Second COVID Infections Linked to Sharply Higher Odds of New Diagnoses, Canadian Data Show

Catching COVID-19 more than once carries risk beyond the infection itself, a pattern public health researchers have tracked across several countries since 2022. A new analysis from the Public Health Agency of Canada and Statistics Canada, published this week, puts numbers on that pattern using Canadian survey data. Adults who reported two or more infections had sharply higher odds of a new back-problem diagnosis than adults who reported none. The size of that jump is the story.

How the study was built

The report ran in the Canada Communicable Disease Report, the agency's peer-reviewed surveillance journal, in its September 2026 issue. Lead author Dianne Zakaria and four coauthors analyzed 8,717 adults with complete survey data. All five work at the agency's Centre for Surveillance and Applied Research. Participants filled out a survey and a follow-up questionnaire between January 2020 and June 2023, the Canadian COVID-19 Antibody and Health Survey. They then reported how many confirmed or suspected infections they'd had, and whether a doctor diagnosed them afterward with any of 21 new conditions.

The researchers adjusted for sex, age, education, income, obesity, smoking status, preexisting conditions, and time at risk using multivariable logistic regression. That statistical step is meant to isolate the effect of reinfection itself, apart from other factors that independently raise disease risk.

What the numbers actually show

Adults with two or more infections had 76 percent higher odds of any new diagnosis compared with adults who reported none. The adjusted odds ratio was 1.76, with a 95 percent confidence interval of 1.12 to 2.76. Respiratory diagnoses climbed further, by a factor of 5.03. Back problems came in third among the headline findings, at 3.79 times the odds.

The back-problem finding is also the least statistically precise of the three.

Its 95 percent confidence interval ran from 1.12 to 12.89. That means the true effect could be as small as a 12 percent increase, or as large as nearly thirteenfold. The respiratory finding's interval, by contrast, started at 1.72, so even its lower bound pointed to a clear increase. Zakaria and her coauthors flagged the wide range themselves rather than let the headline number stand alone.

“...reliance on participation and self-reporting, which make it susceptible to survivor, selection, misclassification, recall and surveillance biases.”

What could cut the other way

The authors are explicit that the study shows an association, not a proven cause. Every infection count and diagnosis came from self-report, not medical records. That opens the door to recall bias, and to people who felt sicker seeking more care, regardless of what actually caused the condition.

The paper also names a specific alternative explanation for the symptom findings: reverse causation. A new health condition can prompt more medical visits and, indirectly, more opportunities for diagnosis, rather than infection driving the condition. The authors say that possibility cannot be ruled out with a single cross-sectional survey.

How common is back pain to begin with

Back pain is already common in the general population, independent of any COVID history. A CDC survey of American adults found 39 percent reported back pain in the prior three months in 2019, before the pandemic began. That is a different country, a different measure, pain versus a new diagnosis, and a different time period. None of that confirms or contradicts the Canadian findings, but it does show how high the baseline already is. That is exactly why this study leaned on statistical adjustment, and on caution about wide confidence intervals.