Pittsburgh's Drowning Gap: Suspicious Deaths in the Three Rivers

The Pattern in the Water

The South Side neighborhood of Pittsburgh hosts one of the densest concentrations of bars in the northeastern United States. East Carson Street runs parallel to the Monongahela River, separated from the waterline by a few hundred feet of parking lots, railroad tracks, and scrub vegetation that thin out as you walk away from the commercial strip. On weekend nights, thousands of people in their twenties cycle through establishments that line both sides of the street for nearly a mile. The river is always close. Between 2008 and 2017, at least four young men left those bars and ended up dead in water, and the official record treats each death as though it occurred in isolation.

The geographic concentration is difficult to dismiss. Tommy Booth disappeared from a bar in Woodland, Pennsylvania in January 2008 and was found fourteen days later in Ridley Creek with zero decomposition, full rigor mortis, and lividity patterns consistent with dying on a flat surface rather than in water. Paul Kochu, a 22-year-old ICU nurse, left the South Side bars after an overnight shift in December 2014 and surfaced nude in the Ohio River near Wheeling, West Virginia, ninety miles downstream, more than three months later, with fractured ribs and a scalp wound the autopsy report could not explain. Dakota James, a 23-year-old Duquesne MBA student, walked out of Cupka’s Cafe 2 on East Carson Street in January 2017 and was recovered from the Ohio River forty days later with ligature marks on his neck and bilateral fingernail bed discoloration that forensic pathologist Cyril Wecht identified as consistent with strangulation.

What connects the cases is not geography alone but the institutional response. The Allegheny County Medical Examiner’s Office, under Dr. Karl Williams, classified each death as accidental drowning or undetermined, because drowning is a diagnosis of exclusion, meaning there is no definitive test that confirms it. When a body spends weeks in water, soft tissue evidence of strangulation, blunt force trauma, or injection marks degrades or vanishes entirely. The ME’s office applied the same default classification to each case, and that classification closed investigative avenues before detectives could pursue them. A 2025 Maryland audit, in which a panel of forensic pathologists unanimously reclassified 36 of 87 water-recovery deaths as homicides, proved that this pattern of misclassification is not theoretical but systemic.

The cases that follow examine each death on its own forensic terms, because the pattern only becomes visible when the individual evidence is laid out with the specificity that the official record withheld.