The prognostic impact of abdominal surgery in cancer patients with neutropenic enterocolitis: a systematic review and meta-analysis, on behalf the Groupe de Recherche en Réanimation Respiratoire du patient d’Onco-Hématologie (GRRR-OH)

Neutropenic enterocolitis (NE) or typhlitis is a serious complication of neutropenia characterized by segmental ulceration and inflammation with necrosis of ileum, cecum and ascending colon [1]. NE was initially described in an autopsy study of children with acute leukemia [2] and evolved to an entity encountered in neutropenic patients [3‐8]. The pathogenesis of NE is poorly understood and probably multifactorial. Immunosuppression induced by neutropenia, combined with chemotherapy toxicity, tumoral infiltration, intramural hemorrhage and inflammatory reaction lead to direct mucosal injury, up to necrotizing damages and microbial translocation. Patients typically present with gastrointestinal (GI) symptoms, in a context of neutropenia, usually following chemotherapy, with bowel wall thickening and positive microbiological samples. Recently, revised diagnostic criteria have been proposed [9]. NE incidence is unknown, reports ranging from 0.8 to 26% [8]. NE carries a poor prognosis, with mortality rates up to 80%, due to complications such as bowel perforation, ischemia, necrosis and septic shock evolution [5, 9, 10].

NE optimal management is controversial, with some advising abdominal surgery [4, 11‐16], and others advocating medical conservative treatment including broad-spectrum antibiotherapy, bowel rest and general supportive care [8, 17, 18]. Physicians are often reluctant to surgery, because of neutropenia and thrombopenia. When surgery is indicated, the question of delaying it until neutropenia resolution arises.

Major advances have been made in the last decade in onco-hematology patients, particularly in the management of septic shock [19, 20], critically ill onco-hematology patients admitted to the intensive care unit (ICU) [21], neutropenic cancer patients [12, 22] and organ failures including acute respiratory failure [23‐27]. Surprisingly, no major improvements have been reported in neutropenic cancer patients presenting with surgical acute abdominal syndrome [28]. Surgical treatment has never been evaluated neither compared to medical treatment, NE being rare, literature scarce and mainly based on small observational reports, case series or case reports. Surgeons and onco-hematologists are usually reluctant to select surgery as the first-choice treatment, concerns being raised regarding the potential risks associated with abdominal surgery during neutropenia, which is furthermore frequently associated with thrombopenia. Nevertheless, no published studies comforted this idea. Moreover, neutropenia is not considered anymore as an unfavorable prognostic factor in critically ill cancer patients, as recently published in a large meta-analysis [22]. Surgery even appeared to be associated with a good prognosis in a recent publication in neutropenic cancer patients with acute abdominal pain [12].

To determine the prognostic impact of abdominal surgery, compared to medical conservative treatment, on short-term mortality of neutropenic cancer patients presenting with NE, we conducted a systematic review and meta-analysis. Secondary objectives were to assess the influence of surgery on outcome in pre-specified subgroups, according to underlying malignancy, period of admission and the presence of neutropenia during the surgery procedure.

The initial search yielded 270 citations, of which 52 were excluded for duplication. Among these records, 129 were excluded as irrelevant to the scope of this review. For the 89 remaining records, abstracts were carefully checked, and 74 full-text articles focusing on NE cancer patients’ management were selected for further evaluation. Articles considered as irrelevant, redundant, with insufficient outcome data reported or less than three patients (including at least one in each treatment arm), or including patients under 18 were excluded. Finally, 20 studies, with a total of 385 patients fulfilled our eligibility criteria and were included (Fig. 1) [9‐11, 35‐50].

This systematic review and meta-analysis, including 385 patients, assessed the prognostic association of abdominal surgery on outcome in neutropenic cancer patients presenting with NE compared to medical conservative treatment. It suggested that surgery was not associated with an increased mortality. According to our results, surgery was not deleterious, regardless of underlying malignancy, time period and the presence of neutropenia at the time of surgery. Interestingly, NE overall mortality was 42.2% (95% IC = 40.2–44.2), which is particularly encouraging compared to the literature from the 1980s. Moreover, recent data supported the good prognosis associated with NE in a large prospective study of critically ill neutropenic cancer patients admitted to the ICU [12].

The optimal management of NE has been a matter of debate [1, 8, 11, 51]. Physicians are frequently reluctant to select surgery as the first-choice treatment in neutropenic patients, based on a potential risk of higher infectious and hemorrhagic complications, although no publications support this idea. Interestingly, an appropriately early indication for appendectomy or cholecystectomy in neutropenic hematology patients was not associated with problematic postoperative course [52, 53]. Similarly, in 85 hematology patients who underwent surgery for acute abdominal complication, neutropenia and thrombopenia were not associated with outcome [54]. Moreover, data obtained in non-cancer patients with thrombocytopenia suggest that even high-risk hemorrhage surgical intervention such as splenectomy carried a low risk of morbidity and mortality [55].

Due to improvements in general supportive care, recent studies reported the success of conservative non-surgical management in most patients diagnosed with NE. It includes immediate broad-spectrum antimicrobial therapy adapted to local microbiological ecology and patients’ colonization [56‐59], general supportive care (intravenous fluid support, parenteral nutrition and nasogastric suction if necessary, platelet transfusions in patients with severe thrombocytopenia, antalgic treatment) and bowel rest [8]. We could not analyze the impact of granulocyte colony stimulating factor (G-CSF) due to insufficient data. Its routine use remains of uncertain benefit and cannot be recommended [60]. Patients should be carefully monitored using repeated imaging to assess bowel wall thickness in addition to clinical response, as relapses can occur [61]. We found that the protective association of abdominal surgery with outcome tended to decrease over time compared to conservative treatment, probably because major advances have been made in the last decade in the medical management of severe sepsis and septic shock [19, 20], management of onco-hematology patients including in the ICU setting [21, 62] and including neutropenic patients [12, 22] and organ failures management [24‐27]. Interestingly, surgery did not become deleterious, whereas medical management improved. Surgical interventions are generally reserved for selected cases of NE based on criteria first proposed by Shamberger et al., including: (a) the persistence of gastrointestinal bleeding despite correction of coagulopathy, thrombocytopenia and neutropenia; (b) free air in the intraperitoneal cavity indicative of bowel perforation; (c) clinical deterioration despite optimal medical management; and (d) the development of other indications for surgery such as appendicitis [63]. However, these criteria have never been evaluated. Another indication should be evaluated, concerning patients with bowel wall thickness greater than 10 mm, who carry a high mortality rate, because they may benefit from a surgical management [38].

Even when the surgery indication is clear, the optimal timing of surgery is debated. For symptomatic septic neutropenic patients, neutropenia recovery represents a high-risk period in which the clinical status is likely to worsen [64]. Waiting for neutropenia resolution remains debated because this approach might expose patients to a septic degradation toward septic shock. Interestingly, Badgwell recently suggested to delay surgery until neutropenia recovery, although he demonstrated in the same publication that surgery was independently associated with a good outcome, regardless of the duration of neutropenia, which appears as a conflicting message [11]. An expert panel from the French Intensive Care Society stated that neutropenia and thrombocytopenia should not modify the timing of surgery in patients with suspicion of digestive tract perforation [16], without any robust publication to rely on. Recent data demonstrated that preoperative septic shock and renal replacement therapy were independently associated with an increased mortality in hematology patients who underwent surgery for an acute abdominal complication [54]. We showed that surgery during the neutropenic period did not modify the prognosis, suggesting that surgery should probably not be delayed. It is important to note that some patients included in the meta-analysis underwent surgery lately at the stage of septic shock and multi-organ failures. Despite these severe situations, abdominal surgery was not associated with an increased mortality, suggesting that the prognostic impact of surgical management may be underestimated. We could not analyze early versus delayed surgical procedures. The influence of an early surgical strategy on outcome deserves to be evaluated, as we know that an early management is associated with a better prognosis [65, 66].

Our results indicated that surgery was not deleterious. Considering that inadequately treated typhlitis carries a high risk of death [6] and that the lack of surgical management was found to be a significant adverse prognostic factor [9, 11], larger indications of abdominal surgery should probably be evaluated. In tricky situations, exploratory laparotomy could probably be performed, as it seems not to be associated with an increased mortality, and represents an effective way to treat NE, perform microbiological samples and remove infectious inoculum. Pathological reports revealed that white laparotomy was uncommon. Infectious documentation is crucial in these patients, as the absence of diagnosis is a well-known adverse prognostic factor [27]. In the absence of microbial diagnosis, the place of empirical antifungal treatment is questionable, at the light of reported microbiological data.

We acknowledge several limitations. The main one is the strength of evidence in the literature concerning NE therapy, which is extremely poor. Available data are limited to low-quality studies, which are all retrospective, single-center, small-sample cohorts, case reports or case series. Moreover, there is a wide heterogeneity in patients, underlying malignancy, neutropenia duration and immunosuppression. There is also a bias in treatment allocation arm according to centers experience and case-volume, surgical indications differing among the studies. The wide admission period did not reflect all recent improvements and results can therefore be influenced. Moreover, study inclusion period was estimated using median inclusion period. This surrogate is, however, imperfect, a few studies being performed over large period. Lastly, several concerns existed with respect to the terminology of NE, because definition criteria evolved over time. It has been shown that clinical impressions are frequently inaccurate, initial clinical diagnosis being correct in only 53% of cases after autopsy or surgery confirmation [10]. Lastly, this study included various types of abdominal surgery, ranging from cholecystectomy to bowel necrosis with peritonitis, with different ranges of severity (no organ dysfunctions to multi-organ failure) prior to surgery, which can represent important cofounder factors.

However, in the absence of prospective studies or large retrospective cohorts, this meta-analysis may represent the best evidence supporting the absence of increased mortality associated with abdominal surgery in neutropenic cancer patients with NE. We do not know whether surgery is superior or comparable to medical treatment, but it did not appear as deleterious. However, surgical therapy can be useful. Delaying surgical therapy due to neutropenia, thrombocytopenia, or other chemotherapy or malignancies associated reasons is not recommended.

These data strengthen the indications of surgical management in the cases of GI or septic complications and question the place of surgery in other cases. These results may lead to conduct future clinical trials, including homogeneous cohorts of patients in terms of abdominal surgery and organ failure severity, in order to determine optimal surgery indications and evaluate the place of early surgical management in this context.

NE is a diagnostic and therapeutic challenge associated with a high mortality rate, with controversial opinions on its optimal management. This systematic review and meta-analysis suggested the absence of excess risk of abdominal surgery on outcome versus conservative medical treatment in neutropenic cancer patients presenting with NE. Major advances have been made in the management of sepsis and supportive care in onco-hematology patients, making medical treatment essential in all cases. However, surgery appeared to be associated with a favorable outcome when indicated. Additional studies are needed to confirm these results and investigate the best indications of surgical treatment.