Retrospective comparison of costs between medical and surgical treatment of canine pyothorax

Abstract

This study compared costs of treating dogs with pyothorax medically versus surgically. Medical records from the University of Wisconsin School of Veterinary Medicine were searched for cases of pyothorax that underwent either medical or surgical treatment. Patients undergoing surgery were subdivided into early (ES; < 48 h) and late (LS; > 48 h) surgery groups. Costs and length of stay were compared between treatment groups. Treatment costs were adjusted for inflation. Nineteen dogs were included in analysis; 7 in the medical group (MG), 5 in the ES group, and 7 in the LS group. Total costs were significantly lower in the MG than in the LS group. Total costs were less for the MG than the ES group, and for ES than LS, but the differences did not achieve significance. Preoperative costs were higher in the LS than the ES group. We conclude that surgery for canine pyothorax is less costly if pursued earlier than later.

Introduction

Pyothorax is defined as an accumulation of purulent material within the pleural space (1). It is a relatively uncommon but potentially life-threatening disease. The most frequently described routes of infection in dogs and cats include migrating foreign bodies such as plant awns, pulmonary neoplasia or abscessation, penetrating wounds, and esophageal perforation; however, the source of infection is not evident in all cases (1–5). A diverse population of microorganisms has been implicated in cases of pyothorax, including obligate anaerobes, Gram-positive filamentous organisms and, occasionally, Gram-negative organisms (2–7).

The clinical diagnosis of pyothorax is usually based on history and physical examination findings including tachypnea, respiratory distress, pyrexia, and lethargy. Thoracic imaging techniques including radiography, ultrasound, or computed tomography (CT) confirm the presence of pleural effusion. Following thoracocentesis, a confirmatory diagnosis of pyothorax can be made by thoracic fluid analysis and bacterial culture. The thoracic fluid is typically exudative, and often intracellular organisms are visualized within neutrophils. Positive culture results are diagnostic for pyothorax; however, positive bacterial culture results may not always be obtained, particularly if anaerobic organisms are present (1).

Treatment of canine pyothorax is multi-modal, and may involve thoracocentesis, intermittent pleural lavage and drainage, administration of antimicrobials, supportive care such as supplemental oxygen, intravenous fluids and analgesics, and surgery in some cases (2–5,7,8). Determination of the most appropriate treatment approach for any given case of pyothorax is challenging and may depend on several factors including severity of disease as manifested by clinical signs, underlying cause if known, and financial limitations of the owners (1,9). Retrospective studies comparing medical versus surgical management of pyothorax (2,3,5,7) have been reported. A clear advantage of medical or surgical treatment has yet to be determined. The financial aspect of these treatment approaches has not been previously described.

Materials and methods

Medical records of dogs from the University of Wisconsin’s School of Veterinary Medicine evaluated with pyothorax between 2000 and 2010 were reviewed. Inclusion criteria were confirmation of a diagnosis of pyothorax either by documentation of intracellular bacteria on cytological examination or a positive bacterial culture of pleural fluid. Cases were excluded from the study if medical records were incomplete, if treatment and further diagnostics were declined after initial confirmation of pyothorax, if the patient survived less than 24 h from admission, or if no clear diagnosis of pyothorax was obtained.

The following data were recorded for each case: signalment, presenting clinical signs, initial management at presentation, imaging modalities used, other diagnostic tests (cytology, bacterial culture), final case management (medical management or surgical management), duration of hospitalization, survival to discharge, and causes of in-hospital deaths (death due to clinical deterioration or euthanasia).

Dogs were divided into 3 groups: medical (MG), early surgery (ES), and late surgery (LS). The MG included dogs that underwent thoracocentesis, thoracostomy tube placement (unilateral or bilateral), and intermittent pleural lavage and drainage for treatment of pyothorax. The ES group included dogs that underwent thoracotomy within 48 h of hospitalization (median sternotomy or lateral thoracotomy) for treatment of pyothorax. The LS group underwent thoracotomy (median sternotomy or lateral thoracotomy) after 48 h of hospitalization. The ES and LS groups were devised to account for the likelihood that late surgical intervention may have occurred from failure to improve significantly with medical management (thoracostomy tubes with pleural lavage and drainage) (1).

Antibiotic therapy, both initial in-hospital therapy and antibiotic therapy at the time of discharge were recorded. Based on in vitro susceptibility results, treatment success with an empirical combination of enrofloxacin (Baytril; Bayer Animal Health, Shawnee Mission, Kansas, USA) and ampicillin (Ampicillin; Sagent Pharmaceuticals, Schaumburg, Illinois, USA), which is a common empirical combination in our hospital, was assessed.

To allow comparison, costs associated with treatment were recorded and adjusted for inflation using an online calculator (CPI Inflation Calculator; http://146.142.4.24/cgi-bin/cpicalc.pl). Due to the small numbers in each group nonparametric analyses were performed. Cost and length of stay comparisons between the 3 groups was made using Kruskal-Wallis analysis of variance and Dunn’s post hoc analysis. Mann Whitney U-test was used to compare length of stay before and following surgery in the ES and LS groups, and to compare preoperative and post-operative costs between the ES and LS groups. Significance was set at P < 0.05.

Results

Medical records of 64 dogs with possible pyothorax were reviewed. Of these, 35 were excluded because a definitive diagnosis of pyothorax was not obtained. Six additional cases were excluded because owners declined all treatment after the initial diagnosis. Four cases were excluded because the dogs died within 24 h of presentation to the hospital. Nineteen cases remained for analysis, with hunting breeds most frequently affected. Breeds included English springer spaniels (n = 6), Labrador retrievers (n = 4), and 1 each of beagle, German short-haired pointer, golden retriever, Doberman, bull mastiff, English setter, St. Bernard, Siberian husky, and border collie. There were 9 male dogs (5 intact, 4 neutered) and 10 female dogs (2 intact, 8 spayed). The median age of the dogs was 4 y (range: 1 to 7 y). Presenting clinical signs included tachypnea (14/19), lethargy (11/19), anorexia (6/19), fever (6/19), weight loss (4/19), exercise intolerance (3/19), and cough (1/19).

Every animal had a confirmed diagnosis of pyothorax that was based on either a positive bacterial culture or visualization of intracellular bacteria on cytology of pleural fluid, or both. Bacterial culture of pleural effusion was performed in all 19 patients. One culture did not result in bacterial growth; 18 cultures were positive for bacteria including: Fusobacterium sp. (n = 7), Prevotella sp. (n = 7), Actinomyces sp. (n = 5), Bacteroides sp. (n = 4), Pasteurella sp. (n = 2), Escherichia coli (n = 2), and 1 isolate each of Staphylococcus sp., Mycoplasma sp., Nocardia sp., Clostridium sp., and Peptostreptococcus sp. Cytology of pleural effusion was performed in 19 patients, and intracellular bacteria were observed in 15.

Bilateral thoracostomy tubes were placed in 13/19 dogs; intermittent pleural lavage and drainage was performed in all dogs that had thoracostomy tubes placed. Intermittent pleural lavage and drainage consisted of infusing warmed crystalloid solutions (mean volume: 19.0 ± 12.1 mL/kg) 2 to 4 times daily, and promptly draining the lavage solution. Of the remaining 6 dogs that did not initially have thoracostomy tubes placed, 5 were treated promptly with surgery, and 1 dog was treated as an outpatient with antibiotics following thoracocentesis.

A total of 12 dogs underwent surgical intervention; 5 were in the ES group and 7 in the LS group. Surgical approaches included median sternotomy (n = 11) or lateral thoracotomy (n = 1). The sole dog treated via lateral thoracotomy had an abscessed lung lobe diagnosed via thoracic CT scan; a total of 8 dogs underwent a thoracic CT scan. Dogs treated with a median sternotomy did not have apparent lateraling lesions.

Initial empirical intravenous antimicrobial therapy was initiated in all 19 dogs; 18 dogs received a combination of more than one antimicrobial agent. All dogs were discharged on oral antimicrobial therapy for a median of 33 d (range: 10 to 120 d). Enrofloxacin and ampicillin are a common empirical choice in our hospital and would have been effective in 18 of 19 cases for which culture and susceptibility results were available.

Hospitalization costs and length of stay results are presented in Table 1. Median costs were significantly higher for the LS group than the MG. Total costs were less for the MG than the ES group, but did not achieve statistical significance. Similarly, total costs for ES were less than LS, but did not achieve statistical significance. Length of stay was significantly longer for the LS group when compared to both the ES and MG. Median preoperative costs were significantly higher in the LS than the ES group. All patients in the 3 groups survived to discharge. None of the patients experienced recurrence of pyothorax within 6 mo of discharge.

Table 1.

All values are reported as median and range and were determined with a CPI Inflation Calculator, http://146.142.4.24/cgi-bin/cpicalc.pl

	MG (n = 7)	ES (n = 5)	LS (n = 7)
Total cost	$3094 ($410 to $3818)a	$4689 ($4079 to $8393)	$9694 ($6053 to $11 533)a
	£1933 (£256 to £2386)a	£2930 (£2549 to £5246)	£6059 (£3783 to £7208)a
Total days of hospitalization	5.5 (1 to 8)a	7 (4 to 7)b	10 (8 to 12)a,b
Days hospitalized postoperatively	n/a	5 (3 to 5)	5 (5 to 7)
Days hospitalized preoperatively	n/a	2 (1 to 2)a	4 (3 to 7)a
Preoperative costs	n/a	$1339 ($1153 to $2398)a	$3928 ($1650 to $4805)a
		£837 (£720 to £1498)a	£2455 (£1031 to £3003)a
Postoperative costs	n/a	$3461 ($2914 to $5995)	$5816 ($2873 to $6727)
		£2163 (£1821 to £3747)	£3635 (£1795 to £4202)

^a,bDenote significant difference between groups: MG — medical group; ES — early surgery group; LS — late surgery group. $ are US dollars. n/a — Not available.

Discussion

Canine pyothorax is often a challenging and costly disease. Patients treated via LS had significantly greater expenses and length of stay than did those in the MG. The significantly longer hospital stays in the LS-treated patients contributed to this difference. Preoperative costs in the LS group were higher than in the ES group. Both the higher costs and length of stay in the LS are an expected finding attributable to study design. The decision to divide the surgical patients into ES and LS groups at the 48-hour mark was largely arbitrary. In the authors’ experience, cases that were initially planned to be treated surgically typically received surgery within the first 48 h after admission to the hospital. However conversion from medical to surgical treatment has been recommended if there is no improvement following 48 to 72 h of intermittent pleural lavage and drainage with thoracostomy tubes (1). Treatment costs in the ES group were lower than those in the LS group, and treatment costs of the MG were less than in the ES group. These differences were not statistically significant but do warrant attention as they may be clinically significant. The groups were small and if more cases could have been included perhaps these differences would have achieved statistical significance.

It is interesting that there were no deaths in any patients while hospitalized. Part of this likely occurred from the selection criteria whereby dogs that died within 24 h of admission were excluded from analysis. If surgery is under consideration, it appears financially advantageous to pursue this course within the first 48 h of hospitalization so long as the patient’s condition is deemed stable for general anesthesia and surgery.

Outcomes of patients treated for pyothorax in our study appear similar to previous reports (2,3,5,7). Boothe et al (3) reported a 77% discharge rate for patients treated with thoracostomy tubes compared to 92% for surgically treated patients. Rooney and Monnet (2) reported 12 of 26 cases that failed medical management and ultimately underwent surgical treatment. Johnson and Martin (7) reported treatment success of 100% in 15 dogs treated with medical management, and Demetriou et al (5) reported a successful outcome in 86% of canine and feline patients treated for pyothorax. No patients had recurrence of pyothorax in this study. In prior studies recurrence rates have varied from 0% to 15% (2–4,7). Our lack of recurrences may be due to the small number of cases in this study.

Treatment approaches in this study appear similar to other reports (1–6,8) and, given the retrospective nature and limited number of cases, it is difficult to compare this to other studies. When considering the various treatment protocols (pleural drainage, pleural drainage plus lavage, or surgical debridement and pleural lavage), clients should be informed that when an initial medical course fails and transitions to surgical debridement the total costs are greater than if surgery had been pursued within the first 48 h of care. Most cases of pyothorax do not have a clear indication for surgical debridement (e.g., focal pulmonary or mediastinal lesions) rendering the decision to pursue surgery challenging.

Advanced thoracic imaging was performed infrequently and at the discretion of attending clinicians. As only 8 dogs underwent thoracic CT scan in this report, it is difficult to comment on the value of CT. Others have reported CT as a useful tool for guiding case management and surgical planning (10).

A variety of bacterial organisms were cultured from our patient population which is consistent with other reports (2–4). It was not the aim of this report to describe all the culture results and respective susceptibilities. However, as noted earlier, enrofloxacin and ampicillin would have been effective in 18 of 19 cases for which culture and susceptibility results were available.

There are several weaknesses in this study. As a retrospective study, unknown variability exists with treatment approach and case selection (medical versus surgical). We were unable to determine, from the available records, why particular cases were transitioned from a medical management approach to a surgical approach. Another weakness was that data for illness severity scores (11) were not available to be calculated for these cases. Hence it is unknown if patients’ severity of illness was comparable between the 3 treatment groups. Patients with more severe disease may require more intensive monitoring and treatment resulting in higher costs. It is possible that animals with more severe disease were more likely to undergo surgical management than medical management. Illness severity may also have had a bearing on how cases were managed. The small number of cases included in this study also makes it difficult to draw valid conclusions with the probability of a type II statistical error being high. Ultimately, larger prospective, controlled trials investigating the various available treatment modalities for canine pyothorax are necessary to establish significant differences between medical and surgical management of this disease, including associated costs.

In conclusion, surgical treatment of pyothorax that occurred late in hospitalization (LS) was significantly more costly and resulted in significantly longer hospitalization than with medical therapy alone or early surgical intervention (ES). CVJ