By Michelle van Slobbe
Introduction
In innovative therapies, you have read about several treatments that may or may not prevent preterm labour. One of the possibly successful therapies that have been discussed is progesterone gel. As you have read this progesterone gel treatment has been investigated in several trials, of which one was done by Hassan et al.[1] In this trial the effect of progesterone treatment was investigated in women with a sonographic (echographic) short cervix. This particular group was chosen because these women have a higher risk of preterm labour. Research done by Andersen et al.[2] shows that women with a cervix shorter than 39 mm have a 25% risk of preterm labour, while women with a longer cervix only have a 6.7% risk. The results of the trial done by Hassan et al. were published in 2011 in Ultrasound in Obstetrics & Gynecology with the title: “Vaginal progesterone reduces the rate of preterm birth in women with a sonographic short cervix: a multicenter, randomized, double-blind, placebo-controlled trial”. In the innovative therapies, this trial is discussed to comment on the effectivity of treatment with progesterone gel. However, when using articles to comment on subjects, it is always important to find out how reliable a certain article is. In this critical appraisal the reliability of the article is discussed so we can answer the question: can, from this research, be concluded that progesterone treatment is effective in preventing preterm labour in women with a sonographic short cervix?
Summary of the article
Hassan et. al randomized 465 women with a sonographic short cervix into two groups: one control group and one treatment group. To be included in the trial women had to be pregnant with only one child, be between 19 weeks and 23 weeks and 6 days of gestation, had to have a sonographic cervix of between 10 and 20 mm and could not have any signs or symptoms of preterm labour. The control group received treatment with a placebo gel, and the treatment group received progesterone gel. The treatment was started at 20 to 23 and 6 days of gestation and was stopped at 36 weeks and 6 days of gestation or when rupture of membranes or delivery occurred. The outcomes that were looked at were primarily preterm birth before 33 weeks of gestation and secondarily neonatal morbidity (e.g. sepsis, respiratory distress syndrome, perinatal mortality). The main analysis of the results was made with an intention to treat analysis, but also treated patient analysis was done and a compliant analysis. Of the 465 women that were randomized, 7 women were lost to follow up so eventually 235 women were in the control arm and 223 in the treatment arm. The primary outcome was that 8.9% of the women in the treatment arm delivered their baby preterm, and 16.1% of the women in the placebo group (relative risk 0.55 P = 0.02). The secondary outcomes such as frequency of respiratory distress syndrome (RDS), sepsis and neonatal death were also better in the treatment group than in the placebo group. The authors calculated that number-needed-to-treat (NNT) to prevent one preterm birth before 33 weeks of gestation is 14, and the NNT to prevent one child developing RDS is 22. The authors thus concluded: “The administration of vaginal progesterone gel to women with a sonographic short cervix in the mid-trimester is associated with a 45% reduction in the rate of preterm birth before 33 weeks of gestation and with improved neonatal outcome”.[1]
Critique
To be able to comment on the reliability of this article the method and result representation will be discussed.
Randomization and blinding
Other important factors to analyse in order to determine the quality of this trial are the randomization and blinding procedure. In this trial, randomization was carried out using an interactive voice response system and concealment of treatment allocation was accomplished. The patients and caregivers were blinded and this was maintained by the medication casing and medication itself having exactly the same appearance. The primary and secondary outcomes were all administered into the database before the blinding was lifted. The randomization and blinding are carried out very thoroughly and this makes the trail more reliable.
Statistical analysis
The next important factor that determines the quality of the article is the statistical analysis was carried out. In this article, three analyses were done: an intention to treat, a treated patient, and a compliant analysis. In the intention to treat analysis the patients that were lost to follow up were not included. According to Montori and Guyatt [3]: “inferences from studies with appreciable loss to follow-up are usually weaker.” In the case of this article 7 women of the 465 that were randomized were lost to follow up, which is in my opinion, not an appreciable loss. This makes the intention to treat analysis still reliable. The intention to treat analysis is the most “safe” analysis to draw your conclusion on because it shows the results that are the closest to what would happen when the drug would be used in practice. In this article, the intention to treat analysis showed significant results on which the conclusion of the article could be drawn. In the article, the two additional analysis demonstrated the same overall conclusion as the intention to treat analysis.
Power and sample size
What is also important to discuss regarding a trial is whether the sample size was large enough to draw a certain conclusion. The authors calculated that they needed 450 women in their trial (225 in every treatment arm) to reach a power of more than 90%. Unfortunately, the control group in the intention to treat analysis only contained 223 patients. However, this does not mean the results are not reliable at all because the power was almost reached and the power was set at 90% which is very high (usually the power is set at 80% )[4]. The power in the study was, however, not high enough to prove differences in the endpoints according to risk strata[1], which is a limitation of this study.
Representation of results
To be able to assess the results as a reader of the article, it is important that all the results are shown clearly in the article. In this article, all results were mentioned with the corresponding confidence interval and P-value. Due to this, the reader can form his/her own opinion based on the objective results. This is thus a strength of the article.
Conflict of interest
The last factor that one should always be cautious of when assessing the reliability of a trial, is the conflict of interest that might have occurred. In this research, the manufacturing company of the drug that was investigated was one of the major sponsors. In some cases, the reliability of the trial may be doubted as a result of this. However, in this article was mentioned that “The funding sources (NICHD/NIH and Columbia Laboratories, Inc.) were not involved in writing the report or the decision to submit the paper for publication”. [1] In addition to that, the authors also mention that the outcomes of the trial were independent of the payment that was received. Because they are honest about the way the trail was sponsored, and the drug manufacturer is not in any way involved in the research, the results can be assessed as reliable.
Conclusion
To conclude, this trial has a lot of strengths, and also some weaknesses. The strengths of this trial are the well-performed randomization and blinding, the big external validity and the clear representation of the results. The weaknesses are the surrogate primary endpoint and the insufficient power to prove differenced according to variable risk strata. Overall, the results of this trail can be assessed as reliable and that treatment with progesterone gel in pregnant women with a sonographic short cervix is effective.
Introduction
In innovative therapies, you have read about several treatments that may or may not prevent preterm labour. One of the possibly successful therapies that have been discussed is progesterone gel. As you have read this progesterone gel treatment has been investigated in several trials, of which one was done by Hassan et al.[1] In this trial the effect of progesterone treatment was investigated in women with a sonographic (echographic) short cervix. This particular group was chosen because these women have a higher risk of preterm labour. Research done by Andersen et al.[2] shows that women with a cervix shorter than 39 mm have a 25% risk of preterm labour, while women with a longer cervix only have a 6.7% risk. The results of the trial done by Hassan et al. were published in 2011 in Ultrasound in Obstetrics & Gynecology with the title: “Vaginal progesterone reduces the rate of preterm birth in women with a sonographic short cervix: a multicenter, randomized, double-blind, placebo-controlled trial”. In the innovative therapies, this trial is discussed to comment on the effectivity of treatment with progesterone gel. However, when using articles to comment on subjects, it is always important to find out how reliable a certain article is. In this critical appraisal the reliability of the article is discussed so we can answer the question: can, from this research, be concluded that progesterone treatment is effective in preventing preterm labour in women with a sonographic short cervix?
Summary of the article
Hassan et. al randomized 465 women with a sonographic short cervix into two groups: one control group and one treatment group. To be included in the trial women had to be pregnant with only one child, be between 19 weeks and 23 weeks and 6 days of gestation, had to have a sonographic cervix of between 10 and 20 mm and could not have any signs or symptoms of preterm labour. The control group received treatment with a placebo gel, and the treatment group received progesterone gel. The treatment was started at 20 to 23 and 6 days of gestation and was stopped at 36 weeks and 6 days of gestation or when rupture of membranes or delivery occurred. The outcomes that were looked at were primarily preterm birth before 33 weeks of gestation and secondarily neonatal morbidity (e.g. sepsis, respiratory distress syndrome, perinatal mortality). The main analysis of the results was made with an intention to treat analysis, but also treated patient analysis was done and a compliant analysis. Of the 465 women that were randomized, 7 women were lost to follow up so eventually 235 women were in the control arm and 223 in the treatment arm. The primary outcome was that 8.9% of the women in the treatment arm delivered their baby preterm, and 16.1% of the women in the placebo group (relative risk 0.55 P = 0.02). The secondary outcomes such as frequency of respiratory distress syndrome (RDS), sepsis and neonatal death were also better in the treatment group than in the placebo group. The authors calculated that number-needed-to-treat (NNT) to prevent one preterm birth before 33 weeks of gestation is 14, and the NNT to prevent one child developing RDS is 22. The authors thus concluded: “The administration of vaginal progesterone gel to women with a sonographic short cervix in the mid-trimester is associated with a 45% reduction in the rate of preterm birth before 33 weeks of gestation and with improved neonatal outcome”.[1]
Critique
To be able to comment on the reliability of this article the method and result representation will be discussed.
Primary and secondary outcomes
What is firstly important to discuss are the outcomes that the trial is investigating. The primary outcome of this trial is preterm birth before 33 weeks of gestation. The primary goal of the treatment that is investigated in this study is, however, not to prevent preterm labour, but to improve neonatal health. The primary outcome in this study is thus a so-called surrogate outcome, which is a weakness of this trial.[1]
What is firstly important to discuss are the outcomes that the trial is investigating. The primary outcome of this trial is preterm birth before 33 weeks of gestation. The primary goal of the treatment that is investigated in this study is, however, not to prevent preterm labour, but to improve neonatal health. The primary outcome in this study is thus a so-called surrogate outcome, which is a weakness of this trial.[1]
Randomization and blinding
Other important factors to analyse in order to determine the quality of this trial are the randomization and blinding procedure. In this trial, randomization was carried out using an interactive voice response system and concealment of treatment allocation was accomplished. The patients and caregivers were blinded and this was maintained by the medication casing and medication itself having exactly the same appearance. The primary and secondary outcomes were all administered into the database before the blinding was lifted. The randomization and blinding are carried out very thoroughly and this makes the trail more reliable.
Statistical analysis
The next important factor that determines the quality of the article is the statistical analysis was carried out. In this article, three analyses were done: an intention to treat, a treated patient, and a compliant analysis. In the intention to treat analysis the patients that were lost to follow up were not included. According to Montori and Guyatt [3]: “inferences from studies with appreciable loss to follow-up are usually weaker.” In the case of this article 7 women of the 465 that were randomized were lost to follow up, which is in my opinion, not an appreciable loss. This makes the intention to treat analysis still reliable. The intention to treat analysis is the most “safe” analysis to draw your conclusion on because it shows the results that are the closest to what would happen when the drug would be used in practice. In this article, the intention to treat analysis showed significant results on which the conclusion of the article could be drawn. In the article, the two additional analysis demonstrated the same overall conclusion as the intention to treat analysis.
Power and sample size
What is also important to discuss regarding a trial is whether the sample size was large enough to draw a certain conclusion. The authors calculated that they needed 450 women in their trial (225 in every treatment arm) to reach a power of more than 90%. Unfortunately, the control group in the intention to treat analysis only contained 223 patients. However, this does not mean the results are not reliable at all because the power was almost reached and the power was set at 90% which is very high (usually the power is set at 80% )[4]. The power in the study was, however, not high enough to prove differences in the endpoints according to risk strata[1], which is a limitation of this study.
Representation of results
To be able to assess the results as a reader of the article, it is important that all the results are shown clearly in the article. In this article, all results were mentioned with the corresponding confidence interval and P-value. Due to this, the reader can form his/her own opinion based on the objective results. This is thus a strength of the article.
External validity
What is also very important to discuss when appraising a trial, is whether the results of the study are applicable to the general population. Because this trial was conducted in several countries there was a substantial representation of several ethnic groups so that the results can be applied to African-Americans, Asians, and Caucasian people.
What is also very important to discuss when appraising a trial, is whether the results of the study are applicable to the general population. Because this trial was conducted in several countries there was a substantial representation of several ethnic groups so that the results can be applied to African-Americans, Asians, and Caucasian people.
Conflict of interest
The last factor that one should always be cautious of when assessing the reliability of a trial, is the conflict of interest that might have occurred. In this research, the manufacturing company of the drug that was investigated was one of the major sponsors. In some cases, the reliability of the trial may be doubted as a result of this. However, in this article was mentioned that “The funding sources (NICHD/NIH and Columbia Laboratories, Inc.) were not involved in writing the report or the decision to submit the paper for publication”. [1] In addition to that, the authors also mention that the outcomes of the trial were independent of the payment that was received. Because they are honest about the way the trail was sponsored, and the drug manufacturer is not in any way involved in the research, the results can be assessed as reliable.
Conclusion
To conclude, this trial has a lot of strengths, and also some weaknesses. The strengths of this trial are the well-performed randomization and blinding, the big external validity and the clear representation of the results. The weaknesses are the surrogate primary endpoint and the insufficient power to prove differenced according to variable risk strata. Overall, the results of this trail can be assessed as reliable and that treatment with progesterone gel in pregnant women with a sonographic short cervix is effective.
References of critical appraisal
[1] Hassan SS, Romero R, Vidyadhari D, Fusey S, Baxter JK, Khandelwal M, Vijayaraghavan J, Trivedi Y, Soma-Pillay P, Sambarey P, Dayal A, Potapov V, O'Brien J, Astakhov V, Yuzko O, Kinzler W, Dattel B, Sehdev H, Mazheika L, Manchulenko D, Gervasi MT, Sullivan L, Conde-Agudelo A, Phillips JA, Creasy GW; PREGNANT Trial. Vaginal progesterone reduces the rate of preterm birth in women with a sonographic short cervix: a multicenter, randomized, double-blind, placebo-controlled trial. Ultrasound Obstet Gynecol. 2011 Jul;38(1):18-31.
[2] Andersen HF1, Nugent CE, Wanty SD, Hayashi RH. Prediction of risk for preterm delivery by ultrasonographic measurement of cervical length. Am J Obstet Gynecol. 1990 Sep;163(3):859-67.
[3] Montori VM, Guyatt GH. Intention-to-treat principle. CMAJ. 2001 Nov 13; 165(10): 1339–1341
[4] Kadam P, Bhalerao S. Sample size calculation. Int J Ayurveda Res. 2010 Jan-Mar; 1(1): 55–57.
No comments:
Post a Comment