Splenic injuries secondary to colonoscopy: Rare but serious complication

Waqas Ullah, Mamoon Ur Rashid, Asif Mehmood, Yousaf Zafar, Ishtiaq Hussain, Deepika Sarvepalli,Muhammad Khalid Hasan

Abstract

Key words: Colonoscopy; Spleen; Splenic rupture; Systematic review; Splenic injuries

INTRODUCTION

Colonoscopy is a safe and routine procedure performed for evaluation of large bowel diseases. The most common post procedure complications include bleeding and perforation. Splenic Injury, a rare complication after colonoscopy was first reported by Wherry and Zehner[1]in 1974. Since then, the reported incidence is on the rise and is estimated to be 1 in 100000 procedures. Possible reasons include increased utilization of colonoscopy for management of gastrointestinal disorders including polypectomy and increased rate of screening colonoscopies from 38% in 2000 to 59%in 2013 in adults ≥ 50 years[2]. Previously reported risks of splenic injury include female sex, advanced age, history of abdominal/pelvic surgeries, anti-platelet or anticoagulant therapy and polypectomies/biopsies performed during colonoscopies, but it is unclear if these factors are purely associations or represent a contributing cause[3].

Even though, splenic injury is a rare complication, it is associated with significantly high mortality rate (5%)[1,4-6]. In order to ensure the timely management of splenic injuries there should be a high index of suspicion of splenic trauma in any patient presenting with severe abdominal pain and hemodynamic instability after colonoscopy. Here, we performed a systematic review to identify the risks, diagnostic factors, clinical course and management of splenic injury with colonoscopy that may help predict splenic complications after colonoscopy.

MATERIALS AND METHODS

Search strategy and selection criteria

A literature search for relevant articles was performed through April 25, 2019, using MEDLINE (PubMed, Ovid), Embase and Cochrane databases (Figure 1). There was no language or time restrictions placed on the search. The search strategies included various combinations of text-words and medical subject headings (MeSH) to generate two subsets of citations: One for spleen injury, using the MeSH and terms like “spleen rupture”, “spleen injury”, “spleen laceration”, “spleen trauma”, “spleen insult”,“spleen” and “spleen disease” and the other for colonoscopy using terms and MeSH like “colonoscopy”, “endoscopy”, “sigmoidoscopy”, “rigid colonoscopy”, “flexible endoscopy”, “rigid endoscopy”, “flexible colonoscopy”. The terms from the two subsets were combined in 1:1 combination and finally results from all the possible combinations were downloaded. Based on our research question, we also manually searched the references in all known articles to identify studies that were missed by the initial search.

The selection criteria for the included studies was: (1) Recruited subjects with any type of spleen injury secondary to colonoscopy; and (2) Discussed the mechanism,diagnosis and outcome of spleen injury. Studies with insufficient data, posters and conference papers were excluded, as were studies with not enough description of its subjects.

Study selection

The titles and abstracts of the selected articles were reviewed independently by three authors and the articles which met the inclusion criteria were reviewed by the fourth author. Full-text articles that were potentially relevant to the study were also reviewed by all the four authors to confirm the eligibility. Disagreements were resolved by mutual consensus and after a detailed group discussion.

Data abstraction and analysis

Three reviewers extracted data[3,4,7-63](Table 1) into the excel sheet identifying the reason of colonoscopy, presentation of patient with spleen injury, types of injury,diagnosis, management and outcome of all patients with spleen injury. After carefully assessing the extracted data, all categorical data was coded into binary numbers.Frequencies and proportions were obtained for nominal and ordinal data while means and standard deviation was calculated for continuous data. Inferential statistics for categorical data was performed using Pearson Chi-square test and the alpha criterion for significance was set at a value less than 0.05. Analysis was performed using SPSS v22.

RESULTS

The mean age of the patients was 62.7 (Standard deviation 15.7 years). Of the reported cases of splenic injury majority were females 64% (n= 41/64), there was no gender data available for four patients. Twenty two percent of the patients (n= 15/68) had a complete rupture of the spleen secondary to colonoscopy while the remaining 63% (n= 43/68) had a variety of spleen injuries such as subcapsular hematoma, spleen laceration and spleen avulsion. Details on the type of spleen injury was not available in 15% patients (n= 10/68). The diagnosis of any kind of spleen insult was made using the computed tomography (CT) in all patients (98.5%,n= 67/68) except one patient who was diagnosed with CT angiography.

The most common reason for colonoscopy was screening for colon cancer 46% (n=31/68) followed by evaluation for the cause of gastrointestinal bleeding 28% (n=19/68). Other rationales for colonoscopy were polypectomy (13%,n= 9/68),abdominal pain (3%,n= 2/68) and weight loss (1.5%,n= 1/68). There was no mention of indication of colonoscopy in 6 (9%) (Table 2).

The most common clinical presentation of splenic insult was abdominal pain either generalized or in the left hypochondrium in 87% of patients (n= 59/68). Other minor presentations included dizziness, syncope, back pain, chest pain and symptoms of anemia (Table 2, Figure 2).

Figure 1 Flow chart for literature search.

Management in all cases was tailored according to the clinical presentation and CT scan findings. Majority of the patients required laparotomy and splenectomy 47% (n=32/68). While patients with minor spleen hematomas and lacerations were managed conservatively with intravenous fluids, pain control and vitals monitoring 38% (n=26/68). Interestingly, 6% (n= 4/68) patients were managed with proximal splenic artery splenic embolization (PSAE) and 4% (n= 3/68) were managed with laparoscopy. There was no outcome reported in 4.4% (n= 3/68) patients.

Despite the wide array of spleen insults ranging from mild hematoma to rupture,majority of patients survived with different management modalities. The overall mortality rate however, was significant at 10% (n= 7/68) while 77% (n= 52/68) had successful recovery. There was no outcome reported in 13% (n= 9/68) patients (Table 3).

The reason of colonoscopy against presentation specifically abdominal pain showed no statistical significanceP= 0.69. For all patients with spleen rupture and injury the most common reason was screening colonoscopy followed by colonoscopy done for gastrointestinal bleeding evaluation. This association between screening colonoscopy and spleen injury was analyzed by likelihood ratio as the number of expected count in more than 20% of the cells for more than 2 variables was less than 5.The association was robust with a value of 2.3, degree of freedom (df) of 4 but not statistically significant (P0.667). Similarly, Chi-Square analysis and cross tabulation for the management of spleen injury was tested for any association with reason of colonoscopy. The likelihood ratio was interpreted based on the same rationale as mentioned above, there was no significant association found between the all type of colonoscopy and the management of spleen injury (splenectomyP= 0.89 and splenorrhaphy = 0.91). The alpha criteria of less than 0.05 was taken as a cutoff for significance (Figure 3).

To assess the association of outcome (survival/mortality) with reason of colonoscopy we cross tabulated these groups and did a chi-square test. The likelihood ratio was 3.17,df4 andPvalue was not significant (screening colonoscopy 0.52,diagnostic colonoscopy 0.61) indicating there was no association between the two variables (Table 4).

We compared the type of spleen injury and its association with the type of management performed. It is interesting to note that 74% of the spleen rupture patients were managed with splenectomy through exploratory laparotomy. About 45% for spleen injury group had splenectomy, an equal number of patients (45%)were managed with conservative management. These results are indicating that laparotomy was a common management strategy in both groups. Only 14% patientsin spleen rupture group had conservative management and PSAE due to multiple comorbidities or surgery denial. The likelihood ratio for splenectomy association with spleen rupture was 8.1,df3 and thePvalue was significant at 0.04 (< 0.05). The strength of association between these variables was was moderately strong at at a cramer’s V 0.34.

Table 1 Details of case reports and Clavein-Dindo Classification of Post-Operative complication (computed tomography)

Han et al[34] 77 Male Polypectomy Splenic rupture CT Laparotomy/splenectomy Grade III Brennan et al[38] 75 Female Screening Splenic injury/hematoma CT Proximal splenic artery embolization Grade III Herreros de Tejada et al[39]65 Female Submucosal Dissection of tumor Splenic rupture CT Laparotomy/splenectomy Grade III Lauretta et al[40] 61 Male Screening Splenic rupture CT Laparotomy/splenectomy Grade III Jamorabo et al[42] 63 Male Screening Splenic rupture CT Laparotomy/splenectomy Grade III Guner et al[44] 53 Male Abdominal Pain Splenic injury/hematoma CT Bed rest, blood transfusion,observation Grade II McBride et al[45] 64 Female Bleeding Splenic injury/hematoma CT Laparotomy/splenectomy Grade III Rex et al[46] 68 Female NA NA CT Conservative management Grade I Rex et al[46] 52 Female Polypectomy Splenic injury/hematoma CT Conservative management Grade I Rex et al[46] 61 Female NA NA CT Conservative management Grade I Rex et al[46] 58 Female NA NA CT Conservative management Grade I Rex et al[46] 74 Female NA NA CT Conservative management Grade I Rex et al[46] 85 Female NA NA CT Conservative management Grade I Chow et al[47] 84 Female Bleeding Splenic injury/hematoma CT Fluids/blood/splenic artery embolization Grade III Pineda et al[48] 62 Female Screening Splenic rupture CT Laparotomy/splenectomy Grade III Corcillo et al[49] 66 Male Bleeding Splenic rupture CT Proximal splenic artery embolization Grade III Corcillo et al[49] 77 Female Polypectomy Splenic rupture CT Proximal splenic artery embolization Grade III Abunnaja et al[50] 62 Female Screening Splenic injury/hematoma CT Ringers lactate,blood,splenectomy,vaccinations Grade II González-Soler et al[51]67 Male Screening Splenic rupture CT Laparotomy/splenectomy Grade III Seifman et al[52] 41 Female Screening Splenic injury/hematoma CT Laparotomy/splenectomy Grade III Singla et al[54] NA NA Screening NA CT NA NA Shankar et al[56] 47 Female Bleeding Splenic injury/hematoma CT Laparotomy/splenectomy Grade III Fishback et al[58] 64 Female Screening Splenic injury/hematoma CT Conservative management Grade I Fishback et al[58] 59 Female Bleeding Splenic injury/hematoma CT Conservative management Grade I Fishback et al[58] 55 Female Bleeding Splenic injury/hematoma CT Laparotomy/splenectomy Grade III Fishback et al[58] 65 Male Polypectomy NA CT Laparotomy/splenectomy Grade III Fishback et al[58] 56 Female Screening Splenic injury/hematoma CT Conservative management Grade I Fishback et al[58] 51 Female Screening Splenic injury/hematoma CT Laparotomy/splenectomy Grade III Fishback et al[58] 54 Female Screening Splenic injury/hematoma CT Conservative management Grade I Fishback et al[58] 68 Female Screening Splenic injury/hematoma CT Laparotomy/splenectomy Grade III

NA: Not available; CT: Computed tomography.

The type of spleen injury was analyzed with the outcomes. An 80% of the spleen rupture and 97% of the spleen injury group survived. However, in the remaining patients, the mortality was 3 times more with spleen rupture rather than minor injury about 75%n= ? and 25%n= ? respectively. The Fisher exact test was used as in 50%cells the expected count was less than 5 and thePvalue was 0.028 (statistically significant) indicating higher mortality with spleen rupture. The estimate of measure size was calculated using the Phi test and it was 0.28 (< 0.3) indicating a weak association between the type of spleen rupture and mortality (Table 5, Figure 4).

DISCUSSION

Colonoscopy is considered a safe procedure with low risk of complications. Common procedure related complications include perforation and bleeding[6]. A retrospective study by Levinet al[64]reported 16318 colonoscopies of which the overall complication rate was 0.5% (perforation rate 0.09%, post- biopsy/polypectomy bleeding rate 0.48%). Splenic injury after colonoscopy procedure is rare but a serious complication.During a study at a single institution, the incidence of splenic injury was 0.2 in 10000 procedures (7 out of 296248) for all the colonoscopies done between 1980 and 2008[65].Another study reported a rate of 0.72 per 10000 procedures (12 of 165527 colonoscopies)[66]. Furthermore, a recent study by Laananiet al[10]reported an incidence of 0.20–0.34 splenic injuries per 10000 first colonoscopy procedures. The discordance of incidence rates in various reports possibly could be due to underreporting and underdiagnosis of the cases. Even though there are a few explanations for splenic trauma during colonoscopy, the mechanism is yet to be fully understood. One of the hypotheses is excessive traction on the splenocolic ligament or on splenocolic adhesions, secondary to previous abdominal surgery or intraabdominal inflammatory processes[1,49,58,67,68]. Another possible reason could be a direct blunt trauma when navigating the colonoscope through the splenic flexure[56]. Other known risk factors for splenic trauma include older age, less experienced endoscopists, being female, polypectomy and biopsies during the procedure[1,3,5,6,69].

Nearly 2/3 of the patients (64%) with splenic trauma were females. The higher rate in females can be partly explained by the slightly better participation of women in the colorectal cancer screening process[2]. Saunderset al[70]described colonoscopy being a more complicated procedure in females compared to males because of inherently long transverse colon in females. Nonetheless, there is no evidence to support that anatomic difference could be a reason for higher incidence in females.

In our systematic review, 46% of the people that presented with splenic injury underwent colonoscopy for routine surveillance while an equal number of cases (46%)occurred when colonoscopy was done for diagnostic reasons (gastrointestinal bleed,abdominal pain, weight loss) and polypectomy together. The higher incidence in screening group can be related to increased colonoscopy rates due to public awareness[2]. In regard to the extent of injury to spleen, complete splenic rupture was seen only in 22%, whereas majority (63%) suffered a variety of spleen injuries such assub-capsular hematoma, spleen laceration and spleen avulsion. Moreover, we did not find any association of reason for colonoscopy with either presentation, management or outcome of the patients.

Table 2 Reason for colonoscopy and types of injury

In our review, CT was sensitive in diagnosing almost all cases (98.5%) of splenic trauma. This was higher than the findings of previous literature, which was 81.8%[22].These findings reinforce the importance of performing a timely CT abdomen in all symptomatic patients suspected of post colonoscopy splenic trauma. Early diagnosis can aid in prompt management of the patients, which in turn may translate into lower mortality rate. Interestingly, until 1987 majority of splenic injuries were diagnosed during laparotomy[71,72]. It was since 1989, when CT and ultrasound came into existence, these became the major diagnostic tools[71,73]. CT is very sensitive and specific in diagnosing and grading the extent of splenic injury and thereby determines the management options. Nevertheless, the single most factor that appears to determine the management is the hemodynamic status of the patient.

Abdominal pain either generalized or in left upper quadrant was the chief complaint in most patients (85%), only 15% presented with signs of hemodynamic instability such as dizziness and syncope. Nearly half of the patients (47%) required laparotomy with splenectomy, 38% had conservative treatment (IV fluids, pain medication and blood transfusion), 4% had laparoscopy and only 6% received spleen saving splenic artery embolization. Here, we further classified management based on the degree of injury. Whereas 75% of patients with splenic rupture required laparotomy with splenectomy, the other group received either splenectomy or conservative management in equal proportions of 45% each. Here, a point worth mentioning was the favorable results shown by PSAE in grade I-II splenic injuries,though it is not widely available but reiterates the idea of increased utilization of minimally invasive treatment strategies[49,74].

The mortality rate of bleeding and perforation after colonoscopy have either trended down or stayed stable in the last few years. However, our review indicated an alarming rise in the mortality rate from splenic trauma post-colonoscopy. The overall mortality rose to 10% with 7 deaths out of 68 reviewed cases. This was twice the previously reported rate of 5.4%[56]. In addition, the mortality was 3 times higher with spleen rupture than with other forms of splenic injury. The survival rate of these groups was 80% and 97% respectively. To understand the rising mortality rates, we reviewed existing literature on the incidence of the complication but unfortunately it was limited to case reports and case series.

We could not do meta analysis as there were not enough evidence available and our review was mostly based on the individual reported cases. There were no randomized control trials or large scale observational studies available to justify further comprehensive analysis.

In conclusion, colonoscopy overall is a safer procedure and is rarely associated with splenic injuries. The incidence and management of splenic complications is not affected by the indication of colonoscopy. However, the management of spleen injury should be tailored according to the type of injury with splenectomy being reserved mostly for unstable patients with spleen rupture as it has been associated with a significantly higher mortality rate.

Table 3 Presentation, diagnosis and management of splenic injury

Table 4 Likelihood ratios for different comparisons

Table 5 Likelihood ratios and Fisher’s test for different comparisons

Figure 2 Graphical representation of splenic injury and reason for colonoscopy.

Figure 3 Graphical representation of reason for colonoscopy and management of splenic injury. PSAE: Proximal splenic artery splenic embolization.

Figure 4 Relation of splenic injury with outcome.

ARTICLE HIGHLIGHTS

Research background

Colonoscopy is a routine diagnostic and therapeutic procedure. Rarely, colonoscopy can cause splenic Injury.

Research motivation

Splenic injury is a rare but fatal complication of colonoscopy. We wanted to study the various research manuscripts published on splenic injuries during colonoscopy and find out the most common indications for colonoscopy, various presentations of patient with spleen injury,different types of injury, diagnosis and management of splenic injury.

Research objectives

The main objectives were to investigate the reason for colonoscopy, presentation of patient with spleen injury, types of injury, diagnosis, management and outcomes of patients

Research methods

A structured search on four databases was done and 45 articles with 68 patients were selected.and analyzed using SPSS. A literature search for relevant articles was performed through April 25, 2019, using MEDLINE (PubMed, Ovid), Embase and Cochrane databases. We selected manuscripts which inlcuded subjects with any type of spleen injury secondary to colonoscopy and discussed the mechanism, diagnosis and outcome of spleen injury.

Research results

We found that the mean age of the patients was 62.7 years with females predominance. Some of the patient (20%) had a complete splenic rupture, while majority (63%) had subcapsular hematoma, spleen laceration and spleen avulsion. We noticed that the most common reason for colonoscopy was screening (46%) followed by diagnostic colonoscopy (28%). Most common presentation was with abdominal pain. Patients with spleen rupture mostly required splenectomy (47%), while minor spleen hematomas and lacerations were managed conservatively (38%). Few patients (6%) were managed with proximal splenic artery splenic embolization and 4% were managed with laparoscopic repair. The overall mortality was 10%while 77% had complete recovery. Majority of the patients with splenic rupture were managed with splenectomy while the rest were managed conservatively (P= 0.04). This association was moderately strong at a cramer’s V test (0.34). The Fisher exact test showed a higher mortality with spleen rupture (P= 0.028).

Research conclusions

We found that the most common reason for colonoscopy among patients with splenic injuries was screening colonoscopy. The most common presentation was with abdominal pain.Computed tomography abdomen was diagnostic mode of choice. Majority of the patients with splenic rupture were managed with splenectomy and overall mortality was 10%. Recently,monitored anesthesia care has been used in majority of the patients for anesthesia during colonoscopy. Propofol has been used as a part of the protocol. It has been postulated that compared to conscious sedation, deep sedation is expected to blunt patient responses to painful stimuli which can lead to traumatic injuries during colonoscopy like splenic injury and perforation. Majority of the manuscripts did not have information on anesthesia protocol, which would have provided valuable information.

Research perspectives

Further studies are needed to find the likely etiology of splenic injury during. Anesthesia with propofol has been postulated to be one of the reasons for splenic injury as it might blunt patient responses to painful stimuli.