Papers

Cohort study of multiple brain lesions in sport divers: role of a patent foramen ovale

BMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7082.701 (Published 08 March 1997) Cite this as: BMJ 1997;314:701
  1. Michael Knauth, neuroradiologista,
  2. Stefan Ries, neurologistb,
  3. Stefan Pohimann, biologistc,
  4. Tina Kerby, techniciana,
  5. Michael Forsting, neuroradiologista,
  6. Michael Daffertshofer, neurologistb,
  7. Michael Hennerici, neurologistb,
  8. Klaus Sartora
  1. a Department of Neuroradiology University of Heidelberg Klinikum Heidelberg Im Neuenheimer Feld 400 D 69120 Heidelberg Germany
  2. b Department of Neurology, University of Heidelberg, Klinikum Mannheim, Germany
  3. c Department of Biology, University of Frankfurt, Klinikum Frankfurt, Germany
  1. Correspondence and reprint requests to: Dr Knauth
  • Accepted 16 January 1997

Abstract

Objective: To investigate the role of a patent foramen ovale in the pathogenesis of multiple brain lesions acquired by sport divers in the absence of reported decompression symptoms.

Design: Prospective double blind cohort study.

Setting: Diving clubs around Heidelberg and departments of neuroradiology and neurology.

Subjects: 87 sport divers with a minimum of 160 scuba dives (dives with self contained underwater breathing apparatus).

Main outcome measures: Presence of multiple brain lesions visualised by cranial magnetic resonance imaging and presence and size of patent foramen ovale as documented by echocontrast transcranial Doppler ultrasonography.

Results: 25 subjects were found to have a right-to-left shunt, 13 with a patent foramen ovale of high haemodynamic relevance. A total of 41 brain lesions were detected in 11 divers. There were seven brain lesions in seven divers without a right-to-left shunt and 34 lesions in four divers with a right-to-left shunt. Multiple brain lesions occurred exclusively in three divers with a large patent foramen ovale (P=0.004).

Conclusions: Multiple brain lesions in sport divers were associated with presence of a large patent foramen ovale. This association suggests paradoxical gas embolism as the pathological mechanism. A patent foramen ovale of high haemodynamic relevance seems to be an important risk factor for developing multiple brain lesions in sport divers.

Key messages

  • An increased prevalence of multiple brain lesions has been reported in scuba divers compared with non-diving controls

  • It has been suggested that the brain lesions were due to arterial gas embolism and that the gas emboli could have entered the arterial circulation via a patent foramen ovale

  • We investigated this hypothesis in volunteer sport divers who had made at least 160 scuba dives

  • Brain lesions occurred in divers even in the absence of reported decompression sickness

  • Multiple brain lesions occurred exclusively in divers with a large patent foramen ovale

  • The association of multiple brain lesions with a large patent foramen ovale suggests paradoxical gas embolism as the pathological mechanism

Introduction

Scuba diving (diving with self contained underwater breathing apparatus) has become a popular sport. There are about three million certified sport divers in the United States and about one million in Europe. While neuroimaging studies exist for professional divers,1 2 divers with decompression accidents,3 and compressed air tunnel workers,4 it was only recently that a larger population of sport divers, most of whom never had a decompression accident, was investigated by magnetic resonance imaging.

Reul et al found an increased prevalence of brain lesions in sport divers compared with non-diving controls.5 However, about 80% of all brain lesions were found in a subgroup of 27% of the divers. Whereas single lesions in the white matter of the brain were common and were also seen in 20% of the controls, multiple brain lesions (more than four) were detected exclusively in divers. While the authors stated that the brain lesions in divers were probably due to arterial gas embolism during or after surfacing, they did not address the question of how the gas emboli entered the arterial circulation.

Gas bubbles are known to be present in venous blood after ascents from water depths as shallow as three metres.6 However, the lungs usually constitute a competent filter for these bubbles,7 and they do not normally enter arterial circulation. A patent foramen ovale or other right-to-left shunt as a pathway for the arterialisation of venous bubbles is present in about 10-30% of the general population.8 9 10 11 This prevalence roughly equals the percentage of divers with multiple brain lesions in the study of Reul et al. This led to the hypothesis, most recently expressed by Wilmshurst et al,12 that divers with a patent foramen ovale might have a high risk of developing brain lesions.

Whereas the prevalence of patent foramen ovale has been found to be increased in divers with decompression sickness,13 14 especially in cases of early onset decompression sickness with neurological symptoms (type 2 decompression sickness),14 no study has addressed the question of whether divers with a patent foramen ovale are prone to developing multiple brain lesions in the absence of decompression sickness. This is important for estimating individual risk in sport diving, and we therefore conducted the present study to answer this question.

Methods

Subjects

We informed adjacent diving clubs about our study and asked for volunteers, with the proviso that subjects must have made a minimum of 160 scuba dives. We subsequently enrolled 100 sport divers, although nine failed to attend. We obtained a medical history to identify and exclude divers with diseases known to potentially cause cerebral lesions, such as multiple sclerosis or cerebrovascular disease. Four divers had to be excluded: two had a history of cerebral stroke, one had a history suggestive of multiple sclerosis, and one had suffered severe perinatal asphyxia.

We questioned the remaining subjects about their alcohol consumption, smoking habits, and presence of hypertension, hypercholesterolaemia, and diabetes mellitus. From each diver we obtained a diving history, including total number of scuba dives, number of decompression dives, and events with signs and symptoms suggestive of decompression sickness. We also obtained details of what method the divers used to equalise pressure in the middle ear. This is because pressure equalisation by the Valsalva manoeuvre might lead to an increased arterialisation of nitrogen bubbles in multi-level dives with two or more descents or in repetitive dives with a short surface interval. All the divers but one (who turned out not to have brain lesions) were unaware of whether they had a patent foramen ovale.

Before performing the examinations, we obtained all subjects' informed consent according to the Helsinki declaration. The study was approved by the local ethics committee.

Magnetic resonance imaging

We performed magnetic resonance imaging with a 1.5 Tesla magnetic resonance scanner and acquired axial proton density weighted and T2 weighted images covering the whole brain (24 slices, TR 3000 ms, TE 20 ms and 80 ms, matrix 192x256, slice thickness 5 mm). We also included a FLAIR (fluid attenuated inversion recovery) sequence (TR 8300 ms, TE 80 ms, T1 1500 ms, with identical slice parameters) as part of the protocol in order to improve differentiation between widened perivascular (Virchow-Robin) spaces and true brain lesions (see fig 1). The magnetic resonance images were reviewed independently by two neuroradiologists blinded as to whether a patent foramen ovale was present. A lesion was counted if it was hyperintense on proton density weighted and T2 weighted images and the FLAIR sequence. When the neuroradiologists' interpretations differed they reached a consensus by joint review.

Fig 1
Fig 1

Magnetic resonance images of brains of two divers: proton density weighted (left), T2 weighted (middle), and FLAIR sequence (right). Signal changes in subcortical white matter are identical in left and middle images (arrows), but FLAIR sequences allow differentiation between true brain lesions (top right, arrow) and widened perivascular spaces (bottom right, arrow)

Transcranial Doppler ultrasonography

Transcranial Doppler ultrasonography was performed after the magnetic resonance imaging to avoid possible confounding of the results by cerebral echocontrast embolism. Two transcranial pulsed wave 2 MHz ultrasound probes were attached bilaterally to the patient's head and focused at the M1 segment of the middle cerebral artery. During transcranial monitoring, 5 ml of a commercially available sonographic contrast medium (Echovist) containing air microbubbles (diameter <8 μm) was injected twice into an antecubital vein.

Tests were performed at rest and after the subject had performed the Valsalva manoeuvre to increase or provoke a right-to-left shunt.15 We considered a right-to-left shunt to be present when five or more air microbubble signals occurred in the Doppler spectra of either middle cerebral artery after the Valsalva manoeuvre. The haemodynamic relevance of the right-to-left shunt was classified as low if <20 air microbubble signals occurred after the Valsalva manoeuvre and high if Embedded Image20 signals were detected. Transcranial Doppler ultrasonography has proved to be superior to transthoracic echocardiography for detecting a patent foramen ovale and similar to transoesophageal echocardiography.16 17

Statistical analysis

For continuous variables, we used the Mann-Whitney U test to compare the groups (subjects with or without a patent foramen ovale). We compared the group proportions of multiple brain lesions with Fisher's test of exact probability and obtained 95% confidence intervals from scientific statistical tables.18 The level of significance was set at P<0.05.

Results

Subjects

Of the 87 divers who entered the study, 67 were men and 20 women with a mean age of 35.7 (SD 8.9) years, mean height 176.4 (7.6) cm, and mean body weight 77.1 (12) kg. They had performed an average of 565.3 (509.1) scuba dives.

Presence of patent foramen ovale

Twenty five of the subjects were found to have a right-to-left shunt, and the haemodynamic relevance of the shunt was classified as high in 13 cases. The group of divers with a patent foramen ovale was not significantly different from the group without the condition in any variable except body weight (table 1)). None of the divers had reported a history of type 2 decompression sickness.

Table 1

Characteristics of 87 sport divers by presence or absence of right-to-left shunt (values are mean (SD) (mean rank) unless stated otherwise)

View this table:

Presence of brain lesions

A total of 41 brain lesions were detected in 11 of the 87 divers. Seven of the divers without a right-to-left shunt had seven brain lesions (one each), while four of the divers with right-to-left shunt had 34 lesions. Of these four divers, one, with a small patent foramen ovale, had one lesion whereas the other three divers had multiple brain lesions and a patent foramen ovale of high haemodynamic relevance. The lesions had a typical size of 2-3 mm, and all of them were located in the anterior cerebral circulation. The prevalence of brain lesions (single or multiple) was higher in the divers with a patent foramen ovale than in those without (16% v 11%), but the difference was not significant (Fisher's exact test P=0.72).

The three divers with multiple brain lesions were non-smokers, aged in their 40s, and did not have any other vascular risk factors (see fig 2). They had a tendency to have a higher total number of dives than did the other 84 divers (757 v 558), but this difference was not significant. Of these three divers, one had 16 lesions, one had 12 lesions, and the third had five lesions. The prevalence of multiple brain lesions was significantly higher in the 25 divers with a patent foramen ovale than in the 62 divers without (12% (95% confidence interval 3% to 31%) v 0% (0% to 6%), Fisher's exact test P=0.022). The difference was even greater for a patent foramen ovale of high haemodynamic relevance: the occurrence of multiple brain lesions in the 13 divers with a large patent foramen ovale was 23% (5% to 54%) compared with 0% (0% to 4%) in all the other divers (Fisher's exact test P=0.004).

Fig 2
Fig 2

Results for 46 year old male diver (856 dives) without symptoms, no history of decompression incidents, and no history of cerebrovascular disease or known vascular risk factors. Top: magnetic resonance images show multiple lesions in central white matter of brain (hyperintense spots visible on all three images). Bottom: transcranial Doppler ultrasonography after injection of contrast medium shows multiple air microbubbles in left middle cerebral artery after Valsalva manoeuvre (arrow)

Discussion

The aim of our study was to determine whether a patent foramen ovale (or other right-to-left shunt) increased sport divers' risk of developing brain lesions. We therefore examined the prevalence of multiple brain lesions in divers with or without a right-to-left shunt and did not examine non-diving controls. We found that multiple brain lesions occurred exclusively in divers with a patent foramen ovale of high haemodynamic relevance. Our data therefore suggest that the presence of a patent foramen ovale represents an important risk factor for the development of multiple brain lesions in sport divers even in the absence of type 2 decompression sickness. Whereas single brain lesions are common in non-diving control subjects,5 multiple brain lesions are unusual.

The association of multiple brain lesions with a large patent foramen ovale supports the hypothesis that the brain lesions were due to arterial gas embolism during or shortly after decompression. A thromboembolic pathogenesis is unlikely because, if that were the case, brain lesions should have been evenly distributed between divers and non-diving controls in the study of Reul et al.5 Furthermore, multiple small brain lesions in the white matter are an unusual finding in patients found to have a patent foramen ovale during cerebrovascular investigation.19

It should be remembered, however, that in our study 22 of the 25 divers with a right-to-left shunt, and 10 of the 13 divers with a shunt of high haemodynamic relevance, did not have multiple brain lesions detectable by magnetic resonance imaging. Thus, while a large patent foramen ovale seems to be a risk factor for developing multiple brain lesions, other cofactors are likely to play a role. On the other hand, lesions detectable by magnetic resonance imaging may represent only the tip of an iceberg of cerebral gas embolism in divers. Even in divers with neurological decompression sickness, brain lesions are not always detectable by magnetic resonance imaging.3 This is similar to results found with patients with a source of cerebral thromboembolism: some patients develop cerebral lesions in the absence of clinical signs and symptoms (“silent infarcts”), while others have no detectable lesions after reversible neurological symptoms (such as transient ischaemic attack).

Reliability of study

Concern was expressed about the study of Reul et al5 in that the self selection of divers might have led to the preferential recruitment of divers who thought that they had cause for concern because they had had minor decompression symptoms, thus creating a bias towards higher numbers of brain lesions in divers.12 20 This argument is not applicable to our study, as all but one diver (who was found not to have any brain lesions) were unaware of whether they had a patent foramen ovale. The difference in the prevalence of multiple brain lesions between the groups of divers with and without patent foramen ovale cannot be explained by this hypothetical bias from self selection.

The group of divers with right-to-left shunts did not differ significantly from those without shunts in any other variable except that they had a lower body weight. Nitrogen-which forms bubbles in divers' tissues during decompression-is 4.5 times more soluble in fat than in water, leading to a higher gas load in divers with a high proportion of body fat. There is, as yet, no hard evidence that this also leads to a higher amount of venous gas bubbles, but even if it did it could not explain our results since the divers with the higher prevalence of multiple brain lesions had lower body weight.

Comparison with other studies

Other studies have found an increased prevalence of a patent foramen ovale or other right-to-left shunt in divers with decompression sickness.14 The prevalence of a patent foramen ovale was as high as 66% in a subgroup of divers with type 2 decompression sickness of early onset thought to be due to gas embolism. In the absence of reported type 2 decompression sickness, the multiple brain lesions we found in the divers of our study are most likely the consequence of subclinical arterial gas embolism.

Our results are also in accordance with those of Reul et al,5 who found multiple brain lesions in sport divers, most of whom did not have a history of decompression sickness. However, we could not confirm their findings of 86 brain lesions in 52 sport divers: we found only 41 lesions in 87 divers. Moreover, Reul et al found at least one brain lesion in more than half of their divers (27/52), whereas we detected brain lesions in only 13% (11/87). Since our study had almost identical inclusion criteria, patient selection can hardly be entirely responsible for such a large difference. For detection of brain lesions, Reul et al used axial proton density weighted and T2 weighted magnetic resonance images, in which fluid filled perivascular spaces can be mistaken for brain lesions (see fig 1).21 22 23 As the signal of the cerebrospinal fluid is suppressed in FLAIR sequences, sensitivity and specificity in the diagnosis of brain lesions is increased.23 24 25 To improve the specificity of our diagnosis, we counted a hyperintense change in signal as a lesion only if it was hyperintense to white matter in T2 weighted and proton density weighted images and in the FLAIR sequence. Thus, the differences between ourselves and Reul et al in the reported prevalence of brain lesions in divers may have been due to the different criteria used in the definition of brain lesions. Furthermore, the total number of dives is not apparent from the publication of Reul et al, and some of the difference in the frequency of brain lesions might be attributable to different total numbers of dives.

Studies in professional divers have failed to demonstrate an increased risk of developing cerebral lesions. Todnem et al and Rinck et al found significantly more abnormal neurological and electroencephalographic findings in the professional divers, but they did not find an increased prevalence of hyperintense spots on the magnetic resonance images of the divers' brains.1 2 All the control subjects were examined with a 1.5 Tesla imager, whereas many of the divers were examined with an imager with a lower field strength (0.5 Tesla). Moreover, the sequence parameters, and therefore the T2 weighting of the images, were different. Thus, the comparability of the groups could be questioned. The authors used only proton density weighted and T2 weighted images, in which perivascular spaces and true brain lesions can be confounded.21 22 23 The extremely high prevalence of “spots of high signal intensity” (42%) in the control subjects raises the question of whether many Virchow-Robin spaces were counted, rendering possible differences in the prevalences of true brain lesions insignificant.

While Todnem et al's and Rinck et al's results seem to contradict those of Reul et al and our study, Fueredi et al found that compressed air tunnel workers had an increased prevalence of brain lesions.4 Compressed air tunnel workers are subject to compression and decompression, breathe compressed air, and are at risk of decompression sickness just as divers are.

Conclusion

Our study shows an increased prevalence of multiple brain lesions in divers with a patent foramen ovale compared with divers without a patent foramen ovale. In the absence of reported type 2 decompression sickness this finding supports the hypothesis that these lesions are the consequence of subclinical cerebral gas embolism. As about a quarter of all sport divers will have a patent foramen ovale, its association with an increased prevalence of multiple brain lesions means there is an urgent need for further discussion of how a right-to-left shunt affects people's fitness to dive and whether different decompression tables should be developed for such divers.

Acknowledgments

We thank A Baier and P Kuprion for their support in performing transcranial Doppler ultrasonography and all of the divers who participated in the study. Dr P Müller gave us important advice about the pathophysiology of decompression sickness. We are indebted to Dr H-P Altenburg for his advice on statistical methodology and to M Garcia for her editorial comments.

Funding: University research grant of the University of Heidelberg

Conflict of interest: None

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
View Abstract