Criteria abstracted from The Users' Guides to the Medical Literature series in JAMA
Impact of Rapid Leukodepletion on the Outcome of Severe Clinical Pertussis in Young Infants
Rowlands HE, Goldman AP, Harrington K, Karimova A, Brierley J, Cross N, Skellett S, Peters MJ.
Pediatrics 2010 126;816-827.[abstract]
Review posted March 28, 2014
- What is being studied?:
Infants less than 3 months old with PCR or culture confirmed B. pertussis infection, who were consecutively admitted to intensive care in 2 separate time periods: from Jan. 2001 to Dec. 2004 (pre leukodepletion protocol) and from Jan. 2005 to August 2009 (Post leukodepletion protocol)
- The study objective:
Enumerate benefits of protocolized, aggressive reduction in white blood cells in infants less than 3 months old with severe Bordetella pertussis compared to historical controls.
-
The study design:
The authors describe this as a case series. However, this design could be considered a retrospective cohort. This article selected patients based on the presence of the disease not based on the outcome (1). The results also allow for calculation of the odds ratio (1).
- The patients included:
- The patients excluded:
The authors sought to describe the clinical course of a consecutive series of critically ill patients with B. pertussis infections and did not exclude patients meeting study entry criteria.
- The interventions compared:
The selected time periods were based on available therapies during each time period. In January 2005, a novel protocol based on leukodepletion was implemented. Exchange transfusion was performed in patients not requiring ECMO while leukofiltration was performed through a WBC filter. The goal was to obtain WBC count < 50,000/mL with a hematocrit between 0.4 to 0.45 in patients not requiring ECMO and a WBC of < 15,000/mL in patients on ECMO. Prior to Jan. 2005, the infants with pertussis were managed without aggressive therapy aimed at leukodepletion.
-
The outcomes evaluated:
The primary outcome was survival to ICU discharge based individual cases and a case mix analysis. Secondary outcomes included reduction in white blood cell count both absolute and percentage fall in white blood cell count.
- The study objective:
- Are the results of the study valid?
Primary questions:
- Was the assignment of patients to treatments randomized?
This is an observational study. Therefore, patients were not randomized to a treatment assignment.
- Were all patients who entered the trial properly accounted for and attributed at its conclusion?
Was follow-up complete?
All patients followed after initiation of the leukodepletion protocol were accounted for. The authors were able to obtain all information for the leukodepletion cohort and the historical comparison cohort.
Were patients analyzed in the groups to which they were randomized?
This is not a randomized study. The patients were analyzed according to the time frame in which they were admitted.Secondary questions:
- Were patients, health workers, and study personnel "blind" to treatment?
No, study personnel were not blinded to treatment.
- Were the groups similar at the start of the trial?
The groups were similar in many ways at the time of enrollment. Among the patients receiving ECMO in the latter era, the median number of ventilator days before transport to the regional referral center was significantly smaller.
- Aside from the experimental intervention, were the groups treated equally?
The patients admitted after Jan. 2005 were managed with low tidal volume strategy and transitioned to oscillatory ventilation if they required mean airway pressure > 16 cm H2O. It is unclear if the patients admitted between Jan. 2001 to Dec. 2004 were ventilated with the same strategies. ECMO management, ventilator management, and other general ICU management techniques vary greatly over an 8 year period. Therefore, it is difficult to assess if the 2 groups were treated equally aside from the implantation of the leukodepletion protocol. However, given these patients were all admitted to a single center over the 8 year time frame is an attempt to minimize differences between institutions.
- What were the results?
- How large was the treatment effect?
Those infants requiring ECMO between 2001 and 2004 had a mean reduction in leukocyte count of 55% by 10 hours. This was significantly less than an 83% mean reduction by 10 hours which was demonstrated in the latter cohort requiring ECMO with leukofiltration (p = 0.017). Infants not requiring ECMO admitted from 2001 to 2004 demonstrated a rise of between 15-20%. Only 3 infants admitted between 2005 to 2009 required leukodepletion by means of transfusion. Those 3 infants had a percent fall similar to those receiving ECMO. The odds ratio of mortality is 7.2 comparing 9 infants with 4 deaths from 2001 to 2004 and 10 infants with 1 death from 2005 to 2009.
When comparing the case mix of the earlier cohort, the expected mortality and observed mortality was not significantly different with a p = 0.87. However, the latter cohort exhibited on 10% mortality which was significantly less than the expected mortality of 47%.
- How precise was the estimate of the treatment effect?
The 95% confidence interval for the odds ratio of death is 0.47 to 227. This is not significant as it crosses 1. The 95% confidence interval for the actual mortality of 44% from 2001 to 2004 was 19% - 73%. The 95% confidence interval for the mortality rate of 10% from 2005 - 2009 was 1.8% - 40%. These wide intervals suggest poor precision most likely related to the small sample size.
- How large was the treatment effect?
- Will the results help me in caring for my patients?
- Can the results be applied to my patient care?
Our patients with B. pertussis appear to be similar to the patients described in this study. Ours tend to be young with a significantly elevated WBC. Despite the evidence of statistical significance, I feel these results are important and are applicable to our patients.
- Were all clinically important outcomes considered?
From Jan. 2001 to Dec. 2004, 4 infants had significant pulmonary hypertension with 3 of those 4 dying. From Jan. 2005 to Dec. 2009, 5 infants had significant pulmonary hypertension with only 1 of those dying. Analyzing this data may show improved survival in infants with pertussis receiving leukodepletion compared to infants with pertussis not aggressively leukodepleted. Halasa has shown the severe mortality of infants with pertussis and suprasystemic pulmonary hypertension requiring ECMO (2). Paddock described the relationship of severe pulmonary hypertension and intraluminal pulmonary artery deposits of leukocytes (3).
It would be interesting to know if there were more ventilator free days or a shortened length of ICU stay following their protocol initiation.A comparison of oxygenation between the groups would be interesting. Romano identified an
improved oxygenation after leukodepletion (4). - Are the likely treatment benefits worth the potential harms and costs?
Despite the clarity of statistically demonstrable benefit, I feel it is worth the risk of double volume exchange transfusion or leukofiltration to attempt to minimize risks associated with leukocytosis in infants with pertussis. These risks predominately include transfusion related acute lung injury, infection from blood transfusion, hemolytic reactions, and non-hemolytic reaction including fever. Small infants requiring ICU management are at risk of requiring blood transfusions mostly from iatrogenic etiologies. The apparent benefit includes potential improved mortality and potentially improved pulmonary hypertension. These potential benefits outweigh the risks. The small sample size significantly limits the power of this study. Performing a multi-center, prospective, observational, cohort study after initiation of a uniform protocol could improve the power.
- Can the results be applied to my patient care?
References
- Dekkers OM, Egger M, Altman DG, Vandenbroucke JP. Distinguishing case series from cohort studies. Annals of internal medicine. Jan 3 2012;156(1 Pt 1):37-40.
- Halasa NB, Barr FE, Johnson JE, Edwards KM. Fatal pulmonary hypertension associated with pertussis in infants: does extracorporeal membrane oxygenation have a role? Pediatrics. Dec 2003;112(6 Pt 1):1274-1278.
- Paddock CD, Sanden GN, Cherry JD, et al. Pathology and pathogenesis of fatal Bordetella pertussis infection in infants. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. Aug 1 2008;47(3):328-338.
- Romano MJ, Weber MD, Weisse ME, Siu BL. Pertussis pneumonia, hypoxemia, hyperleukocytosis, and pulmonary hypertension: improvement in oxygenation after a double volume exchange transfusion. Pediatrics. Aug 2004;114(2):e264-266.