bs_bs_banner
Asian J Endosc Surg ISSN 1758-5902
O R I G I N A L A RT I C L E
Predicting prolonged hospital stay after laparoscopic cholecystectomy Yoshikazu Morimoto, Hitoshi Mizuno, Yusuke Akamaru, Keigo Yasumasa, Hiroshi Noro, Emiko Kono & Yoshio Yamasaki Department of Surgery, Japan Community Healthcare Organization (JCHO), Osaka Hospital, Osaka, Japan
Keywords Laparoscopic cholecystectomy; length of stay; prediction score Correspondence Yoshikazu Morimoto, 4-2-78 Fukushima, Fukushima-ku, Osaka 553-0003, Japan. Tel: +81 6 6441 5451 Fax: +81 6 6445 8900 Email: [email protected] Received 9 November 2014; revised 26 December 2014; accepted 7 February 2015 DOI:10.1111/ases.12183
Abstract Introduction: Widespread application of laparoscopic cholecystectomy (LC) has resulted in a high complication rate and leads to prolonged hospital stays. This study aimed to investigate the preoperative and intraoperative clinical factors that relate to prolongation of hospital stay. Methods: We studied 370 patients who underwent LC for gallbladder disease between 2008 and 2012. Clinical risk factors were retrospectively collected. The clinical pathway for LC was indicated for all patients, and they were divided into two groups according to postoperative length of stay (LOS): the normal duration group (LOS ≤5 days) and the long duration (LD) group (LOS ≥6 days). Multiple regression analysis was used to predict risk factors that identified hospital prolongation to create a LOS prediction score. Results: The normal duration group was 236 patients and the LD group was 134. Seventeen patients (4.6%) required conversion from laparoscopic to open surgery. LOS was 4.82 days in the normal duration group and 12.08 days in the LD group. In the LD group, 18.7% of the patients stayed more than 14 days, but no patients were readmitted. Thirteen clinical factors were statistically different between the two groups. ASA score and LC difficulty were the most predictive risk factors for LOS prolongation. LOS prediction score consisted of eight variables selected from 13 factors; it helped determine the likelihood of whether a patients’ hospital stay was prolonged (sensitivity, 82.1%; specificity, 75.0%). Conclusion: Thirteen factors closely related to hospital stay duration and LOS prediction score could predict the prolongation of a patient’s hospital stay.
Introduction Laparoscopic cholecystectomy (LC) is the established procedure of choice for the treatment of gallbladder disease (1–3). As experience with the operation has grown, its indications have expanded to include progressively more complex and high-risk patients (4–7). With the procedure’s more aggressive application, it is expected that a greater number of complications and prolonged hospital stays will result. Most reports have documented conversion factors of LC to open cholecystectomy, but the prolongation of postoperative hospital stay has not been evaluated (8–10). Furthermore, reports describing clinical risk factors that prolong the hospital stay are rare.
Our objectives were to investigate the preoperative and intraoperative factors that relate to prolongation of hospital stays and to elucidate a length-of-stay (LOS) prediction score that can determine hospital prolongation.
Materials and Methods Between January 2008 and December 2012, 370 consecutive LC were attempted in our department; the patients included 178 men and 192 women aged between 10 and 90 years (mean, 59.3 ± 14.2 years). All patients underwent clinical examination and had preoperative blood tests of liver function and pancreatic enzyme levels. Clinical symptoms, preoperative comorbidities, and ASA score
Asian J Endosc Surg 8 (2015) 289–295 © 2015 Japan Society for Endoscopic Surgery, Asia Endosurgery Task Force and Wiley Publishing Asia Pty Ltd.
289
Predicting prolonged hospital stay
Y Morimoto et al.
were studied. Histories of acute cholecystitis, common bile duct stones (CBDS), and previous abdominal surgery were examined. In imaging diagnosis, ultrasonography, drip intravenous cholangiography-CT (DIC-CT), magnetic resonance cholangiopancreatography (MRCP), and ERCP were selectively performed. ERCP was performed in patients with suspected CBDS. Gallbladder (GB) wall thickness was defined to be more than 5 mm in all imaging studies. Technical difficulty of LC based on the image diagnosis was estimated using the classification established by Ohashi et al (11). Operative information included operation time, blood loss, and performance of an intraoperative cholangiography (IOC). All patients with primary GB diseases underwent LC expect for those with GB cancer, necrotic cholecystitis, and previous upper laparotomy with lymph nodes dissection for malignancy. Before LC, patients with CBDS underwent endoscopic extirpation, which was done endoscopically using a balloon catheter and/or baskettype forceps. Patients with acute cholecystitis underwent emergent LC or GB drainage according to the guidelines for acute cholecystitis in Japan (12). IOC was routinely performed expect in allergic patients. Our institutes’ clinical pathway for LC was applied to all patients. Our clinical pathway, which is similar to that of other institutions in Japan (9), determines the LOS. Namely, discharge is planned for postoperative day 5. According to LOS, two categories of patients were formed: the normal duration (ND) group (LOS ≤5 days) and the long duration (LD) group (LOS ≥6 days). All patients were discharged under the following three conditions in the pathway: (i) the levels of laboratory data were less than twice what they were before LC; (ii) they could return to preoperative lifestyle without disability; and (iii) there was no need for additional treatment of GB disease. In statistical analyses, χ2 test, Fisher’s test, and Student’s t-test were used. Values are expressed as mean ± SD. P < 0.01 was accepted as statistically significant. Multiple regression analysis was used to predict the discrimination prediction system (13). We calculated the regression coefficient and the critical rate (P-value) of each variable, examined the explanatory variable selection criteria (Ru), and then elucidated the discrimination prediction system for identifying prolonged LOS. Receiver operating characteristic was used for evaluating the accuracy of the prediction system. We analyzed the statistics by using MedCalc (MedCalc, Ostend, Belgium) software and Excel software (Microsoft, Redmond, USA) (13).
Results LC was completed in 353 patients (95.4%), with conversion to open cholecystectomy in 17 patients. The ND
290
group had 236 patients and the LD group had 134; all converted patients were included in the LD group. Of the 370 patients, 77.0% had clinical symptoms such as pain, fever, and jaundice, and 64.9% had preoperative comorbidities. The ASA score distribution was as follows: (i) ASA1, 38.4%; (ii) ASA2, 55.1%; and (iii) ASA3, 6.5%. A history of acute cholecystitis and CBDS was noted in 32.4% and 19.2%, respectively. Previous upper and lower abdominal surgery was performed in 3.2% and 18.6%, respectively. Table 1 summarizes the characteristics of the two groups. With regard to previous history and ASA score distribution, a remarkable statistical difference was found between the two groups (P < 0.001). Before LC, 71 patients with CBDS underwent endoscopic extirpation. Preoperative laboratory data are shown in Table 1, and differences in the levels of alkaline phosphatase (ALP) and C-reactive protein (CRP) between the two groups were statistically significant (P < 0.01). In imaging studies, thickened GB wall was present in 39.0% of the ND group and in 62.7% of the LD group (P < 0.001). Table 2 shows the distribution of preoperative clinical diagnoses. The most common admitting diagnosis was GB stones. The rate of acute cholecystitis was 5.9% in the ND group and 21.6% in the LD group (P < 0.0001). The LD group included patients with other complex diseases such as pancreatitis, cholangitis, hepatolithiasis, GB volvulus, intrahepatic abscess, Mirizzi syndrome, and choledocoduodenal fistel. LC difficulty based on the imaging studies showed a significant difference between the two groups (P < 0.0001). With regard to intraoperative factors, statistical differences were found in operation time and blood loss. IOC was performed in 83.1% of the ND group and in 91.0% of the LD group (Table 3). The mean LOS of the ND group was 4.82 ± 0.41 days (range, 3–5 days), whereas that of the LD group was 12.08 ± 10.63 (range, 6–84 days). Figure 1 shows a graph of all patients’ LOS. In the LD group, 18.7% of patients needed to stay more than 14 days after operation. The reasons for prolonged LOS, defined as variances in the clinical pathway, are shown in Table 3. All variances were due to patients-related factors. For example, 39 patients (29.1%) had abnormal aspartate aminotransferase (AST), alanine aminotransferase (ALT), ALP, white blood cell (WBC), or CRP values on postoperative day 5. Clinical symptoms, such as fever or pain, prevented patients from returning to their preoperative lifestyles. Complications with conversion, cholangitis, bleeding, wound infection, bile leakage, or abscess formation delayed discharge. IOC revealed CBDS in five patients, all of whom underwent endoscopic extirpation during their
Asian J Endosc Surg 8 (2015) 289–295 © 2015 Japan Society for Endoscopic Surgery, Asia Endosurgery Task Force and Wiley Publishing Asia Pty Ltd.
Predicting prolonged hospital stay
Y Morimoto et al.
Table 1 Demographic characteristics and laboratory data Variable
ND group (n = 236)
LD group (n = 134)
P-value
Age, mean ± SD (years) Sex, male/female (n) Body length, mean ± SD (cm) Body weight, mean ± SD (kg) BMI, mean ± SD (kg/m2) Clinical symptom, yes/no (n) Preoperative comorbidities,† yes/no (n) ASA score ASA1 ASA2 ASA3 Acute cholecystitis,‡ yes/no (n) Common bile duct stones, yes/no (n) Laparotomy (n) Upper abdomen Lower abdomen No Laboratory data, mean ± SD White blood cell count (/mm3) Total bilirubin (mg/mL) AST (U/L) ALT (U/L) ALP (U/L) LDH (U/L) CRP (mg/mL) Serum amylase (U/L)
56.7 ± 13.2 101/135 (42.8%/57.2%) 160.9 ± 9.3 61.9 ± 12.9 23.8 ± 3.7 167/69 (70.8%/29.2%) 131/105 (55.5%/44.5%)
64.0 ± 14.7 77/57 (57.5%/42.5%) 159.7 ± 9.3 61.2 ± 12.9 23.9 ± 3.9 118/16 (88.1%/11.9%) 109/25 (81.3%/18.7%)
Asian J Endosc Surg ISSN 1758-5902
O R I G I N A L A RT I C L E
Predicting prolonged hospital stay after laparoscopic cholecystectomy Yoshikazu Morimoto, Hitoshi Mizuno, Yusuke Akamaru, Keigo Yasumasa, Hiroshi Noro, Emiko Kono & Yoshio Yamasaki Department of Surgery, Japan Community Healthcare Organization (JCHO), Osaka Hospital, Osaka, Japan
Keywords Laparoscopic cholecystectomy; length of stay; prediction score Correspondence Yoshikazu Morimoto, 4-2-78 Fukushima, Fukushima-ku, Osaka 553-0003, Japan. Tel: +81 6 6441 5451 Fax: +81 6 6445 8900 Email: [email protected] Received 9 November 2014; revised 26 December 2014; accepted 7 February 2015 DOI:10.1111/ases.12183
Abstract Introduction: Widespread application of laparoscopic cholecystectomy (LC) has resulted in a high complication rate and leads to prolonged hospital stays. This study aimed to investigate the preoperative and intraoperative clinical factors that relate to prolongation of hospital stay. Methods: We studied 370 patients who underwent LC for gallbladder disease between 2008 and 2012. Clinical risk factors were retrospectively collected. The clinical pathway for LC was indicated for all patients, and they were divided into two groups according to postoperative length of stay (LOS): the normal duration group (LOS ≤5 days) and the long duration (LD) group (LOS ≥6 days). Multiple regression analysis was used to predict risk factors that identified hospital prolongation to create a LOS prediction score. Results: The normal duration group was 236 patients and the LD group was 134. Seventeen patients (4.6%) required conversion from laparoscopic to open surgery. LOS was 4.82 days in the normal duration group and 12.08 days in the LD group. In the LD group, 18.7% of the patients stayed more than 14 days, but no patients were readmitted. Thirteen clinical factors were statistically different between the two groups. ASA score and LC difficulty were the most predictive risk factors for LOS prolongation. LOS prediction score consisted of eight variables selected from 13 factors; it helped determine the likelihood of whether a patients’ hospital stay was prolonged (sensitivity, 82.1%; specificity, 75.0%). Conclusion: Thirteen factors closely related to hospital stay duration and LOS prediction score could predict the prolongation of a patient’s hospital stay.
Introduction Laparoscopic cholecystectomy (LC) is the established procedure of choice for the treatment of gallbladder disease (1–3). As experience with the operation has grown, its indications have expanded to include progressively more complex and high-risk patients (4–7). With the procedure’s more aggressive application, it is expected that a greater number of complications and prolonged hospital stays will result. Most reports have documented conversion factors of LC to open cholecystectomy, but the prolongation of postoperative hospital stay has not been evaluated (8–10). Furthermore, reports describing clinical risk factors that prolong the hospital stay are rare.
Our objectives were to investigate the preoperative and intraoperative factors that relate to prolongation of hospital stays and to elucidate a length-of-stay (LOS) prediction score that can determine hospital prolongation.
Materials and Methods Between January 2008 and December 2012, 370 consecutive LC were attempted in our department; the patients included 178 men and 192 women aged between 10 and 90 years (mean, 59.3 ± 14.2 years). All patients underwent clinical examination and had preoperative blood tests of liver function and pancreatic enzyme levels. Clinical symptoms, preoperative comorbidities, and ASA score
Asian J Endosc Surg 8 (2015) 289–295 © 2015 Japan Society for Endoscopic Surgery, Asia Endosurgery Task Force and Wiley Publishing Asia Pty Ltd.
289
Predicting prolonged hospital stay
Y Morimoto et al.
were studied. Histories of acute cholecystitis, common bile duct stones (CBDS), and previous abdominal surgery were examined. In imaging diagnosis, ultrasonography, drip intravenous cholangiography-CT (DIC-CT), magnetic resonance cholangiopancreatography (MRCP), and ERCP were selectively performed. ERCP was performed in patients with suspected CBDS. Gallbladder (GB) wall thickness was defined to be more than 5 mm in all imaging studies. Technical difficulty of LC based on the image diagnosis was estimated using the classification established by Ohashi et al (11). Operative information included operation time, blood loss, and performance of an intraoperative cholangiography (IOC). All patients with primary GB diseases underwent LC expect for those with GB cancer, necrotic cholecystitis, and previous upper laparotomy with lymph nodes dissection for malignancy. Before LC, patients with CBDS underwent endoscopic extirpation, which was done endoscopically using a balloon catheter and/or baskettype forceps. Patients with acute cholecystitis underwent emergent LC or GB drainage according to the guidelines for acute cholecystitis in Japan (12). IOC was routinely performed expect in allergic patients. Our institutes’ clinical pathway for LC was applied to all patients. Our clinical pathway, which is similar to that of other institutions in Japan (9), determines the LOS. Namely, discharge is planned for postoperative day 5. According to LOS, two categories of patients were formed: the normal duration (ND) group (LOS ≤5 days) and the long duration (LD) group (LOS ≥6 days). All patients were discharged under the following three conditions in the pathway: (i) the levels of laboratory data were less than twice what they were before LC; (ii) they could return to preoperative lifestyle without disability; and (iii) there was no need for additional treatment of GB disease. In statistical analyses, χ2 test, Fisher’s test, and Student’s t-test were used. Values are expressed as mean ± SD. P < 0.01 was accepted as statistically significant. Multiple regression analysis was used to predict the discrimination prediction system (13). We calculated the regression coefficient and the critical rate (P-value) of each variable, examined the explanatory variable selection criteria (Ru), and then elucidated the discrimination prediction system for identifying prolonged LOS. Receiver operating characteristic was used for evaluating the accuracy of the prediction system. We analyzed the statistics by using MedCalc (MedCalc, Ostend, Belgium) software and Excel software (Microsoft, Redmond, USA) (13).
Results LC was completed in 353 patients (95.4%), with conversion to open cholecystectomy in 17 patients. The ND
290
group had 236 patients and the LD group had 134; all converted patients were included in the LD group. Of the 370 patients, 77.0% had clinical symptoms such as pain, fever, and jaundice, and 64.9% had preoperative comorbidities. The ASA score distribution was as follows: (i) ASA1, 38.4%; (ii) ASA2, 55.1%; and (iii) ASA3, 6.5%. A history of acute cholecystitis and CBDS was noted in 32.4% and 19.2%, respectively. Previous upper and lower abdominal surgery was performed in 3.2% and 18.6%, respectively. Table 1 summarizes the characteristics of the two groups. With regard to previous history and ASA score distribution, a remarkable statistical difference was found between the two groups (P < 0.001). Before LC, 71 patients with CBDS underwent endoscopic extirpation. Preoperative laboratory data are shown in Table 1, and differences in the levels of alkaline phosphatase (ALP) and C-reactive protein (CRP) between the two groups were statistically significant (P < 0.01). In imaging studies, thickened GB wall was present in 39.0% of the ND group and in 62.7% of the LD group (P < 0.001). Table 2 shows the distribution of preoperative clinical diagnoses. The most common admitting diagnosis was GB stones. The rate of acute cholecystitis was 5.9% in the ND group and 21.6% in the LD group (P < 0.0001). The LD group included patients with other complex diseases such as pancreatitis, cholangitis, hepatolithiasis, GB volvulus, intrahepatic abscess, Mirizzi syndrome, and choledocoduodenal fistel. LC difficulty based on the imaging studies showed a significant difference between the two groups (P < 0.0001). With regard to intraoperative factors, statistical differences were found in operation time and blood loss. IOC was performed in 83.1% of the ND group and in 91.0% of the LD group (Table 3). The mean LOS of the ND group was 4.82 ± 0.41 days (range, 3–5 days), whereas that of the LD group was 12.08 ± 10.63 (range, 6–84 days). Figure 1 shows a graph of all patients’ LOS. In the LD group, 18.7% of patients needed to stay more than 14 days after operation. The reasons for prolonged LOS, defined as variances in the clinical pathway, are shown in Table 3. All variances were due to patients-related factors. For example, 39 patients (29.1%) had abnormal aspartate aminotransferase (AST), alanine aminotransferase (ALT), ALP, white blood cell (WBC), or CRP values on postoperative day 5. Clinical symptoms, such as fever or pain, prevented patients from returning to their preoperative lifestyles. Complications with conversion, cholangitis, bleeding, wound infection, bile leakage, or abscess formation delayed discharge. IOC revealed CBDS in five patients, all of whom underwent endoscopic extirpation during their
Asian J Endosc Surg 8 (2015) 289–295 © 2015 Japan Society for Endoscopic Surgery, Asia Endosurgery Task Force and Wiley Publishing Asia Pty Ltd.
Predicting prolonged hospital stay
Y Morimoto et al.
Table 1 Demographic characteristics and laboratory data Variable
ND group (n = 236)
LD group (n = 134)
P-value
Age, mean ± SD (years) Sex, male/female (n) Body length, mean ± SD (cm) Body weight, mean ± SD (kg) BMI, mean ± SD (kg/m2) Clinical symptom, yes/no (n) Preoperative comorbidities,† yes/no (n) ASA score ASA1 ASA2 ASA3 Acute cholecystitis,‡ yes/no (n) Common bile duct stones, yes/no (n) Laparotomy (n) Upper abdomen Lower abdomen No Laboratory data, mean ± SD White blood cell count (/mm3) Total bilirubin (mg/mL) AST (U/L) ALT (U/L) ALP (U/L) LDH (U/L) CRP (mg/mL) Serum amylase (U/L)
56.7 ± 13.2 101/135 (42.8%/57.2%) 160.9 ± 9.3 61.9 ± 12.9 23.8 ± 3.7 167/69 (70.8%/29.2%) 131/105 (55.5%/44.5%)
64.0 ± 14.7 77/57 (57.5%/42.5%) 159.7 ± 9.3 61.2 ± 12.9 23.9 ± 3.9 118/16 (88.1%/11.9%) 109/25 (81.3%/18.7%)
