Splenic rupture or infarction associated with Epstein-Barr virus infectious mononucleosis: a systematic literature review

BACKGROUND
Epstein-Barr virus (EBV), also known as human herpesvirus 4, is one of the most common pathogenic viruses in humans. EBV mononucleosis always involves the spleen and as such it predisposes to splenic rupture, often without a trauma, and splenic infarction. Nowadays the goal of management is to preserve the spleen, thereby eliminating the risk of post-splenectomy infections.


METHODS
To characterise these complications and their management, we performed a systematic review (PROSPERO CRD42022370268) following PRISMA guidelines in three databases: Excerpta Medica, the United States National Library of Medicine, and Web of Science. Articles listed in Google Scholar were also considered. Eligible articles were those describing splenic rupture or infarction in subjects with Epstein-Barr virus mononucleosis.


RESULTS
In the literature, we found 171 articles published since 1970, documenting 186 cases with splenic rupture and 29 with infarction. Both conditions predominantly occurred in males, 60% and 70% respectively. Splenic rupture was preceded by a trauma in 17 (9.1%) cases. Approximately 80% (n = 139) of cases occurred within three weeks of the onset of mononucleosis symptoms. A correlation was found between the World Society of Emergency Surgery splenic rupture score, which was retrospectively calculated, and surgical management: splenectomy in 84% (n = 44) of cases with a severe score and in 58% (n = 70) of cases with a moderate or minor score (p = 0.001). The mortality rate of splenic rupture was 4.8% (n = 9). In splenic infarction, an underlying haematological condition was observed in 21% (n = 6) of cases. The treatment of splenic infarction was always conservative without any fatal outcomes.


CONCLUSIONS
Similarly to traumatic splenic rupture, splenic preservation is increasingly common in the management of mononucleosis-associated cases as well. This complication is still occasionally fatal. Splenic infarction often occurs in subjects with a pre-existing haematological condition.


Introduction
Epstein-Barr virus, also known as herpesvirus 4, is one of the most common human viruses [1,2].Primary infectious mononucleosis, the best-known presentation of this virus, typically affects adolescents and young adults; presents with fatigue, malaise, sore throat and enlarged cervical lymph nodes, liver or spleen; and generally spontaneously resolves over a few weeks [1,2].
The spleen is always involved in mononucleosis.Although not always palpable, splenomegaly is detected by ultrasound in all affected individuals [1][2][3].The splenic architecture is distorted because the parenchyma is infiltrated by atypical lymphocytes that compromise the fibrous support system and thin the capsule.Splenomegaly and the distorted architecture predispose to rupture, which is often not associated with a notable trauma but perhaps exclusively heralded by triggers such as coughing, sneezing, straining during defecation or muscular exertion [1,2,4].It is also traditionally assumed that rupture may result from vigorous palpation [4].There might be an increased tendency to splenic infarction as well [5].
The features of mononucleosis-associated splenic rupture and infarction have not been comprehensively investigated in the recent past.Furthermore, it is currently unclear whether spleen-preserving management is considered a viable alternative to splenectomy.
Since mononucleosis-associated splenic rupture and infarction are rare, available knowledge on these complications is mainly based on case reports.To address these issues, we carried out a systematic review of the literature.

Data sources and search strategy
This review was pre-registered in the Prospective Register of Systematic Reviews (PROSPERO; CRD42022370268) and carried out in agreement with the second edition of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations [10].Data sources included Excerpta Medica, the United States National Library of Medicine, and Web of Science from 1 January 1970, without any further limitation.The search strategy incorporated the following terms entered in separate pairs: (Epstein-Barr virus OR glandular fever OR infectious mononucleosis OR herpesvirus 4) AND (hematoperitoneum OR splenic hematoma OR splenic infarction OR splenic rupture).Articles listed within reference lists of the retrieved records, reports published in Google Scholar and articles already known to the authors were also considered.Searches were conducted in April 2022 and repeated before article submission.

Inclusion criteria
Patients had to meet three criteria to be included: (i) an individually documented case who presented with a positive Paul-Bunnell-Davidsohn test, a specific acute Epstein-Barr virus serology response or both [1,2]; (ii) having clinical features of primary mononucleosis; and (iii) either a splenic rupture or a splenic infarction.Cases with splenic rupture or infarction not supported by serological evidence of an existing Epstein-Barr virus infection were excluded.

Data extraction
For each case, the following information was extracted in a pilot-tested spreadsheet: (a) relevant past and recent medical history with emphasis on any pre-existing haematological diseases, recent abdominal trauma or vigorous abdominal palpation; (b) the clinical features of mononucleosis [1,2], i.e. fever (38.0 °C or more), fatigue, malaise, sore throat, yellowish scleral discolouration (referred to as "jaundice" in the remainder of this article), cervical adenopathy (lymph nodes felt to be 1 cm or larger in diameter), hepatomegaly or splenomegaly (a palpable liver edge or spleen); and (c) clinical and laboratory features of splenic rupture or infarction at presentation, including blood pressure, heart rate and signs of haemodynamic instability (shortness of breath, prolonged capillary refill time, mottling of cool and moist extremities, peripheral cyanosis, altered mentation), pain (classified as diffuse abdominal pain, left upper abdominal pain or left shoulder pain), a full blood count (with leucocyte differential) [11], enzyme values (aminotransferases and lactate dehydrogenase) and imaging studies.
The diagnosis of splenic rupture or splenic infarction was suspected clinically and always confirmed by means of appropriate imaging studies or, in haemodynamically unstable cases, a diagnostic laparotomy.The time elapsed from onset of symptoms of mononucleosis to the diagnosis of splenic rupture or infarction was recorded.The diagnosis of mild mononucleosis was made in cases with two or fewer of the following five typical features: cervical adenopathy, fever, malaise, sore throat, splenomegaly.The diagnosis of mononucleosis-associated transaminitis was made in cases with a more than 2-fold elevation in aminotransferase ratio compared to the laboratory's reference value.It has been speculated that in mononucleosis a more than 2-fold elevation in lactate dehydrogenase ratio indicates splenic infarction [5].Consequently, the value of the latter test was verified.Cases with a platelet count of 20-70 × 10 9 /l or <20 × 10 9 /l were recorded.
The clinical and imaging data of each case of mononucleosis-associated splenic rupture was used to score the rupture as minor, moderate or severe according to the spleen trauma classification recommended in 2017 by the World Society of Emergency Surgery (WSES) [12].This classification takes into account both the anatomy of splenic lesions and the patient's haemodynamic condition and has proved to be a reliable tool in the decision-making process in splenic trauma treatment [12].The management was stratified using the following well-established terminology [6][7][8]: "operative management" was used for cases of splenic rupture who underwent an immediate surgical splenectomy; "failure of non-operative management " for cases who underwent a surgical splenectomy after an initial unsuccessful conservative approach; "non-operative management" for the remaining patients -this term was used both for cases with conservative care alone and for cases with an adjunctive treatment such as transcatheter arterial embolisation.

Comprehensiveness of reporting
The comprehensiveness of included cases was evaluated using the following seven components: 1. Characterisation of the patient; 2. Clinical presentation; 3. Disease duration; 4. Vital parameters; 5. Full blood count; 6. Haemodynamic instability; 7. Management and outcome.Each component was rated as 0, 1 or 2 and the reporting quality was graded according to the sum of each item as satisfactory, good or excellent [13].
Two authors (JMAT/BG) separately and in duplicate performed the literature search, selected studies for inclusion, extracted data and evaluated the comprehensiveness of each case.Any disagreements were discussed, and a senior author (MGB) was consulted to resolve any outstanding issues.One author (JMAT) entered the data into a pre-defined worksheet and another (BG) verified the accuracy of data entry.

Analysis
Pairwise deletion was used to handle missing data.Categorical variables are presented as proportions and continuous variables as medians with interquartile range.Dichotomous categorical variables were compared using the Fisher exact test; ordered categorical variables using the Kruskal-Wallis test and the post-hoc Tukey correction; and continuous variables using the Mann-Whitney-Wilcoxon rank-sum test.Linear regressions with the Spearman nonparametric coefficient of correlation r s were also calculated.A two-sided significance level of 0.05 was used.GraphPad Prism 9.5.1 (GraphPad Software, San Diego, California, USA) was used for statistics.

Ethics approval and consent
The study was a systematic review and as such did not require specific ethics approval at our institutions.

General findings
The characteristics of the 215 patients are presented in table 1.A positive Paul-Bunnell-Davidsohn test (n = 90), a specific acute Epstein-Barr virus serology response (n = 87) or both a Paul-Bunnell-Davidsohn test and a specific Epstein-Barr virus serology response (n = 30) supported the diagnosis of mononucleosis in 207 cases.The diagnosis of mononucleosis was made uniquely on a histological basis (spleen showing largenumbers of atypical lymphocytes) in the remaining 8 cases, who died within hours after admission.A pre-existing haematological disease was significantly more frequent (p = 0.001) in splenic infarction as compared to splenic rupture.The prevalence of recent history of abdominal trauma was not statistically different between rupture and infarction patients.In no cases was rupture or infarction preceded by overeager palpation.As compared to cases affected by splenic infarction, cases with splenic rupture presented less frequently with fever or fatigue (table 1).Furthermore, left shoulder pain and haemodynamic instability were more frequent in the splenic rupture group (p = 0.001).A transaminitis was significantly more common in the splenic infarction group (p = 0.001).The prevalence of cases with lactatedehydrogenase level ≥600 U/l was similar in patients with splenic rupture and in those with splenic infarction.The tendency towards a platelet count 20-70 × 10 9 /l or <20 × 10 9 /l was not statistically different in the two study groups.

Splenic rupture
The relationship between the World Society of Emergency Surgery splenic rupture score and management in patients with splenic rupture is shown in table 2. While almost 85% of cases with a severe rupture score were splenectomised, this figure was about 50% for cases with a minor or moderate rupture score (p = 0.001).
Nine patients with a severe splenic rupture score (6 female and 3 male subjects aged 15 to 29 years, median 20 years)

Splenic infarction
An underlying haematological disease was observed in 6 (21%) of the 29 patients with splenic infarction: hereditary spherocytosis in 4 and sickle cell trait in 2. One patient had coeliac disease.Furthermore, one patient concomitantly had Crohn's disease, sacroiliitis and Hashimoto's thyroiditis.Finally, one female subject was taking hormonal contraception.Interestingly, a transient elevation of the antiphospholipid level was noted in one patient.
The management was non-operative in all 29 patients.A splenectomy was performed one month or more after diagnosis in two patients with spherocytosis and in one patient with long-lasting abdominal pain and fatigue.None of the 29 patients died.

Discussion
This systematic review of the literature focuses on two complications of mononucleosis [1][2][3]: splenic rupture and splenic infarction.The results may be summarised as follows.Both rupture and infarction predominantly occur in males, usually 15 to 30 years of age, and present one to three weeks after the onset of mononucleosis symptoms (which are mild in about 20% of cases) with acute abdominal pain, which is mostly diffuse (but often predominates in the left upper quadrant), and left shoulder pain.Haemodynamic instability secondary to a circulatory shock occurs in approximately one-third of cases of splenic rupture, which is still potentially fatal.Moreover, cases with infarction quite frequently have a pre-existing haematological condition such as spherocytosis or sickle cell trait.Finally, and especially clinically relevant, a spleen-preserving

Systematic review
Swiss Med Wkly.2023;153:40081 treatment is nowadays a viable alternative to splenectomy in mononucleosis-associated splenic rupture.
The cases included in this review were categorised as minor, moderate or severe on the basis of the three-point scoring system recommended by the World Society of Emergency Surgery (WSES) for patients with splenic trauma [12].In the patients with Epstein-Barr virus-associated splenic rupture included in this analysis, this scoring system, which is simpler than the five-point spleen injury scale endorsed by the American Association for the Surgery of Trauma and includes both clinical and imaging data [185], correlates well with required management.
Given the rarity of mononucleosis-associated splenic rupture, there is no clear consensus on treatmentstrategy.Nonoperative management of haemodynamically stable cases, i.e. with a minor or moderate WSES rupture score, is currently the standard of care.In addition to an expectant attitude, non-operative management currently includes splenic artery embolisation [6-8, 117, 186].Partial splenectomy and splenic repair are no longer recommended [187].The recommended treatment must be approached with caution given the risk of ongoing bleeding and the potential for late rupture.These patients should be cared for by an experienced multidisciplinary team, with physical activity restriction after discharge.Specifically, no activity more vigorous than walking is recommended until splenomegaly has resolved on clinical examination, followed by a period of no contact sports for six months or until the splenic architecture normalises on imaging evaluation.In view of radiation concerns with CT scans, ultrasound should be the main imaging modality in children, adolescents and women of childbearing age [188].
In splenectomised patients, prevention of infections is crucial.Recommended methods to decrease the infection risk include patient education, vaccination and antimicrobial prophylaxis [7,9].Although this review challenges the long-standing belief [4,28] that palpation may induce mononucleosis-associated splenic rupture, it still seems judicious to avoid vigorous abdominal palpation in this setting.
Generally, splenic infarction is an uncommon diagnosis [5,189].Thromboembolism -either of cardiovascular origin or as the result of a thrombophilia-and a rapidly enlarging spleen -such as in the case of oncological or non-oncological haematological diseases and acute infections-are the main causes.The results of our review suggest that mononucleosis-associated splenic infarction is often not caused by the Epstein-Barr virus infection alone but also by a pre-existing haematological condition.A high index of suspicion for splenic infarction is appropriate in subjects with the mentioned predisposing conditions presenting with left upper abdominal pain, with or without associated left shoulder pain.The data of this review does not support the use of the lactatedehydrogenase test as a diagnostic tool in cases with suspected infarction.The most appropriate diagnostic imaging is CT with intravenous contrast [5].Regrettably, Doppler ultrasound is of limited diagnostic value.
The main limitation of this systematic review relates to the rarity of these two complications.Hence, we collected information from cases published between 1970 and 2022, which were sometimes not thoroughly documented.Three well-accepted databases and Google Scholar were used for our literature search.It seems to us highly improbable that substantially different results would have been obtained if additional databases had been searched.Furthermore, the recommended treatment strategy does not arise from welldesigned studies but is mainly extrapolated from current guidelines on the management of traumatic splenic rupture [6][7][8].The main strength of the study relates to the relatively high number of included cases.Furthermore, this is the first report which investigates the literature on mononucleosis-associated splenic infarction.
In conclusion, this systematic review of the literature documents that, like with traumatic splenic rupture, splenic preservation is increasingly common in the management of mononucleosis-associated splenic rupture, with mainly good success.The treatment strategy is dictated by haemodynamic parameters.Sadly, even in the third millennium, this disease is still occasionally fatal.

Figure 1 :
Figure 1: Splenic rupture or infarction associated with Epstein-Barr virus infectious mononucleosis.PRISMA flow chart of the literature search.

Figure 2 :
Figure 2: Time from onset of symptoms of Epstein-Barr virus infectious mononucleosis and onset of symptoms of splenic rupture or infarction.

Figure 3 :
Figure 3: World Society of Emergency Surgery splenic rupture score in patients with splenic rupture associated with Epstein-Barr virus infectious mononucleosis.

Table 1 :
Characteristics of 215 patients (82 females and 133 males) with Epstein-Barr virus mononucleosis complicated by splenic rupture or infarction.Results are presented as frequency (and percentage) or as median (with interquartile range).Significant p values are given in bold.

Table 2 :
Relationship between World Society of Emergency Surgery splenic rupture score and management in patients with mononucleosis-associated splenic rupture.
* While 46 out of 55 cases with a severe rupture score were splenectomised, this figure (70 out of 121 cases) was lower (p = 0.001) for cases with a minor or moderate rupture score.