Despite the advances in screening and Rh immune globulin administration, 1% of Rh(D)-negative mothers still become alloimmunized during pregnancy.1 The indirect antiglobulin (Indirect Coombs) test is currently recognized as the reference standard for detecting the anti-D antibody in pregnant women who have become sensitized to Rh(D)-positive blood of the fetus.2 When an indirect antiglobulin test is positive, it is followed by a red blood cell antibody panel.3 Currently, the U.S. Preventative Service Task Force recommends Rh(D) blood typing and anti-D antibody screening for pregnant women at their first prenatal visit and repeat screening at 24–28 weeks of gestation for Rh(D)-negative mothers who are not isoimmunized.4 The primary rationale for performing the second antibody screening is that identification of Rh(D) alloimmunization by 28 weeks of gestation will improve management of isoimmunized pregnancies and prevent unnecessary administration of Rh immune globulin.5,6
However, as a result of a lack of evidence regarding the cost–benefit of the second antibody screen, there is little consensus among professional societies with regard to the best practice. In particular, the American College of Obstetricians and Gynecologists (the College) suggests that the decision regarding a second antibody screen be left to the health care provider.7 The College stated that the efficacy of the second antibody screen is questionable, because the incidence of Rh alloimmunization occurring before 28 weeks of gestation is less than 0.187%,8 and no studies have yet analyzed the cost-effectiveness of the screen.7 In 1978, Bowman et al8 found that 2 of 1,086 (0.187%) Rh(D)-negative women demonstrated Rh(D) isoimmunization at the time of injection of Rh immune globulin at 28 weeks of gestation. Those who do seroconvert between their first prenatal visit and 28 weeks of gestation experience minimal risk to the fetus during the index pregnancy.9–12
The rates of isoimmunization of Rh(D)-negative women before the 28th week have likely continued to decrease since 1978 as a result of enhanced monitoring and Rh immune globulin prophylaxis.13 Thus, in our study, we recalculated this seroconversion rate before the 28th week and constructed a decision tree model to analyze the costs and benefits and under what circumstances elimination of the second indirect antiglobulin test for Rh(D)-negative mothers during pregnancy would be cost-beneficial.
MATERIALS AND METHODS
After obtaining institutional review board approval from the University of Washington, we reviewed patient charts to calculate an updated seroconversion rate before 28 weeks of gestation. A chart review was completed reviewing all Rh(D)-negative mothers who delivered at the University of Washington Medical Center and affiliated hospitals over the 10-year period from January 1, 2002, to December 31, 2011. The initial database search was conducted to select Rh(D)-negative mothers with one or more positive indirect antiglobulin screens during their pregnancy. Patients who were seen for a consult at the hospital but delivered elsewhere were not included. All charts were initially reviewed electronically, and for those charts where incomplete information was available, the paper chart was ordered.
Rh(D)-negative mothers with positive indirect antiglobulin screens were further divided into: 1) nonsensitized: anti-Rh(D) positive from Rh(D) immune globulin administration or Rh(D)-negative fetus (anti-Rh(D) sensitization impossible); 2) sensitized: prior sensitization (positive anti-Rh[D] antibody screens before pregnancy) or gestational sensitization (conversion to positive anti-Rh[D] titers during pregnancy); and 3) insufficient information.
Pregnancies with positive indirect antiglobulin tests at delivery and negative screens at 28 weeks of gestation were not included in the rate, because they were either sensitized after 28 weeks of gestation or sensitized as a result of Rh immune globulin. Those patients with insufficient data in their patient charts such that antibody or Rh immune globulin administration status could not be determined were excluded from the calculations. The updated seroconversion rate was necessary to obtain before the cost–benefit analysis of the second antibody screen at 28 weeks of gestation could be conducted.
The percent of women with a false-positive indirect antiglobulin test at 28 weeks of gestation was also calculated from the patient chart review. False-positives were defined as women who tested positive for the indirect antiglobulin test at 28 weeks of gestation but had one or more of the following explanations for their positive test: amniocentesis or chorionic villus sampling followed by Rh immune globulin, vaginal bleeding, presumed intrauterine bleeding or trauma followed by Rh immune globulin, or other unexplained use of Rh immune globulin before 28 weeks of gestation.
To begin our cost–benefit analysis of antibody screening at 28 weeks of gestation, we used a decision tree model to compare two strategic approaches to anti-Rh(D) antibody testing during pregnancy (Fig. 1). The first strategy is the current standard of care advocated by the U.S. Preventative Services Task Force: screening for anti-Rh(D) antibodies in all Rh(D)-negative women at their first prenatal visit and again at 24–28 weeks of gestation if not isoimmunized and without a known Rh(D)-negative father.4 If no anti-Rh(D) antibodies are found in the antiglobulin assay at 28 weeks of gestation, the woman receives Rh immune globulin at that time. If antibodies are detected, the woman receives an antibody panel to identify the type of antibody. In our hypothetical second strategy, the second anti-Rh(D) antibody screen at 28 weeks of gestation is not performed unless the woman is already known to be isoimmunized (in which case her titers are followed throughout pregnancy). In this strategy, anti-Rh(D) immune globulin is administered to every Rh(D)-negative, previously antibody-negative woman at 28 weeks of gestation unless otherwise indicated (eg, the woman has an Rh[D]-negative fetus known from amniocentesis).
The baseline probabilities and outcomes for each strategy were derived from original research (described previously) and performance of a literature review followed by a bibliographic survey. We used the following search terms in PubMed: Rh isoimmunization at 28 weeks of gestation, isoimmunization, seroconversion, Rh immune globulin, pregnancy, and hemolytic disease of the newborn. The baseline seroconversion rate of 0.099% was obtained by patient chart review analysis at the University of Washington. The rate was varied in the sensitivity analysis from 0.099% to 0.187% based on the literature review. Only one other rate of seroconversion before 28 weeks of gestation was obtained in the literature review: 0.187%.8 The baseline rate of false-positive indirect antiglobulin tests at 28 weeks of gestation was also obtained by patient chart review. The rate was normally distributed with mean 8.02% and range ±5%. The number of Rh(D)-negative mothers delivering at the University of Washington from 2002 to 2012 was obtained through use of the University of Washington Perinatal Database.
The base-case estimates regarding U.S. birth rate and the number of Rh(D)-negative deliveries were obtained using 2010 U.S. Census Data and Rh(D)-negativity rates as reported by the College. Rh(D)-negative deliveries in the United States per year were extrapolated given total U.S. births in 2010 (4.2 million) and the percentage of the population that is Rh(D)-negative (8–15%, depending on race).14,15
Assumptions in our model were made as indicated by current standards: women receiving amniocentesis, invasive procedures, experiencing threatened miscarriages, antepartum hemorrhages, or other indicated medical conditions before 28 weeks of gestation are administered Rh immune globulin at the time of the procedure,16 causing a positive indirect antiglobulin test result at the second antibody screen. Most laboratory blood testing centers in the United States perform a red cell antibody panel test when a positive indirect antiglobulin test is obtained to identify the antibody in question.3 Our proposed second strategy of the decision tree—elimination of the second indirect antiglobulin test—also eliminates the need for the reflexive red cell antibody panel at week 28.
The decision tree was created using TreeAge Pro 2010, and analysis was performed in TreeAge Pro and Microsoft Excel 2010. All probability estimates used in the analysis were obtained through chart review or derived from published literature (Table 1).
Charge data were also derived from published literature and from personal communication with local and national blood centers (Table 1). Screening charges for the indirect antiglobulin test and the red cell antibody panel were determined from available data from the regional blood testing centers. Charges for the indirect antiglobulin test ranged from $37 to $73, depending on the blood center. Charges for the red cell antibody panel ranged from $70 to $271. Other hospitals contacted indicated a range of charges for the indirect antiglobulin test from $38 to $73 (Puget Sound Blood Center, University of Colorado Hospital, The University of Texas Medical Branch at Galveston, personal communication, 2012).
Charges related to Rh immune globulin administration were determined from available hospital data (University of Washington Medical Center, Swedish Hospital, personal communication, 2012). The charge for the immune globulin started at a baseline of $141.50, according to the Red Book, which provides average drug pricing information to health care providers.17 The highest immune globulin charge to the patient was $205 (Swedish Hospital, personal communication, June 5, 2012). The immune globulin and antibody screening charges do not include the personnel costs of administering the drug, because we have assumed that the patient must already visit their health care provider and receive other prenatal tests at that time.
For a cohort of 100,000 women, we calculated for each strategy the cost and benefits of continuing with the current method of antibody screening at two intervals per pregnancy (strategy 1) compared with eliminating the second antibody screen in favor of universal Rh immune globulin administration for Rh(D)-negative women who are antibody-negative (strategy 2) at 28 weeks of gestation. The outcome of the model was the net present value (benefits minus costs) of the proposed strategies with strategy 2 defined as “net benefits” and strategy 1 as “net costs.” Univariable sensitivity analyses were performed to vary input parameters to their logical extremes.
A Monte Carlo analysis was conducted to capture a range of uncertain parameters (eg, Rh immune globulin cost, indirect antiglobulin test cost) simultaneously and thereby construct a range of cost-savings values. The Monte Carlo then randomly “draws” from each parameter distribution, calculating a range of net benefits (eg, cost savings). This random “draw” is conducted 10,000 times, each time generating a net benefit estimate.
The Monte Carlo proceeded as follows within each iteration. The percentage of women false-positive was sampled from a normal distribution with a mean of 8.02%. The bounds on the percentage are ±5%. To accomplish the bounding, the distribution was truncated at 3.02% and 13.02% and the spread was set such that 5% is 3 standard deviations from the mean to include the majority of values (approximately 99.7%). This iteration-specific false-positive rate was used to determine a random binomial flag for each woman indicating prior Rh immune globulin injection. Each woman received a randomly selected indirect antiglobulin test cost and red cell antibody panel cost rounded to the nearest penny. Given these inputs, the women then received two random seroconversion flags based on the percent of women isoimmunized. Finally, these parameter values and flags are combined with each Rh immune globulin cost.
Strategy 1 and strategy 2 costs are calculated per 100,000 women as follows with indirect antiglobulin test indicating the indirect antiglobulin test cost at 8 and 28 weeks of gestation, ABP indicating the red cell antibody panel cost, Rh immune globulin indicating Rh immune globulin injection cost, FP indicating false-positive flag, and SI indicating the seroconversion flag. Strategy 1 had an 8-week indirect antiglobulin test for all women, Rh immune globulin injection for those undergoing a Rh immune globulin-necessitating event before 28 weeks of gestation (eg, amniocentesis) and thus are false-positive at 28 weeks of gestation, 28-week indirect antiglobulin test for all women, red cell antibody panel for false-positive and isoimmunized women, and Rh immune globulin injection for mothers who did not seroconvert. Strategy 2 has an 8-week indirect antiglobulin test for all women, Rh immune globulin injection for those undergoing a Rh immune globulin-necessitating event before 28 weeks of gestation (eg, amniocentesis), and Rh immune globulin injection for all women at 28 weeks of gestation.
- Strategy 1: IAT+Rh immune globulin×FP+IAT+ABP×max(FP+SI)+(1-SI)×Rh immune globulin
- Strategy 2: IAT+Rh immune globulin×FP+Rh immune globulin
We calculated an Rh(D)-antibody seroconversion rate before 28 weeks of gestation through a patient chart review. In a population of 20,891 deliveries (Table 2) recorded at the University of Washington Medical Center from the 10-year period of 2002–2011, 2,029 were to Rh(D)-negative women (9.8%). This percent is lower than the national average of 12.1%,14 potentially as a result of the large Asian population served by the University of Washington Medical Center. Of these mothers, 997 recorded a positive indirect antiglobulin test. Within this population, 2.02% displayed irregular antibodies (nonanti-Rh[D] antibodies). Of the 982 mothers with anti-Rh(D) antibodies, 881 (89.7%) mothers had a positive indirect antiglobulin test result attributable to Rh immune globulin administration, 52 (5.3%) were sensitized during prior pregnancies, 31 (3.2%) were ruled out as a result of Rh(D)-negative fetuses, and 11 were missing information needed for classification (1.1%). Those missing information were primarily transfer patients to the University of Washington Medical Center, whose records were incomplete and thus excluded from analysis.
Only two patients out of the 2,029 Rh(D)-negative deliveries underwent gestational sensitization before 28 weeks of gestation, yielding a seroconversion rate of 0.099% (Table 2). The first sensitization occurred before 28 weeks of gestation, with a documented negative indirect antiglobulin test screen on April 1, 2006, and a positive indirect antiglobulin test screen at 28 weeks of gestation, on July 13, 2006. The patient delivered on September 29, 2006. A positive direct Coombs test of the neonate indicated antibodies bound to neonatal red blood cells in vivo. The second seroconversion occurred in a patient with a negative indirect antiglobulin test screen on December 29, 2003, and a positive screen at 27 weeks of gestation on March 22, 2004. Neither patient had received blood transfusions. These two cases represent seroconversions before the traditional Rh immune globulin administration at 28 weeks of gestation. This represents a 0.099% seroconversion rate.
Using our calculated seroconversion rate, we conducted a cost–benefit analysis of Rh(D)-antibody screening at 28 weeks of gestation. Using our baseline probability and cost estimates, the Monte Carlo analysis indicates average savings of approximately $6.8 million with a standard deviation of approximately $281,000 for 100,000 women (Table 3). Implementation of the strategy 2 (proposed strategy) results in an average cost of $27,690,629, whereas continuation of strategy 1 (current protocol) results in an average cost of $34,548,836 (Table 3). Strategy 2 has the same total cost within a given Rh immune globulin cost because the calculation does not involve Rh immune globulin administration based on seroconversion status. The total strategy 1 cost changes slightly based on Rh immune globulin cost because it includes a woman's isoimmunization status. The smallest savings of $6,798,836 occurs when Rh immune globulin cost is $500, whereas the largest savings of $6,865,905 occurs when Rh immune globulin cost is $141.50.
Implementing the decreased antibody screening strategy in the United States for 1 year will result in cost savings of approximately $35 million. This is an average cost savings of approximately $68 per Rh(D)-negative delivery. As a result of the low seroconversion rate of 0.099%, in 100,000 women only 99 unnecessary Rh immune globulin shots would be administered per year to anti-D-positive women, whereas 100,000 second antibody screens at 28 weeks of gestation would be avoided.
An analysis comparing the total cost savings across different Rh immune globulin prices within each seroconversion rate was done. The median, quantiles, and distribution are similar for Rh immune globulin costs of $141.50 and $205.48. The analyses show a shift to slightly lower savings when the Rh immune globulin price is $500. An Rh immune globulin cost of approximately $69,000 per injection would be needed to equalize the costs of the two strategies. The results of the Monte Carlo analysis can also be analyzed based on groupings, each representing the 10,000 iterations (or random “draws”) conducted using our baseline parameters and given a particular Rh immune globulin price and seroconversion rate. The range of cost savings seen in this analysis is approximately $6 million to $7.7 million. The smallest observed cost savings, $5,956,000, occurred with $500.00 Rh immune globulin cost and 0.187% seroconversion rate. The largest observed cost savings, $7,712,000, occurred with an Rh immune globulin cost $141.50 and seroconversion rates 0.099% and 0.187%.
Interestingly, if the cost of the Rh immune globulin injection is less than the cost of the red cell antibody panel (eg, Rh immune globulin cost of $141.50 and red cell antibody panel cost of $170.50), increasing the seroconversion rate will result in greater cost savings. This is the result of the fact that under these conditions, the money saved from eliminating the red cell antibody panel at 28 weeks of gestation for women who seroconvert outpaces the increased unnecessary Rh immune globulin injections delivered to those women who seroconvert.
A series of univariate sensitivity analyses were performed to estimate the change in the primary outcome (cost savings/y in U.S. dollars) by varying the input parameters to their logical extremes. These analyses revealed that the proposed strategy 2 continues to dominate under all plausible ranges of seroconversion rates (Fig. 2). The net benefits accrued to the United States per year from implementing the decreased screening strategy would remain above $6.8 million in a 100,000-woman cohort, even assuming a seroconversion rate reaching 0.5%.
To accurately assess costs of the two strategies presented, this study recalculates the rate of seroconversion between the first trimester and 28-week antibody screen. We identified a lower rate of seroconversion than previously found. However, given the small number of seroconversions per study (2/2,029 compared with 2/1,086), these rates are not statistically different. We suspect fewer cases were identified in our study as a result of several factors. Given the sensitivity of human chorionic gonadotropin testing in serum and urine, we are identifying many pregnancies that experience bleeding or even go on to miscarriage and treating these women with Rh immune globulin. Additionally, the number of first- and second-trimester procedures (ultrasonography, chorionic villous sampling, and amniocentesis) is increasing, leading to large numbers of women receiving Rh immune globulin before 28 weeks of gestation. Our numbers represent a tertiary care center with an active prenatal diagnosis program, and we may see more use of early Rh immune globulin than at community centers. There may be some variation in the rate of seroconversion before 28 weeks of gestation; likely the national rate lies somewhere between the two estimates.
In this study, we evaluated the cost benefit of a reduced Rh(D) antibody screening regimen before 28 weeks of gestation for Rh(D)-negative women. The cost savings for implementing this plan would range from approximately $6 to $7.7 million for a cohort of 100,000 women. This strategy does not hinge on our calculated seroconversion rate of 0.099% between the first prenatal visit and 28 weeks of gestation, at which point Rh immune globulin would be administered to all Rh(D)-negative, antibody-negative women. These strong cost savings remain high when all parameters are varied to their logical extremes. Therefore, we propose eliminating the 28-week antibody screen as a result of the low seroconversion rate (0.099–0.187%) before 28 weeks of gestation and hence the low additional doses of Rh immune globulin that would be administered to potentially Rh(D)-negative, anti-Rh(D)-antibody-positive women.
There are a number of limitations to this study. In recommending the elimination of the 28-week antibody test, we assume that administration of Rh immune globulin has negligible side effects and that there is minimal risk to the fetus from a seroconversion before 28 weeks of gestation during the index pregnancy.9–12 Adverse effects of Rh immune globulin administration to Rh(D)-negative women are rare and usually mild.18–20 Rh immune globulin, however, may be a limited resource. Our study advocates for additional Rh immune globulin administration, albeit only 99–187 injections per 100,000 Rh(D)-negative women, which translates to between 400–900 additional administrations in the United States per year. Innovations such as fetal Rh(D) status determination through maternal blood sampling before 28 weeks of gestation will hopefully continue to lower unnecessary Rh immune globulin administration.21
Additionally, although the 0.099% of women who seroconvert before 28 weeks of gestation have minimal risks to the neonate during the index pregnancy, we must consider the potential psychological effects and subsequent fertility decisions by Rh(D)-negative women who do seroconvert during that time period. This seroconversion may well be discovered at postpartum testing in the index pregnancy, but it is possible it could be missed, especially with Rh immune globulin antibodies present. For example, would a woman who knew she had developed anti-D antibodies alter her future pregnancy decisions in light of her altered Rh status at 28 weeks? Whereas this question is beyond the scope of this study, future investigations could further clarify this interesting point. Finally, all modeling studies are limited by their simplification of a complicated clinical picture and individual circumstances that can never be fully incorporated into a cost–benefit model.
Only one previous study could be identified that addressed the utility of Rh(D) antibody screening. This Swedish study also suggested that antibody screening in the first trimester is sufficient.22 Other studies investigating Rh(D)-positive women developing non-D antibodies that also result in hemolytic disease of the newborn have concluded that the first-trimester antibody screen is sufficient.23–25 Only one study could be found suggesting the need for repeat anti-D antibody screening at week 28.24 However, the protocol in the country of analysis for this study, Croatia, does not require prophylactic use of Rh immune globulin at 28 weeks of gestation for Rh(D)-negative women.26
The cost savings demonstrated by this study will likely continue to increase if medical costs in the United States continue rapidly expanding at their current rate.27 According to the Institute of Medicine, the United States spends more than $210 billion per year on unnecessary tests, including many blood tests such as the indirect antiglobulin test.28 Our analysis indicates that through a simple change in the national protocol for antibody testing in Rh(D)-negative women, a significant number of blood tests may be eliminated, thus saving millions of dollars per year in unnecessary medical services.
1. Bowman JM. The prevention of Rh immunization. Transfus Med Rev 1988;2:129–50.
2. Walker RH, editor. Technical manual. 10th ed. Arlington (VA): American Association of Blood Banks; 1990.
4. U.S. Preventive Services Task Force. Recommendation statement: screening for Rh(D) incompatibility. Rockville (MD): U.S. Preventive Services Task Force; 2004.
5. Franklin IM. Prevention of rhesus haemolytic disease of the fetus and newborn. Lancet 2009;373:1082.
6. Snyder EL. Prevention of hemolytic disease of the newborn due to anti-D. Prenatal/perinatal testing and Rh immune globulin administration. Am Assoc Blood Banks Bull 1998;2:1–6.
7. American College of Obstetricians and Gynecologists. ACOG Practice Bulletin No 4. Prevention of Rho(D) alloimmunization. Washington (DC): ACOG; 1999.
8. Bowman JM, Chown B, Lewis M, Pollock JM. Rh isoimmunization during pregnancy: antenatal prophylaxis. Can Med Assoc J 1978;118:623–7.
9. Parham P. The immune system. New York (NY): Garland Pub; 2000.
10. Dean L. Hemolytic disease of the newborn. In: Dean L, editor. Blood groups and red cell antigens. Bethesda (MD): National Center for Biotechnology Information; 2005.
11. Walker W, Murray S. Haemolytic disease of the newborn as a family problem. Br Med J 1956;1:187–93.
12. Pilgrim H, Lloyd-Jones M, Rees A. Routine antenatal anti-D prophylaxis for RhD-negative women: a systematic review and economic evaluation. Health Technol Assess 2009;13:iii, ix–xi, 1–103.
13. Hughes RG, Craig JI, Murphy WG, Greer IA. Causes and clinical consequences of rhesus (d) haemolytic disease of the newborn: a study of a Scottish population, 1985–1990. Br J Obstet Gynaecol 1994;101:297–300.
14. U.S. National Center for Health Statistics. National Vital Statistics Reports (NVSR), Births: Final Data For 2008. Vol. 59. Atlanta (GA): National Center for Health Statistics; 2010.
15. American College of Obstetricians and Gynecologists. Management of alloimmunization during pregnancy. ACOG Practice Bulletin No. 75. Obstet Gynecol 2006;108:457–64.
16. Fung Kee Fung K, Eason E, Crane J, Armson A, De La Ronde S, Farine D, et al.; Maternal-Fetal Medicine Committee, Genetics Committee. Prevention of Rh alloimmunization. J Obstet Gynaecol Can 2003;25:765–73.
17. Drug topics red book. Montvale (NJ): Thomson Healthcare; 2010. p. 711.
18. Hong F, Ruiz R, Price H, Griffiths A, Malinoski F, Woloski M. Safety profile of WinRho anti-D. Semin Hematol 1998;35(suppl 1):9–13.
19. Neal JL. RhD isoimmunisation and current management modalities. J Obstet Gynecol Neonatal Nurs 2001;30:589–606.
20. Hirose TG, Mays DA. The safety of RhIG in the prevention of haemolytic disease of the newborn. J Obstet Gynaecol 2007;27:545–57.
21. Wikman AT, Tiblad E, Karlsson A, Olsson ML, Westgren M, Reilly M. Noninvasive single-exon fetal RHD determination in a routine screening program in early pregnancy. Obstet Gynecol 2012;120:227–34.
22. Filbey D, Hanson U, Wesström G. The prevalence of red cell antibodies in pregnancy correlated to the outcome of the newborn: a 12 year study in central Sweden. Acta Obstet Gynecol Scand 1995;74:687–92.
23. Adenij AA, Fuller I, Dale T, Lindow SW. Should we continue screening rhesus D positive women for the development of atypical antibodies in late pregnancy? J Matern Fetal Neonatal Med 2007;20:59–61.
24. Rothenberg JM, Weirermiller B, Dirig K, Hurd WW, Schilder J, Golichowski A. Is a third trimester antibody screen in Rh+ women necessary? Am J Manag Care 1999;5:1145–50.
25. Lurie S, Eliezer E, Piper I, Woliovitch I. Is antibody screening in Rh (D)-positive women necesary? J Matern Fetal Neonatal Med 2003;14:404–6.
26. Dajak S, Stefanović V, Capkun V. Severe hemolytic disease of fetus and newborn caused by red blood cell antibodies undetected at first-trimester screening. Immunohematology 2011;51:1380–8.
27. Cutler DM, Rosen AB, Vijan S. The value of medical spending in the United States, 1960-2000. N Engl J Med 2006;355:920–7.
28. Young PL, Olsen LA. Roundtable on Evidence-Based Medicine, Institute of Medicine. “Front Matter.” The Healthcare Imperative: Lowering Costs and Improving Outcomes: Workshop Series Summary. Washington (DC): The National Academies Press; 2010.