Drenaje con tubo en T versus cierre primario después de la exploración abierta del colédoco
Resumen
Antecedentes
Entre el 5% y el 11% de los pacientes a los que se les realiza una colecistectomía presentan cálculos en el colédoco. Los cálculos se pueden eliminar en el momento de la colecistectomía al abrir y despejar el colédoco. No está claro cuál es la técnica óptima.
Objetivos
El objetivo es evaluar los efectos beneficiosos y perjudiciales del drenaje con tubo en T versus el cierre primario sin stent biliar después de la exploración abierta del colédoco para los cálculos del colédoco.
Métodos de búsqueda
Se hicieron búsquedas en el registro de ensayos controlados del Grupo Cochrane Hepatobiliar (Cochrane Hepato‐Biliar Group), el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials, CENTRAL) en The Cochrane Library, MEDLINE, EMBASE, y Science Citation Index Expanded hasta abril de 2013.
Criterios de selección
Se incluyeron todos los ensayos clínicos aleatorios que compararon el drenaje con tubo en T versus el cierre primario después de la exploración abierta del colédoco.
Obtención y análisis de los datos
Dos de los cuatro revisores identificaron los estudios para inclusión y extrajeron los datos de forma independiente. Los datos se analizaron con el modelo de efectos fijos y con el modelo de efectos aleatorios mediante análisis con Review Manager (RevMan). Para cada resultado, se calculó el cociente de riesgos (CR), el cociente de tasas (CT) o la diferencia de medias (DM) con los intervalos de confianza (IC) del 95%, según el análisis por intención de tratar.
Resultados principales
Se incluyeron seis ensayos que asignaron al azar a 359 participantes, 178 a drenaje con tubo en T y 181 a cierre primario. Todos los ensayos presentaron un alto riesgo de sesgo. No hubo diferencias significativas en la mortalidad entre los dos grupos (4/178 [porcentaje ponderado del 1,2%] en el grupo con tubo en T versus 1/181 [0,6%] en el grupo con cierre primario; CR 2,25; IC del 95%: 0,55 a 9,25; seis ensayos). No hubo diferencias significativas en la tasa de morbilidad grave entre los dos grupos (24/136 [tasa de morbilidad grave ponderada, 145 eventos por 1000 pacientes] en el grupo con tubo en T versus 9/136 (tasa de morbilidad grave ponderada, 66 eventos por 1000 pacientes) en el grupo con cierre primario; CT 2,19; IC del 95%: 0,98 a 4,91; cuatro ensayos). La calidad de vida y el retorno al trabajo no se informaron en los ensayos. El tiempo quirúrgico fue significativamente mayor en el grupo de drenaje con tubo en T comparado con el grupo de cierre primario (DM 28,90 minutos; IC del 95%: 17,18 a 40,62 minutos; un ensayo). La estancia hospitalaria fue significativamente más larga en el grupo de drenaje con tubo en T comparado con el grupo de cierre primario (DM 4,72 días; IC del 95%: 0,83 a 8,60 días; cinco ensayos).
Conclusiones de los autores
El drenaje con tubo en T pareció dar lugar a un tiempo quirúrgico y a una estancia hospitalaria significativamente mayores en comparación con el cierre primario, sin pruebas evidentes de efectos beneficiosos sobre resultados clínicamente importantes después de la exploración abierta del colédoco. Según las pruebas actualmente disponibles, no hay justificación para el uso sistemático del drenaje con tubo en T después de la exploración abierta del colédoco en pacientes con cálculos del colédoco. El drenaje con tubo en T no se debe utilizar fuera de ensayos clínicos aleatorios bien diseñados. Es posible que se necesiten más ensayos aleatorios que comparen los efectos del drenaje con tubo en T versus el cierre primario después de la exploración abierta del colédoco. Dichos ensayos deben tener bajo riesgo de sesgo y evaluar efectos beneficiosos y perjudiciales a largo plazo del drenaje con tubo en T, incluidas las complicaciones a largo plazo como la estenosis biliar y la recurrencia de los cálculos del colédoco.
Resumen en términos sencillos
Drenaje con tubo en T versus ningún drenaje con tubo en T después de la exploración abierta del colédoco
El hígado tiene diversas funciones. La producción de bilis es una de estas funciones. La bilis es necesaria para la digestión de la grasa y la eliminación de ciertos subproductos de desecho del hígado. La bilis producida en el hígado se almacena temporalmente en la vesícula biliar. Al comer alimentos grasos, la vesícula biliar libera la bilis en el intestino delgado. El colédoco es el conducto a través del que la bilis fluye del hígado a la vesícula biliar, y de allí al intestino delgado. Los cálculos pueden dificultar el flujo de la bilis de la vesícula biliar al intestino delgado. Generalmente dichos cálculos se forman en la vesícula biliar y migran al colédoco. La obstrucción del flujo de la bilis puede provocar ictericia. Dichos cálculos generalmente se eliminan con el uso de un endoscopio (al introducir un instrumento equipado con una cámara a través de la boca y en el intestino delgado) antes de la extracción por mínimo acceso de los cálculos biliares (colecistectomía laparoscópica) o como parte de la extracción por mínimo acceso de los cálculos biliares (exploración laparoscópica del colédoco). La exploración laparoscópica del colédoco solamente se puede realizar en centros muy especializados, por lo que el método utilizado habitualmente para tratar los cálculos en el colédoco es la eliminación endoscópica de los cálculos del colédoco. Sin embargo, cuando tal tratamiento endoscópico falla, el paciente tiene que someterse a la exploración abierta del colédoco. Lo anterior incluye la exploración del colédoco mediante instrumentos o una cámara, o ambos, que se introducen en el colédoco generalmente mediante un corte realizado en el mismo. Después que se eliminan los cálculos, se sutura el orificio en el colédoco. Tradicionalmente, los cirujanos han utilizado un tubo en T a través del corte en el colédoco. El tubo en T tiene la forma de la letra inglesa "T" como su nombre lo indica. La parte superior de la letra "T" se coloca dentro del colédoco, mientras la parte inferior larga de la letra "T" se sitúa fuera del abdomen mediante un corte pequeño y se conecta a una bolsa. Esta sonda se inserta con la intención de prevenir la acumulación de bilis en el colédoco debido al edema temporal, que es frecuente después de cualquier corte en cualquier parte del cuerpo. La acumulación de bilis junto con el edema puede potencialmente impedir la cicatrización del conducto biliar, lo que provoca la pérdida de bilis del colédoco hacia el abdomen. La pérdida biliar no controlada en la cavidad abdominal puede ser potencialmente mortal si no se detecta y no se trata de forma apropiada. Además de actuar como un drenaje de la bilis del colédoco al exterior, se puede inyectar colorante en el tubo en T y utilizar una radiografía para detectar cualquier cálculo residual. Una vez que se confirma la ausencia de cálculos residuales, se extrae el tubo en T. Sin embargo, los cirujanos tienen preocupación acerca del pequeño orificio que el tubo en T deja después de la extracción. Normalmente, este pequeño orificio en el colédoco sana sin secuelas, pero en algunos pacientes la bilis puede drenar a través del orificio y provocar el mismo problema que se trata de prevenir con el tubo en T. Por lo tanto, el uso del tubo en T después de la exploración abierta del colédoco es un tema polémico. Se intentó responder a la pregunta de si el cierre primario (suturar el corte en el conducto biliar sin colocar una tubo en T) es mejor que utilizar una tubo en T después de la exploración abierta del colédoco, y se revisó toda la información disponible en la bibliografía de ensayos clínicos aleatorios. Los ensayos clínicos aleatorios son tipos especiales de estudios clínicos que proporcionan las respuestas más válidas si se realizan correctamente.
Se identificaron seis ensayos que incluyeron a 359 participantes de los que 178 pacientes se asignaron a cierre primario y 181 a drenaje con tubo en T después de la exploración abierta del colédoco. Los seis ensayos tuvieron alto riesgo de sesgo (riesgo de sobrestimar los efectos beneficiosos y subestimar los efectos perjudiciales de la intervención o el control). No hubo diferencias significativas en la mortalidad (12 muertes por 1000 participantes del grupo de drenaje con tubo en T versus seis muertes por 1000 participantes del grupo de cierre primario) ni en la tasa de complicación grave después de la cirugía entre los dos grupos (aproximadamente 145 complicaciones por 1000 participantes del grupo de drenaje con tubo en T versus 66 complicaciones por 1000 participantes del grupo cierre primario). Aunque el número de muertes y las tasas de complicación en el grupo de cierre primario parecieron ser menos de la mitad de los del grupo con tubo en T, existe la posibilidad de que dicha observación no sea real, sino que la diferencia hay ocurrido por azar (similar a que hay una oportunidad en ocho de lanzar una moneda y que salga cara o cruz tres veces seguidas). Por este motivo, no puede haber suficiente seguridad de que exista un efecto verdadero y esta diferencia no se considera estadísticamente significativa. Ninguno de los ensayos informó la calidad de vida ni el tiempo hasta el retorno al trabajo. El tiempo quirúrgico promedio fue significativamente mayor en el grupo de drenaje con tubo en T que en el grupo de cierre primario (en cerca de 30 minutos). La estancia hospitalaria promedio fue significativamente más larga en el grupo con tubo en T que en el grupo de cierre primario (en cerca de cinco días). El uso del tubo en T parece aumentar el costo sin proporcionar efectos beneficiosos a los pacientes. Se necesitan ensayos aleatorios adicionales con bajo riesgo de sesgo (bajas probabilidades de establecer conclusiones equivocadas debido a la parcialidad de los profesionales sanitarios, los investigadores o los pacientes) y bajo riesgo de errores aleatorios (establecer conclusiones equivocadas debido al azar) para confirmar si aún se justifica el uso de el tubo en T. Hasta que los resultados de dichos ensayos estén disponibles, no se aconseja el uso habitual del tubo en T después de la exploración abierta del colédoco.
Authors' conclusions
Summary of findings
| T‐tube drainage compared with primary closure for patients undergoing open common bile duct exploration | ||||||
| Patient or population: patients undergoing open common bile duct exploration. | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Primary closure | T‐tube drainage | |||||
| Peri‐operative mortality | 6 per 1000 | 12 per 1000 | RR 2.25 | 359 | ⊕⊝⊝⊝ | |
| Serious morbidity (rate) | 66 per 1000 | 145 per 1000 | Rate ratio 2.19 | 272 | ⊕⊝⊝⊝ | |
| Serious morbidity (proportion) | 62 per 1000 | 158 per 1000 | RR 2.53 | 35 | ⊕⊝⊝⊝ | |
| Operating time | The mean operating time in the control groups was | The mean operating time in the intervention groups was | 40 | ⊕⊝⊝⊝ | ||
| Hospital stay | The mean hospital stay in the control groups was | The mean hospital stay in the intervention groups was | 333 | ⊕⊝⊝⊝ | ||
| *The basis for the assumed risk was the overall control group risk across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Most trials were at high risk of bias. | ||||||
Background
About 5% to 25% of the adult western population have gallstones (GREPCO 1984; GREPCO 1988; Bates 1992; Halldestam 2004). The annual incidence of gallstones is about 1 in 200 people (NIH 1992). Only 2% to 4% of people with gallstones become symptomatic with biliary colic (pain), acute cholecystitis (inflammation), obstructive jaundice, or gallstone pancreatitis in a year (Attili 1995; Halldestam 2004). Cholecystectomy (removal of gallstones) is the preferred option in the treatment of symptomatic gallstones (Strasberg 1993) and every year 1.5 million cholecystectomies are performed in the United States and 60,000 in the United Kingdom (Dolan 2009; HES 2011). Approximately 80% of cholecystectomies are performed laparoscopically (keyhole) (Ballal 2009). The reported proportion of patients with common bile duct (CBD) stones at the time of cholecystectomy varies between 5% (Kama 2001; Hemli 2004) and 11% (Pitluk 1979; Duensing 2000). Laparoscopic cholecystectomy is currently preferred over open cholecystectomy for elective cholecystectomy (Livingston 2004; Ballal 2009; Keus 2010). Although laparoscopic common bile duct exploration can be performed successfully in the majority of the patients (Hemli 2004), choledocholithiasis is one of the important causes of conversion to an open procedure (Bingener‐Casey 2002; Livingston 2004). Endoscopic retrograde cholangio‐pancreatography (ERCP) has decreased the indications for common bile duct exploration (Williams 1994), but open common bile duct exploration is still an important procedure if ERCP fails or where expertise for laparoscopic exploration of the common bile duct is not available.
After the common bile duct exploration is performed and the stones removed, the choice is made between primary duct closure and T‐tube drainage (Lygidakis 1983; Payne 1986; Williams 1994). Recently, biliary stent insertion has also been used as an alternative to T‐tube drainage (Isla 2004; Kim 2004; Griniatsos 2005). T‐tube drainage of the common bile duct is performed for the following reasons (Williams 1994).
-
Prevention of extravasation of bile from the common bile duct.
-
Post‐operative visualisation of the common bile duct.
-
Availability of a T‐tube to extract common bile duct stones with a Burhenne steerable catheter (Burhenne 1973).
However, the availability of ERCP (Cotton 1980; Worthley 1989) and choledochoscopy (Berci 1978) have reduced the importance of these indications (Payne 1986; Williams 1994).
There have been concerns about higher mortality and morbidity after T‐tube drainage (Lygidakis 1983), with reports of tube migration causing common bile duct obstruction (Bernstein 1994) and bile leaks following removal of the T‐tube (Kacker 1995).
This is an update of our previous review (Gurusamy 2007) with a revision in the methodology of the review and interpretation of information according to the current Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Objectives
To assess the benefits and harms of T‐tube drainage versus primary closure after open common bile duct exploration for common bile duct stones.
Methods
Criteria for considering studies for this review
Types of studies
We included all randomised clinical trials examining the effects of T‐tube drainage versus primary closure without a biliary stent after open common bile duct stone exploration, irrespective of language, blinding, or publication status.
Quasi‐randomised studies (where the method of allocating participants to a treatment was not strictly random, for example, date of birth, hospital record number, alternation) were not included in this review. However, we did include reports of harmful effects from quasi‐randomised studies and observational studies.
Types of participants
Participants who were about to undergo open exploration of the common bile duct for possible common bile duct stones.
Types of interventions
Trials comparing T‐tube drainage versus primary duct closure without biliary stent after open operative exploration of the common bile duct.
We considered all common bile duct explorations that were not performed through laparoscopy (keyhole) as open operative exploration of the common bile duct irrespective of the size of the incision used.
Types of outcome measures
Primary outcomes
-
Mortality.
-
Mortality at maximal follow‐up, usually assessed by hazard ratios.
-
Mortality within 30 days (procedure or operative mortality).
-
-
Serious morbidity, corresponding to the grade III or grade IV Clavien‐Dindo classification (Dindo 2004; Clavien 2009) or as classified by the authors. The Clavien‐Dindo classification approximately corresponds to the definition of serious adverse events by the International Conference on Harmonisation of technical requirements for registration of pharmaceuticals for human use ‐ Good Clinical Practice (ICH‐GCP) (ICH‐GCP 1997).
-
Quality of life.
Secondary outcomes
-
Operating time.
-
Length of hospital stay.
-
Return to work.
-
Recurrent or retained common bile duct stones at maximal follow‐up.
The summary of findings was reported for all the outcomes reported by at least one trial in a summary of findings Table for the main comparison (http://ims.cochrane.org/revman/other‐resources/gradepro).
Search methods for identification of studies
We searched the Cochrane Hepato‐Biliary Group Controlled Trials Register, Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library, MEDLINE, EMBASE, and Science Citation Index Expanded (Royle 2003) until April 2013. We also searched the World Health Organization International Clinical Trials Registry Platform (ICTRP) (http://apps.who.int/trialsearch/Default.aspx). This search portal allows searching of ClinicalTrials.gov and ISRCTN trial registers among various other trials registers. We have given the search strategies with the time spans for the searches in Appendix 1. Reference lists of the identified studies were also searched for further trials.
Data collection and analysis
Trial selection and extraction of data
KSG and RK, independently identified the trials for inclusion. KSG and RK also listed the excluded studies with the reasons for their exclusion. BRD adjudicated any differences in opinion.
Both authors independently extracted the following data.
-
Year and language of publication.
-
Country.
-
Year of conduct of the trial.
-
Inclusion and exclusion criteria.
-
Sample size.
-
Population characteristics such as age and sex ratio.
-
Antibiotic prophylaxis.
-
Abdominal drain use.
-
Duration of T‐tube drainage.
-
Whether routine cholangiogram was performed.
-
Outcomes (mentioned above).
-
Risk of bias (described below).
We sought clarification of any unclear or missing information by contacting the authors of the individual trials. If there was any doubt whether the trials shared the same participants, completely or partially (by identifying common authors and centres), we planned to contact the authors of the trials to clarify whether the trial reports had been duplicated.
We resolved any differences in opinion through discussion. BRD adjudicated any differences in opinion.
Assessment of risk of bias
The authors assessed the risk of bias in the trials independently, without masking of the trial names and following the instructions in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and the Cochrane Hepato‐Biliary Group Module (Gluud 2013). Due to the risk of overestimation of intervention effects in randomised trials with high risk of bias (Schulz 1995; Moher 1998; Kjaergard 2001; Wood 2008; Lundh 2012; Savovic 2012, Savovic 2012a), we assessed the influence of risk of bias on the trial results using the following domains.
Allocation sequence generation
-
Low risk of bias: sequence generation was achieved using computer random number generation or a random number table. Drawing lots, tossing a coin, shuffling cards, and throwing dice are adequate if performed by an independent adjudicator.
-
Uncertain risk of bias: the trial was described as randomised, but the method of sequence generation was not specified.
-
High risk of bias: the sequence generation method was not, or might not be, random. Quasi‐randomised studies, those using dates, names, or admittance numbers in order to allocate patients, were inadequate and were excluded for the assessment of benefits but not for assessing harms.
Allocation concealment
-
Low risk of bias: allocation was controlled by a central and independent randomisation unit, sequentially numbered opaque and sealed envelopes, or similar, so that intervention allocations could not have been foreseen in advance of, or during, enrolment.
-
Uncertain risk of bias: the trial was described as randomised, but the method used to conceal the allocation was not described so that intervention allocations might have been foreseen in advance of, or during, enrolment.
-
High risk of bias: if the allocation sequence was known to the investigators who assigned participants, or if the study was quasi‐randomised. Quasi‐randomised studies were excluded for the assessment of benefits but not for assessing harms.
Blinding of participants and personnel
-
Low risk of bias: blinding was performed adequately, or the outcome measurement was not likely to be influenced by lack of blinding.
-
Uncertain risk of bias: there was insufficient information to assess whether the type of blinding used was likely to induce bias on the estimate of effect.
-
High risk of bias: no blinding or incomplete blinding, and the outcome or the outcome measurement was likely to be influenced by lack of blinding.
Blinding of outcome assessors
-
Low risk of bias: blinding was performed adequately, or the outcome measurement was not likely to be influenced by lack of blinding.
-
Uncertain risk of bias: there was insufficient information to assess whether the type of blinding used was likely to induce bias on the estimate of effect.
-
High risk of bias: no blinding or incomplete blinding, and the outcome or the outcome measurement was likely to be influenced by lack of blinding.
Incomplete outcome data
-
Low risk of bias: the underlying reasons for missing data were unlikely to make treatment effects depart from plausible values, or proper methods had been employed to handle missing data.
-
Uncertain risk of bias: there was insufficient information to assess whether the missing data mechanism in combination with the method used to handle missing data were likely to induce bias on the estimate of effect.
-
High risk of bias: the crude estimate of effects (eg, complete case estimate) were clearly biased due to the underlying reasons for missing data, and the methods used to handle missing data were unsatisfactory.
Selective outcome reporting
-
Low risk of bias: pre‐defined or clinically relevant and reasonably expected outcomes were reported on.
-
Uncertain risk of bias: not all pre‐defined or clinically relevant and reasonably expected outcomes were reported on, or were not reported fully, or it was unclear whether data on these outcomes were recorded or not.
-
High risk of bias: one or more clinically relevant and reasonably expected outcomes were not reported on; data on these outcomes were likely to have been recorded.
Vested interest bias
-
Low risk of bias: the trial was conducted by a party without any vested interest in the outcome of the trial.
-
Uncertain risk of bias: it was not clear if the trial was conducted by a party with a vested interest in the outcome of the trial.
-
High risk of bias: the trial was conducted by a party with one or more vested interests in the outcome of the trial (such as a T‐tube manufacturer).
We considered trials to be at low risk of bias if we assessed all the above domains as being low risk of bias. In all other cases, the trials were considered as having high risk of bias.
Statistical methods
We performed the meta‐analyses according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and the Cochrane Hepato‐Biliary Group Module (Gluud 2013). We used the software package Review Manager 5 for statistical analyses (RevMan 2012). For dichotomous variables, we calculated the risk ratio (RR) with 95% confidence interval (CI) if there were two or more trials for an outcome. For continuous outcomes, we calculated the mean difference (MD) and the standardised mean difference (SMD) if different assessment scales were used, with 95% confidence intervals. For count data outcomes such as serious adverse events, the rate ratio was calculated using the methods shown in section 9.4.8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). For such a calculation, one needs the time over which the patients were exposed to the risk of serious morbidity in each of the groups. We considered that both groups were exposed to the risk of serious morbidity for the same time period, which is a reasonable assumption considering that the patients were followed up for the same time in both groups.
We used a random‐effects model (DerSimonian 1986) and a fixed‐effect model (DeMets 1987) for meta‐analysis in the presence of two or more trials included in the outcomes. In the case of a discrepancy between the two models, we reported both results; otherwise we reported the results of the fixed‐effect model. Heterogeneity was explored by a Chi2 test with significance set at P < 0.10, and the quantity of heterogeneity was measured by the I2 statistic (Higgins 2002; Higgins 2011).
The analyses were performed on an intention‐to‐treat basis (Newell 1992) whenever possible. Otherwise we adopted the 'available‐patient analysis' (Higgins 2011). We did not impute any data for post‐randomisation dropouts for any of the continuous outcomes. In the absence of summary information such as means and standard deviations for continuous outcomes, we used the median for the meta‐analysis and imputed the standard deviation from P values according to the instructions given in the Cochrane Handbook for Systematic Reviews of Intervention (Higgins 2011) (See Sensitivity analysis below). If it was not possible to calculate the standard deviation from the P value or the confidence interval, we imputed the standard deviation as the highest standard deviation in the other trials included under that outcome, fully recognising that this form of imputation would decrease the weight of the study for the calculation of mean differences and bias the effect estimate to no effect in the case of standardised mean differences (Higgins 2011).
Trial sequential analysis
Trial sequential analysis (TSA) was applied because cumulative meta‐analyses are at risk of producing random errors due to sparse data and repetitive testing of the accumulated data (www.ctu.dk/tsa). To minimise random errors, we calculated the required information size (that is, the number of participants needed in a meta‐analysis to detect or reject a certain intervention effect). The required information size calculation should also account for the heterogeneity or diversity present in the meta‐analysis. The underlying assumption of trial sequential analysis is that testing for significance may be performed each time a new trial is added to the meta‐analysis. We added the trials according to the year of publication, and if more than one trial was published in a year the trials were added alphabetically according to the last name of the first author. On the basis of the required information size, trial sequential monitoring boundaries were constructed. These boundaries determine the statistical inference one may draw regarding the cumulative meta‐analysis that has not reached the required information size; if the trial sequential monitoring boundary is crossed before the required information size is reached, firm evidence may perhaps be established and further trials may turn out to be superfluous. On the other hand, if the boundary is not surpassed it is most probably necessary to continue doing trials in order to detect or reject a certain intervention effect.
We performed trial sequential analysis (CTU 2011; Thorlund 2011) for binary outcomes and continuous outcomes in which at least two trials were present to determine if further trials were necessary for the outcome. We performed trial sequential analysis using an alpha error of 5%, beta error of 20%, control event proportion obtained from the results, and a relative risk reduction of 20% for binary outcomes (Brok 2008; Wetterslev 2008; Brok 2009; Thorlund 2009; Wetterslev 2009; Thorlund 2010). For continuous outcomes, we conducted trial sequential analysis using an alpha error of 5%, beta error of 20%, the variance of the included trials, and a minimal clinically important difference of one day for hospital stay.
Reporting bias
We planned to use a funnel plot to explore reporting bias (Egger 1997; Macaskill 2001) in the presence of at least 10 trials. We planned to use asymmetry in the funnel plot of study size against treatment effect to identify bias. We also planned to perform the linear regression approach described by Egger (Egger 1997) to determine the funnel plot asymmetry.
Subgroup analysis
We planned to perform the following subgroup analyses.
-
Trials at low risk of bias compared to those at high risk of bias.
-
Trials in which routine antibiotic prophylaxis was used compared to those in which routine antibiotic prophylaxis was not used.
-
Trials in which patients underwent open common bile duct exploration in the presence of acute cholecystitis compared to those in which patients underwent open common bile duct exploration in the absence of acute cholecystitis.
We planned to use the test for subgroup differences available through RevMan.
Sensitivity analysis
We planned to perform a sensitivity analysis by imputing the outcomes for binary outcomes under different scenarios, namely best‐best analysis, worst‐worst analysis, best‐worst analysis, and worst‐best analysis (Gurusamy 2009; Gluud 2013) for any significant binary outcomes. We also performed a sensitivity analysis by excluding the trials in which the median or standard deviation was imputed for the continuous outcomes. Both these sensitivity analyses were planned a priori (Gurusamy 2007). We also performed post hoc sensitivity analyses by considering the complications for which the severity could not be determined as 'mild' complications (best‐best scenario) and then we considered these 'mild' complications as severe complications (worst‐worst scenario). For example, bile leaks are post‐operative complications some of which resolve spontaneously without any prolongation of hospital stay, and hence were categorised as 'mild' adverse events. Other bile leaks need radiological or endoscopic interventions and hence were classified as 'serious' adverse events. If the trial authors did not report the way the bile leaks were treated, it was not possible to determine their severity. So we performed two analyses: one excluding the bile leaks which resolved spontaneously and those for which the treatment was not reported, and another analysis including the bile leaks for which the treatment was not reported.
Results
Description of studies
Results of the search
We identified a total of 1752 references through electronic searches of CENTRAL in The Cochrane Library (n = 179), MEDLINE (n = 759), EMBASE (n = 369), and Science Citation Index Expanded (n = 445). We excluded 542 duplicates and 1195 clearly irrelevant references through reading abstracts. Fifteen references were retrieved for further assessment. No trials were identified from the ICTRP portal or through scanning reference lists of the identified randomised trials. Of the 15 references, we excluded nine because of the reasons listed in the table 'Characteristics of excluded studies'. In total, six references of six randomised trials fulfilled the inclusion criteria. The reference flow is shown in Figure 1. All six trials were completed trials and could provide data for the analyses (Lygidakis 1983; Payne 1986; Makinen 1989; Williams 1994; Marwah 2004; Ambreen 2009). Details of the trials are shown in the table 'Characteristics of included studies'.
Study flow diagram.
Included trials
A total of 359 participants who underwent open common bile duct exploration were randomised in six trials. The number of participants in each trial ranged from 24 to 117. The average age of patients ranged between 44 years and 67 years in the trials that stated this. The proportion of females in the trials ranged between 16.2% and 85.7%.
In four trials, antibiotics were used routinely (Lygidakis 1983; Williams 1994; Marwah 2004; Ambreen 2009). In one trial, antibiotics were used only in patients with acute cholecystitis (Makinen 1989). The information on the use of antibiotics was not available in one trial (Payne 1986). The usage of an intra‐abdominal drain, timing of removal of the T‐tube, and whether a T‐tube cholangiogram was performed are described in the Characteristics of included studies table.
Risk of bias in included studies
All the trials were assessed as having high risk of bias (Figure 2; Figure 3).
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
Random sequence generation and allocation concealment
Only one trial (17%) was free from selection bias (Marwah 2004).
Blinding
None of the trials were free from performance bias or detection bias.
Incomplete outcome data
Two trials (33%) were free from attrition bias (Lygidakis 1983; Marwah 2004).
Selective reporting
Two trials (33%) were free from reporting bias (Marwah 2004; Ambreen 2009).
Other potential sources of bias
One trial (17%) was free from vested interest bias (Marwah 2004).
Effects of interventions
The results are summarised in summary of findings Table for the main comparison.
Mortality
Mortality at maximal follow‐up
Only three trials reported the outcomes beyond the peri‐operative period (Williams 1994; Marwah 2004; Ambreen 2009). The average follow‐up of patients was about 2.5 years in one trial (Williams 1994) and six months in the other two trials (Marwah 2004; Ambreen 2009). There were four deaths after the peri‐operative period that were unrelated to common bile duct stones in the first trial (Williams 1994). The groups to which these patients belonged were not reported. There was no long‐term mortality after the peri‐operative period in the remaining two trials (Marwah 2004; Ambreen 2009).
Mortality within 30 days (procedure or operative mortality)
All six trials reported this outcome. There was no significant difference in the peri‐operative mortality between the two groups (4/178 (adjusted proportion 1.2%) versus 1/181 (0.6%); RR 2.25; 95% CI 0.55 to 9.25; I2 = 0%) (Analysis 1.1).
Trial sequential analysis
The proportion of information accrued was only 0.26% of the required information size and so trial sequential monitoring boundaries were not drawn (Figure 4).
Trial sequential analysis of mortality
The diversity‐adjusted required information size (DARIS) was calculated to 140,645 patients, based on the proportion of patients in the control group with the outcome of 0.5%, a relative risk reduction of 20%, an alpha of 5%, a beta of 20%, and a diversity of 0%. After accruing 359 patients in the six trials, only 0.26% of the DARIS has been reached. Accordingly, the trial sequential analysis does not show the required information size and the trial sequential monitoring boundaries. As shown, the conventional boundaries have also not been crossed.
Serious morbidity (serious complications)
Results on serious complications were available from two trials (Marwah 2004; Ambreen 2009). These two trials were included in the outcome 'serious morbidity' (Analysis 1.2). The severity of the morbidity was not reported in three trials (Lygidakis 1983; Payne 1986; Makinen 1989). Of these three trials, two trials had at least a few adverse events which were clearly serious (Lygidakis 1983; Payne 1986). These two trials were included in the outcome 'serious morbidity' (Analysis 1.2) and in the outcome 'morbidity (severity unknown)' (Analysis 1.3). In one trial, it was not clear whether any of the morbidity was serious (Makinen 1989). This trial was included under the outcome 'morbidity (severity unknown)' (Analysis 1.3). In the last trial, which was not included for any of the two analyses, the morbidity was not reported separately in the two groups (Williams 1994).
There was no significant difference in the rate of serious morbidity between the two groups (rate ratio 2.19; 95% CI 0.98 to 4.91; P = 0.06; I2 = 26%) (Analysis 1.2). The morbidity (unknown severity) was significantly higher in the T‐tube group than the primary closure group by the fixed‐effect model (rate ratio 2.43; 95% CI 1.14 to 5.18; P = 0.02; I2 = 92%) but not by the random‐effects model (rate ratio 1.44; 95% CI 0.08 to 24.57; P = 0.80; I2 = 92%) (Analysis 1.3).
The proportion of patients who developed serious morbidity was reported in one trial (Ambreen 2009). There was no significant difference in the proportion of patients who developed serious morbidity between the groups (3/19 (15.8%) in T‐tube group versus 1/16 (6.3%) in primary closure group; Fisher's exact test P = 0.61) (Analysis 1.4).
Trial sequential analysis
Trial sequential analysis was not performed because of the presence of only trial measuring the proportion with serious morbidity. Trial sequential analysis (TSA) cannot be performed for rate ratios with the TSA Programme.
Quality of life
None of the included trials reported on quality of life.
Operating time
This information was available from only one trial (Marwah 2004). The operating time was significantly longer in the T‐tube drainage group versus the primary closure group (MD 28.90 minutes; 95% CI 17.18 to 40.62 minutes; P < 0.0001) (Analysis 1.5).
Hospital stay
This information was available in five trials (Payne 1986; Makinen 1989; Williams 1994; Marwah 2004; Ambreen 2009). The standard deviation was not reported and was imputed from the P value in one trial (Williams 1994). The meta‐analysis showed that the hospital stay was significantly longer in the T‐tube group than the primary closure group (MD 6.78 days; 95% CI 6.01 to 7.55; P < 0.00001; I2 = 96%) (Analysis 1.6). When the random‐effects model was adopted, the hospital stay was still significantly longer in the T‐tube group than in the primary closure group although the magnitude of the difference was decreased by two days (MD 4.72 days; 95% CI 0.83 to 8.60; P = 0.02; I2 = 96%). Exclusion of the trial in which the standard deviation was imputed from the P value (Williams 1994) did not alter the results.
Trial sequential analysis
The proportion of information accrued was only 1.62% of the required information size and so alpha‐spending boundaries were not drawn (Figure 5).
Trial sequential analysis of duration of hospital stay
The diversity‐adjusted required information size (DARIS) was 20,538 patients, based on a minimal relevant difference (MIRD) of one day, a variance (VAR) of 25.59, an alpha (a) of 5%, a beta (b) of 20%, and a diversity (D2) of 96.09%. After accruing 333 patients in the five trials, only 1.62% of the DARIS has been reached. Accordingly, the trial sequential analysis does not show the required information size and the trial sequential monitoring boundaries. As shown, the conventional boundary has been crossed for longer stay favouring primary closure.
Return to work
None of the included trials reported return to work.
Recurrent or retained common bile duct stones at long‐term follow‐up
Only three trials reported the outcomes beyond the peri‐operative period (Williams 1994; Marwah 2004; Ambreen 2009). The average follow‐up of patients was about 2.5 years in one trial (Williams 1994) and six months in the other two trials (Marwah 2004; Ambreen 2009). There were no recurrent or residual stones in either group in any of the three trials.
Subgroup analysis
All the trials were at high risk of bias. None of the trials separately reported the outcomes for patients with acute cholecystitis. For these reasons, the planned subgroup analyses were not performed. Only one trial was included under the outcome 'operating time' and so a subgroup analysis was not performed for this outcome. No subgroup analysis was performed for the outcome 'morbidity (severity unknown)' as this was not one of the outcomes chosen a priori. The Chi2 test for subgroup differences was not significant for the subgroup analysis of peri‐operative mortality or serious morbidity (Analysis 1.1; Analysis 1.2). However, it should be noted that the rate of serious morbidity was higher in the T‐tube group than the primary closure group in the subgroup of trials in which antibiotic prophylaxis was used (rate ratio 3.26; 95% CI 1.21 to 8.79; P = 0.02; I2 = 9%). The Chi2 test for subgroup differences was significant for hospital stay, by both the fixed‐effect model and random‐effects model (P < 0.00001 for fixed‐effect model, P < 0.007 for random‐effects model) (Analysis 1.6). The hospital stay was significantly longer in the T‐tube group than the primary closure group in the subgroup of trials in which routine prophylaxis was used. In the subgroup of trials in which antibiotics were used selectively and in which antibiotic use was not reported, there were no significant differences between the two groups and the mean effect estimate was to either side of the line of no effect; the trial in which an antibiotic was used selectively favoured the T‐tube group and the trial in which the antibiotic use was not reported favoured the primary closure group (Analysis 1.6).
Reporting bias
We did not perform a funnel plot because only six trials were included in this review.
Discussion
Summary of main results
This review has shown that T‐tube drainage after open choledochotomy appeared to increase the operating time and hospital stay compared to primary closure, without any evidence of difference in serious complications. One of the reasons for considering the use of T‐tube drainage is to decompress the common bile duct in the presence of distal obstruction of the common bile duct (Williams 1994). This is to avoid bile leak and bile peritonitis. One patient in the primary closure group developed biliary peritonitis (treatment not stated) (Makinen 1989). None of the other patients in the primary closure group developed biliary peritonitis or bile leak requiring re‐operation or other interventions. On the other hand, four patients developed bile leak or biliary peritonitis in the T‐tube group after removal; one requiring a re‐operation and three requiring percutaneous drainage (Williams 1994; Marwah 2004; Ambreen 2009). One patient developed peritubal leakage and died of necrotizing fascitis (Marwah 2004). Another patient required laparotomy and repositioning of the T‐tube for biliary peritonitis following T‐tube dislodgement (Payne 1986).
Another reason for considering the use of T‐tube drainage is to allow the possibility of extracting residual stones through the T‐tube tract (Williams 1994). However, no retained stones were detected after operative choledochoscopy (Menzies 1992). Even if patients develop symptomatic retained stones, ERCP and endoscopic sphincterotomy can be used for the treatment. Before the advent of ERCP, percutaneous transhepatic biliary drainage and open exploration were the only options available for the treatment of such patients. So, T‐tube drainage and extraction of stones may have been appropriate in the past while it has become redundant with the currently available treatments.
The definitions used for serious morbidity in this review were those used by The International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) and are used by regulatory authorities such as the Food and Drug Administration (FDA), Medicines and Healthcare products Regulatory Agency (MHRA), and Japan Society of Quality Assurance (JSQA) in the US, Europe, and Japan respectively (ICH‐GCP 1997) in determining the adverse events related to a pharmacological intervention. While it is not mandatory to report surgical interventions using these definitions, we have used these definitions in the absence of any other universal definitions for complications. Serious adverse events of the ICH‐GCP classification roughly correspond to grade III or grade IV complications in the Clavien‐Dindo classification (Dindo 2004; Clavien 2009), the only validated classification systems for assessing surgical interventions. Inevitably, some information will be lost by using such classifications, which pool complications together. However, unless such an approach is followed it is difficult to power trials related to surgery using any meaningful outcomes to the patients. There was no significant difference in the rates of serious morbidity between the two intervention groups. A subgroup analysis of patients included in the trials in which antibiotic prophylaxis was used revealed that the T‐tube group may have significantly higher morbidity rates than the primary closure group.
The post‐operative hospital stay was significantly longer in the T‐tube group than the primary closure group. There was significant heterogeneity in the results. Only one of the six trials reported a shorter post‐operative hospital stay in the T‐tube group than the primary closure group (not statistically significant) (Makinen 1989). In this trial, antibiotics were used only in patients with acute cholecystitis. There is some evidence that antibiotics are useful in open elective biliary tract surgery (including cholecystectomies and other operations on the bile duct) although this evidence is not recent and the evidence is generally poor (Meijer 1990). Clinical practice guidelines generally advocate antibiotic prophylaxis after biliary tract surgery (Mangram 1999). Although further investigation is necessary on this front, the selective use of antibiotics only in patients with acute cholecystitis may be a potential reason for the differences in the results noted in this trial (Makinen 1989). The evidence from the other four trials (Payne 1986; Williams 1994; Marwah 2004; Ambreen 2009) cannot be ignored because of the information from this one trial (Makinen 1989) when a potential reason for the difference exists.
The reason for the longer hospital stay in the T‐tube group is not clear. The reports did not state whether the patients stayed in the hospital until maturation of the tract and removal of the T‐tube. The mean and the standard deviation of the hospital stay were reported in four of the five trials that reported hospital stay (Payne 1986; Makinen 1989; Marwah 2004; Ambreen 2009). The standard deviation suggests that at least some of the patients in these four trials were discharged from hospital before the T‐tube was removed. This suggests that the patients did not wait in the hospital till the T‐tube was removed. One other potential reason for the longer hospital stay in the T‐tube group is serious morbidity. While there was no significant difference in the serious morbidity rates between the two intervention groups, this may be due to lack of evidence of effect rather than lack of effect. In the absence of any other explanation for the prolonged post‐operative hospital stay, one has to conclude that the longer hospital stay in the T‐tube group than in the primary closure group is likely due to higher morbidity.
A longer hospital stay is detrimental to both the patient (particularly in a private healthcare funding set‐up) and the healthcare provider (particularly in a state‐funded or insurance company funded healthcare funding set‐up). In addition to the longer hospital stay, the other costs in the T‐tube group include the cost of the T‐tube, increased operating time (although reported in only trial, T‐tube group is likely to result in a longer operating time than primary closure since T‐tube drainage involves additional surgical steps), the cost of a T‐tube cholangiogram, and the cost of removal of the T‐tube. Unless these costs can be offset by decreased complications, the costs associated with the use of a T‐tube cannot be justified. With five of the six trials showing increased morbidity rates in the T‐tube group compared with the primary closure group (not statistically significant), there appears to be little justification in the use of a T‐tube after open common bile duct exploration.
It appears that the T‐tube has a significant potential to harm the patients and increase the costs for the healthcare funder without any notable benefit based on the information obtained from the trials included in this review. There is additional corroborative information. We identified five non‐randomised studies comparing T‐tube to primary closure following common bile duct exploration (De Roover 1989; Sheen 1990; Sorensen 1994; Xu 2002; Joshi 2010) from the search strategy. All these studies reported complications related to T‐tube drainage such as biliary peritonitis (0/11 (0%) (De Roover 1989); 0/15 (0%) (Sheen 1990); 0/17 (0%) (Sorensen 1994); 4/171 (2.3%) (Xu 2002); 1/16 (6.3%) (Piecuch 2004); and 4/31(12.9%) (Joshi 2010)), biliary pancreatitis (2/17 (11.8%) (Sorensen 1994); 1/16 (6.3%) (Piecuch 2004)), and dislodgement of the T‐tube (3/31 (9.7%)) (Joshi 2010). In contrast, in the primary closure group none of the patients had biliary peritonitis (0/11 (0%) (De Roover 1989); 0/15 (0%) (Sheen 1990); 0/10 (0%) (Sorensen 1994); 0/215 (0%) (Xu 2002); 0/36 (0%) (Piecuch 2004); 0/40 (0%) (Joshi 2010)) or biliary pancreatitis (0/10 (0%)) (Sorensen 1994). In addition, in another Cochrane review, which included 295 patients from three trials, we found that T‐tube drainage had a longer operating time and longer post‐operative hospital stay than primary closure of the common bile duct without stent in laparoscopic common bile duct exploration, without significantly affecting the serious morbidity rates (Gurusamy 2013). In that Cochrane review of primary closure versus T‐tube closure in patients undergoing laparoscopic common bile duct exploration, fewer patients developed biliary complications requiring intervention after primary closure (not statistically significant) (Gurusamy 2013), as in the case of this review. The purpose of T‐tube drainage is the same whether the patients undergo open or laparoscopic common bile duct exploration. Thus, there is corroborative evidence that supports the findings of the present review.
Overall completeness and applicability of evidence
The results of this review are applicable only in patients undergoing open common bile duct exploration. The advent of ERCP has decreased the indications for open common bile duct exploration (Williams 1994). However, open common bile duct exploration is indicated in failed ERCP in places where expertise is not available for laparoscopic exploration of the common bile duct. Williams et al (Williams 1994) specifically mention inclusion of six patients with failed ERCP. This trial also found primary closure to be safe. Thus, failed ERCP need not be a contra‐indication for primary closure provided that there is no outflow obstruction. However, further trials are necessary. None of the trials mentioned whether ampullary dilatation was performed as a part of surgery. However, three trials specifically mentioned that they excluded patients with outflow obstruction from their trial (Makinen 1989; Williams 1994; Marwah 2004). One trial mentioned that bile flow into the duodenum was confirmed by flushing the common bile duct (Ambreen 2009). Choledochoscopy or a cholangiogram that confirms the absence of bile duct outflow obstruction is a pre‐requisite for primary closure.
Quality of the evidence
Although the risk of bias in the included trials was high and the overall quality of evidence was very low, one has to put this into perspective. This is currently the best available evidence. The previous routine use of T‐tube was based on clinical opinion and the fact that a significant proportion of the patients had to undergo major re‐operations in the case where stones were left in the bile duct. The advent of ERCP appears to have made the use of T‐tubes redundant. However, one cannot be completely dogmatic and conclude that there is no need for further research on this topic. The trials were at high risk of bias (for example, outcomes such as hospital stay and return to work can be biased because of lack of blinding and absence of any criteria for discharge from hospital or return to work, and a surgeon favouring primary closure might have discharged the patient from hospital earlier or may have advised the patient that he or she can return to work earlier) and new trials are necessary for this comparison as shown by the trial sequential analyses. Until such new trials with low risks of bias show that T‐tube use is safe and effective, the use of T‐tubes after open common bile duct exploration should be confined to randomised clinical trials. Such randomised trials should include quality of life and return to work as these outcomes are useful to determine whether a T‐tube is beneficial to the patient and whether it is cost‐effective.
Potential biases in the review process
We have followed the Cochrane methodology for performing the review. One potential bias was that we imputed the mean and standard deviation from median and other measures such as the P value or inter‐quartile range. This may have introduced an error in the effect estimate. The sensitivity analysis by excluding the trials in which the imputation was performed did not change the results. The alternative to this imputation is to present the information as presented by the authors, but the interpretation of that information can be even more confusing.
Agreements and disagreements with other studies or reviews
In the previous version of the review, we concluded that primary closure was at least as effective as T‐tube drainage after open common bile duct exploration (Gurusamy 2007). In this review, we have gone one step further and advocate against the use of a T‐tube outside well designed randomised clinical trials.
Study flow diagram.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Trial sequential analysis of mortality
The diversity‐adjusted required information size (DARIS) was calculated to 140,645 patients, based on the proportion of patients in the control group with the outcome of 0.5%, a relative risk reduction of 20%, an alpha of 5%, a beta of 20%, and a diversity of 0%. After accruing 359 patients in the six trials, only 0.26% of the DARIS has been reached. Accordingly, the trial sequential analysis does not show the required information size and the trial sequential monitoring boundaries. As shown, the conventional boundaries have also not been crossed.
Trial sequential analysis of duration of hospital stay
The diversity‐adjusted required information size (DARIS) was 20,538 patients, based on a minimal relevant difference (MIRD) of one day, a variance (VAR) of 25.59, an alpha (a) of 5%, a beta (b) of 20%, and a diversity (D2) of 96.09%. After accruing 333 patients in the five trials, only 1.62% of the DARIS has been reached. Accordingly, the trial sequential analysis does not show the required information size and the trial sequential monitoring boundaries. As shown, the conventional boundary has been crossed for longer stay favouring primary closure.
Comparison 1 T‐tube drainage versus primary closure, Outcome 1 Peri‐operative mortality.
Comparison 1 T‐tube drainage versus primary closure, Outcome 2 Serious morbidity.
Comparison 1 T‐tube drainage versus primary closure, Outcome 3 Morbidity (severity unknown).
| Study | T‐tube | Primary closure | Fisher's exact test P‐value |
| Ambreen 2009 | 3/19 (15.8%) | 1/16 (6.3%) | P = 0.6081 |
Comparison 1 T‐tube drainage versus primary closure, Outcome 4 Serious morbidity.
| Study | T‐tube mean (standard deviation) minutes(number of patients = 20) | Primary closure mean (standard deviation) minutes (number of patients = 20) | Mean difference (95% confidence intervals) | P value |
| Marwah 2004 | 116.65 (20.53) | 87.75 (17.12) | 28.90 minutes (17.18 to 40.62) | < 0.00001 |
Comparison 1 T‐tube drainage versus primary closure, Outcome 5 Operating time.
Comparison 1 T‐tube drainage versus primary closure, Outcome 6 Hospital stay.
| T‐tube drainage compared with primary closure for patients undergoing open common bile duct exploration | ||||||
| Patient or population: patients undergoing open common bile duct exploration. | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Primary closure | T‐tube drainage | |||||
| Peri‐operative mortality | 6 per 1000 | 12 per 1000 | RR 2.25 | 359 | ⊕⊝⊝⊝ | |
| Serious morbidity (rate) | 66 per 1000 | 145 per 1000 | Rate ratio 2.19 | 272 | ⊕⊝⊝⊝ | |
| Serious morbidity (proportion) | 62 per 1000 | 158 per 1000 | RR 2.53 | 35 | ⊕⊝⊝⊝ | |
| Operating time | The mean operating time in the control groups was | The mean operating time in the intervention groups was | 40 | ⊕⊝⊝⊝ | ||
| Hospital stay | The mean hospital stay in the control groups was | The mean hospital stay in the intervention groups was | 333 | ⊕⊝⊝⊝ | ||
| *The basis for the assumed risk was the overall control group risk across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Most trials were at high risk of bias. | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Peri‐operative mortality Show forest plot | 6 | 359 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.25 [0.55, 9.25] |
| 1.1 Routine antibiotic prophylaxis | 4 | 255 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.08 [0.43, 10.14] |
| 1.2 Selective antibiotic prophylaxis | 1 | 24 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 1.3 Antibiotic prophylaxis unknown | 1 | 80 | Risk Ratio (M‐H, Fixed, 95% CI) | 3.0 [0.13, 71.51] |
| 2 Serious morbidity Show forest plot | 4 | Rate Ratio (Fixed, 95% CI) | 2.19 [0.98, 4.91] | |
| 2.1 Routine antibiotic prophylaxis | 3 | Rate Ratio (Fixed, 95% CI) | 3.26 [1.21, 8.79] | |
| 2.2 Antibiotic prophylaxis unknown | 1 | Rate Ratio (Fixed, 95% CI) | 1.0 [0.25, 4.02] | |
| 3 Morbidity (severity unknown) Show forest plot | 3 | 221 | Rate Ratio (Fixed, 95% CI) | 2.43 [1.14, 5.18] |
| 4 Serious morbidity Show forest plot | Other data | No numeric data | ||
| 5 Operating time Show forest plot | Other data | No numeric data | ||
| 6 Hospital stay Show forest plot | 5 | 333 | Mean Difference (IV, Random, 95% CI) | 4.72 [0.83, 8.60] |
| 6.1 Routine antibiotic prophylaxis | 3 | 138 | Mean Difference (IV, Random, 95% CI) | 7.51 [3.36, 11.66] |
| 6.2 Selective antibiotic prophylaxis | 1 | 117 | Mean Difference (IV, Random, 95% CI) | ‐3.10 [‐8.85, 2.65] |
| 6.3 Antibiotic prophylaxis unknown | 1 | 78 | Mean Difference (IV, Random, 95% CI) | 1.70 [‐0.39, 3.79] |
