Abstract
Introduction
In recent years, several management options have been used in the management of perforated diverticulitis, ranging from conservative treatment to laparotomy. General surgery has also become increasingly specialised over time. This retrospective cohort study investigated changes in patient outcomes following perforated diverticulitis, management approach and the influence of consultant subspecialisation over time.
Materials and methods
Data was collected on patients admitted with perforated diverticulitis in the North of England between 2002 and 2016. Subspecialisation was categorised as colorectal or other general subspecialties. The primary outcome of interest was overall 30-day mortality; secondary outcomes included surgical approach, stoma and anastomosis rate.
Results
A total of 3394 cases of perforated diverticulitis were analysed (colorectal, n = 1290 and other subspecialists, n = 2104) with a 30-day mortality of 11.6%. There was a significant reduction in mortality over time (2002–2006: 18.6% to 2012–2016: 6.8, P < 0.001).
There was a significant reduction in open surgery (60% to 25.3%, P < 0.001) with increased conservative management (37.4% to 63.5%, P < 0.001), laparoscopic resection (0.1% to 4.9%, P < 0.001) and laparoscopic washout (0.1% to 5.7%, P < 0.001).
Patients admitted under colorectal surgeons had lower mortality than other subspecialists (9.9% vs 12.4%, P = 0.027), which remained significant following multivariate adjustment (hazard ratio 1.44, P = 0.039). These patients had fewer stomas (13.9% vs. 21.0%, P = 0.001) and higher anastomosis rates (22.1% vs 15.8%, P = 0.004).
Conclusion
This study demonstrated considerable improvements in the management of perforated diverticulitis alongside the positive impact of subspecialisation on patient outcomes.
Keywords: Perforated diverticulitis, Outcomes, Management; Subspecialty, Subspecialisation
Introduction
Diverticular disease is a common pathology whose prevalence is increasing throughout the world.1 While the majority of patients remain asymptomatic, diverticulitis and its associated complications can result in significant morbidity and mortality. The most severe presentation of diverticulitis is perforation.
Perforated diverticulitis encompasses a broad spectrum of disease, from localised perforation through to faecal peritonitis. Traditionally, the Hinchey classification was the most widely used grading system for perforated diverticulitis.2 Conventional thought was that, while cases of localised perforation (Hinchey I and II) can be managed conservatively or with percutaneous drainage, more severe cases with generalised peritonitis (Hinchey III or IV) necessitated emergency surgery,1 most commonly in the form of resection and end colostomy (Hartmann’s procedure). A number of recent studies have assessed the efficacy of less invasive management for more severe cases of perforated diverticulitis. In cases of perforated diverticulitis with pneumoperitoneum, studies have reported good outcomes with non-operative management.3–6 Laparoscopic lavage has also been compared with primary resection, for cases of generalised purulent peritonitis (Hinchey III) across a number of meta-analyses.7–10 While some have shown favourable results for laparoscopic lavage,7 others have been inconclusive.8–10 Further debate exists as to whether a primary anastomosis should be performed in cases of generalised peritonitis (Hinchey III or IV). While the results of two published trials favoured primary anastomosis,12,13 the evidence base for this remains inconclusive.14
Subspecialist colorectal surgeons are more likely than other general surgeons to form a primary anastomosis for perforated diverticulitis in the emergency setting.15–17 The third National Emergency Laparotomy Audit (NELA) report highlights that, while colorectal surgeons perform the highest proportion of emergency laparotomies, many other subspecialists still contribute towards emergency on-call rotas.18 There is a paucity of research considering the impact of consultants’ subspecialties within emergency colorectal surgery. Boyce et al saw that the formation and centralisation of a subspecialist colorectal surgery unit led to a significant reduction in operative mortality, increased anastomosis rate and reduced stoma rate.17 Similar findings were described by Biondo et al across all emergency colorectal resection.16 Subspecialist management of emergency surgical admissions is far from common practice however, and more work is needed to establish whether this could improve patient outcomes.
The aim of this study was to evaluate trends in management approach and outcomes for patients with perforated diverticulitis across NHS hospitals in the North of England over a 15-year period. We also assessed how management was influenced by surgical subspecialisation.
Materials and methods
Data were requested for patients admitted with perforated diverticulitis in all acute NHS foundation trusts in the North of England. Caldicott approval was sought and patient data anonymised before being provided to the authors. This information was derived from hospital episode statistics in each trust.19 Data were requested for all emergency admissions under a general surgeon between 1 January 2002 and 31 December 2016. Data fields that were obtained are detailed in Appendix 1 (available online). The data provided enabled calculation of age at admission, day of admission, season of admission, duration of hospital stay, time to procedure from admission, day of procedure and day of in-hospital death. Patients with perforated diverticulitis were identified with the International Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) codes of K57.2, K57.4 and K57.8. Admissions with missing patient characteristics or 30-day in-hospital mortality (0.54% of 490,380 cases) were excluded from the analysis.
Data definitions
The study population included all patients aged 18 years or older admitted as an emergency under a general surgeon in hospitals in the North of England. Calculation of deprivation scores was achieved by conversion of postal codes to Index of Multiple Deprivations scores using the validated online conversion tool.20 ICD-10 diagnostic codes were used to identify relevant comorbidities, which were mapped and converted to Charlson scores using the using weightings employed by the hospital standardised mortality ratio.21–24 To account partially for the severity of patients’ illnesses, a clinical risk grouping was generated by assigning patients’ primary ICD-10 diagnosis to 1 of 259 diagnostic groups using the Clinical Classification software.25 These diagnostic groups were ranked into three equally sized groups based on the 30-day crude in-hospital mortality rates. Primary procedure data were also used to account for any further differences in case mix by additionally adjusting for quartile of risk of procedure in the same manner. Weekends were defined as Saturday, Sunday or a bank holiday. Weekdays were defined as Monday to Friday. Patients were categorised into three discrete five-year groups for analysis of trends over time (2002–2006, 2007–2011, and 2012–2016). The responsible consultant for the relevant hospital spell was provided in the dataset. The subspecialty practised by the consultant during the relevant study period was identified from NHS and subspecialty society websites.26–28 Two categories, colorectal subspecialists and ‘other general subspecialists’, were generated for subspecialty analysis, where ‘other general subspecialists’ consisted of non-colorectal consultants undertaking emergency general surgery, including upper gastrointestinal, breast, vascular and other general surgeons. Trust size was categorised by hospital bed capacity into small/medium or large/very large according to NELA-defined quartiles.29
Management strategies were divided into six groups for analysis: conservative (non-procedural), percutaneous drainage, laparoscopic washout (no resection), laparoscopic resection, laparotomy and resection, and laparotomy without resection. The OPCS classification of interventions and procedures version 4 codes (OPCS-4) that were used to identify procedure type are outlined in Appendix 2 (available online). The primary outcome of interest was in-hospital death within 30 days of admission for conservatively managed patients and 30 days of procedure for patients who underwent any form of operative intervention (including percutaneous drainage and laparoscopic washout). Secondary outcomes included: hospital length of stay, rate of stoma formation, anastomosis and bowel resection. OPCS-4 codes used to identify stoma formation and anastomoses are outlined in Appendix 2 (available online).
Statistical analysis
These retrospective observational data were collected in Microsoft Excel 2010 and analyses were undertaken using Stata® 13.1 software (StataCorp, College Station, TX). Categorical data were summarised using frequencies and percentages, and continuous data using the mean with its 95% confidence interval. These groups were compared using Pearson’s chi squared test for trend for categorical variables, and analysis of variance with post hoc testing for continuous variables. The factors associated with 30-day in-hospital death were determined using a Cox regression model with a time-dependent covariable to describe the hazards associated with admissions with perforated diverticulitis. Factors with P-value less than 0.200 in the univariable models were entered into multivariable models. The multivariable models were built by inclusion of variables that achieved P-value less than 0.050 and significant improvement of model fit (reduction in Akaike’s Information Criterion of at least 4).30,31 Statistical significance was defined in all analyses as P-value less than 0.050.
Results
Change in patient characteristics over time
A total of 3394 cases of perforated diverticulitis were analysed (850 patients in 2002–2006, 1124 in 2007–2011 and 1420 in 2012–2016). A decrease in mean age was observed over time (2002–2006: 66.6 to 2012–2016: 63.5, P < 0.001), alongside a consistent female predominance. There was a significant decrease in clinical (P < 0.001) risk grouping across the study period, with patients increasingly presenting through accident and emergency rather than general practice (P < 0.001; Table 1).
Table 1.
Baseline characteristics of patients presenting with perforated diverticulitis, by study period.
| Characteristic | Study period | P-value | |||
| 2002–2006 (n = 850) | 2007–2011 (n = 1124) | 2012–2016 (n = 1420) | Overall (n = 3394) | ||
| Age (years) mean (95% CI)a | 66.6 (65.6–67.6) | 65.2 (64.3–66.1) | 63.5 (62.7–64.3) | 64.8 (64.3–65.3) | <0.001 |
| Sex, n (%): | 0.048 | ||||
| Male | 341 (40.1) | 500 (44.5) | 568 (40.0) | 1409 (41.5) | |
| Female | 509 (59.9) | 624 (55.5) | 852 (60.0) | 1985 (58.5) | |
| Charlson score, n (%): | 0.016 | ||||
| 0–1 | 707 (83.2) | 899 (80.0) | 1151 (81.1) | 2757 (81.2) | |
| 2–4 | 121 (14.2) | 195 (17.3) | 207 (14.6) | 523 (15.4) | |
| ≥ 5 | 22 (2.6) | 30 (2.7) | 62 (4.4) | 114 (3.4) | |
| Deprivation score, n (%): | 0.017 | ||||
| 1 (most) | 175 (25.7) | 192 (20.3) | 272 (22.8) | 639 (22.7) | |
| 2 | 177 (26.0) | 247 (26.2) | 257 (21.5) | 681 (24.2) | |
| 3 | 123 (18.1) | 180 (19.1) | 224 (18.8) | 527 (18.7) | |
| 4 | 74 (10.9) | 129 (13.7) | 183 (15.3) | 386 (13.7) | |
| 5 (least) | 131 (19.3) | 196 (20.8) | 257 (21.5) | 584 (20.7) | |
| Clinical risk group, n (%): | < 0.001 | ||||
| 1 (lowest) | 105 (12.4) | 140 (12.5) | 168 (11.8) | 413 (12.2) | |
| 2 | 554 (65.2) | 817 (72.7) | 1145 (80.6) | 2516 (74.1) | |
| 3 (highest) | 191 (22.5) | 167 (14.9) | 107 (7.5) | 465 (13.7) | |
| Operative risk group, n (%): | 0.003 | ||||
| 1 (lowest) | 89 (16.7) | 101 (17.6) | 76 (14.6) | 266 (16.4) | |
| 2 | 78 (14.7) | 107 (18.6) | 114 (22.0) | 299 (18.4) | |
| 3 | 231 (43.4) | 270 (47.0) | 219 (42.2) | 720 (44.3) | |
| 4 (highest) | 134 (25.2) | 97 (16.9) | 110 (21.2) | 341 (21.0) | |
| Admission method, n (%): | < 0.001 | ||||
| A&E | 318 (37.4) | 625 (55.6) | 876 (63.9) | 1819 (54.4) | |
| GP | 375 (44.1) | 346 (30.8) | 265 (19.3) | 986 (29.5) | |
| Consultant clinic | 18 (2.1) | 25 (2.2) | 50 (3.6) | 93 (2.8) | |
| Other | 139 (16.4) | 128 (11.4) | 180 (13.1) | 447 (13.4) | |
| Trust size, n (%): | < 0.001 | ||||
| Small/medium | 371 (43.6) | 559 (49.7) | 772 (54.4) | 1702 (50.1) | |
| Large/very large | 479 (56.4) | 565 (50.3) | 648 (45.6) | 1692 (49.4) | |
| Season, n (%): | 0.538 | ||||
| Spring | 223 (26.2) | 285 (25.4) | 357 (25.1) | 865 (25.5) | |
| Summer | 223 (26.2) | 317 (28.2) | 358 (25.2) | 898 (26.5) | |
| Autumn | 206 (24.2) | 278 (24.7) | 353 (24.9) | 837 (24.7) | |
| Winter | 198 (23.3) | 244 (21.7) | 352 (24.8) | 794 (23.4) | |
| Day of admission, n (%): | 0.787 | ||||
| Weekday | 650 (76.5) | 868 (77.2) | 1080 (76.1) | 2598 (76.5) | |
| Weekend/bank holiday | 200 (23.5) | 256 (22.8) | 340 (23.9) | 796 (23.5) | |
| Day of surgery, n (%): | 0.349 | ||||
| Weekday | 378 (78.3) | 458 (79.9) | 447 (76.4) | 1283 (78.2) | |
| Weekend/bank holiday | 105 (21.7) | 115 (20.1) | 138 (23.6) | 358 (21.8) | |
Percentages and proportions were derived by excluding missing data from the variable. Chi square test for difference.
a analysis of variance.
A&E, accident and emergency department; CI, confidence interval; GP, general practitioner.
Change in management and patient outcomes over time
Patient mortality significantly decreased over the study period from 18.6% in 2002–2006 to 6.8% in 2012–2016 (P < 0.001; Table 2). This remained true following adjustment for other covariates with multivariate regression analysis (2002–2006: hazard ratio, HR, 1.00 vs 2012–2016: HR 0.51, P < 0.001; Appendix 6, available online). Overall length of stay also decreased from 21.6 days to 13.7 days, respectively (P < 0.001). Operative approach varied significantly over the 15-year study period, with a significant reduction in the number undergoing open surgery (2002–2006: 60% to 2012–2016: 25.3%, P < 0.001) and consequent rise in alternative management strategies. There was a significant increase in conservative management (37.4% to 63.5%, P < 0.001), laparoscopic resection (0.1% to 4.9%, P < 0.001) and laparoscopic washout (0.1% to 5.7%, P < 0.001). Of those who underwent a resection, there was no change in stoma rate (P = 0.486) or anastomosis rate (P = 0.051) over time (Table 2). Patients who underwent laparoscopic procedures were more than five years younger on average (P < 0.001) and had fewer comorbidities (P = 0.003; Appendix 3, available online). Patients who had laparoscopic procedures had the lowest 30-day mortality (washout 2.8% and resection 5.3%) with open operation the highest (resection 13.6% and no resection 23.3%, P < 0.001; Appendix 4, available online). They were also less likely to have a stoma (P = 0.005). Patients who underwent open operations had the longest length of stay (resection 22.1 days and no resection 24.0 days), with laparoscopic washout the shortest (9.5 days, P < 0.001).
Table 2.
Operative approach and patient outcomes for patients with perforated diverticulitis, comparing differences over time.
| Study period | P-value | ||||
| 2002–2006 (n = 850) | 2007–2011 (n = 1124) | 2012–2016 (n = 1420) | Overall (n = 3394) | ||
| Operative approach, n (%): | |||||
| Open resection | 464 (54.6) | 460 (40.9) | 344 (24.2) | 1268 (37.4) | < 0.001 |
| Open – no resection | 46 (5.4) | 29 (2.6) | 15 (1.1) | 90 (2.7) | < 0.001 |
| Laparoscopy with washout | 1 (0.1) | 27 (2.4) | 81 (5.7) | 109 (3.2) | < 0.001 |
| Laparoscopic resection | 0 (0.0) | 10 (0.9) | 47 (3.3) | 57 (1.7) | < 0.001 |
| Conservative | 318 (37.4) | 549 (48.8) | 901 (63.5) | 1768 (52.1) | < 0.001 |
| Percutaneous drain | 21 (2.5) | 49 (4.4) | 32 (2.3) | 102 (3.0) | 0.005 |
| Overall 30-day mortality | |||||
| Deaths, n (%) | 158 (18.6) | 135 (12.0) | 96 (6.8) | 389 (11.5) | < 0.001 |
| 30-day mortality by operative approach, n (%): | |||||
| Open resection | 83 (17.9) | 54 (11.7) | 36 (10.5) | 173 (13.6) | 0.003 |
| Open – no resection | 14 (30.4) | 4 (13.8) | 3 (20.0) | 21 (23.3) | 0.239 |
| Laparoscopy with washout | 0 (0.0) | 0 (0.0) | 3 (3.7) | 3 (2.8) | 0.587 |
| Laparoscopic resection | n/a | 1 (10.0) | 2 (4.3) | 3 (5.3) | 0.460 |
| Conservative | 59 (18.6) | 70 (12.8) | 50 (5.5) | 179 (10.1) | < 0.001 |
| Percutaneous drain | 2 (9.5) | 6 (12.2) | 2 (6.3) | 10 (9.8) | 0.674 |
| Length of hospital stay, mean days (95% CI):a | 21.6 (19.8–23.4) | 17.1 (15.9–18.4) | 13.7 (12.8–14.5) | 16.8 (16.1–17.5) | < 0.001 |
| Stoma, n (%):b | 0.486 | ||||
| No | 83 (17.9) | 77 (16.4) | 58 (14.8) | 218 (16.5) | |
| Yes | 381 (82.1) | 393 (83.6) | 333 (85.2) | 1107 (83.5) | |
| Anastomosis, n (%):b | 0.051 | ||||
| No | 364 (78.4) | 394 (83.8) | 328 (83.9) | 1086 (82.0) | |
| Yes | 100 (21.6) | 76 (16.2) | 63 (16.1) | 239 (18.0) | |
Percentages and proportions were derived by excluding missing data from the variable. Chi square test for difference. CI, confidence interval.
a Analysis of variance.
b Values calculated from patients who underwent an open resection or laparoscopic resection; patients who underwent another management approach excluded.
Following multivariate Cox regression, increasing age, Charlson score (comorbidity), clinical risk group and operative risk group, with admission on a weekend/bank holiday, were identified as independent predictors of 30-day mortality (Appendix 6, available online).
Effect of consultant specialty on management and patient outcomes
An increasing number of patients were managed by a colorectal subspecialist (2002–2006: 27.1% to 2012-16: 46.3%, P < 0.001; Appendix 5, available online). There was no significant difference in age, sex, Charlson score or clinical risk group between patients based on the subspecialty of the responsible consultant. However, patients under other general subspecialties were in a higher operative risk group (P < 0.001). Patients under a colorectal subspecialist had a significantly lower overall 30-day mortality than their other general surgical colleagues (9.9% vs 12.4%, P = 0.027; Table 3). Multivariate regression confirmed that management by a colorectal subspecialist is an independent predictor of improved 30-day mortality (colorectal: HR 1.00 vs other subspecialists: HR 1.44, P = 0.039; Appendix 6, available online). Colorectal subspecialists used percutaneous drains (P < 0.001) and laparoscopy with washout (P = 0.021) more frequently than their colleagues. Primary open resection was more commonly performed by other general surgeons (P = 0.021). When focusing on patients who underwent an open operation or laparoscopic resection, colorectal surgeons were also more likely to form a primary anastomosis (22.1% vs. 15.8%, P = 0.004) and less likely to form a stoma (21.0% vs. 13.9%, P = 0.001; Table 3).
Table 3.
Operative approach and patient outcomes for patients with perforated diverticulitis, comparing consultant subspecialty.
| Consultant subspecialty | ||||
| Colorectal (n = 1291) | Other (n = 2103) | Overall (n = 3394) | P-value | |
| Operative approach, n (%): | ||||
| Open resection | 450 (34.9) | 818 (38.9) | 1268 (37.4) | 0.021 |
| Open – no resection | 40 (3.1) | 50 (2.4) | 90 (2.7) | 0.204 |
| Laparoscopy with washout | 53 (4.1) | 56 (2.7) | 109 (3.2) | 0.021 |
| Laparoscopic resection | 26 (2.0) | 31 (1.5) | 57 (1.7) | 0.235 |
| Conservative | 665 (51.5) | 1103 (52.4) | 1768 (52.1) | 0.595 |
| Percutaneous drain | 57 (4.4) | 45 (2.1) | 102 (3.0) | < 0.001 |
| Overall 30-day | 128 (9.9) | 261 (12.4) | 389 (11.5) | 0.027 |
| 30-day mortality by operative approach, n (%): | ||||
| Open | 56 (12.4) | 117 (14.3) | 173 (13.6 | 0.356 |
| Open – no resection | 4 (10.0) | 17 (34.0) | 21 (23.3) | 0.007 |
| Laparoscopy with washout | 2 (3.8) | 1 (1.8) | 3 (2.8) | 0.526 |
| Laparoscopic resection | 2 (7.7) | 1 (3.2) | 3 (5.3) | 0.452 |
| Conservative | 57 (8.6) | 122 (11.1) | 179 (10.1) | 0.093 |
| Percutaneous drain | 7 (12.3) | 3 (6.7) | 10 (9.8) | 0.344 |
| Mortality, n (%): | ||||
| 2002–2006 | 36 (15.7) | 122 (19.7) | 158 (18.6) | 0.180 |
| 2007–2011 | 49 (12.1) | 86 (11.9) | 135 (12.0) | 0.927 |
| 2012–2016 | 43 (6.5) | 53 (6.9) | 96 (6.8) | 0.764 |
| Total length of hospital stay, mean days (95% CI)a | 16.0 (14.8–17.2) | 17.3 (16.4–18.2) | 16.8 (16.1–17.5) | 0.086 |
| Time to procedure, mean days (95% CI)a | 3.4 (2.7–4.1) | 3.4 (3.0–3.9) | 3.4 (3.1–3.8) | 0.908 |
| Postoperative length of stay, mean days (95% CI)a | 13.5 (12.4–14.6) | 14.1 (13.2–15.1) | 13.9 (13.1–14.6) | 0.404 |
| Stoma, n (%):c | 0.001 | |||
| No | 100 (21.0) | 118 (13.9) | 218 (16.5) | |
| Yes | 376 (79.0) | 731 (86.1) | 1107 (83.5) | |
| Anastomosis, n (%):c | 0.004 | |||
| No | 371 (77.9) | 715 (84.2) | 1086 (82.0) | |
| Yes | 105 (22.1) | 134 (15.8) | 239 (18.0) | |
Percentages and proportions were derived by excluding missing data from the variable. Chi square test for difference. CI, confidence interval.
a Analysis of variance.
b Values calculated from patients who underwent an open resection or laparoscopic resection; patients who underwent another management approach excluded.
Discussion
This study has shown a significant decline in mortality for patients with perforated diverticulitis over the past 15 years in the North of England. Patients presenting with perforated diverticulitis in 2012–2016 were almost half as likely to die when compared with patients in 2002–2006 and were also spending a significantly shorter time in hospital. These improvements are in keeping with the wider trends seen within emergency surgery,18,32 and it is likely that many factors have contributed towards this change. The high mortality seen in emergency general surgery has led to increasing focus and drive for improvement, not only in operative practice, but also in peri- and postoperative care.18
Less invasive management strategies are increasingly used in perforated diverticulitis. The number of patients undergoing an open operation has decreased considerably between 2002–2006 (60.0%) and 2012–2016 (25.3% P < 0.001). We also observed a significant increase in the number of patients undergoing conservative treatment (2002–2006 37.4%, 2012–16 63.5%, P < 0.001). This is in keeping with a growing body of evidence demonstrating that even severe cases of perforated diverticulitis can be managed conservatively.7–10 Over the studied time period, the number of patients with perforated diverticulitis steadily increased. Rather than attributing these figures to increasing incidence, this is probably reflective of improved imaging availability. There is good evidence to show the increased availability and use of computed tomography in hospital,33,34 which is likely to lead to increased detection of more minor perforations that may otherwise not have been diagnosed. Our results highlighted significant increases in the use of laparoscopic washout (P < 0.001) and resection (P < 0.001). However, as a proportion of the overall cohort, this group remained relatively small (2002–2006: 0.1%, 2012–2016: 9%) Good evidence has been reported for the potential benefits of laparoscopic colorectal surgery in the emergency setting.35 The increase in use of laparoscopic lavage is perhaps more surprising, given that efficacy of the procedure continues to be a contentious issue.7–10
We identified that patients who underwent a laparoscopic, rather than an open procedure had an improved mortality and reduced length of hospital stay. They were also less likely to have a stoma (P = 0.005). However, this is potentially reflective of a favourable patient group, with these patients being younger (P < 0.001) with fewer comorbidities (P = 0.003) and there being a larger proportion of patients in the lowest clinical risk group (washout 5.5% and resection 8.8%).
In patients who underwent an open operation or laparoscopic resection, we did not identify a trend in stoma (P = 0.486) or anastomosis (P = 0.051) rates. This is probably due to a combination of factors. The safety of primary anastomosis in generalised peritonitis is still under debate,12–14 and less severe cases, which were previously anastomosed, are now being managed conservatively.
Our findings identified that cases of perforated diverticulitis managed by a colorectal subspecialist had significantly lower 30-day mortality than patients under the care of other general subspecialists (9.9% vs 12.4%, P = 0.027). This remained true following adjustment for other covariates with multivariate regression, thus suggesting colorectal subspecialist care to be an independent predictor of survival. Patients under the care of colorectal surgeons were also found to be less likely to have a stoma formed (P = 0.001) with increased rates of primary anastomosis (P = 0.004). These findings are comparable with those of Boyce et al, who evaluated patient outcomes before and after centralisation of a subspecialist colorectal surgery unit.17 The group identified a significant reduction in operative mortality with increased anastomosis rate and reduced stoma rate.17 Biondo et al also described similar findings across all emergency colorectal resection.16 A smaller multicentre study also found that colorectal subspecialists were more likely to form a primary anastomosis. Following multivariate analysis, however, surgical specialty was found not to be an independent risk factor for 30-day postoperative mortality, postoperative surgical or medical adverse events.21 We would postulate that the experience gained by colorectal subspecialists through their elective work is probably driving the improved outcomes in these higher-risk emergency patients. Many non-colorectal subspecialists may lack in confidence and practice with formation of primary anastomosis, especially in more complex cases of perforation. This difference may also relate to experience with postoperative care, enhanced recovery protocols and other interventions that can impact on a patient’s outcome.
The hospital episode statistics data used in this work rely on accurate clinical coding, which is in turn dependent on good documentation from medical staff. In large datasets such as this, small individual inaccuracies in coding are likely to be insignificant and not influential towards overall findings. Admissions are coded by consultant episode of care, rather than operative consultant, thus limiting the conclusions that can be drawn on consultants’ operative influence. In a small number of cases, a separate consultant may have operated on the patient during their admission. We would suggest, however, that any possible bias this would introduce would merely dull any observed effect. More complex patients are more likely to be referred to their parent specialty (i.e. colorectal) rather than vice versa. The handover of patients between on-call teams due to time pressure is likely to be negligible and consistent through the study period. The availability of computed tomography reports defining exact Hinchey grade would have contributed to this study, however the use of hospital episode statistics data prohibits classifying the extent of the perforated diverticulitis. We have attempted to partially account for the likely rise in less severe cases of perforated diverticulitis in our analyses through the use of clinical and operative risk grouping. These potential confounders were also included in our multivariate regression analysis.
Conclusions
This study has demonstrated considerable changes and improvements in the management of perforated diverticulitis. Patients are almost half as likely to die from perforated diverticulitis than 15 years ago. We have highlighted the positive impact subspecialisation can have on patient outcomes. Patients under the care of a colorectal surgeon were found to be less likely to have a stoma, more likely to have an anastomosis and more likely to survive. We postulate that continued subspecialisation in the emergency setting will further improve outcomes.
References
- 1.Sartelli M, Catena F, Ansaloni L et al. . WSES Guidelines for the management of acute left sided colonic diverticulitis in the emergency setting. World J Emerg Surg 2016; (1): 1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hinchey EJ, Schaal PG, Richards GK. Treatment of perforated diverticular disease of the colon. Adv Surg 1978; : 85–109. [PubMed] [Google Scholar]
- 3.Costi R, Cauchy F, Le Bian A et al. . Challenging a classic myth: pneumoperitoneum associated with acute diverticulitis is not an indication for open or laparoscopic emergency surgery in hemodynamically stable patients: a 10-year experience with a nonoperative treatment. Surg Endosc 2012; (7): 2,061–2,071. [DOI] [PubMed] [Google Scholar]
- 4.Dharmarajan S, Hunt SR, Birnbaum EH et al. . The efficacy of nonoperative management of acute complicated diverticulitis. Dis Colon Rectum 2011; (6): 663–671. [DOI] [PubMed] [Google Scholar]
- 5.Sallinen VJ, Mentula PJ, Leppäniemi AK. Nonoperative management of perforated diverticulitis with extraluminal air is safe and effective in selected patients. Dis Colon Rectum 2014; (7): 875–881. [DOI] [PubMed] [Google Scholar]
- 6.Titos-García A, Aranda-Narváez JM, Romacho-López L et al. . Nonoperative management of perforated acute diverticulitis with extraluminal air: results and risk factors of failure. Int J Colorectal Dis 2017; (10): 1,503–1,507. [DOI] [PubMed] [Google Scholar]
- 7.Angenete E, Bock D, Rosenberg J, Haglind E. Laparoscopic lavage is superior to colon resection for perforated purulent diverticulitis-a meta-analysis. Int J Colorectal Dis 2017; (2): 163–169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Galbraith N, Carter JV, Netz U et al. . Laparoscopic lavage in the management of perforated diverticulitis: a contemporary meta-analysis. J Gastrointest Surg Off J Soc Surg Aliment Tract 2017; (9): 1,491–1,499. [DOI] [PubMed] [Google Scholar]
- 9.Ceresoli M, Coccolini F, Montori G et al. . Laparoscopic lavage versus resection in perforated diverticulitis with purulent peritonitis: a meta-analysis of randomized controlled trials. World J Emerg Surg 2016; (1): 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cirocchi R, Di Saverio S, Weber DG et al. . Laparoscopic lavage versus surgical resection for acute diverticulitis with generalised peritonitis: a systematic review and meta-analysis. Tech Coloproctol 2017; (2): 93–110. [DOI] [PubMed] [Google Scholar]
- 11.Kohl A, Rosenberg J, Bock D et al. . Two-year results of the randomized clinical trial DILALA comparing laparoscopic lavage with resection as treatment for perforated diverticulitis. Br J Surg 2018; (9): 1,128–1,134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Oberkofler CE, Rickenbacher A, Raptis DA et al. . A multicenter randomized clinical trial of primary anastomosis or Hartmann’s procedure for perforated left colonic diverticulitis with purulent or fecal peritonitis. Ann Surg 2012; (5): 819–827. [DOI] [PubMed] [Google Scholar]
- 13.Bridoux V, Regimbeau JM, Ouaissi M et al. . Hartmann’s procedure or primary anastomosis for generalized peritonitis due to perforated diverticulitis: a prospective multicenter randomized trial (DIVERTI). J Am Coll Surg 2017; (6): 798–805. [DOI] [PubMed] [Google Scholar]
- 14.Cirocchi R, Afshar S, Shaban F et al. . Perforated sigmoid diverticulitis: Hartmann’s procedure or resection with primary anastomosis: a systematic review and meta-analysis of randomised control trials. Tech Coloproctol 2018; (10): 743–753. [DOI] [PubMed] [Google Scholar]
- 15.Constantinides VA, Tekkis PP, Senapati A. Prospective multicentre evaluation of adverse outcomes following treatment for complicated diverticular disease. Br J Surg 2006; (12): 1,503–1,513. [DOI] [PubMed] [Google Scholar]
- 16.Biondo S, Kreisler E, Fraccalvieri D et al. . Impact of surgical specialization on emergency colorectal surgery outcomes. Arch Surg 2010; (1): 79–86. [DOI] [PubMed] [Google Scholar]
- 17.Boyce SA, Bartolo DCC, Paterson HM. Subspecialist emergency management of diverticulitis is associated with reduced mortality and fewer stomas. Color Dis 2013; (4): 442–447. [DOI] [PubMed] [Google Scholar]
- 18.NELA Project Team Third Patient Report of the National Emergency Laparotomy Audit (NELA): December 2015 to November 2016. London: Royal College of Anaesthetists; 2017. [Google Scholar]
- 19.NHS Digital. Hospital Episode Statistics. https://digital.nhs.uk/data-and-information/publications/statistical/hospital-episode-statistics-for-admitted-patient-care-outpatient-and-accident-and-emergency-data (cited May 2019).
- 20.Ministry of Housing, Communities and Local Government English indices of deprivation 2015. Area lookup. http://imd-by-geo.opendatacommunities.org/area (cited May 2019).
- 21.Clinical Indicators Team Analysis of the Impact of Deprivation on the Summary Hospital-level Mortality Indicator (SHMI). London: Health and Social Care Information Centre; 2014. p. 6. [Google Scholar]
- 22.Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol 1994; (11): 1,245–1,251. [DOI] [PubMed] [Google Scholar]
- 23.Quan H, Li B, Couris CM, Fushimi K et al. . Updating and validating the Charlson Comorbidity Index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol 2011; (6): 676–682. [DOI] [PubMed] [Google Scholar]
- 24.Aylin P, Bottle A, Jen MH, Middleton S. Summary Hospital Mortality Indicator (HSMR). Mortality Indicators 2009. https://www1.imperial.ac.uk/resources/3321CA24-A5BC-4A91-9CC9-12C74AA72FDC (cited May 2019).
- 25.Ruiz M, Bottle A, Aylin PP. The Global Comparators project: International comparison of 30-day in-hospital mortality by day of the week. BMJ Qual Saf 2015; (8): 492–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.NHS Find hospital – consultants. https://www.nhs.uk/Service-Search/Hospital/LocationSearch/8/Consultants (cited May 2019).
- 27.AUGIS Find an AUGIS specialist. http://www.augis.org/find-a-specialist (cited May 2019).
- 28.Association of Coloproctology of Great Britain and Ireland Find a surgeon. https://www.acpgbi.org.uk/patients/surgeon-directory (cited May 2019).
- 29.NELA Project Team Figure 19.1 Achievement of key processes in each hospital. NELA Standalone RAG Table. file:///Users/janemoody/NELA%20Patient%20Audit%202018%20-%20Standalone%20RAG%20Table.pdf (cited May 2019).
- 30.Freemantle N, Richardson M, Wood J et al. . Weekend hospitalization and additional risk of death: an analysis of inpatient data. J R Soc Med 2012; (2): 74–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Freemantle N, Richardson M, Wood J et al. . Can we update the Summary Hospital Mortality Index (SHMI) to make a useful measure of the quality of hospital care? An observational study. BMJ Open 2013; (1): e002018–e002018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Barrow E, Anderson ID, Varley S et al. . Current UK practice in emergency laparotomy. Ann R Coll Surg Engl 2013; (8): 599–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Brenner DJ, Hall EJ. Computed tomography: an increasing source of radiation exposure. N Engl J Med 2007; (22): 2,277–2,284. [DOI] [PubMed] [Google Scholar]
- 34.Bellolio F, Heien H, Sangaralingham L et al. . Increased computed tomography utilization in the emergency department and its association with hospital admission. West J Emerg Med 2017; (5): 835–845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Harji DP, Griffiths B, Burke D, Sagar PM. Systematic review of emergency laparoscopic colorectal resection. Br J Surg 2014; (1): 126–133. [DOI] [PubMed] [Google Scholar]