Escolar Documentos
Profissional Documentos
Cultura Documentos
doi: 10.1111/j.1440-1681.2007.04838.x
Original
Articles
Blackwell
Publishing
SM
Nitroglycerine
Verstegui
et
vs nifedipine
al. Asia in pre-eclampsia
SUMMARY
1. Information regarding the use of continuous i.v. administration of nitroglycerine as an antihypertensive agent in the
management of pre-eclampsia is scarce. In the present study, i.v.
nitroglycerine or sublingual nifedipine were administered to 32
women with severe pre-eclampsia who were being managed with
controlled plasma volume expansion and MgSO4 loading and
maintenance doses. Maternal blood pressure and heart rate
responses, fetal heart rate responses and perinatal fetal
maternal adverse effects were evaluated using classical parametric and non-parametric data analysis and data modelling by
mixed models.
2. An important hypotensive response was observed in both
groups, although this reponse was greater, faster and exhibited
less variability (more precision) in the nitroglycerine-treated
group. Heart rate also increased in both the nitroglycerine- and
nifedipine-treated groups (4.6 4.4 vs 8.6 5.3 b.p.m., respectively), although the increase in the nifedipine-treated group was
almost twofold that in the nitroglycerine-treated group. There
were no significant changes in fetal heart rate in response to
vasodilator therapy. The frequency of perinatal fetalmaternal
adverse effects was similar in both groups at 40% and the
adverse effects observed included flushing, headache, palpitations
and nausea.
3. In conclusion, i.v. infusion of nitroglycerine is an effective,
safe and alternative therapy for severe pre-eclampsia.
Key words: antihypertensive therapy, nifedipine, nitroglycerine,
severe pre-eclampsia.
INTRODUCTION
Pre-eclampsia is a multisystem disorder of vascular endothelial
function related to pregnancy that is characterized by persistent
Correspondence: Manuel Rodrguez-Martnez, Department of Physiology
and Pharmacology, Faculty of Medicine, Universidad Autnoma de San Luis
Potos, PO Box 1521-B, 78210 San Luis Potos, SLP Mxico. Email:
rodrigum@uaslp.mx
Received 14 May 2007; revision 12 August 2007; accepted 23 September
2007.
2007 The Authors
Journal compilation 2007 Blackwell Publishing Asia Pty Ltd
581
METHODS
A triple-blind, randomized, controlled clinical study was registered (Mexican
Social Security Institute (IMSS): 2003-474-05) and performed from March
2004 until March 2005 at the General Hospital No. 1, IMSS (San Luis Potos,
SLP Mxico) following CONSORT recommendations.18 The Hospital Ethics
Committee, in full accordance with International Ethical Guidelines for
Biomedical Research Involving Human Subjects (CIOMS/WHO), approved
the study. All women (1540 years of age) who presented with a pregnancy
of greater than 24 weeks, with or without labour and with singleton or
multiple fetuses, were evaluated obstetrically.
When the principal investigator (SMV) diagnosed patients with uncomplicated severe pre-eclampsia (without imminence of eclampsia or clinical
manifestations of target organ damage, as per the ACOG criteria19) and found
no history of chronic hypertension, antihypertensive therapy or life-threatening
fetal heart beat changes, as determined by cardiotachographic monitoring
(Corometrics Medical Systems, Walling Ford, CN, USA), the women were
invited to participate in the study. They were then informed about the nature,
scope and risks of the study and only those who gave written consent to
participate were included.
Selected women were hospitalized and routine blood and urine chemistry
tests were performed. With women in a seated position, blood pressure
and heart rate were measured using a fitted arm cuff (arm circumference
< 30 cm = regular cuff size; arm circumference > 30 cm = large cuff size)
and an automatic self-calibration monitor (Multiparameter Patient Monitor,
MEC 509 B; Mindray, Shenzhen, China), which had been validated previously
over the range 0300 mmHg against a mercury manometer and provided the
measurements of blood pressure (including mean) and heart rate. These
parameters, along with fetal heart rate, were recorded on a data-collection
sheet until the end of the trial.
An i.v. infusion of Ringers lactate solution was initiated (8 mL/kg per h)
in order to induce a small (< 1% of bodyweight) but controlled extracellular
volume expansion. One hour later, the infusion rate was reduced to 1 mL/kg
per h and a loading dose (4 g /250 mL D5W ) of MgSO4 (10%; Pisa, Guadalajara,
Mxico) was i.v. administered over 30 min. This loading dose was followed
by an i.v. infusion of 1 g/h MgSO4 (10 g /1000 mL D5W per 100 mL per h)
for up to 8 h post-partum. Subjects were monitored for recognized signs of
MgSO4 toxicity every 30 min after the loading dose and until the trial
finished. Women were excluded from the study if, during the MgSO4 loading
dose, they showed: (i) a decrease in systolic, diastolic or mean arterial
pressure > 10 mmHg; or (ii) any another sign of MgSO4 toxicity (i.e. sedation,
hypoventilation and hyporeflexia).
Thirty-two women were randomly assigned to either of the two treatment
groups: nitroglycerine or nifedipine. A statistician (PBM) generated the
balanced blocked randomization sequence (4 8) using the R V 2.01
program (http://www.R-project.org). The group to which each woman was
allocated was inscribed in opaque, sealed and progressively numbered
envelopes. An external collaborator assisted by opening the envelopes as each
women entered the study sequentially, preparing the appropriate active
substances and placebos and supplying, in a carefully blinded mode, the i.v.
and sublingual preparations (active and placebo) to the principal investigator
and external auxiliary nurse, respectively, who then administered the
preparations to the patients. The auxiliary nurses instructed women not to
swallow voluntarily for 20 min by asking them to keep their mouths open
and to refrain from discussing taste perceptions, if any.
In the nitroglycerine group, 5 mg/min nitroglycerine (25 mL/min) was
administered by continuous i.v. infusion with increases in dose of 5 mg/min
(25 mL/min) every 5 min until the therapeutic goal was reached, which was
a decrease in systolic blood pressure (SBP) to < 140 mmHg but not < 120 mmHg
and a decrease in diastolic blood pressure (DBP) to < 100 mmHg but not
< 80 mmHg within the first hour of treatment. The nitroglycerine (USP;
American Regent Laboratories, Shirley, NY, USA), was prepared by diluting
the 50 mg/10 mL ampoule in 240 mL D5W and was administered using
a Flow Gard 6201 single-channel pump (Baxter, Derrfield, IL, USA) and
a Lifeshield latex-free infusion set (Abbott, North Chicago, IL, USA). Simultaneously, a sublingual placebo preparation (100 mL normal saline) was
administered every 30 min. In the nifedipine group, the content (100 mL) of
a 10 mg capsule of nifedipine (Adalat; Bayer, Distrito Federal, Mexico, DF)
was drawn up into an insulin syringe and deposited sublingually, every
30 min as required. Simultaneously, a placebo i.v. infusion (D5W) was
administered (25 mL/min), with increases of 25 mL/min every 5 min until
the therapeutic goal had been reached. All the study participants, including
subjects, the principal investigator, the auxiliary nurses and the data analysts,
were blinded to the group allocations and the agents administered during
the trial.
Once the vasodilator therapy was initiated, maternal blood pressure and
maternal and fetal heart rates were measured every 5 min throughout the
following hour. If the therapeutic goal could not be reached in that time, the
trial was suspended and considered a treatment failure, in which case i.v.
hydralazine (hydralazine HCl; Americam Regent Laboratories) was used.20
Previously, a pilot study had been undertaken (by SMV and AGM) to
standardize blood pressure and heart rate measurements with the automatic
self-calibration monitor described above. In that preliminary study, two
measurements (0 and 10 min) were performed in 20 healthy pregnant
volunteers who had maintained a resting sitting position for 10 min. Calculated
blood pressure intraclass correlation coefficients (ICC) for intra- and
interobserver agreements21 were 0.96 and 1.0, respectively. The sample size
was calculated using blood pressure data from the pilot study and the equation
described by Diggle et al.:22
m = (2(za + zQ)2 s2 (1 r))/(n s2x d2)
where a is the probability of Type I error (= 0.05), d is the smallest clinical
diference that it is important to detect (5 mmHg), Q is the probability
type II error (= 1 power = 1 0.8 = 0.2), s2 is the non-explained variation
(20.53 mmHg), n = 13 measurements/patient, r is the serial correlation
coefficient (= 0.5194) and s2x is the variation within the xj (= 0.0972) for
m = 16 subjects per group.
Primary outcomes were the time-course of maternal blood pressure,
maternal heart rate and fetal heart rate. Secondary outcomes were the
perinatal variables, maternal adverse effects caused by vasodilator therapy
and mean arterial pressure (MAP) responses to 30 min MgSO4 loading dose
administration according to the final group allocations.
Statistical analysis
Statistical analysis was based on two randomly allocated groups (an intentionto-treat analysis), to a 95% confidence interval (CI) and using the R V 2.01
statistical program (http://www.R-project.org). All values reported are the
meanSD or proportions as percentages. Classical statistical tests were used
for two independent samples and paired samples after testing for normality
(ShapiroWilk) and homoscedasticity (Brown Forsythe).23 Classical statistical
tests were used for nominal and ordinal data.23 To identify important
explanatory variables of the response variables, the data from the 32 women
were assessed with mixed models24 and recalculated with linear models in
order to obtain r2 (group explained variation) and h2 (individually explained
variation), because such results are impossible to obtain from mixed models.
RESULTS
Figure 1 shows the flow of participants through each stage of the
trial. Baseline characteristics were comparable between the two
groups (Table 1). The effects of the MgSO4 loading dose were
analysed retrospectively according to the final group allocations.
Women finally allocated to the nitroglycerine group had a significantly
greater reduction in MAP (D) after the MgSO4 loading dose than
those who were finally allocated to the nifedipine group, although
absolute MAP values in both groups after the treatment were similar
(Table 2). The modelling of DMAP indicated that 15% of this variation
582
S Manzur-Verstegui et al.
Table 2 Retrospective analysis of mean arterial pressure before and 30 min
after MgSO4 loading dose, according to final allocation of women into the
two groups
Before
After
P value
(before vs after)
DMAP
(before after)
Nitroglycerine-treated
group (n = 16)
Nifedipine-treated
group (n = 16)
P value
131 3
125 4
< 0.0001
130 3
127 4
< 0.0001
0.618
0.252
0.009
Values are meanSD. For the difference in mean arterial pressure (DMAP)
values in parentheses indicate 95% confidence intervals.
Table 3 Mean arterial pressure at 0 min and 60 min and changes in mean
arterial pressure over the 60 min in both groups
Fig. 1 Trial flow diagram showing the number of subjects actively followed
up at each trial step.
Nifedipine-treated
group (n = 16)
P value
30.4 7.5
79.6 6.2
2.7 1.7
36.9 1.6
29.6 6.7
83.5 6.5
2.3 0.9
37.1 2.8
0.749
0.095
0.612
0.757
167 6
114 3
52 5
168 7
112 2
56 6
0.250
0.910
0.356
131 3
130 3
0.376
0.072
5 (31%)
11 (69%)
0 (0%)
6 (37.5%)
1 (6.25%)
14 (87.5%)
1 (6.25%)
5 (31%)
MAP (mmHg)
0 min
60 min
P value
(0 vs 60 min)
DMAP
(060 min)
Nitroglycerine-treated
group (n = 16)
Nifedipine-treated
group (n = 16)
P value
125 4
96 3
7.8e14
127 4
103 4
1.4e10
0.252
5.0e05
29 5 (26.531.3)
24 6 (21.027.4)
0.043
Values are meanSD. For the difference in mean arterial pressure (DMAP)
values in parentheses indicate 95% confidence intervals.
1.000
Primary outcomes
Within the nitroglycerine-treated group, 44% of patients received
5 mg/min, 44% received 10 mg/min and 12% received 15 or 20 mg/min
nitroglycerine, i.v. infusion; however, these doses were unrelated to
a subjects bodyweight. Conversely, within the nifedipine-treated
group none of the subjects required more than 10 mg nifedipine
to reach the therapeutic goal. The magnitude and time-course of
maternal hypotensive responses are indicated in Table 3 and Fig. 2,
respectively.
Mean arterial pressure showed a statistically significant decrease
just 5 min after initiation of the administration of nitroglycerine
(3.3 mmHg; P = 0.018, Wilcoxon signed-rank test) and 10 min after
the initiation of the administration of nifedipine (4.8 mmHg;
583
Secondary outcomes
DISCUSSION
Magnesium sulphate loading induced a small hypotensive response
(prior to that induced by vasodilators) that was more pronounced in
subjects who were finally allocated to the nitroglycerine treatment
group than in those who were finally allocated to the nifedipine
treatment group. The modelling of this hypotensive response
(D = before after) indicated that this difference could be related to
a relatively higher dose of MgSO4 in the nitroglycerine-treated
group, because women in this group tended to weigh approximately
4 kg less than those in the nifedipine-treated group. This also suggests
the need to review the efficacy of MgSO4 as an anticonvulsivant
(as well as its fetal effects) when a loading dose is administered
per kg bodyweight rather than a traditional fixed dose.4 Our analysis
also indicates that the MgSO4 loading dose-induced hypotensive
response did not affect the subsequent vasodilator effects of
nitroglycerine or nifedipine.
Nifedipine, a calcium channel blocker, was chosen because it is
one of the most widely used drugs in the management of pregnancy
induced hypertension in developing countries,25 such as Mexico.26
The route of administration of nifedipine used in the present study
ensured immediate contact of the total drug dose with the sublingual
mucosa, where it is primarily absorbed; however, it is not possible
to rule out partial absorption through the gastrointestinal tract due
to involuntary swallowing, so that the joint effect of both events on
absolute bioavailability is not known. It has been documented27 that
routes of administration of nifedipine that evade first-pass metabolism
in the gastrointestinal tract, such as use of the sublingual perforated
capsule, result in greater bioavailability and a faster and more stable
serum nifedipine concentration than methods that do not evade
first-pass metabolism (oral aspirated capsule content swallowed
and chewed capsule swallowed methods). Furthermore, sublingual
administration methods allow for a more gradual decrease in blood
pressure than oral swallowed methods, so that, clinically, they are
584
S Manzur-Verstegui et al.
Nifedipine-treated
group (n = 16)
P value
4.7 2.0
4.3 2.0
0.575
1.000
5 (31%)
11 (68.8%)
4 (25%)
12 (75%)
1.000
4 (36.4%)
4 (36.4%)
1 (9.1%)
0 (0.0%)
2 (18.2%)
4 (33.3%)
5 (41.7%)
1 (8.3%)
1 (8.3%)
1 (8.3%)
15 (93.8%)
1 (6.3%)
2.8 0.6
36 2
13 (81.3%)
3 (18.8%)
2.9 0.7
36 3
2 (12.5%)
14 (87.5%)
7 (43.8%)
9 (56.3%)
1 (6.3%)
15 (93.8%)
0 (0.0%)
16 (100.0%)
3.0 0.8
3.0 0.8
2.9 0.6
2.9 0.6
0.599
0.779
0.843
0.033
0.043
0.540
1.000
Values are meanSD or the number of subjects with percentages given in parentheses.
Flushing
Headache
Palpitations
Nausea
Total no. patients
Nitroglycerine-treated
group (n = 16)
4 (25.0%)
3 (18.8%)
3 (18.8%)
0 (0.0%)
16 (100.0%)
Nifedipine-treated
group (n = 16)
6 (37.5%)
2 (12.5%)
2 (12.5%)
1 (6.3%)
16 (100.0%)
P value
0.786
ACKNOWLEDGEMENTS
This research was supported by the Instituto Mexicano del Seguro
Social (IMSS) and the Universidad Autnoma de San Luis Potos
(UASLP). The authors thank Dr JL Romero-Santos (external
collaborator; General Hospital No.1, IMSS), Mrs Geraldine
MacDonald (independent English consultant; San Luis Potos,
Mxico) and the nursing personnel of the General Hospital No.1,
IMSS, for their excellent, professional assistance.
13.
14.
15.
16.
17.
18.
19.
20.
REFERENCES
1.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
585