• forensic science;
  • pregnant women;
  • substance abuse;
  • opioid;
  • prescription drugs/poisoning;
  • overdose/mortality


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References

Abuse of prescription and illicit drugs has been rapidly increasing. This study examines the prevalence of drug use in the non-natural deaths of pregnant or recently pregnant women. Records from Florida's Pregnancy Associated Mortality Review conducted between 1999 and 2005 (n = 415) were linked to 385 toxicology reports obtained from Florida medical examiners' offices. The final study sample consisted of 169 drug-positive, pregnancy-associated non-natural deaths. Of these, 86 were positive for both blood and urine, 64 were positive for blood only and five for urine only, and the remainder were positive for some other specimen. Among these deaths, 91 cases (54%) involved prescription drugs, 78 cases (46%) involved illicit drugs, and 69 cases (41%) involved alcohol. Opioids constituted the majority of deaths associated with prescription drugs. Substantial co-use of opioids and benzodiazepines was seen. Pregnant or recently pregnant women may have more interactions with healthcare providers, which may present more opportunities for intervention and prevention.

With the exception of alcohol and marijuana, prescription drugs are now the most commonly abused psychotropic agents in the United States [1]. Opioid painkillers have been identified as a major driver of the observed increase in prescription drug-related mortality [2-9]. In one study, opioids were encountered in 93.2% of pharmaceutical drug overdose fatalities [4]. In at least one state, the majority of drug overdose-related deaths were attributable to the abuse of prescription opioids [10]. Unlike the illicit drug epidemics of the 1970s that were seen as concentrated in urban centers, the rate of increase in drug poisoning mortality in rural areas is three times greater than what has been observed in metropolitan areas [5]. Prescription drug abuse also has significant associated morbidity that has also seen a sharp increase recently [11]. Between 2004 and 2009, the number of emergency department visits related to misuse or abuse of prescription drugs surpassed the number of visits related to illicit drug use, with the largest increases in drug abuse or misuse coming from opioids and anxiolytics [12]. Overdose deaths have important public health implications: for every one overdose death, there are nine abuse treatment admissions, 35 emergency room visits for misuse or abuse, 161 people with abuse/dependence, and 461 nonmedical users [13]. A pattern of multiple drug abuse has been noted among prescription drug abusers, with opioid abusers being more likely to abuse other medications such as benzodiazepines [4, 6].

In Florida, similar trends are being observed [14-16]. From 2003 to 2009, the annual number of deaths in Florida involving lethal drug concentrations increased 61.0%, while the death rate increased from 10.6 to 15.7 per 100,000 individuals—a 47.5% increase [14]. This observed trend is almost entirely attributable to prescription drugs. During the same period, the prescription drug overdose death rate increased by 84.2%, while the illicit drug over dose death rate decreased by 21.4% [14]. Other studies from Florida support this shift from illicit to prescription drugs as a significant source of mortality [15]. By 2009, the deaths involving prescription drugs were four times as common as deaths involving illicit drugs such as heroin and cocaine [14] (See Fig. 1).


Figure 1. Annual drug overdose death rates for selected prescription and illicit drugs - Florida, 2003–2009.

Source: Centers for Disease Control and Prevention (CDC) [14].

Download figure to PowerPoint

Inappropriate prescribing practices at pain clinics—commonly referred to as pill mills—are contributing to the prescription drug abuse epidemic in Florida [15-22]. The increases observed in the overdose death rates of the commonly abused opioid oxycodone and the benzodiazepine alprazolam—264.6% and 233.8% between 2003 and 2009, respectively—seem to indicate that pill mills are a contributing factor to the situation in Florida [14]. Until recently, the state lacked both a prescription drug monitoring program and legislation against inappropriate prescribing practices [22, 23]. According to data supplied to Florida's Department of Children and Families by Florida's Agency for Health Care Administration, there has been more than a 500% increase in newborns treated for drug withdrawal between 2005 and 2010 that has been anecdotally attributed to maternal use of prescription drugs [24, 25] (See Fig. 2). Given the rise in prescription drug abuse and infants requiring treatment for drug withdrawal, there is reason to believe that Florida's pregnant women are abusing prescription drugs.


Figure 2. Cases of neonatal withdrawal syndrome in Florida, 2005–2010.

Source: Florida Agency for Health Care Administration and Florida Department of Children and Families.

Download figure to PowerPoint

The purpose of this study was to examine the role of prescription drug abuse in the non-natural deaths of Florida's pregnant or recently pregnant women. While there has been substantial investigation into the relationship between substance abuse and pregnancy outcome, there is limited research into how drug abuse, especially prescription drug abuse, may be related to pregnancy-associated non-natural death.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References

Case Identification and Inclusion Criteria

Florida's Department of Health has conducted a Pregnancy-Associated Mortality Review (PAMR) since 1996 [26]. Pregnancy-associated death is defined by PAMR as a woman's death from any cause, while she is pregnant or within 1 year of termination of the pregnancy, regardless of duration, outcome, and site of the pregnancy. Cases of pregnancy-associated death were identified by linking maternal death certificates, Healthy Start Prenatal Risk screenings, birth certificates, and fetal death records by Florida Department of Health (DOH), Bureau of Vital Statistics. The PAMR team provided researchers at the University of Florida with a dataset of all pregnancy-associated deaths between 1999 and 2005 where the manner of death was anything but natural (homicide, suicide, accident or undetermined, herein referred to as “non-natural”). Medical examiners' reports, including toxicology, autopsy, and case history information, were obtained through requests to district medical examiner.

Each case of a pregnancy-associated death was matched to a medical examiner's report and then entered into Microsoft Office Access 2007 database. Complete information on all variables of interest was available for 415 pregnancy-associated non-natural deaths between 1999 and 2005. Three cases were identified as natural deaths and removed from the dataset. Cases where a toxicology report was unavailable were also removed from the cohort (n = 27). Cases that had a negative toxicology report (n = 167) were also removed. Positive toxicology reports that contained only caffeine, carbon monoxide/carboxyhemoglobin, or atropine/lidocaine were also excluded (n = 31). Positive toxicology for carbon monoxide or carboxyhemoglobin is not considered indicative of drug abuse, but rather is commonly associated with fire-related accidents. Caffeine is ubiquitous throughout our culture and was not considered a drug that generally poses serious harm to health. Atropine/lidocaine is used in resuscitation and usually administered as part of a resuscitation protocol.

These deletions reduced the dataset to 187 cases of a pregnancy-associated death with a positive toxicology report. Eighteen additional exclusions were made for toxicology reports indicating the presence of only over-the-counter medications, including nicotine, acetaminophen, salicylates (aspirin), and diphenhydramine (antihistamine). Two cases with negative toxicology were included because they were delayed deaths from acetaminophen toxicity. The final cohort consisted of 169 drug-positive, pregnancy-associated non-natural deaths. The construction of the study sample is summarized in Fig. 3.


Figure 3. Construction of final study sample: 169 pregnant or recently pregnant women who died and who had a positive toxicology report.

Download figure to PowerPoint

Case History Review

For cases that matched the inclusion criteria for our cohort, the following information was extracted from medical examiners' reports: age, race, cause of death, manner of death, and drugs found in toxicology analyses. Information was unavailable on the socioeconomic status of the women. Similarly, in cases of positive toxicology involving prescription drugs, the source of the drugs was not indicated.

Toxicologic Testing

The postmortem toxicology data collected for this study were obtained from the toxicology reports submitted by the individual medical examiner offices and included quantitative drug data for blood, vitreous humor, and/or urine. Because this study's aim was evaluating the prevalence of use (and not dose/concentration), a case was considered positive for a drug/toxin if it was detected in any body fluid sample, whether or not it was quantified. Toxicology reports with presumptively positive (nonconfirmed) results were not included. The toxicology data were collected from the reports generated by a number of different laboratories. The state of Florida is divided into 23 Medical Examiner Districts; each has its own policies and procedures regarding how toxicology testing is performed, but in general, most offices utilize an accredited toxicology laboratory, such as UF and Wuestaff, and others utilize local or regional hospital laboratories for screening, with confirmation and quantification performed selectively at a specialized laboratory. Several large medical examiner districts have their own, in-house accredited laboratories, such as Miami-Dade County. The methodologies used were also varied and included gas chromatography, gas chromatography–mass spectrometry, high-performance liquid chromatography, thin-layer chromatography, ELISA, and immunoassays. For these reasons, the toxicology data are heterogeneous in the manner in which they were reported and included confirmed positive, presumptive positive, and quantified drugs. Because we are interested in exposure of this population to drugs, both illicit and prescription, we chose to include cases with both quantified and nonquantified, but confirmed positive results. For the purposes of this paper, this is the definition of a “positive” toxicology report. Of the 169 cases included in this study, 86 had both blood and urine positive for drug(s), 64 had only blood positive for drug(s), and five had only urine positive for drug(s). The remainder [14] had some other specimen positive for drug(s), such as vitreous humor, gastric contents, bile, brain, liver, or decomposition fluid.

Classification of Drugs

Drugs encountered in this study were broadly classified as either prescription or illicit. Prescription drugs were subcategorized as opioids (morphine, hydrocodone, methadone, hydromorphone, oxycodone, propoxyphene, fentanyl, tramadol, and codeine), benzodiazepines (oxazepam, nordiazepam, temazepam, alprazolam, diazepam, midazolam, benzodiazepine, and clonazepam), and antidepressants (sertraline, venlafaxine, nortriptyline, citalopram, amitriptyline, doxepin, wellbutrin, and fluoxetine). A general “other” category was also used to group less-frequently encountered prescription drugs in the toxicology reports (gabapentin, promethazine, barbiturate, butalbital, olanzapine, phenytoin, benzonatate, cyclobenzaprine, cimetidine, and anxiolytic not otherwise specified).

Morphine is difficult to categorize as prescription or illicit, given that morphine is the major metabolite of heroin. Without either substantial investigative information indicating heroin use, or the detection of heroin and/or 6-monoacetylmorphine in postmortem samples, the source of the morphine cannot be definitively ascertained. For the purposes of this study, we chose to place those positive for morphine without heroin or heroin metabolites in the prescription drug category because of the far greater accessibility to prescription opioids and the relatively low availability of heroin.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References

Table 1 lists the major groups of causes of death for the 169 decedents included in this study. These were grouped into categories based on mechanism of death, such as drug toxicity, motor vehicle crash, and sharp force injury. The largest groups of causes of deaths were drug toxicity (both single and mixed) and motor vehicle crashes, each accounting for approximately one-third of the deaths. Gunshot wounds were listed as the cause of death in 24 of the 169 deaths, or 14%. The remainder of the causes of death each accounted for <10%.

Table 1. Causes of death as reported by the medical examiner on the death certificate
Cause of DeathNumber%
  1. a

    Other: Homicide NOS (2), carbon monoxide (1), drowning (1).

Drug toxicity5532.5
Single drug toxicity3461.8
Mixed drug toxicity2138.2
Motor vehicle crash5432.0
Blunt trauma NOS137.7
Sharp force injury42.4

Table 2 indicates that at least one prescription drug was found in 91 of the 169 cases (54%) of drug-positive, pregnancy-associated non-natural deaths; 78 cases (46%) were positive for at least one illicit drug, and alcohol, whether detected alone or in combination with other substances, was detected in 69 cases (41%). Table 3 indicates that prescription opioids were detected alone or in conjunction with other drugs in 50 cases (30% of toxicology-positive cases). Table 4 indicates that benzodiazepines were the second most frequently detected class of prescription drug with 42 cases (25%) positive for benzodiazepines alone or in conjunction with other drugs. Evidence of co-use/co-abuse with other drugs was detected. Twenty cases (12%) were positive for both opioids and benzodiazepines. Of cases involving benzodiazepines, 18 were also positive for illicit drugs. Table 5 lists the type and frequency of illicit drugs detected with/without involvement of other drug types. Illicit drugs were involved in 69 (41%) of the 169 cases of pregnancy-associated non-natural deaths.

Table 2. Incidence of prescription and illicit drugs, plus alcohol detected and frequency co-detected with other substances in pregnant and recently pregnant women who died, Florida, 1999–2005 (= 169)
Substance(s) DetectedNumber of Positive Toxicology ReportsPercentage Among Pregnant and Recently Pregnant Women Who Died
Prescription drugs, +/− other9154
Prescription drugs, alone4325
Prescription drugs and illicit drugs2213
Prescription drugs and alcohol148
Prescription drugs, alcohol, and illicit drugs127
Illicit drugs, +/− other7846
Illicit drugs, alone3319
Illicit drugs and prescription drugs2213
Illicit drugs and alcohol116
Illicit drugs, alcohol, and prescription drugs127
Alcohol, +/− other6941
Alcohol and illicit drugs116
Alcohol and prescription drugs148
Alcohol and illicit drugs127
Table 3. Frequency of cases involving prescription opioids with/without involvement of other drug types
Prescription Opioids Detected, with/without Other SubstancesNumber of CasesPercentage of Cases
  1. a

    With/without involvement of other prescription drugs.

Opioids, without other drugs2448
Opioids, with benzodiazepinesa2040
Opioids, with antidepressantsa1020
Opioids, with >1 other type of drug816
Opioids, with illicit drugsa1632
Opioids, total cases involving50100
Table 4. Frequency of cases involving benzodiazepines with/without involvement of other drug types
Benzodiazepines Detected, with/without Other SubstancesNumber of CasesPercentage of Cases
  1. a

    With/without involvement of other prescription drugs.

Benzodiazepines, without other drugs1945
Benzodiazepines, with opioidsa2048
Benzodiazepines, with antidepressantsa717
Benzodiazepines, with >1 other type of drug717
Benzodiazepines, with illicit drugsa1843
Benzodiazepines, total cases involving42100
Table 5. Type and frequency of illicit drugs detected with/without involvement of other drug types
Illicit Drugs DetectedaNumber of CasesPercentage of Cases
  1. a

    Single case may be positive for >1 drug/group of drugs.

γ-Hydroxybutyric acid11
Illicit drugs, with alcohol2333
Illicit drugs, with prescription drugs3449
Illicit drugs, total cases69100

Table 6 indicates that 32 of the 55 toxicology-positive decedents (58%) were pregnant at the time of their death. Toxicology reports indicated that 11 of the 32 decedents (34%) had prescription drugs in their systems at the time of their death. Sixteen cases (50%) showed evidence of illicit drug use, and twelve cases (32%) evidence of alcohol use.

Table 6. Substances detected in women who were pregnant at the time of their death (N = 32)
Substance(s) Detected in Pregnant DecedentsNumber of CasesaPercentage of Cases
  1. a

    Single case may be positive for >1 drug/group of drugs.

Prescription drugs1134
Illicit drugs1650
γ-Hydroxybutyric acid13
Total pregnant decedents32100

Because there are numerous challenges in determining intent in cases of overdose, accidental and suicidal overdoses were considered together. Table 7 indicates that 47 of the 169 deaths (28%) were identified as due to the direct toxic affect of drugs. Of these, single or multiple prescription drugs were attributed as the cause of overdose in 33 cases (70%). In six additional cases, involvement of illicit drug use was also detected. Single or multiple illicit drugs alone accounted for 18 cases of overdose (38%). Two other cases of overdose were detected; one overdose was attributed to alcohol toxicity. The other cause of overdose was determined to be acetaminophen toxicity, although opioids were also detected.

Table 7. Cases of pregnancy-associated accidental and intentional overdose (N = 47)
Drug Type Attributed to OverdoseNumber of CasesPercentage of Cases
  1. a

    Includes cases where toxicology report also indicated prescription drug use.

Prescription drugs3370
Single prescription drug class1430
Multiple, with other classes of prescription drugs1328
Multiple, with illicit drugs613
Single or multiple illicit drugsa1838
Total Overdoses47100

Table 8 indicates that alcohol was the drug most commonly found in the toxicology reports of pregnant or recently pregnant women who died from non-natural manners (69 of 169 cases or 41%). Approximately two-thirds of those involved the admixture of alcohol with prescription drugs and one-third with illicit drugs.

Table 8. Frequency of cases involving alcohol with/without involvement of other drug types
Alcohol Detected, with/without Other SubstancesNumber of CasesPercentage of Cases
  1. a

    With/without involvement of other prescription drugs.

Alcohol, without other drugs3246
Alcohol, with opioidsa1217
Alcohol, with benzodiazepinesa1319
Alcohol, with antidepressantsa812
Alcohol, with other prescription drugsa46
Alcohol, with >1 type of prescription drugs913
Alcohol, with illicit drugsa2333
Alcohol, total cases involving69100


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References

Our study represents the first investigation exploring the relationship between prescription and illicit drug use and, pregnancy-associated non-natural death, keeping in mind that “pregnancy-associated death” indicates a death in a woman who was pregnant or recently pregnant at the time of her death. Also important to note is that just because a drug is categorized in the prescription drug group does not mean it was prescribed to the decedent.

Prescription drugs were detected in a majority (91 of 169 or 54%) of toxicology-positive cases of pregnancy-associated non-natural death from Florida between 1999 and 2005. Opioids and benzodiazepines were the most frequently encountered prescription drug classes: 50 total cases were positive for opioids, and 42 total cases were positive for benzodiazepines. In cases of drug-positive overdose, without regard to overdose intent, prescription drugs were the most common drugs encountered. Substantial co-use of opioids and benzodiazepines was seen. Sixteen cases were positive for both benzodiazepines and opioids.

There are several significant limitations in the ascertainment of prescription and illicit drug abuse during and in the year after pregnancy. Most notably, this study investigated a very narrowly defined population: cases of pregnancy-associated non-natural death with positive postmortem toxicology or investigative information directly linking the death to a specific drug or toxin. This study is not able to capture those individuals whose bodies may have completely metabolized any recently ingested drugs, nor able to identify occasional or binge use of drugs and medications.

It is possible that some of those cases positive for morphine (without detectable heroin and/or 6-MAM) were from exposure to heroin rather than a prescription opioid. However, given the historically low use of heroin throughout the general population and the easy access to prescription forms, we chose to group morphine-positive cases with prescription opioids. Even if we took the opposite approach and assumed that ALL morphine detected in our population sample originated from heroin, a staggering number of cases positive for some other opioid remain.

This method of categorizing heroin and morphine deaths is similar to that of the Florida Medical Examiner Commission in their annual drug surveillance reports ( such that our data can be directly compared and correlated with statewide medical examiner data. The statewide data reported by medical examiners show that within the group of all opioids, heroin and morphine together make up a small fraction of opioid-positive deaths, both in this study and across the state of Florida: in 2005, 369 of 1747 opioid-positive deaths, or 21% of opioid-positive deaths, were positive for heroin or morphine combined, as compared to 1378 of 1747 opioid-positive deaths, or 79% that were positive for any other opioid drug. For 2010, the numbers are even more staggering: 699 of 6358 opioid-positive deaths (11%) were positive for heroin or morphine, compared with 5659 of 6358 deaths (89%) that were positive for an opioid other than morphine or heroin. This represents a more than 300% increase in prescription drug exposure in the entire, statewide medical examiner population.

The data in our study sample of pregnant and recently pregnant women are similar to that reported by the Florida Medical Examiner Commission: of the 50 cases that were positive for any opioid, only 8 deaths, or 14%, were positive for morphine without any other prescription opioid. Thus, even if we assumed all the morphine detected in these decedents came from heroin, it would not change the bottom line—prescription opioid use and abuse is epidemic in our population.

The prevalence of opioid exposure in our study population of non-natural deaths in mothers is similar to that in the general medical examiner population. In this comparison of two groups suffering non-natural deaths during the same time frame, roughly 12% were opioid positive. Of note, in 2010, the proportion of opioid-positive non-natural deaths in Florida was 44.5%. This is consistent with the observed overall increase in use and misuse of opioids. Based both on this data and on the rising incidence of neonatal abstinence syndrome, we infer that women are continuing to use opioids during pregnancy.

Further study of the more recent 2006–2011 time period may be of interest. The Centers for Disease Control recently reported that the number of deaths related to prescription opioids rose nearly fourfold between 1999 and 2008 [27]. Comparison studies in states with prescription drug control legislation would be useful to evaluate the efficacy of prescription drug monitoring programs and establish best practices.

Other limitations of this study include sparseness of socioeconomic data. Most of the women in the study population were White; however, their socioeconomic status was not reported. Ethnic or socioeconomic difference in drug use may limit the applicability of these data to non-White women, even those who are or recently were pregnant. Limited data prevented the ascertainment of the source of prescription drugs detected. There was no way of determining the extent of prescription drug diversion from Florida's pain clinics in the study population. Some medical examiners reporting on cases of homicide during the 1999–2005 study period were not made available to the investigators because they are currently under criminal investigation and access to these files was very limited.

One factor worth emphasizing is that this study population is quite unique in that it is composed of relatively young women (of reproductive age), and because they were pregnant or recently pregnant at the time of their death, they would have been expected to have more interactions with healthcare providers and more opportunities for intervention and prevention of high-risk behaviors. However, as shown in this study, they have very similar exposures to both prescription and illicit drugs and are dying from their toxic effects. These cases may represent missed opportunities for counseling and prevention. Another possibility is that women who die of trauma are frequently women who engage in high-risk behaviors including drug use and unprotected sex resulting in pregnancy.

Despite these caveats, this study does provide an important snapshot of the trends in drug use and misuse in a unique population, that is, women who were pregnant or recently pregnant at the time of their death.


The authors thank Rajeeb Das, M.S.P.H., at the UF Family Data Center and P.V. Rao, Ph.D., for their assistance with the study's data collection and analysis.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. References