Abstract

Background: This study aims to determine the optimal postoperative oral feeding initiating time with the lowest postoperative pulmonary complication rate in thoracotomy patients and compare cardiac and psychiatric complication rates caused by different feeding schemes.

Methods: The study included 107 consecutive patients (84 males, 23 females; mean age 53.9 years; range 17 to 81 years) planned to undergo lung resection via elective thoracotomy for both benign and malignant pathologies in a single institution during a time period of two years. Patients were prospectively randomized into three groups according to postoperative oral intake initiation time: oral intake was initiated on the postoperative sixth hour in group 1, 24th hour in group 2, and when bowel functions resumed in group 3. Groups were then compared in terms of postoperative complication rates.

Results: Groups were homogenous according to demographic properties. Twenty patients (18.7%) developed postoperative pulmonary complications: four (11.1%) in group 1, eight (22.2%) in each of groups 2 and 3. Median oral intake initiation time for group 3 was 47 hours (range 27 to 82 hours). There was no significant difference between the groups in terms of postoperative pulmonary and cardiac complications (p=0.358 and p=0.175, respectively). While postoperative incidence of delirium was significantly increased in group 3 (n=5, 14.3%, p=0.032), it was not observed in group 1 and it was observed in two patients (5.6%) in group 2. This complication was directly correlated with development of postoperative pulmonary complications (odds ratio=14.2; p=0.002).

Conclusion: Early (sixth hour) initiation of postoperative oral feeding is not related with increased pulmonary complications. On the contrary, early initiation may enable rapid recovery of postoperative mental and physical conditions, prevent psychiatric disorders, and reduce pulmonary complication rates. Thus this scheme can be administrated safely in all thoracotomy patients without potential risk for preoperative aspiration.

Postoperative pulmonary aspiration of gastric content is an increasingly renowned complication after thoracotomy for lung resection.[1]

The rationale of nil by mouth is to prevent postoperative nausea and vomiting, thus a possible aspiration.[2] There are controversies on the oral intake starting time after a general anesthesia operation. It has been shown that gag reflex is regained on second to sixth hour after extubation.[3] Other researchers have shown that after a long duration of anesthesia, tracheal aspiration risk is relatively higher, and they advocate that normal feeding must be started on postoperative day one to avoid aspiration and pulmonary complications.[4,5] In this study, we aimed to determine the optimal postoperative oral feeding initiating time with the lowest postoperative pulmonary complication (PPC) rate in thoracotomy patients and compare cardiac and psychiatric complication rates caused by different feeding schemes.

Methods

The study was conducted with 108 consecutive patients (85 males, 23 females; mean age 53.9 years; range 17 to 81 years) planned to undergo lung resection via elective thoracotomy for both benign and malignant pathologies in Marmara University Medical Faculty Hospital, between June 2008 and May 2010. Based on their admission order, patients were prospectively randomized into three groups according to postoperative oral intake initiation time: oral intake was initiated on the postoperative sixth hour in group 1 (n=36), 24th hour in group 2 (n=36), and when bowel functions resumed in group 3 (n=36). One patient in group 3 did not want to continue fasting on his 28th hour and was excluded accordingly due to research committee’s decision on consent withdrawal. Thus, the study was completed with 107 patients (84 males, 23 females; mean age 53.9 years; range 17 to 81 years) (Figure 1). The study protocol was approved by the Marmara University Medical Faculty Ethics Committee. A written informed consent was obtained from each patient. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Patient allocation was done by a single surgeon. Preoperative assessment included medical history (demographics, symptoms, and their duration), presence of comorbidities (pulmonary, cardiac, and metabolic), tobacco or alcohol habits, recent chest X-ray, electrocardiogram (ECG), spirometry, blood count, biochemistry and arterial blood gas analysis, with predictive perfusion scintigraphy and an echocardiogram, where indicated. All patients were operated via standard serratus-sparing posterolateral thoracotomy, by three consultant surgeons, whom were also blinded to the group allocation. Clinical data, as well as presence of postoperative complications, were recorded prospectively postoperatively.

Figure 1: Consolidated standards of reporting trials flow diagram.

Postoperative complications were defined as pulmonary (pneumonia, atelectasis requiring bronchoscopy, or acute respiratory failure), cardiac (arrhythmias or myocardial infarction, documented by ECG and/or necessary laboratory tests, and confirmed by a cardiologist), and psychiatric problems (anxiety, organic brain syndrome - cognitive disorder, or delirium). Psychiatric complications were assessed and diagnosed by an expert psychiatrist, called for a consultation over an atypical witnessed behavior or an illusion/hallucination of the patient. Diagnosis was made after confirmation of normal blood electrolyte levels and with the mini-mental state examination, the patient was revisited daily by the consultant, and was called for a minimum of six-months for clinical followup after discharge. Postoperative pneumonia was defined as (i) onset of new radiographic parenchymal infiltrates on routine postoperative daily X-rays, determined by the operating consultant surgeon, (ii) fever over 38 °C, and (iii) one o f t he following: increased C-reactive protein or leukocyte count within 24 hours (>10x109/L), purulent sputum, and positive blood or sputum culture (>107 cfu/mL).

Standard “nil-by-mouth from midnight” fasting policy was applied preoperatively. Patients were randomized into three groups with different timings of postoperative oral feeding initiation. Term “feeding” included administration of water and other liquids of unlimited volume, directly advanced to semi-solid foods subsequent to comfortable liquid intake without any delay, followed by a normal (solid) diet given for the next meal, under the supervision of a nurse or a resident. In case of excessive nausea or vomiting, intake would be postponed. Feeding was started on the postoperative sixth hour in group 1, on the 24th hour in group 2, and when the bowel sounds or flatus was present in group 3. All patients were given oral analgesics after postoperative day three.

Exclusion criteria included: (i) p atients with any type of antibiotic usage one week preoperatively for any reason, or (ii) pre-existing pulmonary infection, documented clinically (purulent sputum or fever >38 °C) or radiologically; (iii) a ny p atient i n w hom the routine cefazolin antibiotic prophylaxis was not possible due to drug allergy, or (iv) had a history of any previous gastric operation; (v) patients on neoadjuvant chemo/radiotherapy, (vi) under any medication possible to alter esophageal motility, or with a history of cerebrovascular disease or stroke because of increased aspiration risk; (vii) p atients i n w hom p reoperative thoracic epidural catheter placement was not achieved were not eligible for the study since they received different postoperative analgesics regimen which may influence the gastrointestinal motility; (viii) patients in whom the planned lung resection was cancelled due a to perioperative inoperability criterion, (ix) not strictly conforming the feeding timetable protocol, or (x) not willing to attend at all.

Statistical analysis
Normally distributed parametrical scale variables were reported as mean and standard deviation within 95% confidence interval and were compared using Student’s t-test or analysis of variance. Mann-Whitney U test was used for non-parametrical variables. Categorical variables were presented as percentages with ranges and were assessed using Fisher’s exact test or Pearson’s chi-square. Odds ratio (OR) was used to quantify the risk of postoperative complication rates according to known patient variables. In this study, the maximum type I error was 0.05, and the level of significance was accepted as p<0.05. All analyses were performed using a registered MedCalc Statistical Software Package (MedCalc Software bvba, Belgium).