⃰*Tumor of the head of the pancreas; Cecal tumor; Peritoneal carcinoma.

Discussion

The main difficulties in carrying out this study were the nonsystematic evaluation of patients in pre-anesthetic consultation, the non-performance of ECB of the pus due to the high cost which was an additional burden in almost all surgical site infections and finally the poor collaboration of some patients as part of followup after discharge from hospital. However, we used as much information as we could gather to carry out this study. In our study, the prevalence of MPO was 3.5%. This prevalence is close to that of 4% found by Pearse et al . in 2012 in Europe [4]. On the other hand, it is significantly lower than that of 9.3% reported by Touré et al. in 2012 in the same department [5]. This difference could be explained by the improvement of the technical platform and the continuing training of the nursing staff which took place between our two studies. Advanced age is frequently described as a factor associated with POD. This was confirmed in our study, through multivariate analysis by logistic regression, where the age of 60 years and over which represented 57% of deaths had demonstrated a significant association with MPO and a risk multiplied by 10 (P = 0.001; OR=9.8; CI 2.5 - 37.7). Marzougui et al . found a similar result in 2014 in Tunis in a study of 102 patients with PPO, where age over 60 years was identified as an independent risk factor for mortality (p < 0.001) [6].

Other studies have reported that advanced age was a predictor of MPO from the age of 70 ( Buisson et al. 2013, p=0.003 [7], Moorjani et al. 2013, p=0.519 OR=1.02 (0.97-1.07) [8], Alves et al., 2007, p=0.0001 [9]) We conclude in our study that age greater than or equal to 60 years is a factor associated with PDO . This could be explained by the fragility of the elderly linked to their background (age, underlying pathologies). Although taken into account among the comorbidities classified according to their degree of severity in the ASA score, arterial hypertension is less cited in its association with POD. In a single-center study published in 2013 focusing on early MPO factors in patients who underwent radical surgery for gastric carcinoma, Vural et al. (2013) [10] observed an absence of a statistically significant link between high blood pressure (hypertension) and POD. In our study, of the six hypertensive patients, 2 died, i.e. 14% of deaths. Hypertension was associated with MPO in the univariate analysis (p=0.002) but not in the multivariate analysis . Few authors that we consulted have evaluated the association oftobaccowith the MPO. Eslami et al. in 2016 concluded that there was no statistically significant relationship between tobacco and MPO [11]. This corroborates our result which did not identify tobacco as a factor associated with MPO in multivariate analysis ; but it should be noted that in univariate analysis Tobacco demonstrated a significant association with MPO (p=0.035), which was also the case in the study by Afshar et Al. (p=0.04) [12].

Preparation for surgery is an important part of prognosis. A group of patients with so-called ‘’poor’’ surgical preparation had recorded 26.5% of deaths versus 11.5% of deaths in the group having benefited from so-called ‘’good’’ surgical preparation according to Elias et al. [13]. In this comparison, the association was statistically established between “poor” surgical preparation and MPO (p=0.001). The ASA score ≥ III has often been associated with significant mortality in different studies. In a multicenter retrospective trial conducted by Pearse et al. [4] about 46,539 cases of non-cardiac surgery, an ASA score ≥ III represented an independent risk factor for 30-day POD (p=0.0001) in multivariate analysis . In our study, the ASA score was evaluated in 8 out of 399 patients. , low knowledge of the tool by state nurses acting as anesthetists who were the majority (97.4%). The mortality rate was 100% in the 2 patients classified ASA ≥ III . The six (6) patients classified ASA< III were all alive. Univariate analysis found a significant association between an ASA score ≥ III and MPO (p=0.036). Emergency surgery is one of the most recurrent risk factors for POD in the literature. In multivariate analysis , emergency surgery was found to be an independent predictive factor of POD (with a risk sometimes multiplied by 7) according to the results of Glance et al. (p=0.001) [14], Sapra et al.(p=0.0005) [15], Elias et al. (p=0.001) [13], Slim et al. (p=0.000; OR: 4.3) [16], Alves et al. (p=0.0001; OR: 7.6) [9]. Of the patients in our study, 64.7% (249) were operated on for emergency surgical pathologies. The latter accounted for 71.4% of the deaths observed. The delay in consultation which would complicate the clinical picture of patients could explain this mortality rate from surgical emergencies. To this could be added the lack of pre-anesthetic consultation (78.6%) during which the surgical risk is supposed to be assessed. However, we did not find a statistically significant relationship between emergency surgery and PDO.

acute generalized peritonitis was at the top of the list (28.6%), followed by abdominal tumors (21.4%) as reported by Touré A et al. in 2012 in the same department [5] as well as Doumbouya N et al. in Guinea [17]. In the statistical association of surgical pathologies with MPO in general surgery, tumors were among the most reported. In their retrospective cohort of 1,618 patients aged 80 or over who underwent at least one major surgical procedure, Afshar et al. [12] evaluated the existence of a link between disseminated cancers and POM, which revealed a statistically significant association between the two (p=0.001; OR: 2.10) on multivariate analysis . In the series by Marzougui et al. [12], the mortality rate was significantly higher when surgery was dictated by abdominal neoplastic pathology (p < 0.001). On their side, Vural et al. (p=0.018) [10] reported an absence of significant relationship between mortality after surgery and stage I tumors. This last result corroborates ours with regard to abdominal tumors, but acute generalized peritonitis was associated with MPO, with a 6-fold increased risk (p=0.004; OR: 6.0) in univariate analysis . Neither of these two diagnoses was statistically associated with POD in multivariate analysis.

Toileting the abdominal cavity and resection of loops were the surgical procedures that caused the most deaths in our study, accounting respectively for 4/14 deaths (28.6%), 4/14 deaths (28.6%). Pearse et al. [4] conducted a large multicenter European cohort focusing on 30-day MPO, which established a significant link in univariate analysis between MPO and gastrointestinal resection (p=0.0001). The duodenum and the jejuno -ileal also occupied respectively the 1st and 4th rank of the operated organs causing the most deaths, including the pancreas and the peritoneum in this same study. In our study, there was a significant association between toileting the abdominal cavity and POM in univariate analysis. Loop resection, in multivariate analysis , demonstrated a highly significant link with postoperative death, with an 8-fold increased risk (p=0.023; OR: 8.3; CI 1.3 - 52.4). Loop resection was retained as a factor associated with MPO in our study. Resection of loops during typhoid perforations known for its high mortality could explain this in our context. The association of surgical duration with MPO is variously discussed. There was no statistically established link between MPO and duration of surgical intervention according to Elias et al. [13] in a series of 416 patients operated on in general surgery, but also Vural et al. [10] in multivariate analysis in a single-center cohort of 160 patients operated on for gastric carcinoma (mean duration of surgical intervention = 3 hours).

In contrast, two studies on colorectal surgery demonstrated a significant association between the duration of the surgical procedure and POD. Slim et al. [16] established a statistical link between a mean surgery duration of 2.1 hours and MPO (p=0.000) in univariate analysis , just like Alves et al. [9] who established the link of MPO with an average duration of 2.4 hours. These latter results were superimposable to those of our study in univariate analysis where surgical intervention lasting 2 hours or more was strongly associated with MPO (p=0.000). In multivariate analysis , surgical intervention lasting 2 hours or more was not associated with POD in our study. It was not uncommon to find that the responsibility for surgical interventions was assumed by operators of different academic levels. In our study, 51.9% of patients were operated on by doctors specializing in general surgery versus 48.1% operated on by general surgery specialists (DES). Among the 14 patients who died, 92.9% were operated on by general surgery specialists working in our department compared to 7.1% (1/14) operated on by DES doctors. This could be explained by the fact that the practice of Surgical interventions with high operative risk were reserved for general surgery specialists. In univariate analysis , we found in our study a significant association between MPO and surgical interventions performed by specialists (p=0.001); which was not the case in multivariate analysis.

In our study, infections, particularly sepsis, occupied first place (33.3%) among the postoperative complications encountered (50% in the group of deceased). Septic shock was also the predominant cause of death (57.1%). In univariate analysis , we found a significant association between sepsis and POD. Sepsis was usually the result of a surgical site infection . A larger study by Elias et al. [13] on 416 patients concluded that there was a statistical link between MPO and SSI in general (p=0.001). Let us note here that there was a notable failure in the management of postoperative complications that occurred in our patients. Most of the explorations for the diagnosis of these complications and the adaptation of postoperative therapy were lacking. Our data showed that in our sample, germ isolation was only performed in 3 patients among surgical site infections. In the group of deceased, none of the 7 cases of SSI benefited from a cytobacteriological examination and continued probabilistic antibiotic therapy (unsuitable).

Revision surgery represents an indirect sign of faulty surgery or inadequate patient control in the postoperative period. In a recent study, Launey et al. found a significant relationship between surgical revision and MPO. These authors noted in multivariate analysis that the number of surgical revisions represented a risk factor for mortality (p=0.015) [3]. In our study, 6 patients (1.5%), including 2 who died, underwent surgical revision; the latter did not emerge as a factor associated with MPO. In our study, 92.9% of deaths occurred during custody. This result raises the question of the level of vigilance observed by nursing staff during on-call duty, when we know that this period includes the usual hours of sleep for off-duty staff.

Conclusion

The analysis of mortality factors in this study allowed us to note a high risk of postoperative mortality in patients of advanced age and in those subject to resection of small bowel loops. Optimizing the perioperative management of patients presenting these identified factors could improve the surgical prognosis. Case-control studies or specific cohorts would be necessary to validate these results and identify more factors likely to influence postoperative mortality.

References

1. Dindo D, Demartines N, Clavien PA (2004) Classification of Surgical Complications: A New Proposal With Evaluation in a Cohort of 6336 Patients and Results of a Survey. Ann Surg 240: 205-213.
2. Khan PS, Dar LA, Hayat H (2013) Predictors of mortality and morbidity in peritonitis in a developing country. Ulusal Cer Derg 29: 124-130.
3. Launey Y, Duteurtre B, Larmet R, Nesseler N, Tawa A, et al. (2017) Risk factors for mortality in postoperative peritonitis in critically ill patients. World J Crit Care Med 2017 February 6: 48-55.
4. Pearse RM, Moreno RP, Bauer P, Pelosi P, Metnitz P, et al. (2012) Mortality after surgery in Europe: a 7 day cohort study. Lancet 380: 1059-1065.
5. Toure A, Toure FB, Soumahoro LT, Nabe D, Diakite S, et al. (2012) Perioperative mortality in the general surgery department at the Ignace Deen national hospital, Conakry University Hospital. African Journal of Surgery 2: 23.
6. Marzougui Y (2014) Postoperative peritonitis risk factors for mortality. The Archives of the Pasteur Institute of Tunis 91: 67-76.
7. Buisson LS, Andrieu G, Sanders V, Chalons N, Girardet E, et al. (2014) Perioperative factors predictive of 6-month mortality in elderly subjects undergoing digestive cancer surgery. World J Surg 38: 1531-1541.
8. Moorjani RG, Marchena -Gomez J, Casimiro -Perez J, RoqueCastellano C, Ramirez- Fellipe J (2013) Morbidity- and mortalityrelated prognostic factors of nontraumatic splenectomies . Asian Journal of Surgery 37: 73-79.
9. Alves A, Panis Y, Mantion G, Slim K, Kwiatkowski F, et al. (2007) The AFC Score: Validation of a 4-Item Predicting Score of Postoperative Mortality After Colorectal Resection for Cancer or Diverticulitis: Results of a Prospective Multicenter Study in 1049 Patients. Ann Surg 246: 9196.
10. Vural S, Civil O, Kement M, Altuntas YE (2013) Risk factors for early postoperative morbidity and mortality in patients undergoing radical surgery for gastric carcinoma: A single center experience. International Journal of Surgery 11: 1103-1109.
11. Eslami MH, Rybin DV, Doros G, Farber A (2016) Description of a risk predictive model of 30-day postoperative mortality after elective abdominal aortic aneurysm repair. J Vasc Surg 2016: 1-10.
12. Afshar AH, Virk N, Porhomayon J, Pourafkari L, Dosluoglu HH, et al. (2015) The validity of the VA surgical risk tool in predicting postoperative mortality among octogenarians. The American Journal of Surgery 209: 274-279.
13. Elias ACGP, Matsuo T, Grion CMC, Cardoso LTQ, Verri PH (2009) POSSUM scoring system for predicting mortality in surgical patients. Rev Esc Enferm USP 43: 22-27.
14. Glance LG, Lustik SJ, Hannan EL, Osler TM, Mukamel DB, et al. (2012) The Surgical Mortality Probability Model: Derivation and Validation of a Simple Risk Prediction Rule for Noncardiac Surgery. Ann Surg 255: 696-702.
15. Sapra RL, Mehrotra S, Lalwani S, Mangla V, Mehta N, et al. (2017) Predicting mortality following gastrointestinal surgery. Curr Med Res Pract 8: 8-12.
16. Slim K, Panis Y, Alves A, Kwiatkowski F, Mathieu P, et al. (2006) Predicting Postoperative Mortality in Patients Undergoing Colorectal Surgery. World J Surg 30: 100-106.
17. Guenaga KF, Matos D, Castro AA, Atallah AN, Wille-Jørgensen P (2003) Mechanical bowel preparation for elective colorectal surgery. The Cochrane Database of Systematic Reviews. PubMed 2003: CD001544.