First, why might this be important to we lung doctors?
In house, one of the three patients that we admitted on MICU Blue in the past 10 days had mild transient diarrhea prior to admission but none since, and is out of the MICU.
 
Two real time cases from today’s phone calls:

1. Brigham/Harvard Pulmonary alum, now a senior clinician-teacher at a major academic medical center in NYC, called me to report presenting to his ED 3 days ago with nausea, vomiting, and diarrhea, COVID19 tested negative, received 3 liters of IVF, no respiratory symptoms though CXR had a streak of atelectasis in LLL, stool+ for enterovirus, sent home, recovering. Three of his patients have been diagnosed and hospitalized with COVID19 in the past 5 days. His division’s schedules have already completely reorganized to handle the current overload, and he is scheduled for duty this week. Living alone, will he progress suddenly? Re-test before he goes back to work?
2. 50 yo patient of mine with bronchiectasis, called last week with a typical flare, given doxy for MSSA. She called yesterday because her 18 yo daughter home for a week from college developed fever, nausea, vomiting, diarrhea, aches, malaise but no cough or dyspnea. I sent her daughter to NWH for COVID/ Flu testing – now pending. This morning, my patient now has chest tightness and a sense of a “new cold”, while her daughter continues with vomiting, diarrhea, and low grade fevers, COVID19 result “will return in 3 days”.

Let's turn to a preprint article in the AJG...
Below is a synopsis from preprint of “Clinical characteristics of COVID-19 patients with digestive symptoms in Hubei, China: a descriptive, cross-sectional, multicenter study.” Pan L Mu M et al, Am J Gastroenterol 2020 (not yet in Pubmed, but on AJG website).

Background  SARS-CoV2 is known to be to be capable of invading the GI tract, and in one small study, virus was detected in stool specimens of 50% of infected patients, but not much has been reported about GI symptoms as part of the COVID-19 syndrome.

1. Original cohort according to the text was n = 300 (vs 310 according to figure 1), and 96 were excluded due to lack of CT or absence of PCR confirmation or lack of complete lab data, leaving the reported cohort of 204 patients.
2. Severity of illness in the overall cohort – hard to be certain because of overlapping terminology especially the use of the term critical. 36% or 74 of the 204 were described as “critically ill”, with 16 or 8% transferred to ICU.
3. GI symptoms in 99, 48% overall in the cohort of 204
   1. No difference in demographics (mean age 55 yr, 55% male, 5% smoking) contact history, or underlying diseases compared to non-GI
   2. Anorexia was considered as a GI symptom and found in 83% of the GI subgroup.
   3. Diarrhea 29% of GI subgroup, Vomiting in 8%, Abdominal pain in 4%
   4. Most of us would consider anorexia not a GI symptom per se, but a common, systemic symptom of illness especially an infection. Per authors, if anorexia were not considered to be a GI symptom, 20% of entire cohort of 204 had GI symptoms of diarrhea, vomiting or abdominal pain, but no separate analysis was performed on this subgroup. Diarrhea was most commonly non-dehydrating loose stools and usually not high volume.
   5. Fever in 94%; Concomitant Respiratory symptoms in 93%  (81% in non-GI)
   6. Pure GI without respiratory symptoms in 7%
   7. Presented later: 9.0 days vs 7.3 days of illness in non-GI
   8. No difference in CBC/diff, LFTs, coags, d-dimer, CK or other labs except PT 14.1 +/- 5.0 vs 12.5 +/- 2.3
   9. Prognosis per authors was worse in GI subgroup, but the numbers from Table 1 don’t bear this out. Mortality was 18% overall, 19.2% in GI subgroup vs 16.2% in non-GI, p = 0.57. There was also a trend for less ICU transfer need, 5.1% in GI subgroup compared to 10.5% in non-GI. Table 1 indicates 22 of 99 were “critical” but that was not defined.
   10. The authors’ basis for increased severity in the GI subgroup seemed to be that they were less likely to be discharged by end of study period – 46% of GI subgroup were still in house, vs 24% of non-GI patients. 

Conclusions – in this early, small series of patients in Hubei province who were ill enough to be hospitalized and limited to those with PCR-confirmed infection and complete lab/chest CT data, 48% had GI symptoms, mainly anorexia.
In terms of prominent GI symptoms, 20% of the overall cohort had diarrhea, vomiting, or abdominal pain. The presence of GI symptoms (when anorexia was included) was correlated with a delayed presentation, possibly more frequent respiratory symptoms, and with a prolonged hospitalization. GI symptoms did not correlate with LFTs or any lab indices of inflammation or severity including creatinine, except a mildly elevated PT, nor with ICU transfer or increased risk of death.
 
How to apply this preliminary data to these outpatients?
The pulmonary/critical care faculty member could certainly have GI-predominant COVID-19 in addition to enterovirus, with implications for his own health if he develops respiratory disease consistent with his chest Xray, and for his division’s staffing.
The mother with bronchiectasis has a reasonable chance of exposure to COVID-19 and incipient respiratory illness, or her daughter may have norovirus or another enterovirus.

I would also recommend an excellent podcast featuring one of Miranda’s favorite PennMed professors, Paul Offit, an ID physician and co-developer of the rotavirus vaccine, who gives a fascinating, if contrarian, perspective on SARS-CoV2 as potentially being fecal-oral spread: https://zdoggmd.com/podcasting/paul-offit-covid/.

1. Selection Bias: In the US and Italy, due to limited resources, hospitalized patients and those with severe symptoms are preferentially tested. As a result, the confirmed population (denominator) reflects the sickest subset of the population rather than the true prevalence of the disease. As recognition and testing of COVID-19 becomes widespread, the crude fatality rate is expected decrease. In fact, in China the initial CFR decreased from 17.3% in January to 0.7% in February as tests became more accessible.
2. Population Specific Factors: There are unique features of the populations affected in both Italy and Washington. Italy has the second oldest population in the world with 23% of people older than 65 years (compared to 15% in the US). In fact, people ≥70 years old account for 37.6% of cases in Italy but only 11.9% of cases in China.⁶ This may account for higher CFR given the association between older age and COVID-19 mortality. Similarly, in Washington, COVID-19 differentially impacted an elderly population given an early outbreak in a local nursing home, the Life Care Center, resulting in at least 37 deaths. While this highlights the risk to older patients with co-morbidities, it is difficult to extrapolate these results to the general population.
3. Definitions: For the data reported from Italy, case fatalities included anyone diagnosed with COVID-19 by RT-PCR who died. However, dying with COVID-19 (association) is not equivalent to dying from COVID-19 (causation). Although yet to be published in a peer-reviewed journal, Prof Walter Riccardi, the Scientific Advisor to the Italian Minister of Health, reported on March 20, 2020 that on reevaluation, “only 12% of death certificates have shown direct causality.” Furthermore, in a retrospective review of 355 deceased Italian patients with RT-PCR proven COVID-19 infection, only 3 (0.2%) had no comorbidities or identifiable alternative causes of death (Onder 2020).