- Email: [email protected]
Magnesium sulfate therapy for the prevention of cerebral palsy in preterm infants: a decision-analytic and economic analysis
Research
www. AJOG.org
OBSTETRICS
Magnesium sulfate therapy for the prevention of cerebral palsy in preterm infants: a decisionanalytic and economic analysis Alison G. Cahill, MD, MSCI; Anthony O. Odibo, MD, MSCE; Molly J. Stout, MD; William A. Grobman, MD, MBA; George A. Macones, MD; Aaron B. Caughey, MD, PhD OBJECTIVE: We sought to estimate the cost-effectiveness of magne-
RESULTS: Magnesium for neuroprophylaxis led to lower costs ($1739
sium neuroprophylaxis for all women at risk for preterm birth ⬍32 weeks.
vs $1917) and better outcomes (56.684 vs 56.678 quality-adjusted life years). However, sensitivity analysis revealed the model to be sensitive to estimates of effect of magnesium on risk of moderate or severe cerebral palsy as well as neonatal death.
STUDY DESIGN: A decision analytic and cost-effectiveness model was designed to compare use of magnesium for neuroprophylaxis vs no treatment for women at risk for preterm birth ⬍32 weeks due to preterm premature rupture of membranes or preterm labor from 24-32 weeks. Outcomes included neonatal death and moderatesevere cerebral palsy. Effectiveness was reported in quality-adjusted life years.
CONCLUSION: Based on currently published evidence for efficacy,
magnesium for neuroprophylaxis in women at risk to deliver preterm is cost-effective. Key words: cerebral palsy, magnesium, neuroprophylaxis, preterm birth
Cite this article as: Cahill AG, Odibo AO, Stout MJ, et al. Magnesium sulfate therapy for the prevention of cerebral palsy in preterm infants: a decision-analytic and economic analysis. Am J Obstet Gynecol 2011;205:542.e1-7.
A
fter years of published observational data suggesting a relationship between maternal exposure to magnesium sulfate and a decrease in adverse neonatal neurologic outcomes in infants born preterm,1-3 4 trials were published over the past 8 years with conflicting results regarding the impact of magnesium in the setting of pregnancies delivering preterm.4-7 The most recent of these trials by Rouse et al7 was a multicenter trial From the Departments of Obstetrics and Gynecology at Washington University in St. Louis School of Medicine, St. Louis, MO (Drs Cahill, Odibo, Stout, and Macones); Feinberg School of Medicine, Northwestern University, Chicago, IL (Dr Grobman); and Oregon Health and Science University, Portland, OR (Dr Caughey). Received May 27, 2011; revised Aug. 14, 2011; accepted Sept. 7, 2011. The authors report no conflict of interest. Presented as a poster at the 30th annual meeting of the Society for Maternal–Fetal Medicine, Chicago, IL, Feb. 4-6, 2010. Reprints not available from the authors. 0002-9378/$36.00 © 2011 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2011.09.004
542.e1
conducted in the United States, and the results offered the most promising evidence that magnesium decreased the risk of moderate or severe cerebral palsy (CP) in infants born preterm who survived. However, CP was not the primary outcome of this trial, but was part of a composite with death due to the design challenge of competing risks and there was a trend toward an increase in death in the infants exposed to magnesium compared to those exposed to placebo. In light of these recent and complex findings as well as the rare nature of these outcomes, it seems reasonable to include a formal estimate of cost-effectiveness of this therapy when considering a sweeping change in practice. We undertook this study to test the hypothesis that magnesium was cost-effective in women at risk for preterm birth for the prevention of CP, and further, to explore whether the cost-effectiveness was limited to particular subgroups of women.
M ATERIALS AND M ETHODS A decision-analytic and cost-effectiveness model was designed to compare 2 strategic approaches for the prevention
American Journal of Obstetrics & Gynecology DECEMBER 2011
of CP in preterm infants from a societal prospective. The use of magnesium for neonatal neuroprophylaxis was compared to nonuse, with the latter representative of the accepted standard of care prior to recent publications.7,8 The model was constructed for women carrying singleton pregnancies between 24 weeks 0/7 days’ and 31 weeks 6/7 days’ gestation at high risk for spontaneous preterm birth ⬍32 weeks’ gestation due to either preterm labor (PTL) or preterm premature rupture of membranes (PPROM) (Figure 1). Two additional models were constructed to estimate the cost-effectiveness of magnesium for neuroprophylaxis in at-risk subpopulations: PPROM only between 24 weeks 0/7 days’ and 31 weeks 6/7 days’ gestation, and patients at risk for preterm birth ⬍28 weeks (due to either PPROM or PTL). Cost was estimated in US dollars, effectiveness was reported as quality-adjusted lifeyears (QALYs), and strategies were compared in terms of number of cases of CP prevented, number of neonatal or infant deaths, and cost-effectiveness. Since only one of the placebo-controlled trials demonstrated statistically significant magnesium efficacy for the
Obstetrics
www.AJOG.org
TABLE 1
Probability, utility, and cost estimates for model Variable
Point estimate
Range
Source
Probabilities
..................................................................................................................................................................................................................................... 13-16
PTB ⬍32 wk
0.025
0.015–0.038
Proportion of PTB ⬍28 wk
0.250
0.200–0.300
PPROM ⬍32 wk
0.011
0.006–0.018
Preterm labor ⬍32 wk
0.020
0.010–0.080
Proportion of PTB after preterm labor ⬍32 wk
0.250
0.100–0.400
CP after PTB ⬍32 wk
0.035
0.020–0.061
CP after birth ⱖ32 wk
0.001
0.0004–0.003
Death after PTB ⬍32 wk
0.119
0.033–0.171
Death after birth ⱖ32 wk
0.0003
0.0002–0.0006
Relative risk of CP with magnesium
0.71
0.50–0.82
Relative risk of death with magnesium
1.06
1.00–1.15
Severe adverse magnesium reaction
0.047
0.001–0.081
Mild adverse magnesium reaction
0.696
0.540–0.920
..................................................................................................................................................................................................................................... 13,16,17 ..................................................................................................................................................................................................................................... 14,16,17 ..................................................................................................................................................................................................................................... 13 ..................................................................................................................................................................................................................................... 16,18,19
..................................................................................................................................................................................................................................... 4,5,7,20,21 ..................................................................................................................................................................................................................................... 22 ..................................................................................................................................................................................................................................... 4,5,7,20,21 ..................................................................................................................................................................................................................................... 22 ..................................................................................................................................................................................................................................... 4,5,7,20,21 ..................................................................................................................................................................................................................................... 4,5,7,20,21
..................................................................................................................................................................................................................................... 5,7,23-28 ..................................................................................................................................................................................................................................... 5,7,23-28 ..................................................................................................................................................................................................................................... 29
Maternal life expectancy, y
55.4
50.1–58.2
Neonatal life expectancy, y
77.2
70.3–82.0
Neonatal life expectancy, with CP, y
28.7
20.6–36.8
..................................................................................................................................................................................................................................... 29 ..................................................................................................................................................................................................................................... 30,31 ..............................................................................................................................................................................................................................................
Utilities
..................................................................................................................................................................................................................................... 32,33
Neonatal death
0.01
0.001–0.02
Moderate-severe CP
0.55
0.50–0.60
Normal child
1.0
..................................................................................................................................................................................................................................... 32,33 ..................................................................................................................................................................................................................................... 32,33
—
..............................................................................................................................................................................................................................................
Costs, $US
..................................................................................................................................................................................................................................... 34
Magnesium treatment
124.24
75.00–250.00
Severe magnesium reaction
145.40
80.95–287.88
..................................................................................................................................................................................................................................... 34 ..................................................................................................................................................................................................................................... 34
Mild magnesium reaction
2.52
0.16–5.04
..................................................................................................................................................................................................................................... 35
Moderate-severe CP
955,940
700,000–1,300,000
..................................................................................................................................................................................................................................... 36
Neonatal death after PTB ⬍32 wk
71,001
452–351,237
Neonatal survival after PTB ⬍32 wk
45,710
13,560–123,207
Neonatal death after birth ⱖ32 wk
67,758
466–283,910
..................................................................................................................................................................................................................................... 36 ..................................................................................................................................................................................................................................... 36 ..............................................................................................................................................................................................................................................
CP, cerebral palsy; PPROM, preterm premature rupture of membranes; PTB, preterm birth. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
reduction in risk of CP,7 magnesium exposure was modeled with the dosing protocol used in that trial. Magnesium treatment was initiated at presentation, and given as a 6-g bolus followed by 2 g/h infusion for 12 hours or until delivery. If delivery did not occur, a patient was eligible for retreatment when preterm delivery threatened again. It was assumed that all women eligible for retreatment
received magnesium again for purposes of cost estimates in the model. Clinical outcomes considered in the models were the occurrence of mild or severe adverse maternal reaction to magnesium, birth of a neonate with moderate or severe CP diagnosed by age 2 years, neonatal or infant death by 1 year, and unaffected neonate. The model was constructed with some notable definitions
Research
and assumptions. CP was diagnosed, and its severity determined, by accepted pediatric standards.9-11 PPROM was defined by clinical criteria described by the American College of Obstetricians and Gynecologists (ACOG),12 and all patients with PPROM delivered preterm (⬍32 weeks). PTL, defined clinically by contractions that change the cervical examination presenting between 24-32 weeks, was assumed to result in preterm delivery by 32 weeks 0 days in only a proportion of cases. The proportion of patients admitted with PTL going on to deliver preterm was based on available published evidence (Table 1). Baseline probabilities and utilities (or values given to a particular health-related outcome) were derived from the literature after an English-language PubMed search with the terms: “magnesium,” “neuroprophylaxis,” “preterm birth,” and “cerebral palsy.” Studies without control groups and reviews were excluded. Point estimates were calculated as weighted means by sample size, and ranges represented by the lowest and highest value for a given estimate in the published literature. For estimates derived from a single source, a range was defined by the 95% confidence intervals that was calculated from the binomial distribution. To derive the estimate of magnesium efficacy, the 4 trials in which magnesium was primarily given for the prevention of CP4-7,20 were used. Cost estimates and ranges were derived from published literature or based on Medicaid reimbursement rates.37 In cases of the latter, charges were multiplied by a cost-charge ratio of 0.6 as an approximation to third-party reimbursements, and regional variation was accounted for in sensitivity analyses by widely varying locally derived estimates. Effectiveness was calculated by the product of the utility value and discounted life expectancy in years. In accordance with standard assumptions, we assumed a life expectancy of 75 years and an annual discount rate for costs and QALYs of 3%.38 A threshold of $100,000 per QALY was chosen as the willingness-to-pay threshold.39 To enable comparison of a number of outcomes of interest across strategies, we considered a hypothetical cohort of 100,000 women at risk for preterm birth. Sensitivity analyses, performed by vary-
DECEMBER 2011 American Journal of Obstetrics & Gynecology
542.e2
Research
Obstetrics
ing ⱖ1 of the estimates along their entire range of potential values, were performed to test which value variation, if any, altered our results. Monte Carlo simulation, performed as a form of multivariable analysis that simultaneously incorporates the uncertainty in all point estimates by sampling from the distributions around these estimates, further tested the robustness of the model. Given the nature of the primary data, and specifically the debates regarding interpretation of the trials individually or in combination regarding the association of magnesium with CP as well as fetal death, we widely varied the estimates of effect in our sensitivity analyses to investigate robustness of the model and investigate threshold values. We conducted similar threshold analyses by varying the background rate of preterm birth within the model, in an effort to indirectly mimic the strategies for a population at high risk for preterm birth. Models were constructed and analyzed using TreeAge Pro 2007 (TreeAge Software, Williamstown, MA). This study was exempt from institutional review board approval.
R ESULTS Under base-case assumptions, magnesium for neuroprophylaxis for all women at high risk for spontaneous preterm birth was the dominant (least costly and most effective) strategy when compared to no magnesium. For every 10,000 women at risk for preterm birth treated with magnesium, $1.8 million were saved and 52 QALYs were gained. Compared to no magnesium, magnesium was also the dominant strategy when given to only women with PPROM as well as when given only to women at risk to deliver ⬍28 weeks. For every 10,000 women with PPROM treated with magnesium, $1.4 million were saved and 51 QALYs were gained (Table 2). Considering a hypothetical cohort of 100,000 at high risk for preterm birth, magnesium treatment resulted in the greatest reduction in the number of cases of moderate or severe CP (n ⫽ 32). While the strategy of magnesium compared to none within the subgroups of women with PPROM only and those at 542.e3
www.AJOG.org
TABLE 2
Base-case cost-effectiveness of magnesium for neuroprophylaxis Strategy
Cost, $
Incremental cost, $
Efficacy (QALY)
Incremental efficacy
Incr C/E (ICER)
All women
.....................................................................................................................................................................................................................................
Magnesium
1739.00
—
56.6836
—
—
No magnesium
1917.20
178.20
56.6784
⫺0.0052
(Dominated)
..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................
PPROM only
.....................................................................................................................................................................................................................................
Magnesium
1462.60
—
56.7022
—
—
No magnesium
1607.50
144.90
56.6972
⫺0.0050
(Dominated)
..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................
⬍28 wk
.....................................................................................................................................................................................................................................
Magnesium
920.60
—
56.7411
—
—
56.7355
⫺0.0056
(Dominated)
.....................................................................................................................................................................................................................................
No magnesium
1019.00
98.50
..............................................................................................................................................................................................................................................
C/E, cost-effectiveness; ICER, incremental cost-effectiveness ratio; Incr, incremental; PPROM, preterm premature rupture of membranes; QALY, quality-adjusted life-year. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
risk to deliver ⬍28 weeks resulted in reductions in the annual number of cases of CP, the absolute number prevented was 12% fewer when applied to the PPROM-only group (n ⫽ 28) and 28% fewer when applied to only those women at risk for preterm birth ⬍28 weeks (n ⫽ 23) (Table 3). Sensitivity analyses showed our results to be robust to the wide ranges of values considered for many inputs. However,
sensitivity analysis revealed the model to be sensitive to estimates of effect of magnesium on risk of moderate or severe CP as well as neonatal death. When the risk reduction in moderate to severe CP conferred by magnesium treatment was ⬍14%, magnesium was no longer costeffective (Figure 2). Similarly, a relative risk of neonatal death associated with magnesium exposure of ⱖ1.16 resulted in no magnesium sulfate treatment be-
TABLE 3
Cases of moderate-severe cerebral palsy and death prevented per dollar spent
Strategy
No. of cases of moderate or severe CP
Total cost ($10 million)
No. of cases of moderate or severe CP prevented
Total cost saved ($10 million)
All women
.....................................................................................................................................................................................................................................
Magnesium
78
17.4
32
1.8
110
19.2
Reference
Reference
.....................................................................................................................................................................................................................................
No magnesium (reference)
..............................................................................................................................................................................................................................................
PPROM only
.....................................................................................................................................................................................................................................
Magnesium
70
14.6
28
1.5
No magnesium (reference)
98
16.1
Reference
Reference
.....................................................................................................................................................................................................................................
..............................................................................................................................................................................................................................................
⬍28 wk only
.....................................................................................................................................................................................................................................
Magnesium
58
9.2
No magnesium (reference)
81
10.2
23
1.0
Reference
Reference
.....................................................................................................................................................................................................................................
..............................................................................................................................................................................................................................................
CP, cerebral palsy; PPROM, preterm premature rupture of membranes. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
American Journal of Obstetrics & Gynecology DECEMBER 2011
Obstetrics
www.AJOG.org
Research
FIGURE 1
Schematic of decision-analytic tree for magnesium for neuroprophylaxis
Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
coming the dominant strategy. Lastly, when the background rate of preterm birth was varied widely, even when the background risk of preterm birth was as high as 15%, magnesium for neuroprophylaxis remained the dominant strategy, holding all other assumptions constant. To test the hypothesis that it is not cost-effective to give magnesium for neuroprophylaxis to all women admitted with PTL because many of these women will not go on to have a subsequent preterm birth, we explored the threshold value of preterm birth after PTL between 24-32 weeks. Our model revealed that even if only 10% of women admitted with a diagnosis of
PTL went on to have a preterm birth ⬍32 weeks, magnesium therapy for all women at risk for preterm birth due to PTL was cost-effective. In Monte Carlo simulation, which functions as a form of multivariate analysis allowing all estimates in the model to be simultaneously varied across their plausible ranges, magnesium for all pregnancies at risk for preterm birth ⬍32 weeks was the preferred strategy 86.7% of the time and was dominant 63.4% of the time (Figure 3). In a 10,000 sample Monte Carlo simulation for both subanalyses (PPROM and women delivering ⬍28 weeks), and a willingness-topay threshold of $100,000/QALY, mag-
nesium was the preferred strategy 100% of the time.
C OMMENT Despite the fact that the correlation between admissions for PTL and subsequent preterm birth is not high,18,40-42 our model demonstrated that magnesium for neuroprophylaxis is cost-effective in women at high risk for spontaneous preterm delivery ⬍32 weeks. And, while magnesium therapy, compared to a strategy of no magnesium, was the dominant strategy for all women at risk for preterm delivery as well as within the subgroup of women diagnosed with
DECEMBER 2011 American Journal of Obstetrics & Gynecology
542.e4
Research
Obstetrics
FIGURE 2
Sensitivity analysis on relative risk reduction of cerebral palsy from magnesium (MS04)
QALY, quality-adjusted life-year. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
PPROM and the subgroup who deliver ⬍28 weeks, magnesium for all women at risk for spontaneous preterm delivery resulted in the greatest reduction in number of cases of moderate or severe CP and the largest number of dollars saved. However, sensitivity analysis revealed the model to be sensitive to estimates of effect of magnesium on risk of moderate or severe CP as well as neonatal death. Additionally, in multivariable iterative analyses, the magnesium strategy compared with none was chosen 100% of the time for the subgroups of women with PPROM and those at risk for delivery ⬍28 weeks; for all women at risk for preterm delivery, it was only the preferred strategy 86.7% of the time, reflective of both the poor correlation between the diagnosis of PTL and subsequent preterm birth as well as the published data regarding magnesium for neuroprophylaxis. Since the body of evidence regarding magnesium for neuroprophylaxis is challenging to interpret, as many authors have illustrated,43-46 and the extrapolation into clinical practice complex, several investigators have used the technique of metaanalysis to formally pool the results from the primary trials and better estimate the effect of magnesium on risk for CP and neonatal death.8,47 However, one criticism that has been 542.e5
www.AJOG.org made with regard to interpretation of these metaanalyses is that the patient populations studied in each of the 4 trials varied significantly. Specifically, 88% of the patients enrolled in the trial by Rouse et al7 had a diagnosis of PPROM, while the most common diagnosis in the study by Crowther et al5 was PTL. In this analysis, we were able to specifically consider whether the application of magnesium therapy for neuroprophylaxis would be most cost-effectively applied to only women with PPROM as has been suggested by some, or to all women at risk for spontaneous preterm birth ⬍32 weeks including women admitted with PTL and cervical dilation. Our findings that magnesium for all women at high risk for spontaneous preterm birth was cost-effective even when only 10% of PTL admissions went on to deliver preterm supports the use of magnesium for all women at risk. It is notable that while magnesium, compared to no therapy, was the dominant strategy in base-case analyses, our Monte Carlo simulation demonstrated magnesium to be the preferred strategy only 86.7% of the time in the ⬍32-week model. This can be viewed in 2 ways. One would be to consider these findings not entirely robust as they do not strictly comply with the 95% confidence interval that is generally preferred. The second would be to realize that, unfortunately, the data on which this analysis is based contain some uncertainty. Thus, we would offer that these findings could be considered a numeric estimate of the complexities and uncertainty of the pooled data. In practical terms, the total number of women and neonates in all of these studies combined remains relatively small with respect to the rare incidence of the outcomes considered. Slight variation in the estimates of effect of magnesium on risk for CP and death can impact the cost-effectiveness of magnesium treatment. These findings from our cost-effectiveness model are consistent with the recent committee opinion from ACOG48 on the use of magnesium sulfate before anticipated preterm birth for neuroprotection. The committee recognized that “none of the individual studies found a
American Journal of Obstetrics & Gynecology DECEMBER 2011
FIGURE 3
Multivariate sensitivity analyses
Outcomes of 10,000 trials from Monte Carlo simulation. Each blue dot represents single trial outcome. Ellipse marks 95% confidence interval. Dashed line represents willingness-to-pay threshold of $100,000; all dots below threshold are cost-effective. QALY, quality-adjusted life-year. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
benefit with regard to their primary outcome” and that “comparison is made difficult by differences in inclusion and exclusion criteria, populations studied, magnesium sulfate regimens, gestational age with treatment, and outcomes studied between the trials.” However, the committee further went on to conclude that “the available evidence suggests that magnesium sulfate given before anticipated early preterm birth reduces the risk of cerebral palsy in surviving infants.” As with all theoretical models, there are important potential limitations that should be considered carefully when interpreting our results. The data modeled are limited to the available published literature. In addition, models such as these are based on a set of assumptions that can impact both internal validity as well as generalizability, and with this in mind we have outlined these assumptions carefully in our “Material and Methods” section. Specific to our question of the effect of magnesium on risk for CP and death in neonates born preterm, the magnitude of effect was assumed to be fixed across preterm gestational ages of exposure. However, it is possible that magnesium has a greater or
Obstetrics
www.AJOG.org lesser effect depending on the gestational age at which the therapy is applied. But currently published data did not allow us to model this potential variation in an evidence-based manner beyond our stratification at 28 weeks of gestation. While the existing evidence is complex, it appears unlikely that additional randomized trials of magnesium to prevent CP will be done. With this in mind, we believe our model aids in the consideration and interpretation of use of magnesium for neuroprophylaxis in pregnancies at high risk to deliver preterm, and allowed us to use sensitivity analyses to explore the uncertainty in the published efficacy data. First, magnesium compared to no therapy is cost-effective. Second, based on our model, it is best applied to all women at high risk for spontaneous preterm birth ⬍32 weeks as defined by the inclusion criteria of the recently published efficacy trial,7 despite the fact that many women admitted and treated for PTL between 24-32 weeks go f on to deliver ⬎34 weeks. REFERENCES 1. van de Bor M, Verloove-Vanhorick SP, Brand R, Keirse MJ, Ruys JH. Incidence and prediction of periventricular-intraventricular hemorrhage in very preterm infants. J Perinat Med 1987;15:333-9. 2. Kuban KC, Leviton A, Pagano M, Fenton T, Strassfeld R, Wolff M. Maternal toxemia is associated with reduced incidence of germinal matrix hemorrhage in premature babies. J Child Neurol 1992;7:70-6. 3. Nelson KB, Grether JK. Can magnesium sulfate reduce the risk of cerebral palsy in very low birthweight infants? Pediatrics 1995;95:263-9. 4. Mittendorf R, Dambrosia J, Pryde PG, et al. Association between the use of antenatal magnesium sulfate in preterm labor and adverse health outcomes in infants. Am J Obstet Gynecol 2002;186:1111-8. 5. Crowther CA, Hiller JE, Doyle LW, Haslam RR. Effect of magnesium sulfate given for neuroprotection before preterm birth: a randomized controlled trial. JAMA 2003;290: 2669-76. 6. Marret S, Marpeau L, Zupan-Simunek V, et al. Magnesium sulphate given before very-preterm birth to protect infant brain: the randomized controlled PREMAG trial*. BJOG 2007; 114:310-8. 7. Rouse DJ, Hirtz DG, Thom E, et al. A randomized, controlled trial of magnesium sulfate for the prevention of cerebral palsy. N Engl J Med 2008;359:895-905.
8. Doyle LW, Crowther CA, Middleton P, Marret S, Rouse D. Magnesium sulphate for women at risk of preterm birth for neuroprotection of the fetus. Cochrane Database Syst Rev 2009;1: CD004661. 9. Blasco PA. Primitive reflexes: their contribution to the early detection of cerebral palsy. Clin Pediatr (Phila) 1994;33:388-97. 10. Capute AJ, Shapiro BK. The motor quotient: a method for the early detection of motor delay. Am J Dis Child 1985;139:940-2. 11. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997;39:214-23. 12. American College of Obstetricians and Gynecologists. ACOG Committee on Practice Bulletins–Obstetrics. ACOG practice bulletin no. 80: premature rupture of membranes; clinical management guidelines for obstetrician-gynecologists. Obstet Gynecol 2007;109:1007-19. 13. Martin JA, Kochanek KD, Strobino DM, Guyer B, MacDorman MF. Annual summary of vital statistics–2003. Pediatrics 2005;115: 619-34. 14. Moster D, Lie RT, Markestad T. Long-term medical and social consequences of preterm birth. N Engl J Med 2008;359:262-73. 15. Tucker JM, Goldenberg RL, Davis RO, Copper RL, Winkler CL, Hauth JC. Etiologies of preterm birth in an indigent population: is prevention a logical expectation? Obstet Gynecol 1991;77:343-7. 16. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008;371:75-84. 17. Mercer BM. Preterm premature rupture of the membranes. Obstet Gynecol 2003;101: 178-93. 18. Macones GA, Segel SY, Stamilio DM, Morgan MA. Prediction of delivery among women with early preterm labor by means of clinical characteristics alone. Am J Obstet Gynecol 1999;181:1414-8. 19. Iams JD. Prediction and early detection of preterm labor. Obstet Gynecol 2003;101: 402-12. 20. Marret S, Marpeau L, Benichou J. Benefit of magnesium sulfate given before very preterm birth to protect infant brain. Pediatrics 2008; 121:225-6. 21. Magpie Trial Follow-Up Study Collaborative Group. The Magpie trial: a randomized trial comparing magnesium sulphate with placebo for pre-eclampsia; outcome for children at 18 months. BJOG 2007;114:289-99. 22. McIntire DD, Leveno KJ. Neonatal mortality and morbidity rates in late preterm births compared with births at term. Obstet Gynecol 2008;111:35-41. 23. Chau AC, Gabert HA, Miller JM Jr. A prospective comparison of terbutaline and magnesium for tocolysis. Obstet Gynecol 1992;80:847-51. 24. El-Sayed YY, Riley ET, Holbrook RH Jr, Cohen SE, Chitkara U, Druzin ML. Randomized comparison of intravenous nitroglycerin and
Research
magnesium sulfate for treatment of preterm labor. Obstet Gynecol 1999;93:79-83. 25. Glock JL, Morales WJ. Efficacy and safety of nifedipine versus magnesium sulfate in the management of preterm labor: a randomized study. Am J Obstet Gynecol 1993;169:960-4. 26. Larmon JE, Ross BS, May WL, Dickerson GA, Fischer RG, Morrison JC. Oral nicardipine versus intravenous magnesium sulfate for the treatment of preterm labor. Am J Obstet Gynecol 1999;181:1432-7. 27. McWhorter J, Carlan SJ, OLeary TD, Richichi K, OBrien WF. Rofecoxib versus magnesium sulfate to arrest preterm labor: a randomized trial. Obstet Gynecol 2004;103:923-30. 28. Terrone DA, Rinehart BK, Kimmel ES, May WL, Larmon JE, Morrison JC. A prospective, randomized, controlled trial of high and low maintenance doses of magnesium sulfate for acute tocolysis. Am J Obstet Gynecol 2000; 182:1477-82. 29. Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/nchs/data/ nvsr/nvsr59/nvsr59_04.pdf. Accessed Aug. 10, 2011. 30. Blair E, Watson L, Badawi N, Stanley FJ. Life expectancy among people with cerebral palsy in Western Australia. Dev Med Child Neurol 2001;43:508-15. 31. Strauss D, Brooks J, Rosenbloom L, Shavelle R. Life expectancy in cerebral palsy: an update. Dev Med Child Neurol 2008;50:487-93. 32. Pham CT, Crowther CA. Birth outcomes: utility values that postnatal women, midwives and medical staff express. BJOG 2003;110:121-7. 33. Vandenbussche FP, De Jong-Potjer LC, Stiggelbout AM, Le Cessie S, Keirse MJ. Differences in the valuation of birth outcomes among pregnant women, mothers, and obstetricians. Birth 1999;26:178-83. 34. Hayes E, Moroz L, Pizzi L, Baxter J. A cost decision analysis of 4 tocolytic drugs. Am J Obstet Gynecol 2007;197:383.e1-6. 35. Waitzman NJ, Romano PS, Scheffler RM. Estimates of the economic costs of birth defects. Inquiry 1994;31:188-205. 36. Phibbs CS, Schmitt SK. Estimates of the cost and length of stay changes that can be attributed to one-week increases in gestational age for premature infants. Early Hum Dev 2006;82:85-95. 37. Medicare and Medicaid statistical supplement, 1995. US Department of Health and Human Services, Health Care Financing Administration. Health Care Financ Rev Stat Suppl 1995:1-388. 38. Drummond MF, Barbieri M, Wong JB. Analytic choices in economic models of treatments for rheumatoid arthritis: what makes a difference? Med Decis Making 2005;25:520-33. 39. Owens DK. Interpretation of cost-effectiveness analyses. J Gen Intern Med 1998; 13:716-7. 40. Hueston WJ. Preterm contractions in community settings, II: predicting preterm birth in women with preterm contractions. Obstet Gynecol 1998;92:43-6.
DECEMBER 2011 American Journal of Obstetrics & Gynecology
542.e6
Research
Obstetrics
41. Hueston WJ. Preterm contractions in community settings, I: treatment of preterm contractions. Obstet Gynecol 1998;92:3842. 42. King JF, Grant A, Keirse MJ, Chalmers I. Beta-mimetics in preterm labor: an overview of the randomized controlled trials. Br J Obstet Gynaecol 1988;95:211-22. 43. Macones GA. MgSO4 for CP prevention: too good to be true? Am J Obstet Gynecol 2009;200:589.
542.e7
www.AJOG.org 44. Cahill AG, Caughey AB. Magnesium for neuroprophylaxis: fact or fiction? Am J Obstet Gynecol 2009;200:590-4. 45. Rouse DJ. Magnesium sulfate for the prevention of cerebral palsy. Am J Obstet Gynecol 2009;200:610-2. 46. Stanley FJ, Crowther C. Antenatal magnesium sulfate for neuroprotection before preterm birth? N Engl J Med 2008;359:962-4. 47. Conde-Agudelo A, Romero R. Antenatal magnesium sulfate for the prevention of cere-
American Journal of Obstetrics & Gynecology DECEMBER 2011
bral palsy in preterm infants less than 34 weeks’ gestation: a systematic review and metaanalysis. Am J Obstet Gynecol 2009; 200:595-609. 48. American College of Obstetricians and Gynecologists Committee on Obstetric Practice; Society for Maternal-Fetal Medicine. Committee opinion No. 455: magnesium sulfate before anticipated preterm birth for neuroprotection. Obstet Gynecol 2010;115: 669-71.
www. AJOG.org
OBSTETRICS
Magnesium sulfate therapy for the prevention of cerebral palsy in preterm infants: a decisionanalytic and economic analysis Alison G. Cahill, MD, MSCI; Anthony O. Odibo, MD, MSCE; Molly J. Stout, MD; William A. Grobman, MD, MBA; George A. Macones, MD; Aaron B. Caughey, MD, PhD OBJECTIVE: We sought to estimate the cost-effectiveness of magne-
RESULTS: Magnesium for neuroprophylaxis led to lower costs ($1739
sium neuroprophylaxis for all women at risk for preterm birth ⬍32 weeks.
vs $1917) and better outcomes (56.684 vs 56.678 quality-adjusted life years). However, sensitivity analysis revealed the model to be sensitive to estimates of effect of magnesium on risk of moderate or severe cerebral palsy as well as neonatal death.
STUDY DESIGN: A decision analytic and cost-effectiveness model was designed to compare use of magnesium for neuroprophylaxis vs no treatment for women at risk for preterm birth ⬍32 weeks due to preterm premature rupture of membranes or preterm labor from 24-32 weeks. Outcomes included neonatal death and moderatesevere cerebral palsy. Effectiveness was reported in quality-adjusted life years.
CONCLUSION: Based on currently published evidence for efficacy,
magnesium for neuroprophylaxis in women at risk to deliver preterm is cost-effective. Key words: cerebral palsy, magnesium, neuroprophylaxis, preterm birth
Cite this article as: Cahill AG, Odibo AO, Stout MJ, et al. Magnesium sulfate therapy for the prevention of cerebral palsy in preterm infants: a decision-analytic and economic analysis. Am J Obstet Gynecol 2011;205:542.e1-7.
A
fter years of published observational data suggesting a relationship between maternal exposure to magnesium sulfate and a decrease in adverse neonatal neurologic outcomes in infants born preterm,1-3 4 trials were published over the past 8 years with conflicting results regarding the impact of magnesium in the setting of pregnancies delivering preterm.4-7 The most recent of these trials by Rouse et al7 was a multicenter trial From the Departments of Obstetrics and Gynecology at Washington University in St. Louis School of Medicine, St. Louis, MO (Drs Cahill, Odibo, Stout, and Macones); Feinberg School of Medicine, Northwestern University, Chicago, IL (Dr Grobman); and Oregon Health and Science University, Portland, OR (Dr Caughey). Received May 27, 2011; revised Aug. 14, 2011; accepted Sept. 7, 2011. The authors report no conflict of interest. Presented as a poster at the 30th annual meeting of the Society for Maternal–Fetal Medicine, Chicago, IL, Feb. 4-6, 2010. Reprints not available from the authors. 0002-9378/$36.00 © 2011 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2011.09.004
542.e1
conducted in the United States, and the results offered the most promising evidence that magnesium decreased the risk of moderate or severe cerebral palsy (CP) in infants born preterm who survived. However, CP was not the primary outcome of this trial, but was part of a composite with death due to the design challenge of competing risks and there was a trend toward an increase in death in the infants exposed to magnesium compared to those exposed to placebo. In light of these recent and complex findings as well as the rare nature of these outcomes, it seems reasonable to include a formal estimate of cost-effectiveness of this therapy when considering a sweeping change in practice. We undertook this study to test the hypothesis that magnesium was cost-effective in women at risk for preterm birth for the prevention of CP, and further, to explore whether the cost-effectiveness was limited to particular subgroups of women.
M ATERIALS AND M ETHODS A decision-analytic and cost-effectiveness model was designed to compare 2 strategic approaches for the prevention
American Journal of Obstetrics & Gynecology DECEMBER 2011
of CP in preterm infants from a societal prospective. The use of magnesium for neonatal neuroprophylaxis was compared to nonuse, with the latter representative of the accepted standard of care prior to recent publications.7,8 The model was constructed for women carrying singleton pregnancies between 24 weeks 0/7 days’ and 31 weeks 6/7 days’ gestation at high risk for spontaneous preterm birth ⬍32 weeks’ gestation due to either preterm labor (PTL) or preterm premature rupture of membranes (PPROM) (Figure 1). Two additional models were constructed to estimate the cost-effectiveness of magnesium for neuroprophylaxis in at-risk subpopulations: PPROM only between 24 weeks 0/7 days’ and 31 weeks 6/7 days’ gestation, and patients at risk for preterm birth ⬍28 weeks (due to either PPROM or PTL). Cost was estimated in US dollars, effectiveness was reported as quality-adjusted lifeyears (QALYs), and strategies were compared in terms of number of cases of CP prevented, number of neonatal or infant deaths, and cost-effectiveness. Since only one of the placebo-controlled trials demonstrated statistically significant magnesium efficacy for the
Obstetrics
www.AJOG.org
TABLE 1
Probability, utility, and cost estimates for model Variable
Point estimate
Range
Source
Probabilities
..................................................................................................................................................................................................................................... 13-16
PTB ⬍32 wk
0.025
0.015–0.038
Proportion of PTB ⬍28 wk
0.250
0.200–0.300
PPROM ⬍32 wk
0.011
0.006–0.018
Preterm labor ⬍32 wk
0.020
0.010–0.080
Proportion of PTB after preterm labor ⬍32 wk
0.250
0.100–0.400
CP after PTB ⬍32 wk
0.035
0.020–0.061
CP after birth ⱖ32 wk
0.001
0.0004–0.003
Death after PTB ⬍32 wk
0.119
0.033–0.171
Death after birth ⱖ32 wk
0.0003
0.0002–0.0006
Relative risk of CP with magnesium
0.71
0.50–0.82
Relative risk of death with magnesium
1.06
1.00–1.15
Severe adverse magnesium reaction
0.047
0.001–0.081
Mild adverse magnesium reaction
0.696
0.540–0.920
..................................................................................................................................................................................................................................... 13,16,17 ..................................................................................................................................................................................................................................... 14,16,17 ..................................................................................................................................................................................................................................... 13 ..................................................................................................................................................................................................................................... 16,18,19
..................................................................................................................................................................................................................................... 4,5,7,20,21 ..................................................................................................................................................................................................................................... 22 ..................................................................................................................................................................................................................................... 4,5,7,20,21 ..................................................................................................................................................................................................................................... 22 ..................................................................................................................................................................................................................................... 4,5,7,20,21 ..................................................................................................................................................................................................................................... 4,5,7,20,21
..................................................................................................................................................................................................................................... 5,7,23-28 ..................................................................................................................................................................................................................................... 5,7,23-28 ..................................................................................................................................................................................................................................... 29
Maternal life expectancy, y
55.4
50.1–58.2
Neonatal life expectancy, y
77.2
70.3–82.0
Neonatal life expectancy, with CP, y
28.7
20.6–36.8
..................................................................................................................................................................................................................................... 29 ..................................................................................................................................................................................................................................... 30,31 ..............................................................................................................................................................................................................................................
Utilities
..................................................................................................................................................................................................................................... 32,33
Neonatal death
0.01
0.001–0.02
Moderate-severe CP
0.55
0.50–0.60
Normal child
1.0
..................................................................................................................................................................................................................................... 32,33 ..................................................................................................................................................................................................................................... 32,33
—
..............................................................................................................................................................................................................................................
Costs, $US
..................................................................................................................................................................................................................................... 34
Magnesium treatment
124.24
75.00–250.00
Severe magnesium reaction
145.40
80.95–287.88
..................................................................................................................................................................................................................................... 34 ..................................................................................................................................................................................................................................... 34
Mild magnesium reaction
2.52
0.16–5.04
..................................................................................................................................................................................................................................... 35
Moderate-severe CP
955,940
700,000–1,300,000
..................................................................................................................................................................................................................................... 36
Neonatal death after PTB ⬍32 wk
71,001
452–351,237
Neonatal survival after PTB ⬍32 wk
45,710
13,560–123,207
Neonatal death after birth ⱖ32 wk
67,758
466–283,910
..................................................................................................................................................................................................................................... 36 ..................................................................................................................................................................................................................................... 36 ..............................................................................................................................................................................................................................................
CP, cerebral palsy; PPROM, preterm premature rupture of membranes; PTB, preterm birth. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
reduction in risk of CP,7 magnesium exposure was modeled with the dosing protocol used in that trial. Magnesium treatment was initiated at presentation, and given as a 6-g bolus followed by 2 g/h infusion for 12 hours or until delivery. If delivery did not occur, a patient was eligible for retreatment when preterm delivery threatened again. It was assumed that all women eligible for retreatment
received magnesium again for purposes of cost estimates in the model. Clinical outcomes considered in the models were the occurrence of mild or severe adverse maternal reaction to magnesium, birth of a neonate with moderate or severe CP diagnosed by age 2 years, neonatal or infant death by 1 year, and unaffected neonate. The model was constructed with some notable definitions
Research
and assumptions. CP was diagnosed, and its severity determined, by accepted pediatric standards.9-11 PPROM was defined by clinical criteria described by the American College of Obstetricians and Gynecologists (ACOG),12 and all patients with PPROM delivered preterm (⬍32 weeks). PTL, defined clinically by contractions that change the cervical examination presenting between 24-32 weeks, was assumed to result in preterm delivery by 32 weeks 0 days in only a proportion of cases. The proportion of patients admitted with PTL going on to deliver preterm was based on available published evidence (Table 1). Baseline probabilities and utilities (or values given to a particular health-related outcome) were derived from the literature after an English-language PubMed search with the terms: “magnesium,” “neuroprophylaxis,” “preterm birth,” and “cerebral palsy.” Studies without control groups and reviews were excluded. Point estimates were calculated as weighted means by sample size, and ranges represented by the lowest and highest value for a given estimate in the published literature. For estimates derived from a single source, a range was defined by the 95% confidence intervals that was calculated from the binomial distribution. To derive the estimate of magnesium efficacy, the 4 trials in which magnesium was primarily given for the prevention of CP4-7,20 were used. Cost estimates and ranges were derived from published literature or based on Medicaid reimbursement rates.37 In cases of the latter, charges were multiplied by a cost-charge ratio of 0.6 as an approximation to third-party reimbursements, and regional variation was accounted for in sensitivity analyses by widely varying locally derived estimates. Effectiveness was calculated by the product of the utility value and discounted life expectancy in years. In accordance with standard assumptions, we assumed a life expectancy of 75 years and an annual discount rate for costs and QALYs of 3%.38 A threshold of $100,000 per QALY was chosen as the willingness-to-pay threshold.39 To enable comparison of a number of outcomes of interest across strategies, we considered a hypothetical cohort of 100,000 women at risk for preterm birth. Sensitivity analyses, performed by vary-
DECEMBER 2011 American Journal of Obstetrics & Gynecology
542.e2
Research
Obstetrics
ing ⱖ1 of the estimates along their entire range of potential values, were performed to test which value variation, if any, altered our results. Monte Carlo simulation, performed as a form of multivariable analysis that simultaneously incorporates the uncertainty in all point estimates by sampling from the distributions around these estimates, further tested the robustness of the model. Given the nature of the primary data, and specifically the debates regarding interpretation of the trials individually or in combination regarding the association of magnesium with CP as well as fetal death, we widely varied the estimates of effect in our sensitivity analyses to investigate robustness of the model and investigate threshold values. We conducted similar threshold analyses by varying the background rate of preterm birth within the model, in an effort to indirectly mimic the strategies for a population at high risk for preterm birth. Models were constructed and analyzed using TreeAge Pro 2007 (TreeAge Software, Williamstown, MA). This study was exempt from institutional review board approval.
R ESULTS Under base-case assumptions, magnesium for neuroprophylaxis for all women at high risk for spontaneous preterm birth was the dominant (least costly and most effective) strategy when compared to no magnesium. For every 10,000 women at risk for preterm birth treated with magnesium, $1.8 million were saved and 52 QALYs were gained. Compared to no magnesium, magnesium was also the dominant strategy when given to only women with PPROM as well as when given only to women at risk to deliver ⬍28 weeks. For every 10,000 women with PPROM treated with magnesium, $1.4 million were saved and 51 QALYs were gained (Table 2). Considering a hypothetical cohort of 100,000 at high risk for preterm birth, magnesium treatment resulted in the greatest reduction in the number of cases of moderate or severe CP (n ⫽ 32). While the strategy of magnesium compared to none within the subgroups of women with PPROM only and those at 542.e3
www.AJOG.org
TABLE 2
Base-case cost-effectiveness of magnesium for neuroprophylaxis Strategy
Cost, $
Incremental cost, $
Efficacy (QALY)
Incremental efficacy
Incr C/E (ICER)
All women
.....................................................................................................................................................................................................................................
Magnesium
1739.00
—
56.6836
—
—
No magnesium
1917.20
178.20
56.6784
⫺0.0052
(Dominated)
..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................
PPROM only
.....................................................................................................................................................................................................................................
Magnesium
1462.60
—
56.7022
—
—
No magnesium
1607.50
144.90
56.6972
⫺0.0050
(Dominated)
..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................
⬍28 wk
.....................................................................................................................................................................................................................................
Magnesium
920.60
—
56.7411
—
—
56.7355
⫺0.0056
(Dominated)
.....................................................................................................................................................................................................................................
No magnesium
1019.00
98.50
..............................................................................................................................................................................................................................................
C/E, cost-effectiveness; ICER, incremental cost-effectiveness ratio; Incr, incremental; PPROM, preterm premature rupture of membranes; QALY, quality-adjusted life-year. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
risk to deliver ⬍28 weeks resulted in reductions in the annual number of cases of CP, the absolute number prevented was 12% fewer when applied to the PPROM-only group (n ⫽ 28) and 28% fewer when applied to only those women at risk for preterm birth ⬍28 weeks (n ⫽ 23) (Table 3). Sensitivity analyses showed our results to be robust to the wide ranges of values considered for many inputs. However,
sensitivity analysis revealed the model to be sensitive to estimates of effect of magnesium on risk of moderate or severe CP as well as neonatal death. When the risk reduction in moderate to severe CP conferred by magnesium treatment was ⬍14%, magnesium was no longer costeffective (Figure 2). Similarly, a relative risk of neonatal death associated with magnesium exposure of ⱖ1.16 resulted in no magnesium sulfate treatment be-
TABLE 3
Cases of moderate-severe cerebral palsy and death prevented per dollar spent
Strategy
No. of cases of moderate or severe CP
Total cost ($10 million)
No. of cases of moderate or severe CP prevented
Total cost saved ($10 million)
All women
.....................................................................................................................................................................................................................................
Magnesium
78
17.4
32
1.8
110
19.2
Reference
Reference
.....................................................................................................................................................................................................................................
No magnesium (reference)
..............................................................................................................................................................................................................................................
PPROM only
.....................................................................................................................................................................................................................................
Magnesium
70
14.6
28
1.5
No magnesium (reference)
98
16.1
Reference
Reference
.....................................................................................................................................................................................................................................
..............................................................................................................................................................................................................................................
⬍28 wk only
.....................................................................................................................................................................................................................................
Magnesium
58
9.2
No magnesium (reference)
81
10.2
23
1.0
Reference
Reference
.....................................................................................................................................................................................................................................
..............................................................................................................................................................................................................................................
CP, cerebral palsy; PPROM, preterm premature rupture of membranes. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
American Journal of Obstetrics & Gynecology DECEMBER 2011
Obstetrics
www.AJOG.org
Research
FIGURE 1
Schematic of decision-analytic tree for magnesium for neuroprophylaxis
Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
coming the dominant strategy. Lastly, when the background rate of preterm birth was varied widely, even when the background risk of preterm birth was as high as 15%, magnesium for neuroprophylaxis remained the dominant strategy, holding all other assumptions constant. To test the hypothesis that it is not cost-effective to give magnesium for neuroprophylaxis to all women admitted with PTL because many of these women will not go on to have a subsequent preterm birth, we explored the threshold value of preterm birth after PTL between 24-32 weeks. Our model revealed that even if only 10% of women admitted with a diagnosis of
PTL went on to have a preterm birth ⬍32 weeks, magnesium therapy for all women at risk for preterm birth due to PTL was cost-effective. In Monte Carlo simulation, which functions as a form of multivariate analysis allowing all estimates in the model to be simultaneously varied across their plausible ranges, magnesium for all pregnancies at risk for preterm birth ⬍32 weeks was the preferred strategy 86.7% of the time and was dominant 63.4% of the time (Figure 3). In a 10,000 sample Monte Carlo simulation for both subanalyses (PPROM and women delivering ⬍28 weeks), and a willingness-topay threshold of $100,000/QALY, mag-
nesium was the preferred strategy 100% of the time.
C OMMENT Despite the fact that the correlation between admissions for PTL and subsequent preterm birth is not high,18,40-42 our model demonstrated that magnesium for neuroprophylaxis is cost-effective in women at high risk for spontaneous preterm delivery ⬍32 weeks. And, while magnesium therapy, compared to a strategy of no magnesium, was the dominant strategy for all women at risk for preterm delivery as well as within the subgroup of women diagnosed with
DECEMBER 2011 American Journal of Obstetrics & Gynecology
542.e4
Research
Obstetrics
FIGURE 2
Sensitivity analysis on relative risk reduction of cerebral palsy from magnesium (MS04)
QALY, quality-adjusted life-year. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
PPROM and the subgroup who deliver ⬍28 weeks, magnesium for all women at risk for spontaneous preterm delivery resulted in the greatest reduction in number of cases of moderate or severe CP and the largest number of dollars saved. However, sensitivity analysis revealed the model to be sensitive to estimates of effect of magnesium on risk of moderate or severe CP as well as neonatal death. Additionally, in multivariable iterative analyses, the magnesium strategy compared with none was chosen 100% of the time for the subgroups of women with PPROM and those at risk for delivery ⬍28 weeks; for all women at risk for preterm delivery, it was only the preferred strategy 86.7% of the time, reflective of both the poor correlation between the diagnosis of PTL and subsequent preterm birth as well as the published data regarding magnesium for neuroprophylaxis. Since the body of evidence regarding magnesium for neuroprophylaxis is challenging to interpret, as many authors have illustrated,43-46 and the extrapolation into clinical practice complex, several investigators have used the technique of metaanalysis to formally pool the results from the primary trials and better estimate the effect of magnesium on risk for CP and neonatal death.8,47 However, one criticism that has been 542.e5
www.AJOG.org made with regard to interpretation of these metaanalyses is that the patient populations studied in each of the 4 trials varied significantly. Specifically, 88% of the patients enrolled in the trial by Rouse et al7 had a diagnosis of PPROM, while the most common diagnosis in the study by Crowther et al5 was PTL. In this analysis, we were able to specifically consider whether the application of magnesium therapy for neuroprophylaxis would be most cost-effectively applied to only women with PPROM as has been suggested by some, or to all women at risk for spontaneous preterm birth ⬍32 weeks including women admitted with PTL and cervical dilation. Our findings that magnesium for all women at high risk for spontaneous preterm birth was cost-effective even when only 10% of PTL admissions went on to deliver preterm supports the use of magnesium for all women at risk. It is notable that while magnesium, compared to no therapy, was the dominant strategy in base-case analyses, our Monte Carlo simulation demonstrated magnesium to be the preferred strategy only 86.7% of the time in the ⬍32-week model. This can be viewed in 2 ways. One would be to consider these findings not entirely robust as they do not strictly comply with the 95% confidence interval that is generally preferred. The second would be to realize that, unfortunately, the data on which this analysis is based contain some uncertainty. Thus, we would offer that these findings could be considered a numeric estimate of the complexities and uncertainty of the pooled data. In practical terms, the total number of women and neonates in all of these studies combined remains relatively small with respect to the rare incidence of the outcomes considered. Slight variation in the estimates of effect of magnesium on risk for CP and death can impact the cost-effectiveness of magnesium treatment. These findings from our cost-effectiveness model are consistent with the recent committee opinion from ACOG48 on the use of magnesium sulfate before anticipated preterm birth for neuroprotection. The committee recognized that “none of the individual studies found a
American Journal of Obstetrics & Gynecology DECEMBER 2011
FIGURE 3
Multivariate sensitivity analyses
Outcomes of 10,000 trials from Monte Carlo simulation. Each blue dot represents single trial outcome. Ellipse marks 95% confidence interval. Dashed line represents willingness-to-pay threshold of $100,000; all dots below threshold are cost-effective. QALY, quality-adjusted life-year. Cahill. Magnesium sulfate for prevention of CP. Am J Obstet Gynecol 2011.
benefit with regard to their primary outcome” and that “comparison is made difficult by differences in inclusion and exclusion criteria, populations studied, magnesium sulfate regimens, gestational age with treatment, and outcomes studied between the trials.” However, the committee further went on to conclude that “the available evidence suggests that magnesium sulfate given before anticipated early preterm birth reduces the risk of cerebral palsy in surviving infants.” As with all theoretical models, there are important potential limitations that should be considered carefully when interpreting our results. The data modeled are limited to the available published literature. In addition, models such as these are based on a set of assumptions that can impact both internal validity as well as generalizability, and with this in mind we have outlined these assumptions carefully in our “Material and Methods” section. Specific to our question of the effect of magnesium on risk for CP and death in neonates born preterm, the magnitude of effect was assumed to be fixed across preterm gestational ages of exposure. However, it is possible that magnesium has a greater or
Obstetrics
www.AJOG.org lesser effect depending on the gestational age at which the therapy is applied. But currently published data did not allow us to model this potential variation in an evidence-based manner beyond our stratification at 28 weeks of gestation. While the existing evidence is complex, it appears unlikely that additional randomized trials of magnesium to prevent CP will be done. With this in mind, we believe our model aids in the consideration and interpretation of use of magnesium for neuroprophylaxis in pregnancies at high risk to deliver preterm, and allowed us to use sensitivity analyses to explore the uncertainty in the published efficacy data. First, magnesium compared to no therapy is cost-effective. Second, based on our model, it is best applied to all women at high risk for spontaneous preterm birth ⬍32 weeks as defined by the inclusion criteria of the recently published efficacy trial,7 despite the fact that many women admitted and treated for PTL between 24-32 weeks go f on to deliver ⬎34 weeks. REFERENCES 1. van de Bor M, Verloove-Vanhorick SP, Brand R, Keirse MJ, Ruys JH. Incidence and prediction of periventricular-intraventricular hemorrhage in very preterm infants. J Perinat Med 1987;15:333-9. 2. Kuban KC, Leviton A, Pagano M, Fenton T, Strassfeld R, Wolff M. Maternal toxemia is associated with reduced incidence of germinal matrix hemorrhage in premature babies. J Child Neurol 1992;7:70-6. 3. Nelson KB, Grether JK. Can magnesium sulfate reduce the risk of cerebral palsy in very low birthweight infants? Pediatrics 1995;95:263-9. 4. Mittendorf R, Dambrosia J, Pryde PG, et al. Association between the use of antenatal magnesium sulfate in preterm labor and adverse health outcomes in infants. Am J Obstet Gynecol 2002;186:1111-8. 5. Crowther CA, Hiller JE, Doyle LW, Haslam RR. Effect of magnesium sulfate given for neuroprotection before preterm birth: a randomized controlled trial. JAMA 2003;290: 2669-76. 6. Marret S, Marpeau L, Zupan-Simunek V, et al. Magnesium sulphate given before very-preterm birth to protect infant brain: the randomized controlled PREMAG trial*. BJOG 2007; 114:310-8. 7. Rouse DJ, Hirtz DG, Thom E, et al. A randomized, controlled trial of magnesium sulfate for the prevention of cerebral palsy. N Engl J Med 2008;359:895-905.
8. Doyle LW, Crowther CA, Middleton P, Marret S, Rouse D. Magnesium sulphate for women at risk of preterm birth for neuroprotection of the fetus. Cochrane Database Syst Rev 2009;1: CD004661. 9. Blasco PA. Primitive reflexes: their contribution to the early detection of cerebral palsy. Clin Pediatr (Phila) 1994;33:388-97. 10. Capute AJ, Shapiro BK. The motor quotient: a method for the early detection of motor delay. Am J Dis Child 1985;139:940-2. 11. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997;39:214-23. 12. American College of Obstetricians and Gynecologists. ACOG Committee on Practice Bulletins–Obstetrics. ACOG practice bulletin no. 80: premature rupture of membranes; clinical management guidelines for obstetrician-gynecologists. Obstet Gynecol 2007;109:1007-19. 13. Martin JA, Kochanek KD, Strobino DM, Guyer B, MacDorman MF. Annual summary of vital statistics–2003. Pediatrics 2005;115: 619-34. 14. Moster D, Lie RT, Markestad T. Long-term medical and social consequences of preterm birth. N Engl J Med 2008;359:262-73. 15. Tucker JM, Goldenberg RL, Davis RO, Copper RL, Winkler CL, Hauth JC. Etiologies of preterm birth in an indigent population: is prevention a logical expectation? Obstet Gynecol 1991;77:343-7. 16. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008;371:75-84. 17. Mercer BM. Preterm premature rupture of the membranes. Obstet Gynecol 2003;101: 178-93. 18. Macones GA, Segel SY, Stamilio DM, Morgan MA. Prediction of delivery among women with early preterm labor by means of clinical characteristics alone. Am J Obstet Gynecol 1999;181:1414-8. 19. Iams JD. Prediction and early detection of preterm labor. Obstet Gynecol 2003;101: 402-12. 20. Marret S, Marpeau L, Benichou J. Benefit of magnesium sulfate given before very preterm birth to protect infant brain. Pediatrics 2008; 121:225-6. 21. Magpie Trial Follow-Up Study Collaborative Group. The Magpie trial: a randomized trial comparing magnesium sulphate with placebo for pre-eclampsia; outcome for children at 18 months. BJOG 2007;114:289-99. 22. McIntire DD, Leveno KJ. Neonatal mortality and morbidity rates in late preterm births compared with births at term. Obstet Gynecol 2008;111:35-41. 23. Chau AC, Gabert HA, Miller JM Jr. A prospective comparison of terbutaline and magnesium for tocolysis. Obstet Gynecol 1992;80:847-51. 24. El-Sayed YY, Riley ET, Holbrook RH Jr, Cohen SE, Chitkara U, Druzin ML. Randomized comparison of intravenous nitroglycerin and
Research
magnesium sulfate for treatment of preterm labor. Obstet Gynecol 1999;93:79-83. 25. Glock JL, Morales WJ. Efficacy and safety of nifedipine versus magnesium sulfate in the management of preterm labor: a randomized study. Am J Obstet Gynecol 1993;169:960-4. 26. Larmon JE, Ross BS, May WL, Dickerson GA, Fischer RG, Morrison JC. Oral nicardipine versus intravenous magnesium sulfate for the treatment of preterm labor. Am J Obstet Gynecol 1999;181:1432-7. 27. McWhorter J, Carlan SJ, OLeary TD, Richichi K, OBrien WF. Rofecoxib versus magnesium sulfate to arrest preterm labor: a randomized trial. Obstet Gynecol 2004;103:923-30. 28. Terrone DA, Rinehart BK, Kimmel ES, May WL, Larmon JE, Morrison JC. A prospective, randomized, controlled trial of high and low maintenance doses of magnesium sulfate for acute tocolysis. Am J Obstet Gynecol 2000; 182:1477-82. 29. Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/nchs/data/ nvsr/nvsr59/nvsr59_04.pdf. Accessed Aug. 10, 2011. 30. Blair E, Watson L, Badawi N, Stanley FJ. Life expectancy among people with cerebral palsy in Western Australia. Dev Med Child Neurol 2001;43:508-15. 31. Strauss D, Brooks J, Rosenbloom L, Shavelle R. Life expectancy in cerebral palsy: an update. Dev Med Child Neurol 2008;50:487-93. 32. Pham CT, Crowther CA. Birth outcomes: utility values that postnatal women, midwives and medical staff express. BJOG 2003;110:121-7. 33. Vandenbussche FP, De Jong-Potjer LC, Stiggelbout AM, Le Cessie S, Keirse MJ. Differences in the valuation of birth outcomes among pregnant women, mothers, and obstetricians. Birth 1999;26:178-83. 34. Hayes E, Moroz L, Pizzi L, Baxter J. A cost decision analysis of 4 tocolytic drugs. Am J Obstet Gynecol 2007;197:383.e1-6. 35. Waitzman NJ, Romano PS, Scheffler RM. Estimates of the economic costs of birth defects. Inquiry 1994;31:188-205. 36. Phibbs CS, Schmitt SK. Estimates of the cost and length of stay changes that can be attributed to one-week increases in gestational age for premature infants. Early Hum Dev 2006;82:85-95. 37. Medicare and Medicaid statistical supplement, 1995. US Department of Health and Human Services, Health Care Financing Administration. Health Care Financ Rev Stat Suppl 1995:1-388. 38. Drummond MF, Barbieri M, Wong JB. Analytic choices in economic models of treatments for rheumatoid arthritis: what makes a difference? Med Decis Making 2005;25:520-33. 39. Owens DK. Interpretation of cost-effectiveness analyses. J Gen Intern Med 1998; 13:716-7. 40. Hueston WJ. Preterm contractions in community settings, II: predicting preterm birth in women with preterm contractions. Obstet Gynecol 1998;92:43-6.
DECEMBER 2011 American Journal of Obstetrics & Gynecology
542.e6
Research
Obstetrics
41. Hueston WJ. Preterm contractions in community settings, I: treatment of preterm contractions. Obstet Gynecol 1998;92:3842. 42. King JF, Grant A, Keirse MJ, Chalmers I. Beta-mimetics in preterm labor: an overview of the randomized controlled trials. Br J Obstet Gynaecol 1988;95:211-22. 43. Macones GA. MgSO4 for CP prevention: too good to be true? Am J Obstet Gynecol 2009;200:589.
542.e7
www.AJOG.org 44. Cahill AG, Caughey AB. Magnesium for neuroprophylaxis: fact or fiction? Am J Obstet Gynecol 2009;200:590-4. 45. Rouse DJ. Magnesium sulfate for the prevention of cerebral palsy. Am J Obstet Gynecol 2009;200:610-2. 46. Stanley FJ, Crowther C. Antenatal magnesium sulfate for neuroprotection before preterm birth? N Engl J Med 2008;359:962-4. 47. Conde-Agudelo A, Romero R. Antenatal magnesium sulfate for the prevention of cere-
American Journal of Obstetrics & Gynecology DECEMBER 2011
bral palsy in preterm infants less than 34 weeks’ gestation: a systematic review and metaanalysis. Am J Obstet Gynecol 2009; 200:595-609. 48. American College of Obstetricians and Gynecologists Committee on Obstetric Practice; Society for Maternal-Fetal Medicine. Committee opinion No. 455: magnesium sulfate before anticipated preterm birth for neuroprotection. Obstet Gynecol 2010;115: 669-71.