CDC recommends (MMWR Recomm Rep. 2005 Jul 29;54(RR-8):1-40) hospitalized patients with influenza A be placed under standard and droplet isolation precautions for 5 days after the onset of their symptoms. This is based on studies of volunteers who received live attenuated flu vaccine drops in their noses. After 7 days only 1 of 18 were shedding virus. One might wonder if attenuated flu vaccine in healthy volunteers is the best way to estimate the length of viral shedding. A new paper of viral shedding in hospitalized elderly patients at the Mayo Clinic suggests it isn't. Sensitive methods for detecting virus were used in 41 patients who met the criteria for the study and over half were found to be shedding virus beyond 7 days.
This finding raises the concern that hospitalized patients who are older and/or have chronic illnesses could shed influenza A virus beyond the 5- to 7-day period that has traditionally been considered to be the time of infectivity. A period of droplet isolation precautions limited to 5 days after symptom onset, as is currently recommended by the Centers for Disease Control and Prevention to control influenza in acute care settings,10 could be insufficient for such patients, and prolonging isolation measures for the entire duration of the hospital stay might be more prudent to prevent outbreaks in hospitals during the influenza season.[snip]
Some limitations of our study need to be recognized. First, our patient population was clearly a selected group of mostly older patients with chronic medical conditions. Sicker patients with prolonged hospital stays were more likely to be recruited into the study, compared with those who were dismissed sooner because of milder disease or fewer complications. However, we believe that the results of our study are applicable to similar patients hospitalized with influenza A. It is unclear if these results are generalizable to all adults with influenza, particularly younger, otherwise healthy adults treated in the outpatient setting. Second, regardless of the source of the initial diagnostic specimen, all follow-up specimens were throat swab samples. It has been suggested that the sensitivity of throat swab samples for influenza diagnosis might be low, and if some throat swab specimens gave false-negative results, we could have underestimated the duration of viral shedding in those cases. Third, we were able to detect viral shedding for longer durations by PCR, compared with the durations we could detect by culture. It is possible that PCR could amplify inactive viral RNA, but it is also possible that culture results could be false-negative because of a lower level of virus excretion or virus inactivation during transportation. Because we did not obtain specimens from other patients and healthcare personnel to monitor for transmission of infectious viruses, it is unclear whether the fact that influenza A could be detected by PCR meant that the patient was infective. Finally, our determination of the total duration of viral shedding and our comparisons of hospitalized patients with and without prolonged viral shedding are subject to bias. Patients were not followed up after discharge from the hospital, and the final duration of viral shedding was unknown for several patients; in many cases, a single negative sample was used to identify the end of viral shedding, which could lead to underestimation of duration if those samples had false-negative results. Small numbers of subjects also precluded adjustment for potential confounders, such as age or comorbid conditions. Future studies should attempt to follow up all patients until the end of viral shedding to minimize bias. Surveillance of contacts should also be attempted to help assess whether prolonged shedding is associated with continued infectivity. (Leekha et al., Infect Control Hosp Epidemiol 2007;28:000)
I have included the long list of limitations and caveats to emphasize the difficulty in studying this problem. Things that seem easy to study are not. After reading it you will have to decide for yourself what the implications are for homecare of family.
Two other points in this paper deserve some comment. Neither a history of vaccination nor the use of antivirals were said to make a difference in whether a person was a "prolonged shedder" or not. The vaccination results might be interpreted as a decline in immune response previously observed amongst the elderly. The results on antivirals are harder to interpret than for vaccination. Which antiviral was used (M2 or neuraminidase inhibitor) was not specified. Examination of Table 4 shows evidence of an effect on shedding beyond 7 days and decreased length of shedding but the differences didn't reach the point where a chance difference could be considered unlikely (i.e., the results weren't "statistically significant"). Lack of statistical significance does not mean the results are due to chance. It only means chance could not be ruled out. The data in the paper suggest an effect on shedding of antiviral therapy, which would be consistent with a number of other papers in the published literature.
Once again, something we thought we knew about influenza must be modified. While this is an "merely" adjustment in a quantitative estimate, it is of some practical significance.
- Log in to post comments
The obvious, careful detail in this long list of "limitations and caveats" should be the norm...
...but instead is a 'refreshing change'.
Even if only one person in 18 (for healthy people) is shedding the virus, consider that in a pandemic involving millions of people, the number of apparently healthy people shedding viruses after one week could also be in the millions. (Case in point: Our local state university occasionally has a single case of bacterial meningitis, in an otherwise healthy population. I appreciate the immediate subsequent action taken by university officials to prevent other cases.) Whether the number is "only" 1 in 18, or whether it is actually higher, offical pandemic plans must include adequate boundaries to prevent infecting others. So what if someone spends another three-eight days in isolation, if it prevents 100% of subsequent infections to others?
I really wonder though about what will happen if its H5N1 HPAI. Historically speaking you are pretty much a dead man/woman after those 3-5 days or recovering.7 days in a HPAI environment and maybe longer you would be shedding? Costs of maintaining one person will be off the scale and I dont see how we can even supply for 1000 cases that last in a rolling number for more than 30 days if we isolate primary and secondary suspected cases such as family.
Is this that reverse isolation thing that I have read about where people who are recovering are staged out as they were in? How would we test for it? Run ferrets, cats and dogs into rooms with them? PCR's take a week to get the results back I think, plus who can do them and how many are there? Think about the logistics of moving hundreds of thousands into warehouses that might have military cots or regular beds in them. We might not even have the fuel to do it. How to feed, treat that many people who are actually sick, recovering or that might be shedding?
I think Randolph's thinking is along the right lines. You might need to progressively lower the degree of isolation as time progresses. If the amount being shed is decreasing, then the probability of spread per patient per day goes down as recovery time increases, and you may have only a limited amount of high isolation units available. So you may end up with highly infectious patients in tier1 isolation, and moderately infectious in tier2. Not ideal, but given likely cost/resource constraints it may be the best you can do.
This study is valuable by specifically delineating the time elderly ill enough with seasonal influenza to warrant hospitalization remain contagious after contracting this illness.
It is well established that normal children, the elderly and those with immunocompromised states remain contagious after seasonal flu for significantly longer than healthy adults with the same infection. Elderly people often suffer from immune senescence, an immunocompromised state, which is the probably reason for their prolonged contagiousness.
We have no idea how long healthy adults and children, the elderly and the immunocompromised will remain contagious after recovery from the acute symptoms of infection with pandemic influenza have passed. In my opinion, it is unwise to extrapolate findings like those in the article under discussion to what will happen with pandemic flu.
Grattan Woodson, MD
Gratt: Agreed. But you use what information you have and correct what you have if warranted.
what is the current proportion of those over age 65 and under age 15 among the entire US population? 25%? More?
Whether or not these results may be properly extrapolated to healthy adults is an open question, but the increase in infectiousness of the elderly and the young could be a significant factor in managing the impact of an influenza pandemic both on the macro and the micro scale.