Last weekend's frightening and widely reported string of overdoses in Philadelphia — nine deaths in 36 hours, according to police — was just part of what officials suspect was a devastating five days that left 35 people dead.
It started Dec. 1, when 12 people died between 6 a.m. and 10 p.m. "We have never seen that before," said Sam P. Gulino, the city's chief medical examiner.
Then came four more deaths last Friday, seven on Saturday, nine on Sunday, and three on Monday. The total could still rise, as deaths that initially appeared natural are investigated for drug links.
The overdoses may have resulted from heroin that was cut with more powerful synthetic opioids such as fentanyl. But police said there were other possibilities, such as the appearance of heroin that was more pure than the users were accustomed to.
"It's a lengthy process of figuring it out," said Lt. John Stanford, a police spokesman. It was clear from narcotics packaging found at the scene that the users who died did not all buy their drugs from the same location or person, he said.
Many of the deaths occurred in and around Kensington, where people come from elsewhere in the city and the suburbs to buy and use heroin. Bucks County authorities said there had been five suspected overdose deaths this week in Falls Township, although it was not clear if any were linked to Philadelphia. Other suburban counties said there had been no discernible uptick in fatalities recently.
Surges in drug overdoses, many related to fentanyl, have been reported around the country over the last year. Seven people died on one September day in Cleveland, according to news reports. Three were found within five hours in separate parts of York County, Pa., on Nov. 30. In Washington County, south of Pittsburgh, with a population of 208,000, eight overdoses were recorded in 70 minutes in August 2015, contributing to a total of 25 over two days. Three of the overdoses were fatal.
But Gulino and others said they could not remember a cluster as extreme as Philadelphia's 35 deaths anywhere in the country.
The cases are considered "suspected" drug deaths until toxicology tests determine what drugs were present in each body, a process that typically takes about eight weeks. Results are not needed to determine overdose as the likely cause of death, however.
" 'Suspected fatal drug deaths' means that the circumstances, scene investigations, and autopsy findings strongly indicate that the death was a drug-related death," said Jeff Moran, a spokesman for the Philadelphia Department of Public Health, which includes the Medical Examiner's Office.
Gulino said that the number of fatalities over that November period was not significantly higher than usual.
The federal Drug Enforcement Administration has an active criminal investigation into the source of the November overdoses, said Patrick J. Trainor, a spokesman for the federal agency's Philadelphia division. Based on the numbers, he said, the latest spate of deaths would be "of very great interest, no question," to DEA investigators.
The Philadelphia Medical Examiner's Office confirmed 702 drug-related deaths last year, a number that has been incorrectly reported as 720 based on a DEA report over the summer. The city is on track to end this year with 25 percent to 30 percent more deaths than in 2015, Gulino said, "almost entirely on the back of heroin and fentanyl."
Anecdotal reports from county coroners around Pennsylvania and New Jersey suggest significant increases are likely this year statewide, on top of jumps of roughly 22 percent in both states last year. Deaths involving fentanyl nearly doubled in Pennsylvania and tripled in New Jersey in 2015. National data lag more than a year behind state reports.
Some chemical variations of fentanyl that are not necessarily more dangerous than the original have appeared in Philadelphia. But Gulino is worried about the possible emergence here of another fentanyl relative, known as carfentanil, that has received attention elsewhere.
With a potency that has been reported up to 10,000 times greater than morphine, deadly amounts of carfentanil might not even be detectable in toxicology testing, Gulino said.
Despite the latest surge, he said, he is confident carfentanil has not arrived in Philadelphia because "were it to show up. it would cause a large number of fatalities. and it wouldn't be over in a few days."
When dangerous drugs are detected, authorities struggle with how specific to get in public warnings, because announcing that stronger drugs are on the street can attract more users and increase the death count.
"When people are addicted and they are seeking that high and they hear about a drug that is leading to overdoses ... they try to acquire that drug for the ultimate high. So warning the public is a double-edged sword," said Jeanine M. Buchanich, who studies overdose mortality trends at the University of Pittsburgh's Graduate School of Public Health.
Buchanich is used to seeing disturbing statistics from around the country. But nothing as high as Philadelphia's 35 deaths in five days, which she heard from a reporter.