Chest radiographs (CXRs) are frequently obtained routinely in intensive care unit (ICU) patients.1 CXRs are also obtained routinely after interventions or surgical procedures. Several studies investigated whether a more restricted use of CXRs is safe for ICU patients. Obtaining CXRs on an on-demand instead of a routine basis may have several advantages, such as a reduction in CXRs with false-positive results, lower costs, and less irradiation to the patient. However, if important findings are missed, the more restricted use may possibly delay therapy and could therefore increase the length of ICU stay, increase the number of ICU readmissions, or even increase mortality.
Some previous studies concluded that CXRs should still be performed on a routine basis for ICU patients because of the frequent incidence of new findings,2 the poor association with clinical examination,3 the frequent incidence of changes in therapy based on the CXR findings,4 and because it is probably more cost effective to catch new findings at an early stage.5 Other studies concluded that routine CXRs should be abandoned because of the low incidence of clinically important findings6–9 or because of the high sensitivity of clinical examination for the more serious conditions diagnosed on CXRs.10 There are also conflicting results regarding the usefulness of routine CXRs after procedures such as endotracheal intubation11,12 or central venous catheter insertion.13–15
Several investigators reported no difference in mortality rate, length of ICU stay, length of hospital stay, or the number of ICU readmissions after the elimination of routine CXRs.9,16–21 Furthermore, there are no indications that the reduction in CXRs was accompanied by a subsequent increase in the number of computed tomography and ultrasound studies.20 As a result, changing the protocol led to a substantial cost reduction.22 Despite this evidence, routine CXRs are still common practice in many ICU departments.23
The majority of the mentioned studies refer to a general ICU population. Only 1 study refers to a postcardiac surgery population.18 Therefore, the aim of this study was to investigate the incidence and clinical significance of abnormalities found on routine postoperative CXRs in cardiac surgery patients and whether a restricted use of CXRs would influence the number of significant findings. The postoperative care for cardiac surgery patients often includes 2 or 3 CXRs within the first 24 hours after surgery. The first CXR is obtained at ICU admission, a second on the morning of the first postoperative day, and eventually, a third CXR is performed after removal of pleural space drains. Our hypothesis is that reducing the number of CXRs to only 1 routine CXR on the morning of the first postoperative day will not lead to a significant underdiagnosis of clinically significant abnormalities.
This study was performed in a tertiary 24-bed closed-format ICU, admitting medical, surgical, and cardiothoracic surgical patients. The medical staff consisted of 16 intensivists and 8 residents in ICU medicine. The study protocol was approved by the local ethics committee. Informed consent was deemed not necessary because no patient interventions were necessary.
We prospectively included all consecutive cardiothoracic patients who underwent cardiac surgery during a 2-month period. All patients were admitted to the ICU directly after surgery and a first CXR was then obtained (admission CXR, CXR 1). A second CXR was performed on the morning of the first postoperative day (postoperative day CXR, CXR 2) and if pleural space drains were present, a third CXR was performed after removal of these drains (drain removal CXR, CXR 3).
The admission CXR (CXR 1) and the drain removal CXR (CXR 3) were both preceded by a clinical investigation performed by an ICU physician. The attending physician was then asked whether the CXR was deemed necessary. This decision was based on (a) physical examination including auscultation of heart and lungs, (b) interpretation of the patients' vital signs, and (c) the results of the first arterial blood gas sample. In case the CXR was deemed necessary, it was marked “clinically indicated”; otherwise, it was also obtained and marked “not clinically indicated.” The postoperative day CXR (CXR 2) was not preceded by a clinical assessment because we preferred to perform this CXR on a routine basis, according to our hypothesis.
The first and second CXRs were both included in the study for all patients, except when it was occasionally not performed because of ICU arrival in the morning (no CXR 1) and early death of 1 patient (no CXR 2). The third CXR was only included when pleural space drains were removed within the first 24 hours after surgery.
Demographic data were collected for all patients. The mean age, mean number of CXRs per patient, and the median duration of ICU and hospital stay were calculated. All CXRs were assessed by both a radiologist and an ICU physician. All CXR abnormalities were noted, and it was also noted whether an abnormality led to an intervention. Only new findings were analyzed and abnormalities already present on the previous CXR were not considered again. The diagnostic efficacy (the number of abnormalities divided by the total number of CXRs) and therapeutic efficacy (the number of interventions based on CXR abnormalities divided by the total number of CXRs) were calculated for all 3 CXRs. The results of CXRs 1 and 3 were classified according to whether these CXRs were marked “clinically indicated” or “not clinically indicated” by the physician beforehand, and the association with the presence of CXR abnormalities was depicted. False negatives, false positives, sensitivity, and specificity were calculated. Findings that led to an intervention and that would have been missed if the latter CXRs were not performed routinely were noted separately.
Data analysis was performed using SPSS version 17.0 (SPSS Inc., Chicago, IL) for Windows.
Two hundred fourteen successive postcardiac surgery patients were included (Table 1). The mean age was 67 years and the majority of patients were men (74%). Most patients underwent coronary arterial bypass grafting (60%), heart valve surgery (21%), or a combination of these (14%). The median length of ICU stay was 1 day.
A total of 534 CXRs were performed (a mean of 2.5 per patient) of which the results are shown in Table 2. One hundred seventy-nine abnormalities were found, resulting in an overall diagnostic efficacy of 33.5% for all CXRs. Pleural effusion, atelectasis, consolidation, a widened mediastinum, and malposition of invasive devices were the most frequent abnormalities on CXRs 1 and 2. CXR 3 showed a low incidence of a widened mediastinum and pneumothorax. The overall therapeutic efficacy was 2.4%.
The association between the physician's answer to whether a CXR was clinically indicated beforehand and the results of these CXRs is depicted for the admission and drain removal CXRs in Table 3. The label “clinically indicated” or “not clinically indicated” had a sensitivity and specificity of respectively 19% and 93% for CXR 1 and 0% and 92% for CXR 3. If the admission and drain removal CXRs that were marked “not clinically indicated” would not have been performed, 68 abnormalities would have been missed. Five of these abnormalities led to an intervention (Table 4): respectively, a change in endotracheal tube position in 2 cases, the start of diuretic therapy, a change in intraaortic balloon pump position, and an ultrasound-guided drainage of pleural effusion. None of these interventions was initiated by abnormalities on CXR 3. Only 32 (10%) of all 321 admission and drain removal CXRs were clinically indicated beforehand according to the physician's opinion. Therefore, when routine CXRs would not have been performed, a reduction of 289 CXRs would have occurred in a 2-month period.
We found an overall diagnostic efficacy of 33.5% for all CXRs performed in the first 24 hours after cardiac surgery. Because the majority of findings did not lead to a change in therapy, the overall therapeutic efficacy was only 2.4%. These results correspond with the findings of studies referring to a general ICU population.2,3,5–9,17,19 In our study, the association between clinically indicated CXRs and the presence of CXR abnormalities was poor. This also confirms previous results.3 However, although the value of a clinical examination for predicting CXR findings may be limited, this simple procedure can still identify some abnormalities that may lead to complications.
As mentioned above, several investigators have studied the effect of eliminating routine CXRs in a general ICU population.9,16–22 A recent multicenter cluster-randomized trial in a general mechanically ventilated ICU population found a 32% reduction in CXRs within the on-demand strategy group compared with the routine strategy group,21 and as in all previous studies comparing these strategies, they did not find any difference in secondary outcome measures. Mets et al.18 found comparable results in a population of postcardiothoracic surgery patients. However, they did not investigate the clinical consequence of the CXR abnormalities found, and neither study investigated what findings were missed in the on-demand strategy group.
In this study, we investigated the incidence and clinical significance of CXR abnormalities found by postoperative CXRs in cardiac surgery patients. We also investigated whether it would be possible to reduce the number of routine CXRs to only 1 in the first 24 hours of ICU stay. According to our study design, in which a routine CXR was performed in every case after clinical examination, no findings could be missed and special emphasis was placed at the clinical consequence of abnormalities found. Beforehand, we aimed for elimination of routine CXRs 1 and 3 because CXR 1 is taken shortly after surgical closure of the chest and CXR 3 is taken shortly after CXR 2 following the limited-risk procedure of pleural space drain removal. CXR 2 would then still be performed routinely because most of the patients will be transferred to the ward shortly after. However, our results show that clinically significant abnormalities were found on 4.2% of CXRs 1 followed by another 1.9% of CXRs 2. Although it may not seem reasonable to delay the first postoperative CXR, we did not investigate the possible consequence of treatment delay of the abnormalities we found. To perform a CXR just before transfer to the ward has a certain safety benefit, but if this is the first postoperative CXR, it will delay diagnosis of abnormalities already present shortly after surgery. The additive benefit of CXR 3 seems to be limited, although delaying CXR 2 until after pleural space drain removal can still ensure finding a rare pneumothorax.
Our study is limited by the fact that this was a single center study. Because we only investigated the value of routine CXRs in the first 24 hours of ICU stay, we did not address the long-term (>24 hours) safety issue of changing the CXR protocol.
For the majority of patients, it seems reasonable to reduce the number of routine CXRs in the first 24 hours of ICU stay, but the optimal timing of these CXRs remains unclear. A new on-demand strategy may lead to a substantial reduction in CXRs, but we also showed that clinical assessment only is not sufficient to identify those patients who are at risk for underdiagnosis in this new strategy. The risk/benefit ratio might be unfavorable in certain patient groups. A more detailed analysis of this topic is beyond the scope of this study and further studies are necessary to identify those categories of patients at risk to address the safety issue adequately.
1. Trotman-Dickenson B. Radiology in the intensive care unit (part I). J Intensive Care Med 2003;18:198–210
2. Bekemeyer WB, Crapo RO, Calhoon S, Cannon CY, Clayton PD. Efficacy of chest radiography in a respiratory intensive care unit: a prospective study. Chest 1985;88:691–6
3. Hall JB, White SR, Karrison T. Efficacy of daily routine chest radiographs in intubated, mechanically ventilated patients. Crit Care Med 1991;19:689–93
4. Marik PE, Janower ML. The impact of routine chest radiography on ICU management decisions: an observational study. Am J Crit Care 1997;6:95–8
5. Brainsky A, Fletcher RH, Glick HA, Lanken PN, Williams SV, Kundel HL. Routine portable chest radiographs in the medical intensive care unit: effects and costs. Crit Care Med 1997;25:801–5
6. Silverstein DS, Livingston DH, Elcavage J, Kovar L, Kelly KM. The utility of routine daily chest radiography in the surgical intensive care unit. J Trauma 1993;35:643–6
7. Fong Y, Whalen GF, Hariri RJ, Barie PS. Utility of routine chest radiographs in the surgical intensive care unit: a prospective study. Arch Surg 1995;130:764–8
8. Graat ME, Choi G, Wolthuis EK, Korevaar JC, Spronk PE, Stoker J, Vroom MB, Schultz MJ. The clinical value of daily routine chest radiographs in a mixed medical-surgical intensive care unit is low. Crit Care 2006;10:R11
9. Hendrikse KA, Gratama JW, Hove W, Rommes JH, Schultz MJ, Spronk PE. Low value of routine chest radiographs in a mixed medical-surgical ICU. Chest 2007;132:823–8
10. Bhagwanjee S, Muckart DJ. Routine daily chest radiography is not indicated for ventilated patients in a surgical ICU. Intensive Care Med 1996;22:1335–8
11. Brunel W, Coleman DL, Schwartz DE, Peper E, Cohen NH. Assessment of routine chest roentgenograms and the physical examination to confirm endotracheal tube position. Chest 1989;96:1043–5
12. Lotano R, Gerber D, Aseron C, Santarelli R, Pratter M. Utility of postintubation chest radiographs in the intensive care unit. Crit Care 2000;4:50–3
13. Sanabria A, Henao C, Bonilla R, Castrillón C, Cruz H, Ramírez W, Navarro P, González M, Díaz A. Routine chest roentgenogram after central venous catheter insertion is not always necessary. Am J Surg 2003;186:35–9
14. Lessnau KD. Is chest radiography necessary after uncomplicated insertion of a triple-lumen catheter in the right internal jugular vein, using the anterior approach? Chest 2005;127:220–3
15. Abood GJ, Davis KA, Esposito TJ, Luchette FA, Gamelli RL. Comparison of routine chest radiograph versus clinician judgment to determine adequate central line placement in critically ill patients. J Trauma 2007;63:50–6
16. Krivopal M, Shlobin OA, Schwartzstein RM. Utility of daily routine portable chest radiographs in mechanically ventilated patients in the medical ICU. Chest 2003;123:1607–14
17. Graat ME, Kroner A, Spronk PE, Korevaar JC, Stoker J, Vroom MB, Schultz MJ. Elimination of daily routine chest radiographs in a mixed medical-surgical intensive care unit. Intensive Care Med 2007;33:639–44
18. Mets O, Spronk PE, Binnekade J, Stoker J, de Mol BA, Schultz MJ. Elimination of daily routine chest radiographs does not change on-demand radiography practice in post-cardiothoracic surgery patients. J Thorac Cardiovasc Surg 2007;134:139–44
19. Clec'h C, Simon P, Hamdi A, Hamza L, Karoubi P, Fosse JP, Gonzalez F, Vincent F, Cohen Y. Are daily routine chest radiographs useful in critically ill, mechanically ventilated patients? A randomized study. Intensive Care Med 2008;34:264–70
20. Kröner A, Binnekade JM, Graat ME, Vroom MB, Stoker J, Spronk PE, Schultz MJ. On-demand rather than daily-routine chest radiography prescription may change neither the number nor the impact of chest computed tomography and ultrasound studies in a multidisciplinary intensive care unit. Anesthesiology 2008; 108:40–5
21. Hejblum G, Chalumeau-Lemoine L, Ioos V, Boëlle PY, Salomon L, Simon T, Vibert JF, Guidet B. Comparison of routine and on-demand prescription of chest radiographs in mechanically ventilated adults: a multicentre, cluster-randomised, two-period crossover study. Lancet 2009; 374:1687–93
22. Price MB, Grant MJ, Welkie K. Financial impact of elimination of routine chest radiographs in a pediatric intensive care unit. Crit Care Med 1999;27:1588–93
23. Graat ME, Hendrikse KA, Spronk PE, Korevaar JC, Stoker J, Schultz MJ. Chest radiography practice in critically ill patients: a postal survey in the Netherlands. BMC Med Imaging 2006;6:8
MT and NJMvdM helped design and conduct the study, analyze the data, and write the manuscript. AK helped analyze the data and write the manuscript. CLMvdH helped conduct the study and write the manuscript. PMJR and TAR helped design and conduct the study. HAJD helped conduct the study. MB helped design the study. MJS helped design the study, analyze the data, and write the manuscript. All authors have seen the original study data, reviewed the analysis of the data, and approved the final manuscript; MT is the author responsible for archiving the study files.
This report will be presented, in part, at EACTA 2010 (June 9–11) and NVA (Dutch National Anesthesiology Conference) 2010 (May 20–21). This report describes an observational clinical study. This study includes hundreds or thousands of subjects. This report describes a cohort observational clinical study. The author states that the report includes every item in the STROBE checklist for cohort observational clinical studies.