Thread by @DrZoeHyde

(1/22) Study of household transmission of #SARSCoV2 in the US in March to April. Extremely thorough investigation by the CDC showing an equal attack rate for children and adults, among other things. Lots to discuss here... #COVID19
doi.org/10.1093/cid/ci…

(2/22) The investigation took place in Utah and Wisconsin. These states were chosen because of the low #COVID19 prevalence at the time, reducing the risk of additional exposure to household members.

(3/22) Index patients were identified via outpatient testing. After this, a list of households to which index patients belonged was established.

(4/22) A total of 58 households were investigated, each with one index patient. There were 188 household contacts.

(5/22) Households were visited at least twice, separated by 14 days. Both swab testing and blood (antibody) testing of all household members was performed at each visit.

(6/22) Each household member completed a symptom diary. If a person developed symptoms, an extra visit was scheduled, during which all household members were tested.

(7/22) In each household, a primary case was identified. The primary case was the first person to show symptoms. This was not always the same as the index case which had been picked up initially.

(8/22) Overall, the secondary attack rate (SAR) was 28%.

Notably, the use of both RT-PCR and antibody testing picked up an extra 8 cases. Using RT-PCR data alone, the SAR would have been 23%.

(9/22) Children (aged <18 years) of the primary case had the highest SAR (42%), followed by adult children (>=18 years) of the primary case (35%), and partners of the primary case (33%).

(10/22) For children aged under 10 years, the secondary attack rate was:

18% when they were a child of the primary case;
0% when they were not a child of the primary case.

(11/22) For children aged 10-17 years, the secondary attack rate was:

58% when they were a child of the primary case;
8% when they were not a child of the primary case.

(12/22) This could suggest that the risk to children in the household is largely dependent on whether the person caring for them is infected.

(13/22) That is, if a parent has the virus, children are much more likely to be subsequently infected than if the virus is introduced to the household by a different family member.

(14/22) This may be one reason why the results of household contact studies differ with regard to the risk posed to children (some saying equal risk, and some saying less).

(15/22) Similarly, the attack rate was particularly high in spouses/partners of the primary case. This probably reflects increased proximity to cases and increased duration of contact.

(16/22) Overall, the secondary attack rate in children and adults was essentially the same:

19/68 children = 27.9%
33/120 adults = 27.5%

(17/22) However, younger children were less likely to be secondary cases than older children:

3/29 children aged under 10 years = 10.3%
16/39 children aged 10-17 years = 41.0%

(18/22) Household contacts with diabetes were more likely to be secondary cases than contacts without diabetes (SAR: 80% vs. 26%).

(19/22) Household contacts of male primary patients were more likely to be secondary cases than those of female primary patients (36% vs. 18%).

(20/22) Household contacts of primary patients who were immunocompromised were more likely to be secondary cases (88% vs. 25%), possibly due to increased or prolonged viral shedding by the primary case.

(21/22) In conclusion, this study suggests that children can be at a similar risk of being infected in the household compared to adults, but this may depend on who introduces the virus (with parents seemingly conferring greatest risk in this study).

(22/22) It also suggests that men may be more likely to transmit the virus to others than women in a household setting, and that people with diabetes are at increased risk of infection (not just getting more sick if they do become infected).