Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Impact of respiratory syncytial virus infection on surgery for congenital heart disease: postoperative course and outcome.

Khongphatthanayothin A, Wong PC, Samara Y, et al.

Crit Care Med 1999;27:1974-81. [abstract]

Reviewed by Christopher J. Babbitt MD, Jonathan D. Feldman MD, UCLA Children's Hospital

Review posted January 25, 2000

I. What is being studied?:

The study objective:

To describe the postoperative course and outcome of cardiac surgery in children with recent RSV infection. To evaluate whether the timing of surgery has any impact on outcome.

The study design:

A retrospective cohort evaluation.

The outcomes assessed:

Each patient's postoperative course was reviewed. Overall mortality was assessed for 2 different subgroups of patients. Also assessed was the development of pulmonary hypertension and if treatment was required, the duration of mechanical ventilatory or inotropic support, and the duration of postoperative hospitalization. Finally, a variety of postoperative morbidities including: sepsis, DIC, ARDS, pneumonia, chylothorax, SVC syndrome, stridor, tracheal stenosis, RV failure, UTI, reactive airway disease, and pericardial effusion were listed.

II. Are the results in the study valid?

Primary questions:

1. Was there a representative and well-defined sample of patients at a similar point in the course of the disease?

Yes and no. The 25 patients used in the final analysis were all children with an age range of 25 days to 40 months. The patients were identified by a retrospective chart review and all patients that had the discharge diagnosis of CHD and RSV infection from January 1990 until April 1995 were selected. The initial review yielded 50 patients. However, 23 patients were excluded from the study because of; 1) 10 patients developed RSV infection after surgery, 2) 4 patients had surgery > 6 months after RSV infection - the predetermined time frame by the authors, 3) 3 patients died without surgery, 4) 4 patients had "hemodynamically insignificant CHD" - it is not clear if these patients had surgery for small PDA's, ASD's, etc., or did not have surgery, 5) 2 patients were lost to follow up. This resulted in 27 patients, but 2 more were excluded because of "other underlying illnesses" which were thought to contribute more to their outcome then the RSV infection.

The authors then divided the patients into 2 groups - Group 1 consisted of patients who had cardiac surgery during the same admission as the RSV infection and Group 2 consisted of patients who had cardiac surgery after discharge home following the hospitalization for the RSV infection. Group 1, however, did contain 3 patients who were discharged home and then readmitted (within 1 - 4 days) despite what the authors initially state were the criteria for dividing the groups.

Group 2 patients appear to be at a later time course in their RSV infection, older, and weigh more at the time of surgery. (The authors did not state statistical significance)

2. Was follow-up sufficiently long and complete?

Yes. Every patient's hospital course during the immediate postoperative period until his or her death or discharge was reviewed.

Secondary questions:

3. Were objective and unbiased outcome criteria used?

Yes and No. There were some clear objective criteria, including: mortality, duration of mechanical ventilatory support after surgery, duration of inotropic support after surgery, and duration of hospitalization after surgery. However, other outcome criteria are simply a list of clinical problems encountered postoperatively, including: pulmonary hypertension, ARDS, pneumonia, sepsis, DIC, stridor, SVC thrombosis, and chylothorax. Patients in Group 2 are reported as not having pulmonary hypertension by clinical assessment, but they did not have pulmonary artery catheters or undergo postoperative catheterization as 7 Group 1 patients did.

4. Was there adjustment for important prognostic factors?

No. Several potential preoperative prognostic factors, including: diagnosis, cardiac medications at the time of surgery, age at the time of surgery, preoperative pulmonary hypertension, weight at the time of surgery, and respiratory failure at the time of surgery were identified. Other potentially important factors, such as: duration of cardiopulmonary bypass and aortic cross-clamp were not mentioned. There was no attempt to use logistic regression analysis to see if these factors were predictive of outcome.

Furthermore, there was no attempt to elucidate why some patients went on to have cardiac surgery during the same admission as their RSV illness, and why others went home first. Was there something else different between these two groups that is not reported? Clearly the patients were not randomly assigned to the two groups, and controlling for this variable (or these variables) would be important.

III. What are the results?

1. How large is the likelihood of the outcome event(s) in a specified period of time?

At first glance, patients who had cardiac surgery performed during the same admission as when they had a symptomatic RSV infection (Group 1) appear to have increased morbidity and mortality. Likewise, patients that were well enough to be discharged home for at least 2 weeks after an RSV infection and then electively admitted to undergo cardiac surgery (Group 2) appear to have less morbidity and mortality. There was no statistical analysis done on the data by the authors. Below are our analyses based on the data in the paper.

Death	Group 1	Group 2	Total
yes	2	0	2
no	11	12	23
Total	13	12	25

Applying Chi-square testing to the data on mortality for a of 0.05 and degrees of freedom of 1 yields a value of 2.006. This is less than the 3.84 value needed to reject the null hypothesis. Based on these results, there is no statistical difference in the mortality of the 2 groups following cardiac surgery. (The formula used is X^2(1)= n(ad-bc)^2/(a+c)(b+d)(a+b)(c+d))

A relative risk for postoperative "bad outcomes" can also calculated. We have arbitrarily designated a "bad outcome" as death or postoperative ventilatory support greater than 4 days.

	Bad outcome	Good outcome	Total
Group 1	A (7)	B (6)	13
Group 2	C (0)	D (12)	12

The relative risk can be found using the formula:

RR= [A/ (A+B)] / [C/(C+D)]
RR= 6.75

It should be noted, however, that 1 was used rather than 0 for "C". We are applying (loosely) the "rule of threes" so that a meaningful calculation can be done.

Finally, comparing the duration of mechanical ventilation for Group 1 vs. Group 2 using the data presented and a student's t test yields a p value of 0.03. For duration of inotropic support the p value is 0.04. However, 2 patients in Group 1 with particularly long postoperative courses do allow the results to reach statistical significance when the data otherwise would not. Statistical significance cannot be calculated for the duration of postoperative hospitalization with the data given.

It must be noted that there is a striking difference in reported complications between the two groups, as Group I patients had episodes of sepsis, pneumonia, ARDS, LV dysfunction and more (9 of 13 with a complication listed), while the Group II patients had only a transient SVC syndrome and reactive airway disease listed (2 of 12 patients).

2. How precise are the estimates of likelihood?

A confidence interval can be calculated from the relative risk for "bad outcomes." The following formula for a 95% confidence interval for relative risk was used:

exp[ln(RR) +/- 1.96(sq rt {1/A + 1/C - 1/(A+B) - 1/(C+D)}]

This yields a 95% confidence interval of 1.03 and 46.9 for a relative risk of 6.75. Therefore, a statistical significance for "bad outcomes" is just barely reached because the confidence interval does not include 1.

IV. Will the results help me in caring for my patients?

1. Were the study patients and their management similar to my own?

Yes, 30-67% of patients hospitalized for RSV and CHD require PICU admission and 18-24% require mechanical ventilation.(1-4)

2. Will the results lead directly to selecting or avoiding therapy?

Possibly. This is a retrospective study with a small number of patients and it is difficult to draw conclusions. Previous studies have shown increased risks for patients undergoing intubation and general anesthesia within 6 weeks of a URI.(5) Although the current study does not show a statistically significant difference in mortality between the two groups, it appears that delaying surgery until after a symptomatic RSV infection has resolved may be warranted. By our statistical analyses, Group 2 patients do have a decrease in postoperative mechanical ventilatory and inotropic support when compared to Group 1 patients. (p value < 0.05) Group 1 patients also have more "bad outcomes" when compared to Group 2 patients.

3. Are the results useful for reassuring or counseling patients?

No. What makes the data difficult to interpret is that there is no explanation as to why some patients were operated on during their hospitalization for RSV infection. The factors that prompted surgery may have lead to the different outcomes rather than the surgery itself. As the authors mention, this study should be a stimulus to undertake larger prospective studies looking at the management of CHD requiring surgery and RSV infection.

References

1. Macdonald NE, Hall CB, Suffin SC, et al. Respiratory syncytial viral infection in infants with congenital heart disease. N Engl J. Med 1982:307397-400. [abstract]
2. Moler FW, Khan AS, Meliones JN, et al. Respiratory syncyntial virus morbidity and mortality estimates in congenital heart disease patients: A recent experience. Crit Care Med 1992;20:1406-1413. [abstract]
3. Navas L, Wang E, Carvalho V, et al. Improved outcome of respiratory syncyntial virus infection in a high-risk patient hopitalized population of Canadian children. J Pediatr 1992;121:348-354. [abstract]
4. Wang EEL. Law BJ, Stephens D. Pediatric investigator collaborative network on infections in Canada (PICNIC) prospective study of the risk factors and outcomes in patients hospitalized with respiratory syncyntial viral lower respiratory tract infection in children. J. Pediatr 1995;126:212-219. [abstract]
5. Martin LD. Anesthetic implications of an upper respiratory infection in children. Ped Clin NA 1994;41(1)121-131

 

Comments

Back to the EB Journal Club Index