Fatal pulmonary hypertension associated with pertussis in infants: does extracorporeal membrane oxygenation have a role?

OBJECTIVE: The deaths of 4 infants who had confirmed pertussis infection at our hospital during the past year motivated us to review their cases as well as the use of extracorporeal membrane oxygenation (ECMO) in respiratory failure caused by pertussis.

METHODS: Retrospective chart reviews of the 4 infants who had pertussis and died at Vanderbilt University Medical Center (VUMC) from May 2001 to May 2002 were conducted. The Extracorporeal Life Support Organization (ELSO) database is an international voluntary registry established in 1986 to compile passive reports of adult and pediatric patients placed on ECMO. We searched this database for pertussis cases and analyzed the clinical outcomes by age, ventilatory management, and measurements of cardiopulmonary status.

RESULTS: All 4 infants who died from pertussis infection at VUMC were younger than 3 months and had severe pulmonary hypertension, and ECMO therapy was considered for respiratory failure. Review of the international ELSO database, focusing only on children from 1986 to July 2002, revealed a total of 23,970 patients placed on ECMO. Since the first pertussis case treated with ECMO in 1990, a total of 61 children with pertussis have been treated with ECMO, representing an increase from 0.09% to 1% of the total cases. Mean age of the pertussis patients placed on ECMO was 88 days (1 day-2.7 years). The overall mortality was 70.5% (43 of 61) but was significantly higher for infants who were younger than 6 weeks (84%) compared with infants who were older than 6 weeks (61%). When evaluating pre-ECMO management, survivors had received significantly higher mean positive end expiratory pressures than nonsurvivors (11.1 +/- 4.5 vs 7.3 +/- 3.1 cmH(2)O) and had significantly higher serum pH than nonsurvivors (7.31 +/- 0.14 vs 7.14 +/- 0.19). There was no effect of duration of ECMO, positive inspiratory pressure, mean arterial pressure, ECMO mode (venoarterial vs venovenous mode), or sex on mortality.

CONCLUSIONS: Although the ECMO registry is not an active surveillance system, it suggests that the numbers of children who require ECMO for pertussis have significantly increased. The high fatality rates of pertussis patients who have placed on ECMO are alarming and should stimulate reevaluation of whether this high-risk intervention is beneficial to patients with pertussis. This review, coupled with the 4 infants who died of pertussis at VUMC, suggests that improved measures are needed to prevent pertussis in infants.