Quantitation of the Postoperative Vascular Response in Four Dorsal Bipedicle Flaps in the Rat

Poor perfusion and chronic tissue ischemia are common clinical scenarios particularly among older patients. Inciting pathologies ranging from peripheral vascular disease, venous congestion, and diabetes result in a compromised tissue bed with poor wound healing capacity, greater susceptibility to ulceration, and increased rates of infection. Efforts to better understand this ischemic condition have long employed the use of surgical flaps in the rat. In a classic paper, McFarlane et al¹ described a 4-cm x 10-cm cranially based flap designed to produce varying amount of necrosis at the tip. Modified versions of this flap have been used to investigate everything from wound bed interactions to therapeutic anticoagulant therapy with outcomes assessed based on the ratio of surviving and necrotic tissue.^2–8
Most clinical scenarios involving compromised blood flow do not lead directly to necrosis. Bipedicle flaps have been used to induce a less severely ischemic tissue bed without subsequent full-thickness necrosis. This lesser ischemic insult has been used to investigate the induction of the delay phenomena² as well as incisional^9–11and excisional^12–14 wound healing. Quirinia et al^9–11 investigated incisional wound healing in the setting of ischemia using his unique H flap model in the rat. By creating an incision, which transected a bipedicle flap at its center, they were able to test healing wounds created in an ischemic, but not necrotic environment. Similar bipedicle models have been used to test the healing of full-thickness biopsy wounds, with conflicting results. An initial paper by Schwartz et al¹² reported that wounds created within a bipedicle flap showed no signs of healing over 12 days. However, in a similar study by Chen et al,¹³ wound healing curves demonstrated an early delay in the healing of ischemic wounds which appeared to normalize rapidly. In a third study, Gould et al¹⁴ found delayed wound healing in bipedicle flaps 2 cm in width, but not in flaps 2.5 cm wide.
The variability inherent in these results prompted the authors’ lab to attempt to quantify the perfusion dynamics along the flap and over time in 4 distinct bipedicle configurations. Each flap had 2-cm pedicles at either end with 2 flaps oriented vertically parallel to the spine, and 2 transverse across the spine. Traditional rectangular shaped flaps created from 2 straight incisions were tested as well as a modified version of each in which curvilinear incisions were used. Curved incisions resulted in elliptical shaped flaps with substantially increased area, but constant 2-cm pedicles. Flap perfusion was documented serially with a scanning Doppler as well as fluorescein dye distribution. This study attempts to better characterize the dynamic changes in perfusion kinetics based on both the location within and the geometry of 4 distinct bipedicle flaps over time. This study highlights the remarkable resilience against acute insult seen in young healthy laboratory animals as well as the continuing challenges encountered when using them to model chronic clinical conditions, such as ischemia.

Methods

Animals and surgical flap models. Female Sprague-Dawley rats weighing 250–275 g were purchased from Charles River Laboratories/SACSO and cared for in accordance with the Guide for the Care and Use of Laboratory Animals prepared by the Institute for Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication no. 86–23, revised 1985). The University of Virginia’s Animal Care and Use Committee approved all procedures.
Twenty-four rats divided into 4 equal groups were used to examine the postoperative vascular response in 4 bipedicle dorsal flap designs (Figure 1). Group A: vertical straight flap 2 cm x 9 cm (18 cm²); Group B: vertical elliptical flap 2 cm x 9 cm with a central width of 4 cm (32.67 cm²); Group C: horizontal straight flap 6 cm x 2 cm (12 cm²); Group D: horizontal elliptical flap 6 cm x 2 cm with a central width of 4.5 cm (23.13 cm²). Anatomic landmarks were used to position each flap. For vertically oriented flaps A and B, the cranial border of each flap was centered on a line connecting the scapular tips. For flaps C and D, a line connecting the caudal most corner of each pedicle was located 2 cm above the iliac crests. Animals in groups A and B were anesthetized with 50/5 mg/kg intraperitoneal ketamine/xylazine for surgery. After concerns were raised regarding the comparison of baseline Doppler measures taken under injectable anesthetic with postoperative measure to be taken under an inhalational anesthetic, operative anesthetic for groups C and D were changed to 3% isoflurane with 2% supplemental oxygen. Based on repeat baselines, which were all taken under 3% isoflurane and 2% oxygen on a new group of animals, the change in operative anesthetic did not appear to have affected perfusion as measured by the scanning Doppler. Once anesthetized, the animals were clipped and depilated. Using a template, the flap was traced on the dorsal skin with a surgical marker. Baseline scanning laser Doppler measurements were taken and the area scrubbed and prepped with betadine and ethanol. Using sterile technique, 2 full-thickness incisions through the panniculus were made according to each respective flap design, the tissue was undermined, and a sterile impermeable barrier was placed between the flap and the underlying muscle to eliminate the possibility of wound bed support. A nylon barrier was used in group A. In subsequent groups, 2-mm polyethylene sheeting was used because of its increased pliability. The flap was returned to its original position and sutured in place with interrupted 4.0 synthetic absorbable sutures. Based on detailed arteriograms,^15,16 vertically oriented and undermined flaps left intact the deep circumflex iliac and circumflex scapular arteries, while severing the posterior intercostals and thoracodorsal arteries. Transverse flaps transected each of these primary perforating arteries but left undisturbed secondary intercostal perforators arranged in longitudinal rows along the length of the spine. Following postoperative Doppler and fluorescein measurements, animals were bandaged with gauze and cloth tape and were allowed to recover on a warming pad. Subcutaneous Buprenex (0.1 cc) was given for postoperative pain control. The rats where housed in separate cages with free access to food and water. At each subsequent analysis, the animals were anesthetized with 3% isoflurane and 2% oxygen for repeat Doppler and fluorescein measurements. At the conclusion of the study, animals were euthanized under isoflurane with an intracardiac injection of pentobarbital sodium.
Doppler measurements. A Lisca PIM Laser Doppler Perfusion Imager (Linköping, Sweden) was used with the manufacturer’s software package Lisca LDPIwin 2.0 for all Doppler measurements. Serial scanning laser Doppler measurements were taken immediately pre- and postoperatively and under 3% isoflurane with 2% supplemental oxygen on either postoperative days 1, 2, 4, 6, 8, 11, and 14 or 1, 2, 4, 7, 10, and 14 as dictated by surgical scheduling constraints. The area surrounding the flap was masked with black tape to readily identify the flap area on the larger image recorded. Scanning Doppler images consisted of a matrix of 64 x 64 pixels or 2,816 individual readings taking over a square area. Black tape, which outlined the flap and absorbed the red Doppler laser, allowed for the precise distinction between the tissue of the flap, and that lying outside the flap. Preoperative readings of the same tissue area to be incorporated into the flap were used as baseline controls. Doppler scans of each flap were recorded as visual images and text file matrixes of the voltage values representing the perfusion were recorded at each point on the 64 x 64 pixel image. Text files were imported into Microsoft Excel® 2002 and values corresponding to the black tape surrounding the flap were erased. The remaining data was divided to correspond to multiple quadrants along the long axis of each flap (9 quadrants positioned along the 9-cm vertical flaps, and 7 quadrants across the 6-cm horizontal flaps, [Figure 2]). All data points within each quadrant were averaged resulting in an estimation of the relative perfusion in each quadrant of the flap. The vertical ellipse flap trial was terminated at 7 days after failing to demonstrate increased ischemia compared with the vertical straight flap. All other groups were followed for 14 days.
Composite Doppler graphs. Doppler results as processed above yielded a single averaged value of perfusion for each flap quadrant on each animal at every time point. This value was derived from approximately 25 individual readings or pixels on the original Doppler scan. The data across animals was then averaged creating a data set consisting of the spatial positioning within the flap, time in days postoperative, and level of perfusion. This data was graphed using Matlab® 7 (The Mathworks, Natick, Mass) as an interpolated 2 dimensional image with the perfusion values represented by color. All graphs were created to the same color scale (Figure 2).
Fluorescein perfusion measurements. Fluorescein is a nontoxic intravascular dye, which diffuses out of vessels at the capillary level and can be visualized with an ultraviolet lamp to document capillary perfusion. Animals were injected with 0.2–0.4 cc 10% fluorescein dye intravenously (IV) or intraperitoneally (IP). Intraperitoneal injections were used as necessary particularly at later time points after tail vein thrombosis had occluded ready IV access. Photographs were taken under ultraviolet lighting and under normal lighting once the dye evenly perfused the tissue outside the flap. This was generally within 5 minutes for IV injections, but could vary from 20–60 min for IP injections based on differential absorption rates. No difference was noted in the pattern of perfusion between IV and IP injections when sufficient time was given to allow even staining of non-flap skin. Areas that did not stain with fluorescein were considered ischemic. The interval between time points was sufficient for the rats to metabolize the previous dose of fluorescein such that no residual stain was visible in the skin prior to each injection. Fluorescein measures were made at each time point concurrent with Doppler readings with one exception. A preoperative baseline was not taken for fluorescein tests since this test quantifies perfusion as either present or absent and could be expected to demonstrate complete penetration preoperatively.

Results

Doppler imaging. Doppler data was analyzed as previously described depicting both the perfusion dynamically over time and geographically along the length of each flap (Figure 2). Relative perfusion (volts), as measured by the scanning laser Doppler, is represented by the color scale with lower levels of perfusion represented on the blue end of the scale and higher levels of perfusion depicted in orange and red. The horizontal axis of each graph corresponds to the position along each flap based on quadrant. The vertical axis represents change over time with the zero axis set immediately postoperative. Preoperative baseline perfusion levels are represented at the bottom of each graph. These data were not normalized to allow for direct comparison between each flap.
Several key observations were made based on the Doppler results. In all cases, the baseline perfusion levels recorded preoperatively corresponded to relatively low levels of perfusion when compared with levels over the subsequent 2 weeks. Repeat baseline perfusion measurements on a second set of 6 animals confirmed these observations (data not shown, see Discussion). For the purposes of discussion, these measured baselines were considered representative of the basal level of perfusion in uninjured tissue. Ischemia was defined as perfusion levels below baseline, and hyperemia was defined as perfusion levels above this baseline. In the immediate postoperative period, an initial ischemic response is seen centered in each flap. When comparing the vertical straight and elliptical flaps, the zone of ischemia appeared to last approximately 3 days in the straight flap as compared to 1 day in the elliptical flap. The transverse straight flap demonstrated a similar ischemic zone over the first 1–2 postoperative days. By contrast, the transverse ellipse graph showed only a short-lived ischemic zone, which resolved in approximately half the time as its straight counterpart. Following the resolution of these ischemic zones, a pronounced and sustained hyperemic response was found particularly in the first 3 flap configurations. This hyperemic response represented perfusion levels well above baseline, and corresponded to those flap designs that demonstrated the most defined ischemic zone immediately postoperatively.
This ischemic and subsequent hyperemic response is shown in Figure 3. The perfusion values for the central 3 quadrants in each flap were averaged and graphed over time. An acute drop in perfusion postoperatively was followed by a rapid rise in perfusion. This rise went on to produce a marked hyperemic state in the vertical straight and elliptical flaps, as well as the transverse straight flap. Slightly hyperemic perfusion levels were found in the transverse ellipse, but this effect was far less dramatic than that seen in the other 3 configurations.
Contrary to the authors’ original assumptions, elliptical flaps, which include a larger tissue volume supported on the same 2-cm pedicles, did not become more ischemic than their straight flap counterparts. A representation of the entire flap width can be seen over time (Figure 3). Based on these images, elliptical flaps appeared to induce a smaller, more brief, and somewhat less severe zone of ischemia than straight flaps. This difference is more noticeable with the transverse flaps where the elliptical flap induced a less pronounced ischemic zone as well as a muted hyperemic response.
Fluorescein penetration. Fluorescein is a useful diagnostic molecule, which remains intravascular until it reaches the walls of capillary beds. It has been used repeatedly to demarcate viable and nonviable tissue^2,4,6,17,18 in single pedicle flaps by visibly identifying the extent of tissue receiving nutritive capillary flow. Representative images taken immediately after surgery and at postoperative days 2 and 4 show the rapid return of capillary perfusion in each flap (Figure 4). An ischemic band can be seen immediately postoperatively in all but the transverse ellipse flap. This zone of ischemia recovers rapidly, as seen by the return of fluorescein perfusion in this zone in the first few postoperative days. As early as 2 days postoperatively, 55% of flaps showed complete resolution of all ischemic areas with 68% reaching that mark by day 4. All flaps showed a marked reduction in the size of their ischemic zone at each time point with only 1 flap persisting to day 7. The transverse ellipse flaps, which did not create an ischemic zone detectable by fluorescein immediately postoperatively, did not go on to develop an ischemic zone at any subsequent time point.

Discussion

This study failed to demonstrate prolonged ischemia in any of the 4 distinct bipedicle flaps. All flaps demonstrated a rapid return to baseline levels of perfusion. This recovery period lasted approximately 1–3 days, with faster recovery in the shorter transverse flaps and the wider elliptical flaps. The Doppler data concur with fluorescein penetration results, which also demonstrated a rapid return to complete dye penetration of the flap within 4 days of surgery. Both vertical flaps, as well as the transverse straight flap, subsequently demonstrated a marked hyperemic response, while the transverse elliptical flap showed a similar, but far less extreme increase in flow. This hyperemic response was entirely unexpected, but highly consistent across animals. It was hypothesized that the preoperative baseline perfusion measures may have been artificially low, thus accounting for the appearance of a return to supranormal postoperative perfusion levels. During the original trials, initial baseline Doppler scans were performed after clipping, depilating, and rinsing the dorsal skin. As the skin dried, it may have cooled, resulting in capillary shunting and leading to artificially low perfusion measures. To investigate this possibility, the dorsal skin of 6 new rats was similarly prepared and bandaged as a to separate in time the initial surgical preparation from the baseline measures. Repeat baseline Doppler scans were performed following a 48-h delay to allow time for the skin to return to its normal temperature. When compared with original preoperative baselines, repeat scans found no significant difference in mean perfusion levels. The baseline values continued to remain well below the ultimate perfusion levels achieved during postoperative follow up, thus confirming true postoperative hyperemia.
One explanation of this finding is the natural inflammatory response associated with wound healing at the incision sites. This dramatic hyperemic response was not seen in the transverse elliptical model despite similar incisional wounds necessary to create the flap. Alternatively, this hyperemic response may represent a rapid rebound to supranormal levels of perfusion associated with the initial creation of an ischemic zone. This may explain the difference in perfusion dynamics between the transverse ellipse, which failed to induce pronounced ischemic or hyperemic responses, and the 3 remaining designs, which produced brief but identifiable ischemic zones followed by a markedly hyperemic rebound. When taken as a whole, the most acutely ischemic flaps demonstrated the most significant subsequent hyperemic rebound in the immediate postoperative period.
The most significant differences between the transverse elliptical flap and its straight counterpart were equally unexpected. Both flaps were based on identical 2-cm pedicles, 6 cm in total length, undermined, placed atop an impermeable membrane, and closed with interrupted sutures. The main difference between the flaps was their central width and thus, total flap area. This suggests that something other than the pedicle may determine vascular flow into the flap. It is possible that an increased ischemic bed would release higher levels of proangiogenic cytokines, however, it is unlikely that this effect could mitigate the acute ischemia immediately postoperatively and over the first 24-hour period. Rapid vasodilation is the most likely cause of the near immediate recovery seen in the transverse elliptical flap. It is possible that the greater width at the center of the flap allowed the branching vascular network to maintain better connectivity with feeder vessels originating from the pedicle. Perhaps, the relatively narrow straight flap effectively fragmented the vascular network. This may cut off the central vascular supply despite intact dermis, particularly in a network with widely branching and circuitous perfusion paths.
To the best of the authors’ knowledge, this is the first quantitative analysis of the rapidly changing perfusion dynamics within bipedicle flaps. These results may explain some of the variability commonly seen between experiments and laboratories. In particular, these results may shed new light on the ischemic wound healing delay reported by Chen et al.¹³ In their study, wounds were placed within a 2.5 cm x 11 cm bipedicle flap very similar to the vertical straight flap used in the present study. Wound healing trajectories demonstrated an early delay in healing, which appeared to normalize rapidly. Overall delays averaged 4 days. Data from the present study suggest that it may take approximately 3–4 days for the flap to recover baseline levels of perfusion, after which perfusion levels remain elevated for at least 10 days. This correlates well with the healing curves of Chen et al.¹³ It may be that there is a minimum threshold of blood flow below, which healing is impaired, and above which it is roughly normal. In that case, the central zone within the vertical flap may provide an impaired environment for the first 3–4 days. However, a similar study recently published by Gould et al,¹⁴ also examined 2.5-cm x 11-cm bipedicle flaps and failed to find a delay in wound healing under those conditions. The study examined both 2.0-cm and 2.5-cm wide flaps and found delayed healing only in the narrower flap. Similarly, a sustained depression of subcutaneous oxygen tension levels was found in the more narrow flaps, but not the wider ones.
Studies examining the importance of width in a flap have been plagued with inconsistency. In contrast to Gould’s findings, Quirinia et al⁹ found no difference in blood flow when comparing 8-cm long bipedicle flaps with widths ranging 1.0 cm–2.5 cm. Milton¹⁹ was the first to report no association between flap width and surviving length in a classic study using single pedicled flaps in pig skin. Milton concluded, “Flaps made under similar conditions of blood supply survive to the same length regardless of width.” This conclusion assumes a random distribution of vessels. As flaps become smaller, however, the exact branching pattern of the vascular network may become increasingly important. In any tissue, the pattern of arterial and capillary branching would determine the extent to which any flap would leave an intact vascular bed. This is well recognized in association with axial pattern flaps in which a surgeon can be confident that a long and narrow flap based on an identified cutaneous vessel will retain an intact vascular supply. Random pattern flaps, which are fed by the dermal-subdermal plexus in fixed skin animals, may function differently in loose skinned animals. In these animals, vessels course much longer distances within the skin so it is able to slide over the muscle beneath.^15,20
Anatomic studies using vascular castings of the dorsal skin in rats showed a highly branched network of vessels feeding the skin in which the detailed pattern of branching within any angiosome is likely to vary somewhat between animals.^15,16 In each animal, any pair of incisions that divided this vascular network of small unnamed vessels would leave different vessels intact within the flap. This may be an important factor in understanding the divergent perfusion patterns seen in the results of the present study when comparing straight and elliptical flaps. The results, particularly in the case of the transverse flaps, showed a trend toward minimizing the ischemic zone when the elliptical flaps were used. The original prediction that more tissue supplied by the same sized pedicles would become more ischemic was not supported. By contrast, the use of elliptical flaps seemed to attenuate rather than magnify the creation of an ischemic zone. It may be that a narrow flap was more likely to sever important branches in the microvasculature, effectively disconnecting the network, while a flap, which was wider in the center, was more likely to leave an intact vascular network. If the wider elliptical flaps preserved a greater percentage of the vascular bed, the beneficial effects of an intact network may have outweighed the increased metabolic demands created by the larger tissue area. Together, these results underscore the importance of slight changes in flap design on the subsequent perfusion of the tissue bed. The convention of assuming a “random” vascular network exists in the dorsal skin in rats may be an oversimplification that breaks down in the setting of small scale flaps. The exact dimensions of such flaps and the relative size of the animal are likely to remain crucial to the reproducibility of results using bipedicle flaps.

Conclusion

To the best of the authors’ knowledge this is the first study to directly quantitate perfusion over time and position in dynamically changing bipedicle flaps. In addition to the rapid recovery from ischemia seen in this study, a marked and prolonged hyperemic response was clearly demonstrated within a tissue bed otherwise presumed to be ischemic. The internal consistency demonstrated between the Doppler and fluorescein measurements underscore the transient nature of ischemia created within these flaps. It is possible that longer 11-cm flaps, such as those used by Chen et al¹³ and Gould et al,¹⁴ may have provided a more substantial and prolonged ischemic challenge to the tissue. Further work would be needed to investigate the perfusion dynamics of longer flaps. This study also reports the unexpected attenuation of the postoperative ischemic zone when elliptical flaps are compared with relatively narrower straight flaps. This reduction was despite an increased metabolic demand generated by the larger tissue bed and may indicate an unrecognized importance of the finer vascular anatomy in random pattern flaps.
Given the highly transient nature of the ischemia seen in this study, the results highlight the difficulties in designing in-vivo models of chronic ischemia in young, healthy laboratory animals. Their remarkable ability to recover from acute surgical insults may be overcome only through a multifactorial systemic approach, which more closely simulates the complex medical histories of older patients with poor circulation and chronic tissue ischemia.