The Role of Lymphatics in the Pathogenesis of Pneumatosis Experimental Bowel Ischemia
in
By P. Sibbons, L. Spitz, and D. van Velzen London, England and Liverpool, England 6 Twelve spontaneously delivered, normally suckled, smallfor-gestational-age (weighing 756 to 1,213 g) neonatal piglets were used to assess the role of the mesenteric vasculature in the pathogenesis of neonatal necrotizing enterocolitis (NEC) by producing intestinal ischemia. Component vessels (arteries, veins, lymphatics) of the mesenteric vascular arcades were variously occluded by ligation for 48 hours. Nine adjacent vessels of the same type or nine adjacent combinations of vessels were occluded in piglets 12 to 18 hours postpartum. Arterial plus lymphatic ligation induced lesions showing the complete histopathological spectrum of NEC (mucosal stripping, hemorrhage, submucosal disruption and destruction, full-thickness necrosis, inflammatory infiltration) including pneumatosis intestinalis. Two of the lowest birth weight animals produced complete NEC in response to lymphatic ligation alone. A condition consistent with “prepneumatosis” was found when lymphatics only were ligated. The distended lymphatic vessels in the submucosa resembled
pneumatosis
with
reference
to shape size and
distribution, but contained milk-derived lipids, some proteins and lymph but no gas. Arterial ligation alone induced NEC-like lesions without pneumatosis. Venous ligation alone induced minor congestive/ hemorrhagic lesions. Pneumatosis appears to originate in the lymphatic vessels of the submucosa in this experimental model of NEC. Lymphatic occlusion alone can cause complete NEC in very SGA neonatal piglets. Arterial plus lymphatic occlusion produces a unique combination of specific pathology resembling human NEC. Copyright
Q 1992 by WA
Saunders Company
INDEX WORDS: Necrotizing enterocolitis; tinalis.
Hemodynamic flow studies using chromium 51labeled autologous erythrocytes showed that the blood flow in the distal ileum was significantly reduced compared with the proximal ileum in both normal birth weight (NBW) and SGA piglets. The reduction was more pronounced in the SGA animals who, in addition, were unable to produce any compensatory collateral flow in the distal ileum in response to the occlusion of vessels in the mesenteric arcades.” Direct ligation of groups of vessels (1 group = artery plus vein plus I or 2 lymphatics) in the mesenteric arcades supplying the distal ileum in SGA piglets similarly induced lesions comparable to NEC including pneumatosis.’ The severity of the pathological process increased with proximity to the ileocecal region. Associated with many of the lesions induced by vascular occlusion was the marked lymphatic distension that assumed the shape and distribution of clinical histopathological pneumatosis but contained lymph and cellular debris and not gas. We termed this feature “prepneumatosis.” The aim of this study was to determine the role of the various vascular components in the mesentery of the ileum in the pathogenesis of NEC and its associated pneumatosis intestinalis. MATERIALS AND METHODS
pneumatosis intes-
Animals
NEUMATOSIS intestinalis is the single most important hallmark in the radiological diagnosis of neonatal necrotizing enterocolitis (NEC). However, speculation remains concerning the precise origin of the gas within the bowel wall in this condition. Analysis of the gas aspirated during surgical exploration has shown that it contains 30% hydrogen, a gas not normally present in the neonatal intestine.’ It has been suggested that the gas is produced by facultative anaerobic bacteria that invade the bowel wall through breaches in the mucosal layer secondary to an ischemic incident.2.’ We have previously demonstrated that NEC can be reliably induced experimentally in small-for-gestational-age (SGA) neonatal piglets exposed to a combination of hypoxia to a PO, of 30 mm Hg and hyperviscosity to a hematocrit in excess of 75%. The lesions induced showed a marked predeliction for the distal ileum and proximal co1on.4
P
Journalof Pediatric Surgery, Vol 27, No 3 (March), 1992: pp 339-343
Twelve SGA (weighing 756 to 1,213 g) (NBW for piglets is approximately 2,000 g), spontaneously delivered, full-term neonatal piglets formed the experimental group. The piglets were permitted to suckle as normal and kept in a controlled environment of 20°C with 50% humidity until removal from the litter 12 to 18 hours after birth.
Procedure Under general inhalation anesthesia with halothane over oxygen and nitrous oxide, and using aseptic technique, ligation of various combinations of component vessels in the mesenteric vascular
From the Department of Paediatric Surgery, Institute of Child Health, London, England, and the Depanrnent of Pathology. Alder Hey Children’s Hospital, Liverpool, England. Presented at the 22nd Annual Meeting of the American Pediatric SurgicalAssociation, Lake Buena vista, Florida, May IS-l& 1991. Address reprint requests to P. Sibbons, Department of Paediatric Surgery, Institute of Child Health, London WClN IEH, England. Copyright o I992 by W.B. Saunders Company 0022-3468192/2703-0013$03.00/O 339
SIBBONS, SPITZ, AND VAN VELZEN
Proximal No
2 1
Position
>
6
5
4
3
Art + Vn
Lvm
Vn + Lym
Vn
Lym
Vn + Lym
3
_
v+n
>
5
6
Art + Lym
1
7 Vn 6
Lym v+n
Vn +
Art
Lym
1
11 Lym 12
Art +
Vn
Vn
Art + Lym
Vn +
Vn
Lym
Art
Art + Vn
Lym
Art
Art t Vn
Art +
Vn
Lym
Lym \
Art
Art + Vn
Lym Vn +
A;’
1
Lym
Lym
r
9 10
Art Art
Art +
2
Lym
Art Art
4
RESULTS
Distal
Animal
Vn + Lym
Y 1
sequence
Fig 1. Sequence of ligations for each animal. Art, arteries alone; Vn, veins alone; Lym, lymphatics elone; Art + Lym, arteries + lymphatics; Vn + Lym, veins + lymphatics; Art + Vn, arteries + veins.
arcades was carried out aided by the use of an operating microscope (magnification x8). Commencing at the ileocecal junction and continuing proximally along the ileum, nine adjacent groups of vessels were selected for ligation separated by 10 intact groups of vessels. The ligation procedures consisted oh (1) artery plus vein (Art + Vn); (2) lymphatics alone (Lym); (3) vein plus lymphatic (Vn + Lym); (4) vein alone (Vn); (5) artery plus lymphatic (Art + Lym); and (6) artery alone (Art). The ligations were performed according to a strict protocol for subsequent analysis (Fig 1) of three sequences comprising six ligations each and occupying the distal 25 cm of ileum. The ligations were performed midway between the serosal surface of the intestine and the mesenteric lymph node. The animals were returned to the sow and littermates on recovery from anesthetic. All these piglets were killed 48 hours postoperatively. At necropsy, the distal ileum together with the associated mesenteric vascular arcades were removed. Tissue for histopathological examination was taken from the center of each bowel segment supplied by the ligated vessels, fixed in formal-calcium for 16 hours, and processed to paraffin wax embedding by routine schedules. Microscopic sections were cut at 4 pm and stained with hematoxylin and eosin.
The histopathological scores for each of the ligation procedures according to position in the distal ileum are shown diagrammatically in Fig 2. Arterial ligation resulted in vascular congestion with severe mucosal stripping and submucosa1 disruption. Necrosis affecting the mucosa and submucosa was often present and in some cases the muscularis was affected. There were only minor degrees of lymphatic engorgement (Fig 3). Ligation of lymphatic vessels induced gross lymphatic engorgement that assumed the typical pattern and distribution of pneumatosis but did not contain gas (prepneumatosis) (Fig 4). This was most often seen in the submucosa but in some cases on the serosal surface as well. This entity of prepneumatosis was found only when lymphatic vessels, alone or in combination, were included in the ligation procedure (Table 1). The complete picture of NEC developed in two of the lowest birth weight piglets who underwent lymphatic ligation only. The combination of arterial and lymphatic ligation induced the most severe pathology. This consisted of complete NEC in all the distal ligation positions and
I
0 .&a
Microscopic Grading Stained sections were evaluated by three separate observers using a histopathological severity score system.4 The system comprised the grading from 1 to 5 of the histopathological severity of defined tissue changes such as vascular congestion, mucosal stripping, necrosis, ulceration, perforation, prepneumatosis, and pneumatosis intestinalis for each lesion. The mean lesion grade was calculated as histopathological severity score for each ligation procedure.
Art
Art+ Lym 40
Vn
BE 00 !
*\ 20
sequences Fig 2. Total pathology score by sequence for each animal according to the type of vessels ligated. Prox, proximal sequence; mid, middle sequence; dist, distal sequence.
LYMPHATICS AND PNEUMATOSIS
341
IN BOWEL ISCHEMIA
Table 1. Presence of Prepneumatosis
in Each Animal According to
the Types of Vessels Ligated Animal NO.
1
Type of Vessels Ligated
Birth Weight(g)
Art
Art+
Lym
Vn
Vn+Lym
Lw
+
2
792
3
1,013
4
1,213
5
894
6
1,131
7
981
t
+
8
940
t
t
9
756
+
10
1,160
11
853
12
997
+ +
Art+Vn
+ + + +
922
Bled +
t Bled
t
+ +
*AbbreviaGons: Art, artery; Lym, lymphatics; Vn, vein. Fig 3. Arterial ligation lesion (H&E, original magnification x40) showing vascular congestion [a}, submucosal disruption (b), some lymphatic distension (cl, and almost full-thickness necrosis.
pneumatosis in 7 of 12 animals in the midileal positions (Fig 5). All animals except one developed moderate prepneumatosis. It would appear that the lesions produced by arterial plus lymphatic ligation amounted to more than merely the sum of each combined. The mean difference in total histopathological severity scores for arterial plus lymphatic ligation was significant at all positions compared with the other five ligation procedures (P I .Ol). DISCUSSION
Although the etiology and precise pathogenesis of neonatal NEC remains open to debate, the most accepted final common pathway to the development of the condition is reduced mucosal blood flow. The
Fig 4. Typical lymphatic ligation lesion (H&E, original magnification x40) showing prepneumatosis (distended lymphatics with the shape and distribution of PI) (c).
area of intestine most frequently affected is the distal ileum and proximal colon. We have previously shown that the blood flow in the distal ileum of neonatal piglets is significantly reduced compared with more proximal ileum. The reduction was greatest in SGA piglets.’ In response to occlusion of vessels in the mesenteric arcades, compensatory collateral flow developed in NBW piglets in both the proximal and distal ileum, whereas in the SGA animals this response failed to occur and at 24 hours postligation blood flow in the distal ileum had further reduced.’ The experiments reported herein were carried out to examine the role of the various component vessels in the mesenteric arcades, singly or in combination, in the induction of the NEC-like lesions. Varying degrees of histopathological changes occurred with the different types of ligation procedures and, with the exception of lymphatic alone, there was a noticeable progression of lesions with proximity to the ileocecal junction.
Fig 5. Lesion produced by arterial plus lymphatic ligation (H&E, original magnification x40) showing full-thickness necrosis, total mucosal stripping, severe submucosal disruption, prepneumatosis (cl, and pneumatosis (d).
SIBBONS. SPITZ, AND VAN VELZEN
342
Arterial plus lymphatic ligation resulted in the most severe damage, which amounted to more than the sum of each ligation procedure alone. All the histopathological features of complete NEC were present in all the distal ligation sequences. Pneumatosis was observed in all these as well as in the middle sequence and moderate to severe prepneumatosis was found at all except one ligation positions. Two piglets undergoing lymphatic ligation only developed complete NEC. These were two of the lowest birth weight animals (weighing 853 g and 894 g). No previous work has been carried out specifically directed at experimental occlusion of the lymphatic vessels in the mesentery of newborn animals. Prepneumatosis and pneumatosis developed only when the ligation procedure included the lymphatic vessels either singly or in combination with the adjacent blood vessels. The distended lymphatics in the submucosa assumed the shape and distribution of typical pneumatosis. We postulate that anaerobic bacteria gain access to the distended lymphatics, which contain an ideal substrate for proliferation and produce the hydrogen gas that results in the typical pneumatosis seen in neonatal NEC. Shut down of the mesenteric lymph node by bacterial translocation would further exacerbate the lymphatic obstruction and cause a progression of the disease process. Jackson et al7 clearly demonstrated that bacterial translocation to the mesenteric lymph node occurred in neonatal rabbits whose alimentary tracts were inoculated with Escherichia coli.
Although the experimental animals in this study were not hemodynamically monitored during the operative procedures, all interligation “control” tissue was histopathologically normal. Therefore, we have no basis for implicating any perioperative stress factor in the pathogenesis of the lesions induced. In the light of these observations, we hypothesize that, in susceptible animals, NEC is initiated by an incident or combination of incidents that cause a reduction of arterial inflow to specific areas of the intestine, particularly the terminal ileum and proximal colon. This is followed by mucosal ischemia, which leads to mucosal breach and the creation of a microenvironment conducive to bacterial proliferation and invasion, particularly of the lymphatics. It is possible that translocation of bacteria via the lymphatits presents an overwhelming antigenic challenge to the local mesenteric lymph node resulting in lymph node shut down.’ The subsequent distension of lymphatics proximal to the lymph-node constitute prepneumatosis in the bowel wall. The stagnant material in the distended lymphatics (high molecular weight milk fats and some proteins) provides an excellent substrate for the production of gas by proliferating facultative anaerobes. This gas is radiologically and histopathologically observed as pneumatosis intestinalis. The fat/protein substrate is also a good medium for lipid peroxidation and the generation of oxygen free radicals, damage from which may be integral to the pathogenesis of NEC.
REFERENCES 1. Engel RR, Virnig NL, Hunt CE, et al: Origin of mural gas in necrotising enterocolitis. Pediatr Res 7:292-296,1973 2. Smith MF, Boriello SP, Clayden GS, et al: Clinical and bacteriological findings in necrotising enterocolitis: A controlled study. J Infect 2:23-31, 1980 3. Popoff MR: Are anaerobes involved in neonatal necrotising enterocolitis?, in Boriello SP (ed): Clinical and Molecular Aspects of Anaerobes. Petersfield, England, Wrightson Biomedical, 1990, pp 49-57 4. Sibbons P, van Velzen D, Spitz L, et al: Relationship of birth-weight to the pathogenesis of necrotising enterocolitis in the neonatal piglet. Pediatr Path01 8:151-162, 1988
5. Sibbons PD, Spitz L, van Velzen D: Collateral blood flow in the distal ileum of neonatal piglets. A clue to the pathogenesis of necrotising enterocolitis. Pediatr Path01 (in press) 6. Sibbons PD, Spitz L, van Velzen D, et al: Necrotising enterocolitis induced by local circulatory interruption in the ileum of neonatal piglets, in Boriello SP (ed): Clinical and Molecular Aspects of Anaerobes. Petersfield, England, Wrightson Biomedical, 1990, pp 59-67 7. Jackson RJ, Smith SD, Widowsky RM, et al: The effect of E. coli virulence on bacterial translocation and systemic sepsis in the neonatal rabbit model. J Pediatr Surg 26:483-486, 1991 8. McConnell I: The Immune System fed 2). London, England, Mosby, 1981, pp 181-192
Discussion M.Z. Rowe (Pittsburgh PA): This is an important question. It is interesting if you look at the literature of necrotizing enterocolitis; there are really very few experimental studies of what pneumatosis is really about, about how it really develops, and I compliment the authors on their determination to try to get some
more insights into the problem. In listening to the papers this morning, it begs a question I think 1’11 mention, and the question is, a lot of research and a lot of questions have been centered on the whole idea of blood flow to the intestine as one of the pathogenic factors of NEC, yet there is beginning to be devel-
LYMPHATICS
AND PNEUMATOSIS
IN BOWEL ISCHEMIA
oped a huge amount of data that suggest that maybe vascular compromise might not have anything to do with necrotizing enterocolitis. For example, as Dr Langer mentioned, the match control studies have over and over again shown that vascular compromise, perinatal shock, patent ductus arteriosus, and umbilical catheters are just as common in match controls as in NEC and these events simply define a vulnerable population rather then suggest an etiologic agent. The study of steroids in a multiinstitutional study showed that children who were born after their mothers got steroids for lung maturity had a significant lower incidence of NEC than did the controls and it was found that these children had an increase in their gut maturity. It was interesting that in the vitamin-E study, again a multiinstitution study in which they were looking at prevention of retrolental fibroplasia, it was shown strikingly that this interfered with white cell function of these babies and there was an increased incidence of NEC, which suggested an infection etiology. And, of course, we are all familiar with the extremely interesting new studies that seem to suggest that prophylactically giving IgA, at least in Vienna, significantly reduces NEC. And then we also note from our own interests that there are clusters of NECs that suggest not perfusion but an infectious origin, so I think this whole question is difficult, but specifically as relates to infection, there are several specific questions that should be addressed. One, what is a low birth weight pig-is that a premature pig or is it small for gestational age or a runt? These are different because a low birth weight is just a small pig. A premature is immature, while a small for gestational age might show intrauterine malnutrition, and I think has different physiologic effects. There are no data that I know that suggest translocation of the bacteria into the lymph nodes will block the lymphatits, so I am confused on how lymphatic obstruction could be an etiologic source in NEC. I am not sure just how that could take place. Just bacteria in the lymphatics usually does not block the lymphatics. If the lymphatics are the source of pneumatosis, do you see air in the lymph nodes since these are bacteria that do form gases?
343
A. Kosloske (Baltimore, MD): I have a couple of questions for Mr Spitz. First, about the bacteria. How much bacterial colonization has occurred at 12 to 18 hours in the neonatal piglet? Are they more like the germ-free animals or are they heavily colonized by that time? I wondered whether you did qualitative cultures or even quantitative cultures that would very helpful in sorting this out. You also wonder what is happening in the lymphatics. As I listen to this, I thought that perhaps rather than the lymphatics being a place where the bacteria stop and proliferate in this model, that by ligating the lymphatics you have artificially created a block to the run-off of bacterial toxins which then goes on and leads to the formation of pneumatosis intestinalis. A.A. de Lorimier (San Francisco, CA): Have you looked at the histology of your human newborn infants with NEC to see evidence of lymphatic distention? L. Spitz (response): In answer to Marc Rowe’s question, these animals were small for gestational age, they were the runts of the group. There were 10 delivered spontaneously. The normal birth weights were the full-grown animals in the same litter. So I think there is a difference there. I think the reason we are getting the lesions we see in these animals is that the mucosa is more susceptible in these small for gestational animals and they are particularly the ones we see in clinical practices as being more prone to develop NEC. I would agree that the mesenteric lymph node blockage is mere speculation on our part. We haven’t probed into this, but it is interesting to note that all the animals who had lymphatic obstruction were the only ones who developed this pneumatosis. In answer to Dr Kosloske’s questions, we have not looked at the bacteria specifically, That study is still in progress and I can’t give you any results of those studies as yet, but these animals were allowed to suckle normally and presumably the intestine was colonized by the time we experimented on them 12 to 18 hours later. We have looked at the histological studies in the human NEC and examined the tissue carefully; they appear to have free pneumatosis as well as pneumatosis lesions.
in
By P. Sibbons, L. Spitz, and D. van Velzen London, England and Liverpool, England 6 Twelve spontaneously delivered, normally suckled, smallfor-gestational-age (weighing 756 to 1,213 g) neonatal piglets were used to assess the role of the mesenteric vasculature in the pathogenesis of neonatal necrotizing enterocolitis (NEC) by producing intestinal ischemia. Component vessels (arteries, veins, lymphatics) of the mesenteric vascular arcades were variously occluded by ligation for 48 hours. Nine adjacent vessels of the same type or nine adjacent combinations of vessels were occluded in piglets 12 to 18 hours postpartum. Arterial plus lymphatic ligation induced lesions showing the complete histopathological spectrum of NEC (mucosal stripping, hemorrhage, submucosal disruption and destruction, full-thickness necrosis, inflammatory infiltration) including pneumatosis intestinalis. Two of the lowest birth weight animals produced complete NEC in response to lymphatic ligation alone. A condition consistent with “prepneumatosis” was found when lymphatics only were ligated. The distended lymphatic vessels in the submucosa resembled
pneumatosis
with
reference
to shape size and
distribution, but contained milk-derived lipids, some proteins and lymph but no gas. Arterial ligation alone induced NEC-like lesions without pneumatosis. Venous ligation alone induced minor congestive/ hemorrhagic lesions. Pneumatosis appears to originate in the lymphatic vessels of the submucosa in this experimental model of NEC. Lymphatic occlusion alone can cause complete NEC in very SGA neonatal piglets. Arterial plus lymphatic occlusion produces a unique combination of specific pathology resembling human NEC. Copyright
Q 1992 by WA
Saunders Company
INDEX WORDS: Necrotizing enterocolitis; tinalis.
Hemodynamic flow studies using chromium 51labeled autologous erythrocytes showed that the blood flow in the distal ileum was significantly reduced compared with the proximal ileum in both normal birth weight (NBW) and SGA piglets. The reduction was more pronounced in the SGA animals who, in addition, were unable to produce any compensatory collateral flow in the distal ileum in response to the occlusion of vessels in the mesenteric arcades.” Direct ligation of groups of vessels (1 group = artery plus vein plus I or 2 lymphatics) in the mesenteric arcades supplying the distal ileum in SGA piglets similarly induced lesions comparable to NEC including pneumatosis.’ The severity of the pathological process increased with proximity to the ileocecal region. Associated with many of the lesions induced by vascular occlusion was the marked lymphatic distension that assumed the shape and distribution of clinical histopathological pneumatosis but contained lymph and cellular debris and not gas. We termed this feature “prepneumatosis.” The aim of this study was to determine the role of the various vascular components in the mesentery of the ileum in the pathogenesis of NEC and its associated pneumatosis intestinalis. MATERIALS AND METHODS
pneumatosis intes-
Animals
NEUMATOSIS intestinalis is the single most important hallmark in the radiological diagnosis of neonatal necrotizing enterocolitis (NEC). However, speculation remains concerning the precise origin of the gas within the bowel wall in this condition. Analysis of the gas aspirated during surgical exploration has shown that it contains 30% hydrogen, a gas not normally present in the neonatal intestine.’ It has been suggested that the gas is produced by facultative anaerobic bacteria that invade the bowel wall through breaches in the mucosal layer secondary to an ischemic incident.2.’ We have previously demonstrated that NEC can be reliably induced experimentally in small-for-gestational-age (SGA) neonatal piglets exposed to a combination of hypoxia to a PO, of 30 mm Hg and hyperviscosity to a hematocrit in excess of 75%. The lesions induced showed a marked predeliction for the distal ileum and proximal co1on.4
P
Journalof Pediatric Surgery, Vol 27, No 3 (March), 1992: pp 339-343
Twelve SGA (weighing 756 to 1,213 g) (NBW for piglets is approximately 2,000 g), spontaneously delivered, full-term neonatal piglets formed the experimental group. The piglets were permitted to suckle as normal and kept in a controlled environment of 20°C with 50% humidity until removal from the litter 12 to 18 hours after birth.
Procedure Under general inhalation anesthesia with halothane over oxygen and nitrous oxide, and using aseptic technique, ligation of various combinations of component vessels in the mesenteric vascular
From the Department of Paediatric Surgery, Institute of Child Health, London, England, and the Depanrnent of Pathology. Alder Hey Children’s Hospital, Liverpool, England. Presented at the 22nd Annual Meeting of the American Pediatric SurgicalAssociation, Lake Buena vista, Florida, May IS-l& 1991. Address reprint requests to P. Sibbons, Department of Paediatric Surgery, Institute of Child Health, London WClN IEH, England. Copyright o I992 by W.B. Saunders Company 0022-3468192/2703-0013$03.00/O 339
SIBBONS, SPITZ, AND VAN VELZEN
Proximal No
2 1
Position
>
6
5
4
3
Art + Vn
Lvm
Vn + Lym
Vn
Lym
Vn + Lym
3
_
v+n
>
5
6
Art + Lym
1
7 Vn 6
Lym v+n
Vn +
Art
Lym
1
11 Lym 12
Art +
Vn
Vn
Art + Lym
Vn +
Vn
Lym
Art
Art + Vn
Lym
Art
Art t Vn
Art +
Vn
Lym
Lym \
Art
Art + Vn
Lym Vn +
A;’
1
Lym
Lym
r
9 10
Art Art
Art +
2
Lym
Art Art
4
RESULTS
Distal
Animal
Vn + Lym
Y 1
sequence
Fig 1. Sequence of ligations for each animal. Art, arteries alone; Vn, veins alone; Lym, lymphatics elone; Art + Lym, arteries + lymphatics; Vn + Lym, veins + lymphatics; Art + Vn, arteries + veins.
arcades was carried out aided by the use of an operating microscope (magnification x8). Commencing at the ileocecal junction and continuing proximally along the ileum, nine adjacent groups of vessels were selected for ligation separated by 10 intact groups of vessels. The ligation procedures consisted oh (1) artery plus vein (Art + Vn); (2) lymphatics alone (Lym); (3) vein plus lymphatic (Vn + Lym); (4) vein alone (Vn); (5) artery plus lymphatic (Art + Lym); and (6) artery alone (Art). The ligations were performed according to a strict protocol for subsequent analysis (Fig 1) of three sequences comprising six ligations each and occupying the distal 25 cm of ileum. The ligations were performed midway between the serosal surface of the intestine and the mesenteric lymph node. The animals were returned to the sow and littermates on recovery from anesthetic. All these piglets were killed 48 hours postoperatively. At necropsy, the distal ileum together with the associated mesenteric vascular arcades were removed. Tissue for histopathological examination was taken from the center of each bowel segment supplied by the ligated vessels, fixed in formal-calcium for 16 hours, and processed to paraffin wax embedding by routine schedules. Microscopic sections were cut at 4 pm and stained with hematoxylin and eosin.
The histopathological scores for each of the ligation procedures according to position in the distal ileum are shown diagrammatically in Fig 2. Arterial ligation resulted in vascular congestion with severe mucosal stripping and submucosa1 disruption. Necrosis affecting the mucosa and submucosa was often present and in some cases the muscularis was affected. There were only minor degrees of lymphatic engorgement (Fig 3). Ligation of lymphatic vessels induced gross lymphatic engorgement that assumed the typical pattern and distribution of pneumatosis but did not contain gas (prepneumatosis) (Fig 4). This was most often seen in the submucosa but in some cases on the serosal surface as well. This entity of prepneumatosis was found only when lymphatic vessels, alone or in combination, were included in the ligation procedure (Table 1). The complete picture of NEC developed in two of the lowest birth weight piglets who underwent lymphatic ligation only. The combination of arterial and lymphatic ligation induced the most severe pathology. This consisted of complete NEC in all the distal ligation positions and
I
0 .&a
Microscopic Grading Stained sections were evaluated by three separate observers using a histopathological severity score system.4 The system comprised the grading from 1 to 5 of the histopathological severity of defined tissue changes such as vascular congestion, mucosal stripping, necrosis, ulceration, perforation, prepneumatosis, and pneumatosis intestinalis for each lesion. The mean lesion grade was calculated as histopathological severity score for each ligation procedure.
Art
Art+ Lym 40
Vn
BE 00 !
*\ 20
sequences Fig 2. Total pathology score by sequence for each animal according to the type of vessels ligated. Prox, proximal sequence; mid, middle sequence; dist, distal sequence.
LYMPHATICS AND PNEUMATOSIS
341
IN BOWEL ISCHEMIA
Table 1. Presence of Prepneumatosis
in Each Animal According to
the Types of Vessels Ligated Animal NO.
1
Type of Vessels Ligated
Birth Weight(g)
Art
Art+
Lym
Vn
Vn+Lym
Lw
+
2
792
3
1,013
4
1,213
5
894
6
1,131
7
981
t
+
8
940
t
t
9
756
+
10
1,160
11
853
12
997
+ +
Art+Vn
+ + + +
922
Bled +
t Bled
t
+ +
*AbbreviaGons: Art, artery; Lym, lymphatics; Vn, vein. Fig 3. Arterial ligation lesion (H&E, original magnification x40) showing vascular congestion [a}, submucosal disruption (b), some lymphatic distension (cl, and almost full-thickness necrosis.
pneumatosis in 7 of 12 animals in the midileal positions (Fig 5). All animals except one developed moderate prepneumatosis. It would appear that the lesions produced by arterial plus lymphatic ligation amounted to more than merely the sum of each combined. The mean difference in total histopathological severity scores for arterial plus lymphatic ligation was significant at all positions compared with the other five ligation procedures (P I .Ol). DISCUSSION
Although the etiology and precise pathogenesis of neonatal NEC remains open to debate, the most accepted final common pathway to the development of the condition is reduced mucosal blood flow. The
Fig 4. Typical lymphatic ligation lesion (H&E, original magnification x40) showing prepneumatosis (distended lymphatics with the shape and distribution of PI) (c).
area of intestine most frequently affected is the distal ileum and proximal colon. We have previously shown that the blood flow in the distal ileum of neonatal piglets is significantly reduced compared with more proximal ileum. The reduction was greatest in SGA piglets.’ In response to occlusion of vessels in the mesenteric arcades, compensatory collateral flow developed in NBW piglets in both the proximal and distal ileum, whereas in the SGA animals this response failed to occur and at 24 hours postligation blood flow in the distal ileum had further reduced.’ The experiments reported herein were carried out to examine the role of the various component vessels in the mesenteric arcades, singly or in combination, in the induction of the NEC-like lesions. Varying degrees of histopathological changes occurred with the different types of ligation procedures and, with the exception of lymphatic alone, there was a noticeable progression of lesions with proximity to the ileocecal junction.
Fig 5. Lesion produced by arterial plus lymphatic ligation (H&E, original magnification x40) showing full-thickness necrosis, total mucosal stripping, severe submucosal disruption, prepneumatosis (cl, and pneumatosis (d).
SIBBONS. SPITZ, AND VAN VELZEN
342
Arterial plus lymphatic ligation resulted in the most severe damage, which amounted to more than the sum of each ligation procedure alone. All the histopathological features of complete NEC were present in all the distal ligation sequences. Pneumatosis was observed in all these as well as in the middle sequence and moderate to severe prepneumatosis was found at all except one ligation positions. Two piglets undergoing lymphatic ligation only developed complete NEC. These were two of the lowest birth weight animals (weighing 853 g and 894 g). No previous work has been carried out specifically directed at experimental occlusion of the lymphatic vessels in the mesentery of newborn animals. Prepneumatosis and pneumatosis developed only when the ligation procedure included the lymphatic vessels either singly or in combination with the adjacent blood vessels. The distended lymphatics in the submucosa assumed the shape and distribution of typical pneumatosis. We postulate that anaerobic bacteria gain access to the distended lymphatics, which contain an ideal substrate for proliferation and produce the hydrogen gas that results in the typical pneumatosis seen in neonatal NEC. Shut down of the mesenteric lymph node by bacterial translocation would further exacerbate the lymphatic obstruction and cause a progression of the disease process. Jackson et al7 clearly demonstrated that bacterial translocation to the mesenteric lymph node occurred in neonatal rabbits whose alimentary tracts were inoculated with Escherichia coli.
Although the experimental animals in this study were not hemodynamically monitored during the operative procedures, all interligation “control” tissue was histopathologically normal. Therefore, we have no basis for implicating any perioperative stress factor in the pathogenesis of the lesions induced. In the light of these observations, we hypothesize that, in susceptible animals, NEC is initiated by an incident or combination of incidents that cause a reduction of arterial inflow to specific areas of the intestine, particularly the terminal ileum and proximal colon. This is followed by mucosal ischemia, which leads to mucosal breach and the creation of a microenvironment conducive to bacterial proliferation and invasion, particularly of the lymphatics. It is possible that translocation of bacteria via the lymphatits presents an overwhelming antigenic challenge to the local mesenteric lymph node resulting in lymph node shut down.’ The subsequent distension of lymphatics proximal to the lymph-node constitute prepneumatosis in the bowel wall. The stagnant material in the distended lymphatics (high molecular weight milk fats and some proteins) provides an excellent substrate for the production of gas by proliferating facultative anaerobes. This gas is radiologically and histopathologically observed as pneumatosis intestinalis. The fat/protein substrate is also a good medium for lipid peroxidation and the generation of oxygen free radicals, damage from which may be integral to the pathogenesis of NEC.
REFERENCES 1. Engel RR, Virnig NL, Hunt CE, et al: Origin of mural gas in necrotising enterocolitis. Pediatr Res 7:292-296,1973 2. Smith MF, Boriello SP, Clayden GS, et al: Clinical and bacteriological findings in necrotising enterocolitis: A controlled study. J Infect 2:23-31, 1980 3. Popoff MR: Are anaerobes involved in neonatal necrotising enterocolitis?, in Boriello SP (ed): Clinical and Molecular Aspects of Anaerobes. Petersfield, England, Wrightson Biomedical, 1990, pp 49-57 4. Sibbons P, van Velzen D, Spitz L, et al: Relationship of birth-weight to the pathogenesis of necrotising enterocolitis in the neonatal piglet. Pediatr Path01 8:151-162, 1988
5. Sibbons PD, Spitz L, van Velzen D: Collateral blood flow in the distal ileum of neonatal piglets. A clue to the pathogenesis of necrotising enterocolitis. Pediatr Path01 (in press) 6. Sibbons PD, Spitz L, van Velzen D, et al: Necrotising enterocolitis induced by local circulatory interruption in the ileum of neonatal piglets, in Boriello SP (ed): Clinical and Molecular Aspects of Anaerobes. Petersfield, England, Wrightson Biomedical, 1990, pp 59-67 7. Jackson RJ, Smith SD, Widowsky RM, et al: The effect of E. coli virulence on bacterial translocation and systemic sepsis in the neonatal rabbit model. J Pediatr Surg 26:483-486, 1991 8. McConnell I: The Immune System fed 2). London, England, Mosby, 1981, pp 181-192
Discussion M.Z. Rowe (Pittsburgh PA): This is an important question. It is interesting if you look at the literature of necrotizing enterocolitis; there are really very few experimental studies of what pneumatosis is really about, about how it really develops, and I compliment the authors on their determination to try to get some
more insights into the problem. In listening to the papers this morning, it begs a question I think 1’11 mention, and the question is, a lot of research and a lot of questions have been centered on the whole idea of blood flow to the intestine as one of the pathogenic factors of NEC, yet there is beginning to be devel-
LYMPHATICS
AND PNEUMATOSIS
IN BOWEL ISCHEMIA
oped a huge amount of data that suggest that maybe vascular compromise might not have anything to do with necrotizing enterocolitis. For example, as Dr Langer mentioned, the match control studies have over and over again shown that vascular compromise, perinatal shock, patent ductus arteriosus, and umbilical catheters are just as common in match controls as in NEC and these events simply define a vulnerable population rather then suggest an etiologic agent. The study of steroids in a multiinstitutional study showed that children who were born after their mothers got steroids for lung maturity had a significant lower incidence of NEC than did the controls and it was found that these children had an increase in their gut maturity. It was interesting that in the vitamin-E study, again a multiinstitution study in which they were looking at prevention of retrolental fibroplasia, it was shown strikingly that this interfered with white cell function of these babies and there was an increased incidence of NEC, which suggested an infection etiology. And, of course, we are all familiar with the extremely interesting new studies that seem to suggest that prophylactically giving IgA, at least in Vienna, significantly reduces NEC. And then we also note from our own interests that there are clusters of NECs that suggest not perfusion but an infectious origin, so I think this whole question is difficult, but specifically as relates to infection, there are several specific questions that should be addressed. One, what is a low birth weight pig-is that a premature pig or is it small for gestational age or a runt? These are different because a low birth weight is just a small pig. A premature is immature, while a small for gestational age might show intrauterine malnutrition, and I think has different physiologic effects. There are no data that I know that suggest translocation of the bacteria into the lymph nodes will block the lymphatits, so I am confused on how lymphatic obstruction could be an etiologic source in NEC. I am not sure just how that could take place. Just bacteria in the lymphatics usually does not block the lymphatics. If the lymphatics are the source of pneumatosis, do you see air in the lymph nodes since these are bacteria that do form gases?
343
A. Kosloske (Baltimore, MD): I have a couple of questions for Mr Spitz. First, about the bacteria. How much bacterial colonization has occurred at 12 to 18 hours in the neonatal piglet? Are they more like the germ-free animals or are they heavily colonized by that time? I wondered whether you did qualitative cultures or even quantitative cultures that would very helpful in sorting this out. You also wonder what is happening in the lymphatics. As I listen to this, I thought that perhaps rather than the lymphatics being a place where the bacteria stop and proliferate in this model, that by ligating the lymphatics you have artificially created a block to the run-off of bacterial toxins which then goes on and leads to the formation of pneumatosis intestinalis. A.A. de Lorimier (San Francisco, CA): Have you looked at the histology of your human newborn infants with NEC to see evidence of lymphatic distention? L. Spitz (response): In answer to Marc Rowe’s question, these animals were small for gestational age, they were the runts of the group. There were 10 delivered spontaneously. The normal birth weights were the full-grown animals in the same litter. So I think there is a difference there. I think the reason we are getting the lesions we see in these animals is that the mucosa is more susceptible in these small for gestational animals and they are particularly the ones we see in clinical practices as being more prone to develop NEC. I would agree that the mesenteric lymph node blockage is mere speculation on our part. We haven’t probed into this, but it is interesting to note that all the animals who had lymphatic obstruction were the only ones who developed this pneumatosis. In answer to Dr Kosloske’s questions, we have not looked at the bacteria specifically, That study is still in progress and I can’t give you any results of those studies as yet, but these animals were allowed to suckle normally and presumably the intestine was colonized by the time we experimented on them 12 to 18 hours later. We have looked at the histological studies in the human NEC and examined the tissue carefully; they appear to have free pneumatosis as well as pneumatosis lesions.
