The major limitation of this study is that being a cohort study, some patients recruited were lost to follow up; as such, dynamic cohorts were used. It will also have been more appropriate to do the phIGFBP-1 testing at 4-weekly intervals in order to determine the most appropriate gestational age for testing. This however was limited by cost. See appendix III.
CHAPTER FOUR RESULTS
A total of 123 women were recruited into the study over the study period. Thirty-one women were lost to follow up and another eight (8) had indicated preterm delivery mostly from severe preeclampsia (and one case of polyhydramnios). Data from a total of 84 women were therefore analyzed.
Table 1: Socio-Demographic and Obstetric Characteristics of Asymptomatic Women at Risk of Preterm Delivery at Ahmadu Bello University Teaching Hospital, Shika, 2017 Socio-Demographic
Characteristic Frequency Percentage
Age Group (years)
15 – 19 5 6.0
20 – 24 15 17.9
25 – 29 22 26.2
30 – 34 24 28.6
35 – 39 17 20.2
> 40 1 1.2
Ethnicity
Igbo 7 8.3
Yoruba 6 7.1
Others 16 19.0
Occupation
House wife 39 46.4
Students 15 17.9
Traders 12 14.3
Civil servant 4 4.8
Artisan 3 3.6
Health worker 1 1.2
Others 10 11.9
Total number of Pregnancy 1 – 4
> 5
Estimated Gestational Age 22 weeks
23 weeks 24 weeks
Total number of parity Nulliparous
1 – 4
> 5
55 29
65.5 34.5
56 18 10
66.7 21.4 11.9
26 42 16
31.0 50.0 19.0
The age of participants ranged between 16 to 42 years with a mean age of 28.74 ± 5.868 years. Majority were Hausa housewives. Of the study participants, 31% were nulliparous, and 50% with parity less than 4. Most (66.7%) were recruited at 22weeks gestation.
Figure 1: Risk Factors for Preterm Delivery in Asymptomatic Women at Risk of Preterm Delivery at Ahmadu Bello University Teaching Hospital, Shika, 2017
The commonest risk factor for preterm delivery was age less than 18 or greater than 35 as seen in 23.8% of cases. Other risk factors that were commonly encountered among participants were history suggestive of bacterial vaginosis (20.2%) and previous spontameous preterm delivery (10.7%). There was no single patient that smoked among the participants.
However, history suggestive of bacterial vaginosis was the only risk factor that showed significant association with preterm delivery (p=0.03).
0 5 10 15 20 25
Percent
Figure 2: Incidence of Preterm Delivery among Asymptomatic susceptible women at Ahmadu Bello University Teaching Hospital, Shika, 2017
Incidence of preterm delivery among asymptomatic women at risk of preterm delivery was 193/1000 women. Most (80.7%) women with a risk factor for preterm delivery ended up having term deliveries meaning having a risk factor is not sufficient to have preterm delivery.
The average gestational age at delivery among asymptomatic susceptible women is mean 38.65 ±2.14 weeks with a range of 30 to 43 weeks.
The mean cervical length at 22-24 weeks in this study was 33.71± 9.61mm with a range of 14-61mm.
Table 2: Distribution of Time of Delivery by phIGFBP-1 Status and Cervical Length among Asymptomatic Women at Risk of Preterm Delivery at Ahmadu Bello University Teaching Hospital, Shika, 2017
Time of Delivery
Total Preterm n(%) Term n(%)
phIGFBP-1 status
Positive 10 (17.9) 46 (82.1) 56 (100.0) Negative 6 (22.2) 21 (77.8) 27 (100.0)
Preterm Delivery 19%
Term Delivery 81%
Total 16 (19.3) 67 (80.7) 83 (100.0) Cervical length
Short 3 (16.7) 15 (83.3)
18 (100.0)
Normal 13 (20.0) 52 (80.0)
65 (100.0)
Total 16 (19.3) 67 (80.7)
83 (100.0) X2 (2, N=84) 0.223 p=0.637 for cervical length
X2 (2, N=84) 0.101 p=0.751 for phIGFBP-1
Majority of women with a positive phIGFBP-1 (82.1%) had term deliveries and similarly, majority of women with short cervices ≤ 25mm had term deliveries. The relative risk (RR) of having a preterm delivery with a positive phIGFBP-1 is 0.8 while the risk ratio of having preterm delivery with a short cervix is 0.83. in both instances, the risk ratio is close to 1 meaning there is little difference in risk of preterm delivery among women with positive or negative phIGFBP-1 and short or normal cervices.
Table 3: Validity of Cervical Length and phIGFBP-1 Status in predicting preterm delivery before 37 completed weeks among Asymptomatic Women at Risk of Preterm Delivery at Ahmadu Bello University Teaching Hospital, Shika, 2017
Measure of Validity
Screening Test
Cervical Length
% (95% Confidence Interval)
phIGFBP-1
% (95% Confidence Interval)
Sensitivity 18.8% (4.1% - 45.7%)
62.5% (35.4% - 84.8%)
Specificity 77.6% (65.8% - 86.9%)
31.3% (20.6% - 43.8%) Positive Predictive
Value 16.7% (6.2% - 37.8%)
17.9% (12.6% - 24.7%) Negative Predictive
Value 80.0% (75.4% - 84.0%)
77.8% (62.9% - 87.9%)
Cervical length measurement has a high specificity of 77.6% (which was statistically significant) meaning it has the ability to correctly identify those who will deliver at term;
and a high negative predictive value meaning those with a normal cervix will truly carry pregnancies to term. Using 95% confidence interval means that we are 95% certain that
of cases that a short cervix correctly identifies those who will deliver preterm; and also a low positive predictive value of 16.7% meaning that those with a short cervix wont necessarily have preterm delivery.
Table 4: Agreement of Cervical Length and phIGFBP-1 Status among Asymptomatic Women at Risk of Preterm Delivery at Ahmadu Bello University Teaching Hospital, Shika, 2017
Cervical Length
phIGFBP-1 Status
Kappa Statistic
p-value Positive Negative Total
Short 14 4 18
Normal 43 23 66 0.071
0.309
Total 57 27 84
The above table shows the measure of agreement of both tests in predicting preterm delivery in susceptible asymptomatic patients. A kappa statistic of 0.071 (which is less than 0.2) means there is poor agreement between the cervical length measurement and the result of the bedside phIGFBP-1 test.
Area: 0.507; p-value: 0.931
Figure 3: Receiver Operatin Characteristics Curve (ROC) for Evaluating the Diagnostic Ability of Cervical length in diagnosing Preterm Delivery among Asymptomatic Women at Risk of Preterm Delivery at Ahmadu Bello University Teaching Hospital, Shika, 2017.
Receiver – operating characteristics (ROC) curves was constructed and the area under the curve (AUC) was used to compare the predictive value of cervical length at 22 – 24 weeks, in predicting spontaneous preterm delivery. Here, the true positive rates (sensitivity) were plotted in function of the false positive rate (100% specificity) for the cutoff point of cervical length. The AUC obtained was 0.507. This showed that in asymptomatic patients at risk of preterm delivery, cervical length measurement is not important in predicting occurrence of preterm delivery.
CHAPTER FIVE
DISCUSSION, RECOMMENDATION, CONCLUSION 5.1 DISCUSSION
The mean age of 28.7years among study participants showed that most women were young. They were also mainly Hausa housewives, which is not unexpected considering the study was conducted in northwestern Nigeria where most women are not occupationally empowered. [12]
From this study, incidence of preterm delivery is 193/1000 (or 19.3%). This finding is lower than the 23.8% documented in a previous study [7] and higher than the quoted average of 5-13%. [9] The mean cervical length at 22-24 weeks in this study was 33.7mm which is longer than the 27mm obtained in a similar study [7] but similar to the findings of 38mm and 35mm by Heath [57] and Iams [45]
Common risk factors for preterm delivery in this study included young or late maternal age, history suggestive of bacterial vaginosis, and previous preterm delivery. Of these, only history suggestive of bacterial vaginosis was found to be statistically significant in preterm birth prediction. This finding has been corroborated in previous studies [26] where up to a nine-fold increase has been reported.
The major findings in this study show that the use of cervical length at 22-24 weeks and phIGFBP at 30 weeks, singly or in combination, do not significantly predict the occurrence of preterm delivery in asymptomatic patients at risk in our environment. This finding conflicts with other documented studies using phIGFBP-1 and cervical length. [7]
Although prior preterm birth is the most important risk factor for a subsequent preterm delivery, this study showed that majority of such patients will end up delivering at term.
This has been shown in previous studies. [7, 58]
For cervical length, with a risk ratio of having a preterm delivery in patients with a short cervix is 0.83, which is close to 1 suggests that there is little difference in risk of preterm delivery among women with short or normal cervices. Previous studies [44,46,59] have shown that cervical length is inversely related to the risk of preterm delivery in asymptomatic women but none of these studies were done among native African blacks.
The ROC curve, which is mostly to the right in this study, with AUC of 0.507, also further showed that in asymptomatic patients at risk of preterm delivery, cervical length measurement is not important in predicting occurrence of preterm delivery.
The poor sensitivities and low positive predictive values observed in this study (because preterm births were infrequent) were also reported in similar studies. [60]
Routine screening using cervical length as a predictor of preterm delivery is not recommended in low risk populations. [61]
Previous studies [62-64] have shown that transvaginal ultrasonographic cervical length measurement has been found to be effective in predicting preterm delivery in asymptomatic high risk groups such as uterine anomalies, excisional cervical procedures such as cone biopsy, and multiple dilatation and evacuation procedures greater than 13 weeks. None of these high-risk groups were assessed in this study. This is probably the reason why a statistically insignificant relationship was obtained between short cervix and preterm delivery. Similarly, the 2013 Cochrane Review for Cervical Assessment by Ultrasound for Preventing Preterm Delivery finds that cervical length measurement using transvaginal ultrasound is one of the best predictors of preterm delivery in all populations so far. They concluded that there is currently insufficient evidence to recommend routine screening of asymptomatic or symptomatic pregnant women with transvaginal ultrasound for cervical length without a specific intervention. [65]
Therefore, routine use of ultrasound for cervical length measurement remains controversial in asymptomatic women and the American College of Obstetricians and Gynecologists (ACOG) does not explicitly recommend this form of screening. [66]
Results from this study showed that there was no significant association between a positive phIGFBP-1 and the occurrence of preterm delivery (risk ratio of 0.8). many statistical tests were applied to check for any significance. This is similar to the finding of Kekki et al in 2001 [67] where no single asymptomatic woman with a positive phIGFBP-1 using quantitative assay delivered preterm whereas 41% of the symptomatic women delivered preterm. These may mean that this biomarker may be better for the prediction of preterm delivery in symptomatic women. Phosphorylated IGFBP-1 has been tested in many symptomatic women [67-70] but few studies have tested its significance in asymptomatic women. [7,67]
In checking the measure of agreement of both short cervical length of 25mm or less and
measurement is not important in predicting occurrence of preterm delivery. This was conflicting with the finding in a previous similar study. [7]
Prediction of a risk scoring system could not be done because most parameters found were not statistically significant. Research has focused on combined risk scoring systems that use multiple serum markers, ultrasound, and maternal but these have not been fully validated in large-scale studies. [45,61,71-75]