$TITLE="Hyperemesis Gravidarum Survey Results";
$MP = "/usr/lib/sendmail -t";
$TO = "frederic@stat.ucla.edu,MFejzo@mednet.ucla.edu";
$CC = "";
?>
Below are some preliminary results, as of March 7, 2002.
(For Feb 20, 2001 results, click
here .)
(Please note that his survey is not supported by, endorsed by, or
affiliated
with UCLA.)
The following is based on the first 267 survey responses (from 208
subjects). We are currently about halfway through the remaining 220-or-so
additional responses. We feel this is a pretty substantial dataset
considering that few previous studies have looked at more than a few dozen
pregnancies.
Of the 267 pregnancies reported, 53 may not qualify as HG based on the
subjects' reported weight loss and the subjects' response to the question
of whether or not they were diagnosed by a physician. In what follows,
these 53 responses are ignored, leaving 214 pregnancies with HG.
1) Effectiveness of medications.
It is very difficult to assess the effectiveness of medications from survey
data, so we encourage readers to refer whenever possible to clinical
studies rather than the results presented here. Our preliminary reports on
the effectiveness of various drugs are listed in the
table
.
The drugs most commonly used appear to be Promethazine, Prochlorperazine,
Pyridoxine, Ondansetron, and Trimethobenzamide. Of these, our data suggest
that Ondansetron (Zofran) may be the most effective, followed by
Promethazine (Phenergan). The other 3 drugs were effective only for a very
few respondents. Pyridoxine (vitamin B6) appears to be very ineffective.
Of the less commonly used medications, Metoclopramide (Reglan) and
Doxylamine (Unisom) show signs of effectiveness, with 40% and 34%,
respectively, of respondents claiming that they were at least slightly
effective. However, this is based on very few respondents (17 and 26,
respectively). Meclizine (Antivert), Corticosteroids, Diclectin, and
medical Marijuana had similarly high percentages reporting effectiveness,
but for these medications the sample sizes are extremely small.
2) Safety of medications.
It is extremely difficult to assess the safety of medications from survey
data, so we encourage readers to refer whenever possible to clinical
studies rather than the results presented here. (Note however that clinical
trials of drugs on pregnant women with HG are scarce; often these drugs are
given safety ratings based on the results of their use in clinical trials
on patients with other conditions.)
a) Side Effects.
From the preliminary survey results it appears that all the drugs have
few side effects: only four medications (Promethazine, Cyclizine, Scophlamine, and
Prochlorperazine) had reports of adverse reactions in more than 10% of
pregnancies, and even among these three drugs none except Promethazine had a rate higher than
12.5%. Further, reactions to Cyclizine and Scophlamine were only reported
in 1 out of 8 cases each, so the estimated rate of 12.5% may well be too
high; the data are certainly insufficient so far to indicate any major
cause for concern. Pyridoxine, Doxylamine, Metoclopramide,
Trimethobenzamide, and Zantac all seemed surprisingly free of side-effects,
with no adverse reactions reported out of a combined 179 reports of
use.
b) Pregancy outcomes.
Though all the drugs appear to be safe in terms of adverse side effects,
there remains the question of the safety of the drugs on the fetus. This
is especially difficult to investigate, since pregnant women with more
severe HG will likely receive different prescriptions (in terms of
medications and doses) than those with less severe cases. Hence the effect
of the drug on the fetus is confounded with the effect of the severity of
HG on the fetus. On the other hand, since studies have shown there is no
positive correlation between nausea and vomiting in pregnancy and fetal
loss (in fact a slight negative correlation has been reported), and since,
to our knowledge, no study has indicated a correlation between the severity
of HG and fetal loss, the extent of the impact of this confounding factor
is unclear.
One way to investigate safety is to investigate the proportion of
pregnancies
that make it to term and the proportion that are lost in the 2nd or 3rd
trimesters --
however, note that we are not addressing possible long term drug effects
to the fetus such as cognitive impairment and other developmental
abnormalities that may reveal themselves after birth. We present a
summary of the preliminary results here, but again the reader should bear
in mind that it is unclear whether the fetal losses
reported here are causally related to the drug prescribed, rather than
severity of HG or some other confounding factor, i.e. some difference
between the patients receiving this drug from other patients. For
instance, 3 of the 4 pregnancies where the patient took Scophlamine
resulted in fetal loss or abortion. However, Scophlamine is rarely
prescribed: it seems plausible that these patients were only prescribed
Scophlamine because nothing else worked; the severity of their HG condition
may well be the culprit, not Scophlamine. By contrast, the results for
Chlorpromazine (Thorazine) are quite impressive: out of 17 completed
pregnancies in which this drug was used, only one did not result in
successful birth and that one was aborted. For most of the other drugs,
the successful birthrate is between 75% and 85%. Note that we distinguish
fetal losses occurring after the first 12 weeks from those in first 12
weeks because the latter are so common among ALL pregnancies and are thus
very difficult to attribute to HG or to any particular medication. One
final note: the successful birth rate for Pyridoxine (vitamin B6) in our
survey results is 90%. Of the 10% that are unsuccessful, only 3 (4.8%) are
attributed to fetal loss, and the other 5 (7.9%) were aborted. The results
suggest that perhaps vitamin B6 is very safe but very ineffective, and
since it is commonly prescribed to patients with HG, perhaps this
medication can be viewed as a sort of placebo, i.e. as a control group with
which to compare other medications. If such a comparison is fair, then the
preliminary survey results suggest that several of the prescriptions have
significant positive correlations with fetal loss, especially
Trimethobenzamide, Promethazine, and Zantac. Again, we must emphasize that
it is unclear whether this correlation may be attributed to the use of the
medication rather than some other confounding factor such as HG
severity.
3 Other summaries of our survey results
a) Job Discrimination.
49 (22.9%) reported discrimination based on HG at work. 80 (37.4%)
reported no discrimination, 1 (0.47%) said she quit work, and 84 (39.3%)
did not respond to this question.
b) Planning.
63 (29.4%) of the pregnancies were reportedly unplanned, and of these 6
(9.5%) were unwanted.
c) Psychology.
16 respondents reported having had a psychological disorder prior to their
pregnancy with HG. 112 said they had none, and 86 (40.2%) had no
comment.
d) Eating Disorders.
Only 8 respondents reported ever having had an eating disorder, with 111
respondents reporting never having had an eating disorder and the remaining
95 (44%) not commenting.
e) Health.
37 respondents reported chronic health problems, 130 reported no chronic
health problems, and 47 (22%) did not respond.
f) The pill.
154 respondents said they had taken the birth control pill, 18 said they
never took it, and 42 (19.6%) did not respond. Out of the 154 who took the
pill, 54 (35.1%) said they had no bad reaction, 19 (12.3%) had no response,
and the remaining 52.6% reported various negative reactions such as
nausea.
g) Changing plans.
99 respondents said they have changed their pregnancy plans because of HG.
30 said they have not, 36 indicated they are not sure, and 49 (22%) did not
respond.
h) Hospitalization.
69 respondents were hospitalized more than once due to HG. 48 said they
were not hospitalized more than once with HG, and 97 (45.3%) did not
respond.
i) Racial identity.
Out of 167 qualified respondents, 120 (71%) are white, 5 (2.9%) are black,
9 (5.4%) are Hispanic, 6 (3.6%) are Asian, and 27 (16%) did not indicate
racial identity.
j) NVP after pregnancy.
Out of 214 qualified pregnancies with HG, 80 said they still had nausea and
vomiting after the pregnancy was over, 85 said it stopped after the
pregnancy was over, and 49 (22.9%) has no comment.
k) Smoking.
10 responses indicated smoking between 2 and 10 cigarettes per day, 3
smoked 20/day, 1 smoked 40/day, 184 responses indicated no smoking before
the pregnancy, and 16 (7.5%) did not respond. During pregnancy, 1 subject
reported smoking 40 cigarettes per day, 1 smoked 5/day, 1 smoked 2/day, 160
reported not smoking, and the remaining 51 (23.8%) did not respond.
l) Alcohol.
23 responses indicated drinking between 21/4 and 1 alcoholic drink per day
prior to pregnancy, 3 reported 2 drinks/day, 1 reported 2.5 drinks/day, 169
responses indicated no alcohol drinking before the pregnancy, and 18 (8.4%)
did not respond. All 174 subjects who responded said they did not drink
any alcohol during pregnancy.
m) Severity Trends.
Of the 167 qualified respondents, 51 respondents said the HG got more
severe in subsequent pregnancies, 40 said not more severe, 2 said the same,
and the remaining 74 did not respond.
4) Drug responses by fetal gender.
To investigate whether the safety and effectiveness of certain
prescriptions may depend on the gender of the fetus, we divided Table 1
into 2 other tables (see
here), one for cases where the fetus was a girl and one for
cases where the fetus was a boy. No major differences were apparent from
our preliminary results.
(Please note that his survey is not supported by, endorsed by, or
affiliated
with UCLA.)