Are daily routine chest radiographs necessary after pulmonary surgery in adult patients?

Abstract

A best evidence topic was constructed according to a structured protocol. The question addressed was whether daily routine (DR) chest radiographs (CXRs) are necessary after pulmonary surgery in adult patients. Of the 66 papers found using a report search, seven presented the best evidence to answer the clinical question. Four of these seven studies specifically addressed post-cardiothoracic adult patients. Three of these seven studies addressed intensive care unit (ICU) patients and included post-cardiothoracic adult patients in well-designed studies. Six of these seven studies compared the DR CXRs strategy to the clinically indicated, on-demand (OD) CXRs strategy. Another study analysed the clinical impact of ceasing to perform the DR, postoperative, post-chest tubes removal CXRs. The authors, journal, date and country of publication, study type, group studied, relevant outcomes and results of these papers are given. We conclude that, on the whole, the seven studies are unanimously in favour of forgoing DR CXRs after lung resection and advocate OD CXRs. One study suggested that hypoxic patients could benefit from a DR CXRs strategy, while other studies failed to identify any subgroup for whom performing DR CXRs was beneficial. Indeed, DR CXRs, commonly taken after thoracic surgery, have poor diagnostic and therapeutic value. Eliminating them for adult patients having undergone thoracic surgery significantly decreases the number of CXRs per patient without increasing mortality rates, length of hospital stays (LOSs), readmission rates and adverse events. Hence, current evidence shows that DR CXRs could be forgone after lung resection because OD CXRs, recommended by clinical monitoring, have a better impact on management and have not been proved to negatively affect patient outcomes. Moreover, an OD CXRs strategy lowers the cost of care. Nevertheless, an OD CXRs strategy requires close clinical monitoring by experienced surgeons and dedicated intensivists. However, given the published studies' low level of evidence, prospective and randomized trials, specifically after thoracic surgery, are necessary in order to confirm these results.

INTRODUCTION

A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].

THREE-PART QUESTION

In [post-thoracic surgery adult patients] is [daily routine chest radiographs strategy] or [on-demand strategy] the best strategy for [clinical management (diagnostic and therapeutic values), length of stay, mortality rate, readmission rate and adverse events]?

CLINICAL SCENARIO

You have a 76-year old emphysematous man under your care. The admission chest radiography and the tomodensitometry show bilateral pulmonary emphysema with a right, compressive bullous lesion. You decide to perform a bullectomy by videothoracoscopy. The first postoperative days are characterized by air leaks without clinical consequences. However, the patient remained in your unit with an apical chest tube that objectified prolonged air leaks. The patient asked about the utility of daily routine (DR) chest radiographs (CXRs). You, therefore, decided to look up the evidence in the literature.

SEARCH STRATEGY

The Medline 1990 to December 2012 database, with Ovid interface, was used, with results limited to English language articles: (‘daily chest radiographs’ [MeSH Terms]) AND (‘thoracic surgery’ [All Fields]) OR (‘intensive care unit’ [All Fields]). Finally, a manual search was used to follow-up on references from the retrieved studies.

SEARCH OUTCOME

A total of 66 abstracts were found, from which seven papers were selected for providing the best evidence on the topic. These papers are documented in Table 1.

Table 1:

Overview of the studies

Author, date, journal and country Study type (level of evidence)	Patient group	Outcomes	Key results	Comments/weaknesses
Graham et al. (1998), J Thorac Cardiovasc Surg, USA [2] Cohort study (6-month period) (level 3)	Study group: 99 patients undergoing 100 open thoracotomies Surgical procedures: 12 pneumonectomies, 37 lobectomies, 33 wedge resections, 18 other indications	Efficacy and clinical utility of DR CXR Cost	Minimal impact on management Elimination of DR CXRs (except the 1st postoperative CXR) allows a reduction of cost by 82%	Single centre and small sample size No randomization group
Mets et al. (2007), J Thorac Cardiovasc Surg, Netherlands [3] Cohort study (6-month period) (level 3)	Study group: 338 post- cardiothoracic patients over two 3-month period Phase 1 (DR CXRs): 169 patients in ICU, 167 patients in post-ICU ward Phase 2 (OD CXRs): 163 patients in ICU, 161 patients in post-ICU ward Surgical procedures: Phase 1: 88 CABG, 39 valve replacement, 27 CABG + valve replacement, 21 other indications Phase 2: 68 CABG, 47 valve replacement, 22 CABG + valve replacement, 26 other indications	Number of OD CXRs OD CXRs efficacy LOS in ICU LOS in hospital Readmission rate	Number of OD CXR slightly increased in ICU or in post-ICU ward OD CXR revealed more unexpected new abnormalities No difference No difference No difference	Single centre and historical comparison
Whitehouse et al. (2009), Surgeon, UK [4] Cohort study (5-month period) (level 3)	Study group: 74 post-thoracic surgery patients; 66 (89%) had DR postoperative CXRs; 25 (34%) had DR post-chest tubes removal CXRs Surgical procedures: Open thoracotomy: 1 pneumonectomy, 11 lobectomies, 20 wedge resections, 6 pleural surgeries VATS: 9 wedge resections, 18 pleural biopsies Other indications: 8	Clinical impact of DR postoperative CXRs Clinical impact of DR post-chest tube removal CXRs	Minimal impact: No complication for patients with OD strategy vs 3 clinically management change for patients with DR strategy Minimal impact: No complication for patients with OD strategy vs 1 clinically management change for patients with DR strategy	Single centre, small sample size Only postoperative and post-chest tube removal CXRs
Cerfolio et al. (2011), Ann Thorac Surg, USA [5] Cohort study (3-year period) (level 3)	Study group: 1037 thoracotomy and pulmonary resection patients Surgical procedures: 609 lobectomies or bilobectomies, 146 segmentectomies, 282 wedge resections	Clinical benefit of DR CXRs for hypoxic patients Clinical benefit of DR CXRs for non-hypoxic patients	Majority of hypoxic patients had pneumothorax in recovery room 79% of hypoxic patients had CXR findings that influence their care 27% of non-hypoxic patients had CXR findings that influence with a minor degree the management of their care	Single centre Observational retrospective study
Graat et al. (2007), Intensive Care Med, Netherlands [6] Cohort study (10-month period) (level 3)	Study group: 1376 patients over two 5-month period Phase 1 (DR CXRs): 754 patients Phase 2 (OD CXRs): 622 patients Type of patients: Phase 1: cardiac surgery 42%, medical 26%, surgical 19%, neurosurgical or neurology 9%, other 4% Phase 2: cardiac surgery 49%, medical 19%, surgical 21%, neurosurgical or neurology 7%, other 3%	Number of OD CXRs Diagnostic and therapeutic values of OD CXRs LOS in ICU Readmission rate ICU and mortality rates	Minimal increased in Phase 2 vs Phase 1 Increased with the intervention No difference No difference No difference	Single centre Unspecific study to the post-thoracic surgery patients
Hejblum et al. (2009), Lancet, France [7] Cohort, multicentre, cluster randomized study (level 2)	Study group: 849 patients DR CXRs strategy: 424 patients OD CXRs strategy: 425 patients Type of patients: - Medical 88% - Surgical 12%	Number of CXRs per patient-day of mechanical ventilation Length of mechanical ventilation LOS in ICU ICU mortality rate	Substantial decrease of the number of CXRs with the OD strategy (32%) No difference No difference No difference	Unspecific study to the post-thoracic surgery patients Specific study to mechanically ventilated adult patients
Oba and Zaza (2010), Radiology, USA [8] Meta-analysis (level 1)	Study group: 7078 patients DR CXRs strategy: 3429 patients OD CXRs strategy: 3649 patients Type of patients: - Medical 59% - Surgical 41%	ICU and hospital mortality rates ICU and hospital LOS Adverse events Subgroup in which DR CXRs might be beneficial	No difference No difference No difference No subgroup	Unspecific study to the post-thoracic surgery patients Specific study to ICU adult patients Meta-analysis included both randomized control and observation studies

DR: daily routine; OD: on-demand; CXRs: chest radiographs; CABG: coronary artery bypass grafting; VATS: video-assisted thoracoscopic surgery; ICU: intensive care unit; LOS: length of stay.

RESULTS

Graham et al. [2] evaluated the efficacy, cost and clinical utility of CXRs during the entire postoperative period of 99 patients undergoing 100 open thoracotomies. The study of the CXRs resulted in the modification of the anticipated management in 33 of 731 DR films (4.5%) and in 10 of 38 on-demand (OD) films (26%). There were no differences (P = 0.454) in the severity of the CXRs findings between DR and OD CXRs.

Mets et al. [3] prospectively analysed the impact of forgoing DR CXRs for cardiothoracic surgery patients (two periods were compared that differed in the number of DR CXRs). Before this intervention (Phase 1), 353 DR CXRs and 261 OD CXRs were taken for 175 patients (3.5 ± 4.3 CXRs per patient) in the ICU; after the intervention (Phase 2), 275 OD CXRs were taken for 163 patients (1.7 ± 1.2; mean difference, 1.8; 95% CI, 1.1–2.5). In the post-ICU period, 413 OD CXRs were taken for 167 patients in Phase 1, whereas 445 OD CXRs were taken for 161 patients in Phase 2 (2.4 ± 1.4 and 2.8 ± 1.4, respectively; mean difference, −0.4; 95% CI, −0.7 to 0.1).

Whitehouse et al. [4] prospectively studied the effectiveness of DR postoperative CXRs and DR post-thoracotomy tube removal CXRs, in 74 post-thoracic surgery patients. Sixty-six patients (89%) had DR postoperative CXRs; 8 (11%) did not. None of the 8 above-mentioned patients suffered from any adverse complications. On the other hand, the management of 3 of the 66 DR CXRs patients changed (with clinical findings that indicated and justified this change). Twenty-five patients (34%) had CXRs routinely performed after the removal of their chest tubes, whereas 49 (66%) did not. Those 49 patients did not experience adverse consequences. Only 1 of the 25 DR CXRs had a change in management (also clinically indicated).

Cerfolio et al. [5] examined the clinical impact of DR CXRs on patient care after elective pulmonary resection by thoracotomy. The authors recorded a total of 1037 patients in the recovery room, some of whom had pneumothorax and some did not. Patients were either stratified by the development of hypoxia during their hospital stay or lack thereof. The majority of patients (n = 834; 80%) did not have pneumothorax in the recovery room. Of these patients, 20 (2%) developed hypoxia. Two hundred and three patients (20%) did have pneumothorax in the recovery room. Forty-two of these patients (20%) developed hypoxia.

Graat et al. [6] evaluated the diagnostic and therapeutic value of OD CXRs in a multidisciplinary ICU, before (Phase 1) and after (Phase 2) forgoing DR CXRs. In Phase 1, 754 patients had a total of 3894 CXRs (2457 DR CXRs and 1437 OD CXRs) taken. In Phase 2, 622 patients had a total of 1267 OD CXRs taken. The ratio of CXRs per patient/day decreased from 1.1 ± 0.3 to 0.6 ± 0.4 after the intervention (P < 0.05). The diagnostic and therapeutic value of OD CXRs increased as the intervention unfolded. Length of stay (LOS) in ICU, readmission rates, ICU and hospital mortality rates were not different between the two phases.

Hejblum et al. [7] studied the efficacy of DR vs OD CXRs, for mechanically ventilated patients. A DR strategy determined that 424 patients should have CXRs, while an OD strategy determined that 425 patients should have CXRs. A total of 7755 CXRs were taken. The mean number of CXRs per patient/day of ventilation was significantly lower with the OD strategy than with the DR strategy (0.75 (0.67–0.83) vs 1.09 (1.05–1.14), P < 0.0001). However, the duration of mechanical ventilation, LOS in ICU or ICU mortality rate did not differ from one strategy to the other.

Oba and Zaza [8] performed a meta-analysis to assess whether forgoing DR CXRs would decrease mortality and LOS outcomes and to identify a subgroup for which DR CXRs might be beneficial. Eight studies of adult medical or surgical ICU patients were included (7078 total patients). A DR CXRs strategy was used for 3429 patients, whereas an OD CXRs strategy was used for 3649 patients. The mean number of CXRs per patient was significantly lower in the OD group than in the DR CXRs group (mean difference: 3.15; 95% CI: 0.88–5.43; P < 0.01). The forgoing of DR CXRs did not affect hospital or ICU mortality (P = ns), nor ICU or hospital LOS (P = ns).

CLINICAL BOTTOM LINE

On the whole, all of the seven studies presented were clearly in favour of forgoing DR CXRs. Three of the ICU settings included in the above-mentioned studies involved post-cardiothoracic adult patients, and four dealt specifically with post-cardiothoracic adult patients. It has been shown that DR CXRs have poor diagnostic and therapeutic value and that forgoing them significantly decreases the number of CXRs per patient, without increasing mortality rates, hospital LOS or adverse events. Furthermore, current evidence indicates that DR CXRs could be forgone after lung resection because OD CXRs, indicated by clinical monitoring, have a better impact on management. Moreover, the cost of care could decrease with an OD CXRs strategy. Nevertheless, this OD strategy requires close clinical monitoring by experienced surgeons and dedicated intensivists. However, prospective and randomized trials are necessary to confirm these results within the context of thoracic surgery.

Conflict of interest: none declared.