Intended for healthcare professionals

Rapid response to:

Practice Easily Missed?

Measles in older children and adults

BMJ 2017; 356 doi: (Published 16 February 2017) Cite this as: BMJ 2017;356:j426

Rapid Response:

Oral versus parenteral administration of antibiotics.

I thank Dr Anand for his comment to my rapid response of 5 March 2017, raising very welcome relevant questions about the deleterious effects of the way antibiotics are administered and reported on. I soon realised that even papers that have antibiotics’ administration in their title may not disclose the mode of administration, and many even fail to disclose what type of antibiotic was administered. So the mode of administration is rarely correlated with the type of observed reactions, including untoward reactions. The reader is left to assume, possibly falsely, that the information applies to all antibiotics and to any mode of administration.

As a general rule, small babies in hospitals are administered antibiotics parenterally because oral administration would be extremely time consuming, difficult and even dangerous. The same applies to adults who are very sick, or unconscious, or immobile, or unable to swallow, or susceptible to vomiting and/or other problems, such as diarrhoea, which would create its own unpleasant practical problems.

As a general rule intravenous antibiotics (in hospital) are later switched to oral administration, again no doubt for practical reasons especially in domestic environment, and very likely also because the patient is mobile and can swallow easily.

I provide, as follows, some of the rare examples of research in which the mode of antibiotics administration is specified, and the effect of the specified administration is described or apparent.

Bolton and Culshaw (1986) wrote, ”Faecal metronidazole and hydroxymetronidazole concentrations measured by high pressure liquid chromatography are reported during 10 episodes of Clostridium difficile colitis in nine patients. [treatment was oral in seven and intravenous in three.] Bacterial faecal concentrations were present in all patients with acute disease receiving oral or intravenous metronidazole, and all responded to therapy. Metronidazole and hydroxymetronidazole concentrations fell as the diarrhoea improved and neither substance [presumably the administered antibiotics] was detectable in the faeces of five [out of the total 9] patients after recovery. This demonstration of intracolonic therapeutic concentrations of metronidazole supports the clinical experience of oral metronidazole being effective in the treatment of antibiotic associated diarrhoea caused by C difficile and also suggests a potential role for intravenous metronidazole in this disease.”

Even though IV administration may not directly adversely affect the bowels and its biota, it affects other internal organs, such as the heart.

Williams et al. (1998) were quoted by Scheibner (2008), “None of the [four] babies were very ill on admission to respective hospitals. The highest temperature was 37.6 C and none suffered paroxysmal cough. They were doing well until they were administered intravenous antibiotics cefotaxime, erythromycin and/or ceftriaxone. Based on the internal consistency of timing, as outlined below, all quite obviously started deteriorating, and died, after the administration of the above antibiotics.”

These four babies’ histories were as follows:

“Baby 1 was admitted to hospital after 48 hours of lethargy, poor feeding, tachypnoea and cough. He remained stable for five days, with satisfactory breast feeding and occasional coughing. On day 3, he was administered intravenous cefotaxime, and, erythromycin on day 5, after which time he immediately developed tachycardia and hypercapnia and required supplemental oxygen, he also developed ‘severe pulmonary hypertension and cardiovascular compromise’, unresponsive to inhaled oxide and died 72 hours after admission to another hospital.

Baby 2 developed progressive tachypnoea with respiratory distress, tachycardia, and hypercapnia unresponsive to different ventilatory regimens 18 hours after being given cefotaxime (from day 1 of the admission) and IV erythromycin (from day 2). Circulatory compromise developed and was not ameliorated with infusion of adrenaline or inhalation of nitric oxide given to treat presumed pulmonary hypertension. Systemic hypotension and severe metabolic acidosis developed, and the infant died following an asystolic arrest 48 hours after initial presentation.

Baby 3 developed intensified respiratory distress ten hours after admission and intravenous cefotaxime administration, with associated hypercapnia and high white blood cell count. Perfusion deteriorated gradually, and the baby died 25 hours after admission.

Baby 4. Within hours of admission to hospital and administration of ceftriaxone, the baby’s respiratory distress worsened and he developed poor perfusion requiring artificial ventilation, together with substantial colloid and ionotropic support of the circulation. On transfer to another hospital, the x-ray showed bilateral consolidation. Echocardiography confirmed severe refractory pulmonary hypertension, and severe metabolic acidosis which culminated in cardiac arrest six hours later.”

The above histories are among the best accounts of the sequence of events enabling the discerning reader to see the connection between the administered IV antibiotics and the observed cardiac deaths.

Some examples of the dangers of IV administration are in Ray et al. (2004) who demonstrated that oral erythromycin prolongs cardiac repolarisation and is associated with reports of torsades de pointes [polymorphic ventricular tachycardia in the context of QT prolongation] and the death rate among users of erythromycin was twice as high as that among non users of any of the study antibiotic medication. IV use increased the rate of ventricular arrhythmias implicated in sudden cardiac deaths, especially when erythromycin is used concurrently with CYP3A inhibitors.

In contrast, Ray et al. (2012) evaluated the use of azithromycin and clarithromycin and the risk of cardiovascular deaths which are often rapidly fatal. Azithromycin, a broad spectrum macrolide antibiotic, has been reported as relatively free of cardiotoxic effects. “However, the closely related drugs erythromycin and clarithromycin can increase the risk of serious ventricular arrhythmias and are associated with increase risk of sudden cardiac deaths.” The mode of administration was not mentioned, but the words ‘used’, ‘course’ and ‘prescriptions” presumably mean oral use.

In summary, both orally (PO) and IV administered antibiotics are not free of serious adverse events.

I would not consider any IV antibiotics completely safe or less reactogenic than those administered PO.

I stopped using any antibiotics in 1985 when I started using sufficiently large daily doses of vitamin C; I haven’t experienced tonsillitis or even a runny nose since. I also noticed that I don’t bruise and occasional abrasions and kitchen cuts heal very quickly. Most spectacular event was being bitten by a nasty spider on my left eyelid in my garden: a momentary sharp pain and then nothing. No swelling, not even a red spot, no delayed reaction.


Bolton and Culshaw. 1986. Faecal metronidazole concentrations during oral and intravenous therapy for antibiotic associated colitis due to Clostridium difficile. Gut; 27: 1169-1172.

Williams et al. 1998. Infant pertussis deaths in New South Wales 1996-1997. Med J Aust; 168:281-283.

Scheibner 2008. What is really the cause of deaths in newborn infants who usually contracted pertussis from their vaccinated family members suffering whooping cough at the time of these babies death ( rapid response; 9 December 2008).

Ray et al. 2004. Oral erythromycin and the risk of sudden death from cardiac causes. NEJM;351:1089-1096.

Ray et al. 2012. Azithromycin and the risk of cardiovascular death. NEJM; 366: 1881-1890.

Competing interests: No competing interests

09 March 2017
Dr Viera Scheibner (PhD)
scientist/author retired
Blackheath NSW Australia