Correspondence

Posterior Communicating Artery and Ischemic Stroke

N Engl J Med 1994; 331:1020-1021October 13, 1994

Article

To the Editor:

In their article on cerebral vascular anatomy as a risk factor for ischemic infarction (June 2 issue),1 Schomer et al. incorrectly imply that unilateral “watershed” infarcts can occur in juxtaventricular white matter. We wish to correct this error. We also urge that the term “watershed” be banished from the medical literature. “Watershed” is a mistranslation of die letzte Wiese, or “last meadow,” where a billabong may form. A true watershed is a ridge dividing two areas drained by different river systems. Hence, it is a source, rather than a destination, of flow.

Eighty-seven years ago, Beevor2 studied cerebral arterial border zones by injecting dyes under equal pressure into major arteries. He described “null points” where no flow occurred, but Zulch3 first suggested that selective damage occurs in these zones when perfusion pressure is lowest, describing additional border zones in the basal ganglia, cerebellum, and spinal cord.

The anatomical locations of cerebral arterial border zones are well known.4 None correspond with the unilateral “watershed infarct” shown in Figure 3B of Schomer et al. There is no anatomical evidence that any border zones occur in deep white matter adjacent to the lateral ventricles. Border-zone lesions are usually cortical when there is hypoperfusion. When they involve white matter, they occur bilaterally and extend in a wedge shape from the cortex to the ventricular wall.5

Arterial-border-zone lesions result in distinct clinical syndromes, none of which were reported in the paper. Schomer et al. made radiologic observations without neuropathologic correlation and failed to supply other explanations for the white-matter lesions they used to demonstrate their case. They also offered no other explanations for their failure to identify posterior communicating arteries on magnetic resonance angiography. If the perfusion pressures in carotid and basilar arteries were the same, there would be no flow in a large posterior communicating artery, and no signal would register on magnetic resonance angiography.

William D. Brown, M.D.
Floyd H. Gilles, M.D.
Marvin D. Nelson, Jr., M.D.
Children's Hospital Los Angeles, Los Angeles, CA 90054-0700

5 References

1. 1

   Schomer DF, Marks MP, Steinberg GK, et al. The anatomy of the posterior communicating artery as a risk factor for ischemic cerebral infarction. N Engl J Med 1994;330:1565-1570
   Full Text | Web of Science | Medline

2. 2

   Beevor CE. The cerebral arterial supply. Brain 1907;30:403-425
   CrossRef | Web of Science

3. 3

   Zulch KJ. Uber die Enstehung und Lokalization der Hirninfarkte. Zentralbl Neurochir 1961;21:158-178
   Medline

4. 4

   Howard R, Trend P, Russell RWR. Clinical features of ischemia in cerebral arterial border zones after periods of reduced cerebral blood flow. Arch Neurol 1987;44:934-940
   Web of Science | Medline

5. 5

   Lindenberg R. Patterns of CNS vulnerability in acute hypoxaemia, including anaesthesia accidents. In: Schade JP, McMenemey WH, eds. Selective vulnerability of the brain in hypoxaemia. Oxford, England: Blackwell Scientific, 1963:189-209.
   

To the Editor:

The conclusion that “a small (< 1 mm in diameter) or absent ipsilateral posterior communicating artery is a risk factor for ischemic cerebral infarction in patients with internal-carotid-artery occlusion” is of vast importance. Sadly, on the basis of the evidence presented by Schomer et al., this conclusion remains unproved.

The anatomical findings presented after the event (presumably an ischemic event of some type, although that is not stated) can have little predictive power, since cause and effect cannot be determined from a single “snapshot.” What, for example, is the effect of a “watershed infarction” on intimately related, narrow-caliber atherosclerotic vessels? Could not these findings represent a secondary effect of propagated thrombosis from a nearby area of intimal damage or ischemic cerebral tissue? To justify the conclusion of Schomer et al., a cohort with known cerebrovascular anatomy, in which the direction of flow was known but in which no demonstrable infarction was present, would have to be observed over time and compared with a matched population of controls. If ischemic cerebral infarction occurred more often in the group with a small or absent posterior communicating artery (ipsilateral to the occluded internal carotid artery), then the conclusion of Schomer et al. would seem valid. Such a study is a daunting prospect, but if therapeutic interventions are being seriously entertained, nothing less should suffice.

J.R.C. Bowen, M.R.C.P.
Princess Alexandra's Hospital, Swindon SN4 0QJ, United Kingdom

Author/Editor Response

The authors reply:

To the Editor: We agree with Dr. Bowen that a statistically significant correlation between ischemic infarction and the presence and size of a posterior communicating artery is important, but cause and effect cannot be determined from a retrospective epidemiologic study such as ours. We selected patients with carotid-artery occlusion as an anatomical model and studied the collateral pathways to the hemisphere at risk. The end point was the radiologic observation of an infarct in the tissues vulnerable to hypoperfusion. A statistically significant correlation between a small or absent ipsilateral posterior communicating artery and the occurrence of such an infarct was observed. This is thought to indicate a risk factor even though a specific cause and effect cannot be proved. Although the number of patients is limited, this finding is predictive of infarction in patients with disease progressing from carotid stenosis to occlusion and also of tolerance to presurgical carotid-occlusion tests. We also believe that the study Dr. Bowen suggests to demonstrate causation would be daunting, would add little to increase confidence in the observation, and is probably not clinically feasible.

We defer to Dr. Brown et al. with regard to the correct translation of the term “die letzte Wiese.” We used the term “watershed infarction” in the manner that is widely accepted -- namely, to identify cerebral infarction resulting from low perfusion rather than thromboembolic disease.

Moody et al.1 defined a “border zone” in the lower centrum semiovale supplied by long arteries and arterioles with exclusive non-interdigitating territories as being vulnerable to circulatory challenge. A border zone may be defined anatomically or physiologically. Nelson et al.2 support this concept, postulating the existence of a “vascular border zone” between the major cerebral arteries that is “functionally related to systemic pressure and to perfusion pressure within each major cerebral vessel.” There is no anatomical basis for this border zone. The consistent finding of an abnormal signal in the deep white matter (and extending to the cortex) on magnetic resonance scans does have an anatomical basis as a border zone and may have a physiologic basis in perfusion pressures.

Donald F. Schomer, M.D.
University of Texas M.D. Anderson Cancer Center, Houston, TX 77030

D.R. Enzmann, M.D.
Stanford University Medical Center, Stanford, CA 94305-5105

2 References

1. 1

   Moody DM, Bell MA, Challa VR. Features of the cerebral vascular pattern that predict vulnerability to perfusion or oxygenation deficiency: an anatomic study. AJNR Am J Neuroradiol 1990;11:431-439
   Web of Science | Medline

2. 2

   Nelson MD Jr, Gonzalez-Gomez I, Gilles FH. The search for human telencephalic ventriculofugal arteries. AJNR Am J Neuroradiol 1991;12:215-222
   Web of Science | Medline

Citing Articles (1)

Citing Articles

1. 1

   Pasi A. Korhonen, Nan M. Laird, Juni Palmgren. (1999) Correcting for non-compliance in randomized trials: an application to the ATBC study. Statistics in Medicine 18:21, 2879-2897
   CrossRef