Nicotine & Tobacco Research, 2016, 79–83 doi:10.1093/ntr/ntv087 Brief report Advance Access publication April 20, 2015
Brief report
Cotinine Validation of Self-Reported Smoking During Pregnancy in the Swedish Medical Birth Register Downloaded from http://ntr.oxfordjournals.org/ at University of Wollongong on May 21, 2016
Kristina Mattsson MSc1, Karin Källén PhD1, Anna Rignell-Hydbom PhD1, Christian H. Lindh PhD1, Bo A.G. Jönsson PhD1, Peik Gustafsson PhD2, Per Olofsson PhD3, Sten A. Ivarsson PhD4, Lars Rylander PhD1 Division of Occupational and Environmental Medicine, Institute of Laboratory Medicine, Lund University, Lund, Sweden; 2Department of Child and Adolescent Psychiatry, Institution of Clinical Sciences, Lund University, Lund, Sweden; 3Department of Obstetrics and Gynecology, Institution of Clinical Sciences, Skåne University Hospital, Lund University, Malmö, Sweden; 4Department of Clinical Sciences, Unit of Pediatric Endocrinology, Lund University/ Clinical Research Centre, Malmö, Sweden 1
Corresponding Author: Kristina Mattsson, MSc, Division of Occupational and Environmental Medicine, Institute of Laboratory Medicine, Lund University, SE-221 85 Lund, Sweden. Telephone: 46-46-222-1638; Fax: 46-46-17-36-69; E-mail: [email protected]
Abstract Introduction: Self-reported data on smoking during pregnancy from the Medical Birth Register of Sweden (MBR) are widely used. However, underreporting of such behavior may occur, leading to biases. It is of importance to validate the smoking data in the MBR. The main objective was to investigate the agreement between self-reported smoking data from the MBR and cotinine levels in maternal serum among women from the general population in the region of Skåne, Sweden. We also estimated the transfer of cotinine from mother to fetus. Methods: From a cohort used previously to investigate the relationship between intrauterine environmental exposures and offspring neuropsychiatric outcomes, there were 204 control children retrieved from the MBR with data on maternal smoking in early pregnancy registered. Data on maternal and umbilical cord cotinine at delivery were available for these children from a regional biobank. Results: There was a high agreement between cotinine levels and MBR smoking data (κ = 0.82) and a high correlation between cotinine levels in maternal and umbilical cord serum (rs = 0.90, P < .001). Of the self-reported nonsmokers, 95% (95% confidence interval: 89% to 97%) were classified as nonsmokers after cotinine measurements. Conclusion: In these data, we found that the agreement between mothers’ self-reported smoking habits during pregnancy and their levels of serum cotinine was high, as was the transfer of cotinine from mother to fetus. This indicates that birth register data on pregnancy smoking in Sweden could be considered a valid measure.
Introduction
the Swedish Medical Birth Register (MBR), information on first-
For the study of different health outcomes related to maternal smoking during pregnancy, self-reported data from birth registers are often used, due to their accessibility and cost-effectiveness. In
trimester pregnancy smoking has been collected by midwives during the maternal health care visits since 1982. Due to the stigma surrounding pregnancy smoking, there is a risk of underreporting
© The Author 2015. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected].
79
80
Methods Participants This study uses data from the Fetal Environment and Neurodevelopment Disorders in Epidemiological Research project (The FENDER project). No woman in the present population was asked to participate in a specified study for the acquisition of these data. The cohort for this analysis is comprised by an earlier control group and data come exclusively from national registries and a regional biobank. A more comprehensive description of methods can be found elsewhere.9 In Malmö, a city of circa 300 000 inhabitants, nearly all deliveries occur at the Malmö University Hospital Maternity Unit. Malmö is the largest town/municipality in Skåne, which is the southernmost county in Sweden. The Malmö Maternity Unit Serum Biobank has been storing umbilical cord serum and maternal serum, both taken at delivery from women giving birth at this unit, between 1969 and 2000, and now has samples from approximately 70 000 deliveries. The control children used in this study were randomly selected from the MBR to match children with attention-deficit-hyperactivity disorder born 1978–2000, according to year of birth and maternal country of origin. The children were chosen in pools of 10 matched eligible controls, and the first baby with an available serum sample in the biobank was chosen. As smoking during pregnancy was recorded in the MBR first in 1982, we did not include children born prior to that (n = 2). In Table 1, background characteristics and smoking prevalence for the included women are shown, as compared to regional birth register data from the same region of Sweden (Skåne) during the corresponding period. MBR records data on maternal smoking during pregnancy. Almost all women attend the free antenatal visits offered in Sweden, and out of the typically circa 10 visits per woman, two contain lengthier interviews from which information is reported to the National Board of Health and Welfare. The women are interviewed by trained midwives. The first visit usually occurs at 8–12 weeks of gestation, and the second at 30–32 weeks. A standardized questionnaire is used and the women report their current
smoking behavior, which is categorized in the MBR as nonsmoking, 1–9 cigarettes/day or more than 9 cigarettes/day. The register has a high level of completeness, covering 98%–99% of all births in Sweden.10 Verbal informed consent was obtained before serum sampling for the biobank, and the women were informed that the samples could be used for future research. Apart from the drawing of blood at delivery, no additional commitment was required by the women. Written informed consent was first implemented in 2005 and was not used for sample collection in the biobank. The study protocol was approved by the Ethics committee at Lund University.
Analyses of Cotinine Cotinine levels were measured in maternal serum and umbilical cord serum from the Malmö Maternity Unit Serum Biobank. The laboratory procedure is described in more detail elsewhere.9 In brief, aliquots of 100 µL sera were added with isotopically labeled internal standards and analysis was then performed using a hybrid triple quadruple linear ion-trap mass spectrometer (LC/MS/MS; UFLCXR, Shimadzu Corporation, Kyoto, Japan; QTRAP 5500; Sciex, Framingham, MA). The limit of detection was 0.2 ng/mL. To increase the accuracy, the value reported is the average of two measurements from the same sample worked up and analyzed on different days. As the optimal cotinine cutoff to reliably distinguish active smokers from nonsmokers exposed to environmental tobacco smoke (ETS) varies according to the distribution of cotinine levels in respective group, we believe that the higher cutoff suggested previously in the literature was arguably the most appropriate, as during the time frame studied (1982–2000), few regulations were in place in order to minimize the population’s exposure to ETS.3,11,12 Accordingly, we classified cotinine levels as follows: as the laboratory had a detection limit of 0.2 ng/mL, individuals with cotinine levels below 0.2 ng/mL were classified as nonsmoking, 0.2–14.9 ng/mL as being exposed to ETS (passive smoking) and 15 ng/mL and higher as smoking. According to these categories, the distribution of maternal and child cotinine levels are presented in Table 2.
Statistical Analyses The correlation between maternal and offspring cotinine levels was explored through Spearman’s rank correlation coefficient (rs). In addition, a ratio between umbilical cord and maternal serum levels was calculated for those child/mother-pairs where the mothers, according to the cotinine levels, were exposed to ETS and where mothers were active smokers, respectively. Kappa coefficients (κ) were calculated as measures of agreement between different smoking variables. First, the agreement between maternal and umbilical cord cotinine levels was investigated. Secondly, we investigated the agreement between MBR-data and maternal cotinine levels. In this analysis we dichotomized the MBR-data (nonsmoking/smoking: ≥1 cigarette/day) as well as the maternal serum cotinine levels (nonsmoking: 9 cigarettes/day 1982–1991 Nonsmoking 1–9 cigarettes/day >9 cigarettes/day 1992–2000 Nonsmoker 1–9 cigarettes/day >9 cigarettes/day Maternal characteristic Maternal age at childbirth (years)
Brief report
Cotinine Validation of Self-Reported Smoking During Pregnancy in the Swedish Medical Birth Register Downloaded from http://ntr.oxfordjournals.org/ at University of Wollongong on May 21, 2016
Kristina Mattsson MSc1, Karin Källén PhD1, Anna Rignell-Hydbom PhD1, Christian H. Lindh PhD1, Bo A.G. Jönsson PhD1, Peik Gustafsson PhD2, Per Olofsson PhD3, Sten A. Ivarsson PhD4, Lars Rylander PhD1 Division of Occupational and Environmental Medicine, Institute of Laboratory Medicine, Lund University, Lund, Sweden; 2Department of Child and Adolescent Psychiatry, Institution of Clinical Sciences, Lund University, Lund, Sweden; 3Department of Obstetrics and Gynecology, Institution of Clinical Sciences, Skåne University Hospital, Lund University, Malmö, Sweden; 4Department of Clinical Sciences, Unit of Pediatric Endocrinology, Lund University/ Clinical Research Centre, Malmö, Sweden 1
Corresponding Author: Kristina Mattsson, MSc, Division of Occupational and Environmental Medicine, Institute of Laboratory Medicine, Lund University, SE-221 85 Lund, Sweden. Telephone: 46-46-222-1638; Fax: 46-46-17-36-69; E-mail: [email protected]
Abstract Introduction: Self-reported data on smoking during pregnancy from the Medical Birth Register of Sweden (MBR) are widely used. However, underreporting of such behavior may occur, leading to biases. It is of importance to validate the smoking data in the MBR. The main objective was to investigate the agreement between self-reported smoking data from the MBR and cotinine levels in maternal serum among women from the general population in the region of Skåne, Sweden. We also estimated the transfer of cotinine from mother to fetus. Methods: From a cohort used previously to investigate the relationship between intrauterine environmental exposures and offspring neuropsychiatric outcomes, there were 204 control children retrieved from the MBR with data on maternal smoking in early pregnancy registered. Data on maternal and umbilical cord cotinine at delivery were available for these children from a regional biobank. Results: There was a high agreement between cotinine levels and MBR smoking data (κ = 0.82) and a high correlation between cotinine levels in maternal and umbilical cord serum (rs = 0.90, P < .001). Of the self-reported nonsmokers, 95% (95% confidence interval: 89% to 97%) were classified as nonsmokers after cotinine measurements. Conclusion: In these data, we found that the agreement between mothers’ self-reported smoking habits during pregnancy and their levels of serum cotinine was high, as was the transfer of cotinine from mother to fetus. This indicates that birth register data on pregnancy smoking in Sweden could be considered a valid measure.
Introduction
the Swedish Medical Birth Register (MBR), information on first-
For the study of different health outcomes related to maternal smoking during pregnancy, self-reported data from birth registers are often used, due to their accessibility and cost-effectiveness. In
trimester pregnancy smoking has been collected by midwives during the maternal health care visits since 1982. Due to the stigma surrounding pregnancy smoking, there is a risk of underreporting
© The Author 2015. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected].
79
80
Methods Participants This study uses data from the Fetal Environment and Neurodevelopment Disorders in Epidemiological Research project (The FENDER project). No woman in the present population was asked to participate in a specified study for the acquisition of these data. The cohort for this analysis is comprised by an earlier control group and data come exclusively from national registries and a regional biobank. A more comprehensive description of methods can be found elsewhere.9 In Malmö, a city of circa 300 000 inhabitants, nearly all deliveries occur at the Malmö University Hospital Maternity Unit. Malmö is the largest town/municipality in Skåne, which is the southernmost county in Sweden. The Malmö Maternity Unit Serum Biobank has been storing umbilical cord serum and maternal serum, both taken at delivery from women giving birth at this unit, between 1969 and 2000, and now has samples from approximately 70 000 deliveries. The control children used in this study were randomly selected from the MBR to match children with attention-deficit-hyperactivity disorder born 1978–2000, according to year of birth and maternal country of origin. The children were chosen in pools of 10 matched eligible controls, and the first baby with an available serum sample in the biobank was chosen. As smoking during pregnancy was recorded in the MBR first in 1982, we did not include children born prior to that (n = 2). In Table 1, background characteristics and smoking prevalence for the included women are shown, as compared to regional birth register data from the same region of Sweden (Skåne) during the corresponding period. MBR records data on maternal smoking during pregnancy. Almost all women attend the free antenatal visits offered in Sweden, and out of the typically circa 10 visits per woman, two contain lengthier interviews from which information is reported to the National Board of Health and Welfare. The women are interviewed by trained midwives. The first visit usually occurs at 8–12 weeks of gestation, and the second at 30–32 weeks. A standardized questionnaire is used and the women report their current
smoking behavior, which is categorized in the MBR as nonsmoking, 1–9 cigarettes/day or more than 9 cigarettes/day. The register has a high level of completeness, covering 98%–99% of all births in Sweden.10 Verbal informed consent was obtained before serum sampling for the biobank, and the women were informed that the samples could be used for future research. Apart from the drawing of blood at delivery, no additional commitment was required by the women. Written informed consent was first implemented in 2005 and was not used for sample collection in the biobank. The study protocol was approved by the Ethics committee at Lund University.
Analyses of Cotinine Cotinine levels were measured in maternal serum and umbilical cord serum from the Malmö Maternity Unit Serum Biobank. The laboratory procedure is described in more detail elsewhere.9 In brief, aliquots of 100 µL sera were added with isotopically labeled internal standards and analysis was then performed using a hybrid triple quadruple linear ion-trap mass spectrometer (LC/MS/MS; UFLCXR, Shimadzu Corporation, Kyoto, Japan; QTRAP 5500; Sciex, Framingham, MA). The limit of detection was 0.2 ng/mL. To increase the accuracy, the value reported is the average of two measurements from the same sample worked up and analyzed on different days. As the optimal cotinine cutoff to reliably distinguish active smokers from nonsmokers exposed to environmental tobacco smoke (ETS) varies according to the distribution of cotinine levels in respective group, we believe that the higher cutoff suggested previously in the literature was arguably the most appropriate, as during the time frame studied (1982–2000), few regulations were in place in order to minimize the population’s exposure to ETS.3,11,12 Accordingly, we classified cotinine levels as follows: as the laboratory had a detection limit of 0.2 ng/mL, individuals with cotinine levels below 0.2 ng/mL were classified as nonsmoking, 0.2–14.9 ng/mL as being exposed to ETS (passive smoking) and 15 ng/mL and higher as smoking. According to these categories, the distribution of maternal and child cotinine levels are presented in Table 2.
Statistical Analyses The correlation between maternal and offspring cotinine levels was explored through Spearman’s rank correlation coefficient (rs). In addition, a ratio between umbilical cord and maternal serum levels was calculated for those child/mother-pairs where the mothers, according to the cotinine levels, were exposed to ETS and where mothers were active smokers, respectively. Kappa coefficients (κ) were calculated as measures of agreement between different smoking variables. First, the agreement between maternal and umbilical cord cotinine levels was investigated. Secondly, we investigated the agreement between MBR-data and maternal cotinine levels. In this analysis we dichotomized the MBR-data (nonsmoking/smoking: ≥1 cigarette/day) as well as the maternal serum cotinine levels (nonsmoking: 9 cigarettes/day 1982–1991 Nonsmoking 1–9 cigarettes/day >9 cigarettes/day 1992–2000 Nonsmoker 1–9 cigarettes/day >9 cigarettes/day Maternal characteristic Maternal age at childbirth (years)
