British ff,ottrnal of Plastic Surgery (s97z), 25, 300-3 I4

C L O S U R E OF R H O M B O I D SKIN D E F E C T S : T H E F L A P S
OF L I M B E R G A N D D U F O U R M E N T E L

By G. D. LISTER,M.B., F.R.C.S.(Ed. & Eng.)

and T. GIBSON,D.Sc., M.B., F.R.C.S.(Ed. & Glasg.)
West of Scotland Regional Plastic and Oral Surgery Unit,
Canniesburn Hospital, Bearsden, Glasgow

IN excising the majority of skin lesions, surgeons create an elliptical defect which
can be closed directly. As the lesions increase in size, however, there comes a stage
when either the long axis of the ellipse becomes too long for the local anatomy or
cosmetic result, or the short axis too wide to permit direct suture. In many such
instances the excisional outline may be replanned and closure obtained with a local
flap, until of course the defect becomes so large that cover can only be obtained with
imported skin.
Designing such local flaps, particularly when the secondary defect is to be closed
directly, can tax the skill of the most experienced plastic surgeon. Only after much
trial and error does he learn intuitively to judge the maximum tension under which he
can suture the partly devascularised flap and ~the absolute limits to which he may
stretch the surrounding skin ; the art of designing .such flaps takes long to learn and
is hard to teach.
There are, however, 2 exceptions to which we would like to draw renewed attention :
the flaps designed by Limberg (1946, 1966, and 1967) and by Dufourmentel (1962, 1963)
to close rhomboid defects. A rhombus is an equilateral parallelogram and may be
regarded as a n " ellipse with straight sides ". Rhomboid excision has some advantages
over elliptical excision particularly when the defect is to be closed with a local flap.
For example, less normal skin needs to be excised in the long axis, the design and closure
of a straight-sided flap is far simpler than that of the round flap and rhomboid excision
lends itself to the technique recommended by Borghouts (1964) of histological examina-
tion of the specimen margins to check tumour clearance.
The Limberg flap and the Dufourmentel flap are different in design and application,
and will be described and discussed separately.

THE LIMBERG FLAP

The flap which Limberg designed for a rhomboid defect is one extension of his
classical studies on transposed triangular flaps. Thus in Figures I and 2 the defect is
closed by interchanging the unequal flaps T U V and UVW. The design may also be
regarded as a rhomboid flap XUVW of the same sizc and shape as the defect into which
it is to be rotated.
The flap when used singly is suitable for closure only of rhomboid defects with
angles of 6o ° and 12o ° ; in this paper such a defect is called a 6o ° rhomboid. The
rhombus is thus composed of 2 equilateral triangles and the short axis is the same
length as each side.
In constructing the flap (Fig. I) the short diagonal is extended in one or other
direction by its own length to the point V. A further incision VW parallel to U X and
equal to each of the sides completes the design. It will be evident that the distance
UW is also equal to the short diagonal of the defect and therefore to all other incisions
in the plan. An attractive aspect of this Limberg flap is that, once the length of the short
diagonal has been determined, the remainder of the design may be completed by using
calipers set to that length.
 CLOSURE OF RHOMBOID SKIN DEFECTS ]OI

For any given 6o ° rhomboid defect there are theoretically 4 Limberg flaps available
as shown in Figure 3. Once the most appropriate flap has been chosen and raised it is
rotated through 6o ° and placed in the defect (Fig. 2). It will be seen that U and W
are approximated in dosing the secondary defect. If follows that the ease with which
these points can be apposed determines whether or not the design is appropriate ; this
in turn depends on the relative availability and extensibility of the skin. A major
consideration in planning should therefore be to seek out the direction of maximum
extensibility in the skin around the defect (Gibson et al., I969). This line is usually
at right angles to Langer's lines (Gibson et al., I97I) and on the face it may be regarded
as at right angles to the crease lines. It may be readily determined clinically by picking

T~

U )<

FIG. I. Design of a Limberg flap. S T U X represents a 60 ° rhomboid defect. SU is extended by

its own length to V. VW is then drawn parallel and equal to U X . Note that the distance between
any z adjacent points in the design is identical.

up the skin between finger and thumb. Having found that line, the points U and W
should if possible be placed on it ; they lie on the same line as one of the sides of the
primary defect and 2 flaps can be designed in this fashion for any one 60 ° rhomboid
(Fig. 4, A). If the lesion is circular, the rhomboid can be rotated giving 2 more flaps
with the points U and W in the same line of maximum extensibility (Figs. 4, B and 4, c).
However, when the shape of the lesion already determines the positioning of the rhom-
boid these remarks are not appropriate and one selects the most suitable flap available
for the rhomboid so positioned.
Limberg pointed out that moving any flap involves but 2 processes, opening wound
angles and closing wound angles. Closing a wound angle produces a dog-ear or, in
Limberg's more precise term, a standing cone. Opening an angle in Limberg's paper
models produces a lying cone. This cannot form in skin ; what usually occurs is a
taut, slightly depressed area around the angle. I f for example one were to close the
:rhomboid defect shown in Figure I directly, dog-ears would form at points T and X
302 BRITISH JOURNAL OF PLASTIC SURGERY
and depressed areas at S and U. By adding the rhomboid flap for closure Limberg
avoids any movement of angles T and S. This can be of cosmetic value, particularly
if the deformations around the other angles can be placed in less obtrusive positions.

X S

T X

B C
FIG. 2 . A~ the flap designed in Figure I has been cut and transposed into the defect. B, 6 days
post-operatively. C, 2 weeks post-operatively.

PRACTICAL APPLICATIONS
Excisional Surgery. In most cases it is possible to decide in advance whether a
skin lesion will require elliptical excision and direct closure, excision with local flap
closure or excision and distant skin cover.
Where doubt exists about the practicability of direct closure it is wise to excise
only the lesion with appropriate clearance as the initial step. I f apposition then shows
that direct closure is feasible further excision to create an ellipse suitably long to avoid
 CLOSURE OF RHOMBOID SKIN DEFECTS 303

1 I 2

FIG. 3. In theory a
choice of 4 Limberg
flaps is available for
any 6o ° rhomboid
defect.

3 I 4

. a
LME

V J
V
C

LME
A
\
k

/
A
/ \
FIG. 4. With roughly circular lesions the rhombus should be so positioned that points U and W
(Fig. z) lie on the line of maximum extensibility. There are 2 possible ways in which the rhombus
can be so sited (A and B) and thus 4 flaps of this kind are available (C). This of course is not
applicable if the shape of the lesion dictates the position of the rhombus.
2Q
 30-'~ BRITISH JOURNAL OF PLASTIC SURGERY

FIG. 5" A, Two alterna-

tive Limberg flaps have
been designed. The 2
others theoretically avail-
able would encroach upon
the ear. B, Direct closure
was felt to involve sutur-
ing under undue tension.
C, The more suitable flap
was chosen, cut, transposed
and sutured, (D and E)
with good result.

the formation of permanent dog-ears is performed. If direct closure is found to be

inapplicable no normal tissue has been discarded needlessly and a flap previously
~esigned can be cut and rotated into position (Fig. 5).
At the other extreme of applicability, it may be found that U and W cannot be
FIG. 6. A defect in the form of a parallel-
ogram, having acute angles of 60 ° and long
sides twice that of the short, can be closed
with z Limberg flaps. The 5 possible designs
are shown.
A

B C
FIG. 7. A parallelogram defect created by excision of a rodent ulcer closed by z Limberg flaps. Aw
distortion of the eyebrow or irregularity of the hairline was avoided by the choice of flaps.
306 ~RITISH J O U R N A L OF P L A S T I C SURGERY

approximated even after undermining. The flap must then be replaced and some
alternative means of skin cover found. With a little experience, however, this rarely
Occurs.
The Simultaneous Use of 2 Limberg Flaps. Any defect which can be considered
as a parallelogram the long side of which is twice as long as the short side, and the acute
angles of which are 6o °, may be closed with 2 Limberg flaps.

FIG. 8. A circular defect m a y

be considered as a hexagon.
E a c h side equals t h e radius o f
the or!ginal circle in length.
FIG. 9. All equilateral h e x a g o n s
are m a d e u p o f t h r e e 60 ° r h o m -
boids. T h r e e L i m b e r g flaps can
be c o n s t r u c t e d to close a hexagonal
defect. Preferably the peripheral
limbs s h o u l d point in the same
direction.
FIG. IO. C o n s t r u c t i o n where 2
of t h e flaps h a v e peripheral l i m b s
pointing towards one another.
T h i s m a y be necessary a n a t o m i c -
ally, but the angle closed at A is
12o ° a n d will p r o d u c e a dog-ear.

FIG. 8

FIG. 9 FIG, io

There are 5 possible constructions (Fig. 6) ; the choice will depend on the local
anatomy and the availability of skin. A clinical example is shown in Figure 7.
The Simultaneous Use of 3 Limberg Flaps. A circular defect closely approximates
to the form of a hexagon, the length of the radius equalling the length of each of the 6
sides (Fig. 8). All such equilateral hexagons are made up of three 60 ° rhomboids and
3 Limberg flaps cart therefore be designed to close the defect (Fig. 9). In this design
the calipers are invaluable ; set to the radius of the circular defect I8 distances are
 CLOSURE O1~ RHOMBOID SKIN DEFECTS 307
lsured and marked and the design completed. The flaps should have the peripheral
bs pointing in the same direction as in Figure 9. Only in special circumstances
uld 2 share a common base (as at A in Figure Io) since, when they are rotated into
ition, an angle of I2o ° is closed and a pronounced and probably permanent dog-ear
reared. In the case illustrated in Figure I I, excision and grafting of a radionecrotic
~r on the back had previously been performed, but the graft failed and further necrosis
urred. The ulcer was again excised and successfully closed with 3 Limberg flaps.
On occasion closure of a defect may be quite feasible technically but may have
~lesirable results. In the case shown in Figure I2 the patient had previously had a
flies fan flap rotated at the left angle of his mouth and now presented with active

FIG. I I . A~ A circular radionecrotic ulcer of the back was excised as an equilateral hexagon and 3
Limberg flaps designed as in Figure Io. B, T h e defect excised and the flaps cut. C, Flaps tran-
sposed and sutured with, later, (D) sound healing.

eukoplakia of the opposite commissure. While excision was indicated, it was important
hat no further narrowing of the mouth should result. Three Limberg flaps: one of
vhich was mucosal, gave an acceptable solution. The 2 skin flaps were designed with
Lcommon base, however, and this resulted in the creation of a dog-ear as predicted.
~lthough it cannot be easily appreciated in black and white photographs, the use of the
tap of buccal mucosa served to reconstitute the vermilion in a manner which would
aave been difficult to achieve by other means.
Finger Webs. Moderate degrees of finger webbing are usually treated by release
and free skin graft or by Z-plasty. However, a free graft has certain inherent draw-
backs while a Z-plasty may not introduce sufficient tissue into the line of the web
because of the limitation in length of the central limb. It was noted that, when a web
is divided in the midline (Fig. I3) and the fingers separated, the defect is rhomboid
and further that there is frequently sufficient skin on the sides of the fingers adjacent
FIG. zz. A, Excision of leukoplakia of the commissure is planned. Two skin flaps and (B) one buccal
mucosal flap designed. C, The ~ skin flaps with a common base as in Figure zI has resulted in a noticeable
dog-ear. D, Further narrowing of the stoma has been avoided by reconstructing the vermilion of the
commissure with the buccal mucosal flap.
 CLOSURE OF RHOMBOID SKIN DETECTS 309
O the web to supply a Limberg flap large enough to give good correction, The limiting
hctor is again the amount of skin available, but this is easily assessed in advance.

DISCUSSION
In our experience of over 5o cases in which the Limberg flap has been used, we
have found it safe, reliable and versatile. One great merit of its geometrically precise
design and the fact that only one measurement is needed to construct both the defect
and the flap, is that it is so easily taught. The trainee has the assurance that the blood
supply is adequate and that the edges to be apposed will fit together precisely without
any of the adjustment so often required in placing other flaps ; in other words, the
flap will " work ".
The major limitation of the Limberg flap is that the secondary defect must be
closed directly and flap placement cannot be eased by the use of free grafts on the
secondary defect ; the size depends directly on the availability of skin in the area. A
further limitation, common to all local flaps, is that the correct quality of skin may not
be readily available, for example after excision of lesions of or near the eyelids and close
to hair margins. At the same time because the Limberg flap can be designed in so
many different directions, it is more often possible to construct a flap of the right skin
than with less versatile designs.
The final scar follows a slightly bizarre line which cannot all be lost in crease lines,
but in most cases the quality of the scar has been good without spreading or hyper-
trophy and the result has certainly been better than that obtainable by the simplest
alternative, a full thickness graft. Theoretically the shape of the Limberg flap should
encourage the formation of a raised "trap-door " scar. Whether the increased tension
imposed on the flap prevents this is not known, but we have only seen it on one occasion,
where a rather small flap was used.
If correctly executed and if good primary healing occurs, the Limberg flap produces
results which, for colour and texture match and overall appearance, are excellent.
31o BRITISH JOURNAL OF PLASTIC SURGERY

THE DUFOURMENTEL FLAP

Whereas the Limberg flap can only be constructed for a 6o ° rhomboid defect, the
Dufourmentel flap, " ' le lambeau en L pour losange ' dit ' L L L ' ", can, in theory, be
used for any rhomboid. The rhombus is thus composed of two isosceles triangles and,
unlike the Limberg construction, the short axis of the defect need not equal each of
its sides, The procedure may be regarded again as transposition of two triangular
flaps or more obviously the rotation of an irregular quadrilateral flap into a rhomboid
defect.
In planning the Dufourmentel flap, the short diagonal L N (Fig. I4) and one or
other of the adjacent sides of the defect K N are extended and the angle so formed
bisected by a line (NQ) equal in length to each of the sides of the defect. The final
line (QR) is then drawn parallel to the long axis of the defect and again equals in length
each side of the defect. Calipers are of some use in constructing the flap since all
incisions are of equal length, but a protractor is necessary to place the flap margins
accurately. Once raised, the flap KNQR is transposed into the defect. As shown
in Figure I5, 4 Dufourmentel flaps can be planned for any rhomboid defect. As
the 4 angles of the defect come to equal one another, as the rhomboid approaches the
shape of a square, flaps may be designed at each of the 4 angles giving 8 alternatives
(Fig. I6).
The merit of placing the equivalent points N and R in the line of maximum skin
extensibility has already been discussed in considering the Limberg flap. This is more
complicated with the Dufourmentel design ; indeed, on initial consideration, it appears
to be quite impractical, since the relationship of the baseline of the triangular flap to the
defect varies as the shape of the rhomboid varies. However, as the geometrical analysis
of the design given as an appendix to this paper shows, the angle between the short
axis of the flap (NR) and the short axis of the defect (LN) changes by less than four
degrees from I5 o° as the acute angle of the defect varies from 6o ° to 90 °. This change
may be ignored and, in practice, having determined the line of maximum skin extensi-
bility, points N and R may be placed upon it and the short axis of the defect at an
angle of I5 °° to that line. This applies only to those cases in which the position of
the rhombus is not dictated by the shape of the lesion.
As the acute angle of the defect falls below 6o °, the Dufourmentel flap becomes
progressively wider than the primary defect. Little is to be gained therefore from the
use of the flap in these circumstances and direct closure is to be preferred.
As the acute angle of the defect increases from 6o °, the defect becomes progressively
wider than the flap and when it reaches a square the short axis of the flap is only three-
quarters the length of the short axis of the defect. It is in this range that the flap is of
clinical value.

PRACTICAL APPLICATIONS
As shown above, the Dufourmentel flap has no practical value where the acute
angle is less than 6o °. In closing a 6o ° rhomboid defect, it has no advantage over the
Limberg flap, although it has been argued in its favour that it has a safer blood supply
than the Limberg flap because its base is wider. Embarrassment of the blood supply
of a Limberg flap, however, is rarely seen anywhere on the body and never on the face.
It is in closing defects the acute angle of which lies between 60 ° and 9°0 , therefore,
that the Dufourmentel flap is of use. Such defects: are not often created as additional
excision is usually possible to convert the defect to a 6o ° rhomboid which can be closed
with the simpler Limberg flap. Where, however, such additional excision is contra-
indicated, for example by the proximity of vital structures, then a Dufourmentel flap
 CLOSURE OF R H O M B O I D SKIN DEFECTS 311

ay provide a satisfactory solution. Such a procedure for a defect which could not be
)sed directly without undue tension is shown in Figure 17.
I_

....~ , ~ " .;:,2~

GI R
FIG. 14 FIG. 15

FIG. 14. Design of a D u f o u r -

mentel flap. L M N K represents
a rhomboid defect. L N and
K N are extended and the angle "-~ .t~;~ 7
t h u s formed bisected by a line
N Q equal to each of the sides of
the rhomboid. QR is t h e n drawn
of equal length and parallel to ~ i i ~'~
the long axis of the defect M K .
FIG. 15. I n theory a choice of
4 Dufourmentel flaps is available
for any defect approximating to
a 6o ° rhomboid.
FIG. 16. I f the defect is square
in shape a choice of 8 flaps is
available.
FIG. 16
DISCUSSION

One major merit of the Limberg flap is the simplicity of the design, all angles
being 60 ° or 12o °, all sides being equal. This makes it easy to learn and apply. T h e
Dufourmentel flap is by no means complicated but it is certainly more difficult to
learn and to plan than the Limberg design. Indeed, at first sight it seems that the
Dufourmentel flap with its angles rarely matching those of the defect is inferior to the
Limber~ design, but this is too suoerficial a view. In both instance.~ z defect is
312 BRITISH JOURNAL OF PLASTIC SURGERY

A B

FIG. 17. A, A D u f o u r -
m e n t e l flap h a s been de-
signed for a p l a n n e d
r h o m b o i d excision h a v i n g
an acute angle o f 75 ° . B,
Direct closure w o u l d have
involved s u t u r i n g u n d e r
considerable tension.
FIG. 17. C, T h e flap was
cut, t r a n s p o s e d a n d sut-
u r e d (D) with good result
(E).

E
 CLOSURE OF RHOMBOID SKIN DEFECTS 313

closed and in closing it the surrounding skin is stretched and put under increased
tension ; the angles carefully planned and measured may have changed considerably
by the end of the procedure.
In its range of usefulness between 6o ° and 9o°, the acute angle of the Dufourmentel
flap is always more acute than that of the defect, while the obtuse angle of the flap is
more obtuse than that of the defect. But imagine point L being pulled to point N as
the defect is closed ; the acute angle of the flap would tend to become less acute, the
obtuse angle less obtuse, while the acute angle of the defect becomes more acute and
the obtuse angle more obtuse. In other words the angles of the flap and the defect
bccome much more congruous. Unfortunately this variation in angle size is compli-
cated by the closure of the secondary defect ; approximating point N to point R would
seem to have the opposite cffects on the angles than bringing L to N.
The matter is obviously very complex ; even if we had simple reliable methods
of measuring extensibility of skin at any one site and the effect of varying tensions on
it, the analysis and prediction of the exact behaviour of either flap would require a
computer rather than a plastic surgeon.
The Dufourmentel flap is probably the better design in so far as it takes into con-
sideration the effects of increased tension on the skin, but until we have more precise
knowledge of the latter both designs are still to some extent empirical. It is perhaps
fortunate that the majority of cascs in which the flaps have been used are in the older
age group and old skin absorbs unequal angles, unequal sides, too-far-opened and too-
far-closed angles in a way which makes such flaps, with their various shortcomings,
cosmetically so successful.

APPENDIX
MATHEMATICAL CONSIDERATIONS IN THE DUFOURMENTEL FLAP

As the angles of the defect vary so do the measurements of the flap. The general
trend has been given in the text, but the mathematical variations are readily derived
from Figure 18, A.
Variation of the Acute Angle of the Flap ([3) with that of the Defect (c~). /3 is one
angle of a right-angled triangle the other angle of which is equal to 7/4. Since
7= I8O-~ .'. /3= 45+~/4
This has been plotted on Figure 18, B. For each degree of change in %/3 changes
by one quarter of a degree. They are only equal when both are 60 °.
Variation of the Obtuse Angle of the Flap (3) with that of the Defect (y). It can be
readily seen that 3 = 18o-Y/4 and this relationship is also plotted on Figure 18, B.
With each degree change of 7, ~ varies by one quarter of a degree and the angles are
equal when both are 144 °.
Since/3 fits c~ precisely only when both are 60 ° and 3 fits y only when both are
I44 °, the angles of the flap cannot both equal their respective angles in one construction.
By plotting ~//3 against y/3 in Figure 18, c it is possible to determine that the angle
sizes which give the closest approximation to perfect fit are ~ 5o °, y 13o ° ; /3 58 °, 3 148°.
Variation of the Angle (0) between the Short Axis of the Defect (a) and that of the
Flap (b).
0 can be readily shown to be equal to 18o °-/5/2.
Substituting/3 = 45 ° + "/4 and symplifying
0 = i57½ °- ~/8.
In other words 0 changes at a rate 8 times slower than c~ and in the clinical range
this is negligible.
314 BRITISH JOURNAL OF PLASTIC SURGERY

Variation in Lengths of Short Axes of Flaps and Defect. Since the construction is
made u p of isosceles triangles the same relationship exists between a and b and ~ and
/3. I n other words a and b are equal when a is 60 °. As ~ increases towards 9 °o the
short axis of the defect (a) increases at a rate 4 times as rapid as (b).

&
120

g0

#
40
., 0
2~

, b

Zt~ 40 ~o ~o ~oo 12o Lao

A B

~.~

~ O,~7N FIG. 18. A, See appendix. B,

By plotting the acute angles of
the flap and the defect, ~ and fl
against one another a direct re-
4.0
lationship is shown. A similar
relationship is obtained plotting
the obtuse angles y and ~ against
one another. C, I f the ratio of
0.5 the acute angles one to the other,
~/fl is plotted against the ratio of
the obtuse angles y/8 the value
at which these ratios are equal,
and therefore at which the angles
are most congruent, can be
0.5 4.0 obtained.

C
REFERENCES
BORGHOUTS, J. M. H. M. (1964). Surgical treatment of basal-cell carcinoma and squamous-
cell carcinoma of the skin. Archivum Chirurgicum Neerlandicum, 16, 19-3 o.
DUFOURMENTEL, C. (1962). Le fermeture des pertes de substance cutan6e limit6es " Le
lambeau de rotation en L pour losange " dit " L L L ". Annales de Chirurgie Plastique,
7, 61-66.
DUFOURMENTEL, C. (1963). An L-shaped flap for lozenge-shaped defects. Transactions of
the Third International Congress of Plastic Surgery, p. 772. Amsterdam: Excerpta
Medical Foundation.
GIBSON, T., STARK, H. and EVANS, J. H. (1969). Directional variations in extensibility
of human skin in vivo. Journal of Biomechanics, 2, 2Ol-2O4.
GIBSON, T., STARK, H. and KENEDI, R. M. (1971). The significance of Langer's lines.
Transactions of the Fifth International Congress of Plastic and Reconstructive Surgery,
p. 1213. Melbourne : Butterworths.
LIMBERG, A. A. (1946). Mathematical principles of local plastic procedures on the surface
of the human body. Leningrad : Medgis.
LIMBERG, A. A. (1966). Design of local flaps. In " Modern Trends in Plastic Surgery ",
Second Edition, ed. by Gibson, T. London : Butterworths.
LIMBERG, A. A. (1967). Planimetrie und Stereometrie der Hautplastik. Jena : VEB Gustav
Fischer Verlag.