Waghalsige chirurgische Studienergebnisse: Ohne Antibiotikatherapie würde eine akute unkomplizierte Divertikulitis genauso harmlos und unauffällig verlaufen, wie mit dem Einsatz von Amoxycillin/Clavulansäure bzw. Ciprofloxacin/Metronidazol. Das wollte ein niederländisches Forschungsteam herausgefunden haben.
Lidewine Daniels et al. hatten 528 Patienten mit computertomografisch gesicherter und erstmals aufgetretener linksseitiger akuter unkomplizierter Divertikulitis vom Typ 1a oder 1b zufällig in zwei Gruppen geteilt: Die erste Gruppe bekamen Amoxicillin/Clavulansäure und bei Allergie Ciprofloxacin/Metronidazol. Die zweite Gruppe wurde nur überwacht.
Verlaufsbeobachtungen
Bis zum Abklingen der Divertikulitis dauerte es ohne Antibiotika 14 Tage, mit Antibiotika zwölf Tage. Angeblich war bei den Patienten ohne Antibiotika die Erkrankung nicht häufiger im Verlauf komplizierter geworden als bei denjenigen mit antibiotischer Therapie, obwohl der Unterschied 3,8 vs. 2,6 Prozent betrug.
Methodische Schwächen
Seltsamerweise war von einer doppelten Verblindung („double blind“) an keiner Stelle der Studie die Rede. Es heißt lediglich, sie (die Patienten) wurden willkürlich einer Beobachtungs- oder einer Antibiotikabehandlungsstrategie zugeteilt [„and assigned randomly to an observational or antibiotic treatment strategy“]. Stutzig macht auch, dass im Gegensatz zu allen anderen Ergebnissen, die tendenziell schlechter in der Beobachtungsgruppe als in der Antibiotika-Gruppe waren, ausgerechnet die stationäre Verweildauer in der Beobachtungsgruppe signifikant kürzer als in der Antibiotikagruppe gewesen sein soll [„Hospital stay was signicantly shorter in the observation group (2 versus 3 days; P = 0.006)“].
Widersprüche
Im krassen Widerspruch dazu steht die Rate der stationären Wiederaufnahme [„readmission (17.6 versus 12.0 per cent; P = 0.148)“] um 46 Prozent höher in der Gruppe ohne Antibiotika. Bei allen anderen Kriterien schnitt die Beobachtungsgruppe zum Teil deutlich oder dramatisch schlechter ab. Die angeblich fehlende Signifikanz bei den sekundären Endpunkten „komplizierte Divertikulitis“, „weiter bestehende Divertikulitis“, „rezidivierende Divertikulitis“, „operative Sigmoid-Resektion“ (!), die bereits beschriebene „stationäre Wiederaufnahme“, „Nebenwirkungen“ und schlussendlich „die Mortalität“ können nicht darüber hinwegtäuschen, dass der generelle Verzicht auf eine primäre Antibiotikatherapie auch bei angeblich akuter unkomplizierter Divertikulitis ein brandgefährliches Vorgehen darstellt [„secondary endpoints: complicated diverticulitis (3.8 versus 2.6 per cent respectively; P = 0.377), ongoing diverticulitis (7.3 versus 4.1 per cent; P = 0.183), recurrent diverticulitis (3.4 versus 3.0 per cent; P = 0.494), sigmoid resection (3.8 versus 2.3 per cent; P = 0.323), readmission (17.6 versus 12.0 per cent; P = 0.148), adverse events (48.5 versus 54.5 per cent; P = 0.221) and mortality (1.1 versus 0.4 per cent; P = 0.432)“].
Sigma-Resektionsrate und Mortalität ohne Antibiose erhöht!
Nicht nur die operative Interventionsrate lag mit der Sigma-Resektion in der Beobachtungsgruppe um 65 Prozent höher als in der Antibiotikagruppe [„3.8 versus 2.3 per cent“]. Die Mortalität lag in der Beobachtungsstrategie-Gruppe um 175 Prozent höher als in der Antibiotikabehandlungsstrategie-Gruppe [„1.1 versus 0.4 per cent“]. Diese Publikation kann mit ihrer apodiktischen Schlussfolgerung, dass man ohne Nachteile für unsere Patienten auf Antibiotika bei unkomplizierter Divertikulitis verzichten könne, [„Conclusion – Observational treatment without antibiotics did not prolong recovery and can be considered appropriate in patients with uncomplicated diverticulitis. Registration number: NCT01111253 (http://www.clinicaltrials.gov).“] so nicht stehen bleiben!
E-Mail an das Autorenteam
Dear Sir, I am very concerned about your publication “Randomized clinical trial of observational versus antibiotic treatment for a first episode of CT-proven uncomplicated acute diverticulitis” Published online 30 September 2016 in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.10309
The study population was “assigned randomly to an observational or antibiotic treatment strategy". It seems to be very doubtful that "hospital stay was significantly shorter in the observation group (2 versus 3 days; P = 0⋅006)" whereas the “readmission (17.6 versus 12.0 per cent; P = 0⋅148)" was 46 per cent higher in the observational group without antibiotic treatment. In all other criteria the observational treatment strategy never had been better than the antibiotic treatment strategy: "secondary endpoints: complicated diverticulitis (3.8 versus 2.6 per cent respectively; P = 0.377), ongoing diverticulitis (7.3 versus 4.1 per cent; P = 0.183), recurrent diverticulitis (3.4 versus 3.0 per cent; P = 0.494), sigmoid resection (3.8 versus 2.3 per cent; P = 0.323), readmission (17.6 versus 12.0 per cent; P = 0.148), adverse events (48.5 versus 54.5 per cent; P = 0.221) and mortality (1.1 versus 0.4 per cent; P = 0.432)"]. Due to your statistics the interventional sigmoid resection happened 65 per cent more often in the observational group "3.8 versus 2.3 per cent". The mortality rate under the observational treatment strategy was 175% higher than under the antibiotic treatment strategy "1.1 versus 0.4 per cent". What is wrong with your publication? Sincerely yours Dr. med. Thomas G. Schätzler, FAfAM Dortmund GERMANY
Die Antwort darauf
Dear Dr Schätzler, First of all, thank you for sharing your concerns. When reading the harshness of your blog, we fear that no valid arguments will ever convince you to embrace new insights in optimal clinical practice what contradicts what used to be common treatment. ‘Manipulierte Studie’ is a gross allegation which is very unprofessional. Nevertheless, we will address your concerns.
You are questioning the significantly shorter hospital stay in the observational group because the readmission rate was higher in this group. However, the hospital stay you are referring to is the length of stay of the initial admission and is therefore not associated with the readmission rate. In fact, in almost all research papers the reported length of stay concerns the initial admission. Also, this is explicitly stated in the full text version twice; “median duration of initial hospital stay was shorter owing to the intravenous administration of antibiotics in the antibiotic group (2 versus 3 days)” in the secondary outcomes section and “duration of initial admission” in table 3 of the results section. Moreover, our publication already provides an answer to your concern about hospital stay. The proportion of days spent outside the hospital during the first 6 months is significantly higher in the observational group. So even with the (non-significantly) higher readmission rate in the observational group, the total hospital stay during the first 6 months was still lower than in the antibiotic group.
Next, you’re questioning the results of almost all secondary outcomes, in fact without acknowledging the comparable results of the primary outcome measure for which this RCT was powered. We agree the non-significant differences you’re citing seem large at first, but when you (again) read the full text version of our publication you can see the number of events is small. Despite we acknowledge these differences may represent a trend, all these secondary outcomes are not significantly different.
Because our RCT was powered on the primary outcome measure, the size of this trial may not have been sufficient to detect differences in outcome measures with a low number of events. Or when increasing the number these events may also be comparable among groups. We can only draw conclusions based on the results we have.
When respecting the statistical rules in medical science, we did not find significant differences in those outcome measures and therefore we cannot conclude there is a difference. Trends do not count, as these comprise a scientifically unacceptable doubt. In our study we have proven that there is no difference in recovery between the observational and antibiotic group and we have not found proof of negative consequences of omitting antibiotics until 12 months of follow up. However, we already planned future research to deepen the small differences in secondary outcomes to see if there may be a subset of patients, if any, that benefits from antibiotics.
We would like to place a few remarks to consider. You’re criticizing non-significant differences in our study, however they are in line with the other RCT on this topic (the AVOD trial) which you presumably also have read. Moreover, there is no evidence at all for benefits of antibiotic treatment of uncomplicated diverticulitis. Thus far, multiple earlier studies and now 2 RCT’s (DIABOLO trial and AVOD trial) show no consequences of omitting antibiotics. And last but not least, people seem to forget that antibiotics are not innocent drugs. As you can see in our publication, 8.3% of the patients in the antibiotic group experienced antibiotic related adverse events which differ from a simple allergic reaction to anaphylactic shock. These numbers do not even include other discomfort patients experience such as abdominal pain and diarrhoea which they certainly do not need in addition to their diverticulitis symptoms. Also, looking at the bigger picture, we should take every opportunity to fight the rising threat of microbial resistance.
To put it more explicit in broader context, we should only treat patients with antibiotics when there is evidence in favour of antibiotics. At this point, there is no such evidence. In fact, there only is evidence in favour of omitting antibiotics.
We hope your questions about our paper are answered and you might consider implementing the results of these two randomized clinical trials in daily practice.
Sincerely, Professor Marja Boermeester (principal investigator), Lidewine Daniels (co-investigator), Stefan van Dijk, PhD Fellow, Department of Surgery, Academic Medical Center, Amsterdam, NL