might not even implant (Fig. 9-12, gh). The same fate awaits conceptions from cross-overs in the inverted within-arm shift, if crossing-over happens within the inverted segment (Fig. 9-12, op). If crossing-over is in the noninverted segment, we see the same imbalances (Fig. 9-12, mn) as in the direct within-arm shift (Fig. 9-12, ij). For example, Rethoré et al. (1989) describe a child with a duplication for the very short segment 5p13.32–p14.2 due to a parental inv ins(5)(p13.31p14.3p15.12) with recombination in the even shorter segment p14.3–p15.11, reflecting the scenario set out in either Figure 9-12, n or Figure 9-11.

Figure 9-12. Range of possible recombinants from crossing-over in one or the other insertion loop following complete synapsis of the intrachromosomal insertion. The four panels show, from the top down, direct between-arm insertion, inverted between-arm insertion, direct within-arm insertion, and inverted within-arm insertion. In the loop diagrams, the dots signify the centromere, and the × shows the point of crossover. The insertion segment DE is shown as a thick line in the loop and in the recombinant chromosomes. Circled letters provide reference points for text comments. (Adapted from Madan and Menko, 1992.)

Figure 9-13. Recombination from a direct within-arm shift. Partial karyotypes of an insertion heterozygote mother and her recombinant child are shown. The karyotypes are 46,dir ins(5)(p14.1p14.3p15.1), and 46,rec(5)dup(5p)dir ins(5)(p14.1p14.3p15.1)mat.5 The child is duplicated for 5p14.3–p15.1, shown as the larger cross-hatched segment. The recombination may have arisen from crossing-over within band p14.1 (smaller crosshatched segment) at either partial synapsis with ballooning out of segments p14.3–p15.1, as in Figure 9-11, or from complete synapsis following double loop formation, as in Figure 9-12, k. (Case of L. E. Voullaire; Webb et al., 1988).

The reader may have discerned a pattern in the above construction. Whichever segment recombination takes place in (the active segment), it is the other(passive) segment that comes to be duplicated or deleted. This is logical. A crossover will create a new version of the active segment that contains  a portion from each contributing chromosome, but it will be the same length as it was before. The other, non–crossing-over segments follow, as it were, passively along.

Abnormal Phenotype in the Carrier

If an abnormal phenotype segregates with the insertion, a causal link is possible. Roberts et al. (1986) describe an apparently balanced inv ins(13)(q21.3q32q31) in four members of a family, three of whom had mental defect and psychiatric disorder. A “brain locus” at one of the three breakpoints may have been disrupted or subject to position effect, or there may have been a cryptic imbalance. Alternatively, and given subnormality in some other karyotypically normal relatives, the association could have been coincidental.

Concerning the X chromosome, if the critical region of the X is involved, an insertion may, in spite of being balanced, produce gonadal dysfunction in the female (Grass et al., 1981).


The risk of having an abnormal recombinant child in the 27 families reviewed by Madan and Menko (1992) was 15%, although they considered this quite possibly to be an underestimate. This is an average figure. We may presume a range from near 50% to zero in the individual case. A high risk is likely if one of the segments is small and the other long, so that (1) there is a high survivability in both the duplicated and deleted state for the small segment, and (2) with one long segment, recombination may be more likely to take place. In this situation, a figure of 30%–40% may be the appropriate one to offer. Given that the partial aneuploid states will involve interstitial regions of the chromosome, very little data, quite possibly none, may be on record for the viability and phenotype of the particular segment; and an educated assessment will have to be made.

Figure 9-14. Recombination from an inverted between-arm shift. Partial karyotypes of an insertion heterozygote mother, and her recombinant child are shown. The karyotypes are 46,inv ins(5)(p13q22q33), and 46,rec(5)dup(5q) inv ins(5)(p13q22q33)mat. The child is duplicated for 5q22–q33 (indicated by the cross-hatched segment). The recombination may have arisen from crossing-over anywhere between 5p13 and 5q22 at either partial synapsis with ballooning out of segments 5q22–q33, as in Figure 9-10, or from complete synapsis following double loop formation, as in Figure 9-12, f. (Case of N. J. Martin; Martin et al., 1985).

Risks are presumably less, and possibly zero, if both segments are long (that is, no recombinants are viable). The risks may also be less— say, below 10%—if both segments are short, which might weigh against recombination; but we have no firm data to buttress this suggestion. As always, a private segregation analysis, if the family offers that opportunity, may provide the best estimate of risk. For one specific insertion, Allderdice et al. (1983) calculated a risk of 31% for female inv ins(9)(q22q34.3q34.1) heterozygotes. One short-segment between-arm shift, 46,dir ins(7)(p22.1p21.4q36.1), with a long centromeric segment for which a high risk might have been predicted, in fact produced no liveborn recombinant child in a three-generation family, although some first-and second-trimester pregnancy losses may have been due to unbalanced forms (Farrell and Chow, 1992).


1. This looping out is also described as ballooning out, folding out, or translocation loops.

2. Within-arm insertion is also called intra-arm, paracentric, and intraradial insertion.

3. If both segments were inverted, the result would be indistinguishable from a paracentric inversion.

4. Between–arm insertion is also called inter-arm, pericentric, and extraradial insertion, and centromere shift.

5. This karyotype stretches the limits of the short nomenclature, since “dup p” could refer to either 5p14.1 or 5p14.3–15.1. The full nomenclature is as follows: 46,XX,5,rec(5)(pter–p14.1::p15.1–p14.3::p13.3–qter), dir ins(5)(p14.1p14.3p15.1)mat.