PEDIATRICS Vol. 106 No. 6 December 2000, pp. 1334-1338

The Changing Demographics of Neonatal Extracorporeal Membrane Oxygenation Patients Reported to the Extracorporeal Life Support Organization (ELSO) Registry

Beverly Jane Roy, MD^*, Peter Rycus, MPH, Steven A. Conrad, MD^§, Reese H. Clark, MD, and for the Extracorporeal Life Support Organization (ELSO) Registry

From ^* Emory University, Department of Pediatrics, Atlanta, Georgia;  University of Michigan, Ann Arbor, Michigan; ^§ Louisiana State University, Shreveport, Louisiana; and  Pediatrix Medical Group, Inc, Sunrise, Florida.

	ABSTRACT

Top Abstract Methods Results Discussion References

Background.  Extracorporeal membrane oxygenation (ECMO) is an important treatment tool in the management of near-term and term neonates with severe hypoxemic respiratory failure. To better understand how health care for patients treated with ECMO has changed, we studied the demographic and treatment data reported to the Extracorporeal Life Support Organization (ELSO) registry from January 1, 1988, through January 1, 1998.

Methods.  We used data stored in the ELSO registry and evaluated the changes in demographics, use of alternate therapies before ECMO, severity of illness, duration of ECMO therapy, and mortality over a 10-year period. All data on neonates reported between January 1, 1988, and January 1, 1998 were used. Verification checks were performed on all fields to eliminate nonsense outliers. We separated the neonates into 2 groupsthose with and those without a congenital diaphragmatic hernia (CDH). All analyses were performed on the total group and each subgroup separately. Changes in continuous data were analyzed by year using analysis of variance. Year differences in categorical data were evaluated with ² analysis. We also used the linear trend test and the Cochran-Armitage trend test to evaluate time-related changes.

Results.  We reviewed 12 175 neonates. Over the decade, there were no changes in mean gestational age, gender, age at which ECMO was started, pH, or PaCO₂ just before ECMO. The proportion of neonates with CDH increased from 18% to 26%, while the proportion with respiratory distress syndrome decreased from 15% to 4%. Other diagnostic categories remained constant. The use of surfactant, high-frequency ventilation, and inhaled nitric oxide increased from 0% in 1988 to 36%, 46%, and 24%, respectively, in 1997. The mean peak pressure being used just before ECMO decreased (47 ± 10 in 1988 to 39 ± 12 in 1997), and the mean PaO₂/FIO₂ ratio increased (38 ± 23 in 1988 to 48 ± 36 in 1997). The primary mode of ECMO remains venoarterial; however, the use of venovenous ECMO increased from 1% to 32% over the decade. Duration of ECMO treatment increased overall, and this trend was seen for patients with and without CDH (124 ± 67 to 141 ± 104 hours for the non-CDH group, 161 ± 99 to 238 ± 141 hours for the CDH group). The number of centers reporting neonatal data to the ELSO registry increased from 52 in 1988 to a peak of 100 in 1993. In 1997, 96 centers reported data to ELSO. The average number of neonatal patients reported from each site decreased from a peak of 18 in 1991 to 9 in 1997. Mortality increased from 18% to 22%; however, when corrected for the relative increase in neonates with CDH, this trend disappeared. Diagnoses-specific mortality rates remained constant. The occurrence of intracranial hemorrhage and/or infarct also stayed constant at 16%.

Conclusions.  The population of neonates treated with ECMO in 1997 was very different from patients treated in the 1980s and early 1990s. They were exposed to an ever-expanding group of new therapies, appeared to be healthier based on indices of gas exchange, and were cared for at centers that reported fewer cases per year.  Key words:  ECMO, inhaled nitric oxide, neonate, high-frequency ventilation.

Extracorporeal membrane oxygenation (ECMO) has been used in the management of neonates with life-threatening cardiorespiratory diseases since 1982.¹ Several studies have demonstrated that ECMO promotes normal survival in neonates with respiratory failure.^2-5 Over the past several years, other therapies such as surfactant replacement, high-frequency ventilation (HFV), and inhaled nitric oxide (iNO) have been introduced and used increasingly in the management of critically ill infants.^6,7 These therapies impact the type of patients treated with ECMO and potentially change the outcome of patents treated with ECMO.^6,7 We studied how the demographics and health care practices have changed for neonates reported to the Extracorporeal Life Support Organization (ELSO) registry from January 1, 1988, through January 1, 1998 to better understand the changing health care practices offered to neonates with severe hypoxemia.

	METHODS

Top Abstract Methods Results Discussion References

We reviewed data collected into the ELSO registry between January 1, 1988, and January 1, 1998. With these data, we evaluated the changes in demographics, use of alternative therapies before ECMO, severity of illness, duration of ECMO, and mortality. Verification checks were performed on all fields to eliminate nonsense outliers. The neonates were separated into 2 groupsthose with congenital diaphragmatic hernia (CDH) and those without CDH. We report data on CDH separately because these infants have a major anomaly that dramatically affects their outcome. Collected data included total number of neonates placed on ECMO, mean gestational age, gender, age at which ECMO was started, pH or PaCO₂ just before ECMO, average number of patients reported per center, primary diagnoses, mortality, use of alternative therapies before ECMO (HFV, surfactant, iNO), use of venovenous (VV) ECMO, length of time on ECMO, mean PaO₂/FIO₂ ratio just before ECMO, mean peak inspiratory pressure just before ECMO, incidence of intracranial hemorrhage or infarct, Apgar scores, race, and mean birth weight.

Statistical analyses were performed on the total group and each subgroup separately. Analysis of variance was used to analyze changes in continuous data by year. Differences in categorical data by year were analyzed with ². Time-related changes were evaluated using the linear trend test and the Cochran-Armitage trend test.

	RESULTS

Top Abstract Methods Results Discussion References

Data on 12 175 neonates was analyzed. Over the 10-year study period, there were no changes in mean gestational age (38.1-38.5 weeks), gender (55%-60%), age at which ECMO was started (48-51 hours), pH (7.38-7.42), or PaCO₂ (38-40) just before ECMO. The number of neonatal patients reported per year to the ELSO registry increased until 1992 to a peak number of 1499 and then has gradually decreased (Fig 1A). The average number of neonates reported per ECMO center peaked in 1991 and then declined to a low of 9 in 1997 (Fig 1B). The number of centers reporting 10 or more patients also showed a decline over the decade (Fig 1B).

View larger version (53K): [in this window] [in a new window]   Fig. 1.   A, The change in the number of neonatal patients reported and the number of centers reporting data to the ELSO registry between 1988 and 1997. B, The change in the average number of neonatal patients per ECMO center and the percent of centers reporting >10 neonatal cases per year.

The diagnostic mix of patients treated with ECMO has also significantly changed. The percentage of patients with respiratory distress syndrome was 15% in 1988 and declined to 4% in 1997 (Fig 2). The percentage of patients with CDH was 18% in 1988 and increased to 26% in 1997. This change reflects a decrease in the number of patients with respiratory distress syndrome, not an absolute increase in the number of patients with CDH. The number of neonates with CDH has remained relatively constant at 250 to 280 per year. The percentage of patients with primary diagnoses of meconium aspiration syndrome or persistent pulmonary hypertension of the newborn remained stable over the study period. Meconium aspiration syndrome remained the most frequently reported diagnosis (Fig 2).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   The change in the relative percents of the patients with a specific admitting diagnosis between 1988 and 1997.

The use of alternative rescue therapies (HFV, surfactant, and iNO) before ECMO dramatically increased over the study period (Fig 3A). HFV and surfactant were reported in 0% of patients in 1988 and had increased to 46% and 36%, respectively, in 1997. iNO was reported in 0% of patients from 1988 to 1991 and showed a yearly increase to 24% in 1997. There was no significant relationship between age at initiation of ECMO and surfactant, HFV, or iNO use.

View larger version (40K): [in this window] [in a new window]   Fig. 3.   A, The increase in the reported use of HFV, surfactant, and iNO in neonates reported to the ELSO registry between 1988 and 1997. B, The relative increased use of VV ECMO in support of neonates treated with ECMO for patients with CDH and those without CDH.

More centers reported using VV ECMO by the end of the study period in both CDH and non-CDH patients (Fig 3B). The number of hours on ECMO for each patient also increased over the study period for both CDH (161 ± 99 to 238 ± 141) and non-CDH (124 ± 67 to 141 ± 104) patients. Duration of ECMO treatment was most influenced by diagnosis. When diagnosis was included in regression models, the type of ECMO did not independently influence length of ECMO treatment. In addition, pre-ECMO use of HFV, iNO, and surfactant did not influence length of ECMO treatment.

The PaO₂/FIO₂ ratio recorded just before initiation of ECMO was higher at the end of the study period for non-CDH patients and remained stable in patients with CDH (Fig 4A). The mean peak inspiratory pressure just before initiation of ECMO decreased over time in both CDH and non-CDH patients (Fig 4B).

View larger version (47K): [in this window] [in a new window]   Fig. 4.   The increase in the average PaO2/FIO2 ratio (A) and the decrease in the average peak inspiratory pressure reported (B) just before start of ECMO for patients with and without CDH.

The percentage of patients who developed intracranial hemorrhage did not change over time, and pretreatment with HFV, surfactant, or iNO was not associated with an increased risk of intracranial hemorrhage.

Mortality did increase with time from 18% to 22%. Mortality was highest in neonates with CDH. When the mortality rate was corrected for the relative increase in the number of patients with CDH, the trend for the overall increase in mortality disappeared. Diagnoses-specific mortality rates for patients with CDH, meconium aspiration syndrome, respiratory distress syndrome, and sepsis have not changed over the last decade.

	DISCUSSION

Top Abstract Methods Results Discussion References

The results of this study indicate that the types of neonatal patients placed on ECMO in 1997 were different from those placed on ECMO from 1988 to 1993. The percentage of patients with CDH was higher, and this likely explains the increase in overall mortality for neonates treated with ECMO. Diagnoses-specific mortality rates for patients with CDH, meconium aspiration syndrome, respiratory distress syndrome, and sepsis have not changed over the last decade. The average number of patients treated per year at each ECMO center decreased, and the overall number of patients was lower in 1993 to 1997 than in 1988. The infants also appeared to be healthier at the start of ECMO based on the indices of gas exchange just before the initiation of ECMO.

Experience reported by a large ECMO program supports our observations. Wilson et al⁷ reported data on 455 patients treated at Boston Children's Hospital from 1984 to 1996. In their population, HFV and iNO were introduced as treatment options in 1991 and 1992, respectively. Over the study period, the annual number of CDH patients was stable, but the number of other types of neonatal diagnoses declined, confirming our observations. The relative percentage of pediatric and cardiac cases and infants with CDH increased between 1984 and 1997 in both the Wilson report and in our data set. This correlated with an increase in mortality. An increase in the mean duration of ECMO per patient and a significant increase in the incidences of patient and circuit complications were also reported.

Wilson et al⁷ speculated that the decrease in the number of neonatal ECMO cases was most likely secondary to improved in-house management of infants with severe respiratory distress, and that the decrease parallels the time when surfactant, HFV, and iNO were being increasingly used. They also speculated that the increased duration of ECMO might, in part, be attributable to the relative increase in patients who traditionally require a longer period on ECMO (CDH and pediatric patients). However, they also noted a trend toward increased time on ECMO in their meconium aspiration syndrome, persistent pulmonary hypertension of the newborn, and sepsis groups. They feel that their data suggest that the neonatal patients being placed on ECMO were sicker in the last 3 years as compared with their earlier ECMO patients.

In contrast, our results from the ELSO registry review suggest that neonatal ECMO patients are healthier based on a lower mean peak inspiratory pressure and an increase in the mean PaO₂/FIO₂ ratio just before ECMO. However, it is also clear that ECMO patients are being treated with more therapies such as HFV, surfactant, and iNO. Thus, it may be that the patients are not healthier but instead are treated more intensively. The fact that the average age at which ECMO was started has not increased significantly suggests that, despite being offered several rescue therapies, the age at which patients are identified as needing ECMO has not changed.

Like Wilson et al, we also found an increased duration of ECMO in both non-CDH and CDH patients. It is possible that patients who have received alternative therapies appear healthier based on physiological measures, but that these may not be reliable measures of severity of lung disease in these patients. This conjecture demands that we reevaluate ECMO criteria in the health care era where HFV, surfactant, and iNO are commonly used to avert the use of ECMO. Perhaps even more importantly, we need data on the patients who have hypoxemic respiratory failure and are successfully treated with alternative rescue treatments to determine if the outcomes of these patients suggest that they are avoiding ECMO safely.

	FOOTNOTES

Received for publication Aug 3, 1999; accepted Feb 28, 2000.

Reprint requests to (R.H.C.) Director of Research, Pediatrix Medical Group, Inc, 1301 Concord Terrace, Sunrise, FL 33323-2825. E-mail: reese_clark{at}pediatrix.com

	ABBREVIATIONS

ECMO, extracorporeal membrane oxygenation; HFV, high-frequency ventilation; iNO, inhaled nitric oxide; ELSO, Extracorporeal Life Support Organization; CDH, congenital diaphragmatic hernia; VV, venovenous.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Bartlett RH Extracorporeal life support for cardiopulmonary failure. Curr Probl Surg. 1990; 27:621-705 [Medline]
2. The UK Collaborative ECMO Trial Group UK collaborative randomised trial of neonatal extracorporeal membrane oxygenation. Lancet. 1996; 348:75-82 [CrossRef][Medline]
3. The UK Collaborative ECMO Group. The collaborative UK ECMO (Extracorporeal Membrane Oxygenation) Trial: follow-up to 1 year of age. Pediatrics. 1998;101(4). URL: http://www.pediatrics.org/cgi/content/full/101/4/e1
4. Cornish JD, Heiss KF, Clark RH, Strieper MJ, Boecler B, Kesser K Efficacy of venovenous extracorporeal membrane oxygenation for neonates with respiratory and circulatory compromise. J Pediatr. 1993; 122:105-109 [Medline]
5. O'Rourke PP, Crone RK, Vacanti JP, Extracorporeal membrane oxygenation and conventional medical therapy in neonates with persistent pulmonary hypertension of the newborn: a prospective randomized study. Pediatrics. 1989; 84:957-963 [Abstract/Free Full Text]
6. Kanto WP Jr A decade of experience with neonatal extracorporeal membrane oxygenation. J Pediatr. 1994; 124:335-347 [CrossRef][Medline]
7. Wilson JM, Bower LK, Thompson JE, Fauza DO, Fackler JC ECMO in evolution: the impact of changing patient demographics and alternative therapies on ECMO. J Pediatr Surg. 1996; 31:1116-1122 [CrossRef][Medline]