Sunday, 4 August 2013

MERS-CoV: Family cluster, untested case, delayed reporting, missing data

Hat tip to Crof's story for drawing my attention to this article.

Omrani and colleagues describe a family cluster of Middle East Respiratory Coronavirus (MERS-CoV) infections. The study, from the Kingdom of Saudi Arabia Ministry of Health (KSA MOH) is interesting for a few reasons:
  1. The first case (51-year old male with type II diabetes; 51M) is believed to have acquired the infection while he was in hospital for investigations into 2 months of back pain, limb weakness and lack of bladder control which lead to diagnosis of multiple myeloma. 14-days after admission to a medical ward he developed fever, cough, shortness of breath (dyspnoea)and hypoxia for which he was treated with antibiotics and oseltamivir (for influenza). He died.
    • Because 14.7 days is the suspected latest time between infection and symptoms of infection, the authors believe MERS-CoV was acquired while in the hospital but there is no clear source. The authors note that it could have been another patient or an asymptomatic healthcare worker (HCW). he hospital is not identified.
    • There were a number of clinical features shared by previous cases that were known to be MERS-CoV positive: fever and cough are found in 91% of 33 previously published MERS-CoV cases; acute renal failure and requirement for mechanical breathing assistance in 79%; chest X-ray infiltrates in 76%; dyspnoea in 60% etc.
    • Numerous friends  relatives and a hospital roommate did not develop symptoms. Two of his brothers, from a household of 10 adults, did.
    • No respiratory virus testing of samples from 51M (Patient #1) was undertaken. It's not clear whether any appropriate samples were collected. 
  2. Patient 2, (39M; brother of 51M) was admitted to a different hospital on Feb 28th, 10-days after becoming unwell. He had developed pneumonia. He died March 2nd.
  3. Patient 3 (40M; brother of 51M and 39M) was admitted to the same hospital as 39M, March 4th. He had developed pneumonia. He had been previously announced by the WHO but without any detail (he seems to be FluTracker's case #16). He developed pneumonia but recovered and was released.
  4. Nasopharyngeal swabs from 51M and 39M were negative for influenzavirus, respiratory syncytial virus, adenovirus, rhinovirus and endemic coronaviruses. No parainfluenzavirus, metapneumovirus or bocavirus testing was mentioned. 39M was positive for MERS-CoV in upper and lower respiratory tract samples; 40M was negative from an upper airway sample but positive from a lower respiratory tract sample.
  5. The authors note that the high mortality rate appears to be die to under-diagnosis of mild and asymptomatic cases. Presumably, such as the one(s) that lead to infection of 51M.
Its nice to have another data-hole filled (40M). Shame its taken since late March when this case was reported by the WHO, to find out about this case and about the possibility of silent in-hospital transmission events that could explain the strangely disseminated appearances of cases across the KSA, often associated with hospitals. 

Still, this cluster doesn't explain why so few exposures resulted in disease - and why the 2 contacts of 51M, neither having underlying diseases, became ill even though other family members and HCWs had longer or more intimate exposures, some quite early on when viral loads may have been at their highest. Some piece(s) of the puzzle is still missing here. 

  • Were more actually infected, but not tested because they were not suitably symptomatic? 
  • Could there be underlying levels of cross-protective pre-existing antibodies in the KSA? 
  • All 3 cases here were males. Is there a risk factor specific to males in this region?
  • Should 51Ms farm be considered more closely? Have those animals been tested? 
  • 39M like to climb palm trees to pollinate the dates - could he have been the real index case of the cluster, but 51M,  infected by him, showed signs more rapidly because of his underlying disease?

I'm not a medical doctor and I don't pretend to be but even I am unclear on why a patient with clear signs and symptoms of a respiratory disease, which was in part treated as a respiratory infection (influenza antivirals were given), was not tested for viruses or bacteria to understand what the patient had. 

Further, this cluster started off in February of this year in a country that is on alert for, and living with, MERS-CoV infections. The World Health Organisation noted in its novel CoV update in November of 2012... 
"Until more information is available, it is prudent to consider that the virus is likely more widely distributed than just the two countries which have identified cases. Member States should consider testing of patients with unexplained pneumonias for the new coronavirus even in the absence of travel or other associations with the two affected countries. In addition, any clusters of SARI or SARI in health care workers should be thoroughly investigated regardless of where in the world they occur."
As Omrani et al show, 51M fit the requirement for being a MERS-CoV case very well; an unknown pneumonia. And yet the patient wasn't tested for any virus at all. And the other 2 cases relied on the shipment of samples to the UK (Public Health England, UK Laboratories) for testing before being confirmed. If this level of pneumonia labwork is the norm, then we have very little understanding of the amount of virus-related pneumonias in the KSA. How do we know whether the MERS-CoV is an exceptional disease-causing agent, or just one of many viruses that might be involved in such disease? That very basic information is essential for a world trying to understand how seriously to take this novel virus. 

This publication speaks to a central problem in trying to understand MERS-CoV in its current site of greatest activity, the KSA- too little testing and too heavy a reliance on signs and symptoms. If 51M got his infection from an asymptomatic HCW or another patient, that happened because testing was not conducted on that/those case(s). Of course, we may yet be wrong about the incubation period given that testing of asymptomatics and non-contacts is so limited - perhaps 51M did bring his infection in with him.

If ever there was an example of the need to test prospectively, regardless of signs and symptoms (something I've harped on about before), this paper gives it more than enough credibility.

Test, test, test. You can't test too much.