Skip to content

Advertisement

  • Research article
  • Open Access
  • Open Peer Review

A pilot study to understand feasibility and acceptability of stool and cord blood sample collection for a large-scale longitudinal birth cohort

  • 1,
  • 1,
  • 2,
  • 2,
  • 3,
  • 4,
  • 3,
  • 5,
  • 6,
  • 7, 8 and
  • 6Email author
BMC Pregnancy and ChildbirthBMC series – open, inclusive and trusted201717:439

https://doi.org/10.1186/s12884-017-1627-7

  • Received: 11 March 2016
  • Accepted: 13 December 2017
  • Published:
Open Peer Review reports

Abstract

Background

Few data are available to guide biological sample collection around the time of birth for large-scale birth cohorts. We are designing a large UK birth cohort to investigate the role of infection and the developing immune system in determining future health and disease. We undertook a pilot to develop methodology for the main study, gain practical experience of collecting samples, and understand the acceptability of sample collection to women in late pregnancy.

Methods

Between February–July 2014, we piloted the feasibility and acceptability of collecting maternal stool, baby stool and cord blood samples from participants recruited at prolonged pregnancy and planned pre-labour caesarean section clinics at University College London Hospital. Participating women were asked to complete acceptability questionnaires.

Results

Overall, 265 women were approached and 171 (65%) participated, with ≥1 sample collected from 113 women or their baby (66%). Women had a mean age of 34 years, were primarily of white ethnicity (130/166, 78%), and half were nulliparous (86/169, 51%). Women undergoing planned pre-labour caesarean section were more likely than those who delivered vaginally to provide ≥1 sample (98% vs 54%), but less likely to provide maternal stool (10% vs 43%). Pre-sample questionnaires were completed by 110/171 women (64%). Most women reported feeling comfortable with samples being collected from their baby (<10% uncomfortable), but were less comfortable about their own stool (19% uncomfortable) or a vaginal swab (24% uncomfortable).

Conclusions

It is possible to collect a range of biological samples from women around the time of delivery, and this was acceptable for most women. These data inform study design and protocol development for large-scale birth cohorts.

Keywords

  • Biological samples
  • Infant faeces
  • Cord blood
  • Feasibility
  • Acceptability
  • Large-scale birth cohorts
  • Bioarchive

Background

Large-scale cohort studies increasingly collect biological samples which, combined with survey data, might provide valuable insights into the biological mechanisms of disease [14]. Storage of samples in large long-term bioarchives has become common, with the aim of using both current and anticipated future molecular technologies to analyse the samples [5, 6]. Reductions in costs of molecular testing, in particular for genomics, proteomics and metabolomics applications, have made such approaches attractive to public health researchers [2, 7]. However, collecting such samples poses multiple challenges, including how to ensure that the proportion of participants providing samples is sufficiently high, and practical challenges associated with working in clinical settings where medical considerations often take priority over research [8, 9].

We are designing a new large-scale UK birth cohort study to investigate how exposure to microorganisms and immunological events during pregnancy and early life influence health and disease outcomes across the life course [10].

We undertook the experimental pilot study described here to: (1) inform decisions about methodology for the main study, (2) develop experience about the practicalities and logistical challenges of collecting samples, and (3) gain understanding of the acceptability of sample collection to women in the late stages of pregnancy and around the time of delivery.

Methods

Study design

We aimed to collect samples from approximately 100 women around the time of delivery at University College London Hospital (UCLH). Women were recruited between 17th February 2014 and 4th July 2014 by two research midwives, working normal office hours (9 am to 5 pm). All pregnant women with singleton pregnancies attending either a prolonged pregnancy clinic (gestation 41+5 or 6) or a planned pre-labour caesarean section clinic up to a week prior to expected delivery were eligible. We selected these groups as containing women who attend outpatient appointments and who might therefore be approached by study midwives in the week prior to labour to discuss the study. Women with multiple pregnancies, those who needed a translator to consent to participate, those under 16 years of age and non-UK residents who intended to return abroad immediately after delivery were excluded.

We aimed for a sample size of 100 participants to estimate with reasonable precision the proportion returning each of the sample types. The study was not designed or powered to detect differences by subgroups.

Sample collection

Cord blood and stool were chosen for collection as these samples were subject to the greatest uncertainty regarding the optimum processes required for collection, as well as acceptability to women. Study protocols developed as a result of this pilot study are available in Additional file 1.

Cord blood was collected in the maternity unit soon after delivery by a research midwife or by the attending midwife if the delivery occurred outside normal office hours. Cord blood banking was available at UCLH for those women who wanted to donate cord blood and this service was given precedence where women agreed to both.

Women were asked to provide a stool sample in the maternity unit before or after delivery, or stool was collected during delivery by midwives. The first baby stool sample (meconium) was collected while the baby remained in hospital. For a subset of ten women recruited in May and June 2014, an additional baby stool collection protocol was established and tested for home collection of stool every other day after birth: samples were taken by women and collected from their home by courier. These samples were used for experimental validation of our procedures. After home collection of baby stool commenced, women not enrolled in home collection were only asked to consent to collection of cord blood and maternal stool. Home collection of baby stool ended on 9th June 2014, after which all collection of maternal stool and baby stool collection stopped; only cord blood was collected from this date until the end of the study (4th July 2014). We were not able to collect reasons for failing to get a sample.

Questionnaires

Women were asked to self-complete paper-based questionnaires before (all women) and after (only women who provided samples) sample collection to assess the acceptability of collecting samples in this pilot as well as the acceptability of additional samples planned for the main study (maternal and baby urine, vaginal swabs, placental tissue and baby saliva). The pre-sample and post-sample questionnaires are available in Additional files 2 and 3, respectively.

Women were asked for their opinions about and satisfaction with the study overall as well as about the acceptability of providing each sample using a five point Likert scale (“very satisfied / comfortable” to “very unsatisfied / uncomfortable”). Women could also provide free-text comments.

Data analysis

Comparisons were made between women who did and not provide at least one sample. Comparisons were also made by sample type for women who provided at least one sample. For maternal stool, the denominator for these analyses included only those women who were approached to provide a stool sample (n = 85).

Comparisons of continuous variables, such as maternal age, were made using Student’s t-tests. Categorical variables were compared using either Chi-squared or Fisher’s exact tests. Means across groups were compared using one-way analysis of variance (ANOVA). All analyses were carried out using STATA version 12.0.

Results

Of 265 women approached, 171 (65%) of women consented to participate and 113 (66%) of these women gave at least one sample (Fig. 1). Seventeen women consented to participate but did not provide any samples or complete a pre-sample questionnaire.
Fig. 1
Fig. 1

Study recruitment and sample collection

The mean age of women who consented to participate was 34 years. Women were primarily of white ethnicity (130/166, 78%) and half were nulliparous (86/169, 51%) (Table 1). Among 168 women where mode of delivery was recorded, 57% delivered vaginally (unassisted or instrumental), nearly one third (29%) by planned pre-labour caesarean section, and 15% by intrapartum caesarean section. For 140 women who consented to participate and where information on day and time of birth were recorded, approximately half delivered outside of normal office hours (75/140, 54%).
Table 1

Demographic characteristics of women who consented to participate, and the proportion providing ≥1 sample

 

All

Sample given

P-value

Characteristic

n

%

n

%

 

Woman’s ethnicitya

 White

130

78

85

65

 

 Asian

12

7

8

67

 

 Black

14

8

11

79

 

 Mixed/Chinese/other

10

6

5

50

 

 Total

166

109

0.56

Paritya

 0

86

51

57

66

 

 1

63

37

42

67

 

 >1

20

12

14

70

 

 Total

169

 

1.00

Mode of deliverya

 Vaginal

95

57

51

54

 

  Unassisted

70

42

35

50

 

  Forceps

15

9

11

73

 

  Ventouse

10

6

5

50

 

 Planned pre-labour caesarean

48

29

47

98

 

 Intrapartum caesarean

25

15

15

60

 

 Total

168

113

< 0.001

Woman given antibiotics at any timea

 Yes

68

 

59

87

 

 No

39

 

33

85

 

 Total

107

 

92

0.76

Infant given antibioticsa

 Yes

12

 

11

92

 

 No

117

 

91

78

 

 Total

129

 

102

0.26

aIn participants where data were available

Women who delivered by planned pre-labour caesarean section were most likely to provide at least one sample (98%) and those delivering by unassisted spontaneous vaginal (50%) or ventouse were least likely (50%) to provide a sample (Table 1). In 110 women who provided at least one sample and where information on day and time of birth were recorded, 47% (52/110) delivered outside of normal office hours. We observed no significant difference between whether a sample was given and maternal age (p = 0.09, continuous variable), ethnicity (p = 0.56) or parity (p = 1.00), although we note the relatively small sample size in this pilot study, which may have been underpowered to detect any difference.

Of the 113 women from whom at least one sample was collected, maternal stool was the least frequently collected sample (26/85, 31%), while cord blood collection was most frequently collected (89/113, 78%). Midwives stopped collecting baby meconium after 6th May 2014 because this sample was found to contain no bacterial nucleic acid signature. Up to this date, 45 out of 70 women (64%) had provided a meconium sample from their baby.

Whether a maternal stool sample was collected was associated with delivery mode (p = 0.01) (Table 2). Maternal stool samples were more likely to be given by women who delivered by spontaneous vaginal delivery (14/29) and least likely to be given by women who had planned pre-labour caesarean sections (3/31) (Table 2).
Table 2

Association between demographic characteristics and sample collection for women who provided ≥1 sample

Characteristic

Total who gave any sample

Cord blood given

P-value

Total asked to provide any stool sampleb

Maternal stool given

P-value

n

%

 

n

%

 

Woman’s ethnicitya

 White

85

66

78

 

65

21

  

 Asian

8

5

63

 

5

2

  

 Black

11

10

91

 

8

1

  

 Mixed/ Chinese/other

5

5

100

 

4

1

  

 Total

109

86

0.37

84

25

0.77

Paritya

 0

57

44

77

 

45

13

  

 1

42

32

76

 

32

11

  

 >1

14

13

93

 

9

2

  

 Total

113

89

0.47

85

26

0.74

Mode of deliverya

 Vaginal

  Unassisted

35

25

71

 

29

14

48

 

  Forceps

11

7

64

 

9

4

44

 

  Ventouse

5

3

60

 

5

1

20

 

 Planned pre-labour caesarean

47

42

89

 

31

3

10

 

 Intrapartum caesarean

15

12

80

 

11

4

36

 

 Total

113

89

0.09

85

26

0.01

aIn participants where data were available

bA total of 85 women were asked to provide maternal stool

We observed no evidence of an association between collection of cord blood and any of the demographic characteristics measured (Table 2). The overall mean volume of blood collected was 12.6 ml (range: 1 to 40 ml), and this was not associated with either mode of delivery (p = 0.10: one way ANOVA).

One hundred and ten of 171 (64%) women who agreed to take part in the study completed a pre-sample questionnaire, including 41 women who did not provide a sample (Fig. 1). Over 95% of women reported that they were satisfied with the explanation about the study, why samples needed to be collected, and what taking part involved (105/110, 95%; 108/110, 98%; and 110/110, 100%, respectively). Women generally felt comfortable about samples being collected from their baby (Table 3). Most women (67%) also reported being comfortable or very comfortable with giving their own stool sample (Table 3). Fewer women felt comfortable about giving a vaginal swab (53%) (Table 3).
Table 3

Pre- and post-sample responses to the question, ‘Overall, how comfortable do you feel with the idea of providing the following samples?’

 

Pre-sample

Post-sample

Sample

n

%

n

%

Maternal stool

 Very comfortable / comfortable

72

67

27

57

 Neither

16

15

3

6

 Uncomfortable / very uncomfortable

19

18

17

36

 Total

107

 

47

 

Maternal urine

 Very comfortable / comfortable

96

97

41

93

 Neither

2

2

2

5

 Uncomfortable/Very uncomfortable

1

1

1

2

 Total

99

 

44

 

Maternal vaginal swab

 Very comfortable / comfortable

52

53

26

59

 Neither

22

22

7

16

 Uncomfortable / very uncomfortable

24

25

11

25

 Total

98

 

44

 

Cord blood

 Very comfortable / comfortable

97

87

47

96

 Neither

7

7

2

4

 Uncomfortable / very uncomfortable

7

7

0

0

 Total

108

 

49

 

Stool from baby’s nappy

 Very comfortable / comfortable

100

93

43

92

 Neither

3

3

3

6

 Uncomfortable / very uncomfortable

5

5

1

2

 Total

108

 

47

 

Urine from baby’s nappy

 Very comfortable / comfortable

91

93

 Neither

4

4

 Uncomfortable / very uncomfortable

3

3

 Total

98

 

Saliva from baby’s mouth

 Very comfortable / comfortable

82

84

 Neither

7

7

 Uncomfortable / very uncomfortable

9

9

 Total

98

 

34 women responded to both questionnaires

Forty-nine of the 113 (43%) women who gave at least one sample returned a post-sample questionnaire. Overall satisfaction with study explanations was very similar to the pre-sample questionnaires (data not shown), and most women reported that they were comfortable with sample collection in the post-sample questionnaire (Table 3).

Free text feedback was provided by 64 women. Several women reported feeling uncomfortable about the idea of having a vaginal swab collected, which they felt might be invasive and painful. Eleven women (17%) reported feeling uncomfortable about providing their own stool due to embarrassment and because this is not a routine procedure (in contrast to collection of urine). Seven women (11%) raised concerns about posting baby stool samples due to the demands of caring for a newborn baby.

Discussion

This study provides evidence about biological sample collection rates from women in late pregnancy, at around the time of delivery and from their babies, and provides insight into whether or not women feel comfortable with sample collection. Overall, sample collection success was good, with 66% (113/171) of women recruited providing at least one sample and most reporting that they found the study acceptable.

In general, samples were more likely to be provided when delivery was by planned pre-labour caesarean section (except maternal stool), probably because pre-labour caesarean sections were easier to plan and prepare for, as the time of delivery was known in advance. The Monday to Friday working pattern of the study’s research midwives may also have impacted on sample collection success; 54% of women who consented to participate delivered outside of normal office hours. A major constraint on sample collection may therefore have been this practical issue rather than a lack of willingness from women having vaginal deliveries. This is reflected in both the percent of those approached who consented to participate (171/265, 65%) and in generally high levels of acceptability for most samples in responses to pre-sample questionnaires.

The collection rate for maternal stool samples was low at 31%. This may have been due in part to our deliberate selection of women who delivered by planned pre-labour caesarean section, from whom fewer stool samples were collected, probably because passing of stool during C-section delivery is unlikely. Since fewer deliveries would be by caesarean section in a population-based study [9], we anticipate that maternal stool sample collection may be higher in our main study.

It was encouraging that the proportion of women from whom cord blood was collected was high (79%), despite the fact that cord blood banking was given precedence over our own sample collection. In the absence of cord blood banking, samples could potentially be collected from a higher number of participants, and/or larger volumes obtained.

Overall, the study and procedures were acceptable to most participants, with particularly high acceptability of collection of maternal and baby urine, baby stool and cord blood. We were surprised that the collection of vaginal swabs was perceived to be painful to collect or too invasive because such swabs have been shown to be highly acceptable in other settings such as sexual health clinics and for cervical smear screening [11, 12]. Such concerns may be alleviated by enabling self-collection and providing more information about what the procedures involve.

A limitation of this study is that it was carried out in specific late pregnancy clinics in a central London teaching hospital, and the women recruited might therefore not be representative of pregnant women in the general population. Moreover, some selection bias is possible, with those women who participated being more likely to find the study acceptable. For a small number of participants, some data items were missing. Due to the scale of this pilot we were not able to include women who needed a translator to understand the study materials. For women who consented to take part but did not provide a samples, we were unable to determine whether this was because their changed their minds or because they lacked an opportunity to provide the sample.

Our pilot found similar or somewhat better cord blood collection rates than other studies. For example, a pilot study for the ELFE longitudinal cohort in France reported cord blood collection for 82% of participating mothers, which was similar to our study, [13] whereas a feasibility study in Germany reported cord blood collection for 54% of women who consented [6]. Few gut microbiota studies report response rates or information about acceptability of stool collection, and most to date have been small in scale [14, 15]. Where reported, stool sample response rates have varied considerably between studies. For example, the Canadian CHILD study reported 3-month and 1-year stool samples collected at home by visiting nurses were available for 33% (422/1264) babies, [16] while stool samples collected from children aged 1–11 months were available for 24% of participants in the American WHEAL cohort study, [17] and a feasibility study in Germany in 75 1- to 3-year-old children reported stool sample for 65.3% of participants [9]. Overall, we found few examples in the literature of dedicated studies describing the acceptability, responses rates and challenges when collecting cord blood and stool from mothers and babies around the time of birth, which emphasises the importance of the data reported here.

Conclusions

Overall, our study indicates that it is acceptable and possible to collect biological samples from pregnant women in a clinical setting using research midwives and routine National Health Service (NHS) staff, which has important implications for study design and protocol development for large-scale birth cohorts collecting biological samples around the time of birth. The major constraints to sample collection are likely to be logistical rather than related to acceptability and willingness to participate.

Abbreviations

ANOVA: 

Analysis of variance

NHS: 

National Health Service

UCLH: 

University College London Hospital

Declarations

Acknowledgements

The authors would like to thank the study participants and staff at University College London Hospital for their time and contribution to the study.

Funding

This pilot study was funded by a Wellcome Trust Strategic Award to University College London (UCL), grant number WT101169MA. NF was funded by a National Institute for Health Research (NIHR) Academic Clinical Lectureship.

Availability of data and materials

The dataset supporting the conclusions of this article is not being made publically available to avoid potential deductive disclosure because the data are from a defined time period at one named hospital with small numbers of patients.

Authors’ contributions

SRB co-wrote the study protocol, managed the data and performed all data analysis, and wrote the first draft of this article. CLT obtained funding, co-conceived the study, contributed to interpretation of results and revisions of the article. HD and CM co-wrote the study protocol and collected all participant data. ET co-wrote the study protocol and contributed to revisions of the article. EMR obtained funding, co-conceived the study, contributed to study design and protocols and revisions of the article. MN and TL obtained funding, co-conceived the study and contributed to study design and protocols. AR and PB (principal investigator) obtained funding, co-conceived the study, contributed to study protocols, contributed to the analysis, interpretation of results and revisions of the article. NF obtained funding co-conceived the study, supervised data collection, contributed to the analysis, interpretation of results and revising the manuscript critically for intellectual content and co-wrote study protocols. All authors read and approved the final version of the manuscript for publication.

Ethics approval and consent to participate

The study was approved by the London City and East Research Ethics Committee (REC number 12/LO/1492). Participants provided written consent to take part in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

(1)
UCL Institute of Child Health, University College London, London, UK
(2)
University College London Hospital, London, UK
(3)
Division of Infection and Immunity, University College London, London, UK
(4)
Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, UK
(5)
Wellcome Trust Sanger Institute, Cambridge, UK
(6)
Institute for Global Health, University College London, London, UK
(7)
Institute for Women’s Health, University College London, London, UK
(8)
Institute of Applied Health Research, University of Birmingham, Birmingham, UK

References

  1. Hobcraft J. Reflections on the incorporation of biomeasures into longitudinal social surveys: an international perspective. Biodemography Soc Biol. 2009;55:252–69. 10.1080/19485560903382460.View ArticlePubMedGoogle Scholar
  2. Richmond RC, Al-Amin A, Smith GD, et al. Approaches for drawing causal inferences from epidemiological birth cohorts: a review. Early Hum Dev. 2014;90:769–80. 10.1016/j.earlhumdev.2014.08.023.View ArticlePubMedPubMed CentralGoogle Scholar
  3. Kuh D, Pierce M, Adams J, et al. Cohort profile: updating the cohort profile for the MRC National Survey of health and development: a new clinic-based data collection for ageing research. Int J Epidemiol. 2011;40:e1–9. 10.1093/ije/dyq231.View ArticlePubMedPubMed CentralGoogle Scholar
  4. Palmer LJ. UK Biobank: bank on it. Lancet Lond Engl. 2007;369:1980–2. 10.1016/S0140-3656736(07)60924-6.View ArticleGoogle Scholar
  5. Jackson C, Best N, Elliott P. UK biobank pilot study: stability of haematological and clinical chemistry analytes. Int J Epidemiol. 2008;37:i16–22. 10.1093/ije/dym280.View ArticlePubMedGoogle Scholar
  6. Medical Research Council. Maximising the value of UK population cohorts MRC strategic review of the largest UK population cohort studies. 2014. https://www.mrc.ac.uk/publications/browse/maximising-the-value-of-uk-population-cohorts/. Accessed 8 Jun 2017.
  7. Ng JWY, Barrett LM, Wong A, et al. The role of longitudinal cohort studies in epigenetic epidemiology: challenges and opportunities. Genome Biol. 2012;13:246. 10.1186/gb-2012-13-6-246.View ArticlePubMedPubMed CentralGoogle Scholar
  8. Ernst SA, Günther K, Frambach T, et al. Prenatal recruitment of participants for a birth cohort study including cord blood collection: results of a feasibility study in Bremen, Germany. Ger Med Sci GMS E-J. 2015;13:Doc04. 10.3205/000208.Google Scholar
  9. Kirkland SA, Raina PS, Wolfson C, et al. Exploring the acceptability and feasibility of conducting a large longitudinal population-based study in Canada. Can J Aging Rev Can Vieil. 2009;28:231–42. 10.1017/S0714980809990043.View ArticleGoogle Scholar
  10. Field N, Townsend CL, Rodger AJ, et al. Infection and immunity from a lifecourse perspective: life study enhancement. Lancet. 2013;382:S35. 10.1016/S0140-6736(13)62460-5.View ArticleGoogle Scholar
  11. Braz NSDF, Lorenzi NPC, Sorpreso ICE, et al. The acceptability of vaginal smear self-collection for screening for cervical cancer: a systematic review. Clinics. 2017;72:183–7. 10.6061/clinics/2017(03)09.View ArticlePubMedPubMed CentralGoogle Scholar
  12. Paudyal P, Llewellyn C, Lau J, et al. Obtaining self-samples to diagnose curable sexually transmitted infections: a systematic review of patients’ experiences. PLoS One. 2015;10:e0124310. 10.1371/journal.pone.0124310.View ArticlePubMedPubMed CentralGoogle Scholar
  13. Oleko A, Betsou F, Sarter H, et al. A pilot study of the ELFE longitudinal cohort: feasibility and preliminary evaluation of biological collection. Biopreservation Biobanking. 2011;9:223–7. 10.1089/bio.2010.0032.View ArticlePubMedPubMed CentralGoogle Scholar
  14. Chu DM, Ma J, Prince AL, et al. Maturation of the infant microbiome community structure and function across multiple body sites and in relation to mode of delivery. Nat Med. 2017;23:314–94. 26View ArticlePubMedPubMed CentralGoogle Scholar
  15. Yassour M, Vatanen T, Siljander H, et al. Natural history of the infant gut microbiome and impact of antibiotic treatment on bacterial strain diversity and stability. Sci Transl Med. 2016;8:343ra81. 10.1126/scitranslmed.aad0917.View ArticlePubMedPubMed CentralGoogle Scholar
  16. Arrieta M-C, Stiemsma LT, Dimitriu PA, et al. Early infancy microbial and metabolic alterations affect risk of childhood asthma. Sci Transl Med. 2015;7:307ra152. 10.1126/scitranslmed.aab2271.View ArticlePubMedGoogle Scholar
  17. Fujimura KE, Sitarik AR, Havstad S, et al. Neonatal gut microbiota associates with childhood multisensitized atopy and T cell differentiation. Nat Med. 2016;22:1187–91. 10.1038/nm.4176.View ArticlePubMedPubMed CentralGoogle Scholar

Copyright

© The Author(s). 2017

Advertisement