Br. J . Surg. 1991, Vol. 78, June, 71 6-721

G. L. Cullingford*, D. N. Watkinst, A. D. J. Wattst and D. F. Mallon? *Department of Surgery and ?Division of Social and Preventative Medicine, Department of Medicine. University of Western Australia, Nedlands, 6009, Western A ustralia Correspondence to: Mr G . L. Cullingford

Severe late postsplenectomy infection In all, 1490 patients underwent splenectomy in Western Australia between 1971 and 1983, giving 7825 person years exposure. Thirtythree patients developed severe late postsplenectomy infection (septicaemia, meningitis or pneurnococcal pneunzonia requiring hospitalization) and three developed overwhelming postsplenectomy infection. The incidence and mortality rates o f severe late postsplenectomy infection were 0.42 and 0.08 per 100 person years exposure respectively and f o r overwhelming postsplenectomy infection the incidence and mortality rates were 0.04 per 100 person years exposure. There were 628 splenectomies after trauma, giving 3922 person years exposure. Eight patients developed severe late postsplenectomy infection of whom one had overwhelming postsplenectomy infection. Following trauma, the incidence of severe late postsplenectomy infection was 0.21 per 100 person years exposure, with the incidence and mortality rates of overwhelming postsplenectomy infection being 0.03 per 100 person years exposure. Patients undergoing splenectomy have a 1 2 6 f o l d increased risk of developing late septicaemia compured with the general population. Splenectomy following trauma gives an 8.6-jold increased risk of late septicaemia. The majority of severe late postsplenectomy injections did not occur within the first 2 years and 42 per cent of severe late postsplenectomy infkctions occurred > 5 years after splenectomy. The low incidence of severe late postsplenectomy infection and overwhelming postsplenectomy injection makes statistical evaluation of the effectiveness of prophylactic antibiotics, vaccination and splenic repair most dificult.

In 1919 Morris and Bullock’ noted the importance of the spleen in resistance to infection and found an increased mortality in splenectomized rats inoculated with the bacillus of rat plague. In 1952, King and Shumacker2 reported the severe, often fatal, septic episodes in children after splenectomy. The first report in adults was by Whitaker3 in 1969. There is an increasing awareness of the risk of severe late postspienectomy infection but the extent of this risk is not well defined. The duration of a patient’s exposure to potential severe late infection following splenectomy has not been taken into account in population studies. This study was conducted to establish the incidence and mortality of severe late postsplenectomy infection in Western Australia.

Patients and methods Di.finitions

The infections selected were: (1 ) septicaemia, as this is recognized to be the most severe infection occurring in the asplenic individual4, (2) meningitis, to determine whether asplenic individuals are more prone to severe infections outside the vascular compartment, and (3) pneumococcal pneumonia, t o determine the susceptibility to pneumococcal infections per se, compared with those associated with septicaemia. If patients with meningitis or pneumonia had associated septicaernia, they were placed in the septicaemia category. Severe late postsplenectomy infections were those requiring a separate subsequent hospital admission to that for splenectomy. The term ‘overwhelming postsplenectomy infection’ describes the particularly fulminant form of septicaemia occurring after splenectomy, which is the most severe of all postsplenectomy infections5. It has a rapid onset and progression, often to death within 48 h,despite intensive therapy. Study poppulntion

Western Australia has a highly stable population’, isolated by large distances from the remainder of the continent. The mean annual

emigration rate from Western Australia during the study period was 1.57 per cent of its mean population of 1.17 million. The study population was patients who had undergone splenectomy in Western Australia hospitals between January 1971 and December 1983. Using the Western Australian Hospital Morbidity System of the Health Department of Western Australia and record linkage programmes (unpublished data), those patients who had undergone splenectomy and had subsequently died or required a separate hospital admission for meningitis, pneumococccal pneumonia or septicaemia were identified and their hospital records reviewed. The period from 1971 to 1983 was selected for analysis to maximize the detection rate of severe late postsplenectomy infection by minimizing the effect of prophylactic antibiotic therapy (rarely prescribed before 1980) and of pneumococcal vaccination (not available in Western Australia until 1980 and not routinely used until 1982). ‘Person years exposure’ after splenectomy were calculated from the date of splenectomy until December 1983. or until death if earlier, using the Man Years Calculation Programme of the Department of Medicine. The incidence and mortality rates of meningitis, pneumococcal pneumonia, septicaemia and overwhelming postsplenectomy infection were calculated according to person years exposure and were standardized for age using the method of direct standardization and the Western Australian population, 1981-1983, as standard weights (unpublished data). The relative risk of developing infection following splenectomy was compared with the rate in the Western Australian population. The number of expected cases was determined as the sum of the products of the age-specific hospital admission rates for the relevant infections in the general population times the number of person years exposure contributed by that age group. The relative risk of infection was the ratio of the observed rates in the splenectomized cohort to the expected rates defined above. In this way a cohort analysis of the risk of infection as a function of years following splenectomy was defined. Statistical significance was calculated utilizing 95% confidence intervals. Caregori:ation of splenectonrj. The indication for splenectomy was determined as follows: category I, following external trauma; category 11, during surgery for malignant

~~~

716

~

~.

0007-1323/91/060716- 06 (’ 1991 Butterworth-Heinemdnn

~.

Ltd

Infection after splenectomy: G. L. Cullingford et al.

Table 1 Incidence und rnortuliiy o/’severe lute postsplenectomy infections in Western Australia 1971-1983

Incidence of infection Overwhelming

Category

Meningitis

Pneurnococcal pneumonia

Sept icaemia

postsplenectomy

All severe

infection

infections

~

I

Trauma, H = 628

3855 patient years I1 Malignant. n = 142 967 patient years

O.OS(0 01 ) 131

0.13(0.01) [51

0.10(0~02)

0.1O(0.02)

111

I11

0.03(0.01)

111

rrr

Haematological, n =443 2385 patients years Overall n = 1490 7825 patient years

0.2 l(0-0I 1 [81

0.03(0.0 I )

[I1

0.3 l(0.02J 131

Non-malignant, n =231 374 patients years IV

Mortality* ~~~

-

1.60(0-53) 161

0.53(0.01) [21

0~08(0~01) c21

0.17(0.01 J [41

0.42(0.02) [I01

0.08(0.01 ) 121

0.67(0-01) 1161

0.13(0.01)

0.04(0.005) c31

0~10(0.01)

0.28(0.01) 1221

0.04(0.005) 131

0.42(0.01) 1331

0.08(0,01) 161

[Sl

c31

Age standardized rates per 100 person years exposure (95% confidence intervals). Numbers of patients infected are shown in square brackets. * A l l late postsplenectomy infectious deaths were from septicaernia

conditions; category 111, surgery for non-malignant conditions; category IV, haematological or lymphoproliferative disorders; and category V. those whose splenectomy could not be classified into the other four categories.

Results During the 13-year study period, 1490 patients underwent splenectomy in Western Australia; 903 males and 587 females contributed 7825 person years exposure with a mean of 5.25 years. Included in these figures were 183 children (under 15 years of age) contributing 779 person years exposure. The numbers undergoing splenectomy and the person years exposure contributed by each category are listed in Table 1 . Forty-six patients were in category V, contributing 244 person years exposure. The age of patients ranged from 19 months to 84 years, with a median of 31 years for both males and females. In those undergoing splenectomy for trauma there was a peak incidence in the 15-29-year age group. There were 323 splenectomies performed in this subgroup accounting for 51.4 per cent of all splenectomies and of these, 77.7 per cent were males. The age-standardized rate of splenectomy per 100 000 population in Western Australia between 1971 and 1983 was 11.7 (95% confidence intervals 10.6-12.6) for males and 7.6 (95% confidence intervals 6.6-8.6) for females. During the review period the splenectomy rate in children fell from 5.2 (95% confidence intervals 3.8-6.8) in 1971-1974 to 2.4 (95% confidence intervals 1.31-3.6) in 1981-1983 which may reflect altered indications and increased conservative management of splenic trauma. Of the 1490 splenectomized patients, 33 (2.2 per cent) developed severe late postsplenectomy infection: three (0.2 per cent) with meningitis, eight (0.5 per cent) with pneurnococcal pneumonia and 22 (1.5 per cent) with septicaemia of whom three (0.2 per cent) had overwhelming postsplenectomy infection. Of the 22 with septicaemia, five (0.8 per cent) occurred after splenectomy for trauma, one (0.7 per cent) after splenectomy during surgery for malignant disease, six (2.6 per cent) after splenectomy during surgery for non-malignant disease and ten (2.3 per cent) after splenectomy for haematological or lymphoproliferative disorders (Table I). None of the patients in the unclassified category V developed severe late postsplenectomy infection. The incidence of severe late postsplenectomy infection and overwhelming postsplenectomy infection were age-standard-

Br. J. Surg., Vol. 78, No. 6, June 1991

ized per 100 person years exposure. The overall incidence rate of severe late postsplenectomy infection was 0.42, of which meningitis accounted for 0.04, pneumococcal pneumonia 0.10 and septicaemia 0.28 per 100 person years exposure (Table 1). In the paediatric population only one patient with severe late postsplenectomy infection was found, a case of overwhelming postsplenectomy infection, giving an incidence and mortality rate of 0.13 per 100 person years exposure. After splenectomy for trauma, the incidence of severe late postsplenectomy infection was lower than for the other splenectomy categories, being 0.21 per 100 person years exposure, with an incidence of septicaemia of 0.1 3 and of overwhelming postsplenectomy infection of 0.03 per 100 person years exposure. Six patients died from septicaemia and three had overwhelming postsplenectomy infection. All three with overwhelming postsplenectomy infection had had a prodromal illness (in two for >48 h), one had undergone splenectomy for trauma and two had undergone splenectomy for idiopathic thrombocytopaenic purpura. The mortality from late postsplenectomy septicaemia was 0.08 per 100 person years exposure with a mortality rate from overwhelming postsplenectomy infection after splenectomy for trauma of 0.03 per 100 person years exposure (Table I ) . The increased risk of developing septicaemia after splenectomy was 12.6 times that of an age-matched non-splenectomized cohort (Table 2). Of the six patients developing septicaemia after splenectomy for non-malignant disease, two had chronic liver disease and three had developed malignant conditions since their splenectomy (two adenocarcinoma and one non-Hodgkin’s lymphoma). When the results were corrected for the subsequent change in diagnosis, the increased risk of developing septicaemia was 8.6-fold after splenectomy for trauma, 7.1 -fold for malignant disease, 15.0-fold for non-malignant disease and 24.4-fold after splenectomy for haematological or lymphoproliferative disorders. The death rate from overwhelming postsplenectomy infection was more than 300 times that expected from septicaemia in the age-standardized population but there was no increased frequency of meningitis or pneumococcal pneumonia. Bacteriology In the 22 patients with a clinical diagnosis of septicaemia, I9 positive blood cultures were obtained, Streptococcus

717

Infection after splenectorny: G. L. Cullingford et al.

rable 2 Increased risk of developing septicaemia after

splenectomy compared with uyr-spec$c rates J i ~ rWesterti A iistruliiiti populution

Corrected figures' Category

Observed

I Trauma I1 Malignant I11 Non-malignant IV Haematological

Overall

22

Expected

O/E-Risk

Observed

8.6(7.5)

5

2.4(4.8)

3

O/E-Risk 8.6 7. I

1

0.58 0.42

6*

0.20

30.1(24.0)

3

15.0

10

0.45

22.1 ( 13.7)

11

24.4

1.75

IX(5-3)

77 --

12.6

5

Values in parentheses are + 95% confidence intervals. *Corrected figures for subsequent change in diagnostic categories after splenectomy : two patients in category 111 subsequently developed adenocarcinoma and one developed non-Hodgkin's lymphoma before their episode of septicaemia. 0, observed; E. expected

Table 3 Barteriologv

of

postsplenectoniy septicaemia

Organism

No. of patients

Streptococcus pneumoniae Ewherichia coli Klebsiella species Proteus species Haemophilus species Serratia species Streptococcus viridans Gram-negative rod, not identified

Total Values in parentheses are percentages

pneumoniae (pneumococcus) being the most common organism (32 per cent of cases) (Table 3). Encapsulated bacteria (including Escherichia coli which possess a microcapsule of acidic polysaccharide) were responsible for 79 per cent of positive blood cultures. Of the 33 severe late postsplenectomy infections, 12 (36 per cent) occurred within 24 months of splenectomy, seven (21 per cent) between 2 and 5 years after splenectomy and 14 (42 per cent) > 5 years after splenectomy. There was no correlation between the seventy of septic episode and interval since splenectomy. None of the patients developing severe late postsplenectomy infection had received antibiotic prophylaxis or pneumococcal vaccine. Two separate pneumococcal infections occurred in a patient with non-Hodgkin's lymphoma despite vaccination with pneumococcal vaccine.

Discussion In this study no significant increase in incidence of pneumococcal pneumonia or meningitis was found after splenectomy but all splenectomized patients had an increased risk of developing late septicaemia. Splenectomy gave a I2.6-fold (95% confidence intervals 7.3-1 7.9) increased risk of developing septicaemia when compared with the general population. The incidence and mortality rates of severe late postsplenectomy infection were 0.42 and 0.08 per 100 person years exposure respectively. The reported incidence of severe late postsplenectomy infection and overwhelming postsplenectomy infection varies considerably in the literature (Table 4 ) owing to differences in referral patterns and surgical practice (e.g. different percentages of children and trauma). One study7 included infections such as pneumonia, urinary tract infections and subphrenic abscesses in the perioperative period. These were complications of the trauma sustained, treatment given and operative procedure rather than the asplenic state. Differences exist as to what constitutes late postsplenectomy infection: miliary tuberculosis',

718

osteomyelitis, rheumatic fever, respiratory tract infections', peritonitis'", late wound infections' I , intra-abdominal abscesses, Herpes zoster12 and pyrexia of unknown origin13 have been included. give insufficient detail to allow categorization of the infections. Some s t ~ d i e s ' " . ' ~ ~ ~ ~ ~ ' ~ evaluated the mortality from severe late postsplenectomy infection but not its incidence. A fewl3.I5 did not differentiate between indications for splenectomy when calculating rates for severe late postsplenectomy infection. Among studies examining late infection after splenectomy for trauma (Table 5 ) , did not exclude iatrogenic trauma or splenectomy incidental to the surgical procedure. Because the larger studies4.I6 have derived incidence rates from other published reports, mostly single centre studies, their results may be biased. Some studies document the person years exposure (Table 6), enabling the calculation of the incidence of severe late postsplenectomy infection as an expression of time exposure. Several gave insufficient detail to allow calculation of the incidence after splenectomy following trauma' o.14.15.19. The indication for splenectomy affected the incidence of severe late postsplenectomy infection and splenectomy for haematological or lymphoproliferative disorders was associated with a higher incidence of severe late postsplenectomy infection than that after trauma, in keeping with other s t ~ d i e s ~ ~ ~ ~ ' ~ ~ ' ~ ~ " After splenectomy for trauma there was an 8.6-fold increased risk of developing late septicaemia but the incidence of severe late postsplenectomy infection was low at 0.21 per 100 person years exposure, similar to that found by Malangoni rt a / . 2 o . There may be a partial protective effect from splenosis, owing to growth of self-implanted remnants of the traumatized spleen but splenosis does not guarantee protection from overwhelming postsplenectomy infection". In this study, after splenectomy for trauma, the mortality from severe late postsplenectomy infection was 0.03 per 100 person years exposure, a crude incidence of 0.2 per cent. There are many case reports of severe late postsplenectomy infection or overwhelming postsplenectomy infection where numbers have not been evaluated22. Excluding these, only 23 deaths have been reported from severe late postsplenectomy infection in 4221 splenectomies following trauma. a collective mortality rate of only 0.5 per cent despite the bias from the high proportion (51 per cent) of children reviewed (Table 5 ) . In those studies documenting person years exposure after splenectomy for trauma (Table6). only three cases of overwhelming postsplenectomy infection have been reported in 7246 person years exposure. an incidence rate of 0.041 per 100 person years exposure. Following any form of splenectomy. only ten cases of overwhelming postsplenectomy infection have been reported, an incidence rate of 0.058 per 100 person years exposure. Increased surgical awareness of overwhelming postsplenectomy infection is leading to modifications of surgical practice. Surgeons attempt repair of iatrogenic splenic injuries and in

Br. J. Surg., Vol. 78, No. 6, June 1991

Infection after splenectomy: G. L. Cullingford at al. Table 4 Summary of studies on late psrsplenectomy infection

Sepsis -

First author*

--.-

Lowdon' (P) Kiesewetter' Singer* ONeal14 Schwartz' Standageas Ellisona6(M) Pedersen" (P) Malangoni'O Carlstedtl3 Chaikof" Pate3' Sekikawa' VichardJ3 Green'' Greed' DiCataldo" Selby" Present study (P)

No. of patients

Children

Trauma?

("/.I

("/I

1118 205 2795 256 193 175 5579** 384 140 133 637 316 242

17 100 57

228 144 18 148 131 1490

Nil ns Nil 75 I00

35tt

ns

ns

ns ns Nil Nil 100

Incidence All All SMP

ns

ns 81

("/.I

Mortality rate (%)

OpsI$

2.0 I .S$ 2.5 2.7 0-5 11.4 1.7** 1.6 0 0 5.3 0.3 0 I *7@ 0.7

0.2 1.o

8.0

6% 4.3

S

ns

All

26.9

LH

ns

ns

3*1** 5.5 21 12.8

SM SMP

LH LH

ns ns

S

LH

2.5 12.0$$ 9.7

All

LH LH S S

11.1

0

ns

0 1.5 04

4.6 2.2

SMP

42

12

~~

Definition#

21 10 25 50 25 23 53 100 26 15 50 100 100 100 100 47

Nil

~

("/.I

ns a3 05

ns ns ns 0

0 06 03

0 0 0.7$ 55 0 0 02

* Superscript number is the reference; all studies from single centres except: M,multiple review; P, population study. t Trauma excludes iatrogenic trauma or incidental splenectomy unless indicated. #Definition of sepsis: All, all infections; LH,sepsis requiring late hospitalization; M, meningitis; P,pneumonia; S,septicaemia; OPSI,overwhelming postsplenectomy infection, patient numbers adjusted to conform to accepted definition'; ns, not stated. $Adjusted to exclude infections within 1 month of splenectomy. **Includes data from Singer' and O'Neall*. ttIncludes incidental or iatrogenic splenectomy. $# Mostly viral infections. All deaths from sepsis occurred in patients with multiple trauma, the time interval following splenectomy is not specified. ***Combined total corrected for Ellison'sa6 inclusion of data from Singer4 and ONeal". tttcorrected to exclude studies where sepsis type not stated. Some symbols appear only in Tables 5 or 6 Tabk 5 Summary of studies on late infection afrer spbectomy for trauma? Sepsis

First author*

No. of patients

Lowdon' (P) Singer* ONeal" Schwartz' Ellisoni6 (M) Peder~en'~ (P) Malangoni'O Chaikofl PateJ2 Sekikawa' VichardJ3 Greena2 Green3* Selby" Present study (P)

287 688 128t 48 1954**tt

Combined***

422 1

mtt 140 96 158 242 228 144 18 74 628

Children

(%I 34 56 0

ns 87 100 0 20

Incidence Definition$

SMP SMP

ns

All ns

204 1*4**tt

SM SMP

LH S

0 0 100

All

ns

S

15

0.7 1*4

S

ns ns

51

(%I

LH LH SMP

OPSI#

0.3 0.6 0 0

0 4 0 0

0*8**tt

ns ns

(%I

23tt

o5tt

2.1

0 0 0 0 1-7g 0.7 0 0 0.2

0 1

0.5

7

ns ns

LH

Mortality rate (%)

2.5 12.0$$ 9.7 11.1 0 1-3

0 0 0 0 0

I

See footnotes to Table 4

some diseases, especially Hodgkin's lyphoma, the frequency of elective splenectomy has been reduced. After trauma attempts should be made at splenic repair in stable patients but should not otherwise prolong the operation. Splenic autotransplantation is not fully evaluated and

Br. J. Surg.. Vot. 78, No. 6, June 1 H 1

protection is not guaranteedZ3ez4.It carries the risk of graft necrosis with abscess formation and may result in an increase in adhesive bowel obstructionz5. Asplenic patients should be considered to have a life-long risk of developing severe late postsplenectomy infection,

719

Infection after splenectomy: G.

L. Cullingford et al.

Table 6 Summary q/‘studie.s on late postsplenectomy infection rates per 100 person years exposure (PYE) Splenectomy for trauma?

All splenectomy indications First author* O’NealI4 Schwartz’ Standagels Pedersen’” Malangoni” Chaikof’ Sekikawa” Green I Green”4 Present study (P)

’

Combined

PYE 944 1090

879 2381 1203 4837 1406 734 105 7825

21 404

Incidence sepsis/100 PYE

Mortality sepsis/100 PYE

Incidence OPSI$/100 PYE

ns 4.77 ns 0.88 0.25 ns 0.57 1.91 1.90 0.42

0.74 0.09 2.28 0.25 0 0.70 0 0.14 0 0.08

ns 0.09 ns ns 0 0.08 0 0.14 0.95 0.04

0.89ttt

0.35

0.06ttt

Incidence sepsis/100 PYE

Mortality sepsis/100 PYE

Incidence OPSI$/lOO PYE

303

3.30

0

0

1203

0.25

0

0

1406 734 105 3855

0.57 1.91 1.90 0.21

0 0.14 0 0.03

0 0.141 0.95 0.03

7246

0.59

0.03

0.04

PYE

See footnotes to Table 4

42 per cent of cases in this study occurred > 5 years after splenectomy. All who have undergone splenectomy should receive polyvalent pneumococcal vaccine. This should be given in elective situations as soon as a decision is made to proceed to splenectomy and in an emergency just before discharge from hospital. The protection from vaccination is not complete and there have been many cases of pneumococcal infections following vaccination, some fatal’ ‘ , 1 9 . 2 3 . 2 6 - 3 0 . Meningococcal and H(ic.mophilus vaccines may be considered, but they have a narrow serotype coverage. Long-term prophylactic antibiotics (such as phenoxymethyl penicillin) may not prevent severe late postsplenectomy infection as not all infections are caused by organisms sensitive to penicillin. Forty-two per cent of severe late postsplenectomy infections occurred > 5 years after splenectomy ; long-term patient compliance is a problem and there are case of fatal pneumococcal sepsis despite penicillin prophylaxis. Early recognition and treatment of infection, as advocated by , be more effective. The patient is given Duncombe et u / . ~ ’may a supply of amoxycillin 500 mg tablets with instructions to take two tablets at the onset of the first symptom of any illlness or fever. Amoxycillin has a better absorption and a wider spectrum of activity than phenoxymethyl penicillin. Education of the medical profession and patients is essential to minimize delay in referral as only by having a high level of suspicion can overwhelming postsplenectomy infection be recognized and aggressive therapy instituted, thereby reducing the severity of infection and hopefully halting progression to overwhelming postsplenectomy infection. The low incidence of severe late postsplenectomy infection and overwhelming postsplenectomy infection, especially after splenectomy for trauma, makes statistical evaluation of the efficacy of prophylactic antibiotics, vaccination and splenic salvage virtually impossible because of the large number of patients required to detect statistically significant differences in incidence.

3. 4. 5.

6. 7. 8. 9. 10. 11.

12.

13. 14. 15. 16. 17. 18.

Acknowledgements

19.

The authors thank two members of the Division of Social and Preventative Medicine, Department of Medicine, University of Western Australia: Associate Professor M . S. Hobbs for advice, and Mr R . Hockey for technical assistance.

20. 21.

References I.

2.

720

Morris DH, Bullock FD. The importance of the spleen in resistance to infection. Ann Surg 1919; 70: 513-21. King H , Shumacker HB. Splenic studies. I. Susceptibility to

22. 23.

infection after splenectomy performed in infancy. Ann Sury 1952; 136: 23942. Whitaker AN. Infection and the spleen: association between hyposplenism, pneumococcal sepsis and disseminated intravascular coagulation. M e d J Aust 1969; 1: 1213-19. Singer DB. Postsplenectomy sepsis. In: Rosenberg HS, Bolande RP, eds. Perspecthw in Pediatric Pathulngy. Chicago: Year Book Medical Publishers, 1973, 1 : 285-31 1. Diamond LK. Splenectomy in childhood and the hazard of overwhelming infection. Pediatrics 1969; 43: 886-9. Australian Demogrupphic Statistics. Canberra, Australian Bureau of Statistics, Catalogue 3 101.O. Schwartz PE, Sterioff S, Mucha P, Melton LJ, Offord KP. Postsplenectomy sepsis and mortality in adults. J A M A 1982; 248: 2299-83. Lowdon AGR, Stewart RHM, Walker W. Risk of serious infection following splenectomy. Br Med J 1966; 1: 446-50. Kiesewetter WB, Patrick DB. Childhood splenectomy: indications for and results from. A m Surg 1971; 35: 13540. Chaikof EL, McCabe CJ. Fatal overwhelming postsplenectomy infection. Ani J Surg 1985; 149: 534-9. Sekikawa T, Shatney CH. Septic sequelae after splenectomy for trauma in adults. Ani J Surg 1983; 145: 667-73. Green JB, Shackford SR, Sise MJ, Fridlund P. Late septic complicaions in adults following splenectomy for trauma: a prospective analysis in 144 patients. J Trauma 1986; 26: 999-1004. Carlstedt A, Tholin B. Infectious complications after splenectomy. Acta Chir Scund 1984; 150: 607-10. O’Neal BJ, McDonald JC. The risk of sepsis of the asplenic adult. Ann Surg 1981; 194: 775-8. Standage BA, Coss JC. Outcome and sepsis after splenectomy in adults. Am J Surg 1982; 143: 545-8. Ellison EC, Fabri PJ. Complications of splenectomy: etiology, prevention, and management. Surg Clin North An1 1983; 63: 1313-30. Selby C, Hart S, Ispahani P, Toghill PJ. Bacteraemia in adults after splenectomy or splenic irradiation. Q J Metf 1987; 63: 523-30. Di Cataido A, Puleo S, Li Destri G e/ a/. Splenic trauma and overwhelming postsplenectomy infection. Br J Sury 1987: 74: 343-5. Pedersen FK. Postsplenectomy infections in Danish children splenectomised 1969-1978. Acta Puediutr Scand 1983; 72: 589-95. Malangoni MA, Dillon LD, Klamer TW, Condon RE. Factors influencing the risk of early and late serious infection in adults after splenectomy for trauma. Surgery 1984; 96: 775-83. Rice HM, James PD. Ectopic splenic tissue failed to prevent fatal pneumococcal septicaemia after splenectomy for trauma. Lunce/ 1980; I : 565-6. Gopal V, Bisno AL. Fulminant pneumococcal infections in ‘normal’ asplenic hosts. Arrh In/ Med 1977; 137: 1526-30. Moore GE, Stevens RE, Moore EE, Aragon GE. Failure of

Br. J . Surg.. Vol. 78, No. 6, June 1991

Infection after splenectomy: G. L. Cullingford et al.

24.

25. 26. 27. 28. 29.

splenic implants to protect against fatal postsplenectomy infection. Am J Surg 1983; 146: 413-14. Tesluk GC, Thomas CG, Benjamin JT, McMillan CW. Fatal overwhelming postsplenectomy sepsis following autologous splenic transplantation in severe congenital osteopetrosis. J Pediacr Surg 1984; 19: 269-72. Cullingford GL, Surveyor I, Edis AJ. Demonstration of functioning heterotopic splenic autografts by scintigraphy. Aust NZ J Surg 1983; 53: 343-7. Appelbaum PC, Shaikh BS, Widome MD, Gordon RA, Austrian R. Fatal pneumococcal bacteremia in a vaccinated, splenectomised child. N Engl J Med 1979; 300: 2 0 3 4 . Giebink GS, Schiffman G , Krivin W, Quie PG. Vaccine-type pneumococcal pneumonia: occurrence after vaccination in an asplenic patient. J A M A 1979; 241: 2736-7. Nielsen JL, Pedersen FK, Ellegaard J. Failure of pneumococcal vaccination in a splenectomized child. Acia Puediair Scand 1982; 71: 331-3. Evans DIK. Fatal post-splenectomy sepsis despite prophylaxis

Br. J. Surg., Vol. 78, No. 6, June 1991

30. 31. 32. 33. 34.

with penicillin and pneumococcal vaccine. Lancet 1984; i:1124. Brivet F, Herer 9, Fremaux A, Dormont J , Tchernia G . Fatal post-splenectomy pneumococcal sepsis despite pneumococcal vaccine and penicillin prophylaxis. Lancet 1984; ii: 356-7. Duncombe AS, Dudley J M , Slater NGP, Treacher DF. Overwhelming pneumococcal sepsis post-splenectomy. Lanccr 1987; i:570. Pate JW, Peters T G , Andrews CR. Postsplenectomy complications. Am Sury 1985; 51: 43741. Vichard PH, Dreyfus-Schmidt G . Aspects actuels de la splenectomie pour traumatisme chez I'adulte. Med Chir Dig 1985; 3: 1 9 3 4 . Green JB, Shackford SR, Sise MJ, Powell RW. Postsplenectomy sepsis in pediatric patients following splenectomy for trauma: a proposal for a multi-instutitional study. J Pediair Surg 1986; 21: 1084-6.

Paper accepted 16 December 1990

721