What is the most relevant standard of success in assisted reproduction?

Challenges in measuring and reporting success rates for assisted reproductive technology treatments: What is optimal?

Laura A. Schieve1 and Meredith A. Reynolds

Division of Reproductive Health, Centers for Disease Control and Prevention, Atlanta, GA, USA

1 To whom correspondence should be addressed at: Division of Reproductive Health, Centers for Disease Control and Prevention, Mailstop K-34, 4770 Buford Highway NE, Atlanta, GA 30341–3717, USA. e-mail: lschieve{at}cdc.gov


    Abstract
 Top
 Abstract
 Introduction
 US ART registry—rationale...
 Adverse outcomes beyond multiple...
 Conclusion
 References
 
For assisted reproductive technology (ART) treatments, measures of success that move beyond traditional measures of pregnancy and live birth and narrow the numerator to infant outcomes with an optimal short- and long-term prognosis are needed. Hence, presentation of singleton live birth delivery rates is warranted. Twins have greatly increased risks for morbidity and mortality in comparison with singletons. Success rates based on singleton live births will more completely inform patients evaluating which ART treatment options will maximize their chance for a healthy infant. Additionally, providers who limit embryos transferred can feel they are on an even playing field in reporting their success rates. Measures of success that narrow the numerator further to exclude preterm or low birth weight singleton births might also be informative. However, the utility of such measures is less clear because the aetiologies of preterm birth and low birth weight among singletons are probably multifactorial. While it may be desirable to consider adverse outcomes such as congenital anomalies in defining treatment success, it is unfeasible to collect complete and accurate data on anomalies in current ART registries. As ART use increases, continual re-examination and critique of the manner in which success is defined and presented to the public is critical.

Key words: assisted reproduction/live birth/singleton birth/success rates


    Introduction
 Top
 Abstract
 Introduction
 US ART registry—rationale...
 Adverse outcomes beyond multiple...
 Conclusion
 References
 
Min et al. (2004Go) propose that the success of an assisted reproductive technology (ART) cycle is most appropriately measured as ‘birth emphasizing a successful singleton at term (BESST)’. This measure, which narrows the numerator of ART success rates from the traditional measures of pregnancy and live birth delivery, highlights infant outcomes with an optimal short- and long-term prognosis. We agree with the growing chorus of professionals, including Min et al., calling for presentation of singleton live birth rates as a distinct indicator of ART success (Cohen and Jones, 2001Go; ESHRE, 2001Go; Evers, 2002Go; Hogue, 2002Go; World Health Organization, 2002Go). We also agree that the BESST measure, which narrows the numerator even further to exclude not only multiple births, but preterm singleton births as well, might be informative to ART patients, providers, policymakers and scientists. However, the BESST measure has certain limitations and thus should not be presented as the sole measure of ART success. In particular, the BESST measure is more difficult to present simply because the aetiology of preterm birth among singletons is largely unknown and probably multifactorial (Lumley, 2003Go; Moutquin, 2003Go). A measure such as BESST that incorporates gestational age at birth must be presented in the context of patient and treatment characteristics associated with increased risk for preterm delivery.


    US ART registry—rationale for presenting singleton live birth rates
 Top
 Abstract
 Introduction
 US ART registry—rationale...
 Adverse outcomes beyond multiple...
 Conclusion
 References
 
Public health importance
The Centers for Disease Control and Prevention (CDC) is charged with implementing the Fertility Clinic Success Rate and Certification Act (FCSRCA) of 1992 which mandates all medical centres performing ART in the USA or its territories to report data related to ART success rates to the CDC annually (US Congress, 1992Go). CDC published the first report under the law in 1997 (Schieve et al., 2002aGo). That report documented national and clinic level success rates for ART procedures initiated in 1995. Since that time, we have published annual reports for ART procedures initiated in each year from 1996 to 2001. In all reports, two measures of success were presented: (i) pregnancy and (ii) birth of one or more living infants (live birth delivery). Data on risks for multiple gestation pregnancies and multiple births were also presented, both nationally and per clinic, such that prospective ART patients and providers were presented with a balanced view of benefits and risks. However, in all reports up to and including the 2000 ART report, the actual percentages of procedures that resulted in singleton live birth were not directly presented but could be estimated from the data presented for total live births and the percentage that were multiples.

The most recent report, covering ART procedures initiated in 2001, introduces singleton live birth rates (Table I) as an additional measure of success (Centers for Disease Control and Prevention, American Society for Reproductive Medicine, and Society for Assisted Reproductive Technology, 2003Go). We believe presentation of this specific rate will draw attention more directly to the need to consider optimal infant outcomes. It is well established that multiple birth infants have greatly increased risks for adverse health outcomes such as low birth weight, preterm delivery, infant death, and disability among survivors (Kiely, 1998Go; Martin and Park, 1999Go; Martin et al., 2002Go; Pharoah, 2002Go). Additionally, the relationship between ART and multiple birth among US infants has been well documented (Reynolds et al., 2003Go). Despite concerted efforts by the medical community and infertility advocacy groups to focus attention on the risk for multiple birth and related sequelae via practice guidelines (American Society for Reproductive Medicine, 1999Go) and education campaigns, the proportion of live birth deliveries that are multiple births among ART patients in the US has changed little since 1996 (Table II). While triplet and higher order multiples have decreased somewhat, the overall multiple birth rate has been stable and the proportion of twins has been stable or actually increased for some ART types.


View this table:
[in this window]
[in a new window]
 
Table I. Live birth rates and singleton live birth rates among ART cycles that progressed to embryo transfer, by the source of the oocyte and the status of the embryo; USA, 2001
 

View this table:
[in this window]
[in a new window]
 
Table II. Percentage of twin, triplet or higher, and total multiple-birth deliveries among live birth deliveries resulting from ART by the source of the oocyte and the status of the embryo; USA, 1996–2001
 
Twins, albeit to a lesser extent than triplet or more infants, are still a high-risk group when it comes to infant morbidity and mortality. The risks for low birth weight and preterm birth both exceed 50% for twins, and the risk for very low birth weight is 10% (Martin and Park, 1999Go; Martin et al., 2002Go). These risks are 6–9 times greater than the comparable risks for singletons. Moreover, twins are at greatly increased risk for perinatal and infant mortality (Kiely, 1998Go; Martin and Park, 1999Go; ESHRE Capri Workshop Group, 2000Go) and neurological impairments, including cerebral palsy (ESHRE Capri Workshop Group, 2000Go; Pharoah, 2002Go). Carrying and delivering a twin gestation is also riskier than a singleton gestation for women. Higher risks of maternal mortality, haemorrhage, pregnancy-induced hypertension and anaemia are all associated with twin pregnancies (Senat et al., 1998Go; ESHRE Capri Workshop Group, 2000Go). Clearly greater progress must be made in reducing the total multiple gestation rate associated with infertility treatments.

In counties such as the USA where insurance companies often do not cover infertility treatments, particularly ART, many patients may feel pressure to maximize the opportunity for live birth delivery by having multiple embryos transferred. Additionally, anecdotal reports suggest that if success is defined solely as live birth deliveries, some ART providers also may feel pressure to transfer multiple embryos in an effort to maximize their publicly reported success rates. This viewpoint has been elaborated in a recent debate in which several leaders in the field, including several past presidents of the Society for Assisted Reproductive Technology (SART), question the value of clinic-specific reporting at all. This viewpoint stems from the concern that reporting success in terms of total pregnancy and live birth rates is resulting in increased competition among ART clinics, with multiple gestation an unfortunate consequence (Grifo et al., 2001Go; Jones and Schnorr, 2001Go). Indeed, in the USA, high-order embryo transfer is still common practice. In 2001, only 6% of ART cycles that used fresh, non-donor oocytes which progressed to the embryo transfer stage involved single embryo transfer. Approximately 66% of ART cycles involved the transfer of three or more embryos, 32% of cycles involved the transfer of four or more embryos, and 11% of cycles involved the transfer of five or more embryos (Figure 1).



View larger version (58K):
[in this window]
[in a new window]
 
Figure 1. Number of embryos transferred during ART cycles using freshly fertilized embryos derived from the patients’ oocytes; USA, 2001.

 
While we do not agree with eliminating clinic-specific reporting, we do agree that reporting must be responsive to such concerns. Success rates based on singleton live birth deliveries will provide patients with a direct measure that will better enable them to make informed decisions about whether and how to undergo ART treatment in a way that maximizes their chance for a healthy infant. Additionally, providers who limit the number of embryos transferred can feel they are on an even playing field in reporting their success rate data. Some may argue that even among births from ART cycles with single embryo transfer, the proportion of singletons may not be 100% as ART treatments also appear to pose a risk for monozygotic multiple birth (Blickstein et al., 1999Go; Schieve et al., 2000aGo; Derom et al., 2001Go). However, in the current US medical climate in which three or more embryos transferred is the norm, the overwhelming risk for multiple birth in ART stems from cycles with multiple embryos transferred (Schieve et al., 1999Go; Reynolds et al., 2001Go; Centers for Disease Control and Prevention, American Society for Reproductive Medicine, and Society for Assisted Reproductive Technology, 2003Go; Vahratian et al., 2003Go; Wright et al., 2003Go). Indeed the difference in risk for multiple birth between two embryos transferred and one embryo transferred is very large. Among ART cycles that used freshly fertilized embryos derived from non-donor oocytes in 2001, the percentage of live birth deliveries that were multiple births was <1% overall when a single embryo was transferred and nearly 34% when two embryos were transferred (Centers for Disease Control and Prevention, American Society for Reproductive Medicine, and Society for Assisted Reproductive Technology, 2003Go). Thus, most multiple births resulting from ART today are probably di-, tri- or higher zygotic and are directly preventable. Also, as continued research provides data on the specific facets of ART that predispose an embryo to monozygotic twinning, these multiple births should also become increasingly preventable.

Feasibility of singleton success rate measure
The feasibility of collecting complete and accurate data to report singleton live birth rates is comparable with the feasibility of data collection for total live birth delivery rates. Pregnancy outcomes are reported for >99% of the pregnancies reported to the US ART Registry, and plurality at birth is reported for 100% of the live births reported. Additionally, results from on-site data validation chart reviews in a random sample of clinics reporting data to the registry reveal very low error rates (<1%) for plurality (Centers for Disease Control and Prevention, American Society for Reproductive Medicine, and Society for Assisted Reproductive Technology, 2003Go).

Singleton live birth rate can also be presented to the general public in a manner that plainly delineates the utility of this measure. There is clear biological plausibility for a direct treatment effect between the transfer of multiple embryos and multiple birth risk. Moreover, the association between ART and multiple birth is strong. In the USA, the proportion of ART infants born in a multiple birth (53%) is 18 times the proportion of multiple births in the general population (3%) (Wright et al., 2003Go). Additionally, as stated earlier, the increased risk for adverse sequalae among multiple births is dramatic, even when considering only twins versus singletons. Thus, the implications of multiple birth and the rationale for presenting singleton live births as opposed to (or at least in addition to) total live births in ART success rates reports can be easily conveyed to both lay and professional audiences.


    Adverse outcomes beyond multiple birth
 Top
 Abstract
 Introduction
 US ART registry—rationale...
 Adverse outcomes beyond multiple...
 Conclusion
 References
 
We also agree with Min et al. (2004Go) that even among singletons, adverse health outcomes, such as preterm delivery, must be considered. However, we do not believe preterm delivery risk is the only important factor to consider and believe it is not feasible to roll this or other birth outcomes into a single success rate measure that can be clearly understood and interpreted by the public. Numerous studies suggest that singleton infants born after ART treatment are at increased risk for low birth weight, very low birth weight, preterm delivery and fetal growth restriction in comparison with naturally conceived infants (Olivennes et al., 1993Go; Gissler et al., 1995Go; Verlaenen et al., 1995Go; Bergh et al., 1999Go; Dhont et al., 1999Go; Westergaard et al., 1999Go; Perri et al., 2001Go; Klemetti et al., 2002Go; Schieve et al., 2002bGo; Wang et al., 2002Go). However, questions remain about whether these risks stem from the ART treatment or from the underlying infertility of the couples using these treatments. Moreover, these outcomes must be considered in the context of other maternal risk factors such as advanced maternal age, primiparity and non-White race (Martin et al., 2002Go).

We suggest that rather than incorporating preterm delivery risk into a success rate measure, it will be more informative to supplement ART success rate measures based on singleton live births with detailed analyses of the additional risks for adverse birth outcomes among ART singletons. Such analyses should include examination of risks within subgroups of ART patients based on demographic, pregnancy history, infertility diagnosis and underlying aetiological factors, such as reproductive tract infections, to the extent possible. A major challenge for such analyses is to obtain accurate and complete data on a population basis for classification of ART patients into meaningful subgroups. For example, diagnostic protocols can vary greatly from clinic to clinic, and therefore a reported diagnosis of, for example, male factor infertility or endometriosis will not necessarily represent the exact same grouping of patients at all ART clinics. Thus, all analyses must be presented in the appropriate context.

Recently, several studies have suggested the need to consider outcomes such as congenital anomalies, even among ART singletons (Ericson and Kallen, 2001Go; Hansen et al., 2002Go). Tracking and reporting these outcomes among ART singletons as well as among multiples would be desirable. However, the feasibility of obtaining accurate and unbiased data must be carefully considered. As the numerator moves along the continuum from pregnancy to healthy child, it becomes increasingly difficult to collect complete and accurate data at the population level. The paramount barrier is that in many countries, including the USA, medical care often is not co-ordinated between infertility, obstetric and paediatric providers. In the USA, the responsibility for reporting ART success rate data rests with the infertility clinic that provided the ART treatment. These clinics do not typically provide care for patients beyond the first trimester of pregnancy. Thus, clinic personnel must work to track patients once they are released from the clinic’s care to ascertain the outcome of each pregnancy. Birth outcome data are primarily from parental report.

While studies demonstrate that birth occurrence, date of birth and birth weight can be reported accurately (Lederman and Paxton, 1998Go), maternal report for even severe birth defects has been shown to have low sensitivity (Rasmussen et al., 1990Go). Thus, while we believe the current US system accurately captures data needed to assess outcomes such as multiple birth and low birth weight, it should not be used to evaluate birth defects. Population-based birth defects registries in the USA use a much more rigorous active surveillance methodology with multiple clinical sources of data (Lynberg and Edmonds, 1994Go). To provide useful information, any system to collect birth defect data among ART births must be similarly designed (Schieve et al., 2000bGo).


    Conclusion
 Top
 Abstract
 Introduction
 US ART registry—rationale...
 Adverse outcomes beyond multiple...
 Conclusion
 References
 
Continual re-examination and critique of the manner in which ART success is defined and presented to the public is a critical component of the public health surveillance process in this rapidly changing field. While we do not believe the new measure proposed by Min et al. (2004Go) is entirely feasible, we welcome the debate and the increased dialogue on the need to consider the health of children conceived using ART. Whether or not a single success rate measure can address this objective entirely or whether supplemental analyses of risks will more accurately and appropriately convey some of the information should continue to be discussed. This debate should consider the balance between presenting multiple measures of success and/or supplemental analyses of risks that provide the maximum information possible, and the added confusion that is often introduced when presenting many separate measures in reports intended for a non-scientific audience. The use of ART continues to increase in the USA and worldwide. Coupled with increasing ART pregnancy rates, this is leading to large increases in the number of children born as a result of ART. As scientists, health care providers and public health professionals, we have an obligation to ensure that the data we present on ART efficacy is consistent with our overarching objectives of promoting safe motherhood and the birth of healthy children.


    References
 Top
 Abstract
 Introduction
 US ART registry—rationale...
 Adverse outcomes beyond multiple...
 Conclusion
 References
 
American Society for Reproductive Medicine (1999) Guidelines on Number of Embryos Transferred. A Practice Committee Report: A Committee Opinion. ASRM, Birmingham, AL.

Bergh T, Ericson A, Hillensjo T, Nygren KG and Wennerholm UB (1999) Deliveries and children born after in-vitro fertilisation in Sweden 1982–95: a retrospective cohort study. Lancet 354,1579–1585.[CrossRef][ISI][Medline]

Blickstein I, Verhoeven HC and Keith LG (1999) Zygotic splitting after assisted reproduction. N Engl J Med 340,738–739.[Free Full Text]

Centers for Disease Control and Prevention, American Society for Reproductive Medicine, and Society for Assisted Reproductive Technology (2003) 2001 Assisted Reproductive Technology Success Rates. US Department of Health and Human Services, CDC: Atlanta, USA.

Cohen J and Jones HW (2001) How to avoid multiple pregnancies in assistive reproductive technologies. Semin Reprod Med 19,269–278.[CrossRef][ISI][Medline]

Derom R, Derom C and Vlietinck R (2001) The risk of monozygotic twinning. In: Blickstein I and Keith LG (eds), Iatrogenic Multiple Pregnancy: Clinical Implications. The Parthenon Publishing Group, New York.

Dhont M, De Sutter P, Ruyssink G, Martens G and Bekaert A (1999) Perinatal outcome of pregnancies after assisted reproduction: a case–control study. Am. J. Obstet Gynecol 181,688–695.[ISI][Medline]

Ericson A and Kallen B (2001) Congenital malformations in infants born after IVF: a population-based study. Hum Reprod 16,504–509.[Abstract/Free Full Text]

ESHRE(2001) Campus Course Report. Prevention of twin pregnancy after IVF/ICSI by single embryo transfer. Hum Reprod 16,790–800.[Abstract/Free Full Text]

ESHRE Capri Workshop Group(2000) Multiple gestation pregnancy. Hum Reprod 15,1856–1864.[Abstract/Free Full Text]

Evers JLH (2002) Female subfertility. Lancet 360,151–159.[CrossRef][ISI][Medline]

Gissler M, Silvero MM and Hemminki E (1995) In-vitro fertilization pregnancies and perinatal health in Finland 1991–1993. Hum Reprod 10,1856–1861.[Abstract]

Grifo J, Hoffman D and McNamee PI (2001) We are due for a correction... and we are working to achieve one. Fertil Steril 75,14.

Hansen M, Kurinczuk JJ, Bower C and Webb S (2002) The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization. N Engl J Med 346,725–730.[Abstract/Free Full Text]

Hogue CJR (2002) Successful assisted reproductive technology: the beauty of one. Obstet Gynecol 100,1017–1019.[Abstract/Free Full Text]

Jones HW and Schnorr JA (2001) Multiple pregnancies: a call for action. Fertil Steril 75,11–13.[ISI][Medline]

Kiely JL (1998) What is the population-based risk of preterm birth among twins and other multiples? Clin Obstet Gynecol 41,3–11.[CrossRef][ISI][Medline]

Klemetti R, Gissler M and Hemminki E (2002) Comparison of perinatal health of children born from IVF in Finland in the early and late 1990s. Hum Reprod 17,2192–2198.[Abstract/Free Full Text]

Lederman SA and Paxton A (1998) Maternal reporting of prepregnancy weight and birth outcome: consistency and completeness compared with the clinical record. Maternal Child Health J 2,123–126.[CrossRef]

Lumley J (2003) Defining the problem: the epidemiology of preterm birth. Br J Obstet Gynecol 110 (Suppl 20),3–7.

Lynberg MC and Edmonds LD (1994) State use of birth defects surveillance. In: Wilcox LS and Marks JS (eds), From Data to Action: CDC’s Public Health Surveillance for Women, Infants, and Children. Centers for Disease Control and Prevention, US Department of Health and Human Services, Atlanta, GA, pp. 217–226.

Martin JA and Park MM (1999) Trends in twin and triplet births: 1980–1997. Natl Vital Stat Rep 47,1–16.[Medline]

Martin JA, Hamilton BE and Ventura SJ (2002) Births: final data for 2001. Natl Vital Stat Rep 51,1–104.[Medline]

Min JK, Breheny SA, MacLachlan V and Healy DL (2004) The singleton, term gestation, live birth rate per cycle initiated: the BESST endpoint for assisted reproduction. Hum Reprod 19,3–7.[Abstract/Free Full Text]

Moutquin J (2003) Classification and heterogeneity of preterm birth. Br J Obstet Gynecol 110 (Suppl 20),30–33.

Olivennes F, Rufat P, Andre B, Poourade A, Quiros MC and Frydman R (1993) The increased risk of complication observed in singleton pregnancies resulting from in-vitro fertilization (IVF) does not seem to be related to the IVF method itself. Hum Reprod 8,1297–1300.[Abstract]

Perri T, Chen R, Yoeli R, Merlob P, Orvieto R, Shalev Y, Ben-Rafael Z and Bar-Hava I (2001) Are singleton assisted reproductive technology pregnancies at risk of prematurity? J Assist Reprod Genet 18,245–249.[CrossRef][ISI][Medline]

Pharoah PO (2002) Neurological outcome in twins. Semin Neonatol 7,223–230.[CrossRef][Medline]

Rasmussen SA, Mulinare J, Khoury MJ and Maloney EK (1990) Evaluation of birth defect histories obtained through maternal interviews. Am J Hum Genet 46,478–85.[ISI][Medline]

Reynolds MA, Schieve LA, Jeng G, Peterson HB and Wilcox LS (2001) Risk of multiple birth associated with in vitro fertilization using donor eggs. Am J Epidemiol 154,1043–1050.[Abstract/Free Full Text]

Reynolds MA, Schieve LA, Martin JA, Jeng G and Macaluso M. (2003) Trends in multiple births conceived using assisted reproductive technologies, United States, 1997–2000. Pediatrics 11,1159–1162.

Schieve LA, Peterson HB, Meikle SF, Jeng G, Danel I, Burnett NM and Wilcox LS (1999) Live-birth rates and multiple-birth risk using in vitro fertilization. J Am Med Assoc 282,1832–1838.[Abstract/Free Full Text]

Schieve LA, Meikle SF, Peterson HB, Jeng G, Burnett NM and Wilcox LS (2000a) Does assisted hatching pose a risk for monozygotic twinning in pregnancies conceived through in vitro fertilization? Fertil Steril 74,288–294.[CrossRef][ISI][Medline]

Schieve LA, Rasmussen SA, Correa A, Santelli J, Tatham L and Wilcox LS. (2000b) Commentary—Assisted reproductive technology in the United States: 1997 results generated from the American Society for Reproductive Medicine/Society for Assisted Reproductive Technology Registry. Fertil Steril 74,653–654.[CrossRef][ISI][Medline]

Schieve LA, Wilcox LS, Zeitz J, Jeng G, Hoffman D, Brzyski R., Toner J, Grainger D, Tatham L and Younger B (2002a) Assessment of outcomes for assisted reproductive technology: overview of issues and the US experience in establishing a surveillance system. In: Vayena E, Rowe PJ and Griffin PD (eds), Medical, Ethical and Social Aspects of Assisted Reproduction (2001: Geneva, Switzerland) Current Practices and Controversies in Assisted Reproduction: Report of a WHO Meeting. WHO Publications, Geneva, Switzerland, pp. 363–376.

Schieve LA, Meikle SF, Ferre C, Peterson HB, Jeng G and Wilcox LS (2002b) Low and very low birth weight in infants conceived with use of assisted reproductive technology. N Engl J Med 346,731–737.[Abstract/Free Full Text]

Senat MV, Ancel PY, Bouvier-Colle MH and Breart G (1998) How does multiple pregnancy affect maternal mortality and morbidity? Clin Obstet Gynecol 41,78–83.[Medline]

US Congress(1992) Fertility Clinic Success Rate and Certification Act. (Publication No. 102-493 October 24, 1992). Washington, DC.

Vahratian A, Schieve LA, Reynolds MA and Jeng G. (2003) Live birth rates and multiple-birth risk of assisted reproductive technology pregnancies conceived using thawed embryos USA 1999–2000. Hum Reprod 18,1442–1448.[Abstract/Free Full Text]

Verlaenen H, Cammu H, Derde MP and Amy JJ (1995) Singleton pregnancy after in vitro fertilization expectations and outcome. Obstet Gynecol 86,906–910.[Abstract/Free Full Text]

Wang JX, Norman RJ and Kristiansson PK (2002) The effect of various infertility treatments on the risk of preterm birth. Hum Reprod 17,945–949.[Abstract/Free Full Text]

Westergaard HB, Johansen AM, Erb K and Anderson AN (1999) Danish National In-Vitro Fertilization Registry 1994 and 1995: a controlled study of births, malformations and cytogenetic findings. Hum Reprod 14,1896–1902.[Abstract/Free Full Text]

World Health Organization(2002) WHO Recommendations. In: Vayena E, Rowe PJ and Griffin PD (eds), Medical, Ethical and Social Aspects of Assisted Reproduction (2001: Geneva, Switzerland) Current Practices and Controversies in Assisted Reproduction: Report of a WHO Meeting. WHO Publications, Geneva, Switzerland, pp. 381–396.

Wright VC, Schieve LA, Reynolds MA and Jeng G (2003) Assisted reproductive technology surveillance—United States, 2000. In Surveillance Summaries, August 29, 2003 MMWR 52(SS-9), 1–16.