published between 2014 and 2017
We agree with Professors Salemi and Zoorob (1) that the 2016 UK guidelines’ message regarding the consumption of alcohol during pregnancy is clear: it is best to be avoided. However, this has not always been the case and previous guidelines implicitly suggested that drinking up to 4 units per week was likely to be safe. We also recognise the methodological challenges of studying long term effects of low maternal alcohol consumption in pregnancy. Like Salemi and Zoorob, we feel that the paucity of evidence is unfortunate and should be addressed. We also maintain that in light of the evidence that is available, basing guidelines on the precautionary principle is reasonable.
Specifically, Salemi and Zoorob raise concerns about the inclusion of a study by Salihu et al. (2, reference 27 in the review) in the review (3), because maternal alcohol consumption was determined retrospectively, i.e. after delivery. Our predefined eligibility criterion was explicit - we required alcohol consumption to be ascertained prior to pregnancy. Salemi and Zoorob therefore question the inclusion of Salihu et al.’s study in our review. In the Methods section, Salihu et al. state that they used “the Missouri linked cohort data files…1989 through 2005.” No further details are provided The Discussion states that “Alcohol was based on maternal recall and therefore subject to bias. Women may underreport actual alcohol consumption because of societal stigmas, biasing study results toward the nu...
Specifically, Salemi and Zoorob raise concerns about the inclusion of a study by Salihu et al. (2, reference 27 in the review) in the review (3), because maternal alcohol consumption was determined retrospectively, i.e. after delivery. Our predefined eligibility criterion was explicit - we required alcohol consumption to be ascertained prior to pregnancy. Salemi and Zoorob therefore question the inclusion of Salihu et al.’s study in our review. In the Methods section, Salihu et al. state that they used “the Missouri linked cohort data files…1989 through 2005.” No further details are provided The Discussion states that “Alcohol was based on maternal recall and therefore subject to bias. Women may underreport actual alcohol consumption because of societal stigmas, biasing study results toward the null.” Only later in the Discussion do they also note that “…women with an adverse pregnancy outcome may be even more likely to underreport alcohol consumption,” suggesting that alcohol consumption was ascertained retrospectively.
Two independent reviewers read this paper and deemed it eligible for inclusion. As stated in our Methods section, we did not contact authors for further information. Having been alerted by Salemi and Zoorob to the use of birth certificates as the data source in Salihu et al., we are still unsure as to the where and when these data were ascertained. Were they abstracted from antenatal records (and hence prospectively assessed?)? Were they self-reported post-delivery (and hence retrospectively ascertained)? Do all hospitals in Missouri collect these data in the same manner? We have now contacted the authors of the original paper for clarification but have not yet had a response.
In light of these concerns, carried out sensitivity analyses omitting the Salihu et al. study in the meta-analyses for preterm delivery and SGA. For preterm delivery, the pooled odds ratio including Salihu et al. and reported in the review was: 1.10 (95%CI: 0.95, 1.28, I2 = 59%). Excluding Salihu et al. it is 1.13 (0.92, 1.38; I2 = 41%). Respective results for SGA are: with Slaihu et al. 1.08 (95%CI: 1.02, 1.14; I2 = 0%) and 0.96 (95%CI: 0.80, 1.15, I2 = 0%) without. These changes would not affect our main findings as reported in the original review: “the two main findings are: (1) a surprisingly limited number of prospective studies specifically addressing the question of whether light maternal alcohol consumption (ie, up to 32g/week (or 4 UK units) has any causal effect (adverse or beneficial) on infant and later offspring outcomes and pregnancy outcomes, and, as a result, (2) a paucity of evidence demonstrating a clear detrimental effect, or safe limit, of light alcohol consumption on outcomes.”
Salemi and Zoorob also question why we included specific results from Ernhart et al. (4, reference 59 in the review) in Table 2. We presented results of analyses in which Ernhart et al. used maternal self-reported alcohol consumption during pregnancy. We did not include results of analyses in which alcohol consumption prior to or just after conception was estimated from an equation developed in a separate sample. We feel that prioritising the results based on prospective maternal report, which is comparable to other studies, was a sensible decision.
Loubaba Mamluk, Luisa Zuccolo and Abigail Fraser on behalf of all co-authors
1. Salemi JL and Zoorob RJ. The importance of methodological rigor and communication of information. http://bmjopen.bmj.com/content/7/7/e015410.responses#the-importance-of-m....
2. Salihu HM, Kornosky JL, Lynch O, Alio AP, August EM, Marty PJ. Impact of prenatal alcohol consumption on placenta-associated syndromes. Alcohol (Fayetteville, NY). 2011;45(1):73-79.
3. Mamluk L, Edwards HB, Savović J, et al. Low alcohol consumption and pregnancy and childhood outcomes: time to change guidelines indicating apparently ‘safe’ levels of alcohol during pregnancy? A systematic review and meta-analyses. BMJ Open 2017;7:e015410. doi: 10.1136/bmjopen-2016-015410
4. Ernhart CB, Sokol RJ, Ager JW, Morrow-Tlucak M, Martier S. Alcohol-related birth defects: assessing the risk. Annals of the New York Academy of Sciences. 1989;562:159-172.
During the peer review process of this manuscript, it came to our attention that one of the reviewers, Erling Solheim, may have an undeclared conflict of interest. We were told that he had worked in the same research group as the authors from June 2006 to September 2007 and had published research with one of the authors in 2008.
We attempted to contact Erling Solheim a number of times to verify these claims, but he did not respond to our emails at the time. We would like to point out, however, that we do not feel that this undeclared conflict of interest (and one from ten years ago) would have compromised the peer review process or altered our decision to publish the manuscript.
In their recent paper,1 Mamluk et al describe the results of a systematic review and meta-analyses of the association between low alcohol intake in pregnancy and several adverse birth outcomes and long-term outcomes in children. The authors conclude that there is “limited evidence for a causal role of light drinking in pregnancy, compared with abstaining, on most of the outcomes examined” and that their “extensive review shows that this specific question is not being researched thoroughly enough, if at all.”
I sympathize with the thorough work performed by the authors. Even so, a word of caution is needed with respect to the second part of their conclusion. The authors defined the intake of interest as a maximum of 32 g of alcohol per week corresponding to 4 standard UK drinks/week. While this makes sense from a British point of view, the definition of a standard drink being 8 g of alcohol, this cut-off makes little sense in most non-British countries, where a standard drink is typically defined as containing 10-12 g. Hence, a large number of non-British studies using three or four drinks as their cut-off for low intake have not been included in the present review, only because their definition did not correspond with the British.
Several meta-analyses in recent years have used slightly more embracing definitions.2 3 These meta-analyses included far more studies, simply because a limit of 32 g fits poorly with international definitions. Using...
Several meta-analyses in recent years have used slightly more embracing definitions.2 3 These meta-analyses included far more studies, simply because a limit of 32 g fits poorly with international definitions. Using slightly higher cut-offs would have enabled the authors to include far more studies. The conclusion from all the reviews and meta-analyses remain the same, ie. little evidence of any association between low alcohol intake on a weekly basis with any adverse outcome.1 2 3 But the conclusion that the evidence is sparse does not seem correct. There is, in fact, quite a few studies if the limit for low intake is set slightly higher; and if (the majority of) studies with a higher cut-off of e.g. 36, 48 g/week or something similar generally show no association with the outcomes of interest, the conclusion should not be that there is paucity of evidence for intake up to 32 g/week. There is plenty of evidence of no association, only the evidence does not fit the British definition of 32 g but mainly with slightly higher cut-off points.
Ulrik Schiøler Kesmodel
Professor in obstetrics and gynaecology
Herlev University Hospital, Copenhagen
Institute for Clinical Medicine, University of Copenhagen
1. Mamluk L, Edwards HB, Savović J et al. Low alcohol consumption and pregnancy and childhood outcomes: time to change guidelines indicating apparently ‘safe’ levels of alcohol during pregnancy? A systematic review and meta-analyses. BMJ Open 2017;7:e015410. doi:10.1136/bmjopen-2016-015410.
2. Patra J, Bakker R, Irving H, et al. Dose-repsonse relationship between alcohol consumption before and during pregnancy and the risks of low birthweight, preterm birth and small for gestational age (SGA) - a systematic review and meta-analyses. BJOG 2011; 118: 1411-1421. DOI: 10.1111/j.1471-0528.2011.03050.x
3. Flak AL, Su S, Bertrand J, Denny CH, Kesmodel US, Cogswell ME. The association between mild, moderate, and binge prenatal alcohol exposure and child
Thank you for the opportunity to respond to Lassman et al.’s re-analysis of our study titled, “Clinical trial registration, reporting, publication and FDAAA Compliance: A cross-sectional analysis and ranking of new drugs approved by the FDA in 2012”.1 Our original study assessed the clinical trial transparency of novel drugs approved by the FDA in 2012 that were sponsored by large drug companies. We assessed the drugs by two sets of transparency standards: U.S. legal requirements and an ethical standard that all human subjects research should be publicly accessible to contribute to generalizable knowledge.
Our original analysis included a review of 15 drugs, sponsored by 10 large companies, involving 342 trials. Lassman and colleagues’ reassessment examined 69 of these 342 trials and focused on only the U.S. legal requirements standard. Lassman et al did not elaborate on why they limited their assessment to this subset of trials and on compliance with legal requirements. As a reminder, US clinical trial disclosure requirements are defined by the Food and Drug Administration Amendments Act (FDAAA), passed in 2007.2
We applaud efforts to replicate studies. We are glad that our policy of publicly sharing data through the Dryad Digital Repository enabled replication and re-analysis.3 Additionally, we generally agree with the Lassman and colleagues re-assessment of our study using today’s new and updated knowledge base and world-view....
We applaud efforts to replicate studies. We are glad that our policy of publicly sharing data through the Dryad Digital Repository enabled replication and re-analysis.3 Additionally, we generally agree with the Lassman and colleagues re-assessment of our study using today’s new and updated knowledge base and world-view. However, we do not agree with their implications that our study was lacking in some way. Our work reflected prevailing standards at the time.
It is important to note, that, when we conducted our analysis (starting around 2013), implementation and interpretation of FDAAA were very different than today. Our study used the leading interpretation of the law at the time of our study. Our interpretation was shared by many drug companies, researchers, and reflected in top academic publications, including the banner NEJM publication on the subject by Anderson et. al. titled, “Compliance with Results Reporting at ClinicalTrials.gov.”4
As such, Lassman et al paper isn’t so much a re-analysis of our work, but rather a reflection and update on our work given the learning and evolution of practices that have emerged since the time our original study was conducted. Updated interpretations of FDAAA were generally catalyzed by the NIH issuance of a Final Rule, on September 21, 2016, clarifying key ambiguous language in FDAAA.5 In particular, thanks in large part to the Final Rule, there is now general consensus that under FDAAA, there was no requirement to report trial results for unapproved drugs (for trials studying initial use indications).
Prior interpretations involved results reporting by one year after a trial’s primary completion date, unless a certificate of delay was filed asking for an extension of time to report. In the case of a certificate being filed, results reporting could be delayed until 30 days after FDA approval of the indication. Under the Final Rule, this will change. Results will be required to be reported for both approved and unapproved drugs.
Lassman et al acknowledge that interpretations of FDAAA have been varied and evolved over time. They state: “FDAAA is complex, and after its passage “a spectrum of interpretations” emerged, particularly with respect to the deadlines for results reporting and the necessity of “certificates of delay” (CODs)….The COD provisions… are ambiguous in some cases.”
Below is a table showing our original 2012 analysis, along with a re-analysis of our 2012 sample of drugs using today’s interpretation of FDAAA. Lassman and colleagues 2017 reanalysis is included in the table as well (Table 1). Our analyses are in general agreement.
Our aim as investigators on the Good Pharma Scorecard project is to ensure that we employ an iterative, open and adaptive learning system when establishing standards and reporting on clinical trial transparency. We strive to continuously improve and refine our methods as appropriate. We also continuously reach out to and convene stakeholders, soliciting feedback to ensure the accuracy and timeliness of our work and that the project is advancing the needs of patients and public health broadly.
We recently published an assessment of clinical trial transparency for novel drugs approved by the FDA in 2014 that were sponsored by large companies.6 It uses the updated and current interpretations of FDAAA catalyzed by the passing of the Final Rule. They are similar to those of Lassman and colleagues. There is less debate at this point about how to interpret most sections of FDAAA reporting requirements. Experts should generally be in agreement on requirements and be better positioned to advance a more open and transparent healthcare innovation and drug development enterprise that strives towards trustworthiness and patient-centricity.
Table 1 (found here: http://blogs.bmj.com/bmjopen/files/2017/09/Miller-1.png): Table showing our (Miller et al) original analysis, along with our re-analysis of our 2012 drugs using current FDAAA interpretations and Lassman and colleagues’ analysis.
The Good Pharma Scorecard is funded by a grant from the Laura and John Arnold Foundation to Bioethics International. In the past 36 months, Dr. Ross has received support through Yale University from Johnson and Johnson to develop methods of clinical trial data sharing, from Medtronic, Inc. and the Food and Drug Administration (FDA) to develop methods for post-market surveillance of medical devices, from the Food and Drug Administration (FDA) to establish the Yale-Mayo Center for Excellence in Regulatory Science and Innovation (CERSI), from the Blue Cross Blue Shield Association to better understand medical technology evaluation, from the Centers of Medicare and Medicaid Services (CMS) to develop and maintain performance measures that are used for public reporting, from the Agency for Healthcare Research and Quality, and from the Laura and John Arnold Foundation to support the Collaboration on Research Integrity and Transparency (CRIT) at Yale.
Miller JE, Korn D, Ross JS. Clinical trial registration, reporting, publication and FDAAA compliance: a cross-sectional analysis and ranking of new drugs approved by the FDA in 2012. BMJ Open 2015; 5: e009758. doi: 10.1136/bmjopen-2015-009758
2 Food and Drug Administration Amendments Act of 2007 (FDAAA), Public Law No. 110-85, § 801, 121 Stat. 904 (2007) (codified at 42 U.S.C. § 282).
3 Miller JE, Korn D, Ross JS (2015) Data from: Clinical trial registration, reporting, publication and FDAAA compliance: a cross-sectional analysis and ranking of new drugs approved by the FDA in 2012. Dryad Digital Repository. http://dx.doi.org/10.5061/dryad.t8n07.
4 Anderson ML, Chiswell K, Peterson ED, et al. Compliance with results reporting at ClinicalTrials.gov. N Engl J Med2015;372:1031–9. doi:10.1056/NEJMsa1409364
5 Clinical Trials Registration and Results Information Submission; Final Rule (NIH Final Rule), 81 Fed. Reg. 64982 (Sept. 21, 2016) (codified at 42 C.F.R. Part 11).
6 Miller JE, Wilenzick M, Ritcey N, Ross JS, Mello MM. Measuring clinical trial transparency: An empirical analysis of newly approved drugs and large pharmaceutical companies. BMJ Open (in press).
We have read the article by Cui et al. with interest, however we believe that some issues should be clarified. The authors present a cross-sectional study aimed to determine the risk factors for diabetic retinopathy (DR) in the Chinese population.
On a cohort of 17 985 individuals from Beijing, China, the prevalence of DR was 1.5% in the general study population and 8.1% among individuals with diabetes mellitus (DM). It is presented that since the Beijing population is composed of Han people as well as minorities from the country, it could be considered as representative of the Chinese population. It should be underlined that the assessed prevalence is far lower than that of other studies, including those from Asian countries. For example the prevalence of DR among primary care Chinese patients with DM in Hong Kong was 39.0% , and 37% among Chinese patients aged over 40 years with a self-reported diagnosis of diabetes in the Beijing Eye Study .
Thus, questions about the methodology can be raised. The American Diabetes Association clearly states the criteria for the diagnosis of diabetes: 1) fasting plasma glucose ≥ 126 mg/dL (≥ 7.0 mmol/L) 2) 2-hour plasma glucose after a 75-g oral glucose tolerance test ≥ 200 mg/dL (≥ 11.1 mmol/L) 3) A1C ≥ 6.5% 4) In a patient with classic symptoms of hyperglycemia or hyperglycemic crisis, a random plasma glucose ≥200 mg/dL (11.1 mmol/L) . Thus, it should be clarified if the diagnosis of diabetes in this study w...
Thus, questions about the methodology can be raised. The American Diabetes Association clearly states the criteria for the diagnosis of diabetes: 1) fasting plasma glucose ≥ 126 mg/dL (≥ 7.0 mmol/L) 2) 2-hour plasma glucose after a 75-g oral glucose tolerance test ≥ 200 mg/dL (≥ 11.1 mmol/L) 3) A1C ≥ 6.5% 4) In a patient with classic symptoms of hyperglycemia or hyperglycemic crisis, a random plasma glucose ≥200 mg/dL (11.1 mmol/L) . Thus, it should be clarified if the diagnosis of diabetes in this study was based on the aforementioned criteria and what stratification method was applied,.
The “Methods” section reveals that fundus examination was conducted in mydriasis by two experienced ophthalmologists. In the “Discussion” it is presented that fundus examinations were performed by experienced ophthalmic technicians and ophthalmologists, and what is claimed as limitation to the study, without mydriasis.
1. Cui J, Ren J-P, Chen D-N, et al. Prevalence and associated factors of diabetic retinopathy in Beijing, China: a cross-sectional study. BMJ Open. 2017;7(8):e015473.
2. Lian JX, Gangwani RA, McGhee SM, et al. Systematic screening for diabetic retinopathy (DR) in Hong Kong: prevalence of DR and visual impairment among diabetic population. Br J Ophthalmol. 2015;100(2):151-155.
3. Xie XW, Xu L, Jonas JB, Wang YX. Prevalence of diabetic retinopathy among subjects with known diabetes in China: the Beijing Eye Study. Eur J Ophthalmol. 2009;19(1):91-99.
4. American Diabetes Association. 2. Classification and Diagnosis of Diabetes. Diabetes Care. 2016;39 Suppl 1:S13-S22.
Dr. Mamluk and colleagues’ recent systematic review and publication in BMJ Open (1) carries with it a provocative title and over the past week its conclusions have garnered substantial attention by news organizations such as CNN (2), Newsweek (3), and the New York Post (4). Although we agree completely with the authors’ assertion that there is a significant gap in the literature concerning low levels of maternal alcohol consumption during pregnancy, we have some substantial concerns regarding misleading statements in the paper and inconsistencies between the methodology that was stated versus that which was conducted.
First, in the introduction, the authors claim that “Internationally, clinical guidelines recommend that pregnant women should abstain from heavy or ‘binge’ drinking”. Although technically correct, it is misleading since the messaging reported in the authors’ cited reference actually shows that nearly every country advises abstaining from any alcohol consumption during pregnancy – not just heavy or binge drinking. Furthermore, the authors state that “…UK guidelines advised women to avoid drinking alcohol while trying to conceive, and in the first trimester, but at the same time indicated that consumption should be restricted to within ‘1 to 2 UK units, one or twice a week.” Their statement makes it seem as though the UK message is contradictory regarding maternal alcohol consumption during pregnancy – it is not. In the 2016 UK Chief Medical Officers’ Lo...
First, in the introduction, the authors claim that “Internationally, clinical guidelines recommend that pregnant women should abstain from heavy or ‘binge’ drinking”. Although technically correct, it is misleading since the messaging reported in the authors’ cited reference actually shows that nearly every country advises abstaining from any alcohol consumption during pregnancy – not just heavy or binge drinking. Furthermore, the authors state that “…UK guidelines advised women to avoid drinking alcohol while trying to conceive, and in the first trimester, but at the same time indicated that consumption should be restricted to within ‘1 to 2 UK units, one or twice a week.” Their statement makes it seem as though the UK message is contradictory regarding maternal alcohol consumption during pregnancy – it is not. In the 2016 UK Chief Medical Officers’ Low Risk Drinking Guidelines (5) it is stated that “If you are pregnant or think you could become pregnant, the safest approach is not to drink alcohol at all, to keep risks to your baby to a minimum.” In regards to women who wish to stay below 1-2 units per day during pregnancy, they state, “Women who wish to stay below these levels need to be careful to avoid underestimating their actual consumption. The safer option is not to drink alcohol at all during pregnancy.” We feel the Chief Medical Officers’ message is quite clear.
We are more concerned with seeming inconsistencies between the authors’ stated methodology and decisions made during the conduct of their systematic review. The methods section is well articulated and thorough and the authors [appropriately] state that eligible studies [for their review] must have had “prospective assessment of prenatal alcohol exposure (ie, before birth).” In fact, the entire case-control study design is ineligible for this reason, and the increased potential for recall bias, possibly differential based on the pregnancy outcome, that is associated with retrospective assessment of alcohol consumption. However, a single US study (6) that the authors admit accounts for >90% of participants included in the meta-analysis for SGA and preterm birth, and therefore drives the meta-analyses on these two outcomes, and is either the only or one of two studies for stillbirth, placental abruption, placenta previa, and pre-eclampsia, does not meet the authors’ own criteria. In fact, the dataset used in the study assesses maternal alcohol consumption based on what is documented on the birth certificate (6). Not only is this not prospective assessment of maternal alcohol consumption and prone to substantial bias, but birth certificates have been reported repeatedly to be one of the worst sources of such information, with extremely poor sensitivity and positive predictive value (7-10). Even the authors of that study acknowledge the 1.3% prevalence of maternal alcohol consumption during pregnancy grossly underestimates what was to be expected based on national data (10-16%). Clearly, this study fails to meet the criteria for this systematic review; regardless, the poor quality of maternal alcohol consumption data from this study would preclude its inclusion in a systematic review searching for reliable evidence to address an important public health question. Considering the overwhelming influence this study has on the meta-analysis conducted by Mamluk and colleagues, we feel this calls into question any conclusions drawn for preterm birth and SGA, and well as non-meta-analysis summaries for stillbirth, placental abruption, placenta previa, and pre-eclampsia.
We are also a bit confused as to the authors’ decisions regarding the inclusion/exclusion of data from one of only three studies on the association between low maternal consumption of alcohol during pregnancy and FASD malformations. This 1989 study by Ernhart et al (11) assessed varying levels of maternal alcohol consumption and the “tally” of both craniofacial and total birth defects. Mamluk and colleagues report in their systematic review the results from Tables 2 and 3 from Ernhart et al., which shows no statistically significant increased ‘tally’ of birth defects for low alcohol consumers (>0.0, ≤0.1 average ounces of alcohol per day) compared to abstainers. Alcohol exposure was assessed prospectively during pregnancy and met Mamluk’s criteria. However, Ernhart and colleagues acknowledge that this method may not reflect alcohol use during the period at, or shortly after, conception (ie, before a woman knows she is pregnant). Therefore, they used an additional method – ‘estimated embryonic average ounces of alcohol per day’ – to estimate exposure during a time window that is critical for alcohol-related dysmorphology (11). This technique, too, seemed to fit Mamluk’s systematic review criteria since it was based on responses elicited prior to delivery/birth. Futhermore, Tables 6 and 7 from Ernhart et al. reveal a statistically significantly (p=0.042) increased ‘tally’ of craniofacial anomalies for low alcohol consumers (>0.0, ≤0.1 average ounces of alcohol per day) compared to abstainers, and a borderline significant results for total anomalies (p=0.06). We are curious as to why one seemingly relevant approach was included in their review, but not another.
We appreciate the desire for Mamluk and colleagues to contribute to this very important evidence base. We also agree that in epidemiology and clinical practice, we often do a poor job of communicating absolute risk; therefore, we were pleased to read the UK Chief Medical Officers’, statement, “The risk of harm to the baby is likely to be low if you have drunk only small amounts of alcohol before you knew you were pregnant or during pregnancy” to prevent undue stress for women who drank small amounts prior to realizing they were pregnant or even during early pregnancy (5). Clearly, Mamluk’s systematic review should be a call to action for researchers to provide more reliable evidence to all stakeholders. However, in light of the clear evidence for adverse effects of higher levels of alcohol consumption, and the fact that we do not know of a safe amount of alcohol use during pregnancy, we feel hesitant to propose changes to nearly every country’s messaging in support of abstinence. We are particularly concerned by the New York Post’s title of “Light drinking during pregnancy does not harm baby: study” that covered the findings of this review.
1. Mamluk L, Edwards HB, Savovic J, et al. Low alcohol consumption and pregnancy and childhood outcomes: time to change guidelines indicating apparently 'safe' levels of alcohol during pregnancy? A systematic review and meta-analyses. BMJ Open. 2017;7(7):e015410.
2. Howard J. CNN: Is light drinking while pregnant OK? 2017; http://www.cnn.com/2017/09/11/health/drinking-alcohol-pregnant-study/ind.... Accessed September 12, 2017.
3. Dovey D. Newsweek: Can You Drink While Pregnant? Even Light Drinking Can Cause Problems 2017; http://www.newsweek.com/can-you-drink-while-pregnant-even-light-drinking.... Accessed September 13, 2017.
4. Whittfield T. New York Post: Light drinking during pregnancy does not harm baby: study. 2017; http://nypost.com/2017/09/11/light-drinking-during-pregnancy-does-not-ha.... Accessed September 12, 2017.
5. Health UDo. UK Chief Medical Officers’ Low Risk Drinking Guidelines. . 2016; https://www.gov.uk/government/uploads/system/uploads/attachment_data/fil.... Accessed September 15, 2017.
6. Salihu HM, Kornosky JL, Lynch O, Alio AP, August EM, Marty PJ. Impact of prenatal alcohol consumption on placenta-associated syndromes. Alcohol (Fayetteville, NY). 2011;45(1):73-79.
7. Northam S, Knapp TR. The reliability and validity of birth certificates. J Obstet Gynecol Neonatal Nurs. 2006;35(1):3-12.
8. Piper JM, Mitchel EF, Jr., Snowden M, Hall C, Adams M, Taylor P. Validation of 1989 Tennessee birth certificates using maternal and newborn hospital records. Am J Epidemiol. 1993;137(7):758-768.
9. Reichman NE, Hade EM. Validation of birth certificate data. A study of women in New Jersey's HealthStart program. Ann Epidemiol. 2001;11(3):186-193.
10. Zollinger TW, Przybylski MJ, Gamache RE. Reliability of Indiana birth certificate data compared to medical records. Ann Epidemiol. 2006;16(1):1-10.
11. Ernhart CB, Sokol RJ, Ager JW, Morrow-Tlucak M, Martier S. Alcohol-related birth defects: assessing the risk. Annals of the New York Academy of Sciences. 1989;562:159-172.
To the BMJ Editor:
A recent analysis in BMJ Open accompanied by a clinical practice guideline on antiretroviral therapy (ART) in pregnant women living with HIV by Siemieniuk et al has concluded that “tenofovir/emtricitabine is likely to increase stillbirth/early neonatal death and early premature delivery compared with zidovudine/lamivudine” (1,2). While the clinical practice guideline was based on a systematic review, in reality, the conclusion was based solely on results of the PROMISE study (3). Evidence from large observational studies did not support this recommendation, but was viewed as too low quality to consider in making recommendations. As coauthors of the PROMISE study, we would like to comment on the authors’ conclusion.
The objective of the PROMISE trial was to compare the efficacy of zidovudine/single-dose nevirapine (AZT-alone) with protease inhibitor-based (lopinavir-ritonavir) combination ART (AZT-based ART [AZT-ART] or tenofovir-based ART [TDF-ART]) to prevent mother-to-child transmission in women with CD4 cell count >350 cells/mm3 (3).The study enrolled during two periods, and the comparisons of the TDF-ART arm to the other arms are restricted to the women who were concomitantly randomized among the three study arms, which occurred only in the second period of the study. In Period 1 – accounting for about two-thirds of enrollment - only hepatitis B virus (HBV)-coinfected women (~1% of overall enrollment) were randomized to TDF-ART vs...
The objective of the PROMISE trial was to compare the efficacy of zidovudine/single-dose nevirapine (AZT-alone) with protease inhibitor-based (lopinavir-ritonavir) combination ART (AZT-based ART [AZT-ART] or tenofovir-based ART [TDF-ART]) to prevent mother-to-child transmission in women with CD4 cell count >350 cells/mm3 (3).The study enrolled during two periods, and the comparisons of the TDF-ART arm to the other arms are restricted to the women who were concomitantly randomized among the three study arms, which occurred only in the second period of the study. In Period 1 – accounting for about two-thirds of enrollment - only hepatitis B virus (HBV)-coinfected women (~1% of overall enrollment) were randomized to TDF-ART vs AZT-ART vs AZT-alone. It was only in Period 2, enrolling 35% of the overall population, that all women, regardless of HBV status, were randomized to one of the three arms. Comparisons of TDF-ART with AZT-ART or AZT-alone were therefore restricted to this second period of time.
PROMISE did not analyze stillbirth in combination with early neonatal death. Contrary to the authors’ statement, the pathophysiology of stillbirth and early neonatal death are not necessarily the same and hence the PROMISE team did not feel it was appropriate to combine these endpoints. It is important to note that the rates of spontaneous abortion and stillbirth were not significantly different between AZT alone, AZT-ART and TDF-ART arms (Table 1). In their meta-analysis, the authors combined data on stillbirth/early infant death from two hepatitis B mono-infection studies (Pan and Wang) which had very few events and did not include HIV-infected pregnant women. The only two events in the HBV studies were from the Pan study: 1 stillbirth in the TDF-alone group in a woman who had a prior stillbirth, and one full-term newborn delivered with the use of forceps from a mother in the TDF-alone group who had declined a cesarean section and died on day 2 from intracerebral hemorrhage and aspiration pneumonia.
The PROMISE team noted that both the AZT-ART and TDF-ART regimens were associated with increased preterm delivery <37 weeks compared to the AZT-alone arm (Table 1); the rate of preterm delivery <37 weeks during Period 2 was not significantly different between AZT-ART and TDF-ART arms (p=0.77). It was only in the evaluation of very preterm delivery (<34 weeks) that a difference in ART arms was observed, with a higher rate of very preterm delivery observed in the TDF-ART compared to AZT-ART arm (p=0.04), although the TDF-ART very preterm delivery rate was not significantly different than AZT-alone arm (p=0.10) (Table 1). This would suggest that both ART regimens may be associated with prematurity, with AZT-ART increasing preterm delivery between 34-36 weeks and TDF-ART possibly increasing very preterm delivery <34 weeks. Given the lack of a biologic rationale for this difference and the small number of very preterm delivery events, the PROMISE team was not willing to draw a definitive conclusion from these data.
The PROMISE team noted in the manuscript discussion that there was an imbalance in the neonatal deaths occurring during Period 1 and 2 in the AZT-ART group: 15 of 17 deaths, 88%, occurred during Period 1, and only 2 (12%) during Period 2, the time of comparison with TDF-ART (Table 1). This resulted in a very low rate of neonatal mortality in AZT-ART group during period 2. In contrast, in the AZT-alone comparison group, 39% of neonatal deaths (11/28) occurred during Period 1 and 61% during Period 2. Consequently, the AZT-ART group appears to have a very low rate of infant mortality during Period 2 (0.6%), the time it was compared to TDF-ART; this is supported by the fact that neonatal mortality was not significantly different between AZT-alone and TDF-ART arms (3.2% vs 4.4%, respectively).
The PROMISE team noted concerns about potential pharmacokinetic interactions between lopinavir and TDF; the BMJ authors dismiss the potential pharmacokinetic interaction as “unlikely”. The concern was about potential increase in tenofovir, not lopinavir levels, even though an increased dose of lopinavir was given in the third trimester in the PROMISE study. The use of the higher lopinavir dose was based on potentially lower levels in late pregnancy with standard dosing in studies in US women (4,5). However, some other studies have not noted a decrease in levels with standard dosing, particularly in Thai women with low body weight (6,7); pharmacokinetic studies have not been done in African women, the group enrolled into PROMISE. In studies of concomitant administration of tenofovir with standard lopinavir dosing (in non-pregnant adults), decreased renal clearance of tenofovir and increased plasma and intracellular levels, particularly in women, have been reported with concomitant administration (8-10). There have not been pharmacokinetic studies of tenofovir given with increased lopinavir dosing. Studies of stored specimens will need to be done to address this issue.
The PROMISE team did not feel it was appropriate to draw definitive conclusions from the study regarding the use of TDF-ART in pregnancy. The team specifically noted that because the study only included protease-inhibitor based ART, one cannot extrapolate data on TDF-ART based on other classes of drugs such as the efavirenz (EFV)-based ART regimen currently recommended in pregnancy by the World Health Organization (11).
While observational data are viewed as lower quality compared to data from randomized trials, the PROMISE team notes that it is unlikely that there will be other randomized trials and that there have been numerous observational studies suggesting the safety of TDF-ART in combination with non-nucleoside reverse transcriptase inhibitor (NNRTI) drugs. The BMJ authors state that observational data comparing ART are problematic because AZT-ART regimens are older and clinical practices might have been different than in TDF-ART era. However, a recent study from Botswana compared birth outcomes, including preterm delivery and neonatal death, among HIV-infected women starting 3-drug ART regimens and who delivered between August 2014 and August 2016; hence clinical practice would be the same for all women. Compared with a regimen of TDF-emtricitabine (FTC)-EFV, all other regimens, including AZT-based ART, were associated with higher risk of adverse outcome; increased risk of preterm birth, very preterm birth and neonatal death were observed for infants exposed to AZT-lamivudine (3TC)-lopinavir-ritonavir.
Thus, while the PROMISE team strongly supports further evaluation of the safety of ART regimens in pregnancy for the woman and her infant in order to find the optimal ART regimen, the PROMISE team does not agree that the PROMISE trial results support a recommendation against using a TDF-based ART regimen in pregnancy.
Mary Glenn Fowler MD, MPH, Johns Hopkins University School of Medicine, Baltimore, MD
Lynne Mofenson MD, Elizabeth Glaser Pediatric AIDS Foundation, Washington DC
Pat Flynn MD, St. Jude Children’s Research Hospital, Memphis, TN
Taha Taha MD, Johns Hopkins University School of Public Health, Baltimore, MD
1. Siemieniuk RA, Foroutan F, Mirza R et al. Antiretroviral therapy for pregnant women living with HIV or hepatitis B: a systematic review and meta-analysis. BMJ Open 207;7:e019022.
2. Siemieniuk RA, Lytuyn L, Ming JM et al. Antiretroviral therapy in pregnant women living with HIV: a clinical practice guideline. BMJ 2017;358:j3961.
3. Fowler MG, Qin M, Fiscus SA, et al. Benefits and risks of antiretroviral therapy for perinatal prevention. N Engl J Med 2016;375:1726-37.
4. Stek AM, Mirochnick M, Capparelli E, et al. Reduced lopinavir exposure during pregnancy. AIDS 2006; 20: 1931-9. 17.
5. Mirochnick M, Best BM, Stek AM, et al. Lopinavir exposure with an increased dose during pregnancy. J Acquir Immune Defic Syndr 2008; 49: 485-91.
6. Calza L, Manfredi R, Trapani F, et al. Lopinavir/ritonavir trough concentrations with the tablet formulation in HIV-1-infected women during the third trimester of pregnancy. Scand J Infect Dis 2012;44:381-7.
7. Cressey TR, Jourdain G, Rawangban B, et al. Pharmacokinetics and virologic response of zidovudine/lopinavir/ritonavir initiated during the third trimester of pregnancy. AIDS 2010;24:2193-200.
8. Kiser JJ, Carten ML, Aquilante CL, et al. The effect of lopinavir/ritonavir on the renal clearance of tenofovir in HIV-infected patients. Clin Pharmacol Ther 2008; 83: 265-72.
9. Kearney BP, Mathias A, Mittan A, Sayre J, Ebrahimi R, Cheng AK. Pharmacokinetics and safety of tenofovir disoproxil fumarate on coadministration with lopinavir/ritonavir. J Acquir Immune Defic Syndr 2006; 43: 278-83.
10. Pruvost A, Negredo E, Théodoro F, et al. Pilot pharmacokinetic study of human immunodeficiency virus-infected patients receiving tenofovir disoproxil fumarate (TDF): investigation of systemic and intracellular interactions between TDF and abacavir, lamivudine, or lopinavir-ritonavir. Antimicrob Agents Chemother 2009; 53: 1937-43.
11. World Health Organization. Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection – recommendations for a public health approach, second edition. June 2016; Geneva, Switzerland.
Table 1 Adverse Outcomes in PROMISE by Study Arm
Spontaneous abortion (<20 wks) or stillbirth (>20 wks)
Total, N=3239 in analysis
Total spontaneous abortion/stillbirth, N=95 (2.9%)
Period 1+2, AZT vs AZT-ART comparisons (total events, N=89)
AZT alone: 41/1451 (2.8%)
AZT-ART: 48/1447 (3.3%), AZT alone vs AZT-ART, p=0.45
Period 1, N=2201 (total events, N=78)
AZT alone: 3/1102 (3.0%)
AZT-ART: 45/1099 (4.1%)
Period 2, N=1038, 3 arm comparisons (total events, N=15):
AZT-alone: 8/349 (2.3%)
AZT-ART: 3/348 (0.9%)
TDF-ART: 6/341 (1.6%), AZT alone vs TDF ART, p=0.79; AZT-ART vs TDF-ART, p=034
Total events, N=5
Period 1 (total events 5)
AZT alone: 3 (<0.5%)
AZT-ART: 2 (<0.5%)
Period 2 (total events 0)
AZT alone: 0%
Total events, N=90
AZT alone: 38 (37 singleton, 1 multiple gestation)
AZT-ART: 46 (39 singleton, 7 multiple gestation)
TDF-ART: 6 (6 singleton)
AZT alone: 2.8%
AZT alone: 2.3%
Preterm birth (<37 weeks)
Total with data, N=3152
Total preterm <37 weeks, N=535 (17.0%)
Period 1+2 AZT vs AZT-ART comparisons (total events, N=473)
AZT alone: 185/1411 (13.1%)
AZT-ART: 288/1406 (20.5%), AZT alone vs AZT-ART, p<0.001
Period 1 (total events, N=359)
AZT alone: 139/1070 (13.0%)
AZT-ART: 220/1060 (20.8%)
Period 2, 3 arm comparisons (total events, N=176):
AZT-alone: 46/341 (13.5%)
AZT-ART: 68/346 (19.7%)
TDF-ART: 62/335 (18.5%), AZT alone vs TDF ART, p=0.09; AZT-ART vs TDF-ART, p=0.77
Very Preterm birth (<34 weeks)
Total with data, N=3152
Total very preterm <34 weeks, N=101 (3.2%)
Period 1+2 AZT vs AZT-ART comparisons (total events, N=81)
AZT alone: 37/1411 (2.6%)
AZT-ART: 44/1406 (3.1%), AZT alone vs AZT-ART, p=0.43
Period 1 (total events, N=61)
AZT alone: 26/1070 (2.4%)
AZT-ART: 35/1060 (3.3%)
Period 2, 3 arm comparisons (total events, N=20):
AZT-alone: 11/341 (3.2%)
AZT-ART: 9/346 (2.6%)
TDF-ART: 20/335 (6.0%), AZT alone vs TDF ART, p=0.10; AZT-ART vs TDF-ART, p=0.04
Infant mortality at <14 days1
Overall live births, N=3202
Total infant mortality, N=60 (1.9%)
Period 1+2 AZT vs AZT-ART comparisons (total events, N=45)
AZT alone: 28/1432 (2.0%)
AZT-ART: 17/1419 (1.2%), AZT alone vs AZT-ART, p=0.132
Period 1 (total events, N=32)
AZT alone: 17/1083 (1.6%); 17/28 deaths, 60.7% of deaths occurred in period 1
AZT-ART: 15/1073 (1.4%); 15/17 deaths, 88.2% of deaths occurred in period 1
Period 2, 3 arm comparisons (total events, N=28):
AZT-alone: 11/341 (3.2%); 11/28, 39.3% of deaths occurred in period 2
AZT-ART: 2/346 (0.6%); 2/17 deaths, only 11.8% of deaths occurred in period 2
TDF-ART: 15/341 (4.4%), AZT alone vs TDF ART, p=0.429; AZT-ART vs TDF-ART, p=0.001
1 Reported causes of neonatal mortality were prematurity (n=23), lower respiratory conditions (n=22), unknown causes (n=7), congenital malformations (n=5), sepsis (n=1), vitamin K deficiency (n=1), and persistent pulmonary hypertension of the newborn (n=1).
Stuart and colleagues protocol for a systematic review of psychological treatment for methamphetamine use  is a timely and relevant response to a significant gap in guidance for treatment of psychological sequelae arising from psychostimulant use. Given that levels of psychological distress and clinical mental health symptoms are very common in people who use methamphetamine, there is an urgent need for psychological treatment guidelines for this population.
While many psychological interventions to date have also targeted comorbid anxiety and depressive symptoms, psychotic symptoms are also highly prevalent, clinically challenging and often poorly addressed. Methamphetamine-associated psychosis has been associated with significant morbidity, including increased agitation and violence, presentation to acute health services, and considerable health service burden . A subset of individuals may also be at risk of developing more persistent psychotic symptoms, later transitioning to a diagnosis of chronic psychotic disorder.
To date, the presence of psychotic symptoms has either been excluded, or not routinely screened for, in the majority of studies of psychological treatment for methamphetamine use. For instance, some studies have excluded participants with recent antipsychotic medication use, while measurement of psychotic symptoms or clinical assessments of psychotic disorder are typically absent [4-7]. On the other hand, many pharmacological inte...
To date, the presence of psychotic symptoms has either been excluded, or not routinely screened for, in the majority of studies of psychological treatment for methamphetamine use. For instance, some studies have excluded participants with recent antipsychotic medication use, while measurement of psychotic symptoms or clinical assessments of psychotic disorder are typically absent [4-7]. On the other hand, many pharmacological interventions targeting methamphetamine-associated psychosis do not detail the psychological components of treatment. This therefore poses a challenge for reviewers seeking to evaluate psychosis as an outcome measure.
Clinicians working with this population struggle with the management of this syndrome, and need guidance in evidence-based approaches. The findings of this review will be important in shaping best practice for co-morbidity in this population, so we encourage the authors to consider how to include guidance on psychological interventions for methamphetamine-associated psychosis within the scope of their protocol.
1. Stuart A, Baker AL, Bowman J, McCarter K, Denham AMJ, Lee N, Colyvas K, Dunlop A, Protocol for a systematic review of psychological treatment for methamphetamine use: an analysis of methamphetamine use and mental health symptom outcomes. BMJ Open 2017; 7.
2. McKetin R, Degenhardt L, Shanahan M, Baker AL, Lee NK, Lubman DI, Health service utilisation attributable to methamphetamine use in Australia: Patterns, predictors and national impact. Drug and Alcohol Review 2017.
3. Shoptaw S, Reback CJ, Peck JA, Yang X, Rotheram-Fuller E, Larkins S, Veniegas RC, Freese TE, Hucks-Ortiz C, Behavioral treatment approaches for methamphetamine dependence and HIV-related sexual risk behaviors among urban gay and bisexual men. Drug and alcohol dependence 2005; 78:125-134.
4. Roll JM, Petry NM, Stitzer ML, Brecht ML, Peirce JM, McCann MJ, Blaine J, MacDonald M, DiMaria J, Lucero L, Contingency management for the treatment of methamphetamine use disorders. American Journal of Psychiatry 2006; 163:1993-1999.
5. Smout MF, Longo M, Harrison S, Minniti R, Wickes W, White JM, Psychosocial treatment for methamphetamine use disorders: a preliminary randomized controlled trial of cognitive behavior therapy and acceptance and commitment therapy. Substance abuse 2010; 31:98-107.
6. Rawson RA, McCann MJ, Flammino F, Shoptaw S, Miotto K, Reiber C, Ling W, A comparison of contingency management and cognitive‐behavioral approaches for stimulant‐dependent individuals. Addiction 2006; 101:267-274.
7. Baker A, Boggs TG, Lewin TJ, Randomized controlled trial of brief cognitive‐behavioural interventions among regular users of amphetamine. Addiction 2001; 96:1279-1287.
It is concerning to read on page 1 of the introduction that “Internationally, clinical guidelines recommend that pregnant women should abstain from heavy or ‘binge’ drinking'". The reference cites the liquor industry social aspects organisation [http://www.iard.org/] as their source, which we admit find rather unusual. What is more unfortunate is the authors then misrepresent the content found on the industry website. Rather than the guidelines recommending the avoidance of heavy drinking, all of the 54 countries’ guidelines for alcohol and pregnancy advise abstinence, and the only exception is Bulgaria which offers a low-risk level for pregnant women who do choose to consume. This appears contradictory to the statement made by the authors in their introduction.
We would like to thank the authors for their informative and innovative research, which interrogated electronic medical record free text, to demonstrate the significant impact of childhood respiratory illness on primary care workload in New Zealand. The findings indicate that almost half (46%) of all child-general practitioner consultations between January 2008 and December 2013 were attributed to the management of respiratory conditions.
The findings of this research are not surprising given that children in developed and developing countries report six to eight episodes of acute respiratory tract infection every year (1). The high rate of consultations may be partly explained by the fact that there is no cure for acute respiratory infections, only symptomatic relief (1). For some parents, this might necessitate frequent and repeat consultations with general practitioners.
While this research sheds light on the role of general practitioners in the management of childhood respiratory illness, there continues to persist ongoing knowledge gaps in an equally important field. There is some evidence to indicate that many parents manage their child’s acute respiratory infections at home, using a range of therapies, including complementary and alternative medicines (CAM)(2). With the increasing popularity of CAM and emerging high level evidence to support the role of CAM therapies for acute respiratory tract infection (e.g. Sambucus nigra (3), Pelargonium sidoides...
While this research sheds light on the role of general practitioners in the management of childhood respiratory illness, there continues to persist ongoing knowledge gaps in an equally important field. There is some evidence to indicate that many parents manage their child’s acute respiratory infections at home, using a range of therapies, including complementary and alternative medicines (CAM)(2). With the increasing popularity of CAM and emerging high level evidence to support the role of CAM therapies for acute respiratory tract infection (e.g. Sambucus nigra (3), Pelargonium sidoides (4), probiotics(5), Echinacea spp.(6)), it is probable that parents also may be accessing CAM through primary care settings, with practitioners of CAM typically based in primary care settings and readily accessible to parents. Notwithstanding, there is little to no robust data regarding this activity.
Primary care is a complex puzzle and it is important to recognise the diverse and varied services provided by a range of practitioners. Research such as this helps to solve crucial parts of this complex puzzle. But missing pieces of the puzzle remain and this requires similar concerted efforts in other areas of primary care. Dowell and colleagues have demonstrated the usefulness and impact of “big data” from primary care in understanding general practitioner workload. It is imperative that similar research explore other commonly utilised, increasingly popular services in primary care, such as CAM. In doing so, we will be able to move one step closer to solving the complex puzzle of primary care.
1. World Health Organization (WHO). Coughs and cold remedies for the treatment of acute respiratory infections in young children. Geneva, Switzerland:: WHO; 2001.
2. MacLennan A, Myers S, Taylor A. The continuing use of complementary and alternative medicine in South Australia: costs and beliefs in 2004. Medical Journal Australia. 2006;184(1):27-31.
3. Ulbricht C, Basch E, Cheung L, Goldberg H, Hammerness P, Isaac R, et al. An evidence-based systematic review of elderberry and elderflower (Sambucus nigra) by the Natural Standard Research Collaboration. . Journal of Dietary Supplements. 2014;11(1):80-120.
4. Timmer A, Günther J, Motschall E, Rücker G, Antes G, Kern WV. Pelargonium sidoides extract for treating acute respiratory tract infections. Cochrane Database Syst Rev. 2013;10:CD006323.
5. Hao Q, Dong BR, Wu T. Probiotics for preventing acute upper respiratory tract infections. Cochrane Database Syst Rev. 2015;2:CD006895.
6. Shah SA, Sander S, White CM, Rinaldi M, Coleman CI. Evaluation of echinacea for the prevention and treatment of the common cold: a meta-analysis. The Lancet Infectious Diseases. 2007;7(9):580.