Summary. A series of women who aborted their pregnancies at King. Khalid University Hospital (KKUH), Riyadh, Saudi Ara- bia, over a period of 12 months (1 ...
Journal of Obstetrics and Gynaecology (1998) Vol. 18, No. 2, 159± 163
GYNAECOLOGY
Pregnancy trends after abortion B. ADELUSI, E. A. BAMGBOYE, N. CHOWDHURY and L. AL-NUAIM King Khalid University Hospital Riyadh, Saudi Arabia
Sum mary A serie s of wom en who aborte d their pregnan cies at King K halid Universit y H ospita l (KKU H), Riyadh , Saudi A rabia, over a perio d of 12 month s (1 Januar y 199 2 to 3 D ecem ber 1992), were follow ed up for 6±48 m onths after w ards to exam ine their pregnan cy trends. Descripti ve statistic s for data presenta tion w ith Kaplan- Meier survivor ship functio n estim ates tim e from abortio n to the nex t pregnan cy and Cox proport ional hazar d regressio n analysi s w ere used to identif y progno stic factors. Eighty-o ne per cent of the w omen achieve d pregnan cy over a 4-year follow -up. T he m edian tim e to pregnan cy w as 7´6 m onths, w ith a 95% con® denc e interva l of 6´4±10´0 m onths. Age and national ity were foun d to be signi® cant prognos tic factor s related to tim e to achiev e a pregnan cy. F urther more, the outcom e of the subseque nt pregnan cy was found to be statistica lly relate d to the num ber of previou s abortion s. T here is a high chance of fertilit y after an abortio n, w ith 75´2% of these occurri ng within the ® rst 12 m onth s of the post-abo rtio n follo w-up perio d and dwindlin g to only 4´1% by the end of 36 m onths. U sing m ultivari ate regressi on analysis , only age seem ed to be an im portan t deter m inan t factor.
Introduction Spontaneous abortion remains a problem for many wom en. It is estimated that this constitutes a major obstetric disaster in about 15% of wom en in the reproductive age (Wilcox et al., 1988; Knudsen, et al., 1991) . Other reports have shown as high as 31±78% cases of early pregnancy losses in some instances (Wilcox et al., 1988; Modvig et al., 1990; McBrede, 1991) . The psychological traum a to which the patient who has recently sustained a spontaneous fetal loss can be subjected, may be such as to require some psychotherapy (Poland et al., 1977). Increase in the risk of abortion after a previous abortion has been reported consistently (Goldzieher and Benigno 1958 ; Risch et al., 1988 ; Knuolsen et al., 1991; Hogberg et al., 1992) . Even then, the reported risks of recurrence vary from report to report, with no generally accepted ® gures (Parazzini et al., 1988; Regan, 1988) . These differences may be due to the differences in de® nition of spontaneous abortion, patient selection, the data collection methods and the often small study num bers which consequently give broad con® dence limits (Knudsen et al., 1991). On the other hand, the pregnancy trends in wom en after a previous abortion have hardly been looked
into, as far as our literature search could show. T he objective of the present study, which is based on a prospective follow-up of patients who spontaneously aborted their pregnancies over a period of 12 months, was to examine trends in pregnancies in these wom en. Furthermore, the viability or non-viability of the subsequent pregnancies achieved were examined.
Materials and methods The cohort of wom en whose pregnancies terminated in spontaneous abortions in the Obstetric Unit of the King Khalid University Hospital (KKUH) between 1 January 1992 and 31 December 1992 , and who were desirous of getting pregnant were followed up by two of the authors (B. A., L. N.) for periods varying from 6 to 48 months until December 1995 . A structured data collection form was designed to collect information on the demographic and obstetric history of the patients. The interval (in months) between the current abortion under study and the subsequent pregnancy was determined, and the outcome of the pregnancy was categorised as spontaneous abortion, or viable pregnancy. Spontaneous abortion, as used in this study, refers to the unexpected unplan ned spontaneous expulsion of a fetus before viability (less than 24 weeks of gestation or weighing less than 500 g). Any pregnancy that reaches 24 weeks or more ( $ 24 weeks) of gestation, or weighs 500 g or more ( $ 500 g) is regarded as viable. Statistical analysis The pregnancy rates in relation to the demographic characteristics of wom en who achieved pregnancy and those who were not known to have been pregnant during the study period were assessed. Cumulative survival after the abortion was calculated, with the Kaplan-Meier procedure. T he duration of survival (tim e taken to achieve pregnancy) was de® ned as the tim e from the index abortion to the date pregnancy was achieved, or when pregnancy was con® rmed biochemically, in cases where there was no menstrual period after the abortion. Cox’ s propor tional hazard regression model was used to study predictors of pregnancy survival. A test of the association of propor tional hazards of independent variables showed no signi® cant evidence of the violation of non-pr oportionality for the independent variables. The analyses were limited to the following information: age, nationality, education, occupation,
Correspo ndence to: D r B. Adelusi , Departm ent of O bstetric s and G ynaecol ogy , K ing Khalid Universit y Hospital , P.O . Box 7805, Riyad h 11472 , Saudi Arabia. 0144-3615/98/020159-05 $9.50
ã
Institute of Obstetrics and Gynaecology Trust, 1998
160
B. Adelusi et al.
parity, history of previous abortions, number of previous abortions, medical complications of previous pregnancy, gestational period at abortion and the outcome of the current pregnancy. All analyses were performed with the BMDP statistical package.
Results There were a total of 273 wom en with recent abortions in the hospital during the study period, all of whom were recruited into the study. T he gestational age of the pregnancies when abortion occurred ranged from 5 weeks to 23 weeks, with a mean of 11´9 6 3´8 weeks. The younge st was 15 years old and the oldest 45 years old (mean 5 28´4 6 6´4 years). Their parity ranged from 0 to 12 with a mean of 2´8 6 2´7. A large propor tion (83´8% ) were Saudis, and a lesser proportio n (26´0% ) had no formal education. Nearly
81´2% were housewives and 118 (43´2% ) had a history of previous abortions. The cumulative pregnancy rate after 48 months, being the longest observed time interval to pregnancy from the beginning of the follow-up period, for all the patients was 81´3% . T he median time to pregnancy was 7´6 months, 95% con® dence interval (6´4 months, 10´0 months). Table I shows the propor tion of the wom en who achieved pregnancy in relation to the different categories of their personal demographic characteristics. The pregnancy rates fell with age, but not with education or occupation. In each category, the greatest number (167 or 75´2% ) of pregnancies occurred within the ® rst 12 months of follow up, and the longer the follow up withou t pregnancy, the fewer the number (9 or 4´1% ) who achieved pregnancy. After 36 months, no pregnancies were recorded amongst those who had not becom e pregnant. Overall in any
Table I. Distribution of pregnancy by personal characteristics in women with abortion Pregnancy time interval (months) Characteristics
No. of women (273)
No. pregnant
12
24
36
% pregnant
Age (years) , 20 20± 24 25± 29 30± 34 $ 35
20 60 77 64 52
18 54 64 53 33
15 41 51 40 20
1 11 13 11 10
2 2 0 2 3
90´0 90´0 83´1 82´8 63´5
Nationality Saudi Non-Saudi
204 69
171 51
135 32
30 16
6 3
83´8 73´9
Education None Primary Secondary Tertiary
71 69 80 53
51 60 66 45
40 44 52 31
10 13 10 13
1 3 4 1
71´8 87´0 82´5 84´9
213 60
173 49
129 38
38 8
6 3
81´2 81´7
Occupation Housewife Working woman
Table II. Pregnancy distribution by previous reproductive history in women with abortion Pregnancy time interval (months) No. of women (273)
No. pregnant
12
24
36
% pregnant
Parity 0 1± 4 . 5
66 143 64
56 118 48
42 92 33
12 22 12
2 4 3
84´8 82´5 75´0
History of antecedent abortions Yes No
118 155
88 134
68 99
18 28
2 7
74´6 86´5
81 37
61 27
48 20
12 6
1 1
75´3 73´0
235 38
193 29
145 22
40 6
8 1
82´1 76´3
Characteristics
Number of previous abortions 1 $ 2 Medical complications of previous pregnancy None Yes
Pregnancy trends after abortion
Table III. Summary of time taken to achieve cumulative pregnancy by personal characteristics in women with abortion
Variable
No. of women (273)
Percentile of time to pregnancy (month) 25%
50%
75%
Age (years) , 20 21± 24 25± 29 30± 34 $ 35
20 60 77 64 52
2´8 2´5 2´5 3´3 5´3
4´7 4´6 5´4 6´4 11´0
10´7 11´3 9´9 12´0 17´7
Nationality Saudi Non-Saudi
204 69
3´3 4´8
6´9 10´6
14´1 22´2
Education None Primary Secondary Tertiary
71 69 80 53
2´5 3´6 2´4 2´6
6´5 7´4 6´7 5´6
14´3 15´5 10´9 14´0
Occupation Housewife Worker
213 60
2´8 2´6
6´8 4´9
14´0 13´3
category of the variables, the propor tion that achieved pregnancy over the entire study period of 48 months was more than 60% . In relation to the previous reproductive characteristics, on the other hand (Table II) the propor tion that achieved pregnancy over the study period was more than 70% . T he pregnancy rates fell in relation to the parity, number of previous abortions, and medical complications. As before, the number of pregnancies in the ® rst 12 months of follow-up were highest in each category. After 36 months of follow-up, there were no further pregnancies in those who had not become pregnant. Table III shows the percentiles of the time to pregnancy in each of the categories of the personal characteristic variables. Strati® cation of pregnancy time by some demographic variables revealed signi® cant differences between the various groups. The 25th, 50th and 75th percentiles of time to pregnancy increased with increasing age, and were higher in non-Saudis. How ever, there was no consistent trend in these with regards to the level of education. W ith regards to the previou s reproductive history, there was a trend towards increased percentiles of time to pregnancy in relation to increasing parity, and presence of medical complications in the previous pregnancy (Table IV). Those who had tw o or more previous abortions, on the other hand, had shorter pregnancy time interval. There was no consistent pattern however when the question is whether or not there was ever a history of abortion. Table V shows the outcome of subsequent pregnancies in relation to the number of previous abortions. There were more (73´5% ) viable pregnancies among the wom en who have had no history of abortions, and the number of viable pregnancies decreased with an increase in the number of previous abortions. On the other hand, the propor tion of abortions increased with the number of previous abortions. There was a statistically signi® cant association
161
between the outcome of pregnancy and the number of previous abortions (P 5 0´04). In a forward stepw ise Cox regression analysis; age, history of im mediate antecedent abortion, number of previous abortions, parity, presence of any pregnancy complications, nationality and type of abortion, (whether complete or incomplete) were entered as covariates. Only age and nationality were found to have statistically signi® cant independent progno stic effects on the time it takes to achieve pregnancy, after adjustm ent for all other variables (P , 0´01). Older wom en and non-Saudis had delayed time interval to achieve pregnancy, and lower propor tion of them achieved pregnancy at the end of the follow-up period.
Discussion W ith a 81´3% of the wom en achieving pregnancy over the 48-month period of follow-up, the study shows that the chances of fertility in these wom en after an abortion, are very high. This high percentage of pregnancies following an abortion in this area may be predicated on the absence of induced (criminal) abortions amongst these patients, and the post-abortal morbidity associated with such abortions (Rogo, 1993) . For example, it has been shown that wom en becoming pregnant after termination faced 10 times the normal risks of mid-trimester abortion (Of® ce of Population Cesuses and Surveys, 1995) and ectopic pregnancy (Panayotou et al., 1972) . Furthermore, the aseptic and antiseptic conditions prevailing in the management of the cases of abortion in the hospital would tend to reduce m orbidit y, and therefore im prove the pregnancy potential and fertility following the abortion. It was observed that cumulative pregnancy rate was highest in the ® rst 12 months after the abortion, where 75´2% of pregnancies were achieved within this period. T his is in agreement with the ® nding s of Oelsner et al. (1987 ) in their study pertaining to reproductive performance after conservative treatment for tubal pregnancy. After this period, the rates dwindled to 4´1% by the end of 36 months of followup, and all those who have not been able to achieve pregnancy by this time failed to do this even by the end of the study. It is probable that the age of the wom en may play a part in this (Hull, 1995) , especially when it is realised that 90% of pregnancies occurred in the younge r wom en compared with 63´5% in wom en $ 35 years. This factor of age is further highlig hted when it is realised that a larger propor tion of the younge r wom en achieved pregnancy at a faster rate than the older ones, especially when viewed from the point of a median (50th percentile) time to pregnancy of 7´6 months. For example, wom en aged , 25 years had a median time interval much less than the median of 7´6 months, and achieved a higher propor tion of pregnancies, when compared with the older wom en $ 30 years who had a longer interval than the median and a lesser propor tion of pregnancies. The same higher propor tion of pregnancies goes for those with one previous abortion, and those who had no antecedent abortion prior to the current one, although those who had multiple abortions had shorter pregnancy times.
162
B. Adelusi et al.
Table IV. Summary of time taken to achieve cumulative pregnancy by previous reproductive history in women with abortion Percentiles of time to pregnancy (months)
No. of women (273)
Variable
25%
50%
75%
Parity 0 1± 4 . 5
66 143 64
2´4 2´8 3´6
4´7 6´7 10´2
12´4 12´5 16´6
Medical complications None Yes
235 38
2´8 3´2
6´3 8´2
13´3 14´6
History of abortion Yes No
118 155
3´2 2´5
7´4 4´9
14´0 4´0
81 37
3´3 2´9
8´2 7´1
14´3 11´8
Number of previous abortions 1 $ 2
Table V. Pregnancy outcome in relation to history of previous abortions Pregnancy outcome None
Abortion
Viable
n
%
n
%
n
%
Total
Previous abortion None 1 $ 2
21 20 10
(13´5) (24´7) (27´0)
20 12 8
(12´9) (14´8) (21´6)
114 49 19
(73´5) (60´5) (51´4)
155 81 37
Total
51
(18´7)
40
(14´6)
182
(66´7)
273
Chi-squared 5
9´55; DF 5 4; P 5
0´04.
The im plication of this may be the fact that the ovaries in the younge r patients still contain the optimum number of follicles and they are more likely to ovulate faster, and therefore get pregnant faster, all other conditions being equal. It is known for example that natural fecundity in wom en declines gradually with age, and this decline accelerates sometime between 35 and 40 years, with fecundity reaching almost zero by 45 years (Hull, 1995) . Sim ilarly, it is possible that the ones with multiple abortions may be taking urgent steps to compensate for their losses by getting pregnant im mediately. Since most of the patients who got pregnant did so during the ® rst 12 months of follow up, it may be appropriate therefore to recommend that patients should try to conceive as soon as possible after an abortion. The ability of the Saudi wom en to achieve pregnancy at a shorter interval than the median time of 7´6 months may have to do with the fact that the majority of them are housewives who do not work, and therefore take reproduction as the main pre-occupation, especially with the desire for large families in this popula tion. This is in contrast to the non-Saudis who work, and m ore often than not, would like to space and limit their families. Indeed, m any of these may be on contraception and therefore, their pregnancy interval would be expected to be longer as a result. Many
of these wom en may not wish to have a large family in any case. It is no wonder therefore that not only do the older wom en and Non-Saudis have delayed time to pregnancy, a lower propor tion of these wom en also achieved pregnancy at the end of the follow-up period. Unfortun ately, we have no similar studies with which to compare these results. The confounding effect of previous abortions on the outcome of subsequent pregnancies, particularly when this is multiple, has been aptly demonstrated in this study. There was a tendency for a higher proportio n of abortions in those patients with history of multiple abortions in the past compared with those with only one or no abortion. T his is in agreement with the studies of others (W arburton and Fraser, 1964; Risch et al., 1988) . The suggestion is that recurrence of abortion may have im munologi cal or endocrinolog ical explanations (Regar et al., 1989.) whereby the achievement of a successful pregnancy prior to the current one, tends to exert a protective in¯ uence on it. Of course, one must take note of the small number of patients in the analysis, for statistical purposes. W hatever the explanations for our ® ndings , there may be a need for a larger number of patients to be studied, and over a longer period, in order to do justice to this issue. However, it would appear, using
Pregnancy trends after abortion
regression analysis, that age is a very signi® cant determinant factor in achieving pregnancy after a history of previous abortion, at least in Saudi Arabia.
Acknowledgement W e w ould like to than k Am ir S . Marzou k for his technica l assistance .
References Counci l on Scienti® c Affairs, Am erican M edical Associatio n (1992) Induce d termination of pregnan cy befor e and after Roe and W ade: T rends in the mortalit y and morbidit y of wom en. Journa l of the A merica n M edica l Associatio n, 268 , 3231 ±3239 . G oldziehe r J. W . and Benign o B. B. (1958 ) The treatm ent of threaten ed and recurre nt abortion : a critica l review . Am erica n Journa l of Obstetric s and Gynecolo gy, 75, 1202±1214. H ogber g U ., Sandstro m A. and N ilsson N . G. (1992 ) Reproductive pattern s am ong S wedish wom en born 1936±1960. A cta Obstetrici a et Gynecolog ica Scandi navica , 71, 201±204. H ull M . G. R. Fem ale age and fecundit y. In: F ertility and Sterility Ð A Curren t Overvie w. P roceedin gs of the 15th W orld Congres s of F ertility, edited by H endo n B., Bringer, J. and Manes, P . pp. 181 ±185 . K nudse n U. B., Hansen V ., Juul S. and S echer N. J. (1991 ) Prognosi s of a new pregnan cy follo wing previou s spontane ous abortio ns. E uropea n Journa l of O bstetrics , Gynecolog y, and R eproduct ive B iolog y, 39, 31±36. M cBride W . Z. (1991 ) Spontane ous abortio n. A merican Family P hysicia n, 43, 175±182. M odvig J., Schm idt and L . D am sgaarc l M. T . (1990 ) Measure m ent of total risk of spontane ous abortion : the virtu e of conditi onal risk estim ation. Am erica n Journa l of E pidem iolog y, 132, 1021 ±1038. O elsne r G ., Morad J., Carp H., M ashiac h S. and Sean D. M. (1987) Reproduc tive perfor m ance follo wing conservativ e micro-sur gical m anagem ent of tubal preg-
163
nancy. British Journa l of Obstetric s and G ynaecolo gy, 94, 1078 ±1083 . Of® ce of Populatio n Censuse s and Survey s (1995 ) A bortion Statistics , Legal abortion s carrie d out unde r the 196 7 A bortio n A ct in England and W ales, 1993 . OPCS S erie s AB, N o. 20. L ondon , Her Majesty ’ s Stationer y O f® ce. Panayoto u P. P ., Kaskareli s D. B., Miettine n O. S., T richo poulos D . B. and Kalandid i A. K. (1972 ) Induce d abortion and ectopi c pregnan cy. A merica n Journa l of O bstetric s and Gynecolog y, 114 , 507±510. Parazzin i F ., Acaia B., Ricciaidie llo O., Fedele L., L iati P . and Candian i G . B. (1988 ) Short-te rm reprodu ctiv e prog nosis when no cause can be found for recurre nt m iscar riage. British Journa l of Obstetric s and G ynaecolo gy, 95, 654±658. Polan d B. J., Miller J. R., Jone s D . C. and Trim ble B. K. (1977 ) Reproduc tive counsell ing in patient s w ho have had a spontane ous abortio n. A merica n Journa l of Obstetrics and Gynecolo gy, 127 , 685±691. Regan L . (1988 ) A prospect ive stud y of spontane ous abortion . Early pregnan cy loss: m echanis m s and treatment. In: Proceedin gs of the 18th Stud y Group of the R oyal Colleg e of O bstetrici ans and Gynecolog ists, edited by Bead R. W . and Sharp F., pp. 23±37. Regan L ., Braude P. R. and T rem bath P. L. (1989 ) In¯ uence of past reprodu ctive perfor mance on risk of spontan eous abortion s. British M edica l Journa l, 299, 541±545. Risch H . A ., W eiss N. S., Clarke E . A. and M iller A . B. (1988 ) Risk factor s for spontane ous abortio n and its recurre nce. Am erica n Journa l of E pide miolog y, 128, 420±430. Rogo K. (1993) Induce d abortio n in sub- Saharan Africa . E ast A frica n M edical Journa l, 70, 386±395. W arburto n D . and Frase r F. C. (1964 ) Spontane ous abortion risk s in m an: data from reprodu ctiv e historie s collected in a m edical genetic s unit. A merica n Journa l of H um an Genetic s, 16, 1±25. W ilco x A. J., W einber g C. R., O ’ Conor J. F., Baird D. D., S chlattere r J. P., Can® eld R. E ., A rm stron g E . G. and N isula B. G. (1988 ) Inciden ce of early loss of pregnan cy. New E nglan d Journa l of M edicin e, 319 , 189±194.
