Content uploaded by Richard Vann
Author content
All content in this area was uploaded by Richard Vann on Mar 28, 2014
Content may be subject to copyright.
UNDERSEA
&
HYPERBARIC
MEDICINE
Volume 29, Number 4
Winter 2002
The Journal of the
Undersea and Hyperbaric Medical Society, Inc.
Rubicon Foundation Archive (http://rubicon-foundation.org)
VOLUME 29, NUMBER 4
WINTER 2002
Undersea &
Hyperbaric
Medicine
CONTENTS
Multiple Sclerosis Forum
Mini-Forum on Multiple Sclerosis (MS) and Hyperbaric Oxygen Therapy
R.E. Moon, Clinical Editor .............................................................................................................................235
Case Report
Diver with acute abdominal pain, right leg paresthesias and weakness: A case report
J.Wang, K. Corson, K. Minky, J. Mader ..........................................................................................................242
Research Papers
The incidence of venous gas emboli in recreational diving
R.G. Dunford, R.D. Vann, W.A. Gerth, C.F. Pieper, K. Huggins
C. Wacholtz, P. B. Bennett ...............................................................................................................................247
Alternobaric oxygen therapy in long term treatment of Ménière’s Disease
B. Fattori, G. De Iaco, A. Nacci, A. Casani, F. Ursino....................................................................................260
Levels of anxiety and hostility in South African Navy Divers
C. H. van Wijk..................................................................................................................................................271
Hyperbaric oxygen improves healing in experimental rat colitis
M.L. Akin, B.M.Gulluoglu, H. Uluutku, C. Erenoglu, E. Elbuken, S. Yildirim, T. Celenk ...............................279
A method of patch clamp recording in hyperbaric oxygen
C.V. Jamieson, A.G. MacDonald......................................................................................................................286
Prescription medication use aboard U.S. Submarines during periods underway
M.H. Jan, T.L. Thomas, T.I. Hooper ................................................................................................................294
Author Indexes...................................................................................................................................307
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
The incidence of venous gas emboli in
recreational diving.
R.G. DUNFORD
1,2
, R.D. VANN
2,3
, W.A. GERTH
2,3,4
, C.F. PIEPER
5,6
, K. HUGGINS
7
,
C. WACHOLTZ
2
, and P.B. BENNETT
2,3
1
Hyperbaric Center, Virginia Mason Medical Center, Seattle, WA
2
Divers Alert Network, Duke Medical Center, Durham, NC
3
Center for Environmental Physiology and Medicine, Duke Medical Center, Durham, NC
4
Navy Experimental Diving Unit, Panama City, FL
5
Center for Aging, Duke Medical Center, Durham, NC
6
Division of Biostatistics, Department of Community and Family Health
7
USC Wrigley Marine Science Center, Avalon, CA
Dunford RG, Vann RD, Gerth WA, Pieper CF, Huggins K, Wachholz C, Bennett PB, The incidence of
venous gas emboli in recreational divers. Undersea Hyperb Med 2002; 29(4):247-259- From 1989-91, the
Divers Alert Network monitored recreational divers for Doppler-detected venous gas emboli (VGE) and
depth-time profiles following multi-day, repetitive, multi-level exposures. A Spencer score >0 occurred in 61
of 67 subjects (91%) and 205 of 281 dives (73%). No subject developed decompression sickness (DCS) on
monitored days although 102 dives (36.3%) scored at Spencer Grades 2 or 3 (High Bubble Grade, HBG). We
recorded the depth-time profiles with Suunto dive computers and estimated exposure severity with a
probabilistic decompression algorithm. The HBG incidence increased 53% over the range of exposure
severity (p<0.001) in the divers, was approximately 20% higher for repetitive dives than for first dives, and
decreased approximately 25% over the 6-8 days of multi-day diving (p<0.001) suggesting a phenomenon
similar to DCS adaptation. The observed HBG incidence was approximately 20% higher for males than
females. Older male divers had a 25% increase in observed incidence of HBG while older female divers
showed a 55% increase when compared to their younger counterparts.
venous gas emboli, Doppler bubble detection, probabilistic modeling, multi-level diving, repetitive
diving, multi-day diving, age, sex, adaptation, dive computers.
INTRODUCTION
Decompression sickness (DCS) is a pathological event thought caused by intravascular
(1) and extravascular (2) gas bubbles. The bubble theory of DCS is supported by empirical
evidence, and venous gas emboli (VGE) can be detected after diving using non-invasive Doppler
ultrasonic methods, but the etiology of DCS is complex, and VGE are frequent after symptom-
free dives (3). High VGE levels, however, are associated statistically with increased DCS
incidence suggesting that VGE and DCS may share a common origin (4). Further, VGE might
initiate DCS should they cross or bypass the pulmonary capillary filter and be transported by the
arterial circulation to organs such as the brain or spinal cord (5-11).
While depth-time exposure is the cause of VGE, other factors may affect their occurrence
Copyright © 2002 Undersea and Hyperbaric Medical Society, Inc. 247
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
including individual susceptibility (3,12), daily variability (3,13), rapid ascent (14,15,16),
exercise at depth (12), and exercise during decompression (17). Divers who were cold at depth
were less likely to have Doppler-detectable VGE than warmer divers over a three-hour post dive
surveillance period (18). In altitude studies, the profusion of Doppler-detectable VGE in males
was greater than in females (13).
There have been studies that monitored the occurrence of Doppler-detected VGE after
controlled open-water diving (19,20), but none have investigated the incidence of VGE after
uncontrolled exposures that are the norm for recreational diving. Our purpose was to explore the
effect of depth-time exposure on the probability of VGE. We used dive computers to record the
multilevel, multi-day depth-time profiles of recreational divers and applied a probabilistic
decompression algorithm to estimate the exposure severity of the dives. This measure of severity
accounted for the effects of dive time, dive depth, and repetitive diving on VGE but not for the
effects of sex, age, and multi-day diving that also influence VGE incidence.
METHODS
From 1989 to 1991, the Divers Alert Network (DAN) sponsored six dive trips on live-
aboard dive boats in the Caribbean and Pacific to investigate the presence of VGE in recreational
scuba divers at warm-water dive sites. The trips were announced in nationwide publications, and
volunteer subjects paid their own expenses. Upon reporting aboard, subjects received a
presentation describing the project before the first dive, and all divers willing to participate were
followed for all dives undertaken during at least one day of diving. Each subject chose the day or
days to be monitored depending on the availability of monitoring slots. Less than 10% of divers
declined to participate.
The divers’ activities were unrestricted, but they were expected to follow the guidelines
of the dive operator. The Duke University Institutional Review Board approved the research
protocol, and the subjects provided written informed consent.
One of two technicians accompanied each trip. Both were trained and experienced in the
use of Doppler equipment for detecting VGE in the field. Doppler signals were acquired using a
2.5-MHz Techno Scientific instrument and dual-ear headphones. Signals were recorded on a
Marantz model PMD430 portable tape deck, and monitoring sequences were annotated with a
Shure "push-to-talk" microphone. Not more than four subjects were followed on any day to
avoid congestion and technician fatigue. To increase the size of the subject population, different
subjects were monitored on subsequent days.
On the first day of a trip before diving began, one-minute precordial Doppler signals
were recorded for each subject as examples of bubble-free heart sounds. These served as base
line comparisons during subsequent signal grading after the trip was complete. The Doppler
probe was placed at the left sternal border and manipulated until the flow sounds were strong
with valve sounds audible in the background. A subject was monitored once 20-40 minutes
post-dive although monitoring was sometimes delayed due to normal subject activity. During
precordial monitoring, the subject stood quietly for one minute before performing three full
knee-bends at 30 sec intervals. Subsequently, the subject’s right and left subclavian veins were
monitored for 30 sec during quiet standing followed by three hand squeezes at 20 sec intervals.
Subjects used the hand on the side opposite that being monitored to steady themselves against
boat movement.
248
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
On a day a subject was monitored, he or she wore a commercially available Suunto
Model SML dive computer that recorded the deepest depth attained in each 3 min interval and
the time of surfacing at 1 min resolution. Dive computers were calibrated and loaned by
SeaQuest Industries (San Diego, CA). After the dives, the technician downloaded each depth-
time profile by hand from the Suunto display. The total dive time and maximum depth were
crosschecked against the subject's personal dive computer and/or watch and depth gauge. Total
dive time represented the time below 5 feet of seawater (fsw) (116 kPa).
The subjects were blinded to the Doppler findings to minimize the impact of the
investigation on subsequent diving. One of the two technicians analyzed the Doppler signals at
least six months after the trip. Signals were scored according to the 5-point Spencer system (4).
The score assigned for each dive represented the highest score at any of the three evaluation sites
(precordial, left and right subclavian). A score was rejected if the presence or absence of
bubbles was equivocal or if the signal was of unacceptable quality. If signals from two of the
three monitored sites were rejected, the dive was removed from analysis. Four monitored dives
by one subject were rejected for such reasons.
Spencer scores were collapsed into a binary outcome variable called High Bubble Grade
(HBG) that was defined as 0 for Spencer Scores of 0 or 1 (HBG=0) and 1 for Scores of 2, 3, or 4
(HBG=1). This dichotomy was selected because higher grades were associated with a 19-fold
greater DCS incidence than were lower grades (4). To explore the consistency of the data, we
also tested the dichotomies: (a) no bubbles (Spencer Score 0) versus bubbles (Spencer Score 1,
2, or 3); and (b) Spencer Scores 0,1,2 versus Spencer Score 3. There were no grade 4 scores.
Statistical Analysis
Logistic regression was used to test the association of the binary outcome variable HBG
with the main effect variable, exposure severity, and with potentially confounding covariates. A
p-value <0.05 was considered significant.
The traditional, although tacit, assumption in the analysis of decompression data where
subjects make multiple exposures has been that each observation for a given individual is
independent of every other observation for that individual (21,22). Since Nishi indicated that
individuals might have different propensities to bubble (4), we tested the validity of the
independence assumption in our data where there were numerous replicate measurements.
There are standard methods that account for subject variability in balanced designs (i.e.,
identical observations for all individuals), but as our data were unbalanced, we used the method
of generalized estimating equations (23) to extend logistic regression into a repeated measures
structure that was appropriate for our non-independent, clustered observations (24). This
allowed us to test the main effect variable (exposure severity) and the covariates (age, sex, etc.)
for association with HBG while allowing for multiple observations for each person. We used
step-wise selection to arrive at a final, parsimonious model and the Wald test to evaluate the
statistical significance of the independent variables. The generalized estimating equations and
logistic regression were implemented by Stata (Version 7, Stata Corporation, College Station,
TX) and crosschecked by SAS (Version 8, Cary, NC).
The potentially confounding covariates that were tested for association with HBG
included Trip Number, Monitoring Time, Trip Day, Subject, Sex, Age, BMI, and three
variables representing ascent from 40, 30 and 20 fsw to the surface (Ascent40, Ascent30, and
Ascent20). Trip Number was an indicator variable to account for possible differences between
trips. Monitoring Time represented the time in minutes at which the single post-dive Doppler
249
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
monitoring session occurred. An ordinal variable Trip Day accounted for the day of the trip on
which monitoring took place. Because ascent rate and demographic factors appear to influence
DCS and VGE susceptibility (16, 25), we tested Sex, Age, and BMI (weight (kg)/height (m)
2
) as
well as the three variables representing ascent from 40, 30 and 20 fsw to the surface. For
subjects who made multiple trips, Age and BMI were adjusted to reflect appropriate values.
Subject was an ordinal variable defined separately for each individual in the study. We also
tested a first order interaction between Age and Sex as previous data had suggested that the
relationship of age to VGE was different for males than for females (13).
The main effect variable, exposure severity, was taken as the conditional probability of
DCS (cP
DCS
) for each dive, estimated using a probabilistic decompression algorithm from depth-
time profiles recorded by the Suunto dive computers. The algorithm, based upon a mathematical
simulation of in vivo bubble growth (21,26), had been calibrated by the method of maximum
likelihood to the observed DCS outcomes of 3,322 chamber dive profiles completed by air- or
nitrox-breathing, water-immersed, working subjects (21). Profiles that included repetitive dives
were each completed within a 24 hr period.
The conditional probability of DCS for a dive is the probability of DCS during or after
the dive, subject to the condition that the diver is DCS-free when first leaving the surface on the
dive. For this study, an individual dive was defined as any exposure to >5 fsw after a surface
interval of >10 minutes. The conditional probability of DCS for the dive was thus computed as a
function of the integral of the instantaneous DCS risk over the time period beginning with first
descent through 5 fsw on the dive, and ending with first descent through 5 fsw on the next dive,
or with decay of the instantaneous DCS risk to zero after the last dive in a day (21,26,27). For
each dive, cP
DCS
increased monotonically from zero at dive start to a maximum value at start of
the next dive, or at a time within 12 hrs of surfacing from the last dive of the day (Fig. 1). We
used the post-dive cP
DCS
maximum as the measure of each dive’s exposure severity.
g
0 100 200 300 400
0
50
100
0.0
0.5
1.0
depth(fsw)
cPDCS
Time (min)
Depth (fsw)
cP
DCS
Fig. 1. Depth-time profile recorded by Suunto SME dive computers for one subject on one dive
day. For dives 1 and 2, the maximum depths were 74 and 70 fsw, the dive times were 51 and 60
min, and the maximum cP
DCS
values were 0.81 and 0.31. The surface interval was 120 minutes
For repetitive dives, cP
DCS
took into account all prior dives and surface intervals beginning with
the first dive of the day. cP
DCS
for repetitive dives could be greater or less than for previous
dives depending on the duration of the surface interval and the severity of the repetitive dive.
250
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
RESULTS
Table 1 lists the 6 trips undertaken in the study. A total of 281 individual dives were
successfully monitored in 67 subjects during 101 exposure days out of 497 possible subject dive
days. Dive trips were 6-8 days in duration and the subjects usually dived every day.
Occasionally, a few divers skipped a day in mid-trip and, on one trip (Saba), there was no diving
on Day 4. On rare occasions of illness or injury, more than one dive day may have been missed,
but we did not keep count of non-diving days. Subjects were monitored after at least one dive
and after as many as 13 dives since 12 subjects made multiple trips. There were 46 males (189
monitored dives) and 21 females (92 monitored dives). Mean age was 44.8+
8.5 years (range 32-
68) for males and 44.5+
11.5 years (range 22-64) for females. The median BMI for males was
24.7 kg/m
2
(interquarterial range 23.6-26.0) and 21.0 kg/m
2
(interquarterial range 19.8-21.8) for
females.
Table 1. Trip location, total subjects and total dives monitored in this study by trip and number
of trip days available for monitoring.
Trip
Subjects Dives Trip Days
Galapagos Is., Ecuador 11 43 8
Cayman Is., British West Indies 13 45 5
Cocos Is., Costa Rica (1) 14 47 6
Cocos Is., Costa Rica ( 2) 15 50 7
Saba Is. Netherlands Antilles 13 44 6*
Truk Lagoon, Federated States of Micronesia 15 52 6
* Day 4 of this 6-day trip involved no diving for all divers.
Table 2 lists parameters for the 281 monitored dives. The maximum depth and dive time
were measured by the computers and also by the subjects’ personal watches and depth gauges.
The measurements agreed within 0.2±2.2 fsw and 0.1±2.6 min except for one time discrepancy
(9 min) and three depth discrepancies (8, 9, and 12 fsw) that were beyond three standard
deviations and were unexplained. These may have represented typographical errors or errors in
the subjects’ personal depth gauges since SeaQuest Industries provided calibrated dive
computers.
As illustrated in Fig. 1, the dive profiles were multilevel. In dives to >50 fsw, for
example, 27% of the 3 min recording intervals were within 10 fsw of maximum depth, and 24%
occurred on the final ascent from 40 fsw to the surface. The remaining 49% were to depths
shallower than maximum depth by 11 fsw or more. Only 22 (8%) of exposures were ≤50 fsw.
Of 183 surface intervals between repetitive dives, 33% were 1-2 hrs, 42% were 2-3 hrs, and none
were longer than 6-7 hrs. Few surface intervals were shorter than one hour.
251
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
Table 2. Median (interquartile range) for maximum depth (fsw), dive time (min), and
cP
DCS
by sex, age, and first or repetitive dives.
Males
Females First Dives Repetitive Dives
N
189 92 101 180
Maximum
Depth
90 (74 -107) 77 (62 - 92) 95 (74 - 111) 80 (68 - 99)
Dive Time
42 (36 -51) 42 (33 - 48) 39 (30 - 45) 45 (36 - 51)
cP
DCS
0.8 (0.5 - 1.3) 0.6 (0.4 - 1.0) 0.5 (0.4 - 0.7) 1.1 (0.6 - 1.5)
The median of the maximum depth for all subjects was 85 fsw (interquartile range
70-130 fsw). The median maximum depth for males was 13 fsw deeper than for females
(Table 2), and males accounted for 29 of the 30 exposures in excess of 120 fsw. For
depths to less than 120 fsw, the median maximum depth for males was 8 fsw deeper than
for females. The median maximum depth for the first dive of the day was 15 fsw deeper
than for repetitive dives.
Table 3 lists the Spencer Doppler scores by sex for the 281 exposures. A score
greater than 0 was noted for 61 of 67 subjects (91%) and after 205 of the 281 dives
(73%). Doppler signals were graded as 2 or 3 for 102 dives (36.3%). No Grade 4 VGE
was detected. The median post-dive time of Doppler monitoring was 43 min
(interquartile range 35-54 min).
No subject developed DCS on any monitored day although one subject was
treated for suspected DCS following several dives on a day that was not monitored.
Table 3. Distribution of Doppler scores by sex.
Spencer
Score
Male Female Total
0 38 (20%) 38 (41%) 76 (27%)
1 70 (37%) 33 (36%) 103 (37%)
2 61 (32%) 10 (11%) 71 (25%)
3
20 (11%) 11 (12%) 31 (11%)
Total 189 92 281
Associations with HBG
Variables significantly associated with HGB included the main effect, cP
DCS
, and
the covariates Trip Number, Age, Sex, Trip Day, and Monitoring Time. A square root
transformation of cP
DCS
improved its association with HGB. There was a first order
interaction between Age and Sex and a small but significant association with two of the
six trips. HGB was not associated with BMI, or with ascent time from 40, 30 or 20 fsw.
The Suunto computer recorded ascent time to the nearest minute and may have been too
imprecise to show any effect.
252
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
When we repeated the above analysis for Spencer Score 0 versus Spencer Score
1, 2, or 3, the results were similar as for HBG (Spencer Scores 0 or 1 versus 2 or 3)
indicating that the observed relationships of HGB to cP
DCS
and covariates were
applicable at several levels of bubbling. There was an insufficient number of Grade 3
scores to assess Spencer Scores 0,1, or 2 versus Spencer Score 3.
Dive Profile Severity
The observed incidence of HBG=1 (expressed in percent as the number of dives
with HBG=1 divided by total dives) against the quartile medians of cP
DCS
is shown in
Figure 2. The incidence of HBG=1 increased over the range of cP
DCS
from 13% to 56%.
The relationships of cP
DCS
to HGB=1 were coincident for first and repetitive dives (Fig.
2) indicating that cP
DCS
successfully accounted for HBG=1 after repetitive diving during
a single day. While cP
DCS
was generally greater for repetitive dives than for first dives
(Fig. 2), this was not necessarily so (see Fig.1) and may have reflected the diving style of
the observed population sample rather than a general characteristic of repetitive diving.
The relative odds of HBG=1 increased 2.04 times (95% CI=1.38-14.20) for each unit
increase in cP
DCS
(p<0.001).
g
0.0 0.5 1.0 1.5 2.0
0
20
40
60
80
100
First Dives
Repetitive Dives
26
25
25
25
n=4545
45
45
Exposure Severity (cP
DCS
)
% High Bubble Grade
Fig. 2. The relationship of exposure severity (cP
DCS
) to the observed incidence of
HBG=1 for first and repetitive. Bars represent the standard error of the binomial
distribution. The number of dives on which each point is based is noted next to the bar.
Multiday Diving
The observed incidence of HBG=1 for each Trip Day is shown in Fig 3. (Trip Days 7
and 8 were omitted as there were only 9 total dives on Trip Day 7 and 6 dives on Trip Day 8.)
HBG=1 was about 20% greater for repetitive dives than for first dives on all but Trip Day 4.
The crossover on Trip Day 4 is unexplained as the first dives on that day appeared similar to
first dives on other days. For both first and repetitive dives, the % HBG=1 decreased by 20-30%
over the duration of the trip. The relative odds of HBG=1 decreased 0.68 times (95% CI=0.56-
0.84) for each succeeding Trip Day (p<0.001).
253
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
The observed incidence of HBG=1 decreased for a given cP
DCS
as Trip Day increased is
shown in Fig 4. In other words, cP
DCS
did not account for the decrease in HBG=1 that occurred
with Trip Day. At a given cP
DCS
, the incidence of HBG=1 was always lower for Trip Days 5-8
than for Trip Days 1-2 with the highest observed incidence in Trip Days 5-8 approximately
equal to lowest for Trip Days 1-2.
Fig 3
0123456
0
20
40
60
80
100
First dives
Repetitive dives
n=34
29
22
21
31
23
34
11
16
14
17
14
Day of Trip
% High Bubble Grade
Fig. 3. Effect of multi-day diving (Trip Day)
on the observed incidence of HBG=1 for first
and repetitive dives. Day 7 and 8 were
omitted as there were only 9 dives on Day 7
and 6 dives on Day 8.
Fig 4
0.0 0.5 1.0 1.5 2.0
0
20
40
60
80
100
Day 1-2
24
Day 5-8
Day 3-4
Exposure Severity (cP
DCS
)
n=25
25
25
24
24
23
23
22
22
22
22
% High Bubble Grade
Fig. 4. Effect of exposure severity (cP
DCS
)
on the observed incidence of HBG=1 by
Trip Day
.
Age and Sex
The observed incidence of HBG=1 for males and females against age is shown in Fig 5.
The incidence of HBG=1 increased from 0-52% for (quartile mean age range 31-57) and from
35-60% for men (quartile mean age range 36-54). There was an effect of age on gender (p=0.02)
such that the relative odds of HBG=1 was significantly greater for women at 5.85 fold per
decade (95% CI=2.57-13.32) compared to a 1.53 fold increase per decade for men (95% CI=1.2-
1.95).
As the mean maximum depth for males was 13 fsw greater than for females and since
males undertook 29 of the 30 dives >120 fsw, we repeated the analysis after removing all dives
254
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
to >120 fsw. We found no substantive change in the results indicating that cP
DCS
successfully
accounted for changes in exposure severity due to depth.
To estimate the amount of unique subject variance, we computed a coefficient of
determination (R
2
) for the Subject variable and contrasted this with R
2
computed for the
demographic variables Age and Sex. The R
2
due to Subject was 0.533 while that due to Age and
Sex was 0.12. This indicates that 22.5% of the explained variance was attributable to Age and
Sex while the remaining 77.5% was unique to the individual or due to factors not in the model.
g
30 40 50 60
0
20
40
60
80
100
Males
Females
n=40
40
23
40
40
23
23
23
Age (years)
% High Bubble Grade
Fig. 5. Effects of age and sex on the observed incidence of HBG=1.
Delay to monitoring
The observed incidence of HBG=1 as a function of quartile medians for time to Doppler
monitoring is shown in Fig 6. As before, the incidence of HBG=1 was greater for repetitive
dives than for first dives. The relative odds of HBG=1 decreased by a factor of 0.78 for each 10
minute increase in Monitoring Time (p=0.019; 95% CI 0.63-0.96). The maximum observed
incidence of HBG=1 occurred at 35-45 minutes after surfacing, in agreement with published
reports that maximum Doppler scores occurred within the first hour (4). Figure 6 indicates that if
monitoring is delayed 25-30 min after peak VGE occurrence, the VGE incidence could be
underestimated by as much as 20%.
255
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
30 35 40 45 50 55 60 65
0
20
40
60
80
100
First Dives
Repetitive Dives
Delay to Doppler (min)
n=45
25
45
45
45
25
25
26
% High Bubble Grade
Fig. 6. Effects of time to Doppler monitoring on the observed incidence of HBG=1 for first and
repetitive dives.
DISCUSSION
We used dive computers to record the depth-time profiles of recreational divers and
monitored the divers with Doppler ultrasound for post-dive VGE. VGE signals were detected in
91% of the divers and in 73% of the monitored dives.
In 1,726 DCIEM dives (4), the DCS incidence was 1.1% DCS for Spencer Grades 1 or 2,
6.3% for Grade 3, and 9.7% for Grade 4. If the same relationship of VGE to DCS had held in our
study, we would have expected about four DCS incidents. However, many of these DCIEM
dives required decompression stops, while our dives required no stops. Thus, the type of diving
may condition the relationship of VGE to DCS.
cP
DCS
was associated with HBG (p<0.001) despite being based on DCS data from
chamber dive trials rather than on VGE data. While cP
DCS
was predictive of VGE, we do not
suggest a physiological relationship between estimated DCS probability and Doppler score
although the two may be related to a common third entity such as depth-time exposure. cP
DCS
was equally successful in describing VGE after first and repetitive dives in a single day (Fig. 2),
and cP
DCS
was higher for repetitive dives than for first dives indicating that repetitive dives, as
conducted in the study, had greater exposure severity than did first dives. The 1.5-fold decrease
per day in the incidence of HBG=1 for a given cP
DCS
over the 6-8 days of the trips, on the other
hand, demonstrated that cP
DCS
could not account for the effects of multi-day diving (Fig. 3). This
was not unexpected since the model used to estimate cP
DCS
was calibrated to data that included
no multi-day profiles of this kind (21).
The observed decrease in incidence of HBG=1 recalls the daily reduction in DCS
incidence reported for repeated exposures of compressed air workers (28) and suggests a similar
phenomenon for VGE. Our observations disagree with a previous experiment that failed to find
reduced Doppler scores in subjects who dived each day for 12 consecutive days (29). This
256
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
discrepancy may reflect the different nature of the dive exposures that were single-level chamber
dives to 148 fsw for 28 minutes followed by 32 minutes of decompression.
Figure 5 indicates that males and older divers had a higher incidence of HBG=1 than did
females and younger divers. This agreed with previous reports that males have higher Doppler
scores and that Doppler scores increase with age (30,31). We also found an interactive effect of
age on gender suggesting that the increase in incidence of HBG=1 in females with age is higher
than that for males. In this study, subjects differed in age from other diving populations. For
example, our subjects were older (mean age 45 years) than 1,111 patients treated for diving
injuries in 1989-91 (mean age 36 years) (32,33,34).
BMI was not associated with HBG=1 although other reports have shown bubble score to
be associated with weight, percent body fat, and maximum oxygen uptake (19). Age was found
to be positively associated with BMI (Spearman rho=0.123, p<0.05), suggesting that age and
BMI were collinear and that age absorbed the variance contained in BMI.
The practical significance of Doppler-detected bubbles in the etiology of DCS has been
debated since the 1970s when VGE were first shown to be common after diving. Patent foramen
ovale (9,10,11) or other routes of VGE passage into the arterial circulation (5,6,7) are potential
mechanisms linking VGE with neurological DCS, and observational studies of VGE such as
reported here might ultimately prove useful for decreasing the incidence of neurological DCS. In
addition, we suggest that confounding variables age, sex or multi-day diving found to be
associated with VGE incidence here might also be useful in improving the predictive power of
DCS probability models.
Limitations
In this field study, we did not follow the same divers on consecutive days, we monitored
the divers only once after each dive and minimally interfere with divers’ activity. In doing so,
we widen the sample of divers who could be studied and observed divers close to their normal
activity. However, we were then unable to employ methodological design controls as would be
in found in a laboratory study but, instead, used statistical methods to account for the variations
inherent in field studies.
Subjects were monitored only once after each dive which might have underestimated the
maximum bubble score when delay to monitoring exceeded 30-45 minutes (Fig. 6). As subjects
were not monitored every day, the effects of multi-day diving could not be observed directly,
although statistical analysis allowed indirect assessment.
Estimated cP
DCS
values are based on a probabilistic model of DCS incidence in which
DCS risk is a function of prevailing bubble volumes in a series of modeled tissue compartments.
These volumes, and the associated DCS risk, decay to zero within any 12-hour post-dive surface
interval, but some modeled tissue compartments may remain gas-supersaturated beyond 12
hours. Such “residual gas” effects would increase the risk of a given dive from day-to-day, and
thereby worsen the model’s agreement with the observed decrease in HBG=1 incidence by Trip
Day (Figure 3). However, only single-day data were available from any given diver to make the
cP
DCS
estimates, which required the assumption that a given day’s diving was independent of
previous days’ diving. This obviated the influence of any potential day-to-day residual gas
effect in the estimates.
Subjects occasionally skipped a day of diving (most often mid-week) although we did not
keep track of days that were skipped. The effect of a skipped day would have been to lessen the
decrease in the incidence of HBG=1 that was observed to occur during successive Trip Days
257
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
(Fig. 3). That a significant multi-day effect was apparent despite skipped days is evidence of the
strength of the effect.
The Suunto decompression computers recorded the deepest depth every three minutes
and produce a reported mean depth greater than true mean depth. Resulting cP
DCS
estimates
were based on the assumption that all the time in any 3 min segment was spent at the maximum
depth recorded of that segment.
The results were based on a small population sample in an uncontrolled study design and
require independent verification.
CONCLUSION
Recreational divers were monitored for Doppler detected venous gas emboli following
multi-day, multi-level, repetitive dive exposures. A probabilistic model of DCS risk provided an
estimate of decompression stress based on depth-time profiles recorded by dive computers.
VGE increased with decompression stress as well as diver age but females showed a stronger age
effect than males. Detectable bubbles decreased with multi-day diving suggesting an adaptive
effect.
ACKNOWLEDGMENTS
This study was carried with the support of the Divers Alert Network. The authors would like to
thank the recreational divers who willingly volunteered as subjects. Without their commitment,
this study would not have been possible.
LITERATURE CITED
1. Spencer MP, Campbell SD. Development of bubbles in venous and arterial blood during hyperbaric
decompression. Bull Mason Clinic. 1968; 22:26-32.
2. Catchpole HR, Gersh I. Pathologic factors and pathological consequences of decompression
sickness. Physiol Rev. 1942;27:360-397.
3. Spencer MP. Decompression limits for compressed air determine by ultrasonically detected blood
bubbles. J Appl Physiol. 1976; 40:229- 235.
4. Nishi RY. Doppler evaluation and bubble detection. In Bennett PB, Elliott, DH. eds. The
physiology and medicine of diving, 4
th
ed. W. B. Saunders Company Ltd. London, pp 433-453.
1993.
5. Gait D, Miller KW, Paton WDM, Smith EB, Welch B. The redistribution of vascular bubbles in
multiple dives. Undersea Biomed Res. 1975; 2:42-50.
6. Butler BD, Hills BA. The lung as a filter for microbubbles. J. Appl. Physiol. 1979;47:537-543.
7. Vik A, Jenssen BM, Brubakk AO. Arterial gas bubbles after decompression in pigs with patent
foramen ovale. Undersea & Hyperbar Med. 1993;20:121-31.
8. Pearson AC, Labovitz AJ, Tatineni S, Gomez CR. Superiority of transesophageal echocardiography
in detecting cardiac source of embolism in patients with cerebral ischemia of uncertain etiology. J
Am Coll Cardio. 1991;17:66-72.
9. Moon RE, Camporesi EM, Kisslo JA. Patent foramen ovale and decompression sickness in divers.
Lancet 1989;1(8637):513-514.
10. Wilmshurst PT, Bryne JC, Webb-Pole MM. Relation between interatrial shunts and decompression
sickness in divers. Lancet 1989;2(8675):1302-1306.
11. Germonpre P, Dendale P, Unger P, Balestra C. Patent foramen ovale and decompression sickness in
sports divers. J. Appl. Physiol. 1998; 84: 1622–1626.
258
Rubicon Foundation Archive (http://rubicon-foundation.org)
UHM 2002, Vol. 29, No. 4 – VGE in recreational divers
259
12. Pilmanis AA. Ascent and silent bubbles. In Lang MA, Egstrom GH, eds. Proceedings of the AAUS
Biomechanics of Safe Ascents Workshop. Costa Masa, CA: American Academy of Underwater
Sciences, 1990:65-71.
13. Dixon GA, Krutz RW, Fischer JR. Decompression sickness and bubble formation in females
exposed to a simulated 7.8 PSIA suit environment. Aviat. Space Environ Med. 1988; 59:1146-1149.
14. Smith KH, Stayton RL. Hyperbaric decompression by means of bubble detection. ONR contract
#N000 14 69-C-0402. 1978.
15. Neuman TS, Hall DA, Linaweaver PG. Gas phase separation during decompression in man:
ultrasound monitoring. Undersea Biomed Res. 1976; 3:121-130.
16. Uguccioni DM, Vann RD, Smith LR, Butler BD, Rove DB, Roer RD. Effect of safety stops on
venous gas emboli after no-stop diving. Undersea Hyperbar Med. 1995;22 (Suppl):53.
17. Jankowski LW, Nishi RY, Eaton DJ, Griffin AP. Exercise during decompression reduces the amount
of venous gas emboli. Undersea Hyperbar Med. 1997; 24: 59-65.
18. Dunford R, Hayward J. Venous gas bubble production following cold stress during a no-
decompression dive. Undersea Biomed Res. 1981; 8:41-49.
19. Carturan D, Boussuges A, Burnet H, Fondarai J, Vanuxem P, Gardette B. Circulating venous
bubbles in recreational diving: relationships with age , weight, maximal oxygen uptake and body fat
percentage. Int J Sports Med. 1999;20:410-414.
20. Hamilton RW, Rogers RE, Powell MR, Vann RD, Dunford RG, Spencer MP, Richardson D. The
DSAT recreational dive planner. Diving Science and Technology, Inc. and Hamilton Research, Ltd.
1994.
21. Gerth WA, Vann RD. Probabilistic gas and bubble dynamics models of decompression sickness
occurrence in air and nitrogen-oxygen diving. Undersea Hyperbar Med. 1997; 24:275-292.
22. Weathersby PK, Survanshi SS, Homer LD, Parker E, Thalmann ED. Predicting the time of
occurrence of decompression sickness. J Appl Physiol. 1992;72:1541-1548.
23. Burton P, Gurrin L, Sly P. Extending the simple linear regression model to account for correlated
responses: an introduction to generalized estimating equations and multi-level mixed modeling. Stat.
Med. 1998; 17: 1261-1291.
24. Liang KV, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;
73: 13-22.
25. Vann RD, Thalmann ED. Decompression physiology and practice. In Bennett PB, Elliott, DH. eds.
The physiology and medicine of diving, 4
th
ed. W. B. Saunders Company Ltd. London, pp 376-432.
1993.
26. Gerth WA, Vann RD. Development of iso-DCS risk air and nitrox decompression tables using
statistical bubble dynamic models. National Oceanic and Atmospheric Administration, Office of
Undersea Research, Rep NA46RU0505, Bethesda MD. 1996.
27. Gerth WA. Overview of survival functions and methodology. In: Weathersby PK and Gerth WA, eds.
Survival analysis and maximum likelihood techniques as applied to physiological modeling. Undersea and
Hyperbaric Medical Society, Bethesda, MD (in press).
28. Walder, DN. Adaptation to decompression sickness in caisson work. Biometerology. 1968;11: 350-
359.
29. Eckenhoff RG, Hughes JS. Acclimatization to decompression stress. In Proc. VIIIth Symp.
Underwater Physiology, pp. 93-100. Ed. A.J. Bachrach & M. M.Matzen. Bethesda, MD.: Undersea
Medical Society. 1984
30. Dunford RD, Wachholz C, Huggins KE, Vann RD, Gerth W, Bennett PB. DCS risk and Doppler
bubbles in sport divers. Undersea Biomed Res. 1993;20(Suppl):127.
31. Dunford RG, Wachholz C, Fabus S, Huggins K, Mitchell, P, Bennett P. Doppler analysis of sport
diver profiles. Undersea Biomed Res. 1991;18(Suppl):95.
32. 1989 Report on Diving Accidents and Fatalities. Divers Alert Network, Durham, NC. 1991.
33. 1990 Report on Diving Accidents and Fatalities. Divers Alert Network, Durham, NC. 1992.
34. 1991 Report on Diving Accidents and Fatalities. Divers Alert Network, Durham, NC. 1993.
Rubicon Foundation Archive (http://rubicon-foundation.org)
Rubicon Foundation Archive (http://rubicon-foundation.org)
UNDERSEA & HYPERBARIC MEDICAL SOCIETY
10531 METROPOLITAN AVENUE
KENSINGTON, MARYLAND 20895
USA
Phone: 301.942.2980
Fax: 301.942.7804
Email: uhms@uhms.org
Website: www.uhms.org
Rubicon Foundation Archive (http://rubicon-foundation.org)