Arteriolar vasodilation with relative underfilling of the arterial vascular compartment can occur in early pregnancy and leads to circulatory dysfunction characterized by arterial hypotension, high cardiac output, and stimulation of the renin-angiotensin system [7]. Because these hemodynamic changes have been reproduced by systemic estradiol administration in animal experiments, researchers suggest that arteriolar vasodilation during pregnancy is related to the very high plasma estrogen levels occurring with this condition [8]. Recent in vitro investigations showing that estradiol increases endothelial nitric oxide and prostacyclin release, and in vivo studies identifying increased nitric oxide biosynthesis in pregnant animals suggest that the vasodilator effect of estradiol may be mediated by endothelial vasodilators [9-11]. The ovarian hyperstimulation syndrome is also associated with dramatically increased circulating plasma estradiol concentration [3-6]. Therefore, arteriolar vasodilation might also occur in OHSS and contribute to the circulatory dysfunction that characterizes it.

We report the results of a prospective investigation evaluating systemic hemodynamics, endogenous vasoactive neurohumoral factors (renin-angiotensin and sympathetic nervous systems, antidiuretic hormone, atrial natriuretic factor, and renal prostaglandins), and renal function in 31 consecutive patients with severe OHSS. Our results indicate that the syndrome is consistently associated with peripheral arteriolar vasodilation, which may play an important pathogenetic role.

Methods

Top Methods Results Discussion Author & Article Info References

We included 31 consecutive patients admitted to our hospital during a 5-year period for severe OHSS according to the classification proposed by Golan and colleagues [5]. Nine of these patients were referred from other centers working in close connection with our assisted-reproduction unit and using the same in vitro fertilization protocol used in our hospital. They had induced multiple follicular growth for in vitro fertilization. All patients gave informed consent to participate in the study, which was approved by the Investigation and Ethics Committee of the Hospital Clnic i Provincial of Barcelona. Because a woman with severe OHSS may be pregnant, no invasive procedures or tests requiring the administration of exogenous substances were allowed by the Committee.

We stimulated patients' ovaries with follicle-stimulating hormone (75 IU per ampule) and human menopausal gonadotropin (75 IU of follicle-stimulating hormone and 75 IU of luteinizing hormone per ampule) under pituitary suppression with a gonadotropin-releasing hormone analog (buserelin acetate or leuprolide acetate) [12]. On days 1 and 2 of ovarian stimulation, we administered 3 ampules per day of follicle-stimulating hormone with 3 ampules of human menopausal gonadotropin. On days 3 to 7 of ovarian stimulation, we administered 2 ampules per day of human menopausal gonadotropin to each patient. From day 8 onward, we administered human menopausal gonadotropin on an individual basis according to the ovarian response. Transvaginal ultrasonography and plasma estradiol measurement were done daily from the seventh day of ovarian stimulation to assess follicular development. We did follicular aspiration 36 hours after an ovulatory dose of 5000 IU of human chorionic gonadotropin. Forty-eight hours later, we transferred three to four embryos per patient into the uterus. We gave additional doses of 5000 and 2500 IU of human chorionic gonadotropin on the days of follicular aspiration and embryo transfer, respectively, to supplement the luteal phase.

After diagnosis of severe OHSS, patients were hospitalized and given a 60-mmol sodium diet. Urine was carefully collected to measure electrolytes and concentrations of prostaglandin E₂ and 6-keto-prostaglandin F₁, the stable metabolite of prostacyclin. The next morning, after overnight fasting from food and after 2 hours of bed rest, arterial blood pressure and heart rate were measured and an antecubital vein was catheterized; 30 minutes later, blood samples were taken to measure plasma renin activity; plasma aldosterone, norepinephrine, antidiuretic hormone, atrial natriuretic peptide, serum creatinine, electrolyte, and hematocrit concentrations. Subsequently, 24 patients were transferred to an echocardiographic unit, where cardiac output and arterial pressure were measured.

Patients were then confined to bed rest and given low-sodium diets. We added diuretics (20 mg of furosemide given intravenously every 12 hours) and plasma volume expansion with albumin (50 g/d of salt-poor albumin) for those patients with urine volume less than 20 mL/h or sodium excretion less than 20 mmol/d. No patient was treated using paracentesis. Four to 5 weeks after recovery from OHSS, the patients were readmitted to repeat all these measurements, and thus each patient was her own control for hemodynamic, hormonal, and renal excretory studies. Seven women were pregnant during the second study. However, hematocrit concentrations and hemodynamic and neurohormonal parameters in these cases were similar to those of nonpregnant women. All nonpregnant women had the second study during the follicular phase of their menstrual cycles.

We measured plasma renin activity; plasma concentrations of aldosterone, antidiuretic hormone, atrial natriuretic peptide, and estradiol; and urinary concentrations of prostaglandin E₂ and 6-keto-prostaglandin F₁ by radioimmunoassay according to methods previously described [13-16]. We measured plasma norepinephrine concentration using the radioenzymatic assay of Peuler and Johnson [17]. We estimated cardiac output using two-dimensional and Doppler echocardiographic examinations (SSH-65A; Toshiba, Tokyo, Japan) [18], and we measured arterial pressure using sphygmomanometry. We calculated mean arterial pressure as diastolic blood pressure plus one third of the difference between the systolic and diastolic blood pressures. We used the following formula to estimate peripheral vascular resistance: mean arterial pressure minus right atrial pressure divided by cardiac output x 80. Because we did not measure right atrial pressure, we considered it to be zero in this calculation.

Results are presented as the mean and the 95% CI of the mean or as the mean of paired differences and the 95% CI estimate of the mean paired difference when appropriate. We used paired and unpaired Student t-tests and two-variable regression for the statistical analysis of results. We used percentage changes in peripheral vascular resistance, cardiac output, mean arterial pressure, and hematocrit concentration during OHSS relative to baseline recovery values to calculate correlations within these parameters and between them and absolute values of neurohormonal measurements (peak estradiol level and plasma renin activity, aldosterone, norepinephrine, and antidiuretic hormone during the syndrome). Correlations between neurohormonal measurements were made using absolute values.

Results

Top Methods Results Discussion Author & Article Info References

Patient Clinical Characteristics

The mean patient age was 31.5 years (range, 27 to 36 years). Thirteen patients had polycystic ovarian disease, a condition predisposing them to development of OHSS [5]. The mean peak plasma estradiol level during ovarian stimulation was 17 018 pmol/L (range, 11 563 to 37 554 pmol/L; CI, 14 757 to 19 280 pmol/L). The mean number of follicles observed by transvaginal ultrasonography on the day of human chorionic gonadotropin administration was 33.3 (range, 16 to 67; CI, 28.4 to 38.2). At hospital admission, mean size of the ovaries was 13.3 x 14.3 cm (range, 12 x 12 to 20 x 18), and all patients had marked abdominal distention because of enlarged ovaries and ascites. Six patients also had pleural effusion. No patient had peripheral edema. The amount of fluid retained, estimated by loss of body weight during hospitalization, was 4.5 kg (range, 2.7 to 7.5 kg; CI, 3.7 to 5.3 kg). No patient had renal failure during the syndrome. All patients recovered; seven became pregnant and five carried successfully to full term.

Systemic Hemodynamic Features, Hormonal Changes, and Renal Function during Severe Ovarian Hyperstimulation Syndrome

Table 1 shows the systemic hemodynamic, renal function, and hormonal measurements recorded during severe OHSS and 4 to 5 weeks after recovery. Compared with recovery values, arterial hypotension, tachycardia, high cardiac output, and low estimated peripheral vascular resistance characterized severe OHSS. These hemodynamic changes occurred in all patients (Figure 1). As expected, percentage changes in peripheral vascular resistance during the syndrome relative to baseline recovery values correlated with percentage changes in cardiac output (r = –0.72;P < 0.001) and mean arterial pressure (r = 0.52; P < 0.001).

View larger version (21K): [in this window] [in a new window]   Figure 1. Cardiac output and peripheral vascular resistance during the ovarian hyperstimulation syndrome (OHSS) and after recovery measured in 24 patients.

The mean hematocrit concentration decreased from 0.436 during OHSS to 0.389 during recovery (P < 0.001) (Table 1). The hematocrit concentration was greater during OHSS than during recovery in all but 4 patients, although in 9 cases the difference was only 0.02 or less. We found no correlation between percentage changes in hematocrit during the syndrome and percentage changes in mean arterial pressure, cardiac output, and peripheral vascular resistance. Hemoconcentration during OHSS, arbitrarily defined as a hematocrit concentration 0.1 over baseline recovery values, developed in 16 (51%) patients. In these patients, mean hematocrit concentrations during the syndrome and after recovery were 0.471 and 0.388, respectively (mean of the paired difference, 0.082; CI, 0.627 to 0.101). The corresponding values in patients without hemoconcentration were 0.399 and 0.39, respectively (mean of the paired difference, 0.0093; CI, 0.003 to 0.0216). We established the definition of hemoconcentration when we analyzed the results.

Plasma renin activity and plasma concentrations of aldosterone, norepinephrine, and antidiuretic hormone were increased markedly during OHSS and were normal after recovery in all patients (Table 1). We found a close, direct relation between plasma renin activity and aldosterone (r = 0.82; P < 0.001) and between plasma norepinephrine concentration and plasma renin activity and antidiuretic hormone concentration Figure 2 during the syndrome. Figure 3 shows the plasma atrial natriuretic peptide concentration during OHSS and after recovery for our patients. In most cases, OHSS was associated with high plasma levels of this hormone. No correlation was evident between percentage changes in peripheral vascular resistance and plasma renin activity, aldosterone, norepinephrine, and antidiuretic hormone during the syndrome. In contrast, percentage changes in hematocrit concentration correlated directly with plasma renin activity (r = 0.43; P < 0.05), aldosterone (r = 0.41; P < 0.05), norepinephrine (r = 0.55; P < 0.01), and antidiuretic hormone (r = 0.52; P < 0.001) during the syndrome.

View larger version (13K): [in this window] [in a new window]   Figure 2. The relation between plasma norepinephrine concentration and the plasma renin activity and antidiuretic hormone concentration.

View larger version (27K): [in this window] [in a new window]   Figure 3. Plasma concentration of atrial natriuretic peptide during the ovarian hyperstimulation syndrome and after recovery measured in 31 patients.

Peak plasma estradiol levels did not correlate with percentage changes in peripheral vascular resistance, cardiac output, and hematocrit concentration or with values of neurohormonal variables during the syndrome.

At admission, all patients had oliguria, sodium retention, and marked dilutional hyponatremia (Table 1). Although the serum creatinine concentration was normal in all patients (upper normal limit in our laboratory, 115 µmol/L), this parameter was 37% [P < 0.001] greater during the syndrome than after recovery, suggesting a small decrease in the glomerular filtration rate. Despite the low urine volume, urinary excretion of prostaglandin E₂ and 6-keto-prostaglandin F₁ )was substantially greater during the syndrome than after recovery.

Clinical Characteristics and Hemodynamic and Neurohormonal Measurements in Patients with and without Hemoconcentration

Table 2 shows the results obtained in patients with and without hemoconcentration. We found no statistical differences between groups regarding systemic hemodynamic measurements, renal function, plasma concentration of atrial natriuretic peptide, and urinary excretion of prostaglandin E₂ and 6-keto-prostaglandin F₁ . However, plasma renin activity and plasma aldosterone, norepinephrine, and antidiuretic hormone concentrations were significantly higher in patients with hemoconcentration.

Table 3 shows that there were no statistical differences between patients with and without hemoconcentration regarding age, presence of polycystic ovarian disease, gonadotropin dosage, peak estradiol level, number of follicles observed by transvaginal ultrasonography, and subsequent pregnancy. However, the difference in the number of follicles was almost statistically significant.

Discussion

Top Methods Results Discussion Author & Article Info References

During the last decade, severe OHSS has been studied increasingly because of its frequent occurrence as a result of the growing number of in vitro fertilization programs. We believe severe OHSS results from acutely increased capillary permeability leading to massive loss of intravascular fluid to the abdominal cavity, profound intravascular volume depletion, hemoconcentration, and renal impairment [3-6]. The site and mechanism of the increased capillary permeability have not been defined properly. Although the enlarged ovaries, which can have diameters of more than 10 to 15 cm, and presence of marked edema and capillary proliferation were first proposed as the site of ascites formation [19], recent experimental studies suggest that ascites may originate from extra-ovarian sites [20]. On the other hand, despite the fact that several substances of ovarian origin, such as prostaglandins, cytokines, and prorenin, have been proposed as mediators of increased vascular permeability [3, 4, 6], their role in the pathogenesis of severe OHSS is unclear.

Another aspect that has been insufficiently investigated is the role of the endogenous neurohormonal vasoactive systems in OHSS. Most studies have focused their attention on the renin-aldosterone system [6]. However, to our knowledge, besides a preliminary investigation by our group [21], no other study assessing sympathetic nervous activity and antidiuretic hormone in severe OHSS has been published. Severe OHSS is associated with increased plasma renin activity and aldosterone concentration, and some investigators suggest that it is a primary rather than a secondary phenomenon [5, 6, 22]. The following data have been offered to support this contention. First, evidence suggests that a renin-angiotensin system exists in the ovary that is activated by gonadotropins [23]. During gonadotropin ovarian stimulation, plasma prorenin markedly increases in close correlation with the number of ovarian follicles [24]. On the other hand, the preovulatory follicular fluid concentration of prorenin is as much as 12 times greater than that of plasma prorenin [25]. Second, plasma renin levels parallel the severity of OHSS [22]. Finally, in a recent study, plasma volume expansion did not normalize plasma renin activity and aldosterone concentration in patients with severe OHSS [26]. On the basis of these findings, hyperstimulated ovaries are considered the main source of hyperreninism in OHSS [5, 6, 26]. Moreover, because angiotensin II increases vascular permeability [27], it is considered the most probable ovarian factor responsible for the syndrome [6].

Our results confirm that the renin-angiotensin-aldosterone system is markedly activated in severe OHSS. However, they strongly suggest that this is a secondary rather than a primary event because plasma renin activity was closely correlated with the plasma levels of other volume-dependent endogenous vasoactive substances, such as norepinephrine and antidiuretic hormone, which were also markedly increased. Plasma norepinephrine concentration is a sensitive index of baroreceptor activity and sympathetic discharge [28]. On the other hand, both renal renin release and antidiuretic hormone secretion are greatly stimulated by the sympathetic nervous system [29, 30]. Therefore, the most plausible explanation for increased plasma levels of renin and antidiuretic hormone in severe OHSS is baroreceptor-mediated stimulation of the sympathetic nervous system secondary to circulatory dysfunction. The contribution of the ovarian renin-angiotensin system, however, cannot be overlooked.

In our study, we included only patients with severe OHSS, as defined by marked ovarian enlargement, clinically detectable ascites, arterial hypotension, tachycardia, oliguria, sodium retention, and dilutional hyponatremia. Our results confirm that increased vascular permeability is an important mechanism in the pathogenesis of circulatory and renal dysfunction in OHSS. The hematocrit concentration was greater during the syndrome than after recovery and correlated closely with the plasma levels of renin, aldosterone, norepinephrine, and antidiuretic hormone. However, hematocrit concentration during the syndrome was equal to or less than that during recovery in four cases, and in another nine patients the increase in hematocrit concentration during the syndrome was very small, suggesting that circulatory dysfunction cannot be explained based solely on hypovolemia.

This contention is supported by the hemodynamic changes we observed. If circulatory dysfunction in OHSS was due solely to a shift of fluid from the intravascular compartment to the peritoneal cavity, leading to contraction of the circulating blood volume, reduced cardiac output and increased peripheral vascular resistance should be anticipated. In contrast, cardiac output was increased and estimated peripheral vascular resistance was markedly reduced in all patients studied. Although we did not measure right atrial pressure in our patients, and therefore had to calculate peripheral vascular resistance without considering this parameter, this would probably not influence the results. In fact, right atrial pressure is normal or decreased in other edematous disorders associated with hyperdynamic circulation [31]. Our findings, therefore, indicate that severe OHSS is consistently associated with marked peripheral arteriolar vasodilation and suggest that this abnormality also may play an important role in the pathogenesis of the circulatory and renal dysfunction that characterizes this syndrome.

The plasma concentration of atrial natriuretic peptide was markedly increased during severe OHSS in our patients. This observation supports the argument that circulatory dysfunction in OHSS cannot be explained satisfactorily based only on hypovolemia. Right atrial pressure (or atrial distention), which is the primary stimulus of atrial natriuretic peptide release [32], is decreased when intravascular volume is depleted. Therefore, if circulatory dysfunction in severe OHSS was due solely to a shift of intravascular fluid to the abdominal cavity, the plasma concentration of atrial natriuretic peptide should not increase.

Although the best indicator of severity in OHSS is thought to be the hematocrit concentration [6, 33], no objective data have been reported to support this contention, which derives from the general belief that severe OHSS is due to intravascular volume depletion and that the hematocrit concentration is directly proportional to the change in plasma volume. Our study provides the first evidence indicating that hematocrit concentration correlates with severity of OHSS because patients with increased hematocrit concentration had the highest plasma levels of renin, norepinephrine, and antidiuretic hormone, which are sensitive markers of effective intra-arterial blood volume. The absence of differences in arterial pressure and peripheral vascular resistance between patients with and without hemoconcentration reflects an adequate homeostatic pressor response to these endogenous vasoconstrictor systems in the former group of patients.

Despite arterial hypotension and intense stimulation of the renin-angiotensin and sympathetic nervous systems, which are powerful renal vasoconstrictors, renal failure developed in none of our patients during OHSS. This finding is probably related to the markedly increased renal production of prostaglandin E₂ and prostacyclin observed in all patients, which we estimated by measuring the urinary excretion of prostaglandin E₂ and 6-keto-prostaglandin F₁ , respectively. Prostaglandin E₂ and prostacyclin are renal vasodilators and antagonize the vasoconstrictor effect of angiotensin II and norepinephrine [34].

Our results have important therapeutic implications. Nonsteroidal anti-inflammatory drugs have been used widely in patients with OHSS to correct increased vascular permeability [4, 35]. Recently, researchers proposed the use of angiotensin-converting enzyme inhibitors for the same purpose [36]. Our study strongly suggests that angiotensin II and renal prostaglandins are important mechanisms for maintaining arterial pressure and renal perfusion, respectively, in OHSS. Therefore, administration of nonsteroidal anti-inflammatory drugs or angiotensin-converting enzyme inhibitors in patients with severe OHSS may induce acute renal failure or severe arterial hypotension. Most patients with OHSS recover with simple measures such as bed rest, a low-sodium diet, and diuretic therapy. In the most severe cases, however, treatments to maintain circulatory function may be needed.

The site and mechanism of arterial vasodilation and increased capillary permeability, and why the peritoneal cavity is the predominant site where the fluid-shifting phenomena occur in OHSS cannot be determined from this study. Because arteriolar vasodilation induces the formation of interstitial edema by increasing capillary surface area and capillary hydrostatic pressure [37, 38], a link between arteriolar vasodilation and fluid leakage in OHSS may exist. However, future studies are needed to clarify these points.

Our study indicates that the pathogenesis of OHSS is more complex than currently understood. In addition to hemoconcentration caused by the escape of fluid to extravascular spaces, OHSS is consistently associated with marked arteriolar vasodilation. The simultaneous occurrence of both disorders leads to a hyperdynamic circulatory dysfunction characterized by arterial hypotension, increased cardiac output, reduced peripheral vascular resistance, and intense stimulation of the renin-angiotensin and sympathetic nervous systems and antidiuretic hormone. The renal vasoconstrictor effect of these endogenous neurohormonal systems is counterbalanced by increased renal production of vasodilator prostaglandins, which maintains renal perfusion and glomerular filtration rate within normal limits. However, these systems promote renal sodium and water retention, which contribute to edema formation.