Previous ectopic pregnancy should be considered a contraindication for microsurgery

Previous ectopic pregnancy should be considered a contraindication for . microsurgery ANNIKA STRANDELL AND JANE THORBURN

From the Department of Obstetrics and Gynecology, The Sahlgrenska University Hospital, Goteborg, Sweden

t ( tc; O/)\/cr ( J \ , ~ i ~ ~ ~ d Sctr~zd 1996: 75: 394-399. 0 Acta Obstet Gynecol Scand 1996

BM ~ , ~ Y I M ~ M / . lo ewmate the risk of subsequent ectopic pregnancy (EP) after tubal surgery. gi\.cn thtir the woninn becomes pregnant. by means of a logistic model. a retrospective study \\;I\ i i i i t i a ted . . M d l d \ . Dui-ing the period 1986- 1990. 22 1 women with tubal infertility underwent microsur- g q . Suhwyucnt fertility was evaluated in 1991. Ninety women conceived, of whom 84 were ~ncluded in the \ttidy (30 with EP and 54 with intra-uterine pregnancy as the only outcome). C‘iriiicaf b a c h g m n d factors of importance. surgical procedures used. scoring systems for u b a l Ie\ion\. adnexal adhesions and risk of EP were analysed for possible correlation to sub- wyuciit EP, T!ie\e factors were further used in a logistic model to estimate the risk of subse- yiient Ef’ a \ only outcome. Kc.\ir/r\. Thc r i k of EP after microsurgery is minimum 15% without any risk factors. Pre\ious Ef’ and c.nciclnietrio\is could be identified as factors with prognostic power in the logistic model. ( h i e pre\,icxi$ EP implies a 60% risk, whereas two previous EPs and endometriosis

COIM / u \ i o / i . Patient\ L\ ith previous EP yhould generally not be considered for microsurgery o\\ in? i t ) the tiiph r k k of recurrence and to the reduced chance of intra-uterine pregnancy.

KC,! woi.t/\: ectopic pxcgnancy: logistic model: pregnancy outcome: risk estimation: tubal ! n l c ! ~ ~ ~ \ l i l ~ c l >

, 4 c ~ ~ q ~ i i , ~ / 26 , I W ! C . 1095

The incidence of ectopic pregnancy (EP) appears still t o be on the rise in many countries, with a four- to fix-fold increase over the past two dec- ades 1.2). The increase can be explained by changes i n risk factor patterns in the population and by demographic changes and increased diag- nostic accuracy (2). A more conservative approach to the treatment of EP with a risk of recurrence in the p i w m cd tube and more widespread accept- ance of tubal microsurgery may also have contrib-

. A h / ~ i w r c r r i o i i i. EP: ectopic p x g a n c y : PID: pelvic inflammatory disease: IVF: i / i \ ir1.1, fi.rtiliratIon: [UP: intrauterine pregnancy: TUCD: intra- uterine coiiri-aceptivc device: LC: linear combination: P(EP): ilir protxibilir! of hubscyuent ectopic pregnancy.

~~ ~ ~~ ~~~ ~

uted to the increase. A recent report though, shows a tendency of declining incidence of EP in Sweden since the beginning of the nineties, suggesting that the ‘epidemic’ might be over (26).

Known risk factors for EP are pelvic inflammatory disease (PID), infertility. previous EP (regard- less of treatment procedure) and pelvic operations per se (7,8,9,10,11). Patients undergoing tubal microsurgery generally present known risk factors for EP and are potential high risk patients for subsequent EP even before surgery. The surgical procedure per se may increase the risk to an even greater level. EP rates after tubal microsurgery are reported to be in the range of 2% to 17% (4,5,6).

However, the risk of EP must aln.ays be connid-

Risk of ectopic pregnancy after microsurgen 395

Table I Factors analysed for possible correlation to subsequent ectolic pregnancy (EP) and their frequency distribution or mean values in 84 patients, 30 with ectopic pregnancy and 54 with intra-uterine pregnancy (IUP) as the only outcome ~ - - ~ ~ ~ ~ ~ ~ ~~ ~ ~~ ~ ~

EP IUP Factor pva lue (n=30) (n=54)

Tuba1 score n s grade 1 - 1 1 78% 93% grade Il l-IV 22% 7 Yo

Adhesion score grade 1 - 1 1 grade Il l-IV

n.s 67% 6 7 '10 33% 3 3 '/o

~ - ~ - ~ ~~ ~ ~~ ~ ~ ~~ ~ - ~-

Risk score for EP

Surgical procedure n s

<o 001 3 6 (s e mean 0 3) 2 6 (s e mean 0 1)

salpingostomy (n=38) 40 0% 47 2 % adhesiolysis (n=32) 40 0% 37 7% anastomosis (n=9) 13 3% 9 4% implantation (n=5) 6 7% 5 7% - ~ ~ ~ ~ ~ -~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

Background factors No of events - ~ - ~~ ~ ~ ~ __ ~ ~ ~ ~ ~ ~ ~~

previous EP t o 001 0 50% 88 9% 1 30 0% 9 2% 2 13 4% 1 9% 3 0 0 4 3 3% 0 5 3 3% 0

previous delivery

previous abdominal surgery*

n.s.

n.s.

76.7% 72.2% 16.7% 13.0% 6.6% 7.4% 0 7.4%

36.7% 50.0% 36.7% 35.3% 13.3% 11.1% 6.7% 1.8% 0 0 3.3% 1.8% 3.3% 0

previous endometriosis n s 17% 6 Yo

previous salpingitis n s 30% 51 '/o previous intrauterine device n s 23% 2 6 '10

Age (years) Duration of infertility

(years)

n s

n s

32 9 (s e mean 0 6)

3 6 (s e mean 0 4)

32 1 (s e mean 0 5)

4 1 (s e mean 0 3) ~ ~ ~ ~ ~ ~~ ~ __ ~ ~ ~ ~ ~ ~ ~ ~ -

*including surgery for ectopic pregnancy, fertility surgery, cesarean sx t i on , appendectomy and cholecystectorny

ered in relation to the likelihood of a subsequent tle- livery.

I12 v i m fertilization (IVF) as an alternative to microsurgery should therefore be considered in cases of extreme risk of subsequent EP after tubal surgery, although the risk of EP after IVF treatment

ate. The aim of this study was to use a logistic model to identify patients with such an increased risk of subsequent EP and reduced chance of intra- uterine pregnancy (IUP), so that tubal microsurgery should be avoided.

Patients and methods is not negligible. To estimate the risk of subsequent EP after tubal

microsurgery a logistic model taking confounding and interaction into consideration seems appropri-

During the period 1986 to 1990,221 patients underwent tubal microsurgery at Sahlgrenska University

0 Actu Ohstet Gynecol Scarid 75 (1996)

Table iI Result of the stepwise logistic regression analysis to assess the risk of subsequent ectopic pregnancy (EP)

~~

Regression Standard Wald coefficient error Chi-

Variable ifnctori b se (b) square p-value ~

X W d P ' -1 36 033 l u m b e r of previctis EP 1 64 0 49 11 2 0 0008 Pndoii iPl i insis 1 7 2 0 8 1 4 5 0 0336

-

Hospital. Slbeden. The follow-up time varied between one and six years. The subsequent fertility i n relation t o background factors and scoring systenis ha5 recentlj been reported (12). In that study, a modified lifetable analysis demonstrated a reduced pregnancy uate already after one year and only EPs \bere v ~ n after four years. Of the 221 patients, 90 iwncei\ ed (40 .75 ). resulting in 132 pregnancies, of ivhicli 51 ere EPs. The EP rate was 16.3% ( ~ = 3 6 ) and EPs represented 40.9% of all pregnancies. Thirty-six \A.omen had at least one subsequent ectopic pregnancy, 23 had one, nine had two, three had three and one patient had four EPs. Thirty women had EP iis the only fertility outcome, while six also had additional IUPs. Fifty-four patients had IUP as the only outcomr and among these 43 had a full term pregnancy.

Eighty-four patients were included in this study atid di\,idetl into two groups: 30 with EP and 54 with It:€' a h the only outcome. Women not experi-

0 . 0 0 LC - - 2 - 1 0 1 2 3 4 5 6

I /:< I l:\iiiiwc.cl probability of ectopic pregnancy after micro- . t i i . p ! l l w c,diulated Iiiiear combination (LC) is plotted on thr. \ -a \ i \ TIw prohahilit:, of subsequent ectopic pregnancy iPtF,P:) a t i c i inicrnwrgcry can he read on the y-axis. The prob- .rbilit~ o t 1 .I) c t l t i a l ~ ;I 100V ri\k of ectopic pregnancy.

encing subsequent pregnancies ( P I = 13 I ) or with both EP and IUP (11=6) were excluded.

The following factors were analysed to estimate the risk of subsequent EP: background factors. scores for tubal lesions, scores for adnexal adhesions, risk score of EP and surgical procedures.

Background factors, e.g. previous obstetric outcome, abdominal surgery, anamnestic data of PID, history of endometriosis and previous use of an intra-uterine contraceptive device (IUCD). were regis- tered for each patient.

A risk score for EP (14). based on four background factors: previous EP. previous abdominal surgery, history of infertility and the presence of an IUCD was calculated for each patient. By multiplying a given value for each factor and adding the products, a risk score value can be obtained and the absolute risk for EP calculated. A risk score above 1.75 implies an increased risk for EP in the general population as for the individual.

Tuba1 scores and adhesion scores to predict the subsequent EP and IUP rates after tubal microsurgery were calculated separately for each patient. us- ing the classification systems according to Mage et al. (5 ) . The scoring system includes four different grades, where grade IV represents the most severe pathology.

Four different surgical procedures were used. ap- plying atraumatic techniques ( I 3): salpingostomy ( ? I = 109), adhesiolysis (tz=77), anastomosis ( ? I = 18) and implantation (ti = 17).

Statistics

For comparisons between groups of women with EP and IUP respectively, Fisher's nonparametric permutation test was used (15,24). For comparison of proportions Fisher's exact test was used. All factors were thereafter included in a stepwise logistic model (16). In the model, the probability of (the risk of) subsequent EP as the only outcome in case of pregnancy after microsurgery was estimated.

The linear combination (LC value) in the logistic model, can be determined by multiplying a given coefficient by the numerical value of the respective factor and then adding the products (Table 11). The LC value is thereafter used to estimate the risk of EP (P(EP)) as the only outcome after tubal microsurgery. By plotting the LC-value in Fig. 1 , which is the graph of the probability formula P(EP)=l/( I+e-"). the probability of EP is obtained.

Results

Thirty women with subsequent EP were compared with 54 women with IUP as the only outcome.

Risk of ectopic pregnancy after rnicrosurgev 397

To predict EP and IUP rates after fertility surgery several scoring systems have been described (5,20) and we have used the Mage systems. We found his scoring systems to be more valuable to predict the subsequent IUP rate (12), than the EP rate. Mage reported in his tubal score EP rates of 9.5% in grades 1-11 as compared with 8.8% in grades 111-IV, suggesting no correlation between tubal damage and EP-rates. Furthermore, in present- ing his adhesion score, he reported EP rates of 16.6% in patients with no adhesions as well as in patients with severe adhesions, also indicating that subsequent EP rates are not influenced by the extent of adhesions.

In this study, patients with subsequent EP presented with higher tubal scores than those with subsequent IUP. However, in agreement with Mage, there was no correlation between the extent of adhesions and the risk of a subsequent EP. The logistic model shows that neither tubal nor adhesion score contain prognostic information after allowance for the effects of previous EP and endometriosis.

In the model, the risk score for EP was not identified to have prognostic power, probably because of interaction between the four factors included in the score and the separately analysed factor 'previous EP'.

Among background factors, a history of EP was more frequently present in patients with subsequent EP, being present in every second patient, and previous EP was identified as one of the prognostic factors in the logistic model.

Previous abdominal surgery, including surgery for EP, did not have prognostic power in the model. However, previous EP and consequently previous EP surgery, had prognostic power. Therefore we suggest that all other abdominal surgical procedures are of minor importance to predict the risk of subsequent EP.

Previous endometriosis was identified to be a factor with prognostic power in the logistic model. However, the higher p-value (Table 11) of previous endometriosis compared with the extremely low p - value of the factor previous EP shows that the factor previous EP is the most important factor in the model. Several of the other factors tested in the model interact and were consequently excluded as factors possessing prognostic power themselves. Endometriosis may be a more independent factor without interaction with any other factor and thus it is included in the model. Furthermore, the factor itself can be extremely difficult to interpret since endometriosis has a wide variety of appearances and treatments. However, a history of endometriosis was found as infrequent as in 6% of patients with subsequent IUP (Table I).

Factors of importance were analyzed for possible correlation to EP. Mean values and the frequexy distribution of all factors in the two groups can be seen in Table I. Two factors were associated with a significant correlation: previous EP and risk sccre for EP.

The results of the stepwise logistic regression analysis is presented in Table 11. Two factors were selected by the logistic procedure: previous EP and endometriosis. The linear combination in the logj s- tic model is presented as follows:

LC = -1.36 + 1.64 x EP + 1.72 x endometriosis, where EP = number of previous EP and endometriosis is coded yes = 1 and no = 0.

For example, a patient with two previous EPs and endometriosis has the following LC value:

LC = -1.36 + 1.64 x 2 + 1.72 x 1 = 3.64. When the LC value is plotted in Fig. 1, the risk of

a subsequent EP as the only fertility outcome, pro- vided the patient becomes pregnant, can be esi:i- mated to be more than 95%. In a patient with one previous EP and no endometriosis, the correspon'd- ing risk is almost 60%. A women without previous EP and endometriosis still has an estimated risk (3f subsequent EP above 15%.

Discussion

Patients considered for fertility surgery may generally be regarded as high risk patients for EP, owing to the frequent presence of identical risk factors (9,10,17). High subsequent EP rates can thus be expected in a loaded study population undergoing microsurgery, as demonstrated in this study. Tuba1 surgery itself as a further risk factor for EP will add to the risk of a subsequent EP (6,19). Salp- ingostomies are reported in one study, to entail a six-fold risk and other surgical methods a 2.8-fold risk (6). Other studies do not support this increased risk of subsequent EP after salpingostomies. Win- ston reported on 9.5% EPs after salpingostomim compared with 7.5% after surgery for cornual blockage (4). Furthermore, Singhal had an EP rate of 5% after adhesiolysis, 6% after salpingostomies and 8% after tubo-cornual anastomosis (25). Ako in our report, no increased risk of EP after salpingostomies could be found as compared with other surgical procedures. Women with subsequent IUP did not differ in respect to surgical procedures compared with women with subsequent EP (Table I). However, procedures like implantation and anastomosis were less frequently performed than salpingostomies and adhesiolysis in both groups. We are not surprised by this result since the El'- rates in all four surgical procedure groups were high, as recently presented (12).

0 Actn Ohstet Gvnecol Scmrid 75 (1 996)

398 .A. Strirritlcll m i d J . T ~ ~ L ~ I - ~ I

Prc\,ioiis PID was not found to be an important factor to predict wbsequent EP after microsurgery. On the other hand, the data included only patients' reports o t clinical infections and the frequency of undiaposed subclinical infections could not be de- termi tied.

Formet- t.tsers of IUCDs are not considered to be ;it an increavd risk of later EP ( 18) and are, according to the results of this study. not at increased risk of EP ;iftei. mici-osurgery either. Since neither PID nor pre\,ioux use of IUCD was included in the model. their predictive power for a subsequent EP must be expected to be low.

Two f.:tciors with prognostic power could be iden- titied in the model: previous EP and endometriosis of Lvhich we consider previous EP to be the most important clinical factor. A prerequisite for the present model is conception and the risk of EP among patients \vho will not conceive cannot be estimated. The model is of clinical importance and can easilq he used as a tool for excluding patients tins ti i t ;I b le i or mi crosurg ery.

I t m u \ t hc: pointed out that only 40% of all our pa- ticnts undergoing tuba1 surgery conceived and EPs represcnted 40 c/i of all pregnancies. Assuming conception in ;I patient with a previous single episode of EP. there is an almost 60% risk of EP as the only fu- turc outcome. In case of both ;I previous EP and endometriosis. the risk is further increased. A note- northq finding is the increased risk (>15%) of a subsequent EP even i n patients without any of the predictite tactors. as compared with the risk in the p i e r a l population (~2%) (2 1 1.

The EP rate among IVF patients in oiir unit is be- low 554 and identification of high risk patients for EP after IVF-ET is currently undertaken in a similar study with risk estimation in a logistic model. Fac- tors predisposing to EP after IVF-ET have been sug- g e m d from other centers to be prior reconstructive surgery and pre-existing tubal pathology (22.23).

The conclusion of this investigation is that patients L b i t h previous EP should not generally be considered for microsurgery. If the same patient would ha\,c ;I reduced risk of EP after IVF-ET, this treat- inent would be a better option. Local results after IVF treatment must be considered. however. Pa- tients ~ i t h two previous EPs and a 90% risk of a fu- ture EP in case of presnancy should henceforth not be considered for tubal surgery.

Acknowledgments

T l i z sitid! u ,I\ \upported by the Toi-e Nilsson Fund. the Univer- \I[! oi' Gotelvii-~. [lie 'Medical Society of Gliteborg' and the S\\cdisli I le~licai Rewarch Council. grant No BY4- 17X-05978- t i \ ,

References

I .

2.

3.

3 .

5.

6.

7.

8.

9.

10.

1 1 .

12.

13.

14.

IS.

16.

17.

18.

19.

20.

21.

22.

23.

Morbidity and Mortality Weekly Report: Ectopic pregnancy - United States 1987. 1990: 39: 401-1. Makinen J . Ectopic pregnancy in Finland in 3966-1986. Thesis, Turku Finland 19x8 ISBN 951-880-180-0. Hallat JG. Tubal conservation in ectopic pregnancq: A study of 200 cases. Am J Obstet Gynecol 1986: 1 54: I ? 16- 21. Winston RML. Microsurgery of the fallopian tube: from fantasy to reality. Fertil Steril 1980: 33: 521-10. Mage G. Pouly JL. Bouquet de Joliniere J. Chabrand S. Ri- ouallon A. Bruhat MA. A pre-operatiic classification to predict the intraurerine and ectopic rates after distal tubal microwrgery. Fcrtil Steril 1986: 46: 807- 10. Tuomivaara L. Riinnberg L. Ectopic pregnancy and infertility following treatment of couple\: a follow-up of 929 cases. Eur J Obstet Gynecol Reprod Biol 199 1: 42: 33-8. Levin AA. Schoebauni SC. Stubblefield PG. Zirnicki S. Monson RR. Ryan KJ. Ectopic pregnancy and prior in- duced abortion. AJPH 1982: 72: 253-6. WHO. A multinational case-control study of ectopic pregnancy. Clin Reprod Fertil 1985: 3: 131-3. Thorhurn J. Philipson M. Lindblom €3. Fertilit! after ectopic pregnancy in relation to background lactors and surgical treatment. Fertil Steril 1988: 49: 595-601. Tuomivaara L. Kauppila A. Ectopic pregnancy: A case- control study of aetiological risk factors. Arch Gynecol Obstet 1988: 24-3: 5-11. Ory SJ, Nnadi E. Herrman R. O'Brien PS. Melton LJ. Fer- tility after ectopic surgery. Fertil Steril 1993: 0: 2.3 1-5. Strandell A. Bryrnan I. Janson P-0. Thorburn J . Back- ground factors and scoring systems in relation to pregnancy outcome after fertility surgery. Acta Obqet Gynecol Scand

Swolin K. The use of large doses of intraperitoneal glucocorticoids for pre\ ention of post opemi\ e adhe\ion. Acta Obstet Gynecol Scand 1967: 46: 204- 18. Thorburn J , Philipson M. Lindbloni B. Background factors of ectopic pregnancy. Risk estimation by means of a lopi\- tic model. Eur J Obstet Gynecol 1986: 23: 333-40. Bradley JV. Distribution-free \tatistical t a t s . Prentice Hall New York 1968: 78-80. SAS Institute Inc. SAYSTAT User's Guide. Cary NC. SAS Institute 1nc.Version 6 Fourth Edition. 1989: 2: 107 I - 126. Yang CP. Chow WH. Daling JR. Weis5 NS. Moore DE. Does prior infertility increase the risk of tubal pregnancy? Fertil Steril 1987: 8: 62-6. Pouly JL. Chapron C. C'aiii\ M. Mage G. W-attiez A. Man- hes H, Bruhat MA. Subsequent fertility for patients pre- senting with an ectopic pregnan nd having an intra-uterine deb ice in sir ir . Human Reprod 1991 : 6: 999- 100 1 . Parazzini F. Tozzi L. Ferraroni M. Bocciolone LA. La Vec- chia C. Fedele L. Risk factors for ectopic pregnancy: An italian case-control study. Obstet Gynecol 1992: 80: 82 1-6. Hulka JF. Adnexal adhesions: A prognostic staging and classification system based on a fi\wyear s u r ~ e q of fertility surgery results at Chapel Hill. North Carolina. . h i J Obstrt Gynecol 1982: 144: 141-7. Thorburn J. Ectopic pregnancy - clinical Feature\. r ~ s k fnc- tors and fertility prognosis. The\i\. S\4eclen I987 ISBN 9 I -

Zouves C. Erenus M. Come1 V. Tubal ectopic prepanc! after in i ,itm fertilization and embryo tran\fer: a role for proximal occlusion or salpingectomy alier failed distal tubal surgery'? Fertil Steril 1991: 56: 691-4. Dubuisson JB. Aubriot FX. Mathieu L. Foulol H. hlande-

1995: 74: 281-7.

7900-193-7.

brot L. Bouquet de Jolinikre J. Risk factors for ectopic pregnancy in 556 pregnancies after in vitro fertilization: implications for preventive management. Fertil Steril 199 1 : 56: 686-9.

24. Oden A, Wedel H. Arguments for Fishers permutation lest. Annals of Statistics 1975; 3: 518-20.

25. Singhal V. Li TC, Cooke ID. An analysis of factors influ- encing the outcome of 232 consecutive tuba1 microsurgery cases. Br J Obstet Gynaecol 1991; 98: 628-36.

26. Thorburn I. Is the ‘epidemic’ of ectopic pregnancy over in

Risk qf ectopic pregnancy after microsurger? 399

Sweden? A report on present incidence and mortality. LBkartidningen 1995; 92: 4701-6.

Address for correspondence: Aiinika Strandell, M.D Department of Obstetrics and Gynecology The Sahlgrenska University Hospital S-4 13 45 Goteborg Sweden

0 Act0 Obstet Gynecnl Scaizd 75 (1996)

Documents

Previous ectopic pregnancy should be considered a contraindication for microsurgery